Patch-2.6.17.14-vs2.0.2.1-grsec2.1.9.diff

From Linix VServer
Revision as of 00:23, 11 November 2025 by 192.168.65.1 (talk) (Restored content from Wayback Machine)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search

diff -NurpP --minimal linux-2.6.17.13/Makefile linux-2.6.17.13-g-v/Makefile --- linux-2.6.17.13/Makefile 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/Makefile 2006-09-20 15:00:12.000000000 +0200 @@ -518,7 +518,7 @@ export MODLIB ifeq ($(KBUILD_EXTMOD),) -core-y += kernel/ mm/ fs/ ipc/ security/ crypto/ block/ +core-y += kernel/ mm/ fs/ ipc/ security/ crypto/ block/ grsecurity/ vmlinux-dirs := $(patsubst %/,%,$(filter %/, $(init-y) $(init-m) \ $(core-y) $(core-m) $(drivers-y) $(drivers-m) \ diff -NurpP --minimal linux-2.6.17.13/arch/alpha/Kconfig linux-2.6.17.13-g-v/arch/alpha/Kconfig --- linux-2.6.17.13/arch/alpha/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/alpha/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -632,6 +632,8 @@ source "arch/alpha/oprofile/Kconfig" source "arch/alpha/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/alpha/kernel/entry.S linux-2.6.17.13-g-v/arch/alpha/kernel/entry.S --- linux-2.6.17.13/arch/alpha/kernel/entry.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/alpha/kernel/entry.S 2006-09-20 15:00:12.000000000 +0200 @@ -874,24 +874,15 @@ sys_getxgid: .globl sys_getxpid .ent sys_getxpid sys_getxpid: + lda $sp, -16($sp) + stq $26, 0($sp) .prologue 0 - ldq $2, TI_TASK($8) - /* See linux/kernel/timer.c sys_getppid for discussion - about this loop. */ - ldq $3, TASK_GROUP_LEADER($2) - ldq $4, TASK_REAL_PARENT($3) - ldl $0, TASK_TGID($2) -1: ldl $1, TASK_TGID($4) -#ifdef CONFIG_SMP - mov $4, $5 - mb - ldq $3, TASK_GROUP_LEADER($2) - ldq $4, TASK_REAL_PARENT($3) - cmpeq $4, $5, $5 - beq $5, 1b -#endif - stq $1, 80($sp) + lda $16, 96($sp) + jsr $26, do_getxpid + ldq $26, 0($sp) + + lda $sp, 16($sp) ret .end sys_getxpid diff -NurpP --minimal linux-2.6.17.13/arch/alpha/kernel/module.c linux-2.6.17.13-g-v/arch/alpha/kernel/module.c --- linux-2.6.17.13/arch/alpha/kernel/module.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/alpha/kernel/module.c 2006-09-20 15:00:12.000000000 +0200 @@ -177,7 +177,7 @@ apply_relocate_add(Elf64_Shdr *sechdrs, /* The small sections were sorted to the end of the segment. The following should definitely cover them. */ - gp = (u64)me->module_core + me->core_size - 0x8000; + gp = (u64)me->module_core_rw + me->core_size_rw - 0x8000; got = sechdrs[me->arch.gotsecindex].sh_addr; for (i = 0; i < n; i++) { diff -NurpP --minimal linux-2.6.17.13/arch/alpha/kernel/osf_sys.c linux-2.6.17.13-g-v/arch/alpha/kernel/osf_sys.c --- linux-2.6.17.13/arch/alpha/kernel/osf_sys.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/alpha/kernel/osf_sys.c 2006-09-20 15:00:12.000000000 +0200 @@ -38,6 +38,7 @@ #include #include #include +#include #include #include @@ -399,18 +400,20 @@ asmlinkage int osf_utsname(char __user *name) { int error; + struct new_utsname *ptr; down_read(&uts_sem); + ptr = vx_new_utsname(); error = -EFAULT; - if (copy_to_user(name + 0, system_utsname.sysname, 32)) + if (copy_to_user(name + 0, ptr->sysname, 32)) goto out; - if (copy_to_user(name + 32, system_utsname.nodename, 32)) + if (copy_to_user(name + 32, ptr->nodename, 32)) goto out; - if (copy_to_user(name + 64, system_utsname.release, 32)) + if (copy_to_user(name + 64, ptr->release, 32)) goto out; - if (copy_to_user(name + 96, system_utsname.version, 32)) + if (copy_to_user(name + 96, ptr->version, 32)) goto out; - if (copy_to_user(name + 128, system_utsname.machine, 32)) + if (copy_to_user(name + 128, ptr->machine, 32)) goto out; error = 0; @@ -439,6 +442,7 @@ osf_getdomainname(char __user *name, int { unsigned len; int i; + char *domainname; if (!access_ok(VERIFY_WRITE, name, namelen)) return -EFAULT; @@ -448,9 +452,10 @@ osf_getdomainname(char __user *name, int len = 32; down_read(&uts_sem); + domainname = vx_new_uts(domainname); for (i = 0; i < len; ++i) { - __put_user(system_utsname.domainname[i], name + i); - if (system_utsname.domainname[i] == '\0') + __put_user(domainname[i], name + i); + if (domainname[i] == '\0') break; } up_read(&uts_sem); @@ -607,30 +612,30 @@ osf_sigstack(struct sigstack __user *uss asmlinkage long osf_sysinfo(int command, char __user *buf, long count) { - static char * sysinfo_table[] = { - system_utsname.sysname, - system_utsname.nodename, - system_utsname.release, - system_utsname.version, - system_utsname.machine, - "alpha", /* instruction set architecture */ - "dummy", /* hardware serial number */ - "dummy", /* hardware manufacturer */ - "dummy", /* secure RPC domain */ - }; unsigned long offset; char *res; long len, err = -EINVAL; offset = command-1; - if (offset >= sizeof(sysinfo_table)/sizeof(char *)) { + if (offset >= 9) { /* Digital UNIX has a few unpublished interfaces here */ printk("sysinfo(%d)", command); goto out; } down_read(&uts_sem); - res = sysinfo_table[offset]; + switch (offset) + { + case 0: res = vx_new_uts(sysname); break; + case 1: res = vx_new_uts(nodename); break; + case 2: res = vx_new_uts(release); break; + case 3: res = vx_new_uts(version); break; + case 4: res = vx_new_uts(machine); break; + case 5: res = "alpha"; break; + default: + res = "dummy"; + break; + } len = strlen(res)+1; if (len > count) len = count; @@ -1273,6 +1278,10 @@ arch_get_unmapped_area(struct file *filp merely specific addresses, but regions of memory -- perhaps this feature should be incorporated into all ports? */ +#ifdef CONFIG_PAX_RANDMMAP + if (!(current->mm->pax_flags & MF_PAX_RANDMMAP) || !filp) +#endif + if (addr) { addr = arch_get_unmapped_area_1 (PAGE_ALIGN(addr), len, limit); if (addr != (unsigned long) -ENOMEM) @@ -1280,8 +1289,8 @@ arch_get_unmapped_area(struct file *filp } /* Next, try allocating at TASK_UNMAPPED_BASE. */ - addr = arch_get_unmapped_area_1 (PAGE_ALIGN(TASK_UNMAPPED_BASE), - len, limit); + addr = arch_get_unmapped_area_1 (PAGE_ALIGN(current->mm->mmap_base), len, limit); + if (addr != (unsigned long) -ENOMEM) return addr; diff -NurpP --minimal linux-2.6.17.13/arch/alpha/kernel/ptrace.c linux-2.6.17.13-g-v/arch/alpha/kernel/ptrace.c --- linux-2.6.17.13/arch/alpha/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/alpha/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,7 @@ #include #include #include +#include #include #include @@ -283,6 +284,14 @@ do_sys_ptrace(long request, long pid, lo goto out_notsk; } + if (gr_handle_ptrace(child, request)) + goto out; + + if (!vx_check(vx_task_xid(child), VX_WATCH|VX_IDENT)) { + ret = -EPERM; + goto out; + } + if (request == PTRACE_ATTACH) { ret = ptrace_attach(child); goto out; diff -NurpP --minimal linux-2.6.17.13/arch/alpha/kernel/systbls.S linux-2.6.17.13-g-v/arch/alpha/kernel/systbls.S --- linux-2.6.17.13/arch/alpha/kernel/systbls.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/alpha/kernel/systbls.S 2006-09-20 15:00:12.000000000 +0200 @@ -447,7 +447,7 @@ sys_call_table: .quad sys_stat64 /* 425 */ .quad sys_lstat64 .quad sys_fstat64 - .quad sys_ni_syscall /* sys_vserver */ + .quad sys_vserver /* sys_vserver */ .quad sys_ni_syscall /* sys_mbind */ .quad sys_ni_syscall /* sys_get_mempolicy */ .quad sys_ni_syscall /* sys_set_mempolicy */ diff -NurpP --minimal linux-2.6.17.13/arch/alpha/mm/fault.c linux-2.6.17.13-g-v/arch/alpha/mm/fault.c --- linux-2.6.17.13/arch/alpha/mm/fault.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/alpha/mm/fault.c 2006-09-20 15:00:12.000000000 +0200 @@ -25,6 +25,7 @@ #include #include #include +#include #include #include @@ -56,6 +57,124 @@ __load_new_mm_context(struct mm_struct * __reload_thread(pcb); } +#ifdef CONFIG_PAX_PAGEEXEC +/* + * PaX: decide what to do with offenders (regs->pc = fault address) + * + * returns 1 when task should be killed + * 2 when patched PLT trampoline was detected + * 3 when unpatched PLT trampoline was detected + */ +static int pax_handle_fetch_fault(struct pt_regs *regs) +{ + +#ifdef CONFIG_PAX_EMUPLT + int err; + + do { /* PaX: patched PLT emulation #1 */ + unsigned int ldah, ldq, jmp; + + err = get_user(ldah, (unsigned int *)regs->pc); + err |= get_user(ldq, (unsigned int *)(regs->pc+4)); + err |= get_user(jmp, (unsigned int *)(regs->pc+8)); + + if (err) + break; + + if ((ldah & 0xFFFF0000U) == 0x277B0000U && + (ldq & 0xFFFF0000U) == 0xA77B0000U && + jmp == 0x6BFB0000U) + { + unsigned long r27, addr; + unsigned long addrh = (ldah | 0xFFFFFFFFFFFF0000UL) << 16; + unsigned long addrl = ldq | 0xFFFFFFFFFFFF0000UL; + + addr = regs->r27 + ((addrh ^ 0x80000000UL) + 0x80000000UL) + ((addrl ^ 0x8000UL) + 0x8000UL); + err = get_user(r27, (unsigned long*)addr); + if (err) + break; + + regs->r27 = r27; + regs->pc = r27; + return 2; + } + } while (0); + + do { /* PaX: patched PLT emulation #2 */ + unsigned int ldah, lda, br; + + err = get_user(ldah, (unsigned int *)regs->pc); + err |= get_user(lda, (unsigned int *)(regs->pc+4)); + err |= get_user(br, (unsigned int *)(regs->pc+8)); + + if (err) + break; + + if ((ldah & 0xFFFF0000U)== 0x277B0000U && + (lda & 0xFFFF0000U) == 0xA77B0000U && + (br & 0xFFE00000U) == 0xC3E00000U) + { + unsigned long addr = br | 0xFFFFFFFFFFE00000UL; + unsigned long addrh = (ldah | 0xFFFFFFFFFFFF0000UL) << 16; + unsigned long addrl = lda | 0xFFFFFFFFFFFF0000UL; + + regs->r27 += ((addrh ^ 0x80000000UL) + 0x80000000UL) + ((addrl ^ 0x8000UL) + 0x8000UL); + regs->pc += 12 + (((addr ^ 0x00100000UL) + 0x00100000UL) << 2); + return 2; + } + } while (0); + + do { /* PaX: unpatched PLT emulation */ + unsigned int br; + + err = get_user(br, (unsigned int *)regs->pc); + + if (!err && (br & 0xFFE00000U) == 0xC3800000U) { + unsigned int br2, ldq, nop, jmp; + unsigned long addr = br | 0xFFFFFFFFFFE00000UL, resolver; + + addr = regs->pc + 4 + (((addr ^ 0x00100000UL) + 0x00100000UL) << 2); + err = get_user(br2, (unsigned int *)addr); + err |= get_user(ldq, (unsigned int *)(addr+4)); + err |= get_user(nop, (unsigned int *)(addr+8)); + err |= get_user(jmp, (unsigned int *)(addr+12)); + err |= get_user(resolver, (unsigned long *)(addr+16)); + + if (err) + break; + + if (br2 == 0xC3600000U && + ldq == 0xA77B000CU && + nop == 0x47FF041FU && + jmp == 0x6B7B0000U) + { + regs->r28 = regs->pc+4; + regs->r27 = addr+16; + regs->pc = resolver; + return 3; + } + } + } while (0); +#endif + + return 1; +} + +void pax_report_insns(void *pc, void *sp) +{ + unsigned long i; + + printk(KERN_ERR "PAX: bytes at PC: "); + for (i = 0; i < 5; i++) { + unsigned int c; + if (get_user(c, (unsigned int*)pc+i)) + printk("???????? "); + else + printk("%08x ", c); + } + printk("\n"); +} +#endif /* * This routine handles page faults. It determines the address, @@ -133,8 +252,29 @@ do_page_fault(unsigned long address, uns good_area: si_code = SEGV_ACCERR; if (cause < 0) { - if (!(vma->vm_flags & VM_EXEC)) + if (!(vma->vm_flags & VM_EXEC)) { + +#ifdef CONFIG_PAX_PAGEEXEC + if (!(mm->pax_flags & MF_PAX_PAGEEXEC) || address != regs->pc) + goto bad_area; + + up_read(&mm->mmap_sem); + switch(pax_handle_fetch_fault(regs)) { + +#ifdef CONFIG_PAX_EMUPLT + case 2: + case 3: + return; +#endif + + } + pax_report_fault(regs, (void*)regs->pc, (void*)rdusp()); + do_exit(SIGKILL); +#else goto bad_area; +#endif + + } } else if (!cause) { /* Allow reads even for write-only mappings */ if (!(vma->vm_flags & (VM_READ | VM_WRITE))) diff -NurpP --minimal linux-2.6.17.13/arch/alpha/mm/init.c linux-2.6.17.13-g-v/arch/alpha/mm/init.c --- linux-2.6.17.13/arch/alpha/mm/init.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/alpha/mm/init.c 2006-09-20 15:00:12.000000000 +0200 @@ -21,6 +21,7 @@ #include #include /* max_low_pfn */ #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/arch/arm/Kconfig linux-2.6.17.13-g-v/arch/arm/Kconfig --- linux-2.6.17.13/arch/arm/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/arm/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -873,6 +873,8 @@ source "arch/arm/oprofile/Kconfig" source "arch/arm/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/arm/kernel/calls.S linux-2.6.17.13-g-v/arch/arm/kernel/calls.S --- linux-2.6.17.13/arch/arm/kernel/calls.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/arm/kernel/calls.S 2006-09-20 15:00:12.000000000 +0200 @@ -322,7 +322,7 @@ /* 310 */ CALL(sys_request_key) CALL(sys_keyctl) CALL(ABI(sys_semtimedop, sys_oabi_semtimedop)) -/* vserver */ CALL(sys_ni_syscall) + CALL(sys_vserver) CALL(sys_ioprio_set) /* 315 */ CALL(sys_ioprio_get) CALL(sys_inotify_init) diff -NurpP --minimal linux-2.6.17.13/arch/arm/mm/mmap.c linux-2.6.17.13-g-v/arch/arm/mm/mmap.c --- linux-2.6.17.13/arch/arm/mm/mmap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/arm/mm/mmap.c 2006-09-20 15:00:12.000000000 +0200 @@ -62,6 +62,10 @@ arch_get_unmapped_area(struct file *filp if (len > TASK_SIZE) return -ENOMEM; +#ifdef CONFIG_PAX_RANDMMAP + if (!(mm->pax_flags & MF_PAX_RANDMMAP) || !filp) +#endif + if (addr) { if (do_align) addr = COLOUR_ALIGN(addr, pgoff); @@ -76,7 +80,7 @@ arch_get_unmapped_area(struct file *filp if (len > mm->cached_hole_size) { start_addr = addr = mm->free_area_cache; } else { - start_addr = addr = TASK_UNMAPPED_BASE; + start_addr = addr = mm->mmap_base; mm->cached_hole_size = 0; } @@ -93,8 +97,8 @@ full_search: * Start a new search - just in case we missed * some holes. */ - if (start_addr != TASK_UNMAPPED_BASE) { - start_addr = addr = TASK_UNMAPPED_BASE; + if (start_addr != mm->mmap_base) { + start_addr = addr = mm->mmap_base; mm->cached_hole_size = 0; goto full_search; } diff -NurpP --minimal linux-2.6.17.13/arch/arm26/Kconfig linux-2.6.17.13-g-v/arch/arm26/Kconfig --- linux-2.6.17.13/arch/arm26/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/arm26/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -234,6 +234,8 @@ source "drivers/usb/Kconfig" source "arch/arm26/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/arm26/kernel/calls.S linux-2.6.17.13-g-v/arch/arm26/kernel/calls.S --- linux-2.6.17.13/arch/arm26/kernel/calls.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/arm26/kernel/calls.S 2006-09-20 15:00:12.000000000 +0200 @@ -257,6 +257,11 @@ __syscall_start: .long sys_lremovexattr .long sys_fremovexattr .long sys_tkill + + .rept 313 - (. - __syscall_start) / 4 + .long sys_ni_syscall + .endr + .long sys_vserver /* 313 */ __syscall_end: .rept NR_syscalls - (__syscall_end - __syscall_start) / 4 diff -NurpP --minimal linux-2.6.17.13/arch/arm26/kernel/traps.c linux-2.6.17.13-g-v/arch/arm26/kernel/traps.c --- linux-2.6.17.13/arch/arm26/kernel/traps.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/arm26/kernel/traps.c 2006-09-20 15:00:12.000000000 +0200 @@ -186,8 +186,9 @@ NORET_TYPE void die(const char *str, str printk("Internal error: %s: %x\n", str, err); printk("CPU: %d\n", smp_processor_id()); show_regs(regs); - printk("Process %s (pid: %d, stack limit = 0x%p)\n", - current->comm, current->pid, end_of_stack(tsk)); + printk("Process %s (pid: %d[#%u], stack limit = 0x%p)\n", + current->comm, current->pid, + current->xid, end_of_stack(tsk)); if (!user_mode(regs) || in_interrupt()) { __dump_stack(tsk, (unsigned long)(regs + 1)); diff -NurpP --minimal linux-2.6.17.13/arch/cris/Kconfig linux-2.6.17.13-g-v/arch/cris/Kconfig --- linux-2.6.17.13/arch/cris/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/cris/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -181,6 +181,8 @@ source "drivers/usb/Kconfig" source "arch/cris/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/frv/mm/mmu-context.c linux-2.6.17.13-g-v/arch/frv/mm/mmu-context.c --- linux-2.6.17.13/arch/frv/mm/mmu-context.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/frv/mm/mmu-context.c 2006-09-20 15:00:12.000000000 +0200 @@ -11,6 +11,7 @@ #include #include +#include #include #define NR_CXN 4096 diff -NurpP --minimal linux-2.6.17.13/arch/h8300/Kconfig linux-2.6.17.13-g-v/arch/h8300/Kconfig --- linux-2.6.17.13/arch/h8300/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/h8300/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -199,6 +199,8 @@ source "fs/Kconfig" source "arch/h8300/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/i386/Kconfig linux-2.6.17.13-g-v/arch/i386/Kconfig --- linux-2.6.17.13/arch/i386/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -948,7 +948,7 @@ config PCI choice prompt "PCI access mode" depends on PCI && !X86_VISWS - default PCI_GOANY + default PCI_GODIRECT ---help--- On PCI systems, the BIOS can be used to detect the PCI devices and determine their configuration. However, some old PCI motherboards @@ -980,7 +980,7 @@ endchoice config PCI_BIOS bool - depends on !X86_VISWS && PCI && (PCI_GOBIOS || PCI_GOANY) + depends on !X86_VISWS && PCI && PCI_GOBIOS default y config PCI_DIRECT @@ -1088,6 +1088,8 @@ endmenu source "arch/i386/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/i386/Kconfig.cpu linux-2.6.17.13-g-v/arch/i386/Kconfig.cpu --- linux-2.6.17.13/arch/i386/Kconfig.cpu 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/Kconfig.cpu 2006-09-20 15:00:12.000000000 +0200 @@ -251,7 +251,7 @@ config X86_PPRO_FENCE config X86_F00F_BUG bool - depends on M586MMX || M586TSC || M586 || M486 || M386 + depends on (M586MMX || M586TSC || M586 || M486 || M386) && !PAX_KERNEXEC default y config X86_WP_WORKS_OK @@ -281,7 +281,7 @@ config X86_CMPXCHG64 config X86_ALIGNMENT_16 bool - depends on MWINCHIP3D || MWINCHIP2 || MWINCHIPC6 || MCYRIXIII || X86_ELAN || MK6 || M586MMX || M586TSC || M586 || M486 || MVIAC3_2 || MGEODEGX1 + depends on MWINCHIP3D || MWINCHIP2 || MWINCHIPC6 || MCYRIXIII || X86_ELAN || MK8 || MK7 || MK6 || MPENTIUM4 || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || M586 || M486 || MVIAC3_2 || MGEODEGX1 default y config X86_GOOD_APIC diff -NurpP --minimal linux-2.6.17.13/arch/i386/Kconfig.debug linux-2.6.17.13-g-v/arch/i386/Kconfig.debug --- linux-2.6.17.13/arch/i386/Kconfig.debug 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/Kconfig.debug 2006-09-20 15:00:12.000000000 +0200 @@ -53,7 +53,7 @@ config DEBUG_PAGEALLOC config DEBUG_RODATA bool "Write protect kernel read-only data structures" - depends on DEBUG_KERNEL + depends on DEBUG_KERNEL && 0 help Mark the kernel read-only data as write-protected in the pagetables, in order to catch accidental (and incorrect) writes to such const diff -NurpP --minimal linux-2.6.17.13/arch/i386/boot/compressed/head.S linux-2.6.17.13-g-v/arch/i386/boot/compressed/head.S --- linux-2.6.17.13/arch/i386/boot/compressed/head.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/boot/compressed/head.S 2006-09-20 15:00:12.000000000 +0200 @@ -39,11 +39,13 @@ startup_32: movl %eax,%gs lss stack_start,%esp + movl 0x000000,%ecx xorl %eax,%eax 1: incl %eax # check that A20 really IS enabled movl %eax,0x000000 # loop forever if it isn't cmpl %eax,0x100000 je 1b + movl %ecx,0x000000 /* * Initialize eflags. Some BIOS's leave bits like NT set. This would diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/acpi/sleep.c linux-2.6.17.13-g-v/arch/i386/kernel/acpi/sleep.c --- linux-2.6.17.13/arch/i386/kernel/acpi/sleep.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/acpi/sleep.c 2006-09-20 15:00:12.000000000 +0200 @@ -10,6 +10,7 @@ #include #include #include +#include /* address in low memory of the wakeup routine. */ unsigned long acpi_wakeup_address = 0; @@ -24,11 +25,22 @@ static void init_low_mapping(pgd_t * pgd { int pgd_ofs = 0; +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + while ((pgd_ofs < pgd_limit) && (pgd_ofs + USER_PTRS_PER_PGD < PTRS_PER_PGD)) { set_pgd(pgd, *(pgd + USER_PTRS_PER_PGD)); pgd_ofs++, pgd++; } + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + flush_tlb_all(); } @@ -55,7 +67,18 @@ int acpi_save_state_mem(void) */ void acpi_restore_state_mem(void) { +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + zap_low_mappings(); + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + } /** diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/alternative.c linux-2.6.17.13-g-v/arch/i386/kernel/alternative.c --- linux-2.6.17.13/arch/i386/kernel/alternative.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/alternative.c 2006-09-20 15:00:12.000000000 +0200 @@ -3,6 +3,7 @@ #include #include #include +#include #define DEBUG 0 #if DEBUG @@ -101,71 +102,128 @@ void apply_alternatives(struct alt_instr struct alt_instr *a; int diff, i, k; +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + DPRINTK("%s: alt table %p -> %p\n", __FUNCTION__, start, end); for (a = start; a < end; a++) { BUG_ON(a->replacementlen > a->instrlen); if (!boot_cpu_has(a->cpuid)) continue; - memcpy(a->instr, a->replacement, a->replacementlen); + memcpy(a->instr + __KERNEL_TEXT_OFFSET, a->replacement, a->replacementlen); diff = a->instrlen - a->replacementlen; /* Pad the rest with nops */ for (i = a->replacementlen; diff > 0; diff -= k, i += k) { k = diff; if (k > ASM_NOP_MAX) k = ASM_NOP_MAX; - memcpy(a->instr + i, noptable[k], k); + memcpy(a->instr + i + __KERNEL_TEXT_OFFSET, noptable[k], k); } } + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + } static void alternatives_smp_save(struct alt_instr *start, struct alt_instr *end) { struct alt_instr *a; +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + DPRINTK("%s: alt table %p-%p\n", __FUNCTION__, start, end); for (a = start; a < end; a++) { memcpy(a->replacement + a->replacementlen, - a->instr, + a->instr + __KERNEL_TEXT_OFFSET, a->instrlen); } + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + } static void alternatives_smp_apply(struct alt_instr *start, struct alt_instr *end) { struct alt_instr *a; +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + for (a = start; a < end; a++) { - memcpy(a->instr, + memcpy(a->instr + __KERNEL_TEXT_OFFSET, a->replacement + a->replacementlen, a->instrlen); } + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + } static void alternatives_smp_lock(u8 **start, u8 **end, u8 *text, u8 *text_end) { - u8 **ptr; + u8 *ptr; - for (ptr = start; ptr < end; ptr++) { - if (*ptr < text) +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + + for (; start < end; start++) { + ptr = *start + __KERNEL_TEXT_OFFSET; + if (ptr < text) continue; - if (*ptr > text_end) + if (ptr > text_end) continue; - **ptr = 0xf0; /* lock prefix */ + *ptr = 0xf0; /* lock prefix */ }; + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + } static void alternatives_smp_unlock(u8 **start, u8 **end, u8 *text, u8 *text_end) { unsigned char **noptable = find_nop_table(); - u8 **ptr; + u8 *ptr; - for (ptr = start; ptr < end; ptr++) { - if (*ptr < text) +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + + for (; start < end; start++) { + ptr = *start + __KERNEL_TEXT_OFFSET; + if (ptr < text) continue; - if (*ptr > text_end) + if (ptr > text_end) continue; - **ptr = noptable[1][0]; + *ptr = noptable[1][0]; }; + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + } struct smp_alt_module { diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/apic.c linux-2.6.17.13-g-v/arch/i386/kernel/apic.c --- linux-2.6.17.13/arch/i386/kernel/apic.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/apic.c 2006-09-20 15:00:12.000000000 +0200 @@ -1176,7 +1176,7 @@ inline void smp_local_timer_interrupt(st { profile_tick(CPU_PROFILING, regs); #ifdef CONFIG_SMP - update_process_times(user_mode_vm(regs)); + update_process_times(user_mode(regs)); #endif /* diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/apm.c linux-2.6.17.13-g-v/arch/i386/kernel/apm.c --- linux-2.6.17.13/arch/i386/kernel/apm.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/apm.c 2006-09-20 15:00:12.000000000 +0200 @@ -235,7 +235,7 @@ #include "io_ports.h" extern unsigned long get_cmos_time(void); -extern void machine_real_restart(unsigned char *, int); +extern void machine_real_restart(const unsigned char *, unsigned int); #if defined(CONFIG_APM_DISPLAY_BLANK) && defined(CONFIG_VT) extern int (*console_blank_hook)(int); @@ -589,9 +589,18 @@ static u8 apm_bios_call(u32 func, u32 eb struct desc_struct save_desc_40; struct desc_struct *gdt; +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; +#endif + cpus = apm_save_cpus(); cpu = get_cpu(); + +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + gdt = get_cpu_gdt_table(cpu); save_desc_40 = gdt[0x40 / 8]; gdt[0x40 / 8] = bad_bios_desc; @@ -603,6 +612,11 @@ static u8 apm_bios_call(u32 func, u32 eb APM_DO_RESTORE_SEGS; local_irq_restore(flags); gdt[0x40 / 8] = save_desc_40; + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + put_cpu(); apm_restore_cpus(cpus); @@ -633,9 +647,18 @@ static u8 apm_bios_call_simple(u32 func, struct desc_struct save_desc_40; struct desc_struct *gdt; +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; +#endif + cpus = apm_save_cpus(); cpu = get_cpu(); + +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + gdt = get_cpu_gdt_table(cpu); save_desc_40 = gdt[0x40 / 8]; gdt[0x40 / 8] = bad_bios_desc; @@ -647,6 +670,11 @@ static u8 apm_bios_call_simple(u32 func, APM_DO_RESTORE_SEGS; local_irq_restore(flags); gdt[0x40 / 8] = save_desc_40; + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + put_cpu(); apm_restore_cpus(cpus); return error; @@ -910,7 +938,7 @@ recalc: static void apm_power_off(void) { - unsigned char po_bios_call[] = { + const unsigned char po_bios_call[] = { 0xb8, 0x00, 0x10, /* movw $0x1000,ax */ 0x8e, 0xd0, /* movw ax,ss */ 0xbc, 0x00, 0xf0, /* movw $0xf000,sp */ diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/asm-offsets.c linux-2.6.17.13-g-v/arch/i386/kernel/asm-offsets.c --- linux-2.6.17.13/arch/i386/kernel/asm-offsets.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/asm-offsets.c 2006-09-20 15:00:12.000000000 +0200 @@ -68,5 +68,6 @@ void foo(void) sizeof(struct tss_struct)); DEFINE(PAGE_SIZE_asm, PAGE_SIZE); + DEFINE(PTRS_PER_PTE_asm, PTRS_PER_PTE); DEFINE(VSYSCALL_BASE, __fix_to_virt(FIX_VSYSCALL)); } diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/cpu/common.c linux-2.6.17.13-g-v/arch/i386/kernel/cpu/common.c --- linux-2.6.17.13/arch/i386/kernel/cpu/common.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/cpu/common.c 2006-09-20 15:00:12.000000000 +0200 @@ -4,7 +4,6 @@ #include #include #include -#include #include #include #include @@ -19,16 +18,18 @@ #include "cpu.h" -DEFINE_PER_CPU(struct Xgt_desc_struct, cpu_gdt_descr); -EXPORT_PER_CPU_SYMBOL(cpu_gdt_descr); - DEFINE_PER_CPU(unsigned char, cpu_16bit_stack[CPU_16BIT_STACK_SIZE]); EXPORT_PER_CPU_SYMBOL(cpu_16bit_stack); static int cachesize_override __cpuinitdata = -1; static int disable_x86_fxsr __cpuinitdata; static int disable_x86_serial_nr __cpuinitdata = 1; + +#ifdef CONFIG_PAX_NOVSYSCALL +static int disable_x86_sep __cpuinitdata = 1; +#else static int disable_x86_sep __cpuinitdata; +#endif struct cpu_dev * cpu_devs[X86_VENDOR_NUM] = {}; @@ -590,11 +591,10 @@ void __init early_cpu_init(void) void __cpuinit cpu_init(void) { int cpu = smp_processor_id(); - struct tss_struct * t = &per_cpu(init_tss, cpu); + struct tss_struct * t = init_tss + cpu; struct thread_struct *thread = &current->thread; - struct desc_struct *gdt; + struct desc_struct *gdt = get_cpu_gdt_table(cpu); __u32 stk16_off = (__u32)&per_cpu(cpu_16bit_stack, cpu); - struct Xgt_desc_struct *cpu_gdt_descr = &per_cpu(cpu_gdt_descr, cpu); if (cpu_test_and_set(cpu, cpu_initialized)) { printk(KERN_WARNING "CPU#%d already initialized!\n", cpu); @@ -612,29 +612,11 @@ void __cpuinit cpu_init(void) } /* - * This is a horrible hack to allocate the GDT. The problem - * is that cpu_init() is called really early for the boot CPU - * (and hence needs bootmem) but much later for the secondary - * CPUs, when bootmem will have gone away - */ - if (NODE_DATA(0)->bdata->node_bootmem_map) { - gdt = (struct desc_struct *)alloc_bootmem_pages(PAGE_SIZE); - /* alloc_bootmem_pages panics on failure, so no check */ - memset(gdt, 0, PAGE_SIZE); - } else { - gdt = (struct desc_struct *)get_zeroed_page(GFP_KERNEL); - if (unlikely(!gdt)) { - printk(KERN_CRIT "CPU%d failed to allocate GDT\n", cpu); - for (;;) - local_irq_enable(); - } - } - - /* * Initialize the per-CPU GDT with the boot GDT, * and set up the GDT descriptor: */ - memcpy(gdt, cpu_gdt_table, GDT_SIZE); + if (cpu) + memcpy(gdt, cpu_gdt_table, GDT_SIZE); /* Set up GDT entry for 16bit stack */ *(__u64 *)(&gdt[GDT_ENTRY_ESPFIX_SS]) |= @@ -642,10 +624,10 @@ void __cpuinit cpu_init(void) ((((__u64)stk16_off) << 32) & 0xff00000000000000ULL) | (CPU_16BIT_STACK_SIZE - 1); - cpu_gdt_descr->size = GDT_SIZE - 1; - cpu_gdt_descr->address = (unsigned long)gdt; + cpu_gdt_descr[cpu].size = GDT_SIZE - 1; + cpu_gdt_descr[cpu].address = (unsigned long)gdt; - load_gdt(cpu_gdt_descr); + load_gdt(&cpu_gdt_descr[cpu]); load_idt(&idt_descr); /* @@ -660,7 +642,7 @@ void __cpuinit cpu_init(void) load_esp0(t, thread); set_tss_desc(cpu,t); load_TR_desc(); - load_LDT(&init_mm.context); + _load_LDT(&init_mm.context); #ifdef CONFIG_DOUBLEFAULT /* Set up doublefault TSS pointer in the GDT */ @@ -668,7 +650,7 @@ void __cpuinit cpu_init(void) #endif /* Clear %fs and %gs. */ - asm volatile ("xorl %eax, %eax; movl %eax, %fs; movl %eax, %gs"); + asm volatile ("movl %0, %%fs; movl %0, %%gs" : : "r"(0)); /* Clear all 6 debug registers: */ set_debugreg(0, 0); diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/crash.c linux-2.6.17.13-g-v/arch/i386/kernel/crash.c --- linux-2.6.17.13/arch/i386/kernel/crash.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/crash.c 2006-09-20 15:00:12.000000000 +0200 @@ -105,7 +105,7 @@ static int crash_nmi_callback(struct pt_ return 1; local_irq_disable(); - if (!user_mode_vm(regs)) { + if (!user_mode(regs)) { crash_fixup_ss_esp(&fixed_regs, regs); regs = &fixed_regs; } diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/doublefault.c linux-2.6.17.13-g-v/arch/i386/kernel/doublefault.c --- linux-2.6.17.13/arch/i386/kernel/doublefault.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/doublefault.c 2006-09-20 15:00:12.000000000 +0200 @@ -11,7 +11,7 @@ #define DOUBLEFAULT_STACKSIZE (1024) static unsigned long doublefault_stack[DOUBLEFAULT_STACKSIZE]; -#define STACK_START (unsigned long)(doublefault_stack+DOUBLEFAULT_STACKSIZE) +#define STACK_START (unsigned long)(doublefault_stack+DOUBLEFAULT_STACKSIZE-2) #define ptr_ok(x) ((x) > PAGE_OFFSET && (x) < PAGE_OFFSET + 0x1000000) @@ -56,10 +56,10 @@ struct tss_struct doublefault_tss __cach .eip = (unsigned long) doublefault_fn, .eflags = X86_EFLAGS_SF | 0x2, /* 0x2 bit is always set */ .esp = STACK_START, - .es = __USER_DS, + .es = __KERNEL_DS, .cs = __KERNEL_CS, .ss = __KERNEL_DS, - .ds = __USER_DS, + .ds = __KERNEL_DS, .__cr3 = __pa(swapper_pg_dir) }; diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/efi.c linux-2.6.17.13-g-v/arch/i386/kernel/efi.c --- linux-2.6.17.13/arch/i386/kernel/efi.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/efi.c 2006-09-20 15:00:12.000000000 +0200 @@ -64,82 +64,43 @@ extern void * boot_ioremap(unsigned long static unsigned long efi_rt_eflags; static DEFINE_SPINLOCK(efi_rt_lock); -static pgd_t efi_bak_pg_dir_pointer[2]; +static pgd_t __initdata efi_bak_pg_dir_pointer[KERNEL_PGD_PTRS] __attribute__ ((aligned (4096))); -static void efi_call_phys_prelog(void) +static void __init efi_call_phys_prelog(void) { - unsigned long cr4; - unsigned long temp; - struct Xgt_desc_struct *cpu_gdt_descr; - spin_lock(&efi_rt_lock); local_irq_save(efi_rt_eflags); - cpu_gdt_descr = &per_cpu(cpu_gdt_descr, 0); - - /* - * If I don't have PSE, I should just duplicate two entries in page - * directory. If I have PSE, I just need to duplicate one entry in - * page directory. - */ - cr4 = read_cr4(); - - if (cr4 & X86_CR4_PSE) { - efi_bak_pg_dir_pointer[0].pgd = - swapper_pg_dir[pgd_index(0)].pgd; - swapper_pg_dir[0].pgd = - swapper_pg_dir[pgd_index(PAGE_OFFSET)].pgd; - } else { - efi_bak_pg_dir_pointer[0].pgd = - swapper_pg_dir[pgd_index(0)].pgd; - efi_bak_pg_dir_pointer[1].pgd = - swapper_pg_dir[pgd_index(0x400000)].pgd; - swapper_pg_dir[pgd_index(0)].pgd = - swapper_pg_dir[pgd_index(PAGE_OFFSET)].pgd; - temp = PAGE_OFFSET + 0x400000; - swapper_pg_dir[pgd_index(0x400000)].pgd = - swapper_pg_dir[pgd_index(temp)].pgd; - } + clone_pgd_range(efi_bak_pg_dir_pointer, swapper_pg_dir, KERNEL_PGD_PTRS); + clone_pgd_range(swapper_pg_dir, swapper_pg_dir + USER_PGD_PTRS, + USER_PGD_PTRS >= KERNEL_PGD_PTRS ? KERNEL_PGD_PTRS : USER_PGD_PTRS); /* * After the lock is released, the original page table is restored. */ - local_flush_tlb(); + __flush_tlb_all(); - cpu_gdt_descr->address = __pa(cpu_gdt_descr->address); - load_gdt(cpu_gdt_descr); + cpu_gdt_descr[0].address = __pa(cpu_gdt_descr[0].address); + load_gdt((struct Xgt_desc_struct *) __pa(&cpu_gdt_descr[0])); } -static void efi_call_phys_epilog(void) +static void __init efi_call_phys_epilog(void) { - unsigned long cr4; - struct Xgt_desc_struct *cpu_gdt_descr = &per_cpu(cpu_gdt_descr, 0); - - cpu_gdt_descr->address = (unsigned long)__va(cpu_gdt_descr->address); - load_gdt(cpu_gdt_descr); - - cr4 = read_cr4(); + cpu_gdt_descr[0].address = (unsigned long) __va(cpu_gdt_descr[0].address); + load_gdt(&cpu_gdt_descr[0]); - if (cr4 & X86_CR4_PSE) { - swapper_pg_dir[pgd_index(0)].pgd = - efi_bak_pg_dir_pointer[0].pgd; - } else { - swapper_pg_dir[pgd_index(0)].pgd = - efi_bak_pg_dir_pointer[0].pgd; - swapper_pg_dir[pgd_index(0x400000)].pgd = - efi_bak_pg_dir_pointer[1].pgd; - } + clone_pgd_range(swapper_pg_dir, efi_bak_pg_dir_pointer, KERNEL_PGD_PTRS); /* * After the lock is released, the original page table is restored. */ - local_flush_tlb(); + __flush_tlb_all(); local_irq_restore(efi_rt_eflags); spin_unlock(&efi_rt_lock); } -static efi_status_t +static efi_status_t __init phys_efi_set_virtual_address_map(unsigned long memory_map_size, unsigned long descriptor_size, u32 descriptor_version, @@ -155,7 +116,7 @@ phys_efi_set_virtual_address_map(unsigne return status; } -static efi_status_t +static efi_status_t __init phys_efi_get_time(efi_time_t *tm, efi_time_cap_t *tc) { efi_status_t status; diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/efi_stub.S linux-2.6.17.13-g-v/arch/i386/kernel/efi_stub.S --- linux-2.6.17.13/arch/i386/kernel/efi_stub.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/efi_stub.S 2006-09-20 15:00:12.000000000 +0200 @@ -7,6 +7,7 @@ #include #include +#include #include #include @@ -22,7 +23,7 @@ * service functions will comply with gcc calling convention, too. */ -.text +__INIT ENTRY(efi_call_phys) /* * 0. The function can only be called in Linux kernel. So CS has been @@ -38,9 +39,7 @@ ENTRY(efi_call_phys) * The mapping of lower virtual memory has been created in prelog and * epilog. */ - movl $1f, %edx - subl $__PAGE_OFFSET, %edx - jmp *%edx + jmp 1f-__PAGE_OFFSET 1: /* @@ -49,14 +48,8 @@ ENTRY(efi_call_phys) * parameter 2, ..., param n. To make things easy, we save the return * address of efi_call_phys in a global variable. */ - popl %edx - movl %edx, saved_return_addr - /* get the function pointer into ECX*/ - popl %ecx - movl %ecx, efi_rt_function_ptr - movl $2f, %edx - subl $__PAGE_OFFSET, %edx - pushl %edx + popl (saved_return_addr) + popl (efi_rt_function_ptr) /* * 3. Clear PG bit in %CR0. @@ -75,9 +68,8 @@ ENTRY(efi_call_phys) /* * 5. Call the physical function. */ - jmp *%ecx + call *(efi_rt_function_ptr-__PAGE_OFFSET) -2: /* * 6. After EFI runtime service returns, control will return to * following instruction. We'd better readjust stack pointer first. @@ -87,37 +79,29 @@ ENTRY(efi_call_phys) /* * 7. Restore PG bit */ - movl %cr0, %edx - orl $0x80000000, %edx - movl %edx, %cr0 - jmp 1f -1: /* * 8. Now restore the virtual mode from flat mode by * adding EIP with PAGE_OFFSET. */ - movl $1f, %edx - jmp *%edx + movl %cr0, %edx + orl $0x80000000, %edx + movl %edx, %cr0 + jmp 1f+__PAGE_OFFSET 1: /* * 9. Balance the stack. And because EAX contain the return value, * we'd better not clobber it. */ - leal efi_rt_function_ptr, %edx - movl (%edx), %ecx - pushl %ecx + pushl (efi_rt_function_ptr) /* - * 10. Push the saved return address onto the stack and return. + * 10. Return to the saved return address. */ - leal saved_return_addr, %edx - movl (%edx), %ecx - pushl %ecx - ret + jmpl *(saved_return_addr) .previous -.data +__INITDATA saved_return_addr: .long 0 efi_rt_function_ptr: diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/entry.S linux-2.6.17.13-g-v/arch/i386/kernel/entry.S --- linux-2.6.17.13/arch/i386/kernel/entry.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/entry.S 2006-09-20 15:00:12.000000000 +0200 @@ -82,7 +82,7 @@ VM_MASK = 0x00020000 #define resume_kernel restore_nocheck #endif -#define SAVE_ALL \ +#define __SAVE_ALL(_DS) \ cld; \ pushl %es; \ pushl %ds; \ @@ -93,10 +93,24 @@ VM_MASK = 0x00020000 pushl %edx; \ pushl %ecx; \ pushl %ebx; \ - movl $(__USER_DS), %edx; \ + movl $(_DS), %edx; \ movl %edx, %ds; \ movl %edx, %es; +#ifdef CONFIG_PAX_KERNEXEC +#define SAVE_ALL \ + __SAVE_ALL(__KERNEL_DS) \ + movl %cr0, %edx; \ + movl %edx, %esi; \ + orl $0x10000, %edx; \ + xorl %edx, %esi; \ + movl %edx, %cr0; +#elif defined(CONFIG_PAX_NOVSYSCALL) +#define SAVE_ALL __SAVE_ALL(__KERNEL_DS) +#else +#define SAVE_ALL __SAVE_ALL(__USER_DS) +#endif + #define RESTORE_INT_REGS \ popl %ebx; \ popl %ecx; \ @@ -146,7 +160,19 @@ ret_from_intr: movl EFLAGS(%esp), %eax # mix EFLAGS and CS movb CS(%esp), %al testl $(VM_MASK | 3), %eax + +#ifdef CONFIG_PAX_KERNEXEC + jnz resume_userspace + + movl %cr0, %edx + xorl %esi, %edx + movl %edx, %cr0 + jmp resume_kernel +#else jz resume_kernel +#endif + + ENTRY(resume_userspace) cli # make sure we don't miss an interrupt # setting need_resched or sigpending @@ -190,9 +216,17 @@ sysenter_past_esp: * Load the potential sixth argument from user stack. * Careful about security. */ + +#ifdef CONFIG_PAX_MEMORY_UDEREF + pushl $(__USER_DS) + pop %ds +1: movl %ds:(%ebp),%ebp +#else cmpl $__PAGE_OFFSET-3,%ebp jae syscall_fault 1: movl (%ebp),%ebp +#endif + .section __ex_table,"a" .align 4 .long 1b,syscall_fault @@ -213,13 +247,33 @@ sysenter_past_esp: movl TI_flags(%ebp), %ecx testw $_TIF_ALLWORK_MASK, %cx jne syscall_exit_work + +#ifdef CONFIG_PAX_RANDKSTACK + pushl %eax + call pax_randomize_kstack + popl %eax +#endif + /* if something modifies registers it must also disable sysexit */ movl EIP(%esp), %edx movl OLDESP(%esp), %ecx +1: mov DS(%esp), %ds +2: mov ES(%esp), %es xorl %ebp,%ebp sti sysexit +.section .fixup,"ax" +3: movl $0,DS(%esp) + jmp 1b +4: movl $0,ES(%esp) + jmp 2b +.previous +.section __ex_table,"a" + .align 4 + .long 1b,3b + .long 2b,4b +.previous # system call handler stub ENTRY(system_call) @@ -247,6 +301,10 @@ syscall_exit: testw $_TIF_ALLWORK_MASK, %cx # current->work jne syscall_exit_work +#ifdef CONFIG_PAX_RANDKSTACK + call pax_randomize_kstack +#endif + restore_all: movl EFLAGS(%esp), %eax # mix EFLAGS, SS and CS # Warning: OLDSS(%esp) contains the wrong/random values if we @@ -402,7 +460,7 @@ syscall_badsys: * Build the entry stubs and pointer table with * some assembler magic. */ -.data +.section .rodata,"a",@progbits ENTRY(interrupt) .text @@ -412,7 +470,7 @@ ENTRY(irq_entries_start) ALIGN 1: pushl $vector-256 jmp common_interrupt -.data +.section .rodata,"a",@progbits .long 1b .text vector=vector+1 @@ -459,10 +517,19 @@ error_code: movl ORIG_EAX(%esp), %edx # get the error code movl %eax, ORIG_EAX(%esp) movl %ecx, ES(%esp) - movl $(__USER_DS), %ecx + movl $(__KERNEL_DS), %ecx movl %ecx, %ds movl %ecx, %es movl %esp,%eax # pt_regs pointer + +#ifdef CONFIG_PAX_KERNEXEC + movl %cr0, %ecx + movl %ecx, %esi + orl $0x10000, %ecx + xorl %ecx, %esi + movl %ecx, %cr0 +#endif + call *%edi jmp ret_from_exception @@ -558,6 +625,13 @@ nmi_stack_correct: xorl %edx,%edx # zero error code movl %esp,%eax # pt_regs pointer call do_nmi + +#ifdef CONFIG_PAX_KERNEXEC + movl %cr0, %edx + xorl %esi, %edx + movl %edx, %cr0 +#endif + jmp restore_all nmi_stack_fixup: @@ -588,6 +662,13 @@ nmi_16bit_stack: FIXUP_ESPFIX_STACK # %eax == %esp xorl %edx,%edx # zero error code call do_nmi + +#ifdef CONFIG_PAX_KERNEXEC + movl %cr0, %edx + xorl %esi, %edx + movl %edx, %cr0 +#endif + RESTORE_REGS lss 12+4(%esp), %esp # back to 16bit stack 1: iret @@ -663,7 +744,6 @@ ENTRY(spurious_interrupt_bug) pushl $do_spurious_interrupt_bug jmp error_code -.section .rodata,"a" #include "syscall_table.S" syscall_table_size=(.-sys_call_table) diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/head.S linux-2.6.17.13-g-v/arch/i386/kernel/head.S --- linux-2.6.17.13/arch/i386/kernel/head.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/head.S 2006-09-20 15:00:12.000000000 +0200 @@ -46,6 +46,16 @@ */ #define INIT_MAP_BEYOND_END (128*1024) +#ifdef CONFIG_PAX_KERNEXEC +/* PaX: fill first page in .text with int3 to catch NULL derefs in kernel mode */ +.fill 4096,1,0xcc +#endif + +/* + * Real beginning of normal "text" segment + */ +ENTRY(stext) +ENTRY(_stext) /* * 32-bit kernel entrypoint; only used by the boot CPU. On entry, @@ -67,6 +77,26 @@ ENTRY(startup_32) movl %eax,%fs movl %eax,%gs +#ifdef CONFIG_PAX_MEMORY_UDEREF + movl $((((__PAGE_OFFSET-1) & 0xf0000000) >> 12) | 0x00c09700),%eax + movl %eax,(cpu_gdt_table - __PAGE_OFFSET + GDT_ENTRY_KERNEL_DS * 8 + 4) + movl $((((__PAGE_OFFSET-1) & 0xf0000000) >> 12) | 0x00c0f300),%eax + movl %eax,(cpu_gdt_table - __PAGE_OFFSET + GDT_ENTRY_DEFAULT_USER_DS * 8 + 4) +#endif + +#ifdef CONFIG_PAX_KERNEXEC + movl $ __KERNEL_TEXT_OFFSET,%eax + movw %ax,(cpu_gdt_table - __PAGE_OFFSET + __KERNEL_CS + 2) + rorl $16,%eax + movb %al,(cpu_gdt_table - __PAGE_OFFSET + __KERNEL_CS + 4) + movb %ah,(cpu_gdt_table - __PAGE_OFFSET + __KERNEL_CS + 7) + + movb %al,(boot_gdt_table - __PAGE_OFFSET + __BOOT_CS + 4) + movb %ah,(boot_gdt_table - __PAGE_OFFSET + __BOOT_CS + 7) + rorl $16,%eax + movw %ax,(boot_gdt_table - __PAGE_OFFSET + __BOOT_CS + 2) +#endif + /* * Clear BSS first so that there are no surprises... * No need to cld as DF is already clear from cld above... @@ -114,24 +144,42 @@ ENTRY(startup_32) * Warning: don't use %esi or the stack in this code. However, %esp * can be used as a GPR if you really need it... */ -page_pde_offset = (__PAGE_OFFSET >> 20); - +#ifdef CONFIG_X86_PAE +page_pde_offset = ((__PAGE_OFFSET >> 21) * (4096 / PTRS_PER_PTE_asm)); +#else +page_pde_offset = ((__PAGE_OFFSET >> 22) * (4096 / PTRS_PER_PTE_asm)); +#endif movl $(pg0 - __PAGE_OFFSET), %edi +#ifdef CONFIG_X86_PAE + movl $(swapper_pm_dir - __PAGE_OFFSET), %edx +#else movl $(swapper_pg_dir - __PAGE_OFFSET), %edx - movl $0x007, %eax /* 0x007 = PRESENT+RW+USER */ +#endif + movl $0x063, %eax /* 0x063 = DIRTY+ACCESSED+PRESENT+RW */ 10: - leal 0x007(%edi),%ecx /* Create PDE entry */ + leal 0x063(%edi),%ecx /* Create PDE entry */ movl %ecx,(%edx) /* Store identity PDE entry */ movl %ecx,page_pde_offset(%edx) /* Store kernel PDE entry */ +#ifdef CONFIG_X86_PAE + movl $0,4(%edx) + movl $0,page_pde_offset+4(%edx) + addl $8,%edx + movl $512, %ecx +#else addl $4,%edx movl $1024, %ecx +#endif 11: stosl +#ifdef CONFIG_X86_PAE + movl $0,(%edi) + addl $4,%edi +#endif addl $0x1000,%eax loop 11b /* End condition: we must map up to and including INIT_MAP_BEYOND_END */ - /* bytes beyond the end of our own page tables; the +0x007 is the attribute bits */ - leal (INIT_MAP_BEYOND_END+0x007)(%edi),%ebp + /* bytes beyond the end of our own page tables; the +0x063 is the attribute bits */ + leal (INIT_MAP_BEYOND_END+0x063)(%edi),%ebp cmpl %ebp,%eax jb 10b movl %edi,(init_pg_tables_end - __PAGE_OFFSET) @@ -154,6 +202,11 @@ ENTRY(startup_32_smp) movl %eax,%fs movl %eax,%gs + /* This is a secondary processor (AP) */ + xorl %ebx,%ebx + incl %ebx +#endif /* CONFIG_SMP */ + /* * New page tables may be in 4Mbyte page mode and may * be using the global pages. @@ -169,26 +222,27 @@ ENTRY(startup_32_smp) * not yet offset PAGE_OFFSET.. */ #define cr4_bits mmu_cr4_features-__PAGE_OFFSET +3: movl cr4_bits,%edx andl %edx,%edx - jz 6f + jz 5f movl %cr4,%eax # Turn on paging options (PSE,PAE,..) orl %edx,%eax movl %eax,%cr4 - btl $5, %eax # check if PAE is enabled - jnc 6f +#ifdef CONFIG_X86_PAE + movl %ebx,%edi /* Check if extended functions are implemented */ movl $0x80000000, %eax cpuid cmpl $0x80000000, %eax - jbe 6f + jbe 4f mov $0x80000001, %eax cpuid /* Execute Disable bit supported? */ btl $20, %edx - jnc 6f + jnc 4f /* Setup EFER (Extended Feature Enable Register) */ movl $0xc0000080, %ecx @@ -197,14 +251,12 @@ ENTRY(startup_32_smp) btsl $11, %eax /* Make changes effective */ wrmsr + btsl $63,__supported_pte_mask-__PAGE_OFFSET -6: - /* This is a secondary processor (AP) */ - xorl %ebx,%ebx - incl %ebx - -3: -#endif /* CONFIG_SMP */ +4: + movl %edi,%ebx +#endif +5: /* * Enable paging @@ -229,9 +281,7 @@ ENTRY(startup_32_smp) #ifdef CONFIG_SMP andl %ebx,%ebx - jz 1f /* Initial CPU cleans BSS */ - jmp checkCPUtype -1: + jnz checkCPUtype /* Initial CPU cleans BSS */ #endif /* CONFIG_SMP */ /* @@ -308,8 +358,6 @@ is386: movl $2,%ecx # set MP ljmp $(__KERNEL_CS),$1f 1: movl $(__KERNEL_DS),%eax # reload all the segment registers movl %eax,%ss # after changing gdt. - - movl $(__USER_DS),%eax # DS/ES contains default USER segment movl %eax,%ds movl %eax,%es @@ -383,63 +431,71 @@ rp_sidt: /* This is the default interrupt "handler" :-) */ ALIGN ignore_int: - cld #ifdef CONFIG_PRINTK - pushl %eax - pushl %ecx - pushl %edx - pushl %es - pushl %ds + cld movl $(__KERNEL_DS),%eax movl %eax,%ds movl %eax,%es - pushl 16(%esp) - pushl 24(%esp) - pushl 32(%esp) - pushl 40(%esp) + pushl 12(%esp) + pushl 12(%esp) + pushl 12(%esp) + pushl 12(%esp) pushl $int_msg #ifdef CONFIG_EARLY_PRINTK call early_printk #else call printk #endif - addl $(5*4),%esp - popl %ds - popl %es - popl %edx - popl %ecx - popl %eax #endif - iret - -/* - * Real beginning of normal "text" segment - */ -ENTRY(stext) -ENTRY(_stext) +1: hlt + jmp 1b -/* - * BSS section - */ -.section ".bss.page_aligned","w" +.section .swapper_pg_dir,"a",@progbits ENTRY(swapper_pg_dir) +#ifdef CONFIG_X86_PAE + .long swapper_pm_dir-__PAGE_OFFSET+1 + .long 0 + .long swapper_pm_dir+512*8-__PAGE_OFFSET+1 + .long 0 + .long swapper_pm_dir+512*16-__PAGE_OFFSET+1 + .long 0 + .long swapper_pm_dir+512*24-__PAGE_OFFSET+1 + .long 0 +#else .fill 1024,4,0 +#endif + +#ifdef CONFIG_X86_PAE +.section .swapper_pm_dir,"a",@progbits +ENTRY(swapper_pm_dir) + .fill 512,8,0 + .fill 512,8,0 + .fill 512,8,0 + .fill 512,8,0 +#endif + +.section .empty_zero_page,"a",@progbits ENTRY(empty_zero_page) .fill 4096,1,0 /* - * This starts the data section. + * The IDT has to be page-aligned to simplify the Pentium + * F0 0F bug workaround.. We have a special link segment + * for this. */ -.data +.section .idt,"a",@progbits +ENTRY(idt_table) + .fill 256,8,0 + +.section .rodata,"a",@progbits +ready: .byte 0 ENTRY(stack_start) - .long init_thread_union+THREAD_SIZE + .long init_thread_union+THREAD_SIZE-8 .long __BOOT_DS -ready: .byte 0 - int_msg: - .asciz "Unknown interrupt or fault at EIP %p %p %p\n" + .asciz "Unknown interrupt, stack: %p %p %p %p\n" /* * The IDT and GDT 'descriptors' are a strange 48-bit object @@ -465,10 +521,12 @@ idt_descr: # boot GDT descriptor (later on used by CPU#0): .word 0 # 32 bit align gdt_desc.address -cpu_gdt_descr: +ENTRY(cpu_gdt_descr) .word GDT_ENTRIES*8-1 .long cpu_gdt_table + .fill NR_CPUS*8-6,1,0 # space for the other GDT descriptors + /* * The boot_gdt_table must mirror the equivalent in setup.S and is * used only for booting. @@ -476,13 +534,13 @@ cpu_gdt_descr: .align L1_CACHE_BYTES ENTRY(boot_gdt_table) .fill GDT_ENTRY_BOOT_CS,8,0 - .quad 0x00cf9a000000ffff /* kernel 4GB code at 0x00000000 */ - .quad 0x00cf92000000ffff /* kernel 4GB data at 0x00000000 */ + .quad 0x00cf9b000000ffff /* kernel 4GB code at 0x00000000 */ + .quad 0x00cf93000000ffff /* kernel 4GB data at 0x00000000 */ /* * The Global Descriptor Table contains 28 quadwords, per-CPU. */ - .align L1_CACHE_BYTES + .align PAGE_SIZE_asm ENTRY(cpu_gdt_table) .quad 0x0000000000000000 /* NULL descriptor */ .quad 0x0000000000000000 /* 0x0b reserved */ @@ -497,10 +555,10 @@ ENTRY(cpu_gdt_table) .quad 0x0000000000000000 /* 0x53 reserved */ .quad 0x0000000000000000 /* 0x5b reserved */ - .quad 0x00cf9a000000ffff /* 0x60 kernel 4GB code at 0x00000000 */ - .quad 0x00cf92000000ffff /* 0x68 kernel 4GB data at 0x00000000 */ - .quad 0x00cffa000000ffff /* 0x73 user 4GB code at 0x00000000 */ - .quad 0x00cff2000000ffff /* 0x7b user 4GB data at 0x00000000 */ + .quad 0x00cf9b000000ffff /* 0x60 kernel 4GB code at 0x00000000 */ + .quad 0x00cf93000000ffff /* 0x68 kernel 4GB data at 0x00000000 */ + .quad 0x00cffb000000ffff /* 0x73 user 4GB code at 0x00000000 */ + .quad 0x00cff3000000ffff /* 0x7b user 4GB data at 0x00000000 */ .quad 0x0000000000000000 /* 0x80 TSS descriptor */ .quad 0x0000000000000000 /* 0x88 LDT descriptor */ @@ -510,24 +568,30 @@ ENTRY(cpu_gdt_table) * They code segments and data segments have fixed 64k limits, * the transfer segment sizes are set at run time. */ - .quad 0x00409a000000ffff /* 0x90 32-bit code */ - .quad 0x00009a000000ffff /* 0x98 16-bit code */ - .quad 0x000092000000ffff /* 0xa0 16-bit data */ - .quad 0x0000920000000000 /* 0xa8 16-bit data */ - .quad 0x0000920000000000 /* 0xb0 16-bit data */ + .quad 0x00409b000000ffff /* 0x90 32-bit code */ + .quad 0x00009b000000ffff /* 0x98 16-bit code */ + .quad 0x000093000000ffff /* 0xa0 16-bit data */ + .quad 0x0000930000000000 /* 0xa8 16-bit data */ + .quad 0x0000930000000000 /* 0xb0 16-bit data */ /* * The APM segments have byte granularity and their bases * are set at run time. All have 64k limits. */ - .quad 0x00409a000000ffff /* 0xb8 APM CS code */ - .quad 0x00009a000000ffff /* 0xc0 APM CS 16 code (16 bit) */ - .quad 0x004092000000ffff /* 0xc8 APM DS data */ + .quad 0x00409b000000ffff /* 0xb8 APM CS code */ + .quad 0x00009b000000ffff /* 0xc0 APM CS 16 code (16 bit) */ + .quad 0x004093000000ffff /* 0xc8 APM DS data */ - .quad 0x0000920000000000 /* 0xd0 - ESPFIX 16-bit SS */ + .quad 0x0000930000000000 /* 0xd0 - ESPFIX 16-bit SS */ .quad 0x0000000000000000 /* 0xd8 - unused */ .quad 0x0000000000000000 /* 0xe0 - unused */ .quad 0x0000000000000000 /* 0xe8 - unused */ .quad 0x0000000000000000 /* 0xf0 - unused */ .quad 0x0000000000000000 /* 0xf8 - GDT entry 31: double-fault TSS */ + /* Be sure this is zeroed to avoid false validations in Xen */ + .fill PAGE_SIZE_asm / 8 - GDT_ENTRIES,8,0 + +#ifdef CONFIG_SMP + .fill (NR_CPUS-1) * (PAGE_SIZE_asm / 8),8,0 /* other CPU's GDT */ +#endif diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/i386_ksyms.c linux-2.6.17.13-g-v/arch/i386/kernel/i386_ksyms.c --- linux-2.6.17.13/arch/i386/kernel/i386_ksyms.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/i386_ksyms.c 2006-09-20 15:00:12.000000000 +0200 @@ -3,12 +3,16 @@ #include #include +EXPORT_SYMBOL_GPL(cpu_gdt_table); + EXPORT_SYMBOL(__down_failed); EXPORT_SYMBOL(__down_failed_interruptible); EXPORT_SYMBOL(__down_failed_trylock); EXPORT_SYMBOL(__up_wakeup); /* Networking helper routines. */ EXPORT_SYMBOL(csum_partial_copy_generic); +EXPORT_SYMBOL(csum_partial_copy_generic_to_user); +EXPORT_SYMBOL(csum_partial_copy_generic_from_user); EXPORT_SYMBOL(__get_user_1); EXPORT_SYMBOL(__get_user_2); diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/init_task.c linux-2.6.17.13-g-v/arch/i386/kernel/init_task.c --- linux-2.6.17.13/arch/i386/kernel/init_task.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/init_task.c 2006-09-20 15:00:12.000000000 +0200 @@ -42,5 +42,5 @@ EXPORT_SYMBOL(init_task); * per-CPU TSS segments. Threads are completely 'soft' on Linux, * no more per-task TSS's. */ -DEFINE_PER_CPU(struct tss_struct, init_tss) ____cacheline_internodealigned_in_smp = INIT_TSS; +struct tss_struct init_tss[NR_CPUS] ____cacheline_internodealigned_in_smp = { [0 ... NR_CPUS-1] = INIT_TSS }; diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/ioport.c linux-2.6.17.13-g-v/arch/i386/kernel/ioport.c --- linux-2.6.17.13/arch/i386/kernel/ioport.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/ioport.c 2006-09-20 15:00:12.000000000 +0200 @@ -16,6 +16,7 @@ #include #include #include +#include /* Set EXTENT bits starting at BASE in BITMAP to value TURN_ON. */ static void set_bitmap(unsigned long *bitmap, unsigned int base, unsigned int extent, int new_value) @@ -64,9 +65,16 @@ asmlinkage long sys_ioperm(unsigned long if ((from + num <= from) || (from + num > IO_BITMAP_BITS)) return -EINVAL; +#ifdef CONFIG_GRKERNSEC_IO + if (turn_on) { + gr_handle_ioperm(); +#else if (turn_on && !capable(CAP_SYS_RAWIO)) +#endif return -EPERM; - +#ifdef CONFIG_GRKERNSEC_IO + } +#endif /* * If it's the first ioperm() call in this thread's lifetime, set the * IO bitmap up. ioperm() is much less timing critical than clone(), @@ -88,7 +96,7 @@ asmlinkage long sys_ioperm(unsigned long * because the ->io_bitmap_max value must match the bitmap * contents: */ - tss = &per_cpu(init_tss, get_cpu()); + tss = init_tss + get_cpu(); set_bitmap(t->io_bitmap_ptr, from, num, !turn_on); @@ -142,8 +150,13 @@ asmlinkage long sys_iopl(unsigned long u return -EINVAL; /* Trying to gain more privileges? */ if (level > old) { +#ifdef CONFIG_GRKERNSEC_IO + gr_handle_iopl(); + return -EPERM; +#else if (!capable(CAP_SYS_RAWIO)) return -EPERM; +#endif } t->iopl = level << 12; regs->eflags = (regs->eflags & ~X86_EFLAGS_IOPL) | t->iopl; diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/irq.c linux-2.6.17.13-g-v/arch/i386/kernel/irq.c --- linux-2.6.17.13/arch/i386/kernel/irq.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/irq.c 2006-09-20 15:00:12.000000000 +0200 @@ -91,7 +91,7 @@ fastcall unsigned int do_IRQ(struct pt_r int arg1, arg2, ebx; /* build the stack frame on the IRQ stack */ - isp = (u32*) ((char*)irqctx + sizeof(*irqctx)); + isp = (u32*) ((char*)irqctx + sizeof(*irqctx)) - 2; irqctx->tinfo.task = curctx->tinfo.task; irqctx->tinfo.previous_esp = current_stack_pointer; @@ -119,10 +119,10 @@ fastcall unsigned int do_IRQ(struct pt_r * gcc's 3.0 and earlier don't handle that correctly. */ static char softirq_stack[NR_CPUS * THREAD_SIZE] - __attribute__((__aligned__(THREAD_SIZE))); + __attribute__((__aligned__(THREAD_SIZE), __section__(".bss.page_aligned"))); static char hardirq_stack[NR_CPUS * THREAD_SIZE] - __attribute__((__aligned__(THREAD_SIZE))); + __attribute__((__aligned__(THREAD_SIZE), __section__(".bss.page_aligned"))); /* * allocate per-cpu stacks for hardirq and for softirq processing @@ -182,7 +182,7 @@ asmlinkage void do_softirq(void) irqctx->tinfo.previous_esp = current_stack_pointer; /* build the stack frame on the softirq stack */ - isp = (u32*) ((char*)irqctx + sizeof(*irqctx)); + isp = (u32*) ((char*)irqctx + sizeof(*irqctx)) - 2; asm volatile( " xchgl %%ebx,%%esp \n" diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/ldt.c linux-2.6.17.13-g-v/arch/i386/kernel/ldt.c --- linux-2.6.17.13/arch/i386/kernel/ldt.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/ldt.c 2006-09-20 15:00:12.000000000 +0200 @@ -103,6 +103,19 @@ int init_new_context(struct task_struct retval = copy_ldt(&mm->context, &old_mm->context); up(&old_mm->context.sem); } + +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) + if (!mm->context.user_cs_limit) { + mm->context.user_cs_base = 0UL; + mm->context.user_cs_limit = ~0UL; + +#if defined(CONFIG_PAX_PAGEEXEC) && defined(CONFIG_SMP) + cpus_clear(mm->context.cpu_user_cs_mask); +#endif + + } +#endif + return retval; } @@ -160,7 +173,7 @@ static int read_default_ldt(void __user { int err; unsigned long size; - void *address; + const void *address; err = 0; address = &default_ldt[0]; @@ -215,6 +228,13 @@ static int write_ldt(void __user * ptr, } } +#ifdef CONFIG_PAX_SEGMEXEC + if ((mm->pax_flags & MF_PAX_SEGMEXEC) && (ldt_info.contents & MODIFY_LDT_CONTENTS_CODE)) { + error = -EINVAL; + goto out_unlock; + } +#endif + entry_1 = LDT_entry_a(&ldt_info); entry_2 = LDT_entry_b(&ldt_info); if (oldmode) diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/module.c linux-2.6.17.13-g-v/arch/i386/kernel/module.c --- linux-2.6.17.13/arch/i386/kernel/module.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/module.c 2006-09-20 15:00:12.000000000 +0200 @@ -21,6 +21,7 @@ #include #include #include +#include #if 0 #define DEBUGP printk @@ -32,9 +33,30 @@ void *module_alloc(unsigned long size) { if (size == 0) return NULL; + +#ifdef CONFIG_PAX_KERNEXEC + return vmalloc(size); +#else return vmalloc_exec(size); +#endif + } +#ifdef CONFIG_PAX_KERNEXEC +void *module_alloc_exec(unsigned long size) +{ + struct vm_struct *area; + + if (size == 0) + return NULL; + + area = __get_vm_area(size, 0, (unsigned long)&MODULES_VADDR, (unsigned long)&MODULES_END); + if (area) + return area->addr; + + return NULL; +} +#endif /* Free memory returned from module_alloc */ void module_free(struct module *mod, void *module_region) @@ -44,6 +66,45 @@ void module_free(struct module *mod, voi table entries. */ } +#ifdef CONFIG_PAX_KERNEXEC +void module_free_exec(struct module *mod, void *module_region) +{ + struct vm_struct **p, *tmp; + + if (!module_region) + return; + + if ((PAGE_SIZE-1) & (unsigned long)module_region) { + printk(KERN_ERR "Trying to module_free_exec() bad address (%p)\n", module_region); + WARN_ON(1); + return; + } + + write_lock(&vmlist_lock); + for (p = &vmlist ; (tmp = *p) != NULL ;p = &tmp->next) + if (tmp->addr == module_region) + break; + + if (tmp) { + unsigned long cr0; + + pax_open_kernel(cr0); + memset(tmp->addr, 0xCC, tmp->size); + pax_close_kernel(cr0); + + *p = tmp->next; + kfree(tmp); + } + write_unlock(&vmlist_lock); + + if (!tmp) { + printk(KERN_ERR "Trying to module_free_exec() nonexistent vm area (%p)\n", + module_region); + WARN_ON(1); + } +} +#endif + /* We don't need anything special. */ int module_frob_arch_sections(Elf_Ehdr *hdr, Elf_Shdr *sechdrs, @@ -62,14 +123,16 @@ int apply_relocate(Elf32_Shdr *sechdrs, unsigned int i; Elf32_Rel *rel = (void *)sechdrs[relsec].sh_addr; Elf32_Sym *sym; - uint32_t *location; + uint32_t *plocation, location; DEBUGP("Applying relocate section %u to %u\n", relsec, sechdrs[relsec].sh_info); for (i = 0; i < sechdrs[relsec].sh_size / sizeof(*rel); i++) { /* This is where to make the change */ - location = (void *)sechdrs[sechdrs[relsec].sh_info].sh_addr - + rel[i].r_offset; + plocation = (void *)sechdrs[sechdrs[relsec].sh_info].sh_addr + rel[i].r_offset; + location = (uint32_t)plocation; + if (sechdrs[sechdrs[relsec].sh_info].sh_flags & SHF_EXECINSTR) + plocation = (void *)plocation + __KERNEL_TEXT_OFFSET; /* This is the symbol it is referring to. Note that all undefined symbols have been resolved. */ sym = (Elf32_Sym *)sechdrs[symindex].sh_addr @@ -78,11 +141,11 @@ int apply_relocate(Elf32_Shdr *sechdrs, switch (ELF32_R_TYPE(rel[i].r_info)) { case R_386_32: /* We add the value into the location given */ - *location += sym->st_value; + *plocation += sym->st_value; break; case R_386_PC32: /* Add the value, subtract its postition */ - *location += sym->st_value - (uint32_t)location; + *plocation += sym->st_value - location; break; default: printk(KERN_ERR "module %s: Unknown relocation: %u\n", diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/process.c linux-2.6.17.13-g-v/arch/i386/kernel/process.c --- linux-2.6.17.13/arch/i386/kernel/process.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/process.c 2006-09-20 15:00:12.000000000 +0200 @@ -69,7 +69,7 @@ EXPORT_SYMBOL(boot_option_idle_override) */ unsigned long thread_saved_pc(struct task_struct *tsk) { - return ((unsigned long *)tsk->thread.esp)[3]; + return tsk->thread.eip; } /* @@ -294,7 +294,7 @@ void show_regs(struct pt_regs * regs) printk("EIP: %04x:[<%08lx>] CPU: %d\n",0xffff & regs->xcs,regs->eip, smp_processor_id()); print_symbol("EIP is at %s\n", regs->eip); - if (user_mode_vm(regs)) + if (user_mode(regs)) printk(" ESP: %04x:%08lx",0xffff & regs->xss,regs->esp); printk(" EFLAGS: %08lx %s (%s %.*s)\n", regs->eflags, print_tainted(), system_utsname.release, @@ -343,8 +343,8 @@ int kernel_thread(int (*fn)(void *), voi regs.ebx = (unsigned long) fn; regs.edx = (unsigned long) arg; - regs.xds = __USER_DS; - regs.xes = __USER_DS; + regs.xds = __KERNEL_DS; + regs.xes = __KERNEL_DS; regs.orig_eax = -1; regs.eip = (unsigned long) kernel_thread_helper; regs.xcs = __KERNEL_CS; @@ -366,7 +366,7 @@ void exit_thread(void) /* The process may have allocated an io port bitmap... nuke it. */ if (unlikely(NULL != t->io_bitmap_ptr)) { int cpu = get_cpu(); - struct tss_struct *tss = &per_cpu(init_tss, cpu); + struct tss_struct *tss = init_tss + cpu; kfree(t->io_bitmap_ptr); t->io_bitmap_ptr = NULL; @@ -386,6 +386,9 @@ void flush_thread(void) { struct task_struct *tsk = current; + __asm__("mov %0,%%fs\n" + "mov %0,%%gs\n" + : : "r" (0) : "memory"); memset(tsk->thread.debugreg, 0, sizeof(unsigned long)*8); memset(tsk->thread.tls_array, 0, sizeof(tsk->thread.tls_array)); /* @@ -418,7 +421,7 @@ int copy_thread(int nr, unsigned long cl struct task_struct *tsk; int err; - childregs = task_pt_regs(p); + childregs = task_stack_page(p) + THREAD_SIZE - sizeof(struct pt_regs) - 8; *childregs = *regs; childregs->eax = 0; childregs->esp = esp; @@ -461,6 +464,11 @@ int copy_thread(int nr, unsigned long cl if (idx < GDT_ENTRY_TLS_MIN || idx > GDT_ENTRY_TLS_MAX) goto out; +#ifdef CONFIG_PAX_SEGMEXEC + if ((current->mm->pax_flags & MF_PAX_SEGMEXEC) && (info.contents & MODIFY_LDT_CONTENTS_CODE)) + goto out; +#endif + desc = p->thread.tls_array + idx - GDT_ENTRY_TLS_MIN; desc->a = LDT_entry_a(&info); desc->b = LDT_entry_b(&info); @@ -625,7 +633,11 @@ struct task_struct fastcall * __switch_t struct thread_struct *prev = &prev_p->thread, *next = &next_p->thread; int cpu = smp_processor_id(); - struct tss_struct *tss = &per_cpu(init_tss, cpu); + struct tss_struct *tss = init_tss + cpu; + +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; +#endif /* never put a printk in __switch_to... printk() calls wake_up*() indirectly */ @@ -648,11 +660,23 @@ struct task_struct fastcall * __switch_t savesegment(fs, prev->fs); savesegment(gs, prev->gs); +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + +#ifdef CONFIG_PAX_MEMORY_UDEREF + __set_fs(get_fs(), cpu); +#endif + /* * Load the per-thread Thread-Local Storage descriptor. */ load_TLS(next, cpu); +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + /* * Restore %fs and %gs if needed. * @@ -806,8 +830,18 @@ asmlinkage int sys_set_thread_area(struc struct desc_struct *desc; int cpu, idx; +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; +#endif + if (copy_from_user(&info, u_info, sizeof(info))) return -EFAULT; + +#ifdef CONFIG_PAX_SEGMEXEC + if ((current->mm->pax_flags & MF_PAX_SEGMEXEC) && (info.contents & MODIFY_LDT_CONTENTS_CODE)) + return -EINVAL; +#endif + idx = info.entry_number; /* @@ -839,8 +873,17 @@ asmlinkage int sys_set_thread_area(struc desc->a = LDT_entry_a(&info); desc->b = LDT_entry_b(&info); } + +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + load_TLS(t, cpu); +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + put_cpu(); return 0; @@ -896,9 +939,27 @@ asmlinkage int sys_get_thread_area(struc return 0; } -unsigned long arch_align_stack(unsigned long sp) +#ifdef CONFIG_PAX_RANDKSTACK +asmlinkage void pax_randomize_kstack(void) { - if (randomize_va_space) - sp -= get_random_int() % 8192; - return sp & ~0xf; + struct tss_struct *tss = init_tss + smp_processor_id(); + unsigned long time; + + if (!randomize_va_space) + return; + + rdtscl(time); + + /* P4 seems to return a 0 LSB, ignore it */ +#ifdef CONFIG_MPENTIUM4 + time &= 0x1EUL; + time <<= 2; +#else + time &= 0xFUL; + time <<= 3; +#endif + + tss->esp0 ^= time; + current->thread.esp0 = tss->esp0; } +#endif diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/ptrace.c linux-2.6.17.13-g-v/arch/i386/kernel/ptrace.c --- linux-2.6.17.13/arch/i386/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -17,6 +17,7 @@ #include #include #include +#include #include #include @@ -342,6 +343,11 @@ ptrace_set_thread_area(struct task_struc if (copy_from_user(&info, user_desc, sizeof(info))) return -EFAULT; +#ifdef CONFIG_PAX_SEGMEXEC + if ((child->mm->pax_flags & MF_PAX_SEGMEXEC) && (info.contents & MODIFY_LDT_CONTENTS_CODE)) + return -EINVAL; +#endif + if (idx < GDT_ENTRY_TLS_MIN || idx > GDT_ENTRY_TLS_MAX) return -EINVAL; @@ -432,6 +438,17 @@ long arch_ptrace(struct task_struct *chi if(addr == (long) &dummy->u_debugreg[5]) break; if(addr < (long) &dummy->u_debugreg[4] && ((unsigned long) data) >= TASK_SIZE-3) break; + +#ifdef CONFIG_GRKERNSEC + if(addr >= (long) &dummy->u_debugreg[0] && + addr <= (long) &dummy->u_debugreg[3]){ + long reg = (addr - (long) &dummy->u_debugreg[0]) >> 2; + long type = (child->thread.debugreg[7] >> (DR_CONTROL_SHIFT + 4*reg)) & 3; + long align = (child->thread.debugreg[7] >> (DR_CONTROL_SHIFT + 2 + 4*reg)) & 3; + if((type & 1) && (data & align)) + break; + } +#endif /* Sanity-check data. Take one half-byte at once with * check = (val >> (16 + 4*i)) & 0xf. It contains the @@ -645,7 +662,7 @@ void send_sigtrap(struct task_struct *ts info.si_code = TRAP_BRKPT; /* User-mode eip? */ - info.si_addr = user_mode_vm(regs) ? (void __user *) regs->eip : NULL; + info.si_addr = user_mode(regs) ? (void __user *) regs->eip : NULL; /* Send us the fakey SIGTRAP */ force_sig_info(SIGTRAP, &info, tsk); diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/reboot.c linux-2.6.17.13-g-v/arch/i386/kernel/reboot.c --- linux-2.6.17.13/arch/i386/kernel/reboot.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/reboot.c 2006-09-20 15:00:12.000000000 +0200 @@ -25,7 +25,7 @@ void (*pm_power_off)(void); EXPORT_SYMBOL(pm_power_off); -static int reboot_mode; +static unsigned short reboot_mode; static int reboot_thru_bios; #ifdef CONFIG_SMP @@ -138,18 +138,18 @@ core_initcall(reboot_init); doesn't work with at least one type of 486 motherboard. It is easy to stop this code working; hence the copious comments. */ -static unsigned long long +static const unsigned long long real_mode_gdt_entries [3] = { 0x0000000000000000ULL, /* Null descriptor */ - 0x00009a000000ffffULL, /* 16-bit real-mode 64k code at 0x00000000 */ - 0x000092000100ffffULL /* 16-bit real-mode 64k data at 0x00000100 */ + 0x00009b000000ffffULL, /* 16-bit real-mode 64k code at 0x00000000 */ + 0x000093000100ffffULL /* 16-bit real-mode 64k data at 0x00000100 */ }; static struct { unsigned short size __attribute__ ((packed)); - unsigned long long * base __attribute__ ((packed)); + const unsigned long long * base __attribute__ ((packed)); } real_mode_gdt = { sizeof (real_mode_gdt_entries) - 1, real_mode_gdt_entries }, real_mode_idt = { 0x3ff, NULL }, @@ -175,7 +175,7 @@ no_idt = { 0, NULL }; More could be done here to set up the registers as if a CPU reset had occurred; hopefully real BIOSs don't assume much. */ -static unsigned char real_mode_switch [] = +static const unsigned char real_mode_switch [] = { 0x66, 0x0f, 0x20, 0xc0, /* movl %cr0,%eax */ 0x66, 0x83, 0xe0, 0x11, /* andl $0x00000011,%eax */ @@ -189,7 +189,7 @@ static unsigned char real_mode_switch [] 0x24, 0x10, /* f: andb $0x10,al */ 0x66, 0x0f, 0x22, 0xc0 /* movl %eax,%cr0 */ }; -static unsigned char jump_to_bios [] = +static const unsigned char jump_to_bios [] = { 0xea, 0x00, 0x00, 0xff, 0xff /* ljmp $0xffff,$0x0000 */ }; @@ -199,10 +199,14 @@ static unsigned char jump_to_bios [] = * specified by the code and length parameters. * We assume that length will aways be less that 100! */ -void machine_real_restart(unsigned char *code, int length) +void machine_real_restart(const unsigned char *code, unsigned int length) { unsigned long flags; +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; +#endif + local_irq_disable(); /* Write zero to CMOS register number 0x0f, which the BIOS POST @@ -223,8 +227,16 @@ void machine_real_restart(unsigned char from the kernel segment. This assumes the kernel segment starts at virtual address PAGE_OFFSET. */ - memcpy (swapper_pg_dir, swapper_pg_dir + USER_PGD_PTRS, - sizeof (swapper_pg_dir [0]) * KERNEL_PGD_PTRS); +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + + clone_pgd_range(swapper_pg_dir, swapper_pg_dir + USER_PGD_PTRS, + USER_PGD_PTRS >= KERNEL_PGD_PTRS ? KERNEL_PGD_PTRS : USER_PGD_PTRS); + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif /* * Use `swapper_pg_dir' as our page directory. @@ -237,7 +249,7 @@ void machine_real_restart(unsigned char REBOOT.COM programs, and the previous reset routine did this too. */ - *((unsigned short *)0x472) = reboot_mode; + __put_user(reboot_mode, (unsigned short __user *)0x472); /* For the switch to real mode, copy some code to low memory. It has to be in the first 64k because it is running in 16-bit mode, and it @@ -245,9 +257,9 @@ void machine_real_restart(unsigned char off paging. Copy it near the end of the first page, out of the way of BIOS variables. */ - memcpy ((void *) (0x1000 - sizeof (real_mode_switch) - 100), + flags = __copy_to_user ((void __user *) (0x1000 - sizeof (real_mode_switch) - 100), real_mode_switch, sizeof (real_mode_switch)); - memcpy ((void *) (0x1000 - 100), code, length); + flags = __copy_to_user ((void __user *) (0x1000 - 100), code, length); /* Set up the IDT for real mode. */ @@ -329,7 +341,7 @@ void machine_emergency_restart(void) __asm__ __volatile__("int3"); } /* rebooting needs to touch the page at absolute addr 0 */ - *((unsigned short *)__va(0x472)) = reboot_mode; + __put_user(reboot_mode, (unsigned short __user *)0x472); for (;;) { mach_reboot_fixups(); /* for board specific fixups */ mach_reboot(); diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/setup.c linux-2.6.17.13-g-v/arch/i386/kernel/setup.c --- linux-2.6.17.13/arch/i386/kernel/setup.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/setup.c 2006-09-20 15:00:12.000000000 +0200 @@ -88,7 +88,11 @@ struct cpuinfo_x86 new_cpu_data __initda struct cpuinfo_x86 boot_cpu_data __read_mostly = { 0, 0, 0, 0, -1, 1, 0, 0, -1 }; EXPORT_SYMBOL(boot_cpu_data); +#ifdef CONFIG_X86_PAE +unsigned long mmu_cr4_features = X86_CR4_PAE; +#else unsigned long mmu_cr4_features; +#endif #ifdef CONFIG_ACPI int acpi_disabled = 0; @@ -1493,14 +1497,14 @@ void __init setup_arch(char **cmdline_p) if (!MOUNT_ROOT_RDONLY) root_mountflags &= ~MS_RDONLY; - init_mm.start_code = (unsigned long) _text; - init_mm.end_code = (unsigned long) _etext; + init_mm.start_code = (unsigned long) _text + __KERNEL_TEXT_OFFSET; + init_mm.end_code = (unsigned long) _etext + __KERNEL_TEXT_OFFSET; init_mm.end_data = (unsigned long) _edata; init_mm.brk = init_pg_tables_end + PAGE_OFFSET; - code_resource.start = virt_to_phys(_text); - code_resource.end = virt_to_phys(_etext)-1; - data_resource.start = virt_to_phys(_etext); + code_resource.start = virt_to_phys(_text + __KERNEL_TEXT_OFFSET); + code_resource.end = virt_to_phys(_etext + __KERNEL_TEXT_OFFSET)-1; + data_resource.start = virt_to_phys(_data); data_resource.end = virt_to_phys(_edata)-1; parse_cmdline_early(cmdline_p); diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/signal.c linux-2.6.17.13-g-v/arch/i386/kernel/signal.c --- linux-2.6.17.13/arch/i386/kernel/signal.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/signal.c 2006-09-20 15:00:12.000000000 +0200 @@ -351,7 +351,17 @@ static int setup_frame(int sig, struct k goto give_sigsegv; } +#ifdef CONFIG_PAX_NOVSYSCALL + restorer = frame->retcode; +#else restorer = &__kernel_sigreturn; + +#ifdef CONFIG_PAX_SEGMEXEC + if (current->mm->pax_flags & MF_PAX_SEGMEXEC) + restorer -= SEGMEXEC_TASK_SIZE; +#endif +#endif + if (ka->sa.sa_flags & SA_RESTORER) restorer = ka->sa.sa_restorer; @@ -447,7 +457,18 @@ static int setup_rt_frame(int sig, struc goto give_sigsegv; /* Set up to return from userspace. */ + +#ifdef CONFIG_PAX_NOVSYSCALL + restorer = frame->retcode; +#else restorer = &__kernel_rt_sigreturn; + +#ifdef CONFIG_PAX_SEGMEXEC + if (current->mm->pax_flags & MF_PAX_SEGMEXEC) + restorer -= SEGMEXEC_TASK_SIZE; +#endif +#endif + if (ka->sa.sa_flags & SA_RESTORER) restorer = ka->sa.sa_restorer; err |= __put_user(restorer, &frame->pretcode); @@ -580,7 +601,7 @@ static void fastcall do_signal(struct pt * before reaching here, so testing against kernel * CS suffices. */ - if (!user_mode(regs)) + if (!user_mode_novm(regs)) return; if (test_thread_flag(TIF_RESTORE_SIGMASK)) diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/smpboot.c linux-2.6.17.13-g-v/arch/i386/kernel/smpboot.c --- linux-2.6.17.13/arch/i386/kernel/smpboot.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/smpboot.c 2006-09-20 15:00:12.000000000 +0200 @@ -1069,7 +1069,7 @@ static int __cpuinit __smp_prepare_cpu(i /* init low mem mapping */ clone_pgd_range(swapper_pg_dir, swapper_pg_dir + USER_PGD_PTRS, - KERNEL_PGD_PTRS); + USER_PGD_PTRS >= KERNEL_PGD_PTRS ? KERNEL_PGD_PTRS : USER_PGD_PTRS); flush_tlb_all(); schedule_work(&task); wait_for_completion(&done); diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/sys_i386.c linux-2.6.17.13-g-v/arch/i386/kernel/sys_i386.c --- linux-2.6.17.13/arch/i386/kernel/sys_i386.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/sys_i386.c 2006-09-20 15:00:12.000000000 +0200 @@ -19,6 +19,7 @@ #include #include #include +#include #include #include @@ -99,6 +100,191 @@ out: return err; } +unsigned long +arch_get_unmapped_area(struct file *filp, unsigned long addr, + unsigned long len, unsigned long pgoff, unsigned long flags) +{ + struct mm_struct *mm = current->mm; + struct vm_area_struct *vma; + unsigned long start_addr, task_size = TASK_SIZE; + +#ifdef CONFIG_PAX_SEGMEXEC + if (mm->pax_flags & MF_PAX_SEGMEXEC) + task_size = SEGMEXEC_TASK_SIZE; +#endif + + if (len > task_size) + return -ENOMEM; + +#ifdef CONFIG_PAX_RANDMMAP + if (!(mm->pax_flags & MF_PAX_RANDMMAP) || !filp) +#endif + + if (addr) { + addr = PAGE_ALIGN(addr); + vma = find_vma(mm, addr); + if (task_size - len >= addr && + (!vma || addr + len <= vma->vm_start)) + return addr; + } + if (len > mm->cached_hole_size) { + start_addr = addr = mm->free_area_cache; + } else { + start_addr = addr = mm->mmap_base; + mm->cached_hole_size = 0; + } + +#ifdef CONFIG_PAX_PAGEEXEC + if ((mm->pax_flags & MF_PAX_PAGEEXEC) && (flags & MAP_EXECUTABLE) && start_addr >= mm->mmap_base) { + start_addr = 0x00110000UL; + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + start_addr += mm->delta_mmap & 0x03FFF000UL; +#endif + + if (mm->start_brk <= start_addr && start_addr < mm->mmap_base) + start_addr = addr = mm->mmap_base; + else + addr = start_addr; + } +#endif + +full_search: + for (vma = find_vma(mm, addr); ; vma = vma->vm_next) { + /* At this point: (!vma || addr < vma->vm_end). */ + if (task_size - len < addr) { + /* + * Start a new search - just in case we missed + * some holes. + */ + if (start_addr != mm->mmap_base) { + start_addr = addr = mm->mmap_base; + mm->cached_hole_size = 0; + goto full_search; + } + return -ENOMEM; + } + if (!vma || addr + len <= vma->vm_start) { + /* + * Remember the place where we stopped the search: + */ + mm->free_area_cache = addr + len; + return addr; + } + if (addr + mm->cached_hole_size < vma->vm_start) + mm->cached_hole_size = vma->vm_start - addr; + addr = vma->vm_end; + if (mm->start_brk <= addr && addr < mm->mmap_base) { + start_addr = addr = mm->mmap_base; + goto full_search; + } + } +} + +unsigned long +arch_get_unmapped_area_topdown(struct file *filp, const unsigned long addr0, + const unsigned long len, const unsigned long pgoff, + const unsigned long flags) +{ + struct vm_area_struct *vma; + struct mm_struct *mm = current->mm; + unsigned long base = mm->mmap_base, addr = addr0, task_size = TASK_SIZE; + +#ifdef CONFIG_PAX_SEGMEXEC + if (mm->pax_flags & MF_PAX_SEGMEXEC) + task_size = SEGMEXEC_TASK_SIZE; +#endif + + /* requested length too big for entire address space */ + if (len > task_size) + return -ENOMEM; + +#ifdef CONFIG_PAX_PAGEEXEC + if ((mm->pax_flags & MF_PAX_PAGEEXEC) && (flags & MAP_EXECUTABLE)) + goto bottomup; +#endif + +#ifdef CONFIG_PAX_RANDMMAP + if (!(mm->pax_flags & MF_PAX_RANDMMAP) || !filp) +#endif + + /* requesting a specific address */ + if (addr) { + addr = PAGE_ALIGN(addr); + vma = find_vma(mm, addr); + if (task_size - len >= addr && + (!vma || addr + len <= vma->vm_start)) + return addr; + } + + /* check if free_area_cache is useful for us */ + if (len <= mm->cached_hole_size) { + mm->cached_hole_size = 0; + mm->free_area_cache = mm->mmap_base; + } + + /* either no address requested or can't fit in requested address hole */ + addr = mm->free_area_cache; + + /* make sure it can fit in the remaining address space */ + if (addr > len) { + vma = find_vma(mm, addr-len); + if (!vma || addr <= vma->vm_start) + /* remember the address as a hint for next time */ + return (mm->free_area_cache = addr-len); + } + + if (mm->mmap_base < len) + goto bottomup; + + addr = mm->mmap_base-len; + + do { + /* + * Lookup failure means no vma is above this address, + * else if new region fits below vma->vm_start, + * return with success: + */ + vma = find_vma(mm, addr); + if (!vma || addr+len <= vma->vm_start) + /* remember the address as a hint for next time */ + return (mm->free_area_cache = addr); + + /* remember the largest hole we saw so far */ + if (addr + mm->cached_hole_size < vma->vm_start) + mm->cached_hole_size = vma->vm_start - addr; + + /* try just below the current vma->vm_start */ + addr = vma->vm_start-len; + } while (len < vma->vm_start); + +bottomup: + /* + * A failed mmap() very likely causes application failure, + * so fall back to the bottom-up function here. This scenario + * can happen with large stack limits and large mmap() + * allocations. + */ + mm->mmap_base = TASK_UNMAPPED_BASE; + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base += mm->delta_mmap; +#endif + + mm->free_area_cache = mm->mmap_base; + mm->cached_hole_size = ~0UL; + addr = arch_get_unmapped_area(filp, addr0, len, pgoff, flags); + /* + * Restore the topdown base: + */ + mm->mmap_base = base; + mm->free_area_cache = base; + mm->cached_hole_size = ~0UL; + + return addr; +} struct sel_arg_struct { unsigned long n; @@ -210,7 +396,7 @@ asmlinkage int sys_uname(struct old_utsn if (!name) return -EFAULT; down_read(&uts_sem); - err=copy_to_user(name, &system_utsname, sizeof (*name)); + err=copy_to_user(name, vx_new_utsname(), sizeof (*name)); up_read(&uts_sem); return err?-EFAULT:0; } @@ -218,6 +404,7 @@ asmlinkage int sys_uname(struct old_utsn asmlinkage int sys_olduname(struct oldold_utsname __user * name) { int error; + struct new_utsname *ptr; if (!name) return -EFAULT; @@ -226,15 +413,16 @@ asmlinkage int sys_olduname(struct oldol down_read(&uts_sem); - error = __copy_to_user(&name->sysname,&system_utsname.sysname,__OLD_UTS_LEN); + ptr = vx_new_utsname(); + error = __copy_to_user(&name->sysname,ptr->sysname,__OLD_UTS_LEN); error |= __put_user(0,name->sysname+__OLD_UTS_LEN); - error |= __copy_to_user(&name->nodename,&system_utsname.nodename,__OLD_UTS_LEN); + error |= __copy_to_user(&name->nodename,ptr->nodename,__OLD_UTS_LEN); error |= __put_user(0,name->nodename+__OLD_UTS_LEN); - error |= __copy_to_user(&name->release,&system_utsname.release,__OLD_UTS_LEN); + error |= __copy_to_user(&name->release,ptr->release,__OLD_UTS_LEN); error |= __put_user(0,name->release+__OLD_UTS_LEN); - error |= __copy_to_user(&name->version,&system_utsname.version,__OLD_UTS_LEN); + error |= __copy_to_user(&name->version,ptr->version,__OLD_UTS_LEN); error |= __put_user(0,name->version+__OLD_UTS_LEN); - error |= __copy_to_user(&name->machine,&system_utsname.machine,__OLD_UTS_LEN); + error |= __copy_to_user(&name->machine,ptr->machine,__OLD_UTS_LEN); error |= __put_user(0,name->machine+__OLD_UTS_LEN); up_read(&uts_sem); diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/syscall_table.S linux-2.6.17.13-g-v/arch/i386/kernel/syscall_table.S --- linux-2.6.17.13/arch/i386/kernel/syscall_table.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/syscall_table.S 2006-09-20 15:00:12.000000000 +0200 @@ -1,3 +1,4 @@ +.section .rodata,"a",@progbits ENTRY(sys_call_table) .long sys_restart_syscall /* 0 - old "setup()" system call, used for restarting */ .long sys_exit @@ -272,7 +273,7 @@ ENTRY(sys_call_table) .long sys_tgkill /* 270 */ .long sys_utimes .long sys_fadvise64_64 - .long sys_ni_syscall /* sys_vserver */ + .long sys_vserver .long sys_mbind .long sys_get_mempolicy .long sys_set_mempolicy diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/sysenter.c linux-2.6.17.13-g-v/arch/i386/kernel/sysenter.c --- linux-2.6.17.13/arch/i386/kernel/sysenter.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/sysenter.c 2006-09-20 15:00:12.000000000 +0200 @@ -24,7 +24,7 @@ extern asmlinkage void sysenter_entry(vo void enable_sep_cpu(void) { int cpu = get_cpu(); - struct tss_struct *tss = &per_cpu(init_tss, cpu); + struct tss_struct *tss = init_tss + cpu; if (!boot_cpu_has(X86_FEATURE_SEP)) { put_cpu(); @@ -48,6 +48,7 @@ extern const char vsyscall_sysenter_star int __init sysenter_setup(void) { +#ifndef CONFIG_PAX_NOVSYSCALL void *page = (void *)get_zeroed_page(GFP_ATOMIC); __set_fixmap(FIX_VSYSCALL, __pa(page), PAGE_READONLY_EXEC); @@ -62,6 +63,7 @@ int __init sysenter_setup(void) memcpy(page, &vsyscall_sysenter_start, &vsyscall_sysenter_end - &vsyscall_sysenter_start); +#endif return 0; } diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/traps.c linux-2.6.17.13-g-v/arch/i386/kernel/traps.c --- linux-2.6.17.13/arch/i386/kernel/traps.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/traps.c 2006-09-20 15:00:12.000000000 +0200 @@ -28,6 +28,7 @@ #include #include #include +#include #ifdef CONFIG_EISA #include @@ -53,23 +54,19 @@ #include #include +#include #include "mach_traps.h" asmlinkage int system_call(void); -struct desc_struct default_ldt[] = { { 0, 0 }, { 0, 0 }, { 0, 0 }, +const struct desc_struct default_ldt[] = { { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0, 0 } }; /* Do we ignore FPU interrupts ? */ char ignore_fpu_irq = 0; -/* - * The IDT has to be page-aligned to simplify the Pentium - * F0 0F bug workaround.. We have a special link segment - * for this. - */ -struct desc_struct idt_table[256] __attribute__((__section__(".data.idt"))) = { {0, 0}, }; +extern struct desc_struct idt_table[256]; asmlinkage void divide_error(void); asmlinkage void debug(void); @@ -144,18 +141,22 @@ static inline unsigned long print_contex { unsigned long addr; int printed = 0; /* nr of entries already printed on current line */ + int i = kstack_depth_to_print; #ifdef CONFIG_FRAME_POINTER while (valid_stack_ptr(tinfo, (void *)ebp)) { addr = *(unsigned long *)(ebp + 4); printed = print_addr_and_symbol(addr, log_lvl, printed); ebp = *(unsigned long *)ebp; + --i; } #else while (valid_stack_ptr(tinfo, stack)) { addr = *stack++; - if (__kernel_text_address(addr)) + if (__kernel_text_address(addr + __KERNEL_TEXT_OFFSET)) { printed = print_addr_and_symbol(addr, log_lvl, printed); + --i; + } } #endif if (printed) @@ -249,7 +250,7 @@ void show_registers(struct pt_regs *regs esp = (unsigned long) (&regs->esp); savesegment(ss, ss); - if (user_mode_vm(regs)) { + if (user_mode(regs)) { in_kernel = 0; esp = regs->esp; ss = regs->xss & 0xffff; @@ -268,21 +269,24 @@ void show_registers(struct pt_regs *regs regs->esi, regs->edi, regs->ebp, esp); printk(KERN_EMERG "ds: %04x es: %04x ss: %04x\n", regs->xds & 0xffff, regs->xes & 0xffff, ss); - printk(KERN_EMERG "Process %s (pid: %d, threadinfo=%p task=%p)", - current->comm, current->pid, current_thread_info(), current); + printk(KERN_EMERG "Process %s (pid: %d[#%u], threadinfo=%p task=%p)", + current->comm, current->pid, current->xid, + current_thread_info(), current); /* * When in-kernel, we also print out the stack and code at the * time of the fault.. */ if (in_kernel) { u8 __user *eip; + mm_segment_t old_fs = get_fs(); printk("\n" KERN_EMERG "Stack: "); show_stack_log_lvl(NULL, (unsigned long *)esp, KERN_EMERG); printk(KERN_EMERG "Code: "); - eip = (u8 __user *)regs->eip - 43; + set_fs(KERNEL_DS); + eip = (u8 __user *)regs->eip - 43 + __KERNEL_TEXT_OFFSET; for (i = 0; i < 64; i++, eip++) { unsigned char c; @@ -290,11 +294,12 @@ void show_registers(struct pt_regs *regs printk(" Bad EIP value."); break; } - if (eip == (u8 __user *)regs->eip) + if (eip == (u8 __user *)regs->eip + __KERNEL_TEXT_OFFSET) printk("<%02x> ", c); else printk("%02x ", c); } + set_fs(old_fs); } printk("\n"); } @@ -306,29 +311,34 @@ static void handle_BUG(struct pt_regs *r char *file; char c; unsigned long eip; + mm_segment_t old_fs = get_fs(); - eip = regs->eip; + eip = regs->eip + __KERNEL_TEXT_OFFSET; + set_fs(KERNEL_DS); if (eip < PAGE_OFFSET) goto no_bug; if (__get_user(ud2, (unsigned short __user *)eip)) goto no_bug; if (ud2 != 0x0b0f) goto no_bug; - if (__get_user(line, (unsigned short __user *)(eip + 2))) + if (__get_user(line, (unsigned short __user *)(eip + 7))) goto bug; - if (__get_user(file, (char * __user *)(eip + 4)) || - (unsigned long)file < PAGE_OFFSET || __get_user(c, file)) + if (__get_user(file, (char * __user *)(eip + 3)) || file < _text + __KERNEL_TEXT_OFFSET) + goto bug; + if (__get_user(c, file)) file = ""; printk(KERN_EMERG "------------[ cut here ]------------\n"); printk(KERN_EMERG "kernel BUG at %s:%d!\n", file, line); no_bug: + set_fs(old_fs); return; /* Here we know it was a BUG but file-n-line is unavailable */ bug: + set_fs(old_fs); printk(KERN_EMERG "Kernel BUG\n"); } @@ -351,6 +361,8 @@ void die(const char * str, struct pt_reg oops_enter(); + vxh_throw_oops(); + if (die.lock_owner != raw_smp_processor_id()) { console_verbose(); spin_lock_irqsave(&die.lock, flags); @@ -387,9 +399,9 @@ void die(const char * str, struct pt_reg if (nl) printk("\n"); if (notify_die(DIE_OOPS, str, regs, err, - current->thread.trap_no, SIGSEGV) != - NOTIFY_STOP) { + current->thread.trap_no, SIGSEGV) != NOTIFY_STOP) { show_registers(regs); + vxh_dump_history(); /* Executive summary in case the oops scrolled away */ esp = (unsigned long) (&regs->esp); savesegment(ss, ss); @@ -430,7 +442,7 @@ void die(const char * str, struct pt_reg static inline void die_if_kernel(const char * str, struct pt_regs * regs, long err) { - if (!user_mode_vm(regs)) + if (!user_mode(regs)) die(str, regs, err); } @@ -448,7 +460,7 @@ static void __kprobes do_trap(int trapnr goto trap_signal; } - if (!user_mode(regs)) + if (!user_mode_novm(regs)) goto kernel_trap; trap_signal: { @@ -536,7 +548,7 @@ fastcall void __kprobes do_general_prote long error_code) { int cpu = get_cpu(); - struct tss_struct *tss = &per_cpu(init_tss, cpu); + struct tss_struct *tss = &init_tss[cpu]; struct thread_struct *thread = &current->thread; /* @@ -572,9 +584,25 @@ fastcall void __kprobes do_general_prote if (regs->eflags & VM_MASK) goto gp_in_vm86; - if (!user_mode(regs)) + if (!user_mode_novm(regs)) goto gp_in_kernel; +#ifdef CONFIG_PAX_PAGEEXEC + if (current->mm && (current->mm->pax_flags & MF_PAX_PAGEEXEC)) { + struct mm_struct *mm = current->mm; + unsigned long limit; + + down_write(&mm->mmap_sem); + limit = mm->context.user_cs_limit; + if (limit < TASK_SIZE) { + track_exec_limit(mm, limit, TASK_SIZE, PROT_EXEC); + up_write(&mm->mmap_sem); + return; + } + up_write(&mm->mmap_sem); + } +#endif + current->thread.error_code = error_code; current->thread.trap_no = 13; force_sig(SIGSEGV, current); @@ -590,6 +618,13 @@ gp_in_kernel: if (notify_die(DIE_GPF, "general protection fault", regs, error_code, 13, SIGSEGV) == NOTIFY_STOP) return; + +#ifdef CONFIG_PAX_KERNEXEC + if ((regs->xcs & 0xFFFF) == __KERNEL_CS) + die("PAX: suspicious general protection fault", regs, error_code); + else +#endif + die("general protection fault", regs, error_code); } } @@ -663,7 +698,7 @@ void die_nmi (struct pt_regs *regs, cons /* If we are in kernel we are probably nested up pretty bad * and might aswell get out now while we still can. */ - if (!user_mode_vm(regs)) { + if (!user_mode(regs)) { current->thread.trap_no = 2; crash_kexec(regs); } @@ -816,7 +851,7 @@ fastcall void __kprobes do_debug(struct * check for kernel mode by just checking the CPL * of CS. */ - if (!user_mode(regs)) + if (!user_mode_novm(regs)) goto clear_TF_reenable; } @@ -1106,7 +1141,19 @@ do { \ */ void set_intr_gate(unsigned int n, void *addr) { + +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + _set_gate(idt_table+n,14,0,addr,__KERNEL_CS); + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + } /* diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/vm86.c linux-2.6.17.13-g-v/arch/i386/kernel/vm86.c --- linux-2.6.17.13/arch/i386/kernel/vm86.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/vm86.c 2006-09-20 15:00:12.000000000 +0200 @@ -123,7 +123,7 @@ struct pt_regs * fastcall save_v86_state do_exit(SIGSEGV); } - tss = &per_cpu(init_tss, get_cpu()); + tss = init_tss + get_cpu(); current->thread.esp0 = current->thread.saved_esp0; current->thread.sysenter_cs = __KERNEL_CS; load_esp0(tss, &current->thread); @@ -297,7 +297,7 @@ static void do_sys_vm86(struct kernel_vm savesegment(fs, tsk->thread.saved_fs); savesegment(gs, tsk->thread.saved_gs); - tss = &per_cpu(init_tss, get_cpu()); + tss = init_tss + get_cpu(); tsk->thread.esp0 = (unsigned long) &info->VM86_TSS_ESP0; if (cpu_has_sep) tsk->thread.sysenter_cs = 0; diff -NurpP --minimal linux-2.6.17.13/arch/i386/kernel/vmlinux.lds.S linux-2.6.17.13-g-v/arch/i386/kernel/vmlinux.lds.S --- linux-2.6.17.13/arch/i386/kernel/vmlinux.lds.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/kernel/vmlinux.lds.S 2006-09-20 15:00:12.000000000 +0200 @@ -4,10 +4,19 @@ #define LOAD_OFFSET __PAGE_OFFSET +#include + #include #include #include #include +#include + +#ifdef CONFIG_X86_PAE +#define PMD_SHIFT 21 +#else +#define PMD_SHIFT 22 +#endif OUTPUT_FORMAT("elf32-i386", "elf32-i386", "elf32-i386") OUTPUT_ARCH(i386) @@ -16,59 +25,14 @@ jiffies = jiffies_64; SECTIONS { . = __KERNEL_START; - phys_startup_32 = startup_32 - LOAD_OFFSET; - /* read-only */ - _text = .; /* Text and read-only data */ - .text : AT(ADDR(.text) - LOAD_OFFSET) { - *(.text) - SCHED_TEXT - LOCK_TEXT - KPROBES_TEXT - *(.fixup) - *(.gnu.warning) - } = 0x9090 - - _etext = .; /* End of text section */ - - . = ALIGN(16); /* Exception table */ - __start___ex_table = .; - __ex_table : AT(ADDR(__ex_table) - LOAD_OFFSET) { *(__ex_table) } - __stop___ex_table = .; - - RODATA + phys_startup_32 = startup_32 - LOAD_OFFSET + __KERNEL_TEXT_OFFSET; - /* writeable */ - .data : AT(ADDR(.data) - LOAD_OFFSET) { /* Data */ - *(.data) - CONSTRUCTORS + .text.startup : AT(ADDR(.text.startup) - LOAD_OFFSET) { + BYTE(0xEA) /* jmp far */ + LONG(phys_startup_32) + SHORT(__BOOT_CS) } - . = ALIGN(4096); - __nosave_begin = .; - .data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) { *(.data.nosave) } - . = ALIGN(4096); - __nosave_end = .; - - . = ALIGN(4096); - .data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) { - *(.data.idt) - } - - . = ALIGN(32); - .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) { - *(.data.cacheline_aligned) - } - - /* rarely changed data like cpu maps */ - . = ALIGN(32); - .data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) { *(.data.read_mostly) } - _edata = .; /* End of data section */ - - . = ALIGN(THREAD_SIZE); /* init_task */ - .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) { - *(.data.init_task) - } - /* might get freed after init */ . = ALIGN(4096); __smp_alt_begin = .; @@ -92,11 +56,6 @@ SECTIONS /* will be freed after init */ . = ALIGN(4096); /* Init code and data */ __init_begin = .; - .init.text : AT(ADDR(.init.text) - LOAD_OFFSET) { - _sinittext = .; - *(.init.text) - _einittext = .; - } .init.data : AT(ADDR(.init.data) - LOAD_OFFSET) { *(.init.data) } . = ALIGN(16); __setup_start = .; @@ -128,9 +87,7 @@ SECTIONS .altinstr_replacement : AT(ADDR(.altinstr_replacement) - LOAD_OFFSET) { *(.altinstr_replacement) } - /* .exit.text is discard at runtime, not link time, to deal with references - from .altinstructions and .eh_frame */ - .exit.text : AT(ADDR(.exit.text) - LOAD_OFFSET) { *(.exit.text) } + .exit.data : AT(ADDR(.exit.data) - LOAD_OFFSET) { *(.exit.data) } . = ALIGN(4096); __initramfs_start = .; @@ -140,10 +97,108 @@ SECTIONS __per_cpu_start = .; .data.percpu : AT(ADDR(.data.percpu) - LOAD_OFFSET) { *(.data.percpu) } __per_cpu_end = .; + + /* read-only */ + . = ALIGN(4096); - __init_end = .; + .init.text (. - __KERNEL_TEXT_OFFSET) : AT(ADDR(.init.text) - LOAD_OFFSET + __KERNEL_TEXT_OFFSET) { + _sinittext = .; + *(.init.text) + _einittext = .; + } + + /* .exit.text is discard at runtime, not link time, to deal with references + from .altinstructions and .eh_frame */ + .exit.text : AT(ADDR(.exit.text) - LOAD_OFFSET + __KERNEL_TEXT_OFFSET) { *(.exit.text) } + +#ifdef CONFIG_PAX_KERNEXEC + .text.align : AT(ADDR(.text.align) - LOAD_OFFSET + __KERNEL_TEXT_OFFSET) { + . = ALIGN(__KERNEL_TEXT_OFFSET - LOAD_OFFSET) - 1; + BYTE(0) + } +#else + . = ALIGN(4096); +#endif + + __init_end = . + __KERNEL_TEXT_OFFSET; /* freed after init ends here */ - + + _text = .; /* Text and read-only data */ + .text : AT(ADDR(.text) - LOAD_OFFSET + __KERNEL_TEXT_OFFSET) { + *(.text) + SCHED_TEXT + LOCK_TEXT + KPROBES_TEXT + *(.fixup) + *(.gnu.warning) + } = 0x9090 + + _etext = .; /* End of text section */ + . += __KERNEL_TEXT_OFFSET; + . = ALIGN(16); /* Exception table */ + __start___ex_table = .; + __ex_table : AT(ADDR(__ex_table) - LOAD_OFFSET) { *(__ex_table) } + __stop___ex_table = .; + + . = ALIGN(4096); + .rodata.page_aligned : AT(ADDR(.rodata.page_aligned) - LOAD_OFFSET) { + *(.empty_zero_page) + +#ifdef CONFIG_X86_PAE + *(.swapper_pm_dir) +#endif + + *(.swapper_pg_dir) + *(.idt) + } + + RODATA + +#ifdef CONFIG_PAX_KERNEXEC + . = ALIGN(4096); + MODULES_VADDR = .; + + .module.text : AT(ADDR(.module.text) - LOAD_OFFSET) { + . += (4 * 1024 * 1024); + . = ALIGN(1 << PMD_SHIFT) - 1; + BYTE(0) + } + + MODULES_END = .; +#else + . = ALIGN(32); +#endif + + /* writeable */ + .data : AT(ADDR(.data) - LOAD_OFFSET) { /* Data */ + _data = .; + *(.data) + CONSTRUCTORS + } + + . = ALIGN(4096); + __nosave_begin = .; + .data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) { *(.data.nosave) } + . = ALIGN(4096); + __nosave_end = .; + + . = ALIGN(32); + .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) { + *(.data.cacheline_aligned) + } + + /* rarely changed data like cpu maps */ + . = ALIGN(32); + .data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) { *(.data.read_mostly) } + + . = ALIGN(THREAD_SIZE); /* init_task */ + .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) { + *(.data.init_task) + } + + _edata = .; /* End of data section */ + + . = ALIGN(4096); __bss_start = .; /* BSS */ .bss.page_aligned : AT(ADDR(.bss.page_aligned) - LOAD_OFFSET) { *(.bss.page_aligned) diff -NurpP --minimal linux-2.6.17.13/arch/i386/lib/checksum.S linux-2.6.17.13-g-v/arch/i386/lib/checksum.S --- linux-2.6.17.13/arch/i386/lib/checksum.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/lib/checksum.S 2006-09-20 15:00:12.000000000 +0200 @@ -27,7 +27,8 @@ #include #include - +#include + /* * computes a partial checksum, e.g. for TCP/UDP fragments */ @@ -281,12 +282,23 @@ unsigned int csum_partial_copy_generic ( .align 4 .globl csum_partial_copy_generic - +.globl csum_partial_copy_generic_to_user +.globl csum_partial_copy_generic_from_user + #ifndef CONFIG_X86_USE_PPRO_CHECKSUM #define ARGBASE 16 #define FP 12 - + +csum_partial_copy_generic_to_user: + pushl $(__USER_DS) + popl %es + jmp csum_partial_copy_generic + +csum_partial_copy_generic_from_user: + pushl $(__USER_DS) + popl %ds + csum_partial_copy_generic: subl $4,%esp pushl %edi @@ -305,7 +317,7 @@ csum_partial_copy_generic: jmp 4f SRC(1: movw (%esi), %bx ) addl $2, %esi -DST( movw %bx, (%edi) ) +DST( movw %bx, %es:(%edi) ) addl $2, %edi addw %bx, %ax adcl $0, %eax @@ -317,30 +329,30 @@ DST( movw %bx, (%edi) ) SRC(1: movl (%esi), %ebx ) SRC( movl 4(%esi), %edx ) adcl %ebx, %eax -DST( movl %ebx, (%edi) ) +DST( movl %ebx, %es:(%edi) ) adcl %edx, %eax -DST( movl %edx, 4(%edi) ) +DST( movl %edx, %es:4(%edi) ) SRC( movl 8(%esi), %ebx ) SRC( movl 12(%esi), %edx ) adcl %ebx, %eax -DST( movl %ebx, 8(%edi) ) +DST( movl %ebx, %es:8(%edi) ) adcl %edx, %eax -DST( movl %edx, 12(%edi) ) +DST( movl %edx, %es:12(%edi) ) SRC( movl 16(%esi), %ebx ) SRC( movl 20(%esi), %edx ) adcl %ebx, %eax -DST( movl %ebx, 16(%edi) ) +DST( movl %ebx, %es:16(%edi) ) adcl %edx, %eax -DST( movl %edx, 20(%edi) ) +DST( movl %edx, %es:20(%edi) ) SRC( movl 24(%esi), %ebx ) SRC( movl 28(%esi), %edx ) adcl %ebx, %eax -DST( movl %ebx, 24(%edi) ) +DST( movl %ebx, %es:24(%edi) ) adcl %edx, %eax -DST( movl %edx, 28(%edi) ) +DST( movl %edx, %es:28(%edi) ) lea 32(%esi), %esi lea 32(%edi), %edi @@ -354,7 +366,7 @@ DST( movl %edx, 28(%edi) ) shrl $2, %edx # This clears CF SRC(3: movl (%esi), %ebx ) adcl %ebx, %eax -DST( movl %ebx, (%edi) ) +DST( movl %ebx, %es:(%edi) ) lea 4(%esi), %esi lea 4(%edi), %edi dec %edx @@ -366,12 +378,12 @@ DST( movl %ebx, (%edi) ) jb 5f SRC( movw (%esi), %cx ) leal 2(%esi), %esi -DST( movw %cx, (%edi) ) +DST( movw %cx, %es:(%edi) ) leal 2(%edi), %edi je 6f shll $16,%ecx SRC(5: movb (%esi), %cl ) -DST( movb %cl, (%edi) ) +DST( movb %cl, %es:(%edi) ) 6: addl %ecx, %eax adcl $0, %eax 7: @@ -382,7 +394,7 @@ DST( movb %cl, (%edi) ) 6001: movl ARGBASE+20(%esp), %ebx # src_err_ptr - movl $-EFAULT, (%ebx) + movl $-EFAULT, %ss:(%ebx) # zero the complete destination - computing the rest # is too much work @@ -395,11 +407,15 @@ DST( movb %cl, (%edi) ) 6002: movl ARGBASE+24(%esp), %ebx # dst_err_ptr - movl $-EFAULT,(%ebx) + movl $-EFAULT,%ss:(%ebx) jmp 5000b .previous + pushl %ss + popl %ds + pushl %ss + popl %es popl %ebx popl %esi popl %edi @@ -411,17 +427,28 @@ DST( movb %cl, (%edi) ) /* Version for PentiumII/PPro */ #define ROUND1(x) \ + nop; nop; nop; \ SRC(movl x(%esi), %ebx ) ; \ addl %ebx, %eax ; \ - DST(movl %ebx, x(%edi) ) ; + DST(movl %ebx, %es:x(%edi)); #define ROUND(x) \ + nop; nop; nop; \ SRC(movl x(%esi), %ebx ) ; \ adcl %ebx, %eax ; \ - DST(movl %ebx, x(%edi) ) ; + DST(movl %ebx, %es:x(%edi)); #define ARGBASE 12 - + +csum_partial_copy_generic_to_user: + pushl $(__USER_DS) + popl %es + jmp csum_partial_copy_generic + +csum_partial_copy_generic_from_user: + pushl $(__USER_DS) + popl %ds + csum_partial_copy_generic: pushl %ebx pushl %edi @@ -440,7 +467,7 @@ csum_partial_copy_generic: subl %ebx, %edi lea -1(%esi),%edx andl $-32,%edx - lea 3f(%ebx,%ebx), %ebx + lea 3f(%ebx,%ebx,2), %ebx testl %esi, %esi jmp *%ebx 1: addl $64,%esi @@ -461,19 +488,19 @@ csum_partial_copy_generic: jb 5f SRC( movw (%esi), %dx ) leal 2(%esi), %esi -DST( movw %dx, (%edi) ) +DST( movw %dx, %es:(%edi) ) leal 2(%edi), %edi je 6f shll $16,%edx 5: SRC( movb (%esi), %dl ) -DST( movb %dl, (%edi) ) +DST( movb %dl, %es:(%edi) ) 6: addl %edx, %eax adcl $0, %eax 7: .section .fixup, "ax" 6001: movl ARGBASE+20(%esp), %ebx # src_err_ptr - movl $-EFAULT, (%ebx) + movl $-EFAULT, %ss:(%ebx) # zero the complete destination (computing the rest is too much work) movl ARGBASE+8(%esp),%edi # dst movl ARGBASE+12(%esp),%ecx # len @@ -481,10 +508,14 @@ DST( movb %dl, (%edi) ) rep; stosb jmp 7b 6002: movl ARGBASE+24(%esp), %ebx # dst_err_ptr - movl $-EFAULT, (%ebx) + movl $-EFAULT, %ss:(%ebx) jmp 7b .previous + pushl %ss + popl %ds + pushl %ss + popl %es popl %esi popl %edi popl %ebx diff -NurpP --minimal linux-2.6.17.13/arch/i386/lib/getuser.S linux-2.6.17.13-g-v/arch/i386/lib/getuser.S --- linux-2.6.17.13/arch/i386/lib/getuser.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/lib/getuser.S 2006-09-20 15:00:12.000000000 +0200 @@ -9,6 +9,7 @@ * return value. */ #include +#include /* @@ -30,8 +31,12 @@ __get_user_1: GET_THREAD_INFO(%edx) cmpl TI_addr_limit(%edx),%eax jae bad_get_user + pushl $(__USER_DS) + popl %ds 1: movzbl (%eax),%edx xorl %eax,%eax + pushl %ss + pop %ds ret .align 4 @@ -42,7 +47,11 @@ __get_user_2: GET_THREAD_INFO(%edx) cmpl TI_addr_limit(%edx),%eax jae bad_get_user + pushl $(__USER_DS) + popl %ds 2: movzwl -1(%eax),%edx + pushl %ss + pop %ds xorl %eax,%eax ret @@ -54,11 +63,17 @@ __get_user_4: GET_THREAD_INFO(%edx) cmpl TI_addr_limit(%edx),%eax jae bad_get_user + pushl $(__USER_DS) + popl %ds 3: movl -3(%eax),%edx + pushl %ss + pop %ds xorl %eax,%eax ret bad_get_user: + pushl %ss + pop %ds xorl %edx,%edx movl $-14,%eax ret diff -NurpP --minimal linux-2.6.17.13/arch/i386/lib/mmx.c linux-2.6.17.13-g-v/arch/i386/lib/mmx.c --- linux-2.6.17.13/arch/i386/lib/mmx.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/lib/mmx.c 2006-09-20 15:00:12.000000000 +0200 @@ -48,14 +48,30 @@ void *_mmx_memcpy(void *to, const void * " prefetch 256(%0)\n" "2: \n" ".section .fixup, \"ax\"\n" - "3: movw $0x1AEB, 1b\n" /* jmp on 26 bytes */ + "3: \n" + +#ifdef CONFIG_PAX_KERNEXEC + " cli\n" + " movl %%cr0, %%eax\n" + " andl $0xFFFEFFFF, %%eax\n" + " movl %%eax, %%cr0\n" +#endif + + " movw $0x1AEB, 1b\n" /* jmp on 26 bytes */ + +#ifdef CONFIG_PAX_KERNEXEC + " orl $0x00010000, %%eax\n" + " movl %%eax, %%cr0\n" + " sti\n" +#endif + " jmp 2b\n" ".previous\n" ".section __ex_table,\"a\"\n" " .align 4\n" " .long 1b, 3b\n" ".previous" - : : "r" (from) ); + : : "r" (from) : "ax"); for(; i>5; i--) @@ -79,14 +95,30 @@ void *_mmx_memcpy(void *to, const void * " movq %%mm2, 48(%1)\n" " movq %%mm3, 56(%1)\n" ".section .fixup, \"ax\"\n" - "3: movw $0x05EB, 1b\n" /* jmp on 5 bytes */ + "3:\n" + +#ifdef CONFIG_PAX_KERNEXEC + " cli\n" + " movl %%cr0, %%eax\n" + " andl $0xFFFEFFFF, %%eax\n" + " movl %%eax, %%cr0\n" +#endif + + " movw $0x05EB, 1b\n" /* jmp on 5 bytes */ + +#ifdef CONFIG_PAX_KERNEXEC + " orl $0x00010000, %%eax\n" + " movl %%eax, %%cr0\n" + " sti\n" +#endif + " jmp 2b\n" ".previous\n" ".section __ex_table,\"a\"\n" " .align 4\n" " .long 1b, 3b\n" ".previous" - : : "r" (from), "r" (to) : "memory"); + : : "r" (from), "r" (to) : "memory", "ax"); from+=64; to+=64; } @@ -179,14 +211,30 @@ static void fast_copy_page(void *to, voi " prefetch 256(%0)\n" "2: \n" ".section .fixup, \"ax\"\n" - "3: movw $0x1AEB, 1b\n" /* jmp on 26 bytes */ + "3: \n" + +#ifdef CONFIG_PAX_KERNEXEC + " cli\n" + " movl %%cr0, %%eax\n" + " andl $0xFFFEFFFF, %%eax\n" + " movl %%eax, %%cr0\n" +#endif + + " movw $0x1AEB, 1b\n" /* jmp on 26 bytes */ + +#ifdef CONFIG_PAX_KERNEXEC + " orl $0x00010000, %%eax\n" + " movl %%eax, %%cr0\n" + " sti\n" +#endif + " jmp 2b\n" ".previous\n" ".section __ex_table,\"a\"\n" " .align 4\n" " .long 1b, 3b\n" ".previous" - : : "r" (from) ); + : : "r" (from) : "ax"); for(i=0; i<(4096-320)/64; i++) { @@ -209,14 +257,30 @@ static void fast_copy_page(void *to, voi " movq 56(%0), %%mm7\n" " movntq %%mm7, 56(%1)\n" ".section .fixup, \"ax\"\n" - "3: movw $0x05EB, 1b\n" /* jmp on 5 bytes */ + "3:\n" + +#ifdef CONFIG_PAX_KERNEXEC + " cli\n" + " movl %%cr0, %%eax\n" + " andl $0xFFFEFFFF, %%eax\n" + " movl %%eax, %%cr0\n" +#endif + + " movw $0x05EB, 1b\n" /* jmp on 5 bytes */ + +#ifdef CONFIG_PAX_KERNEXEC + " orl $0x00010000, %%eax\n" + " movl %%eax, %%cr0\n" + " sti\n" +#endif + " jmp 2b\n" ".previous\n" ".section __ex_table,\"a\"\n" " .align 4\n" " .long 1b, 3b\n" ".previous" - : : "r" (from), "r" (to) : "memory"); + : : "r" (from), "r" (to) : "memory", "ax"); from+=64; to+=64; } @@ -309,14 +373,30 @@ static void fast_copy_page(void *to, voi " prefetch 256(%0)\n" "2: \n" ".section .fixup, \"ax\"\n" - "3: movw $0x1AEB, 1b\n" /* jmp on 26 bytes */ + "3: \n" + +#ifdef CONFIG_PAX_KERNEXEC + " cli\n" + " movl %%cr0, %%eax\n" + " andl $0xFFFEFFFF, %%eax\n" + " movl %%eax, %%cr0\n" +#endif + + " movw $0x1AEB, 1b\n" /* jmp on 26 bytes */ + +#ifdef CONFIG_PAX_KERNEXEC + " orl $0x00010000, %%eax\n" + " movl %%eax, %%cr0\n" + " sti\n" +#endif + " jmp 2b\n" ".previous\n" ".section __ex_table,\"a\"\n" " .align 4\n" " .long 1b, 3b\n" ".previous" - : : "r" (from) ); + : : "r" (from) : "ax"); for(i=0; i<4096/64; i++) { @@ -339,14 +419,30 @@ static void fast_copy_page(void *to, voi " movq %%mm2, 48(%1)\n" " movq %%mm3, 56(%1)\n" ".section .fixup, \"ax\"\n" - "3: movw $0x05EB, 1b\n" /* jmp on 5 bytes */ + "3:\n" + +#ifdef CONFIG_PAX_KERNEXEC + " cli\n" + " movl %%cr0, %%eax\n" + " andl $0xFFFEFFFF, %%eax\n" + " movl %%eax, %%cr0\n" +#endif + + " movw $0x05EB, 1b\n" /* jmp on 5 bytes */ + +#ifdef CONFIG_PAX_KERNEXEC + " orl $0x00010000, %%eax\n" + " movl %%eax, %%cr0\n" + " sti\n" +#endif + " jmp 2b\n" ".previous\n" ".section __ex_table,\"a\"\n" " .align 4\n" " .long 1b, 3b\n" ".previous" - : : "r" (from), "r" (to) : "memory"); + : : "r" (from), "r" (to) : "memory", "ax"); from+=64; to+=64; } diff -NurpP --minimal linux-2.6.17.13/arch/i386/lib/putuser.S linux-2.6.17.13-g-v/arch/i386/lib/putuser.S --- linux-2.6.17.13/arch/i386/lib/putuser.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/lib/putuser.S 2006-09-20 15:00:12.000000000 +0200 @@ -9,6 +9,7 @@ * return value. */ #include +#include /* @@ -33,7 +34,11 @@ __put_user_1: ENTER cmpl TI_addr_limit(%ebx),%ecx jae bad_put_user + pushl $(__USER_DS) + popl %ds 1: movb %al,(%ecx) + pushl %ss + popl %ds xorl %eax,%eax EXIT @@ -45,7 +50,11 @@ __put_user_2: subl $1,%ebx cmpl %ebx,%ecx jae bad_put_user + pushl $(__USER_DS) + popl %ds 2: movw %ax,(%ecx) + pushl %ss + popl %ds xorl %eax,%eax EXIT @@ -57,7 +66,11 @@ __put_user_4: subl $3,%ebx cmpl %ebx,%ecx jae bad_put_user + pushl $(__USER_DS) + popl %ds 3: movl %eax,(%ecx) + pushl %ss + popl %ds xorl %eax,%eax EXIT @@ -69,12 +82,18 @@ __put_user_8: subl $7,%ebx cmpl %ebx,%ecx jae bad_put_user + pushl $(__USER_DS) + popl %ds 4: movl %eax,(%ecx) 5: movl %edx,4(%ecx) + pushl %ss + popl %ds xorl %eax,%eax EXIT bad_put_user: + pushl %ss + popl %ds movl $-14,%eax EXIT diff -NurpP --minimal linux-2.6.17.13/arch/i386/lib/usercopy.c linux-2.6.17.13-g-v/arch/i386/lib/usercopy.c --- linux-2.6.17.13/arch/i386/lib/usercopy.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/lib/usercopy.c 2006-09-20 15:00:12.000000000 +0200 @@ -33,6 +33,7 @@ do { \ int __d0, __d1, __d2; \ might_sleep(); \ __asm__ __volatile__( \ + " movw %w10,%%ds\n" \ " testl %1,%1\n" \ " jz 2f\n" \ "0: lodsb\n" \ @@ -43,6 +44,8 @@ do { \ " jnz 0b\n" \ "1: subl %1,%0\n" \ "2:\n" \ + " pushl %%ss\n" \ + " popl %%ds\n" \ ".section .fixup,\"ax\"\n" \ "3: movl %5,%0\n" \ " jmp 2b\n" \ @@ -53,7 +56,8 @@ do { \ ".previous" \ : "=d"(res), "=c"(count), "=&a" (__d0), "=&S" (__d1), \ "=&D" (__d2) \ - : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), "4"(dst) \ + : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), "4"(dst),\ + "r"(__USER_DS) \ : "memory"); \ } while (0) @@ -123,10 +127,13 @@ do { \ int __d0; \ might_sleep(); \ __asm__ __volatile__( \ + " movw %w6,%%es\n" \ "0: rep; stosl\n" \ " movl %2,%0\n" \ "1: rep; stosb\n" \ "2:\n" \ + " pushl %%ss\n" \ + " popl %%es\n" \ ".section .fixup,\"ax\"\n" \ "3: lea 0(%2,%0,4),%0\n" \ " jmp 2b\n" \ @@ -137,7 +144,8 @@ do { \ " .long 1b,2b\n" \ ".previous" \ : "=&c"(size), "=&D" (__d0) \ - : "r"(size & 3), "0"(size / 4), "1"(addr), "a"(0)); \ + : "r"(size & 3), "0"(size / 4), "1"(addr), "a"(0), \ + "r"(__USER_DS)); \ } while (0) /** @@ -198,14 +206,17 @@ long strnlen_user(const char __user *s, might_sleep(); __asm__ __volatile__( + " movw %w8,%%es\n" " testl %0, %0\n" " jz 3f\n" - " andl %0,%%ecx\n" + " movl %0,%%ecx\n" "0: repne; scasb\n" " setne %%al\n" " subl %%ecx,%0\n" " addl %0,%%eax\n" "1:\n" + " pushl %%ss\n" + " popl %%es\n" ".section .fixup,\"ax\"\n" "2: xorl %%eax,%%eax\n" " jmp 1b\n" @@ -217,7 +228,7 @@ long strnlen_user(const char __user *s, " .long 0b,2b\n" ".previous" :"=r" (n), "=D" (s), "=a" (res), "=c" (tmp) - :"0" (n), "1" (s), "2" (0), "3" (mask) + :"0" (n), "1" (s), "2" (0), "3" (mask), "r" (__USER_DS) :"cc"); return res & mask; } @@ -229,6 +240,7 @@ __copy_user_intel(void __user *to, const { int d0, d1; __asm__ __volatile__( + " movw %w6, %%es\n" " .align 2,0x90\n" "1: movl 32(%4), %%eax\n" " cmpl $67, %0\n" @@ -237,36 +249,36 @@ __copy_user_intel(void __user *to, const " .align 2,0x90\n" "3: movl 0(%4), %%eax\n" "4: movl 4(%4), %%edx\n" - "5: movl %%eax, 0(%3)\n" - "6: movl %%edx, 4(%3)\n" + "5: movl %%eax, %%es:0(%3)\n" + "6: movl %%edx, %%es:4(%3)\n" "7: movl 8(%4), %%eax\n" "8: movl 12(%4),%%edx\n" - "9: movl %%eax, 8(%3)\n" - "10: movl %%edx, 12(%3)\n" + "9: movl %%eax, %%es:8(%3)\n" + "10: movl %%edx, %%es:12(%3)\n" "11: movl 16(%4), %%eax\n" "12: movl 20(%4), %%edx\n" - "13: movl %%eax, 16(%3)\n" - "14: movl %%edx, 20(%3)\n" + "13: movl %%eax, %%es:16(%3)\n" + "14: movl %%edx, %%es:20(%3)\n" "15: movl 24(%4), %%eax\n" "16: movl 28(%4), %%edx\n" - "17: movl %%eax, 24(%3)\n" - "18: movl %%edx, 28(%3)\n" + "17: movl %%eax, %%es:24(%3)\n" + "18: movl %%edx, %%es:28(%3)\n" "19: movl 32(%4), %%eax\n" "20: movl 36(%4), %%edx\n" - "21: movl %%eax, 32(%3)\n" - "22: movl %%edx, 36(%3)\n" + "21: movl %%eax, %%es:32(%3)\n" + "22: movl %%edx, %%es:36(%3)\n" "23: movl 40(%4), %%eax\n" "24: movl 44(%4), %%edx\n" - "25: movl %%eax, 40(%3)\n" - "26: movl %%edx, 44(%3)\n" + "25: movl %%eax, %%es:40(%3)\n" + "26: movl %%edx, %%es:44(%3)\n" "27: movl 48(%4), %%eax\n" "28: movl 52(%4), %%edx\n" - "29: movl %%eax, 48(%3)\n" - "30: movl %%edx, 52(%3)\n" + "29: movl %%eax, %%es:48(%3)\n" + "30: movl %%edx, %%es:52(%3)\n" "31: movl 56(%4), %%eax\n" "32: movl 60(%4), %%edx\n" - "33: movl %%eax, 56(%3)\n" - "34: movl %%edx, 60(%3)\n" + "33: movl %%eax, %%es:56(%3)\n" + "34: movl %%edx, %%es:60(%3)\n" " addl $-64, %0\n" " addl $64, %4\n" " addl $64, %3\n" @@ -280,6 +292,8 @@ __copy_user_intel(void __user *to, const "36: movl %%eax, %0\n" "37: rep; movsb\n" "100:\n" + " pushl %%ss\n" + " popl %%es\n" ".section .fixup,\"ax\"\n" "101: lea 0(%%eax,%0,4),%0\n" " jmp 100b\n" @@ -326,7 +340,7 @@ __copy_user_intel(void __user *to, const " .long 99b,101b\n" ".previous" : "=&c"(size), "=&D" (d0), "=&S" (d1) - : "1"(to), "2"(from), "0"(size) + : "1"(to), "2"(from), "0"(size), "r"(__USER_DS) : "eax", "edx", "memory"); return size; } @@ -336,6 +350,7 @@ __copy_user_zeroing_intel(void *to, cons { int d0, d1; __asm__ __volatile__( + " movw %w6, %%ds\n" " .align 2,0x90\n" "0: movl 32(%4), %%eax\n" " cmpl $67, %0\n" @@ -344,36 +359,36 @@ __copy_user_zeroing_intel(void *to, cons " .align 2,0x90\n" "2: movl 0(%4), %%eax\n" "21: movl 4(%4), %%edx\n" - " movl %%eax, 0(%3)\n" - " movl %%edx, 4(%3)\n" + " movl %%eax, %%es:0(%3)\n" + " movl %%edx, %%es:4(%3)\n" "3: movl 8(%4), %%eax\n" "31: movl 12(%4),%%edx\n" - " movl %%eax, 8(%3)\n" - " movl %%edx, 12(%3)\n" + " movl %%eax, %%es:8(%3)\n" + " movl %%edx, %%es:12(%3)\n" "4: movl 16(%4), %%eax\n" "41: movl 20(%4), %%edx\n" - " movl %%eax, 16(%3)\n" - " movl %%edx, 20(%3)\n" + " movl %%eax, %%es:16(%3)\n" + " movl %%edx, %%es:20(%3)\n" "10: movl 24(%4), %%eax\n" "51: movl 28(%4), %%edx\n" - " movl %%eax, 24(%3)\n" - " movl %%edx, 28(%3)\n" + " movl %%eax, %%es:24(%3)\n" + " movl %%edx, %%es:28(%3)\n" "11: movl 32(%4), %%eax\n" "61: movl 36(%4), %%edx\n" - " movl %%eax, 32(%3)\n" - " movl %%edx, 36(%3)\n" + " movl %%eax, %%es:32(%3)\n" + " movl %%edx, %%es:36(%3)\n" "12: movl 40(%4), %%eax\n" "71: movl 44(%4), %%edx\n" - " movl %%eax, 40(%3)\n" - " movl %%edx, 44(%3)\n" + " movl %%eax, %%es:40(%3)\n" + " movl %%edx, %%es:44(%3)\n" "13: movl 48(%4), %%eax\n" "81: movl 52(%4), %%edx\n" - " movl %%eax, 48(%3)\n" - " movl %%edx, 52(%3)\n" + " movl %%eax, %%es:48(%3)\n" + " movl %%edx, %%es:52(%3)\n" "14: movl 56(%4), %%eax\n" "91: movl 60(%4), %%edx\n" - " movl %%eax, 56(%3)\n" - " movl %%edx, 60(%3)\n" + " movl %%eax, %%es:56(%3)\n" + " movl %%edx, %%es:60(%3)\n" " addl $-64, %0\n" " addl $64, %4\n" " addl $64, %3\n" @@ -387,6 +402,8 @@ __copy_user_zeroing_intel(void *to, cons " movl %%eax,%0\n" "7: rep; movsb\n" "8:\n" + " pushl %%ss\n" + " popl %%ds\n" ".section .fixup,\"ax\"\n" "9: lea 0(%%eax,%0,4),%0\n" "16: pushl %0\n" @@ -421,7 +438,7 @@ __copy_user_zeroing_intel(void *to, cons " .long 7b,16b\n" ".previous" : "=&c"(size), "=&D" (d0), "=&S" (d1) - : "1"(to), "2"(from), "0"(size) + : "1"(to), "2"(from), "0"(size), "r"(__USER_DS) : "eax", "edx", "memory"); return size; } @@ -441,6 +458,7 @@ __copy_user_intel(void __user *to, const do { \ int __d0, __d1, __d2; \ __asm__ __volatile__( \ + " movw %w8,%%es\n" \ " cmp $7,%0\n" \ " jbe 1f\n" \ " movl %1,%0\n" \ @@ -456,6 +474,8 @@ do { \ " movl %3,%0\n" \ "1: rep; movsb\n" \ "2:\n" \ + " pushl %%ss\n" \ + " popl %%es\n" \ ".section .fixup,\"ax\"\n" \ "5: addl %3,%0\n" \ " jmp 2b\n" \ @@ -469,7 +489,7 @@ do { \ " .long 1b,2b\n" \ ".previous" \ : "=&c"(size), "=&D" (__d0), "=&S" (__d1), "=r"(__d2) \ - : "3"(size), "0"(size), "1"(to), "2"(from) \ + : "3"(size), "0"(size), "1"(to), "2"(from), "r"(__USER_DS)\ : "memory"); \ } while (0) @@ -477,6 +497,7 @@ do { \ do { \ int __d0, __d1, __d2; \ __asm__ __volatile__( \ + " movw %w8,%%ds\n" \ " cmp $7,%0\n" \ " jbe 1f\n" \ " movl %1,%0\n" \ @@ -492,6 +513,8 @@ do { \ " movl %3,%0\n" \ "1: rep; movsb\n" \ "2:\n" \ + " pushl %%ss\n" \ + " popl %%ds\n" \ ".section .fixup,\"ax\"\n" \ "5: addl %3,%0\n" \ " jmp 6f\n" \ @@ -511,7 +534,7 @@ do { \ " .long 1b,6b\n" \ ".previous" \ : "=&c"(size), "=&D" (__d0), "=&S" (__d1), "=r"(__d2) \ - : "3"(size), "0"(size), "1"(to), "2"(from) \ + : "3"(size), "0"(size), "1"(to), "2"(from), "r"(__USER_DS)\ : "memory"); \ } while (0) @@ -640,3 +663,45 @@ copy_from_user(void *to, const void __us return n; } EXPORT_SYMBOL(copy_from_user); + +#ifdef CONFIG_PAX_MEMORY_UDEREF +void __set_fs(mm_segment_t x, int cpu) +{ + unsigned long limit = x.seg; + + current_thread_info()->addr_limit = x; + if (likely(limit)) { + limit -= 1UL; + limit >>= 12; + } + + get_cpu_gdt_table(cpu)[GDT_ENTRY_DEFAULT_USER_DS].a = (limit & 0xFFFFUL); + get_cpu_gdt_table(cpu)[GDT_ENTRY_DEFAULT_USER_DS].b = (limit & 0xF0000UL) | 0xC0F300UL; +} + +void set_fs(mm_segment_t x) +{ + int cpu = get_cpu(); + +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + + __set_fs(x, cpu); + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + + put_cpu_no_resched(); +} +#else +void set_fs(mm_segment_t x) +{ + current_thread_info()->addr_limit = x; +} +#endif + +EXPORT_SYMBOL(set_fs); diff -NurpP --minimal linux-2.6.17.13/arch/i386/mach-voyager/voyager_smp.c linux-2.6.17.13-g-v/arch/i386/mach-voyager/voyager_smp.c --- linux-2.6.17.13/arch/i386/mach-voyager/voyager_smp.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/mach-voyager/voyager_smp.c 2006-09-20 15:00:12.000000000 +0200 @@ -1295,7 +1295,7 @@ smp_local_timer_interrupt(struct pt_regs per_cpu(prof_counter, cpu); } - update_process_times(user_mode_vm(regs)); + update_process_times(user_mode(regs)); } if( ((1< -#undef CONFIG_X86_PAE #include #include #include #include #include -/* - * I'm cheating here. It is known that the two boot PTE pages are - * allocated next to each other. I'm pretending that they're just - * one big array. - */ - -#define BOOT_PTE_PTRS (PTRS_PER_PTE*2) -#define boot_pte_index(address) \ - (((address) >> PAGE_SHIFT) & (BOOT_PTE_PTRS - 1)) - -static inline boot_pte_t* boot_vaddr_to_pte(void *address) -{ - boot_pte_t* boot_pg = (boot_pte_t*)pg0; - return &boot_pg[boot_pte_index((unsigned long)address)]; -} /* * This is only for a caller who is clever enough to page-align * phys_addr and virtual_source, and who also has a preference * about which virtual address from which to steal ptes */ -static void __boot_ioremap(unsigned long phys_addr, unsigned long nrpages, - void* virtual_source) +static void __init __boot_ioremap(unsigned long phys_addr, unsigned int nrpages, + char* virtual_source) { - boot_pte_t* pte; - int i; - char *vaddr = virtual_source; + pgd_t *pgd; + pud_t *pud; + pmd_t *pmd; + pte_t* pte; + unsigned int i; + unsigned long vaddr = (unsigned long)virtual_source; + + pgd = pgd_offset_k(vaddr); + pud = pud_offset(pgd, vaddr); + pmd = pmd_offset(pud, vaddr); + pte = pte_offset_kernel(pmd, vaddr); - pte = boot_vaddr_to_pte(virtual_source); for (i=0; i < nrpages; i++, phys_addr += PAGE_SIZE, pte++) { set_pte(pte, pfn_pte(phys_addr>>PAGE_SHIFT, PAGE_KERNEL)); - __flush_tlb_one(&vaddr[i*PAGE_SIZE]); + __flush_tlb_one(&virtual_source[i*PAGE_SIZE]); } } diff -NurpP --minimal linux-2.6.17.13/arch/i386/mm/extable.c linux-2.6.17.13-g-v/arch/i386/mm/extable.c --- linux-2.6.17.13/arch/i386/mm/extable.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/mm/extable.c 2006-09-20 15:00:12.000000000 +0200 @@ -12,7 +12,7 @@ int fixup_exception(struct pt_regs *regs const struct exception_table_entry *fixup; #ifdef CONFIG_PNPBIOS - if (unlikely((regs->xcs & ~15) == (GDT_ENTRY_PNPBIOS_BASE << 3))) + if (unlikely(!(regs->eflags & VM_MASK) && ((regs->xcs & 0xFFFCU) == PNP_CS32 || (regs->xcs & 0xFFFCU) == PNP_CS16))) { extern u32 pnp_bios_fault_eip, pnp_bios_fault_esp; extern u32 pnp_bios_is_utter_crap; diff -NurpP --minimal linux-2.6.17.13/arch/i386/mm/fault.c linux-2.6.17.13-g-v/arch/i386/mm/fault.c --- linux-2.6.17.13/arch/i386/mm/fault.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/mm/fault.c 2006-09-20 15:00:12.000000000 +0200 @@ -22,6 +22,9 @@ #include #include #include +#include +#include +#include #include #include @@ -82,11 +85,13 @@ static inline unsigned long get_segment_ /* Unlikely, but must come before segment checks. */ if (unlikely((regs->eflags & VM_MASK) != 0)) - return eip + (seg << 4); + return (eip & 0xFFFF) + (seg << 4); /* By far the most common cases. */ - if (likely(seg == __USER_CS || seg == __KERNEL_CS)) + if (likely(seg == __USER_CS)) return eip; + if (likely(seg == __KERNEL_CS)) + return eip + __KERNEL_TEXT_OFFSET; /* Check the segment exists, is within the current LDT/GDT size, that kernel/user (ring 0..3) has the appropriate privilege, @@ -214,6 +219,30 @@ static noinline void force_sig_info_faul fastcall void do_invalid_op(struct pt_regs *, unsigned long); +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) +static int pax_handle_fetch_fault(struct pt_regs *regs); +#endif + +#ifdef CONFIG_PAX_PAGEEXEC +static inline pmd_t * pax_get_pmd(struct mm_struct *mm, unsigned long address) +{ + pgd_t *pgd; + pud_t *pud; + pmd_t *pmd; + + pgd = pgd_offset(mm, address); + if (!pgd_present(*pgd)) + return NULL; + pud = pud_offset(pgd, address); + if (!pud_present(*pud)) + return NULL; + pmd = pmd_offset(pud, address); + if (!pmd_present(*pmd)) + return NULL; + return pmd; +} +#endif + static inline pmd_t *vmalloc_sync_one(pgd_t *pgd, unsigned long address) { unsigned index = pgd_index(address); @@ -295,13 +324,20 @@ fastcall void __kprobes do_page_fault(st struct mm_struct *mm; struct vm_area_struct * vma; unsigned long address; - unsigned long page; int write, si_code; +#ifdef CONFIG_PAX_PAGEEXEC + pmd_t *pmd; + pte_t *pte; + spinlock_t *ptl; + unsigned char pte_mask; +#endif + /* get the address */ address = read_cr2(); tsk = current; + mm = tsk->mm; si_code = SEGV_MAPERR; @@ -340,14 +376,12 @@ fastcall void __kprobes do_page_fault(st if (regs->eflags & (X86_EFLAGS_IF|VM_MASK)) local_irq_enable(); - mm = tsk->mm; - /* * If we're in an interrupt, have no user context or are running in an * atomic region then we must not take the fault.. */ if (in_atomic() || !mm) - goto bad_area_nosemaphore; + goto bad_area_nopax; /* When running in the kernel we expect faults to occur only to * addresses in user space. All other faults represent errors in the @@ -367,10 +401,101 @@ fastcall void __kprobes do_page_fault(st if (!down_read_trylock(&mm->mmap_sem)) { if ((error_code & 4) == 0 && !search_exception_tables(regs->eip)) - goto bad_area_nosemaphore; + goto bad_area_nopax; down_read(&mm->mmap_sem); } +#ifdef CONFIG_PAX_PAGEEXEC + if (unlikely((error_code & 5) != 5 || + (regs->eflags & X86_EFLAGS_VM) || + !(mm->pax_flags & MF_PAX_PAGEEXEC))) + goto not_pax_fault; + + /* PaX: it's our fault, let's handle it if we can */ + + /* PaX: take a look at read faults before acquiring any locks */ + if (unlikely(!(error_code & 2) && (regs->eip == address))) { + /* instruction fetch attempt from a protected page in user mode */ + up_read(&mm->mmap_sem); + switch (pax_handle_fetch_fault(regs)) { + +#ifdef CONFIG_PAX_EMUTRAMP + case 2: + return; +#endif + + } + pax_report_fault(regs, (void*)regs->eip, (void*)regs->esp); + do_exit(SIGKILL); + } + + pmd = pax_get_pmd(mm, address); + if (unlikely(!pmd)) + goto not_pax_fault; + + pte = pte_offset_map_lock(mm, pmd, address, &ptl); + if (unlikely(!(pte_val(*pte) & _PAGE_PRESENT) || pte_user(*pte))) { + pte_unmap_unlock(pte, ptl); + goto not_pax_fault; + } + + if (unlikely((error_code & 2) && !pte_write(*pte))) { + /* write attempt to a protected page in user mode */ + pte_unmap_unlock(pte, ptl); + goto not_pax_fault; + } + +#ifdef CONFIG_SMP + if (likely(address > get_limit(regs->xcs) && cpu_isset(smp_processor_id(), mm->context.cpu_user_cs_mask))) +#else + if (likely(address > get_limit(regs->xcs))) +#endif + { + set_pte(pte, pte_mkread(*pte)); + __flush_tlb_one(address); + pte_unmap_unlock(pte, ptl); + up_read(&mm->mmap_sem); + return; + } + + pte_mask = _PAGE_ACCESSED | _PAGE_USER | ((error_code & 2) << (_PAGE_BIT_DIRTY-1)); + + /* + * PaX: fill DTLB with user rights and retry + */ + __asm__ __volatile__ ( + "movw %w4,%%ds\n" + "orb %2,%%ss:(%1)\n" +#if defined(CONFIG_M586) || defined(CONFIG_M586TSC) +/* + * PaX: let this uncommented 'invlpg' remind us on the behaviour of Intel's + * (and AMD's) TLBs. namely, they do not cache PTEs that would raise *any* + * page fault when examined during a TLB load attempt. this is true not only + * for PTEs holding a non-present entry but also present entries that will + * raise a page fault (such as those set up by PaX, or the copy-on-write + * mechanism). in effect it means that we do *not* need to flush the TLBs + * for our target pages since their PTEs are simply not in the TLBs at all. + + * the best thing in omitting it is that we gain around 15-20% speed in the + * fast path of the page fault handler and can get rid of tracing since we + * can no longer flush unintended entries. + */ + "invlpg (%0)\n" +#endif + "testb $0,(%0)\n" + "xorb %3,%%ss:(%1)\n" + "pushl %%ss\n" + "popl %%ds\n" + : + : "q" (address), "r" (pte), "q" (pte_mask), "i" (_PAGE_USER), "r" (__USER_DS) + : "memory", "cc"); + pte_unmap_unlock(pte, ptl); + up_read(&mm->mmap_sem); + return; + +not_pax_fault: +#endif + vma = find_vma(mm, address); if (!vma) goto bad_area; @@ -456,6 +581,37 @@ bad_area: up_read(&mm->mmap_sem); bad_area_nosemaphore: + +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) + if (mm && (error_code & 4) && !(regs->eflags & X86_EFLAGS_VM)) { + +#ifdef CONFIG_PAX_PAGEEXEC + if ((mm->pax_flags & MF_PAX_PAGEEXEC) && !(error_code & 3) && (regs->eip == address)) { + pax_report_fault(regs, (void*)regs->eip, (void*)regs->esp); + do_exit(SIGKILL); + } +#endif + +#ifdef CONFIG_PAX_SEGMEXEC + if ((mm->pax_flags & MF_PAX_SEGMEXEC) && !(error_code & 3) && (regs->eip + SEGMEXEC_TASK_SIZE == address)) { + + switch (pax_handle_fetch_fault(regs)) { + +#ifdef CONFIG_PAX_EMUTRAMP + case 2: + return; +#endif + + } + pax_report_fault(regs, (void*)regs->eip, (void*)regs->esp); + do_exit(SIGKILL); + } +#endif + + } +#endif + +bad_area_nopax: /* User mode accesses just cause a SIGSEGV */ if (error_code & 4) { /* @@ -523,6 +679,21 @@ no_context: if (address < PAGE_SIZE) printk(KERN_ALERT "BUG: unable to handle kernel NULL " "pointer dereference"); + +#ifdef CONFIG_PAX_KERNEXEC +#ifdef CONFIG_MODULES + else if (init_mm.start_code <= address && address < (unsigned long)MODULES_END) +#else + else if (init_mm.start_code <= address && address < init_mm.end_code) +#endif + if (tsk->signal->curr_ip) + printk(KERN_ERR "PAX: From %u.%u.%u.%u: %s:%d, uid/euid: %u/%u, attempted to modify kernel code", + NIPQUAD(tsk->signal->curr_ip), tsk->comm, tsk->pid, tsk->uid, tsk->euid); + else + printk(KERN_ERR "PAX: %s:%d, uid/euid: %u/%u, attempted to modify kernel code", + tsk->comm, tsk->pid, tsk->uid, tsk->euid); +#endif + else printk(KERN_ALERT "BUG: unable to handle kernel paging" " request"); @@ -530,24 +701,34 @@ no_context: printk(KERN_ALERT " printing eip:\n"); printk("%08lx\n", regs->eip); } - page = read_cr3(); - page = ((unsigned long *) __va(page))[address >> 22]; - if (oops_may_print()) - printk(KERN_ALERT "*pde = %08lx\n", page); - /* - * We must not directly access the pte in the highpte - * case, the page table might be allocated in highmem. - * And lets rather not kmap-atomic the pte, just in case - * it's allocated already. - */ + + if (oops_may_print()) { + unsigned long index = pgd_index(address); + pgd_t *pgd; + pud_t *pud; + pmd_t *pmd; + pte_t *pte; + + pgd = index + (pgd_t *)__va(read_cr3()); + printk(KERN_ALERT "*pgd = %*llx\n", sizeof(*pgd), (unsigned long long)pgd_val(*pgd)); + if (pgd_present(*pgd)) { + pud = pud_offset(pgd, address); + pmd = pmd_offset(pud, address); + printk(KERN_ALERT "*pmd = %*llx\n", sizeof(*pmd), (unsigned long long)pmd_val(*pmd)); + /* + * We must not directly access the pte in the highpte + * case, the page table might be allocated in highmem. + * And lets rather not kmap-atomic the pte, just in case + * it's allocated already. + */ #ifndef CONFIG_HIGHPTE - if ((page & 1) && oops_may_print()) { - page &= PAGE_MASK; - address &= 0x003ff000; - page = ((unsigned long *) __va(page))[address >> PAGE_SHIFT]; - printk(KERN_ALERT "*pte = %08lx\n", page); - } + if (pmd_present(*pmd) && !pmd_large(*pmd)) { + pte = pte_offset_kernel(pmd, address); + printk(KERN_ALERT "*pte = %*llx\n", sizeof(*pte), (unsigned long long)pte_val(*pte)); + } #endif + } + } tsk->thread.cr2 = address; tsk->thread.trap_no = 14; tsk->thread.error_code = error_code; @@ -624,3 +805,105 @@ void vmalloc_sync_all(void) } } #endif + +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) +/* + * PaX: decide what to do with offenders (regs->eip = fault address) + * + * returns 1 when task should be killed + * 2 when gcc trampoline was detected + */ +static int pax_handle_fetch_fault(struct pt_regs *regs) +{ + +#ifdef CONFIG_PAX_EMUTRAMP + static const unsigned char trans[8] = {6, 1, 2, 0, 13, 5, 3, 4}; + int err; +#endif + + if (regs->eflags & X86_EFLAGS_VM) + return 1; + +#ifdef CONFIG_PAX_EMUTRAMP + if (!(current->mm->pax_flags & MF_PAX_EMUTRAMP)) + return 1; + + do { /* PaX: gcc trampoline emulation #1 */ + unsigned char mov1, mov2; + unsigned short jmp; + unsigned long addr1, addr2; + + err = get_user(mov1, (unsigned char __user *)regs->eip); + err |= get_user(addr1, (unsigned long __user *)(regs->eip + 1)); + err |= get_user(mov2, (unsigned char __user *)(regs->eip + 5)); + err |= get_user(addr2, (unsigned long __user *)(regs->eip + 6)); + err |= get_user(jmp, (unsigned short __user *)(regs->eip + 10)); + + if (err) + break; + + if ((mov1 & 0xF8) == 0xB8 && + (mov2 & 0xF8) == 0xB8 && + (mov1 & 0x07) != (mov2 & 0x07) && + (jmp & 0xF8FF) == 0xE0FF && + (mov2 & 0x07) == ((jmp>>8) & 0x07)) + { + ((unsigned long *)regs)[trans[mov1 & 0x07]] = addr1; + ((unsigned long *)regs)[trans[mov2 & 0x07]] = addr2; + regs->eip = addr2; + return 2; + } + } while (0); + + do { /* PaX: gcc trampoline emulation #2 */ + unsigned char mov, jmp; + unsigned long addr1, addr2; + + err = get_user(mov, (unsigned char __user *)regs->eip); + err |= get_user(addr1, (unsigned long __user *)(regs->eip + 1)); + err |= get_user(jmp, (unsigned char __user *)(regs->eip + 5)); + err |= get_user(addr2, (unsigned long __user *)(regs->eip + 6)); + + if (err) + break; + + if ((mov & 0xF8) == 0xB8 && + jmp == 0xE9) + { + ((unsigned long *)regs)[trans[mov & 0x07]] = addr1; + regs->eip += addr2 + 10; + return 2; + } + } while (0); +#endif + + return 1; /* PaX in action */ +} +#endif + +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) +void pax_report_insns(void *pc, void *sp) +{ + long i; + + printk(KERN_ERR "PAX: bytes at PC: "); + for (i = 0; i < 20; i++) { + unsigned char c; + if (get_user(c, (unsigned char __user *)pc+i)) + printk("?? "); + else + printk("%02x ", c); + } + printk("\n"); + + printk(KERN_ERR "PAX: bytes at SP-4: "); + for (i = -1; i < 20; i++) { + unsigned long c; + if (get_user(c, (unsigned long __user *)sp+i)) + printk("???????? "); + else + printk("%08lx ", c); + } + printk("\n"); +} +#endif diff -NurpP --minimal linux-2.6.17.13/arch/i386/mm/hugetlbpage.c linux-2.6.17.13-g-v/arch/i386/mm/hugetlbpage.c --- linux-2.6.17.13/arch/i386/mm/hugetlbpage.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/mm/hugetlbpage.c 2006-09-20 15:00:12.000000000 +0200 @@ -121,7 +121,12 @@ static unsigned long hugetlb_get_unmappe { struct mm_struct *mm = current->mm; struct vm_area_struct *vma; - unsigned long start_addr; + unsigned long start_addr, task_size = TASK_SIZE; + +#ifdef CONFIG_PAX_SEGMEXEC + if (mm->pax_flags & MF_PAX_SEGMEXEC) + task_size = SEGMEXEC_TASK_SIZE; +#endif if (len > mm->cached_hole_size) { start_addr = mm->free_area_cache; @@ -135,7 +140,7 @@ full_search: for (vma = find_vma(mm, addr); ; vma = vma->vm_next) { /* At this point: (!vma || addr < vma->vm_end). */ - if (TASK_SIZE - len < addr) { + if (task_size - len < addr) { /* * Start a new search - just in case we missed * some holes. @@ -163,9 +168,8 @@ static unsigned long hugetlb_get_unmappe { struct mm_struct *mm = current->mm; struct vm_area_struct *vma, *prev_vma; - unsigned long base = mm->mmap_base, addr = addr0; + unsigned long base = mm->mmap_base, addr; unsigned long largest_hole = mm->cached_hole_size; - int first_time = 1; /* don't allow allocations above current base */ if (mm->free_area_cache > base) @@ -175,7 +179,7 @@ static unsigned long hugetlb_get_unmappe largest_hole = 0; mm->free_area_cache = base; } -try_again: + /* make sure it can fit in the remaining address space */ if (mm->free_area_cache < len) goto fail; @@ -217,16 +221,6 @@ try_again: fail: /* - * if hint left us with no space for the requested - * mapping then try again: - */ - if (first_time) { - mm->free_area_cache = base; - largest_hole = 0; - first_time = 0; - goto try_again; - } - /* * A failed mmap() very likely causes application failure, * so fall back to the bottom-up function here. This scenario * can happen with large stack limits and large mmap() @@ -252,16 +246,23 @@ hugetlb_get_unmapped_area(struct file *f { struct mm_struct *mm = current->mm; struct vm_area_struct *vma; + unsigned long task_size = TASK_SIZE; if (len & ~HPAGE_MASK) return -EINVAL; - if (len > TASK_SIZE) + +#ifdef CONFIG_PAX_SEGMEXEC + if (mm->pax_flags & MF_PAX_SEGMEXEC) + task_size = SEGMEXEC_TASK_SIZE; +#endif + + if (len > task_size || addr > task_size - len) return -ENOMEM; if (addr) { addr = ALIGN(addr, HPAGE_SIZE); vma = find_vma(mm, addr); - if (TASK_SIZE - len >= addr && + if (task_size - len >= addr && (!vma || addr + len <= vma->vm_start)) return addr; } diff -NurpP --minimal linux-2.6.17.13/arch/i386/mm/init.c linux-2.6.17.13-g-v/arch/i386/mm/init.c --- linux-2.6.17.13/arch/i386/mm/init.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/mm/init.c 2006-09-20 15:00:12.000000000 +0200 @@ -41,6 +41,7 @@ #include #include #include +#include unsigned int __VMALLOC_RESERVE = 128 << 20; @@ -50,30 +51,6 @@ unsigned long highstart_pfn, highend_pfn static int noinline do_test_wp_bit(void); /* - * Creates a middle page table and puts a pointer to it in the - * given global directory entry. This only returns the gd entry - * in non-PAE compilation mode, since the middle layer is folded. - */ -static pmd_t * __init one_md_table_init(pgd_t *pgd) -{ - pud_t *pud; - pmd_t *pmd_table; - -#ifdef CONFIG_X86_PAE - pmd_table = (pmd_t *) alloc_bootmem_low_pages(PAGE_SIZE); - set_pgd(pgd, __pgd(__pa(pmd_table) | _PAGE_PRESENT)); - pud = pud_offset(pgd, 0); - if (pmd_table != pmd_offset(pud, 0)) - BUG(); -#else - pud = pud_offset(pgd, 0); - pmd_table = pmd_offset(pud, 0); -#endif - - return pmd_table; -} - -/* * Create a page table and place a pointer to it in a middle page * directory entry. */ @@ -81,7 +58,11 @@ static pte_t * __init one_page_table_ini { if (pmd_none(*pmd)) { pte_t *page_table = (pte_t *) alloc_bootmem_low_pages(PAGE_SIZE); +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) + set_pmd(pmd, __pmd(__pa(page_table) | _KERNPG_TABLE)); +#else set_pmd(pmd, __pmd(__pa(page_table) | _PAGE_TABLE)); +#endif if (page_table != pte_offset_kernel(pmd, 0)) BUG(); @@ -116,8 +97,6 @@ static void __init page_table_range_init pgd = pgd_base + pgd_idx; for ( ; (pgd_idx < PTRS_PER_PGD) && (vaddr != end); pgd++, pgd_idx++) { - if (pgd_none(*pgd)) - one_md_table_init(pgd); pud = pud_offset(pgd, vaddr); pmd = pmd_offset(pud, vaddr); for (; (pmd_idx < PTRS_PER_PMD) && (vaddr != end); pmd++, pmd_idx++) { @@ -130,11 +109,22 @@ static void __init page_table_range_init } } -static inline int is_kernel_text(unsigned long addr) +static inline int is_kernel_text(unsigned long start, unsigned long end) { - if (addr >= PAGE_OFFSET && addr <= (unsigned long)__init_end) - return 1; - return 0; + unsigned long etext; + +#if defined(CONFIG_MODULES) && defined(CONFIG_PAX_KERNEXEC) + etext = (unsigned long)&MODULES_END - __KERNEL_TEXT_OFFSET; +#else + etext = (unsigned long)&_etext; +#endif + + if ((start > etext + __KERNEL_TEXT_OFFSET || + end <= (unsigned long)_stext + __KERNEL_TEXT_OFFSET) && + (start > (unsigned long)_einittext + __KERNEL_TEXT_OFFSET || + end <= (unsigned long)_sinittext + __KERNEL_TEXT_OFFSET)) + return 0; + return 1; } /* @@ -146,26 +136,24 @@ static void __init kernel_physical_mappi { unsigned long pfn; pgd_t *pgd; + pud_t *pud; pmd_t *pmd; pte_t *pte; - int pgd_idx, pmd_idx, pte_ofs; + unsigned int pgd_idx, pmd_idx, pte_ofs; pgd_idx = pgd_index(PAGE_OFFSET); pgd = pgd_base + pgd_idx; pfn = 0; - for (; pgd_idx < PTRS_PER_PGD; pgd++, pgd_idx++) { - pmd = one_md_table_init(pgd); - if (pfn >= max_low_pfn) - continue; + for (; pgd_idx < PTRS_PER_PGD && pfn < max_low_pfn; pgd++, pgd_idx++) { + pud = pud_offset(pgd, 0); + pmd = pmd_offset(pud, 0); for (pmd_idx = 0; pmd_idx < PTRS_PER_PMD && pfn < max_low_pfn; pmd++, pmd_idx++) { - unsigned int address = pfn * PAGE_SIZE + PAGE_OFFSET; + unsigned long address = pfn * PAGE_SIZE + PAGE_OFFSET; /* Map with big pages if possible, otherwise create normal page tables. */ if (cpu_has_pse) { - unsigned int address2 = (pfn + PTRS_PER_PTE - 1) * PAGE_SIZE + PAGE_OFFSET + PAGE_SIZE-1; - - if (is_kernel_text(address) || is_kernel_text(address2)) + if (is_kernel_text(address, address + PMD_SIZE)) set_pmd(pmd, pfn_pmd(pfn, PAGE_KERNEL_LARGE_EXEC)); else set_pmd(pmd, pfn_pmd(pfn, PAGE_KERNEL_LARGE)); @@ -174,7 +162,7 @@ static void __init kernel_physical_mappi pte = one_page_table_init(pmd); for (pte_ofs = 0; pte_ofs < PTRS_PER_PTE && pfn < max_low_pfn; pte++, pfn++, pte_ofs++) { - if (is_kernel_text(address)) + if (is_kernel_text(address, address + PAGE_SIZE)) set_pte(pte, pfn_pte(pfn, PAGE_KERNEL_EXEC)); else set_pte(pte, pfn_pte(pfn, PAGE_KERNEL)); @@ -341,13 +329,6 @@ static void __init pagetable_init (void) unsigned long vaddr; pgd_t *pgd_base = swapper_pg_dir; -#ifdef CONFIG_X86_PAE - int i; - /* Init entries of the first-level page table to the zero page */ - for (i = 0; i < PTRS_PER_PGD; i++) - set_pgd(pgd_base + i, __pgd(__pa(empty_zero_page) | _PAGE_PRESENT)); -#endif - /* Enable PSE if available */ if (cpu_has_pse) { set_in_cr4(X86_CR4_PSE); @@ -371,17 +352,6 @@ static void __init pagetable_init (void) page_table_range_init(vaddr, 0, pgd_base); permanent_kmaps_init(pgd_base); - -#ifdef CONFIG_X86_PAE - /* - * Add low memory identity-mappings - SMP needs it when - * starting up on an AP from real-mode. In the non-PAE - * case we already have these mappings through head.S. - * All user-space mappings are explicitly cleared after - * SMP startup. - */ - set_pgd(&pgd_base[0], pgd_base[USER_PTRS_PER_PGD]); -#endif } #ifdef CONFIG_SOFTWARE_SUSPEND @@ -423,7 +393,6 @@ void zap_low_mappings (void) flush_tlb_all(); } -static int disable_nx __initdata = 0; u64 __supported_pte_mask __read_mostly = ~_PAGE_NX; /* @@ -437,11 +406,9 @@ u64 __supported_pte_mask __read_mostly = void __init noexec_setup(const char *str) { if (!strncmp(str, "on",2) && cpu_has_nx) { - __supported_pte_mask |= _PAGE_NX; - disable_nx = 0; + nx_enabled = 1; } else if (!strncmp(str,"off",3)) { - disable_nx = 1; - __supported_pte_mask &= ~_PAGE_NX; + nx_enabled = 0; } } @@ -450,17 +417,13 @@ int nx_enabled = 0; static void __init set_nx(void) { - unsigned int v[4], l, h; + if (!nx_enabled && cpu_has_nx) { + unsigned l, h; - if (cpu_has_pae && (cpuid_eax(0x80000000) > 0x80000001)) { - cpuid(0x80000001, &v[0], &v[1], &v[2], &v[3]); - if ((v[3] & (1 << 20)) && !disable_nx) { - rdmsr(MSR_EFER, l, h); - l |= EFER_NX; - wrmsr(MSR_EFER, l, h); - nx_enabled = 1; - __supported_pte_mask |= _PAGE_NX; - } + __supported_pte_mask &= ~_PAGE_NX; + rdmsr(MSR_EFER, l, h); + l &= ~EFER_NX; + wrmsr(MSR_EFER, l, h); } } @@ -512,14 +475,6 @@ void __init paging_init(void) load_cr3(swapper_pg_dir); -#ifdef CONFIG_X86_PAE - /* - * We will bail out later - printk doesn't work right now so - * the user would just see a hanging kernel. - */ - if (cpu_has_pae) - set_in_cr4(X86_CR4_PAE); -#endif __flush_tlb_all(); kmap_init(); @@ -611,7 +566,7 @@ void __init mem_init(void) set_highmem_pages_init(bad_ppro); codesize = (unsigned long) &_etext - (unsigned long) &_text; - datasize = (unsigned long) &_edata - (unsigned long) &_etext; + datasize = (unsigned long) &_edata - (unsigned long) &_data; initsize = (unsigned long) &__init_end - (unsigned long) &__init_begin; kclist_add(&kcore_mem, __va(0), max_low_pfn << PAGE_SHIFT); @@ -628,10 +583,6 @@ void __init mem_init(void) (unsigned long) (totalhigh_pages << (PAGE_SHIFT-10)) ); -#ifdef CONFIG_X86_PAE - if (!cpu_has_pae) - panic("cannot execute a PAE-enabled kernel on a PAE-less CPU!"); -#endif if (boot_cpu_data.wp_works_ok < 0) test_wp_bit(); @@ -761,6 +712,37 @@ void free_init_pages(char *what, unsigne void free_initmem(void) { + +#ifdef CONFIG_PAX_KERNEXEC + /* PaX: limit KERNEL_CS to actual size */ + unsigned long addr, limit; + int cpu; + pgd_t *pgd; + pud_t *pud; + pmd_t *pmd; + +#ifdef CONFIG_MODULES + limit = (unsigned long)&MODULES_END - __KERNEL_TEXT_OFFSET; +#else + limit = (unsigned long)&_etext; +#endif + limit = (limit - 1UL) >> PAGE_SHIFT; + + for (cpu = 0; cpu < NR_CPUS; cpu++) { + get_cpu_gdt_table(cpu)[GDT_ENTRY_KERNEL_CS].a = (get_cpu_gdt_table(cpu)[GDT_ENTRY_KERNEL_CS].a & 0xFFFF0000UL) | (limit & 0x0FFFFUL); + get_cpu_gdt_table(cpu)[GDT_ENTRY_KERNEL_CS].b = (get_cpu_gdt_table(cpu)[GDT_ENTRY_KERNEL_CS].b & 0xFFF0FFFFUL) | (limit & 0xF0000UL); + } + + /* PaX: make KERNEL_CS read-only */ + for (addr = __KERNEL_TEXT_OFFSET; addr < (unsigned long)&_data; addr += PMD_SIZE) { + pgd = pgd_offset_k(addr); + pud = pud_offset(pgd, addr); + pmd = pmd_offset(pud, addr); + set_pmd(pmd, __pmd(pmd_val(*pmd) & ~_PAGE_RW)); + } + flush_tlb_all(); +#endif + free_init_pages("unused kernel memory", (unsigned long)(&__init_begin), (unsigned long)(&__init_end)); diff -NurpP --minimal linux-2.6.17.13/arch/i386/mm/mmap.c linux-2.6.17.13-g-v/arch/i386/mm/mmap.c --- linux-2.6.17.13/arch/i386/mm/mmap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/mm/mmap.c 2006-09-20 15:00:12.000000000 +0200 @@ -34,12 +34,18 @@ * Leave an at least ~128 MB hole. */ #define MIN_GAP (128*1024*1024) -#define MAX_GAP (TASK_SIZE/6*5) +#define MAX_GAP (task_size/6*5) static inline unsigned long mmap_base(struct mm_struct *mm) { unsigned long gap = current->signal->rlim[RLIMIT_STACK].rlim_cur; unsigned long random_factor = 0; + unsigned long task_size = TASK_SIZE; + +#ifdef CONFIG_PAX_SEGMEXEC + if (mm->pax_flags & MF_PAX_SEGMEXEC) + task_size = SEGMEXEC_TASK_SIZE; +#endif if (current->flags & PF_RANDOMIZE) random_factor = get_random_int() % (1024*1024); @@ -49,7 +55,7 @@ static inline unsigned long mmap_base(st else if (gap > MAX_GAP) gap = MAX_GAP; - return PAGE_ALIGN(TASK_SIZE - gap - random_factor); + return PAGE_ALIGN(task_size - gap - random_factor); } /* @@ -66,10 +72,22 @@ void arch_pick_mmap_layout(struct mm_str (current->personality & ADDR_COMPAT_LAYOUT) || current->signal->rlim[RLIMIT_STACK].rlim_cur == RLIM_INFINITY) { mm->mmap_base = TASK_UNMAPPED_BASE; + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base += mm->delta_mmap; +#endif + mm->get_unmapped_area = arch_get_unmapped_area; mm->unmap_area = arch_unmap_area; } else { mm->mmap_base = mmap_base(mm); + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base -= mm->delta_mmap + mm->delta_stack; +#endif + mm->get_unmapped_area = arch_get_unmapped_area_topdown; mm->unmap_area = arch_unmap_area_topdown; } diff -NurpP --minimal linux-2.6.17.13/arch/i386/mm/pageattr.c linux-2.6.17.13-g-v/arch/i386/mm/pageattr.c --- linux-2.6.17.13/arch/i386/mm/pageattr.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/mm/pageattr.c 2006-09-20 15:00:12.000000000 +0200 @@ -14,6 +14,7 @@ #include #include #include +#include static DEFINE_SPINLOCK(cpa_lock); static struct list_head df_list = LIST_HEAD_INIT(df_list); @@ -84,7 +85,18 @@ static void set_pmd_pte(pte_t *kpte, uns struct page *page; unsigned long flags; +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + set_pte_atomic(kpte, pte); /* change init_mm */ + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + if (PTRS_PER_PMD > 1) return; @@ -111,7 +123,7 @@ static inline void revert_page(struct pa pte_t *linear; ref_prot = - ((address & LARGE_PAGE_MASK) < (unsigned long)&_etext) + ((address & LARGE_PAGE_MASK) < (unsigned long)&_etext + __KERNEL_TEXT_OFFSET) ? PAGE_KERNEL_LARGE_EXEC : PAGE_KERNEL_LARGE; linear = (pte_t *) @@ -143,7 +155,7 @@ __change_page_attr(struct page *page, pg struct page *split; ref_prot = - ((address & LARGE_PAGE_MASK) < (unsigned long)&_etext) + ((address & LARGE_PAGE_MASK) < (unsigned long)&_etext + __KERNEL_TEXT_OFFSET) ? PAGE_KERNEL_EXEC : PAGE_KERNEL; split = split_large_page(address, prot, ref_prot); if (!split) diff -NurpP --minimal linux-2.6.17.13/arch/i386/oprofile/backtrace.c linux-2.6.17.13-g-v/arch/i386/oprofile/backtrace.c --- linux-2.6.17.13/arch/i386/oprofile/backtrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/oprofile/backtrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -116,7 +116,7 @@ x86_backtrace(struct pt_regs * const reg head = (struct frame_head *)regs->ebp; #endif - if (!user_mode_vm(regs)) { + if (!user_mode(regs)) { while (depth-- && valid_kernel_stack(head, regs)) head = dump_kernel_backtrace(head); return; diff -NurpP --minimal linux-2.6.17.13/arch/i386/power/cpu.c linux-2.6.17.13-g-v/arch/i386/power/cpu.c --- linux-2.6.17.13/arch/i386/power/cpu.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/i386/power/cpu.c 2006-09-20 15:00:12.000000000 +0200 @@ -62,7 +62,7 @@ static void do_fpu_end(void) static void fix_processor_context(void) { int cpu = smp_processor_id(); - struct tss_struct * t = &per_cpu(init_tss, cpu); + struct tss_struct * t = init_tss + cpu; set_tss_desc(cpu,t); /* This just modifies memory; should not be necessary. But... This is necessary, because 386 hardware has concept of busy TSS or some similar stupidity. */ diff -NurpP --minimal linux-2.6.17.13/arch/ia64/Kconfig linux-2.6.17.13-g-v/arch/ia64/Kconfig --- linux-2.6.17.13/arch/ia64/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -506,6 +506,8 @@ endmenu source "arch/ia64/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/ia64/ia32/binfmt_elf32.c linux-2.6.17.13-g-v/arch/ia64/ia32/binfmt_elf32.c --- linux-2.6.17.13/arch/ia64/ia32/binfmt_elf32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/ia32/binfmt_elf32.c 2006-09-20 15:00:12.000000000 +0200 @@ -46,6 +46,17 @@ randomize_stack_top(unsigned long stack_ #define elf_read_implies_exec(ex, have_pt_gnu_stack) (!(have_pt_gnu_stack)) +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) ((tsk)->personality == PER_LINUX32 ? 0x08048000UL : 0x4000000000000000UL) + +#define PAX_DELTA_MMAP_LSB(tsk) IA32_PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) ((tsk)->personality == PER_LINUX32 ? 16 : 3*PAGE_SHIFT - IA32_PAGE_SHIFT) +#define PAX_DELTA_EXEC_LSB(tsk) IA32_PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) ((tsk)->personality == PER_LINUX32 ? 16 : 3*PAGE_SHIFT - IA32_PAGE_SHIFT) +#define PAX_DELTA_STACK_LSB(tsk) IA32_PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) ((tsk)->personality == PER_LINUX32 ? 16 : 3*PAGE_SHIFT - IA32_PAGE_SHIFT) +#endif + /* Ugly but avoids duplication */ #include "../../../fs/binfmt_elf.c" @@ -239,7 +250,8 @@ ia32_setup_arg_pages (struct linux_binpr kmem_cache_free(vm_area_cachep, mpnt); return ret; } - current->mm->stack_vm = current->mm->total_vm = vma_pages(mpnt); + vx_vmpages_sub(current->mm, current->mm->total_vm - vma_pages(mpnt)); + current->mm->stack_vm = current->mm->total_vm; } for (i = 0 ; i < MAX_ARG_PAGES ; i++) { diff -NurpP --minimal linux-2.6.17.13/arch/ia64/ia32/ia32_entry.S linux-2.6.17.13-g-v/arch/ia64/ia32/ia32_entry.S --- linux-2.6.17.13/arch/ia64/ia32/ia32_entry.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/ia32/ia32_entry.S 2006-09-20 15:00:12.000000000 +0200 @@ -483,7 +483,7 @@ ia32_syscall_table: data8 sys_tgkill /* 270 */ data8 compat_sys_utimes data8 sys32_fadvise64_64 - data8 sys_ni_syscall + data8 sys32_vserver data8 sys_ni_syscall data8 sys_ni_syscall /* 275 */ data8 sys_ni_syscall diff -NurpP --minimal linux-2.6.17.13/arch/ia64/ia32/ia32priv.h linux-2.6.17.13-g-v/arch/ia64/ia32/ia32priv.h --- linux-2.6.17.13/arch/ia64/ia32/ia32priv.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/ia32/ia32priv.h 2006-09-20 15:00:12.000000000 +0200 @@ -305,7 +305,14 @@ struct old_linux32_dirent { #define ELF_DATA ELFDATA2LSB #define ELF_ARCH EM_386 -#define IA32_STACK_TOP IA32_PAGE_OFFSET +#ifdef CONFIG_PAX_RANDUSTACK +#define __IA32_DELTA_STACK (current->mm->delta_stack) +#else +#define __IA32_DELTA_STACK 0UL +#endif + +#define IA32_STACK_TOP (IA32_PAGE_OFFSET - __IA32_DELTA_STACK) + #define IA32_GATE_OFFSET IA32_PAGE_OFFSET #define IA32_GATE_END IA32_PAGE_OFFSET + PAGE_SIZE diff -NurpP --minimal linux-2.6.17.13/arch/ia64/kernel/entry.S linux-2.6.17.13-g-v/arch/ia64/kernel/entry.S --- linux-2.6.17.13/arch/ia64/kernel/entry.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/kernel/entry.S 2006-09-20 15:00:12.000000000 +0200 @@ -1577,7 +1577,7 @@ sys_call_table: data8 sys_mq_notify data8 sys_mq_getsetattr data8 sys_ni_syscall // reserved for kexec_load - data8 sys_ni_syscall // reserved for vserver + data8 sys_vserver data8 sys_waitid // 1270 data8 sys_add_key data8 sys_request_key diff -NurpP --minimal linux-2.6.17.13/arch/ia64/kernel/module.c linux-2.6.17.13-g-v/arch/ia64/kernel/module.c --- linux-2.6.17.13/arch/ia64/kernel/module.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/kernel/module.c 2006-09-20 15:00:12.000000000 +0200 @@ -322,7 +322,7 @@ module_alloc (unsigned long size) void module_free (struct module *mod, void *module_region) { - if (mod->arch.init_unw_table && module_region == mod->module_init) { + if (mod->arch.init_unw_table && module_region == mod->module_init_rx) { unw_remove_unwind_table(mod->arch.init_unw_table); mod->arch.init_unw_table = NULL; } @@ -500,15 +500,39 @@ module_frob_arch_sections (Elf_Ehdr *ehd } static inline int +in_init_rx (const struct module *mod, uint64_t addr) +{ + return addr - (uint64_t) mod->module_init_rx < mod->init_size_rx; +} + +static inline int +in_init_rw (const struct module *mod, uint64_t addr) +{ + return addr - (uint64_t) mod->module_init_rw < mod->init_size_rw; +} + +static inline int in_init (const struct module *mod, uint64_t addr) { - return addr - (uint64_t) mod->module_init < mod->init_size; + return in_init_rx(mod, value) || in_init_rw(mod, value); +} + +static inline int +in_core_rx (const struct module *mod, uint64_t addr) +{ + return addr - (uint64_t) mod->module_core_rx < mod->core_size_rx; +} + +static inline int +in_core_rw (const struct module *mod, uint64_t addr) +{ + return addr - (uint64_t) mod->module_core_rw < mod->core_size_rw; } static inline int in_core (const struct module *mod, uint64_t addr) { - return addr - (uint64_t) mod->module_core < mod->core_size; + return in_core_rx(mod, value) || in_core_rw(mod, value); } static inline int @@ -692,7 +716,14 @@ do_reloc (struct module *mod, uint8_t r_ break; case RV_BDREL: - val -= (uint64_t) (in_init(mod, val) ? mod->module_init : mod->module_core); + if (in_init_rx(mod, val)) + val -= (uint64_t) mod->module_init_rx; + else if (in_init_rw(mod, val)) + val -= (uint64_t) mod->module_init_rw; + else if (in_core_rx(mod, val)) + val -= (uint64_t) mod->module_core_rx; + else if (in_core_rw(mod, val)) + val -= (uint64_t) mod->module_core_rw; break; case RV_LTV: @@ -826,15 +857,15 @@ apply_relocate_add (Elf64_Shdr *sechdrs, * addresses have been selected... */ uint64_t gp; - if (mod->core_size > MAX_LTOFF) + if (mod->core_size_rx + mod->core_size_rw > MAX_LTOFF) /* * This takes advantage of fact that SHF_ARCH_SMALL gets allocated * at the end of the module. */ - gp = mod->core_size - MAX_LTOFF / 2; + gp = mod->core_size_rx + mod->core_size_rw - MAX_LTOFF / 2; else - gp = mod->core_size / 2; - gp = (uint64_t) mod->module_core + ((gp + 7) & -8); + gp = (mod->core_size_rx + mod->core_size_rw) / 2; + gp = (uint64_t) mod->module_core_rx + ((gp + 7) & -8); mod->arch.gp = gp; DEBUGP("%s: placing gp at 0x%lx\n", __FUNCTION__, gp); } diff -NurpP --minimal linux-2.6.17.13/arch/ia64/kernel/perfmon.c linux-2.6.17.13-g-v/arch/ia64/kernel/perfmon.c --- linux-2.6.17.13/arch/ia64/kernel/perfmon.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/kernel/perfmon.c 2006-09-20 15:00:12.000000000 +0200 @@ -41,6 +41,8 @@ #include #include #include +#include +#include #include #include @@ -2355,7 +2357,7 @@ pfm_smpl_buffer_alloc(struct task_struct */ insert_vm_struct(mm, vma); - mm->total_vm += size >> PAGE_SHIFT; + vx_vmpages_add(mm, size >> PAGE_SHIFT); vm_stat_account(vma->vm_mm, vma->vm_flags, vma->vm_file, vma_pages(vma)); up_write(&task->mm->mmap_sem); diff -NurpP --minimal linux-2.6.17.13/arch/ia64/kernel/ptrace.c linux-2.6.17.13-g-v/arch/ia64/kernel/ptrace.c --- linux-2.6.17.13/arch/ia64/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -18,6 +18,8 @@ #include #include #include +#include +#include #include #include @@ -1443,10 +1445,16 @@ sys_ptrace (long request, pid_t pid, uns read_unlock(&tasklist_lock); if (!child) goto out; + if (!vx_check(vx_task_xid(child), VX_WATCH|VX_IDENT)) + goto out_tsk; + ret = -EPERM; if (pid == 1) /* no messing around with init! */ goto out_tsk; + if (gr_handle_ptrace(child, request)) + goto out_tsk; + if (request == PTRACE_ATTACH) { ret = ptrace_attach(child); goto out_tsk; diff -NurpP --minimal linux-2.6.17.13/arch/ia64/kernel/sys_ia64.c linux-2.6.17.13-g-v/arch/ia64/kernel/sys_ia64.c --- linux-2.6.17.13/arch/ia64/kernel/sys_ia64.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/kernel/sys_ia64.c 2006-09-20 15:00:12.000000000 +0200 @@ -38,6 +38,13 @@ arch_get_unmapped_area (struct file *fil if (REGION_NUMBER(addr) == RGN_HPAGE) addr = 0; #endif + +#ifdef CONFIG_PAX_RANDMMAP + if ((mm->pax_flags & MF_PAX_RANDMMAP) && addr && filp) + addr = mm->free_area_cache; + else +#endif + if (!addr) addr = mm->free_area_cache; @@ -56,9 +63,9 @@ arch_get_unmapped_area (struct file *fil for (vma = find_vma(mm, addr); ; vma = vma->vm_next) { /* At this point: (!vma || addr < vma->vm_end). */ if (TASK_SIZE - len < addr || RGN_MAP_LIMIT - len < REGION_OFFSET(addr)) { - if (start_addr != TASK_UNMAPPED_BASE) { + if (start_addr != mm->mmap_base) { /* Start a new search --- just in case we missed some holes. */ - addr = TASK_UNMAPPED_BASE; + addr = mm->mmap_base; goto full_search; } return -ENOMEM; diff -NurpP --minimal linux-2.6.17.13/arch/ia64/mm/fault.c linux-2.6.17.13-g-v/arch/ia64/mm/fault.c --- linux-2.6.17.13/arch/ia64/mm/fault.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/mm/fault.c 2006-09-20 15:00:12.000000000 +0200 @@ -10,6 +10,8 @@ #include #include #include +#include +#include #include #include @@ -51,6 +53,23 @@ mapped_kernel_page_is_present (unsigned return pte_present(pte); } +#ifdef CONFIG_PAX_PAGEEXEC +void pax_report_insns(void *pc, void *sp) +{ + unsigned long i; + + printk(KERN_ERR "PAX: bytes at PC: "); + for (i = 0; i < 8; i++) { + unsigned int c; + if (get_user(c, (unsigned int*)pc+i)) + printk("???????? "); + else + printk("%08x ", c); + } + printk("\n"); +} +#endif + void __kprobes ia64_do_page_fault (unsigned long address, unsigned long isr, struct pt_regs *regs) { @@ -116,9 +135,23 @@ ia64_do_page_fault (unsigned long addres | (((isr >> IA64_ISR_W_BIT) & 1UL) << VM_WRITE_BIT) | (((isr >> IA64_ISR_R_BIT) & 1UL) << VM_READ_BIT)); - if ((vma->vm_flags & mask) != mask) + if ((vma->vm_flags & mask) != mask) { + +#ifdef CONFIG_PAX_PAGEEXEC + if (!(vma->vm_flags & VM_EXEC) && (mask & VM_EXEC)) { + if (!(mm->pax_flags & MF_PAX_PAGEEXEC) || address != regs->cr_iip) + goto bad_area; + + up_read(&mm->mmap_sem); + pax_report_fault(regs, (void*)regs->cr_iip, (void*)regs->r12); + do_exit(SIGKILL); + } +#endif + goto bad_area; + } + survive: /* * If for any reason at all we couldn't handle the fault, make diff -NurpP --minimal linux-2.6.17.13/arch/ia64/mm/init.c linux-2.6.17.13-g-v/arch/ia64/mm/init.c --- linux-2.6.17.13/arch/ia64/mm/init.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/mm/init.c 2006-09-20 15:00:12.000000000 +0200 @@ -20,8 +20,8 @@ #include #include #include +#include -#include #include #include #include diff -NurpP --minimal linux-2.6.17.13/arch/ia64/sn/kernel/xpc_main.c linux-2.6.17.13-g-v/arch/ia64/sn/kernel/xpc_main.c --- linux-2.6.17.13/arch/ia64/sn/kernel/xpc_main.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ia64/sn/kernel/xpc_main.c 2006-09-20 15:00:12.000000000 +0200 @@ -108,6 +108,7 @@ static ctl_table xpc_sys_xpc_hb_dir[] = 0644, NULL, &proc_dointvec_minmax, + NULL, &sysctl_intvec, NULL, &xpc_hb_min_interval, @@ -121,6 +122,7 @@ static ctl_table xpc_sys_xpc_hb_dir[] = 0644, NULL, &proc_dointvec_minmax, + NULL, &sysctl_intvec, NULL, &xpc_hb_check_min_interval, @@ -145,6 +147,7 @@ static ctl_table xpc_sys_xpc_dir[] = { 0644, NULL, &proc_dointvec_minmax, + NULL, &sysctl_intvec, NULL, &xpc_disengage_request_min_timelimit, diff -NurpP --minimal linux-2.6.17.13/arch/m32r/kernel/sys_m32r.c linux-2.6.17.13-g-v/arch/m32r/kernel/sys_m32r.c --- linux-2.6.17.13/arch/m32r/kernel/sys_m32r.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/m32r/kernel/sys_m32r.c 2006-09-20 15:00:12.000000000 +0200 @@ -21,6 +21,7 @@ #include #include #include +#include #include #include @@ -206,7 +207,7 @@ asmlinkage int sys_uname(struct old_utsn if (!name) return -EFAULT; down_read(&uts_sem); - err=copy_to_user(name, &system_utsname, sizeof (*name)); + err=copy_to_user(name, vx_new_utsname(), sizeof (*name)); up_read(&uts_sem); return err?-EFAULT:0; } diff -NurpP --minimal linux-2.6.17.13/arch/m68k/Kconfig linux-2.6.17.13-g-v/arch/m68k/Kconfig --- linux-2.6.17.13/arch/m68k/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/m68k/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -654,6 +654,8 @@ source "fs/Kconfig" source "arch/m68k/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/m68k/kernel/ptrace.c linux-2.6.17.13-g-v/arch/m68k/kernel/ptrace.c --- linux-2.6.17.13/arch/m68k/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/m68k/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -280,6 +280,8 @@ long arch_ptrace(struct task_struct *chi ret = ptrace_request(child, request, addr, data); break; } + if (!vx_check(vx_task_xid(child), VX_WATCH|VX_IDENT)) + goto out_tsk; return ret; out_eio: diff -NurpP --minimal linux-2.6.17.13/arch/m68knommu/Kconfig linux-2.6.17.13-g-v/arch/m68knommu/Kconfig --- linux-2.6.17.13/arch/m68knommu/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/m68knommu/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -654,6 +654,8 @@ source "fs/Kconfig" source "arch/m68knommu/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/mips/Kconfig linux-2.6.17.13-g-v/arch/mips/Kconfig --- linux-2.6.17.13/arch/mips/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -1852,6 +1852,8 @@ source "arch/mips/oprofile/Kconfig" source "arch/mips/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/mips/kernel/binfmt_elfn32.c linux-2.6.17.13-g-v/arch/mips/kernel/binfmt_elfn32.c --- linux-2.6.17.13/arch/mips/kernel/binfmt_elfn32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/kernel/binfmt_elfn32.c 2006-09-20 15:00:12.000000000 +0200 @@ -50,6 +50,17 @@ typedef elf_fpreg_t elf_fpregset_t[ELF_N #undef ELF_ET_DYN_BASE #define ELF_ET_DYN_BASE (TASK32_SIZE / 3 * 2) +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 0x00400000UL : 0x00400000UL) + +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 27-PAGE_SHIFT : 36-PAGE_SHIFT) +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 27-PAGE_SHIFT : 36-PAGE_SHIFT) +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 27-PAGE_SHIFT : 36-PAGE_SHIFT) +#endif + #include #include #include diff -NurpP --minimal linux-2.6.17.13/arch/mips/kernel/binfmt_elfo32.c linux-2.6.17.13-g-v/arch/mips/kernel/binfmt_elfo32.c --- linux-2.6.17.13/arch/mips/kernel/binfmt_elfo32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/kernel/binfmt_elfo32.c 2006-09-20 15:00:12.000000000 +0200 @@ -52,6 +52,17 @@ typedef elf_fpreg_t elf_fpregset_t[ELF_N #undef ELF_ET_DYN_BASE #define ELF_ET_DYN_BASE (TASK32_SIZE / 3 * 2) +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 0x00400000UL : 0x00400000UL) + +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 27-PAGE_SHIFT : 36-PAGE_SHIFT) +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 27-PAGE_SHIFT : 36-PAGE_SHIFT) +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 27-PAGE_SHIFT : 36-PAGE_SHIFT) +#endif + #include #include #include diff -NurpP --minimal linux-2.6.17.13/arch/mips/kernel/linux32.c linux-2.6.17.13-g-v/arch/mips/kernel/linux32.c --- linux-2.6.17.13/arch/mips/kernel/linux32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/kernel/linux32.c 2006-09-20 15:00:12.000000000 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include #include @@ -1040,7 +1041,7 @@ asmlinkage long sys32_newuname(struct ne int ret = 0; down_read(&uts_sem); - if (copy_to_user(name,&system_utsname,sizeof *name)) + if (copy_to_user(name, vx_new_utsname(), sizeof *name)) ret = -EFAULT; up_read(&uts_sem); diff -NurpP --minimal linux-2.6.17.13/arch/mips/kernel/mips-mt.c linux-2.6.17.13-g-v/arch/mips/kernel/mips-mt.c --- linux-2.6.17.13/arch/mips/kernel/mips-mt.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/kernel/mips-mt.c 2006-09-20 15:00:12.000000000 +0200 @@ -7,6 +7,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/arch/mips/kernel/ptrace.c linux-2.6.17.13-g-v/arch/mips/kernel/ptrace.c --- linux-2.6.17.13/arch/mips/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -490,6 +490,8 @@ asmlinkage void do_syscall_trace(struct goto out; if (!test_thread_flag(TIF_SYSCALL_TRACE)) goto out; + if (!vx_check(vx_task_xid(child), VX_WATCH|VX_IDENT)) + goto out_tsk; /* The 0x80 provides a way for the tracing parent to distinguish between a syscall stop and SIGTRAP delivery */ diff -NurpP --minimal linux-2.6.17.13/arch/mips/kernel/scall32-o32.S linux-2.6.17.13-g-v/arch/mips/kernel/scall32-o32.S --- linux-2.6.17.13/arch/mips/kernel/scall32-o32.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/kernel/scall32-o32.S 2006-09-20 15:00:12.000000000 +0200 @@ -618,7 +618,7 @@ einval: li v0, -EINVAL sys sys_mq_timedreceive 5 sys sys_mq_notify 2 /* 4275 */ sys sys_mq_getsetattr 3 - sys sys_ni_syscall 0 /* sys_vserver */ + sys sys_vserver 3 sys sys_waitid 5 sys sys_ni_syscall 0 /* available, was setaltroot */ sys sys_add_key 5 /* 4280 */ diff -NurpP --minimal linux-2.6.17.13/arch/mips/kernel/scall64-64.S linux-2.6.17.13-g-v/arch/mips/kernel/scall64-64.S --- linux-2.6.17.13/arch/mips/kernel/scall64-64.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/kernel/scall64-64.S 2006-09-20 15:00:12.000000000 +0200 @@ -433,7 +433,7 @@ sys_call_table: PTR sys_mq_timedreceive PTR sys_mq_notify PTR sys_mq_getsetattr /* 5235 */ - PTR sys_ni_syscall /* sys_vserver */ + PTR sys_vserver PTR sys_waitid PTR sys_ni_syscall /* available, was setaltroot */ PTR sys_add_key diff -NurpP --minimal linux-2.6.17.13/arch/mips/kernel/scall64-n32.S linux-2.6.17.13-g-v/arch/mips/kernel/scall64-n32.S --- linux-2.6.17.13/arch/mips/kernel/scall64-n32.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/kernel/scall64-n32.S 2006-09-20 15:00:12.000000000 +0200 @@ -359,7 +359,7 @@ EXPORT(sysn32_call_table) PTR compat_sys_mq_timedreceive PTR compat_sys_mq_notify PTR compat_sys_mq_getsetattr - PTR sys_ni_syscall /* 6240, sys_vserver */ + PTR sys32_vserver /* 6240 */ PTR sysn32_waitid PTR sys_ni_syscall /* available, was setaltroot */ PTR sys_add_key diff -NurpP --minimal linux-2.6.17.13/arch/mips/kernel/scall64-o32.S linux-2.6.17.13-g-v/arch/mips/kernel/scall64-o32.S --- linux-2.6.17.13/arch/mips/kernel/scall64-o32.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/kernel/scall64-o32.S 2006-09-20 15:00:12.000000000 +0200 @@ -481,7 +481,7 @@ sys_call_table: PTR compat_sys_mq_timedreceive PTR compat_sys_mq_notify /* 4275 */ PTR compat_sys_mq_getsetattr - PTR sys_ni_syscall /* sys_vserver */ + PTR sys32_vserver PTR sys32_waitid PTR sys_ni_syscall /* available, was setaltroot */ PTR sys_add_key /* 4280 */ diff -NurpP --minimal linux-2.6.17.13/arch/mips/kernel/syscall.c linux-2.6.17.13-g-v/arch/mips/kernel/syscall.c --- linux-2.6.17.13/arch/mips/kernel/syscall.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/kernel/syscall.c 2006-09-20 15:00:12.000000000 +0200 @@ -29,6 +29,7 @@ #include #include #include +#include #include #include @@ -89,6 +90,11 @@ unsigned long arch_get_unmapped_area(str do_color_align = 0; if (filp || (flags & MAP_SHARED)) do_color_align = 1; + +#ifdef CONFIG_PAX_RANDMMAP + if (!(current->mm->pax_flags & MF_PAX_RANDMMAP) || !filp) +#endif + if (addr) { if (do_color_align) addr = COLOUR_ALIGN(addr, pgoff); @@ -99,7 +105,7 @@ unsigned long arch_get_unmapped_area(str (!vmm || addr + len <= vmm->vm_start)) return addr; } - addr = TASK_UNMAPPED_BASE; + addr = current->mm->mmap_base; if (do_color_align) addr = COLOUR_ALIGN(addr, pgoff); else @@ -232,7 +238,7 @@ out: */ asmlinkage int sys_uname(struct old_utsname __user * name) { - if (name && !copy_to_user(name, &system_utsname, sizeof (*name))) + if (name && !copy_to_user(name, vx_new_utsname(), sizeof (*name))) return 0; return -EFAULT; } @@ -243,21 +249,23 @@ asmlinkage int sys_uname(struct old_utsn asmlinkage int sys_olduname(struct oldold_utsname __user * name) { int error; + struct new_utsname *ptr; if (!name) return -EFAULT; if (!access_ok(VERIFY_WRITE,name,sizeof(struct oldold_utsname))) return -EFAULT; - error = __copy_to_user(&name->sysname,&system_utsname.sysname,__OLD_UTS_LEN); + ptr = vx_new_utsname(); + error = __copy_to_user(&name->sysname,ptr->sysname,__OLD_UTS_LEN); error -= __put_user(0,name->sysname+__OLD_UTS_LEN); - error -= __copy_to_user(&name->nodename,&system_utsname.nodename,__OLD_UTS_LEN); + error -= __copy_to_user(&name->nodename,ptr->nodename,__OLD_UTS_LEN); error -= __put_user(0,name->nodename+__OLD_UTS_LEN); - error -= __copy_to_user(&name->release,&system_utsname.release,__OLD_UTS_LEN); + error -= __copy_to_user(&name->release,ptr->release,__OLD_UTS_LEN); error -= __put_user(0,name->release+__OLD_UTS_LEN); - error -= __copy_to_user(&name->version,&system_utsname.version,__OLD_UTS_LEN); + error -= __copy_to_user(&name->version,ptr->version,__OLD_UTS_LEN); error -= __put_user(0,name->version+__OLD_UTS_LEN); - error -= __copy_to_user(&name->machine,&system_utsname.machine,__OLD_UTS_LEN); + error -= __copy_to_user(&name->machine,ptr->machine,__OLD_UTS_LEN); error = __put_user(0,name->machine+__OLD_UTS_LEN); error = error ? -EFAULT : 0; diff -NurpP --minimal linux-2.6.17.13/arch/mips/kernel/sysirix.c linux-2.6.17.13-g-v/arch/mips/kernel/sysirix.c --- linux-2.6.17.13/arch/mips/kernel/sysirix.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/kernel/sysirix.c 2006-09-20 15:00:12.000000000 +0200 @@ -31,6 +31,7 @@ #include #include #include +#include #include #include @@ -884,7 +885,7 @@ asmlinkage int irix_getdomainname(char _ down_read(&uts_sem); if (len > __NEW_UTS_LEN) len = __NEW_UTS_LEN; - err = copy_to_user(name, system_utsname.domainname, len) ? -EFAULT : 0; + err = copy_to_user(name, vx_new_uts(domainname), len) ? -EFAULT : 0; up_read(&uts_sem); return err; @@ -1127,11 +1128,11 @@ struct iuname { asmlinkage int irix_uname(struct iuname __user *buf) { down_read(&uts_sem); - if (copy_from_user(system_utsname.sysname, buf->sysname, 65) - || copy_from_user(system_utsname.nodename, buf->nodename, 65) - || copy_from_user(system_utsname.release, buf->release, 65) - || copy_from_user(system_utsname.version, buf->version, 65) - || copy_from_user(system_utsname.machine, buf->machine, 65)) { + if (copy_from_user(vx_new_uts(sysname), buf->sysname, 65) + || copy_from_user(vx_new_uts(nodename), buf->nodename, 65) + || copy_from_user(vx_new_uts(release), buf->release, 65) + || copy_from_user(vx_new_uts(version), buf->version, 65) + || copy_from_user(vx_new_uts(machine), buf->machine, 65)) { return -EFAULT; } up_read(&uts_sem); diff -NurpP --minimal linux-2.6.17.13/arch/mips/mm/fault.c linux-2.6.17.13-g-v/arch/mips/mm/fault.c --- linux-2.6.17.13/arch/mips/mm/fault.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/mips/mm/fault.c 2006-09-20 15:00:12.000000000 +0200 @@ -27,6 +27,23 @@ #include #include /* For VMALLOC_END */ +#ifdef CONFIG_PAX_PAGEEXEC +void pax_report_insns(void *pc) +{ + unsigned long i; + + printk(KERN_ERR "PAX: bytes at PC: "); + for (i = 0; i < 5; i++) { + unsigned int c; + if (get_user(c, (unsigned int*)pc+i)) + printk("???????? "); + else + printk("%08x ", c); + } + printk("\n"); +} +#endif + /* * This routine handles page faults. It determines the address, * and the problem, and then passes it off to one of the appropriate diff -NurpP --minimal linux-2.6.17.13/arch/parisc/Kconfig linux-2.6.17.13-g-v/arch/parisc/Kconfig --- linux-2.6.17.13/arch/parisc/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/parisc/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -253,6 +253,8 @@ source "arch/parisc/oprofile/Kconfig" source "arch/parisc/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/parisc/hpux/sys_hpux.c linux-2.6.17.13-g-v/arch/parisc/hpux/sys_hpux.c --- linux-2.6.17.13/arch/parisc/hpux/sys_hpux.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/parisc/hpux/sys_hpux.c 2006-09-20 15:00:12.000000000 +0200 @@ -33,6 +33,7 @@ #include #include #include +#include #include #include @@ -266,15 +267,15 @@ static int hpux_uname(struct hpux_utsnam down_read(&uts_sem); - error = __copy_to_user(&name->sysname,&system_utsname.sysname,HPUX_UTSLEN-1); + error = __copy_to_user(&name->sysname,vx_new_uts(sysname),HPUX_UTSLEN-1); error |= __put_user(0,name->sysname+HPUX_UTSLEN-1); - error |= __copy_to_user(&name->nodename,&system_utsname.nodename,HPUX_UTSLEN-1); + error |= __copy_to_user(&name->nodename,vx_new_uts(nodename),HPUX_UTSLEN-1); error |= __put_user(0,name->nodename+HPUX_UTSLEN-1); - error |= __copy_to_user(&name->release,&system_utsname.release,HPUX_UTSLEN-1); + error |= __copy_to_user(&name->release,vx_new_uts(release),HPUX_UTSLEN-1); error |= __put_user(0,name->release+HPUX_UTSLEN-1); - error |= __copy_to_user(&name->version,&system_utsname.version,HPUX_UTSLEN-1); + error |= __copy_to_user(&name->version,vx_new_uts(version),HPUX_UTSLEN-1); error |= __put_user(0,name->version+HPUX_UTSLEN-1); - error |= __copy_to_user(&name->machine,&system_utsname.machine,HPUX_UTSLEN-1); + error |= __copy_to_user(&name->machine,vx_new_uts(machine),HPUX_UTSLEN-1); error |= __put_user(0,name->machine+HPUX_UTSLEN-1); up_read(&uts_sem); @@ -373,8 +374,8 @@ int hpux_utssys(char *ubuf, int n, int t /* TODO: print a warning about using this? */ down_write(&uts_sem); error = -EFAULT; - if (!copy_from_user(system_utsname.sysname, ubuf, len)) { - system_utsname.sysname[len] = 0; + if (!copy_from_user(vx_new_uts(sysname), ubuf, len)) { + vx_new_uts(sysname)[len] = 0; error = 0; } up_write(&uts_sem); @@ -400,8 +401,8 @@ int hpux_utssys(char *ubuf, int n, int t /* TODO: print a warning about this? */ down_write(&uts_sem); error = -EFAULT; - if (!copy_from_user(system_utsname.release, ubuf, len)) { - system_utsname.release[len] = 0; + if (!copy_from_user(vx_new_uts(release), ubuf, len)) { + vx_new_uts(release)[len] = 0; error = 0; } up_write(&uts_sem); @@ -422,13 +423,13 @@ int hpux_getdomainname(char *name, int l down_read(&uts_sem); - nlen = strlen(system_utsname.domainname) + 1; + nlen = strlen(vx_new_uts(domainname)) + 1; if (nlen < len) len = nlen; if(len > __NEW_UTS_LEN) goto done; - if(copy_to_user(name, system_utsname.domainname, len)) + if(copy_to_user(name, vx_new_uts(domainname), len)) goto done; err = 0; done: diff -NurpP --minimal linux-2.6.17.13/arch/parisc/kernel/module.c linux-2.6.17.13-g-v/arch/parisc/kernel/module.c --- linux-2.6.17.13/arch/parisc/kernel/module.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/parisc/kernel/module.c 2006-09-20 15:00:12.000000000 +0200 @@ -72,16 +72,38 @@ /* three functions to determine where in the module core * or init pieces the location is */ +static inline int is_init_rx(struct module *me, void *loc) +{ + return (loc >= me->module_init_rx && + loc < (me->module_init_rx + me->init_size_rx)); +} + +static inline int is_init_rw(struct module *me, void *loc) +{ + return (loc >= me->module_init_rw && + loc < (me->module_init_rw + me->init_size_rw)); +} + static inline int is_init(struct module *me, void *loc) { - return (loc >= me->module_init && - loc <= (me->module_init + me->init_size)); + return is_init_rx(me, loc) || is_init_rw(me, loc); +} + +static inline int is_core_rx(struct module *me, void *loc) +{ + return (loc >= me->module_core_rx && + loc < (me->module_core_rx + me->core_size_rx)); +} + +static inline int is_core_rw(struct module *me, void *loc) +{ + return (loc >= me->module_core_rw && + loc < (me->module_core_rw + me->core_size_rw)); } static inline int is_core(struct module *me, void *loc) { - return (loc >= me->module_core && - loc <= (me->module_core + me->core_size)); + return is_core_rx(me, loc) || is_core_rw(me, loc); } static inline int is_local(struct module *me, void *loc) @@ -289,21 +311,21 @@ int module_frob_arch_sections(CONST Elf_ } /* align things a bit */ - me->core_size = ALIGN(me->core_size, 16); - me->arch.got_offset = me->core_size; - me->core_size += gots * sizeof(struct got_entry); + me->core_size_rw = ALIGN(me->core_size_rw, 16); + me->arch.got_offset = me->core_size_rw; + me->core_size_rw += gots * sizeof(struct got_entry); - me->core_size = ALIGN(me->core_size, 16); - me->arch.fdesc_offset = me->core_size; - me->core_size += fdescs * sizeof(Elf_Fdesc); + me->core_size_rw = ALIGN(me->core_size_rw, 16); + me->arch.fdesc_offset = me->core_size_rw; + me->core_size_rw += fdescs * sizeof(Elf_Fdesc); - me->core_size = ALIGN(me->core_size, 16); - me->arch.stub_offset = me->core_size; - me->core_size += stubs * sizeof(struct stub_entry); + me->core_size_rx = ALIGN(me->core_size_rx, 16); + me->arch.stub_offset = me->core_size_rx; + me->core_size_rx += stubs * sizeof(struct stub_entry); - me->init_size = ALIGN(me->init_size, 16); - me->arch.init_stub_offset = me->init_size; - me->init_size += init_stubs * sizeof(struct stub_entry); + me->init_size_rx = ALIGN(me->init_size_rx, 16); + me->arch.init_stub_offset = me->init_size_rx; + me->init_size_rx += init_stubs * sizeof(struct stub_entry); me->arch.got_max = gots; me->arch.fdesc_max = fdescs; @@ -323,7 +345,7 @@ static Elf64_Word get_got(struct module BUG_ON(value == 0); - got = me->module_core + me->arch.got_offset; + got = me->module_core_rw + me->arch.got_offset; for (i = 0; got[i].addr; i++) if (got[i].addr == value) goto out; @@ -341,7 +363,7 @@ static Elf64_Word get_got(struct module #ifdef __LP64__ static Elf_Addr get_fdesc(struct module *me, unsigned long value) { - Elf_Fdesc *fdesc = me->module_core + me->arch.fdesc_offset; + Elf_Fdesc *fdesc = me->module_core_rw + me->arch.fdesc_offset; if (!value) { printk(KERN_ERR "%s: zero OPD requested!\n", me->name); @@ -359,7 +381,7 @@ static Elf_Addr get_fdesc(struct module /* Create new one */ fdesc->addr = value; - fdesc->gp = (Elf_Addr)me->module_core + me->arch.got_offset; + fdesc->gp = (Elf_Addr)me->module_core_rw + me->arch.got_offset; return (Elf_Addr)fdesc; } #endif /* __LP64__ */ @@ -373,12 +395,12 @@ static Elf_Addr get_stub(struct module * if(init_section) { i = me->arch.init_stub_count++; BUG_ON(me->arch.init_stub_count > me->arch.init_stub_max); - stub = me->module_init + me->arch.init_stub_offset + + stub = me->module_init_rx + me->arch.init_stub_offset + i * sizeof(struct stub_entry); } else { i = me->arch.stub_count++; BUG_ON(me->arch.stub_count > me->arch.stub_max); - stub = me->module_core + me->arch.stub_offset + + stub = me->module_core_rx + me->arch.stub_offset + i * sizeof(struct stub_entry); } @@ -721,7 +743,7 @@ register_unwind_table(struct module *me, table = (unsigned char *)sechdrs[me->arch.unwind_section].sh_addr; end = table + sechdrs[me->arch.unwind_section].sh_size; - gp = (Elf_Addr)me->module_core + me->arch.got_offset; + gp = (Elf_Addr)me->module_core_rw + me->arch.got_offset; DEBUGP("register_unwind_table(), sect = %d at 0x%p - 0x%p (gp=0x%lx)\n", me->arch.unwind_section, table, end, gp); diff -NurpP --minimal linux-2.6.17.13/arch/parisc/kernel/ptrace.c linux-2.6.17.13-g-v/arch/parisc/kernel/ptrace.c --- linux-2.6.17.13/arch/parisc/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/parisc/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -18,6 +18,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/arch/parisc/kernel/sys_parisc.c linux-2.6.17.13-g-v/arch/parisc/kernel/sys_parisc.c --- linux-2.6.17.13/arch/parisc/kernel/sys_parisc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/parisc/kernel/sys_parisc.c 2006-09-20 15:00:12.000000000 +0200 @@ -105,7 +105,7 @@ unsigned long arch_get_unmapped_area(str if (len > TASK_SIZE) return -ENOMEM; if (!addr) - addr = TASK_UNMAPPED_BASE; + addr = current->mm->mmap_base; if (filp) { addr = get_shared_area(filp->f_mapping, addr, len, pgoff); diff -NurpP --minimal linux-2.6.17.13/arch/parisc/kernel/sys_parisc32.c linux-2.6.17.13-g-v/arch/parisc/kernel/sys_parisc32.c --- linux-2.6.17.13/arch/parisc/kernel/sys_parisc32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/parisc/kernel/sys_parisc32.c 2006-09-20 15:00:12.000000000 +0200 @@ -599,6 +599,7 @@ asmlinkage int sys32_sysinfo(struct sysi do { seq = read_seqbegin(&xtime_lock); + /* FIXME: requires vx virtualization */ val.uptime = jiffies / HZ; val.loads[0] = avenrun[0] << (SI_LOAD_SHIFT - FSHIFT); diff -NurpP --minimal linux-2.6.17.13/arch/parisc/kernel/syscall_table.S linux-2.6.17.13-g-v/arch/parisc/kernel/syscall_table.S --- linux-2.6.17.13/arch/parisc/kernel/syscall_table.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/parisc/kernel/syscall_table.S 2006-09-20 15:00:12.000000000 +0200 @@ -368,7 +368,7 @@ ENTRY_COMP(mbind) /* 260 */ ENTRY_COMP(get_mempolicy) ENTRY_COMP(set_mempolicy) - ENTRY_SAME(ni_syscall) /* 263: reserved for vserver */ + ENTRY_DIFF(vserver) ENTRY_SAME(add_key) ENTRY_SAME(request_key) /* 265 */ ENTRY_SAME(keyctl) diff -NurpP --minimal linux-2.6.17.13/arch/parisc/kernel/traps.c linux-2.6.17.13-g-v/arch/parisc/kernel/traps.c --- linux-2.6.17.13/arch/parisc/kernel/traps.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/parisc/kernel/traps.c 2006-09-20 15:00:12.000000000 +0200 @@ -711,9 +711,7 @@ void handle_interruption(int code, struc down_read(&current->mm->mmap_sem); vma = find_vma(current->mm,regs->iaoq[0]); - if (vma && (regs->iaoq[0] >= vma->vm_start) - && (vma->vm_flags & VM_EXEC)) { - + if (vma && (regs->iaoq[0] >= vma->vm_start)) { fault_address = regs->iaoq[0]; fault_space = regs->iasq[0]; diff -NurpP --minimal linux-2.6.17.13/arch/parisc/mm/fault.c linux-2.6.17.13-g-v/arch/parisc/mm/fault.c --- linux-2.6.17.13/arch/parisc/mm/fault.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/parisc/mm/fault.c 2006-09-20 15:00:12.000000000 +0200 @@ -16,6 +16,8 @@ #include #include #include +#include +#include #include #include @@ -57,7 +59,7 @@ DEFINE_PER_CPU(struct exception_data, ex static unsigned long parisc_acctyp(unsigned long code, unsigned int inst) { - if (code == 6 || code == 16) + if (code == 6 || code == 7 || code == 16) return VM_EXEC; switch (inst & 0xf0000000) { @@ -143,6 +145,116 @@ parisc_acctyp(unsigned long code, unsign } #endif +#ifdef CONFIG_PAX_PAGEEXEC +/* + * PaX: decide what to do with offenders (instruction_pointer(regs) = fault address) + * + * returns 1 when task should be killed + * 2 when rt_sigreturn trampoline was detected + * 3 when unpatched PLT trampoline was detected + */ +static int pax_handle_fetch_fault(struct pt_regs *regs) +{ + +#ifdef CONFIG_PAX_EMUPLT + int err; + + do { /* PaX: unpatched PLT emulation */ + unsigned int bl, depwi; + + err = get_user(bl, (unsigned int*)instruction_pointer(regs)); + err |= get_user(depwi, (unsigned int*)(instruction_pointer(regs)+4)); + + if (err) + break; + + if (bl == 0xEA9F1FDDU && depwi == 0xD6801C1EU) { + unsigned int ldw, bv, ldw2, addr = instruction_pointer(regs)-12; + + err = get_user(ldw, (unsigned int*)addr); + err |= get_user(bv, (unsigned int*)(addr+4)); + err |= get_user(ldw2, (unsigned int*)(addr+8)); + + if (err) + break; + + if (ldw == 0x0E801096U && + bv == 0xEAC0C000U && + ldw2 == 0x0E881095U) + { + unsigned int resolver, map; + + err = get_user(resolver, (unsigned int*)(instruction_pointer(regs)+8)); + err |= get_user(map, (unsigned int*)(instruction_pointer(regs)+12)); + if (err) + break; + + regs->gr[20] = instruction_pointer(regs)+8; + regs->gr[21] = map; + regs->gr[22] = resolver; + regs->iaoq[0] = resolver | 3UL; + regs->iaoq[1] = regs->iaoq[0] + 4; + return 3; + } + } + } while (0); +#endif + +#ifdef CONFIG_PAX_EMUTRAMP + +#ifndef CONFIG_PAX_EMUSIGRT + if (!(current->mm->pax_flags & MF_PAX_EMUTRAMP)) + return 1; +#endif + + do { /* PaX: rt_sigreturn emulation */ + unsigned int ldi1, ldi2, bel, nop; + + err = get_user(ldi1, (unsigned int *)instruction_pointer(regs)); + err |= get_user(ldi2, (unsigned int *)(instruction_pointer(regs)+4)); + err |= get_user(bel, (unsigned int *)(instruction_pointer(regs)+8)); + err |= get_user(nop, (unsigned int *)(instruction_pointer(regs)+12)); + + if (err) + break; + + if ((ldi1 == 0x34190000U || ldi1 == 0x34190002U) && + ldi2 == 0x3414015AU && + bel == 0xE4008200U && + nop == 0x08000240U) + { + regs->gr[25] = (ldi1 & 2) >> 1; + regs->gr[20] = __NR_rt_sigreturn; + regs->gr[31] = regs->iaoq[1] + 16; + regs->sr[0] = regs->iasq[1]; + regs->iaoq[0] = 0x100UL; + regs->iaoq[1] = regs->iaoq[0] + 4; + regs->iasq[0] = regs->sr[2]; + regs->iasq[1] = regs->sr[2]; + return 2; + } + } while (0); +#endif + + return 1; +} + +void pax_report_insns(void *pc, void *sp) +{ + unsigned long i; + + printk(KERN_ERR "PAX: bytes at PC: "); + for (i = 0; i < 5; i++) { + unsigned int c; + if (get_user(c, (unsigned int*)pc+i)) + printk("???????? "); + else + printk("%08x ", c); + } + printk("\n"); +} +#endif + void do_page_fault(struct pt_regs *regs, unsigned long code, unsigned long address) { @@ -168,8 +280,33 @@ good_area: acc_type = parisc_acctyp(code,regs->iir); - if ((vma->vm_flags & acc_type) != acc_type) + if ((vma->vm_flags & acc_type) != acc_type) { + +#ifdef CONFIG_PAX_PAGEEXEC + if ((mm->pax_flags & MF_PAX_PAGEEXEC) && (acc_type & VM_EXEC) && + (address & ~3UL) == instruction_pointer(regs)) + { + up_read(&mm->mmap_sem); + switch(pax_handle_fetch_fault(regs)) { + +#ifdef CONFIG_PAX_EMUPLT + case 3: + return; +#endif + +#ifdef CONFIG_PAX_EMUTRAMP + case 2: + return; +#endif + + } + pax_report_fault(regs, (void*)instruction_pointer(regs), (void*)regs->gr[30]); + do_exit(SIGKILL); + } +#endif + goto bad_area; + } /* * If for any reason at all we couldn't handle the fault, make diff -NurpP --minimal linux-2.6.17.13/arch/powerpc/Kconfig linux-2.6.17.13-g-v/arch/powerpc/Kconfig --- linux-2.6.17.13/arch/powerpc/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/powerpc/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -1018,6 +1018,8 @@ endmenu source "arch/powerpc/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" config KEYS_COMPAT diff -NurpP --minimal linux-2.6.17.13/arch/powerpc/kernel/module_32.c linux-2.6.17.13-g-v/arch/powerpc/kernel/module_32.c --- linux-2.6.17.13/arch/powerpc/kernel/module_32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/powerpc/kernel/module_32.c 2006-09-20 15:00:12.000000000 +0200 @@ -123,7 +123,7 @@ int module_frob_arch_sections(Elf32_Ehdr me->arch.core_plt_section = i; } if (!me->arch.core_plt_section || !me->arch.init_plt_section) { - printk("Module doesn't contain .plt or .init.plt sections.\n"); + printk("Module %s doesn't contain .plt or .init.plt sections.\n", me->name); return -ENOEXEC; } @@ -164,11 +164,16 @@ static uint32_t do_plt_call(void *locati DEBUGP("Doing plt for call to 0x%x at 0x%x\n", val, (unsigned int)location); /* Init, or core PLT? */ - if (location >= mod->module_core - && location < mod->module_core + mod->core_size) + if ((location >= mod->module_core_rx && location < mod->module_core_rx + mod->core_size_rx) || + (location >= mod->module_core_rw && location < mod->module_core_rw + mod->core_size_rw)) entry = (void *)sechdrs[mod->arch.core_plt_section].sh_addr; - else + else if ((location >= mod->module_init_rx && location < mod->module_init_rx + mod->init_size_rx) || + (location >= mod->module_init_rw && location < mod->module_init_rw + mod->init_size_rw)) entry = (void *)sechdrs[mod->arch.init_plt_section].sh_addr; + else { + printk(KERN_ERR "%s: invalid R_PPC_REL24 entry found\n", mod->name); + return ~0UL; + } /* Find this entry, or if that fails, the next avail. entry */ while (entry->jump[0]) { diff -NurpP --minimal linux-2.6.17.13/arch/powerpc/kernel/process.c linux-2.6.17.13-g-v/arch/powerpc/kernel/process.c --- linux-2.6.17.13/arch/powerpc/kernel/process.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/powerpc/kernel/process.c 2006-09-20 15:00:12.000000000 +0200 @@ -432,8 +432,9 @@ void show_regs(struct pt_regs * regs) trap = TRAP(regs); if (trap == 0x300 || trap == 0x600) printk("DAR: "REG", DSISR: "REG"\n", regs->dar, regs->dsisr); - printk("TASK = %p[%d] '%s' THREAD: %p", - current, current->pid, current->comm, task_thread_info(current)); + printk("TASK = %p[%d,#%u] '%s' THREAD: %p", + current, current->pid, current->xid, + current->comm, task_thread_info(current)); #ifdef CONFIG_SMP printk(" CPU: %d", smp_processor_id()); diff -NurpP --minimal linux-2.6.17.13/arch/powerpc/kernel/syscalls.c linux-2.6.17.13-g-v/arch/powerpc/kernel/syscalls.c --- linux-2.6.17.13/arch/powerpc/kernel/syscalls.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/powerpc/kernel/syscalls.c 2006-09-20 15:00:12.000000000 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include #include @@ -260,7 +261,7 @@ long ppc_newuname(struct new_utsname __u int err = 0; down_read(&uts_sem); - if (copy_to_user(name, &system_utsname, sizeof(*name))) + if (copy_to_user(name, vx_new_utsname(), sizeof(*name))) err = -EFAULT; up_read(&uts_sem); if (!err) @@ -273,7 +274,7 @@ int sys_uname(struct old_utsname __user int err = 0; down_read(&uts_sem); - if (copy_to_user(name, &system_utsname, sizeof(*name))) + if (copy_to_user(name, vx_new_utsname(), sizeof(*name))) err = -EFAULT; up_read(&uts_sem); if (!err) @@ -284,25 +285,22 @@ int sys_uname(struct old_utsname __user int sys_olduname(struct oldold_utsname __user *name) { int error; + struct new_utsname *ptr; if (!access_ok(VERIFY_WRITE, name, sizeof(struct oldold_utsname))) return -EFAULT; down_read(&uts_sem); - error = __copy_to_user(&name->sysname, &system_utsname.sysname, - __OLD_UTS_LEN); + ptr = vx_new_utsname(); + error = __copy_to_user(&name->sysname, ptr->sysname, __OLD_UTS_LEN); error |= __put_user(0, name->sysname + __OLD_UTS_LEN); - error |= __copy_to_user(&name->nodename, &system_utsname.nodename, - __OLD_UTS_LEN); + error |= __copy_to_user(&name->nodename, ptr->nodename, __OLD_UTS_LEN); error |= __put_user(0, name->nodename + __OLD_UTS_LEN); - error |= __copy_to_user(&name->release, &system_utsname.release, - __OLD_UTS_LEN); + error |= __copy_to_user(&name->release, ptr->release, __OLD_UTS_LEN); error |= __put_user(0, name->release + __OLD_UTS_LEN); - error |= __copy_to_user(&name->version, &system_utsname.version, - __OLD_UTS_LEN); + error |= __copy_to_user(&name->version, ptr->version, __OLD_UTS_LEN); error |= __put_user(0, name->version + __OLD_UTS_LEN); - error |= __copy_to_user(&name->machine, &system_utsname.machine, - __OLD_UTS_LEN); + error |= __copy_to_user(&name->machine, ptr->machine, __OLD_UTS_LEN); error |= override_machine(name->machine); up_read(&uts_sem); diff -NurpP --minimal linux-2.6.17.13/arch/powerpc/kernel/systbl.S linux-2.6.17.13-g-v/arch/powerpc/kernel/systbl.S --- linux-2.6.17.13/arch/powerpc/kernel/systbl.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/powerpc/kernel/systbl.S 2006-09-20 15:00:12.000000000 +0200 @@ -296,7 +296,7 @@ COMPAT_SYS(fstatfs64) SYSX(sys_ni_syscall, ppc_fadvise64_64, ppc_fadvise64_64) PPC_SYS(rtas) OLDSYS(debug_setcontext) -SYSCALL(ni_syscall) +SYSX(sys_vserver, sys32_vserver, sys_vserver) SYSCALL(ni_syscall) COMPAT_SYS(mbind) COMPAT_SYS(get_mempolicy) diff -NurpP --minimal linux-2.6.17.13/arch/powerpc/kernel/vdso.c linux-2.6.17.13-g-v/arch/powerpc/kernel/vdso.c --- linux-2.6.17.13/arch/powerpc/kernel/vdso.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/powerpc/kernel/vdso.c 2006-09-20 15:00:12.000000000 +0200 @@ -23,6 +23,7 @@ #include #include #include +#include #include #include @@ -293,7 +294,7 @@ int arch_setup_additional_pages(struct l kmem_cache_free(vm_area_cachep, vma); return -ENOMEM; } - mm->total_vm += (vma->vm_end - vma->vm_start) >> PAGE_SHIFT; + vx_vmpages_add(mm, (vma->vm_end - vma->vm_start) >> PAGE_SHIFT); up_write(&mm->mmap_sem); return 0; diff -NurpP --minimal linux-2.6.17.13/arch/powerpc/mm/fault.c linux-2.6.17.13-g-v/arch/powerpc/mm/fault.c --- linux-2.6.17.13/arch/powerpc/mm/fault.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/powerpc/mm/fault.c 2006-09-20 15:00:12.000000000 +0200 @@ -29,6 +29,7 @@ #include #include #include +#include #include #include @@ -103,6 +104,38 @@ static void do_dabr(struct pt_regs *regs } #endif /* !(CONFIG_4xx || CONFIG_BOOKE)*/ +#ifdef CONFIG_PAX_PAGEEXEC +/* + * PaX: decide what to do with offenders (regs->nip = fault address) + * + * returns 1 when task should be killed + */ +static int pax_handle_fetch_fault(struct pt_regs *regs) +{ + +#if defined(CONFIG_PAX_EMUPLT) || defined(CONFIG_PAX_EMUSIGRT) + int err; +#endif + + return 1; +} + +void pax_report_insns(void *pc, void *sp) +{ + unsigned long i; + + printk(KERN_ERR "PAX: bytes at PC: "); + for (i = 0; i < 5; i++) { + unsigned int c; + if (get_user(c, (unsigned int*)pc+i)) + printk("???????? "); + else + printk("%08x ", c); + } + printk("\n"); +} +#endif + /* * For 600- and 800-family processors, the error_code parameter is DSISR * for a data fault, SRR1 for an instruction fault. For 400-family processors @@ -335,6 +368,19 @@ bad_area: bad_area_nosemaphore: /* User mode accesses cause a SIGSEGV */ if (user_mode(regs)) { + +#ifdef CONFIG_PAX_PAGEEXEC + if (mm->pax_flags & MF_PAX_PAGEEXEC) { + if (is_exec && (error_code & DSISR_PROTFAULT)) { + switch (pax_handle_fetch_fault(regs)) { + } + + pax_report_fault(regs, (void*)regs->nip, (void*)regs->gpr[1]); + do_exit(SIGKILL); + } + } +#endif + _exception(SIGSEGV, regs, code, address); return 0; } diff -NurpP --minimal linux-2.6.17.13/arch/powerpc/mm/mmap.c linux-2.6.17.13-g-v/arch/powerpc/mm/mmap.c --- linux-2.6.17.13/arch/powerpc/mm/mmap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/powerpc/mm/mmap.c 2006-09-20 15:00:12.000000000 +0200 @@ -74,10 +74,22 @@ void arch_pick_mmap_layout(struct mm_str */ if (mmap_is_legacy()) { mm->mmap_base = TASK_UNMAPPED_BASE; + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base += mm->delta_mmap; +#endif + mm->get_unmapped_area = arch_get_unmapped_area; mm->unmap_area = arch_unmap_area; } else { mm->mmap_base = mmap_base(); + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base -= mm->delta_mmap; +#endif + mm->get_unmapped_area = arch_get_unmapped_area_topdown; mm->unmap_area = arch_unmap_area_topdown; } diff -NurpP --minimal linux-2.6.17.13/arch/ppc/Kconfig linux-2.6.17.13-g-v/arch/ppc/Kconfig --- linux-2.6.17.13/arch/ppc/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ppc/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -1414,6 +1414,8 @@ source "arch/powerpc/oprofile/Kconfig" source "arch/ppc/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/ppc/mm/fault.c linux-2.6.17.13-g-v/arch/ppc/mm/fault.c --- linux-2.6.17.13/arch/ppc/mm/fault.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/ppc/mm/fault.c 2006-09-20 15:00:12.000000000 +0200 @@ -26,6 +26,11 @@ #include #include #include +#include +#include +#include +#include +#include #include #include @@ -49,6 +54,364 @@ unsigned long pte_misses; /* updated by unsigned long pte_errors; /* updated by do_page_fault() */ unsigned int probingmem; +#ifdef CONFIG_PAX_EMUSIGRT +void pax_syscall_close(struct vm_area_struct * vma) +{ + vma->vm_mm->call_syscall = 0UL; +} + +static struct page* pax_syscall_nopage(struct vm_area_struct *vma, unsigned long address, int *type) +{ + struct page* page; + unsigned int *kaddr; + + page = alloc_page(GFP_HIGHUSER); + if (!page) + return NOPAGE_OOM; + + kaddr = kmap(page); + memset(kaddr, 0, PAGE_SIZE); + kaddr[0] = 0x44000002U; /* sc */ + __flush_dcache_icache(kaddr); + kunmap(page); + if (type) + *type = VM_FAULT_MAJOR; + return page; +} + +static struct vm_operations_struct pax_vm_ops = { + .close = pax_syscall_close, + .nopage = pax_syscall_nopage, +}; + +static int pax_insert_vma(struct vm_area_struct *vma, unsigned long addr) +{ + int ret; + + memset(vma, 0, sizeof(*vma)); + vma->vm_mm = current->mm; + vma->vm_start = addr; + vma->vm_end = addr + PAGE_SIZE; + vma->vm_flags = VM_READ | VM_EXEC | VM_MAYREAD | VM_MAYEXEC; + vma->vm_page_prot = protection_map[vma->vm_flags & 0x0f]; + vma->vm_ops = &pax_vm_ops; + + ret = insert_vm_struct(current->mm, vma); + if (ret) + return ret; + + ++current->mm->total_vm; + return 0; +} +#endif + +#ifdef CONFIG_PAX_PAGEEXEC +/* + * PaX: decide what to do with offenders (regs->nip = fault address) + * + * returns 1 when task should be killed + * 2 when patched GOT trampoline was detected + * 3 when patched PLT trampoline was detected + * 4 when unpatched PLT trampoline was detected + * 5 when sigreturn trampoline was detected + * 7 when rt_sigreturn trampoline was detected + */ +static int pax_handle_fetch_fault(struct pt_regs *regs) +{ + +#if defined(CONFIG_PAX_EMUPLT) || defined(CONFIG_PAX_EMUSIGRT) + int err; +#endif + +#ifdef CONFIG_PAX_EMUPLT + do { /* PaX: patched GOT emulation */ + unsigned int blrl; + + err = get_user(blrl, (unsigned int*)regs->nip); + + if (!err && blrl == 0x4E800021U) { + unsigned long temp = regs->nip; + + regs->nip = regs->link & 0xFFFFFFFCUL; + regs->link = temp + 4UL; + return 2; + } + } while (0); + + do { /* PaX: patched PLT emulation #1 */ + unsigned int b; + + err = get_user(b, (unsigned int *)regs->nip); + + if (!err && (b & 0xFC000003U) == 0x48000000U) { + regs->nip += (((b | 0xFC000000UL) ^ 0x02000000UL) + 0x02000000UL); + return 3; + } + } while (0); + + do { /* PaX: unpatched PLT emulation #1 */ + unsigned int li, b; + + err = get_user(li, (unsigned int *)regs->nip); + err |= get_user(b, (unsigned int *)(regs->nip+4)); + + if (!err && (li & 0xFFFF0000U) == 0x39600000U && (b & 0xFC000003U) == 0x48000000U) { + unsigned int rlwinm, add, li2, addis2, mtctr, li3, addis3, bctr; + unsigned long addr = b | 0xFC000000UL; + + addr = regs->nip + 4 + ((addr ^ 0x02000000UL) + 0x02000000UL); + err = get_user(rlwinm, (unsigned int*)addr); + err |= get_user(add, (unsigned int*)(addr+4)); + err |= get_user(li2, (unsigned int*)(addr+8)); + err |= get_user(addis2, (unsigned int*)(addr+12)); + err |= get_user(mtctr, (unsigned int*)(addr+16)); + err |= get_user(li3, (unsigned int*)(addr+20)); + err |= get_user(addis3, (unsigned int*)(addr+24)); + err |= get_user(bctr, (unsigned int*)(addr+28)); + + if (err) + break; + + if (rlwinm == 0x556C083CU && + add == 0x7D6C5A14U && + (li2 & 0xFFFF0000U) == 0x39800000U && + (addis2 & 0xFFFF0000U) == 0x3D8C0000U && + mtctr == 0x7D8903A6U && + (li3 & 0xFFFF0000U) == 0x39800000U && + (addis3 & 0xFFFF0000U) == 0x3D8C0000U && + bctr == 0x4E800420U) + { + regs->gpr[PT_R11] = 3 * (((li | 0xFFFF0000UL) ^ 0x00008000UL) + 0x00008000UL); + regs->gpr[PT_R12] = (((li3 | 0xFFFF0000UL) ^ 0x00008000UL) + 0x00008000UL); + regs->gpr[PT_R12] += (addis3 & 0xFFFFU) << 16; + regs->ctr = (((li2 | 0xFFFF0000UL) ^ 0x00008000UL) + 0x00008000UL); + regs->ctr += (addis2 & 0xFFFFU) << 16; + regs->nip = regs->ctr; + return 4; + } + } + } while (0); + +#if 0 + do { /* PaX: unpatched PLT emulation #2 */ + unsigned int lis, lwzu, b, bctr; + + err = get_user(lis, (unsigned int *)regs->nip); + err |= get_user(lwzu, (unsigned int *)(regs->nip+4)); + err |= get_user(b, (unsigned int *)(regs->nip+8)); + err |= get_user(bctr, (unsigned int *)(regs->nip+12)); + + if (err) + break; + + if ((lis & 0xFFFF0000U) == 0x39600000U && + (lwzu & 0xU) == 0xU && + (b & 0xFC000003U) == 0x48000000U && + bctr == 0x4E800420U) + { + unsigned int addis, addi, rlwinm, add, li2, addis2, mtctr, li3, addis3, bctr; + unsigned long addr = b | 0xFC000000UL; + + addr = regs->nip + 12 + ((addr ^ 0x02000000UL) + 0x02000000UL); + err = get_user(addis, (unsigned int*)addr); + err |= get_user(addi, (unsigned int*)(addr+4)); + err |= get_user(rlwinm, (unsigned int*)(addr+8)); + err |= get_user(add, (unsigned int*)(addr+12)); + err |= get_user(li2, (unsigned int*)(addr+16)); + err |= get_user(addis2, (unsigned int*)(addr+20)); + err |= get_user(mtctr, (unsigned int*)(addr+24)); + err |= get_user(li3, (unsigned int*)(addr+28)); + err |= get_user(addis3, (unsigned int*)(addr+32)); + err |= get_user(bctr, (unsigned int*)(addr+36)); + + if (err) + break; + + if ((addis & 0xFFFF0000U) == 0x3D6B0000U && + (addi & 0xFFFF0000U) == 0x396B0000U && + rlwinm == 0x556C083CU && + add == 0x7D6C5A14U && + (li2 & 0xFFFF0000U) == 0x39800000U && + (addis2 & 0xFFFF0000U) == 0x3D8C0000U && + mtctr == 0x7D8903A6U && + (li3 & 0xFFFF0000U) == 0x39800000U && + (addis3 & 0xFFFF0000U) == 0x3D8C0000U && + bctr == 0x4E800420U) + { + regs->gpr[PT_R11] = + regs->gpr[PT_R11] = 3 * (((li | 0xFFFF0000UL) ^ 0x00008000UL) + 0x00008000UL); + regs->gpr[PT_R12] = (((li3 | 0xFFFF0000UL) ^ 0x00008000UL) + 0x00008000UL); + regs->gpr[PT_R12] += (addis3 & 0xFFFFU) << 16; + regs->ctr = (((li2 | 0xFFFF0000UL) ^ 0x00008000UL) + 0x00008000UL); + regs->ctr += (addis2 & 0xFFFFU) << 16; + regs->nip = regs->ctr; + return 4; + } + } + } while (0); +#endif + + do { /* PaX: unpatched PLT emulation #3 */ + unsigned int li, b; + + err = get_user(li, (unsigned int *)regs->nip); + err |= get_user(b, (unsigned int *)(regs->nip+4)); + + if (!err && (li & 0xFFFF0000U) == 0x39600000U && (b & 0xFC000003U) == 0x48000000U) { + unsigned int addis, lwz, mtctr, bctr; + unsigned long addr = b | 0xFC000000UL; + + addr = regs->nip + 4 + ((addr ^ 0x02000000UL) + 0x02000000UL); + err = get_user(addis, (unsigned int*)addr); + err |= get_user(lwz, (unsigned int*)(addr+4)); + err |= get_user(mtctr, (unsigned int*)(addr+8)); + err |= get_user(bctr, (unsigned int*)(addr+12)); + + if (err) + break; + + if ((addis & 0xFFFF0000U) == 0x3D6B0000U && + (lwz & 0xFFFF0000U) == 0x816B0000U && + mtctr == 0x7D6903A6U && + bctr == 0x4E800420U) + { + unsigned int r11; + + addr = (addis << 16) + (((li | 0xFFFF0000UL) ^ 0x00008000UL) + 0x00008000UL); + addr += (((lwz | 0xFFFF0000UL) ^ 0x00008000UL) + 0x00008000UL); + + err = get_user(r11, (unsigned int*)addr); + if (err) + break; + + regs->gpr[PT_R11] = r11; + regs->ctr = r11; + regs->nip = r11; + return 4; + } + } + } while (0); +#endif + +#ifdef CONFIG_PAX_EMUSIGRT + do { /* PaX: sigreturn emulation */ + unsigned int li, sc; + + err = get_user(li, (unsigned int *)regs->nip); + err |= get_user(sc, (unsigned int *)(regs->nip+4)); + + if (!err && li == 0x38000000U + __NR_sigreturn && sc == 0x44000002U) { + struct vm_area_struct *vma; + unsigned long call_syscall; + + down_read(&current->mm->mmap_sem); + call_syscall = current->mm->call_syscall; + up_read(&current->mm->mmap_sem); + if (likely(call_syscall)) + goto emulate; + + vma = kmem_cache_alloc(vm_area_cachep, SLAB_KERNEL); + + down_write(&current->mm->mmap_sem); + if (current->mm->call_syscall) { + call_syscall = current->mm->call_syscall; + up_write(&current->mm->mmap_sem); + if (vma) kmem_cache_free(vm_area_cachep, vma); + goto emulate; + } + + call_syscall = get_unmapped_area(NULL, 0UL, PAGE_SIZE, 0UL, MAP_PRIVATE); + if (!vma || (call_syscall & ~PAGE_MASK)) { + up_write(&current->mm->mmap_sem); + if (vma) kmem_cache_free(vm_area_cachep, vma); + return 1; + } + + if (pax_insert_vma(vma, call_syscall)) { + up_write(&current->mm->mmap_sem); + kmem_cache_free(vm_area_cachep, vma); + return 1; + } + + current->mm->call_syscall = call_syscall; + up_write(&current->mm->mmap_sem); + +emulate: + regs->gpr[PT_R0] = __NR_sigreturn; + regs->nip = call_syscall; + return 5; + } + } while (0); + + do { /* PaX: rt_sigreturn emulation */ + unsigned int li, sc; + + err = get_user(li, (unsigned int *)regs->nip); + err |= get_user(sc, (unsigned int *)(regs->nip+4)); + + if (!err && li == 0x38000000U + __NR_rt_sigreturn && sc == 0x44000002U) { + struct vm_area_struct *vma; + unsigned int call_syscall; + + down_read(&current->mm->mmap_sem); + call_syscall = current->mm->call_syscall; + up_read(&current->mm->mmap_sem); + if (likely(call_syscall)) + goto rt_emulate; + + vma = kmem_cache_alloc(vm_area_cachep, SLAB_KERNEL); + + down_write(&current->mm->mmap_sem); + if (current->mm->call_syscall) { + call_syscall = current->mm->call_syscall; + up_write(&current->mm->mmap_sem); + if (vma) kmem_cache_free(vm_area_cachep, vma); + goto rt_emulate; + } + + call_syscall = get_unmapped_area(NULL, 0UL, PAGE_SIZE, 0UL, MAP_PRIVATE); + if (!vma || (call_syscall & ~PAGE_MASK)) { + up_write(&current->mm->mmap_sem); + if (vma) kmem_cache_free(vm_area_cachep, vma); + return 1; + } + + if (pax_insert_vma(vma, call_syscall)) { + up_write(&current->mm->mmap_sem); + kmem_cache_free(vm_area_cachep, vma); + return 1; + } + + current->mm->call_syscall = call_syscall; + up_write(&current->mm->mmap_sem); + +rt_emulate: + regs->gpr[PT_R0] = __NR_rt_sigreturn; + regs->nip = call_syscall; + return 6; + } + } while (0); +#endif + + return 1; +} + +void pax_report_insns(void *pc, void *sp) +{ + unsigned long i; + + printk(KERN_ERR "PAX: bytes at PC: "); + for (i = 0; i < 5; i++) { + unsigned int c; + if (get_user(c, (unsigned int*)pc+i)) + printk("???????? "); + else + printk("%08x ", c); + } + printk("\n"); +} +#endif + /* * Check whether the instruction at regs->nip is a store using * an update addressing form which will update r1. @@ -109,7 +472,7 @@ int do_page_fault(struct pt_regs *regs, * indicate errors in DSISR but can validly be set in SRR1. */ if (TRAP(regs) == 0x400) - error_code &= 0x48200000; + error_code &= 0x58200000; else is_write = error_code & 0x02000000; #endif /* CONFIG_4xx || CONFIG_BOOKE */ @@ -204,15 +567,14 @@ good_area: pte_t *ptep; pmd_t *pmdp; -#if 0 +#if 1 /* It would be nice to actually enforce the VM execute permission on CPUs which can do so, but far too much stuff in userspace doesn't get the permissions right, so we let any page be executed for now. */ if (! (vma->vm_flags & VM_EXEC)) goto bad_area; -#endif - +#else /* Since 4xx/Book-E supports per-page execute permission, * we lazily flush dcache to icache. */ ptep = NULL; @@ -235,6 +597,7 @@ good_area: pte_unmap_unlock(ptep, ptl); } #endif +#endif /* a read */ } else { /* protection fault */ @@ -280,6 +643,33 @@ bad_area: /* User mode accesses cause a SIGSEGV */ if (user_mode(regs)) { + +#ifdef CONFIG_PAX_PAGEEXEC + if (mm->pax_flags & MF_PAX_PAGEEXEC) { + if ((TRAP(regs) == 0x400) && (regs->nip == address)) { + switch (pax_handle_fetch_fault(regs)) { + +#ifdef CONFIG_PAX_EMUPLT + case 2: + case 3: + case 4: + return 0; +#endif + +#ifdef CONFIG_PAX_EMUSIGRT + case 5: + case 6: + return 0; +#endif + + } + + pax_report_fault(regs, (void*)regs->nip, (void*)regs->gpr[1]); + do_exit(SIGKILL); + } + } +#endif + _exception(SIGSEGV, regs, code, address); return 0; } diff -NurpP --minimal linux-2.6.17.13/arch/s390/Kconfig linux-2.6.17.13-g-v/arch/s390/Kconfig --- linux-2.6.17.13/arch/s390/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/s390/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -478,6 +478,8 @@ source "arch/s390/oprofile/Kconfig" source "arch/s390/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/s390/kernel/module.c linux-2.6.17.13-g-v/arch/s390/kernel/module.c --- linux-2.6.17.13/arch/s390/kernel/module.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/s390/kernel/module.c 2006-09-20 15:00:12.000000000 +0200 @@ -164,11 +164,11 @@ module_frob_arch_sections(Elf_Ehdr *hdr, /* Increase core size by size of got & plt and set start offsets for got and plt. */ - me->core_size = ALIGN(me->core_size, 4); - me->arch.got_offset = me->core_size; - me->core_size += me->arch.got_size; - me->arch.plt_offset = me->core_size; - me->core_size += me->arch.plt_size; + me->core_size_rw = ALIGN(me->core_size_rw, 4); + me->arch.got_offset = me->core_size_rw; + me->core_size_rw += me->arch.got_size; + me->arch.plt_offset = me->core_size_rx; + me->core_size_rx += me->arch.plt_size; return 0; } @@ -254,7 +254,7 @@ apply_rela(Elf_Rela *rela, Elf_Addr base if (info->got_initialized == 0) { Elf_Addr *gotent; - gotent = me->module_core + me->arch.got_offset + + gotent = me->module_core_rw + me->arch.got_offset + info->got_offset; *gotent = val; info->got_initialized = 1; @@ -278,7 +278,7 @@ apply_rela(Elf_Rela *rela, Elf_Addr base else if (r_type == R_390_GOTENT || r_type == R_390_GOTPLTENT) *(unsigned int *) loc = - (val + (Elf_Addr) me->module_core - loc) >> 1; + (val + (Elf_Addr) me->module_core_rw - loc) >> 1; else if (r_type == R_390_GOT64 || r_type == R_390_GOTPLT64) *(unsigned long *) loc = val; @@ -292,7 +292,7 @@ apply_rela(Elf_Rela *rela, Elf_Addr base case R_390_PLTOFF64: /* 16 bit offset from GOT to PLT. */ if (info->plt_initialized == 0) { unsigned int *ip; - ip = me->module_core + me->arch.plt_offset + + ip = me->module_core_rx + me->arch.plt_offset + info->plt_offset; #ifndef CONFIG_64BIT ip[0] = 0x0d105810; /* basr 1,0; l 1,6(1); br 1 */ @@ -314,7 +314,7 @@ apply_rela(Elf_Rela *rela, Elf_Addr base val = me->arch.plt_offset - me->arch.got_offset + info->plt_offset + rela->r_addend; else - val = (Elf_Addr) me->module_core + + val = (Elf_Addr) me->module_core_rx + me->arch.plt_offset + info->plt_offset + rela->r_addend - loc; if (r_type == R_390_PLT16DBL) @@ -334,7 +334,7 @@ apply_rela(Elf_Rela *rela, Elf_Addr base case R_390_GOTOFF32: /* 32 bit offset to GOT. */ case R_390_GOTOFF64: /* 64 bit offset to GOT. */ val = val + rela->r_addend - - ((Elf_Addr) me->module_core + me->arch.got_offset); + ((Elf_Addr) me->module_core_rw + me->arch.got_offset); if (r_type == R_390_GOTOFF16) *(unsigned short *) loc = val; else if (r_type == R_390_GOTOFF32) @@ -344,7 +344,7 @@ apply_rela(Elf_Rela *rela, Elf_Addr base break; case R_390_GOTPC: /* 32 bit PC relative offset to GOT. */ case R_390_GOTPCDBL: /* 32 bit PC rel. off. to GOT shifted by 1. */ - val = (Elf_Addr) me->module_core + me->arch.got_offset + + val = (Elf_Addr) me->module_core_rw + me->arch.got_offset + rela->r_addend - loc; if (r_type == R_390_GOTPC) *(unsigned int *) loc = val; diff -NurpP --minimal linux-2.6.17.13/arch/s390/kernel/process.c linux-2.6.17.13-g-v/arch/s390/kernel/process.c --- linux-2.6.17.13/arch/s390/kernel/process.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/s390/kernel/process.c 2006-09-20 15:00:12.000000000 +0200 @@ -165,9 +165,9 @@ void show_regs(struct pt_regs *regs) struct task_struct *tsk = current; printk("CPU: %d %s\n", task_thread_info(tsk)->cpu, print_tainted()); - printk("Process %s (pid: %d, task: %p, ksp: %p)\n", - current->comm, current->pid, (void *) tsk, - (void *) tsk->thread.ksp); + printk("Process %s (pid: %d[#%u], task: %p, ksp: %p)\n", + current->comm, current->pid, current->xid, + (void *) tsk, (void *) tsk->thread.ksp); show_registers(regs); /* Show stack backtrace if pt_regs is from kernel mode */ diff -NurpP --minimal linux-2.6.17.13/arch/s390/kernel/ptrace.c linux-2.6.17.13-g-v/arch/s390/kernel/ptrace.c --- linux-2.6.17.13/arch/s390/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/s390/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -723,7 +723,13 @@ sys_ptrace(long request, long pid, long goto out; } + if (!vx_check(vx_task_xid(child), VX_WATCH|VX_IDENT)) { + ret = -EPERM; + goto out_tsk; + } + ret = do_ptrace(child, request, addr, data); +out_tsk: put_task_struct(child); out: unlock_kernel(); diff -NurpP --minimal linux-2.6.17.13/arch/s390/kernel/syscalls.S linux-2.6.17.13-g-v/arch/s390/kernel/syscalls.S --- linux-2.6.17.13/arch/s390/kernel/syscalls.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/s390/kernel/syscalls.S 2006-09-20 15:00:12.000000000 +0200 @@ -271,7 +271,7 @@ SYSCALL(sys_clock_settime,sys_clock_sett SYSCALL(sys_clock_gettime,sys_clock_gettime,sys32_clock_gettime_wrapper) /* 260 */ SYSCALL(sys_clock_getres,sys_clock_getres,sys32_clock_getres_wrapper) SYSCALL(sys_clock_nanosleep,sys_clock_nanosleep,sys32_clock_nanosleep_wrapper) -NI_SYSCALL /* reserved for vserver */ +SYSCALL(sys_vserver,sys_vserver,sys32_vserver) SYSCALL(s390_fadvise64_64,sys_ni_syscall,sys32_fadvise64_64_wrapper) SYSCALL(sys_statfs64,sys_statfs64,compat_sys_statfs64_wrapper) SYSCALL(sys_fstatfs64,sys_fstatfs64,compat_sys_fstatfs64_wrapper) diff -NurpP --minimal linux-2.6.17.13/arch/sh/Kconfig linux-2.6.17.13-g-v/arch/sh/Kconfig --- linux-2.6.17.13/arch/sh/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sh/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -646,6 +646,8 @@ source "arch/sh/oprofile/Kconfig" source "arch/sh/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/sh/kernel/kgdb_stub.c linux-2.6.17.13-g-v/arch/sh/kernel/kgdb_stub.c --- linux-2.6.17.13/arch/sh/kernel/kgdb_stub.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sh/kernel/kgdb_stub.c 2006-09-20 15:00:12.000000000 +0200 @@ -412,7 +412,7 @@ static struct task_struct *get_thread(in if (pid == PID_MAX) pid = 0; /* First check via PID */ - thread = find_task_by_pid(pid); + thread = find_task_by_real_pid(pid); if (thread) return thread; diff -NurpP --minimal linux-2.6.17.13/arch/sh/kernel/setup.c linux-2.6.17.13-g-v/arch/sh/kernel/setup.c --- linux-2.6.17.13/arch/sh/kernel/setup.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sh/kernel/setup.c 2006-09-20 15:00:12.000000000 +0200 @@ -21,6 +21,7 @@ #include #include #include +#include #include #include #include @@ -481,7 +482,7 @@ static int show_cpuinfo(struct seq_file seq_printf(m, "machine\t\t: %s\n", get_system_type()); seq_printf(m, "processor\t: %d\n", cpu); - seq_printf(m, "cpu family\t: %s\n", system_utsname.machine); + seq_printf(m, "cpu family\t: %s\n", vx_new_uts(machine)); seq_printf(m, "cpu type\t: %s\n", get_cpu_subtype()); show_cpuflags(m); diff -NurpP --minimal linux-2.6.17.13/arch/sh/kernel/sys_sh.c linux-2.6.17.13-g-v/arch/sh/kernel/sys_sh.c --- linux-2.6.17.13/arch/sh/kernel/sys_sh.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sh/kernel/sys_sh.c 2006-09-20 15:00:12.000000000 +0200 @@ -21,6 +21,7 @@ #include #include #include +#include #include #include @@ -267,7 +268,7 @@ asmlinkage int sys_uname(struct old_utsn if (!name) return -EFAULT; down_read(&uts_sem); - err=copy_to_user(name, &system_utsname, sizeof (*name)); + err=copy_to_user(name, vx_new_utsname(), sizeof (*name)); up_read(&uts_sem); return err?-EFAULT:0; } diff -NurpP --minimal linux-2.6.17.13/arch/sh64/kernel/sys_sh64.c linux-2.6.17.13-g-v/arch/sh64/kernel/sys_sh64.c --- linux-2.6.17.13/arch/sh64/kernel/sys_sh64.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sh64/kernel/sys_sh64.c 2006-09-20 15:00:12.000000000 +0200 @@ -29,6 +29,7 @@ #include #include #include +#include #include #include #include @@ -279,7 +280,7 @@ asmlinkage int sys_uname(struct old_utsn if (!name) return -EFAULT; down_read(&uts_sem); - err=copy_to_user(name, &system_utsname, sizeof (*name)); + err=copy_to_user(name, vx_new_utsname(), sizeof (*name)); up_read(&uts_sem); return err?-EFAULT:0; } diff -NurpP --minimal linux-2.6.17.13/arch/sparc/Kconfig linux-2.6.17.13-g-v/arch/sparc/Kconfig --- linux-2.6.17.13/arch/sparc/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -291,6 +291,8 @@ source "fs/Kconfig" source "arch/sparc/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/sparc/Makefile linux-2.6.17.13-g-v/arch/sparc/Makefile --- linux-2.6.17.13/arch/sparc/Makefile 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc/Makefile 2006-09-20 15:00:12.000000000 +0200 @@ -34,7 +34,7 @@ libs-y += arch/sparc/prom/ arch/sparc/li # Renaming is done to avoid confusing pattern matching rules in 2.5.45 (multy-) INIT_Y := $(patsubst %/, %/built-in.o, $(init-y)) CORE_Y := $(core-y) -CORE_Y += kernel/ mm/ fs/ ipc/ security/ crypto/ block/ +CORE_Y += kernel/ mm/ fs/ ipc/ security/ crypto/ block/ grsecurity/ CORE_Y := $(patsubst %/, %/built-in.o, $(CORE_Y)) DRIVERS_Y := $(patsubst %/, %/built-in.o, $(drivers-y)) NET_Y := $(patsubst %/, %/built-in.o, $(net-y)) diff -NurpP --minimal linux-2.6.17.13/arch/sparc/kernel/ptrace.c linux-2.6.17.13-g-v/arch/sparc/kernel/ptrace.c --- linux-2.6.17.13/arch/sparc/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -19,6 +19,7 @@ #include #include #include +#include #include #include @@ -300,6 +301,16 @@ asmlinkage void do_ptrace(struct pt_regs goto out; } + if (!vx_check(vx_task_xid(child), VX_WATCH|VX_IDENT)) { + pt_error_return(regs, ESRCH); + goto out_tsk; + } + + if (gr_handle_ptrace(child, request)) { + pt_error_return(regs, EPERM); + goto out_tsk; + } + if ((current->personality == PER_SUNOS && request == PTRACE_SUNATTACH) || (current->personality != PER_SUNOS && request == PTRACE_ATTACH)) { if (ptrace_attach(child)) { diff -NurpP --minimal linux-2.6.17.13/arch/sparc/kernel/sys_sparc.c linux-2.6.17.13-g-v/arch/sparc/kernel/sys_sparc.c --- linux-2.6.17.13/arch/sparc/kernel/sys_sparc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc/kernel/sys_sparc.c 2006-09-20 15:00:12.000000000 +0200 @@ -21,6 +21,7 @@ #include #include #include +#include #include #include @@ -56,7 +57,7 @@ unsigned long arch_get_unmapped_area(str if (ARCH_SUN4C_SUN4 && len > 0x20000000) return -ENOMEM; if (!addr) - addr = TASK_UNMAPPED_BASE; + addr = current->mm->mmap_base; if (flags & MAP_SHARED) addr = COLOUR_ALIGN(addr); @@ -473,13 +474,13 @@ asmlinkage int sys_getdomainname(char __ down_read(&uts_sem); - nlen = strlen(system_utsname.domainname) + 1; + nlen = strlen(vx_new_uts(domainname)) + 1; if (nlen < len) len = nlen; if (len > __NEW_UTS_LEN) goto done; - if (copy_to_user(name, system_utsname.domainname, len)) + if (copy_to_user(name, vx_new_uts(domainname), len)) goto done; err = 0; done: diff -NurpP --minimal linux-2.6.17.13/arch/sparc/kernel/sys_sunos.c linux-2.6.17.13-g-v/arch/sparc/kernel/sys_sunos.c --- linux-2.6.17.13/arch/sparc/kernel/sys_sunos.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc/kernel/sys_sunos.c 2006-09-20 15:00:12.000000000 +0200 @@ -35,6 +35,7 @@ #include #include #include +#include #include @@ -482,14 +483,16 @@ struct sunos_utsname { asmlinkage int sunos_uname(struct sunos_utsname __user *name) { int ret; + struct new_utsname *ptr; down_read(&uts_sem); - ret = copy_to_user(&name->sname[0], &system_utsname.sysname[0], sizeof(name->sname) - 1); + ptr = vx_new_utsname(); + ret = copy_to_user(&name->sname[0], ptr->sysname, sizeof(name->sname) - 1); if (!ret) { - ret |= __copy_to_user(&name->nname[0], &system_utsname.nodename[0], sizeof(name->nname) - 1); + ret |= __copy_to_user(&name->nname[0], ptr->nodename, sizeof(name->nname) - 1); ret |= __put_user('\0', &name->nname[8]); - ret |= __copy_to_user(&name->rel[0], &system_utsname.release[0], sizeof(name->rel) - 1); - ret |= __copy_to_user(&name->ver[0], &system_utsname.version[0], sizeof(name->ver) - 1); - ret |= __copy_to_user(&name->mach[0], &system_utsname.machine[0], sizeof(name->mach) - 1); + ret |= __copy_to_user(&name->rel[0], ptr->release, sizeof(name->rel) - 1); + ret |= __copy_to_user(&name->ver[0], ptr->version, sizeof(name->ver) - 1); + ret |= __copy_to_user(&name->mach[0], ptr->machine, sizeof(name->mach) - 1); } up_read(&uts_sem); return ret ? -EFAULT : 0; diff -NurpP --minimal linux-2.6.17.13/arch/sparc/kernel/systbls.S linux-2.6.17.13-g-v/arch/sparc/kernel/systbls.S --- linux-2.6.17.13/arch/sparc/kernel/systbls.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc/kernel/systbls.S 2006-09-20 15:00:12.000000000 +0200 @@ -72,7 +72,7 @@ sys_call_table: /*250*/ .long sparc_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl /*255*/ .long sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep /*260*/ .long sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun -/*265*/ .long sys_timer_delete, sys_timer_create, sys_nis_syscall, sys_io_setup, sys_io_destroy +/*265*/ .long sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy /*270*/ .long sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink /*275*/ .long sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid /*280*/ .long sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat diff -NurpP --minimal linux-2.6.17.13/arch/sparc/mm/fault.c linux-2.6.17.13-g-v/arch/sparc/mm/fault.c --- linux-2.6.17.13/arch/sparc/mm/fault.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc/mm/fault.c 2006-09-20 15:00:12.000000000 +0200 @@ -21,6 +21,10 @@ #include #include #include +#include +#include +#include +#include #include #include @@ -217,6 +221,252 @@ static unsigned long compute_si_addr(str return safe_compute_effective_address(regs, insn); } +#ifdef CONFIG_PAX_PAGEEXEC +void pax_emuplt_close(struct vm_area_struct * vma) +{ + vma->vm_mm->call_dl_resolve = 0UL; +} + +static struct page* pax_emuplt_nopage(struct vm_area_struct *vma, unsigned long address, int *type) +{ + struct page* page; + unsigned int *kaddr; + + page = alloc_page(GFP_HIGHUSER); + if (!page) + return NOPAGE_OOM; + + kaddr = kmap(page); + memset(kaddr, 0, PAGE_SIZE); + kaddr[0] = 0x9DE3BFA8U; /* save */ + flush_dcache_page(page); + kunmap(page); + if (type) + *type = VM_FAULT_MAJOR; + + return page; +} + +static struct vm_operations_struct pax_vm_ops = { + .close = pax_emuplt_close, + .nopage = pax_emuplt_nopage, +}; + +static int pax_insert_vma(struct vm_area_struct *vma, unsigned long addr) +{ + int ret; + + memset(vma, 0, sizeof(*vma)); + vma->vm_mm = current->mm; + vma->vm_start = addr; + vma->vm_end = addr + PAGE_SIZE; + vma->vm_flags = VM_READ | VM_EXEC | VM_MAYREAD | VM_MAYEXEC; + vma->vm_page_prot = protection_map[vma->vm_flags & 0x0f]; + vma->vm_ops = &pax_vm_ops; + + ret = insert_vm_struct(current->mm, vma); + if (ret) + return ret; + + ++current->mm->total_vm; + return 0; +} + +/* + * PaX: decide what to do with offenders (regs->pc = fault address) + * + * returns 1 when task should be killed + * 2 when patched PLT trampoline was detected + * 3 when unpatched PLT trampoline was detected + */ +static int pax_handle_fetch_fault(struct pt_regs *regs) +{ + +#ifdef CONFIG_PAX_EMUPLT + int err; + + do { /* PaX: patched PLT emulation #1 */ + unsigned int sethi1, sethi2, jmpl; + + err = get_user(sethi1, (unsigned int*)regs->pc); + err |= get_user(sethi2, (unsigned int*)(regs->pc+4)); + err |= get_user(jmpl, (unsigned int*)(regs->pc+8)); + + if (err) + break; + + if ((sethi1 & 0xFFC00000U) == 0x03000000U && + (sethi2 & 0xFFC00000U) == 0x03000000U && + (jmpl & 0xFFFFE000U) == 0x81C06000U) + { + unsigned int addr; + + regs->u_regs[UREG_G1] = (sethi2 & 0x003FFFFFU) << 10; + addr = regs->u_regs[UREG_G1]; + addr += (((jmpl | 0xFFFFE000U) ^ 0x00001000U) + 0x00001000U); + regs->pc = addr; + regs->npc = addr+4; + return 2; + } + } while (0); + + { /* PaX: patched PLT emulation #2 */ + unsigned int ba; + + err = get_user(ba, (unsigned int*)regs->pc); + + if (!err && (ba & 0xFFC00000U) == 0x30800000U) { + unsigned int addr; + + addr = regs->pc + ((((ba | 0xFFC00000U) ^ 0x00200000U) + 0x00200000U) << 2); + regs->pc = addr; + regs->npc = addr+4; + return 2; + } + } + + do { /* PaX: patched PLT emulation #3 */ + unsigned int sethi, jmpl, nop; + + err = get_user(sethi, (unsigned int*)regs->pc); + err |= get_user(jmpl, (unsigned int*)(regs->pc+4)); + err |= get_user(nop, (unsigned int*)(regs->pc+8)); + + if (err) + break; + + if ((sethi & 0xFFC00000U) == 0x03000000U && + (jmpl & 0xFFFFE000U) == 0x81C06000U && + nop == 0x01000000U) + { + unsigned int addr; + + addr = (sethi & 0x003FFFFFU) << 10; + regs->u_regs[UREG_G1] = addr; + addr += (((jmpl | 0xFFFFE000U) ^ 0x00001000U) + 0x00001000U); + regs->pc = addr; + regs->npc = addr+4; + return 2; + } + } while (0); + + do { /* PaX: unpatched PLT emulation step 1 */ + unsigned int sethi, ba, nop; + + err = get_user(sethi, (unsigned int*)regs->pc); + err |= get_user(ba, (unsigned int*)(regs->pc+4)); + err |= get_user(nop, (unsigned int*)(regs->pc+8)); + + if (err) + break; + + if ((sethi & 0xFFC00000U) == 0x03000000U && + ((ba & 0xFFC00000U) == 0x30800000U || (ba & 0xFFF80000U) == 0x30680000U) && + nop == 0x01000000U) + { + unsigned int addr, save, call; + + if ((ba & 0xFFC00000U) == 0x30800000U) + addr = regs->pc + 4 + ((((ba | 0xFFC00000U) ^ 0x00200000U) + 0x00200000U) << 2); + else + addr = regs->pc + 4 + ((((ba | 0xFFF80000U) ^ 0x00040000U) + 0x00040000U) << 2); + + err = get_user(save, (unsigned int*)addr); + err |= get_user(call, (unsigned int*)(addr+4)); + err |= get_user(nop, (unsigned int*)(addr+8)); + if (err) + break; + + if (save == 0x9DE3BFA8U && + (call & 0xC0000000U) == 0x40000000U && + nop == 0x01000000U) + { + struct vm_area_struct *vma; + unsigned long call_dl_resolve; + + down_read(&current->mm->mmap_sem); + call_dl_resolve = current->mm->call_dl_resolve; + up_read(&current->mm->mmap_sem); + if (likely(call_dl_resolve)) + goto emulate; + + vma = kmem_cache_alloc(vm_area_cachep, SLAB_KERNEL); + + down_write(&current->mm->mmap_sem); + if (current->mm->call_dl_resolve) { + call_dl_resolve = current->mm->call_dl_resolve; + up_write(&current->mm->mmap_sem); + if (vma) kmem_cache_free(vm_area_cachep, vma); + goto emulate; + } + + call_dl_resolve = get_unmapped_area(NULL, 0UL, PAGE_SIZE, 0UL, MAP_PRIVATE); + if (!vma || (call_dl_resolve & ~PAGE_MASK)) { + up_write(&current->mm->mmap_sem); + if (vma) kmem_cache_free(vm_area_cachep, vma); + return 1; + } + + if (pax_insert_vma(vma, call_dl_resolve)) { + up_write(&current->mm->mmap_sem); + kmem_cache_free(vm_area_cachep, vma); + return 1; + } + + current->mm->call_dl_resolve = call_dl_resolve; + up_write(&current->mm->mmap_sem); + +emulate: + regs->u_regs[UREG_G1] = (sethi & 0x003FFFFFU) << 10; + regs->pc = call_dl_resolve; + regs->npc = addr+4; + return 3; + } + } + } while (0); + + do { /* PaX: unpatched PLT emulation step 2 */ + unsigned int save, call, nop; + + err = get_user(save, (unsigned int*)(regs->pc-4)); + err |= get_user(call, (unsigned int*)regs->pc); + err |= get_user(nop, (unsigned int*)(regs->pc+4)); + if (err) + break; + + if (save == 0x9DE3BFA8U && + (call & 0xC0000000U) == 0x40000000U && + nop == 0x01000000U) + { + unsigned int dl_resolve = regs->pc + ((((call | 0xC0000000U) ^ 0x20000000U) + 0x20000000U) << 2); + + regs->u_regs[UREG_RETPC] = regs->pc; + regs->pc = dl_resolve; + regs->npc = dl_resolve+4; + return 3; + } + } while (0); +#endif + + return 1; +} + +void pax_report_insns(void *pc, void *sp) +{ + unsigned long i; + + printk(KERN_ERR "PAX: bytes at PC: "); + for (i = 0; i < 5; i++) { + unsigned int c; + if (get_user(c, (unsigned int*)pc+i)) + printk("???????? "); + else + printk("%08x ", c); + } + printk("\n"); +} +#endif + asmlinkage void do_sparc_fault(struct pt_regs *regs, int text_fault, int write, unsigned long address) { @@ -280,6 +530,24 @@ good_area: if(!(vma->vm_flags & VM_WRITE)) goto bad_area; } else { + +#ifdef CONFIG_PAX_PAGEEXEC + if ((mm->pax_flags & MF_PAX_PAGEEXEC) && text_fault && !(vma->vm_flags & VM_EXEC)) { + up_read(&mm->mmap_sem); + switch (pax_handle_fetch_fault(regs)) { + +#ifdef CONFIG_PAX_EMUPLT + case 2: + case 3: + return; +#endif + + } + pax_report_fault(regs, (void*)regs->pc, (void*)regs->u_regs[UREG_FP]); + do_exit(SIGKILL); + } +#endif + /* Allow reads even for write-only mappings */ if(!(vma->vm_flags & (VM_READ | VM_EXEC))) goto bad_area; diff -NurpP --minimal linux-2.6.17.13/arch/sparc/mm/init.c linux-2.6.17.13-g-v/arch/sparc/mm/init.c --- linux-2.6.17.13/arch/sparc/mm/init.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc/mm/init.c 2006-09-20 15:00:12.000000000 +0200 @@ -333,17 +333,17 @@ void __init paging_init(void) /* Initialize the protection map with non-constant, MMU dependent values. */ protection_map[0] = PAGE_NONE; - protection_map[1] = PAGE_READONLY; - protection_map[2] = PAGE_COPY; - protection_map[3] = PAGE_COPY; + protection_map[1] = PAGE_READONLY_NOEXEC; + protection_map[2] = PAGE_COPY_NOEXEC; + protection_map[3] = PAGE_COPY_NOEXEC; protection_map[4] = PAGE_READONLY; protection_map[5] = PAGE_READONLY; protection_map[6] = PAGE_COPY; protection_map[7] = PAGE_COPY; protection_map[8] = PAGE_NONE; - protection_map[9] = PAGE_READONLY; - protection_map[10] = PAGE_SHARED; - protection_map[11] = PAGE_SHARED; + protection_map[9] = PAGE_READONLY_NOEXEC; + protection_map[10] = PAGE_SHARED_NOEXEC; + protection_map[11] = PAGE_SHARED_NOEXEC; protection_map[12] = PAGE_READONLY; protection_map[13] = PAGE_READONLY; protection_map[14] = PAGE_SHARED; diff -NurpP --minimal linux-2.6.17.13/arch/sparc/mm/srmmu.c linux-2.6.17.13-g-v/arch/sparc/mm/srmmu.c --- linux-2.6.17.13/arch/sparc/mm/srmmu.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc/mm/srmmu.c 2006-09-20 15:00:12.000000000 +0200 @@ -2161,6 +2161,13 @@ void __init ld_mmu_srmmu(void) BTFIXUPSET_INT(page_shared, pgprot_val(SRMMU_PAGE_SHARED)); BTFIXUPSET_INT(page_copy, pgprot_val(SRMMU_PAGE_COPY)); BTFIXUPSET_INT(page_readonly, pgprot_val(SRMMU_PAGE_RDONLY)); + +#ifdef CONFIG_PAX_PAGEEXEC + BTFIXUPSET_INT(page_shared_noexec, pgprot_val(SRMMU_PAGE_SHARED_NOEXEC)); + BTFIXUPSET_INT(page_copy_noexec, pgprot_val(SRMMU_PAGE_COPY_NOEXEC)); + BTFIXUPSET_INT(page_readonly_noexec, pgprot_val(SRMMU_PAGE_RDONLY_NOEXEC)); +#endif + BTFIXUPSET_INT(page_kernel, pgprot_val(SRMMU_PAGE_KERNEL)); page_kernel = pgprot_val(SRMMU_PAGE_KERNEL); diff -NurpP --minimal linux-2.6.17.13/arch/sparc64/Kconfig linux-2.6.17.13-g-v/arch/sparc64/Kconfig --- linux-2.6.17.13/arch/sparc64/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc64/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -423,6 +423,8 @@ endmenu source "arch/sparc64/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/sparc64/kernel/binfmt_aout32.c linux-2.6.17.13-g-v/arch/sparc64/kernel/binfmt_aout32.c --- linux-2.6.17.13/arch/sparc64/kernel/binfmt_aout32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc64/kernel/binfmt_aout32.c 2006-09-20 15:00:12.000000000 +0200 @@ -27,6 +27,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/arch/sparc64/kernel/ptrace.c linux-2.6.17.13-g-v/arch/sparc64/kernel/ptrace.c --- linux-2.6.17.13/arch/sparc64/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc64/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -22,6 +22,7 @@ #include #include #include +#include #include #include @@ -213,6 +214,16 @@ asmlinkage void do_ptrace(struct pt_regs goto out; } + if (!vx_check(vx_task_xid(child), VX_WATCH|VX_IDENT)) { + pt_error_return(regs, ESRCH); + goto out_tsk; + } + + if (gr_handle_ptrace(child, (long)request)) { + pt_error_return(regs, EPERM); + goto out_tsk; + } + if ((current->personality == PER_SUNOS && request == PTRACE_SUNATTACH) || (current->personality != PER_SUNOS && request == PTRACE_ATTACH)) { if (ptrace_attach(child)) { diff -NurpP --minimal linux-2.6.17.13/arch/sparc64/kernel/sys_sparc.c linux-2.6.17.13-g-v/arch/sparc64/kernel/sys_sparc.c --- linux-2.6.17.13/arch/sparc64/kernel/sys_sparc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc64/kernel/sys_sparc.c 2006-09-20 15:00:12.000000000 +0200 @@ -26,6 +26,7 @@ #include #include #include +#include #include #include @@ -140,6 +141,10 @@ unsigned long arch_get_unmapped_area(str if (filp || (flags & MAP_SHARED)) do_color_align = 1; +#ifdef CONFIG_PAX_RANDMMAP + if (!(mm->pax_flags & MF_PAX_RANDMMAP) || !filp) +#endif + if (addr) { if (do_color_align) addr = COLOUR_ALIGN(addr, pgoff); @@ -153,9 +158,9 @@ unsigned long arch_get_unmapped_area(str } if (len > mm->cached_hole_size) { - start_addr = addr = mm->free_area_cache; + start_addr = addr = mm->free_area_cache; } else { - start_addr = addr = TASK_UNMAPPED_BASE; + start_addr = addr = mm->mmap_base; mm->cached_hole_size = 0; } @@ -175,8 +180,8 @@ full_search: vma = find_vma(mm, VA_EXCLUDE_END); } if (unlikely(task_size < addr)) { - if (start_addr != TASK_UNMAPPED_BASE) { - start_addr = addr = TASK_UNMAPPED_BASE; + if (start_addr != mm->mmap_base) { + start_addr = addr = mm->mmap_base; mm->cached_hole_size = 0; goto full_search; } @@ -379,6 +384,12 @@ void arch_pick_mmap_layout(struct mm_str current->signal->rlim[RLIMIT_STACK].rlim_cur == RLIM_INFINITY || sysctl_legacy_va_layout) { mm->mmap_base = TASK_UNMAPPED_BASE + random_factor; + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base += mm->delta_mmap; +#endif + mm->get_unmapped_area = arch_get_unmapped_area; mm->unmap_area = arch_unmap_area; } else { @@ -393,6 +404,12 @@ void arch_pick_mmap_layout(struct mm_str gap = (task_size / 6 * 5); mm->mmap_base = PAGE_ALIGN(task_size - gap - random_factor); + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base -= mm->delta_mmap + mm->delta_stack; +#endif + mm->get_unmapped_area = arch_get_unmapped_area_topdown; mm->unmap_area = arch_unmap_area_topdown; } @@ -711,13 +728,13 @@ asmlinkage long sys_getdomainname(char _ down_read(&uts_sem); - nlen = strlen(system_utsname.domainname) + 1; + nlen = strlen(vx_new_uts(domainname)) + 1; if (nlen < len) len = nlen; if (len > __NEW_UTS_LEN) goto done; - if (copy_to_user(name, system_utsname.domainname, len)) + if (copy_to_user(name, vx_new_uts(domainname), len)) goto done; err = 0; done: diff -NurpP --minimal linux-2.6.17.13/arch/sparc64/kernel/sys_sunos32.c linux-2.6.17.13-g-v/arch/sparc64/kernel/sys_sunos32.c --- linux-2.6.17.13/arch/sparc64/kernel/sys_sunos32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc64/kernel/sys_sunos32.c 2006-09-20 15:00:12.000000000 +0200 @@ -35,6 +35,7 @@ #include #include #include +#include #include #include @@ -437,18 +438,20 @@ struct sunos_utsname { asmlinkage int sunos_uname(struct sunos_utsname __user *name) { int ret; + struct new_utsname *ptr; down_read(&uts_sem); - ret = copy_to_user(&name->sname[0], &system_utsname.sysname[0], + ptr = vx_new_utsname(); + ret = copy_to_user(&name->sname[0], ptr->sysname, sizeof(name->sname) - 1); - ret |= copy_to_user(&name->nname[0], &system_utsname.nodename[0], + ret |= copy_to_user(&name->nname[0], ptr->nodename, sizeof(name->nname) - 1); ret |= put_user('\0', &name->nname[8]); - ret |= copy_to_user(&name->rel[0], &system_utsname.release[0], + ret |= copy_to_user(&name->rel[0], ptr->release, sizeof(name->rel) - 1); - ret |= copy_to_user(&name->ver[0], &system_utsname.version[0], + ret |= copy_to_user(&name->ver[0], ptr->version, sizeof(name->ver) - 1); - ret |= copy_to_user(&name->mach[0], &system_utsname.machine[0], + ret |= copy_to_user(&name->mach[0], ptr->machine, sizeof(name->mach) - 1); up_read(&uts_sem); return (ret ? -EFAULT : 0); diff -NurpP --minimal linux-2.6.17.13/arch/sparc64/kernel/systbls.S linux-2.6.17.13-g-v/arch/sparc64/kernel/systbls.S --- linux-2.6.17.13/arch/sparc64/kernel/systbls.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc64/kernel/systbls.S 2006-09-20 15:00:12.000000000 +0200 @@ -73,7 +73,7 @@ sys_call_table32: /*250*/ .word sys32_mremap, sys32_sysctl, sys32_getsid, sys_fdatasync, sys32_nfsservctl .word sys32_sync_file_range, compat_sys_clock_settime, compat_sys_clock_gettime, compat_sys_clock_getres, sys32_clock_nanosleep /*260*/ .word compat_sys_sched_getaffinity, compat_sys_sched_setaffinity, sys32_timer_settime, compat_sys_timer_gettime, sys_timer_getoverrun - .word sys_timer_delete, compat_sys_timer_create, sys_ni_syscall, compat_sys_io_setup, sys_io_destroy + .word sys_timer_delete, compat_sys_timer_create, sys32_vserver, compat_sys_io_setup, sys_io_destroy /*270*/ .word sys32_io_submit, sys_io_cancel, compat_sys_io_getevents, sys32_mq_open, sys_mq_unlink .word compat_sys_mq_timedsend, compat_sys_mq_timedreceive, compat_sys_mq_notify, compat_sys_mq_getsetattr, compat_sys_waitid /*280*/ .word sys32_tee, sys_add_key, sys_request_key, sys_keyctl, compat_sys_openat @@ -143,7 +143,7 @@ sys_call_table: /*250*/ .word sys64_mremap, sys_sysctl, sys_getsid, sys_fdatasync, sys_nfsservctl .word sys_sync_file_range, sys_clock_settime, sys_clock_gettime, sys_clock_getres, sys_clock_nanosleep /*260*/ .word sys_sched_getaffinity, sys_sched_setaffinity, sys_timer_settime, sys_timer_gettime, sys_timer_getoverrun - .word sys_timer_delete, sys_timer_create, sys_ni_syscall, sys_io_setup, sys_io_destroy + .word sys_timer_delete, sys_timer_create, sys_vserver, sys_io_setup, sys_io_destroy /*270*/ .word sys_io_submit, sys_io_cancel, sys_io_getevents, sys_mq_open, sys_mq_unlink .word sys_mq_timedsend, sys_mq_timedreceive, sys_mq_notify, sys_mq_getsetattr, sys_waitid /*280*/ .word sys_tee, sys_add_key, sys_request_key, sys_keyctl, sys_openat diff -NurpP --minimal linux-2.6.17.13/arch/sparc64/mm/fault.c linux-2.6.17.13-g-v/arch/sparc64/mm/fault.c --- linux-2.6.17.13/arch/sparc64/mm/fault.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc64/mm/fault.c 2006-09-20 15:00:12.000000000 +0200 @@ -19,6 +19,10 @@ #include #include #include +#include +#include +#include +#include #include #include @@ -253,6 +257,369 @@ cannot_handle: unhandled_fault (address, current, regs); } +#ifdef CONFIG_PAX_PAGEEXEC +#ifdef CONFIG_PAX_EMUPLT +static void pax_emuplt_close(struct vm_area_struct * vma) +{ + vma->vm_mm->call_dl_resolve = 0UL; +} + +static struct page* pax_emuplt_nopage(struct vm_area_struct *vma, unsigned long address, int *type) +{ + struct page* page; + unsigned int *kaddr; + + page = alloc_page(GFP_HIGHUSER); + if (!page) + return NOPAGE_OOM; + + kaddr = kmap(page); + memset(kaddr, 0, PAGE_SIZE); + kaddr[0] = 0x9DE3BFA8U; /* save */ + flush_dcache_page(page); + kunmap(page); + if (type) + *type = VM_FAULT_MAJOR; + return page; +} + +static struct vm_operations_struct pax_vm_ops = { + .close = pax_emuplt_close, + .nopage = pax_emuplt_nopage, +}; + +static int pax_insert_vma(struct vm_area_struct *vma, unsigned long addr) +{ + int ret; + + memset(vma, 0, sizeof(*vma)); + vma->vm_mm = current->mm; + vma->vm_start = addr; + vma->vm_end = addr + PAGE_SIZE; + vma->vm_flags = VM_READ | VM_EXEC | VM_MAYREAD | VM_MAYEXEC; + vma->vm_page_prot = protection_map[vma->vm_flags & 0x0f]; + vma->vm_ops = &pax_vm_ops; + + ret = insert_vm_struct(current->mm, vma); + if (ret) + return ret; + + ++current->mm->total_vm; + return 0; +} +#endif + +/* + * PaX: decide what to do with offenders (regs->tpc = fault address) + * + * returns 1 when task should be killed + * 2 when patched PLT trampoline was detected + * 3 when unpatched PLT trampoline was detected + */ +static int pax_handle_fetch_fault(struct pt_regs *regs) +{ + +#ifdef CONFIG_PAX_EMUPLT + int err; + + do { /* PaX: patched PLT emulation #1 */ + unsigned int sethi1, sethi2, jmpl; + + err = get_user(sethi1, (unsigned int*)regs->tpc); + err |= get_user(sethi2, (unsigned int*)(regs->tpc+4)); + err |= get_user(jmpl, (unsigned int*)(regs->tpc+8)); + + if (err) + break; + + if ((sethi1 & 0xFFC00000U) == 0x03000000U && + (sethi2 & 0xFFC00000U) == 0x03000000U && + (jmpl & 0xFFFFE000U) == 0x81C06000U) + { + unsigned long addr; + + regs->u_regs[UREG_G1] = (sethi2 & 0x003FFFFFU) << 10; + addr = regs->u_regs[UREG_G1]; + addr += (((jmpl | 0xFFFFFFFFFFFFE000UL) ^ 0x00001000UL) + 0x00001000UL); + regs->tpc = addr; + regs->tnpc = addr+4; + return 2; + } + } while (0); + + { /* PaX: patched PLT emulation #2 */ + unsigned int ba; + + err = get_user(ba, (unsigned int*)regs->tpc); + + if (!err && (ba & 0xFFC00000U) == 0x30800000U) { + unsigned long addr; + + addr = regs->tpc + ((((ba | 0xFFFFFFFFFFC00000UL) ^ 0x00200000UL) + 0x00200000UL) << 2); + regs->tpc = addr; + regs->tnpc = addr+4; + return 2; + } + } + + do { /* PaX: patched PLT emulation #3 */ + unsigned int sethi, jmpl, nop; + + err = get_user(sethi, (unsigned int*)regs->tpc); + err |= get_user(jmpl, (unsigned int*)(regs->tpc+4)); + err |= get_user(nop, (unsigned int*)(regs->tpc+8)); + + if (err) + break; + + if ((sethi & 0xFFC00000U) == 0x03000000U && + (jmpl & 0xFFFFE000U) == 0x81C06000U && + nop == 0x01000000U) + { + unsigned long addr; + + addr = (sethi & 0x003FFFFFU) << 10; + regs->u_regs[UREG_G1] = addr; + addr += (((jmpl | 0xFFFFFFFFFFFFE000UL) ^ 0x00001000UL) + 0x00001000UL); + regs->tpc = addr; + regs->tnpc = addr+4; + return 2; + } + } while (0); + + do { /* PaX: patched PLT emulation #4 */ + unsigned int mov1, call, mov2; + + err = get_user(mov1, (unsigned int*)regs->tpc); + err |= get_user(call, (unsigned int*)(regs->tpc+4)); + err |= get_user(mov2, (unsigned int*)(regs->tpc+8)); + + if (err) + break; + + if (mov1 == 0x8210000FU && + (call & 0xC0000000U) == 0x40000000U && + mov2 == 0x9E100001U) + { + unsigned long addr; + + regs->u_regs[UREG_G1] = regs->u_regs[UREG_RETPC]; + addr = regs->tpc + 4 + ((((call | 0xFFFFFFFFC0000000UL) ^ 0x20000000UL) + 0x20000000UL) << 2); + regs->tpc = addr; + regs->tnpc = addr+4; + return 2; + } + } while (0); + + do { /* PaX: patched PLT emulation #5 */ + unsigned int sethi1, sethi2, or1, or2, sllx, jmpl, nop; + + err = get_user(sethi1, (unsigned int*)regs->tpc); + err |= get_user(sethi2, (unsigned int*)(regs->tpc+4)); + err |= get_user(or1, (unsigned int*)(regs->tpc+8)); + err |= get_user(or2, (unsigned int*)(regs->tpc+12)); + err |= get_user(sllx, (unsigned int*)(regs->tpc+16)); + err |= get_user(jmpl, (unsigned int*)(regs->tpc+20)); + err |= get_user(nop, (unsigned int*)(regs->tpc+24)); + + if (err) + break; + + if ((sethi1 & 0xFFC00000U) == 0x03000000U && + (sethi2 & 0xFFC00000U) == 0x0B000000U && + (or1 & 0xFFFFE000U) == 0x82106000U && + (or2 & 0xFFFFE000U) == 0x8A116000U && + sllx == 0x83287020 && + jmpl == 0x81C04005U && + nop == 0x01000000U) + { + unsigned long addr; + + regs->u_regs[UREG_G1] = ((sethi1 & 0x003FFFFFU) << 10) | (or1 & 0x000003FFU); + regs->u_regs[UREG_G1] <<= 32; + regs->u_regs[UREG_G5] = ((sethi2 & 0x003FFFFFU) << 10) | (or2 & 0x000003FFU); + addr = regs->u_regs[UREG_G1] + regs->u_regs[UREG_G5]; + regs->tpc = addr; + regs->tnpc = addr+4; + return 2; + } + } while (0); + + do { /* PaX: patched PLT emulation #6 */ + unsigned int sethi1, sethi2, sllx, or, jmpl, nop; + + err = get_user(sethi1, (unsigned int*)regs->tpc); + err |= get_user(sethi2, (unsigned int*)(regs->tpc+4)); + err |= get_user(sllx, (unsigned int*)(regs->tpc+8)); + err |= get_user(or, (unsigned int*)(regs->tpc+12)); + err |= get_user(jmpl, (unsigned int*)(regs->tpc+16)); + err |= get_user(nop, (unsigned int*)(regs->tpc+20)); + + if (err) + break; + + if ((sethi1 & 0xFFC00000U) == 0x03000000U && + (sethi2 & 0xFFC00000U) == 0x0B000000U && + sllx == 0x83287020 && + (or & 0xFFFFE000U) == 0x8A116000U && + jmpl == 0x81C04005U && + nop == 0x01000000U) + { + unsigned long addr; + + regs->u_regs[UREG_G1] = (sethi1 & 0x003FFFFFU) << 10; + regs->u_regs[UREG_G1] <<= 32; + regs->u_regs[UREG_G5] = ((sethi2 & 0x003FFFFFU) << 10) | (or & 0x3FFU); + addr = regs->u_regs[UREG_G1] + regs->u_regs[UREG_G5]; + regs->tpc = addr; + regs->tnpc = addr+4; + return 2; + } + } while (0); + + do { /* PaX: patched PLT emulation #7 */ + unsigned int sethi, ba, nop; + + err = get_user(sethi, (unsigned int*)regs->tpc); + err |= get_user(ba, (unsigned int*)(regs->tpc+4)); + err |= get_user(nop, (unsigned int*)(regs->tpc+8)); + + if (err) + break; + + if ((sethi & 0xFFC00000U) == 0x03000000U && + (ba & 0xFFF00000U) == 0x30600000U && + nop == 0x01000000U) + { + unsigned long addr; + + addr = (sethi & 0x003FFFFFU) << 10; + regs->u_regs[UREG_G1] = addr; + addr = regs->tpc + ((((ba | 0xFFFFFFFFFFF80000UL) ^ 0x00040000UL) + 0x00040000UL) << 2); + regs->tpc = addr; + regs->tnpc = addr+4; + return 2; + } + } while (0); + + do { /* PaX: unpatched PLT emulation step 1 */ + unsigned int sethi, ba, nop; + + err = get_user(sethi, (unsigned int*)regs->tpc); + err |= get_user(ba, (unsigned int*)(regs->tpc+4)); + err |= get_user(nop, (unsigned int*)(regs->tpc+8)); + + if (err) + break; + + if ((sethi & 0xFFC00000U) == 0x03000000U && + ((ba & 0xFFC00000U) == 0x30800000U || (ba & 0xFFF80000U) == 0x30680000U) && + nop == 0x01000000U) + { + unsigned long addr; + unsigned int save, call; + + if ((ba & 0xFFC00000U) == 0x30800000U) + addr = regs->tpc + 4 + ((((ba | 0xFFFFFFFFFFC00000UL) ^ 0x00200000UL) + 0x00200000UL) << 2); + else + addr = regs->tpc + 4 + ((((ba | 0xFFFFFFFFFFF80000UL) ^ 0x00040000UL) + 0x00040000UL) << 2); + + err = get_user(save, (unsigned int*)addr); + err |= get_user(call, (unsigned int*)(addr+4)); + err |= get_user(nop, (unsigned int*)(addr+8)); + if (err) + break; + + if (save == 0x9DE3BFA8U && + (call & 0xC0000000U) == 0x40000000U && + nop == 0x01000000U) + { + struct vm_area_struct *vma; + unsigned long call_dl_resolve; + + down_read(&current->mm->mmap_sem); + call_dl_resolve = current->mm->call_dl_resolve; + up_read(&current->mm->mmap_sem); + if (likely(call_dl_resolve)) + goto emulate; + + vma = kmem_cache_alloc(vm_area_cachep, SLAB_KERNEL); + + down_write(&current->mm->mmap_sem); + if (current->mm->call_dl_resolve) { + call_dl_resolve = current->mm->call_dl_resolve; + up_write(&current->mm->mmap_sem); + if (vma) kmem_cache_free(vm_area_cachep, vma); + goto emulate; + } + + call_dl_resolve = get_unmapped_area(NULL, 0UL, PAGE_SIZE, 0UL, MAP_PRIVATE); + if (!vma || (call_dl_resolve & ~PAGE_MASK)) { + up_write(&current->mm->mmap_sem); + if (vma) kmem_cache_free(vm_area_cachep, vma); + return 1; + } + + if (pax_insert_vma(vma, call_dl_resolve)) { + up_write(&current->mm->mmap_sem); + kmem_cache_free(vm_area_cachep, vma); + return 1; + } + + current->mm->call_dl_resolve = call_dl_resolve; + up_write(&current->mm->mmap_sem); + +emulate: + regs->u_regs[UREG_G1] = (sethi & 0x003FFFFFU) << 10; + regs->tpc = call_dl_resolve; + regs->tnpc = addr+4; + return 3; + } + } + } while (0); + + do { /* PaX: unpatched PLT emulation step 2 */ + unsigned int save, call, nop; + + err = get_user(save, (unsigned int*)(regs->tpc-4)); + err |= get_user(call, (unsigned int*)regs->tpc); + err |= get_user(nop, (unsigned int*)(regs->tpc+4)); + if (err) + break; + + if (save == 0x9DE3BFA8U && + (call & 0xC0000000U) == 0x40000000U && + nop == 0x01000000U) + { + unsigned long dl_resolve = regs->tpc + ((((call | 0xFFFFFFFFC0000000UL) ^ 0x20000000UL) + 0x20000000UL) << 2); + + regs->u_regs[UREG_RETPC] = regs->tpc; + regs->tpc = dl_resolve; + regs->tnpc = dl_resolve+4; + return 3; + } + } while (0); +#endif + + return 1; +} + +void pax_report_insns(void *pc, void *sp) +{ + unsigned long i; + + printk(KERN_ERR "PAX: bytes at PC: "); + for (i = 0; i < 5; i++) { + unsigned int c; + if (get_user(c, (unsigned int*)pc+i)) + printk("???????? "); + else + printk("%08x ", c); + } + printk("\n"); +} +#endif + asmlinkage void __kprobes do_sparc64_fault(struct pt_regs *regs) { struct mm_struct *mm = current->mm; @@ -295,8 +662,10 @@ asmlinkage void __kprobes do_sparc64_fau goto intr_or_no_mm; if (test_thread_flag(TIF_32BIT)) { - if (!(regs->tstate & TSTATE_PRIV)) + if (!(regs->tstate & TSTATE_PRIV)) { regs->tpc &= 0xffffffff; + regs->tnpc &= 0xffffffff; + } address &= 0xffffffff; } @@ -313,6 +682,29 @@ asmlinkage void __kprobes do_sparc64_fau if (!vma) goto bad_area; +#ifdef CONFIG_PAX_PAGEEXEC + /* PaX: detect ITLB misses on non-exec pages */ + if ((mm->pax_flags & MF_PAX_PAGEEXEC) && vma->vm_start <= address && + !(vma->vm_flags & VM_EXEC) && (fault_code & FAULT_CODE_ITLB)) + { + if (address != regs->tpc) + goto good_area; + + up_read(&mm->mmap_sem); + switch (pax_handle_fetch_fault(regs)) { + +#ifdef CONFIG_PAX_EMUPLT + case 2: + case 3: + return; +#endif + + } + pax_report_fault(regs, (void*)regs->tpc, (void*)(regs->u_regs[UREG_FP] + STACK_BIAS)); + do_exit(SIGKILL); + } +#endif + /* Pure DTLB misses do not tell us whether the fault causing * load/store/atomic was a write or not, it only says that there * was no match. So in such a case we (carefully) read the diff -NurpP --minimal linux-2.6.17.13/arch/sparc64/solaris/fs.c linux-2.6.17.13-g-v/arch/sparc64/solaris/fs.c --- linux-2.6.17.13/arch/sparc64/solaris/fs.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc64/solaris/fs.c 2006-09-20 15:00:12.000000000 +0200 @@ -363,7 +363,7 @@ static int report_statvfs(struct vfsmoun int j = strlen (p); if (j > 15) j = 15; - if (IS_RDONLY(inode)) i = 1; + if (IS_RDONLY(inode) || MNT_IS_RDONLY(mnt)) i = 1; if (mnt->mnt_flags & MNT_NOSUID) i |= 2; if (!sysv_valid_dev(inode->i_sb->s_dev)) return -EOVERFLOW; @@ -399,7 +399,7 @@ static int report_statvfs64(struct vfsmo int j = strlen (p); if (j > 15) j = 15; - if (IS_RDONLY(inode)) i = 1; + if (IS_RDONLY(inode) || MNT_IS_RDONLY(mnt)) i = 1; if (mnt->mnt_flags & MNT_NOSUID) i |= 2; if (!sysv_valid_dev(inode->i_sb->s_dev)) return -EOVERFLOW; diff -NurpP --minimal linux-2.6.17.13/arch/sparc64/solaris/misc.c linux-2.6.17.13-g-v/arch/sparc64/solaris/misc.c --- linux-2.6.17.13/arch/sparc64/solaris/misc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/sparc64/solaris/misc.c 2006-09-20 15:00:12.000000000 +0200 @@ -17,6 +17,7 @@ #include #include #include +#include #include #include @@ -239,7 +240,7 @@ asmlinkage int solaris_utssys(u32 buf, u /* Let's cheat */ err = set_utsfield(v->sysname, "SunOS", 1, 0); down_read(&uts_sem); - err |= set_utsfield(v->nodename, system_utsname.nodename, + err |= set_utsfield(v->nodename, vx_new_uts(nodename), 1, 1); up_read(&uts_sem); err |= set_utsfield(v->release, "2.6", 0, 0); @@ -263,7 +264,7 @@ asmlinkage int solaris_utsname(u32 buf) /* Why should we not lie a bit? */ down_read(&uts_sem); err = set_utsfield(v->sysname, "SunOS", 0, 0); - err |= set_utsfield(v->nodename, system_utsname.nodename, 1, 1); + err |= set_utsfield(v->nodename, vx_new_uts(nodename), 1, 1); err |= set_utsfield(v->release, "5.6", 0, 0); err |= set_utsfield(v->version, "Generic", 0, 0); err |= set_utsfield(v->machine, machine(), 0, 0); @@ -295,7 +296,7 @@ asmlinkage int solaris_sysinfo(int cmd, case SI_HOSTNAME: r = buffer + 256; down_read(&uts_sem); - for (p = system_utsname.nodename, q = buffer; + for (p = vx_new_uts(nodename), q = buffer; q < r && *p && *p != '.'; *q++ = *p++); up_read(&uts_sem); *q = 0; diff -NurpP --minimal linux-2.6.17.13/arch/um/Kconfig linux-2.6.17.13-g-v/arch/um/Kconfig --- linux-2.6.17.13/arch/um/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/um/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -279,6 +279,8 @@ source "drivers/connector/Kconfig" source "fs/Kconfig" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/um/drivers/mconsole_kern.c linux-2.6.17.13-g-v/arch/um/drivers/mconsole_kern.c --- linux-2.6.17.13/arch/um/drivers/mconsole_kern.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/um/drivers/mconsole_kern.c 2006-09-20 15:00:12.000000000 +0200 @@ -23,6 +23,7 @@ #include "linux/list.h" #include "linux/mm.h" #include "linux/console.h" +#include "linux/vs_cvirt.h" #include "asm/irq.h" #include "asm/uaccess.h" #include "user_util.h" diff -NurpP --minimal linux-2.6.17.13/arch/um/kernel/syscall_kern.c linux-2.6.17.13-g-v/arch/um/kernel/syscall_kern.c --- linux-2.6.17.13/arch/um/kernel/syscall_kern.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/um/kernel/syscall_kern.c 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,8 @@ #include "linux/unistd.h" #include "linux/slab.h" #include "linux/utime.h" +#include + #include "asm/mman.h" #include "asm/uaccess.h" #include "kern_util.h" @@ -110,7 +112,7 @@ long sys_uname(struct old_utsname __user if (!name) return -EFAULT; down_read(&uts_sem); - err=copy_to_user(name, &system_utsname, sizeof (*name)); + err=copy_to_user(name, vx_new_utsname(), sizeof (*name)); up_read(&uts_sem); return err?-EFAULT:0; } @@ -118,6 +120,7 @@ long sys_uname(struct old_utsname __user long sys_olduname(struct oldold_utsname __user * name) { long error; + struct new_utsname *ptr; if (!name) return -EFAULT; @@ -126,19 +129,20 @@ long sys_olduname(struct oldold_utsname down_read(&uts_sem); - error = __copy_to_user(&name->sysname,&system_utsname.sysname, + ptr = vx_new_utsname(); + error = __copy_to_user(&name->sysname,ptr->sysname, __OLD_UTS_LEN); error |= __put_user(0,name->sysname+__OLD_UTS_LEN); - error |= __copy_to_user(&name->nodename,&system_utsname.nodename, + error |= __copy_to_user(&name->nodename,ptr->nodename, __OLD_UTS_LEN); error |= __put_user(0,name->nodename+__OLD_UTS_LEN); - error |= __copy_to_user(&name->release,&system_utsname.release, + error |= __copy_to_user(&name->release,ptr->release, __OLD_UTS_LEN); error |= __put_user(0,name->release+__OLD_UTS_LEN); - error |= __copy_to_user(&name->version,&system_utsname.version, + error |= __copy_to_user(&name->version,ptr->version, __OLD_UTS_LEN); error |= __put_user(0,name->version+__OLD_UTS_LEN); - error |= __copy_to_user(&name->machine,&system_utsname.machine, + error |= __copy_to_user(&name->machine,ptr->machine, __OLD_UTS_LEN); error |= __put_user(0,name->machine+__OLD_UTS_LEN); diff -NurpP --minimal linux-2.6.17.13/arch/um/sys-x86_64/syscalls.c linux-2.6.17.13-g-v/arch/um/sys-x86_64/syscalls.c --- linux-2.6.17.13/arch/um/sys-x86_64/syscalls.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/um/sys-x86_64/syscalls.c 2006-09-20 15:00:12.000000000 +0200 @@ -9,6 +9,7 @@ #include "linux/shm.h" #include "linux/utsname.h" #include "linux/personality.h" +#include "linux/vs_cvirt.h" #include "asm/uaccess.h" #define __FRAME_OFFSETS #include "asm/ptrace.h" @@ -21,7 +22,7 @@ asmlinkage long sys_uname64(struct new_u { int err; down_read(&uts_sem); - err = copy_to_user(name, &system_utsname, sizeof (*name)); + err = copy_to_user(name, vx_new_utsname(), sizeof (*name)); up_read(&uts_sem); if (personality(current->personality) == PER_LINUX32) err |= copy_to_user(&name->machine, "i686", 5); diff -NurpP --minimal linux-2.6.17.13/arch/v850/Kconfig linux-2.6.17.13-g-v/arch/v850/Kconfig --- linux-2.6.17.13/arch/v850/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/v850/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -326,6 +326,8 @@ source "drivers/usb/Kconfig" source "arch/v850/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/v850/kernel/module.c linux-2.6.17.13-g-v/arch/v850/kernel/module.c --- linux-2.6.17.13/arch/v850/kernel/module.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/v850/kernel/module.c 2006-09-20 15:00:12.000000000 +0200 @@ -150,8 +150,8 @@ static uint32_t do_plt_call (void *locat tramp[1] = ((val >> 16) & 0xffff) + 0x610000; /* ...; jmp r1 */ /* Init, or core PLT? */ - if (location >= mod->module_core - && location < mod->module_core + mod->core_size) + if (location >= mod->module_core_rx + && location < mod->module_core_rx + mod->core_size_rx) entry = (void *)sechdrs[mod->arch.core_plt_section].sh_addr; else entry = (void *)sechdrs[mod->arch.init_plt_section].sh_addr; diff -NurpP --minimal linux-2.6.17.13/arch/v850/kernel/ptrace.c linux-2.6.17.13-g-v/arch/v850/kernel/ptrace.c --- linux-2.6.17.13/arch/v850/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/v850/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -117,6 +117,9 @@ long arch_ptrace(struct task_struct *chi { int rval; + if (!vx_check(vx_task_xid(child), VX_WATCH|VX_IDENT)) + goto out; + switch (request) { unsigned long val, copied; diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/Kconfig linux-2.6.17.13-g-v/arch/x86_64/Kconfig --- linux-2.6.17.13/arch/x86_64/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -617,6 +617,8 @@ endmenu source "arch/x86_64/Kconfig.debug" +source "kernel/vserver/Kconfig" + source "security/Kconfig" source "crypto/Kconfig" diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/boot/compressed/head.S linux-2.6.17.13-g-v/arch/x86_64/boot/compressed/head.S --- linux-2.6.17.13/arch/x86_64/boot/compressed/head.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/boot/compressed/head.S 2006-09-20 15:00:12.000000000 +0200 @@ -41,11 +41,13 @@ startup_32: movl %eax,%gs lss stack_start,%esp + movl 0x000000,%ecx xorl %eax,%eax 1: incl %eax # check that A20 really IS enabled movl %eax,0x000000 # loop forever if it isn't cmpl %eax,0x100000 je 1b + movl %ecx,0x000000 /* * Initialize eflags. Some BIOS's leave bits like NT set. This would diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/ia32/ia32_aout.c linux-2.6.17.13-g-v/arch/x86_64/ia32/ia32_aout.c --- linux-2.6.17.13/arch/x86_64/ia32/ia32_aout.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/ia32/ia32_aout.c 2006-09-20 15:00:12.000000000 +0200 @@ -25,6 +25,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/ia32/ia32_binfmt.c linux-2.6.17.13-g-v/arch/x86_64/ia32/ia32_binfmt.c --- linux-2.6.17.13/arch/x86_64/ia32/ia32_binfmt.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/ia32/ia32_binfmt.c 2006-09-20 15:00:12.000000000 +0200 @@ -186,6 +186,17 @@ struct elf_prpsinfo //#include #include +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) 0x08048000UL + +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) 16 +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) 16 +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) 16 +#endif + typedef struct user_i387_ia32_struct elf_fpregset_t; typedef struct user32_fxsr_struct elf_fpxregset_t; @@ -371,7 +382,8 @@ int ia32_setup_arg_pages(struct linux_bi kmem_cache_free(vm_area_cachep, mpnt); return ret; } - mm->stack_vm = mm->total_vm = vma_pages(mpnt); + vx_vmpages_sub(mm, mm->total_vm - vma_pages(mpnt)); + mm->stack_vm = mm->total_vm; } for (i = 0 ; i < MAX_ARG_PAGES ; i++) { diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/ia32/ia32entry.S linux-2.6.17.13-g-v/arch/x86_64/ia32/ia32entry.S --- linux-2.6.17.13/arch/x86_64/ia32/ia32entry.S 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/ia32/ia32entry.S 2006-09-20 15:00:12.000000000 +0200 @@ -652,7 +652,7 @@ ia32_sys_call_table: .quad sys_tgkill /* 270 */ .quad compat_sys_utimes .quad sys32_fadvise64_64 - .quad quiet_ni_syscall /* sys_vserver */ + .quad sys32_vserver .quad sys_mbind .quad compat_sys_get_mempolicy /* 275 */ .quad sys_set_mempolicy diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/ia32/mmap32.c linux-2.6.17.13-g-v/arch/x86_64/ia32/mmap32.c --- linux-2.6.17.13/arch/x86_64/ia32/mmap32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/ia32/mmap32.c 2006-09-20 15:00:12.000000000 +0200 @@ -68,10 +68,22 @@ void ia32_pick_mmap_layout(struct mm_str (current->personality & ADDR_COMPAT_LAYOUT) || current->signal->rlim[RLIMIT_STACK].rlim_cur == RLIM_INFINITY) { mm->mmap_base = TASK_UNMAPPED_BASE; + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base += mm->delta_mmap; +#endif + mm->get_unmapped_area = arch_get_unmapped_area; mm->unmap_area = arch_unmap_area; } else { mm->mmap_base = mmap_base(mm); + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base -= mm->delta_mmap + mm->delta_stack; +#endif + mm->get_unmapped_area = arch_get_unmapped_area_topdown; mm->unmap_area = arch_unmap_area_topdown; } diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/ia32/sys_ia32.c linux-2.6.17.13-g-v/arch/x86_64/ia32/sys_ia32.c --- linux-2.6.17.13/arch/x86_64/ia32/sys_ia32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/ia32/sys_ia32.c 2006-09-20 15:00:12.000000000 +0200 @@ -61,6 +61,7 @@ #include #include #include +#include #include #include #include @@ -793,6 +794,7 @@ asmlinkage long sys32_mmap2(unsigned lon asmlinkage long sys32_olduname(struct oldold_utsname __user * name) { int error; + struct new_utsname *ptr; if (!name) return -EFAULT; @@ -801,13 +803,14 @@ asmlinkage long sys32_olduname(struct ol down_read(&uts_sem); - error = __copy_to_user(&name->sysname,&system_utsname.sysname,__OLD_UTS_LEN); + ptr = vx_new_utsname(); + error = __copy_to_user(&name->sysname,ptr->sysname,__OLD_UTS_LEN); __put_user(0,name->sysname+__OLD_UTS_LEN); - __copy_to_user(&name->nodename,&system_utsname.nodename,__OLD_UTS_LEN); + __copy_to_user(&name->nodename,ptr->nodename,__OLD_UTS_LEN); __put_user(0,name->nodename+__OLD_UTS_LEN); - __copy_to_user(&name->release,&system_utsname.release,__OLD_UTS_LEN); + __copy_to_user(&name->release,ptr->release,__OLD_UTS_LEN); __put_user(0,name->release+__OLD_UTS_LEN); - __copy_to_user(&name->version,&system_utsname.version,__OLD_UTS_LEN); + __copy_to_user(&name->version,ptr->version,__OLD_UTS_LEN); __put_user(0,name->version+__OLD_UTS_LEN); { char *arch = "x86_64"; @@ -830,7 +833,7 @@ long sys32_uname(struct old_utsname __us if (!name) return -EFAULT; down_read(&uts_sem); - err=copy_to_user(name, &system_utsname, sizeof (*name)); + err=copy_to_user(name, vx_new_utsname(), sizeof (*name)); up_read(&uts_sem); if (personality(current->personality) == PER_LINUX32) err |= copy_to_user(&name->machine, "i686", 5); diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/ia32/syscall32.c linux-2.6.17.13-g-v/arch/x86_64/ia32/syscall32.c --- linux-2.6.17.13/arch/x86_64/ia32/syscall32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/ia32/syscall32.c 2006-09-20 15:00:12.000000000 +0200 @@ -10,6 +10,7 @@ #include #include #include +#include #include #include #include @@ -70,7 +71,7 @@ int syscall32_setup_pages(struct linux_b kmem_cache_free(vm_area_cachep, vma); return ret; } - mm->total_vm += npages; + vx_vmpages_add(mm, npages); up_write(&mm->mmap_sem); return 0; } diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/kernel/process.c linux-2.6.17.13-g-v/arch/x86_64/kernel/process.c --- linux-2.6.17.13/arch/x86_64/kernel/process.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/kernel/process.c 2006-09-20 15:00:12.000000000 +0200 @@ -830,9 +830,3 @@ int dump_task_regs(struct task_struct *t return 1; } -unsigned long arch_align_stack(unsigned long sp) -{ - if (randomize_va_space) - sp -= get_random_int() % 8192; - return sp & ~0xf; -} diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/kernel/ptrace.c linux-2.6.17.13-g-v/arch/x86_64/kernel/ptrace.c --- linux-2.6.17.13/arch/x86_64/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -19,6 +19,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/kernel/setup64.c linux-2.6.17.13-g-v/arch/x86_64/kernel/setup64.c --- linux-2.6.17.13/arch/x86_64/kernel/setup64.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/kernel/setup64.c 2006-09-20 15:00:12.000000000 +0200 @@ -38,7 +38,6 @@ struct desc_ptr idt_descr = { 256 * 16 - char boot_cpu_stack[IRQSTACKSIZE] __attribute__((section(".bss.page_aligned"))); unsigned long __supported_pte_mask __read_mostly = ~0UL; -static int do_not_nx __cpuinitdata = 0; /* noexec=on|off Control non executable mappings for 64bit processes. @@ -50,16 +49,14 @@ int __init nonx_setup(char *str) { if (!strncmp(str, "on", 2)) { __supported_pte_mask |= _PAGE_NX; - do_not_nx = 0; } else if (!strncmp(str, "off", 3)) { - do_not_nx = 1; __supported_pte_mask &= ~_PAGE_NX; } return 1; } __setup("noexec=", nonx_setup); /* parsed early actually */ -int force_personality32 = 0; +int force_personality32; /* noexec32=on|off Control non executable heap for 32bit processes. @@ -173,7 +170,7 @@ void __cpuinit check_efer(void) unsigned long efer; rdmsrl(MSR_EFER, efer); - if (!(efer & EFER_NX) || do_not_nx) { + if (!(efer & EFER_NX)) { __supported_pte_mask &= ~_PAGE_NX; } } diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/kernel/sys_x86_64.c linux-2.6.17.13-g-v/arch/x86_64/kernel/sys_x86_64.c --- linux-2.6.17.13/arch/x86_64/kernel/sys_x86_64.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/kernel/sys_x86_64.c 2006-09-20 15:00:12.000000000 +0200 @@ -16,6 +16,7 @@ #include #include #include +#include #include #include @@ -65,8 +66,8 @@ out: return error; } -static void find_start_end(unsigned long flags, unsigned long *begin, - unsigned long *end) +static void find_start_end(struct mm_struct *mm, unsigned long flags, + unsigned long *begin, unsigned long *end) { if (!test_thread_flag(TIF_IA32) && (flags & MAP_32BIT)) { /* This is usually used needed to map code in small @@ -79,7 +80,7 @@ static void find_start_end(unsigned long *begin = 0x40000000; *end = 0x80000000; } else { - *begin = TASK_UNMAPPED_BASE; + *begin = mm->mmap_base; *end = TASK_SIZE; } } @@ -93,11 +94,15 @@ arch_get_unmapped_area(struct file *filp unsigned long start_addr; unsigned long begin, end; - find_start_end(flags, &begin, &end); + find_start_end(mm, flags, &begin, &end); if (len > end) return -ENOMEM; +#ifdef CONFIG_PAX_RANDMMAP + if (!(mm->pax_flags & MF_PAX_RANDMMAP) || !filp) +#endif + if (addr) { addr = PAGE_ALIGN(addr); vma = find_vma(mm, addr); @@ -148,7 +153,7 @@ asmlinkage long sys_uname(struct new_uts { int err; down_read(&uts_sem); - err = copy_to_user(name, &system_utsname, sizeof (*name)); + err = copy_to_user(name, vx_new_utsname(), sizeof (*name)); up_read(&uts_sem); if (personality(current->personality) == PER_LINUX32) err |= copy_to_user(&name->machine, "i686", 5); diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/kernel/traps.c linux-2.6.17.13-g-v/arch/x86_64/kernel/traps.c --- linux-2.6.17.13/arch/x86_64/kernel/traps.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/kernel/traps.c 2006-09-20 15:00:12.000000000 +0200 @@ -324,8 +324,9 @@ void show_registers(struct pt_regs *regs printk("CPU %d ", cpu); __show_regs(regs); - printk("Process %s (pid: %d, threadinfo %p, task %p)\n", - cur->comm, cur->pid, task_thread_info(cur), cur); + printk("Process %s (pid: %d[#%u], threadinfo %p, task %p)\n", + cur->comm, cur->pid, cur->xid, + task_thread_info(cur), cur); /* * When in-kernel, we also print out the stack and code at the diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/mm/fault.c linux-2.6.17.13-g-v/arch/x86_64/mm/fault.c --- linux-2.6.17.13/arch/x86_64/mm/fault.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/mm/fault.c 2006-09-20 15:00:12.000000000 +0200 @@ -24,6 +24,7 @@ #include #include #include +#include #include #include @@ -294,6 +295,33 @@ static int vmalloc_fault(unsigned long a return 0; } +#ifdef CONFIG_PAX_PAGEEXEC +void pax_report_insns(void *pc, void *sp) +{ + long i; + + printk(KERN_ERR "PAX: bytes at PC: "); + for (i = 0; i < 20; i++) { + unsigned char c; + if (get_user(c, (unsigned char __user *)pc+i)) + printk("?? "); + else + printk("%02x ", c); + } + printk("\n"); + + printk(KERN_ERR "PAX: bytes at SP-8: "); + for (i = -1; i < 10; i++) { + unsigned long c; + if (get_user(c, (unsigned long __user *)sp+i)) + printk("???????????????? "); + else + printk("%016lx ", c); + } + printk("\n"); +} +#endif + int page_fault_trace = 0; int exception_trace = 1; @@ -423,6 +451,8 @@ asmlinkage void __kprobes do_page_fault( good_area: info.si_code = SEGV_ACCERR; write = 0; + if ((error_code & PF_INSTR) && !(vma->vm_flags & VM_EXEC)) + goto bad_area; switch (error_code & (PF_PROT|PF_WRITE)) { default: /* 3: write, present */ /* fall through */ @@ -489,7 +519,14 @@ bad_area_nosemaphore: tsk->comm, tsk->pid, address, regs->rip, regs->rsp, error_code); } - + +#ifdef CONFIG_PAX_PAGEEXEC + if (mm && (mm->pax_flags & MF_PAX_PAGEEXEC) && (error_code & 16)) { + pax_report_fault(regs, (void*)regs->rip, (void*)regs->rsp); + do_exit(SIGKILL); + } +#endif + tsk->thread.cr2 = address; /* Kernel addresses are always protection faults */ tsk->thread.error_code = error_code | (address >= TASK_SIZE); diff -NurpP --minimal linux-2.6.17.13/arch/x86_64/mm/mmap.c linux-2.6.17.13-g-v/arch/x86_64/mm/mmap.c --- linux-2.6.17.13/arch/x86_64/mm/mmap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/x86_64/mm/mmap.c 2006-09-20 15:00:12.000000000 +0200 @@ -24,6 +24,12 @@ void arch_pick_mmap_layout(struct mm_str unsigned rnd = get_random_int() & 0xfffffff; mm->mmap_base += ((unsigned long)rnd) << PAGE_SHIFT; } + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base += mm->delta_mmap; +#endif + mm->get_unmapped_area = arch_get_unmapped_area; mm->unmap_area = arch_unmap_area; } diff -NurpP --minimal linux-2.6.17.13/arch/xtensa/kernel/syscalls.c linux-2.6.17.13-g-v/arch/xtensa/kernel/syscalls.c --- linux-2.6.17.13/arch/xtensa/kernel/syscalls.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/arch/xtensa/kernel/syscalls.c 2006-09-20 15:00:12.000000000 +0200 @@ -35,6 +35,7 @@ #include #include #include +#include #include #include #include @@ -129,7 +130,7 @@ out: int sys_uname(struct old_utsname * name) { - if (name && !copy_to_user(name, &system_utsname, sizeof (*name))) + if (name && !copy_to_user(name, vx_new_utsname(), sizeof (*name))) return 0; return -EFAULT; } diff -NurpP --minimal linux-2.6.17.13/drivers/block/Kconfig linux-2.6.17.13-g-v/drivers/block/Kconfig --- linux-2.6.17.13/drivers/block/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/block/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -315,6 +315,13 @@ config BLK_DEV_CRYPTOLOOP instead, which can be configured to be on-disk compatible with the cryptoloop device. +config BLK_DEV_VROOT + tristate "Virtual Root device support" + depends on QUOTACTL + ---help--- + Saying Y here will allow you to use quota/fs ioctls on a shared + partition within a virtual server without compromising security. + config BLK_DEV_NBD tristate "Network block device support" depends on NET diff -NurpP --minimal linux-2.6.17.13/drivers/block/Makefile linux-2.6.17.13-g-v/drivers/block/Makefile --- linux-2.6.17.13/drivers/block/Makefile 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/block/Makefile 2006-09-20 15:00:12.000000000 +0200 @@ -29,4 +29,5 @@ obj-$(CONFIG_BLK_DEV_CRYPTOLOOP) += cryp obj-$(CONFIG_VIODASD) += viodasd.o obj-$(CONFIG_BLK_DEV_SX8) += sx8.o obj-$(CONFIG_BLK_DEV_UB) += ub.o +obj-$(CONFIG_BLK_DEV_VROOT) += vroot.o diff -NurpP --minimal linux-2.6.17.13/drivers/block/vroot.c linux-2.6.17.13-g-v/drivers/block/vroot.c --- linux-2.6.17.13/drivers/block/vroot.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/drivers/block/vroot.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,288 @@ +/* + * linux/drivers/block/vroot.c + * + * written by Herbert Pötzl, 9/11/2002 + * ported to 2.6.10 by Herbert Pötzl, 30/12/2004 + * + * based on the loop.c code by Theodore Ts'o. + * + * Copyright (C) 2002-2005 by Herbert Pötzl. + * Redistribution of this file is permitted under the + * GNU General Public License. + * + */ + +#include +#include +#include +#include +#include +#include + +#include +#include + + +static int max_vroot = 8; + +static struct vroot_device *vroot_dev; +static struct gendisk **disks; + + +static int vroot_set_dev( + struct vroot_device *vr, + struct file *vr_file, + struct block_device *bdev, + unsigned int arg) +{ + struct block_device *real_bdev; + struct file *file; + struct inode *inode; + int error; + + error = -EBUSY; + if (vr->vr_state != Vr_unbound) + goto out; + + error = -EBADF; + file = fget(arg); + if (!file) + goto out; + + error = -EINVAL; + inode = file->f_dentry->d_inode; + + + if (S_ISBLK(inode->i_mode)) { + real_bdev = inode->i_bdev; + vr->vr_device = real_bdev; + __iget(real_bdev->bd_inode); + } else + goto out_fput; + + vxdprintk(VXD_CBIT(misc, 0), + "vroot[%d]_set_dev: dev=" VXF_DEV, + vr->vr_number, VXD_DEV(real_bdev)); + + vr->vr_state = Vr_bound; + error = 0; + + out_fput: + fput(file); + out: + return error; +} + +static int vroot_clr_dev( + struct vroot_device *vr, + struct file *vr_file, + struct block_device *bdev) +{ + struct block_device *real_bdev; + + if (vr->vr_state != Vr_bound) + return -ENXIO; + if (vr->vr_refcnt > 1) /* we needed one fd for the ioctl */ + return -EBUSY; + + real_bdev = vr->vr_device; + + vxdprintk(VXD_CBIT(misc, 0), + "vroot[%d]_clr_dev: dev=" VXF_DEV, + vr->vr_number, VXD_DEV(real_bdev)); + + bdput(real_bdev); + vr->vr_state = Vr_unbound; + vr->vr_device = NULL; + return 0; +} + + +static int vr_ioctl(struct inode * inode, struct file * file, + unsigned int cmd, unsigned long arg) +{ + struct vroot_device *vr = inode->i_bdev->bd_disk->private_data; + int err; + + down(&vr->vr_ctl_mutex); + switch (cmd) { + case VROOT_SET_DEV: + err = vroot_set_dev(vr, file, inode->i_bdev, arg); + break; + case VROOT_CLR_DEV: + err = vroot_clr_dev(vr, file, inode->i_bdev); + break; + default: + err = -EINVAL; + break; + } + up(&vr->vr_ctl_mutex); + return err; +} + +static int vr_open(struct inode *inode, struct file *file) +{ + struct vroot_device *vr = inode->i_bdev->bd_disk->private_data; + + down(&vr->vr_ctl_mutex); + vr->vr_refcnt++; + up(&vr->vr_ctl_mutex); + return 0; +} + +static int vr_release(struct inode *inode, struct file *file) +{ + struct vroot_device *vr = inode->i_bdev->bd_disk->private_data; + + down(&vr->vr_ctl_mutex); + --vr->vr_refcnt; + up(&vr->vr_ctl_mutex); + return 0; +} + +static struct block_device_operations vr_fops = { + .owner = THIS_MODULE, + .open = vr_open, + .release = vr_release, + .ioctl = vr_ioctl, +}; + +struct block_device *__vroot_get_real_bdev(struct block_device *bdev) +{ + struct inode *inode = bdev->bd_inode; + struct vroot_device *vr; + struct block_device *real_bdev; + int minor = iminor(inode); + + vr = &vroot_dev[minor]; + real_bdev = vr->vr_device; + + vxdprintk(VXD_CBIT(misc, 0), + "vroot[%d]_get_real_bdev: dev=" VXF_DEV, + vr->vr_number, VXD_DEV(real_bdev)); + + if (vr->vr_state != Vr_bound) + return ERR_PTR(-ENXIO); + + __iget(real_bdev->bd_inode); + return real_bdev; +} + +/* + * And now the modules code and kernel interface. + */ + +module_param(max_vroot, int, 0); + +MODULE_PARM_DESC(max_vroot, "Maximum number of vroot devices (1-256)"); +MODULE_LICENSE("GPL"); +MODULE_ALIAS_BLOCKDEV_MAJOR(VROOT_MAJOR); + +MODULE_AUTHOR ("Herbert Pötzl"); +MODULE_DESCRIPTION ("Virtual Root Device Mapper"); + + +int __init vroot_init(void) +{ + int err, i; + + if (max_vroot < 1 || max_vroot > 256) { + max_vroot = MAX_VROOT_DEFAULT; + printk(KERN_WARNING "vroot: invalid max_vroot " + "(must be between 1 and 256), " + "using default (%d)\n", max_vroot); + } + + if (register_blkdev(VROOT_MAJOR, "vroot")) + return -EIO; + + err = -ENOMEM; + vroot_dev = kmalloc(max_vroot * sizeof(struct vroot_device), GFP_KERNEL); + if (!vroot_dev) + goto out_mem1; + memset(vroot_dev, 0, max_vroot * sizeof(struct vroot_device)); + + disks = kmalloc(max_vroot * sizeof(struct gendisk *), GFP_KERNEL); + if (!disks) + goto out_mem2; + + for (i = 0; i < max_vroot; i++) { + disks[i] = alloc_disk(1); + if (!disks[i]) + goto out_mem3; + } + + devfs_mk_dir("vroot"); + + for (i = 0; i < max_vroot; i++) { + struct vroot_device *vr = &vroot_dev[i]; + struct gendisk *disk = disks[i]; + + memset(vr, 0, sizeof(*vr)); + init_MUTEX(&vr->vr_ctl_mutex); + vr->vr_number = i; + disk->major = VROOT_MAJOR; + disk->first_minor = i; + disk->fops = &vr_fops; + sprintf(disk->disk_name, "vroot%d", i); + sprintf(disk->devfs_name, "vroot/%d", i); + disk->private_data = vr; + } + + err = register_vroot_grb(&__vroot_get_real_bdev); + if (err) + goto out_reg; + + for (i = 0; i < max_vroot; i++) + add_disk(disks[i]); + printk(KERN_INFO "vroot: loaded (max %d devices)\n", max_vroot); + return 0; + +out_reg: + devfs_remove("vroot"); +out_mem3: + while (i--) + put_disk(disks[i]); + kfree(disks); +out_mem2: + kfree(vroot_dev); +out_mem1: + unregister_blkdev(VROOT_MAJOR, "vroot"); + printk(KERN_ERR "vroot: ran out of memory\n"); + return err; +} + +void vroot_exit(void) +{ + int i; + + if (unregister_vroot_grb(&__vroot_get_real_bdev)) + printk(KERN_WARNING "vroot: cannot unregister grb\n"); + + for (i = 0; i < max_vroot; i++) { + del_gendisk(disks[i]); + put_disk(disks[i]); + } + devfs_remove("vroot"); + if (unregister_blkdev(VROOT_MAJOR, "vroot")) + printk(KERN_WARNING "vroot: cannot unregister blkdev\n"); + + kfree(disks); + kfree(vroot_dev); +} + +module_init(vroot_init); +module_exit(vroot_exit); + +#ifndef MODULE + +static int __init max_vroot_setup(char *str) +{ + max_vroot = simple_strtol(str, NULL, 0); + return 1; +} + +__setup("max_vroot=", max_vroot_setup); + +#endif + diff -NurpP --minimal linux-2.6.17.13/drivers/char/agp/frontend.c linux-2.6.17.13-g-v/drivers/char/agp/frontend.c --- linux-2.6.17.13/drivers/char/agp/frontend.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/char/agp/frontend.c 2006-09-20 15:00:12.000000000 +0200 @@ -841,7 +841,7 @@ static int agpioc_reserve_wrap(struct ag if (copy_from_user(&reserve, arg, sizeof(struct agp_region))) return -EFAULT; - if ((unsigned) reserve.seg_count >= ~0U/sizeof(struct agp_segment)) + if ((unsigned) reserve.seg_count >= ~0U/sizeof(struct agp_segment_priv)) return -EFAULT; client = agp_find_client_by_pid(reserve.pid); diff -NurpP --minimal linux-2.6.17.13/drivers/char/keyboard.c linux-2.6.17.13-g-v/drivers/char/keyboard.c --- linux-2.6.17.13/drivers/char/keyboard.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/char/keyboard.c 2006-09-20 15:00:12.000000000 +0200 @@ -618,6 +618,16 @@ static void k_spec(struct vc_data *vc, u kbd->kbdmode == VC_MEDIUMRAW) && value != KVAL(K_SAK)) return; /* SAK is allowed even in raw mode */ + +#if defined(CONFIG_GRKERNSEC_PROC) || defined(CONFIG_GRKERNSEC_PROC_MEMMAP) + { + void *func = fn_handler[value]; + if (func == fn_show_state || func == fn_show_ptregs || + func == fn_show_mem) + return; + } +#endif + fn_handler[value](vc, regs); } diff -NurpP --minimal linux-2.6.17.13/drivers/char/mem.c linux-2.6.17.13-g-v/drivers/char/mem.c --- linux-2.6.17.13/drivers/char/mem.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/char/mem.c 2006-09-20 15:00:12.000000000 +0200 @@ -28,6 +28,7 @@ #include #include #include +#include #include #include @@ -36,6 +37,10 @@ # include #endif +#ifdef CONFIG_GRKERNSEC +extern struct file_operations grsec_fops; +#endif + /* * Architectures vary in how they handle caching for addresses * outside of main memory. @@ -175,6 +180,11 @@ static ssize_t write_mem(struct file * f if (!valid_phys_addr_range(p, count)) return -EFAULT; +#ifdef CONFIG_GRKERNSEC_KMEM + gr_handle_mem_write(); + return -EPERM; +#endif + written = 0; #ifdef __ARCH_HAS_NO_PAGE_ZERO_MAPPED @@ -251,6 +261,11 @@ static int mmap_mem(struct file * file, size, vma->vm_page_prot); +#ifdef CONFIG_GRKERNSEC_KMEM + if (gr_handle_mem_mmap(vma->vm_pgoff << PAGE_SHIFT, vma)) + return -EPERM; +#endif + /* Remap-pfn-range will mark the range VM_IO and VM_RESERVED */ if (remap_pfn_range(vma, vma->vm_start, @@ -478,6 +493,11 @@ static ssize_t write_kmem(struct file * ssize_t written; char * kbuf; /* k-addr because vwrite() takes vmlist_lock rwlock */ +#ifdef CONFIG_GRKERNSEC_KMEM + gr_handle_kmem_write(); + return -EPERM; +#endif + if (p < (unsigned long) high_memory) { wrote = count; @@ -618,7 +638,23 @@ static inline size_t read_zero_pagealign count = size; zap_page_range(vma, addr, count, NULL); - zeromap_page_range(vma, addr, count, PAGE_COPY); + zeromap_page_range(vma, addr, count, vma->vm_page_prot); + +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) { + unsigned long addr_m; + struct vm_area_struct * vma_m; + + addr_m = vma->vm_start + vma->vm_mirror; + vma_m = find_vma(mm, addr_m); + if (vma_m && vma_m->vm_start == addr_m && (vma_m->vm_flags & VM_MIRROR)) { + addr_m = addr + vma->vm_mirror; + zap_page_range(vma_m, addr_m, count, NULL); + } else + printk(KERN_ERR "PAX: VMMIRROR: read_zero bug, %08lx, %08lx\n", + addr, vma->vm_start); + } +#endif size -= count; buf += count; @@ -767,6 +803,16 @@ static loff_t memory_lseek(struct file * static int open_port(struct inode * inode, struct file * filp) { +#ifdef CONFIG_GRKERNSEC_KMEM + gr_handle_open_port(); + return -EPERM; +#endif + + return capable(CAP_SYS_RAWIO) ? 0 : -EPERM; +} + +static int open_mem(struct inode * inode, struct file * filp) +{ return capable(CAP_SYS_RAWIO) ? 0 : -EPERM; } @@ -774,7 +820,6 @@ static int open_port(struct inode * inod #define full_lseek null_lseek #define write_zero write_null #define read_full read_zero -#define open_mem open_port #define open_kmem open_mem #define open_oldmem open_mem @@ -897,6 +942,11 @@ static int memory_open(struct inode * in filp->f_op = &oldmem_fops; break; #endif +#ifdef CONFIG_GRKERNSEC + case 13: + filp->f_op = &grsec_fops; + break; +#endif default: return -ENXIO; } @@ -929,6 +979,9 @@ static const struct { #ifdef CONFIG_CRASH_DUMP {12,"oldmem", S_IRUSR | S_IWUSR | S_IRGRP, &oldmem_fops}, #endif +#ifdef CONFIG_GRKERNSEC + {13,"grsec", S_IRUSR | S_IWUGO, &grsec_fops}, +#endif }; static struct class *mem_class; diff -NurpP --minimal linux-2.6.17.13/drivers/char/random.c linux-2.6.17.13-g-v/drivers/char/random.c --- linux-2.6.17.13/drivers/char/random.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/char/random.c 2006-09-20 15:00:12.000000000 +0200 @@ -249,8 +249,13 @@ /* * Configuration information */ +#ifdef CONFIG_GRKERNSEC_RANDNET +#define INPUT_POOL_WORDS 512 +#define OUTPUT_POOL_WORDS 128 +#else #define INPUT_POOL_WORDS 128 #define OUTPUT_POOL_WORDS 32 +#endif #define SEC_XFER_SIZE 512 /* @@ -287,10 +292,17 @@ static struct poolinfo { int poolwords; int tap1, tap2, tap3, tap4, tap5; } poolinfo_table[] = { +#ifdef CONFIG_GRKERNSEC_RANDNET + /* x^512 + x^411 + x^308 + x^208 +x^104 + x + 1 -- 225 */ + { 512, 411, 308, 208, 104, 1 }, + /* x^128 + x^103 + x^76 + x^51 + x^25 + x + 1 -- 105 */ + { 128, 103, 76, 51, 25, 1 }, +#else /* x^128 + x^103 + x^76 + x^51 +x^25 + x + 1 -- 105 */ { 128, 103, 76, 51, 25, 1 }, /* x^32 + x^26 + x^20 + x^14 + x^7 + x + 1 -- 15 */ { 32, 26, 20, 14, 7, 1 }, +#endif #if 0 /* x^2048 + x^1638 + x^1231 + x^819 + x^411 + x + 1 -- 115 */ { 2048, 1638, 1231, 819, 411, 1 }, @@ -1174,7 +1186,7 @@ static char sysctl_bootid[16]; static int proc_do_uuid(ctl_table *table, int write, struct file *filp, void __user *buffer, size_t *lenp, loff_t *ppos) { - ctl_table fake_table; + ctl_table fake_table = {0}; unsigned char buf[64], tmp_uuid[16], *uuid; uuid = table->data; @@ -1658,3 +1670,25 @@ randomize_range(unsigned long start, uns return 0; return PAGE_ALIGN(get_random_int() % range + start); } + +#if defined(CONFIG_PAX_ASLR) || defined(CONFIG_GRKERNSEC) +unsigned long pax_get_random_long(void) +{ + static time_t rekey_time; + static __u32 secret[12]; + time_t t; + + /* + * Pick a random secret every REKEY_INTERVAL seconds. + */ + t = get_seconds(); + if (!rekey_time || (t - rekey_time) > REKEY_INTERVAL) { + rekey_time = t; + get_random_bytes(secret, sizeof(secret)); + } + + secret[1] = half_md4_transform(secret+8, secret); + secret[0] = half_md4_transform(secret+8, secret); + return *(unsigned long *)secret; +} +#endif diff -NurpP --minimal linux-2.6.17.13/drivers/char/tty_io.c linux-2.6.17.13-g-v/drivers/char/tty_io.c --- linux-2.6.17.13/drivers/char/tty_io.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/char/tty_io.c 2006-09-20 15:00:12.000000000 +0200 @@ -103,6 +103,7 @@ #include #include #include +#include #include @@ -2388,13 +2389,16 @@ static int tiocsctty(struct tty_struct * static int tiocgpgrp(struct tty_struct *tty, struct tty_struct *real_tty, pid_t __user *p) { + pid_t pgrp; /* * (tty == real_tty) is a cheap way of * testing if the tty is NOT a master pty. */ if (tty == real_tty && current->signal->tty != real_tty) return -ENOTTY; - return put_user(real_tty->pgrp, p); + + pgrp = vx_map_pid(real_tty->pgrp); + return put_user(pgrp, p); } static int tiocspgrp(struct tty_struct *tty, struct tty_struct *real_tty, pid_t __user *p) @@ -2412,6 +2416,8 @@ static int tiocspgrp(struct tty_struct * return -ENOTTY; if (get_user(pgrp, p)) return -EFAULT; + + pgrp = vx_rmap_pid(pgrp); if (pgrp < 0) return -EINVAL; if (session_of_pgrp(pgrp) != current->signal->session) diff -NurpP --minimal linux-2.6.17.13/drivers/char/vt_ioctl.c linux-2.6.17.13-g-v/drivers/char/vt_ioctl.c --- linux-2.6.17.13/drivers/char/vt_ioctl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/char/vt_ioctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -96,6 +96,12 @@ do_kdsk_ioctl(int cmd, struct kbentry __ case KDSKBENT: if (!perm) return -EPERM; + +#ifdef CONFIG_GRKERNSEC + if (!capable(CAP_SYS_TTY_CONFIG)) + return -EPERM; +#endif + if (!i && v == K_NOSUCHMAP) { /* disallocate map */ key_map = key_maps[s]; @@ -236,6 +242,13 @@ do_kdgkb_ioctl(int cmd, struct kbsentry goto reterr; } +#ifdef CONFIG_GRKERNSEC + if (!capable(CAP_SYS_TTY_CONFIG)) { + ret = -EPERM; + goto reterr; + } +#endif + q = func_table[i]; first_free = funcbufptr + (funcbufsize - funcbufleft); for (j = i+1; j < MAX_NR_FUNC && !func_table[j]; j++) diff -NurpP --minimal linux-2.6.17.13/drivers/ieee1394/hosts.c linux-2.6.17.13-g-v/drivers/ieee1394/hosts.c --- linux-2.6.17.13/drivers/ieee1394/hosts.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/ieee1394/hosts.c 2006-09-20 15:00:12.000000000 +0200 @@ -75,6 +75,7 @@ static int dummy_isoctl(struct hpsb_iso } static struct hpsb_host_driver dummy_driver = { + .name = "dummy", .transmit_packet = dummy_transmit_packet, .devctl = dummy_devctl, .isoctl = dummy_isoctl diff -NurpP --minimal linux-2.6.17.13/drivers/ieee1394/ohci1394.c linux-2.6.17.13-g-v/drivers/ieee1394/ohci1394.c --- linux-2.6.17.13/drivers/ieee1394/ohci1394.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/ieee1394/ohci1394.c 2006-09-20 15:00:12.000000000 +0200 @@ -162,9 +162,9 @@ printk(level "%s: " fmt "\n" , OHCI1394_ printk(level "%s: fw-host%d: " fmt "\n" , OHCI1394_DRIVER_NAME, ohci->host->id , ## args) /* Module Parameters */ -static int phys_dma = 1; +static int phys_dma = 0; module_param(phys_dma, int, 0644); -MODULE_PARM_DESC(phys_dma, "Enable physical dma (default = 1)."); +MODULE_PARM_DESC(phys_dma, "Enable physical dma (default = 0)."); static void dma_trm_tasklet(unsigned long data); static void dma_trm_reset(struct dma_trm_ctx *d); diff -NurpP --minimal linux-2.6.17.13/drivers/infiniband/core/uverbs_mem.c linux-2.6.17.13-g-v/drivers/infiniband/core/uverbs_mem.c --- linux-2.6.17.13/drivers/infiniband/core/uverbs_mem.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/infiniband/core/uverbs_mem.c 2006-09-20 15:00:12.000000000 +0200 @@ -36,6 +36,7 @@ #include #include +#include #include "uverbs.h" @@ -161,7 +162,7 @@ out: if (ret < 0) __ib_umem_release(dev, mem, 0); else - current->mm->locked_vm = locked; + vx_vmlocked_sub(current->mm, current->mm->locked_vm - locked); up_write(&current->mm->mmap_sem); free_page((unsigned long) page_list); @@ -174,8 +175,8 @@ void ib_umem_release(struct ib_device *d __ib_umem_release(dev, umem, 1); down_write(&current->mm->mmap_sem); - current->mm->locked_vm -= - PAGE_ALIGN(umem->length + umem->offset) >> PAGE_SHIFT; + vx_vmlocked_sub(current->mm, + PAGE_ALIGN(umem->length + umem->offset) >> PAGE_SHIFT); up_write(&current->mm->mmap_sem); } @@ -184,7 +185,7 @@ static void ib_umem_account(void *work_p struct ib_umem_account_work *work = work_ptr; down_write(&work->mm->mmap_sem); - work->mm->locked_vm -= work->diff; + vx_vmlocked_sub(work->mm, work->diff); up_write(&work->mm->mmap_sem); mmput(work->mm); kfree(work); diff -NurpP --minimal linux-2.6.17.13/drivers/infiniband/hw/ipath/ipath_user_pages.c linux-2.6.17.13-g-v/drivers/infiniband/hw/ipath/ipath_user_pages.c --- linux-2.6.17.13/drivers/infiniband/hw/ipath/ipath_user_pages.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/infiniband/hw/ipath/ipath_user_pages.c 2006-09-20 15:00:12.000000000 +0200 @@ -32,6 +32,7 @@ #include #include +#include #include "ipath_kernel.h" @@ -71,7 +72,8 @@ static int __get_user_pages(unsigned lon lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur >> PAGE_SHIFT; - if (num_pages > lock_limit) { + if (num_pages > lock_limit || + !vx_vmlocked_avail(current->mm, num_pages)) { ret = -ENOMEM; goto bail; } @@ -88,7 +90,7 @@ static int __get_user_pages(unsigned lon goto bail_release; } - current->mm->locked_vm += num_pages; + vx_vmlocked_add(current->mm, num_pages); ret = 0; goto bail; @@ -157,7 +159,7 @@ void ipath_release_user_pages(struct pag __ipath_release_user_pages(p, num_pages, 1); - current->mm->locked_vm -= num_pages; + vx_vmlocked_sub(current->mm, num_pages); up_write(&current->mm->mmap_sem); } @@ -173,7 +175,7 @@ static void user_pages_account(void *ptr struct ipath_user_pages_work *work = ptr; down_write(&work->mm->mmap_sem); - work->mm->locked_vm -= work->num_pages; + vx_vmlocked_sub(work->mm, work->num_pages); up_write(&work->mm->mmap_sem); mmput(work->mm); kfree(work); diff -NurpP --minimal linux-2.6.17.13/drivers/mtd/devices/doc2001.c linux-2.6.17.13-g-v/drivers/mtd/devices/doc2001.c --- linux-2.6.17.13/drivers/mtd/devices/doc2001.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/mtd/devices/doc2001.c 2006-09-20 15:00:12.000000000 +0200 @@ -423,6 +423,8 @@ static int doc_read_ecc (struct mtd_info /* Don't allow read past end of device */ if (from >= this->totlen) return -EINVAL; + if (!len) + return -EINVAL; /* Don't allow a single read to cross a 512-byte block boundary */ if (from + len > ((from | 0x1ff) + 1)) diff -NurpP --minimal linux-2.6.17.13/drivers/net/pcnet32.c linux-2.6.17.13-g-v/drivers/net/pcnet32.c --- linux-2.6.17.13/drivers/net/pcnet32.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/net/pcnet32.c 2006-09-20 15:00:12.000000000 +0200 @@ -81,7 +81,7 @@ static int cards_found; /* * VLB I/O addresses */ -static unsigned int pcnet32_portlist[] __initdata = +static unsigned int pcnet32_portlist[] __devinitdata = { 0x300, 0x320, 0x340, 0x360, 0 }; static int pcnet32_debug = 0; diff -NurpP --minimal linux-2.6.17.13/drivers/pci/proc.c linux-2.6.17.13-g-v/drivers/pci/proc.c --- linux-2.6.17.13/drivers/pci/proc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/pci/proc.c 2006-09-20 15:00:12.000000000 +0200 @@ -473,7 +473,15 @@ static int __init pci_proc_init(void) { struct proc_dir_entry *entry; struct pci_dev *dev = NULL; +#ifdef CONFIG_GRKERNSEC_PROC_ADD +#ifdef CONFIG_GRKERNSEC_PROC_USER + proc_bus_pci_dir = proc_mkdir_mode("pci", S_IRUSR | S_IXUSR, proc_bus); +#elif CONFIG_GRKERNSEC_PROC_USERGROUP + proc_bus_pci_dir = proc_mkdir_mode("pci", S_IRUSR | S_IXUSR | S_IRGRP | S_IXGRP, proc_bus); +#endif +#else proc_bus_pci_dir = proc_mkdir("pci", proc_bus); +#endif entry = create_proc_entry("devices", 0, proc_bus_pci_dir); if (entry) entry->proc_fops = &proc_bus_pci_dev_operations; diff -NurpP --minimal linux-2.6.17.13/drivers/pnp/pnpbios/bioscalls.c linux-2.6.17.13-g-v/drivers/pnp/pnpbios/bioscalls.c --- linux-2.6.17.13/drivers/pnp/pnpbios/bioscalls.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/pnp/pnpbios/bioscalls.c 2006-09-20 15:00:12.000000000 +0200 @@ -65,7 +65,7 @@ set_base(gdt[(selname) >> 3], (u32)(addr set_limit(gdt[(selname) >> 3], size); \ } while(0) -static struct desc_struct bad_bios_desc = { 0, 0x00409200 }; +static struct desc_struct bad_bios_desc = { 0, 0x00409300 }; /* * At some point we want to use this stack frame pointer to unwind @@ -93,6 +93,10 @@ static inline u16 call_pnp_bios(u16 func struct desc_struct save_desc_40; int cpu; +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; +#endif + /* * PnP BIOSes are generally not terribly re-entrant. * Also, don't rely on them to save everything correctly. @@ -107,6 +111,10 @@ static inline u16 call_pnp_bios(u16 func /* On some boxes IRQ's during PnP BIOS calls are deadly. */ spin_lock_irqsave(&pnp_bios_lock, flags); +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + /* The lock prevents us bouncing CPU here */ if (ts1_size) Q2_SET_SEL(smp_processor_id(), PNP_TS1, ts1_base, ts1_size); @@ -142,9 +150,14 @@ static inline u16 call_pnp_bios(u16 func "i" (0) : "memory" ); - spin_unlock_irqrestore(&pnp_bios_lock, flags); get_cpu_gdt_table(cpu)[0x40 / 8] = save_desc_40; + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + + spin_unlock_irqrestore(&pnp_bios_lock, flags); put_cpu(); /* If we get here and this is set then the PnP BIOS faulted on us. */ diff -NurpP --minimal linux-2.6.17.13/drivers/video/i810/i810_main.c linux-2.6.17.13-g-v/drivers/video/i810/i810_main.c --- linux-2.6.17.13/drivers/video/i810/i810_main.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/video/i810/i810_main.c 2006-09-20 15:00:12.000000000 +0200 @@ -1508,7 +1508,7 @@ static int i810fb_cursor(struct fb_info int size = ((cursor->image.width + 7) >> 3) * cursor->image.height; int i; - u8 *data = kmalloc(64 * 8, GFP_ATOMIC); + u8 *data = kmalloc(64 * 8, GFP_KERNEL); if (data == NULL) return -ENOMEM; diff -NurpP --minimal linux-2.6.17.13/drivers/video/vesafb.c linux-2.6.17.13-g-v/drivers/video/vesafb.c --- linux-2.6.17.13/drivers/video/vesafb.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/drivers/video/vesafb.c 2006-09-20 15:00:12.000000000 +0200 @@ -259,7 +259,7 @@ static int __init vesafb_probe(struct pl size_remap = size_total; vesafb_fix.smem_len = size_remap; -#ifndef __i386__ +#if !defined(__i386__) || defined(CONFIG_PAX_KERNEXEC) screen_info.vesapm_seg = 0; #endif diff -NurpP --minimal linux-2.6.17.13/fs/Kconfig linux-2.6.17.13-g-v/fs/Kconfig --- linux-2.6.17.13/fs/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -794,7 +794,7 @@ config PROC_FS config PROC_KCORE bool "/proc/kcore support" if !ARM - depends on PROC_FS && MMU + depends on PROC_FS && MMU && !GRKERNSEC_PROC_ADD config PROC_VMCORE bool "/proc/vmcore support (EXPERIMENTAL)" diff -NurpP --minimal linux-2.6.17.13/fs/attr.c linux-2.6.17.13-g-v/fs/attr.c --- linux-2.6.17.13/fs/attr.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/attr.c 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,9 @@ #include #include #include +#include +#include +#include /* Taken over from the old code... */ @@ -56,6 +59,28 @@ int inode_change_ok(struct inode *inode, if (current->fsuid != inode->i_uid && !capable(CAP_FOWNER)) goto error; } + + /* Check for evil vserver activity */ + if (vx_check(0, VX_ADMIN)) + goto fine; + + if (IS_BARRIER(inode)) { + vxwprintk(1, "xid=%d messing with the barrier.", + vx_current_xid()); + goto error; + } + switch (inode->i_sb->s_magic) { + case PROC_SUPER_MAGIC: + vxwprintk(1, "xid=%d messing with the procfs.", + vx_current_xid()); + goto error; + case DEVPTS_SUPER_MAGIC: + if (vx_check(inode->i_xid, VX_IDENT)) + goto fine; + vxwprintk(1, "xid=%d messing with the devpts.", + vx_current_xid()); + goto error; + } fine: retval = 0; error: @@ -79,6 +104,8 @@ int inode_setattr(struct inode * inode, inode->i_uid = attr->ia_uid; if (ia_valid & ATTR_GID) inode->i_gid = attr->ia_gid; + if ((ia_valid & ATTR_XID) && IS_TAGXID(inode)) + inode->i_xid = attr->ia_xid; if (ia_valid & ATTR_ATIME) inode->i_atime = timespec_trunc(attr->ia_atime, inode->i_sb->s_time_gran); @@ -153,7 +180,8 @@ int notify_change(struct dentry * dentry error = security_inode_setattr(dentry, attr); if (!error) { if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) || - (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) + (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) || + (ia_valid & ATTR_XID && attr->ia_xid != inode->i_xid)) error = DQUOT_TRANSFER(inode, attr) ? -EDQUOT : 0; if (!error) error = inode_setattr(inode, attr); diff -NurpP --minimal linux-2.6.17.13/fs/binfmt_aout.c linux-2.6.17.13-g-v/fs/binfmt_aout.c --- linux-2.6.17.13/fs/binfmt_aout.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/binfmt_aout.c 2006-09-20 15:00:12.000000000 +0200 @@ -24,6 +24,8 @@ #include #include #include +#include +#include #include #include @@ -123,10 +125,12 @@ static int aout_core_dump(long signr, st /* If the size of the dump file exceeds the rlimit, then see what would happen if we wrote the stack, but not the data area. */ #ifdef __sparc__ + gr_learn_resource(current, RLIMIT_CORE, dump.u_dsize+dump.u_ssize, 1); if ((dump.u_dsize+dump.u_ssize) > current->signal->rlim[RLIMIT_CORE].rlim_cur) dump.u_dsize = 0; #else + gr_learn_resource(current, RLIMIT_CORE, (dump.u_dsize+dump.u_ssize+1) * PAGE_SIZE, 1); if ((dump.u_dsize+dump.u_ssize+1) * PAGE_SIZE > current->signal->rlim[RLIMIT_CORE].rlim_cur) dump.u_dsize = 0; @@ -134,10 +138,12 @@ static int aout_core_dump(long signr, st /* Make sure we have enough room to write the stack and data areas. */ #ifdef __sparc__ + gr_learn_resource(current, RLIMIT_CORE, dump.u_ssize, 1); if ((dump.u_ssize) > current->signal->rlim[RLIMIT_CORE].rlim_cur) dump.u_ssize = 0; #else + gr_learn_resource(current, RLIMIT_CORE, (dump.u_ssize+1) * PAGE_SIZE, 1); if ((dump.u_ssize+1) * PAGE_SIZE > current->signal->rlim[RLIMIT_CORE].rlim_cur) dump.u_ssize = 0; @@ -287,6 +293,8 @@ static int load_aout_binary(struct linux rlim = current->signal->rlim[RLIMIT_DATA].rlim_cur; if (rlim >= RLIM_INFINITY) rlim = ~0; + + gr_learn_resource(current, RLIMIT_DATA, ex.a_data + ex.a_bss, 1); if (ex.a_data + ex.a_bss > rlim) return -ENOMEM; @@ -319,6 +327,28 @@ static int load_aout_binary(struct linux current->mm->mmap = NULL; compute_creds(bprm); current->flags &= ~PF_FORKNOEXEC; + +#if defined(CONFIG_PAX_NOEXEC) || defined(CONFIG_PAX_ASLR) + current->mm->pax_flags = 0UL; +#endif + +#ifdef CONFIG_PAX_PAGEEXEC + if (!(N_FLAGS(ex) & F_PAX_PAGEEXEC)) { + current->mm->pax_flags |= MF_PAX_PAGEEXEC; + +#ifdef CONFIG_PAX_EMUTRAMP + if (N_FLAGS(ex) & F_PAX_EMUTRAMP) + current->mm->pax_flags |= MF_PAX_EMUTRAMP; +#endif + +#ifdef CONFIG_PAX_MPROTECT + if (!(N_FLAGS(ex) & F_PAX_MPROTECT)) + current->mm->pax_flags |= MF_PAX_MPROTECT; +#endif + + } +#endif + #ifdef __sparc__ if (N_MAGIC(ex) == NMAGIC) { loff_t pos = fd_offset; @@ -414,7 +444,7 @@ static int load_aout_binary(struct linux down_write(&current->mm->mmap_sem); error = do_mmap(bprm->file, N_DATADDR(ex), ex.a_data, - PROT_READ | PROT_WRITE | PROT_EXEC, + PROT_READ | PROT_WRITE, MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE | MAP_EXECUTABLE, fd_offset + ex.a_text); up_write(&current->mm->mmap_sem); diff -NurpP --minimal linux-2.6.17.13/fs/binfmt_elf.c linux-2.6.17.13-g-v/fs/binfmt_elf.c --- linux-2.6.17.13/fs/binfmt_elf.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/binfmt_elf.c 2006-09-25 15:27:33.000000000 +0200 @@ -38,15 +38,19 @@ #include #include #include - +#include +#include +#include #include #include #include -#include +#ifdef CONFIG_PAX_SEGMEXEC +#include +#endif -static int load_elf_binary(struct linux_binprm * bprm, struct pt_regs * regs); -static int load_elf_library(struct file*); +static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs); +static int load_elf_library(struct file *); static unsigned long elf_map (struct file *, unsigned long, struct elf_phdr *, int, int); extern int dump_fpu (struct pt_regs *, elf_fpregset_t *); @@ -59,15 +63,15 @@ extern int dump_fpu (struct pt_regs *, e * don't even try. */ #if defined(USE_ELF_CORE_DUMP) && defined(CONFIG_ELF_CORE) -static int elf_core_dump(long signr, struct pt_regs * regs, struct file * file); +static int elf_core_dump(long signr, struct pt_regs *regs, struct file *file); #else #define elf_core_dump NULL #endif #if ELF_EXEC_PAGESIZE > PAGE_SIZE -# define ELF_MIN_ALIGN ELF_EXEC_PAGESIZE +#define ELF_MIN_ALIGN ELF_EXEC_PAGESIZE #else -# define ELF_MIN_ALIGN PAGE_SIZE +#define ELF_MIN_ALIGN PAGE_SIZE #endif #ifndef ELF_CORE_EFLAGS @@ -86,10 +90,12 @@ static struct linux_binfmt elf_format = .min_coredump = ELF_EXEC_PAGESIZE }; -#define BAD_ADDR(x) ((unsigned long)(x) >= TASK_SIZE) +#define BAD_ADDR(x) ((unsigned long)(x) >= TASK_SIZE) static int set_brk(unsigned long start, unsigned long end) { + unsigned long e = end; + start = ELF_PAGEALIGN(start); end = ELF_PAGEALIGN(end); if (end > start) { @@ -100,17 +106,15 @@ static int set_brk(unsigned long start, if (BAD_ADDR(addr)) return addr; } - current->mm->start_brk = current->mm->brk = end; + current->mm->start_brk = current->mm->brk = e; return 0; } - /* We need to explicitly zero any fractional pages after the data section (i.e. bss). This would contain the junk from the file that should not - be in memory */ - - + be in memory + */ static int padzero(unsigned long elf_bss) { unsigned long nbyte; @@ -129,7 +133,9 @@ static int padzero(unsigned long elf_bss #define STACK_ADD(sp, items) ((elf_addr_t __user *)(sp) + (items)) #define STACK_ROUND(sp, items) \ ((15 + (unsigned long) ((sp) + (items))) &~ 15UL) -#define STACK_ALLOC(sp, len) ({ elf_addr_t __user *old_sp = (elf_addr_t __user *)sp; sp += len; old_sp; }) +#define STACK_ALLOC(sp, len) ({ \ + elf_addr_t __user *old_sp = (elf_addr_t __user *)sp; sp += len; \ + old_sp; }) #else #define STACK_ADD(sp, items) ((elf_addr_t __user *)(sp) - (items)) #define STACK_ROUND(sp, items) \ @@ -138,7 +144,7 @@ static int padzero(unsigned long elf_bss #endif static int -create_elf_tables(struct linux_binprm *bprm, struct elfhdr * exec, +create_elf_tables(struct linux_binprm *bprm, struct elfhdr *exec, int interp_aout, unsigned long load_addr, unsigned long interp_load_addr) { @@ -161,7 +167,6 @@ create_elf_tables(struct linux_binprm *b * for userspace to get any other way, in others (i386) it is * merely difficult. */ - u_platform = NULL; if (k_platform) { size_t len = strlen(k_platform) + 1; @@ -171,7 +176,7 @@ create_elf_tables(struct linux_binprm *b * evictions by the processes running on the same package. One * thing we can do is to shuffle the initial stack for them. */ - + p = arch_align_stack(p); u_platform = (elf_addr_t __user *)STACK_ALLOC(p, len); @@ -180,9 +185,12 @@ create_elf_tables(struct linux_binprm *b } /* Create the ELF interpreter info */ - elf_info = (elf_addr_t *) current->mm->saved_auxv; + elf_info = (elf_addr_t *)current->mm->saved_auxv; #define NEW_AUX_ENT(id, val) \ - do { elf_info[ei_index++] = id; elf_info[ei_index++] = val; } while (0) + do { \ + elf_info[ei_index++] = id; \ + elf_info[ei_index++] = val; \ + } while (0) #ifdef ARCH_DLINFO /* @@ -195,21 +203,22 @@ create_elf_tables(struct linux_binprm *b NEW_AUX_ENT(AT_PAGESZ, ELF_EXEC_PAGESIZE); NEW_AUX_ENT(AT_CLKTCK, CLOCKS_PER_SEC); NEW_AUX_ENT(AT_PHDR, load_addr + exec->e_phoff); - NEW_AUX_ENT(AT_PHENT, sizeof (struct elf_phdr)); + NEW_AUX_ENT(AT_PHENT, sizeof(struct elf_phdr)); NEW_AUX_ENT(AT_PHNUM, exec->e_phnum); NEW_AUX_ENT(AT_BASE, interp_load_addr); NEW_AUX_ENT(AT_FLAGS, 0); NEW_AUX_ENT(AT_ENTRY, exec->e_entry); - NEW_AUX_ENT(AT_UID, (elf_addr_t) tsk->uid); - NEW_AUX_ENT(AT_EUID, (elf_addr_t) tsk->euid); - NEW_AUX_ENT(AT_GID, (elf_addr_t) tsk->gid); - NEW_AUX_ENT(AT_EGID, (elf_addr_t) tsk->egid); - NEW_AUX_ENT(AT_SECURE, (elf_addr_t) security_bprm_secureexec(bprm)); + NEW_AUX_ENT(AT_UID, tsk->uid); + NEW_AUX_ENT(AT_EUID, tsk->euid); + NEW_AUX_ENT(AT_GID, tsk->gid); + NEW_AUX_ENT(AT_EGID, tsk->egid); + NEW_AUX_ENT(AT_SECURE, security_bprm_secureexec(bprm)); if (k_platform) { - NEW_AUX_ENT(AT_PLATFORM, (elf_addr_t)(unsigned long)u_platform); + NEW_AUX_ENT(AT_PLATFORM, + (elf_addr_t)(unsigned long)u_platform); } if (bprm->interp_flags & BINPRM_FLAGS_EXECFD) { - NEW_AUX_ENT(AT_EXECFD, (elf_addr_t) bprm->interp_data); + NEW_AUX_ENT(AT_EXECFD, bprm->interp_data); } #undef NEW_AUX_ENT /* AT_NULL is zero; clear the rest too */ @@ -232,7 +241,7 @@ create_elf_tables(struct linux_binprm *b /* Point sp at the lowest address on the stack */ #ifdef CONFIG_STACK_GROWSUP sp = (elf_addr_t __user *)bprm->p - items - ei_index; - bprm->exec = (unsigned long) sp; /* XXX: PARISC HACK */ + bprm->exec = (unsigned long)sp; /* XXX: PARISC HACK */ #else sp = (elf_addr_t __user *)bprm->p; #endif @@ -285,7 +294,7 @@ create_elf_tables(struct linux_binprm *b #ifndef elf_map static unsigned long elf_map(struct file *filep, unsigned long addr, - struct elf_phdr *eppnt, int prot, int type) + struct elf_phdr *eppnt, int prot, int type) { unsigned long map_addr; unsigned long pageoffset = ELF_PAGEOFFSET(eppnt->p_vaddr); @@ -310,16 +319,14 @@ static unsigned long elf_map(struct file is only provided so that we can read a.out libraries that have an ELF header */ -static unsigned long load_elf_interp(struct elfhdr * interp_elf_ex, - struct file * interpreter, - unsigned long *interp_load_addr) +static unsigned long load_elf_interp(struct elfhdr *interp_elf_ex, + struct file *interpreter, unsigned long *interp_load_addr) { struct elf_phdr *elf_phdata; struct elf_phdr *eppnt; - unsigned long load_addr = 0; - int load_addr_set = 0; + unsigned long load_addr = 0, min_addr, max_addr, task_size = TASK_SIZE; unsigned long last_bss = 0, elf_bss = 0; - unsigned long error = ~0UL; + unsigned long error = -EINVAL; int retval, i, size; /* First of all, some simple consistency checks */ @@ -342,15 +349,15 @@ static unsigned long load_elf_interp(str goto out; /* Now read in all of the header information */ - size = sizeof(struct elf_phdr) * interp_elf_ex->e_phnum; if (size > ELF_MIN_ALIGN) goto out; - elf_phdata = (struct elf_phdr *) kmalloc(size, GFP_KERNEL); + elf_phdata = kmalloc(size, GFP_KERNEL); if (!elf_phdata) goto out; - retval = kernel_read(interpreter,interp_elf_ex->e_phoff,(char *)elf_phdata,size); + retval = kernel_read(interpreter, interp_elf_ex->e_phoff, + (char *)elf_phdata,size); error = -EIO; if (retval != size) { if (retval < 0) @@ -358,59 +365,86 @@ static unsigned long load_elf_interp(str goto out_close; } - eppnt = elf_phdata; - for (i=0; ie_phnum; i++, eppnt++) { - if (eppnt->p_type == PT_LOAD) { - int elf_type = MAP_PRIVATE | MAP_DENYWRITE; - int elf_prot = 0; - unsigned long vaddr = 0; - unsigned long k, map_addr; +#ifdef CONFIG_PAX_SEGMEXEC + if (current->mm->pax_flags & MF_PAX_SEGMEXEC) + task_size = SEGMEXEC_TASK_SIZE; +#endif - if (eppnt->p_flags & PF_R) elf_prot = PROT_READ; - if (eppnt->p_flags & PF_W) elf_prot |= PROT_WRITE; - if (eppnt->p_flags & PF_X) elf_prot |= PROT_EXEC; - vaddr = eppnt->p_vaddr; - if (interp_elf_ex->e_type == ET_EXEC || load_addr_set) - elf_type |= MAP_FIXED; + eppnt = elf_phdata; + min_addr = task_size; + max_addr = 0; + error = -ENOMEM; - map_addr = elf_map(interpreter, load_addr + vaddr, eppnt, elf_prot, elf_type); - error = map_addr; - if (BAD_ADDR(map_addr)) - goto out_close; + for (i = 0; i < interp_elf_ex->e_phnum; i++, eppnt++) { + if (eppnt->p_type != PT_LOAD) + continue; - if (!load_addr_set && interp_elf_ex->e_type == ET_DYN) { - load_addr = map_addr - ELF_PAGESTART(vaddr); - load_addr_set = 1; - } + /* + * Check to see if the section's size will overflow the + * allowed task size. Note that p_filesz must always be + * <= p_memsize so it is only necessary to check p_memsz. + */ + if (eppnt->p_filesz > eppnt->p_memsz || eppnt->p_vaddr >= eppnt->p_vaddr + eppnt->p_memsz) + goto out_close; - /* - * Check to see if the section's size will overflow the - * allowed task size. Note that p_filesz must always be - * <= p_memsize so it is only necessary to check p_memsz. - */ - k = load_addr + eppnt->p_vaddr; - if (BAD_ADDR(k) || eppnt->p_filesz > eppnt->p_memsz || - eppnt->p_memsz > TASK_SIZE || TASK_SIZE - eppnt->p_memsz < k) { - error = -ENOMEM; + if (min_addr > ELF_PAGESTART(eppnt->p_vaddr)) + min_addr = ELF_PAGESTART(eppnt->p_vaddr); + if (max_addr < ELF_PAGEALIGN(eppnt->p_vaddr + eppnt->p_memsz)) + max_addr = ELF_PAGEALIGN(eppnt->p_vaddr + eppnt->p_memsz); + } + if (min_addr >= max_addr || max_addr > task_size) goto out_close; - } - /* - * Find the end of the file mapping for this phdr, and keep - * track of the largest address we see for this. - */ - k = load_addr + eppnt->p_vaddr + eppnt->p_filesz; - if (k > elf_bss) - elf_bss = k; + if (interp_elf_ex->e_type == ET_DYN) { + load_addr = get_unmapped_area(interpreter, 0, max_addr - min_addr, 0, MAP_PRIVATE | MAP_EXECUTABLE); - /* - * Do the same thing for the memory mapping - between - * elf_bss and last_bss is the bss section. - */ - k = load_addr + eppnt->p_memsz + eppnt->p_vaddr; - if (k > last_bss) - last_bss = k; - } + if (load_addr >= task_size) + goto out_close; + + load_addr -= min_addr; + } + + eppnt = elf_phdata; + for (i = 0; i < interp_elf_ex->e_phnum; i++, eppnt++) { + int elf_type = MAP_PRIVATE | MAP_DENYWRITE | MAP_FIXED; + int elf_prot = 0; + unsigned long vaddr = 0; + unsigned long k, map_addr; + + if (eppnt->p_type != PT_LOAD) + continue; + + if (eppnt->p_flags & PF_R) + elf_prot = PROT_READ; + if (eppnt->p_flags & PF_W) + elf_prot |= PROT_WRITE; + if (eppnt->p_flags & PF_X) + elf_prot |= PROT_EXEC; + vaddr = eppnt->p_vaddr; + + map_addr = elf_map(interpreter, load_addr + vaddr, + eppnt, elf_prot, elf_type); + error = map_addr; + if (BAD_ADDR(map_addr)) + goto out_close; + + k = load_addr + eppnt->p_vaddr; + + /* + * Find the end of the file mapping for this phdr, and + * keep track of the largest address we see for this. + */ + k = load_addr + eppnt->p_vaddr + eppnt->p_filesz; + if (k > elf_bss) + elf_bss = k; + + /* + * Do the same thing for the memory mapping - between + * elf_bss and last_bss is the bss section. + */ + k = load_addr + eppnt->p_memsz + eppnt->p_vaddr; + if (k > last_bss) + last_bss = k; } /* @@ -424,7 +458,8 @@ static unsigned long load_elf_interp(str goto out_close; } - elf_bss = ELF_PAGESTART(elf_bss + ELF_MIN_ALIGN - 1); /* What we have mapped so far */ + /* What we have mapped so far */ + elf_bss = ELF_PAGESTART(elf_bss + ELF_MIN_ALIGN - 1); /* Map the last of the bss segment */ if (last_bss > elf_bss) { @@ -436,7 +471,9 @@ static unsigned long load_elf_interp(str } *interp_load_addr = load_addr; - error = ((unsigned long) interp_elf_ex->e_entry) + load_addr; + error = ((unsigned long)interp_elf_ex->e_entry) + load_addr; + if (BAD_ADDR(error)) + error = -EFAULT; out_close: kfree(elf_phdata); @@ -444,10 +481,10 @@ out: return error; } -static unsigned long load_aout_interp(struct exec * interp_ex, - struct file * interpreter) +static unsigned long load_aout_interp(struct exec *interp_ex, + struct file *interpreter) { - unsigned long text_data, elf_entry = ~0UL; + unsigned long text_data, elf_entry = -EINVAL; char __user * addr; loff_t offset; @@ -464,7 +501,7 @@ static unsigned long load_aout_interp(st case ZMAGIC: case QMAGIC: offset = N_TXTOFF(*interp_ex); - addr = (char __user *) N_TXTADDR(*interp_ex); + addr = (char __user *)N_TXTADDR(*interp_ex); break; default: goto out; @@ -480,7 +517,6 @@ static unsigned long load_aout_interp(st flush_icache_range((unsigned long)addr, (unsigned long)addr + text_data); - down_write(&current->mm->mmap_sem); do_brk(ELF_PAGESTART(text_data + ELF_MIN_ALIGN - 1), interp_ex->a_bss); @@ -491,6 +527,180 @@ out: return elf_entry; } +#if (defined(CONFIG_PAX_EI_PAX) || defined(CONFIG_PAX_PT_PAX_FLAGS)) && defined(CONFIG_PAX_SOFTMODE) +static unsigned long pax_parse_softmode(const struct elf_phdr * const elf_phdata) +{ + unsigned long pax_flags = 0UL; + +#ifdef CONFIG_PAX_PAGEEXEC + if (elf_phdata->p_flags & PF_PAGEEXEC) + pax_flags |= MF_PAX_PAGEEXEC; +#endif + +#ifdef CONFIG_PAX_SEGMEXEC + if (elf_phdata->p_flags & PF_SEGMEXEC) + pax_flags |= MF_PAX_SEGMEXEC; +#endif + +#ifdef CONFIG_PAX_DEFAULT_PAGEEXEC + if (pax_flags & MF_PAX_PAGEEXEC) + pax_flags &= ~MF_PAX_SEGMEXEC; +#endif + +#ifdef CONFIG_PAX_DEFAULT_SEGMEXEC + if (pax_flags & MF_PAX_SEGMEXEC) + pax_flags &= ~MF_PAX_PAGEEXEC; +#endif + +#ifdef CONFIG_PAX_EMUTRAMP + if (elf_phdata->p_flags & PF_EMUTRAMP) + pax_flags |= MF_PAX_EMUTRAMP; +#endif + +#ifdef CONFIG_PAX_MPROTECT + if (elf_phdata->p_flags & PF_MPROTECT) + pax_flags |= MF_PAX_MPROTECT; +#endif + +#if defined(CONFIG_PAX_RANDMMAP) || defined(CONFIG_PAX_RANDUSTACK) + if (randomize_va_space && (elf_phdata->p_flags & PF_RANDMMAP)) + pax_flags |= MF_PAX_RANDMMAP; +#endif + + return pax_flags; +} +#endif + +#ifdef CONFIG_PAX_PT_PAX_FLAGS +static unsigned long pax_parse_hardmode(const struct elf_phdr * const elf_phdata) +{ + unsigned long pax_flags = 0UL; + +#ifdef CONFIG_PAX_PAGEEXEC + if (!(elf_phdata->p_flags & PF_NOPAGEEXEC)) + pax_flags |= MF_PAX_PAGEEXEC; +#endif + +#ifdef CONFIG_PAX_SEGMEXEC + if (!(elf_phdata->p_flags & PF_NOSEGMEXEC)) + pax_flags |= MF_PAX_SEGMEXEC; +#endif + +#ifdef CONFIG_PAX_DEFAULT_PAGEEXEC + if (pax_flags & MF_PAX_PAGEEXEC) + pax_flags &= ~MF_PAX_SEGMEXEC; +#endif + +#ifdef CONFIG_PAX_DEFAULT_SEGMEXEC + if (pax_flags & MF_PAX_SEGMEXEC) + pax_flags &= ~MF_PAX_PAGEEXEC; +#endif + +#ifdef CONFIG_PAX_EMUTRAMP + if (!(elf_phdata->p_flags & PF_NOEMUTRAMP)) + pax_flags |= MF_PAX_EMUTRAMP; +#endif + +#ifdef CONFIG_PAX_MPROTECT + if (!(elf_phdata->p_flags & PF_NOMPROTECT)) + pax_flags |= MF_PAX_MPROTECT; +#endif + +#if defined(CONFIG_PAX_RANDMMAP) || defined(CONFIG_PAX_RANDUSTACK) + if (randomize_va_space && !(elf_phdata->p_flags & PF_NORANDMMAP)) + pax_flags |= MF_PAX_RANDMMAP; +#endif + + return pax_flags; +} +#endif + +#ifdef CONFIG_PAX_EI_PAX +static unsigned long pax_parse_ei_pax(const struct elfhdr * const elf_ex) +{ + unsigned long pax_flags = 0UL; + +#ifdef CONFIG_PAX_PAGEEXEC + if (!(elf_ex->e_ident[EI_PAX] & EF_PAX_PAGEEXEC)) + pax_flags |= MF_PAX_PAGEEXEC; +#endif + +#ifdef CONFIG_PAX_SEGMEXEC + if (!(elf_ex->e_ident[EI_PAX] & EF_PAX_SEGMEXEC)) + pax_flags |= MF_PAX_SEGMEXEC; +#endif + +#ifdef CONFIG_PAX_DEFAULT_PAGEEXEC + if (pax_flags & MF_PAX_PAGEEXEC) + pax_flags &= ~MF_PAX_SEGMEXEC; +#endif + +#ifdef CONFIG_PAX_DEFAULT_SEGMEXEC + if (pax_flags & MF_PAX_SEGMEXEC) + pax_flags &= ~MF_PAX_PAGEEXEC; +#endif + +#ifdef CONFIG_PAX_EMUTRAMP + if ((pax_flags & (MF_PAX_PAGEEXEC | MF_PAX_SEGMEXEC)) && (elf_ex->e_ident[EI_PAX] & EF_PAX_EMUTRAMP)) + pax_flags |= MF_PAX_EMUTRAMP; +#endif + +#ifdef CONFIG_PAX_MPROTECT + if ((pax_flags & (MF_PAX_PAGEEXEC | MF_PAX_SEGMEXEC)) && !(elf_ex->e_ident[EI_PAX] & EF_PAX_MPROTECT)) + pax_flags |= MF_PAX_MPROTECT; +#endif + +#ifdef CONFIG_PAX_ASLR + if (randomize_va_space && !(elf_ex->e_ident[EI_PAX] & EF_PAX_RANDMMAP)) + pax_flags |= MF_PAX_RANDMMAP; +#endif + + return pax_flags; +} +#endif + +#if defined(CONFIG_PAX_EI_PAX) || defined(CONFIG_PAX_PT_PAX_FLAGS) +static long pax_parse_elf_flags(const struct elfhdr * const elf_ex, const struct elf_phdr * const elf_phdata) +{ + unsigned long pax_flags = 0UL; + +#ifdef CONFIG_PAX_PT_PAX_FLAGS + unsigned long i; +#endif + +#ifdef CONFIG_PAX_EI_PAX + pax_flags = pax_parse_ei_pax(elf_ex); +#endif + +#ifdef CONFIG_PAX_PT_PAX_FLAGS + for (i = 0UL; i < elf_ex->e_phnum; i++) + if (elf_phdata[i].p_type == PT_PAX_FLAGS) { + if (((elf_phdata[i].p_flags & PF_PAGEEXEC) && (elf_phdata[i].p_flags & PF_NOPAGEEXEC)) || + ((elf_phdata[i].p_flags & PF_SEGMEXEC) && (elf_phdata[i].p_flags & PF_NOSEGMEXEC)) || + ((elf_phdata[i].p_flags & PF_EMUTRAMP) && (elf_phdata[i].p_flags & PF_NOEMUTRAMP)) || + ((elf_phdata[i].p_flags & PF_MPROTECT) && (elf_phdata[i].p_flags & PF_NOMPROTECT)) || + ((elf_phdata[i].p_flags & PF_RANDMMAP) && (elf_phdata[i].p_flags & PF_NORANDMMAP))) + return -EINVAL; + +#ifdef CONFIG_PAX_SOFTMODE + if (pax_softmode) + pax_flags = pax_parse_softmode(&elf_phdata[i]); + else +#endif + + pax_flags = pax_parse_hardmode(&elf_phdata[i]); + break; + } +#endif + + if (0 > pax_check_flags(&pax_flags)) + return -EINVAL; + + current->mm->pax_flags = pax_flags; + return 0; +} +#endif + /* * These are the functions used to load ELF style executables and shared * libraries. There is no binary dependent code anywhere else. @@ -519,7 +729,7 @@ static unsigned long randomize_stack_top #endif } -static int load_elf_binary(struct linux_binprm * bprm, struct pt_regs * regs) +static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs) { struct file *interpreter = NULL; /* to shut gcc up */ unsigned long load_addr = 0, load_bias = 0; @@ -527,8 +737,8 @@ static int load_elf_binary(struct linux_ char * elf_interpreter = NULL; unsigned int interpreter_type = INTERPRETER_NONE; unsigned char ibcs2_interpreter = 0; - unsigned long error; - struct elf_phdr * elf_ppnt, *elf_phdata; + unsigned long error = 0; + struct elf_phdr *elf_ppnt, *elf_phdata; unsigned long elf_bss, elf_brk; int elf_exec_fileno; int retval, i; @@ -545,6 +755,7 @@ static int load_elf_binary(struct linux_ struct elfhdr interp_elf_ex; struct exec interp_ex; } *loc; + unsigned long task_size = TASK_SIZE; loc = kmalloc(sizeof(*loc), GFP_KERNEL); if (!loc) { @@ -553,7 +764,7 @@ static int load_elf_binary(struct linux_ } /* Get the exec-header */ - loc->elf_ex = *((struct elfhdr *) bprm->buf); + loc->elf_ex = *((struct elfhdr *)bprm->buf); retval = -ENOEXEC; /* First of all, some simple consistency checks */ @@ -568,7 +779,6 @@ static int load_elf_binary(struct linux_ goto out; /* Now read in all of the header information */ - if (loc->elf_ex.e_phentsize != sizeof(struct elf_phdr)) goto out; if (loc->elf_ex.e_phnum < 1 || @@ -576,18 +786,19 @@ static int load_elf_binary(struct linux_ goto out; size = loc->elf_ex.e_phnum * sizeof(struct elf_phdr); retval = -ENOMEM; - elf_phdata = (struct elf_phdr *) kmalloc(size, GFP_KERNEL); + elf_phdata = kmalloc(size, GFP_KERNEL); if (!elf_phdata) goto out; - retval = kernel_read(bprm->file, loc->elf_ex.e_phoff, (char *) elf_phdata, size); + retval = kernel_read(bprm->file, loc->elf_ex.e_phoff, + (char *)elf_phdata, size); if (retval != size) { if (retval >= 0) retval = -EIO; goto out_free_ph; } - files = current->files; /* Refcounted so ok */ + files = current->files; /* Refcounted so ok */ retval = unshare_files(); if (retval < 0) goto out_free_ph; @@ -598,7 +809,6 @@ static int load_elf_binary(struct linux_ /* exec will make our files private anyway, but for the a.out loader stuff we need to do it earlier */ - retval = get_unused_fd(); if (retval < 0) goto out_free_fh; @@ -620,7 +830,6 @@ static int load_elf_binary(struct linux_ * shared libraries - for now assume that this * is an a.out format binary */ - retval = -ENOEXEC; if (elf_ppnt->p_filesz > PATH_MAX || elf_ppnt->p_filesz < 2) @@ -628,13 +837,13 @@ static int load_elf_binary(struct linux_ retval = -ENOMEM; elf_interpreter = kmalloc(elf_ppnt->p_filesz, - GFP_KERNEL); + GFP_KERNEL); if (!elf_interpreter) goto out_free_file; retval = kernel_read(bprm->file, elf_ppnt->p_offset, - elf_interpreter, - elf_ppnt->p_filesz); + elf_interpreter, + elf_ppnt->p_filesz); if (retval != elf_ppnt->p_filesz) { if (retval >= 0) retval = -EIO; @@ -678,7 +887,8 @@ static int load_elf_binary(struct linux_ retval = PTR_ERR(interpreter); if (IS_ERR(interpreter)) goto out_free_interp; - retval = kernel_read(interpreter, 0, bprm->buf, BINPRM_BUF_SIZE); + retval = kernel_read(interpreter, 0, bprm->buf, + BINPRM_BUF_SIZE); if (retval != BINPRM_BUF_SIZE) { if (retval >= 0) retval = -EIO; @@ -686,8 +896,8 @@ static int load_elf_binary(struct linux_ } /* Get the exec headers */ - loc->interp_ex = *((struct exec *) bprm->buf); - loc->interp_elf_ex = *((struct elfhdr *) bprm->buf); + loc->interp_ex = *((struct exec *)bprm->buf); + loc->interp_elf_ex = *((struct elfhdr *)bprm->buf); break; } elf_ppnt++; @@ -739,7 +949,6 @@ static int load_elf_binary(struct linux_ /* OK, we are done with that, now set up the arg stuff, and then start this sucker up */ - if ((!bprm->sh_bang) && (interpreter_type == INTERPRETER_AOUT)) { char *passed_p = passed_fileno; sprintf(passed_fileno, "%d", elf_exec_fileno); @@ -759,7 +968,6 @@ static int load_elf_binary(struct linux_ /* Discard our unneeded old files struct */ if (files) { - steal_locks(files); put_files_struct(files); files = NULL; } @@ -770,15 +978,89 @@ static int load_elf_binary(struct linux_ current->mm->end_code = 0; current->mm->mmap = NULL; current->flags &= ~PF_FORKNOEXEC; + +#if defined(CONFIG_PAX_NOEXEC) || defined(CONFIG_PAX_ASLR) + current->mm->pax_flags = 0UL; +#endif + +#ifdef CONFIG_PAX_DLRESOLVE + current->mm->call_dl_resolve = 0UL; +#endif + +#if defined(CONFIG_PPC32) && defined(CONFIG_PAX_EMUSIGRT) + current->mm->call_syscall = 0UL; +#endif + +#ifdef CONFIG_PAX_ASLR + current->mm->delta_mmap = 0UL; + current->mm->delta_exec = 0UL; + current->mm->delta_stack = 0UL; +#endif + current->mm->def_flags = def_flags; +#if defined(CONFIG_PAX_EI_PAX) || defined(CONFIG_PAX_PT_PAX_FLAGS) + if (0 > pax_parse_elf_flags(&loc->elf_ex, elf_phdata)) { + send_sig(SIGKILL, current, 0); + goto out_free_dentry; + } +#endif + +#ifdef CONFIG_PAX_HAVE_ACL_FLAGS + pax_set_initial_flags(bprm); +#elif defined(CONFIG_PAX_HOOK_ACL_FLAGS) + if (pax_set_initial_flags_func) + (pax_set_initial_flags_func)(bprm); +#endif + +#ifdef CONFIG_ARCH_TRACK_EXEC_LIMIT + if (current->mm->pax_flags & MF_PAX_PAGEEXEC) + current->mm->context.user_cs_limit = PAGE_SIZE; +#endif + +#ifdef CONFIG_PAX_SEGMEXEC + if (current->mm->pax_flags & MF_PAX_SEGMEXEC) { + int cpu = get_cpu(); + + current->mm->context.user_cs_base = SEGMEXEC_TASK_SIZE; + current->mm->context.user_cs_limit = -SEGMEXEC_TASK_SIZE; + set_user_cs(current->mm, cpu); + put_cpu(); + task_size = SEGMEXEC_TASK_SIZE; + } +#endif + +#ifdef CONFIG_PAX_ASLR + if (current->mm->pax_flags & MF_PAX_RANDMMAP) { +#define pax_delta_mask(delta, lsb, len) (((delta) & ((1UL << (len)) - 1)) << (lsb)) + + current->mm->delta_mmap = pax_delta_mask(pax_get_random_long(), PAX_DELTA_MMAP_LSB(current), PAX_DELTA_MMAP_LEN(current)); + current->mm->delta_exec = pax_delta_mask(pax_get_random_long(), PAX_DELTA_EXEC_LSB(current), PAX_DELTA_EXEC_LEN(current)); + current->mm->delta_stack = pax_delta_mask(pax_get_random_long(), PAX_DELTA_STACK_LSB(current), PAX_DELTA_STACK_LEN(current)); + } +#endif + +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) + if (current->mm->pax_flags & (MF_PAX_PAGEEXEC | MF_PAX_SEGMEXEC)) + executable_stack = EXSTACK_DEFAULT; +#endif + /* Do this immediately, since STACK_TOP as used in setup_arg_pages may depend on the personality. */ SET_PERSONALITY(loc->elf_ex, ibcs2_interpreter); + +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) + if (!(current->mm->pax_flags & (MF_PAX_PAGEEXEC | MF_PAX_SEGMEXEC))) +#endif + if (elf_read_implies_exec(loc->elf_ex, executable_stack)) current->personality |= READ_IMPLIES_EXEC; - if ( !(current->personality & ADDR_NO_RANDOMIZE) && randomize_va_space) +#ifdef CONFIG_PAX_ASLR + if (!(current->mm->pax_flags & MF_PAX_RANDMMAP)) +#endif + + if (!(current->personality & ADDR_NO_RANDOMIZE) && randomize_va_space) current->flags |= PF_RANDOMIZE; arch_pick_mmap_layout(current->mm); @@ -799,8 +1081,8 @@ static int load_elf_binary(struct linux_ the correct location in memory. At this point, we assume that the image should be loaded at fixed address, not at a variable address. */ - - for(i = 0, elf_ppnt = elf_phdata; i < loc->elf_ex.e_phnum; i++, elf_ppnt++) { + for(i = 0, elf_ppnt = elf_phdata; + i < loc->elf_ex.e_phnum; i++, elf_ppnt++) { int elf_prot = 0, elf_flags; unsigned long k, vaddr; @@ -828,30 +1110,44 @@ static int load_elf_binary(struct linux_ load_bias, nbyte)) { /* * This bss-zeroing can fail if the ELF - * file specifies odd protections. So + * file specifies odd protections. So * we don't check the return value */ } } } - if (elf_ppnt->p_flags & PF_R) elf_prot |= PROT_READ; - if (elf_ppnt->p_flags & PF_W) elf_prot |= PROT_WRITE; - if (elf_ppnt->p_flags & PF_X) elf_prot |= PROT_EXEC; + if (elf_ppnt->p_flags & PF_R) + elf_prot |= PROT_READ; + if (elf_ppnt->p_flags & PF_W) + elf_prot |= PROT_WRITE; + if (elf_ppnt->p_flags & PF_X) + elf_prot |= PROT_EXEC; - elf_flags = MAP_PRIVATE|MAP_DENYWRITE|MAP_EXECUTABLE; + elf_flags = MAP_PRIVATE | MAP_DENYWRITE | MAP_EXECUTABLE; vaddr = elf_ppnt->p_vaddr; if (loc->elf_ex.e_type == ET_EXEC || load_addr_set) { elf_flags |= MAP_FIXED; } else if (loc->elf_ex.e_type == ET_DYN) { - /* Try and get dynamic programs out of the way of the default mmap - base, as well as whatever program they might try to exec. This - is because the brk will follow the loader, and is not movable. */ + /* Try and get dynamic programs out of the way of the + * default mmap base, as well as whatever program they + * might try to exec. This is because the brk will + * follow the loader, and is not movable. */ load_bias = ELF_PAGESTART(ELF_ET_DYN_BASE - vaddr); + +#ifdef CONFIG_PAX_RANDMMAP + /* PaX: randomize base address at the default exe base if requested */ + if (current->mm->pax_flags & MF_PAX_RANDMMAP) { + load_bias = ELF_PAGESTART(PAX_ELF_ET_DYN_BASE(current) - vaddr + current->mm->delta_exec); + elf_flags |= MAP_FIXED; + } +#endif + } - error = elf_map(bprm->file, load_bias + vaddr, elf_ppnt, elf_prot, elf_flags); + error = elf_map(bprm->file, load_bias + vaddr, elf_ppnt, + elf_prot, elf_flags); if (BAD_ADDR(error)) { send_sig(SIGKILL, current, 0); goto out_free_dentry; @@ -868,18 +1164,20 @@ static int load_elf_binary(struct linux_ } } k = elf_ppnt->p_vaddr; - if (k < start_code) start_code = k; - if (start_data < k) start_data = k; + if (k < start_code) + start_code = k; + if (start_data < k) + start_data = k; /* * Check to see if the section's size will overflow the * allowed task size. Note that p_filesz must always be * <= p_memsz so it is only necessary to check p_memsz. */ - if (BAD_ADDR(k) || elf_ppnt->p_filesz > elf_ppnt->p_memsz || - elf_ppnt->p_memsz > TASK_SIZE || - TASK_SIZE - elf_ppnt->p_memsz < k) { - /* set_brk can never work. Avoid overflows. */ + if (k >= task_size || elf_ppnt->p_filesz > elf_ppnt->p_memsz || + elf_ppnt->p_memsz > task_size || + task_size - elf_ppnt->p_memsz < k) { + /* set_brk can never work. Avoid overflows. */ send_sig(SIGKILL, current, 0); goto out_free_dentry; } @@ -905,6 +1203,12 @@ static int load_elf_binary(struct linux_ start_data += load_bias; end_data += load_bias; +#ifdef CONFIG_PAX_RANDMMAP + if (current->mm->pax_flags & MF_PAX_RANDMMAP) + elf_brk += PAGE_SIZE + pax_delta_mask(pax_get_random_long(), 4, PAGE_SHIFT); +#undef pax_delta_mask +#endif + /* Calling set_brk effectively mmaps the pages that we need * for the bss and break sections. We must do this before * mapping in the interpreter, to make sure it doesn't wind @@ -966,8 +1270,9 @@ static int load_elf_binary(struct linux_ compute_creds(bprm); current->flags &= ~PF_FORKNOEXEC; - create_elf_tables(bprm, &loc->elf_ex, (interpreter_type == INTERPRETER_AOUT), - load_addr, interp_load_addr); + create_elf_tables(bprm, &loc->elf_ex, + (interpreter_type == INTERPRETER_AOUT), + load_addr, interp_load_addr); /* N.B. passed_fileno might not be initialized? */ if (interpreter_type == INTERPRETER_AOUT) current->mm->arg_start += strlen(passed_fileno) + 1; @@ -981,7 +1286,7 @@ static int load_elf_binary(struct linux_ /* Why this, you ask??? Well SVr4 maps page 0 as read-only, and some applications "depend" upon this behavior. Since we do not have the power to recompile these, we - emulate the SVr4 behavior. Sigh. */ + emulate the SVr4 behavior. Sigh. */ down_write(&current->mm->mmap_sem); error = do_mmap(NULL, 0, PAGE_SIZE, PROT_READ | PROT_EXEC, MAP_FIXED | MAP_PRIVATE, 0); @@ -1036,7 +1341,6 @@ out_free_ph: /* This is really simpleminded and specialized - we are loading an a.out library that is given an ELF header. */ - static int load_elf_library(struct file *file) { struct elf_phdr *elf_phdata; @@ -1046,7 +1350,7 @@ static int load_elf_library(struct file struct elfhdr elf_ex; error = -ENOEXEC; - retval = kernel_read(file, 0, (char *) &elf_ex, sizeof(elf_ex)); + retval = kernel_read(file, 0, (char *)&elf_ex, sizeof(elf_ex)); if (retval != sizeof(elf_ex)) goto out; @@ -1055,7 +1359,7 @@ static int load_elf_library(struct file /* First of all, some simple consistency checks */ if (elf_ex.e_type != ET_EXEC || elf_ex.e_phnum > 2 || - !elf_check_arch(&elf_ex) || !file->f_op || !file->f_op->mmap) + !elf_check_arch(&elf_ex) || !file->f_op || !file->f_op->mmap) goto out; /* Now read in all of the header information */ @@ -1103,7 +1407,8 @@ static int load_elf_library(struct file goto out_free_ph; } - len = ELF_PAGESTART(eppnt->p_filesz + eppnt->p_vaddr + ELF_MIN_ALIGN - 1); + len = ELF_PAGESTART(eppnt->p_filesz + eppnt->p_vaddr + + ELF_MIN_ALIGN - 1); bss = eppnt->p_memsz + eppnt->p_vaddr; if (bss > len) { down_write(&current->mm->mmap_sem); @@ -1156,18 +1461,18 @@ static int dump_seek(struct file *file, * * I think we should skip something. But I am not sure how. H.J. */ -static int maydump(struct vm_area_struct *vma) +static int maydump(struct vm_area_struct *vma, long signr) { /* Do not dump I/O mapped devices or special mappings */ if (vma->vm_flags & (VM_IO | VM_RESERVED)) return 0; - /* Dump shared memory only if mapped from an anonymous file. */ + /* Dump shared memory only if mapped from an anonymous file. */ if (vma->vm_flags & VM_SHARED) return vma->vm_file->f_dentry->d_inode->i_nlink == 0; /* If it hasn't been written to, don't write it out */ - if (!vma->anon_vma) + if (signr != SIGKILL && !vma->anon_vma) return 0; return 1; @@ -1276,11 +1581,11 @@ static void fill_note(struct memelfnote } /* - * fill up all the fields in prstatus from the given task struct, except registers - * which need to be filled up separately. + * fill up all the fields in prstatus from the given task struct, except + * registers which need to be filled up separately. */ static void fill_prstatus(struct elf_prstatus *prstatus, - struct task_struct *p, long signr) + struct task_struct *p, long signr) { prstatus->pr_info.si_signo = prstatus->pr_cursig = signr; prstatus->pr_sigpend = p->pending.signal.sig[0]; @@ -1365,8 +1670,8 @@ struct elf_thread_status /* * In order to add the specific thread information for the elf file format, - * we need to keep a linked list of every threads pr_status and then - * create a single section for them in the final core file. + * we need to keep a linked list of every threads pr_status and then create + * a single section for them in the final core file. */ static int elf_dump_thread_status(long signr, struct elf_thread_status *t) { @@ -1377,19 +1682,23 @@ static int elf_dump_thread_status(long s fill_prstatus(&t->prstatus, p, signr); elf_core_copy_task_regs(p, &t->prstatus.pr_reg); - fill_note(&t->notes[0], "CORE", NT_PRSTATUS, sizeof(t->prstatus), &(t->prstatus)); + fill_note(&t->notes[0], "CORE", NT_PRSTATUS, sizeof(t->prstatus), + &(t->prstatus)); t->num_notes++; sz += notesize(&t->notes[0]); - if ((t->prstatus.pr_fpvalid = elf_core_copy_task_fpregs(p, NULL, &t->fpu))) { - fill_note(&t->notes[1], "CORE", NT_PRFPREG, sizeof(t->fpu), &(t->fpu)); + if ((t->prstatus.pr_fpvalid = elf_core_copy_task_fpregs(p, NULL, + &t->fpu))) { + fill_note(&t->notes[1], "CORE", NT_PRFPREG, sizeof(t->fpu), + &(t->fpu)); t->num_notes++; sz += notesize(&t->notes[1]); } #ifdef ELF_CORE_COPY_XFPREGS if (elf_core_copy_task_xfpregs(p, &t->xfpu)) { - fill_note(&t->notes[2], "LINUX", NT_PRXFPREG, sizeof(t->xfpu), &t->xfpu); + fill_note(&t->notes[2], "LINUX", NT_PRXFPREG, sizeof(t->xfpu), + &t->xfpu); t->num_notes++; sz += notesize(&t->notes[2]); } @@ -1404,7 +1713,7 @@ static int elf_dump_thread_status(long s * and then they are actually written out. If we run out of core limit * we just truncate. */ -static int elf_core_dump(long signr, struct pt_regs * regs, struct file * file) +static int elf_core_dump(long signr, struct pt_regs *regs, struct file *file) { #define NUM_NOTES 6 int has_dumped = 0; @@ -1433,12 +1742,12 @@ static int elf_core_dump(long signr, str /* * We no longer stop all VM operations. * - * This is because those proceses that could possibly change map_count or - * the mmap / vma pages are now blocked in do_exit on current finishing - * this core dump. + * This is because those proceses that could possibly change map_count + * or the mmap / vma pages are now blocked in do_exit on current + * finishing this core dump. * * Only ptrace can touch these memory addresses, but it doesn't change - * the map_count or the pages allocated. So no possibility of crashing + * the map_count or the pages allocated. So no possibility of crashing * exists while dumping the mm->vm_next areas to the core file. */ @@ -1500,7 +1809,7 @@ static int elf_core_dump(long signr, str #endif /* Set up header */ - fill_elf_header(elf, segs+1); /* including notes section */ + fill_elf_header(elf, segs + 1); /* including notes section */ has_dumped = 1; current->flags |= PF_DUMPCORE; @@ -1510,24 +1819,24 @@ static int elf_core_dump(long signr, str * with info from their /proc. */ - fill_note(notes +0, "CORE", NT_PRSTATUS, sizeof(*prstatus), prstatus); - + fill_note(notes + 0, "CORE", NT_PRSTATUS, sizeof(*prstatus), prstatus); fill_psinfo(psinfo, current->group_leader, current->mm); - fill_note(notes +1, "CORE", NT_PRPSINFO, sizeof(*psinfo), psinfo); + fill_note(notes + 1, "CORE", NT_PRPSINFO, sizeof(*psinfo), psinfo); numnote = 2; - auxv = (elf_addr_t *) current->mm->saved_auxv; + auxv = (elf_addr_t *)current->mm->saved_auxv; i = 0; do i += 2; while (auxv[i - 2] != AT_NULL); fill_note(&notes[numnote++], "CORE", NT_AUXV, - i * sizeof (elf_addr_t), auxv); + i * sizeof(elf_addr_t), auxv); /* Try to dump the FPU. */ - if ((prstatus->pr_fpvalid = elf_core_copy_task_fpregs(current, regs, fpu))) + if ((prstatus->pr_fpvalid = + elf_core_copy_task_fpregs(current, regs, fpu))) fill_note(notes + numnote++, "CORE", NT_PRFPREG, sizeof(*fpu), fpu); #ifdef ELF_CORE_COPY_XFPREGS @@ -1572,12 +1881,14 @@ static int elf_core_dump(long signr, str phdr.p_offset = offset; phdr.p_vaddr = vma->vm_start; phdr.p_paddr = 0; - phdr.p_filesz = maydump(vma) ? sz : 0; + phdr.p_filesz = maydump(vma, signr) ? sz : 0; phdr.p_memsz = sz; offset += phdr.p_filesz; phdr.p_flags = vma->vm_flags & VM_READ ? PF_R : 0; - if (vma->vm_flags & VM_WRITE) phdr.p_flags |= PF_W; - if (vma->vm_flags & VM_EXEC) phdr.p_flags |= PF_X; + if (vma->vm_flags & VM_WRITE) + phdr.p_flags |= PF_W; + if (vma->vm_flags & VM_EXEC) + phdr.p_flags |= PF_X; phdr.p_align = ELF_EXEC_PAGESIZE; DUMP_WRITE(&phdr, sizeof(phdr)); @@ -1594,7 +1905,9 @@ static int elf_core_dump(long signr, str /* write out the thread status notes section */ list_for_each(t, &thread_list) { - struct elf_thread_status *tmp = list_entry(t, struct elf_thread_status, list); + struct elf_thread_status *tmp = + list_entry(t, struct elf_thread_status, list); + for (i = 0; i < tmp->num_notes; i++) if (!writenote(&tmp->notes[i], file)) goto end_coredump; @@ -1605,24 +1918,25 @@ static int elf_core_dump(long signr, str for (vma = current->mm->mmap; vma != NULL; vma = vma->vm_next) { unsigned long addr; - if (!maydump(vma)) + if (!maydump(vma, signr)) continue; for (addr = vma->vm_start; addr < vma->vm_end; addr += PAGE_SIZE) { - struct page* page; + struct page *page; struct vm_area_struct *vma; if (get_user_pages(current, current->mm, addr, 1, 0, 1, &page, &vma) <= 0) { - DUMP_SEEK (file->f_pos + PAGE_SIZE); + DUMP_SEEK(file->f_pos + PAGE_SIZE); } else { if (page == ZERO_PAGE(addr)) { - DUMP_SEEK (file->f_pos + PAGE_SIZE); + DUMP_SEEK(file->f_pos + PAGE_SIZE); } else { void *kaddr; - flush_cache_page(vma, addr, page_to_pfn(page)); + flush_cache_page(vma, addr, + page_to_pfn(page)); kaddr = kmap(page); if ((size += PAGE_SIZE) > limit || !dump_write(file, kaddr, @@ -1644,7 +1958,8 @@ static int elf_core_dump(long signr, str if ((off_t)file->f_pos != offset) { /* Sanity check */ - printk(KERN_WARNING "elf_core_dump: file->f_pos (%ld) != offset (%ld)\n", + printk(KERN_WARNING + "elf_core_dump: file->f_pos (%ld) != offset (%ld)\n", (off_t)file->f_pos, offset); } diff -NurpP --minimal linux-2.6.17.13/fs/binfmt_elf_fdpic.c linux-2.6.17.13-g-v/fs/binfmt_elf_fdpic.c --- linux-2.6.17.13/fs/binfmt_elf_fdpic.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/binfmt_elf_fdpic.c 2006-09-20 15:00:12.000000000 +0200 @@ -32,6 +32,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/fs/binfmt_flat.c linux-2.6.17.13-g-v/fs/binfmt_flat.c --- linux-2.6.17.13/fs/binfmt_flat.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/binfmt_flat.c 2006-09-20 15:00:12.000000000 +0200 @@ -37,6 +37,7 @@ #include #include #include +#include #include #include @@ -552,7 +553,9 @@ static int load_flat_file(struct linux_b realdatastart = (unsigned long) -ENOMEM; printk("Unable to allocate RAM for process data, errno %d\n", (int)-datapos); + down_write(&current->mm->mmap_sem); do_munmap(current->mm, textpos, text_len); + up_write(&current->mm->mmap_sem); ret = realdatastart; goto err; } @@ -574,8 +577,10 @@ static int load_flat_file(struct linux_b } if (result >= (unsigned long)-4096) { printk("Unable to read data+bss, errno %d\n", (int)-result); + down_write(&current->mm->mmap_sem); do_munmap(current->mm, textpos, text_len); do_munmap(current->mm, realdatastart, data_len + extra); + up_write(&current->mm->mmap_sem); ret = result; goto err; } @@ -639,8 +644,10 @@ static int load_flat_file(struct linux_b } if (result >= (unsigned long)-4096) { printk("Unable to read code+data+bss, errno %d\n",(int)-result); + down_write(&current->mm->mmap_sem); do_munmap(current->mm, textpos, text_len + data_len + extra + MAX_SHARED_LIBS * sizeof(unsigned long)); + up_write(&current->mm->mmap_sem); ret = result; goto err; } diff -NurpP --minimal linux-2.6.17.13/fs/binfmt_misc.c linux-2.6.17.13-g-v/fs/binfmt_misc.c --- linux-2.6.17.13/fs/binfmt_misc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/binfmt_misc.c 2006-09-20 15:00:12.000000000 +0200 @@ -112,9 +112,11 @@ static int load_misc_binary(struct linux struct files_struct *files = NULL; retval = -ENOEXEC; - if (!enabled) + if (!enabled || bprm->misc) goto _ret; + bprm->misc++; + /* to keep locking time low, we copy the interpreter string */ read_lock(&entries_lock); fmt = check_file(bprm); diff -NurpP --minimal linux-2.6.17.13/fs/binfmt_som.c linux-2.6.17.13-g-v/fs/binfmt_som.c --- linux-2.6.17.13/fs/binfmt_som.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/binfmt_som.c 2006-09-20 15:00:12.000000000 +0200 @@ -28,6 +28,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/fs/buffer.c linux-2.6.17.13-g-v/fs/buffer.c --- linux-2.6.17.13/fs/buffer.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/buffer.c 2006-09-20 15:00:12.000000000 +0200 @@ -42,6 +42,7 @@ #include #include #include +#include static int fsync_buffers_list(spinlock_t *lock, struct list_head *list); static void invalidate_bh_lrus(void); @@ -2167,6 +2168,7 @@ static int __generic_cont_expand(struct err = -EFBIG; limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur; + gr_learn_resource(current, RLIMIT_FSIZE, (unsigned long) size, 1); if (limit != RLIM_INFINITY && size > (loff_t)limit) { send_sig(SIGXFSZ, current, 0); goto out; diff -NurpP --minimal linux-2.6.17.13/fs/compat.c linux-2.6.17.13-g-v/fs/compat.c --- linux-2.6.17.13/fs/compat.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/compat.c 2006-09-20 15:00:12.000000000 +0200 @@ -46,6 +46,7 @@ #include #include #include +#include #include /* siocdevprivate_ioctl */ @@ -1498,6 +1499,11 @@ int compat_do_execve(char * filename, struct file *file; int retval; int i; +#ifdef CONFIG_GRKERNSEC + struct file *old_exec_file; + struct acl_subject_label *old_acl; + struct rlimit old_rlim[RLIM_NLIMITS]; +#endif retval = -ENOMEM; bprm = kzalloc(sizeof(*bprm), GFP_KERNEL); @@ -1515,6 +1521,15 @@ int compat_do_execve(char * filename, bprm->file = file; bprm->filename = filename; bprm->interp = filename; + + gr_learn_resource(current, RLIMIT_NPROC, atomic_read(&current->user->processes), 1); + retval = -EAGAIN; + if (gr_handle_nproc()) + goto out_file; + retval = -EACCES; + if (!gr_acl_handle_execve(file->f_dentry, file->f_vfsmnt)) + goto out_file; + bprm->mm = mm_alloc(); retval = -ENOMEM; if (!bprm->mm) @@ -1553,10 +1568,39 @@ int compat_do_execve(char * filename, if (retval < 0) goto out; + if (!gr_tpe_allow(file)) { + retval = -EACCES; + goto out; + } + + if (gr_check_crash_exec(file)) { + retval = -EACCES; + goto out; + } + + gr_log_chroot_exec(file->f_dentry, file->f_vfsmnt); + + gr_handle_exec_args(bprm, (char __user * __user *)argv); + +#ifdef CONFIG_GRKERNSEC + old_acl = current->acl; + memcpy(old_rlim, current->signal->rlim, sizeof(old_rlim)); + old_exec_file = current->exec_file; + get_file(file); + current->exec_file = file; +#endif + + gr_set_proc_label(file->f_dentry, file->f_vfsmnt); + retval = search_binary_handler(bprm, regs); if (retval >= 0) { free_arg_pages(bprm); +#ifdef CONFIG_GRKERNSEC + if (old_exec_file) + fput(old_exec_file); +#endif + /* execve success */ security_bprm_free(bprm); acct_update_integrals(current); @@ -1564,6 +1608,13 @@ int compat_do_execve(char * filename, return retval; } +#ifdef CONFIG_GRKERNSEC + current->acl = old_acl; + memcpy(current->signal->rlim, old_rlim, sizeof(old_rlim)); + fput(current->exec_file); + current->exec_file = old_exec_file; +#endif + out: /* Something went wrong, return the inode and free the argument pages*/ for (i = 0 ; i < MAX_ARG_PAGES ; i++) { diff -NurpP --minimal linux-2.6.17.13/fs/dcache.c linux-2.6.17.13-g-v/fs/dcache.c --- linux-2.6.17.13/fs/dcache.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/dcache.c 2006-09-20 15:00:12.000000000 +0200 @@ -1389,7 +1389,7 @@ already_unhashed: * * "buflen" should be positive. Caller holds the dcache_lock. */ -static char * __d_path( struct dentry *dentry, struct vfsmount *vfsmnt, +char * __d_path( struct dentry *dentry, struct vfsmount *vfsmnt, struct dentry *root, struct vfsmount *rootmnt, char *buffer, int buflen) { diff -NurpP --minimal linux-2.6.17.13/fs/devpts/inode.c linux-2.6.17.13-g-v/fs/devpts/inode.c --- linux-2.6.17.13/fs/devpts/inode.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/devpts/inode.c 2006-09-20 15:00:12.000000000 +0200 @@ -20,7 +20,19 @@ #include #include -#define DEVPTS_SUPER_MAGIC 0x1cd1 + +static int devpts_permission(struct inode *inode, int mask, struct nameidata *nd) +{ + int ret = -EACCES; + + if (vx_check(inode->i_xid, VX_IDENT)) + ret = generic_permission(inode, mask, NULL); + return ret; +} + +static struct inode_operations devpts_file_inode_operations = { + .permission = devpts_permission, +}; static struct vfsmount *devpts_mnt; static struct dentry *devpts_root; @@ -91,6 +103,24 @@ static int devpts_remount(struct super_b return 0; } +static int devpts_filter(struct dentry *de) +{ + return vx_check(de->d_inode->i_xid, VX_IDENT); +} + +static int devpts_readdir(struct file * filp, void * dirent, filldir_t filldir) +{ + return dcache_readdir_filter(filp, dirent, filldir, devpts_filter); +} + +static struct file_operations devpts_dir_operations = { + .open = dcache_dir_open, + .release = dcache_dir_close, + .llseek = dcache_dir_lseek, + .read = generic_read_dir, + .readdir = devpts_readdir, +}; + static struct super_operations devpts_sops = { .statfs = simple_statfs, .remount_fs = devpts_remount, @@ -117,8 +147,9 @@ devpts_fill_super(struct super_block *s, inode->i_uid = inode->i_gid = 0; inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO | S_IWUSR; inode->i_op = &simple_dir_inode_operations; - inode->i_fop = &simple_dir_operations; + inode->i_fop = &devpts_dir_operations; inode->i_nlink = 2; + inode->i_xid = vx_current_xid(); devpts_root = s->s_root = d_alloc_root(inode); if (s->s_root) @@ -177,6 +208,8 @@ int devpts_pty_new(struct tty_struct *tt inode->i_gid = config.setgid ? config.gid : current->fsgid; inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; init_special_inode(inode, S_IFCHR|config.mode, device); + inode->i_xid = vx_current_xid(); + inode->i_op = &devpts_file_inode_operations; inode->u.generic_ip = tty; dentry = get_node(number); diff -NurpP --minimal linux-2.6.17.13/fs/exec.c linux-2.6.17.13-g-v/fs/exec.c --- linux-2.6.17.13/fs/exec.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/exec.c 2006-09-20 15:00:12.000000000 +0200 @@ -49,6 +49,10 @@ #include #include #include +#include +#include +#include +#include #include #include @@ -67,6 +71,15 @@ EXPORT_SYMBOL(suid_dumpable); static struct linux_binfmt *formats; static DEFINE_RWLOCK(binfmt_lock); +#ifdef CONFIG_PAX_SOFTMODE +unsigned int pax_softmode; +#endif + +#ifdef CONFIG_PAX_HOOK_ACL_FLAGS +void (*pax_set_initial_flags_func)(struct linux_binprm * bprm); +EXPORT_SYMBOL(pax_set_initial_flags_func); +#endif + int register_binfmt(struct linux_binfmt * fmt) { struct linux_binfmt ** tmp = &formats; @@ -312,6 +325,10 @@ void install_arg_page(struct vm_area_str if (unlikely(anon_vma_prepare(vma))) goto out; +#ifdef CONFIG_PAX_SEGMEXEC + if (page_count(page) == 1) +#endif + flush_dcache_page(page); pte = get_locked_pte(mm, address, &ptl); if (!pte) @@ -321,9 +338,21 @@ void install_arg_page(struct vm_area_str goto out; } inc_mm_counter(mm, anon_rss); + +#ifdef CONFIG_PAX_SEGMEXEC + if (page_count(page) == 1) +#endif + lru_cache_add_active(page); set_pte_at(mm, address, pte, pte_mkdirty(pte_mkwrite(mk_pte( page, vma->vm_page_prot)))); + +#ifdef CONFIG_PAX_SEGMEXEC + if (page_count(page) != 1) + page_add_anon_rmap(page, vma, address); + else +#endif + page_add_new_anon_rmap(page, vma, address); pte_unmap_unlock(pte, ptl); @@ -346,6 +375,10 @@ int setup_arg_pages(struct linux_binprm int i, ret; long arg_size; +#ifdef CONFIG_PAX_SEGMEXEC + struct vm_area_struct *mpnt_m = NULL; +#endif + #ifdef CONFIG_STACK_GROWSUP /* Move the argument and environment strings to the bottom of the * stack space. @@ -404,11 +437,19 @@ int setup_arg_pages(struct linux_binprm bprm->loader += stack_base; bprm->exec += stack_base; - mpnt = kmem_cache_alloc(vm_area_cachep, SLAB_KERNEL); + mpnt = kmem_cache_zalloc(vm_area_cachep, SLAB_KERNEL); if (!mpnt) return -ENOMEM; - memset(mpnt, 0, sizeof(*mpnt)); +#ifdef CONFIG_PAX_SEGMEXEC + if ((mm->pax_flags & MF_PAX_SEGMEXEC) && (VM_STACK_FLAGS & VM_MAYEXEC)) { + mpnt_m = kmem_cache_zalloc(vm_area_cachep, SLAB_KERNEL); + if (!mpnt_m) { + kmem_cache_free(vm_area_cachep, mpnt); + return -ENOMEM; + } + } +#endif down_write(&mm->mmap_sem); { @@ -430,13 +471,51 @@ int setup_arg_pages(struct linux_binprm else mpnt->vm_flags = VM_STACK_FLAGS; mpnt->vm_flags |= mm->def_flags; - mpnt->vm_page_prot = protection_map[mpnt->vm_flags & 0x7]; + +#if defined(CONFIG_PAX_PAGEEXEC) && defined(CONFIG_X86_32) + if (!(mm->pax_flags & MF_PAX_PAGEEXEC)) + mpnt->vm_page_prot = protection_map[(mpnt->vm_flags | VM_EXEC) & (VM_READ|VM_WRITE|VM_EXEC)]; + else +#endif + + mpnt->vm_page_prot = protection_map[mpnt->vm_flags & (VM_READ|VM_WRITE|VM_EXEC)]; if ((ret = insert_vm_struct(mm, mpnt))) { up_write(&mm->mmap_sem); kmem_cache_free(vm_area_cachep, mpnt); + +#ifdef CONFIG_PAX_SEGMEXEC + if (mpnt_m) + kmem_cache_free(vm_area_cachep, mpnt_m); +#endif + return ret; } - mm->stack_vm = mm->total_vm = vma_pages(mpnt); + vx_vmpages_sub(mm, mm->total_vm - vma_pages(mpnt)); + mm->stack_vm = mm->total_vm; + +#ifdef CONFIG_PAX_SEGMEXEC + if (mpnt_m) { + *mpnt_m = *mpnt; + if (!(mpnt->vm_flags & VM_EXEC)) { + mpnt_m->vm_flags &= ~(VM_READ | VM_WRITE | VM_EXEC); + mpnt_m->vm_page_prot = PAGE_NONE; + } + mpnt_m->vm_start += SEGMEXEC_TASK_SIZE; + mpnt_m->vm_end += SEGMEXEC_TASK_SIZE; + if ((ret = insert_vm_struct(mm, mpnt_m))) { + up_write(&mm->mmap_sem); + kmem_cache_free(vm_area_cachep, mpnt_m); + return ret; + } + mpnt_m->vm_flags |= VM_MIRROR; + mpnt->vm_flags |= VM_MIRROR; + mpnt_m->vm_mirror = mpnt->vm_start - mpnt_m->vm_start; + mpnt->vm_mirror = mpnt_m->vm_start - mpnt->vm_start; + mpnt_m->vm_pgoff = mpnt->vm_pgoff; + mm->total_vm += vma_pages(mpnt_m); + } +#endif + } for (i = 0 ; i < MAX_ARG_PAGES ; i++) { @@ -444,6 +523,14 @@ int setup_arg_pages(struct linux_binprm if (page) { bprm->page[i] = NULL; install_arg_page(mpnt, page, stack_base); + +#ifdef CONFIG_PAX_SEGMEXEC + if (mpnt_m) { + page_cache_get(page); + install_arg_page(mpnt_m, page, stack_base + SEGMEXEC_TASK_SIZE); + } +#endif + } stack_base += PAGE_SIZE; } @@ -1147,6 +1234,11 @@ int do_execve(char * filename, struct file *file; int retval; int i; +#ifdef CONFIG_GRKERNSEC + struct file *old_exec_file; + struct acl_subject_label *old_acl; + struct rlimit old_rlim[RLIM_NLIMITS]; +#endif retval = -ENOMEM; bprm = kzalloc(sizeof(*bprm), GFP_KERNEL); @@ -1158,10 +1250,29 @@ int do_execve(char * filename, if (IS_ERR(file)) goto out_kfree; + gr_learn_resource(current, RLIMIT_NPROC, atomic_read(&current->user->processes), 1); + + if (gr_handle_nproc()) { + allow_write_access(file); + fput(file); + return -EAGAIN; + } + + if (!gr_acl_handle_execve(file->f_dentry, file->f_vfsmnt)) { + allow_write_access(file); + fput(file); + return -EACCES; + } + sched_exec(); bprm->p = PAGE_SIZE*MAX_ARG_PAGES-sizeof(void *); +#ifdef CONFIG_PAX_RANDUSTACK + if (randomize_va_space) + bprm->p -= (pax_get_random_long() & ~(sizeof(void *)-1)) & ~PAGE_MASK; +#endif + bprm->file = file; bprm->filename = filename; bprm->interp = filename; @@ -1203,8 +1314,38 @@ int do_execve(char * filename, if (retval < 0) goto out; + if (!gr_tpe_allow(file)) { + retval = -EACCES; + goto out; + } + + if (gr_check_crash_exec(file)) { + retval = -EACCES; + goto out; + } + + gr_log_chroot_exec(file->f_dentry, file->f_vfsmnt); + + gr_handle_exec_args(bprm, argv); + +#ifdef CONFIG_GRKERNSEC + old_acl = current->acl; + memcpy(old_rlim, current->signal->rlim, sizeof(old_rlim)); + old_exec_file = current->exec_file; + get_file(file); + current->exec_file = file; +#endif + + retval = gr_set_proc_label(file->f_dentry, file->f_vfsmnt); + if (retval < 0) + goto out_fail; + retval = search_binary_handler(bprm,regs); if (retval >= 0) { +#ifdef CONFIG_GRKERNSEC + if (old_exec_file) + fput(old_exec_file); +#endif free_arg_pages(bprm); /* execve success */ @@ -1214,6 +1355,14 @@ int do_execve(char * filename, return retval; } +out_fail: +#ifdef CONFIG_GRKERNSEC + current->acl = old_acl; + memcpy(current->signal->rlim, old_rlim, sizeof(old_rlim)); + fput(current->exec_file); + current->exec_file = old_exec_file; +#endif + out: /* Something went wrong, return the inode and free the argument pages*/ for (i = 0 ; i < MAX_ARG_PAGES ; i++) { @@ -1337,7 +1486,7 @@ static void format_corename(char *corena case 'h': down_read(&uts_sem); rc = snprintf(out_ptr, out_end - out_ptr, - "%s", system_utsname.nodename); + "%s", vx_new_uts(nodename)); up_read(&uts_sem); if (rc > out_end - out_ptr) goto out; @@ -1374,6 +1523,114 @@ static void format_corename(char *corena *out_ptr = 0; } +int pax_check_flags(unsigned long * flags) +{ + int retval = 0; + +#if !defined(__i386__) || !defined(CONFIG_PAX_SEGMEXEC) + if (*flags & MF_PAX_SEGMEXEC) + { + *flags &= ~MF_PAX_SEGMEXEC; + retval = -EINVAL; + } +#endif + + if ((*flags & MF_PAX_PAGEEXEC) + +#ifdef CONFIG_PAX_PAGEEXEC + && (*flags & MF_PAX_SEGMEXEC) +#endif + + ) + { + *flags &= ~MF_PAX_PAGEEXEC; + retval = -EINVAL; + } + + if ((*flags & MF_PAX_MPROTECT) + +#ifdef CONFIG_PAX_MPROTECT + && !(*flags & (MF_PAX_PAGEEXEC | MF_PAX_SEGMEXEC)) +#endif + + ) + { + *flags &= ~MF_PAX_MPROTECT; + retval = -EINVAL; + } + + if ((*flags & MF_PAX_EMUTRAMP) + +#ifdef CONFIG_PAX_EMUTRAMP + && !(*flags & (MF_PAX_PAGEEXEC | MF_PAX_SEGMEXEC)) +#endif + + ) + { + *flags &= ~MF_PAX_EMUTRAMP; + retval = -EINVAL; + } + + return retval; +} + +EXPORT_SYMBOL(pax_check_flags); + +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) +void pax_report_fault(struct pt_regs *regs, void *pc, void *sp) +{ + struct task_struct *tsk = current; + struct mm_struct *mm = current->mm; + char* buffer_exec = (char*)__get_free_page(GFP_ATOMIC); + char* buffer_fault = (char*)__get_free_page(GFP_ATOMIC); + char* path_exec=NULL; + char* path_fault=NULL; + unsigned long start=0UL, end=0UL, offset=0UL; + + if (buffer_exec && buffer_fault) { + struct vm_area_struct* vma, * vma_exec=NULL, * vma_fault=NULL; + + down_read(&mm->mmap_sem); + vma = mm->mmap; + while (vma && (!vma_exec || !vma_fault)) { + if ((vma->vm_flags & VM_EXECUTABLE) && vma->vm_file) + vma_exec = vma; + if (vma->vm_start <= (unsigned long)pc && (unsigned long)pc < vma->vm_end) + vma_fault = vma; + vma = vma->vm_next; + } + if (vma_exec) { + path_exec = d_path(vma_exec->vm_file->f_dentry, vma_exec->vm_file->f_vfsmnt, buffer_exec, PAGE_SIZE); + if (IS_ERR(path_exec)) + path_exec = ""; + } + if (vma_fault) { + start = vma_fault->vm_start; + end = vma_fault->vm_end; + offset = vma_fault->vm_pgoff << PAGE_SHIFT; + if (vma_fault->vm_file) { + path_fault = d_path(vma_fault->vm_file->f_dentry, vma_fault->vm_file->f_vfsmnt, buffer_fault, PAGE_SIZE); + if (IS_ERR(path_fault)) + path_fault = ""; + } else + path_fault = ""; + } + up_read(&mm->mmap_sem); + } + if (tsk->signal->curr_ip) + printk(KERN_ERR "PAX: From %u.%u.%u.%u: execution attempt in: %s, %08lx-%08lx %08lx\n", NIPQUAD(tsk->signal->curr_ip), path_fault, start, end, offset); + else + printk(KERN_ERR "PAX: execution attempt in: %s, %08lx-%08lx %08lx\n", path_fault, start, end, offset); + printk(KERN_ERR "PAX: terminating task: %s(%s):%d, uid/euid: %u/%u, " + "PC: %p, SP: %p\n", path_exec, tsk->comm, tsk->pid, + tsk->uid, tsk->euid, pc, sp); + free_page((unsigned long)buffer_exec); + free_page((unsigned long)buffer_fault); + pax_report_insns(pc, sp); + do_coredump(SIGKILL, SIGKILL, regs); +} +#endif + static void zap_threads (struct mm_struct *mm) { struct task_struct *g, *p; @@ -1491,6 +1748,10 @@ int do_coredump(long signr, int exit_cod */ clear_thread_flag(TIF_SIGPENDING); + if (signr == SIGKILL || signr == SIGILL) + gr_handle_brute_attach(current); + + gr_learn_resource(current, RLIMIT_CORE, binfmt->min_coredump, 1); if (current->signal->rlim[RLIMIT_CORE].rlim_cur < binfmt->min_coredump) goto fail_unlock; diff -NurpP --minimal linux-2.6.17.13/fs/ext2/balloc.c linux-2.6.17.13-g-v/fs/ext2/balloc.c --- linux-2.6.17.13/fs/ext2/balloc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext2/balloc.c 2006-09-20 15:00:12.000000000 +0200 @@ -17,6 +17,7 @@ #include #include #include +#include /* * balloc.c contains the blocks allocation and deallocation routines @@ -109,6 +110,8 @@ static int reserve_blocks(struct super_b free_blocks = percpu_counter_read_positive(&sbi->s_freeblocks_counter); root_blocks = le32_to_cpu(es->s_r_blocks_count); + DLIMIT_ADJUST_BLOCK(sb, vx_current_xid(), &free_blocks, &root_blocks); + if (free_blocks < count) count = free_blocks; @@ -259,6 +262,7 @@ do_more: } error_return: brelse(bitmap_bh); + DLIMIT_FREE_BLOCK(inode, freed); release_blocks(sb, freed); DQUOT_FREE_BLOCK(inode, freed); } @@ -362,6 +366,10 @@ int ext2_new_block(struct inode *inode, *err = -ENOSPC; goto out_dquot; } + if (DLIMIT_ALLOC_BLOCK(inode, es_alloc)) { + *err = -ENOSPC; + goto out_dlimit; + } ext2_debug ("goal=%lu.\n", goal); @@ -509,6 +517,8 @@ got_block: *err = 0; out_release: group_release_blocks(sb, group_no, desc, gdp_bh, group_alloc); + DLIMIT_FREE_BLOCK(inode, es_alloc); +out_dlimit: release_blocks(sb, es_alloc); out_dquot: DQUOT_FREE_BLOCK(inode, dq_alloc); diff -NurpP --minimal linux-2.6.17.13/fs/ext2/ext2.h linux-2.6.17.13-g-v/fs/ext2/ext2.h --- linux-2.6.17.13/fs/ext2/ext2.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext2/ext2.h 2006-09-20 15:00:12.000000000 +0200 @@ -165,6 +165,7 @@ extern const struct file_operations ext2 extern struct address_space_operations ext2_aops; extern struct address_space_operations ext2_aops_xip; extern struct address_space_operations ext2_nobh_aops; +extern int ext2_sync_flags(struct inode *inode); /* namei.c */ extern struct inode_operations ext2_dir_inode_operations; diff -NurpP --minimal linux-2.6.17.13/fs/ext2/file.c linux-2.6.17.13-g-v/fs/ext2/file.c --- linux-2.6.17.13/fs/ext2/file.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext2/file.c 2006-09-20 15:00:12.000000000 +0200 @@ -81,4 +81,5 @@ struct inode_operations ext2_file_inode_ #endif .setattr = ext2_setattr, .permission = ext2_permission, + .sync_flags = ext2_sync_flags, }; diff -NurpP --minimal linux-2.6.17.13/fs/ext2/ialloc.c linux-2.6.17.13-g-v/fs/ext2/ialloc.c --- linux-2.6.17.13/fs/ext2/ialloc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext2/ialloc.c 2006-09-20 15:00:12.000000000 +0200 @@ -18,6 +18,8 @@ #include #include #include +#include +#include #include "ext2.h" #include "xattr.h" #include "acl.h" @@ -126,6 +128,7 @@ void ext2_free_inode (struct inode * ino ext2_xattr_delete_inode(inode); DQUOT_FREE_INODE(inode); DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); } es = EXT2_SB(sb)->s_es; @@ -465,6 +468,11 @@ struct inode *ext2_new_inode(struct inod if (!inode) return ERR_PTR(-ENOMEM); + inode->i_xid = vx_current_fsxid(sb); + if (DLIMIT_ALLOC_INODE(inode)) { + err = -ENOSPC; + goto fail_dlim; + } ei = EXT2_I(inode); sbi = EXT2_SB(sb); es = sbi->s_es; @@ -579,7 +587,8 @@ got: inode->i_blocks = 0; inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC; memset(ei->i_data, 0, sizeof(ei->i_data)); - ei->i_flags = EXT2_I(dir)->i_flags & ~EXT2_BTREE_FL; + ei->i_flags = EXT2_I(dir)->i_flags & + ~(EXT2_BTREE_FL|EXT2_IUNLINK_FL|EXT2_BARRIER_FL); if (S_ISLNK(mode)) ei->i_flags &= ~(EXT2_IMMUTABLE_FL|EXT2_APPEND_FL); /* dirsync is only applied to directories */ @@ -627,12 +636,15 @@ fail_free_drop: fail_drop: DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); inode->i_flags |= S_NOQUOTA; inode->i_nlink = 0; iput(inode); return ERR_PTR(err); fail: + DLIMIT_FREE_INODE(inode); +fail_dlim: make_bad_inode(inode); iput(inode); return ERR_PTR(err); diff -NurpP --minimal linux-2.6.17.13/fs/ext2/inode.c linux-2.6.17.13-g-v/fs/ext2/inode.c --- linux-2.6.17.13/fs/ext2/inode.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext2/inode.c 2006-09-20 15:00:12.000000000 +0200 @@ -31,6 +31,7 @@ #include #include #include +#include #include "ext2.h" #include "acl.h" #include "xip.h" @@ -1042,25 +1043,70 @@ void ext2_set_inode_flags(struct inode * { unsigned int flags = EXT2_I(inode)->i_flags; - inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC); + inode->i_flags &= ~(S_IMMUTABLE | S_IUNLINK | S_BARRIER | + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC); + + if (flags & EXT2_IMMUTABLE_FL) + inode->i_flags |= S_IMMUTABLE; + if (flags & EXT2_IUNLINK_FL) + inode->i_flags |= S_IUNLINK; + if (flags & EXT2_BARRIER_FL) + inode->i_flags |= S_BARRIER; + if (flags & EXT2_SYNC_FL) inode->i_flags |= S_SYNC; if (flags & EXT2_APPEND_FL) inode->i_flags |= S_APPEND; - if (flags & EXT2_IMMUTABLE_FL) - inode->i_flags |= S_IMMUTABLE; if (flags & EXT2_NOATIME_FL) inode->i_flags |= S_NOATIME; if (flags & EXT2_DIRSYNC_FL) inode->i_flags |= S_DIRSYNC; } +int ext2_sync_flags(struct inode *inode) +{ + unsigned int oldflags, newflags; + + oldflags = EXT2_I(inode)->i_flags; + newflags = oldflags & ~(EXT2_APPEND_FL | + EXT2_IMMUTABLE_FL | EXT2_IUNLINK_FL | + EXT2_BARRIER_FL | EXT2_NOATIME_FL | + EXT2_SYNC_FL | EXT2_DIRSYNC_FL); + + if (IS_APPEND(inode)) + newflags |= EXT2_APPEND_FL; + if (IS_IMMUTABLE(inode)) + newflags |= EXT2_IMMUTABLE_FL; + if (IS_IUNLINK(inode)) + newflags |= EXT2_IUNLINK_FL; + if (IS_BARRIER(inode)) + newflags |= EXT2_BARRIER_FL; + + /* we do not want to copy superblock flags */ + if (inode->i_flags & S_NOATIME) + newflags |= EXT2_NOATIME_FL; + if (inode->i_flags & S_SYNC) + newflags |= EXT2_SYNC_FL; + if (inode->i_flags & S_DIRSYNC) + newflags |= EXT2_DIRSYNC_FL; + + if (oldflags ^ newflags) { + EXT2_I(inode)->i_flags = newflags; + inode->i_ctime = CURRENT_TIME; + mark_inode_dirty(inode); + } + + return 0; +} + void ext2_read_inode (struct inode * inode) { struct ext2_inode_info *ei = EXT2_I(inode); ino_t ino = inode->i_ino; struct buffer_head * bh; struct ext2_inode * raw_inode = ext2_get_inode(inode->i_sb, ino, &bh); + uid_t uid; + gid_t gid; int n; #ifdef CONFIG_EXT2_FS_POSIX_ACL @@ -1071,12 +1117,17 @@ void ext2_read_inode (struct inode * ino goto bad_inode; inode->i_mode = le16_to_cpu(raw_inode->i_mode); - inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low); - inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low); + uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low); + gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low); if (!(test_opt (inode->i_sb, NO_UID32))) { - inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16; - inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16; + uid |= le16_to_cpu(raw_inode->i_uid_high) << 16; + gid |= le16_to_cpu(raw_inode->i_gid_high) << 16; } + inode->i_uid = INOXID_UID(XID_TAG(inode), uid, gid); + inode->i_gid = INOXID_GID(XID_TAG(inode), uid, gid); + inode->i_xid = INOXID_XID(XID_TAG(inode), uid, gid, + le16_to_cpu(raw_inode->i_raw_xid)); + inode->i_nlink = le16_to_cpu(raw_inode->i_links_count); inode->i_size = le32_to_cpu(raw_inode->i_size); inode->i_atime.tv_sec = le32_to_cpu(raw_inode->i_atime); @@ -1174,8 +1225,8 @@ static int ext2_update_inode(struct inod struct ext2_inode_info *ei = EXT2_I(inode); struct super_block *sb = inode->i_sb; ino_t ino = inode->i_ino; - uid_t uid = inode->i_uid; - gid_t gid = inode->i_gid; + uid_t uid = XIDINO_UID(XID_TAG(inode), inode->i_uid, inode->i_xid); + gid_t gid = XIDINO_GID(XID_TAG(inode), inode->i_gid, inode->i_xid); struct buffer_head * bh; struct ext2_inode * raw_inode = ext2_get_inode(sb, ino, &bh); int n; @@ -1210,6 +1261,9 @@ static int ext2_update_inode(struct inod raw_inode->i_uid_high = 0; raw_inode->i_gid_high = 0; } +#ifdef CONFIG_INOXID_INTERN + raw_inode->i_raw_xid = cpu_to_le16(inode->i_xid); +#endif raw_inode->i_links_count = cpu_to_le16(inode->i_nlink); raw_inode->i_size = cpu_to_le32(inode->i_size); raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec); @@ -1296,7 +1350,8 @@ int ext2_setattr(struct dentry *dentry, if (error) return error; if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) || - (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) { + (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) || + (iattr->ia_valid & ATTR_XID && iattr->ia_xid != inode->i_xid)) { error = DQUOT_TRANSFER(inode, iattr) ? -EDQUOT : 0; if (error) return error; diff -NurpP --minimal linux-2.6.17.13/fs/ext2/ioctl.c linux-2.6.17.13-g-v/fs/ext2/ioctl.c --- linux-2.6.17.13/fs/ext2/ioctl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext2/ioctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -11,6 +11,7 @@ #include #include #include +#include #include #include @@ -30,7 +31,8 @@ int ext2_ioctl (struct inode * inode, st case EXT2_IOC_SETFLAGS: { unsigned int oldflags; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) @@ -50,7 +52,9 @@ int ext2_ioctl (struct inode * inode, st * * This test looks nicer. Thanks to Pauline Middelink */ - if ((flags ^ oldflags) & (EXT2_APPEND_FL | EXT2_IMMUTABLE_FL)) { + if ((oldflags & EXT2_IMMUTABLE_FL) || + ((flags ^ oldflags) & (EXT2_APPEND_FL | + EXT2_IMMUTABLE_FL | EXT2_IUNLINK_FL))) { if (!capable(CAP_LINUX_IMMUTABLE)) return -EPERM; } @@ -69,7 +73,8 @@ int ext2_ioctl (struct inode * inode, st case EXT2_IOC_SETVERSION: if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (get_user(inode->i_generation, (int __user *) arg)) return -EFAULT; diff -NurpP --minimal linux-2.6.17.13/fs/ext2/namei.c linux-2.6.17.13-g-v/fs/ext2/namei.c --- linux-2.6.17.13/fs/ext2/namei.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext2/namei.c 2006-09-20 15:00:12.000000000 +0200 @@ -31,6 +31,7 @@ */ #include +#include #include "ext2.h" #include "xattr.h" #include "acl.h" @@ -66,6 +67,7 @@ static struct dentry *ext2_lookup(struct inode = iget(dir->i_sb, ino); if (!inode) return ERR_PTR(-EACCES); + vx_propagate_xid(nd, inode); } return d_splice_alias(inode, dentry); } @@ -391,6 +393,7 @@ struct inode_operations ext2_dir_inode_o #endif .setattr = ext2_setattr, .permission = ext2_permission, + .sync_flags = ext2_sync_flags, }; struct inode_operations ext2_special_inode_operations = { @@ -402,4 +405,5 @@ struct inode_operations ext2_special_ino #endif .setattr = ext2_setattr, .permission = ext2_permission, + .sync_flags = ext2_sync_flags, }; diff -NurpP --minimal linux-2.6.17.13/fs/ext2/super.c linux-2.6.17.13-g-v/fs/ext2/super.c --- linux-2.6.17.13/fs/ext2/super.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext2/super.c 2006-09-20 15:00:12.000000000 +0200 @@ -328,7 +328,7 @@ enum { Opt_err_ro, Opt_nouid32, Opt_nocheck, Opt_debug, Opt_oldalloc, Opt_orlov, Opt_nobh, Opt_user_xattr, Opt_nouser_xattr, Opt_acl, Opt_noacl, Opt_xip, Opt_ignore, Opt_err, Opt_quota, - Opt_usrquota, Opt_grpquota + Opt_usrquota, Opt_grpquota, Opt_tagxid }; static match_table_t tokens = { @@ -356,6 +356,7 @@ static match_table_t tokens = { {Opt_acl, "acl"}, {Opt_noacl, "noacl"}, {Opt_xip, "xip"}, + {Opt_tagxid, "tagxid"}, {Opt_grpquota, "grpquota"}, {Opt_ignore, "noquota"}, {Opt_quota, "quota"}, @@ -419,6 +420,11 @@ static int parse_options (char * options case Opt_nouid32: set_opt (sbi->s_mount_opt, NO_UID32); break; +#ifndef CONFIG_INOXID_NONE + case Opt_tagxid: + set_opt (sbi->s_mount_opt, TAGXID); + break; +#endif case Opt_nocheck: clear_opt (sbi->s_mount_opt, CHECK); break; @@ -720,6 +726,8 @@ static int ext2_fill_super(struct super_ if (!parse_options ((char *) data, sbi)) goto failed_mount; + if (EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_TAGXID) + sb->s_flags |= MS_TAGXID; sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | ((EXT2_SB(sb)->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); @@ -1029,6 +1037,13 @@ static int ext2_remount (struct super_bl goto restore_opts; } + if ((sbi->s_mount_opt & EXT2_MOUNT_TAGXID) && + !(sb->s_flags & MS_TAGXID)) { + printk("EXT2-fs: %s: tagxid not permitted on remount.\n", + sb->s_id); + return -EINVAL; + } + sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | ((sbi->s_mount_opt & EXT2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); diff -NurpP --minimal linux-2.6.17.13/fs/ext2/symlink.c linux-2.6.17.13-g-v/fs/ext2/symlink.c --- linux-2.6.17.13/fs/ext2/symlink.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext2/symlink.c 2006-09-20 15:00:12.000000000 +0200 @@ -38,6 +38,7 @@ struct inode_operations ext2_symlink_ino .listxattr = ext2_listxattr, .removexattr = generic_removexattr, #endif + .sync_flags = ext2_sync_flags, }; struct inode_operations ext2_fast_symlink_inode_operations = { @@ -49,4 +50,5 @@ struct inode_operations ext2_fast_symlin .listxattr = ext2_listxattr, .removexattr = generic_removexattr, #endif + .sync_flags = ext2_sync_flags, }; diff -NurpP --minimal linux-2.6.17.13/fs/ext2/xattr.c linux-2.6.17.13-g-v/fs/ext2/xattr.c --- linux-2.6.17.13/fs/ext2/xattr.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext2/xattr.c 2006-09-20 15:00:12.000000000 +0200 @@ -60,6 +60,7 @@ #include #include #include +#include #include "ext2.h" #include "xattr.h" #include "acl.h" @@ -645,8 +646,12 @@ ext2_xattr_set2(struct inode *inode, str the inode. */ ea_bdebug(new_bh, "reusing block"); + error = -ENOSPC; + if (DLIMIT_ALLOC_BLOCK(inode, 1)) + goto cleanup; error = -EDQUOT; if (DQUOT_ALLOC_BLOCK(inode, 1)) { + DLIMIT_FREE_BLOCK(inode, 1); unlock_buffer(new_bh); goto cleanup; } @@ -740,6 +745,7 @@ ext2_xattr_set2(struct inode *inode, str le32_to_cpu(HDR(old_bh)->h_refcount) - 1); if (ce) mb_cache_entry_release(ce); + DLIMIT_FREE_BLOCK(inode, 1); DQUOT_FREE_BLOCK(inode, 1); mark_buffer_dirty(old_bh); ea_bdebug(old_bh, "refcount now=%d", @@ -804,6 +810,7 @@ ext2_xattr_delete_inode(struct inode *in mark_buffer_dirty(bh); if (IS_SYNC(inode)) sync_dirty_buffer(bh); + DLIMIT_FREE_BLOCK(inode, 1); DQUOT_FREE_BLOCK(inode, 1); } EXT2_I(inode)->i_file_acl = 0; diff -NurpP --minimal linux-2.6.17.13/fs/ext3/balloc.c linux-2.6.17.13-g-v/fs/ext3/balloc.c --- linux-2.6.17.13/fs/ext3/balloc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext3/balloc.c 2006-09-20 15:00:12.000000000 +0200 @@ -20,6 +20,7 @@ #include #include #include +#include /* * balloc.c contains the blocks allocation and deallocation routines @@ -504,8 +505,10 @@ void ext3_free_blocks(handle_t *handle, return; } ext3_free_blocks_sb(handle, sb, block, count, &dquot_freed_blocks); - if (dquot_freed_blocks) + if (dquot_freed_blocks) { + DLIMIT_FREE_BLOCK(inode, dquot_freed_blocks); DQUOT_FREE_BLOCK(inode, dquot_freed_blocks); + } return; } @@ -1162,18 +1165,32 @@ out: return ret; } -static int ext3_has_free_blocks(struct ext3_sb_info *sbi) +static int ext3_has_free_blocks(struct super_block *sb) { - int free_blocks, root_blocks; + struct ext3_sb_info *sbi = EXT3_SB(sb); + int free_blocks, root_blocks, cond; free_blocks = percpu_counter_read_positive(&sbi->s_freeblocks_counter); root_blocks = le32_to_cpu(sbi->s_es->s_r_blocks_count); - if (free_blocks < root_blocks + 1 && !capable(CAP_SYS_RESOURCE) && + + vxdprintk(VXD_CBIT(dlim, 3), + "ext3_has_free_blocks(%p): free=%u, root=%u", + sb, free_blocks, root_blocks); + + DLIMIT_ADJUST_BLOCK(sb, vx_current_xid(), &free_blocks, &root_blocks); + + cond = (free_blocks < root_blocks + 1 && + !capable(CAP_SYS_RESOURCE) && sbi->s_resuid != current->fsuid && - (sbi->s_resgid == 0 || !in_group_p (sbi->s_resgid))) { - return 0; - } - return 1; + (sbi->s_resgid == 0 || !in_group_p (sbi->s_resgid))); + + vxdprintk(VXD_CBIT(dlim, 3), + "ext3_has_free_blocks(%p): %u<%u+1, %c, %u!=%u r=%d", + sb, free_blocks, root_blocks, + !capable(CAP_SYS_RESOURCE)?'1':'0', + sbi->s_resuid, current->fsuid, cond?0:1); + + return (cond ? 0 : 1); } /* @@ -1184,7 +1201,7 @@ static int ext3_has_free_blocks(struct e */ int ext3_should_retry_alloc(struct super_block *sb, int *retries) { - if (!ext3_has_free_blocks(EXT3_SB(sb)) || (*retries)++ > 3) + if (!ext3_has_free_blocks(sb) || (*retries)++ > 3) return 0; jbd_debug(1, "%s: retrying operation after ENOSPC\n", sb->s_id); @@ -1240,6 +1257,8 @@ int ext3_new_blocks(handle_t *handle, st *errp = -EDQUOT; return 0; } + if (DLIMIT_ALLOC_BLOCK(inode, 1)) + goto out_dlimit; sbi = EXT3_SB(sb); es = EXT3_SB(sb)->s_es; @@ -1256,7 +1275,7 @@ int ext3_new_blocks(handle_t *handle, st if (block_i && ((windowsz = block_i->rsv_window_node.rsv_goal_size) > 0)) my_rsv = &block_i->rsv_window_node; - if (!ext3_has_free_blocks(sbi)) { + if (!ext3_has_free_blocks(sb)) { *errp = -ENOSPC; goto out; } @@ -1448,6 +1467,9 @@ allocated: io_error: *errp = -EIO; out: + if (!performed_allocation) + DLIMIT_FREE_BLOCK(inode, 1); +out_dlimit: if (fatal) { *errp = fatal; ext3_std_error(sb, fatal); diff -NurpP --minimal linux-2.6.17.13/fs/ext3/file.c linux-2.6.17.13-g-v/fs/ext3/file.c --- linux-2.6.17.13/fs/ext3/file.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext3/file.c 2006-09-20 15:00:12.000000000 +0200 @@ -133,5 +133,6 @@ struct inode_operations ext3_file_inode_ .removexattr = generic_removexattr, #endif .permission = ext3_permission, + .sync_flags = ext3_sync_flags, }; diff -NurpP --minimal linux-2.6.17.13/fs/ext3/ialloc.c linux-2.6.17.13-g-v/fs/ext3/ialloc.c --- linux-2.6.17.13/fs/ext3/ialloc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext3/ialloc.c 2006-09-20 15:00:12.000000000 +0200 @@ -23,6 +23,8 @@ #include #include #include +#include +#include #include @@ -127,6 +129,7 @@ void ext3_free_inode (handle_t *handle, ext3_xattr_delete_inode(handle, inode); DQUOT_FREE_INODE(inode); DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); is_directory = S_ISDIR(inode->i_mode); @@ -443,6 +446,12 @@ struct inode *ext3_new_inode(handle_t *h inode = new_inode(sb); if (!inode) return ERR_PTR(-ENOMEM); + + inode->i_xid = vx_current_fsxid(sb); + if (DLIMIT_ALLOC_INODE(inode)) { + err = -ENOSPC; + goto out_dlimit; + } ei = EXT3_I(inode); sbi = EXT3_SB(sb); @@ -565,7 +574,8 @@ got: ei->i_dir_start_lookup = 0; ei->i_disksize = 0; - ei->i_flags = EXT3_I(dir)->i_flags & ~EXT3_INDEX_FL; + ei->i_flags = EXT3_I(dir)->i_flags & + ~(EXT3_INDEX_FL|EXT3_IUNLINK_FL|EXT3_BARRIER_FL); if (S_ISLNK(mode)) ei->i_flags &= ~(EXT3_IMMUTABLE_FL|EXT3_APPEND_FL); /* dirsync only applies to directories */ @@ -620,6 +630,8 @@ got: fail: ext3_std_error(sb, err); out: + DLIMIT_FREE_INODE(inode); +out_dlimit: iput(inode); ret = ERR_PTR(err); really_out: @@ -631,6 +643,7 @@ fail_free_drop: fail_drop: DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); inode->i_flags |= S_NOQUOTA; inode->i_nlink = 0; iput(inode); diff -NurpP --minimal linux-2.6.17.13/fs/ext3/inode.c linux-2.6.17.13-g-v/fs/ext3/inode.c --- linux-2.6.17.13/fs/ext3/inode.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext3/inode.c 2006-09-20 15:00:12.000000000 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include "xattr.h" #include "acl.h" @@ -2563,19 +2564,77 @@ void ext3_set_inode_flags(struct inode * { unsigned int flags = EXT3_I(inode)->i_flags; - inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC); + inode->i_flags &= ~(S_IMMUTABLE | S_IUNLINK | S_BARRIER | + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC); + + if (flags & EXT3_IMMUTABLE_FL) + inode->i_flags |= S_IMMUTABLE; + if (flags & EXT3_IUNLINK_FL) + inode->i_flags |= S_IUNLINK; + if (flags & EXT3_BARRIER_FL) + inode->i_flags |= S_BARRIER; + if (flags & EXT3_SYNC_FL) inode->i_flags |= S_SYNC; if (flags & EXT3_APPEND_FL) inode->i_flags |= S_APPEND; - if (flags & EXT3_IMMUTABLE_FL) - inode->i_flags |= S_IMMUTABLE; if (flags & EXT3_NOATIME_FL) inode->i_flags |= S_NOATIME; if (flags & EXT3_DIRSYNC_FL) inode->i_flags |= S_DIRSYNC; } +int ext3_sync_flags(struct inode *inode) +{ + unsigned int oldflags, newflags; + int err = 0; + + oldflags = EXT3_I(inode)->i_flags; + newflags = oldflags & ~(EXT3_APPEND_FL | + EXT3_IMMUTABLE_FL | EXT3_IUNLINK_FL | + EXT3_BARRIER_FL | EXT3_NOATIME_FL | + EXT3_SYNC_FL | EXT3_DIRSYNC_FL); + + if (IS_APPEND(inode)) + newflags |= EXT3_APPEND_FL; + if (IS_IMMUTABLE(inode)) + newflags |= EXT3_IMMUTABLE_FL; + if (IS_IUNLINK(inode)) + newflags |= EXT3_IUNLINK_FL; + if (IS_BARRIER(inode)) + newflags |= EXT3_BARRIER_FL; + + /* we do not want to copy superblock flags */ + if (inode->i_flags & S_NOATIME) + newflags |= EXT3_NOATIME_FL; + if (inode->i_flags & S_SYNC) + newflags |= EXT3_SYNC_FL; + if (inode->i_flags & S_DIRSYNC) + newflags |= EXT3_DIRSYNC_FL; + + if (oldflags ^ newflags) { + handle_t *handle; + struct ext3_iloc iloc; + + handle = ext3_journal_start(inode, 1); + if (IS_ERR(handle)) + return PTR_ERR(handle); + if (IS_SYNC(inode)) + handle->h_sync = 1; + err = ext3_reserve_inode_write(handle, inode, &iloc); + if (err) + goto flags_err; + + EXT3_I(inode)->i_flags = newflags; + inode->i_ctime = CURRENT_TIME; + + err = ext3_mark_iloc_dirty(handle, inode, &iloc); + flags_err: + ext3_journal_stop(handle); + } + return err; +} + void ext3_read_inode(struct inode * inode) { struct ext3_iloc iloc; @@ -2583,6 +2642,8 @@ void ext3_read_inode(struct inode * inod struct ext3_inode_info *ei = EXT3_I(inode); struct buffer_head *bh; int block; + uid_t uid; + gid_t gid; #ifdef CONFIG_EXT3_FS_POSIX_ACL ei->i_acl = EXT3_ACL_NOT_CACHED; @@ -2595,12 +2656,17 @@ void ext3_read_inode(struct inode * inod bh = iloc.bh; raw_inode = ext3_raw_inode(&iloc); inode->i_mode = le16_to_cpu(raw_inode->i_mode); - inode->i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low); - inode->i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low); + uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low); + gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low); if(!(test_opt (inode->i_sb, NO_UID32))) { - inode->i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16; - inode->i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16; + uid |= le16_to_cpu(raw_inode->i_uid_high) << 16; + gid |= le16_to_cpu(raw_inode->i_gid_high) << 16; } + inode->i_uid = INOXID_UID(XID_TAG(inode), uid, gid); + inode->i_gid = INOXID_GID(XID_TAG(inode), uid, gid); + inode->i_xid = INOXID_XID(XID_TAG(inode), uid, gid, + le16_to_cpu(raw_inode->i_raw_xid)); + inode->i_nlink = le16_to_cpu(raw_inode->i_links_count); inode->i_size = le32_to_cpu(raw_inode->i_size); inode->i_atime.tv_sec = le32_to_cpu(raw_inode->i_atime); @@ -2727,6 +2793,8 @@ static int ext3_do_update_inode(handle_t struct ext3_inode *raw_inode = ext3_raw_inode(iloc); struct ext3_inode_info *ei = EXT3_I(inode); struct buffer_head *bh = iloc->bh; + uid_t uid = XIDINO_UID(XID_TAG(inode), inode->i_uid, inode->i_xid); + gid_t gid = XIDINO_GID(XID_TAG(inode), inode->i_gid, inode->i_xid); int err = 0, rc, block; /* For fields not not tracking in the in-memory inode, @@ -2736,29 +2804,32 @@ static int ext3_do_update_inode(handle_t raw_inode->i_mode = cpu_to_le16(inode->i_mode); if(!(test_opt(inode->i_sb, NO_UID32))) { - raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid)); - raw_inode->i_gid_low = cpu_to_le16(low_16_bits(inode->i_gid)); + raw_inode->i_uid_low = cpu_to_le16(low_16_bits(uid)); + raw_inode->i_gid_low = cpu_to_le16(low_16_bits(gid)); /* * Fix up interoperability with old kernels. Otherwise, old inodes get * re-used with the upper 16 bits of the uid/gid intact */ if(!ei->i_dtime) { raw_inode->i_uid_high = - cpu_to_le16(high_16_bits(inode->i_uid)); + cpu_to_le16(high_16_bits(uid)); raw_inode->i_gid_high = - cpu_to_le16(high_16_bits(inode->i_gid)); + cpu_to_le16(high_16_bits(gid)); } else { raw_inode->i_uid_high = 0; raw_inode->i_gid_high = 0; } } else { raw_inode->i_uid_low = - cpu_to_le16(fs_high2lowuid(inode->i_uid)); + cpu_to_le16(fs_high2lowuid(uid)); raw_inode->i_gid_low = - cpu_to_le16(fs_high2lowgid(inode->i_gid)); + cpu_to_le16(fs_high2lowgid(gid)); raw_inode->i_uid_high = 0; raw_inode->i_gid_high = 0; } +#ifdef CONFIG_INOXID_INTERN + raw_inode->i_raw_xid = cpu_to_le16(inode->i_xid); +#endif raw_inode->i_links_count = cpu_to_le16(inode->i_nlink); raw_inode->i_size = cpu_to_le32(ei->i_disksize); raw_inode->i_atime = cpu_to_le32(inode->i_atime.tv_sec); @@ -2911,7 +2982,8 @@ int ext3_setattr(struct dentry *dentry, return error; if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) || - (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) { + (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) || + (ia_valid & ATTR_XID && attr->ia_xid != inode->i_xid)) { handle_t *handle; /* (user+group)*(old+new) structure, inode write (sb, @@ -2933,6 +3005,8 @@ int ext3_setattr(struct dentry *dentry, inode->i_uid = attr->ia_uid; if (attr->ia_valid & ATTR_GID) inode->i_gid = attr->ia_gid; + if ((attr->ia_valid & ATTR_XID) && IS_TAGXID(inode)) + inode->i_xid = attr->ia_xid; error = ext3_mark_inode_dirty(handle, inode); ext3_journal_stop(handle); } diff -NurpP --minimal linux-2.6.17.13/fs/ext3/ioctl.c linux-2.6.17.13-g-v/fs/ext3/ioctl.c --- linux-2.6.17.13/fs/ext3/ioctl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext3/ioctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -8,11 +8,13 @@ */ #include +#include #include #include #include #include #include +#include #include @@ -36,7 +38,8 @@ int ext3_ioctl (struct inode * inode, st unsigned int oldflags; unsigned int jflag; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) @@ -60,7 +63,9 @@ int ext3_ioctl (struct inode * inode, st * * This test looks nicer. Thanks to Pauline Middelink */ - if ((flags ^ oldflags) & (EXT3_APPEND_FL | EXT3_IMMUTABLE_FL)) { + if ((oldflags & EXT3_IMMUTABLE_FL) || + ((flags ^ oldflags) & (EXT3_APPEND_FL | + EXT3_IMMUTABLE_FL | EXT3_IUNLINK_FL))) { if (!capable(CAP_LINUX_IMMUTABLE)) { mutex_unlock(&inode->i_mutex); return -EPERM; @@ -122,7 +127,8 @@ flags_err: if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (get_user(generation, (int __user *) arg)) return -EFAULT; @@ -176,7 +182,8 @@ flags_err: if (!test_opt(inode->i_sb, RESERVATION) ||!S_ISREG(inode->i_mode)) return -ENOTTY; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) @@ -211,7 +218,8 @@ flags_err: if (!capable(CAP_SYS_RESOURCE)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (get_user(n_blocks_count, (__u32 __user *)arg)) @@ -232,7 +240,8 @@ flags_err: if (!capable(CAP_SYS_RESOURCE)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (copy_from_user(&input, (struct ext3_new_group_input __user *)arg, @@ -247,6 +256,38 @@ flags_err: return err; } +#if defined(CONFIG_VSERVER_LEGACY) && !defined(CONFIG_INOXID_NONE) + case EXT3_IOC_SETXID: { + handle_t *handle; + struct ext3_iloc iloc; + int xid; + int err; + + /* fixme: if stealth, return -ENOTTY */ + if (!capable(CAP_CONTEXT)) + return -EPERM; + if (IS_RDONLY(inode)) + return -EROFS; + if (!(inode->i_sb->s_flags & MS_TAGXID)) + return -ENOSYS; + if (get_user(xid, (int __user *) arg)) + return -EFAULT; + + handle = ext3_journal_start(inode, 1); + if (IS_ERR(handle)) + return PTR_ERR(handle); + err = ext3_reserve_inode_write(handle, inode, &iloc); + if (err) + return err; + + inode->i_xid = (xid & 0xFFFF); + inode->i_ctime = CURRENT_TIME; + + err = ext3_mark_iloc_dirty(handle, inode, &iloc); + ext3_journal_stop(handle); + return err; + } +#endif default: return -ENOTTY; diff -NurpP --minimal linux-2.6.17.13/fs/ext3/namei.c linux-2.6.17.13-g-v/fs/ext3/namei.c --- linux-2.6.17.13/fs/ext3/namei.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext3/namei.c 2006-09-20 15:00:12.000000000 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include "namei.h" #include "xattr.h" @@ -1009,6 +1010,7 @@ static struct dentry *ext3_lookup(struct if (!inode) return ERR_PTR(-EACCES); + vx_propagate_xid(nd, inode); } return d_splice_alias(inode, dentry); } @@ -2384,6 +2386,7 @@ struct inode_operations ext3_dir_inode_o .removexattr = generic_removexattr, #endif .permission = ext3_permission, + .sync_flags = ext3_sync_flags, }; struct inode_operations ext3_special_inode_operations = { @@ -2395,4 +2398,5 @@ struct inode_operations ext3_special_ino .removexattr = generic_removexattr, #endif .permission = ext3_permission, + .sync_flags = ext3_sync_flags, }; diff -NurpP --minimal linux-2.6.17.13/fs/ext3/super.c linux-2.6.17.13-g-v/fs/ext3/super.c --- linux-2.6.17.13/fs/ext3/super.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext3/super.c 2006-09-20 15:00:12.000000000 +0200 @@ -675,7 +675,7 @@ enum { Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota, Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_quota, Opt_noquota, Opt_ignore, Opt_barrier, Opt_err, Opt_resize, Opt_usrquota, - Opt_grpquota + Opt_grpquota, Opt_tagxid }; static match_table_t tokens = { @@ -724,6 +724,7 @@ static match_table_t tokens = { {Opt_quota, "quota"}, {Opt_usrquota, "usrquota"}, {Opt_barrier, "barrier=%u"}, + {Opt_tagxid, "tagxid"}, {Opt_err, NULL}, {Opt_resize, "resize"}, }; @@ -816,6 +817,11 @@ static int parse_options (char *options, case Opt_nouid32: set_opt (sbi->s_mount_opt, NO_UID32); break; +#ifndef CONFIG_INOXID_NONE + case Opt_tagxid: + set_opt (sbi->s_mount_opt, TAGXID); + break; +#endif case Opt_nocheck: clear_opt (sbi->s_mount_opt, CHECK); break; @@ -1470,6 +1476,9 @@ static int ext3_fill_super (struct super NULL, 0)) goto failed_mount; + if (EXT3_SB(sb)->s_mount_opt & EXT3_MOUNT_TAGXID) + sb->s_flags |= MS_TAGXID; + sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | ((sbi->s_mount_opt & EXT3_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); @@ -2273,6 +2282,12 @@ static int ext3_remount (struct super_bl if (sbi->s_mount_opt & EXT3_MOUNT_ABORT) ext3_abort(sb, __FUNCTION__, "Abort forced by user"); + if ((sbi->s_mount_opt & EXT3_MOUNT_TAGXID) && + !(sb->s_flags & MS_TAGXID)) { + printk("EXT3-fs: %s: tagxid not permitted on remount.\n", + sb->s_id); + return -EINVAL; + } sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | ((sbi->s_mount_opt & EXT3_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); diff -NurpP --minimal linux-2.6.17.13/fs/ext3/symlink.c linux-2.6.17.13-g-v/fs/ext3/symlink.c --- linux-2.6.17.13/fs/ext3/symlink.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext3/symlink.c 2006-09-20 15:00:12.000000000 +0200 @@ -40,6 +40,7 @@ struct inode_operations ext3_symlink_ino .listxattr = ext3_listxattr, .removexattr = generic_removexattr, #endif + .sync_flags = ext3_sync_flags, }; struct inode_operations ext3_fast_symlink_inode_operations = { @@ -51,4 +52,5 @@ struct inode_operations ext3_fast_symlin .listxattr = ext3_listxattr, .removexattr = generic_removexattr, #endif + .sync_flags = ext3_sync_flags, }; diff -NurpP --minimal linux-2.6.17.13/fs/ext3/xattr.c linux-2.6.17.13-g-v/fs/ext3/xattr.c --- linux-2.6.17.13/fs/ext3/xattr.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ext3/xattr.c 2006-09-20 15:00:12.000000000 +0200 @@ -58,6 +58,7 @@ #include #include #include +#include #include "xattr.h" #include "acl.h" @@ -495,6 +496,7 @@ ext3_xattr_release_block(handle_t *handl ext3_journal_dirty_metadata(handle, bh); if (IS_SYNC(inode)) handle->h_sync = 1; + DLIMIT_FREE_BLOCK(inode, 1); DQUOT_FREE_BLOCK(inode, 1); unlock_buffer(bh); ea_bdebug(bh, "refcount now=%d; releasing", @@ -763,11 +765,14 @@ inserted: if (new_bh == bs->bh) ea_bdebug(new_bh, "keeping"); else { + error = -ENOSPC; + if (DLIMIT_ALLOC_BLOCK(inode, 1)) + goto cleanup; /* The old block is released after updating the inode. */ error = -EDQUOT; if (DQUOT_ALLOC_BLOCK(inode, 1)) - goto cleanup; + goto cleanup_dlimit; error = ext3_journal_get_write_access(handle, new_bh); if (error) @@ -843,6 +848,8 @@ cleanup: cleanup_dquot: DQUOT_FREE_BLOCK(inode, 1); +cleanup_dlimit: + DLIMIT_FREE_BLOCK(inode, 1); goto cleanup; bad_block: diff -NurpP --minimal linux-2.6.17.13/fs/fcntl.c linux-2.6.17.13-g-v/fs/fcntl.c --- linux-2.6.17.13/fs/fcntl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/fcntl.c 2006-09-20 15:00:12.000000000 +0200 @@ -18,6 +18,8 @@ #include #include #include +#include +#include #include #include @@ -63,6 +65,7 @@ static int locate_fd(struct files_struct struct fdtable *fdt; error = -EINVAL; + gr_learn_resource(current, RLIMIT_NOFILE, orig_start, 0); if (orig_start >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) goto out; @@ -83,8 +86,11 @@ repeat: } error = -EMFILE; + gr_learn_resource(current, RLIMIT_NOFILE, newfd, 0); if (newfd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) goto out; + if (!vx_files_avail(1)) + goto out; error = expand_files(files, newfd); if (error < 0) @@ -125,6 +131,7 @@ static int dupfd(struct file *file, unsi FD_SET(fd, fdt->open_fds); FD_CLR(fd, fdt->close_on_exec); spin_unlock(&files->file_lock); + vx_openfd_inc(fd); fd_install(fd, file); } else { spin_unlock(&files->file_lock); @@ -141,6 +148,8 @@ asmlinkage long sys_dup2(unsigned int ol struct files_struct * files = current->files; struct fdtable *fdt; + gr_learn_resource(current, RLIMIT_NOFILE, newfd, 0); + spin_lock(&files->file_lock); if (!(file = fcheck(oldfd))) goto out_unlock; @@ -177,6 +186,9 @@ asmlinkage long sys_dup2(unsigned int ol if (tofree) filp_close(tofree, files); + else + vx_openfd_inc(newfd); /* fd was unused */ + err = newfd; out: return err; @@ -427,7 +439,8 @@ static inline int sigio_perm(struct task return (((fown->euid == 0) || (fown->euid == p->suid) || (fown->euid == p->uid) || (fown->uid == p->suid) || (fown->uid == p->uid)) && - !security_file_send_sigiotask(p, fown, sig)); + !security_file_send_sigiotask(p, fown, sig) && + !gr_check_protected_task(p) && !gr_pid_is_chrooted(p)); } static void send_sigio_to_task(struct task_struct *p, @@ -479,7 +492,7 @@ void send_sigio(struct fown_struct *fown read_lock(&tasklist_lock); if (pid > 0) { - p = find_task_by_pid(pid); + p = find_task_by_real_pid(pid); if (p) { send_sigio_to_task(p, fown, fd, band); } @@ -514,7 +527,7 @@ int send_sigurg(struct fown_struct *fown read_lock(&tasklist_lock); if (pid > 0) { - p = find_task_by_pid(pid); + p = find_task_by_real_pid(pid); if (p) { send_sigurg_to_task(p, fown); } diff -NurpP --minimal linux-2.6.17.13/fs/file_table.c linux-2.6.17.13-g-v/fs/file_table.c --- linux-2.6.17.13/fs/file_table.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/file_table.c 2006-09-20 15:00:12.000000000 +0200 @@ -22,6 +22,8 @@ #include #include #include +#include +#include #include @@ -121,6 +123,8 @@ struct file *get_empty_filp(void) f->f_gid = tsk->fsgid; eventpoll_init_file(f); /* f->f_version: 0 */ + f->f_xid = vx_current_xid(); + vx_files_inc(f); return f; over: @@ -175,6 +179,8 @@ void fastcall __fput(struct file *file) fops_put(file->f_op); if (file->f_mode & FMODE_WRITE) put_write_access(inode); + vx_files_dec(file); + file->f_xid = 0; file_kill(file); file->f_dentry = NULL; file->f_vfsmnt = NULL; @@ -240,6 +246,8 @@ void put_filp(struct file *file) { if (atomic_dec_and_test(&file->f_count)) { security_file_free(file); + vx_files_dec(file); + file->f_xid = 0; file_kill(file); file_free(file); } diff -NurpP --minimal linux-2.6.17.13/fs/hfsplus/ioctl.c linux-2.6.17.13-g-v/fs/hfsplus/ioctl.c --- linux-2.6.17.13/fs/hfsplus/ioctl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/hfsplus/ioctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -16,6 +16,7 @@ #include #include #include +#include #include #include "hfsplus_fs.h" @@ -35,7 +36,8 @@ int hfsplus_ioctl(struct inode *inode, s flags |= EXT2_FLAG_NODUMP; /* EXT2_NODUMP_FL */ return put_user(flags, (int __user *)arg); case HFSPLUS_IOC_EXT2_SETFLAGS: { - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) diff -NurpP --minimal linux-2.6.17.13/fs/inode.c linux-2.6.17.13-g-v/fs/inode.c --- linux-2.6.17.13/fs/inode.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/inode.c 2006-09-20 15:00:12.000000000 +0200 @@ -116,6 +116,9 @@ static struct inode *alloc_inode(struct struct address_space * const mapping = &inode->i_data; inode->i_sb = sb; + + /* essential because of inode slab reuse */ + inode->i_xid = 0; inode->i_blkbits = sb->s_blocksize_bits; inode->i_flags = 0; atomic_set(&inode->i_count, 1); @@ -235,6 +238,8 @@ void __iget(struct inode * inode) inodes_stat.nr_unused--; } +EXPORT_SYMBOL_GPL(__iget); + /** * clear_inode - clear an inode * @inode: inode to clear diff -NurpP --minimal linux-2.6.17.13/fs/ioctl.c linux-2.6.17.13-g-v/fs/ioctl.c --- linux-2.6.17.13/fs/ioctl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ioctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -13,10 +13,19 @@ #include #include #include +#include +#include +#include #include #include + +#ifdef CONFIG_VSERVER_LEGACY +extern int vx_proc_ioctl(struct inode *, struct file *, + unsigned int, unsigned long); +#endif + static long do_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) { @@ -147,6 +156,48 @@ int vfs_ioctl(struct file *filp, unsigne else error = -ENOTTY; break; +#ifdef CONFIG_VSERVER_LEGACY +#ifndef CONFIG_INOXID_NONE + case FIOC_GETXID: { + struct inode *inode = filp->f_dentry->d_inode; + + /* fixme: if stealth, return -ENOTTY */ + error = -EPERM; + if (capable(CAP_CONTEXT)) + error = put_user(inode->i_xid, (int __user *) arg); + break; + } + case FIOC_SETXID: { + struct inode *inode = filp->f_dentry->d_inode; + int xid; + + /* fixme: if stealth, return -ENOTTY */ + error = -EPERM; + if (!capable(CAP_CONTEXT)) + break; + error = -EROFS; + if (IS_RDONLY(inode)) + break; + error = -ENOSYS; + if (!(inode->i_sb->s_flags & MS_TAGXID)) + break; + error = -EFAULT; + if (get_user(xid, (int __user *) arg)) + break; + error = 0; + inode->i_xid = (xid & 0xFFFF); + inode->i_ctime = CURRENT_TIME; + mark_inode_dirty(inode); + break; + } +#endif + case FIOC_GETXFLG: + case FIOC_SETXFLG: + error = -ENOTTY; + if (filp->f_dentry->d_inode->i_sb->s_magic == PROC_SUPER_MAGIC) + error = vx_proc_ioctl(filp->f_dentry->d_inode, filp, cmd, arg); + break; +#endif default: if (S_ISREG(filp->f_dentry->d_inode->i_mode)) error = file_ioctl(filp, cmd, arg); diff -NurpP --minimal linux-2.6.17.13/fs/ioprio.c linux-2.6.17.13-g-v/fs/ioprio.c --- linux-2.6.17.13/fs/ioprio.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/ioprio.c 2006-09-20 15:00:12.000000000 +0200 @@ -24,6 +24,7 @@ #include #include #include +#include static int set_task_ioprio(struct task_struct *task, int ioprio) { @@ -95,7 +96,7 @@ asmlinkage long sys_ioprio_set(int which if (!who) user = current->user; else - user = find_user(who); + user = find_user(vx_current_xid(), who); if (!user) break; @@ -149,7 +150,7 @@ asmlinkage long sys_ioprio_get(int which if (!who) user = current->user; else - user = find_user(who); + user = find_user(vx_current_xid(), who); if (!user) break; diff -NurpP --minimal linux-2.6.17.13/fs/jfs/acl.c linux-2.6.17.13-g-v/fs/jfs/acl.c --- linux-2.6.17.13/fs/jfs/acl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/acl.c 2006-09-20 15:00:12.000000000 +0200 @@ -232,7 +232,8 @@ int jfs_setattr(struct dentry *dentry, s return rc; if ((iattr->ia_valid & ATTR_UID && iattr->ia_uid != inode->i_uid) || - (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid)) { + (iattr->ia_valid & ATTR_GID && iattr->ia_gid != inode->i_gid) || + (iattr->ia_valid & ATTR_XID && iattr->ia_xid != inode->i_xid)) { if (DQUOT_TRANSFER(inode, iattr)) return -EDQUOT; } diff -NurpP --minimal linux-2.6.17.13/fs/jfs/file.c linux-2.6.17.13-g-v/fs/jfs/file.c --- linux-2.6.17.13/fs/jfs/file.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/file.c 2006-09-20 15:00:12.000000000 +0200 @@ -98,6 +98,7 @@ struct inode_operations jfs_file_inode_o .setattr = jfs_setattr, .permission = jfs_permission, #endif + .sync_flags = jfs_sync_flags, }; const struct file_operations jfs_file_operations = { diff -NurpP --minimal linux-2.6.17.13/fs/jfs/inode.c linux-2.6.17.13-g-v/fs/jfs/inode.c --- linux-2.6.17.13/fs/jfs/inode.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/inode.c 2006-09-20 15:00:12.000000000 +0200 @@ -22,6 +22,7 @@ #include #include #include +#include #include "jfs_incore.h" #include "jfs_inode.h" #include "jfs_filsys.h" @@ -144,6 +145,7 @@ void jfs_delete_inode(struct inode *inod DQUOT_INIT(inode); DQUOT_FREE_INODE(inode); DQUOT_DROP(inode); + DLIMIT_FREE_INODE(inode); } clear_inode(inode); diff -NurpP --minimal linux-2.6.17.13/fs/jfs/ioctl.c linux-2.6.17.13-g-v/fs/jfs/ioctl.c --- linux-2.6.17.13/fs/jfs/ioctl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/ioctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -10,6 +10,7 @@ #include #include #include +#include #include #include @@ -65,7 +66,8 @@ int jfs_ioctl(struct inode * inode, stru case JFS_IOC_SETFLAGS: { unsigned int oldflags; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) @@ -85,8 +87,8 @@ int jfs_ioctl(struct inode * inode, stru * the relevant capability. */ if ((oldflags & JFS_IMMUTABLE_FL) || - ((flags ^ oldflags) & - (JFS_APPEND_FL | JFS_IMMUTABLE_FL))) { + ((flags ^ oldflags) & (JFS_APPEND_FL | + JFS_IMMUTABLE_FL | JFS_IUNLINK_FL))) { if (!capable(CAP_LINUX_IMMUTABLE)) return -EPERM; } diff -NurpP --minimal linux-2.6.17.13/fs/jfs/jfs_dinode.h linux-2.6.17.13-g-v/fs/jfs/jfs_dinode.h --- linux-2.6.17.13/fs/jfs/jfs_dinode.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/jfs_dinode.h 2006-09-20 15:00:12.000000000 +0200 @@ -162,9 +162,12 @@ struct dinode { #define JFS_APPEND_FL 0x01000000 /* writes to file may only append */ #define JFS_IMMUTABLE_FL 0x02000000 /* Immutable file */ -#define JFS_FL_USER_VISIBLE 0x03F80000 +#define JFS_BARRIER_FL 0x04000000 /* Barrier for chroot() */ +#define JFS_IUNLINK_FL 0x08000000 /* Immutable unlink */ + +#define JFS_FL_USER_VISIBLE 0x0FF80000 #define JFS_FL_USER_MODIFIABLE 0x03F80000 -#define JFS_FL_INHERIT 0x03C80000 +#define JFS_FL_INHERIT 0x0BC80000 /* These are identical to EXT[23]_IOC_GETFLAGS/SETFLAGS */ #define JFS_IOC_GETFLAGS _IOR('f', 1, long) diff -NurpP --minimal linux-2.6.17.13/fs/jfs/jfs_dtree.c linux-2.6.17.13-g-v/fs/jfs/jfs_dtree.c --- linux-2.6.17.13/fs/jfs/jfs_dtree.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/jfs_dtree.c 2006-09-20 15:00:12.000000000 +0200 @@ -102,6 +102,7 @@ #include #include +#include #include "jfs_incore.h" #include "jfs_superblock.h" #include "jfs_filsys.h" @@ -383,10 +384,10 @@ static u32 add_index(tid_t tid, struct i */ if (DQUOT_ALLOC_BLOCK(ip, sbi->nbperpage)) goto clean_up; - if (dbAlloc(ip, 0, sbi->nbperpage, &xaddr)) { - DQUOT_FREE_BLOCK(ip, sbi->nbperpage); - goto clean_up; - } + if (DLIMIT_ALLOC_BLOCK(ip, sbi->nbperpage)) + goto clean_up_quota; + if (dbAlloc(ip, 0, sbi->nbperpage, &xaddr)) + goto clean_up_dlim; /* * Save the table, we're going to overwrite it with the @@ -479,6 +480,10 @@ static u32 add_index(tid_t tid, struct i return index; + clean_up_dlim: + DLIMIT_FREE_BLOCK(ip, sbi->nbperpage); + clean_up_quota: + DQUOT_FREE_BLOCK(ip, sbi->nbperpage); clean_up: jfs_ip->next_index--; @@ -952,6 +957,7 @@ static int dtSplitUp(tid_t tid, struct tlock *tlck; struct lv *lv; int quota_allocation = 0; + int dlimit_allocation = 0; /* get split page */ smp = split->mp; @@ -1036,6 +1042,12 @@ static int dtSplitUp(tid_t tid, } quota_allocation += n; + if (DLIMIT_ALLOC_BLOCK(ip, n)) { + rc = -ENOSPC; + goto extendOut; + } + dlimit_allocation += n; + if ((rc = dbReAlloc(sbi->ipbmap, xaddr, (s64) xlen, (s64) n, &nxaddr))) goto extendOut; @@ -1309,6 +1321,9 @@ static int dtSplitUp(tid_t tid, freeKeyName: kfree(key.name); + /* Rollback dlimit allocation */ + if (rc && dlimit_allocation) + DLIMIT_FREE_BLOCK(ip, dlimit_allocation); /* Rollback quota allocation */ if (rc && quota_allocation) DQUOT_FREE_BLOCK(ip, quota_allocation); @@ -1376,6 +1391,12 @@ static int dtSplitPage(tid_t tid, struct release_metapage(rmp); return -EDQUOT; } + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) { + DQUOT_FREE_BLOCK(ip, lengthPXD(pxd)); + release_metapage(rmp); + return -ENOSPC; + } jfs_info("dtSplitPage: ip:0x%p smp:0x%p rmp:0x%p", ip, smp, rmp); @@ -1926,6 +1947,12 @@ static int dtSplitRoot(tid_t tid, release_metapage(rmp); return -EDQUOT; } + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) { + DQUOT_FREE_BLOCK(ip, lengthPXD(pxd)); + release_metapage(rmp); + return -ENOSPC; + } BT_MARK_DIRTY(rmp, ip); /* @@ -2292,6 +2319,8 @@ static int dtDeleteUp(tid_t tid, struct xlen = lengthPXD(&fp->header.self); + /* Free dlimit allocation. */ + DLIMIT_FREE_BLOCK(ip, xlen); /* Free quota allocation. */ DQUOT_FREE_BLOCK(ip, xlen); @@ -2368,6 +2397,8 @@ static int dtDeleteUp(tid_t tid, struct xlen = lengthPXD(&p->header.self); + /* Free dlimit allocation */ + DLIMIT_FREE_BLOCK(ip, xlen); /* Free quota allocation */ DQUOT_FREE_BLOCK(ip, xlen); diff -NurpP --minimal linux-2.6.17.13/fs/jfs/jfs_extent.c linux-2.6.17.13-g-v/fs/jfs/jfs_extent.c --- linux-2.6.17.13/fs/jfs/jfs_extent.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/jfs_extent.c 2006-09-20 15:00:12.000000000 +0200 @@ -18,6 +18,7 @@ #include #include +#include #include "jfs_incore.h" #include "jfs_inode.h" #include "jfs_superblock.h" @@ -146,6 +147,13 @@ extAlloc(struct inode *ip, s64 xlen, s64 mutex_unlock(&JFS_IP(ip)->commit_mutex); return -EDQUOT; } + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, nxlen)) { + DQUOT_FREE_BLOCK(ip, nxlen); + dbFree(ip, nxaddr, (s64) nxlen); + mutex_unlock(&JFS_IP(ip)->commit_mutex); + return -ENOSPC; + } /* determine the value of the extent flag */ xflag = (abnr == TRUE) ? XAD_NOTRECORDED : 0; @@ -164,6 +172,7 @@ extAlloc(struct inode *ip, s64 xlen, s64 */ if (rc) { dbFree(ip, nxaddr, nxlen); + DLIMIT_FREE_BLOCK(ip, nxlen); DQUOT_FREE_BLOCK(ip, nxlen); mutex_unlock(&JFS_IP(ip)->commit_mutex); return (rc); @@ -261,6 +270,13 @@ int extRealloc(struct inode *ip, s64 nxl mutex_unlock(&JFS_IP(ip)->commit_mutex); return -EDQUOT; } + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, nxlen)) { + DQUOT_FREE_BLOCK(ip, nxlen); + dbFree(ip, nxaddr, (s64) nxlen); + up(&JFS_IP(ip)->commit_sem); + return -ENOSPC; + } delta = nxlen - xlen; @@ -297,6 +313,7 @@ int extRealloc(struct inode *ip, s64 nxl /* extend the extent */ if ((rc = xtExtend(0, ip, xoff + xlen, (int) nextend, 0))) { dbFree(ip, xaddr + xlen, delta); + DLIMIT_FREE_BLOCK(ip, nxlen); DQUOT_FREE_BLOCK(ip, nxlen); goto exit; } @@ -308,6 +325,7 @@ int extRealloc(struct inode *ip, s64 nxl */ if ((rc = xtTailgate(0, ip, xoff, (int) ntail, nxaddr, 0))) { dbFree(ip, nxaddr, nxlen); + DLIMIT_FREE_BLOCK(ip, nxlen); DQUOT_FREE_BLOCK(ip, nxlen); goto exit; } diff -NurpP --minimal linux-2.6.17.13/fs/jfs/jfs_filsys.h linux-2.6.17.13-g-v/fs/jfs/jfs_filsys.h --- linux-2.6.17.13/fs/jfs/jfs_filsys.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/jfs_filsys.h 2006-09-20 15:00:12.000000000 +0200 @@ -84,6 +84,7 @@ #define JFS_DIR_INDEX 0x00200000 /* Persistant index for */ /* directory entries */ +#define JFS_TAGXID 0x00800000 /* xid tagging */ /* * buffer cache configuration diff -NurpP --minimal linux-2.6.17.13/fs/jfs/jfs_imap.c linux-2.6.17.13-g-v/fs/jfs/jfs_imap.c --- linux-2.6.17.13/fs/jfs/jfs_imap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/jfs_imap.c 2006-09-20 15:00:12.000000000 +0200 @@ -45,6 +45,7 @@ #include #include #include +#include #include "jfs_incore.h" #include "jfs_inode.h" @@ -3075,6 +3076,8 @@ static int copy_from_dinode(struct dinod { struct jfs_inode_info *jfs_ip = JFS_IP(ip); struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb); + uid_t uid; + gid_t gid; jfs_ip->fileset = le32_to_cpu(dip->di_fileset); jfs_ip->mode2 = le32_to_cpu(dip->di_mode); @@ -3094,14 +3097,18 @@ static int copy_from_dinode(struct dinod } ip->i_nlink = le32_to_cpu(dip->di_nlink); - jfs_ip->saved_uid = le32_to_cpu(dip->di_uid); + uid = le32_to_cpu(dip->di_uid); + gid = le32_to_cpu(dip->di_gid); + ip->i_xid = INOXID_XID(XID_TAG(ip), uid, gid, 0); + + jfs_ip->saved_uid = INOXID_UID(XID_TAG(ip), uid, gid); if (sbi->uid == -1) ip->i_uid = jfs_ip->saved_uid; else { ip->i_uid = sbi->uid; } - jfs_ip->saved_gid = le32_to_cpu(dip->di_gid); + jfs_ip->saved_gid = INOXID_GID(XID_TAG(ip), uid, gid); if (sbi->gid == -1) ip->i_gid = jfs_ip->saved_gid; else { @@ -3167,14 +3174,12 @@ static void copy_to_dinode(struct dinode dip->di_size = cpu_to_le64(ip->i_size); dip->di_nblocks = cpu_to_le64(PBLK2LBLK(ip->i_sb, ip->i_blocks)); dip->di_nlink = cpu_to_le32(ip->i_nlink); - if (sbi->uid == -1) - dip->di_uid = cpu_to_le32(ip->i_uid); - else - dip->di_uid = cpu_to_le32(jfs_ip->saved_uid); - if (sbi->gid == -1) - dip->di_gid = cpu_to_le32(ip->i_gid); - else - dip->di_gid = cpu_to_le32(jfs_ip->saved_gid); + + dip->di_uid = cpu_to_le32(XIDINO_UID(XID_TAG(ip), + (sbi->uid == -1) ? ip->i_uid : jfs_ip->saved_uid, ip->i_xid)); + dip->di_gid = cpu_to_le32(XIDINO_GID(XID_TAG(ip), + (sbi->gid == -1) ? ip->i_gid : jfs_ip->saved_gid, ip->i_xid)); + /* * mode2 is only needed for storing the higher order bits. * Trust i_mode for the lower order ones diff -NurpP --minimal linux-2.6.17.13/fs/jfs/jfs_inode.c linux-2.6.17.13-g-v/fs/jfs/jfs_inode.c --- linux-2.6.17.13/fs/jfs/jfs_inode.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/jfs_inode.c 2006-09-20 15:00:12.000000000 +0200 @@ -18,6 +18,8 @@ #include #include +#include +#include #include "jfs_incore.h" #include "jfs_inode.h" #include "jfs_filsys.h" @@ -30,19 +32,59 @@ void jfs_set_inode_flags(struct inode *i { unsigned int flags = JFS_IP(inode)->mode2; - inode->i_flags &= ~(S_IMMUTABLE | S_APPEND | - S_NOATIME | S_DIRSYNC | S_SYNC); + inode->i_flags &= ~(S_IMMUTABLE | S_IUNLINK | S_BARRIER | + S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC); if (flags & JFS_IMMUTABLE_FL) inode->i_flags |= S_IMMUTABLE; + if (flags & JFS_IUNLINK_FL) + inode->i_flags |= S_IUNLINK; + if (flags & JFS_BARRIER_FL) + inode->i_flags |= S_BARRIER; + + if (flags & JFS_SYNC_FL) + inode->i_flags |= S_SYNC; if (flags & JFS_APPEND_FL) inode->i_flags |= S_APPEND; if (flags & JFS_NOATIME_FL) inode->i_flags |= S_NOATIME; if (flags & JFS_DIRSYNC_FL) inode->i_flags |= S_DIRSYNC; - if (flags & JFS_SYNC_FL) - inode->i_flags |= S_SYNC; +} + +int jfs_sync_flags(struct inode *inode) +{ + unsigned int oldflags, newflags; + + oldflags = JFS_IP(inode)->mode2; + newflags = oldflags & ~(JFS_APPEND_FL | + JFS_IMMUTABLE_FL | JFS_IUNLINK_FL | + JFS_BARRIER_FL | JFS_NOATIME_FL | + JFS_SYNC_FL | JFS_DIRSYNC_FL); + + if (IS_APPEND(inode)) + newflags |= JFS_APPEND_FL; + if (IS_IMMUTABLE(inode)) + newflags |= JFS_IMMUTABLE_FL; + if (IS_IUNLINK(inode)) + newflags |= JFS_IUNLINK_FL; + if (IS_BARRIER(inode)) + newflags |= JFS_BARRIER_FL; + + /* we do not want to copy superblock flags */ + if (inode->i_flags & S_NOATIME) + newflags |= JFS_NOATIME_FL; + if (inode->i_flags & S_SYNC) + newflags |= JFS_SYNC_FL; + if (inode->i_flags & S_DIRSYNC) + newflags |= JFS_DIRSYNC_FL; + + if (oldflags ^ newflags) { + JFS_IP(inode)->mode2 = newflags; + inode->i_ctime = CURRENT_TIME; + mark_inode_dirty(inode); + } + return 0; } /* @@ -89,10 +131,17 @@ struct inode *ialloc(struct inode *paren jfs_inode->saved_uid = inode->i_uid; jfs_inode->saved_gid = inode->i_gid; + inode->i_xid = vx_current_fsxid(sb); + if (DLIMIT_ALLOC_INODE(inode)) { + iput(inode); + return NULL; + } + /* * Allocate inode to quota. */ if (DQUOT_ALLOC_INODE(inode)) { + DLIMIT_FREE_INODE(inode); DQUOT_DROP(inode); inode->i_flags |= S_NOQUOTA; inode->i_nlink = 0; diff -NurpP --minimal linux-2.6.17.13/fs/jfs/jfs_inode.h linux-2.6.17.13-g-v/fs/jfs/jfs_inode.h --- linux-2.6.17.13/fs/jfs/jfs_inode.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/jfs_inode.h 2006-09-20 15:00:12.000000000 +0200 @@ -31,6 +31,7 @@ extern void jfs_truncate(struct inode *) extern void jfs_truncate_nolock(struct inode *, loff_t); extern void jfs_free_zero_link(struct inode *); extern struct dentry *jfs_get_parent(struct dentry *dentry); +extern int jfs_sync_flags(struct inode *); extern void jfs_set_inode_flags(struct inode *); extern struct address_space_operations jfs_aops; diff -NurpP --minimal linux-2.6.17.13/fs/jfs/jfs_xtree.c linux-2.6.17.13-g-v/fs/jfs/jfs_xtree.c --- linux-2.6.17.13/fs/jfs/jfs_xtree.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/jfs_xtree.c 2006-09-20 15:00:12.000000000 +0200 @@ -21,6 +21,7 @@ #include #include +#include #include "jfs_incore.h" #include "jfs_filsys.h" #include "jfs_metapage.h" @@ -841,7 +842,12 @@ int xtInsert(tid_t tid, /* transaction hint = 0; if ((rc = DQUOT_ALLOC_BLOCK(ip, xlen))) goto out; + if ((rc = DLIMIT_ALLOC_BLOCK(ip, xlen))) { + DQUOT_FREE_BLOCK(ip, xlen); + goto out; + } if ((rc = dbAlloc(ip, hint, (s64) xlen, &xaddr))) { + DLIMIT_FREE_BLOCK(ip, xlen); DQUOT_FREE_BLOCK(ip, xlen); goto out; } @@ -871,6 +877,7 @@ int xtInsert(tid_t tid, /* transaction /* undo data extent allocation */ if (*xaddrp == 0) { dbFree(ip, xaddr, (s64) xlen); + DLIMIT_FREE_BLOCK(ip, xlen); DQUOT_FREE_BLOCK(ip, xlen); } return rc; @@ -1231,6 +1238,7 @@ xtSplitPage(tid_t tid, struct inode *ip, struct tlock *tlck; struct xtlock *sxtlck = NULL, *rxtlck = NULL; int quota_allocation = 0; + int dlimit_allocation = 0; smp = split->mp; sp = XT_PAGE(ip, smp); @@ -1250,6 +1258,13 @@ xtSplitPage(tid_t tid, struct inode *ip, quota_allocation += lengthPXD(pxd); + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) { + rc = -ENOSPC; + goto clean_up; + } + dlimit_allocation += lengthPXD(pxd); + /* * allocate the new right page for the split */ @@ -1451,6 +1466,9 @@ xtSplitPage(tid_t tid, struct inode *ip, clean_up: + /* Rollback dlimit allocation. */ + if (dlimit_allocation) + DLIMIT_FREE_BLOCK(ip, dlimit_allocation); /* Rollback quota allocation. */ if (quota_allocation) DQUOT_FREE_BLOCK(ip, quota_allocation); @@ -1515,6 +1533,12 @@ xtSplitRoot(tid_t tid, release_metapage(rmp); return -EDQUOT; } + /* Allocate blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, lengthPXD(pxd))) { + DQUOT_FREE_BLOCK(ip, lengthPXD(pxd)); + release_metapage(rmp); + return -ENOSPC; + } jfs_info("xtSplitRoot: ip:0x%p rmp:0x%p", ip, rmp); @@ -3941,6 +3965,8 @@ s64 xtTruncate(tid_t tid, struct inode * else ip->i_size = newsize; + /* update dlimit allocation to reflect freed blocks */ + DLIMIT_FREE_BLOCK(ip, nfreed); /* update quota allocation to reflect freed blocks */ DQUOT_FREE_BLOCK(ip, nfreed); diff -NurpP --minimal linux-2.6.17.13/fs/jfs/namei.c linux-2.6.17.13-g-v/fs/jfs/namei.c --- linux-2.6.17.13/fs/jfs/namei.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/namei.c 2006-09-20 15:00:12.000000000 +0200 @@ -20,6 +20,7 @@ #include #include #include +#include #include "jfs_incore.h" #include "jfs_superblock.h" #include "jfs_inode.h" @@ -1465,6 +1466,7 @@ static struct dentry *jfs_lookup(struct return ERR_PTR(-EACCES); } + vx_propagate_xid(nd, ip); dentry = d_splice_alias(ip, dentry); if (dentry && (JFS_SBI(dip->i_sb)->mntflag & JFS_OS2)) @@ -1517,6 +1519,7 @@ struct inode_operations jfs_dir_inode_op .setattr = jfs_setattr, .permission = jfs_permission, #endif + .sync_flags = jfs_sync_flags, }; const struct file_operations jfs_dir_operations = { diff -NurpP --minimal linux-2.6.17.13/fs/jfs/super.c linux-2.6.17.13-g-v/fs/jfs/super.c --- linux-2.6.17.13/fs/jfs/super.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/super.c 2006-09-20 15:00:12.000000000 +0200 @@ -194,7 +194,8 @@ static void jfs_put_super(struct super_b enum { Opt_integrity, Opt_nointegrity, Opt_iocharset, Opt_resize, Opt_resize_nosize, Opt_errors, Opt_ignore, Opt_err, Opt_quota, - Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask + Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask, + Opt_tagxid }; static match_table_t tokens = { @@ -204,6 +205,7 @@ static match_table_t tokens = { {Opt_resize, "resize=%u"}, {Opt_resize_nosize, "resize"}, {Opt_errors, "errors=%s"}, + {Opt_tagxid, "tagxid"}, {Opt_ignore, "noquota"}, {Opt_ignore, "quota"}, {Opt_usrquota, "usrquota"}, @@ -338,6 +340,11 @@ static int parse_options(char *options, } break; } +#ifndef CONFIG_TAGGING_NONE + case Opt_tagxid: + *flag |= JFS_TAGXID; + break; +#endif default: printk("jfs: Unrecognized mount option \"%s\" " " or missing value\n", p); @@ -368,6 +375,13 @@ static int jfs_remount(struct super_bloc if (!parse_options(data, sb, &newLVSize, &flag)) { return -EINVAL; } + + if ((flag & JFS_TAGXID) && !(sb->s_flags & MS_TAGXID)) { + printk(KERN_ERR "JFS: %s: tagxid not permitted on remount.\n", + sb->s_id); + return -EINVAL; + } + if (newLVSize) { if (sb->s_flags & MS_RDONLY) { printk(KERN_ERR @@ -439,6 +453,9 @@ static int jfs_fill_super(struct super_b #ifdef CONFIG_JFS_POSIX_ACL sb->s_flags |= MS_POSIXACL; #endif + /* map mount option tagxid */ + if (sbi->flag & JFS_TAGXID) + sb->s_flags |= MS_TAGXID; if (newLVSize) { printk(KERN_ERR "resize option for remount only\n"); diff -NurpP --minimal linux-2.6.17.13/fs/jfs/xattr.c linux-2.6.17.13-g-v/fs/jfs/xattr.c --- linux-2.6.17.13/fs/jfs/xattr.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/jfs/xattr.c 2006-09-20 15:00:12.000000000 +0200 @@ -23,6 +23,7 @@ #include #include #include +#include #include "jfs_incore.h" #include "jfs_superblock.h" #include "jfs_dmap.h" @@ -263,9 +264,16 @@ static int ea_write(struct inode *ip, st if (DQUOT_ALLOC_BLOCK(ip, nblocks)) { return -EDQUOT; } + /* Allocate new blocks to dlimit. */ + if (DLIMIT_ALLOC_BLOCK(ip, nblocks)) { + DQUOT_FREE_BLOCK(ip, nblocks); + return -ENOSPC; + } rc = dbAlloc(ip, INOHINT(ip), nblocks, &blkno); if (rc) { + /*Rollback dlimit allocation. */ + DLIMIT_FREE_BLOCK(ip, nblocks); /*Rollback quota allocation. */ DQUOT_FREE_BLOCK(ip, nblocks); return rc; @@ -332,6 +340,8 @@ static int ea_write(struct inode *ip, st failed: /* Rollback quota allocation. */ + DLIMIT_FREE_BLOCK(ip, nblocks); + /* Rollback quota allocation. */ DQUOT_FREE_BLOCK(ip, nblocks); dbFree(ip, blkno, nblocks); @@ -468,6 +478,7 @@ static int ea_get(struct inode *inode, s s64 blkno; int rc; int quota_allocation = 0; + int dlimit_allocation = 0; /* When fsck.jfs clears a bad ea, it doesn't clear the size */ if (ji->ea.flag == 0) @@ -543,6 +554,12 @@ static int ea_get(struct inode *inode, s quota_allocation = blocks_needed; + /* Allocate new blocks to dlimit. */ + rc = -ENOSPC; + if (DLIMIT_ALLOC_BLOCK(inode, blocks_needed)) + goto clean_up; + dlimit_allocation = blocks_needed; + rc = dbAlloc(inode, INOHINT(inode), (s64) blocks_needed, &blkno); if (rc) @@ -599,6 +616,9 @@ static int ea_get(struct inode *inode, s return ea_size; clean_up: + /* Rollback dlimit allocation */ + if (dlimit_allocation) + DLIMIT_FREE_BLOCK(inode, dlimit_allocation); /* Rollback quota allocation */ if (quota_allocation) DQUOT_FREE_BLOCK(inode, quota_allocation); @@ -675,8 +695,10 @@ static int ea_put(tid_t tid, struct inod } /* If old blocks exist, they must be removed from quota allocation. */ - if (old_blocks) + if (old_blocks) { + DLIMIT_FREE_BLOCK(inode, old_blocks); DQUOT_FREE_BLOCK(inode, old_blocks); + } inode->i_ctime = CURRENT_TIME; diff -NurpP --minimal linux-2.6.17.13/fs/libfs.c linux-2.6.17.13-g-v/fs/libfs.c --- linux-2.6.17.13/fs/libfs.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/libfs.c 2006-09-20 15:00:12.000000000 +0200 @@ -124,7 +124,8 @@ static inline unsigned char dt_type(stru * both impossible due to the lock on directory. */ -int dcache_readdir(struct file * filp, void * dirent, filldir_t filldir) +static inline int do_dcache_readdir_filter(struct file * filp, + void * dirent, filldir_t filldir, int (*filter)(struct dentry *dentry)) { struct dentry *dentry = filp->f_dentry; struct dentry *cursor = filp->private_data; @@ -158,6 +159,8 @@ int dcache_readdir(struct file * filp, v next = list_entry(p, struct dentry, d_u.d_child); if (d_unhashed(next) || !next->d_inode) continue; + if (filter && !filter(next)) + continue; spin_unlock(&dcache_lock); if (filldir(dirent, next->d_name.name, next->d_name.len, filp->f_pos, next->d_inode->i_ino, dt_type(next->d_inode)) < 0) @@ -174,6 +177,18 @@ int dcache_readdir(struct file * filp, v return 0; } +int dcache_readdir(struct file * filp, void * dirent, filldir_t filldir) +{ + return do_dcache_readdir_filter(filp, dirent, filldir, NULL); +} + +int dcache_readdir_filter(struct file * filp, void * dirent, filldir_t filldir, + int (*filter)(struct dentry *)) +{ + return do_dcache_readdir_filter(filp, dirent, filldir, filter); +} + + ssize_t generic_read_dir(struct file *filp, char __user *buf, size_t siz, loff_t *ppos) { return -EISDIR; @@ -623,6 +638,7 @@ EXPORT_SYMBOL(dcache_dir_close); EXPORT_SYMBOL(dcache_dir_lseek); EXPORT_SYMBOL(dcache_dir_open); EXPORT_SYMBOL(dcache_readdir); +EXPORT_SYMBOL(dcache_readdir_filter); EXPORT_SYMBOL(generic_read_dir); EXPORT_SYMBOL(get_sb_pseudo); EXPORT_SYMBOL(simple_commit_write); diff -NurpP --minimal linux-2.6.17.13/fs/lockd/clntproc.c linux-2.6.17.13-g-v/fs/lockd/clntproc.c --- linux-2.6.17.13/fs/lockd/clntproc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/lockd/clntproc.c 2006-09-20 15:00:12.000000000 +0200 @@ -18,6 +18,7 @@ #include #include #include +#include #define NLMDBG_FACILITY NLMDBG_CLIENT #define NLMCLNT_GRACE_WAIT (5*HZ) @@ -130,11 +131,11 @@ static void nlmclnt_setlockargs(struct n nlmclnt_next_cookie(&argp->cookie); argp->state = nsm_local_state; memcpy(&lock->fh, NFS_FH(fl->fl_file->f_dentry->d_inode), sizeof(struct nfs_fh)); - lock->caller = system_utsname.nodename; + lock->caller = vx_new_uts(nodename); lock->oh.data = req->a_owner; lock->oh.len = snprintf(req->a_owner, sizeof(req->a_owner), "%u@%s", (unsigned int)fl->fl_u.nfs_fl.owner->pid, - system_utsname.nodename); + vx_new_uts(nodename)); lock->svid = fl->fl_u.nfs_fl.owner->pid; lock->fl.fl_start = fl->fl_start; lock->fl.fl_end = fl->fl_end; diff -NurpP --minimal linux-2.6.17.13/fs/locks.c linux-2.6.17.13-g-v/fs/locks.c --- linux-2.6.17.13/fs/locks.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/locks.c 2006-09-20 15:00:12.000000000 +0200 @@ -125,6 +125,7 @@ #include #include #include +#include #include #include @@ -147,6 +148,8 @@ static kmem_cache_t *filelock_cache __re /* Allocate an empty lock structure. */ static struct file_lock *locks_alloc_lock(void) { + if (!vx_locks_avail(1)) + return NULL; return kmem_cache_alloc(filelock_cache, SLAB_KERNEL); } @@ -172,6 +175,7 @@ static void locks_free_lock(struct file_ BUG_ON(!list_empty(&fl->fl_block)); BUG_ON(!list_empty(&fl->fl_link)); + vx_locks_dec(fl); locks_release_private(fl); kmem_cache_free(filelock_cache, fl); } @@ -191,6 +195,7 @@ void locks_init_lock(struct file_lock *f fl->fl_start = fl->fl_end = 0; fl->fl_ops = NULL; fl->fl_lmops = NULL; + fl->fl_xid = -1; } EXPORT_SYMBOL(locks_init_lock); @@ -248,6 +253,7 @@ void locks_copy_lock(struct file_lock *n new->fl_file = fl->fl_file; new->fl_ops = fl->fl_ops; new->fl_lmops = fl->fl_lmops; + new->fl_xid = fl->fl_xid; locks_copy_private(new, fl); } @@ -286,6 +292,9 @@ static int flock_make_lock(struct file * fl->fl_flags = FL_FLOCK; fl->fl_type = type; fl->fl_end = OFFSET_MAX; + + fl->fl_xid = filp->f_xid; + vx_locks_inc(fl); *lock = fl; return 0; @@ -451,6 +460,7 @@ static int lease_init(struct file *filp, fl->fl_owner = current->files; fl->fl_pid = current->tgid; + fl->fl_xid = vx_current_xid(); fl->fl_file = filp; fl->fl_flags = FL_LEASE; @@ -470,6 +480,8 @@ static int lease_alloc(struct file *filp if (fl == NULL) goto out; + fl->fl_xid = vx_current_xid(); + vx_locks_inc(fl); error = lease_init(filp, type, fl); if (error) { locks_free_lock(fl); @@ -780,6 +792,7 @@ static int flock_lock_file(struct file * goto out; } locks_copy_lock(new_fl, request); + vx_locks_inc(new_fl); locks_insert_lock(&inode->i_flock, new_fl); new_fl = NULL; error = 0; @@ -791,7 +804,8 @@ out: return error; } -static int __posix_lock_file_conf(struct inode *inode, struct file_lock *request, struct file_lock *conflock) +static int __posix_lock_file_conf(struct inode *inode, struct file_lock *request, + struct file_lock *conflock, xid_t xid) { struct file_lock *fl; struct file_lock *new_fl, *new_fl2; @@ -805,7 +819,11 @@ static int __posix_lock_file_conf(struct * so we get them in advance to avoid races. */ new_fl = locks_alloc_lock(); + new_fl->fl_xid = xid; + vx_locks_inc(new_fl); new_fl2 = locks_alloc_lock(); + new_fl2->fl_xid = xid; + vx_locks_inc(new_fl2); lock_kernel(); if (request->fl_type != F_UNLCK) { @@ -986,7 +1004,8 @@ static int __posix_lock_file_conf(struct */ int posix_lock_file(struct file *filp, struct file_lock *fl) { - return __posix_lock_file_conf(filp->f_dentry->d_inode, fl, NULL); + return __posix_lock_file_conf(filp->f_dentry->d_inode, + fl, NULL, filp->f_xid); } EXPORT_SYMBOL(posix_lock_file); @@ -1001,7 +1020,8 @@ EXPORT_SYMBOL(posix_lock_file); int posix_lock_file_conf(struct file *filp, struct file_lock *fl, struct file_lock *conflock) { - return __posix_lock_file_conf(filp->f_dentry->d_inode, fl, conflock); + return __posix_lock_file_conf(filp->f_dentry->d_inode, + fl, conflock, filp->f_xid); } EXPORT_SYMBOL(posix_lock_file_conf); @@ -1091,7 +1111,7 @@ int locks_mandatory_area(int read_write, fl.fl_end = offset + count - 1; for (;;) { - error = __posix_lock_file_conf(inode, &fl, NULL); + error = __posix_lock_file_conf(inode, &fl, NULL, filp->f_xid); if (error != -EAGAIN) break; if (!(fl.fl_flags & FL_SLEEP)) @@ -1653,6 +1673,9 @@ int fcntl_setlk(unsigned int fd, struct if (file_lock == NULL) return -ENOLCK; + file_lock->fl_xid = filp->f_xid; + vx_locks_inc(file_lock); + /* * This might block, so we do it before checking the inode. */ @@ -1796,6 +1819,9 @@ int fcntl_setlk64(unsigned int fd, struc if (file_lock == NULL) return -ENOLCK; + file_lock->fl_xid = filp->f_xid; + vx_locks_inc(file_lock); + /* * This might block, so we do it before checking the inode. */ @@ -2106,6 +2132,10 @@ int get_locks_status(char *buffer, char list_for_each(tmp, &file_lock_list) { struct list_head *btmp; struct file_lock *fl = list_entry(tmp, struct file_lock, fl_link); + + if (!vx_check(fl->fl_xid, VX_IDENT|VX_WATCH)) + continue; + lock_get_status(q, fl, ++i, ""); move_lock_status(&q, &pos, offset); diff -NurpP --minimal linux-2.6.17.13/fs/namei.c linux-2.6.17.13-g-v/fs/namei.c --- linux-2.6.17.13/fs/namei.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/namei.c 2006-09-20 17:30:50.000000000 +0200 @@ -32,6 +32,10 @@ #include #include #include +#include +#include +#include +#include #include #include @@ -225,6 +229,31 @@ int generic_permission(struct inode *ino return -EACCES; } +static inline int vx_barrier(struct inode *inode) +{ + if (IS_BARRIER(inode) && !vx_check(0, VX_ADMIN)) { + vxwprintk(1, "xid=%d did hit the barrier.", + vx_current_xid()); + return 1; + } + return 0; +} + +static inline int xid_permission(struct inode *inode, int mask, struct nameidata *nd) +{ + if (vx_barrier(inode)) + return -EACCES; + if (inode->i_xid == 0) + return 0; + if (vx_check(inode->i_xid, VX_ADMIN|VX_WATCH|VX_IDENT)) + return 0; + + vxwprintk(1, "xid=%d denied access to %p[#%d,%lu] »%s«.", + vx_current_xid(), inode, inode->i_xid, inode->i_ino, + vxd_cond_path(nd)); + return -EACCES; +} + int permission(struct inode *inode, int mask, struct nameidata *nd) { int retval, submask; @@ -235,7 +264,7 @@ int permission(struct inode *inode, int /* * Nobody gets write access to a read-only fs. */ - if (IS_RDONLY(inode) && + if ((IS_RDONLY(inode) || (nd && MNT_IS_RDONLY(nd->mnt))) && (S_ISREG(mode) || S_ISDIR(mode) || S_ISLNK(mode))) return -EROFS; @@ -249,6 +278,8 @@ int permission(struct inode *inode, int /* Ordinary permission routines do not understand MAY_APPEND. */ submask = mask & ~MAY_APPEND; + if ((retval = xid_permission(inode, mask, nd))) + return retval; if (inode->i_op && inode->i_op->permission) retval = inode->i_op->permission(inode, submask, nd); else @@ -403,6 +434,8 @@ static int exec_permission_lite(struct i { umode_t mode = inode->i_mode; + if (vx_barrier(inode)) + return -EACCES; if (inode->i_op && inode->i_op->permission) return -EAGAIN; @@ -611,6 +644,13 @@ static inline int do_follow_link(struct err = security_inode_follow_link(path->dentry, nd); if (err) goto loop; + + if (gr_handle_follow_link(path->dentry->d_parent->d_inode, + path->dentry->d_inode, path->dentry, nd->mnt)) { + err = -EACCES; + goto loop; + } + current->link_count++; current->total_link_count++; nd->depth++; @@ -705,7 +745,8 @@ static __always_inline void follow_dotdo if (nd->dentry == current->fs->root && nd->mnt == current->fs->rootmnt) { read_unlock(&current->fs->lock); - break; + /* for sane '/' avoid follow_mount() */ + return; } read_unlock(&current->fs->lock); spin_lock(&dcache_lock); @@ -742,16 +783,34 @@ static int do_lookup(struct nameidata *n { struct vfsmount *mnt = nd->mnt; struct dentry *dentry = __d_lookup(nd->dentry, name); + struct inode *inode; if (!dentry) goto need_lookup; if (dentry->d_op && dentry->d_op->d_revalidate) goto need_revalidate; + inode = dentry->d_inode; + if (!inode) + goto done; + if (!vx_check(inode->i_xid, VX_WATCH|VX_ADMIN|VX_HOSTID|VX_IDENT)) + goto hidden; + if (inode->i_sb->s_magic == PROC_SUPER_MAGIC) { + struct proc_dir_entry *de = PDE(inode); + + if (de && !vx_hide_check(0, de->vx_flags)) + goto hidden; + } done: path->mnt = mnt; path->dentry = dentry; __follow_mount(path); return 0; +hidden: + vxwprintk(1, "xid=%d did lookup hidden %p[#%d,%lu] »%s«.", + vx_current_xid(), inode, inode->i_xid, inode->i_ino, + vxd_path(dentry, mnt)); + dput(dentry); + return -ENOENT; need_lookup: dentry = real_lookup(nd->dentry, name, nd); @@ -954,11 +1013,18 @@ return_reval: break; } return_base: + if (!gr_acl_handle_hidden_file(nd->dentry, nd->mnt)) { + path_release(nd); + return -ENOENT; + } return 0; out_dput: dput_path(&next, nd); break; } + if (!gr_acl_handle_hidden_file(nd->dentry, nd->mnt)) + err = -ENOENT; + path_release(nd); return_err: return err; @@ -1349,7 +1415,8 @@ static inline int check_sticky(struct in * 10. We don't allow removal of NFS sillyrenamed files; it's handled by * nfs_async_unlink(). */ -static int may_delete(struct inode *dir,struct dentry *victim,int isdir) +static int may_delete(struct inode *dir, struct dentry *victim, + int isdir, struct nameidata *nd) { int error; @@ -1359,13 +1426,13 @@ static int may_delete(struct inode *dir, BUG_ON(victim->d_parent->d_inode != dir); audit_inode_child(victim->d_name.name, victim->d_inode, dir->i_ino); - error = permission(dir,MAY_WRITE | MAY_EXEC, NULL); + error = permission(dir,MAY_WRITE | MAY_EXEC, nd); if (error) return error; if (IS_APPEND(dir)) return -EPERM; if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)|| - IS_IMMUTABLE(victim->d_inode)) + IS_IXORUNLINK(victim->d_inode)) return -EPERM; if (isdir) { if (!S_ISDIR(victim->d_inode->i_mode)) @@ -1512,7 +1579,8 @@ int may_open(struct nameidata *nd, int a return -EACCES; flag &= ~O_TRUNC; - } else if (IS_RDONLY(inode) && (flag & FMODE_WRITE)) + } else if ((IS_RDONLY(inode) || MNT_IS_RDONLY(nd->mnt)) + && (flag & FMODE_WRITE)) return -EROFS; /* * An append-only file must be opened in append mode for writing. @@ -1601,6 +1669,17 @@ int open_namei(int dfd, const char *path nd, flag); if (error) return error; + + if (gr_handle_rawio(nd->dentry->d_inode)) { + error = -EPERM; + goto exit; + } + + if (!gr_acl_handle_open(nd->dentry, nd->mnt, flag)) { + error = -EACCES; + goto exit; + } + goto ok; } @@ -1641,9 +1720,16 @@ do_last: /* Negative dentry, just create the file */ if (!path.dentry->d_inode) { + if (!gr_acl_handle_creat(path.dentry, nd->dentry, nd->mnt, flag, mode)) { + error = -EACCES; + mutex_unlock(&dir->d_inode->i_mutex); + goto exit_dput; + } if (!IS_POSIXACL(dir->d_inode)) mode &= ~current->fs->umask; error = vfs_create(dir->d_inode, path.dentry, mode, nd); + if (!error) + gr_handle_create(path.dentry, nd->mnt); mutex_unlock(&dir->d_inode->i_mutex); dput(nd->dentry); nd->dentry = path.dentry; @@ -1658,6 +1744,23 @@ do_last: /* * It already exists. */ + + if (gr_handle_rawio(path.dentry->d_inode)) { + mutex_unlock(&dir->d_inode->i_mutex); + error = -EPERM; + goto exit_dput; + } + if (!gr_acl_handle_open(path.dentry, nd->mnt, flag)) { + mutex_unlock(&dir->d_inode->i_mutex); + error = -EACCES; + goto exit_dput; + } + if (gr_handle_fifo(path.dentry, nd->mnt, dir, flag, acc_mode)) { + mutex_unlock(&dir->d_inode->i_mutex); + error = -EACCES; + goto exit_dput; + } + mutex_unlock(&dir->d_inode->i_mutex); error = -EEXIST; @@ -1711,6 +1814,13 @@ do_link: error = security_inode_follow_link(path.dentry, nd); if (error) goto exit_dput; + + if (gr_handle_follow_link(path.dentry->d_parent->d_inode, path.dentry->d_inode, + path.dentry, nd->mnt)) { + error = -EACCES; + goto exit_dput; + } + error = __do_follow_link(&path, nd); if (error) { /* Does someone understand code flow here? Or it is only @@ -1790,9 +1900,10 @@ fail: } EXPORT_SYMBOL_GPL(lookup_create); -int vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev) +int vfs_mknod(struct inode *dir, struct dentry *dentry, + int mode, dev_t dev, struct nameidata *nd) { - int error = may_create(dir, dentry, NULL); + int error = may_create(dir, dentry, nd); if (error) return error; @@ -1837,16 +1948,33 @@ asmlinkage long sys_mknodat(int dfd, con if (!IS_POSIXACL(nd.dentry->d_inode)) mode &= ~current->fs->umask; if (!IS_ERR(dentry)) { + if (gr_handle_chroot_mknod(dentry, nd.mnt, mode)) { + error = -EPERM; + dput(dentry); + mutex_unlock(&nd.dentry->d_inode->i_mutex); + path_release(&nd); + goto out; + } + + if (!gr_acl_handle_mknod(dentry, nd.dentry, nd.mnt, mode)) { + error = -EACCES; + dput(dentry); + mutex_unlock(&nd.dentry->d_inode->i_mutex); + path_release(&nd); + goto out; + } + switch (mode & S_IFMT) { case 0: case S_IFREG: error = vfs_create(nd.dentry->d_inode,dentry,mode,&nd); break; case S_IFCHR: case S_IFBLK: - error = vfs_mknod(nd.dentry->d_inode,dentry,mode, - new_decode_dev(dev)); + error = vfs_mknod(nd.dentry->d_inode, dentry, mode, + new_decode_dev(dev), &nd); break; case S_IFIFO: case S_IFSOCK: - error = vfs_mknod(nd.dentry->d_inode,dentry,mode,0); + error = vfs_mknod(nd.dentry->d_inode, dentry, mode, + 0, &nd); break; case S_IFDIR: error = -EPERM; @@ -1854,6 +1982,10 @@ asmlinkage long sys_mknodat(int dfd, con default: error = -EINVAL; } + + if (!error) + gr_handle_create(dentry, nd.mnt); + dput(dentry); } mutex_unlock(&nd.dentry->d_inode->i_mutex); @@ -1869,9 +2001,10 @@ asmlinkage long sys_mknod(const char __u return sys_mknodat(AT_FDCWD, filename, mode, dev); } -int vfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) +int vfs_mkdir(struct inode *dir, struct dentry *dentry, + int mode, struct nameidata *nd) { - int error = may_create(dir, dentry, NULL); + int error = may_create(dir, dentry, nd); if (error) return error; @@ -1908,9 +2041,19 @@ asmlinkage long sys_mkdirat(int dfd, con dentry = lookup_create(&nd, 1); error = PTR_ERR(dentry); if (!IS_ERR(dentry)) { + error = 0; if (!IS_POSIXACL(nd.dentry->d_inode)) mode &= ~current->fs->umask; - error = vfs_mkdir(nd.dentry->d_inode, dentry, mode); + + if (!gr_acl_handle_mkdir(dentry, nd.dentry, nd.mnt)) + error = -EACCES; + + if (!error) + error = vfs_mkdir(nd.dentry->d_inode, dentry, mode, &nd); + + if (!error) + gr_handle_create(dentry, nd.mnt); + dput(dentry); } mutex_unlock(&nd.dentry->d_inode->i_mutex); @@ -1955,9 +2098,10 @@ void dentry_unhash(struct dentry *dentry spin_unlock(&dcache_lock); } -int vfs_rmdir(struct inode *dir, struct dentry *dentry) +int vfs_rmdir(struct inode *dir, struct dentry *dentry, + struct nameidata *nd) { - int error = may_delete(dir, dentry, 1); + int error = may_delete(dir, dentry, 1, nd); if (error) return error; @@ -1994,6 +2138,8 @@ static long do_rmdir(int dfd, const char char * name; struct dentry *dentry; struct nameidata nd; + ino_t saved_ino = 0; + dev_t saved_dev = 0; name = getname(pathname); if(IS_ERR(name)) @@ -2018,7 +2164,21 @@ static long do_rmdir(int dfd, const char dentry = lookup_hash(&nd); error = PTR_ERR(dentry); if (!IS_ERR(dentry)) { - error = vfs_rmdir(nd.dentry->d_inode, dentry); + error = 0; + if (dentry->d_inode) { + if (dentry->d_inode->i_nlink <= 1) { + saved_ino = dentry->d_inode->i_ino; + saved_dev = dentry->d_inode->i_sb->s_dev; + } + + if (!gr_acl_handle_rmdir(dentry, nd.mnt)) + error = -EACCES; + } + + if (!error) + error = vfs_rmdir(nd.dentry->d_inode, dentry, &nd); + if (!error && (saved_dev || saved_ino)) + gr_handle_delete(saved_ino, saved_dev); dput(dentry); } mutex_unlock(&nd.dentry->d_inode->i_mutex); @@ -2034,9 +2194,10 @@ asmlinkage long sys_rmdir(const char __u return do_rmdir(AT_FDCWD, pathname); } -int vfs_unlink(struct inode *dir, struct dentry *dentry) +int vfs_unlink(struct inode *dir, struct dentry *dentry, + struct nameidata *nd) { - int error = may_delete(dir, dentry, 0); + int error = may_delete(dir, dentry, 0, nd); if (error) return error; @@ -2077,6 +2238,8 @@ static long do_unlinkat(int dfd, const c struct dentry *dentry; struct nameidata nd; struct inode *inode = NULL; + ino_t saved_ino = 0; + dev_t saved_dev = 0; name = getname(pathname); if(IS_ERR(name)) @@ -2092,13 +2255,26 @@ static long do_unlinkat(int dfd, const c dentry = lookup_hash(&nd); error = PTR_ERR(dentry); if (!IS_ERR(dentry)) { + error = 0; /* Why not before? Because we want correct error value */ if (nd.last.name[nd.last.len]) goto slashes; inode = dentry->d_inode; - if (inode) + if (inode) { + if (inode->i_nlink <= 1) { + saved_ino = inode->i_ino; + saved_dev = inode->i_sb->s_dev; + } + + if (!gr_acl_handle_unlink(dentry, nd.mnt)) + error = -EACCES; + atomic_inc(&inode->i_count); - error = vfs_unlink(nd.dentry->d_inode, dentry); + } + if (!error) + error = vfs_unlink(nd.dentry->d_inode, dentry, &nd); + if (!error && (saved_ino || saved_dev)) + gr_handle_delete(saved_ino, saved_dev); exit2: dput(dentry); } @@ -2133,9 +2309,10 @@ asmlinkage long sys_unlink(const char __ return do_unlinkat(AT_FDCWD, pathname); } -int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname, int mode) +int vfs_symlink(struct inode *dir, struct dentry *dentry, + const char *oldname, int mode, struct nameidata *nd) { - int error = may_create(dir, dentry, NULL); + int error = may_create(dir, dentry, nd); if (error) return error; @@ -2176,7 +2353,15 @@ asmlinkage long sys_symlinkat(const char dentry = lookup_create(&nd, 0); error = PTR_ERR(dentry); if (!IS_ERR(dentry)) { - error = vfs_symlink(nd.dentry->d_inode, dentry, from, S_IALLUGO); + error = 0; + if (!gr_acl_handle_symlink(dentry, nd.dentry, nd.mnt, from)) + error = -EACCES; + + if (!error) + error = vfs_symlink(nd.dentry->d_inode, dentry, from, S_IALLUGO, &nd); + + if (!error) + gr_handle_create(dentry, nd.mnt); dput(dentry); } mutex_unlock(&nd.dentry->d_inode->i_mutex); @@ -2193,7 +2378,8 @@ asmlinkage long sys_symlink(const char _ return sys_symlinkat(oldname, AT_FDCWD, newname); } -int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry) +int vfs_link(struct dentry *old_dentry, struct inode *dir, + struct dentry *new_dentry, struct nameidata *nd) { struct inode *inode = old_dentry->d_inode; int error; @@ -2201,7 +2387,7 @@ int vfs_link(struct dentry *old_dentry, if (!inode) return -ENOENT; - error = may_create(dir, new_dentry, NULL); + error = may_create(dir, new_dentry, nd); if (error) return error; @@ -2211,7 +2397,7 @@ int vfs_link(struct dentry *old_dentry, /* * A link to an append-only or immutable file cannot be created. */ - if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) + if (IS_APPEND(inode) || IS_IXORUNLINK(inode)) return -EPERM; if (!dir->i_op || !dir->i_op->link) return -EPERM; @@ -2268,7 +2454,20 @@ asmlinkage long sys_linkat(int olddfd, c new_dentry = lookup_create(&nd, 0); error = PTR_ERR(new_dentry); if (!IS_ERR(new_dentry)) { - error = vfs_link(old_nd.dentry, nd.dentry->d_inode, new_dentry); + error = 0; + if (gr_handle_hardlink(old_nd.dentry, old_nd.mnt, + old_nd.dentry->d_inode, + old_nd.dentry->d_inode->i_mode, to)) + error = -EPERM; + if (!gr_acl_handle_link(new_dentry, nd.dentry, nd.mnt, + old_nd.dentry, old_nd.mnt, to)) + error = -EACCES; + if (!error) + error = vfs_link(old_nd.dentry, nd.dentry->d_inode, new_dentry, &nd); + + if (!error) + gr_handle_create(new_dentry, nd.mnt); + dput(new_dentry); } mutex_unlock(&nd.dentry->d_inode->i_mutex); @@ -2400,14 +2599,14 @@ int vfs_rename(struct inode *old_dir, st if (old_dentry->d_inode == new_dentry->d_inode) return 0; - error = may_delete(old_dir, old_dentry, is_dir); + error = may_delete(old_dir, old_dentry, is_dir, NULL); if (error) return error; if (!new_dentry->d_inode) error = may_create(new_dir, new_dentry, NULL); else - error = may_delete(new_dir, new_dentry, is_dir); + error = may_delete(new_dir, new_dentry, is_dir, NULL); if (error) return error; @@ -2485,6 +2684,9 @@ static int do_rename(int olddfd, const c error = -EINVAL; if (old_dentry == trap) goto exit4; + error = -EROFS; + if (MNT_IS_RDONLY(newnd.mnt)) + goto exit4; new_dentry = lookup_hash(&newnd); error = PTR_ERR(new_dentry); if (IS_ERR(new_dentry)) @@ -2494,8 +2696,16 @@ static int do_rename(int olddfd, const c if (new_dentry == trap) goto exit5; - error = vfs_rename(old_dir->d_inode, old_dentry, + error = gr_acl_handle_rename(new_dentry, newnd.dentry, newnd.mnt, + old_dentry, old_dir->d_inode, oldnd.mnt, + newname); + + if (!error) + error = vfs_rename(old_dir->d_inode, old_dentry, new_dir->d_inode, new_dentry); + if (!error) + gr_handle_rename(old_dir->d_inode, newnd.dentry->d_inode, old_dentry, + new_dentry, oldnd.mnt, new_dentry->d_inode ? 1 : 0); exit5: dput(new_dentry); exit4: diff -NurpP --minimal linux-2.6.17.13/fs/namespace.c linux-2.6.17.13-g-v/fs/namespace.c --- linux-2.6.17.13/fs/namespace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/namespace.c 2006-09-20 15:00:12.000000000 +0200 @@ -23,6 +23,10 @@ #include #include #include +#include +#include +#include +#include #include #include #include "pnode.h" @@ -241,6 +245,7 @@ static struct vfsmount *clone_mnt(struct mnt->mnt_root = dget(root); mnt->mnt_mountpoint = mnt->mnt_root; mnt->mnt_parent = mnt; + mnt->mnt_xid = old->mnt_xid; if (flag & CL_SLAVE) { list_add(&mnt->mnt_slave, &old->mnt_slave_list); @@ -349,43 +354,85 @@ static inline void mangle(struct seq_fil seq_escape(m, s, " \t\n\\"); } +static int mnt_is_reachable(struct vfsmount *mnt) +{ + struct vfsmount *root_mnt; + struct dentry *root, *point; + int ret; + + if (mnt == mnt->mnt_namespace->root) + return 1; + + spin_lock(&dcache_lock); + root_mnt = current->fs->rootmnt; + root = current->fs->root; + point = root; + + while ((mnt != mnt->mnt_parent) && (mnt != root_mnt)) { + point = mnt->mnt_mountpoint; + mnt = mnt->mnt_parent; + } + + ret = (mnt == root_mnt) && is_subdir(point, root); + + spin_unlock(&dcache_lock); + + return ret; +} + static int show_vfsmnt(struct seq_file *m, void *v) { struct vfsmount *mnt = v; int err = 0; static struct proc_fs_info { - int flag; - char *str; + int s_flag; + int mnt_flag; + char *set_str; + char *unset_str; } fs_info[] = { - { MS_SYNCHRONOUS, ",sync" }, - { MS_DIRSYNC, ",dirsync" }, - { MS_MANDLOCK, ",mand" }, - { 0, NULL } - }; - static struct proc_fs_info mnt_info[] = { - { MNT_NOSUID, ",nosuid" }, - { MNT_NODEV, ",nodev" }, - { MNT_NOEXEC, ",noexec" }, - { MNT_NOATIME, ",noatime" }, - { MNT_NODIRATIME, ",nodiratime" }, - { 0, NULL } + { MS_RDONLY, MNT_RDONLY, "ro", "rw" }, + { MS_SYNCHRONOUS, 0, ",sync", NULL }, + { MS_DIRSYNC, 0, ",dirsync", NULL }, + { MS_MANDLOCK, 0, ",mand", NULL }, + { MS_TAGXID, 0, ",tagxid", NULL }, + { MS_NOATIME, MNT_NOATIME, ",noatime", NULL }, + { MS_NODIRATIME, MNT_NODIRATIME, ",nodiratime", NULL }, + { 0, MNT_NOSUID, ",nosuid", NULL }, + { 0, MNT_NODEV, ",nodev", NULL }, + { 0, MNT_NOEXEC, ",noexec", NULL }, + { 0, 0, NULL, NULL } }; - struct proc_fs_info *fs_infop; + struct proc_fs_info *p; + unsigned long s_flags = mnt->mnt_sb->s_flags; + int mnt_flags = mnt->mnt_flags; - mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none"); - seq_putc(m, ' '); - seq_path(m, mnt, mnt->mnt_root, " \t\n\\"); - seq_putc(m, ' '); - mangle(m, mnt->mnt_sb->s_type->name); - seq_puts(m, mnt->mnt_sb->s_flags & MS_RDONLY ? " ro" : " rw"); - for (fs_infop = fs_info; fs_infop->flag; fs_infop++) { - if (mnt->mnt_sb->s_flags & fs_infop->flag) - seq_puts(m, fs_infop->str); + if (vx_flags(VXF_HIDE_MOUNT, 0)) + return 0; + if (!mnt_is_reachable(mnt)) + return 0; + + if (!vx_check(0, VX_ADMIN|VX_WATCH) && + mnt == current->fs->rootmnt) { + seq_puts(m, "/dev/root / "); + } else { + mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none"); + seq_putc(m, ' '); + seq_path(m, mnt, mnt->mnt_root, " \t\n\\"); + seq_putc(m, ' '); } - for (fs_infop = mnt_info; fs_infop->flag; fs_infop++) { - if (mnt->mnt_flags & fs_infop->flag) - seq_puts(m, fs_infop->str); + mangle(m, mnt->mnt_sb->s_type->name); + seq_putc(m, ' '); + for (p = fs_info; (p->s_flag | p->mnt_flag) ; p++) { + if ((s_flags & p->s_flag) || (mnt_flags & p->mnt_flag)) { + if (p->set_str) + seq_puts(m, p->set_str); + } else { + if (p->unset_str) + seq_puts(m, p->unset_str); + } } + if (mnt->mnt_flags & MNT_XID) + seq_printf(m, ",xid=%d", mnt->mnt_xid); if (mnt->mnt_sb->s_op->show_options) err = mnt->mnt_sb->s_op->show_options(m, mnt); seq_puts(m, " 0 0\n"); @@ -404,17 +451,27 @@ static int show_vfsstat(struct seq_file struct vfsmount *mnt = v; int err = 0; - /* device */ - if (mnt->mnt_devname) { - seq_puts(m, "device "); - mangle(m, mnt->mnt_devname); - } else - seq_puts(m, "no device"); + if (vx_flags(VXF_HIDE_MOUNT, 0)) + return 0; + if (!mnt_is_reachable(mnt) && !vx_check(0, VX_WATCH)) + return 0; - /* mount point */ - seq_puts(m, " mounted on "); - seq_path(m, mnt, mnt->mnt_root, " \t\n\\"); - seq_putc(m, ' '); + if (!vx_check(0, VX_ADMIN|VX_WATCH) && + mnt == current->fs->rootmnt) { + seq_puts(m, "device /dev/root mounted on / "); + } else { + /* device */ + if (mnt->mnt_devname) { + seq_puts(m, "device "); + mangle(m, mnt->mnt_devname); + } else + seq_puts(m, "no device"); + + /* mount point */ + seq_puts(m, " mounted on "); + seq_path(m, mnt, mnt->mnt_root, " \t\n\\"); + seq_putc(m, ' '); + } /* file system type */ seq_puts(m, "with fstype "); @@ -600,6 +657,8 @@ static int do_umount(struct vfsmount *mn DQUOT_OFF(sb); retval = do_remount_sb(sb, MS_RDONLY, NULL, 0); unlock_kernel(); + + gr_log_remount(mnt->mnt_devname, retval); } up_write(&sb->s_umount); return retval; @@ -620,6 +679,9 @@ static int do_umount(struct vfsmount *mn security_sb_umount_busy(mnt); up_write(&namespace_sem); release_mounts(&umount_list); + + gr_log_unmount(mnt->mnt_devname, retval); + return retval; } @@ -646,7 +708,7 @@ asmlinkage long sys_umount(char __user * goto dput_and_out; retval = -EPERM; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT)) goto dput_and_out; retval = do_umount(nd.mnt, flags); @@ -670,7 +732,7 @@ asmlinkage long sys_oldumount(char __use static int mount_is_safe(struct nameidata *nd) { - if (capable(CAP_SYS_ADMIN)) + if (vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT)) return 0; return -EPERM; #ifdef notyet @@ -899,11 +961,13 @@ static int do_change_type(struct nameida /* * do loopback mount. */ -static int do_loopback(struct nameidata *nd, char *old_name, int recurse) +static int do_loopback(struct nameidata *nd, char *old_name, xid_t xid, + unsigned long flags, int mnt_flags) { struct nameidata old_nd; struct vfsmount *mnt = NULL; int err = mount_is_safe(nd); + int recurse = flags & MS_REC; if (err) return err; if (!old_name || !*old_name) @@ -929,6 +993,12 @@ static int do_loopback(struct nameidata if (!mnt) goto out; + mnt->mnt_flags = mnt_flags; + if (flags & MS_XID) { + mnt->mnt_xid = xid; + mnt->mnt_flags |= MNT_XID; + } + err = graft_tree(mnt, nd); if (err) { LIST_HEAD(umount_list); @@ -937,6 +1007,7 @@ static int do_loopback(struct nameidata spin_unlock(&vfsmount_lock); release_mounts(&umount_list); } + mnt->mnt_flags = mnt_flags; out: up_write(&namespace_sem); @@ -950,12 +1021,12 @@ out: * on it - tough luck. */ static int do_remount(struct nameidata *nd, int flags, int mnt_flags, - void *data) + void *data, xid_t xid) { int err; struct super_block *sb = nd->mnt->mnt_sb; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_REMOUNT)) return -EPERM; if (!check_mnt(nd->mnt)) @@ -989,7 +1060,7 @@ static int do_move_mount(struct nameidat struct nameidata old_nd, parent_nd; struct vfsmount *p; int err = 0; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT)) return -EPERM; if (!old_name || !*old_name) return -EINVAL; @@ -1069,7 +1140,7 @@ static int do_new_mount(struct nameidata return -EINVAL; /* we need capabilities... */ - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT)) return -EPERM; mnt = do_kern_mount(type, flags, name, data); @@ -1307,6 +1378,7 @@ long do_mount(char *dev_name, char *dir_ struct nameidata nd; int retval = 0; int mnt_flags = 0; + xid_t xid = 0; /* Discard magic */ if ((flags & MS_MGC_MSK) == MS_MGC_VAL) @@ -1322,7 +1394,17 @@ long do_mount(char *dev_name, char *dir_ if (data_page) ((char *)data_page)[PAGE_SIZE - 1] = 0; + retval = vx_parse_xid(data_page, &xid, 1); + if (retval) { + mnt_flags |= MNT_XID; + /* bind and re-mounts get xid flag */ + if (flags & (MS_BIND|MS_REMOUNT)) + flags |= MS_XID; + } + /* Separate the per-mountpoint flags */ + if (flags & MS_RDONLY) + mnt_flags |= MNT_RDONLY; if (flags & MS_NOSUID) mnt_flags |= MNT_NOSUID; if (flags & MS_NODEV) @@ -1334,6 +1416,8 @@ long do_mount(char *dev_name, char *dir_ if (flags & MS_NODIRATIME) mnt_flags |= MNT_NODIRATIME; + if (!capable(CAP_SYS_ADMIN)) + mnt_flags |= MNT_NODEV; flags &= ~(MS_NOSUID | MS_NOEXEC | MS_NODEV | MS_ACTIVE | MS_NOATIME | MS_NODIRATIME); @@ -1346,11 +1430,16 @@ long do_mount(char *dev_name, char *dir_ if (retval) goto dput_out; + if (gr_handle_chroot_mount(nd.dentry, nd.mnt, dev_name)) { + retval = -EPERM; + goto dput_out; + } + if (flags & MS_REMOUNT) retval = do_remount(&nd, flags & ~MS_REMOUNT, mnt_flags, - data_page); + data_page, xid); else if (flags & MS_BIND) - retval = do_loopback(&nd, dev_name, flags & MS_REC); + retval = do_loopback(&nd, dev_name, xid, flags, mnt_flags); else if (flags & (MS_SHARED | MS_PRIVATE | MS_SLAVE | MS_UNBINDABLE)) retval = do_change_type(&nd, flags); else if (flags & MS_MOVE) @@ -1360,6 +1449,9 @@ long do_mount(char *dev_name, char *dir_ dev_name, data_page); dput_out: path_release(&nd); + + gr_log_mount(dev_name, dir_name, retval); + return retval; } @@ -1448,7 +1540,7 @@ int copy_namespace(int flags, struct tas if (!(flags & CLONE_NEWNS)) return 0; - if (!capable(CAP_SYS_ADMIN)) { + if (!vx_capable(CAP_SYS_ADMIN, VXC_SECURE_MOUNT)) { err = -EPERM; goto out; } @@ -1612,6 +1704,9 @@ asmlinkage long sys_pivot_root(const cha if (!capable(CAP_SYS_ADMIN)) return -EPERM; + if (gr_handle_chroot_pivot()) + return -EPERM; + lock_kernel(); error = __user_walk(new_root, LOOKUP_FOLLOW | LOOKUP_DIRECTORY, diff -NurpP --minimal linux-2.6.17.13/fs/nfs/dir.c linux-2.6.17.13-g-v/fs/nfs/dir.c --- linux-2.6.17.13/fs/nfs/dir.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/nfs/dir.c 2006-09-20 15:00:12.000000000 +0200 @@ -28,9 +28,11 @@ #include #include #include +#include #include #include #include +#include #include "nfs4_fs.h" #include "delegation.h" @@ -904,6 +906,7 @@ static struct dentry *nfs_lookup(struct res = (struct dentry *)inode; if (IS_ERR(res)) goto out_unlock; + vx_propagate_xid(nd, inode); no_entry: res = d_add_unique(dentry, inode); if (res != NULL) @@ -937,7 +940,8 @@ static int is_atomic_open(struct inode * if (nd->flags & LOOKUP_DIRECTORY) return 0; /* Are we trying to write to a read only partition? */ - if (IS_RDONLY(dir) && (nd->intent.open.flags & (O_CREAT|O_TRUNC|FMODE_WRITE))) + if ((IS_RDONLY(dir) || MNT_IS_RDONLY(nd->mnt)) && + (nd->intent.open.flags & (O_CREAT|O_TRUNC|FMODE_WRITE))) return 0; return 1; } diff -NurpP --minimal linux-2.6.17.13/fs/nfs/inode.c linux-2.6.17.13-g-v/fs/nfs/inode.c --- linux-2.6.17.13/fs/nfs/inode.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/nfs/inode.c 2006-09-20 15:00:12.000000000 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include #include @@ -343,12 +344,16 @@ nfs_sb_init(struct super_block *sb, rpc_ } server->backing_dev_info.ra_pages = server->rpages * NFS_MAX_READAHEAD; + if (server->flags & NFS_MOUNT_TAGXID) + sb->s_flags |= MS_TAGXID; + sb->s_maxbytes = fsinfo.maxfilesize; if (sb->s_maxbytes > MAX_LFS_FILESIZE) sb->s_maxbytes = MAX_LFS_FILESIZE; server->client->cl_intr = (server->flags & NFS_MOUNT_INTR) ? 1 : 0; server->client->cl_softrtry = (server->flags & NFS_MOUNT_SOFT) ? 1 : 0; + server->client->cl_tagxid = (server->flags & NFS_MOUNT_TAGXID) ? 1 : 0; /* We're airborne Set socket buffersize */ rpc_setbufsize(server->client, server->wsize + 100, server->rsize + 100); @@ -423,6 +428,7 @@ nfs_create_client(struct nfs_server *ser clnt->cl_intr = 1; clnt->cl_softrtry = 1; + clnt->cl_tagxid = 1; return clnt; @@ -602,6 +608,7 @@ static void nfs_show_mount_options(struc { NFS_MOUNT_NOAC, ",noac", "" }, { NFS_MOUNT_NONLM, ",nolock", "" }, { NFS_MOUNT_NOACL, ",noacl", "" }, + { NFS_MOUNT_TAGXID, ",tagxid", "" }, { 0, NULL, NULL } }; struct proc_nfs_info *nfs_infop; @@ -903,8 +910,10 @@ nfs_fhget(struct super_block *sb, struct nfsi->change_attr = fattr->change_attr; inode->i_size = nfs_size_to_loff_t(fattr->size); inode->i_nlink = fattr->nlink; - inode->i_uid = fattr->uid; - inode->i_gid = fattr->gid; + inode->i_uid = INOXID_UID(XID_TAG(inode), fattr->uid, fattr->gid); + inode->i_gid = INOXID_GID(XID_TAG(inode), fattr->uid, fattr->gid); + inode->i_xid = INOXID_XID(XID_TAG(inode), fattr->uid, fattr->gid, 0); + /* maybe fattr->xid someday */ if (fattr->valid & (NFS_ATTR_FATTR_V3 | NFS_ATTR_FATTR_V4)) { /* * report the blocks in 512byte units @@ -995,6 +1004,8 @@ void nfs_setattr_update_inode(struct ino inode->i_uid = attr->ia_uid; if ((attr->ia_valid & ATTR_GID) != 0) inode->i_gid = attr->ia_gid; + if ((attr->ia_valid & ATTR_XID) && IS_TAGXID(inode)) + inode->i_xid = attr->ia_xid; spin_lock(&inode->i_lock); NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL; spin_unlock(&inode->i_lock); @@ -1397,6 +1408,9 @@ static int nfs_check_inode_attributes(st struct nfs_inode *nfsi = NFS_I(inode); loff_t cur_size, new_isize; int data_unstable; + uid_t uid; + gid_t gid; + xid_t xid; if ((fattr->valid & NFS_ATTR_FATTR) == 0) @@ -1434,10 +1448,15 @@ static int nfs_check_inode_attributes(st if (cur_size != new_isize && nfsi->npages == 0) nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE; + uid = INOXID_UID(XID_TAG(inode), fattr->uid, fattr->gid); + gid = INOXID_GID(XID_TAG(inode), fattr->uid, fattr->gid); + xid = INOXID_XID(XID_TAG(inode), fattr->uid, fattr->gid, 0); + /* Have any file permissions changed? */ if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO) - || inode->i_uid != fattr->uid - || inode->i_gid != fattr->gid) + || inode->i_uid != uid + || inode->i_gid != gid + || inode->i_xid != xid) nfsi->cache_validity |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL; /* Has the link count changed? */ @@ -1522,6 +1541,9 @@ static int nfs_update_inode(struct inode loff_t cur_isize, new_isize; unsigned int invalid = 0; int data_stable; + uid_t uid; + gid_t gid; + xid_t xid; dfprintk(VFS, "NFS: %s(%s/%ld ct=%d info=0x%x)\n", __FUNCTION__, inode->i_sb->s_id, inode->i_ino, @@ -1591,15 +1613,21 @@ static int nfs_update_inode(struct inode } memcpy(&inode->i_atime, &fattr->atime, sizeof(inode->i_atime)); + uid = INOXID_UID(XID_TAG(inode), fattr->uid, fattr->gid); + gid = INOXID_GID(XID_TAG(inode), fattr->uid, fattr->gid); + xid = INOXID_XID(XID_TAG(inode), fattr->uid, fattr->gid, 0); + if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO) || - inode->i_uid != fattr->uid || - inode->i_gid != fattr->gid) + inode->i_uid != uid || + inode->i_gid != gid || + inode->i_xid != xid) invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL; inode->i_mode = fattr->mode; inode->i_nlink = fattr->nlink; - inode->i_uid = fattr->uid; - inode->i_gid = fattr->gid; + inode->i_uid = uid; + inode->i_gid = gid; + inode->i_xid = xid; if (fattr->valid & (NFS_ATTR_FATTR_V3 | NFS_ATTR_FATTR_V4)) { /* diff -NurpP --minimal linux-2.6.17.13/fs/nfs/nfs3xdr.c linux-2.6.17.13-g-v/fs/nfs/nfs3xdr.c --- linux-2.6.17.13/fs/nfs/nfs3xdr.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/nfs/nfs3xdr.c 2006-09-20 15:00:12.000000000 +0200 @@ -22,6 +22,7 @@ #include #include #include +#include #define NFSDBG_FACILITY NFSDBG_XDR @@ -178,7 +179,7 @@ xdr_decode_fattr(u32 *p, struct nfs_fatt } static inline u32 * -xdr_encode_sattr(u32 *p, struct iattr *attr) +xdr_encode_sattr(u32 *p, struct iattr *attr, int tagxid) { if (attr->ia_valid & ATTR_MODE) { *p++ = xdr_one; @@ -186,15 +187,17 @@ xdr_encode_sattr(u32 *p, struct iattr *a } else { *p++ = xdr_zero; } - if (attr->ia_valid & ATTR_UID) { + if (attr->ia_valid & ATTR_UID || + (tagxid && (attr->ia_valid & ATTR_XID))) { *p++ = xdr_one; - *p++ = htonl(attr->ia_uid); + *p++ = htonl(XIDINO_UID(tagxid, attr->ia_uid, attr->ia_xid)); } else { *p++ = xdr_zero; } - if (attr->ia_valid & ATTR_GID) { + if (attr->ia_valid & ATTR_GID || + (tagxid && (attr->ia_valid & ATTR_XID))) { *p++ = xdr_one; - *p++ = htonl(attr->ia_gid); + *p++ = htonl(XIDINO_GID(tagxid, attr->ia_gid, attr->ia_xid)); } else { *p++ = xdr_zero; } @@ -279,7 +282,8 @@ static int nfs3_xdr_sattrargs(struct rpc_rqst *req, u32 *p, struct nfs3_sattrargs *args) { p = xdr_encode_fhandle(p, args->fh); - p = xdr_encode_sattr(p, args->sattr); + p = xdr_encode_sattr(p, args->sattr, + req->rq_task->tk_client->cl_tagxid); *p++ = htonl(args->guard); if (args->guard) p = xdr_encode_time3(p, &args->guardtime); @@ -370,7 +374,8 @@ nfs3_xdr_createargs(struct rpc_rqst *req *p++ = args->verifier[0]; *p++ = args->verifier[1]; } else - p = xdr_encode_sattr(p, args->sattr); + p = xdr_encode_sattr(p, args->sattr, + req->rq_task->tk_client->cl_tagxid); req->rq_slen = xdr_adjust_iovec(req->rq_svec, p); return 0; @@ -384,7 +389,8 @@ nfs3_xdr_mkdirargs(struct rpc_rqst *req, { p = xdr_encode_fhandle(p, args->fh); p = xdr_encode_array(p, args->name, args->len); - p = xdr_encode_sattr(p, args->sattr); + p = xdr_encode_sattr(p, args->sattr, + req->rq_task->tk_client->cl_tagxid); req->rq_slen = xdr_adjust_iovec(req->rq_svec, p); return 0; } @@ -397,7 +403,8 @@ nfs3_xdr_symlinkargs(struct rpc_rqst *re { p = xdr_encode_fhandle(p, args->fromfh); p = xdr_encode_array(p, args->fromname, args->fromlen); - p = xdr_encode_sattr(p, args->sattr); + p = xdr_encode_sattr(p, args->sattr, + req->rq_task->tk_client->cl_tagxid); p = xdr_encode_array(p, args->topath, args->tolen); req->rq_slen = xdr_adjust_iovec(req->rq_svec, p); return 0; @@ -412,7 +419,8 @@ nfs3_xdr_mknodargs(struct rpc_rqst *req, p = xdr_encode_fhandle(p, args->fh); p = xdr_encode_array(p, args->name, args->len); *p++ = htonl(args->type); - p = xdr_encode_sattr(p, args->sattr); + p = xdr_encode_sattr(p, args->sattr, + req->rq_task->tk_client->cl_tagxid); if (args->type == NF3CHR || args->type == NF3BLK) { *p++ = htonl(MAJOR(args->rdev)); *p++ = htonl(MINOR(args->rdev)); diff -NurpP --minimal linux-2.6.17.13/fs/nfs/nfsroot.c linux-2.6.17.13-g-v/fs/nfs/nfsroot.c --- linux-2.6.17.13/fs/nfs/nfsroot.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/nfs/nfsroot.c 2006-09-20 15:00:12.000000000 +0200 @@ -87,6 +87,7 @@ #include #include #include +#include /* Define this to allow debugging output */ #undef NFSROOT_DEBUG @@ -124,7 +125,7 @@ enum { Opt_soft, Opt_hard, Opt_intr, Opt_nointr, Opt_posix, Opt_noposix, Opt_cto, Opt_nocto, Opt_ac, Opt_noac, Opt_lock, Opt_nolock, Opt_v2, Opt_v3, Opt_udp, Opt_tcp, - Opt_acl, Opt_noacl, + Opt_acl, Opt_noacl, Opt_tagxid, /* Error token */ Opt_err }; @@ -161,6 +162,7 @@ static match_table_t __initdata tokens = {Opt_tcp, "tcp"}, {Opt_acl, "acl"}, {Opt_noacl, "noacl"}, + {Opt_tagxid, "tagxid"}, {Opt_err, NULL} }; @@ -275,6 +277,11 @@ static int __init root_nfs_parse(char *n case Opt_noacl: nfs_data.flags |= NFS_MOUNT_NOACL; break; +#ifndef CONFIG_INOXID_NONE + case Opt_tagxid: + nfs_data.flags |= NFS_MOUNT_TAGXID; + break; +#endif default: printk(KERN_WARNING "Root-NFS: unknown " "option: %s\n", p); @@ -312,7 +319,7 @@ static int __init root_nfs_name(char *na /* Override them by options set on kernel command-line */ root_nfs_parse(name, buf); - cp = system_utsname.nodename; + cp = vx_new_uts(nodename); if (strlen(buf) + strlen(cp) > NFS_MAXPATHLEN) { printk(KERN_ERR "Root-NFS: Pathname for remote directory too long.\n"); return -1; diff -NurpP --minimal linux-2.6.17.13/fs/nfsd/auth.c linux-2.6.17.13-g-v/fs/nfsd/auth.c --- linux-2.6.17.13/fs/nfsd/auth.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/nfsd/auth.c 2006-09-20 15:00:12.000000000 +0200 @@ -9,6 +9,7 @@ #include #include #include +#include #define CAP_NFSD_MASK (CAP_FS_MASK|CAP_TO_MASK(CAP_SYS_RESOURCE)) @@ -41,19 +42,22 @@ int nfsd_setuser(struct svc_rqst *rqstp, get_group_info(cred.cr_group_info); if (cred.cr_uid != (uid_t) -1) - current->fsuid = cred.cr_uid; + current->fsuid = INOXID_UID(XID_TAG_NFSD, cred.cr_uid, cred.cr_gid); else current->fsuid = exp->ex_anon_uid; if (cred.cr_gid != (gid_t) -1) - current->fsgid = cred.cr_gid; + current->fsgid = INOXID_GID(XID_TAG_NFSD, cred.cr_uid, cred.cr_gid); else current->fsgid = exp->ex_anon_gid; + /* this desperately needs a tag :) */ + current->xid = INOXID_XID(XID_TAG_NFSD, cred.cr_uid, cred.cr_gid, 0); + if (!cred.cr_group_info) return -ENOMEM; ret = set_current_groups(cred.cr_group_info); put_group_info(cred.cr_group_info); - if ((cred.cr_uid)) { + if (INOXID_UID(XID_TAG_NFSD, cred.cr_uid, cred.cr_gid)) { cap_t(current->cap_effective) &= ~CAP_NFSD_MASK; } else { cap_t(current->cap_effective) |= (CAP_NFSD_MASK & diff -NurpP --minimal linux-2.6.17.13/fs/nfsd/nfs3xdr.c linux-2.6.17.13-g-v/fs/nfsd/nfs3xdr.c --- linux-2.6.17.13/fs/nfsd/nfs3xdr.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/nfsd/nfs3xdr.c 2006-09-20 15:00:12.000000000 +0200 @@ -21,6 +21,7 @@ #include #include #include +#include #define NFSDDBG_FACILITY NFSDDBG_XDR @@ -111,6 +112,8 @@ static inline u32 * decode_sattr3(u32 *p, struct iattr *iap) { u32 tmp; + uid_t uid = 0; + gid_t gid = 0; iap->ia_valid = 0; @@ -120,12 +123,15 @@ decode_sattr3(u32 *p, struct iattr *iap) } if (*p++) { iap->ia_valid |= ATTR_UID; - iap->ia_uid = ntohl(*p++); + uid = ntohl(*p++); } if (*p++) { iap->ia_valid |= ATTR_GID; - iap->ia_gid = ntohl(*p++); + gid = ntohl(*p++); } + iap->ia_uid = INOXID_UID(XID_TAG_NFSD, uid, gid); + iap->ia_gid = INOXID_GID(XID_TAG_NFSD, uid, gid); + iap->ia_xid = INOXID_XID(XID_TAG_NFSD, uid, gid, 0); if (*p++) { u64 newsize; @@ -163,8 +169,10 @@ encode_fattr3(struct svc_rqst *rqstp, u3 *p++ = htonl(nfs3_ftypes[(stat->mode & S_IFMT) >> 12]); *p++ = htonl((u32) stat->mode); *p++ = htonl((u32) stat->nlink); - *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid)); - *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid)); + *p++ = htonl((u32) nfsd_ruid(rqstp, + XIDINO_UID(XID_TAG(dentry->d_inode), stat->uid, stat->xid))); + *p++ = htonl((u32) nfsd_rgid(rqstp, + XIDINO_GID(XID_TAG(dentry->d_inode), stat->gid, stat->xid))); if (S_ISLNK(stat->mode) && stat->size > NFS3_MAXPATHLEN) { p = xdr_encode_hyper(p, (u64) NFS3_MAXPATHLEN); } else { diff -NurpP --minimal linux-2.6.17.13/fs/nfsd/nfs4recover.c linux-2.6.17.13-g-v/fs/nfsd/nfs4recover.c --- linux-2.6.17.13/fs/nfsd/nfs4recover.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/nfsd/nfs4recover.c 2006-09-20 15:00:12.000000000 +0200 @@ -155,7 +155,7 @@ nfsd4_create_clid_dir(struct nfs4_client dprintk("NFSD: nfsd4_create_clid_dir: DIRECTORY EXISTS\n"); goto out_put; } - status = vfs_mkdir(rec_dir.dentry->d_inode, dentry, S_IRWXU); + status = vfs_mkdir(rec_dir.dentry->d_inode, dentry, S_IRWXU, NULL); out_put: dput(dentry); out_unlock: @@ -259,7 +259,7 @@ nfsd4_remove_clid_file(struct dentry *di return -EINVAL; } mutex_lock(&dir->d_inode->i_mutex); - status = vfs_unlink(dir->d_inode, dentry); + status = vfs_unlink(dir->d_inode, dentry, NULL); mutex_unlock(&dir->d_inode->i_mutex); return status; } @@ -274,7 +274,7 @@ nfsd4_clear_clid_dir(struct dentry *dir, * a kernel from the future.... */ nfsd4_list_rec_dir(dentry, nfsd4_remove_clid_file); mutex_lock(&dir->d_inode->i_mutex); - status = vfs_rmdir(dir->d_inode, dentry); + status = vfs_rmdir(dir->d_inode, dentry, NULL); mutex_unlock(&dir->d_inode->i_mutex); return status; } diff -NurpP --minimal linux-2.6.17.13/fs/nfsd/nfs4xdr.c linux-2.6.17.13-g-v/fs/nfsd/nfs4xdr.c --- linux-2.6.17.13/fs/nfsd/nfs4xdr.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/nfsd/nfs4xdr.c 2006-09-20 15:00:12.000000000 +0200 @@ -57,6 +57,7 @@ #include #include #include +#include #define NFSDDBG_FACILITY NFSDDBG_XDR @@ -1560,14 +1561,18 @@ out_acl: WRITE32(stat.nlink); } if (bmval1 & FATTR4_WORD1_OWNER) { - status = nfsd4_encode_user(rqstp, stat.uid, &p, &buflen); + status = nfsd4_encode_user(rqstp, + XIDINO_UID(XID_TAG(dentry->d_inode), + stat.uid, stat.xid), &p, &buflen); if (status == nfserr_resource) goto out_resource; if (status) goto out; } if (bmval1 & FATTR4_WORD1_OWNER_GROUP) { - status = nfsd4_encode_group(rqstp, stat.gid, &p, &buflen); + status = nfsd4_encode_group(rqstp, + XIDINO_GID(XID_TAG(dentry->d_inode), + stat.gid, stat.xid), &p, &buflen); if (status == nfserr_resource) goto out_resource; if (status) diff -NurpP --minimal linux-2.6.17.13/fs/nfsd/nfsxdr.c linux-2.6.17.13-g-v/fs/nfsd/nfsxdr.c --- linux-2.6.17.13/fs/nfsd/nfsxdr.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/nfsd/nfsxdr.c 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,7 @@ #include #include #include +#include #define NFSDDBG_FACILITY NFSDDBG_XDR @@ -102,6 +103,8 @@ static inline u32 * decode_sattr(u32 *p, struct iattr *iap) { u32 tmp, tmp1; + uid_t uid = 0; + gid_t gid = 0; iap->ia_valid = 0; @@ -115,12 +118,15 @@ decode_sattr(u32 *p, struct iattr *iap) } if ((tmp = ntohl(*p++)) != (u32)-1) { iap->ia_valid |= ATTR_UID; - iap->ia_uid = tmp; + uid = tmp; } if ((tmp = ntohl(*p++)) != (u32)-1) { iap->ia_valid |= ATTR_GID; - iap->ia_gid = tmp; + gid = tmp; } + iap->ia_uid = INOXID_UID(XID_TAG_NFSD, uid, gid); + iap->ia_gid = INOXID_GID(XID_TAG_NFSD, uid, gid); + iap->ia_xid = INOXID_XID(XID_TAG_NFSD, uid, gid, 0); if ((tmp = ntohl(*p++)) != (u32)-1) { iap->ia_valid |= ATTR_SIZE; iap->ia_size = tmp; @@ -164,8 +170,10 @@ encode_fattr(struct svc_rqst *rqstp, u32 *p++ = htonl(nfs_ftypes[type >> 12]); *p++ = htonl((u32) stat->mode); *p++ = htonl((u32) stat->nlink); - *p++ = htonl((u32) nfsd_ruid(rqstp, stat->uid)); - *p++ = htonl((u32) nfsd_rgid(rqstp, stat->gid)); + *p++ = htonl((u32) nfsd_ruid(rqstp, + XIDINO_UID(XID_TAG(dentry->d_inode), stat->uid, stat->xid))); + *p++ = htonl((u32) nfsd_rgid(rqstp, + XIDINO_GID(XID_TAG(dentry->d_inode), stat->gid, stat->xid))); if (S_ISLNK(type) && stat->size > NFS_MAXPATHLEN) { *p++ = htonl(NFS_MAXPATHLEN); diff -NurpP --minimal linux-2.6.17.13/fs/nfsd/vfs.c linux-2.6.17.13-g-v/fs/nfsd/vfs.c --- linux-2.6.17.13/fs/nfsd/vfs.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/nfsd/vfs.c 2006-09-20 15:00:12.000000000 +0200 @@ -1156,13 +1156,13 @@ nfsd_create(struct svc_rqst *rqstp, stru err = vfs_create(dirp, dchild, iap->ia_mode, NULL); break; case S_IFDIR: - err = vfs_mkdir(dirp, dchild, iap->ia_mode); + err = vfs_mkdir(dirp, dchild, iap->ia_mode, NULL); break; case S_IFCHR: case S_IFBLK: case S_IFIFO: case S_IFSOCK: - err = vfs_mknod(dirp, dchild, iap->ia_mode, rdev); + err = vfs_mknod(dirp, dchild, iap->ia_mode, rdev, NULL); break; default: printk("nfsd: bad file type %o in nfsd_create\n", type); @@ -1442,11 +1442,13 @@ nfsd_symlink(struct svc_rqst *rqstp, str else { strncpy(path_alloced, path, plen); path_alloced[plen] = 0; - err = vfs_symlink(dentry->d_inode, dnew, path_alloced, mode); + err = vfs_symlink(dentry->d_inode, dnew, + path_alloced, mode, NULL); kfree(path_alloced); } } else - err = vfs_symlink(dentry->d_inode, dnew, path, mode); + err = vfs_symlink(dentry->d_inode, dnew, + path, mode, NULL); if (!err) if (EX_ISSYNC(fhp->fh_export)) @@ -1504,7 +1506,7 @@ nfsd_link(struct svc_rqst *rqstp, struct dold = tfhp->fh_dentry; dest = dold->d_inode; - err = vfs_link(dold, dirp, dnew); + err = vfs_link(dold, dirp, dnew, NULL); if (!err) { if (EX_ISSYNC(ffhp->fh_export)) { err = nfserrno(nfsd_sync_dir(ddir)); @@ -1666,9 +1668,9 @@ nfsd_unlink(struct svc_rqst *rqstp, stru err = -EPERM; } else #endif - err = vfs_unlink(dirp, rdentry); + err = vfs_unlink(dirp, rdentry, NULL); } else { /* It's RMDIR */ - err = vfs_rmdir(dirp, rdentry); + err = vfs_rmdir(dirp, rdentry, NULL); } dput(rdentry); @@ -1777,7 +1779,8 @@ nfsd_permission(struct svc_export *exp, */ if (!(acc & MAY_LOCAL_ACCESS)) if (acc & (MAY_WRITE | MAY_SATTR | MAY_TRUNC)) { - if (EX_RDONLY(exp) || IS_RDONLY(inode)) + if (EX_RDONLY(exp) || IS_RDONLY(inode) + || MNT_IS_RDONLY(exp->ex_mnt)) return nfserr_rofs; if (/* (acc & MAY_WRITE) && */ IS_IMMUTABLE(inode)) return nfserr_perm; diff -NurpP --minimal linux-2.6.17.13/fs/open.c linux-2.6.17.13-g-v/fs/open.c --- linux-2.6.17.13/fs/open.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/open.c 2006-09-20 15:00:12.000000000 +0200 @@ -28,6 +28,10 @@ #include #include #include +#include +#include +#include +#include #include @@ -46,6 +50,8 @@ int vfs_statfs(struct super_block *sb, s if (retval == 0 && buf->f_frsize == 0) buf->f_frsize = buf->f_bsize; } + if (!vx_check(0, VX_ADMIN|VX_WATCH)) + vx_vsi_statfs(sb, buf); } return retval; } @@ -207,6 +213,9 @@ int do_truncate(struct dentry *dentry, l if (length < 0) return -EINVAL; + if (filp && !gr_acl_handle_truncate(dentry, filp->f_vfsmnt)) + return -EACCES; + newattrs.ia_size = length; newattrs.ia_valid = ATTR_SIZE | time_attrs; if (filp) { @@ -249,7 +258,7 @@ static long do_sys_truncate(const char _ goto dput_and_out; error = -EROFS; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || MNT_IS_RDONLY(nd.mnt)) goto dput_and_out; error = -EPERM; @@ -379,7 +388,7 @@ asmlinkage long sys_utime(char __user * inode = nd.dentry->d_inode; error = -EROFS; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || MNT_IS_RDONLY(nd.mnt)) goto dput_and_out; /* Don't worry, the checks are done in inode_change_ok() */ @@ -407,6 +416,12 @@ asmlinkage long sys_utime(char __user * (error = vfs_permission(&nd, MAY_WRITE)) != 0) goto dput_and_out; } + + if (!gr_acl_handle_utime(nd.dentry, nd.mnt)) { + error = -EACCES; + goto dput_and_out; + } + mutex_lock(&inode->i_mutex); error = notify_change(nd.dentry, &newattrs); mutex_unlock(&inode->i_mutex); @@ -436,7 +451,7 @@ long do_utimes(int dfd, char __user *fil inode = nd.dentry->d_inode; error = -EROFS; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || MNT_IS_RDONLY(nd.mnt)) goto dput_and_out; /* Don't worry, the checks are done in inode_change_ok() */ @@ -460,6 +475,12 @@ long do_utimes(int dfd, char __user *fil (error = vfs_permission(&nd, MAY_WRITE)) != 0) goto dput_and_out; } + + if (!gr_acl_handle_utime(nd.dentry, nd.mnt)) { + error = -EACCES; + goto dput_and_out; + } + mutex_lock(&inode->i_mutex); error = notify_change(nd.dentry, &newattrs); mutex_unlock(&inode->i_mutex); @@ -523,9 +544,14 @@ asmlinkage long sys_faccessat(int dfd, c if (!res) { res = vfs_permission(&nd, mode); /* SuS v2 requires we report a read only fs too */ - if(!res && (mode & S_IWOTH) && IS_RDONLY(nd.dentry->d_inode) + if(!res && (mode & S_IWOTH) + && (IS_RDONLY(nd.dentry->d_inode) || MNT_IS_RDONLY(nd.mnt)) && !special_file(nd.dentry->d_inode->i_mode)) res = -EROFS; + + if (!res && !gr_acl_handle_access(nd.dentry, nd.mnt, mode)) + res = -EACCES; + path_release(&nd); } @@ -554,6 +580,8 @@ asmlinkage long sys_chdir(const char __u if (error) goto dput_and_out; + gr_log_chdir(nd.dentry, nd.mnt); + set_fs_pwd(current->fs, nd.mnt, nd.dentry); dput_and_out: @@ -584,6 +612,13 @@ asmlinkage long sys_fchdir(unsigned int goto out_putf; error = file_permission(file, MAY_EXEC); + + if (!error && !gr_chroot_fchdir(dentry, mnt)) + error = -EPERM; + + if (!error) + gr_log_chdir(dentry, mnt); + if (!error) set_fs_pwd(current->fs, mnt, dentry); out_putf: @@ -609,8 +644,16 @@ asmlinkage long sys_chroot(const char __ if (!capable(CAP_SYS_CHROOT)) goto dput_and_out; + if (gr_handle_chroot_chroot(nd.dentry, nd.mnt)) + goto dput_and_out; + set_fs_root(current->fs, nd.mnt, nd.dentry); set_fs_altroot(); + + gr_handle_chroot_caps(current); + + gr_handle_chroot_chdir(nd.dentry, nd.mnt); + error = 0; dput_and_out: path_release(&nd); @@ -636,14 +679,27 @@ asmlinkage long sys_fchmod(unsigned int audit_inode(NULL, inode, 0); err = -EROFS; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || MNT_IS_RDONLY(file->f_vfsmnt)) goto out_putf; err = -EPERM; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) goto out_putf; + + if (!gr_acl_handle_fchmod(dentry, file->f_vfsmnt, mode)) { + err = -EACCES; + goto out_putf; + } + mutex_lock(&inode->i_mutex); if (mode == (mode_t) -1) mode = inode->i_mode; + + if (gr_handle_chroot_chmod(dentry, file->f_vfsmnt, mode)) { + err = -EPERM; + mutex_unlock(&inode->i_mutex); + goto out_putf; + } + newattrs.ia_mode = (mode & S_IALLUGO) | (inode->i_mode & ~S_IALLUGO); newattrs.ia_valid = ATTR_MODE | ATTR_CTIME; err = notify_change(dentry, &newattrs); @@ -669,16 +725,28 @@ asmlinkage long sys_fchmodat(int dfd, co inode = nd.dentry->d_inode; error = -EROFS; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || MNT_IS_RDONLY(nd.mnt)) goto dput_and_out; error = -EPERM; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) goto dput_and_out; + if (!gr_acl_handle_chmod(nd.dentry, nd.mnt, mode)) { + error = -EACCES; + goto dput_and_out; + }; + mutex_lock(&inode->i_mutex); if (mode == (mode_t) -1) mode = inode->i_mode; + + if (gr_handle_chroot_chmod(nd.dentry, nd.mnt, mode)) { + error = -EACCES; + mutex_unlock(&inode->i_mutex); + goto dput_and_out; + } + newattrs.ia_mode = (mode & S_IALLUGO) | (inode->i_mode & ~S_IALLUGO); newattrs.ia_valid = ATTR_MODE | ATTR_CTIME; error = notify_change(nd.dentry, &newattrs); @@ -695,7 +763,7 @@ asmlinkage long sys_chmod(const char __u return sys_fchmodat(AT_FDCWD, filename, mode); } -static int chown_common(struct dentry * dentry, uid_t user, gid_t group) +static int chown_common(struct dentry * dentry, uid_t user, gid_t group, struct vfsmount *mnt) { struct inode * inode; int error; @@ -707,19 +775,25 @@ static int chown_common(struct dentry * goto out; } error = -EROFS; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || MNT_IS_RDONLY(mnt)) goto out; error = -EPERM; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) goto out; + + if (!gr_acl_handle_chown(dentry, mnt)) { + error = -EACCES; + goto out; + } + newattrs.ia_valid = ATTR_CTIME; if (user != (uid_t) -1) { newattrs.ia_valid |= ATTR_UID; - newattrs.ia_uid = user; + newattrs.ia_uid = vx_map_uid(user); } if (group != (gid_t) -1) { newattrs.ia_valid |= ATTR_GID; - newattrs.ia_gid = group; + newattrs.ia_gid = vx_map_gid(group); } if (!S_ISDIR(inode->i_mode)) newattrs.ia_valid |= ATTR_KILL_SUID|ATTR_KILL_SGID; @@ -737,7 +811,7 @@ asmlinkage long sys_chown(const char __u error = user_path_walk(filename, &nd); if (!error) { - error = chown_common(nd.dentry, user, group); + error = chown_common(nd.dentry, user, group, nd.mnt); path_release(&nd); } return error; @@ -756,7 +830,7 @@ asmlinkage long sys_fchownat(int dfd, co follow = (flag & AT_SYMLINK_NOFOLLOW) ? 0 : LOOKUP_FOLLOW; error = __user_walk_fd(dfd, filename, follow, &nd); if (!error) { - error = chown_common(nd.dentry, user, group); + error = chown_common(nd.dentry, user, group, nd.mnt); path_release(&nd); } out: @@ -770,7 +844,7 @@ asmlinkage long sys_lchown(const char __ error = user_path_walk_link(filename, &nd); if (!error) { - error = chown_common(nd.dentry, user, group); + error = chown_common(nd.dentry, user, group, nd.mnt); path_release(&nd); } return error; @@ -787,7 +861,7 @@ asmlinkage long sys_fchown(unsigned int struct dentry * dentry; dentry = file->f_dentry; audit_inode(NULL, dentry->d_inode, 0); - error = chown_common(dentry, user, group); + error = chown_common(dentry, user, group, file->f_vfsmnt); fput(file); } return error; @@ -995,6 +1069,7 @@ repeat: * N.B. For clone tasks sharing a files structure, this test * will limit the total number of files that can be opened. */ + gr_learn_resource(current, RLIMIT_NOFILE, fd, 0); if (fd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) goto out; @@ -1015,6 +1090,7 @@ repeat: FD_SET(fd, fdt->open_fds); FD_CLR(fd, fdt->close_on_exec); files->next_fd = fd + 1; + vx_openfd_inc(fd); #if 1 /* Sanity check */ if (fdt->fd[fd] != NULL) { @@ -1037,6 +1113,7 @@ static void __put_unused_fd(struct files __FD_CLR(fd, fdt->open_fds); if (fd < files->next_fd) files->next_fd = fd; + vx_openfd_dec(fd); } void fastcall put_unused_fd(unsigned int fd) diff -NurpP --minimal linux-2.6.17.13/fs/pipe.c linux-2.6.17.13-g-v/fs/pipe.c --- linux-2.6.17.13/fs/pipe.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/pipe.c 2006-09-20 15:00:12.000000000 +0200 @@ -842,7 +842,7 @@ void free_pipe_info(struct inode *inode) inode->i_pipe = NULL; } -static struct vfsmount *pipe_mnt __read_mostly; +struct vfsmount *pipe_mnt __read_mostly; static int pipefs_delete_dentry(struct dentry *dentry) { return 1; diff -NurpP --minimal linux-2.6.17.13/fs/proc/array.c linux-2.6.17.13-g-v/fs/proc/array.c --- linux-2.6.17.13/fs/proc/array.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/proc/array.c 2006-09-20 15:00:12.000000000 +0200 @@ -75,6 +75,9 @@ #include #include #include +#include +#include +#include #include #include @@ -135,7 +138,9 @@ static const char *task_state_array[] = "T (stopped)", /* 4 */ "T (tracing stop)", /* 8 */ "Z (zombie)", /* 16 */ - "X (dead)" /* 32 */ + "X (dead)", /* 32 */ + "N (noninteractive)", /* 64 */ + "H (on hold)" /* 128 */ }; static inline const char * get_task_state(struct task_struct *tsk) @@ -144,7 +149,8 @@ static inline const char * get_task_stat TASK_INTERRUPTIBLE | TASK_UNINTERRUPTIBLE | TASK_STOPPED | - TASK_TRACED)) | + TASK_TRACED | + TASK_ONHOLD)) | (tsk->exit_state & (EXIT_ZOMBIE | EXIT_DEAD)); const char **p = &task_state_array[0]; @@ -161,8 +167,13 @@ static inline char * task_state(struct t struct group_info *group_info; int g; struct fdtable *fdt = NULL; + pid_t pid, ptgid, tppid, tgid; read_lock(&tasklist_lock); + tgid = vx_map_tgid(p->tgid); + pid = vx_map_pid(p->pid); + ptgid = vx_map_pid(p->group_leader->real_parent->tgid); + tppid = vx_map_pid(p->parent->pid); buffer += sprintf(buffer, "State:\t%s\n" "SleepAVG:\t%lu%%\n" @@ -174,9 +185,8 @@ static inline char * task_state(struct t "Gid:\t%d\t%d\t%d\t%d\n", get_task_state(p), (p->sleep_avg/1024)*100/(1020000000/1024), - p->tgid, - p->pid, pid_alive(p) ? p->group_leader->real_parent->tgid : 0, - pid_alive(p) && p->ptrace ? p->parent->pid : 0, + tgid, pid, (pid > 1) ? ptgid : 0, + pid_alive(p) && p->ptrace ? tppid : 0, p->uid, p->euid, p->suid, p->fsuid, p->gid, p->egid, p->sgid, p->fsgid); read_unlock(&tasklist_lock); @@ -293,9 +303,30 @@ static inline char *task_cap(struct task cap_t(p->cap_effective)); } +#if defined(CONFIG_PAX_NOEXEC) || defined(CONFIG_PAX_ASLR) +static inline char *task_pax(struct task_struct *p, char *buffer) +{ + if (p->mm) + return buffer + sprintf(buffer, "PaX:\t%c%c%c%c%c\n", + p->mm->pax_flags & MF_PAX_PAGEEXEC ? 'P' : 'p', + p->mm->pax_flags & MF_PAX_EMUTRAMP ? 'E' : 'e', + p->mm->pax_flags & MF_PAX_MPROTECT ? 'M' : 'm', + p->mm->pax_flags & MF_PAX_RANDMMAP ? 'R' : 'r', + p->mm->pax_flags & MF_PAX_SEGMEXEC ? 'S' : 's'); + else + return buffer + sprintf(buffer, "PaX:\t-----\n"); +} +#endif + int proc_pid_status(struct task_struct *task, char * buffer) { char * orig = buffer; +#ifdef CONFIG_VSERVER_LEGACY + struct vx_info *vxi; +#endif +#ifdef CONFIG_VSERVER_LEGACYNET + struct nx_info *nxi; +#endif struct mm_struct *mm = get_task_mm(task); buffer = task_name(task, buffer); @@ -308,12 +339,63 @@ int proc_pid_status(struct task_struct * buffer = task_sig(task, buffer); buffer = task_cap(task, buffer); buffer = cpuset_task_status_allowed(task, buffer); + + if (task_vx_flags(task, VXF_INFO_HIDE, 0)) + goto skip; +#ifdef CONFIG_VSERVER_LEGACY + buffer += sprintf (buffer,"s_context: %d\n", vx_task_xid(task)); + vxi = task_get_vx_info(task); + if (vxi) { + buffer += sprintf (buffer,"ctxflags: %08llx\n" + ,(unsigned long long)vxi->vx_flags); + buffer += sprintf (buffer,"initpid: %d\n" + ,vxi->vx_initpid); + } else { + buffer += sprintf (buffer,"ctxflags: none\n"); + buffer += sprintf (buffer,"initpid: none\n"); + } + put_vx_info(vxi); +#else + buffer += sprintf (buffer,"VxID: %d\n", vx_task_xid(task)); +#endif +#ifdef CONFIG_VSERVER_LEGACYNET + nxi = task_get_nx_info(task); + if (nxi) { + int i; + + buffer += sprintf (buffer,"ipv4root:"); + for (i=0; inbipv4; i++){ + buffer += sprintf (buffer," %08x/%08x" + ,nxi->ipv4[i] + ,nxi->mask[i]); + } + *buffer++ = '\n'; + buffer += sprintf (buffer,"ipv4root_bcast: %08x\n" + ,nxi->v4_bcast); + } else { + buffer += sprintf (buffer,"ipv4root: 0\n"); + buffer += sprintf (buffer,"ipv4root_bcast: 0\n"); + } + put_nx_info(nxi); +#endif +skip: #if defined(CONFIG_S390) buffer = task_show_regs(task, buffer); #endif + +#if defined(CONFIG_PAX_NOEXEC) || defined(CONFIG_PAX_ASLR) + buffer = task_pax(task, buffer); +#endif + return buffer - orig; } +#ifdef CONFIG_GRKERNSEC_PROC_MEMMAP +#define PAX_RAND_FLAGS(_mm) (_mm != NULL && _mm != current->mm && \ + (_mm->pax_flags & MF_PAX_RANDMMAP || \ + _mm->pax_flags & MF_PAX_SEGMEXEC)) +#endif + static int do_task_stat(struct task_struct *task, char * buffer, int whole) { unsigned long vsize, eip, esp, wchan = ~0UL; @@ -322,7 +404,7 @@ static int do_task_stat(struct task_stru sigset_t sigign, sigcatch; char state; int res; - pid_t ppid, pgid = -1, sid = -1; + pid_t pid, ppid, pgid = -1, sid = -1; int num_threads = 0; struct mm_struct *mm; unsigned long long start_time; @@ -386,7 +468,11 @@ static int do_task_stat(struct task_stru stime = cputime_add(stime, task->signal->stime); } } - ppid = pid_alive(task) ? task->group_leader->real_parent->tgid : 0; + pid = vx_info_map_pid(task->vx_info, pid_alive(task) ? task->pid : 0); + ppid = (!(pid > 1)) ? 0 : vx_info_map_tgid(task->vx_info, + task->group_leader->real_parent->tgid); + pgid = vx_info_map_pid(task->vx_info, pgid); + read_unlock(&tasklist_lock); if (!whole || num_threads<2) @@ -398,6 +484,19 @@ static int do_task_stat(struct task_stru stime = task->stime; } +#ifdef CONFIG_GRKERNSEC_PROC_MEMMAP + if (PAX_RAND_FLAGS(mm)) { + eip = 0; + esp = 0; + wchan = 0; + } +#endif +#ifdef CONFIG_GRKERNSEC_HIDESYM + wchan = 0; + eip =0; + esp =0; +#endif + /* scale priority and nice values from timeslices to -20..20 */ /* to make it look like a "normal" Unix priority/nice value */ priority = task_prio(task); @@ -410,10 +509,21 @@ static int do_task_stat(struct task_stru /* convert nsec -> ticks */ start_time = nsec_to_clock_t(start_time); + /* fixup start time for virt uptime */ + if (vx_flags(VXF_VIRT_UPTIME, 0)) { + unsigned long long bias = + current->vx_info->cvirt.bias_clock; + + if (start_time > bias) + start_time -= bias; + else + start_time = 0; + } + res = sprintf(buffer,"%d (%s) %c %d %d %d %d %d %lu %lu \ %lu %lu %lu %lu %lu %ld %ld %ld %ld %d 0 %llu %lu %ld %lu %lu %lu %lu %lu \ %lu %lu %lu %lu %lu %lu %lu %lu %d %d %lu %lu\n", - task->pid, + pid, tcomm, state, ppid, @@ -437,9 +547,15 @@ static int do_task_stat(struct task_stru vsize, mm ? get_mm_rss(mm) : 0, rsslim, +#ifdef CONFIG_GRKERNSEC_PROC_MEMMAP + PAX_RAND_FLAGS(mm) ? 1 : (mm ? mm->start_code : 0), + PAX_RAND_FLAGS(mm) ? 1 : (mm ? mm->end_code : 0), + PAX_RAND_FLAGS(mm) ? 0 : (mm ? mm->start_stack : 0), +#else mm ? mm->start_code : 0, mm ? mm->end_code : 0, mm ? mm->start_stack : 0, +#endif esp, eip, /* The signal information here is obsolete. @@ -485,3 +601,14 @@ int proc_pid_statm(struct task_struct *t return sprintf(buffer,"%d %d %d %d %d %d %d\n", size, resident, shared, text, lib, data, 0); } + +#ifdef CONFIG_GRKERNSEC_PROC_IPADDR +int proc_pid_ipaddr(struct task_struct *task, char * buffer) +{ + int len; + + len = sprintf(buffer, "%u.%u.%u.%u\n", NIPQUAD(task->signal->curr_ip)); + return len; +} +#endif + diff -NurpP --minimal linux-2.6.17.13/fs/proc/base.c linux-2.6.17.13-g-v/fs/proc/base.c --- linux-2.6.17.13/fs/proc/base.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/proc/base.c 2006-09-20 15:00:12.000000000 +0200 @@ -72,6 +72,9 @@ #include #include #include +#include +#include +#include #include "internal.h" /* @@ -122,9 +125,14 @@ enum pid_directory_inos { PROC_TGID_ATTR_EXEC, PROC_TGID_ATTR_FSCREATE, #endif + PROC_TGID_VX_INFO, + PROC_TGID_IP_INFO, #ifdef CONFIG_AUDITSYSCALL PROC_TGID_LOGINUID, #endif +#ifdef CONFIG_GRKERNSEC_PROC_IPADDR + PROC_TGID_IPADDR, +#endif PROC_TGID_OOM_SCORE, PROC_TGID_OOM_ADJUST, PROC_TID_INO, @@ -163,6 +171,8 @@ enum pid_directory_inos { PROC_TID_ATTR_EXEC, PROC_TID_ATTR_FSCREATE, #endif + PROC_TID_VX_INFO, + PROC_TID_IP_INFO, #ifdef CONFIG_AUDITSYSCALL PROC_TID_LOGINUID, #endif @@ -204,6 +214,9 @@ static struct pid_entry tgid_base_stuff[ E(PROC_TGID_EXE, "exe", S_IFLNK|S_IRWXUGO), E(PROC_TGID_MOUNTS, "mounts", S_IFREG|S_IRUGO), E(PROC_TGID_MOUNTSTATS, "mountstats", S_IFREG|S_IRUSR), +#ifdef CONFIG_GRKERNSEC_PROC_IPADDR + E(PROC_TGID_IPADDR, "ipaddr", S_IFREG|S_IRUSR), +#endif #ifdef CONFIG_MMU E(PROC_TGID_SMAPS, "smaps", S_IFREG|S_IRUGO), #endif @@ -219,6 +232,8 @@ static struct pid_entry tgid_base_stuff[ #ifdef CONFIG_CPUSETS E(PROC_TGID_CPUSET, "cpuset", S_IFREG|S_IRUGO), #endif + E(PROC_TGID_VX_INFO, "vinfo", S_IFREG|S_IRUGO), + E(PROC_TGID_IP_INFO, "ninfo", S_IFREG|S_IRUGO), E(PROC_TGID_OOM_SCORE, "oom_score",S_IFREG|S_IRUGO), E(PROC_TGID_OOM_ADJUST,"oom_adj", S_IFREG|S_IRUGO|S_IWUSR), #ifdef CONFIG_AUDITSYSCALL @@ -261,6 +276,8 @@ static struct pid_entry tid_base_stuff[] #ifdef CONFIG_CPUSETS E(PROC_TID_CPUSET, "cpuset", S_IFREG|S_IRUGO), #endif + E(PROC_TID_VX_INFO, "vinfo", S_IFREG|S_IRUGO), + E(PROC_TID_IP_INFO, "ninfo", S_IFREG|S_IRUGO), E(PROC_TID_OOM_SCORE, "oom_score",S_IFREG|S_IRUGO), E(PROC_TID_OOM_ADJUST, "oom_adj", S_IFREG|S_IRUGO|S_IWUSR), #ifdef CONFIG_AUDITSYSCALL @@ -410,7 +427,7 @@ static int proc_task_root_link(struct in (task->parent == current && \ (task->ptrace & PT_PTRACED) && \ (task->state == TASK_STOPPED || task->state == TASK_TRACED) && \ - security_ptrace(current,task) == 0)) + security_ptrace(current,task) == 0 && !gr_handle_proc_ptrace(task))) static int proc_pid_environ(struct task_struct *task, char * buffer) { @@ -547,6 +564,10 @@ static int proc_check_chroot(struct dent struct vfsmount *our_vfsmnt, *mnt; int res = 0; + /* context admin override */ + if (capable(CAP_CONTEXT)) + goto override; + read_lock(&current->fs->lock); our_vfsmnt = mntget(current->fs->rootmnt); base = dget(current->fs->root); @@ -570,6 +591,7 @@ static int proc_check_chroot(struct dent exit: dput(base); mntput(our_vfsmnt); +override: dput(root); mntput(vfsmnt); return res; @@ -591,9 +613,25 @@ static int proc_check_root(struct inode static int proc_permission(struct inode *inode, int mask, struct nameidata *nd) { + int ret = -EACCES; + struct task_struct *task; + if (generic_permission(inode, mask, NULL) != 0) - return -EACCES; - return proc_check_root(inode); + goto out; + + ret = proc_check_root(inode); + if (ret) + goto out; + + task = proc_task(inode); + + if (!task) + goto out; + + ret = gr_acl_handle_procpidmem(task); + +out: + return ret; } static int proc_setattr(struct dentry *dentry, struct iattr *attr) @@ -1286,7 +1324,7 @@ static int proc_pident_readdir(struct fi struct inode *inode = dentry->d_inode; struct pid_entry *p; ino_t ino; - int ret; + int ret, hide; ret = -ENOENT; if (!pid_alive(proc_task(inode))) @@ -1317,11 +1355,20 @@ static int proc_pident_readdir(struct fi goto out; } p = ents + i; + hide = vx_flags(VXF_INFO_HIDE, 0); while (p->name) { + if (hide) { + switch (p->type) { + case PROC_TGID_VX_INFO: + case PROC_TGID_IP_INFO: + goto skip; + } + } if (filldir(dirent, p->name, p->len, filp->f_pos, fake_ino(pid, p->type), p->mode >> 12) < 0) goto out; filp->f_pos++; + skip: p++; } } @@ -1396,6 +1443,10 @@ static struct inode *proc_pid_make_inode inode->i_uid = task->euid; inode->i_gid = task->egid; } +#ifdef CONFIG_GRKERNSEC_PROC_USERGROUP + inode->i_gid = CONFIG_GRKERNSEC_PROC_GID; +#endif + inode->i_xid = vx_task_xid(task); security_task_to_inode(task, inode); out: @@ -1421,10 +1472,17 @@ static int pid_revalidate(struct dentry { struct inode *inode = dentry->d_inode; struct task_struct *task = proc_task(inode); + + if (!vx_check(vx_task_xid(task), VX_IDENT)) + goto out_drop; + /* discard wrong fakeinit */ + if (pid_alive(task)) { if (proc_type(inode) == PROC_TGID_INO || proc_type(inode) == PROC_TID_INO || task_dumpable(task)) { inode->i_uid = task->euid; +#ifndef CONFIG_GRKERNSEC_PROC_USERGROUP inode->i_gid = task->egid; +#endif } else { inode->i_uid = 0; inode->i_gid = 0; @@ -1433,6 +1491,7 @@ static int pid_revalidate(struct dentry security_task_to_inode(task, inode); return 1; } +out_drop: d_drop(dentry); return 0; } @@ -1675,6 +1734,9 @@ static struct file_operations proc_tgid_ static struct inode_operations proc_tgid_attr_inode_operations; #endif +extern int proc_pid_vx_info(struct task_struct *, char *); +extern int proc_pid_nx_info(struct task_struct *, char *); + static int get_tid_list(int index, unsigned int *tids, struct inode *dir); /* SMP-safe */ @@ -1756,6 +1818,12 @@ static struct dentry *proc_pident_lookup inode->i_fop = &proc_info_file_operations; ei->op.proc_read = proc_pid_status; break; +#ifdef CONFIG_GRKERNSEC_PROC_IPADDR + case PROC_TGID_IPADDR: + inode->i_fop = &proc_info_file_operations; + ei->op.proc_read = proc_pid_ipaddr; + break; +#endif case PROC_TID_STAT: inode->i_fop = &proc_info_file_operations; ei->op.proc_read = proc_tid_stat; @@ -1866,15 +1934,33 @@ static struct dentry *proc_pident_lookup inode->i_fop = &proc_loginuid_operations; break; #endif + case PROC_TID_VX_INFO: + case PROC_TGID_VX_INFO: + if (task_vx_flags(task, VXF_INFO_HIDE, 0)) + goto out_noent; + inode->i_fop = &proc_info_file_operations; + ei->op.proc_read = proc_pid_vx_info; + break; + case PROC_TID_IP_INFO: + case PROC_TGID_IP_INFO: + if (task_vx_flags(task, VXF_INFO_HIDE, 0)) + goto out_noent; + inode->i_fop = &proc_info_file_operations; + ei->op.proc_read = proc_pid_nx_info; + break; default: printk("procfs: impossible type (%d)",p->type); - iput(inode); - return ERR_PTR(-EINVAL); + error = -EINVAL; + goto out_put; } dentry->d_op = &pid_dentry_operations; d_add(dentry, inode); return NULL; +out_noent: + error=-ENOENT; +out_put: + iput(inode); out: return ERR_PTR(error); } @@ -1962,14 +2048,14 @@ static int proc_self_readlink(struct den int buflen) { char tmp[30]; - sprintf(tmp, "%d", current->tgid); + sprintf(tmp, "%d", vx_map_tgid(current->tgid)); return vfs_readlink(dentry,buffer,buflen,tmp); } static void *proc_self_follow_link(struct dentry *dentry, struct nameidata *nd) { char tmp[30]; - sprintf(tmp, "%d", current->tgid); + sprintf(tmp, "%d", vx_map_tgid(current->tgid)); return ERR_PTR(vfs_follow_link(nd,tmp)); } @@ -2033,6 +2119,20 @@ void proc_pid_flush(struct dentry *proc_ } } +#define VXF_FAKE_INIT (VXF_INFO_INIT|VXF_STATE_INIT) + +static inline int proc_pid_visible(struct task_struct *task, int pid) +{ + if ((pid == 1) && + !vx_flags(VXF_FAKE_INIT, VXF_FAKE_INIT)) + goto visible; + if (vx_check(vx_task_xid(task), VX_WATCH|VX_IDENT)) + goto visible; + return 0; +visible: + return 1; +} + /* SMP-safe */ struct dentry *proc_pid_lookup(struct inode *dir, struct dentry * dentry, struct nameidata *nd) { @@ -2069,14 +2169,38 @@ struct dentry *proc_pid_lookup(struct in if (!task) goto out; - inode = proc_pid_make_inode(dir->i_sb, task, PROC_TGID_INO); - + if (gr_check_hidden_task(task)) { + put_task_struct(task); + goto out; + } - if (!inode) { +#if defined(CONFIG_GRKERNSEC_PROC_USER) || defined(CONFIG_GRKERNSEC_PROC_USERGROUP) + if (current->uid && (task->uid != current->uid) +#ifdef CONFIG_GRKERNSEC_PROC_USERGROUP + && !in_group_p(CONFIG_GRKERNSEC_PROC_GID) +#endif + ) { put_task_struct(task); goto out; } +#endif + + /* check for context visibility */ + if (!proc_pid_visible(task, tgid)) + goto out_drop_task; + + inode = proc_pid_make_inode(dir->i_sb, task, PROC_TGID_INO); + if (!inode) + goto out_drop_task; + +#ifdef CONFIG_GRKERNSEC_PROC_USER + inode->i_mode = S_IFDIR|S_IRUSR|S_IXUSR; +#elif defined(CONFIG_GRKERNSEC_PROC_USERGROUP) + inode->i_mode = S_IFDIR|S_IRUSR|S_IXUSR|S_IRGRP|S_IXGRP; + inode->i_gid = CONFIG_GRKERNSEC_PROC_GID; +#else inode->i_mode = S_IFDIR|S_IRUGO|S_IXUGO; +#endif inode->i_op = &proc_tgid_base_inode_operations; inode->i_fop = &proc_tgid_base_operations; inode->i_flags|=S_IMMUTABLE; @@ -2104,6 +2228,8 @@ struct dentry *proc_pid_lookup(struct in goto out; } return NULL; +out_drop_task: + put_task_struct(task); out: return ERR_PTR(-ENOENT); } @@ -2119,6 +2245,8 @@ static struct dentry *proc_task_lookup(s tid = name_to_int(dentry); if (tid == ~0U) goto out; + if (vx_current_initpid(tid)) + goto out; read_lock(&tasklist_lock); task = find_task_by_pid(tid); @@ -2130,11 +2258,14 @@ static struct dentry *proc_task_lookup(s if (leader->tgid != task->tgid) goto out_drop_task; - inode = proc_pid_make_inode(dir->i_sb, task, PROC_TID_INO); - + /* check for context visibility */ + if (!proc_pid_visible(task, tid)) + goto out_drop_task; + inode = proc_pid_make_inode(dir->i_sb, task, PROC_TID_INO); if (!inode) goto out_drop_task; + inode->i_mode = S_IFDIR|S_IRUGO|S_IXUGO; inode->i_op = &proc_tid_base_inode_operations; inode->i_fop = &proc_tid_base_operations; @@ -2168,13 +2299,16 @@ out: static int get_tgid_list(int index, unsigned long version, unsigned int *tgids) { struct task_struct *p; +#if defined(CONFIG_GRKERNSEC_PROC_USER) || defined(CONFIG_GRKERNSEC_PROC_USERGROUP) + struct task_struct *tmp = current; +#endif int nr_tgids = 0; index--; read_lock(&tasklist_lock); p = NULL; if (version) { - p = find_task_by_pid(version); + p = find_task_by_real_pid(version); if (p && !thread_group_leader(p)) p = NULL; } @@ -2188,9 +2322,24 @@ static int get_tgid_list(int index, unsi int tgid = p->pid; if (!pid_alive(p)) continue; + if (gr_pid_is_chrooted(p)) + continue; + if (gr_check_hidden_task(p)) + continue; +#if defined(CONFIG_GRKERNSEC_PROC_USER) || defined(CONFIG_GRKERNSEC_PROC_USERGROUP) + if (tmp->uid && (p->uid != tmp->uid) +#ifdef CONFIG_GRKERNSEC_PROC_USERGROUP + && !in_group_p(CONFIG_GRKERNSEC_PROC_GID) +#endif + ) + continue; +#endif + /* check for context visibility */ + if (!proc_pid_visible(p, tgid)) + continue; if (--index >= 0) continue; - tgids[nr_tgids] = tgid; + tgids[nr_tgids] = vx_map_tgid(tgid); nr_tgids++; if (nr_tgids >= PROC_MAXPIDS) break; @@ -2220,10 +2369,13 @@ static int get_tid_list(int index, unsig if (pid_alive(task)) do { int tid = task->pid; + /* check for context visibility */ + if (!proc_pid_visible(task, tid)) + continue; if (--index >= 0) continue; if (tids != NULL) - tids[nr_tids] = tid; + tids[nr_tids] = vx_map_pid(tid); nr_tids++; if (nr_tids >= PROC_MAXPIDS) break; @@ -2299,11 +2451,14 @@ static int proc_task_readdir(struct file unsigned int nr_tids, i; struct dentry *dentry = filp->f_dentry; struct inode *inode = dentry->d_inode; + struct task_struct *task = proc_task(inode); int retval = -ENOENT; ino_t ino; unsigned long pos = filp->f_pos; /* avoiding "long long" filp->f_pos */ - if (!pid_alive(proc_task(inode))) + if (!vx_check(vx_task_xid(task), VX_WATCH|VX_IDENT)) + goto out; + if (!pid_alive(task)) goto out; retval = 0; diff -NurpP --minimal linux-2.6.17.13/fs/proc/generic.c linux-2.6.17.13-g-v/fs/proc/generic.c --- linux-2.6.17.13/fs/proc/generic.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/proc/generic.c 2006-09-20 15:00:12.000000000 +0200 @@ -20,6 +20,7 @@ #include #include #include +#include #include #include "internal.h" @@ -395,12 +396,16 @@ struct dentry *proc_lookup(struct inode for (de = de->subdir; de ; de = de->next) { if (de->namelen != dentry->d_name.len) continue; + if (!vx_hide_check(0, de->vx_flags)) + continue; if (!memcmp(dentry->d_name.name, de->name, de->namelen)) { unsigned int ino = de->low_ino; spin_unlock(&proc_subdir_lock); error = -EINVAL; inode = proc_get_inode(dir->i_sb, ino, de); + /* generic proc entries belong to the host */ + inode->i_xid = 0; spin_lock(&proc_subdir_lock); break; } @@ -476,12 +481,15 @@ int proc_readdir(struct file * filp, } do { + if (!vx_hide_check(0, de->vx_flags)) + goto skip; /* filldir passes info to user space */ spin_unlock(&proc_subdir_lock); if (filldir(dirent, de->name, de->namelen, filp->f_pos, de->low_ino, de->mode >> 12) < 0) goto out; spin_lock(&proc_subdir_lock); + skip: filp->f_pos++; de = de->next; } while (de); @@ -604,6 +612,7 @@ static struct proc_dir_entry *proc_creat ent->namelen = len; ent->mode = mode; ent->nlink = nlink; + ent->vx_flags = IATTR_PROC_DEFAULT; out: return ent; } @@ -624,7 +633,8 @@ struct proc_dir_entry *proc_symlink(cons kfree(ent->data); kfree(ent); ent = NULL; - } + } else + ent->vx_flags = IATTR_PROC_SYMLINK; } else { kfree(ent); ent = NULL; diff -NurpP --minimal linux-2.6.17.13/fs/proc/inode.c linux-2.6.17.13-g-v/fs/proc/inode.c --- linux-2.6.17.13/fs/proc/inode.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/proc/inode.c 2006-09-20 15:00:12.000000000 +0200 @@ -169,8 +169,14 @@ struct inode *proc_get_inode(struct supe if (de->mode) { inode->i_mode = de->mode; inode->i_uid = de->uid; +#ifdef CONFIG_GRKERNSEC_PROC_USERGROUP + inode->i_gid = CONFIG_GRKERNSEC_PROC_GID; +#else inode->i_gid = de->gid; +#endif } + if (de->vx_flags) + PROC_I(inode)->vx_flags = de->vx_flags; if (de->size) inode->i_size = de->size; if (de->nlink) diff -NurpP --minimal linux-2.6.17.13/fs/proc/internal.h linux-2.6.17.13-g-v/fs/proc/internal.h --- linux-2.6.17.13/fs/proc/internal.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/proc/internal.h 2006-09-20 15:00:12.000000000 +0200 @@ -36,6 +36,9 @@ extern int proc_tid_stat(struct task_str extern int proc_tgid_stat(struct task_struct *, char *); extern int proc_pid_status(struct task_struct *, char *); extern int proc_pid_statm(struct task_struct *, char *); +#ifdef CONFIG_GRKERNSEC_PROC_IPADDR +extern int proc_pid_ipaddr(struct task_struct*,char*); +#endif void free_proc_entry(struct proc_dir_entry *de); diff -NurpP --minimal linux-2.6.17.13/fs/proc/proc_misc.c linux-2.6.17.13-g-v/fs/proc/proc_misc.c --- linux-2.6.17.13/fs/proc/proc_misc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/proc/proc_misc.c 2006-09-20 15:00:12.000000000 +0200 @@ -53,6 +53,8 @@ #include #include "internal.h" +#include + #define LOAD_INT(x) ((x) >> FSHIFT) #define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1-1)) * 100) /* @@ -82,17 +84,32 @@ static int proc_calc_metrics(char *page, static int loadavg_read_proc(char *page, char **start, off_t off, int count, int *eof, void *data) { + unsigned int running, threads; int a, b, c; int len; - a = avenrun[0] + (FIXED_1/200); - b = avenrun[1] + (FIXED_1/200); - c = avenrun[2] + (FIXED_1/200); - len = sprintf(page,"%d.%02d %d.%02d %d.%02d %ld/%d %d\n", + if (vx_flags(VXF_VIRT_LOAD, 0)) { + struct vx_info *vxi = current->vx_info; + + a = vxi->cvirt.load[0] + (FIXED_1/200); + b = vxi->cvirt.load[1] + (FIXED_1/200); + c = vxi->cvirt.load[2] + (FIXED_1/200); + + running = atomic_read(&vxi->cvirt.nr_running); + threads = atomic_read(&vxi->cvirt.nr_threads); + } else { + a = avenrun[0] + (FIXED_1/200); + b = avenrun[1] + (FIXED_1/200); + c = avenrun[2] + (FIXED_1/200); + + running = nr_running(); + threads = nr_threads; + } + len = sprintf(page,"%d.%02d %d.%02d %d.%02d %d/%d %d\n", LOAD_INT(a), LOAD_FRAC(a), LOAD_INT(b), LOAD_FRAC(b), LOAD_INT(c), LOAD_FRAC(c), - nr_running(), nr_threads, last_pid); + running, threads, last_pid); return proc_calc_metrics(page, start, off, count, eof, len); } @@ -106,6 +123,9 @@ static int uptime_read_proc(char *page, do_posix_clock_monotonic_gettime(&uptime); cputime_to_timespec(idletime, &idle); + if (vx_flags(VXF_VIRT_UPTIME, 0)) + vx_vsi_uptime(&uptime, &idle); + len = sprintf(page,"%lu.%02lu %lu.%02lu\n", (unsigned long) uptime.tv_sec, (uptime.tv_nsec / (NSEC_PER_SEC / 100)), @@ -143,7 +163,7 @@ static int meminfo_read_proc(char *page, * sysctl_overcommit_ratio / 100) + total_swap_pages; cached = get_page_cache_size() - total_swapcache_pages - i.bufferram; - if (cached < 0) + if (cached < 0 || vx_flags(VXF_VIRT_MEM, 0)) cached = 0; get_vmalloc_info(&vmi); @@ -238,8 +258,9 @@ static int version_read_proc(char *page, { int len; - strcpy(page, linux_banner); - len = strlen(page); + len = sprintf(page, vx_linux_banner, + vx_new_uts(release), + vx_new_uts(version)); return proc_calc_metrics(page, start, off, count, eof, len); } @@ -651,6 +672,8 @@ void create_seq_entry(char *name, mode_t void __init proc_misc_init(void) { struct proc_dir_entry *entry; + int gr_mode = 0; + static struct { char *name; int (*read_proc)(char*,char**,off_t,int,int*,void*); @@ -666,7 +689,9 @@ void __init proc_misc_init(void) {"stram", stram_read_proc}, #endif {"filesystems", filesystems_read_proc}, +#ifndef CONFIG_GRKERNSEC_PROC_ADD {"cmdline", cmdline_read_proc}, +#endif {"locks", locks_read_proc}, {"execdomains", execdomains_read_proc}, {NULL,} @@ -674,19 +699,36 @@ void __init proc_misc_init(void) for (p = simple_ones; p->name; p++) create_proc_read_entry(p->name, 0, NULL, p->read_proc, NULL); +#ifdef CONFIG_GRKERNSEC_PROC_USER + gr_mode = S_IRUSR; +#elif defined(CONFIG_GRKERNSEC_PROC_USERGROUP) + gr_mode = S_IRUSR | S_IRGRP; +#endif +#ifdef CONFIG_GRKERNSEC_PROC_ADD + create_proc_read_entry("cmdline", gr_mode, NULL, &cmdline_read_proc, NULL); +#endif + proc_symlink("mounts", NULL, "self/mounts"); /* And now for trickier ones */ entry = create_proc_entry("kmsg", S_IRUSR, &proc_root); if (entry) entry->proc_fops = &proc_kmsg_operations; +#ifdef CONFIG_GRKERNSEC_PROC_ADD + create_seq_entry("devices", gr_mode, &proc_devinfo_operations); +#else create_seq_entry("devices", 0, &proc_devinfo_operations); +#endif create_seq_entry("cpuinfo", 0, &proc_cpuinfo_operations); create_seq_entry("partitions", 0, &proc_partitions_operations); create_seq_entry("stat", 0, &proc_stat_operations); create_seq_entry("interrupts", 0, &proc_interrupts_operations); #ifdef CONFIG_SLAB +#ifdef CONFIG_GRKERNSEC_PROC_ADD + create_seq_entry("slabinfo",S_IWUSR|gr_mode,&proc_slabinfo_operations); +#else create_seq_entry("slabinfo",S_IWUSR|S_IRUGO,&proc_slabinfo_operations); +#endif #ifdef CONFIG_DEBUG_SLAB_LEAK create_seq_entry("slab_allocators", 0 ,&proc_slabstats_operations); #endif @@ -701,7 +743,7 @@ void __init proc_misc_init(void) #ifdef CONFIG_SCHEDSTATS create_seq_entry("schedstat", 0, &proc_schedstat_operations); #endif -#ifdef CONFIG_PROC_KCORE +#if defined(CONFIG_PROC_KCORE) && !defined(CONFIG_GRKERNSEC_PROC_ADD) proc_root_kcore = create_proc_entry("kcore", S_IRUSR, NULL); if (proc_root_kcore) { proc_root_kcore->proc_fops = &proc_kcore_operations; diff -NurpP --minimal linux-2.6.17.13/fs/proc/root.c linux-2.6.17.13-g-v/fs/proc/root.c --- linux-2.6.17.13/fs/proc/root.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/proc/root.c 2006-09-20 15:00:12.000000000 +0200 @@ -25,6 +25,9 @@ struct proc_dir_entry *proc_net, *proc_n #ifdef CONFIG_SYSCTL struct proc_dir_entry *proc_sys_root; #endif +struct proc_dir_entry *proc_virtual; + +extern void proc_vx_init(void); static struct super_block *proc_get_sb(struct file_system_type *fs_type, int flags, const char *dev_name, void *data) @@ -53,7 +56,13 @@ void __init proc_root_init(void) return; } proc_misc_init(); +#ifdef CONFIG_GRKERNSEC_PROC_USER + proc_net = proc_mkdir_mode("net", S_IRUSR | S_IXUSR, NULL); +#elif defined(CONFIG_GRKERNSEC_PROC_USERGROUP) + proc_net = proc_mkdir_mode("net", S_IRUSR | S_IXUSR | S_IRGRP | S_IXGRP, NULL); +#else proc_net = proc_mkdir("net", NULL); +#endif proc_net_stat = proc_mkdir("net/stat", NULL); #ifdef CONFIG_SYSVIPC @@ -77,7 +86,16 @@ void __init proc_root_init(void) #ifdef CONFIG_PROC_DEVICETREE proc_device_tree_init(); #endif +#ifdef CONFIG_GRKERNSEC_PROC_ADD +#ifdef CONFIG_GRKERNSEC_PROC_USER + proc_bus = proc_mkdir_mode("bus", S_IRUSR | S_IXUSR, NULL); +#elif defined(CONFIG_GRKERNSEC_PROC_USERGROUP) + proc_bus = proc_mkdir_mode("bus", S_IRUSR | S_IXUSR | S_IRGRP | S_IXGRP, NULL); +#endif +#else proc_bus = proc_mkdir("bus", NULL); +#endif + proc_vx_init(); } static int proc_root_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat diff -NurpP --minimal linux-2.6.17.13/fs/proc/task_mmu.c linux-2.6.17.13-g-v/fs/proc/task_mmu.c --- linux-2.6.17.13/fs/proc/task_mmu.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/proc/task_mmu.c 2006-09-20 15:00:12.000000000 +0200 @@ -43,15 +43,27 @@ char *task_mem(struct mm_struct *mm, cha "VmStk:\t%8lu kB\n" "VmExe:\t%8lu kB\n" "VmLib:\t%8lu kB\n" - "VmPTE:\t%8lu kB\n", - hiwater_vm << (PAGE_SHIFT-10), + "VmPTE:\t%8lu kB\n" + +#ifdef CONFIG_ARCH_TRACK_EXEC_LIMIT + "CsBase:\t%8lx\nCsLim:\t%8lx\n" +#endif + + ,hiwater_vm << (PAGE_SHIFT-10), (total_vm - mm->reserved_vm) << (PAGE_SHIFT-10), mm->locked_vm << (PAGE_SHIFT-10), hiwater_rss << (PAGE_SHIFT-10), total_rss << (PAGE_SHIFT-10), data << (PAGE_SHIFT-10), mm->stack_vm << (PAGE_SHIFT-10), text, lib, - (PTRS_PER_PTE*sizeof(pte_t)*mm->nr_ptes) >> 10); + (PTRS_PER_PTE*sizeof(pte_t)*mm->nr_ptes) >> 10 + +#ifdef CONFIG_ARCH_TRACK_EXEC_LIMIT + , mm->context.user_cs_base, mm->context.user_cs_limit +#endif + + ); + return buffer; } @@ -118,6 +130,12 @@ struct mem_size_stats unsigned long private_dirty; }; +#ifdef CONFIG_GRKERNSEC_PROC_MEMMAP +#define PAX_RAND_FLAGS(_mm) (_mm != NULL && _mm != current->mm && \ + (_mm->pax_flags & MF_PAX_RANDMMAP || \ + _mm->pax_flags & MF_PAX_SEGMEXEC)) +#endif + static int show_map_internal(struct seq_file *m, void *v, struct mem_size_stats *mss) { struct task_struct *task = m->private; @@ -136,13 +154,30 @@ static int show_map_internal(struct seq_ } seq_printf(m, "%08lx-%08lx %c%c%c%c %08lx %02x:%02x %lu %n", +#ifdef CONFIG_GRKERNSEC_PROC_MEMMAP + PAX_RAND_FLAGS(mm) ? 0UL : vma->vm_start, + PAX_RAND_FLAGS(mm) ? 0UL : vma->vm_end, +#else vma->vm_start, vma->vm_end, +#endif + +#if 0 + flags & VM_MAYREAD ? flags & VM_READ ? 'R' : '+' : flags & VM_READ ? 'r' : '-', + flags & VM_MAYWRITE ? flags & VM_WRITE ? 'W' : '+' : flags & VM_WRITE ? 'w' : '-', + flags & VM_MAYEXEC ? flags & VM_EXEC ? 'X' : '+' : flags & VM_EXEC ? 'x' : '-', +#else flags & VM_READ ? 'r' : '-', flags & VM_WRITE ? 'w' : '-', flags & VM_EXEC ? 'x' : '-', +#endif + flags & VM_MAYSHARE ? 's' : 'p', +#ifdef CONFIG_GRKERNSEC_PROC_MEMMAP + PAX_RAND_FLAGS(mm) ? 0UL : vma->vm_pgoff << PAGE_SHIFT, +#else vma->vm_pgoff << PAGE_SHIFT, +#endif MAJOR(dev), MINOR(dev), ino, &len); /* @@ -154,13 +189,13 @@ static int show_map_internal(struct seq_ seq_path(m, file->f_vfsmnt, file->f_dentry, "\n"); } else { if (mm) { - if (vma->vm_start <= mm->start_brk && - vma->vm_end >= mm->brk) { + if (vma->vm_start <= mm->brk && vma->vm_end >= mm->start_brk) { pad_len_spaces(m, len); seq_puts(m, "[heap]"); } else { - if (vma->vm_start <= mm->start_stack && - vma->vm_end >= mm->start_stack) { + if ((vma->vm_flags & (VM_GROWSDOWN | VM_GROWSUP)) || + (vma->vm_start <= mm->start_stack && + vma->vm_end >= mm->start_stack)) { pad_len_spaces(m, len); seq_puts(m, "[stack]"); @@ -173,7 +208,25 @@ static int show_map_internal(struct seq_ } seq_putc(m, '\n'); - if (mss) + + if (mss) { +#ifdef CONFIG_GRKERNSEC_PROC_MEMMAP + if (PAX_RAND_FLAGS(mm)) + seq_printf(m, + "Size: %8lu kB\n" + "Rss: %8lu kB\n" + "Shared_Clean: %8lu kB\n" + "Shared_Dirty: %8lu kB\n" + "Private_Clean: %8lu kB\n" + "Private_Dirty: %8lu kB\n", + 0UL, + 0UL, + 0UL, + 0UL, + 0UL, + 0UL); + else +#endif seq_printf(m, "Size: %8lu kB\n" "Rss: %8lu kB\n" @@ -187,6 +240,7 @@ static int show_map_internal(struct seq_ mss->shared_dirty >> 10, mss->private_clean >> 10, mss->private_dirty >> 10); + } if (m->count < m->size) /* vma is copied successfully */ m->version = (vma != get_gate_vma(task))? vma->vm_start: 0; diff -NurpP --minimal linux-2.6.17.13/fs/quota.c linux-2.6.17.13-g-v/fs/quota.c --- linux-2.6.17.13/fs/quota.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/quota.c 2006-09-20 15:00:12.000000000 +0200 @@ -17,6 +17,9 @@ #include #include #include +#include +#include +#include /* Check validity of generic quotactl commands */ static int generic_quotactl_valid(struct super_block *sb, int type, int cmd, qid_t id) @@ -81,11 +84,11 @@ static int generic_quotactl_valid(struct if (cmd == Q_GETQUOTA) { if (((type == USRQUOTA && current->euid != id) || (type == GRPQUOTA && !in_egroup_p(id))) && - !capable(CAP_SYS_ADMIN)) + !vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return -EPERM; } else if (cmd != Q_GETFMT && cmd != Q_SYNC && cmd != Q_GETINFO) - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return -EPERM; return 0; @@ -132,10 +135,10 @@ static int xqm_quotactl_valid(struct sup if (cmd == Q_XGETQUOTA) { if (((type == XQM_USRQUOTA && current->euid != id) || (type == XQM_GRPQUOTA && !in_egroup_p(id))) && - !capable(CAP_SYS_ADMIN)) + !vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return -EPERM; } else if (cmd != Q_XGETQSTAT && cmd != Q_XQUOTASYNC) { - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return -EPERM; } @@ -337,6 +340,43 @@ static int do_quotactl(struct super_bloc return 0; } +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE) + +#include +#include + +static vroot_grb_func *vroot_get_real_bdev = NULL; + +static spinlock_t vroot_grb_lock = SPIN_LOCK_UNLOCKED; + +int register_vroot_grb(vroot_grb_func *func) { + int ret = -EBUSY; + + spin_lock(&vroot_grb_lock); + if (!vroot_get_real_bdev) { + vroot_get_real_bdev = func; + ret = 0; + } + spin_unlock(&vroot_grb_lock); + return ret; +} +EXPORT_SYMBOL(register_vroot_grb); + +int unregister_vroot_grb(vroot_grb_func *func) { + int ret = -EINVAL; + + spin_lock(&vroot_grb_lock); + if (vroot_get_real_bdev) { + vroot_get_real_bdev = NULL; + ret = 0; + } + spin_unlock(&vroot_grb_lock); + return ret; +} +EXPORT_SYMBOL(unregister_vroot_grb); + +#endif + /* * This is the system call interface. This communicates with * the user-level programs. Currently this only supports diskquota @@ -362,6 +402,23 @@ asmlinkage long sys_quotactl(unsigned in putname(tmp); if (IS_ERR(bdev)) return PTR_ERR(bdev); +#if defined(CONFIG_BLK_DEV_VROOT) || defined(CONFIG_BLK_DEV_VROOT_MODULE) + if (bdev && bdev->bd_inode && + imajor(bdev->bd_inode) == VROOT_MAJOR) { + struct block_device *bdnew = (void *)-EINVAL; + + if (vroot_get_real_bdev) + bdnew = vroot_get_real_bdev(bdev); + else + vxdprintk(VXD_CBIT(misc, 0), + "vroot_get_real_bdev not set"); + + bdput(bdev); + if (IS_ERR(bdnew)) + return PTR_ERR(bdnew); + bdev = bdnew; + } +#endif sb = get_super(bdev); bdput(bdev); if (!sb) diff -NurpP --minimal linux-2.6.17.13/fs/readdir.c linux-2.6.17.13-g-v/fs/readdir.c --- linux-2.6.17.13/fs/readdir.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/readdir.c 2006-09-20 15:00:12.000000000 +0200 @@ -16,6 +16,8 @@ #include #include #include +#include +#include #include @@ -65,6 +67,7 @@ struct old_linux_dirent { struct readdir_callback { struct old_linux_dirent __user * dirent; + struct file * file; int result; }; @@ -76,6 +79,10 @@ static int fillonedir(void * __buf, cons if (buf->result) return -EINVAL; + + if (!gr_acl_handle_filldir(buf->file, name, namlen, ino)) + return 0; + buf->result++; dirent = buf->dirent; if (!access_ok(VERIFY_WRITE, dirent, @@ -107,6 +114,7 @@ asmlinkage long old_readdir(unsigned int buf.result = 0; buf.dirent = dirent; + buf.file = file; error = vfs_readdir(file, fillonedir, &buf); if (error >= 0) @@ -133,6 +141,7 @@ struct linux_dirent { struct getdents_callback { struct linux_dirent __user * current_dir; struct linux_dirent __user * previous; + struct file * file; int count; int error; }; @@ -147,6 +156,10 @@ static int filldir(void * __buf, const c buf->error = -EINVAL; /* only used if we fail.. */ if (reclen > buf->count) return -EINVAL; + + if (!gr_acl_handle_filldir(buf->file, name, namlen, ino)) + return 0; + dirent = buf->previous; if (dirent) { if (__put_user(offset, &dirent->d_off)) @@ -191,6 +204,7 @@ asmlinkage long sys_getdents(unsigned in buf.current_dir = dirent; buf.previous = NULL; + buf.file = file; buf.count = count; buf.error = 0; @@ -217,6 +231,7 @@ out: struct getdents_callback64 { struct linux_dirent64 __user * current_dir; struct linux_dirent64 __user * previous; + struct file * file; int count; int error; }; @@ -231,6 +246,10 @@ static int filldir64(void * __buf, const buf->error = -EINVAL; /* only used if we fail.. */ if (reclen > buf->count) return -EINVAL; + + if (!gr_acl_handle_filldir(buf->file, name, namlen, ino)) + return 0; + dirent = buf->previous; if (dirent) { if (__put_user(offset, &dirent->d_off)) @@ -277,6 +296,7 @@ asmlinkage long sys_getdents64(unsigned buf.current_dir = dirent; buf.previous = NULL; + buf.file = file; buf.count = count; buf.error = 0; diff -NurpP --minimal linux-2.6.17.13/fs/reiserfs/bitmap.c linux-2.6.17.13-g-v/fs/reiserfs/bitmap.c --- linux-2.6.17.13/fs/reiserfs/bitmap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/reiserfs/bitmap.c 2006-09-20 15:00:12.000000000 +0200 @@ -13,6 +13,7 @@ #include #include #include +#include #define PREALLOCATION_SIZE 9 @@ -411,8 +412,10 @@ static void _reiserfs_free_block(struct set_sb_free_blocks(rs, sb_free_blocks(rs) + 1); journal_mark_dirty(th, s, sbh); - if (for_unformatted) + if (for_unformatted) { + DLIMIT_FREE_BLOCK(inode, 1); DQUOT_FREE_BLOCK_NODIRTY(inode, 1); + } } void reiserfs_free_block(struct reiserfs_transaction_handle *th, @@ -1021,6 +1024,7 @@ static inline int blocknrs_and_prealloc_ int passno = 0; int nr_allocated = 0; int bigalloc = 0; + int blocks; determine_prealloc_size(hint); if (!hint->formatted_node) { @@ -1030,19 +1034,30 @@ static inline int blocknrs_and_prealloc_ "reiserquota: allocating %d blocks id=%u", amount_needed, hint->inode->i_uid); #endif - quota_ret = - DQUOT_ALLOC_BLOCK_NODIRTY(hint->inode, amount_needed); - if (quota_ret) /* Quota exceeded? */ + quota_ret = DQUOT_ALLOC_BLOCK_NODIRTY(hint->inode, + amount_needed); + if (quota_ret) return QUOTA_EXCEEDED; + if (DLIMIT_ALLOC_BLOCK(hint->inode, amount_needed)) { + DQUOT_FREE_BLOCK_NODIRTY(hint->inode, + amount_needed); + return NO_DISK_SPACE; + } + if (hint->preallocate && hint->prealloc_size) { #ifdef REISERQUOTA_DEBUG reiserfs_debug(s, REISERFS_DEBUG_CODE, "reiserquota: allocating (prealloc) %d blocks id=%u", hint->prealloc_size, hint->inode->i_uid); #endif - quota_ret = - DQUOT_PREALLOC_BLOCK_NODIRTY(hint->inode, - hint->prealloc_size); + quota_ret = DQUOT_PREALLOC_BLOCK_NODIRTY(hint->inode, + hint->prealloc_size); + if (!quota_ret && + DLIMIT_ALLOC_BLOCK(hint->inode, hint->prealloc_size)) { + DQUOT_FREE_BLOCK_NODIRTY(hint->inode, + hint->prealloc_size); + quota_ret = 1; + } if (quota_ret) hint->preallocate = hint->prealloc_size = 0; } @@ -1093,7 +1108,10 @@ static inline int blocknrs_and_prealloc_ nr_allocated, hint->inode->i_uid); #endif - DQUOT_FREE_BLOCK_NODIRTY(hint->inode, amount_needed + hint->prealloc_size - nr_allocated); /* Free not allocated blocks */ + /* Free not allocated blocks */ + blocks = amount_needed + hint->prealloc_size - nr_allocated; + DLIMIT_FREE_BLOCK(hint->inode, blocks); + DQUOT_FREE_BLOCK_NODIRTY(hint->inode, blocks); } while (nr_allocated--) reiserfs_free_block(hint->th, hint->inode, @@ -1125,10 +1143,10 @@ static inline int blocknrs_and_prealloc_ REISERFS_I(hint->inode)->i_prealloc_count, hint->inode->i_uid); #endif - DQUOT_FREE_BLOCK_NODIRTY(hint->inode, amount_needed + - hint->prealloc_size - nr_allocated - - REISERFS_I(hint->inode)-> - i_prealloc_count); + blocks = amount_needed + hint->prealloc_size - nr_allocated - + REISERFS_I(hint->inode)->i_prealloc_count; + DLIMIT_FREE_BLOCK(hint->inode, blocks); + DQUOT_FREE_BLOCK_NODIRTY(hint->inode, blocks); } return CARRY_ON; diff -NurpP --minimal linux-2.6.17.13/fs/reiserfs/file.c linux-2.6.17.13-g-v/fs/reiserfs/file.c --- linux-2.6.17.13/fs/reiserfs/file.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/reiserfs/file.c 2006-09-20 15:00:12.000000000 +0200 @@ -1588,4 +1588,5 @@ struct inode_operations reiserfs_file_in .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .sync_flags = reiserfs_sync_flags, }; diff -NurpP --minimal linux-2.6.17.13/fs/reiserfs/inode.c linux-2.6.17.13-g-v/fs/reiserfs/inode.c --- linux-2.6.17.13/fs/reiserfs/inode.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/reiserfs/inode.c 2006-09-20 15:00:12.000000000 +0200 @@ -17,6 +17,8 @@ #include #include #include +#include +#include extern int reiserfs_default_io_size; /* default io size devuned in super.c */ @@ -57,6 +59,7 @@ void reiserfs_delete_inode(struct inode * stat data deletion */ if (!err) DQUOT_FREE_INODE(inode); + DLIMIT_FREE_INODE(inode); if (journal_end(&th, inode->i_sb, jbegin_count)) { mutex_unlock(&inode->i_mutex); @@ -1125,6 +1128,8 @@ static void init_inode(struct inode *ino struct buffer_head *bh; struct item_head *ih; __u32 rdev; + uid_t uid; + gid_t gid; //int version = ITEM_VERSION_1; bh = PATH_PLAST_BUFFER(path); @@ -1148,12 +1153,13 @@ static void init_inode(struct inode *ino (struct stat_data_v1 *)B_I_PITEM(bh, ih); unsigned long blocks; + uid = sd_v1_uid(sd); + gid = sd_v1_gid(sd); + set_inode_item_key_version(inode, KEY_FORMAT_3_5); set_inode_sd_version(inode, STAT_DATA_V1); inode->i_mode = sd_v1_mode(sd); inode->i_nlink = sd_v1_nlink(sd); - inode->i_uid = sd_v1_uid(sd); - inode->i_gid = sd_v1_gid(sd); inode->i_size = sd_v1_size(sd); inode->i_atime.tv_sec = sd_v1_atime(sd); inode->i_mtime.tv_sec = sd_v1_mtime(sd); @@ -1195,11 +1201,12 @@ static void init_inode(struct inode *ino // (directories and symlinks) struct stat_data *sd = (struct stat_data *)B_I_PITEM(bh, ih); + uid = sd_v2_uid(sd); + gid = sd_v2_gid(sd); + inode->i_mode = sd_v2_mode(sd); inode->i_nlink = sd_v2_nlink(sd); - inode->i_uid = sd_v2_uid(sd); inode->i_size = sd_v2_size(sd); - inode->i_gid = sd_v2_gid(sd); inode->i_mtime.tv_sec = sd_v2_mtime(sd); inode->i_atime.tv_sec = sd_v2_atime(sd); inode->i_ctime.tv_sec = sd_v2_ctime(sd); @@ -1229,6 +1236,10 @@ static void init_inode(struct inode *ino sd_attrs_to_i_attrs(sd_v2_attrs(sd), inode); } + inode->i_uid = INOXID_UID(XID_TAG(inode), uid, gid); + inode->i_gid = INOXID_GID(XID_TAG(inode), uid, gid); + inode->i_xid = INOXID_XID(XID_TAG(inode), uid, gid, 0); + pathrelse(path); if (S_ISREG(inode->i_mode)) { inode->i_op = &reiserfs_file_inode_operations; @@ -1251,13 +1262,15 @@ static void init_inode(struct inode *ino static void inode2sd(void *sd, struct inode *inode, loff_t size) { struct stat_data *sd_v2 = (struct stat_data *)sd; + uid_t uid = XIDINO_UID(XID_TAG(inode), inode->i_uid, inode->i_xid); + gid_t gid = XIDINO_GID(XID_TAG(inode), inode->i_gid, inode->i_xid); __u16 flags; + set_sd_v2_uid(sd_v2, uid); + set_sd_v2_gid(sd_v2, gid); set_sd_v2_mode(sd_v2, inode->i_mode); set_sd_v2_nlink(sd_v2, inode->i_nlink); - set_sd_v2_uid(sd_v2, inode->i_uid); set_sd_v2_size(sd_v2, size); - set_sd_v2_gid(sd_v2, inode->i_gid); set_sd_v2_mtime(sd_v2, inode->i_mtime.tv_sec); set_sd_v2_atime(sd_v2, inode->i_atime.tv_sec); set_sd_v2_ctime(sd_v2, inode->i_ctime.tv_sec); @@ -1788,6 +1801,10 @@ int reiserfs_new_inode(struct reiserfs_t BUG_ON(!th->t_trans_id); + if (DLIMIT_ALLOC_INODE(inode)) { + err = -ENOSPC; + goto out_bad_dlimit; + } if (DQUOT_ALLOC_INODE(inode)) { err = -EDQUOT; goto out_end_trans; @@ -1973,6 +1990,9 @@ int reiserfs_new_inode(struct reiserfs_t DQUOT_FREE_INODE(inode); out_end_trans: + DLIMIT_FREE_INODE(inode); + + out_bad_dlimit: journal_end(th, th->t_super, th->t_blocks_allocated); /* Drop can be outside and it needs more credits so it's better to have it outside */ DQUOT_DROP(inode); @@ -2700,6 +2720,14 @@ void sd_attrs_to_i_attrs(__u16 sd_attrs, inode->i_flags |= S_IMMUTABLE; else inode->i_flags &= ~S_IMMUTABLE; + if (sd_attrs & REISERFS_IUNLINK_FL) + inode->i_flags |= S_IUNLINK; + else + inode->i_flags &= ~S_IUNLINK; + if (sd_attrs & REISERFS_BARRIER_FL) + inode->i_flags |= S_BARRIER; + else + inode->i_flags &= ~S_BARRIER; if (sd_attrs & REISERFS_APPEND_FL) inode->i_flags |= S_APPEND; else @@ -2722,6 +2750,14 @@ void i_attrs_to_sd_attrs(struct inode *i *sd_attrs |= REISERFS_IMMUTABLE_FL; else *sd_attrs &= ~REISERFS_IMMUTABLE_FL; + if (inode->i_flags & S_IUNLINK) + *sd_attrs |= REISERFS_IUNLINK_FL; + else + *sd_attrs &= ~REISERFS_IUNLINK_FL; + if (inode->i_flags & S_BARRIER) + *sd_attrs |= REISERFS_BARRIER_FL; + else + *sd_attrs &= ~REISERFS_BARRIER_FL; if (inode->i_flags & S_SYNC) *sd_attrs |= REISERFS_SYNC_FL; else @@ -2901,6 +2937,22 @@ static ssize_t reiserfs_direct_IO(int rw reiserfs_get_blocks_direct_io, NULL); } +int reiserfs_sync_flags(struct inode *inode) +{ + u16 oldflags, newflags; + + oldflags = REISERFS_I(inode)->i_attrs; + newflags = oldflags; + i_attrs_to_sd_attrs(inode, &newflags); + + if (oldflags ^ newflags) { + REISERFS_I(inode)->i_attrs = newflags; + inode->i_ctime = CURRENT_TIME_SEC; + mark_inode_dirty(inode); + } + return 0; +} + int reiserfs_setattr(struct dentry *dentry, struct iattr *attr) { struct inode *inode = dentry->d_inode; @@ -2945,9 +2997,11 @@ int reiserfs_setattr(struct dentry *dent } error = inode_change_ok(inode, attr); + if (!error) { if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) || - (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) { + (ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid) || + (ia_valid & ATTR_XID && attr->ia_xid != inode->i_xid)) { error = reiserfs_chown_xattrs(inode, attr); if (!error) { @@ -2977,6 +3031,9 @@ int reiserfs_setattr(struct dentry *dent inode->i_uid = attr->ia_uid; if (attr->ia_valid & ATTR_GID) inode->i_gid = attr->ia_gid; + if ((attr->ia_valid & ATTR_XID) && + IS_TAGXID(inode)) + inode->i_xid = attr->ia_xid; mark_inode_dirty(inode); error = journal_end(&th, inode->i_sb, jbegin_count); diff -NurpP --minimal linux-2.6.17.13/fs/reiserfs/ioctl.c linux-2.6.17.13-g-v/fs/reiserfs/ioctl.c --- linux-2.6.17.13/fs/reiserfs/ioctl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/reiserfs/ioctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -4,6 +4,7 @@ #include #include +#include #include #include #include @@ -23,7 +24,7 @@ static int reiserfs_unpack(struct inode int reiserfs_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, unsigned long arg) { - unsigned int flags; + unsigned int flags, oldflags; switch (cmd) { case REISERFS_IOC_UNPACK: @@ -42,12 +43,14 @@ int reiserfs_ioctl(struct inode *inode, flags = REISERFS_I(inode)->i_attrs; i_attrs_to_sd_attrs(inode, (__u16 *) & flags); + flags &= REISERFS_FL_USER_VISIBLE; return put_user(flags, (int __user *)arg); case REISERFS_IOC_SETFLAGS:{ if (!reiserfs_attrs(inode->i_sb)) return -ENOTTY; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if ((current->fsuid != inode->i_uid) @@ -57,10 +60,12 @@ int reiserfs_ioctl(struct inode *inode, if (get_user(flags, (int __user *)arg)) return -EFAULT; - if (((flags ^ REISERFS_I(inode)-> - i_attrs) & (REISERFS_IMMUTABLE_FL | - REISERFS_APPEND_FL)) - && !capable(CAP_LINUX_IMMUTABLE)) + oldflags = REISERFS_I(inode) -> i_attrs; + if (((oldflags & REISERFS_IMMUTABLE_FL) || + ((flags ^ oldflags) & + (REISERFS_IMMUTABLE_FL | REISERFS_IUNLINK_FL | + REISERFS_APPEND_FL))) && + !capable(CAP_LINUX_IMMUTABLE)) return -EPERM; if ((flags & REISERFS_NOTAIL_FL) && @@ -71,6 +76,9 @@ int reiserfs_ioctl(struct inode *inode, if (result) return result; } + + flags = flags & REISERFS_FL_USER_MODIFIABLE; + flags |= oldflags & ~REISERFS_FL_USER_MODIFIABLE; sd_attrs_to_i_attrs(flags, inode); REISERFS_I(inode)->i_attrs = flags; inode->i_ctime = CURRENT_TIME_SEC; @@ -82,7 +90,8 @@ int reiserfs_ioctl(struct inode *inode, case REISERFS_IOC_SETVERSION: if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) return -EPERM; - if (IS_RDONLY(inode)) + if (IS_RDONLY(inode) || + (filp && MNT_IS_RDONLY(filp->f_vfsmnt))) return -EROFS; if (get_user(inode->i_generation, (int __user *)arg)) return -EFAULT; diff -NurpP --minimal linux-2.6.17.13/fs/reiserfs/namei.c linux-2.6.17.13-g-v/fs/reiserfs/namei.c --- linux-2.6.17.13/fs/reiserfs/namei.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/reiserfs/namei.c 2006-09-20 15:00:12.000000000 +0200 @@ -19,6 +19,7 @@ #include #include #include +#include #define INC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) { i->i_nlink++; if (i->i_nlink >= REISERFS_LINK_MAX) i->i_nlink=1; } #define DEC_DIR_INODE_NLINK(i) if (i->i_nlink != 1) i->i_nlink--; @@ -365,6 +366,7 @@ static struct dentry *reiserfs_lookup(st reiserfs_write_unlock(dir->i_sb); return ERR_PTR(-EACCES); } + vx_propagate_xid(nd, inode); /* Propogate the priv_object flag so we know we're in the priv tree */ if (is_reiserfs_priv_object(dir)) @@ -600,6 +602,7 @@ static int new_inode_init(struct inode * } else { inode->i_gid = current->fsgid; } + inode->i_xid = vx_current_fsxid(inode->i_sb); DQUOT_INIT(inode); return 0; } @@ -1546,6 +1549,7 @@ struct inode_operations reiserfs_dir_ino .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .sync_flags = reiserfs_sync_flags, }; /* @@ -1562,6 +1566,7 @@ struct inode_operations reiserfs_symlink .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .sync_flags = reiserfs_sync_flags, }; @@ -1575,5 +1580,6 @@ struct inode_operations reiserfs_special .listxattr = reiserfs_listxattr, .removexattr = reiserfs_removexattr, .permission = reiserfs_permission, + .sync_flags = reiserfs_sync_flags, }; diff -NurpP --minimal linux-2.6.17.13/fs/reiserfs/stree.c linux-2.6.17.13-g-v/fs/reiserfs/stree.c --- linux-2.6.17.13/fs/reiserfs/stree.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/reiserfs/stree.c 2006-09-20 15:00:12.000000000 +0200 @@ -57,6 +57,7 @@ #include #include #include +#include /* Does the buffer contain a disk block which is in the tree. */ inline int B_IS_IN_TREE(const struct buffer_head *p_s_bh) @@ -1298,6 +1299,7 @@ int reiserfs_delete_item(struct reiserfs "reiserquota delete_item(): freeing %u, id=%u type=%c", quota_cut_bytes, p_s_inode->i_uid, head2type(&s_ih)); #endif + DLIMIT_FREE_SPACE(p_s_inode, quota_cut_bytes); DQUOT_FREE_SPACE_NODIRTY(p_s_inode, quota_cut_bytes); /* Return deleted body length */ @@ -1384,6 +1386,7 @@ void reiserfs_delete_solid_item(struct r quota_cut_bytes, inode->i_uid, key2type(key)); #endif + DLIMIT_FREE_SPACE(inode, quota_cut_bytes); DQUOT_FREE_SPACE_NODIRTY(inode, quota_cut_bytes); } @@ -1741,6 +1744,7 @@ int reiserfs_cut_from_item(struct reiser "reiserquota cut_from_item(): freeing %u id=%u type=%c", quota_cut_bytes, p_s_inode->i_uid, '?'); #endif + DLIMIT_FREE_SPACE(p_s_inode, quota_cut_bytes); DQUOT_FREE_SPACE_NODIRTY(p_s_inode, quota_cut_bytes); return n_ret_value; } @@ -1982,6 +1986,11 @@ int reiserfs_paste_into_item(struct reis pathrelse(p_s_search_path); return -EDQUOT; } + if (DLIMIT_ALLOC_SPACE(inode, n_pasted_size)) { + DQUOT_FREE_SPACE_NODIRTY(inode, n_pasted_size); + pathrelse(p_s_search_path); + return -ENOSPC; + } init_tb_struct(th, &s_paste_balance, th->t_super, p_s_search_path, n_pasted_size); #ifdef DISPLACE_NEW_PACKING_LOCALITIES @@ -2034,6 +2043,7 @@ int reiserfs_paste_into_item(struct reis n_pasted_size, inode->i_uid, key2type(&(p_s_key->on_disk_key))); #endif + DLIMIT_FREE_SPACE(inode, n_pasted_size); DQUOT_FREE_SPACE_NODIRTY(inode, n_pasted_size); return retval; } @@ -2071,6 +2081,11 @@ int reiserfs_insert_item(struct reiserfs pathrelse(p_s_path); return -EDQUOT; } + if (DLIMIT_ALLOC_SPACE(inode, quota_bytes)) { + DQUOT_FREE_SPACE_NODIRTY(inode, quota_bytes); + pathrelse(p_s_path); + return -ENOSPC; + } } init_tb_struct(th, &s_ins_balance, th->t_super, p_s_path, IH_SIZE + ih_item_len(p_s_ih)); @@ -2118,7 +2133,9 @@ int reiserfs_insert_item(struct reiserfs "reiserquota insert_item(): freeing %u id=%u type=%c", quota_bytes, inode->i_uid, head2type(p_s_ih)); #endif - if (inode) + if (inode) { + DLIMIT_FREE_SPACE(inode, quota_bytes); DQUOT_FREE_SPACE_NODIRTY(inode, quota_bytes); + } return retval; } diff -NurpP --minimal linux-2.6.17.13/fs/reiserfs/super.c linux-2.6.17.13-g-v/fs/reiserfs/super.c --- linux-2.6.17.13/fs/reiserfs/super.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/reiserfs/super.c 2006-09-20 15:00:12.000000000 +0200 @@ -883,6 +883,9 @@ static int reiserfs_parse_options(struct {"user_xattr",.setmask = 1 << REISERFS_UNSUPPORTED_OPT}, {"nouser_xattr",.clrmask = 1 << REISERFS_UNSUPPORTED_OPT}, #endif +#ifndef CONFIG_INOXID_NONE + {"tagxid",.setmask = 1 << REISERFS_TAGXID}, +#endif #ifdef CONFIG_REISERFS_FS_POSIX_ACL {"acl",.setmask = 1 << REISERFS_POSIXACL}, {"noacl",.clrmask = 1 << REISERFS_POSIXACL}, @@ -1155,6 +1158,12 @@ static int reiserfs_remount(struct super return -EINVAL; } + if ((mount_options & (1 << REISERFS_TAGXID)) && + !(s->s_flags & MS_TAGXID)) { + reiserfs_warning(s, "reiserfs: tagxid not permitted on remount."); + return -EINVAL; + } + handle_attrs(s); /* Add options that are safe here */ @@ -1730,6 +1739,10 @@ static int reiserfs_fill_super(struct su goto error; } + /* map mount option tagxid */ + if (REISERFS_SB(s)->s_mount_opt & (1 << REISERFS_TAGXID)) + s->s_flags |= MS_TAGXID; + rs = SB_DISK_SUPER_BLOCK(s); /* Let's do basic sanity check to verify that underlying device is not smaller than the filesystem. If the check fails then abort and scream, diff -NurpP --minimal linux-2.6.17.13/fs/reiserfs/xattr.c linux-2.6.17.13-g-v/fs/reiserfs/xattr.c --- linux-2.6.17.13/fs/reiserfs/xattr.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/reiserfs/xattr.c 2006-09-20 15:00:12.000000000 +0200 @@ -35,6 +35,7 @@ #include #include #include +#include #include #include #include @@ -824,7 +825,7 @@ int reiserfs_delete_xattrs(struct inode if (dir->d_inode->i_nlink <= 2) { root = get_xa_root(inode->i_sb); reiserfs_write_lock_xattrs(inode->i_sb); - err = vfs_rmdir(root->d_inode, dir); + err = vfs_rmdir(root->d_inode, dir, NULL); reiserfs_write_unlock_xattrs(inode->i_sb); dput(root); } else { diff -NurpP --minimal linux-2.6.17.13/fs/stat.c linux-2.6.17.13-g-v/fs/stat.c --- linux-2.6.17.13/fs/stat.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/stat.c 2006-09-20 15:00:12.000000000 +0200 @@ -27,6 +27,7 @@ void generic_fillattr(struct inode *inod stat->nlink = inode->i_nlink; stat->uid = inode->i_uid; stat->gid = inode->i_gid; + stat->xid = inode->i_xid; stat->rdev = inode->i_rdev; stat->atime = inode->i_atime; stat->mtime = inode->i_mtime; diff -NurpP --minimal linux-2.6.17.13/fs/super.c linux-2.6.17.13-g-v/fs/super.c --- linux-2.6.17.13/fs/super.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/super.c 2006-09-20 15:00:12.000000000 +0200 @@ -38,6 +38,8 @@ #include #include #include +#include +#include #include @@ -803,7 +805,7 @@ struct vfsmount * do_kern_mount(const char *fstype, int flags, const char *name, void *data) { struct file_system_type *type = get_fs_type(fstype); - struct super_block *sb = ERR_PTR(-ENOMEM); + struct super_block *sb; struct vfsmount *mnt; int error; char *secdata = NULL; @@ -811,6 +813,12 @@ do_kern_mount(const char *fstype, int fl if (!type) return ERR_PTR(-ENODEV); + sb = ERR_PTR(-EPERM); + if ((type->fs_flags & FS_BINARY_MOUNTDATA) && + !vx_capable(CAP_SYS_ADMIN, VXC_BINARY_MOUNT)) + goto out; + + sb = ERR_PTR(-ENOMEM); mnt = alloc_vfsmnt(name); if (!mnt) goto out; @@ -832,6 +840,13 @@ do_kern_mount(const char *fstype, int fl sb = type->get_sb(type, flags, name, data); if (IS_ERR(sb)) goto out_free_secdata; + + error = -EPERM; + if (!capable(CAP_SYS_ADMIN) && !sb->s_bdev && + (sb->s_magic != PROC_SUPER_MAGIC) && + (sb->s_magic != DEVPTS_SUPER_MAGIC)) + goto out_sb; + error = security_sb_kern_mount(sb, secdata); if (error) goto out_sb; diff -NurpP --minimal linux-2.6.17.13/fs/sysfs/mount.c linux-2.6.17.13-g-v/fs/sysfs/mount.c --- linux-2.6.17.13/fs/sysfs/mount.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/sysfs/mount.c 2006-09-20 15:00:12.000000000 +0200 @@ -11,8 +11,6 @@ #include "sysfs.h" -/* Random magic number */ -#define SYSFS_MAGIC 0x62656572 struct vfsmount *sysfs_mount; struct super_block * sysfs_sb = NULL; @@ -38,7 +36,7 @@ static int sysfs_fill_super(struct super sb->s_blocksize = PAGE_CACHE_SIZE; sb->s_blocksize_bits = PAGE_CACHE_SHIFT; - sb->s_magic = SYSFS_MAGIC; + sb->s_magic = SYSFS_SUPER_MAGIC; sb->s_op = &sysfs_ops; sb->s_time_gran = 1; sysfs_sb = sb; diff -NurpP --minimal linux-2.6.17.13/fs/xattr.c linux-2.6.17.13-g-v/fs/xattr.c --- linux-2.6.17.13/fs/xattr.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xattr.c 2006-09-20 15:00:12.000000000 +0200 @@ -18,6 +18,7 @@ #include #include #include +#include #include @@ -168,7 +169,7 @@ EXPORT_SYMBOL_GPL(vfs_removexattr); */ static long setxattr(struct dentry *d, char __user *name, void __user *value, - size_t size, int flags) + size_t size, int flags, struct vfsmount *mnt) { int error; void *kvalue = NULL; @@ -195,6 +196,9 @@ setxattr(struct dentry *d, char __user * } } + if (MNT_IS_RDONLY(mnt)) + return -EROFS; + error = vfs_setxattr(d, kname, kvalue, size, flags); kfree(kvalue); return error; @@ -210,7 +214,7 @@ sys_setxattr(char __user *path, char __u error = user_path_walk(path, &nd); if (error) return error; - error = setxattr(nd.dentry, name, value, size, flags); + error = setxattr(nd.dentry, name, value, size, flags, nd.mnt); path_release(&nd); return error; } @@ -225,7 +229,7 @@ sys_lsetxattr(char __user *path, char __ error = user_path_walk_link(path, &nd); if (error) return error; - error = setxattr(nd.dentry, name, value, size, flags); + error = setxattr(nd.dentry, name, value, size, flags, nd.mnt); path_release(&nd); return error; } @@ -243,7 +247,7 @@ sys_fsetxattr(int fd, char __user *name, return error; dentry = f->f_dentry; audit_inode(NULL, dentry->d_inode, 0); - error = setxattr(dentry, name, value, size, flags); + error = setxattr(dentry, name, value, size, flags, f->f_vfsmnt); fput(f); return error; } @@ -416,7 +420,7 @@ sys_flistxattr(int fd, char __user *list * Extended attribute REMOVE operations */ static long -removexattr(struct dentry *d, char __user *name) +removexattr(struct dentry *d, char __user *name, struct vfsmount *mnt) { int error; char kname[XATTR_NAME_MAX + 1]; @@ -427,6 +431,9 @@ removexattr(struct dentry *d, char __use if (error < 0) return error; + if (MNT_IS_RDONLY(mnt)) + return -EROFS; + return vfs_removexattr(d, kname); } @@ -439,7 +446,7 @@ sys_removexattr(char __user *path, char error = user_path_walk(path, &nd); if (error) return error; - error = removexattr(nd.dentry, name); + error = removexattr(nd.dentry, name, nd.mnt); path_release(&nd); return error; } @@ -453,7 +460,7 @@ sys_lremovexattr(char __user *path, char error = user_path_walk_link(path, &nd); if (error) return error; - error = removexattr(nd.dentry, name); + error = removexattr(nd.dentry, name, nd.mnt); path_release(&nd); return error; } @@ -470,7 +477,7 @@ sys_fremovexattr(int fd, char __user *na return error; dentry = f->f_dentry; audit_inode(NULL, dentry->d_inode, 0); - error = removexattr(dentry, name); + error = removexattr(dentry, name, f->f_vfsmnt); fput(f); return error; } diff -NurpP --minimal linux-2.6.17.13/fs/xfs/linux-2.6/xfs_file.c linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_file.c --- linux-2.6.17.13/fs/xfs/linux-2.6/xfs_file.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_file.c 2006-09-20 15:00:12.000000000 +0200 @@ -461,6 +461,11 @@ xfs_file_mmap( vattr_t vattr; int error; +#if defined(CONFIG_PAX_PAGEEXEC) && defined(CONFIG_X86_32) + if ((vma->vm_mm->pax_flags & MF_PAX_PAGEEXEC) && !(vma->vm_flags & VM_EXEC)) + vma->vm_page_prot = __pgprot(pte_val(pte_exprotect(__pte(pgprot_val(vma->vm_page_prot))))); +#endif + vma->vm_ops = &xfs_file_vm_ops; #ifdef CONFIG_XFS_DMAPI diff -NurpP --minimal linux-2.6.17.13/fs/xfs/linux-2.6/xfs_ioctl.c linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_ioctl.c --- linux-2.6.17.13/fs/xfs/linux-2.6/xfs_ioctl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_ioctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -1100,6 +1100,8 @@ xfs_ioc_fsgeometry( #define LINUX_XFLAG_APPEND 0x00000020 /* writes to file may only append */ #define LINUX_XFLAG_NODUMP 0x00000040 /* do not dump file */ #define LINUX_XFLAG_NOATIME 0x00000080 /* do not update atime */ +#define LINUX_XFLAG_BARRIER 0x04000000 /* chroot() barrier */ +#define LINUX_XFLAG_IUNLINK 0x08000000 /* immutable unlink */ STATIC unsigned int xfs_merge_ioc_xflags( @@ -1140,6 +1142,10 @@ xfs_di2lxflags( if (di_flags & XFS_DIFLAG_IMMUTABLE) flags |= LINUX_XFLAG_IMMUTABLE; + if (di_flags & XFS_DIFLAG_IUNLINK) + flags |= LINUX_XFLAG_IUNLINK; + if (di_flags & XFS_DIFLAG_BARRIER) + flags |= LINUX_XFLAG_BARRIER; if (di_flags & XFS_DIFLAG_APPEND) flags |= LINUX_XFLAG_APPEND; if (di_flags & XFS_DIFLAG_SYNC) diff -NurpP --minimal linux-2.6.17.13/fs/xfs/linux-2.6/xfs_iops.c linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_iops.c --- linux-2.6.17.13/fs/xfs/linux-2.6/xfs_iops.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_iops.c 2006-09-20 15:00:12.000000000 +0200 @@ -55,6 +55,7 @@ #include #include #include +#include /* * Get a XFS inode from a given vnode. @@ -409,6 +410,7 @@ xfs_vn_lookup( d_add(dentry, NULL); return NULL; } + vx_propagate_xid(nd, vn_to_inode(cvp)); return d_splice_alias(vn_to_inode(cvp), dentry); } @@ -658,6 +660,10 @@ xfs_vn_setattr( int flags = 0; int error; + error = inode_change_ok(inode, attr); + if (error) + return error; + if (ia_valid & ATTR_UID) { vattr.va_mask |= XFS_AT_UID; vattr.va_uid = attr->ia_uid; @@ -666,6 +672,10 @@ xfs_vn_setattr( vattr.va_mask |= XFS_AT_GID; vattr.va_gid = attr->ia_gid; } + if ((ia_valid & ATTR_XID) && IS_TAGXID(inode)) { + vattr.va_mask |= XFS_AT_XID; + vattr.va_xid = attr->ia_xid; + } if (ia_valid & ATTR_SIZE) { vattr.va_mask |= XFS_AT_SIZE; vattr.va_size = attr->ia_size; @@ -711,6 +721,41 @@ xfs_vn_truncate( } STATIC int +xfs_vn_sync_flags(struct inode *inode) +{ + unsigned int oldflags, newflags; + vattr_t vattr; + int flags = 0; + int error; + vnode_t *vp = vn_from_inode(inode); + + memset(&vattr, 0, sizeof(vattr_t)); + + vattr.va_mask = XFS_AT_XFLAGS; + VOP_GETATTR(vp, &vattr, 0, NULL, error); + if (error) + return error; + oldflags = vattr.va_xflags; + newflags = oldflags & ~(XFS_XFLAG_IMMUTABLE | + XFS_XFLAG_IUNLINK | XFS_XFLAG_BARRIER); + + if (IS_IMMUTABLE(inode)) + newflags |= XFS_XFLAG_IMMUTABLE; + if (IS_IUNLINK(inode)) + newflags |= XFS_XFLAG_IUNLINK; + if (IS_BARRIER(inode)) + newflags |= XFS_XFLAG_BARRIER; + + if (oldflags ^ newflags) { + vattr.va_xflags = newflags; + vattr.va_mask |= XFS_AT_XFLAGS; + VOP_SETATTR(vp, &vattr, flags, NULL, error); + } + vn_revalidate(vp); + return error; +} + +STATIC int xfs_vn_setxattr( struct dentry *dentry, const char *name, @@ -823,6 +868,7 @@ struct inode_operations xfs_inode_operat .getxattr = xfs_vn_getxattr, .listxattr = xfs_vn_listxattr, .removexattr = xfs_vn_removexattr, + .sync_flags = xfs_vn_sync_flags, }; struct inode_operations xfs_dir_inode_operations = { @@ -842,6 +888,7 @@ struct inode_operations xfs_dir_inode_op .getxattr = xfs_vn_getxattr, .listxattr = xfs_vn_listxattr, .removexattr = xfs_vn_removexattr, + .sync_flags = xfs_vn_sync_flags, }; struct inode_operations xfs_symlink_inode_operations = { @@ -855,4 +902,5 @@ struct inode_operations xfs_symlink_inod .getxattr = xfs_vn_getxattr, .listxattr = xfs_vn_listxattr, .removexattr = xfs_vn_removexattr, + .sync_flags = xfs_vn_sync_flags, }; diff -NurpP --minimal linux-2.6.17.13/fs/xfs/linux-2.6/xfs_linux.h linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_linux.h --- linux-2.6.17.13/fs/xfs/linux-2.6/xfs_linux.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_linux.h 2006-09-20 15:00:12.000000000 +0200 @@ -142,6 +142,7 @@ BUFFER_FNS(PrivateStart, unwritten); #define current_pid() (current->pid) #define current_fsuid(cred) (current->fsuid) #define current_fsgid(cred) (current->fsgid) +#define current_fsxid(cred,vp) (vx_current_fsxid(vn_to_inode(vp)->i_sb)) #define NBPP PAGE_SIZE #define DPPSHFT (PAGE_SHIFT - 9) diff -NurpP --minimal linux-2.6.17.13/fs/xfs/linux-2.6/xfs_super.c linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_super.c --- linux-2.6.17.13/fs/xfs/linux-2.6/xfs_super.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_super.c 2006-09-20 15:00:12.000000000 +0200 @@ -160,6 +160,7 @@ xfs_revalidate_inode( inode->i_nlink = ip->i_d.di_nlink; inode->i_uid = ip->i_d.di_uid; inode->i_gid = ip->i_d.di_gid; + inode->i_xid = ip->i_d.di_xid; switch (inode->i_mode & S_IFMT) { case S_IFBLK: @@ -188,6 +189,14 @@ xfs_revalidate_inode( inode->i_flags |= S_IMMUTABLE; else inode->i_flags &= ~S_IMMUTABLE; + if (ip->i_d.di_flags & XFS_DIFLAG_IUNLINK) + inode->i_flags |= S_IUNLINK; + else + inode->i_flags &= ~S_IUNLINK; + if (ip->i_d.di_flags & XFS_DIFLAG_BARRIER) + inode->i_flags |= S_BARRIER; + else + inode->i_flags &= ~S_BARRIER; if (ip->i_d.di_flags & XFS_DIFLAG_APPEND) inode->i_flags |= S_APPEND; else @@ -724,6 +733,12 @@ xfs_fs_remount( int error; VFS_PARSEARGS(vfsp, options, args, 1, error); + if ((args->flags2 & XFSMNT2_TAGXID) && + !(sb->s_flags & MS_TAGXID)) { + printk("XFS: %s: tagxid not permitted on remount.\n", + sb->s_id); + error = EINVAL; + } if (!error) VFS_MNTUPDATE(vfsp, flags, args, error); kmem_free(args, sizeof(*args)); diff -NurpP --minimal linux-2.6.17.13/fs/xfs/linux-2.6/xfs_sysctl.c linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_sysctl.c --- linux-2.6.17.13/fs/xfs/linux-2.6/xfs_sysctl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_sysctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -57,74 +57,74 @@ xfs_stats_clear_proc_handler( STATIC ctl_table xfs_table[] = { {XFS_RESTRICT_CHOWN, "restrict_chown", &xfs_params.restrict_chown.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.restrict_chown.min, &xfs_params.restrict_chown.max}, {XFS_SGID_INHERIT, "irix_sgid_inherit", &xfs_params.sgid_inherit.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.sgid_inherit.min, &xfs_params.sgid_inherit.max}, {XFS_SYMLINK_MODE, "irix_symlink_mode", &xfs_params.symlink_mode.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.symlink_mode.min, &xfs_params.symlink_mode.max}, {XFS_PANIC_MASK, "panic_mask", &xfs_params.panic_mask.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.panic_mask.min, &xfs_params.panic_mask.max}, {XFS_ERRLEVEL, "error_level", &xfs_params.error_level.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.error_level.min, &xfs_params.error_level.max}, {XFS_SYNCD_TIMER, "xfssyncd_centisecs", &xfs_params.syncd_timer.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.syncd_timer.min, &xfs_params.syncd_timer.max}, {XFS_INHERIT_SYNC, "inherit_sync", &xfs_params.inherit_sync.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.inherit_sync.min, &xfs_params.inherit_sync.max}, {XFS_INHERIT_NODUMP, "inherit_nodump", &xfs_params.inherit_nodump.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.inherit_nodump.min, &xfs_params.inherit_nodump.max}, {XFS_INHERIT_NOATIME, "inherit_noatime", &xfs_params.inherit_noatim.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.inherit_noatim.min, &xfs_params.inherit_noatim.max}, {XFS_BUF_TIMER, "xfsbufd_centisecs", &xfs_params.xfs_buf_timer.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.xfs_buf_timer.min, &xfs_params.xfs_buf_timer.max}, {XFS_BUF_AGE, "age_buffer_centisecs", &xfs_params.xfs_buf_age.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.xfs_buf_age.min, &xfs_params.xfs_buf_age.max}, {XFS_INHERIT_NOSYM, "inherit_nosymlinks", &xfs_params.inherit_nosym.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.inherit_nosym.min, &xfs_params.inherit_nosym.max}, {XFS_ROTORSTEP, "rotorstep", &xfs_params.rotorstep.val, sizeof(int), 0644, NULL, &proc_dointvec_minmax, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.rotorstep.min, &xfs_params.rotorstep.max}, /* please keep this the last entry */ #ifdef CONFIG_PROC_FS {XFS_STATS_CLEAR, "stats_clear", &xfs_params.stats_clear.val, sizeof(int), 0644, NULL, &xfs_stats_clear_proc_handler, - &sysctl_intvec, NULL, + NULL, &sysctl_intvec, NULL, &xfs_params.stats_clear.min, &xfs_params.stats_clear.max}, #endif /* CONFIG_PROC_FS */ diff -NurpP --minimal linux-2.6.17.13/fs/xfs/linux-2.6/xfs_vnode.c linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_vnode.c --- linux-2.6.17.13/fs/xfs/linux-2.6/xfs_vnode.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_vnode.c 2006-09-20 15:00:12.000000000 +0200 @@ -103,6 +103,7 @@ vn_revalidate_core( inode->i_nlink = vap->va_nlink; inode->i_uid = vap->va_uid; inode->i_gid = vap->va_gid; + inode->i_xid = vap->va_xid; inode->i_blocks = vap->va_nblocks; inode->i_mtime = vap->va_mtime; inode->i_ctime = vap->va_ctime; @@ -111,6 +112,14 @@ vn_revalidate_core( inode->i_flags |= S_IMMUTABLE; else inode->i_flags &= ~S_IMMUTABLE; + if (vap->va_xflags & XFS_XFLAG_IUNLINK) + inode->i_flags |= S_IUNLINK; + else + inode->i_flags &= ~S_IUNLINK; + if (vap->va_xflags & XFS_XFLAG_BARRIER) + inode->i_flags |= S_BARRIER; + else + inode->i_flags &= ~S_BARRIER; if (vap->va_xflags & XFS_XFLAG_APPEND) inode->i_flags |= S_APPEND; else diff -NurpP --minimal linux-2.6.17.13/fs/xfs/linux-2.6/xfs_vnode.h linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_vnode.h --- linux-2.6.17.13/fs/xfs/linux-2.6/xfs_vnode.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/linux-2.6/xfs_vnode.h 2006-09-20 15:00:12.000000000 +0200 @@ -404,6 +404,7 @@ typedef struct vattr { xfs_nlink_t va_nlink; /* number of references to file */ uid_t va_uid; /* owner user id */ gid_t va_gid; /* owner group id */ + xid_t va_xid; /* owner group id */ xfs_ino_t va_nodeid; /* file id */ xfs_off_t va_size; /* file size in bytes */ u_long va_blocksize; /* blocksize preferred for i/o */ @@ -452,13 +453,15 @@ typedef struct vattr { #define XFS_AT_PROJID 0x04000000 #define XFS_AT_SIZE_NOPERM 0x08000000 #define XFS_AT_GENCOUNT 0x10000000 +#define XFS_AT_XID 0x20000000 #define XFS_AT_ALL (XFS_AT_TYPE|XFS_AT_MODE|XFS_AT_UID|XFS_AT_GID|\ XFS_AT_FSID|XFS_AT_NODEID|XFS_AT_NLINK|XFS_AT_SIZE|\ XFS_AT_ATIME|XFS_AT_MTIME|XFS_AT_CTIME|XFS_AT_RDEV|\ XFS_AT_BLKSIZE|XFS_AT_NBLOCKS|XFS_AT_VCODE|XFS_AT_MAC|\ XFS_AT_ACL|XFS_AT_CAP|XFS_AT_INF|XFS_AT_XFLAGS|XFS_AT_EXTSIZE|\ - XFS_AT_NEXTENTS|XFS_AT_ANEXTENTS|XFS_AT_PROJID|XFS_AT_GENCOUNT) + XFS_AT_NEXTENTS|XFS_AT_ANEXTENTS|XFS_AT_PROJID|XFS_AT_GENCOUNT\ + XFS_AT_XID) #define XFS_AT_STAT (XFS_AT_TYPE|XFS_AT_MODE|XFS_AT_UID|XFS_AT_GID|\ XFS_AT_FSID|XFS_AT_NODEID|XFS_AT_NLINK|XFS_AT_SIZE|\ diff -NurpP --minimal linux-2.6.17.13/fs/xfs/quota/xfs_qm_syscalls.c linux-2.6.17.13-g-v/fs/xfs/quota/xfs_qm_syscalls.c --- linux-2.6.17.13/fs/xfs/quota/xfs_qm_syscalls.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/quota/xfs_qm_syscalls.c 2006-09-20 15:00:12.000000000 +0200 @@ -215,7 +215,7 @@ xfs_qm_scall_quotaoff( xfs_qoff_logitem_t *qoffstart; int nculprits; - if (!force && !capable(CAP_SYS_ADMIN)) + if (!force && !vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return XFS_ERROR(EPERM); /* * No file system can have quotas enabled on disk but not in core. @@ -384,7 +384,7 @@ xfs_qm_scall_trunc_qfiles( int error; xfs_inode_t *qip; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return XFS_ERROR(EPERM); error = 0; if (!XFS_SB_VERSION_HASQUOTA(&mp->m_sb) || flags == 0) { @@ -429,7 +429,7 @@ xfs_qm_scall_quotaon( uint accflags; __int64_t sbflags; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return XFS_ERROR(EPERM); flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD); @@ -600,7 +600,7 @@ xfs_qm_scall_setqlim( int error; xfs_qcnt_t hard, soft; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_QUOTA_CTL)) return XFS_ERROR(EPERM); if ((newlim->d_fieldmask & diff -NurpP --minimal linux-2.6.17.13/fs/xfs/xfs_clnt.h linux-2.6.17.13-g-v/fs/xfs/xfs_clnt.h --- linux-2.6.17.13/fs/xfs/xfs_clnt.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/xfs_clnt.h 2006-09-20 15:00:12.000000000 +0200 @@ -99,5 +99,7 @@ struct xfs_mount_args { */ #define XFSMNT2_COMPAT_IOSIZE 0x00000001 /* don't report large preferred * I/O size in stat(2) */ +#define XFSMNT2_TAGXID 0x80000000 /* context xid tagging */ + #endif /* __XFS_CLNT_H__ */ diff -NurpP --minimal linux-2.6.17.13/fs/xfs/xfs_dinode.h linux-2.6.17.13-g-v/fs/xfs/xfs_dinode.h --- linux-2.6.17.13/fs/xfs/xfs_dinode.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/xfs_dinode.h 2006-09-20 15:00:12.000000000 +0200 @@ -53,7 +53,8 @@ typedef struct xfs_dinode_core __uint32_t di_gid; /* owner's group id */ __uint32_t di_nlink; /* number of links to file */ __uint16_t di_projid; /* owner's project id */ - __uint8_t di_pad[8]; /* unused, zeroed space */ + __uint16_t di_xid; /* vserver context id */ + __uint8_t di_pad[6]; /* unused, zeroed space */ __uint16_t di_flushiter; /* incremented on flush */ xfs_timestamp_t di_atime; /* time last accessed */ xfs_timestamp_t di_mtime; /* time last modified */ @@ -257,6 +258,9 @@ typedef enum xfs_dinode_fmt #define XFS_DIFLAG_NOSYMLINKS_BIT 10 /* disallow symlink creation */ #define XFS_DIFLAG_EXTSIZE_BIT 11 /* inode extent size allocator hint */ #define XFS_DIFLAG_EXTSZINHERIT_BIT 12 /* inherit inode extent size */ +#define XFS_DIFLAG_BARRIER_BIT 13 /* chroot() barrier */ +#define XFS_DIFLAG_IUNLINK_BIT 14 /* immutable unlink */ + #define XFS_DIFLAG_REALTIME (1 << XFS_DIFLAG_REALTIME_BIT) #define XFS_DIFLAG_PREALLOC (1 << XFS_DIFLAG_PREALLOC_BIT) #define XFS_DIFLAG_NEWRTBM (1 << XFS_DIFLAG_NEWRTBM_BIT) @@ -270,12 +274,14 @@ typedef enum xfs_dinode_fmt #define XFS_DIFLAG_NOSYMLINKS (1 << XFS_DIFLAG_NOSYMLINKS_BIT) #define XFS_DIFLAG_EXTSIZE (1 << XFS_DIFLAG_EXTSIZE_BIT) #define XFS_DIFLAG_EXTSZINHERIT (1 << XFS_DIFLAG_EXTSZINHERIT_BIT) +#define XFS_DIFLAG_BARRIER (1 << XFS_DIFLAG_BARRIER_BIT) +#define XFS_DIFLAG_IUNLINK (1 << XFS_DIFLAG_IUNLINK_BIT) #define XFS_DIFLAG_ANY \ (XFS_DIFLAG_REALTIME | XFS_DIFLAG_PREALLOC | XFS_DIFLAG_NEWRTBM | \ XFS_DIFLAG_IMMUTABLE | XFS_DIFLAG_APPEND | XFS_DIFLAG_SYNC | \ XFS_DIFLAG_NOATIME | XFS_DIFLAG_NODUMP | XFS_DIFLAG_RTINHERIT | \ XFS_DIFLAG_PROJINHERIT | XFS_DIFLAG_NOSYMLINKS | XFS_DIFLAG_EXTSIZE | \ - XFS_DIFLAG_EXTSZINHERIT) + XFS_DIFLAG_EXTSZINHERIT | XFS_DIFLAG_BARRIER | XFS_DIFLAG_IUNLINK) #endif /* __XFS_DINODE_H__ */ diff -NurpP --minimal linux-2.6.17.13/fs/xfs/xfs_fs.h linux-2.6.17.13-g-v/fs/xfs/xfs_fs.h --- linux-2.6.17.13/fs/xfs/xfs_fs.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/xfs_fs.h 2006-09-20 15:00:12.000000000 +0200 @@ -67,6 +67,8 @@ struct fsxattr { #define XFS_XFLAG_NOSYMLINKS 0x00000400 /* disallow symlink creation */ #define XFS_XFLAG_EXTSIZE 0x00000800 /* extent size allocator hint */ #define XFS_XFLAG_EXTSZINHERIT 0x00001000 /* inherit inode extent size */ +#define XFS_XFLAG_BARRIER 0x00004000 /* chroot() barrier */ +#define XFS_XFLAG_IUNLINK 0x00008000 /* immutable unlink */ #define XFS_XFLAG_HASATTR 0x80000000 /* no DIFLAG for this */ /* @@ -295,7 +297,8 @@ typedef struct xfs_bstat { __s32 bs_extents; /* number of extents */ __u32 bs_gen; /* generation count */ __u16 bs_projid; /* project id */ - unsigned char bs_pad[14]; /* pad space, unused */ + __u16 bs_xid; /* context id */ + unsigned char bs_pad[12]; /* pad space, unused */ __u32 bs_dmevmask; /* DMIG event mask */ __u16 bs_dmstate; /* DMIG state info */ __u16 bs_aextents; /* attribute number of extents */ diff -NurpP --minimal linux-2.6.17.13/fs/xfs/xfs_inode.c linux-2.6.17.13-g-v/fs/xfs/xfs_inode.c --- linux-2.6.17.13/fs/xfs/xfs_inode.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/xfs_inode.c 2006-09-20 15:00:12.000000000 +0200 @@ -52,6 +52,7 @@ #include "xfs_mac.h" #include "xfs_acl.h" +#include kmem_zone_t *xfs_ifork_zone; kmem_zone_t *xfs_inode_zone; @@ -732,20 +733,35 @@ xfs_xlate_dinode_core( xfs_dinode_core_t *buf_core = (xfs_dinode_core_t *)buf; xfs_dinode_core_t *mem_core = (xfs_dinode_core_t *)dip; xfs_arch_t arch = ARCH_CONVERT; + uint32_t uid = 0, gid = 0; + uint16_t xid = 0; ASSERT(dir); + if (dir < 0) { + xid = mem_core->di_xid; + /* FIXME: supposed to use superblock flag */ + uid = XIDINO_UID(1, mem_core->di_uid, xid); + gid = XIDINO_GID(1, mem_core->di_gid, xid); + xid = XIDINO_XID(1, xid); + } + INT_XLATE(buf_core->di_magic, mem_core->di_magic, dir, arch); INT_XLATE(buf_core->di_mode, mem_core->di_mode, dir, arch); INT_XLATE(buf_core->di_version, mem_core->di_version, dir, arch); INT_XLATE(buf_core->di_format, mem_core->di_format, dir, arch); INT_XLATE(buf_core->di_onlink, mem_core->di_onlink, dir, arch); - INT_XLATE(buf_core->di_uid, mem_core->di_uid, dir, arch); - INT_XLATE(buf_core->di_gid, mem_core->di_gid, dir, arch); + INT_XLATE(buf_core->di_uid, uid, dir, arch); + INT_XLATE(buf_core->di_gid, gid, dir, arch); + INT_XLATE(buf_core->di_xid, xid, dir, arch); INT_XLATE(buf_core->di_nlink, mem_core->di_nlink, dir, arch); INT_XLATE(buf_core->di_projid, mem_core->di_projid, dir, arch); if (dir > 0) { + /* FIXME: supposed to use superblock flag */ + mem_core->di_uid = INOXID_UID(1, uid, gid); + mem_core->di_gid = INOXID_GID(1, uid, gid); + mem_core->di_xid = INOXID_XID(1, uid, gid, xid); memcpy(mem_core->di_pad, buf_core->di_pad, sizeof(buf_core->di_pad)); } else { @@ -794,6 +810,10 @@ _xfs_dic2xflags( flags |= XFS_XFLAG_PREALLOC; if (di_flags & XFS_DIFLAG_IMMUTABLE) flags |= XFS_XFLAG_IMMUTABLE; + if (di_flags & XFS_DIFLAG_IUNLINK) + flags |= XFS_XFLAG_IUNLINK; + if (di_flags & XFS_DIFLAG_BARRIER) + flags |= XFS_XFLAG_BARRIER; if (di_flags & XFS_DIFLAG_APPEND) flags |= XFS_XFLAG_APPEND; if (di_flags & XFS_DIFLAG_SYNC) @@ -1121,6 +1141,7 @@ xfs_ialloc( ASSERT(ip->i_d.di_nlink == nlink); ip->i_d.di_uid = current_fsuid(cr); ip->i_d.di_gid = current_fsgid(cr); + ip->i_d.di_xid = current_fsxid(cr, vp); ip->i_d.di_projid = prid; memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad)); diff -NurpP --minimal linux-2.6.17.13/fs/xfs/xfs_itable.c linux-2.6.17.13-g-v/fs/xfs/xfs_itable.c --- linux-2.6.17.13/fs/xfs/xfs_itable.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/xfs_itable.c 2006-09-20 15:00:12.000000000 +0200 @@ -85,6 +85,7 @@ xfs_bulkstat_one_iget( buf->bs_mode = dic->di_mode; buf->bs_uid = dic->di_uid; buf->bs_gid = dic->di_gid; + buf->bs_xid = dic->di_xid; buf->bs_size = dic->di_size; vn_atime_to_bstime(vp, &buf->bs_atime); buf->bs_mtime.tv_sec = dic->di_mtime.t_sec; @@ -159,6 +160,7 @@ xfs_bulkstat_one_dinode( buf->bs_mode = INT_GET(dic->di_mode, ARCH_CONVERT); buf->bs_uid = INT_GET(dic->di_uid, ARCH_CONVERT); buf->bs_gid = INT_GET(dic->di_gid, ARCH_CONVERT); + buf->bs_xid = INT_GET(dic->di_xid, ARCH_CONVERT); buf->bs_size = INT_GET(dic->di_size, ARCH_CONVERT); buf->bs_atime.tv_sec = INT_GET(dic->di_atime.t_sec, ARCH_CONVERT); buf->bs_atime.tv_nsec = INT_GET(dic->di_atime.t_nsec, ARCH_CONVERT); diff -NurpP --minimal linux-2.6.17.13/fs/xfs/xfs_mount.h linux-2.6.17.13-g-v/fs/xfs/xfs_mount.h --- linux-2.6.17.13/fs/xfs/xfs_mount.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/xfs_mount.h 2006-09-20 15:00:12.000000000 +0200 @@ -445,6 +445,7 @@ typedef struct xfs_mount { #define XFS_MOUNT_NO_PERCPU_SB (1ULL << 23) /* don't use per-cpu superblock counters */ +#define XFS_MOUNT_TAGXID (1ULL << 31) /* context xid tagging */ /* * Default minimum read and write sizes. diff -NurpP --minimal linux-2.6.17.13/fs/xfs/xfs_vfsops.c linux-2.6.17.13-g-v/fs/xfs/xfs_vfsops.c --- linux-2.6.17.13/fs/xfs/xfs_vfsops.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/xfs_vfsops.c 2006-09-20 15:00:12.000000000 +0200 @@ -308,6 +308,8 @@ xfs_start_flags( if (ap->flags2 & XFSMNT2_COMPAT_IOSIZE) mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE; + if (ap->flags2 & XFSMNT2_TAGXID) + mp->m_flags |= XFS_MOUNT_TAGXID; /* * no recovery flag requires a read-only mount @@ -402,6 +404,8 @@ xfs_finish_flags( return XFS_ERROR(EINVAL); } + if (ap->flags2 & XFSMNT2_TAGXID) + vfs->vfs_super->s_flags |= MS_TAGXID; return 0; } @@ -1655,6 +1659,7 @@ xfs_vget( * in stat(). */ #define MNTOPT_ATTR2 "attr2" /* do use attr2 attribute format */ #define MNTOPT_NOATTR2 "noattr2" /* do not use attr2 attribute format */ +#define MNTOPT_TAGXID "tagxid" /* context xid tagging for inodes */ STATIC unsigned long suffix_strtoul(const char *cp, char **endp, unsigned int base) @@ -1829,6 +1834,10 @@ xfs_parseargs( args->flags |= XFSMNT_ATTR2; } else if (!strcmp(this_char, MNTOPT_NOATTR2)) { args->flags &= ~XFSMNT_ATTR2; +#ifndef CONFIG_INOXID_NONE + } else if (!strcmp(this_char, MNTOPT_TAGXID)) { + args->flags2 |= XFSMNT2_TAGXID; +#endif } else if (!strcmp(this_char, "osyncisdsync")) { /* no-op, this is now the default */ printk("XFS: osyncisdsync is now the default, option is deprecated.\n"); diff -NurpP --minimal linux-2.6.17.13/fs/xfs/xfs_vnodeops.c linux-2.6.17.13-g-v/fs/xfs/xfs_vnodeops.c --- linux-2.6.17.13/fs/xfs/xfs_vnodeops.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/fs/xfs/xfs_vnodeops.c 2006-09-20 15:00:12.000000000 +0200 @@ -154,6 +154,7 @@ xfs_getattr( vap->va_mode = ip->i_d.di_mode; vap->va_uid = ip->i_d.di_uid; vap->va_gid = ip->i_d.di_gid; + vap->va_xid = ip->i_d.di_xid; vap->va_projid = ip->i_d.di_projid; /* @@ -254,6 +255,7 @@ xfs_setattr( uint commit_flags=0; uid_t uid=0, iuid=0; gid_t gid=0, igid=0; + xid_t xid=0, ixid=0; int timeflags = 0; vnode_t *vp; xfs_prid_t projid=0, iprojid=0; @@ -310,6 +312,7 @@ xfs_setattr( (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID))) { uint qflags = 0; + /* FIXME: handle xid? */ if ((mask & XFS_AT_UID) && XFS_IS_UQUOTA_ON(mp)) { uid = vap->va_uid; qflags |= XFS_QMOPT_UQUOTA; @@ -390,6 +393,8 @@ xfs_setattr( if (mask & (XFS_AT_MODE|XFS_AT_XFLAGS|XFS_AT_EXTSIZE|XFS_AT_UID| XFS_AT_GID|XFS_AT_PROJID)) { + /* FIXME: handle xid? */ + /* * CAP_FOWNER overrides the following restrictions: * @@ -438,7 +443,7 @@ xfs_setattr( * and can change the group id only to a group of which he * or she is a member. */ - if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID)) { + if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_XID|XFS_AT_PROJID)) { /* * These IDs could have changed since we last looked at them. * But, we're assured that if the ownership did change @@ -446,10 +451,12 @@ xfs_setattr( * would have changed also. */ iuid = ip->i_d.di_uid; - iprojid = ip->i_d.di_projid; igid = ip->i_d.di_gid; - gid = (mask & XFS_AT_GID) ? vap->va_gid : igid; + ixid = ip->i_d.di_xid; + iprojid = ip->i_d.di_projid; uid = (mask & XFS_AT_UID) ? vap->va_uid : iuid; + gid = (mask & XFS_AT_GID) ? vap->va_gid : igid; + xid = (mask & XFS_AT_XID) ? vap->va_xid : ixid; projid = (mask & XFS_AT_PROJID) ? (xfs_prid_t)vap->va_projid : iprojid; @@ -477,6 +484,7 @@ xfs_setattr( if ((XFS_IS_UQUOTA_ON(mp) && iuid != uid) || (XFS_IS_PQUOTA_ON(mp) && iprojid != projid) || (XFS_IS_GQUOTA_ON(mp) && igid != gid)) { + /* FIXME: handle xid? */ ASSERT(tp); code = XFS_QM_DQVOPCHOWNRESV(mp, tp, ip, udqp, gdqp, capable(CAP_FOWNER) ? @@ -694,7 +702,7 @@ xfs_setattr( * and can change the group id only to a group of which he * or she is a member. */ - if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID)) { + if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_XID|XFS_AT_PROJID)) { /* * CAP_FSETID overrides the following restrictions: * @@ -710,6 +718,12 @@ xfs_setattr( * Change the ownerships and register quota modifications * in the transaction. */ + if (ixid != xid) { + if (XFS_IS_GQUOTA_ON(mp)) { + /* FIXME: handle xid quota? */ + } + ip->i_d.di_xid = xid; + } if (iuid != uid) { if (XFS_IS_UQUOTA_ON(mp)) { ASSERT(mask & XFS_AT_UID); @@ -790,6 +804,10 @@ xfs_setattr( di_flags = (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC); if (vap->va_xflags & XFS_XFLAG_IMMUTABLE) di_flags |= XFS_DIFLAG_IMMUTABLE; + if (vap->va_xflags & XFS_XFLAG_IUNLINK) + di_flags |= XFS_DIFLAG_IUNLINK; + if (vap->va_xflags & XFS_XFLAG_BARRIER) + di_flags |= XFS_DIFLAG_BARRIER; if (vap->va_xflags & XFS_XFLAG_APPEND) di_flags |= XFS_DIFLAG_APPEND; if (vap->va_xflags & XFS_XFLAG_SYNC) diff -NurpP --minimal linux-2.6.17.13/grsecurity/Kconfig linux-2.6.17.13-g-v/grsecurity/Kconfig --- linux-2.6.17.13/grsecurity/Kconfig 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,890 @@ +# +# grecurity configuration +# + +menu "Grsecurity" + +config GRKERNSEC + bool "Grsecurity" + select CRYPTO + select CRYPTO_SHA256 + help + If you say Y here, you will be able to configure many features + that will enhance the security of your system. It is highly + recommended that you say Y here and read through the help + for each option so that you fully understand the features and + can evaluate their usefulness for your machine. + +choice + prompt "Security Level" + depends GRKERNSEC + default GRKERNSEC_CUSTOM + +config GRKERNSEC_LOW + bool "Low" + select GRKERNSEC_LINK + select GRKERNSEC_FIFO + select GRKERNSEC_RANDPID + select GRKERNSEC_EXECVE + select GRKERNSEC_RANDNET + select GRKERNSEC_DMESG + select GRKERNSEC_CHROOT_CHDIR + select GRKERNSEC_MODSTOP if (MODULES) + + help + If you choose this option, several of the grsecurity options will + be enabled that will give you greater protection against a number + of attacks, while assuring that none of your software will have any + conflicts with the additional security measures. If you run a lot + of unusual software, or you are having problems with the higher + security levels, you should say Y here. With this option, the + following features are enabled: + + - Linking restrictions + - FIFO restrictions + - Randomized PIDs + - Enforcing RLIMIT_NPROC on execve + - Restricted dmesg + - Enforced chdir("/") on chroot + - Runtime module disabling + +config GRKERNSEC_MEDIUM + bool "Medium" + select PAX + select PAX_EI_PAX + select PAX_PT_PAX_FLAGS + select PAX_HAVE_ACL_FLAGS + select GRKERNSEC_PROC_MEMMAP if (PAX_NOEXEC || PAX_ASLR) + select GRKERNSEC_CHROOT_SYSCTL + select GRKERNSEC_LINK + select GRKERNSEC_FIFO + select GRKERNSEC_RANDPID + select GRKERNSEC_EXECVE + select GRKERNSEC_DMESG + select GRKERNSEC_RANDNET + select GRKERNSEC_FORKFAIL + select GRKERNSEC_TIME + select GRKERNSEC_SIGNAL + select GRKERNSEC_CHROOT + select GRKERNSEC_CHROOT_UNIX + select GRKERNSEC_CHROOT_MOUNT + select GRKERNSEC_CHROOT_PIVOT + select GRKERNSEC_CHROOT_DOUBLE + select GRKERNSEC_CHROOT_CHDIR + select GRKERNSEC_CHROOT_MKNOD + select GRKERNSEC_PROC + select GRKERNSEC_PROC_USERGROUP + select GRKERNSEC_MODSTOP if (MODULES) + select PAX_RANDUSTACK + select PAX_ASLR + select PAX_RANDMMAP + select PAX_NOVSYSCALL if (X86 && !X86_64) + + help + If you say Y here, several features in addition to those included + in the low additional security level will be enabled. These + features provide even more security to your system, though in rare + cases they may be incompatible with very old or poorly written + software. If you enable this option, make sure that your auth + service (identd) is running as gid 1001. With this option, + the following features (in addition to those provided in the + low additional security level) will be enabled: + + - Randomized TCP source ports + - Failed fork logging + - Time change logging + - Signal logging + - Deny mounts in chroot + - Deny double chrooting + - Deny sysctl writes in chroot + - Deny mknod in chroot + - Deny access to abstract AF_UNIX sockets out of chroot + - Deny pivot_root in chroot + - Denied writes of /dev/kmem, /dev/mem, and /dev/port + - /proc restrictions with special GID set to 10 (usually wheel) + - Address Space Layout Randomization (ASLR) + +config GRKERNSEC_HIGH + bool "High" + select GRKERNSEC_LINK + select GRKERNSEC_FIFO + select GRKERNSEC_RANDPID + select GRKERNSEC_EXECVE + select GRKERNSEC_DMESG + select GRKERNSEC_FORKFAIL + select GRKERNSEC_TIME + select GRKERNSEC_SIGNAL + select GRKERNSEC_CHROOT_SHMAT + select GRKERNSEC_CHROOT_UNIX + select GRKERNSEC_CHROOT_MOUNT + select GRKERNSEC_CHROOT_FCHDIR + select GRKERNSEC_CHROOT_PIVOT + select GRKERNSEC_CHROOT_DOUBLE + select GRKERNSEC_CHROOT_CHDIR + select GRKERNSEC_CHROOT_MKNOD + select GRKERNSEC_CHROOT_CAPS + select GRKERNSEC_CHROOT_SYSCTL + select GRKERNSEC_CHROOT_FINDTASK + select GRKERNSEC_PROC + select GRKERNSEC_PROC_MEMMAP if (PAX_NOEXEC || PAX_ASLR) + select GRKERNSEC_HIDESYM + select GRKERNSEC_BRUTE + select GRKERNSEC_SHM if (SYSVIPC) + select GRKERNSEC_PROC_USERGROUP + select GRKERNSEC_KMEM + select GRKERNSEC_RESLOG + select GRKERNSEC_RANDNET + select GRKERNSEC_PROC_ADD + select GRKERNSEC_CHROOT_CHMOD + select GRKERNSEC_CHROOT_NICE + select GRKERNSEC_AUDIT_MOUNT + select GRKERNSEC_MODSTOP if (MODULES) + select PAX + select PAX_RANDUSTACK + select PAX_ASLR + select PAX_RANDMMAP + select PAX_NOEXEC + select PAX_MPROTECT + select PAX_EI_PAX + select PAX_PT_PAX_FLAGS + select PAX_HAVE_ACL_FLAGS + select PAX_KERNEXEC if (!X86_64 && !MODULES && !HOTPLUG_PCI_COMPAQ_NVRAM && !PCI_BIOS) + select PAX_RANDKSTACK if (X86_TSC && !X86_64) + select PAX_SEGMEXEC if (X86 && !X86_64) + select PAX_PAGEEXEC if (!X86) + select PAX_EMUPLT if (ALPHA || PARISC || PPC32 || SPARC32 || SPARC64) + select PAX_DLRESOLVE if (SPARC32 || SPARC64) + select PAX_SYSCALL if (PPC32) + select PAX_EMUTRAMP if (PARISC) + select PAX_EMUSIGRT if (PARISC) + select PAX_NOVSYSCALL if (X86 && !X86_64) + select PAX_ETEXECRELOCS if (ALPHA || IA64 || PARISC) + help + If you say Y here, many of the features of grsecurity will be + enabled, which will protect you against many kinds of attacks + against your system. The heightened security comes at a cost + of an increased chance of incompatibilities with rare software + on your machine. Since this security level enables PaX, you should + view and read about the PaX + project. While you are there, download chpax and run it on + binaries that cause problems with PaX. Also remember that + since the /proc restrictions are enabled, you must run your + identd as gid 1001. This security level enables the following + features in addition to those listed in the low and medium + security levels: + + - Additional /proc restrictions + - Chmod restrictions in chroot + - No signals, ptrace, or viewing of processes outside of chroot + - Capability restrictions in chroot + - Deny fchdir out of chroot + - Priority restrictions in chroot + - Segmentation-based implementation of PaX + - Mprotect restrictions + - Removal of addresses from /proc//[smaps|maps|stat] + - Kernel stack randomization + - Mount/unmount/remount logging + - Kernel symbol hiding + - Destroy unused shared memory + - Prevention of memory exhaustion-based exploits +config GRKERNSEC_CUSTOM + bool "Custom" + help + If you say Y here, you will be able to configure every grsecurity + option, which allows you to enable many more features that aren't + covered in the basic security levels. These additional features + include TPE, socket restrictions, and the sysctl system for + grsecurity. It is advised that you read through the help for + each option to determine its usefulness in your situation. + +endchoice + +menu "Address Space Protection" +depends on GRKERNSEC + +config GRKERNSEC_KMEM + bool "Deny writing to /dev/kmem, /dev/mem, and /dev/port" + help + If you say Y here, /dev/kmem and /dev/mem won't be allowed to + be written to via mmap or otherwise to modify the running kernel. + /dev/port will also not be allowed to be opened. If you have module + support disabled, enabling this will close up four ways that are + currently used to insert malicious code into the running kernel. + Even with all these features enabled, we still highly recommend that + you use the RBAC system, as it is still possible for an attacker to + modify the running kernel through privileged I/O granted by ioperm/iopl. + If you are not using XFree86, you may be able to stop this additional + case by enabling the 'Disable privileged I/O' option. Though nothing + legitimately writes to /dev/kmem, XFree86 does need to write to /dev/mem, + but only to video memory, which is the only writing we allow in this + case. If /dev/kmem or /dev/mem are mmaped without PROT_WRITE, they will + not be allowed to mprotect it with PROT_WRITE later. + It is highly recommended that you say Y here if you meet all the + conditions above. + +config GRKERNSEC_IO + bool "Disable privileged I/O" + depends on X86 + select RTC + help + If you say Y here, all ioperm and iopl calls will return an error. + Ioperm and iopl can be used to modify the running kernel. + Unfortunately, some programs need this access to operate properly, + the most notable of which are XFree86 and hwclock. hwclock can be + remedied by having RTC support in the kernel, so CONFIG_RTC is + enabled if this option is enabled, to ensure that hwclock operates + correctly. XFree86 still will not operate correctly with this option + enabled, so DO NOT CHOOSE Y IF YOU USE XFree86. If you use XFree86 + and you still want to protect your kernel against modification, + use the RBAC system. + +config GRKERNSEC_PROC_MEMMAP + bool "Remove addresses from /proc//[smaps|maps|stat]" + depends on PAX_NOEXEC || PAX_ASLR + help + If you say Y here, the /proc//maps and /proc//stat files will + give no information about the addresses of its mappings if + PaX features that rely on random addresses are enabled on the task. + If you use PaX it is greatly recommended that you say Y here as it + closes up a hole that makes the full ASLR useless for suid + binaries. + +config GRKERNSEC_BRUTE + bool "Deter exploit bruteforcing" + help + If you say Y here, attempts to bruteforce exploits against forking + daemons such as apache or sshd will be deterred. When a child of a + forking daemon is killed by PaX or crashes due to an illegal + instruction, the parent process will be delayed 30 seconds upon every + subsequent fork until the administrator is able to assess the + situation and restart the daemon. It is recommended that you also + enable signal logging in the auditing section so that logs are + generated when a process performs an illegal instruction. + +config GRKERNSEC_MODSTOP + bool "Runtime module disabling" + depends on MODULES + help + If you say Y here, you will be able to disable the ability to (un)load + modules at runtime. This feature is useful if you need the ability + to load kernel modules at boot time, but do not want to allow an + attacker to load a rootkit kernel module into the system, or to remove + a loaded kernel module important to system functioning. You should + enable the /dev/mem protection feature as well, since rootkits can be + inserted into the kernel via other methods than kernel modules. Since + an untrusted module could still be loaded by modifying init scripts and + rebooting the system, it is also recommended that you enable the RBAC + system. If you enable this option, a sysctl option with name + "disable_modules" will be created. Setting this option to "1" disables + module loading. After this option is set, no further writes to it are + allowed until the system is rebooted. + +config GRKERNSEC_HIDESYM + bool "Hide kernel symbols" + help + If you say Y here, getting information on loaded modules, and + displaying all kernel symbols through a syscall will be restricted + to users with CAP_SYS_MODULE. This option is only effective + provided the following conditions are met: + 1) The kernel using grsecurity is not precompiled by some distribution + 2) You are using the RBAC system and hiding other files such as your + kernel image and System.map + 3) You have the additional /proc restrictions enabled, which removes + /proc/kcore + If the above conditions are met, this option will aid to provide a + useful protection against local and remote kernel exploitation of + overflows and arbitrary read/write vulnerabilities. + +endmenu +menu "Role Based Access Control Options" +depends on GRKERNSEC + +config GRKERNSEC_ACL_HIDEKERN + bool "Hide kernel processes" + help + If you say Y here, all kernel threads will be hidden to all + processes but those whose subject has the "view hidden processes" + flag. + +config GRKERNSEC_ACL_MAXTRIES + int "Maximum tries before password lockout" + default 3 + help + This option enforces the maximum number of times a user can attempt + to authorize themselves with the grsecurity RBAC system before being + denied the ability to attempt authorization again for a specified time. + The lower the number, the harder it will be to brute-force a password. + +config GRKERNSEC_ACL_TIMEOUT + int "Time to wait after max password tries, in seconds" + default 30 + help + This option specifies the time the user must wait after attempting to + authorize to the RBAC system with the maximum number of invalid + passwords. The higher the number, the harder it will be to brute-force + a password. + +endmenu +menu "Filesystem Protections" +depends on GRKERNSEC + +config GRKERNSEC_PROC + bool "Proc restrictions" + help + If you say Y here, the permissions of the /proc filesystem + will be altered to enhance system security and privacy. You MUST + choose either a user only restriction or a user and group restriction. + Depending upon the option you choose, you can either restrict users to + see only the processes they themselves run, or choose a group that can + view all processes and files normally restricted to root if you choose + the "restrict to user only" option. NOTE: If you're running identd as + a non-root user, you will have to run it as the group you specify here. + +config GRKERNSEC_PROC_USER + bool "Restrict /proc to user only" + depends on GRKERNSEC_PROC + help + If you say Y here, non-root users will only be able to view their own + processes, and restricts them from viewing network-related information, + and viewing kernel symbol and module information. + +config GRKERNSEC_PROC_USERGROUP + bool "Allow special group" + depends on GRKERNSEC_PROC && !GRKERNSEC_PROC_USER + help + If you say Y here, you will be able to select a group that will be + able to view all processes, network-related information, and + kernel and symbol information. This option is useful if you want + to run identd as a non-root user. + +config GRKERNSEC_PROC_GID + int "GID for special group" + depends on GRKERNSEC_PROC_USERGROUP + default 1001 + +config GRKERNSEC_PROC_ADD + bool "Additional restrictions" + depends on GRKERNSEC_PROC_USER || GRKERNSEC_PROC_USERGROUP + help + If you say Y here, additional restrictions will be placed on + /proc that keep normal users from viewing device information and + slabinfo information that could be useful for exploits. + +config GRKERNSEC_LINK + bool "Linking restrictions" + help + If you say Y here, /tmp race exploits will be prevented, since users + will no longer be able to follow symlinks owned by other users in + world-writable +t directories (i.e. /tmp), unless the owner of the + symlink is the owner of the directory. users will also not be + able to hardlink to files they do not own. If the sysctl option is + enabled, a sysctl option with name "linking_restrictions" is created. + +config GRKERNSEC_FIFO + bool "FIFO restrictions" + help + If you say Y here, users will not be able to write to FIFOs they don't + own in world-writable +t directories (i.e. /tmp), unless the owner of + the FIFO is the same owner of the directory it's held in. If the sysctl + option is enabled, a sysctl option with name "fifo_restrictions" is + created. + +config GRKERNSEC_CHROOT + bool "Chroot jail restrictions" + help + If you say Y here, you will be able to choose several options that will + make breaking out of a chrooted jail much more difficult. If you + encounter no software incompatibilities with the following options, it + is recommended that you enable each one. + +config GRKERNSEC_CHROOT_MOUNT + bool "Deny mounts" + depends on GRKERNSEC_CHROOT + help + If you say Y here, processes inside a chroot will not be able to + mount or remount filesystems. If the sysctl option is enabled, a + sysctl option with name "chroot_deny_mount" is created. + +config GRKERNSEC_CHROOT_DOUBLE + bool "Deny double-chroots" + depends on GRKERNSEC_CHROOT + help + If you say Y here, processes inside a chroot will not be able to chroot + again outside the chroot. This is a widely used method of breaking + out of a chroot jail and should not be allowed. If the sysctl + option is enabled, a sysctl option with name + "chroot_deny_chroot" is created. + +config GRKERNSEC_CHROOT_PIVOT + bool "Deny pivot_root in chroot" + depends on GRKERNSEC_CHROOT + help + If you say Y here, processes inside a chroot will not be able to use + a function called pivot_root() that was introduced in Linux 2.3.41. It + works similar to chroot in that it changes the root filesystem. This + function could be misused in a chrooted process to attempt to break out + of the chroot, and therefore should not be allowed. If the sysctl + option is enabled, a sysctl option with name "chroot_deny_pivot" is + created. + +config GRKERNSEC_CHROOT_CHDIR + bool "Enforce chdir(\"/\") on all chroots" + depends on GRKERNSEC_CHROOT + help + If you say Y here, the current working directory of all newly-chrooted + applications will be set to the the root directory of the chroot. + The man page on chroot(2) states: + Note that this call does not change the current working + directory, so that `.' can be outside the tree rooted at + `/'. In particular, the super-user can escape from a + `chroot jail' by doing `mkdir foo; chroot foo; cd ..'. + + It is recommended that you say Y here, since it's not known to break + any software. If the sysctl option is enabled, a sysctl option with + name "chroot_enforce_chdir" is created. + +config GRKERNSEC_CHROOT_CHMOD + bool "Deny (f)chmod +s" + depends on GRKERNSEC_CHROOT + help + If you say Y here, processes inside a chroot will not be able to chmod + or fchmod files to make them have suid or sgid bits. This protects + against another published method of breaking a chroot. If the sysctl + option is enabled, a sysctl option with name "chroot_deny_chmod" is + created. + +config GRKERNSEC_CHROOT_FCHDIR + bool "Deny fchdir out of chroot" + depends on GRKERNSEC_CHROOT + help + If you say Y here, a well-known method of breaking chroots by fchdir'ing + to a file descriptor of the chrooting process that points to a directory + outside the filesystem will be stopped. If the sysctl option + is enabled, a sysctl option with name "chroot_deny_fchdir" is created. + +config GRKERNSEC_CHROOT_MKNOD + bool "Deny mknod" + depends on GRKERNSEC_CHROOT + help + If you say Y here, processes inside a chroot will not be allowed to + mknod. The problem with using mknod inside a chroot is that it + would allow an attacker to create a device entry that is the same + as one on the physical root of your system, which could range from + anything from the console device to a device for your harddrive (which + they could then use to wipe the drive or steal data). It is recommended + that you say Y here, unless you run into software incompatibilities. + If the sysctl option is enabled, a sysctl option with name + "chroot_deny_mknod" is created. + +config GRKERNSEC_CHROOT_SHMAT + bool "Deny shmat() out of chroot" + depends on GRKERNSEC_CHROOT + help + If you say Y here, processes inside a chroot will not be able to attach + to shared memory segments that were created outside of the chroot jail. + It is recommended that you say Y here. If the sysctl option is enabled, + a sysctl option with name "chroot_deny_shmat" is created. + +config GRKERNSEC_CHROOT_UNIX + bool "Deny access to abstract AF_UNIX sockets out of chroot" + depends on GRKERNSEC_CHROOT + help + If you say Y here, processes inside a chroot will not be able to + connect to abstract (meaning not belonging to a filesystem) Unix + domain sockets that were bound outside of a chroot. It is recommended + that you say Y here. If the sysctl option is enabled, a sysctl option + with name "chroot_deny_unix" is created. + +config GRKERNSEC_CHROOT_FINDTASK + bool "Protect outside processes" + depends on GRKERNSEC_CHROOT + help + If you say Y here, processes inside a chroot will not be able to + kill, send signals with fcntl, ptrace, capget, setpgid, getpgid, + getsid, or view any process outside of the chroot. If the sysctl + option is enabled, a sysctl option with name "chroot_findtask" is + created. + +config GRKERNSEC_CHROOT_NICE + bool "Restrict priority changes" + depends on GRKERNSEC_CHROOT + help + If you say Y here, processes inside a chroot will not be able to raise + the priority of processes in the chroot, or alter the priority of + processes outside the chroot. This provides more security than simply + removing CAP_SYS_NICE from the process' capability set. If the + sysctl option is enabled, a sysctl option with name "chroot_restrict_nice" + is created. + +config GRKERNSEC_CHROOT_SYSCTL + bool "Deny sysctl writes" + depends on GRKERNSEC_CHROOT + help + If you say Y here, an attacker in a chroot will not be able to + write to sysctl entries, either by sysctl(2) or through a /proc + interface. It is strongly recommended that you say Y here. If the + sysctl option is enabled, a sysctl option with name + "chroot_deny_sysctl" is created. + +config GRKERNSEC_CHROOT_CAPS + bool "Capability restrictions" + depends on GRKERNSEC_CHROOT + help + If you say Y here, the capabilities on all root processes within a + chroot jail will be lowered to stop module insertion, raw i/o, + system and net admin tasks, rebooting the system, modifying immutable + files, modifying IPC owned by another, and changing the system time. + This is left an option because it can break some apps. Disable this + if your chrooted apps are having problems performing those kinds of + tasks. If the sysctl option is enabled, a sysctl option with + name "chroot_caps" is created. + +endmenu +menu "Kernel Auditing" +depends on GRKERNSEC + +config GRKERNSEC_AUDIT_GROUP + bool "Single group for auditing" + help + If you say Y here, the exec, chdir, (un)mount, and ipc logging features + will only operate on a group you specify. This option is recommended + if you only want to watch certain users instead of having a large + amount of logs from the entire system. If the sysctl option is enabled, + a sysctl option with name "audit_group" is created. + +config GRKERNSEC_AUDIT_GID + int "GID for auditing" + depends on GRKERNSEC_AUDIT_GROUP + default 1007 + +config GRKERNSEC_EXECLOG + bool "Exec logging" + help + If you say Y here, all execve() calls will be logged (since the + other exec*() calls are frontends to execve(), all execution + will be logged). Useful for shell-servers that like to keep track + of their users. If the sysctl option is enabled, a sysctl option with + name "exec_logging" is created. + WARNING: This option when enabled will produce a LOT of logs, especially + on an active system. + +config GRKERNSEC_RESLOG + bool "Resource logging" + help + If you say Y here, all attempts to overstep resource limits will + be logged with the resource name, the requested size, and the current + limit. It is highly recommended that you say Y here. If the sysctl + option is enabled, a sysctl option with name "resource_logging" is + created. If the RBAC system is enabled, the sysctl value is ignored. + +config GRKERNSEC_CHROOT_EXECLOG + bool "Log execs within chroot" + help + If you say Y here, all executions inside a chroot jail will be logged + to syslog. This can cause a large amount of logs if certain + applications (eg. djb's daemontools) are installed on the system, and + is therefore left as an option. If the sysctl option is enabled, a + sysctl option with name "chroot_execlog" is created. + +config GRKERNSEC_AUDIT_CHDIR + bool "Chdir logging" + help + If you say Y here, all chdir() calls will be logged. If the sysctl + option is enabled, a sysctl option with name "audit_chdir" is created. + +config GRKERNSEC_AUDIT_MOUNT + bool "(Un)Mount logging" + help + If you say Y here, all mounts and unmounts will be logged. If the + sysctl option is enabled, a sysctl option with name "audit_mount" is + created. + +config GRKERNSEC_AUDIT_IPC + bool "IPC logging" + help + If you say Y here, creation and removal of message queues, semaphores, + and shared memory will be logged. If the sysctl option is enabled, a + sysctl option with name "audit_ipc" is created. + +config GRKERNSEC_SIGNAL + bool "Signal logging" + help + If you say Y here, certain important signals will be logged, such as + SIGSEGV, which will as a result inform you of when a error in a program + occurred, which in some cases could mean a possible exploit attempt. + If the sysctl option is enabled, a sysctl option with name + "signal_logging" is created. + +config GRKERNSEC_FORKFAIL + bool "Fork failure logging" + help + If you say Y here, all failed fork() attempts will be logged. + This could suggest a fork bomb, or someone attempting to overstep + their process limit. If the sysctl option is enabled, a sysctl option + with name "forkfail_logging" is created. + +config GRKERNSEC_TIME + bool "Time change logging" + help + If you say Y here, any changes of the system clock will be logged. + If the sysctl option is enabled, a sysctl option with name + "timechange_logging" is created. + +config GRKERNSEC_PROC_IPADDR + bool "/proc//ipaddr support" + help + If you say Y here, a new entry will be added to each /proc/ + directory that contains the IP address of the person using the task. + The IP is carried across local TCP and AF_UNIX stream sockets. + This information can be useful for IDS/IPSes to perform remote response + to a local attack. The entry is readable by only the owner of the + process (and root if he has CAP_DAC_OVERRIDE, which can be removed via + the RBAC system), and thus does not create privacy concerns. + +config GRKERNSEC_AUDIT_TEXTREL + bool 'ELF text relocations logging (READ HELP)' + depends on PAX_MPROTECT + help + If you say Y here, text relocations will be logged with the filename + of the offending library or binary. The purpose of the feature is + to help Linux distribution developers get rid of libraries and + binaries that need text relocations which hinder the future progress + of PaX. Only Linux distribution developers should say Y here, and + never on a production machine, as this option creates an information + leak that could aid an attacker in defeating the randomization of + a single memory region. If the sysctl option is enabled, a sysctl + option with name "audit_textrel" is created. + +endmenu + +menu "Executable Protections" +depends on GRKERNSEC + +config GRKERNSEC_EXECVE + bool "Enforce RLIMIT_NPROC on execs" + help + If you say Y here, users with a resource limit on processes will + have the value checked during execve() calls. The current system + only checks the system limit during fork() calls. If the sysctl option + is enabled, a sysctl option with name "execve_limiting" is created. + +config GRKERNSEC_SHM + bool "Destroy unused shared memory" + depends on SYSVIPC + help + If you say Y here, shared memory will be destroyed when no one is + attached to it. Otherwise, resources involved with the shared + memory can be used up and not be associated with any process (as the + shared memory still exists, and the creating process has exited). If + the sysctl option is enabled, a sysctl option with name + "destroy_unused_shm" is created. + +config GRKERNSEC_DMESG + bool "Dmesg(8) restriction" + help + If you say Y here, non-root users will not be able to use dmesg(8) + to view up to the last 4kb of messages in the kernel's log buffer. + If the sysctl option is enabled, a sysctl option with name "dmesg" is + created. + +config GRKERNSEC_RANDPID + bool "Randomized PIDs" + help + If you say Y here, all PIDs created on the system will be + pseudo-randomly generated. This is extremely effective along + with the /proc restrictions to disallow an attacker from guessing + pids of daemons, etc. PIDs are also used in some cases as part + of a naming system for temporary files, so this option would keep + those filenames from being predicted as well. We also use code + to make sure that PID numbers aren't reused too soon. If the sysctl + option is enabled, a sysctl option with name "rand_pids" is created. + +config GRKERNSEC_TPE + bool "Trusted Path Execution (TPE)" + help + If you say Y here, you will be able to choose a gid to add to the + supplementary groups of users you want to mark as "untrusted." + These users will not be able to execute any files that are not in + root-owned directories writable only by root. If the sysctl option + is enabled, a sysctl option with name "tpe" is created. + +config GRKERNSEC_TPE_ALL + bool "Partially restrict non-root users" + depends on GRKERNSEC_TPE + help + If you say Y here, All non-root users other than the ones in the + group specified in the main TPE option will only be allowed to + execute files in directories they own that are not group or + world-writable, or in directories owned by root and writable only by + root. If the sysctl option is enabled, a sysctl option with name + "tpe_restrict_all" is created. + +config GRKERNSEC_TPE_INVERT + bool "Invert GID option" + depends on GRKERNSEC_TPE + help + If you say Y here, the group you specify in the TPE configuration will + decide what group TPE restrictions will be *disabled* for. This + option is useful if you want TPE restrictions to be applied to most + users on the system. + +config GRKERNSEC_TPE_GID + int "GID for untrusted users" + depends on GRKERNSEC_TPE && !GRKERNSEC_TPE_INVERT + default 1005 + help + If you have selected the "Invert GID option" above, setting this + GID determines what group TPE restrictions will be *disabled* for. + If you have not selected the "Invert GID option" above, setting this + GID determines what group TPE restrictions will be *enabled* for. + If the sysctl option is enabled, a sysctl option with name "tpe_gid" + is created. + +config GRKERNSEC_TPE_GID + int "GID for trusted users" + depends on GRKERNSEC_TPE && GRKERNSEC_TPE_INVERT + default 1005 + help + If you have selected the "Invert GID option" above, setting this + GID determines what group TPE restrictions will be *disabled* for. + If you have not selected the "Invert GID option" above, setting this + GID determines what group TPE restrictions will be *enabled* for. + If the sysctl option is enabled, a sysctl option with name "tpe_gid" + is created. + +endmenu +menu "Network Protections" +depends on GRKERNSEC + +config GRKERNSEC_RANDNET + bool "Larger entropy pools" + help + If you say Y here, the entropy pools used for many features of Linux + and grsecurity will be doubled in size. Since several grsecurity + features use additional randomness, it is recommended that you say Y + here. Saying Y here has a similar effect as modifying + /proc/sys/kernel/random/poolsize. + +config GRKERNSEC_SOCKET + bool "Socket restrictions" + help + If you say Y here, you will be able to choose from several options. + If you assign a GID on your system and add it to the supplementary + groups of users you want to restrict socket access to, this patch + will perform up to three things, based on the option(s) you choose. + +config GRKERNSEC_SOCKET_ALL + bool "Deny any sockets to group" + depends on GRKERNSEC_SOCKET + help + If you say Y here, you will be able to choose a GID of whose users will + be unable to connect to other hosts from your machine or run server + applications from your machine. If the sysctl option is enabled, a + sysctl option with name "socket_all" is created. + +config GRKERNSEC_SOCKET_ALL_GID + int "GID to deny all sockets for" + depends on GRKERNSEC_SOCKET_ALL + default 1004 + help + Here you can choose the GID to disable socket access for. Remember to + add the users you want socket access disabled for to the GID + specified here. If the sysctl option is enabled, a sysctl option + with name "socket_all_gid" is created. + +config GRKERNSEC_SOCKET_CLIENT + bool "Deny client sockets to group" + depends on GRKERNSEC_SOCKET + help + If you say Y here, you will be able to choose a GID of whose users will + be unable to connect to other hosts from your machine, but will be + able to run servers. If this option is enabled, all users in the group + you specify will have to use passive mode when initiating ftp transfers + from the shell on your machine. If the sysctl option is enabled, a + sysctl option with name "socket_client" is created. + +config GRKERNSEC_SOCKET_CLIENT_GID + int "GID to deny client sockets for" + depends on GRKERNSEC_SOCKET_CLIENT + default 1003 + help + Here you can choose the GID to disable client socket access for. + Remember to add the users you want client socket access disabled for to + the GID specified here. If the sysctl option is enabled, a sysctl + option with name "socket_client_gid" is created. + +config GRKERNSEC_SOCKET_SERVER + bool "Deny server sockets to group" + depends on GRKERNSEC_SOCKET + help + If you say Y here, you will be able to choose a GID of whose users will + be unable to run server applications from your machine. If the sysctl + option is enabled, a sysctl option with name "socket_server" is created. + +config GRKERNSEC_SOCKET_SERVER_GID + int "GID to deny server sockets for" + depends on GRKERNSEC_SOCKET_SERVER + default 1002 + help + Here you can choose the GID to disable server socket access for. + Remember to add the users you want server socket access disabled for to + the GID specified here. If the sysctl option is enabled, a sysctl + option with name "socket_server_gid" is created. + +endmenu +menu "Sysctl support" +depends on GRKERNSEC && SYSCTL + +config GRKERNSEC_SYSCTL + bool "Sysctl support" + help + If you say Y here, you will be able to change the options that + grsecurity runs with at bootup, without having to recompile your + kernel. You can echo values to files in /proc/sys/kernel/grsecurity + to enable (1) or disable (0) various features. All the sysctl entries + are mutable until the "grsec_lock" entry is set to a non-zero value. + All features enabled in the kernel configuration are disabled at boot + if you do not say Y to the "Turn on features by default" option. + All options should be set at startup, and the grsec_lock entry should + be set to a non-zero value after all the options are set. + *THIS IS EXTREMELY IMPORTANT* + +config GRKERNSEC_SYSCTL_ON + bool "Turn on features by default" + depends on GRKERNSEC_SYSCTL + help + If you say Y here, instead of having all features enabled in the + kernel configuration disabled at boot time, the features will be + enabled at boot time. It is recommended you say Y here unless + there is some reason you would want all sysctl-tunable features to + be disabled by default. As mentioned elsewhere, it is important + to enable the grsec_lock entry once you have finished modifying + the sysctl entries. + +endmenu +menu "Logging Options" +depends on GRKERNSEC + +config GRKERNSEC_FLOODTIME + int "Seconds in between log messages (minimum)" + default 10 + help + This option allows you to enforce the number of seconds between + grsecurity log messages. The default should be suitable for most + people, however, if you choose to change it, choose a value small enough + to allow informative logs to be produced, but large enough to + prevent flooding. + +config GRKERNSEC_FLOODBURST + int "Number of messages in a burst (maximum)" + default 4 + help + This option allows you to choose the maximum number of messages allowed + within the flood time interval you chose in a separate option. The + default should be suitable for most people, however if you find that + many of your logs are being interpreted as flooding, you may want to + raise this value. + +endmenu + +endmenu diff -NurpP --minimal linux-2.6.17.13/grsecurity/Makefile linux-2.6.17.13-g-v/grsecurity/Makefile --- linux-2.6.17.13/grsecurity/Makefile 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/Makefile 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,20 @@ +# grsecurity's ACL system was originally written in 2001 by Michael Dalton +# during 2001-2005 it has been completely redesigned by Brad Spengler +# into an RBAC system +# +# All code in this directory and various hooks inserted throughout the kernel +# are copyright Brad Spengler, and released under the GPL v2 or higher + +obj-y = grsec_chdir.o grsec_chroot.o grsec_exec.o grsec_fifo.o grsec_fork.o \ + grsec_mount.o grsec_rand.o grsec_sig.o grsec_sock.o grsec_sysctl.o \ + grsec_time.o grsec_tpe.o grsec_ipc.o grsec_link.o grsec_textrel.o + +obj-$(CONFIG_GRKERNSEC) += grsec_init.o grsum.o gracl.o gracl_ip.o gracl_segv.o \ + gracl_cap.o gracl_alloc.o gracl_shm.o grsec_mem.o gracl_fs.o \ + gracl_learn.o grsec_log.o +obj-$(CONFIG_GRKERNSEC_RESLOG) += gracl_res.o + +ifndef CONFIG_GRKERNSEC +obj-y += grsec_disabled.o +endif + diff -NurpP --minimal linux-2.6.17.13/grsecurity/gracl.c linux-2.6.17.13-g-v/grsecurity/gracl.c --- linux-2.6.17.13/grsecurity/gracl.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/gracl.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,3547 @@ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include + +static struct acl_role_db acl_role_set; +static struct name_db name_set; +static struct inodev_db inodev_set; + +/* for keeping track of userspace pointers used for subjects, so we + can share references in the kernel as well +*/ + +static struct dentry *real_root; +static struct vfsmount *real_root_mnt; + +static struct acl_subj_map_db subj_map_set; + +static struct acl_role_label *default_role; + +static u16 acl_sp_role_value; + +extern char *gr_shared_page[4]; +static DECLARE_MUTEX(gr_dev_sem); +rwlock_t gr_inode_lock = RW_LOCK_UNLOCKED; + +struct gr_arg *gr_usermode; + +static unsigned int gr_status = GR_STATUS_INIT; + +extern int chkpw(struct gr_arg *entry, unsigned char *salt, unsigned char *sum); +extern void gr_clear_learn_entries(void); + +#ifdef CONFIG_GRKERNSEC_RESLOG +extern void gr_log_resource(const struct task_struct *task, + const int res, const unsigned long wanted, const int gt); +#endif + +extern char * __d_path(struct dentry *dentry, struct vfsmount *vfsmnt, + struct dentry *root, struct vfsmount *rootmnt, + char *buffer, int buflen); + +unsigned char *gr_system_salt; +unsigned char *gr_system_sum; + +static struct sprole_pw **acl_special_roles = NULL; +static __u16 num_sprole_pws = 0; + +static struct acl_role_label *kernel_role = NULL; + +static unsigned int gr_auth_attempts = 0; +static unsigned long gr_auth_expires = 0UL; + +extern struct vfsmount *sock_mnt; +extern struct vfsmount *pipe_mnt; +extern struct vfsmount *shm_mnt; +static struct acl_object_label *fakefs_obj; + +extern int gr_init_uidset(void); +extern void gr_free_uidset(void); +extern void gr_remove_uid(uid_t uid); +extern int gr_find_uid(uid_t uid); + +__inline__ int +gr_acl_is_enabled(void) +{ + return (gr_status & GR_READY); +} + +char gr_roletype_to_char(void) +{ + switch (current->role->roletype & + (GR_ROLE_DEFAULT | GR_ROLE_USER | GR_ROLE_GROUP | + GR_ROLE_SPECIAL)) { + case GR_ROLE_DEFAULT: + return 'D'; + case GR_ROLE_USER: + return 'U'; + case GR_ROLE_GROUP: + return 'G'; + case GR_ROLE_SPECIAL: + return 'S'; + } + + return 'X'; +} + +__inline__ int +gr_acl_tpe_check(void) +{ + if (unlikely(!(gr_status & GR_READY))) + return 0; + if (current->role->roletype & GR_ROLE_TPE) + return 1; + else + return 0; +} + +int +gr_handle_rawio(const struct inode *inode) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_CAPS + if (inode && S_ISBLK(inode->i_mode) && + grsec_enable_chroot_caps && proc_is_chrooted(current) && + !capable(CAP_SYS_RAWIO)) + return 1; +#endif + return 0; +} + +static int +gr_streq(const char *a, const char *b, const unsigned int lena, const unsigned int lenb) +{ + int i; + unsigned long *l1; + unsigned long *l2; + unsigned char *c1; + unsigned char *c2; + int num_longs; + + if (likely(lena != lenb)) + return 0; + + l1 = (unsigned long *)a; + l2 = (unsigned long *)b; + + num_longs = lena / sizeof(unsigned long); + + for (i = num_longs; i--; l1++, l2++) { + if (unlikely(*l1 != *l2)) + return 0; + } + + c1 = (unsigned char *) l1; + c2 = (unsigned char *) l2; + + i = lena - (num_longs * sizeof(unsigned long)); + + for (; i--; c1++, c2++) { + if (unlikely(*c1 != *c2)) + return 0; + } + + return 1; +} + +static char * +gen_full_path(struct dentry *dentry, struct vfsmount *vfsmnt, + struct dentry *root, struct vfsmount *rootmnt, char *buf, int buflen) +{ + char *end = buf + buflen; + char *retval; + int namelen = 0; + + *--end = '\0'; + + retval = end - 1; + *retval = '/'; + + if (dentry == root && vfsmnt == rootmnt) + return retval; + if (dentry != vfsmnt->mnt_root && !IS_ROOT(dentry)) { + namelen = strlen(dentry->d_name.name); + buflen -= namelen; + if (buflen < 2) + goto err; + if (dentry->d_parent != root || vfsmnt != rootmnt) + buflen--; + } + + retval = __d_path(dentry->d_parent, vfsmnt, root, rootmnt, buf, buflen); + if (unlikely(IS_ERR(retval))) +err: + retval = strcpy(buf, ""); + else if (namelen != 0) { + end = buf + buflen - 1; // accounts for null termination + if (dentry->d_parent != root || vfsmnt != rootmnt) + *end++ = '/'; // accounted for above with buflen-- + memcpy(end, dentry->d_name.name, namelen); + } + + return retval; +} + +static char * +__d_real_path(const struct dentry *dentry, const struct vfsmount *vfsmnt, + char *buf, int buflen) +{ + char *res; + + /* we can use real_root, real_root_mnt, because this is only called + by the RBAC system */ + res = gen_full_path((struct dentry *)dentry, (struct vfsmount *)vfsmnt, real_root, real_root_mnt, buf, buflen); + + return res; +} + +static char * +d_real_path(const struct dentry *dentry, const struct vfsmount *vfsmnt, + char *buf, int buflen) +{ + char *res; + struct dentry *root; + struct vfsmount *rootmnt; + + /* we can't use real_root, real_root_mnt, because they belong only to the RBAC system */ + read_lock(&child_reaper->fs->lock); + root = dget(child_reaper->fs->root); + rootmnt = mntget(child_reaper->fs->rootmnt); + read_unlock(&child_reaper->fs->lock); + + spin_lock(&dcache_lock); + res = gen_full_path((struct dentry *)dentry, (struct vfsmount *)vfsmnt, root, rootmnt, buf, buflen); + spin_unlock(&dcache_lock); + + dput(root); + mntput(rootmnt); + return res; +} + +static char * +gr_to_filename_rbac(const struct dentry *dentry, const struct vfsmount *mnt) +{ + char *ret; + spin_lock(&dcache_lock); + ret = __d_real_path(dentry, mnt, per_cpu_ptr(gr_shared_page[0],smp_processor_id()), + PAGE_SIZE); + spin_unlock(&dcache_lock); + return ret; +} + +char * +gr_to_filename_nolock(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return __d_real_path(dentry, mnt, per_cpu_ptr(gr_shared_page[0],smp_processor_id()), + PAGE_SIZE); +} + +char * +gr_to_filename(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return d_real_path(dentry, mnt, per_cpu_ptr(gr_shared_page[0], smp_processor_id()), + PAGE_SIZE); +} + +char * +gr_to_filename1(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return d_real_path(dentry, mnt, per_cpu_ptr(gr_shared_page[1], smp_processor_id()), + PAGE_SIZE); +} + +char * +gr_to_filename2(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return d_real_path(dentry, mnt, per_cpu_ptr(gr_shared_page[2], smp_processor_id()), + PAGE_SIZE); +} + +char * +gr_to_filename3(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return d_real_path(dentry, mnt, per_cpu_ptr(gr_shared_page[3], smp_processor_id()), + PAGE_SIZE); +} + +__inline__ __u32 +to_gr_audit(const __u32 reqmode) +{ + /* masks off auditable permission flags, then shifts them to create + auditing flags, and adds the special case of append auditing if + we're requesting write */ + return (((reqmode & GR_AUDIT_READ) << 10) | ((reqmode & GR_WRITE) ? GR_AUDIT_APPEND : 0)); +} + +struct acl_subject_label * +lookup_subject_map(const struct acl_subject_label *userp) +{ + unsigned int index = shash(userp, subj_map_set.s_size); + struct subject_map *match; + + match = subj_map_set.s_hash[index]; + + while (match && match->user != userp) + match = match->next; + + if (match != NULL) + return match->kernel; + else + return NULL; +} + +static void +insert_subj_map_entry(struct subject_map *subjmap) +{ + unsigned int index = shash(subjmap->user, subj_map_set.s_size); + struct subject_map **curr; + + subjmap->prev = NULL; + + curr = &subj_map_set.s_hash[index]; + if (*curr != NULL) + (*curr)->prev = subjmap; + + subjmap->next = *curr; + *curr = subjmap; + + return; +} + +static struct acl_role_label * +lookup_acl_role_label(const struct task_struct *task, const uid_t uid, + const gid_t gid) +{ + unsigned int index = rhash(uid, GR_ROLE_USER, acl_role_set.r_size); + struct acl_role_label *match; + struct role_allowed_ip *ipp; + unsigned int x; + + match = acl_role_set.r_hash[index]; + + while (match) { + if ((match->roletype & (GR_ROLE_DOMAIN | GR_ROLE_USER)) == (GR_ROLE_DOMAIN | GR_ROLE_USER)) { + for (x = 0; x < match->domain_child_num; x++) { + if (match->domain_children[x] == uid) + goto found; + } + } else if (match->uidgid == uid && match->roletype & GR_ROLE_USER) + break; + match = match->next; + } +found: + if (match == NULL) { + try_group: + index = rhash(gid, GR_ROLE_GROUP, acl_role_set.r_size); + match = acl_role_set.r_hash[index]; + + while (match) { + if ((match->roletype & (GR_ROLE_DOMAIN | GR_ROLE_GROUP)) == (GR_ROLE_DOMAIN | GR_ROLE_GROUP)) { + for (x = 0; x < match->domain_child_num; x++) { + if (match->domain_children[x] == gid) + goto found2; + } + } else if (match->uidgid == gid && match->roletype & GR_ROLE_GROUP) + break; + match = match->next; + } +found2: + if (match == NULL) + match = default_role; + if (match->allowed_ips == NULL) + return match; + else { + for (ipp = match->allowed_ips; ipp; ipp = ipp->next) { + if (likely + ((ntohl(task->signal->curr_ip) & ipp->netmask) == + (ntohl(ipp->addr) & ipp->netmask))) + return match; + } + match = default_role; + } + } else if (match->allowed_ips == NULL) { + return match; + } else { + for (ipp = match->allowed_ips; ipp; ipp = ipp->next) { + if (likely + ((ntohl(task->signal->curr_ip) & ipp->netmask) == + (ntohl(ipp->addr) & ipp->netmask))) + return match; + } + goto try_group; + } + + return match; +} + +struct acl_subject_label * +lookup_acl_subj_label(const ino_t ino, const dev_t dev, + const struct acl_role_label *role) +{ + unsigned int index = fhash(ino, dev, role->subj_hash_size); + struct acl_subject_label *match; + + match = role->subj_hash[index]; + + while (match && (match->inode != ino || match->device != dev || + (match->mode & GR_DELETED))) { + match = match->next; + } + + if (match && !(match->mode & GR_DELETED)) + return match; + else + return NULL; +} + +static struct acl_object_label * +lookup_acl_obj_label(const ino_t ino, const dev_t dev, + const struct acl_subject_label *subj) +{ + unsigned int index = fhash(ino, dev, subj->obj_hash_size); + struct acl_object_label *match; + + match = subj->obj_hash[index]; + + while (match && (match->inode != ino || match->device != dev || + (match->mode & GR_DELETED))) { + match = match->next; + } + + if (match && !(match->mode & GR_DELETED)) + return match; + else + return NULL; +} + +static struct acl_object_label * +lookup_acl_obj_label_create(const ino_t ino, const dev_t dev, + const struct acl_subject_label *subj) +{ + unsigned int index = fhash(ino, dev, subj->obj_hash_size); + struct acl_object_label *match; + + match = subj->obj_hash[index]; + + while (match && (match->inode != ino || match->device != dev || + !(match->mode & GR_DELETED))) { + match = match->next; + } + + if (match && (match->mode & GR_DELETED)) + return match; + + match = subj->obj_hash[index]; + + while (match && (match->inode != ino || match->device != dev || + (match->mode & GR_DELETED))) { + match = match->next; + } + + if (match && !(match->mode & GR_DELETED)) + return match; + else + return NULL; +} + +static struct name_entry * +lookup_name_entry(const char *name) +{ + unsigned int len = strlen(name); + unsigned int key = full_name_hash(name, len); + unsigned int index = key % name_set.n_size; + struct name_entry *match; + + match = name_set.n_hash[index]; + + while (match && (match->key != key || !gr_streq(match->name, name, match->len, len))) + match = match->next; + + return match; +} + +static struct inodev_entry * +lookup_inodev_entry(const ino_t ino, const dev_t dev) +{ + unsigned int index = fhash(ino, dev, inodev_set.i_size); + struct inodev_entry *match; + + match = inodev_set.i_hash[index]; + + while (match && (match->nentry->inode != ino || match->nentry->device != dev)) + match = match->next; + + return match; +} + +static void +insert_inodev_entry(struct inodev_entry *entry) +{ + unsigned int index = fhash(entry->nentry->inode, entry->nentry->device, + inodev_set.i_size); + struct inodev_entry **curr; + + entry->prev = NULL; + + curr = &inodev_set.i_hash[index]; + if (*curr != NULL) + (*curr)->prev = entry; + + entry->next = *curr; + *curr = entry; + + return; +} + +static void +__insert_acl_role_label(struct acl_role_label *role, uid_t uidgid) +{ + unsigned int index = + rhash(uidgid, role->roletype & (GR_ROLE_USER | GR_ROLE_GROUP), acl_role_set.r_size); + struct acl_role_label **curr; + + role->prev = NULL; + + curr = &acl_role_set.r_hash[index]; + if (*curr != NULL) + (*curr)->prev = role; + + role->next = *curr; + *curr = role; + + return; +} + +static void +insert_acl_role_label(struct acl_role_label *role) +{ + int i; + + if (role->roletype & GR_ROLE_DOMAIN) { + for (i = 0; i < role->domain_child_num; i++) + __insert_acl_role_label(role, role->domain_children[i]); + } else + __insert_acl_role_label(role, role->uidgid); +} + +static int +insert_name_entry(char *name, const ino_t inode, const dev_t device) +{ + struct name_entry **curr, *nentry; + struct inodev_entry *ientry; + unsigned int len = strlen(name); + unsigned int key = full_name_hash(name, len); + unsigned int index = key % name_set.n_size; + + curr = &name_set.n_hash[index]; + + while (*curr && ((*curr)->key != key || !gr_streq((*curr)->name, name, (*curr)->len, len))) + curr = &((*curr)->next); + + if (*curr != NULL) + return 1; + + nentry = acl_alloc(sizeof (struct name_entry)); + if (nentry == NULL) + return 0; + ientry = acl_alloc(sizeof (struct inodev_entry)); + if (ientry == NULL) + return 0; + ientry->nentry = nentry; + + nentry->key = key; + nentry->name = name; + nentry->inode = inode; + nentry->device = device; + nentry->len = len; + + nentry->prev = NULL; + curr = &name_set.n_hash[index]; + if (*curr != NULL) + (*curr)->prev = nentry; + nentry->next = *curr; + *curr = nentry; + + /* insert us into the table searchable by inode/dev */ + insert_inodev_entry(ientry); + + return 1; +} + +static void +insert_acl_obj_label(struct acl_object_label *obj, + struct acl_subject_label *subj) +{ + unsigned int index = + fhash(obj->inode, obj->device, subj->obj_hash_size); + struct acl_object_label **curr; + + + obj->prev = NULL; + + curr = &subj->obj_hash[index]; + if (*curr != NULL) + (*curr)->prev = obj; + + obj->next = *curr; + *curr = obj; + + return; +} + +static void +insert_acl_subj_label(struct acl_subject_label *obj, + struct acl_role_label *role) +{ + unsigned int index = fhash(obj->inode, obj->device, role->subj_hash_size); + struct acl_subject_label **curr; + + obj->prev = NULL; + + curr = &role->subj_hash[index]; + if (*curr != NULL) + (*curr)->prev = obj; + + obj->next = *curr; + *curr = obj; + + return; +} + +/* allocating chained hash tables, so optimal size is where lambda ~ 1 */ + +static void * +create_table(__u32 * len, int elementsize) +{ + unsigned int table_sizes[] = { + 7, 13, 31, 61, 127, 251, 509, 1021, 2039, 4093, 8191, 16381, + 32749, 65521, 131071, 262139, 524287, 1048573, 2097143, + 4194301, 8388593, 16777213, 33554393, 67108859, 134217689, + 268435399, 536870909, 1073741789, 2147483647 + }; + void *newtable = NULL; + unsigned int pwr = 0; + + while ((pwr < ((sizeof (table_sizes) / sizeof (table_sizes[0])) - 1)) && + table_sizes[pwr] <= *len) + pwr++; + + if (table_sizes[pwr] <= *len) + return newtable; + + if ((table_sizes[pwr] * elementsize) <= PAGE_SIZE) + newtable = + kmalloc(table_sizes[pwr] * elementsize, GFP_KERNEL); + else + newtable = vmalloc(table_sizes[pwr] * elementsize); + + *len = table_sizes[pwr]; + + return newtable; +} + +static int +init_variables(const struct gr_arg *arg) +{ + unsigned int stacksize; + + subj_map_set.s_size = arg->role_db.num_subjects; + acl_role_set.r_size = arg->role_db.num_roles + arg->role_db.num_domain_children; + name_set.n_size = arg->role_db.num_objects; + inodev_set.i_size = arg->role_db.num_objects; + + if (!subj_map_set.s_size || !acl_role_set.r_size || + !name_set.n_size || !inodev_set.i_size) + return 1; + + if (!gr_init_uidset()) + return 1; + + /* set up the stack that holds allocation info */ + + stacksize = arg->role_db.num_pointers + 5; + + if (!acl_alloc_stack_init(stacksize)) + return 1; + + /* grab reference for the real root dentry and vfsmount */ + read_lock(&child_reaper->fs->lock); + real_root_mnt = mntget(child_reaper->fs->rootmnt); + real_root = dget(child_reaper->fs->root); + read_unlock(&child_reaper->fs->lock); + + fakefs_obj = acl_alloc(sizeof(struct acl_object_label)); + if (fakefs_obj == NULL) + return 1; + fakefs_obj->mode = GR_FIND | GR_READ | GR_WRITE | GR_EXEC; + + subj_map_set.s_hash = + (struct subject_map **) create_table(&subj_map_set.s_size, sizeof(void *)); + acl_role_set.r_hash = + (struct acl_role_label **) create_table(&acl_role_set.r_size, sizeof(void *)); + name_set.n_hash = (struct name_entry **) create_table(&name_set.n_size, sizeof(void *)); + inodev_set.i_hash = + (struct inodev_entry **) create_table(&inodev_set.i_size, sizeof(void *)); + + if (!subj_map_set.s_hash || !acl_role_set.r_hash || + !name_set.n_hash || !inodev_set.i_hash) + return 1; + + memset(subj_map_set.s_hash, 0, + sizeof(struct subject_map *) * subj_map_set.s_size); + memset(acl_role_set.r_hash, 0, + sizeof (struct acl_role_label *) * acl_role_set.r_size); + memset(name_set.n_hash, 0, + sizeof (struct name_entry *) * name_set.n_size); + memset(inodev_set.i_hash, 0, + sizeof (struct inodev_entry *) * inodev_set.i_size); + + return 0; +} + +/* free information not needed after startup + currently contains user->kernel pointer mappings for subjects +*/ + +static void +free_init_variables(void) +{ + __u32 i; + + if (subj_map_set.s_hash) { + for (i = 0; i < subj_map_set.s_size; i++) { + if (subj_map_set.s_hash[i]) { + kfree(subj_map_set.s_hash[i]); + subj_map_set.s_hash[i] = NULL; + } + } + + if ((subj_map_set.s_size * sizeof (struct subject_map *)) <= + PAGE_SIZE) + kfree(subj_map_set.s_hash); + else + vfree(subj_map_set.s_hash); + } + + return; +} + +static void +free_variables(void) +{ + struct acl_subject_label *s; + struct acl_role_label *r; + struct task_struct *task, *task2; + unsigned int i, x; + + gr_clear_learn_entries(); + + read_lock(&tasklist_lock); + do_each_thread(task2, task) { + task->acl_sp_role = 0; + task->acl_role_id = 0; + task->acl = NULL; + task->role = NULL; + } while_each_thread(task2, task); + read_unlock(&tasklist_lock); + + /* release the reference to the real root dentry and vfsmount */ + if (real_root) + dput(real_root); + real_root = NULL; + if (real_root_mnt) + mntput(real_root_mnt); + real_root_mnt = NULL; + + /* free all object hash tables */ + + FOR_EACH_ROLE_START(r, i) + if (r->subj_hash == NULL) + break; + FOR_EACH_SUBJECT_START(r, s, x) + if (s->obj_hash == NULL) + break; + if ((s->obj_hash_size * sizeof (struct acl_object_label *)) <= PAGE_SIZE) + kfree(s->obj_hash); + else + vfree(s->obj_hash); + FOR_EACH_SUBJECT_END(s, x) + FOR_EACH_NESTED_SUBJECT_START(r, s) + if (s->obj_hash == NULL) + break; + if ((s->obj_hash_size * sizeof (struct acl_object_label *)) <= PAGE_SIZE) + kfree(s->obj_hash); + else + vfree(s->obj_hash); + FOR_EACH_NESTED_SUBJECT_END(s) + if ((r->subj_hash_size * sizeof (struct acl_subject_label *)) <= PAGE_SIZE) + kfree(r->subj_hash); + else + vfree(r->subj_hash); + r->subj_hash = NULL; + FOR_EACH_ROLE_END(r,i) + + acl_free_all(); + + if (acl_role_set.r_hash) { + if ((acl_role_set.r_size * sizeof (struct acl_role_label *)) <= + PAGE_SIZE) + kfree(acl_role_set.r_hash); + else + vfree(acl_role_set.r_hash); + } + if (name_set.n_hash) { + if ((name_set.n_size * sizeof (struct name_entry *)) <= + PAGE_SIZE) + kfree(name_set.n_hash); + else + vfree(name_set.n_hash); + } + + if (inodev_set.i_hash) { + if ((inodev_set.i_size * sizeof (struct inodev_entry *)) <= + PAGE_SIZE) + kfree(inodev_set.i_hash); + else + vfree(inodev_set.i_hash); + } + + gr_free_uidset(); + + memset(&name_set, 0, sizeof (struct name_db)); + memset(&inodev_set, 0, sizeof (struct inodev_db)); + memset(&acl_role_set, 0, sizeof (struct acl_role_db)); + memset(&subj_map_set, 0, sizeof (struct acl_subj_map_db)); + + default_role = NULL; + + return; +} + +static __u32 +count_user_objs(struct acl_object_label *userp) +{ + struct acl_object_label o_tmp; + __u32 num = 0; + + while (userp) { + if (copy_from_user(&o_tmp, userp, + sizeof (struct acl_object_label))) + break; + + userp = o_tmp.prev; + num++; + } + + return num; +} + +static struct acl_subject_label * +do_copy_user_subj(struct acl_subject_label *userp, struct acl_role_label *role); + +static int +copy_user_glob(struct acl_object_label *obj) +{ + struct acl_object_label *g_tmp, **guser; + unsigned int len; + char *tmp; + + if (obj->globbed == NULL) + return 0; + + guser = &obj->globbed; + while (*guser) { + g_tmp = (struct acl_object_label *) + acl_alloc(sizeof (struct acl_object_label)); + if (g_tmp == NULL) + return -ENOMEM; + + if (copy_from_user(g_tmp, *guser, + sizeof (struct acl_object_label))) + return -EFAULT; + + len = strnlen_user(g_tmp->filename, PATH_MAX); + + if (!len || len >= PATH_MAX) + return -EINVAL; + + if ((tmp = (char *) acl_alloc(len)) == NULL) + return -ENOMEM; + + if (copy_from_user(tmp, g_tmp->filename, len)) + return -EFAULT; + + g_tmp->filename = tmp; + + *guser = g_tmp; + guser = &(g_tmp->next); + } + + return 0; +} + +static int +copy_user_objs(struct acl_object_label *userp, struct acl_subject_label *subj, + struct acl_role_label *role) +{ + struct acl_object_label *o_tmp; + unsigned int len; + int ret; + char *tmp; + + while (userp) { + if ((o_tmp = (struct acl_object_label *) + acl_alloc(sizeof (struct acl_object_label))) == NULL) + return -ENOMEM; + + if (copy_from_user(o_tmp, userp, + sizeof (struct acl_object_label))) + return -EFAULT; + + userp = o_tmp->prev; + + len = strnlen_user(o_tmp->filename, PATH_MAX); + + if (!len || len >= PATH_MAX) + return -EINVAL; + + if ((tmp = (char *) acl_alloc(len)) == NULL) + return -ENOMEM; + + if (copy_from_user(tmp, o_tmp->filename, len)) + return -EFAULT; + + o_tmp->filename = tmp; + + insert_acl_obj_label(o_tmp, subj); + if (!insert_name_entry(o_tmp->filename, o_tmp->inode, + o_tmp->device)) + return -ENOMEM; + + ret = copy_user_glob(o_tmp); + if (ret) + return ret; + + if (o_tmp->nested) { + o_tmp->nested = do_copy_user_subj(o_tmp->nested, role); + if (IS_ERR(o_tmp->nested)) + return PTR_ERR(o_tmp->nested); + + /* insert into nested subject list */ + o_tmp->nested->next = role->hash->first; + role->hash->first = o_tmp->nested; + } + } + + return 0; +} + +static __u32 +count_user_subjs(struct acl_subject_label *userp) +{ + struct acl_subject_label s_tmp; + __u32 num = 0; + + while (userp) { + if (copy_from_user(&s_tmp, userp, + sizeof (struct acl_subject_label))) + break; + + userp = s_tmp.prev; + /* do not count nested subjects against this count, since + they are not included in the hash table, but are + attached to objects. We have already counted + the subjects in userspace for the allocation + stack + */ + if (!(s_tmp.mode & GR_NESTED)) + num++; + } + + return num; +} + +static int +copy_user_allowedips(struct acl_role_label *rolep) +{ + struct role_allowed_ip *ruserip, *rtmp = NULL, *rlast; + + ruserip = rolep->allowed_ips; + + while (ruserip) { + rlast = rtmp; + + if ((rtmp = (struct role_allowed_ip *) + acl_alloc(sizeof (struct role_allowed_ip))) == NULL) + return -ENOMEM; + + if (copy_from_user(rtmp, ruserip, + sizeof (struct role_allowed_ip))) + return -EFAULT; + + ruserip = rtmp->prev; + + if (!rlast) { + rtmp->prev = NULL; + rolep->allowed_ips = rtmp; + } else { + rlast->next = rtmp; + rtmp->prev = rlast; + } + + if (!ruserip) + rtmp->next = NULL; + } + + return 0; +} + +static int +copy_user_transitions(struct acl_role_label *rolep) +{ + struct role_transition *rusertp, *rtmp = NULL, *rlast; + + unsigned int len; + char *tmp; + + rusertp = rolep->transitions; + + while (rusertp) { + rlast = rtmp; + + if ((rtmp = (struct role_transition *) + acl_alloc(sizeof (struct role_transition))) == NULL) + return -ENOMEM; + + if (copy_from_user(rtmp, rusertp, + sizeof (struct role_transition))) + return -EFAULT; + + rusertp = rtmp->prev; + + len = strnlen_user(rtmp->rolename, GR_SPROLE_LEN); + + if (!len || len >= GR_SPROLE_LEN) + return -EINVAL; + + if ((tmp = (char *) acl_alloc(len)) == NULL) + return -ENOMEM; + + if (copy_from_user(tmp, rtmp->rolename, len)) + return -EFAULT; + + rtmp->rolename = tmp; + + if (!rlast) { + rtmp->prev = NULL; + rolep->transitions = rtmp; + } else { + rlast->next = rtmp; + rtmp->prev = rlast; + } + + if (!rusertp) + rtmp->next = NULL; + } + + return 0; +} + +static struct acl_subject_label * +do_copy_user_subj(struct acl_subject_label *userp, struct acl_role_label *role) +{ + struct acl_subject_label *s_tmp = NULL, *s_tmp2; + unsigned int len; + char *tmp; + __u32 num_objs; + struct acl_ip_label **i_tmp, *i_utmp2; + struct gr_hash_struct ghash; + struct subject_map *subjmap; + unsigned int i_num; + int err; + + s_tmp = lookup_subject_map(userp); + + /* we've already copied this subject into the kernel, just return + the reference to it, and don't copy it over again + */ + if (s_tmp) + return(s_tmp); + + if ((s_tmp = (struct acl_subject_label *) + acl_alloc(sizeof (struct acl_subject_label))) == NULL) + return ERR_PTR(-ENOMEM); + + subjmap = (struct subject_map *)kmalloc(sizeof (struct subject_map), GFP_KERNEL); + if (subjmap == NULL) + return ERR_PTR(-ENOMEM); + + subjmap->user = userp; + subjmap->kernel = s_tmp; + insert_subj_map_entry(subjmap); + + if (copy_from_user(s_tmp, userp, + sizeof (struct acl_subject_label))) + return ERR_PTR(-EFAULT); + + len = strnlen_user(s_tmp->filename, PATH_MAX); + + if (!len || len >= PATH_MAX) + return ERR_PTR(-EINVAL); + + if ((tmp = (char *) acl_alloc(len)) == NULL) + return ERR_PTR(-ENOMEM); + + if (copy_from_user(tmp, s_tmp->filename, len)) + return ERR_PTR(-EFAULT); + + s_tmp->filename = tmp; + + if (!strcmp(s_tmp->filename, "/")) + role->root_label = s_tmp; + + if (copy_from_user(&ghash, s_tmp->hash, sizeof(struct gr_hash_struct))) + return ERR_PTR(-EFAULT); + + /* copy user and group transition tables */ + + if (s_tmp->user_trans_num) { + uid_t *uidlist; + + uidlist = (uid_t *)acl_alloc(s_tmp->user_trans_num * sizeof(uid_t)); + if (uidlist == NULL) + return ERR_PTR(-ENOMEM); + if (copy_from_user(uidlist, s_tmp->user_transitions, s_tmp->user_trans_num * sizeof(uid_t))) + return ERR_PTR(-EFAULT); + + s_tmp->user_transitions = uidlist; + } + + if (s_tmp->group_trans_num) { + gid_t *gidlist; + + gidlist = (gid_t *)acl_alloc(s_tmp->group_trans_num * sizeof(gid_t)); + if (gidlist == NULL) + return ERR_PTR(-ENOMEM); + if (copy_from_user(gidlist, s_tmp->group_transitions, s_tmp->group_trans_num * sizeof(gid_t))) + return ERR_PTR(-EFAULT); + + s_tmp->group_transitions = gidlist; + } + + /* set up object hash table */ + num_objs = count_user_objs(ghash.first); + + s_tmp->obj_hash_size = num_objs; + s_tmp->obj_hash = + (struct acl_object_label **) + create_table(&(s_tmp->obj_hash_size), sizeof(void *)); + + if (!s_tmp->obj_hash) + return ERR_PTR(-ENOMEM); + + memset(s_tmp->obj_hash, 0, + s_tmp->obj_hash_size * + sizeof (struct acl_object_label *)); + + /* add in objects */ + err = copy_user_objs(ghash.first, s_tmp, role); + + if (err) + return ERR_PTR(err); + + /* set pointer for parent subject */ + if (s_tmp->parent_subject) { + s_tmp2 = do_copy_user_subj(s_tmp->parent_subject, role); + + if (IS_ERR(s_tmp2)) + return s_tmp2; + + s_tmp->parent_subject = s_tmp2; + } + + /* add in ip acls */ + + if (!s_tmp->ip_num) { + s_tmp->ips = NULL; + goto insert; + } + + i_tmp = + (struct acl_ip_label **) acl_alloc(s_tmp->ip_num * + sizeof (struct + acl_ip_label *)); + + if (!i_tmp) + return ERR_PTR(-ENOMEM); + + for (i_num = 0; i_num < s_tmp->ip_num; i_num++) { + *(i_tmp + i_num) = + (struct acl_ip_label *) + acl_alloc(sizeof (struct acl_ip_label)); + if (!*(i_tmp + i_num)) + return ERR_PTR(-ENOMEM); + + if (copy_from_user + (&i_utmp2, s_tmp->ips + i_num, + sizeof (struct acl_ip_label *))) + return ERR_PTR(-EFAULT); + + if (copy_from_user + (*(i_tmp + i_num), i_utmp2, + sizeof (struct acl_ip_label))) + return ERR_PTR(-EFAULT); + + if ((*(i_tmp + i_num))->iface == NULL) + continue; + + len = strnlen_user((*(i_tmp + i_num))->iface, IFNAMSIZ); + if (!len || len >= IFNAMSIZ) + return ERR_PTR(-EINVAL); + tmp = acl_alloc(len); + if (tmp == NULL) + return ERR_PTR(-ENOMEM); + if (copy_from_user(tmp, (*(i_tmp + i_num))->iface, len)) + return ERR_PTR(-EFAULT); + (*(i_tmp + i_num))->iface = tmp; + } + + s_tmp->ips = i_tmp; + +insert: + if (!insert_name_entry(s_tmp->filename, s_tmp->inode, + s_tmp->device)) + return ERR_PTR(-ENOMEM); + + return s_tmp; +} + +static int +copy_user_subjs(struct acl_subject_label *userp, struct acl_role_label *role) +{ + struct acl_subject_label s_pre; + struct acl_subject_label * ret; + int err; + + while (userp) { + if (copy_from_user(&s_pre, userp, + sizeof (struct acl_subject_label))) + return -EFAULT; + + /* do not add nested subjects here, add + while parsing objects + */ + + if (s_pre.mode & GR_NESTED) { + userp = s_pre.prev; + continue; + } + + ret = do_copy_user_subj(userp, role); + + err = PTR_ERR(ret); + if (IS_ERR(ret)) + return err; + + insert_acl_subj_label(ret, role); + + userp = s_pre.prev; + } + + return 0; +} + +static int +copy_user_acl(struct gr_arg *arg) +{ + struct acl_role_label *r_tmp = NULL, **r_utmp, *r_utmp2; + struct sprole_pw *sptmp; + struct gr_hash_struct *ghash; + uid_t *domainlist; + unsigned int r_num; + unsigned int len; + char *tmp; + int err = 0; + __u16 i; + __u32 num_subjs; + + /* we need a default and kernel role */ + if (arg->role_db.num_roles < 2) + return -EINVAL; + + /* copy special role authentication info from userspace */ + + num_sprole_pws = arg->num_sprole_pws; + acl_special_roles = (struct sprole_pw **) acl_alloc(num_sprole_pws * sizeof(struct sprole_pw *)); + + if (!acl_special_roles) { + err = -ENOMEM; + goto cleanup; + } + + for (i = 0; i < num_sprole_pws; i++) { + sptmp = (struct sprole_pw *) acl_alloc(sizeof(struct sprole_pw)); + if (!sptmp) { + err = -ENOMEM; + goto cleanup; + } + if (copy_from_user(sptmp, arg->sprole_pws + i, + sizeof (struct sprole_pw))) { + err = -EFAULT; + goto cleanup; + } + + len = + strnlen_user(sptmp->rolename, GR_SPROLE_LEN); + + if (!len || len >= GR_SPROLE_LEN) { + err = -EINVAL; + goto cleanup; + } + + if ((tmp = (char *) acl_alloc(len)) == NULL) { + err = -ENOMEM; + goto cleanup; + } + + if (copy_from_user(tmp, sptmp->rolename, len)) { + err = -EFAULT; + goto cleanup; + } + +#ifdef CONFIG_GRKERNSEC_ACL_DEBUG + printk(KERN_ALERT "Copying special role %s\n", tmp); +#endif + sptmp->rolename = tmp; + acl_special_roles[i] = sptmp; + } + + r_utmp = (struct acl_role_label **) arg->role_db.r_table; + + for (r_num = 0; r_num < arg->role_db.num_roles; r_num++) { + r_tmp = acl_alloc(sizeof (struct acl_role_label)); + + if (!r_tmp) { + err = -ENOMEM; + goto cleanup; + } + + if (copy_from_user(&r_utmp2, r_utmp + r_num, + sizeof (struct acl_role_label *))) { + err = -EFAULT; + goto cleanup; + } + + if (copy_from_user(r_tmp, r_utmp2, + sizeof (struct acl_role_label))) { + err = -EFAULT; + goto cleanup; + } + + len = strnlen_user(r_tmp->rolename, GR_SPROLE_LEN); + + if (!len || len >= PATH_MAX) { + err = -EINVAL; + goto cleanup; + } + + if ((tmp = (char *) acl_alloc(len)) == NULL) { + err = -ENOMEM; + goto cleanup; + } + if (copy_from_user(tmp, r_tmp->rolename, len)) { + err = -EFAULT; + goto cleanup; + } + r_tmp->rolename = tmp; + + if (!strcmp(r_tmp->rolename, "default") + && (r_tmp->roletype & GR_ROLE_DEFAULT)) { + default_role = r_tmp; + } else if (!strcmp(r_tmp->rolename, ":::kernel:::")) { + kernel_role = r_tmp; + } + + if ((ghash = (struct gr_hash_struct *) acl_alloc(sizeof(struct gr_hash_struct))) == NULL) { + err = -ENOMEM; + goto cleanup; + } + if (copy_from_user(ghash, r_tmp->hash, sizeof(struct gr_hash_struct))) { + err = -EFAULT; + goto cleanup; + } + + r_tmp->hash = ghash; + + num_subjs = count_user_subjs(r_tmp->hash->first); + + r_tmp->subj_hash_size = num_subjs; + r_tmp->subj_hash = + (struct acl_subject_label **) + create_table(&(r_tmp->subj_hash_size), sizeof(void *)); + + if (!r_tmp->subj_hash) { + err = -ENOMEM; + goto cleanup; + } + + err = copy_user_allowedips(r_tmp); + if (err) + goto cleanup; + + /* copy domain info */ + if (r_tmp->domain_children != NULL) { + domainlist = acl_alloc(r_tmp->domain_child_num * sizeof(uid_t)); + if (domainlist == NULL) { + err = -ENOMEM; + goto cleanup; + } + if (copy_from_user(domainlist, r_tmp->domain_children, r_tmp->domain_child_num * sizeof(uid_t))) { + err = -EFAULT; + goto cleanup; + } + r_tmp->domain_children = domainlist; + } + + err = copy_user_transitions(r_tmp); + if (err) + goto cleanup; + + memset(r_tmp->subj_hash, 0, + r_tmp->subj_hash_size * + sizeof (struct acl_subject_label *)); + + err = copy_user_subjs(r_tmp->hash->first, r_tmp); + + if (err) + goto cleanup; + + /* set nested subject list to null */ + r_tmp->hash->first = NULL; + + insert_acl_role_label(r_tmp); + } + + goto return_err; + cleanup: + free_variables(); + return_err: + return err; + +} + +static int +gracl_init(struct gr_arg *args) +{ + int error = 0; + + memcpy(gr_system_salt, args->salt, GR_SALT_LEN); + memcpy(gr_system_sum, args->sum, GR_SHA_LEN); + + if (init_variables(args)) { + gr_log_str(GR_DONT_AUDIT_GOOD, GR_INITF_ACL_MSG, GR_VERSION); + error = -ENOMEM; + free_variables(); + goto out; + } + + error = copy_user_acl(args); + free_init_variables(); + if (error) { + free_variables(); + goto out; + } + + if ((error = gr_set_acls(0))) { + free_variables(); + goto out; + } + + gr_status |= GR_READY; + out: + return error; +} + +/* derived from glibc fnmatch() 0: match, 1: no match*/ + +static int +glob_match(const char *p, const char *n) +{ + char c; + + while ((c = *p++) != '\0') { + switch (c) { + case '?': + if (*n == '\0') + return 1; + else if (*n == '/') + return 1; + break; + case '\\': + if (*n != c) + return 1; + break; + case '*': + for (c = *p++; c == '?' || c == '*'; c = *p++) { + if (*n == '/') + return 1; + else if (c == '?') { + if (*n == '\0') + return 1; + else + ++n; + } + } + if (c == '\0') { + return 0; + } else { + const char *endp; + + if ((endp = strchr(n, '/')) == NULL) + endp = n + strlen(n); + + if (c == '[') { + for (--p; n < endp; ++n) + if (!glob_match(p, n)) + return 0; + } else if (c == '/') { + while (*n != '\0' && *n != '/') + ++n; + if (*n == '/' && !glob_match(p, n + 1)) + return 0; + } else { + for (--p; n < endp; ++n) + if (*n == c && !glob_match(p, n)) + return 0; + } + + return 1; + } + case '[': + { + int not; + char cold; + + if (*n == '\0' || *n == '/') + return 1; + + not = (*p == '!' || *p == '^'); + if (not) + ++p; + + c = *p++; + for (;;) { + unsigned char fn = (unsigned char)*n; + + if (c == '\0') + return 1; + else { + if (c == fn) + goto matched; + cold = c; + c = *p++; + + if (c == '-' && *p != ']') { + unsigned char cend = *p++; + + if (cend == '\0') + return 1; + + if (cold <= fn && fn <= cend) + goto matched; + + c = *p++; + } + } + + if (c == ']') + break; + } + if (!not) + return 1; + break; + matched: + while (c != ']') { + if (c == '\0') + return 1; + + c = *p++; + } + if (not) + return 1; + } + break; + default: + if (c != *n) + return 1; + } + + ++n; + } + + if (*n == '\0') + return 0; + + if (*n == '/') + return 0; + + return 1; +} + +static struct acl_object_label * +chk_glob_label(struct acl_object_label *globbed, + struct dentry *dentry, struct vfsmount *mnt, char **path) +{ + struct acl_object_label *tmp; + + if (*path == NULL) + *path = gr_to_filename_nolock(dentry, mnt); + + tmp = globbed; + + while (tmp) { + if (!glob_match(tmp->filename, *path)) + return tmp; + tmp = tmp->next; + } + + return NULL; +} + +static struct acl_object_label * +__full_lookup(const struct dentry *orig_dentry, const struct vfsmount *orig_mnt, + const ino_t curr_ino, const dev_t curr_dev, + const struct acl_subject_label *subj, char **path) +{ + struct acl_subject_label *tmpsubj; + struct acl_object_label *retval; + struct acl_object_label *retval2; + + tmpsubj = (struct acl_subject_label *) subj; + read_lock(&gr_inode_lock); + do { + retval = lookup_acl_obj_label(curr_ino, curr_dev, tmpsubj); + if (retval) { + if (retval->globbed) { + retval2 = chk_glob_label(retval->globbed, (struct dentry *)orig_dentry, + (struct vfsmount *)orig_mnt, path); + if (retval2) + retval = retval2; + } + break; + } + } while ((tmpsubj = tmpsubj->parent_subject)); + read_unlock(&gr_inode_lock); + + return retval; +} + +static __inline__ struct acl_object_label * +full_lookup(const struct dentry *orig_dentry, const struct vfsmount *orig_mnt, + const struct dentry *curr_dentry, + const struct acl_subject_label *subj, char **path) +{ + return __full_lookup(orig_dentry, orig_mnt, + curr_dentry->d_inode->i_ino, + curr_dentry->d_inode->i_sb->s_dev, subj, path); +} + +static struct acl_object_label * +__chk_obj_label(const struct dentry *l_dentry, const struct vfsmount *l_mnt, + const struct acl_subject_label *subj, char *path) +{ + struct dentry *dentry = (struct dentry *) l_dentry; + struct vfsmount *mnt = (struct vfsmount *) l_mnt; + struct acl_object_label *retval; + + spin_lock(&dcache_lock); + + if (unlikely(mnt == shm_mnt || mnt == pipe_mnt || mnt == sock_mnt)) { + retval = fakefs_obj; + goto out; + } + + for (;;) { + if (dentry == real_root && mnt == real_root_mnt) + break; + + if (dentry == mnt->mnt_root || IS_ROOT(dentry)) { + if (mnt->mnt_parent == mnt) + break; + + retval = full_lookup(l_dentry, l_mnt, dentry, subj, &path); + if (retval != NULL) + goto out; + + dentry = mnt->mnt_mountpoint; + mnt = mnt->mnt_parent; + continue; + } + + retval = full_lookup(l_dentry, l_mnt, dentry, subj, &path); + if (retval != NULL) + goto out; + + dentry = dentry->d_parent; + } + + retval = full_lookup(l_dentry, l_mnt, dentry, subj, &path); + + if (retval == NULL) + retval = full_lookup(l_dentry, l_mnt, real_root, subj, &path); +out: + spin_unlock(&dcache_lock); + return retval; +} + +static __inline__ struct acl_object_label * +chk_obj_label(const struct dentry *l_dentry, const struct vfsmount *l_mnt, + const struct acl_subject_label *subj) +{ + char *path = NULL; + return __chk_obj_label(l_dentry, l_mnt, subj, path); +} + +static __inline__ struct acl_object_label * +chk_obj_create_label(const struct dentry *l_dentry, const struct vfsmount *l_mnt, + const struct acl_subject_label *subj, char *path) +{ + return __chk_obj_label(l_dentry, l_mnt, subj, path); +} + +static struct acl_subject_label * +chk_subj_label(const struct dentry *l_dentry, const struct vfsmount *l_mnt, + const struct acl_role_label *role) +{ + struct dentry *dentry = (struct dentry *) l_dentry; + struct vfsmount *mnt = (struct vfsmount *) l_mnt; + struct acl_subject_label *retval; + + spin_lock(&dcache_lock); + + for (;;) { + if (dentry == real_root && mnt == real_root_mnt) + break; + if (dentry == mnt->mnt_root || IS_ROOT(dentry)) { + if (mnt->mnt_parent == mnt) + break; + + read_lock(&gr_inode_lock); + retval = + lookup_acl_subj_label(dentry->d_inode->i_ino, + dentry->d_inode->i_sb->s_dev, role); + read_unlock(&gr_inode_lock); + if (retval != NULL) + goto out; + + dentry = mnt->mnt_mountpoint; + mnt = mnt->mnt_parent; + continue; + } + + read_lock(&gr_inode_lock); + retval = lookup_acl_subj_label(dentry->d_inode->i_ino, + dentry->d_inode->i_sb->s_dev, role); + read_unlock(&gr_inode_lock); + if (retval != NULL) + goto out; + + dentry = dentry->d_parent; + } + + read_lock(&gr_inode_lock); + retval = lookup_acl_subj_label(dentry->d_inode->i_ino, + dentry->d_inode->i_sb->s_dev, role); + read_unlock(&gr_inode_lock); + + if (unlikely(retval == NULL)) { + read_lock(&gr_inode_lock); + retval = lookup_acl_subj_label(real_root->d_inode->i_ino, + real_root->d_inode->i_sb->s_dev, role); + read_unlock(&gr_inode_lock); + } +out: + spin_unlock(&dcache_lock); + + return retval; +} + +static void +gr_log_learn(const struct task_struct *task, const struct dentry *dentry, const struct vfsmount *mnt, const __u32 mode) +{ + security_learn(GR_LEARN_AUDIT_MSG, task->role->rolename, task->role->roletype, + task->uid, task->gid, task->exec_file ? gr_to_filename1(task->exec_file->f_dentry, + task->exec_file->f_vfsmnt) : task->acl->filename, task->acl->filename, + 1, 1, gr_to_filename(dentry, mnt), (unsigned long) mode, NIPQUAD(task->signal->curr_ip)); + + return; +} + +static void +gr_log_learn_id_change(const struct task_struct *task, const char type, const unsigned int real, + const unsigned int effective, const unsigned int fs) +{ + security_learn(GR_ID_LEARN_MSG, task->role->rolename, task->role->roletype, + task->uid, task->gid, task->exec_file ? gr_to_filename1(task->exec_file->f_dentry, + task->exec_file->f_vfsmnt) : task->acl->filename, task->acl->filename, + type, real, effective, fs, NIPQUAD(task->signal->curr_ip)); + + return; +} + +__u32 +gr_check_link(const struct dentry * new_dentry, + const struct dentry * parent_dentry, + const struct vfsmount * parent_mnt, + const struct dentry * old_dentry, const struct vfsmount * old_mnt) +{ + struct acl_object_label *obj; + __u32 oldmode, newmode; + __u32 needmode; + + if (unlikely(!(gr_status & GR_READY))) + return (GR_CREATE | GR_LINK); + + obj = chk_obj_label(old_dentry, old_mnt, current->acl); + oldmode = obj->mode; + + if (current->acl->mode & (GR_LEARN | GR_INHERITLEARN)) + oldmode |= (GR_CREATE | GR_LINK); + + needmode = GR_CREATE | GR_AUDIT_CREATE | GR_SUPPRESS; + if (old_dentry->d_inode->i_mode & (S_ISUID | S_ISGID)) + needmode |= GR_SETID | GR_AUDIT_SETID; + + newmode = + gr_check_create(new_dentry, parent_dentry, parent_mnt, + oldmode | needmode); + + needmode = newmode & (GR_FIND | GR_APPEND | GR_WRITE | GR_EXEC | + GR_SETID | GR_READ | GR_FIND | GR_DELETE | + GR_INHERIT | GR_AUDIT_INHERIT); + + if (old_dentry->d_inode->i_mode & (S_ISUID | S_ISGID) && !(newmode & GR_SETID)) + goto bad; + + if ((oldmode & needmode) != needmode) + goto bad; + + needmode = oldmode & (GR_NOPTRACE | GR_PTRACERD | GR_INHERIT | GR_AUDITS); + if ((newmode & needmode) != needmode) + goto bad; + + if ((newmode & (GR_CREATE | GR_LINK)) == (GR_CREATE | GR_LINK)) + return newmode; +bad: + needmode = oldmode; + if (old_dentry->d_inode->i_mode & (S_ISUID | S_ISGID)) + needmode |= GR_SETID; + + if (current->acl->mode & (GR_LEARN | GR_INHERITLEARN)) { + gr_log_learn(current, old_dentry, old_mnt, needmode); + return (GR_CREATE | GR_LINK); + } else if (newmode & GR_SUPPRESS) + return GR_SUPPRESS; + else + return 0; +} + +__u32 +gr_search_file(const struct dentry * dentry, const __u32 mode, + const struct vfsmount * mnt) +{ + __u32 retval = mode; + struct acl_subject_label *curracl; + struct acl_object_label *currobj; + + if (unlikely(!(gr_status & GR_READY))) + return (mode & ~GR_AUDITS); + + curracl = current->acl; + + currobj = chk_obj_label(dentry, mnt, curracl); + retval = currobj->mode & mode; + + if (unlikely + ((curracl->mode & (GR_LEARN | GR_INHERITLEARN)) && !(mode & GR_NOPTRACE) + && (retval != (mode & ~(GR_AUDITS | GR_SUPPRESS))))) { + __u32 new_mode = mode; + + new_mode &= ~(GR_AUDITS | GR_SUPPRESS); + + retval = new_mode; + + if (new_mode & GR_EXEC && curracl->mode & GR_INHERITLEARN) + new_mode |= GR_INHERIT; + + if (!(mode & GR_NOLEARN)) + gr_log_learn(current, dentry, mnt, new_mode); + } + + return retval; +} + +__u32 +gr_check_create(const struct dentry * new_dentry, const struct dentry * parent, + const struct vfsmount * mnt, const __u32 mode) +{ + struct name_entry *match; + struct acl_object_label *matchpo; + struct acl_subject_label *curracl; + char *path; + __u32 retval; + + if (unlikely(!(gr_status & GR_READY))) + return (mode & ~GR_AUDITS); + + preempt_disable(); + path = gr_to_filename_rbac(new_dentry, mnt); + match = lookup_name_entry(path); + + if (!match) + goto check_parent; + + curracl = current->acl; + + read_lock(&gr_inode_lock); + matchpo = lookup_acl_obj_label_create(match->inode, match->device, curracl); + read_unlock(&gr_inode_lock); + + if (matchpo) { + if ((matchpo->mode & mode) != + (mode & ~(GR_AUDITS | GR_SUPPRESS)) + && curracl->mode & (GR_LEARN | GR_INHERITLEARN)) { + __u32 new_mode = mode; + + new_mode &= ~(GR_AUDITS | GR_SUPPRESS); + + gr_log_learn(current, new_dentry, mnt, new_mode); + + preempt_enable(); + return new_mode; + } + preempt_enable(); + return (matchpo->mode & mode); + } + + check_parent: + curracl = current->acl; + + matchpo = chk_obj_create_label(parent, mnt, curracl, path); + retval = matchpo->mode & mode; + + if ((retval != (mode & ~(GR_AUDITS | GR_SUPPRESS))) + && (curracl->mode & (GR_LEARN | GR_INHERITLEARN))) { + __u32 new_mode = mode; + + new_mode &= ~(GR_AUDITS | GR_SUPPRESS); + + gr_log_learn(current, new_dentry, mnt, new_mode); + preempt_enable(); + return new_mode; + } + + preempt_enable(); + return retval; +} + +int +gr_check_hidden_task(const struct task_struct *task) +{ + if (unlikely(!(gr_status & GR_READY))) + return 0; + + if (!(task->acl->mode & GR_PROCFIND) && !(current->acl->mode & GR_VIEW)) + return 1; + + return 0; +} + +int +gr_check_protected_task(const struct task_struct *task) +{ + if (unlikely(!(gr_status & GR_READY) || !task)) + return 0; + + if ((task->acl->mode & GR_PROTECTED) && !(current->acl->mode & GR_KILL) && + task->acl != current->acl) + return 1; + + return 0; +} + +void +gr_copy_label(struct task_struct *tsk) +{ + tsk->signal->used_accept = 0; + tsk->acl_sp_role = 0; + tsk->acl_role_id = current->acl_role_id; + tsk->acl = current->acl; + tsk->role = current->role; + tsk->signal->curr_ip = current->signal->curr_ip; + if (current->exec_file) + get_file(current->exec_file); + tsk->exec_file = current->exec_file; + tsk->is_writable = current->is_writable; + if (unlikely(current->signal->used_accept)) + current->signal->curr_ip = 0; + + return; +} + +static void +gr_set_proc_res(struct task_struct *task) +{ + struct acl_subject_label *proc; + unsigned short i; + + proc = task->acl; + + if (proc->mode & (GR_LEARN | GR_INHERITLEARN)) + return; + + for (i = 0; i < (GR_NLIMITS - 1); i++) { + if (!(proc->resmask & (1 << i))) + continue; + + task->signal->rlim[i].rlim_cur = proc->res[i].rlim_cur; + task->signal->rlim[i].rlim_max = proc->res[i].rlim_max; + } + + return; +} + +int +gr_check_user_change(int real, int effective, int fs) +{ + unsigned int i; + __u16 num; + uid_t *uidlist; + int curuid; + int realok = 0; + int effectiveok = 0; + int fsok = 0; + + if (unlikely(!(gr_status & GR_READY))) + return 0; + + if (current->acl->mode & (GR_LEARN | GR_INHERITLEARN)) + gr_log_learn_id_change(current, 'u', real, effective, fs); + + num = current->acl->user_trans_num; + uidlist = current->acl->user_transitions; + + if (uidlist == NULL) + return 0; + + if (real == -1) + realok = 1; + if (effective == -1) + effectiveok = 1; + if (fs == -1) + fsok = 1; + + if (current->acl->user_trans_type & GR_ID_ALLOW) { + for (i = 0; i < num; i++) { + curuid = (int)uidlist[i]; + if (real == curuid) + realok = 1; + if (effective == curuid) + effectiveok = 1; + if (fs == curuid) + fsok = 1; + } + } else if (current->acl->user_trans_type & GR_ID_DENY) { + for (i = 0; i < num; i++) { + curuid = (int)uidlist[i]; + if (real == curuid) + break; + if (effective == curuid) + break; + if (fs == curuid) + break; + } + /* not in deny list */ + if (i == num) { + realok = 1; + effectiveok = 1; + fsok = 1; + } + } + + if (realok && effectiveok && fsok) + return 0; + else { + gr_log_int(GR_DONT_AUDIT, GR_USRCHANGE_ACL_MSG, realok ? (effectiveok ? (fsok ? 0 : fs) : effective) : real); + return 1; + } +} + +int +gr_check_group_change(int real, int effective, int fs) +{ + unsigned int i; + __u16 num; + gid_t *gidlist; + int curgid; + int realok = 0; + int effectiveok = 0; + int fsok = 0; + + if (unlikely(!(gr_status & GR_READY))) + return 0; + + if (current->acl->mode & (GR_LEARN | GR_INHERITLEARN)) + gr_log_learn_id_change(current, 'g', real, effective, fs); + + num = current->acl->group_trans_num; + gidlist = current->acl->group_transitions; + + if (gidlist == NULL) + return 0; + + if (real == -1) + realok = 1; + if (effective == -1) + effectiveok = 1; + if (fs == -1) + fsok = 1; + + if (current->acl->group_trans_type & GR_ID_ALLOW) { + for (i = 0; i < num; i++) { + curgid = (int)gidlist[i]; + if (real == curgid) + realok = 1; + if (effective == curgid) + effectiveok = 1; + if (fs == curgid) + fsok = 1; + } + } else if (current->acl->group_trans_type & GR_ID_DENY) { + for (i = 0; i < num; i++) { + curgid = (int)gidlist[i]; + if (real == curgid) + break; + if (effective == curgid) + break; + if (fs == curgid) + break; + } + /* not in deny list */ + if (i == num) { + realok = 1; + effectiveok = 1; + fsok = 1; + } + } + + if (realok && effectiveok && fsok) + return 0; + else { + gr_log_int(GR_DONT_AUDIT, GR_GRPCHANGE_ACL_MSG, realok ? (effectiveok ? (fsok ? 0 : fs) : effective) : real); + return 1; + } +} + +void +gr_set_role_label(struct task_struct *task, const uid_t uid, const uid_t gid) +{ + struct acl_role_label *role = task->role; + struct acl_subject_label *subj = NULL; + struct acl_object_label *obj; + struct file *filp; + + if (unlikely(!(gr_status & GR_READY))) + return; + + filp = task->exec_file; + + /* kernel process, we'll give them the kernel role */ + if (unlikely(!filp)) { + task->role = kernel_role; + task->acl = kernel_role->root_label; + return; + } else if (!task->role || !(task->role->roletype & GR_ROLE_SPECIAL)) + role = lookup_acl_role_label(task, uid, gid); + + /* perform subject lookup in possibly new role + we can use this result below in the case where role == task->role + */ + subj = chk_subj_label(filp->f_dentry, filp->f_vfsmnt, role); + + /* if we changed uid/gid, but result in the same role + and are using inheritance, don't lose the inherited subject + if current subject is other than what normal lookup + would result in, we arrived via inheritance, don't + lose subject + */ + if (role != task->role || (!(task->acl->mode & GR_INHERITLEARN) && + (subj == task->acl))) + task->acl = subj; + + task->role = role; + + task->is_writable = 0; + + /* ignore additional mmap checks for processes that are writable + by the default ACL */ + obj = chk_obj_label(filp->f_dentry, filp->f_vfsmnt, default_role->root_label); + if (unlikely(obj->mode & GR_WRITE)) + task->is_writable = 1; + obj = chk_obj_label(filp->f_dentry, filp->f_vfsmnt, task->role->root_label); + if (unlikely(obj->mode & GR_WRITE)) + task->is_writable = 1; + +#ifdef CONFIG_GRKERNSEC_ACL_DEBUG + printk(KERN_ALERT "Set role label for (%s:%d): role:%s, subject:%s\n", task->comm, task->pid, task->role->rolename, task->acl->filename); +#endif + + gr_set_proc_res(task); + + return; +} + +int +gr_set_proc_label(const struct dentry *dentry, const struct vfsmount *mnt) +{ + struct task_struct *task = current; + struct acl_subject_label *newacl; + struct acl_object_label *obj; + __u32 retmode; + + if (unlikely(!(gr_status & GR_READY))) + return 0; + + newacl = chk_subj_label(dentry, mnt, task->role); + + task_lock(task); + if (((task->ptrace & PT_PTRACED) && !(task->acl->mode & + GR_POVERRIDE) && (task->acl != newacl) && + !(task->role->roletype & GR_ROLE_GOD) && + !gr_search_file(dentry, GR_PTRACERD, mnt) && + !(task->acl->mode & (GR_LEARN | GR_INHERITLEARN))) || + (atomic_read(&task->fs->count) > 1 || + atomic_read(&task->files->count) > 1 || + atomic_read(&task->sighand->count) > 1)) { + task_unlock(task); + gr_log_fs_generic(GR_DONT_AUDIT, GR_PTRACE_EXEC_ACL_MSG, dentry, mnt); + return -EACCES; + } + task_unlock(task); + + obj = chk_obj_label(dentry, mnt, task->acl); + retmode = obj->mode & (GR_INHERIT | GR_AUDIT_INHERIT); + + if (!(task->acl->mode & GR_INHERITLEARN) && + ((newacl->mode & GR_LEARN) || !(retmode & GR_INHERIT))) { + if (obj->nested) + task->acl = obj->nested; + else + task->acl = newacl; + } else if (retmode & GR_INHERIT && retmode & GR_AUDIT_INHERIT) + gr_log_str_fs(GR_DO_AUDIT, GR_INHERIT_ACL_MSG, task->acl->filename, dentry, mnt); + + task->is_writable = 0; + + /* ignore additional mmap checks for processes that are writable + by the default ACL */ + obj = chk_obj_label(dentry, mnt, default_role->root_label); + if (unlikely(obj->mode & GR_WRITE)) + task->is_writable = 1; + obj = chk_obj_label(dentry, mnt, task->role->root_label); + if (unlikely(obj->mode & GR_WRITE)) + task->is_writable = 1; + + gr_set_proc_res(task); + +#ifdef CONFIG_GRKERNSEC_ACL_DEBUG + printk(KERN_ALERT "Set subject label for (%s:%d): role:%s, subject:%s\n", task->comm, task->pid, task->role->rolename, task->acl->filename); +#endif + return 0; +} + +static void +do_handle_delete(const ino_t ino, const dev_t dev) +{ + struct acl_object_label *matchpo; + struct acl_subject_label *matchps; + struct acl_subject_label *subj; + struct acl_role_label *role; + unsigned int i, x; + + FOR_EACH_ROLE_START(role, i) + FOR_EACH_SUBJECT_START(role, subj, x) + if ((matchpo = lookup_acl_obj_label(ino, dev, subj)) != NULL) + matchpo->mode |= GR_DELETED; + FOR_EACH_SUBJECT_END(subj,x) + FOR_EACH_NESTED_SUBJECT_START(role, subj) + if (subj->inode == ino && subj->device == dev) + subj->mode |= GR_DELETED; + FOR_EACH_NESTED_SUBJECT_END(subj) + if ((matchps = lookup_acl_subj_label(ino, dev, role)) != NULL) + matchps->mode |= GR_DELETED; + FOR_EACH_ROLE_END(role,i) + + return; +} + +void +gr_handle_delete(const ino_t ino, const dev_t dev) +{ + if (unlikely(!(gr_status & GR_READY))) + return; + + write_lock(&gr_inode_lock); + if (unlikely((unsigned long)lookup_inodev_entry(ino, dev))) + do_handle_delete(ino, dev); + write_unlock(&gr_inode_lock); + + return; +} + +static void +update_acl_obj_label(const ino_t oldinode, const dev_t olddevice, + const ino_t newinode, const dev_t newdevice, + struct acl_subject_label *subj) +{ + unsigned int index = fhash(oldinode, olddevice, subj->obj_hash_size); + struct acl_object_label *match; + + match = subj->obj_hash[index]; + + while (match && (match->inode != oldinode || + match->device != olddevice || + !(match->mode & GR_DELETED))) + match = match->next; + + if (match && (match->inode == oldinode) + && (match->device == olddevice) + && (match->mode & GR_DELETED)) { + if (match->prev == NULL) { + subj->obj_hash[index] = match->next; + if (match->next != NULL) + match->next->prev = NULL; + } else { + match->prev->next = match->next; + if (match->next != NULL) + match->next->prev = match->prev; + } + match->prev = NULL; + match->next = NULL; + match->inode = newinode; + match->device = newdevice; + match->mode &= ~GR_DELETED; + + insert_acl_obj_label(match, subj); + } + + return; +} + +static void +update_acl_subj_label(const ino_t oldinode, const dev_t olddevice, + const ino_t newinode, const dev_t newdevice, + struct acl_role_label *role) +{ + unsigned int index = fhash(oldinode, olddevice, role->subj_hash_size); + struct acl_subject_label *match; + + match = role->subj_hash[index]; + + while (match && (match->inode != oldinode || + match->device != olddevice || + !(match->mode & GR_DELETED))) + match = match->next; + + if (match && (match->inode == oldinode) + && (match->device == olddevice) + && (match->mode & GR_DELETED)) { + if (match->prev == NULL) { + role->subj_hash[index] = match->next; + if (match->next != NULL) + match->next->prev = NULL; + } else { + match->prev->next = match->next; + if (match->next != NULL) + match->next->prev = match->prev; + } + match->prev = NULL; + match->next = NULL; + match->inode = newinode; + match->device = newdevice; + match->mode &= ~GR_DELETED; + + insert_acl_subj_label(match, role); + } + + return; +} + +static void +update_inodev_entry(const ino_t oldinode, const dev_t olddevice, + const ino_t newinode, const dev_t newdevice) +{ + unsigned int index = fhash(oldinode, olddevice, inodev_set.i_size); + struct inodev_entry *match; + + match = inodev_set.i_hash[index]; + + while (match && (match->nentry->inode != oldinode || + match->nentry->device != olddevice)) + match = match->next; + + if (match && (match->nentry->inode == oldinode) + && (match->nentry->device == olddevice)) { + if (match->prev == NULL) { + inodev_set.i_hash[index] = match->next; + if (match->next != NULL) + match->next->prev = NULL; + } else { + match->prev->next = match->next; + if (match->next != NULL) + match->next->prev = match->prev; + } + match->prev = NULL; + match->next = NULL; + match->nentry->inode = newinode; + match->nentry->device = newdevice; + + insert_inodev_entry(match); + } + + return; +} + +static void +do_handle_create(const struct name_entry *matchn, const struct dentry *dentry, + const struct vfsmount *mnt) +{ + struct acl_subject_label *subj; + struct acl_role_label *role; + unsigned int i, x; + + FOR_EACH_ROLE_START(role, i) + update_acl_subj_label(matchn->inode, matchn->device, + dentry->d_inode->i_ino, + dentry->d_inode->i_sb->s_dev, role); + + FOR_EACH_NESTED_SUBJECT_START(role, subj) + if ((subj->inode == dentry->d_inode->i_ino) && + (subj->device == dentry->d_inode->i_sb->s_dev)) { + subj->inode = dentry->d_inode->i_ino; + subj->device = dentry->d_inode->i_sb->s_dev; + } + FOR_EACH_NESTED_SUBJECT_END(subj) + FOR_EACH_SUBJECT_START(role, subj, x) + update_acl_obj_label(matchn->inode, matchn->device, + dentry->d_inode->i_ino, + dentry->d_inode->i_sb->s_dev, subj); + FOR_EACH_SUBJECT_END(subj,x) + FOR_EACH_ROLE_END(role,i) + + update_inodev_entry(matchn->inode, matchn->device, + dentry->d_inode->i_ino, dentry->d_inode->i_sb->s_dev); + + return; +} + +void +gr_handle_create(const struct dentry *dentry, const struct vfsmount *mnt) +{ + struct name_entry *matchn; + + if (unlikely(!(gr_status & GR_READY))) + return; + + preempt_disable(); + matchn = lookup_name_entry(gr_to_filename_rbac(dentry, mnt)); + + if (unlikely((unsigned long)matchn)) { + write_lock(&gr_inode_lock); + do_handle_create(matchn, dentry, mnt); + write_unlock(&gr_inode_lock); + } + preempt_enable(); + + return; +} + +void +gr_handle_rename(struct inode *old_dir, struct inode *new_dir, + struct dentry *old_dentry, + struct dentry *new_dentry, + struct vfsmount *mnt, const __u8 replace) +{ + struct name_entry *matchn; + + if (unlikely(!(gr_status & GR_READY))) + return; + + preempt_disable(); + matchn = lookup_name_entry(gr_to_filename_rbac(new_dentry, mnt)); + + /* we wouldn't have to check d_inode if it weren't for + NFS silly-renaming + */ + + write_lock(&gr_inode_lock); + if (unlikely(replace && new_dentry->d_inode)) { + if (unlikely(lookup_inodev_entry(new_dentry->d_inode->i_ino, + new_dentry->d_inode->i_sb->s_dev) && + (old_dentry->d_inode->i_nlink <= 1))) + do_handle_delete(new_dentry->d_inode->i_ino, + new_dentry->d_inode->i_sb->s_dev); + } + + if (unlikely(lookup_inodev_entry(old_dentry->d_inode->i_ino, + old_dentry->d_inode->i_sb->s_dev) && + (old_dentry->d_inode->i_nlink <= 1))) + do_handle_delete(old_dentry->d_inode->i_ino, + old_dentry->d_inode->i_sb->s_dev); + + if (unlikely((unsigned long)matchn)) + do_handle_create(matchn, old_dentry, mnt); + + write_unlock(&gr_inode_lock); + preempt_enable(); + + return; +} + +static int +lookup_special_role_auth(__u16 mode, const char *rolename, unsigned char **salt, + unsigned char **sum) +{ + struct acl_role_label *r; + struct role_allowed_ip *ipp; + struct role_transition *trans; + unsigned int i; + int found = 0; + + /* check transition table */ + + for (trans = current->role->transitions; trans; trans = trans->next) { + if (!strcmp(rolename, trans->rolename)) { + found = 1; + break; + } + } + + if (!found) + return 0; + + /* handle special roles that do not require authentication + and check ip */ + + FOR_EACH_ROLE_START(r, i) + if (!strcmp(rolename, r->rolename) && + (r->roletype & GR_ROLE_SPECIAL)) { + found = 0; + if (r->allowed_ips != NULL) { + for (ipp = r->allowed_ips; ipp; ipp = ipp->next) { + if ((ntohl(current->signal->curr_ip) & ipp->netmask) == + (ntohl(ipp->addr) & ipp->netmask)) + found = 1; + } + } else + found = 2; + if (!found) + return 0; + + if (((mode == SPROLE) && (r->roletype & GR_ROLE_NOPW)) || + ((mode == SPROLEPAM) && (r->roletype & GR_ROLE_PAM))) { + *salt = NULL; + *sum = NULL; + return 1; + } + } + FOR_EACH_ROLE_END(r,i) + + for (i = 0; i < num_sprole_pws; i++) { + if (!strcmp(rolename, acl_special_roles[i]->rolename)) { + *salt = acl_special_roles[i]->salt; + *sum = acl_special_roles[i]->sum; + return 1; + } + } + + return 0; +} + +static void +assign_special_role(char *rolename) +{ + struct acl_object_label *obj; + struct acl_role_label *r; + struct acl_role_label *assigned = NULL; + struct task_struct *tsk; + struct file *filp; + unsigned int i; + + FOR_EACH_ROLE_START(r, i) + if (!strcmp(rolename, r->rolename) && + (r->roletype & GR_ROLE_SPECIAL)) + assigned = r; + FOR_EACH_ROLE_END(r,i) + + if (!assigned) + return; + + read_lock(&tasklist_lock); + read_lock(&grsec_exec_file_lock); + + tsk = current->parent; + if (tsk == NULL) + goto out_unlock; + + filp = tsk->exec_file; + if (filp == NULL) + goto out_unlock; + + tsk->is_writable = 0; + + tsk->acl_sp_role = 1; + tsk->acl_role_id = ++acl_sp_role_value; + tsk->role = assigned; + tsk->acl = chk_subj_label(filp->f_dentry, filp->f_vfsmnt, tsk->role); + + /* ignore additional mmap checks for processes that are writable + by the default ACL */ + obj = chk_obj_label(filp->f_dentry, filp->f_vfsmnt, default_role->root_label); + if (unlikely(obj->mode & GR_WRITE)) + tsk->is_writable = 1; + obj = chk_obj_label(filp->f_dentry, filp->f_vfsmnt, tsk->role->root_label); + if (unlikely(obj->mode & GR_WRITE)) + tsk->is_writable = 1; + +#ifdef CONFIG_GRKERNSEC_ACL_DEBUG + printk(KERN_ALERT "Assigning special role:%s subject:%s to process (%s:%d)\n", tsk->role->rolename, tsk->acl->filename, tsk->comm, tsk->pid); +#endif + +out_unlock: + read_unlock(&grsec_exec_file_lock); + read_unlock(&tasklist_lock); + return; +} + +int gr_check_secure_terminal(struct task_struct *task) +{ + struct task_struct *p, *p2, *p3; + struct files_struct *files; + struct fdtable *fdt; + struct file *our_file = NULL, *file; + int i; + + if (task->signal->tty == NULL) + return 1; + + files = get_files_struct(task); + if (files != NULL) { + rcu_read_lock(); + fdt = files_fdtable(files); + for (i=0; i < fdt->max_fds; i++) { + file = fcheck_files(files, i); + if (file && (our_file == NULL) && (file->private_data == task->signal->tty)) { + get_file(file); + our_file = file; + } + } + rcu_read_unlock(); + put_files_struct(files); + } + + if (our_file == NULL) + return 1; + + read_lock(&tasklist_lock); + do_each_thread(p2, p) { + files = get_files_struct(p); + if (files == NULL || + (p->signal && p->signal->tty == task->signal->tty)) { + if (files != NULL) + put_files_struct(files); + continue; + } + rcu_read_lock(); + fdt = files_fdtable(files); + for (i=0; i < fdt->max_fds; i++) { + file = fcheck_files(files, i); + if (file && S_ISCHR(file->f_dentry->d_inode->i_mode) && + file->f_dentry->d_inode->i_rdev == our_file->f_dentry->d_inode->i_rdev) { + p3 = task; + while (p3->pid > 0) { + if (p3 == p) + break; + p3 = p3->parent; + } + if (p3 == p) + break; + gr_log_ttysniff(GR_DONT_AUDIT_GOOD, GR_TTYSNIFF_ACL_MSG, p); + gr_handle_alertkill(p); + rcu_read_unlock(); + put_files_struct(files); + read_unlock(&tasklist_lock); + fput(our_file); + return 0; + } + } + rcu_read_unlock(); + put_files_struct(files); + } while_each_thread(p2, p); + read_unlock(&tasklist_lock); + + fput(our_file); + return 1; +} + +ssize_t +write_grsec_handler(struct file *file, const char * buf, size_t count, loff_t *ppos) +{ + struct gr_arg_wrapper uwrap; + unsigned char *sprole_salt; + unsigned char *sprole_sum; + int error = sizeof (struct gr_arg_wrapper); + int error2 = 0; + + down(&gr_dev_sem); + + if ((gr_status & GR_READY) && !(current->acl->mode & GR_KERNELAUTH)) { + error = -EPERM; + goto out; + } + + if (count != sizeof (struct gr_arg_wrapper)) { + gr_log_int_int(GR_DONT_AUDIT_GOOD, GR_DEV_ACL_MSG, (int)count, (int)sizeof(struct gr_arg_wrapper)); + error = -EINVAL; + goto out; + } + + + if (gr_auth_expires && time_after_eq(get_seconds(), gr_auth_expires)) { + gr_auth_expires = 0; + gr_auth_attempts = 0; + } + + if (copy_from_user(&uwrap, buf, sizeof (struct gr_arg_wrapper))) { + error = -EFAULT; + goto out; + } + + if ((uwrap.version != GRSECURITY_VERSION) || (uwrap.size != sizeof(struct gr_arg))) { + error = -EINVAL; + goto out; + } + + if (copy_from_user(gr_usermode, uwrap.arg, sizeof (struct gr_arg))) { + error = -EFAULT; + goto out; + } + + if (gr_usermode->mode != SPROLE && gr_usermode->mode != SPROLEPAM && + gr_auth_attempts >= CONFIG_GRKERNSEC_ACL_MAXTRIES && + time_after(gr_auth_expires, get_seconds())) { + error = -EBUSY; + goto out; + } + + /* if non-root trying to do anything other than use a special role, + do not attempt authentication, do not count towards authentication + locking + */ + + if (gr_usermode->mode != SPROLE && gr_usermode->mode != STATUS && + gr_usermode->mode != UNSPROLE && gr_usermode->mode != SPROLEPAM && + current->uid) { + error = -EPERM; + goto out; + } + + /* ensure pw and special role name are null terminated */ + + gr_usermode->pw[GR_PW_LEN - 1] = '\0'; + gr_usermode->sp_role[GR_SPROLE_LEN - 1] = '\0'; + + /* Okay. + * We have our enough of the argument structure..(we have yet + * to copy_from_user the tables themselves) . Copy the tables + * only if we need them, i.e. for loading operations. */ + + switch (gr_usermode->mode) { + case STATUS: + if (gr_status & GR_READY) { + error = 1; + if (!gr_check_secure_terminal(current)) + error = 3; + } else + error = 2; + goto out; + case SHUTDOWN: + if ((gr_status & GR_READY) + && !(chkpw(gr_usermode, gr_system_salt, gr_system_sum))) { + gr_status &= ~GR_READY; + gr_log_noargs(GR_DONT_AUDIT_GOOD, GR_SHUTS_ACL_MSG); + free_variables(); + memset(gr_usermode, 0, sizeof (struct gr_arg)); + memset(gr_system_salt, 0, GR_SALT_LEN); + memset(gr_system_sum, 0, GR_SHA_LEN); + } else if (gr_status & GR_READY) { + gr_log_noargs(GR_DONT_AUDIT, GR_SHUTF_ACL_MSG); + error = -EPERM; + } else { + gr_log_noargs(GR_DONT_AUDIT_GOOD, GR_SHUTI_ACL_MSG); + error = -EAGAIN; + } + break; + case ENABLE: + if (!(gr_status & GR_READY) && !(error2 = gracl_init(gr_usermode))) + gr_log_str(GR_DONT_AUDIT_GOOD, GR_ENABLE_ACL_MSG, GR_VERSION); + else { + if (gr_status & GR_READY) + error = -EAGAIN; + else + error = error2; + gr_log_str(GR_DONT_AUDIT, GR_ENABLEF_ACL_MSG, GR_VERSION); + } + break; + case RELOAD: + if (!(gr_status & GR_READY)) { + gr_log_str(GR_DONT_AUDIT_GOOD, GR_RELOADI_ACL_MSG, GR_VERSION); + error = -EAGAIN; + } else if (!(chkpw(gr_usermode, gr_system_salt, gr_system_sum))) { + lock_kernel(); + gr_status &= ~GR_READY; + free_variables(); + if (!(error2 = gracl_init(gr_usermode))) { + unlock_kernel(); + gr_log_str(GR_DONT_AUDIT_GOOD, GR_RELOAD_ACL_MSG, GR_VERSION); + } else { + unlock_kernel(); + error = error2; + gr_log_str(GR_DONT_AUDIT, GR_RELOADF_ACL_MSG, GR_VERSION); + } + } else { + gr_log_str(GR_DONT_AUDIT, GR_RELOADF_ACL_MSG, GR_VERSION); + error = -EPERM; + } + break; + case SEGVMOD: + if (unlikely(!(gr_status & GR_READY))) { + gr_log_noargs(GR_DONT_AUDIT_GOOD, GR_SEGVMODI_ACL_MSG); + error = -EAGAIN; + break; + } + + if (!(chkpw(gr_usermode, gr_system_salt, gr_system_sum))) { + gr_log_noargs(GR_DONT_AUDIT_GOOD, GR_SEGVMODS_ACL_MSG); + if (gr_usermode->segv_device && gr_usermode->segv_inode) { + struct acl_subject_label *segvacl; + segvacl = + lookup_acl_subj_label(gr_usermode->segv_inode, + gr_usermode->segv_device, + current->role); + if (segvacl) { + segvacl->crashes = 0; + segvacl->expires = 0; + } + } else if (gr_find_uid(gr_usermode->segv_uid) >= 0) { + gr_remove_uid(gr_usermode->segv_uid); + } + } else { + gr_log_noargs(GR_DONT_AUDIT, GR_SEGVMODF_ACL_MSG); + error = -EPERM; + } + break; + case SPROLE: + case SPROLEPAM: + if (unlikely(!(gr_status & GR_READY))) { + gr_log_noargs(GR_DONT_AUDIT_GOOD, GR_SPROLEI_ACL_MSG); + error = -EAGAIN; + break; + } + + if (current->role->expires && time_after_eq(get_seconds(), current->role->expires)) { + current->role->expires = 0; + current->role->auth_attempts = 0; + } + + if (current->role->auth_attempts >= CONFIG_GRKERNSEC_ACL_MAXTRIES && + time_after(current->role->expires, get_seconds())) { + error = -EBUSY; + goto out; + } + + if (lookup_special_role_auth + (gr_usermode->mode, gr_usermode->sp_role, &sprole_salt, &sprole_sum) + && ((!sprole_salt && !sprole_sum) + || !(chkpw(gr_usermode, sprole_salt, sprole_sum)))) { + char *p = ""; + assign_special_role(gr_usermode->sp_role); + read_lock(&tasklist_lock); + if (current->parent) + p = current->parent->role->rolename; + read_unlock(&tasklist_lock); + gr_log_str_int(GR_DONT_AUDIT_GOOD, GR_SPROLES_ACL_MSG, + p, acl_sp_role_value); + } else { + gr_log_str(GR_DONT_AUDIT, GR_SPROLEF_ACL_MSG, gr_usermode->sp_role); + error = -EPERM; + if(!(current->role->auth_attempts++)) + current->role->expires = get_seconds() + CONFIG_GRKERNSEC_ACL_TIMEOUT; + + goto out; + } + break; + case UNSPROLE: + if (unlikely(!(gr_status & GR_READY))) { + gr_log_noargs(GR_DONT_AUDIT_GOOD, GR_UNSPROLEI_ACL_MSG); + error = -EAGAIN; + break; + } + + if (current->role->roletype & GR_ROLE_SPECIAL) { + char *p = ""; + int i = 0; + + read_lock(&tasklist_lock); + if (current->parent) { + p = current->parent->role->rolename; + i = current->parent->acl_role_id; + } + read_unlock(&tasklist_lock); + + gr_log_str_int(GR_DONT_AUDIT_GOOD, GR_UNSPROLES_ACL_MSG, p, i); + gr_set_acls(1); + } else { + gr_log_str(GR_DONT_AUDIT, GR_UNSPROLEF_ACL_MSG, current->role->rolename); + error = -EPERM; + goto out; + } + break; + default: + gr_log_int(GR_DONT_AUDIT, GR_INVMODE_ACL_MSG, gr_usermode->mode); + error = -EINVAL; + break; + } + + if (error != -EPERM) + goto out; + + if(!(gr_auth_attempts++)) + gr_auth_expires = get_seconds() + CONFIG_GRKERNSEC_ACL_TIMEOUT; + + out: + up(&gr_dev_sem); + return error; +} + +int +gr_set_acls(const int type) +{ + struct acl_object_label *obj; + struct task_struct *task, *task2; + struct file *filp; + struct acl_role_label *role = current->role; + __u16 acl_role_id = current->acl_role_id; + + read_lock(&tasklist_lock); + read_lock(&grsec_exec_file_lock); + do_each_thread(task2, task) { + /* check to see if we're called from the exit handler, + if so, only replace ACLs that have inherited the admin + ACL */ + + if (type && (task->role != role || + task->acl_role_id != acl_role_id)) + continue; + + task->acl_role_id = 0; + task->acl_sp_role = 0; + + if ((filp = task->exec_file)) { + task->role = lookup_acl_role_label(task, task->uid, task->gid); + + task->acl = + chk_subj_label(filp->f_dentry, filp->f_vfsmnt, + task->role); + if (task->acl) { + struct acl_subject_label *curr; + curr = task->acl; + + task->is_writable = 0; + /* ignore additional mmap checks for processes that are writable + by the default ACL */ + obj = chk_obj_label(filp->f_dentry, filp->f_vfsmnt, default_role->root_label); + if (unlikely(obj->mode & GR_WRITE)) + task->is_writable = 1; + obj = chk_obj_label(filp->f_dentry, filp->f_vfsmnt, task->role->root_label); + if (unlikely(obj->mode & GR_WRITE)) + task->is_writable = 1; + + gr_set_proc_res(task); + +#ifdef CONFIG_GRKERNSEC_ACL_DEBUG + printk(KERN_ALERT "gr_set_acls for (%s:%d): role:%s, subject:%s\n", task->comm, task->pid, task->role->rolename, task->acl->filename); +#endif + } else { + read_unlock(&grsec_exec_file_lock); + read_unlock(&tasklist_lock); + gr_log_str_int(GR_DONT_AUDIT_GOOD, GR_DEFACL_MSG, task->comm, task->pid); + return 1; + } + } else { + // it's a kernel process + task->role = kernel_role; + task->acl = kernel_role->root_label; +#ifdef CONFIG_GRKERNSEC_ACL_HIDEKERN + task->acl->mode &= ~GR_PROCFIND; +#endif + } + } while_each_thread(task2, task); + read_unlock(&grsec_exec_file_lock); + read_unlock(&tasklist_lock); + return 0; +} + +void +gr_learn_resource(const struct task_struct *task, + const int res, const unsigned long wanted, const int gt) +{ + struct acl_subject_label *acl; + + if (unlikely((gr_status & GR_READY) && + task->acl && (task->acl->mode & (GR_LEARN | GR_INHERITLEARN)))) + goto skip_reslog; + +#ifdef CONFIG_GRKERNSEC_RESLOG + gr_log_resource(task, res, wanted, gt); +#endif + skip_reslog: + + if (unlikely(!(gr_status & GR_READY) || !wanted)) + return; + + acl = task->acl; + + if (likely(!acl || !(acl->mode & (GR_LEARN | GR_INHERITLEARN)) || + !(acl->resmask & (1 << (unsigned short) res)))) + return; + + if (wanted >= acl->res[res].rlim_cur) { + unsigned long res_add; + + res_add = wanted; + switch (res) { + case RLIMIT_CPU: + res_add += GR_RLIM_CPU_BUMP; + break; + case RLIMIT_FSIZE: + res_add += GR_RLIM_FSIZE_BUMP; + break; + case RLIMIT_DATA: + res_add += GR_RLIM_DATA_BUMP; + break; + case RLIMIT_STACK: + res_add += GR_RLIM_STACK_BUMP; + break; + case RLIMIT_CORE: + res_add += GR_RLIM_CORE_BUMP; + break; + case RLIMIT_RSS: + res_add += GR_RLIM_RSS_BUMP; + break; + case RLIMIT_NPROC: + res_add += GR_RLIM_NPROC_BUMP; + break; + case RLIMIT_NOFILE: + res_add += GR_RLIM_NOFILE_BUMP; + break; + case RLIMIT_MEMLOCK: + res_add += GR_RLIM_MEMLOCK_BUMP; + break; + case RLIMIT_AS: + res_add += GR_RLIM_AS_BUMP; + break; + case RLIMIT_LOCKS: + res_add += GR_RLIM_LOCKS_BUMP; + break; + } + + acl->res[res].rlim_cur = res_add; + + if (wanted > acl->res[res].rlim_max) + acl->res[res].rlim_max = res_add; + + security_learn(GR_LEARN_AUDIT_MSG, task->role->rolename, + task->role->roletype, acl->filename, + acl->res[res].rlim_cur, acl->res[res].rlim_max, + "", (unsigned long) res); + } + + return; +} + +#ifdef CONFIG_PAX_HAVE_ACL_FLAGS +void +pax_set_initial_flags(struct linux_binprm *bprm) +{ + struct task_struct *task = current; + struct acl_subject_label *proc; + unsigned long flags; + + if (unlikely(!(gr_status & GR_READY))) + return; + + flags = pax_get_flags(task); + + proc = task->acl; + + if (proc->pax_flags & GR_PAX_DISABLE_PAGEEXEC) + flags &= ~MF_PAX_PAGEEXEC; + if (proc->pax_flags & GR_PAX_DISABLE_SEGMEXEC) + flags &= ~MF_PAX_SEGMEXEC; + if (proc->pax_flags & GR_PAX_DISABLE_RANDMMAP) + flags &= ~MF_PAX_RANDMMAP; + if (proc->pax_flags & GR_PAX_DISABLE_EMUTRAMP) + flags &= ~MF_PAX_EMUTRAMP; + if (proc->pax_flags & GR_PAX_DISABLE_MPROTECT) + flags &= ~MF_PAX_MPROTECT; + + if (proc->pax_flags & GR_PAX_ENABLE_PAGEEXEC) + flags |= MF_PAX_PAGEEXEC; + if (proc->pax_flags & GR_PAX_ENABLE_SEGMEXEC) + flags |= MF_PAX_SEGMEXEC; + if (proc->pax_flags & GR_PAX_ENABLE_RANDMMAP) + flags |= MF_PAX_RANDMMAP; + if (proc->pax_flags & GR_PAX_ENABLE_EMUTRAMP) + flags |= MF_PAX_EMUTRAMP; + if (proc->pax_flags & GR_PAX_ENABLE_MPROTECT) + flags |= MF_PAX_MPROTECT; + + pax_set_flags(task, flags); + + return; +} +#endif + +#ifdef CONFIG_SYSCTL +extern struct proc_dir_entry *proc_sys_root; + +/* the following function is called under the BKL */ + +__u32 +gr_handle_sysctl(const struct ctl_table *table, const void *oldval, + const void *newval) +{ + struct proc_dir_entry *tmp; + struct nameidata nd; + const char *proc_sys = "/proc/sys"; + char *path; + struct acl_object_label *obj; + unsigned short len = 0, pos = 0, depth = 0, i; + __u32 err = 0; + __u32 mode = 0; + + if (unlikely(!(gr_status & GR_READY))) + return 1; + + path = per_cpu_ptr(gr_shared_page[0], smp_processor_id()); + + if (oldval) + mode |= GR_READ; + if (newval) + mode |= GR_WRITE; + + /* convert the requested sysctl entry into a pathname */ + + for (tmp = table->de; tmp != proc_sys_root; tmp = tmp->parent) { + len += strlen(tmp->name); + len++; + depth++; + } + + if ((len + depth + strlen(proc_sys) + 1) > PAGE_SIZE) + return 0; /* deny */ + + memset(path, 0, PAGE_SIZE); + + memcpy(path, proc_sys, strlen(proc_sys)); + + pos += strlen(proc_sys); + + for (; depth > 0; depth--) { + path[pos] = '/'; + pos++; + for (i = 1, tmp = table->de; tmp != proc_sys_root; + tmp = tmp->parent) { + if (depth == i) { + memcpy(path + pos, tmp->name, + strlen(tmp->name)); + pos += strlen(tmp->name); + } + i++; + } + } + + err = path_lookup(path, LOOKUP_FOLLOW, &nd); + + if (err) + goto out; + + obj = chk_obj_label(nd.dentry, nd.mnt, current->acl); + err = obj->mode & (mode | to_gr_audit(mode) | GR_SUPPRESS); + + if (unlikely((current->acl->mode & (GR_LEARN | GR_INHERITLEARN)) && + ((err & mode) != mode))) { + __u32 new_mode = mode; + + new_mode &= ~(GR_AUDITS | GR_SUPPRESS); + + err = new_mode; + gr_log_learn(current, nd.dentry, nd.mnt, new_mode); + } else if ((err & mode) != mode && !(err & GR_SUPPRESS)) { + gr_log_str4(GR_DONT_AUDIT, GR_SYSCTL_ACL_MSG, "denied", + path, (mode & GR_READ) ? " reading" : "", + (mode & GR_WRITE) ? " writing" : ""); + err = 0; + } else if ((err & mode) != mode) { + err = 0; + } else if (((err & mode) == mode) && (err & GR_AUDITS)) { + gr_log_str4(GR_DO_AUDIT, GR_SYSCTL_ACL_MSG, "successful", + path, (mode & GR_READ) ? " reading" : "", + (mode & GR_WRITE) ? " writing" : ""); + } + + path_release(&nd); + + out: + return err; +} +#endif + +int +gr_handle_proc_ptrace(struct task_struct *task) +{ + struct file *filp; + struct task_struct *tmp = task; + struct task_struct *curtemp = current; + __u32 retmode; + + if (unlikely(!(gr_status & GR_READY))) + return 0; + + read_lock(&tasklist_lock); + read_lock(&grsec_exec_file_lock); + filp = task->exec_file; + + while (tmp->pid > 0) { + if (tmp == curtemp) + break; + tmp = tmp->parent; + } + + if (!filp || (tmp->pid == 0 && !(current->acl->mode & GR_RELAXPTRACE))) { + read_unlock(&grsec_exec_file_lock); + read_unlock(&tasklist_lock); + return 1; + } + + retmode = gr_search_file(filp->f_dentry, GR_NOPTRACE, filp->f_vfsmnt); + read_unlock(&grsec_exec_file_lock); + read_unlock(&tasklist_lock); + + if (retmode & GR_NOPTRACE) + return 1; + + if (!(current->acl->mode & GR_POVERRIDE) && !(current->role->roletype & GR_ROLE_GOD) + && (current->acl != task->acl || (current->acl != current->role->root_label + && current->pid != task->pid))) + return 1; + + return 0; +} + +int +gr_handle_ptrace(struct task_struct *task, const long request) +{ + struct task_struct *tmp = task; + struct task_struct *curtemp = current; + __u32 retmode; + + if (unlikely(!(gr_status & GR_READY))) + return 0; + + read_lock(&tasklist_lock); + while (tmp->pid > 0) { + if (tmp == curtemp) + break; + tmp = tmp->parent; + } + + if (tmp->pid == 0 && !(current->acl->mode & GR_RELAXPTRACE)) { + read_unlock(&tasklist_lock); + gr_log_ptrace(GR_DONT_AUDIT, GR_PTRACE_ACL_MSG, task); + return 1; + } + read_unlock(&tasklist_lock); + + read_lock(&grsec_exec_file_lock); + if (unlikely(!task->exec_file)) { + read_unlock(&grsec_exec_file_lock); + return 0; + } + + retmode = gr_search_file(task->exec_file->f_dentry, GR_PTRACERD | GR_NOPTRACE, task->exec_file->f_vfsmnt); + read_unlock(&grsec_exec_file_lock); + + if (retmode & GR_NOPTRACE) { + gr_log_ptrace(GR_DONT_AUDIT, GR_PTRACE_ACL_MSG, task); + return 1; + } + + if (retmode & GR_PTRACERD) { + switch (request) { + case PTRACE_POKETEXT: + case PTRACE_POKEDATA: + case PTRACE_POKEUSR: +#if !defined(CONFIG_PPC32) && !defined(CONFIG_PPC64) && !defined(CONFIG_PARISC) && !defined(CONFIG_ALPHA) && !defined(CONFIG_IA64) + case PTRACE_SETREGS: + case PTRACE_SETFPREGS: +#endif +#ifdef CONFIG_X86 + case PTRACE_SETFPXREGS: +#endif +#ifdef CONFIG_ALTIVEC + case PTRACE_SETVRREGS: +#endif + return 1; + default: + return 0; + } + } else if (!(current->acl->mode & GR_POVERRIDE) && + !(current->role->roletype & GR_ROLE_GOD) && + (current->acl != task->acl)) { + gr_log_ptrace(GR_DONT_AUDIT, GR_PTRACE_ACL_MSG, task); + return 1; + } + + return 0; +} + +static int is_writable_mmap(const struct file *filp) +{ + struct task_struct *task = current; + struct acl_object_label *obj, *obj2; + + if (gr_status & GR_READY && !(task->acl->mode & GR_OVERRIDE) && + !task->is_writable && S_ISREG(filp->f_dentry->d_inode->i_mode)) { + obj = chk_obj_label(filp->f_dentry, filp->f_vfsmnt, default_role->root_label); + obj2 = chk_obj_label(filp->f_dentry, filp->f_vfsmnt, + task->role->root_label); + if (unlikely((obj->mode & GR_WRITE) || (obj2->mode & GR_WRITE))) { + gr_log_fs_generic(GR_DONT_AUDIT, GR_WRITLIB_ACL_MSG, filp->f_dentry, filp->f_vfsmnt); + return 1; + } + } + return 0; +} + +int +gr_acl_handle_mmap(const struct file *file, const unsigned long prot) +{ + __u32 mode; + + if (unlikely(!file || !(prot & PROT_EXEC))) + return 1; + + if (is_writable_mmap(file)) + return 0; + + mode = + gr_search_file(file->f_dentry, + GR_EXEC | GR_AUDIT_EXEC | GR_SUPPRESS, + file->f_vfsmnt); + + if (!gr_tpe_allow(file)) + return 0; + + if (unlikely(!(mode & GR_EXEC) && !(mode & GR_SUPPRESS))) { + gr_log_fs_rbac_generic(GR_DONT_AUDIT, GR_MMAP_ACL_MSG, file->f_dentry, file->f_vfsmnt); + return 0; + } else if (unlikely(!(mode & GR_EXEC))) { + return 0; + } else if (unlikely(mode & GR_EXEC && mode & GR_AUDIT_EXEC)) { + gr_log_fs_rbac_generic(GR_DO_AUDIT, GR_MMAP_ACL_MSG, file->f_dentry, file->f_vfsmnt); + return 1; + } + + return 1; +} + +int +gr_acl_handle_mprotect(const struct file *file, const unsigned long prot) +{ + __u32 mode; + + if (unlikely(!file || !(prot & PROT_EXEC))) + return 1; + + if (is_writable_mmap(file)) + return 0; + + mode = + gr_search_file(file->f_dentry, + GR_EXEC | GR_AUDIT_EXEC | GR_SUPPRESS, + file->f_vfsmnt); + + if (!gr_tpe_allow(file)) + return 0; + + if (unlikely(!(mode & GR_EXEC) && !(mode & GR_SUPPRESS))) { + gr_log_fs_rbac_generic(GR_DONT_AUDIT, GR_MPROTECT_ACL_MSG, file->f_dentry, file->f_vfsmnt); + return 0; + } else if (unlikely(!(mode & GR_EXEC))) { + return 0; + } else if (unlikely(mode & GR_EXEC && mode & GR_AUDIT_EXEC)) { + gr_log_fs_rbac_generic(GR_DO_AUDIT, GR_MPROTECT_ACL_MSG, file->f_dentry, file->f_vfsmnt); + return 1; + } + + return 1; +} + +void +gr_acl_handle_psacct(struct task_struct *task, const long code) +{ + unsigned long runtime; + unsigned long cputime; + unsigned int wday, cday; + __u8 whr, chr; + __u8 wmin, cmin; + __u8 wsec, csec; + + if (unlikely(!(gr_status & GR_READY) || !task->acl || + !(task->acl->mode & GR_PROCACCT))) + return; + + runtime = xtime.tv_sec - task->start_time.tv_sec; + wday = runtime / (3600 * 24); + runtime -= wday * (3600 * 24); + whr = runtime / 3600; + runtime -= whr * 3600; + wmin = runtime / 60; + runtime -= wmin * 60; + wsec = runtime; + + cputime = (task->utime + task->stime) / HZ; + cday = cputime / (3600 * 24); + cputime -= cday * (3600 * 24); + chr = cputime / 3600; + cputime -= chr * 3600; + cmin = cputime / 60; + cputime -= cmin * 60; + csec = cputime; + + gr_log_procacct(GR_DO_AUDIT, GR_ACL_PROCACCT_MSG, task, wday, whr, wmin, wsec, cday, chr, cmin, csec, code); + + return; +} + +void gr_set_kernel_label(struct task_struct *task) +{ + if (gr_status & GR_READY) { + task->role = kernel_role; + task->acl = kernel_role->root_label; + } + return; +} + +int gr_acl_handle_filldir(const struct file *file, const char *name, const unsigned int namelen, const ino_t ino) +{ + struct task_struct *task = current; + struct dentry *dentry = file->f_dentry; + struct vfsmount *mnt = file->f_vfsmnt; + struct acl_object_label *obj, *tmp; + struct acl_subject_label *subj; + unsigned int bufsize; + int is_not_root; + char *path; + + if (unlikely(!(gr_status & GR_READY))) + return 1; + + if (task->acl->mode & (GR_LEARN | GR_INHERITLEARN)) + return 1; + + subj = task->acl; + do { + obj = lookup_acl_obj_label(ino, dentry->d_inode->i_sb->s_dev, subj); + if (obj != NULL) + return (obj->mode & GR_FIND) ? 1 : 0; + } while ((subj = subj->parent_subject)); + + obj = chk_obj_label(dentry, mnt, task->acl); + if (obj->globbed == NULL) + return (obj->mode & GR_FIND) ? 1 : 0; + + is_not_root = ((obj->filename[0] == '/') && + (obj->filename[1] == '\0')) ? 0 : 1; + bufsize = PAGE_SIZE - namelen - is_not_root; + + /* check bufsize > PAGE_SIZE || bufsize == 0 */ + if (unlikely((bufsize - 1) > (PAGE_SIZE - 1))) + return 1; + + preempt_disable(); + path = d_real_path(dentry, mnt, per_cpu_ptr(gr_shared_page[0], smp_processor_id()), + bufsize); + + bufsize = strlen(path); + + /* if base is "/", don't append an additional slash */ + if (is_not_root) + *(path + bufsize) = '/'; + memcpy(path + bufsize + is_not_root, name, namelen); + *(path + bufsize + namelen + is_not_root) = '\0'; + + tmp = obj->globbed; + while (tmp) { + if (!glob_match(tmp->filename, path)) { + preempt_enable(); + return (tmp->mode & GR_FIND) ? 1 : 0; + } + tmp = tmp->next; + } + preempt_enable(); + return (obj->mode & GR_FIND) ? 1 : 0; +} + +EXPORT_SYMBOL(gr_learn_resource); +EXPORT_SYMBOL(gr_set_kernel_label); +#ifdef CONFIG_SECURITY +EXPORT_SYMBOL(gr_check_user_change); +EXPORT_SYMBOL(gr_check_group_change); +#endif + diff -NurpP --minimal linux-2.6.17.13/grsecurity/gracl_alloc.c linux-2.6.17.13-g-v/grsecurity/gracl_alloc.c --- linux-2.6.17.13/grsecurity/gracl_alloc.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/gracl_alloc.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,91 @@ +#include +#include +#include +#include +#include +#include + +static unsigned long alloc_stack_next = 1; +static unsigned long alloc_stack_size = 1; +static void **alloc_stack; + +static __inline__ int +alloc_pop(void) +{ + if (alloc_stack_next == 1) + return 0; + + kfree(alloc_stack[alloc_stack_next - 2]); + + alloc_stack_next--; + + return 1; +} + +static __inline__ void +alloc_push(void *buf) +{ + if (alloc_stack_next >= alloc_stack_size) + BUG(); + + alloc_stack[alloc_stack_next - 1] = buf; + + alloc_stack_next++; + + return; +} + +void * +acl_alloc(unsigned long len) +{ + void *ret; + + if (len > PAGE_SIZE) + BUG(); + + ret = kmalloc(len, GFP_KERNEL); + + if (ret) + alloc_push(ret); + + return ret; +} + +void +acl_free_all(void) +{ + if (gr_acl_is_enabled() || !alloc_stack) + return; + + while (alloc_pop()) ; + + if (alloc_stack) { + if ((alloc_stack_size * sizeof (void *)) <= PAGE_SIZE) + kfree(alloc_stack); + else + vfree(alloc_stack); + } + + alloc_stack = NULL; + alloc_stack_size = 1; + alloc_stack_next = 1; + + return; +} + +int +acl_alloc_stack_init(unsigned long size) +{ + if ((size * sizeof (void *)) <= PAGE_SIZE) + alloc_stack = + (void **) kmalloc(size * sizeof (void *), GFP_KERNEL); + else + alloc_stack = (void **) vmalloc(size * sizeof (void *)); + + alloc_stack_size = size; + + if (!alloc_stack) + return 0; + else + return 1; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/gracl_cap.c linux-2.6.17.13-g-v/grsecurity/gracl_cap.c --- linux-2.6.17.13/grsecurity/gracl_cap.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/gracl_cap.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,110 @@ +#include +#include +#include +#include +#include +#include +#include + +static const char *captab_log[] = { + "CAP_CHOWN", + "CAP_DAC_OVERRIDE", + "CAP_DAC_READ_SEARCH", + "CAP_FOWNER", + "CAP_FSETID", + "CAP_KILL", + "CAP_SETGID", + "CAP_SETUID", + "CAP_SETPCAP", + "CAP_LINUX_IMMUTABLE", + "CAP_NET_BIND_SERVICE", + "CAP_NET_BROADCAST", + "CAP_NET_ADMIN", + "CAP_NET_RAW", + "CAP_IPC_LOCK", + "CAP_IPC_OWNER", + "CAP_SYS_MODULE", + "CAP_SYS_RAWIO", + "CAP_SYS_CHROOT", + "CAP_SYS_PTRACE", + "CAP_SYS_PACCT", + "CAP_SYS_ADMIN", + "CAP_SYS_BOOT", + "CAP_SYS_NICE", + "CAP_SYS_RESOURCE", + "CAP_SYS_TIME", + "CAP_SYS_TTY_CONFIG", + "CAP_MKNOD", + "CAP_LEASE" +}; + +EXPORT_SYMBOL(gr_task_is_capable); + +int +gr_task_is_capable(struct task_struct *task, const int cap) +{ + struct acl_subject_label *curracl; + __u32 cap_drop = 0, cap_mask = 0; + + if (!gr_acl_is_enabled()) + return 1; + + curracl = task->acl; + + cap_drop = curracl->cap_lower; + cap_mask = curracl->cap_mask; + + while ((curracl = curracl->parent_subject)) { + if (!(cap_mask & (1 << cap)) && (curracl->cap_mask & (1 << cap))) + cap_drop |= curracl->cap_lower & (1 << cap); + cap_mask |= curracl->cap_mask; + } + + if (!cap_raised(cap_drop, cap)) + return 1; + + curracl = task->acl; + + if ((curracl->mode & (GR_LEARN | GR_INHERITLEARN)) + && cap_raised(task->cap_effective, cap)) { + security_learn(GR_LEARN_AUDIT_MSG, task->role->rolename, + task->role->roletype, task->uid, + task->gid, task->exec_file ? + gr_to_filename(task->exec_file->f_dentry, + task->exec_file->f_vfsmnt) : curracl->filename, + curracl->filename, 0UL, + 0UL, "", (unsigned long) cap, NIPQUAD(task->signal->curr_ip)); + return 1; + } + + if ((cap >= 0) && (cap < (sizeof(captab_log)/sizeof(captab_log[0]))) && cap_raised(task->cap_effective, cap)) + gr_log_cap(GR_DONT_AUDIT, GR_CAP_ACL_MSG, task, captab_log[cap]); + + return 0; +} + +int +gr_is_capable_nolog(const int cap) +{ + struct acl_subject_label *curracl; + __u32 cap_drop = 0, cap_mask = 0; + + if (!gr_acl_is_enabled()) + return 1; + + curracl = current->acl; + + cap_drop = curracl->cap_lower; + cap_mask = curracl->cap_mask; + + while ((curracl = curracl->parent_subject)) { + cap_drop |= curracl->cap_lower & (cap_mask & ~curracl->cap_mask); + cap_mask |= curracl->cap_mask; + } + + if (!cap_raised(cap_drop, cap)) + return 1; + + return 0; +} + diff -NurpP --minimal linux-2.6.17.13/grsecurity/gracl_fs.c linux-2.6.17.13-g-v/grsecurity/gracl_fs.c --- linux-2.6.17.13/grsecurity/gracl_fs.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/gracl_fs.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,423 @@ +#include +#include +#include +#include +#include +#include +#include +#include +#include + +__u32 +gr_acl_handle_hidden_file(const struct dentry * dentry, + const struct vfsmount * mnt) +{ + __u32 mode; + + if (unlikely(!dentry->d_inode)) + return GR_FIND; + + mode = + gr_search_file(dentry, GR_FIND | GR_AUDIT_FIND | GR_SUPPRESS, mnt); + + if (unlikely(mode & GR_FIND && mode & GR_AUDIT_FIND)) { + gr_log_fs_rbac_generic(GR_DO_AUDIT, GR_HIDDEN_ACL_MSG, dentry, mnt); + return mode; + } else if (unlikely(!(mode & GR_FIND) && !(mode & GR_SUPPRESS))) { + gr_log_fs_rbac_generic(GR_DONT_AUDIT, GR_HIDDEN_ACL_MSG, dentry, mnt); + return 0; + } else if (unlikely(!(mode & GR_FIND))) + return 0; + + return GR_FIND; +} + +__u32 +gr_acl_handle_open(const struct dentry * dentry, const struct vfsmount * mnt, + const int fmode) +{ + __u32 reqmode = GR_FIND; + __u32 mode; + + if (unlikely(!dentry->d_inode)) + return reqmode; + + if (unlikely(fmode & O_APPEND)) + reqmode |= GR_APPEND; + else if (unlikely(fmode & FMODE_WRITE)) + reqmode |= GR_WRITE; + if (likely((fmode & FMODE_READ) && !(fmode & O_DIRECTORY))) + reqmode |= GR_READ; + + mode = + gr_search_file(dentry, reqmode | to_gr_audit(reqmode) | GR_SUPPRESS, + mnt); + + if (unlikely(((mode & reqmode) == reqmode) && mode & GR_AUDITS)) { + gr_log_fs_rbac_mode2(GR_DO_AUDIT, GR_OPEN_ACL_MSG, dentry, mnt, + reqmode & GR_READ ? " reading" : "", + reqmode & GR_WRITE ? " writing" : reqmode & + GR_APPEND ? " appending" : ""); + return reqmode; + } else + if (unlikely((mode & reqmode) != reqmode && !(mode & GR_SUPPRESS))) + { + gr_log_fs_rbac_mode2(GR_DONT_AUDIT, GR_OPEN_ACL_MSG, dentry, mnt, + reqmode & GR_READ ? " reading" : "", + reqmode & GR_WRITE ? " writing" : reqmode & + GR_APPEND ? " appending" : ""); + return 0; + } else if (unlikely((mode & reqmode) != reqmode)) + return 0; + + return reqmode; +} + +__u32 +gr_acl_handle_creat(const struct dentry * dentry, + const struct dentry * p_dentry, + const struct vfsmount * p_mnt, const int fmode, + const int imode) +{ + __u32 reqmode = GR_WRITE | GR_CREATE; + __u32 mode; + + if (unlikely(fmode & O_APPEND)) + reqmode |= GR_APPEND; + if (unlikely((fmode & FMODE_READ) && !(fmode & O_DIRECTORY))) + reqmode |= GR_READ; + if (unlikely((fmode & O_CREAT) && (imode & (S_ISUID | S_ISGID)))) + reqmode |= GR_SETID; + + mode = + gr_check_create(dentry, p_dentry, p_mnt, + reqmode | to_gr_audit(reqmode) | GR_SUPPRESS); + + if (unlikely(((mode & reqmode) == reqmode) && mode & GR_AUDITS)) { + gr_log_fs_rbac_mode2(GR_DO_AUDIT, GR_CREATE_ACL_MSG, dentry, p_mnt, + reqmode & GR_READ ? " reading" : "", + reqmode & GR_WRITE ? " writing" : reqmode & + GR_APPEND ? " appending" : ""); + return reqmode; + } else + if (unlikely((mode & reqmode) != reqmode && !(mode & GR_SUPPRESS))) + { + gr_log_fs_rbac_mode2(GR_DONT_AUDIT, GR_CREATE_ACL_MSG, dentry, p_mnt, + reqmode & GR_READ ? " reading" : "", + reqmode & GR_WRITE ? " writing" : reqmode & + GR_APPEND ? " appending" : ""); + return 0; + } else if (unlikely((mode & reqmode) != reqmode)) + return 0; + + return reqmode; +} + +__u32 +gr_acl_handle_access(const struct dentry * dentry, const struct vfsmount * mnt, + const int fmode) +{ + __u32 mode, reqmode = GR_FIND; + + if ((fmode & S_IXOTH) && !S_ISDIR(dentry->d_inode->i_mode)) + reqmode |= GR_EXEC; + if (fmode & S_IWOTH) + reqmode |= GR_WRITE; + if (fmode & S_IROTH) + reqmode |= GR_READ; + + mode = + gr_search_file(dentry, reqmode | to_gr_audit(reqmode) | GR_SUPPRESS, + mnt); + + if (unlikely(((mode & reqmode) == reqmode) && mode & GR_AUDITS)) { + gr_log_fs_rbac_mode3(GR_DO_AUDIT, GR_ACCESS_ACL_MSG, dentry, mnt, + reqmode & GR_READ ? " reading" : "", + reqmode & GR_WRITE ? " writing" : "", + reqmode & GR_EXEC ? " executing" : ""); + return reqmode; + } else + if (unlikely((mode & reqmode) != reqmode && !(mode & GR_SUPPRESS))) + { + gr_log_fs_rbac_mode3(GR_DONT_AUDIT, GR_ACCESS_ACL_MSG, dentry, mnt, + reqmode & GR_READ ? " reading" : "", + reqmode & GR_WRITE ? " writing" : "", + reqmode & GR_EXEC ? " executing" : ""); + return 0; + } else if (unlikely((mode & reqmode) != reqmode)) + return 0; + + return reqmode; +} + +static __u32 generic_fs_handler(const struct dentry *dentry, const struct vfsmount *mnt, __u32 reqmode, const char *fmt) +{ + __u32 mode; + + mode = gr_search_file(dentry, reqmode | to_gr_audit(reqmode) | GR_SUPPRESS, mnt); + + if (unlikely(((mode & (reqmode)) == (reqmode)) && mode & GR_AUDITS)) { + gr_log_fs_rbac_generic(GR_DO_AUDIT, fmt, dentry, mnt); + return mode; + } else if (unlikely((mode & (reqmode)) != (reqmode) && !(mode & GR_SUPPRESS))) { + gr_log_fs_rbac_generic(GR_DONT_AUDIT, fmt, dentry, mnt); + return 0; + } else if (unlikely((mode & (reqmode)) != (reqmode))) + return 0; + + return (reqmode); +} + +__u32 +gr_acl_handle_rmdir(const struct dentry * dentry, const struct vfsmount * mnt) +{ + return generic_fs_handler(dentry, mnt, GR_WRITE | GR_DELETE , GR_RMDIR_ACL_MSG); +} + +__u32 +gr_acl_handle_unlink(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return generic_fs_handler(dentry, mnt, GR_WRITE | GR_DELETE , GR_UNLINK_ACL_MSG); +} + +__u32 +gr_acl_handle_truncate(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return generic_fs_handler(dentry, mnt, GR_WRITE, GR_TRUNCATE_ACL_MSG); +} + +__u32 +gr_acl_handle_utime(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return generic_fs_handler(dentry, mnt, GR_WRITE, GR_ATIME_ACL_MSG); +} + +__u32 +gr_acl_handle_fchmod(const struct dentry *dentry, const struct vfsmount *mnt, + mode_t mode) +{ + if (unlikely(dentry->d_inode && S_ISSOCK(dentry->d_inode->i_mode))) + return 1; + + if (unlikely((mode != (mode_t)-1) && (mode & (S_ISUID | S_ISGID)))) { + return generic_fs_handler(dentry, mnt, GR_WRITE | GR_SETID, + GR_FCHMOD_ACL_MSG); + } else { + return generic_fs_handler(dentry, mnt, GR_WRITE, GR_FCHMOD_ACL_MSG); + } +} + +__u32 +gr_acl_handle_chmod(const struct dentry *dentry, const struct vfsmount *mnt, + mode_t mode) +{ + if (unlikely((mode != (mode_t)-1) && (mode & (S_ISUID | S_ISGID)))) { + return generic_fs_handler(dentry, mnt, GR_WRITE | GR_SETID, + GR_CHMOD_ACL_MSG); + } else { + return generic_fs_handler(dentry, mnt, GR_WRITE, GR_CHMOD_ACL_MSG); + } +} + +__u32 +gr_acl_handle_chown(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return generic_fs_handler(dentry, mnt, GR_WRITE, GR_CHOWN_ACL_MSG); +} + +__u32 +gr_acl_handle_execve(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return generic_fs_handler(dentry, mnt, GR_EXEC, GR_EXEC_ACL_MSG); +} + +__u32 +gr_acl_handle_unix(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return generic_fs_handler(dentry, mnt, GR_READ | GR_WRITE, + GR_UNIXCONNECT_ACL_MSG); +} + +/* hardlinks require at minimum create permission, + any additional privilege required is based on the + privilege of the file being linked to +*/ +__u32 +gr_acl_handle_link(const struct dentry * new_dentry, + const struct dentry * parent_dentry, + const struct vfsmount * parent_mnt, + const struct dentry * old_dentry, + const struct vfsmount * old_mnt, const char *to) +{ + __u32 mode; + __u32 needmode = GR_CREATE | GR_LINK; + __u32 needaudit = GR_AUDIT_CREATE | GR_AUDIT_LINK; + + mode = + gr_check_link(new_dentry, parent_dentry, parent_mnt, old_dentry, + old_mnt); + + if (unlikely(((mode & needmode) == needmode) && (mode & needaudit))) { + gr_log_fs_rbac_str(GR_DO_AUDIT, GR_LINK_ACL_MSG, old_dentry, old_mnt, to); + return mode; + } else if (unlikely(((mode & needmode) != needmode) && !(mode & GR_SUPPRESS))) { + gr_log_fs_rbac_str(GR_DONT_AUDIT, GR_LINK_ACL_MSG, old_dentry, old_mnt, to); + return 0; + } else if (unlikely((mode & needmode) != needmode)) + return 0; + + return 1; +} + +__u32 +gr_acl_handle_symlink(const struct dentry * new_dentry, + const struct dentry * parent_dentry, + const struct vfsmount * parent_mnt, const char *from) +{ + __u32 needmode = GR_WRITE | GR_CREATE; + __u32 mode; + + mode = + gr_check_create(new_dentry, parent_dentry, parent_mnt, + GR_CREATE | GR_AUDIT_CREATE | + GR_WRITE | GR_AUDIT_WRITE | GR_SUPPRESS); + + if (unlikely(mode & GR_WRITE && mode & GR_AUDITS)) { + gr_log_fs_str_rbac(GR_DO_AUDIT, GR_SYMLINK_ACL_MSG, from, new_dentry, parent_mnt); + return mode; + } else if (unlikely(((mode & needmode) != needmode) && !(mode & GR_SUPPRESS))) { + gr_log_fs_str_rbac(GR_DONT_AUDIT, GR_SYMLINK_ACL_MSG, from, new_dentry, parent_mnt); + return 0; + } else if (unlikely((mode & needmode) != needmode)) + return 0; + + return (GR_WRITE | GR_CREATE); +} + +static __u32 generic_fs_create_handler(const struct dentry *new_dentry, const struct dentry *parent_dentry, const struct vfsmount *parent_mnt, __u32 reqmode, const char *fmt) +{ + __u32 mode; + + mode = gr_check_create(new_dentry, parent_dentry, parent_mnt, reqmode | to_gr_audit(reqmode) | GR_SUPPRESS); + + if (unlikely(((mode & (reqmode)) == (reqmode)) && mode & GR_AUDITS)) { + gr_log_fs_rbac_generic(GR_DO_AUDIT, fmt, new_dentry, parent_mnt); + return mode; + } else if (unlikely((mode & (reqmode)) != (reqmode) && !(mode & GR_SUPPRESS))) { + gr_log_fs_rbac_generic(GR_DONT_AUDIT, fmt, new_dentry, parent_mnt); + return 0; + } else if (unlikely((mode & (reqmode)) != (reqmode))) + return 0; + + return (reqmode); +} + +__u32 +gr_acl_handle_mknod(const struct dentry * new_dentry, + const struct dentry * parent_dentry, + const struct vfsmount * parent_mnt, + const int mode) +{ + __u32 reqmode = GR_WRITE | GR_CREATE; + if (unlikely(mode & (S_ISUID | S_ISGID))) + reqmode |= GR_SETID; + + return generic_fs_create_handler(new_dentry, parent_dentry, parent_mnt, + reqmode, GR_MKNOD_ACL_MSG); +} + +__u32 +gr_acl_handle_mkdir(const struct dentry *new_dentry, + const struct dentry *parent_dentry, + const struct vfsmount *parent_mnt) +{ + return generic_fs_create_handler(new_dentry, parent_dentry, parent_mnt, + GR_WRITE | GR_CREATE, GR_MKDIR_ACL_MSG); +} + +#define RENAME_CHECK_SUCCESS(old, new) \ + (((old & (GR_WRITE | GR_READ)) == (GR_WRITE | GR_READ)) && \ + ((new & (GR_WRITE | GR_READ)) == (GR_WRITE | GR_READ))) + +int +gr_acl_handle_rename(struct dentry *new_dentry, + struct dentry *parent_dentry, + const struct vfsmount *parent_mnt, + struct dentry *old_dentry, + struct inode *old_parent_inode, + struct vfsmount *old_mnt, const char *newname) +{ + __u32 comp1, comp2; + int error = 0; + + if (unlikely(!gr_acl_is_enabled())) + return 0; + + if (!new_dentry->d_inode) { + comp1 = gr_check_create(new_dentry, parent_dentry, parent_mnt, + GR_READ | GR_WRITE | GR_CREATE | GR_AUDIT_READ | + GR_AUDIT_WRITE | GR_AUDIT_CREATE | GR_SUPPRESS); + comp2 = gr_search_file(old_dentry, GR_READ | GR_WRITE | + GR_DELETE | GR_AUDIT_DELETE | + GR_AUDIT_READ | GR_AUDIT_WRITE | + GR_SUPPRESS, old_mnt); + } else { + comp1 = gr_search_file(new_dentry, GR_READ | GR_WRITE | + GR_CREATE | GR_DELETE | + GR_AUDIT_CREATE | GR_AUDIT_DELETE | + GR_AUDIT_READ | GR_AUDIT_WRITE | + GR_SUPPRESS, parent_mnt); + comp2 = + gr_search_file(old_dentry, + GR_READ | GR_WRITE | GR_AUDIT_READ | + GR_DELETE | GR_AUDIT_DELETE | + GR_AUDIT_WRITE | GR_SUPPRESS, old_mnt); + } + + if (RENAME_CHECK_SUCCESS(comp1, comp2) && + ((comp1 & GR_AUDITS) || (comp2 & GR_AUDITS))) + gr_log_fs_rbac_str(GR_DO_AUDIT, GR_RENAME_ACL_MSG, old_dentry, old_mnt, newname); + else if (!RENAME_CHECK_SUCCESS(comp1, comp2) && !(comp1 & GR_SUPPRESS) + && !(comp2 & GR_SUPPRESS)) { + gr_log_fs_rbac_str(GR_DONT_AUDIT, GR_RENAME_ACL_MSG, old_dentry, old_mnt, newname); + error = -EACCES; + } else if (unlikely(!RENAME_CHECK_SUCCESS(comp1, comp2))) + error = -EACCES; + + return error; +} + +void +gr_acl_handle_exit(void) +{ + u16 id; + char *rolename; + struct file *exec_file; + + if (unlikely(current->acl_sp_role && gr_acl_is_enabled())) { + id = current->acl_role_id; + rolename = current->role->rolename; + gr_set_acls(1); + gr_log_str_int(GR_DONT_AUDIT_GOOD, GR_SPROLEL_ACL_MSG, rolename, id); + } + + write_lock(&grsec_exec_file_lock); + exec_file = current->exec_file; + current->exec_file = NULL; + write_unlock(&grsec_exec_file_lock); + + if (exec_file) + fput(exec_file); +} + +int +gr_acl_handle_procpidmem(const struct task_struct *task) +{ + if (unlikely(!gr_acl_is_enabled())) + return 0; + + if (task->acl->mode & GR_PROTPROCFD) + return -EACCES; + + return 0; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/gracl_ip.c linux-2.6.17.13-g-v/grsecurity/gracl_ip.c --- linux-2.6.17.13/grsecurity/gracl_ip.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/gracl_ip.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,313 @@ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#define GR_BIND 0x01 +#define GR_CONNECT 0x02 +#define GR_INVERT 0x04 + +static const char * gr_protocols[256] = { + "ip", "icmp", "igmp", "ggp", "ipencap", "st", "tcp", "cbt", + "egp", "igp", "bbn-rcc", "nvp", "pup", "argus", "emcon", "xnet", + "chaos", "udp", "mux", "dcn", "hmp", "prm", "xns-idp", "trunk-1", + "trunk-2", "leaf-1", "leaf-2", "rdp", "irtp", "iso-tp4", "netblt", "mfe-nsp", + "merit-inp", "sep", "3pc", "idpr", "xtp", "ddp", "idpr-cmtp", "tp++", + "il", "ipv6", "sdrp", "ipv6-route", "ipv6-frag", "idrp", "rsvp", "gre", + "mhrp", "bna", "ipv6-crypt", "ipv6-auth", "i-nlsp", "swipe", "narp", "mobile", + "tlsp", "skip", "ipv6-icmp", "ipv6-nonxt", "ipv6-opts", "unknown:61", "cftp", "unknown:63", + "sat-expak", "kryptolan", "rvd", "ippc", "unknown:68", "sat-mon", "visa", "ipcv", + "cpnx", "cphb", "wsn", "pvp", "br-sat-mon", "sun-nd", "wb-mon", "wb-expak", + "iso-ip", "vmtp", "secure-vmtp", "vines", "ttp", "nfsnet-igp", "dgp", "tcf", + "eigrp", "ospf", "sprite-rpc", "larp", "mtp", "ax.25", "ipip", "micp", + "scc-sp", "etherip", "encap", "unknown:99", "gmtp", "ifmp", "pnni", "pim", + "aris", "scps", "qnx", "a/n", "ipcomp", "snp", "compaq-peer", "ipx-in-ip", + "vrrp", "pgm", "unknown:114", "l2tp", "ddx", "iatp", "stp", "srp", + "uti", "smp", "sm", "ptp", "isis", "fire", "crtp", "crdup", + "sscopmce", "iplt", "sps", "pipe", "sctp", "fc", "unkown:134", "unknown:135", + "unknown:136", "unknown:137", "unknown:138", "unknown:139", "unknown:140", "unknown:141", "unknown:142", "unknown:143", + "unknown:144", "unknown:145", "unknown:146", "unknown:147", "unknown:148", "unknown:149", "unknown:150", "unknown:151", + "unknown:152", "unknown:153", "unknown:154", "unknown:155", "unknown:156", "unknown:157", "unknown:158", "unknown:159", + "unknown:160", "unknown:161", "unknown:162", "unknown:163", "unknown:164", "unknown:165", "unknown:166", "unknown:167", + "unknown:168", "unknown:169", "unknown:170", "unknown:171", "unknown:172", "unknown:173", "unknown:174", "unknown:175", + "unknown:176", "unknown:177", "unknown:178", "unknown:179", "unknown:180", "unknown:181", "unknown:182", "unknown:183", + "unknown:184", "unknown:185", "unknown:186", "unknown:187", "unknown:188", "unknown:189", "unknown:190", "unknown:191", + "unknown:192", "unknown:193", "unknown:194", "unknown:195", "unknown:196", "unknown:197", "unknown:198", "unknown:199", + "unknown:200", "unknown:201", "unknown:202", "unknown:203", "unknown:204", "unknown:205", "unknown:206", "unknown:207", + "unknown:208", "unknown:209", "unknown:210", "unknown:211", "unknown:212", "unknown:213", "unknown:214", "unknown:215", + "unknown:216", "unknown:217", "unknown:218", "unknown:219", "unknown:220", "unknown:221", "unknown:222", "unknown:223", + "unknown:224", "unknown:225", "unknown:226", "unknown:227", "unknown:228", "unknown:229", "unknown:230", "unknown:231", + "unknown:232", "unknown:233", "unknown:234", "unknown:235", "unknown:236", "unknown:237", "unknown:238", "unknown:239", + "unknown:240", "unknown:241", "unknown:242", "unknown:243", "unknown:244", "unknown:245", "unknown:246", "unknown:247", + "unknown:248", "unknown:249", "unknown:250", "unknown:251", "unknown:252", "unknown:253", "unknown:254", "unknown:255", + }; + +static const char * gr_socktypes[11] = { + "unknown:0", "stream", "dgram", "raw", "rdm", "seqpacket", "unknown:6", + "unknown:7", "unknown:8", "unknown:9", "packet" + }; + +const char * +gr_proto_to_name(unsigned char proto) +{ + return gr_protocols[proto]; +} + +const char * +gr_socktype_to_name(unsigned char type) +{ + return gr_socktypes[type]; +} + +int +gr_search_socket(const int domain, const int type, const int protocol) +{ + struct acl_subject_label *curr; + + if (unlikely(!gr_acl_is_enabled())) + goto exit; + + if ((domain < 0) || (type < 0) || (protocol < 0) || (domain != PF_INET) + || (domain >= NPROTO) || (type >= SOCK_MAX) || (protocol > 255)) + goto exit; // let the kernel handle it + + curr = current->acl; + + if (!curr->ips) + goto exit; + + if ((curr->ip_type & (1 << type)) && + (curr->ip_proto[protocol / 32] & (1 << (protocol % 32)))) + goto exit; + + if (curr->mode & (GR_LEARN | GR_INHERITLEARN)) { + /* we don't place acls on raw sockets , and sometimes + dgram/ip sockets are opened for ioctl and not + bind/connect, so we'll fake a bind learn log */ + if (type == SOCK_RAW || type == SOCK_PACKET) { + __u32 fakeip = 0; + security_learn(GR_IP_LEARN_MSG, current->role->rolename, + current->role->roletype, current->uid, + current->gid, current->exec_file ? + gr_to_filename(current->exec_file->f_dentry, + current->exec_file->f_vfsmnt) : + curr->filename, curr->filename, + NIPQUAD(fakeip), 0, type, + protocol, GR_CONNECT, +NIPQUAD(current->signal->curr_ip)); + } else if ((type == SOCK_DGRAM) && (protocol == IPPROTO_IP)) { + __u32 fakeip = 0; + security_learn(GR_IP_LEARN_MSG, current->role->rolename, + current->role->roletype, current->uid, + current->gid, current->exec_file ? + gr_to_filename(current->exec_file->f_dentry, + current->exec_file->f_vfsmnt) : + curr->filename, curr->filename, + NIPQUAD(fakeip), 0, type, + protocol, GR_BIND, NIPQUAD(current->signal->curr_ip)); + } + /* we'll log when they use connect or bind */ + goto exit; + } + + gr_log_str3(GR_DONT_AUDIT, GR_SOCK_MSG, "inet", + gr_socktype_to_name(type), gr_proto_to_name(protocol)); + + return 0; + exit: + return 1; +} + +int check_ip_policy(struct acl_ip_label *ip, __u32 ip_addr, __u16 ip_port, __u8 protocol, const int mode, const int type, __u32 our_addr, __u32 our_netmask) +{ + if ((ip->mode & mode) && + (ip_port >= ip->low) && + (ip_port <= ip->high) && + ((ntohl(ip_addr) & our_netmask) == + (ntohl(our_addr) & our_netmask)) + && (ip->proto[protocol / 32] & (1 << (protocol % 32))) + && (ip->type & (1 << type))) { + if (ip->mode & GR_INVERT) + return 2; // specifically denied + else + return 1; // allowed + } + + return 0; // not specifically allowed, may continue parsing +} + +static int +gr_search_connectbind(const int mode, const struct sock *sk, + const struct sockaddr_in *addr, const int type) +{ + char iface[IFNAMSIZ] = {0}; + struct acl_subject_label *curr; + struct acl_ip_label *ip; + struct net_device *dev; + struct in_device *idev; + unsigned long i; + int ret; + __u32 ip_addr = 0; + __u32 our_addr; + __u32 our_netmask; + char *p; + __u16 ip_port = 0; + + if (unlikely(!gr_acl_is_enabled() || sk->sk_family != PF_INET)) + return 1; + + curr = current->acl; + + if (!curr->ips) + return 1; + + ip_addr = addr->sin_addr.s_addr; + ip_port = ntohs(addr->sin_port); + + if (curr->mode & (GR_LEARN | GR_INHERITLEARN)) { + security_learn(GR_IP_LEARN_MSG, current->role->rolename, + current->role->roletype, current->uid, + current->gid, current->exec_file ? + gr_to_filename(current->exec_file->f_dentry, + current->exec_file->f_vfsmnt) : + curr->filename, curr->filename, + NIPQUAD(ip_addr), ip_port, type, + sk->sk_protocol, mode, NIPQUAD(current->signal->curr_ip)); + return 1; + } + + for (i = 0; i < curr->ip_num; i++) { + ip = *(curr->ips + i); + if (ip->iface != NULL) { + strncpy(iface, ip->iface, IFNAMSIZ - 1); + p = strchr(iface, ':'); + if (p != NULL) + *p = '\0'; + dev = dev_get_by_name(iface); + if (dev == NULL) + continue; + idev = in_dev_get(dev); + if (idev == NULL) { + dev_put(dev); + continue; + } + rcu_read_lock(); + for_ifa(idev) { + if (!strcmp(ip->iface, ifa->ifa_label)) { + our_addr = ifa->ifa_address; + our_netmask = 0xffffffff; + ret = check_ip_policy(ip, ip_addr, ip_port, sk->sk_protocol, mode, type, our_addr, our_netmask); + if (ret == 1) { + rcu_read_unlock(); + in_dev_put(idev); + dev_put(dev); + return 1; + } else if (ret == 2) { + rcu_read_unlock(); + in_dev_put(idev); + dev_put(dev); + goto denied; + } + } + } endfor_ifa(idev); + rcu_read_unlock(); + in_dev_put(idev); + dev_put(dev); + } else { + our_addr = ip->addr; + our_netmask = ip->netmask; + ret = check_ip_policy(ip, ip_addr, ip_port, sk->sk_protocol, mode, type, our_addr, our_netmask); + if (ret == 1) + return 1; + else if (ret == 2) + goto denied; + } + } + +denied: + if (mode == GR_BIND) + gr_log_int5_str2(GR_DONT_AUDIT, GR_BIND_ACL_MSG, NIPQUAD(ip_addr), ip_port, gr_socktype_to_name(type), gr_proto_to_name(sk->sk_protocol)); + else if (mode == GR_CONNECT) + gr_log_int5_str2(GR_DONT_AUDIT, GR_CONNECT_ACL_MSG, NIPQUAD(ip_addr), ip_port, gr_socktype_to_name(type), gr_proto_to_name(sk->sk_protocol)); + + return 0; +} + +int +gr_search_connect(const struct socket *sock, const struct sockaddr_in *addr) +{ + return gr_search_connectbind(GR_CONNECT, sock->sk, addr, sock->type); +} + +int +gr_search_bind(const struct socket *sock, const struct sockaddr_in *addr) +{ + return gr_search_connectbind(GR_BIND, sock->sk, addr, sock->type); +} + +int gr_search_listen(const struct socket *sock) +{ + struct sock *sk = sock->sk; + struct sockaddr_in addr; + + addr.sin_addr.s_addr = inet_sk(sk)->saddr; + addr.sin_port = inet_sk(sk)->sport; + + return gr_search_connectbind(GR_BIND, sock->sk, &addr, sock->type); +} + +int gr_search_accept(const struct socket *sock) +{ + struct sock *sk = sock->sk; + struct sockaddr_in addr; + + addr.sin_addr.s_addr = inet_sk(sk)->saddr; + addr.sin_port = inet_sk(sk)->sport; + + return gr_search_connectbind(GR_BIND, sock->sk, &addr, sock->type); +} + +int +gr_search_udp_sendmsg(const struct sock *sk, const struct sockaddr_in *addr) +{ + if (addr) + return gr_search_connectbind(GR_CONNECT, sk, addr, SOCK_DGRAM); + else { + struct sockaddr_in sin; + const struct inet_sock *inet = inet_sk(sk); + + sin.sin_addr.s_addr = inet->daddr; + sin.sin_port = inet->dport; + + return gr_search_connectbind(GR_CONNECT, sk, &sin, SOCK_DGRAM); + } +} + +int +gr_search_udp_recvmsg(const struct sock *sk, const struct sk_buff *skb) +{ + struct sockaddr_in sin; + + if (unlikely(skb->len < sizeof (struct udphdr))) + return 1; // skip this packet + + sin.sin_addr.s_addr = skb->nh.iph->saddr; + sin.sin_port = skb->h.uh->source; + + return gr_search_connectbind(GR_CONNECT, sk, &sin, SOCK_DGRAM); +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/gracl_learn.c linux-2.6.17.13-g-v/grsecurity/gracl_learn.c --- linux-2.6.17.13/grsecurity/gracl_learn.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/gracl_learn.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,204 @@ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +extern ssize_t write_grsec_handler(struct file * file, const char __user * buf, + size_t count, loff_t *ppos); +extern int gr_acl_is_enabled(void); + +static DECLARE_WAIT_QUEUE_HEAD(learn_wait); +static int gr_learn_attached; + +/* use a 512k buffer */ +#define LEARN_BUFFER_SIZE (512 * 1024) + +static spinlock_t gr_learn_lock = SPIN_LOCK_UNLOCKED; +static DECLARE_MUTEX(gr_learn_user_sem); + +/* we need to maintain two buffers, so that the kernel context of grlearn + uses a semaphore around the userspace copying, and the other kernel contexts + use a spinlock when copying into the buffer, since they cannot sleep +*/ +static char *learn_buffer; +static char *learn_buffer_user; +static int learn_buffer_len; +static int learn_buffer_user_len; + +static ssize_t +read_learn(struct file *file, char __user * buf, size_t count, loff_t * ppos) +{ + DECLARE_WAITQUEUE(wait, current); + ssize_t retval = 0; + + add_wait_queue(&learn_wait, &wait); + set_current_state(TASK_INTERRUPTIBLE); + do { + down(&gr_learn_user_sem); + spin_lock(&gr_learn_lock); + if (learn_buffer_len) + break; + spin_unlock(&gr_learn_lock); + up(&gr_learn_user_sem); + if (file->f_flags & O_NONBLOCK) { + retval = -EAGAIN; + goto out; + } + if (signal_pending(current)) { + retval = -ERESTARTSYS; + goto out; + } + + schedule(); + } while (1); + + memcpy(learn_buffer_user, learn_buffer, learn_buffer_len); + learn_buffer_user_len = learn_buffer_len; + retval = learn_buffer_len; + learn_buffer_len = 0; + + spin_unlock(&gr_learn_lock); + + if (copy_to_user(buf, learn_buffer_user, learn_buffer_user_len)) + retval = -EFAULT; + + up(&gr_learn_user_sem); +out: + set_current_state(TASK_RUNNING); + remove_wait_queue(&learn_wait, &wait); + return retval; +} + +static unsigned int +poll_learn(struct file * file, poll_table * wait) +{ + poll_wait(file, &learn_wait, wait); + + if (learn_buffer_len) + return (POLLIN | POLLRDNORM); + + return 0; +} + +void +gr_clear_learn_entries(void) +{ + char *tmp; + + down(&gr_learn_user_sem); + if (learn_buffer != NULL) { + spin_lock(&gr_learn_lock); + tmp = learn_buffer; + learn_buffer = NULL; + spin_unlock(&gr_learn_lock); + vfree(learn_buffer); + } + if (learn_buffer_user != NULL) { + vfree(learn_buffer_user); + learn_buffer_user = NULL; + } + learn_buffer_len = 0; + up(&gr_learn_user_sem); + + return; +} + +void +gr_add_learn_entry(const char *fmt, ...) +{ + va_list args; + unsigned int len; + + if (!gr_learn_attached) + return; + + spin_lock(&gr_learn_lock); + + /* leave a gap at the end so we know when it's "full" but don't have to + compute the exact length of the string we're trying to append + */ + if (learn_buffer_len > LEARN_BUFFER_SIZE - 16384) { + spin_unlock(&gr_learn_lock); + wake_up_interruptible(&learn_wait); + return; + } + if (learn_buffer == NULL) { + spin_unlock(&gr_learn_lock); + return; + } + + va_start(args, fmt); + len = vsnprintf(learn_buffer + learn_buffer_len, LEARN_BUFFER_SIZE - learn_buffer_len, fmt, args); + va_end(args); + + learn_buffer_len += len + 1; + + spin_unlock(&gr_learn_lock); + wake_up_interruptible(&learn_wait); + + return; +} + +static int +open_learn(struct inode *inode, struct file *file) +{ + if (file->f_mode & FMODE_READ && gr_learn_attached) + return -EBUSY; + if (file->f_mode & FMODE_READ) { + down(&gr_learn_user_sem); + if (learn_buffer == NULL) + learn_buffer = vmalloc(LEARN_BUFFER_SIZE); + if (learn_buffer_user == NULL) + learn_buffer_user = vmalloc(LEARN_BUFFER_SIZE); + if (learn_buffer == NULL) + return -ENOMEM; + if (learn_buffer_user == NULL) + return -ENOMEM; + learn_buffer_len = 0; + learn_buffer_user_len = 0; + gr_learn_attached = 1; + up(&gr_learn_user_sem); + } + return 0; +} + +static int +close_learn(struct inode *inode, struct file *file) +{ + char *tmp; + + if (file->f_mode & FMODE_READ) { + down(&gr_learn_user_sem); + if (learn_buffer != NULL) { + spin_lock(&gr_learn_lock); + tmp = learn_buffer; + learn_buffer = NULL; + spin_unlock(&gr_learn_lock); + vfree(tmp); + } + if (learn_buffer_user != NULL) { + vfree(learn_buffer_user); + learn_buffer_user = NULL; + } + learn_buffer_len = 0; + learn_buffer_user_len = 0; + gr_learn_attached = 0; + up(&gr_learn_user_sem); + } + + return 0; +} + +struct file_operations grsec_fops = { + .read = read_learn, + .write = write_grsec_handler, + .open = open_learn, + .release = close_learn, + .poll = poll_learn, +}; diff -NurpP --minimal linux-2.6.17.13/grsecurity/gracl_res.c linux-2.6.17.13-g-v/grsecurity/gracl_res.c --- linux-2.6.17.13/grsecurity/gracl_res.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/gracl_res.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,45 @@ +#include +#include +#include +#include + +static const char *restab_log[] = { + [RLIMIT_CPU] = "RLIMIT_CPU", + [RLIMIT_FSIZE] = "RLIMIT_FSIZE", + [RLIMIT_DATA] = "RLIMIT_DATA", + [RLIMIT_STACK] = "RLIMIT_STACK", + [RLIMIT_CORE] = "RLIMIT_CORE", + [RLIMIT_RSS] = "RLIMIT_RSS", + [RLIMIT_NPROC] = "RLIMIT_NPROC", + [RLIMIT_NOFILE] = "RLIMIT_NOFILE", + [RLIMIT_MEMLOCK] = "RLIMIT_MEMLOCK", + [RLIMIT_AS] = "RLIMIT_AS", + [RLIMIT_LOCKS] = "RLIMIT_LOCKS", + [RLIMIT_LOCKS + 1] = "RLIMIT_CRASH" +}; + +void +gr_log_resource(const struct task_struct *task, + const int res, const unsigned long wanted, const int gt) +{ + if (res == RLIMIT_NPROC && + (cap_raised(task->cap_effective, CAP_SYS_ADMIN) || + cap_raised(task->cap_effective, CAP_SYS_RESOURCE))) + return; + else if (res == RLIMIT_MEMLOCK && + cap_raised(task->cap_effective, CAP_IPC_LOCK)) + return; + + if (!gr_acl_is_enabled() && !grsec_resource_logging) + return; + + preempt_disable(); + + if (unlikely(((gt && wanted > task->signal->rlim[res].rlim_cur) || + (!gt && wanted >= task->signal->rlim[res].rlim_cur)) && + task->signal->rlim[res].rlim_cur != RLIM_INFINITY)) + gr_log_res_ulong2_str(GR_DONT_AUDIT, GR_RESOURCE_MSG, task, wanted, restab_log[res], task->signal->rlim[res].rlim_cur); + preempt_enable_no_resched(); + + return; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/gracl_segv.c linux-2.6.17.13-g-v/grsecurity/gracl_segv.c --- linux-2.6.17.13/grsecurity/gracl_segv.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/gracl_segv.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,295 @@ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +static struct crash_uid *uid_set; +static unsigned short uid_used; +static spinlock_t gr_uid_lock = SPIN_LOCK_UNLOCKED; +extern rwlock_t gr_inode_lock; +extern struct acl_subject_label * + lookup_acl_subj_label(const ino_t inode, const dev_t dev, + struct acl_role_label *role); +extern int specific_send_sig_info(int sig, struct siginfo *info, struct task_struct *t); + +int +gr_init_uidset(void) +{ + uid_set = + kmalloc(GR_UIDTABLE_MAX * sizeof (struct crash_uid), GFP_KERNEL); + uid_used = 0; + + return uid_set ? 1 : 0; +} + +void +gr_free_uidset(void) +{ + if (uid_set) + kfree(uid_set); + + return; +} + +int +gr_find_uid(const uid_t uid) +{ + struct crash_uid *tmp = uid_set; + uid_t buid; + int low = 0, high = uid_used - 1, mid; + + while (high >= low) { + mid = (low + high) >> 1; + buid = tmp[mid].uid; + if (buid == uid) + return mid; + if (buid > uid) + high = mid - 1; + if (buid < uid) + low = mid + 1; + } + + return -1; +} + +static __inline__ void +gr_insertsort(void) +{ + unsigned short i, j; + struct crash_uid index; + + for (i = 1; i < uid_used; i++) { + index = uid_set[i]; + j = i; + while ((j > 0) && uid_set[j - 1].uid > index.uid) { + uid_set[j] = uid_set[j - 1]; + j--; + } + uid_set[j] = index; + } + + return; +} + +static __inline__ void +gr_insert_uid(const uid_t uid, const unsigned long expires) +{ + int loc; + + if (uid_used == GR_UIDTABLE_MAX) + return; + + loc = gr_find_uid(uid); + + if (loc >= 0) { + uid_set[loc].expires = expires; + return; + } + + uid_set[uid_used].uid = uid; + uid_set[uid_used].expires = expires; + uid_used++; + + gr_insertsort(); + + return; +} + +void +gr_remove_uid(const unsigned short loc) +{ + unsigned short i; + + for (i = loc + 1; i < uid_used; i++) + uid_set[i - 1] = uid_set[i]; + + uid_used--; + + return; +} + +int +gr_check_crash_uid(const uid_t uid) +{ + int loc; + int ret = 0; + + if (unlikely(!gr_acl_is_enabled())) + return 0; + + spin_lock(&gr_uid_lock); + loc = gr_find_uid(uid); + + if (loc < 0) + goto out_unlock; + + if (time_before_eq(uid_set[loc].expires, get_seconds())) + gr_remove_uid(loc); + else + ret = 1; + +out_unlock: + spin_unlock(&gr_uid_lock); + return ret; +} + +static __inline__ int +proc_is_setxid(const struct task_struct *task) +{ + if (task->uid != task->euid || task->uid != task->suid || + task->uid != task->fsuid) + return 1; + if (task->gid != task->egid || task->gid != task->sgid || + task->gid != task->fsgid) + return 1; + + return 0; +} +static __inline__ int +gr_fake_force_sig(int sig, struct task_struct *t) +{ + unsigned long int flags; + int ret; + + spin_lock_irqsave(&t->sighand->siglock, flags); + if (sigismember(&t->blocked, sig) || t->sighand->action[sig-1].sa.sa_handler == SIG_IGN) { + t->sighand->action[sig-1].sa.sa_handler = SIG_DFL; + sigdelset(&t->blocked, sig); + recalc_sigpending_tsk(t); + } + ret = specific_send_sig_info(sig, (void*)1L, t); + spin_unlock_irqrestore(&t->sighand->siglock, flags); + + return ret; +} + +void +gr_handle_crash(struct task_struct *task, const int sig) +{ + struct acl_subject_label *curr; + struct acl_subject_label *curr2; + struct task_struct *tsk, *tsk2; + + if (sig != SIGSEGV && sig != SIGKILL && sig != SIGBUS && sig != SIGILL) + return; + + if (unlikely(!gr_acl_is_enabled())) + return; + + curr = task->acl; + + if (!(curr->resmask & (1 << GR_CRASH_RES))) + return; + + if (time_before_eq(curr->expires, get_seconds())) { + curr->expires = 0; + curr->crashes = 0; + } + + curr->crashes++; + + if (!curr->expires) + curr->expires = get_seconds() + curr->res[GR_CRASH_RES].rlim_max; + + if ((curr->crashes >= curr->res[GR_CRASH_RES].rlim_cur) && + time_after(curr->expires, get_seconds())) { + if (task->uid && proc_is_setxid(task)) { + gr_log_crash1(GR_DONT_AUDIT, GR_SEGVSTART_ACL_MSG, task, curr->res[GR_CRASH_RES].rlim_max); + spin_lock(&gr_uid_lock); + gr_insert_uid(task->uid, curr->expires); + spin_unlock(&gr_uid_lock); + curr->expires = 0; + curr->crashes = 0; + read_lock(&tasklist_lock); + do_each_thread(tsk2, tsk) { + if (tsk != task && tsk->uid == task->uid) + gr_fake_force_sig(SIGKILL, tsk); + } while_each_thread(tsk2, tsk); + read_unlock(&tasklist_lock); + } else { + gr_log_crash2(GR_DONT_AUDIT, GR_SEGVNOSUID_ACL_MSG, task, curr->res[GR_CRASH_RES].rlim_max); + read_lock(&tasklist_lock); + do_each_thread(tsk2, tsk) { + if (likely(tsk != task)) { + curr2 = tsk->acl; + + if (curr2->device == curr->device && + curr2->inode == curr->inode) + gr_fake_force_sig(SIGKILL, tsk); + } + } while_each_thread(tsk2, tsk); + read_unlock(&tasklist_lock); + } + } + + return; +} + +int +gr_check_crash_exec(const struct file *filp) +{ + struct acl_subject_label *curr; + + if (unlikely(!gr_acl_is_enabled())) + return 0; + + read_lock(&gr_inode_lock); + curr = lookup_acl_subj_label(filp->f_dentry->d_inode->i_ino, + filp->f_dentry->d_inode->i_sb->s_dev, + current->role); + read_unlock(&gr_inode_lock); + + if (!curr || !(curr->resmask & (1 << GR_CRASH_RES)) || + (!curr->crashes && !curr->expires)) + return 0; + + if ((curr->crashes >= curr->res[GR_CRASH_RES].rlim_cur) && + time_after(curr->expires, get_seconds())) + return 1; + else if (time_before_eq(curr->expires, get_seconds())) { + curr->crashes = 0; + curr->expires = 0; + } + + return 0; +} + +void +gr_handle_alertkill(struct task_struct *task) +{ + struct acl_subject_label *curracl; + __u32 curr_ip; + struct task_struct *p, *p2; + + if (unlikely(!gr_acl_is_enabled())) + return; + + curracl = task->acl; + curr_ip = task->signal->curr_ip; + + if ((curracl->mode & GR_KILLIPPROC) && curr_ip) { + read_lock(&tasklist_lock); + do_each_thread(p2, p) { + if (p->signal->curr_ip == curr_ip) + gr_fake_force_sig(SIGKILL, p); + } while_each_thread(p2, p); + read_unlock(&tasklist_lock); + } else if (curracl->mode & GR_KILLPROC) + gr_fake_force_sig(SIGKILL, task); + + return; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/gracl_shm.c linux-2.6.17.13-g-v/grsecurity/gracl_shm.c --- linux-2.6.17.13/grsecurity/gracl_shm.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/gracl_shm.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,34 @@ +#include +#include +#include +#include +#include +#include +#include +#include +#include + +int +gr_handle_shmat(const pid_t shm_cprid, const pid_t shm_lapid, + const time_t shm_createtime, const uid_t cuid, const int shmid) +{ + struct task_struct *task; + + if (!gr_acl_is_enabled()) + return 1; + + task = find_task_by_pid(shm_cprid); + + if (unlikely(!task)) + task = find_task_by_pid(shm_lapid); + + if (unlikely(task && (time_before((unsigned long)task->start_time.tv_sec, (unsigned long)shm_createtime) || + (task->pid == shm_lapid)) && + (task->acl->mode & GR_PROTSHM) && + (task->acl != current->acl))) { + gr_log_int3(GR_DONT_AUDIT, GR_SHMAT_ACL_MSG, cuid, shm_cprid, shmid); + return 0; + } + + return 1; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_chdir.c linux-2.6.17.13-g-v/grsecurity/grsec_chdir.c --- linux-2.6.17.13/grsecurity/grsec_chdir.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_chdir.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,19 @@ +#include +#include +#include +#include +#include +#include + +void +gr_log_chdir(const struct dentry *dentry, const struct vfsmount *mnt) +{ +#ifdef CONFIG_GRKERNSEC_AUDIT_CHDIR + if ((grsec_enable_chdir && grsec_enable_group && + in_group_p(grsec_audit_gid)) || (grsec_enable_chdir && + !grsec_enable_group)) { + gr_log_fs_generic(GR_DO_AUDIT, GR_CHDIR_AUDIT_MSG, dentry, mnt); + } +#endif + return; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_chroot.c linux-2.6.17.13-g-v/grsecurity/grsec_chroot.c --- linux-2.6.17.13/grsecurity/grsec_chroot.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_chroot.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,332 @@ +#include +#include +#include +#include +#include +#include +#include +#include + +int +gr_handle_chroot_unix(const pid_t pid) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_UNIX + struct pid *spid = NULL; + + if (unlikely(!grsec_enable_chroot_unix)) + return 1; + + if (likely(!proc_is_chrooted(current))) + return 1; + + read_lock(&tasklist_lock); + + spid = find_pid(pid); + if (spid) { + struct task_struct *p; + p = pid_task(spid, PIDTYPE_PID); + task_lock(p); + if (unlikely(!have_same_root(current, p))) { + task_unlock(p); + read_unlock(&tasklist_lock); + gr_log_noargs(GR_DONT_AUDIT, GR_UNIX_CHROOT_MSG); + return 0; + } + task_unlock(p); + } + read_unlock(&tasklist_lock); +#endif + return 1; +} + +int +gr_handle_chroot_nice(void) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_NICE + if (grsec_enable_chroot_nice && proc_is_chrooted(current)) { + gr_log_noargs(GR_DONT_AUDIT, GR_NICE_CHROOT_MSG); + return -EPERM; + } +#endif + return 0; +} + +int +gr_handle_chroot_setpriority(struct task_struct *p, const int niceval) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_NICE + if (grsec_enable_chroot_nice && (niceval < task_nice(p)) + && proc_is_chrooted(current)) { + gr_log_str_int(GR_DONT_AUDIT, GR_PRIORITY_CHROOT_MSG, p->comm, p->pid); + return -EACCES; + } +#endif + return 0; +} + +int +gr_handle_chroot_rawio(const struct inode *inode) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_CAPS + if (grsec_enable_chroot_caps && proc_is_chrooted(current) && + inode && S_ISBLK(inode->i_mode) && !capable(CAP_SYS_RAWIO)) + return 1; +#endif + return 0; +} + +int +gr_pid_is_chrooted(struct task_struct *p) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_FINDTASK + if (!grsec_enable_chroot_findtask || !proc_is_chrooted(current) || !p) + return 0; + + task_lock(p); + if ((p->exit_state & (EXIT_ZOMBIE | EXIT_DEAD)) || + !have_same_root(current, p)) { + task_unlock(p); + return 1; + } + task_unlock(p); +#endif + return 0; +} + +EXPORT_SYMBOL(gr_pid_is_chrooted); + +#if defined(CONFIG_GRKERNSEC_CHROOT_DOUBLE) || defined(CONFIG_GRKERNSEC_CHROOT_FCHDIR) +int gr_is_outside_chroot(const struct dentry *u_dentry, const struct vfsmount *u_mnt) +{ + struct dentry *dentry = (struct dentry *)u_dentry; + struct vfsmount *mnt = (struct vfsmount *)u_mnt; + struct dentry *realroot; + struct vfsmount *realrootmnt; + struct dentry *currentroot; + struct vfsmount *currentmnt; + int ret = 1; + + read_lock(&child_reaper->fs->lock); + realrootmnt = mntget(child_reaper->fs->rootmnt); + realroot = dget(child_reaper->fs->root); + read_unlock(&child_reaper->fs->lock); + + read_lock(&current->fs->lock); + currentmnt = mntget(current->fs->rootmnt); + currentroot = dget(current->fs->root); + read_unlock(&current->fs->lock); + + spin_lock(&dcache_lock); + for (;;) { + if (unlikely((dentry == realroot && mnt == realrootmnt) + || (dentry == currentroot && mnt == currentmnt))) + break; + if (unlikely(dentry == mnt->mnt_root || IS_ROOT(dentry))) { + if (mnt->mnt_parent == mnt) + break; + dentry = mnt->mnt_mountpoint; + mnt = mnt->mnt_parent; + continue; + } + dentry = dentry->d_parent; + } + spin_unlock(&dcache_lock); + + dput(currentroot); + mntput(currentmnt); + + /* access is outside of chroot */ + if (dentry == realroot && mnt == realrootmnt) + ret = 0; + + dput(realroot); + mntput(realrootmnt); + return ret; +} +#endif + +int +gr_chroot_fchdir(struct dentry *u_dentry, struct vfsmount *u_mnt) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_FCHDIR + if (!grsec_enable_chroot_fchdir) + return 1; + + if (!proc_is_chrooted(current)) + return 1; + else if (!gr_is_outside_chroot(u_dentry, u_mnt)) { + gr_log_fs_generic(GR_DONT_AUDIT, GR_CHROOT_FCHDIR_MSG, u_dentry, u_mnt); + return 0; + } +#endif + return 1; +} + +int +gr_chroot_shmat(const pid_t shm_cprid, const pid_t shm_lapid, + const time_t shm_createtime) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_SHMAT + struct pid *pid = NULL; + time_t starttime; + + if (unlikely(!grsec_enable_chroot_shmat)) + return 1; + + if (likely(!proc_is_chrooted(current))) + return 1; + + read_lock(&tasklist_lock); + + pid = find_pid(shm_cprid); + if (pid) { + struct task_struct *p; + p = pid_task(pid, PIDTYPE_PID); + task_lock(p); + starttime = p->start_time.tv_sec; + if (unlikely(!have_same_root(current, p) && + time_before((unsigned long)starttime, (unsigned long)shm_createtime))) { + task_unlock(p); + read_unlock(&tasklist_lock); + gr_log_noargs(GR_DONT_AUDIT, GR_SHMAT_CHROOT_MSG); + return 0; + } + task_unlock(p); + } else { + pid = find_pid(shm_lapid); + if (pid) { + struct task_struct *p; + p = pid_task(pid, PIDTYPE_PID); + task_lock(p); + if (unlikely(!have_same_root(current, p))) { + task_unlock(p); + read_unlock(&tasklist_lock); + gr_log_noargs(GR_DONT_AUDIT, GR_SHMAT_CHROOT_MSG); + return 0; + } + task_unlock(p); + } + } + + read_unlock(&tasklist_lock); +#endif + return 1; +} + +void +gr_log_chroot_exec(const struct dentry *dentry, const struct vfsmount *mnt) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_EXECLOG + if (grsec_enable_chroot_execlog && proc_is_chrooted(current)) + gr_log_fs_generic(GR_DO_AUDIT, GR_EXEC_CHROOT_MSG, dentry, mnt); +#endif + return; +} + +int +gr_handle_chroot_mknod(const struct dentry *dentry, + const struct vfsmount *mnt, const int mode) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_MKNOD + if (grsec_enable_chroot_mknod && !S_ISFIFO(mode) && !S_ISREG(mode) && + proc_is_chrooted(current)) { + gr_log_fs_generic(GR_DONT_AUDIT, GR_MKNOD_CHROOT_MSG, dentry, mnt); + return -EPERM; + } +#endif + return 0; +} + +int +gr_handle_chroot_mount(const struct dentry *dentry, + const struct vfsmount *mnt, const char *dev_name) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_MOUNT + if (grsec_enable_chroot_mount && proc_is_chrooted(current)) { + gr_log_str_fs(GR_DONT_AUDIT, GR_MOUNT_CHROOT_MSG, dev_name, dentry, mnt); + return -EPERM; + } +#endif + return 0; +} + +int +gr_handle_chroot_pivot(void) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_PIVOT + if (grsec_enable_chroot_pivot && proc_is_chrooted(current)) { + gr_log_noargs(GR_DONT_AUDIT, GR_PIVOT_CHROOT_MSG); + return -EPERM; + } +#endif + return 0; +} + +int +gr_handle_chroot_chroot(const struct dentry *dentry, const struct vfsmount *mnt) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_DOUBLE + if (grsec_enable_chroot_double && proc_is_chrooted(current) && + !gr_is_outside_chroot(dentry, mnt)) { + gr_log_fs_generic(GR_DONT_AUDIT, GR_CHROOT_CHROOT_MSG, dentry, mnt); + return -EPERM; + } +#endif + return 0; +} + +void +gr_handle_chroot_caps(struct task_struct *task) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_CAPS + if (grsec_enable_chroot_caps && proc_is_chrooted(task)) { + task->cap_permitted = + cap_drop(task->cap_permitted, GR_CHROOT_CAPS); + task->cap_inheritable = + cap_drop(task->cap_inheritable, GR_CHROOT_CAPS); + task->cap_effective = + cap_drop(task->cap_effective, GR_CHROOT_CAPS); + } +#endif + return; +} + +int +gr_handle_chroot_sysctl(const int op) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_SYSCTL + if (grsec_enable_chroot_sysctl && proc_is_chrooted(current) + && (op & 002)) + return -EACCES; +#endif + return 0; +} + +void +gr_handle_chroot_chdir(struct dentry *dentry, struct vfsmount *mnt) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_CHDIR + if (grsec_enable_chroot_chdir) + set_fs_pwd(current->fs, mnt, dentry); +#endif + return; +} + +int +gr_handle_chroot_chmod(const struct dentry *dentry, + const struct vfsmount *mnt, const int mode) +{ +#ifdef CONFIG_GRKERNSEC_CHROOT_CHMOD + if (grsec_enable_chroot_chmod && + ((mode & S_ISUID) || ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP))) && + proc_is_chrooted(current)) { + gr_log_fs_generic(GR_DONT_AUDIT, GR_CHMOD_CHROOT_MSG, dentry, mnt); + return -EPERM; + } +#endif + return 0; +} + +#ifdef CONFIG_SECURITY +EXPORT_SYMBOL(gr_handle_chroot_caps); +#endif diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_disabled.c linux-2.6.17.13-g-v/grsecurity/grsec_disabled.c --- linux-2.6.17.13/grsecurity/grsec_disabled.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_disabled.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,419 @@ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#ifdef CONFIG_PAX_HAVE_ACL_FLAGS +void +pax_set_initial_flags(struct linux_binprm *bprm) +{ + return; +} +#endif + +#ifdef CONFIG_SYSCTL +__u32 +gr_handle_sysctl(const struct ctl_table * table, __u32 mode) +{ + return mode; +} +#endif + +int +gr_acl_is_enabled(void) +{ + return 0; +} + +int +gr_handle_rawio(const struct inode *inode) +{ + return 0; +} + +void +gr_acl_handle_psacct(struct task_struct *task, const long code) +{ + return; +} + +int +gr_handle_ptrace(struct task_struct *task, const long request) +{ + return 0; +} + +int +gr_handle_proc_ptrace(struct task_struct *task) +{ + return 0; +} + +void +gr_learn_resource(const struct task_struct *task, + const int res, const unsigned long wanted, const int gt) +{ + return; +} + +int +gr_set_acls(const int type) +{ + return 0; +} + +int +gr_check_hidden_task(const struct task_struct *tsk) +{ + return 0; +} + +int +gr_check_protected_task(const struct task_struct *task) +{ + return 0; +} + +void +gr_copy_label(struct task_struct *tsk) +{ + return; +} + +void +gr_set_pax_flags(struct task_struct *task) +{ + return; +} + +int +gr_set_proc_label(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return 0; +} + +void +gr_handle_delete(const ino_t ino, const dev_t dev) +{ + return; +} + +void +gr_handle_create(const struct dentry *dentry, const struct vfsmount *mnt) +{ + return; +} + +void +gr_handle_crash(struct task_struct *task, const int sig) +{ + return; +} + +int +gr_check_crash_exec(const struct file *filp) +{ + return 0; +} + +int +gr_check_crash_uid(const uid_t uid) +{ + return 0; +} + +void +gr_handle_rename(struct inode *old_dir, struct inode *new_dir, + struct dentry *old_dentry, + struct dentry *new_dentry, + struct vfsmount *mnt, const __u8 replace) +{ + return; +} + +int +gr_search_socket(const int family, const int type, const int protocol) +{ + return 1; +} + +int +gr_search_connectbind(const int mode, const struct socket *sock, + const struct sockaddr_in *addr) +{ + return 1; +} + +int +gr_task_is_capable(struct task_struct *task, const int cap) +{ + return 1; +} + +int +gr_is_capable_nolog(const int cap) +{ + return 1; +} + +void +gr_handle_alertkill(struct task_struct *task) +{ + return; +} + +__u32 +gr_acl_handle_execve(const struct dentry * dentry, const struct vfsmount * mnt) +{ + return 1; +} + +__u32 +gr_acl_handle_hidden_file(const struct dentry * dentry, + const struct vfsmount * mnt) +{ + return 1; +} + +__u32 +gr_acl_handle_open(const struct dentry * dentry, const struct vfsmount * mnt, + const int fmode) +{ + return 1; +} + +__u32 +gr_acl_handle_rmdir(const struct dentry * dentry, const struct vfsmount * mnt) +{ + return 1; +} + +__u32 +gr_acl_handle_unlink(const struct dentry * dentry, const struct vfsmount * mnt) +{ + return 1; +} + +int +gr_acl_handle_mmap(const struct file *file, const unsigned long prot, + unsigned int *vm_flags) +{ + return 1; +} + +__u32 +gr_acl_handle_truncate(const struct dentry * dentry, + const struct vfsmount * mnt) +{ + return 1; +} + +__u32 +gr_acl_handle_utime(const struct dentry * dentry, const struct vfsmount * mnt) +{ + return 1; +} + +__u32 +gr_acl_handle_access(const struct dentry * dentry, + const struct vfsmount * mnt, const int fmode) +{ + return 1; +} + +__u32 +gr_acl_handle_fchmod(const struct dentry * dentry, const struct vfsmount * mnt, + mode_t mode) +{ + return 1; +} + +__u32 +gr_acl_handle_chmod(const struct dentry * dentry, const struct vfsmount * mnt, + mode_t mode) +{ + return 1; +} + +__u32 +gr_acl_handle_chown(const struct dentry * dentry, const struct vfsmount * mnt) +{ + return 1; +} + +void +grsecurity_init(void) +{ + return; +} + +__u32 +gr_acl_handle_mknod(const struct dentry * new_dentry, + const struct dentry * parent_dentry, + const struct vfsmount * parent_mnt, + const int mode) +{ + return 1; +} + +__u32 +gr_acl_handle_mkdir(const struct dentry * new_dentry, + const struct dentry * parent_dentry, + const struct vfsmount * parent_mnt) +{ + return 1; +} + +__u32 +gr_acl_handle_symlink(const struct dentry * new_dentry, + const struct dentry * parent_dentry, + const struct vfsmount * parent_mnt, const char *from) +{ + return 1; +} + +__u32 +gr_acl_handle_link(const struct dentry * new_dentry, + const struct dentry * parent_dentry, + const struct vfsmount * parent_mnt, + const struct dentry * old_dentry, + const struct vfsmount * old_mnt, const char *to) +{ + return 1; +} + +int +gr_acl_handle_rename(const struct dentry *new_dentry, + const struct dentry *parent_dentry, + const struct vfsmount *parent_mnt, + const struct dentry *old_dentry, + const struct inode *old_parent_inode, + const struct vfsmount *old_mnt, const char *newname) +{ + return 0; +} + +int +gr_acl_handle_filldir(const struct file *file, const char *name, + const int namelen, const ino_t ino) +{ + return 1; +} + +int +gr_handle_shmat(const pid_t shm_cprid, const pid_t shm_lapid, + const time_t shm_createtime, const uid_t cuid, const int shmid) +{ + return 1; +} + +int +gr_search_bind(const struct socket *sock, const struct sockaddr_in *addr) +{ + return 1; +} + +int +gr_search_accept(const struct socket *sock) +{ + return 1; +} + +int +gr_search_listen(const struct socket *sock) +{ + return 1; +} + +int +gr_search_connect(const struct socket *sock, const struct sockaddr_in *addr) +{ + return 1; +} + +__u32 +gr_acl_handle_unix(const struct dentry * dentry, const struct vfsmount * mnt) +{ + return 1; +} + +__u32 +gr_acl_handle_creat(const struct dentry * dentry, + const struct dentry * p_dentry, + const struct vfsmount * p_mnt, const int fmode, + const int imode) +{ + return 1; +} + +void +gr_acl_handle_exit(void) +{ + return; +} + +int +gr_acl_handle_mprotect(const struct file *file, const unsigned long prot) +{ + return 1; +} + +void +gr_set_role_label(const uid_t uid, const gid_t gid) +{ + return; +} + +int +gr_acl_handle_procpidmem(const struct task_struct *task) +{ + return 0; +} + +int +gr_search_udp_recvmsg(const struct sock *sk, const struct sk_buff *skb) +{ + return 1; +} + +int +gr_search_udp_sendmsg(const struct sock *sk, const struct sockaddr_in *addr) +{ + return 1; +} + +void +gr_set_kernel_label(struct task_struct *task) +{ + return; +} + +int +gr_check_user_change(int real, int effective, int fs) +{ + return 0; +} + +int +gr_check_group_change(int real, int effective, int fs) +{ + return 0; +} + + +EXPORT_SYMBOL(gr_task_is_capable); +EXPORT_SYMBOL(gr_learn_resource); +EXPORT_SYMBOL(gr_set_kernel_label); +#ifdef CONFIG_SECURITY +EXPORT_SYMBOL(gr_check_user_change); +EXPORT_SYMBOL(gr_check_group_change); +#endif diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_exec.c linux-2.6.17.13-g-v/grsecurity/grsec_exec.c --- linux-2.6.17.13/grsecurity/grsec_exec.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_exec.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,88 @@ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include + +#ifdef CONFIG_GRKERNSEC_EXECLOG +static char gr_exec_arg_buf[132]; +static DECLARE_MUTEX(gr_exec_arg_sem); +#endif + +int +gr_handle_nproc(void) +{ +#ifdef CONFIG_GRKERNSEC_EXECVE + if (grsec_enable_execve && current->user && + (atomic_read(&current->user->processes) > + current->signal->rlim[RLIMIT_NPROC].rlim_cur) && + !capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE)) { + gr_log_noargs(GR_DONT_AUDIT, GR_NPROC_MSG); + return -EAGAIN; + } +#endif + return 0; +} + +void +gr_handle_exec_args(struct linux_binprm *bprm, const char __user *__user *argv) +{ +#ifdef CONFIG_GRKERNSEC_EXECLOG + char *grarg = gr_exec_arg_buf; + unsigned int i, x, execlen = 0; + char c; + + if (!((grsec_enable_execlog && grsec_enable_group && + in_group_p(grsec_audit_gid)) + || (grsec_enable_execlog && !grsec_enable_group))) + return; + + down(&gr_exec_arg_sem); + memset(grarg, 0, sizeof(gr_exec_arg_buf)); + + if (unlikely(argv == NULL)) + goto log; + + for (i = 0; i < bprm->argc && execlen < 128; i++) { + const char __user *p; + unsigned int len; + + if (copy_from_user(&p, argv + i, sizeof(p))) + goto log; + if (!p) + goto log; + len = strnlen_user(p, 128 - execlen); + if (len > 128 - execlen) + len = 128 - execlen; + else if (len > 0) + len--; + if (copy_from_user(grarg + execlen, p, len)) + goto log; + + /* rewrite unprintable characters */ + for (x = 0; x < len; x++) { + c = *(grarg + execlen + x); + if (c < 32 || c > 126) + *(grarg + execlen + x) = ' '; + } + + execlen += len; + *(grarg + execlen) = ' '; + *(grarg + execlen + 1) = '\0'; + execlen++; + } + + log: + gr_log_fs_str(GR_DO_AUDIT, GR_EXEC_AUDIT_MSG, bprm->file->f_dentry, + bprm->file->f_vfsmnt, grarg); + up(&gr_exec_arg_sem); +#endif + return; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_fifo.c linux-2.6.17.13-g-v/grsecurity/grsec_fifo.c --- linux-2.6.17.13/grsecurity/grsec_fifo.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_fifo.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,22 @@ +#include +#include +#include +#include +#include + +int +gr_handle_fifo(const struct dentry *dentry, const struct vfsmount *mnt, + const struct dentry *dir, const int flag, const int acc_mode) +{ +#ifdef CONFIG_GRKERNSEC_FIFO + if (grsec_enable_fifo && S_ISFIFO(dentry->d_inode->i_mode) && + !(flag & O_EXCL) && (dir->d_inode->i_mode & S_ISVTX) && + (dentry->d_inode->i_uid != dir->d_inode->i_uid) && + (current->fsuid != dentry->d_inode->i_uid)) { + if (!generic_permission(dentry->d_inode, acc_mode, NULL)) + gr_log_fs_int2(GR_DONT_AUDIT, GR_FIFO_MSG, dentry, mnt, dentry->d_inode->i_uid, dentry->d_inode->i_gid); + return -EACCES; + } +#endif + return 0; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_fork.c linux-2.6.17.13-g-v/grsecurity/grsec_fork.c --- linux-2.6.17.13/grsecurity/grsec_fork.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_fork.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,15 @@ +#include +#include +#include +#include +#include + +void +gr_log_forkfail(const int retval) +{ +#ifdef CONFIG_GRKERNSEC_FORKFAIL + if (grsec_enable_forkfail && retval != -ERESTARTNOINTR) + gr_log_int(GR_DONT_AUDIT, GR_FAILFORK_MSG, retval); +#endif + return; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_init.c linux-2.6.17.13-g-v/grsecurity/grsec_init.c --- linux-2.6.17.13/grsecurity/grsec_init.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_init.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,236 @@ +#include +#include +#include +#include +#include +#include +#include +#include + +int grsec_enable_shm; +int grsec_enable_link; +int grsec_enable_dmesg; +int grsec_enable_fifo; +int grsec_enable_execve; +int grsec_enable_execlog; +int grsec_enable_signal; +int grsec_enable_forkfail; +int grsec_enable_time; +int grsec_enable_audit_textrel; +int grsec_enable_group; +int grsec_audit_gid; +int grsec_enable_chdir; +int grsec_enable_audit_ipc; +int grsec_enable_mount; +int grsec_enable_chroot_findtask; +int grsec_enable_chroot_mount; +int grsec_enable_chroot_shmat; +int grsec_enable_chroot_fchdir; +int grsec_enable_chroot_double; +int grsec_enable_chroot_pivot; +int grsec_enable_chroot_chdir; +int grsec_enable_chroot_chmod; +int grsec_enable_chroot_mknod; +int grsec_enable_chroot_nice; +int grsec_enable_chroot_execlog; +int grsec_enable_chroot_caps; +int grsec_enable_chroot_sysctl; +int grsec_enable_chroot_unix; +int grsec_enable_tpe; +int grsec_tpe_gid; +int grsec_enable_tpe_all; +int grsec_enable_randpid; +int grsec_enable_socket_all; +int grsec_socket_all_gid; +int grsec_enable_socket_client; +int grsec_socket_client_gid; +int grsec_enable_socket_server; +int grsec_socket_server_gid; +int grsec_resource_logging; +int grsec_lock; + +spinlock_t grsec_alert_lock = SPIN_LOCK_UNLOCKED; +unsigned long grsec_alert_wtime = 0; +unsigned long grsec_alert_fyet = 0; + +spinlock_t grsec_audit_lock = SPIN_LOCK_UNLOCKED; + +rwlock_t grsec_exec_file_lock = RW_LOCK_UNLOCKED; + +char *gr_shared_page[4]; + +char *gr_alert_log_fmt; +char *gr_audit_log_fmt; +char *gr_alert_log_buf; +char *gr_audit_log_buf; + +extern struct gr_arg *gr_usermode; +extern unsigned char *gr_system_salt; +extern unsigned char *gr_system_sum; + +void +grsecurity_init(void) +{ + int j; + /* create the per-cpu shared pages */ + + preempt_disable(); + for (j = 0; j < 4; j++) { + gr_shared_page[j] = (char *)__alloc_percpu(PAGE_SIZE); + if (gr_shared_page[j] == NULL) { + panic("Unable to allocate grsecurity shared page"); + return; + } + } + preempt_enable(); + + /* allocate log buffers */ + gr_alert_log_fmt = kmalloc(512, GFP_KERNEL); + if (!gr_alert_log_fmt) { + panic("Unable to allocate grsecurity alert log format buffer"); + return; + } + gr_audit_log_fmt = kmalloc(512, GFP_KERNEL); + if (!gr_audit_log_fmt) { + panic("Unable to allocate grsecurity audit log format buffer"); + return; + } + gr_alert_log_buf = (char *) get_zeroed_page(GFP_KERNEL); + if (!gr_alert_log_buf) { + panic("Unable to allocate grsecurity alert log buffer"); + return; + } + gr_audit_log_buf = (char *) get_zeroed_page(GFP_KERNEL); + if (!gr_audit_log_buf) { + panic("Unable to allocate grsecurity audit log buffer"); + return; + } + + /* allocate memory for authentication structure */ + gr_usermode = kmalloc(sizeof(struct gr_arg), GFP_KERNEL); + gr_system_salt = kmalloc(GR_SALT_LEN, GFP_KERNEL); + gr_system_sum = kmalloc(GR_SHA_LEN, GFP_KERNEL); + + if (!gr_usermode || !gr_system_salt || !gr_system_sum) { + panic("Unable to allocate grsecurity authentication structure"); + return; + } + +#if !defined(CONFIG_GRKERNSEC_SYSCTL) || defined(CONFIG_GRKERNSEC_SYSCTL_ON) +#ifndef CONFIG_GRKERNSEC_SYSCTL + grsec_lock = 1; +#endif +#ifdef CONFIG_GRKERNSEC_SHM + grsec_enable_shm = 1; +#endif +#ifdef CONFIG_GRKERNSEC_AUDIT_TEXTREL + grsec_enable_audit_textrel = 1; +#endif +#ifdef CONFIG_GRKERNSEC_AUDIT_GROUP + grsec_enable_group = 1; + grsec_audit_gid = CONFIG_GRKERNSEC_AUDIT_GID; +#endif +#ifdef CONFIG_GRKERNSEC_AUDIT_CHDIR + grsec_enable_chdir = 1; +#endif +#ifdef CONFIG_GRKERNSEC_AUDIT_IPC + grsec_enable_audit_ipc = 1; +#endif +#ifdef CONFIG_GRKERNSEC_AUDIT_MOUNT + grsec_enable_mount = 1; +#endif +#ifdef CONFIG_GRKERNSEC_LINK + grsec_enable_link = 1; +#endif +#ifdef CONFIG_GRKERNSEC_DMESG + grsec_enable_dmesg = 1; +#endif +#ifdef CONFIG_GRKERNSEC_FIFO + grsec_enable_fifo = 1; +#endif +#ifdef CONFIG_GRKERNSEC_EXECVE + grsec_enable_execve = 1; +#endif +#ifdef CONFIG_GRKERNSEC_EXECLOG + grsec_enable_execlog = 1; +#endif +#ifdef CONFIG_GRKERNSEC_SIGNAL + grsec_enable_signal = 1; +#endif +#ifdef CONFIG_GRKERNSEC_FORKFAIL + grsec_enable_forkfail = 1; +#endif +#ifdef CONFIG_GRKERNSEC_TIME + grsec_enable_time = 1; +#endif +#ifdef CONFIG_GRKERNSEC_RESLOG + grsec_resource_logging = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_FINDTASK + grsec_enable_chroot_findtask = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_UNIX + grsec_enable_chroot_unix = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_MOUNT + grsec_enable_chroot_mount = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_FCHDIR + grsec_enable_chroot_fchdir = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_SHMAT + grsec_enable_chroot_shmat = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_DOUBLE + grsec_enable_chroot_double = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_PIVOT + grsec_enable_chroot_pivot = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_CHDIR + grsec_enable_chroot_chdir = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_CHMOD + grsec_enable_chroot_chmod = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_MKNOD + grsec_enable_chroot_mknod = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_NICE + grsec_enable_chroot_nice = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_EXECLOG + grsec_enable_chroot_execlog = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_CAPS + grsec_enable_chroot_caps = 1; +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_SYSCTL + grsec_enable_chroot_sysctl = 1; +#endif +#ifdef CONFIG_GRKERNSEC_TPE + grsec_enable_tpe = 1; + grsec_tpe_gid = CONFIG_GRKERNSEC_TPE_GID; +#ifdef CONFIG_GRKERNSEC_TPE_ALL + grsec_enable_tpe_all = 1; +#endif +#endif +#ifdef CONFIG_GRKERNSEC_RANDPID + grsec_enable_randpid = 1; +#endif +#ifdef CONFIG_GRKERNSEC_SOCKET_ALL + grsec_enable_socket_all = 1; + grsec_socket_all_gid = CONFIG_GRKERNSEC_SOCKET_ALL_GID; +#endif +#ifdef CONFIG_GRKERNSEC_SOCKET_CLIENT + grsec_enable_socket_client = 1; + grsec_socket_client_gid = CONFIG_GRKERNSEC_SOCKET_CLIENT_GID; +#endif +#ifdef CONFIG_GRKERNSEC_SOCKET_SERVER + grsec_enable_socket_server = 1; + grsec_socket_server_gid = CONFIG_GRKERNSEC_SOCKET_SERVER_GID; +#endif +#endif + + return; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_ipc.c linux-2.6.17.13-g-v/grsecurity/grsec_ipc.c --- linux-2.6.17.13/grsecurity/grsec_ipc.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_ipc.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,81 @@ +#include +#include +#include +#include +#include +#include + +void +gr_log_msgget(const int ret, const int msgflg) +{ +#ifdef CONFIG_GRKERNSEC_AUDIT_IPC + if (((grsec_enable_group && in_group_p(grsec_audit_gid) && + grsec_enable_audit_ipc) || (grsec_enable_audit_ipc && + !grsec_enable_group)) && (ret >= 0) + && (msgflg & IPC_CREAT)) + gr_log_noargs(GR_DO_AUDIT, GR_MSGQ_AUDIT_MSG); +#endif + return; +} + +void +gr_log_msgrm(const uid_t uid, const uid_t cuid) +{ +#ifdef CONFIG_GRKERNSEC_AUDIT_IPC + if ((grsec_enable_group && in_group_p(grsec_audit_gid) && + grsec_enable_audit_ipc) || + (grsec_enable_audit_ipc && !grsec_enable_group)) + gr_log_int_int(GR_DO_AUDIT, GR_MSGQR_AUDIT_MSG, uid, cuid); +#endif + return; +} + +void +gr_log_semget(const int err, const int semflg) +{ +#ifdef CONFIG_GRKERNSEC_AUDIT_IPC + if (((grsec_enable_group && in_group_p(grsec_audit_gid) && + grsec_enable_audit_ipc) || (grsec_enable_audit_ipc && + !grsec_enable_group)) && (err >= 0) + && (semflg & IPC_CREAT)) + gr_log_noargs(GR_DO_AUDIT, GR_SEM_AUDIT_MSG); +#endif + return; +} + +void +gr_log_semrm(const uid_t uid, const uid_t cuid) +{ +#ifdef CONFIG_GRKERNSEC_AUDIT_IPC + if ((grsec_enable_group && in_group_p(grsec_audit_gid) && + grsec_enable_audit_ipc) || + (grsec_enable_audit_ipc && !grsec_enable_group)) + gr_log_int_int(GR_DO_AUDIT, GR_SEMR_AUDIT_MSG, uid, cuid); +#endif + return; +} + +void +gr_log_shmget(const int err, const int shmflg, const size_t size) +{ +#ifdef CONFIG_GRKERNSEC_AUDIT_IPC + if (((grsec_enable_group && in_group_p(grsec_audit_gid) && + grsec_enable_audit_ipc) || (grsec_enable_audit_ipc && + !grsec_enable_group)) && (err >= 0) + && (shmflg & IPC_CREAT)) + gr_log_int(GR_DO_AUDIT, GR_SHM_AUDIT_MSG, size); +#endif + return; +} + +void +gr_log_shmrm(const uid_t uid, const uid_t cuid) +{ +#ifdef CONFIG_GRKERNSEC_AUDIT_IPC + if ((grsec_enable_group && in_group_p(grsec_audit_gid) && + grsec_enable_audit_ipc) || + (grsec_enable_audit_ipc && !grsec_enable_group)) + gr_log_int_int(GR_DO_AUDIT, GR_SHMR_AUDIT_MSG, uid, cuid); +#endif + return; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_link.c linux-2.6.17.13-g-v/grsecurity/grsec_link.c --- linux-2.6.17.13/grsecurity/grsec_link.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_link.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,39 @@ +#include +#include +#include +#include +#include + +int +gr_handle_follow_link(const struct inode *parent, + const struct inode *inode, + const struct dentry *dentry, const struct vfsmount *mnt) +{ +#ifdef CONFIG_GRKERNSEC_LINK + if (grsec_enable_link && S_ISLNK(inode->i_mode) && + (parent->i_mode & S_ISVTX) && (parent->i_uid != inode->i_uid) && + (parent->i_mode & S_IWOTH) && (current->fsuid != inode->i_uid)) { + gr_log_fs_int2(GR_DONT_AUDIT, GR_SYMLINK_MSG, dentry, mnt, inode->i_uid, inode->i_gid); + return -EACCES; + } +#endif + return 0; +} + +int +gr_handle_hardlink(const struct dentry *dentry, + const struct vfsmount *mnt, + struct inode *inode, const int mode, const char *to) +{ +#ifdef CONFIG_GRKERNSEC_LINK + if (grsec_enable_link && current->fsuid != inode->i_uid && + (!S_ISREG(mode) || (mode & S_ISUID) || + ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP)) || + (generic_permission(inode, MAY_READ | MAY_WRITE, NULL))) && + !capable(CAP_FOWNER) && current->uid) { + gr_log_fs_int2_str(GR_DONT_AUDIT, GR_HARDLINK_MSG, dentry, mnt, inode->i_uid, inode->i_gid, to); + return -EPERM; + } +#endif + return 0; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_log.c linux-2.6.17.13-g-v/grsecurity/grsec_log.c --- linux-2.6.17.13/grsecurity/grsec_log.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_log.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,265 @@ +#include +#include +#include +#include +#include +#include + +#define BEGIN_LOCKS(x) \ + read_lock(&tasklist_lock); \ + read_lock(&grsec_exec_file_lock); \ + if (x != GR_DO_AUDIT) \ + spin_lock(&grsec_alert_lock); \ + else \ + spin_lock(&grsec_audit_lock) + +#define END_LOCKS(x) \ + if (x != GR_DO_AUDIT) \ + spin_unlock(&grsec_alert_lock); \ + else \ + spin_unlock(&grsec_audit_lock); \ + read_unlock(&grsec_exec_file_lock); \ + read_unlock(&tasklist_lock); \ + if (x == GR_DONT_AUDIT) \ + gr_handle_alertkill(current) + +enum { + FLOODING, + NO_FLOODING +}; + +extern char *gr_alert_log_fmt; +extern char *gr_audit_log_fmt; +extern char *gr_alert_log_buf; +extern char *gr_audit_log_buf; + +static int gr_log_start(int audit) +{ + char *loglevel = (audit == GR_DO_AUDIT) ? KERN_INFO : KERN_ALERT; + char *fmt = (audit == GR_DO_AUDIT) ? gr_audit_log_fmt : gr_alert_log_fmt; + char *buf = (audit == GR_DO_AUDIT) ? gr_audit_log_buf : gr_alert_log_buf; + + if (audit == GR_DO_AUDIT) + goto set_fmt; + + if (!grsec_alert_wtime || jiffies - grsec_alert_wtime > CONFIG_GRKERNSEC_FLOODTIME * HZ) { + grsec_alert_wtime = jiffies; + grsec_alert_fyet = 0; + } else if ((jiffies - grsec_alert_wtime < CONFIG_GRKERNSEC_FLOODTIME * HZ) && (grsec_alert_fyet < CONFIG_GRKERNSEC_FLOODBURST)) { + grsec_alert_fyet++; + } else if (grsec_alert_fyet == CONFIG_GRKERNSEC_FLOODBURST) { + grsec_alert_wtime = jiffies; + grsec_alert_fyet++; + printk(KERN_ALERT "grsec: more alerts, logging disabled for %d seconds\n", CONFIG_GRKERNSEC_FLOODTIME); + return FLOODING; + } else return FLOODING; + +set_fmt: + memset(buf, 0, PAGE_SIZE); + if (current->signal->curr_ip && gr_acl_is_enabled()) { + sprintf(fmt, "%s%s", loglevel, "grsec: From %u.%u.%u.%u: (%.64s:%c:%.950s) "); + snprintf(buf, PAGE_SIZE - 1, fmt, NIPQUAD(current->signal->curr_ip), current->role->rolename, gr_roletype_to_char(), current->acl->filename); + } else if (current->signal->curr_ip) { + sprintf(fmt, "%s%s", loglevel, "grsec: From %u.%u.%u.%u: "); + snprintf(buf, PAGE_SIZE - 1, fmt, NIPQUAD(current->signal->curr_ip)); + } else if (gr_acl_is_enabled()) { + sprintf(fmt, "%s%s", loglevel, "grsec: (%.64s:%c:%.950s) "); + snprintf(buf, PAGE_SIZE - 1, fmt, current->role->rolename, gr_roletype_to_char(), current->acl->filename); + } else { + sprintf(fmt, "%s%s", loglevel, "grsec: "); + strcpy(buf, fmt); + } + + return NO_FLOODING; +} + +static void gr_log_middle(int audit, const char *msg, va_list ap) +{ + char *buf = (audit == GR_DO_AUDIT) ? gr_audit_log_buf : gr_alert_log_buf; + unsigned int len = strlen(buf); + + vsnprintf(buf + len, PAGE_SIZE - len - 1, msg, ap); + + return; +} + +static void gr_log_middle_varargs(int audit, const char *msg, ...) +{ + char *buf = (audit == GR_DO_AUDIT) ? gr_audit_log_buf : gr_alert_log_buf; + unsigned int len = strlen(buf); + va_list ap; + + va_start(ap, msg); + vsnprintf(buf + len, PAGE_SIZE - len - 1, msg, ap); + va_end(ap); + + return; +} + +static void gr_log_end(int audit) +{ + char *buf = (audit == GR_DO_AUDIT) ? gr_audit_log_buf : gr_alert_log_buf; + unsigned int len = strlen(buf); + + snprintf(buf + len, PAGE_SIZE - len - 1, DEFAULTSECMSG, DEFAULTSECARGS(current)); + printk("%s\n", buf); + + return; +} + +void gr_log_varargs(int audit, const char *msg, int argtypes, ...) +{ + int logtype; + char *result = (audit == GR_DO_AUDIT) ? "successful" : "denied"; + char *str1, *str2, *str3; + int num1, num2; + unsigned long ulong1, ulong2; + struct dentry *dentry; + struct vfsmount *mnt; + struct file *file; + struct task_struct *task; + va_list ap; + + BEGIN_LOCKS(audit); + logtype = gr_log_start(audit); + if (logtype == FLOODING) { + END_LOCKS(audit); + return; + } + va_start(ap, argtypes); + switch (argtypes) { + case GR_TTYSNIFF: + task = va_arg(ap, struct task_struct *); + gr_log_middle_varargs(audit, msg, NIPQUAD(task->signal->curr_ip), gr_task_fullpath0(task), task->comm, task->pid, gr_parent_task_fullpath0(task), task->parent->comm, task->parent->pid); + break; + case GR_RBAC: + dentry = va_arg(ap, struct dentry *); + mnt = va_arg(ap, struct vfsmount *); + gr_log_middle_varargs(audit, msg, result, gr_to_filename(dentry, mnt)); + break; + case GR_RBAC_STR: + dentry = va_arg(ap, struct dentry *); + mnt = va_arg(ap, struct vfsmount *); + str1 = va_arg(ap, char *); + gr_log_middle_varargs(audit, msg, result, gr_to_filename(dentry, mnt), str1); + break; + case GR_STR_RBAC: + str1 = va_arg(ap, char *); + dentry = va_arg(ap, struct dentry *); + mnt = va_arg(ap, struct vfsmount *); + gr_log_middle_varargs(audit, msg, result, str1, gr_to_filename(dentry, mnt)); + break; + case GR_RBAC_MODE2: + dentry = va_arg(ap, struct dentry *); + mnt = va_arg(ap, struct vfsmount *); + str1 = va_arg(ap, char *); + str2 = va_arg(ap, char *); + gr_log_middle_varargs(audit, msg, result, gr_to_filename(dentry, mnt), str1, str2); + break; + case GR_RBAC_MODE3: + dentry = va_arg(ap, struct dentry *); + mnt = va_arg(ap, struct vfsmount *); + str1 = va_arg(ap, char *); + str2 = va_arg(ap, char *); + str3 = va_arg(ap, char *); + gr_log_middle_varargs(audit, msg, result, gr_to_filename(dentry, mnt), str1, str2, str3); + break; + case GR_FILENAME: + dentry = va_arg(ap, struct dentry *); + mnt = va_arg(ap, struct vfsmount *); + gr_log_middle_varargs(audit, msg, gr_to_filename(dentry, mnt)); + break; + case GR_STR_FILENAME: + str1 = va_arg(ap, char *); + dentry = va_arg(ap, struct dentry *); + mnt = va_arg(ap, struct vfsmount *); + gr_log_middle_varargs(audit, msg, str1, gr_to_filename(dentry, mnt)); + break; + case GR_FILENAME_STR: + dentry = va_arg(ap, struct dentry *); + mnt = va_arg(ap, struct vfsmount *); + str1 = va_arg(ap, char *); + gr_log_middle_varargs(audit, msg, gr_to_filename(dentry, mnt), str1); + break; + case GR_FILENAME_TWO_INT: + dentry = va_arg(ap, struct dentry *); + mnt = va_arg(ap, struct vfsmount *); + num1 = va_arg(ap, int); + num2 = va_arg(ap, int); + gr_log_middle_varargs(audit, msg, gr_to_filename(dentry, mnt), num1, num2); + break; + case GR_FILENAME_TWO_INT_STR: + dentry = va_arg(ap, struct dentry *); + mnt = va_arg(ap, struct vfsmount *); + num1 = va_arg(ap, int); + num2 = va_arg(ap, int); + str1 = va_arg(ap, char *); + gr_log_middle_varargs(audit, msg, gr_to_filename(dentry, mnt), num1, num2, str1); + break; + case GR_TEXTREL: + file = va_arg(ap, struct file *); + ulong1 = va_arg(ap, unsigned long); + ulong2 = va_arg(ap, unsigned long); + gr_log_middle_varargs(audit, msg, file ? gr_to_filename(file->f_dentry, file->f_vfsmnt) : "", ulong1, ulong2); + break; + case GR_PTRACE: + task = va_arg(ap, struct task_struct *); + gr_log_middle_varargs(audit, msg, task->exec_file ? gr_to_filename(task->exec_file->f_dentry, task->exec_file->f_vfsmnt) : "(none)", task->comm, task->pid); + break; + case GR_RESOURCE: + task = va_arg(ap, struct task_struct *); + ulong1 = va_arg(ap, unsigned long); + str1 = va_arg(ap, char *); + ulong2 = va_arg(ap, unsigned long); + gr_log_middle_varargs(audit, msg, ulong1, str1, ulong2, gr_task_fullpath(task), task->comm, task->pid, task->uid, task->euid, task->gid, task->egid, gr_parent_task_fullpath(task), task->parent->comm, task->parent->pid, task->parent->uid, task->parent->euid, task->parent->gid, task->parent->egid); + break; + case GR_CAP: + task = va_arg(ap, struct task_struct *); + str1 = va_arg(ap, char *); + gr_log_middle_varargs(audit, msg, str1, gr_task_fullpath(task), task->comm, task->pid, task->uid, task->euid, task->gid, task->egid, gr_parent_task_fullpath(task), task->parent->comm, task->parent->pid, task->parent->uid, task->parent->euid, task->parent->gid, task->parent->egid); + break; + case GR_SIG: + task = va_arg(ap, struct task_struct *); + num1 = va_arg(ap, int); + gr_log_middle_varargs(audit, msg, num1, gr_task_fullpath0(task), task->comm, task->pid, task->uid, task->euid, task->gid, task->egid, gr_parent_task_fullpath0(task), task->parent->comm, task->parent->pid, task->parent->uid, task->parent->euid, task->parent->gid, task->parent->egid); + break; + case GR_CRASH1: + task = va_arg(ap, struct task_struct *); + ulong1 = va_arg(ap, unsigned long); + gr_log_middle_varargs(audit, msg, gr_task_fullpath(task), task->comm, task->pid, task->uid, task->euid, task->gid, task->egid, gr_parent_task_fullpath(task), task->parent->comm, task->parent->pid, task->parent->uid, task->parent->euid, task->parent->gid, task->parent->egid, task->uid, ulong1); + break; + case GR_CRASH2: + task = va_arg(ap, struct task_struct *); + ulong1 = va_arg(ap, unsigned long); + gr_log_middle_varargs(audit, msg, gr_task_fullpath(task), task->comm, task->pid, task->uid, task->euid, task->gid, task->egid, gr_parent_task_fullpath(task), task->parent->comm, task->parent->pid, task->parent->uid, task->parent->euid, task->parent->gid, task->parent->egid, ulong1); + break; + case GR_PSACCT: + { + unsigned int wday, cday; + __u8 whr, chr; + __u8 wmin, cmin; + __u8 wsec, csec; + char cur_tty[64] = { 0 }; + char parent_tty[64] = { 0 }; + + task = va_arg(ap, struct task_struct *); + wday = va_arg(ap, unsigned int); + cday = va_arg(ap, unsigned int); + whr = va_arg(ap, int); + chr = va_arg(ap, int); + wmin = va_arg(ap, int); + cmin = va_arg(ap, int); + wsec = va_arg(ap, int); + csec = va_arg(ap, int); + ulong1 = va_arg(ap, unsigned long); + + gr_log_middle_varargs(audit, msg, gr_task_fullpath(task), task->comm, task->pid, NIPQUAD(task->signal->curr_ip), tty_name(task->signal->tty, cur_tty), task->uid, task->euid, task->gid, task->egid, wday, whr, wmin, wsec, cday, chr, cmin, csec, (task->flags & PF_SIGNALED) ? "killed by signal" : "exited", ulong1, gr_parent_task_fullpath(task), task->parent->comm, task->parent->pid, NIPQUAD(task->parent->signal->curr_ip), tty_name(task->parent->signal->tty, parent_tty), task->parent->uid, task->parent->euid, task->parent->gid, task->parent->egid); + } + break; + default: + gr_log_middle(audit, msg, ap); + } + va_end(ap); + gr_log_end(audit); + END_LOCKS(audit); +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_mem.c linux-2.6.17.13-g-v/grsecurity/grsec_mem.c --- linux-2.6.17.13/grsecurity/grsec_mem.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_mem.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,71 @@ +#include +#include +#include +#include +#include + +void +gr_handle_ioperm(void) +{ + gr_log_noargs(GR_DONT_AUDIT, GR_IOPERM_MSG); + return; +} + +void +gr_handle_iopl(void) +{ + gr_log_noargs(GR_DONT_AUDIT, GR_IOPL_MSG); + return; +} + +void +gr_handle_mem_write(void) +{ + gr_log_noargs(GR_DONT_AUDIT, GR_MEM_WRITE_MSG); + return; +} + +void +gr_handle_kmem_write(void) +{ + gr_log_noargs(GR_DONT_AUDIT, GR_KMEM_MSG); + return; +} + +void +gr_handle_open_port(void) +{ + gr_log_noargs(GR_DONT_AUDIT, GR_PORT_OPEN_MSG); + return; +} + +int +gr_handle_mem_mmap(const unsigned long offset, struct vm_area_struct *vma) +{ + unsigned long start, end; + + start = offset; + end = start + vma->vm_end - vma->vm_start; + + if (start > end) { + gr_log_noargs(GR_DONT_AUDIT, GR_MEM_MMAP_MSG); + return -EPERM; + } + + /* allowed ranges : ISA I/O BIOS */ + if ((start >= __pa(high_memory)) +#ifdef CONFIG_X86 + || (start >= 0x000a0000 && end <= 0x00100000) + || (start >= 0x00000000 && end <= 0x00001000) +#endif + ) + return 0; + + if (vma->vm_flags & VM_WRITE) { + gr_log_noargs(GR_DONT_AUDIT, GR_MEM_MMAP_MSG); + return -EPERM; + } else + vma->vm_flags &= ~VM_MAYWRITE; + + return 0; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_mount.c linux-2.6.17.13-g-v/grsecurity/grsec_mount.c --- linux-2.6.17.13/grsecurity/grsec_mount.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_mount.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,34 @@ +#include +#include +#include +#include + +void +gr_log_remount(const char *devname, const int retval) +{ +#ifdef CONFIG_GRKERNSEC_AUDIT_MOUNT + if (grsec_enable_mount && (retval >= 0)) + gr_log_str(GR_DO_AUDIT, GR_REMOUNT_AUDIT_MSG, devname ? devname : "none"); +#endif + return; +} + +void +gr_log_unmount(const char *devname, const int retval) +{ +#ifdef CONFIG_GRKERNSEC_AUDIT_MOUNT + if (grsec_enable_mount && (retval >= 0)) + gr_log_str(GR_DO_AUDIT, GR_UNMOUNT_AUDIT_MSG, devname ? devname : "none"); +#endif + return; +} + +void +gr_log_mount(const char *from, const char *to, const int retval) +{ +#ifdef CONFIG_GRKERNSEC_AUDIT_MOUNT + if (grsec_enable_mount && (retval >= 0)) + gr_log_str_str(GR_DO_AUDIT, GR_MOUNT_AUDIT_MSG, from, to); +#endif + return; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_rand.c linux-2.6.17.13-g-v/grsecurity/grsec_rand.c --- linux-2.6.17.13/grsecurity/grsec_rand.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_rand.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,26 @@ +#include +#include +#include +#include +#include + +extern int pid_max; + +int +gr_random_pid(void) +{ +#ifdef CONFIG_GRKERNSEC_RANDPID + int pid; + + if (grsec_enable_randpid && current->fs->root) { + /* return a pid in the range 1 ... pid_max - 1 + optimize this so we don't have to do a real division + */ + pid = 1 + (get_random_long() % pid_max); + if (pid == pid_max) + pid = pid_max - 1; + return pid; + } +#endif + return 0; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_sig.c linux-2.6.17.13-g-v/grsecurity/grsec_sig.c --- linux-2.6.17.13/grsecurity/grsec_sig.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_sig.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,59 @@ +#include +#include +#include +#include + +void +gr_log_signal(const int sig, const struct task_struct *t) +{ +#ifdef CONFIG_GRKERNSEC_SIGNAL + if (grsec_enable_signal && ((sig == SIGSEGV) || (sig == SIGILL) || + (sig == SIGABRT) || (sig == SIGBUS))) { + if (t->pid == current->pid) { + gr_log_int(GR_DONT_AUDIT_GOOD, GR_UNISIGLOG_MSG, sig); + } else { + gr_log_sig(GR_DONT_AUDIT_GOOD, GR_DUALSIGLOG_MSG, t, sig); + } + } +#endif + return; +} + +int +gr_handle_signal(const struct task_struct *p, const int sig) +{ +#ifdef CONFIG_GRKERNSEC + if (current->pid > 1 && gr_check_protected_task(p)) { + gr_log_sig(GR_DONT_AUDIT, GR_SIG_ACL_MSG, p, sig); + return -EPERM; + } else if (gr_pid_is_chrooted((struct task_struct *)p)) { + return -EPERM; + } +#endif + return 0; +} + +void gr_handle_brute_attach(struct task_struct *p) +{ +#ifdef CONFIG_GRKERNSEC_BRUTE + read_lock(&tasklist_lock); + read_lock(&grsec_exec_file_lock); + if (p->parent && p->parent->exec_file == p->exec_file) + p->parent->brute = 1; + read_unlock(&grsec_exec_file_lock); + read_unlock(&tasklist_lock); +#endif + return; +} + +void gr_handle_brute_check(void) +{ +#ifdef CONFIG_GRKERNSEC_BRUTE + if (current->brute) { + set_current_state(TASK_UNINTERRUPTIBLE); + schedule_timeout(30 * HZ); + } +#endif + return; +} + diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_sock.c linux-2.6.17.13-g-v/grsecurity/grsec_sock.c --- linux-2.6.17.13/grsecurity/grsec_sock.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_sock.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,263 @@ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#if defined(CONFIG_IP_NF_MATCH_STEALTH_MODULE) +extern struct sock *udp_v4_lookup(u32 saddr, u16 sport, u32 daddr, u16 dport, int dif); +EXPORT_SYMBOL(udp_v4_lookup); +#endif + +EXPORT_SYMBOL(gr_cap_rtnetlink); + +extern int gr_search_udp_recvmsg(const struct sock *sk, const struct sk_buff *skb); +extern int gr_search_udp_sendmsg(const struct sock *sk, const struct sockaddr_in *addr); + +EXPORT_SYMBOL(gr_search_udp_recvmsg); +EXPORT_SYMBOL(gr_search_udp_sendmsg); + +#ifdef CONFIG_UNIX_MODULE +EXPORT_SYMBOL(gr_acl_handle_unix); +EXPORT_SYMBOL(gr_acl_handle_mknod); +EXPORT_SYMBOL(gr_handle_chroot_unix); +EXPORT_SYMBOL(gr_handle_create); +#endif + +#ifdef CONFIG_GRKERNSEC +#define gr_conn_table_size 32749 +struct conn_table_entry { + struct conn_table_entry *next; + struct signal_struct *sig; +}; + +struct conn_table_entry *gr_conn_table[gr_conn_table_size]; +spinlock_t gr_conn_table_lock = SPIN_LOCK_UNLOCKED; + +extern const char * gr_socktype_to_name(unsigned char type); +extern const char * gr_proto_to_name(unsigned char proto); + +static __inline__ int +conn_hash(__u32 saddr, __u32 daddr, __u16 sport, __u16 dport, unsigned int size) +{ + return ((daddr + saddr + (sport << 8) + (dport << 16)) % size); +} + +static __inline__ int +conn_match(const struct signal_struct *sig, __u32 saddr, __u32 daddr, + __u16 sport, __u16 dport) +{ + if (unlikely(sig->gr_saddr == saddr && sig->gr_daddr == daddr && + sig->gr_sport == sport && sig->gr_dport == dport)) + return 1; + else + return 0; +} + +static void gr_add_to_task_ip_table_nolock(struct signal_struct *sig, struct conn_table_entry *newent) +{ + struct conn_table_entry **match; + unsigned int index; + + index = conn_hash(sig->gr_saddr, sig->gr_daddr, + sig->gr_sport, sig->gr_dport, + gr_conn_table_size); + + newent->sig = sig; + + match = &gr_conn_table[index]; + newent->next = *match; + *match = newent; + + return; +} + +static void gr_del_task_from_ip_table_nolock(struct signal_struct *sig) +{ + struct conn_table_entry *match, *last = NULL; + unsigned int index; + + index = conn_hash(sig->gr_saddr, sig->gr_daddr, + sig->gr_sport, sig->gr_dport, + gr_conn_table_size); + + match = gr_conn_table[index]; + while (match && !conn_match(match->sig, + sig->gr_saddr, sig->gr_daddr, sig->gr_sport, + sig->gr_dport)) { + last = match; + match = match->next; + } + + if (match) { + if (last) + last->next = match->next; + else + gr_conn_table[index] = NULL; + kfree(match); + } + + return; +} + +static struct signal_struct * gr_lookup_task_ip_table(__u32 saddr, __u32 daddr, + __u16 sport, __u16 dport) +{ + struct conn_table_entry *match; + unsigned int index; + + index = conn_hash(saddr, daddr, sport, dport, gr_conn_table_size); + + match = gr_conn_table[index]; + while (match && !conn_match(match->sig, saddr, daddr, sport, dport)) + match = match->next; + + if (match) + return match->sig; + else + return NULL; +} + +#endif + +void gr_update_task_in_ip_table(struct task_struct *task, const struct inet_sock *inet) +{ +#ifdef CONFIG_GRKERNSEC + struct signal_struct *sig = task->signal; + struct conn_table_entry *newent; + + newent = kmalloc(sizeof(struct conn_table_entry), GFP_ATOMIC); + if (newent == NULL) + return; + /* no bh lock needed since we are called with bh disabled */ + spin_lock(&gr_conn_table_lock); + gr_del_task_from_ip_table_nolock(sig); + sig->gr_saddr = inet->rcv_saddr; + sig->gr_daddr = inet->daddr; + sig->gr_sport = inet->sport; + sig->gr_dport = inet->dport; + gr_add_to_task_ip_table_nolock(sig, newent); + spin_unlock(&gr_conn_table_lock); +#endif + return; +} + +void gr_del_task_from_ip_table(struct task_struct *task) +{ +#ifdef CONFIG_GRKERNSEC + spin_lock(&gr_conn_table_lock); + gr_del_task_from_ip_table_nolock(task->signal); + spin_unlock(&gr_conn_table_lock); +#endif + return; +} + +void +gr_attach_curr_ip(const struct sock *sk) +{ +#ifdef CONFIG_GRKERNSEC + struct signal_struct *p, *set; + const struct inet_sock *inet = inet_sk(sk); + + if (unlikely(sk->sk_protocol != IPPROTO_TCP)) + return; + + set = current->signal; + + spin_lock_bh(&gr_conn_table_lock); + p = gr_lookup_task_ip_table(inet->daddr, inet->rcv_saddr, + inet->dport, inet->sport); + if (unlikely(p != NULL)) { + set->curr_ip = p->curr_ip; + set->used_accept = 1; + gr_del_task_from_ip_table_nolock(p); + spin_unlock_bh(&gr_conn_table_lock); + return; + } + spin_unlock_bh(&gr_conn_table_lock); + + set->curr_ip = inet->daddr; + set->used_accept = 1; +#endif + return; +} + +int +gr_handle_sock_all(const int family, const int type, const int protocol) +{ +#ifdef CONFIG_GRKERNSEC_SOCKET_ALL + if (grsec_enable_socket_all && in_group_p(grsec_socket_all_gid) && + (family != AF_UNIX) && (family != AF_LOCAL)) { + gr_log_int_str2(GR_DONT_AUDIT, GR_SOCK2_MSG, family, gr_socktype_to_name(type), gr_proto_to_name(protocol)); + return -EACCES; + } +#endif + return 0; +} + +int +gr_handle_sock_server(const struct sockaddr *sck) +{ +#ifdef CONFIG_GRKERNSEC_SOCKET_SERVER + if (grsec_enable_socket_server && + in_group_p(grsec_socket_server_gid) && + sck && (sck->sa_family != AF_UNIX) && + (sck->sa_family != AF_LOCAL)) { + gr_log_noargs(GR_DONT_AUDIT, GR_BIND_MSG); + return -EACCES; + } +#endif + return 0; +} + +int +gr_handle_sock_server_other(const struct sock *sck) +{ +#ifdef CONFIG_GRKERNSEC_SOCKET_SERVER + if (grsec_enable_socket_server && + in_group_p(grsec_socket_server_gid) && + sck && (sck->sk_family != AF_UNIX) && + (sck->sk_family != AF_LOCAL)) { + gr_log_noargs(GR_DONT_AUDIT, GR_BIND_MSG); + return -EACCES; + } +#endif + return 0; +} + +int +gr_handle_sock_client(const struct sockaddr *sck) +{ +#ifdef CONFIG_GRKERNSEC_SOCKET_CLIENT + if (grsec_enable_socket_client && in_group_p(grsec_socket_client_gid) && + sck && (sck->sa_family != AF_UNIX) && + (sck->sa_family != AF_LOCAL)) { + gr_log_noargs(GR_DONT_AUDIT, GR_CONNECT_MSG); + return -EACCES; + } +#endif + return 0; +} + +__u32 +gr_cap_rtnetlink(void) +{ +#ifdef CONFIG_GRKERNSEC + if (!gr_acl_is_enabled()) + return current->cap_effective; + else if (cap_raised(current->cap_effective, CAP_NET_ADMIN) && + gr_task_is_capable(current, CAP_NET_ADMIN)) + return current->cap_effective; + else + return 0; +#else + return current->cap_effective; +#endif +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_sysctl.c linux-2.6.17.13-g-v/grsecurity/grsec_sysctl.c --- linux-2.6.17.13/grsecurity/grsec_sysctl.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_sysctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,466 @@ +#include +#include +#include +#include +#include + +#ifdef CONFIG_GRKERNSEC_MODSTOP +int grsec_modstop; +#endif + +int +gr_handle_sysctl_mod(const char *dirname, const char *name, const int op) +{ +#ifdef CONFIG_GRKERNSEC_SYSCTL + if (!strcmp(dirname, "grsecurity") && grsec_lock && (op & 002)) { + gr_log_str(GR_DONT_AUDIT, GR_SYSCTL_MSG, name); + return -EACCES; + } +#endif +#ifdef CONFIG_GRKERNSEC_MODSTOP + if (!strcmp(dirname, "grsecurity") && !strcmp(name, "disable_modules") && + grsec_modstop && (op & 002)) { + gr_log_str(GR_DONT_AUDIT, GR_SYSCTL_MSG, name); + return -EACCES; + } +#endif + return 0; +} + +#if defined(CONFIG_GRKERNSEC_SYSCTL) || defined(CONFIG_GRKERNSEC_MODSTOP) +enum {GS_LINK=1, GS_FIFO, GS_EXECVE, GS_EXECLOG, GS_SIGNAL, +GS_FORKFAIL, GS_TIME, GS_CHROOT_SHMAT, GS_CHROOT_UNIX, GS_CHROOT_MNT, +GS_CHROOT_FCHDIR, GS_CHROOT_DBL, GS_CHROOT_PVT, GS_CHROOT_CD, GS_CHROOT_CM, +GS_CHROOT_MK, GS_CHROOT_NI, GS_CHROOT_EXECLOG, GS_CHROOT_CAPS, +GS_CHROOT_SYSCTL, GS_TPE, GS_TPE_GID, GS_TPE_ALL, GS_SIDCAPS, +GS_RANDPID, GS_SOCKET_ALL, GS_SOCKET_ALL_GID, GS_SOCKET_CLIENT, +GS_SOCKET_CLIENT_GID, GS_SOCKET_SERVER, GS_SOCKET_SERVER_GID, +GS_GROUP, GS_GID, GS_ACHDIR, GS_AMOUNT, GS_AIPC, GS_DMSG, +GS_TEXTREL, GS_FINDTASK, GS_SHM, GS_LOCK, GS_MODSTOP, GS_RESLOG}; + + +ctl_table grsecurity_table[] = { +#ifdef CONFIG_GRKERNSEC_SYSCTL +#ifdef CONFIG_GRKERNSEC_LINK + { + .ctl_name = GS_LINK, + .procname = "linking_restrictions", + .data = &grsec_enable_link, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_FIFO + { + .ctl_name = GS_FIFO, + .procname = "fifo_restrictions", + .data = &grsec_enable_fifo, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_EXECVE + { + .ctl_name = GS_EXECVE, + .procname = "execve_limiting", + .data = &grsec_enable_execve, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_EXECLOG + { + .ctl_name = GS_EXECLOG, + .procname = "exec_logging", + .data = &grsec_enable_execlog, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_SIGNAL + { + .ctl_name = GS_SIGNAL, + .procname = "signal_logging", + .data = &grsec_enable_signal, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_FORKFAIL + { + .ctl_name = GS_FORKFAIL, + .procname = "forkfail_logging", + .data = &grsec_enable_forkfail, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_TIME + { + .ctl_name = GS_TIME, + .procname = "timechange_logging", + .data = &grsec_enable_time, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_SHMAT + { + .ctl_name = GS_CHROOT_SHMAT, + .procname = "chroot_deny_shmat", + .data = &grsec_enable_chroot_shmat, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_UNIX + { + .ctl_name = GS_CHROOT_UNIX, + .procname = "chroot_deny_unix", + .data = &grsec_enable_chroot_unix, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_MOUNT + { + .ctl_name = GS_CHROOT_MNT, + .procname = "chroot_deny_mount", + .data = &grsec_enable_chroot_mount, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_FCHDIR + { + .ctl_name = GS_CHROOT_FCHDIR, + .procname = "chroot_deny_fchdir", + .data = &grsec_enable_chroot_fchdir, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_DOUBLE + { + .ctl_name = GS_CHROOT_DBL, + .procname = "chroot_deny_chroot", + .data = &grsec_enable_chroot_double, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_PIVOT + { + .ctl_name = GS_CHROOT_PVT, + .procname = "chroot_deny_pivot", + .data = &grsec_enable_chroot_pivot, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_CHDIR + { + .ctl_name = GS_CHROOT_CD, + .procname = "chroot_enforce_chdir", + .data = &grsec_enable_chroot_chdir, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_CHMOD + { + .ctl_name = GS_CHROOT_CM, + .procname = "chroot_deny_chmod", + .data = &grsec_enable_chroot_chmod, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_MKNOD + { + .ctl_name = GS_CHROOT_MK, + .procname = "chroot_deny_mknod", + .data = &grsec_enable_chroot_mknod, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_NICE + { + .ctl_name = GS_CHROOT_NI, + .procname = "chroot_restrict_nice", + .data = &grsec_enable_chroot_nice, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_EXECLOG + { + .ctl_name = GS_CHROOT_EXECLOG, + .procname = "chroot_execlog", + .data = &grsec_enable_chroot_execlog, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_CAPS + { + .ctl_name = GS_CHROOT_CAPS, + .procname = "chroot_caps", + .data = &grsec_enable_chroot_caps, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_SYSCTL + { + .ctl_name = GS_CHROOT_SYSCTL, + .procname = "chroot_deny_sysctl", + .data = &grsec_enable_chroot_sysctl, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_TPE + { + .ctl_name = GS_TPE, + .procname = "tpe", + .data = &grsec_enable_tpe, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, + { + .ctl_name = GS_TPE_GID, + .procname = "tpe_gid", + .data = &grsec_tpe_gid, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_TPE_ALL + { + .ctl_name = GS_TPE_ALL, + .procname = "tpe_restrict_all", + .data = &grsec_enable_tpe_all, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_RANDPID + { + .ctl_name = GS_RANDPID, + .procname = "rand_pids", + .data = &grsec_enable_randpid, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_SOCKET_ALL + { + .ctl_name = GS_SOCKET_ALL, + .procname = "socket_all", + .data = &grsec_enable_socket_all, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, + { + .ctl_name = GS_SOCKET_ALL_GID, + .procname = "socket_all_gid", + .data = &grsec_socket_all_gid, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_SOCKET_CLIENT + { + .ctl_name = GS_SOCKET_CLIENT, + .procname = "socket_client", + .data = &grsec_enable_socket_client, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, + { + .ctl_name = GS_SOCKET_CLIENT_GID, + .procname = "socket_client_gid", + .data = &grsec_socket_client_gid, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_SOCKET_SERVER + { + .ctl_name = GS_SOCKET_SERVER, + .procname = "socket_server", + .data = &grsec_enable_socket_server, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, + { + .ctl_name = GS_SOCKET_SERVER_GID, + .procname = "socket_server_gid", + .data = &grsec_socket_server_gid, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_AUDIT_GROUP + { + .ctl_name = GS_GROUP, + .procname = "audit_group", + .data = &grsec_enable_group, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, + { + .ctl_name = GS_GID, + .procname = "audit_gid", + .data = &grsec_audit_gid, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_AUDIT_CHDIR + { + .ctl_name = GS_ACHDIR, + .procname = "audit_chdir", + .data = &grsec_enable_chdir, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_AUDIT_MOUNT + { + .ctl_name = GS_AMOUNT, + .procname = "audit_mount", + .data = &grsec_enable_mount, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_AUDIT_IPC + { + .ctl_name = GS_AIPC, + .procname = "audit_ipc", + .data = &grsec_enable_audit_ipc, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_AUDIT_TEXTREL + { + .ctl_name = GS_TEXTREL, + .procname = "audit_textrel", + .data = &grsec_enable_audit_textrel, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_DMESG + { + .ctl_name = GS_DMSG, + .procname = "dmesg", + .data = &grsec_enable_dmesg, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_CHROOT_FINDTASK + { + .ctl_name = GS_FINDTASK, + .procname = "chroot_findtask", + .data = &grsec_enable_chroot_findtask, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_SHM + { + .ctl_name = GS_SHM, + .procname = "destroy_unused_shm", + .data = &grsec_enable_shm, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_RESLOG + { + .ctl_name = GS_RESLOG, + .procname = "resource_logging", + .data = &grsec_resource_logging, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif + { + .ctl_name = GS_LOCK, + .procname = "grsec_lock", + .data = &grsec_lock, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_GRKERNSEC_MODSTOP + { + .ctl_name = GS_MODSTOP, + .procname = "disable_modules", + .data = &grsec_modstop, + .maxlen = sizeof(int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, +#endif + { .ctl_name = 0 } +}; +#endif + +int gr_check_modstop(void) +{ +#ifdef CONFIG_GRKERNSEC_MODSTOP + if (grsec_modstop == 1) { + gr_log_noargs(GR_DONT_AUDIT, GR_STOPMOD_MSG); + return 1; + } +#endif + return 0; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_textrel.c linux-2.6.17.13-g-v/grsecurity/grsec_textrel.c --- linux-2.6.17.13/grsecurity/grsec_textrel.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_textrel.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,16 @@ +#include +#include +#include +#include +#include +#include + +void +gr_log_textrel(struct vm_area_struct * vma) +{ +#ifdef CONFIG_GRKERNSEC_AUDIT_TEXTREL + if (grsec_enable_audit_textrel) + gr_log_textrel_ulong_ulong(GR_DO_AUDIT, GR_TEXTREL_AUDIT_MSG, vma->vm_file, vma->vm_start, vma->vm_pgoff); +#endif + return; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_time.c linux-2.6.17.13-g-v/grsecurity/grsec_time.c --- linux-2.6.17.13/grsecurity/grsec_time.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_time.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,13 @@ +#include +#include +#include + +void +gr_log_timechange(void) +{ +#ifdef CONFIG_GRKERNSEC_TIME + if (grsec_enable_time) + gr_log_noargs(GR_DONT_AUDIT_GOOD, GR_TIME_MSG); +#endif + return; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsec_tpe.c linux-2.6.17.13-g-v/grsecurity/grsec_tpe.c --- linux-2.6.17.13/grsecurity/grsec_tpe.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsec_tpe.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,37 @@ +#include +#include +#include +#include +#include + +extern int gr_acl_tpe_check(void); + +int +gr_tpe_allow(const struct file *file) +{ +#ifdef CONFIG_GRKERNSEC + struct inode *inode = file->f_dentry->d_parent->d_inode; + + if (current->uid && ((grsec_enable_tpe && +#ifdef CONFIG_GRKERNSEC_TPE_INVERT + !in_group_p(grsec_tpe_gid) +#else + in_group_p(grsec_tpe_gid) +#endif + ) || gr_acl_tpe_check()) && + (inode->i_uid || (!inode->i_uid && ((inode->i_mode & S_IWGRP) || + (inode->i_mode & S_IWOTH))))) { + gr_log_fs_generic(GR_DONT_AUDIT, GR_EXEC_TPE_MSG, file->f_dentry, file->f_vfsmnt); + return 0; + } +#ifdef CONFIG_GRKERNSEC_TPE_ALL + if (current->uid && grsec_enable_tpe && grsec_enable_tpe_all && + ((inode->i_uid && (inode->i_uid != current->uid)) || + (inode->i_mode & S_IWGRP) || (inode->i_mode & S_IWOTH))) { + gr_log_fs_generic(GR_DONT_AUDIT, GR_EXEC_TPE_MSG, file->f_dentry, file->f_vfsmnt); + return 0; + } +#endif +#endif + return 1; +} diff -NurpP --minimal linux-2.6.17.13/grsecurity/grsum.c linux-2.6.17.13-g-v/grsecurity/grsum.c --- linux-2.6.17.13/grsecurity/grsum.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/grsecurity/grsum.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,59 @@ +#include +#include +#include +#include +#include +#include + + +#if !defined(CONFIG_CRYPTO) || defined(CONFIG_CRYPTO_MODULE) || !defined(CONFIG_CRYPTO_SHA256) || defined(CONFIG_CRYPTO_SHA256_MODULE) +#error "crypto and sha256 must be built into the kernel" +#endif + +int +chkpw(struct gr_arg *entry, unsigned char *salt, unsigned char *sum) +{ + char *p; + struct crypto_tfm *tfm; + unsigned char temp_sum[GR_SHA_LEN]; + struct scatterlist sg[2]; + volatile int retval = 0; + volatile int dummy = 0; + unsigned int i; + + tfm = crypto_alloc_tfm("sha256", 0); + if (tfm == NULL) { + /* should never happen, since sha256 should be built in */ + return 1; + } + + crypto_digest_init(tfm); + + p = salt; + sg[0].page = virt_to_page(p); + sg[0].offset = ((long) p & ~PAGE_MASK); + sg[0].length = GR_SALT_LEN; + + crypto_digest_update(tfm, sg, 1); + + p = entry->pw; + sg[0].page = virt_to_page(p); + sg[0].offset = ((long) p & ~PAGE_MASK); + sg[0].length = strlen(entry->pw); + + crypto_digest_update(tfm, sg, 1); + + crypto_digest_final(tfm, temp_sum); + + memset(entry->pw, 0, GR_PW_LEN); + + for (i = 0; i < GR_SHA_LEN; i++) + if (sum[i] != temp_sum[i]) + retval = 1; + else + dummy = 1; // waste a cycle + + crypto_free_tfm(tfm); + + return retval; +} diff -NurpP --minimal linux-2.6.17.13/include/asm-alpha/a.out.h linux-2.6.17.13-g-v/include/asm-alpha/a.out.h --- linux-2.6.17.13/include/asm-alpha/a.out.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-alpha/a.out.h 2006-09-20 15:00:12.000000000 +0200 @@ -98,7 +98,7 @@ struct exec set_personality (((BFPM->sh_bang || EX.ah.entry < 0x100000000L \ ? ADDR_LIMIT_32BIT : 0) | PER_OSF4)) -#define STACK_TOP \ +#define __STACK_TOP \ (current->personality & ADDR_LIMIT_32BIT ? 0x80000000 : 0x00120000000UL) #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-alpha/elf.h linux-2.6.17.13-g-v/include/asm-alpha/elf.h --- linux-2.6.17.13/include/asm-alpha/elf.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-alpha/elf.h 2006-09-20 15:00:12.000000000 +0200 @@ -91,6 +91,17 @@ typedef elf_fpreg_t elf_fpregset_t[ELF_N #define ELF_ET_DYN_BASE (TASK_UNMAPPED_BASE + 0x1000000) +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) ((tsk)->personality & ADDR_LIMIT_32BIT ? 0x10000 : 0x120000000UL) + +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) ((tsk)->personality & ADDR_LIMIT_32BIT ? 14 : 28) +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) ((tsk)->personality & ADDR_LIMIT_32BIT ? 14 : 28) +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) ((tsk)->personality & ADDR_LIMIT_32BIT ? 14 : 19) +#endif + /* $0 is set by ld.so to a pointer to a function which might be registered using atexit. This provides a mean for the dynamic linker to call DT_FINI functions for shared libraries that have diff -NurpP --minimal linux-2.6.17.13/include/asm-alpha/kmap_types.h linux-2.6.17.13-g-v/include/asm-alpha/kmap_types.h --- linux-2.6.17.13/include/asm-alpha/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-alpha/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -25,7 +25,8 @@ D(9) KM_IRQ0, D(10) KM_IRQ1, D(11) KM_SOFTIRQ0, D(12) KM_SOFTIRQ1, -D(13) KM_TYPE_NR +D(13) KM_CLEARPAGE, +D(14) KM_TYPE_NR }; #undef D diff -NurpP --minimal linux-2.6.17.13/include/asm-alpha/page.h linux-2.6.17.13-g-v/include/asm-alpha/page.h --- linux-2.6.17.13/include/asm-alpha/page.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-alpha/page.h 2006-09-20 15:00:12.000000000 +0200 @@ -93,6 +93,16 @@ typedef unsigned long pgprot_t; #define VM_DATA_DEFAULT_FLAGS (VM_READ | VM_WRITE | VM_EXEC | \ VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC) + +#ifdef CONFIG_PAX_PAGEEXEC +#ifdef CONFIG_PAX_MPROTECT +#define __VM_STACK_FLAGS (((current->mm->pax_flags & MF_PAX_MPROTECT)?0:VM_MAYEXEC) | \ + ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#else +#define __VM_STACK_FLAGS (VM_MAYEXEC | ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#endif +#endif + #endif /* __KERNEL__ */ #include diff -NurpP --minimal linux-2.6.17.13/include/asm-alpha/pgtable.h linux-2.6.17.13-g-v/include/asm-alpha/pgtable.h --- linux-2.6.17.13/include/asm-alpha/pgtable.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-alpha/pgtable.h 2006-09-20 15:00:12.000000000 +0200 @@ -102,6 +102,17 @@ struct vm_area_struct; #define PAGE_SHARED __pgprot(_PAGE_VALID | __ACCESS_BITS) #define PAGE_COPY __pgprot(_PAGE_VALID | __ACCESS_BITS | _PAGE_FOW) #define PAGE_READONLY __pgprot(_PAGE_VALID | __ACCESS_BITS | _PAGE_FOW) + +#ifdef CONFIG_PAX_PAGEEXEC +# define PAGE_SHARED_NOEXEC __pgprot(_PAGE_VALID | __ACCESS_BITS | _PAGE_FOE) +# define PAGE_COPY_NOEXEC __pgprot(_PAGE_VALID | __ACCESS_BITS | _PAGE_FOW | _PAGE_FOE) +# define PAGE_READONLY_NOEXEC __pgprot(_PAGE_VALID | __ACCESS_BITS | _PAGE_FOW | _PAGE_FOE) +#else +# define PAGE_SHARED_NOEXEC PAGE_SHARED +# define PAGE_COPY_NOEXEC PAGE_COPY +# define PAGE_READONLY_NOEXEC PAGE_READONLY +#endif + #define PAGE_KERNEL __pgprot(_PAGE_VALID | _PAGE_ASM | _PAGE_KRE | _PAGE_KWE) #define _PAGE_NORMAL(x) __pgprot(_PAGE_VALID | __ACCESS_BITS | (x)) diff -NurpP --minimal linux-2.6.17.13/include/asm-arm/a.out.h linux-2.6.17.13-g-v/include/asm-arm/a.out.h --- linux-2.6.17.13/include/asm-arm/a.out.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-arm/a.out.h 2006-09-20 15:00:12.000000000 +0200 @@ -28,7 +28,7 @@ struct exec #define M_ARM 103 #ifdef __KERNEL__ -#define STACK_TOP ((current->personality == PER_LINUX_32BIT) ? \ +#define __STACK_TOP ((current->personality == PER_LINUX_32BIT) ? \ TASK_SIZE : TASK_SIZE_26) #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-arm/elf.h linux-2.6.17.13-g-v/include/asm-arm/elf.h --- linux-2.6.17.13/include/asm-arm/elf.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-arm/elf.h 2006-09-20 15:00:12.000000000 +0200 @@ -56,6 +56,17 @@ typedef struct user_fp elf_fpregset_t; #define ELF_ET_DYN_BASE (2 * TASK_SIZE / 3) +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) 0x00008000UL + +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) ((tsk->personality == PER_LINUX_32BIT) ? 16 : 10) +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) ((tsk->personality == PER_LINUX_32BIT) ? 16 : 10) +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) ((tsk->personality == PER_LINUX_32BIT) ? 16 : 10) +#endif + /* When the program starts, a1 contains a pointer to a function to be registered with atexit, as per the SVR4 ABI. A value of 0 means we have no such handler. */ diff -NurpP --minimal linux-2.6.17.13/include/asm-arm/kmap_types.h linux-2.6.17.13-g-v/include/asm-arm/kmap_types.h --- linux-2.6.17.13/include/asm-arm/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-arm/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -18,6 +18,7 @@ enum km_type { KM_IRQ1, KM_SOFTIRQ0, KM_SOFTIRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-arm/tlb.h linux-2.6.17.13-g-v/include/asm-arm/tlb.h --- linux-2.6.17.13/include/asm-arm/tlb.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-arm/tlb.h 2006-09-20 15:00:12.000000000 +0200 @@ -28,6 +28,7 @@ #else /* !CONFIG_MMU */ #include +#include /* * TLB handling. This allows us to remove pages from the page diff -NurpP --minimal linux-2.6.17.13/include/asm-arm26/kmap_types.h linux-2.6.17.13-g-v/include/asm-arm26/kmap_types.h --- linux-2.6.17.13/include/asm-arm26/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-arm26/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -6,7 +6,8 @@ */ enum km_type { KM_IRQ0, - KM_USER1 + KM_USER1, + KM_CLEARPAGE }; #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-arm26/tlb.h linux-2.6.17.13-g-v/include/asm-arm26/tlb.h --- linux-2.6.17.13/include/asm-arm26/tlb.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-arm26/tlb.h 2006-09-20 15:00:12.000000000 +0200 @@ -3,6 +3,7 @@ #include #include +#include /* * TLB handling. This allows us to remove pages from the page diff -NurpP --minimal linux-2.6.17.13/include/asm-arm26/unistd.h linux-2.6.17.13-g-v/include/asm-arm26/unistd.h --- linux-2.6.17.13/include/asm-arm26/unistd.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-arm26/unistd.h 2006-09-20 15:00:12.000000000 +0200 @@ -304,6 +304,8 @@ #define __NR_mq_getsetattr (__NR_SYSCALL_BASE+279) #define __NR_waitid (__NR_SYSCALL_BASE+280) +#define __NR_vserver (__NR_SYSCALL_BASE+313) + /* * The following SWIs are ARM private. FIXME - make appropriate for arm26 */ diff -NurpP --minimal linux-2.6.17.13/include/asm-cris/kmap_types.h linux-2.6.17.13-g-v/include/asm-cris/kmap_types.h --- linux-2.6.17.13/include/asm-cris/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-cris/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -19,6 +19,7 @@ enum km_type { KM_IRQ1, KM_SOFTIRQ0, KM_SOFTIRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-frv/kmap_types.h linux-2.6.17.13-g-v/include/asm-frv/kmap_types.h --- linux-2.6.17.13/include/asm-frv/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-frv/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -23,6 +23,7 @@ enum km_type { KM_IRQ1, KM_SOFTIRQ0, KM_SOFTIRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-generic/tlb.h linux-2.6.17.13-g-v/include/asm-generic/tlb.h --- linux-2.6.17.13/include/asm-generic/tlb.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-generic/tlb.h 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,7 @@ #include #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/include/asm-h8300/kmap_types.h linux-2.6.17.13-g-v/include/asm-h8300/kmap_types.h --- linux-2.6.17.13/include/asm-h8300/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-h8300/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,7 @@ enum km_type { KM_IRQ1, KM_SOFTIRQ0, KM_SOFTIRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/a.out.h linux-2.6.17.13-g-v/include/asm-i386/a.out.h --- linux-2.6.17.13/include/asm-i386/a.out.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/a.out.h 2006-09-20 15:00:12.000000000 +0200 @@ -19,7 +19,11 @@ struct exec #ifdef __KERNEL__ -#define STACK_TOP TASK_SIZE +#ifdef CONFIG_PAX_SEGMEXEC +#define __STACK_TOP ((current->mm->pax_flags & MF_PAX_SEGMEXEC)?TASK_SIZE/2:TASK_SIZE) +#else +#define __STACK_TOP TASK_SIZE +#endif #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/alternative.h linux-2.6.17.13-g-v/include/asm-i386/alternative.h --- linux-2.6.17.13/include/asm-i386/alternative.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/alternative.h 2006-09-20 15:00:12.000000000 +0200 @@ -47,7 +47,7 @@ extern void alternatives_smp_switch(int " .byte 662b-661b\n" /* sourcelen */ \ " .byte 664f-663f\n" /* replacementlen */ \ ".previous\n" \ - ".section .altinstr_replacement,\"ax\"\n" \ + ".section .altinstr_replacement,\"a\"\n" \ "663:\n\t" newinstr "\n664:\n" /* replacement */\ ".previous" :: "i" (feature) : "memory") @@ -71,7 +71,7 @@ extern void alternatives_smp_switch(int " .byte 662b-661b\n" /* sourcelen */ \ " .byte 664f-663f\n" /* replacementlen */ \ ".previous\n" \ - ".section .altinstr_replacement,\"ax\"\n" \ + ".section .altinstr_replacement,\"a\"\n" \ "663:\n\t" newinstr "\n664:\n" /* replacement */\ ".previous" :: "i" (feature), ##input) @@ -110,7 +110,7 @@ extern void alternatives_smp_switch(int " .byte 662b-661b\n" /* sourcelen */ \ " .byte 664f-663f\n" /* replacementlen */ \ ".previous\n" \ - ".section .smp_altinstr_replacement,\"awx\"\n" \ + ".section .smp_altinstr_replacement,\"aw\"\n" \ "663:\n\t" upinstr "\n" /* replacement */ \ "664:\n\t.fill 662b-661b,1,0x42\n" /* space for original */ \ ".previous" : args) diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/auxvec.h linux-2.6.17.13-g-v/include/asm-i386/auxvec.h --- linux-2.6.17.13/include/asm-i386/auxvec.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/auxvec.h 2006-09-20 15:00:12.000000000 +0200 @@ -5,7 +5,9 @@ * Architecture-neutral AT_ values in 0-17, leave some room * for more of them, start the x86-specific ones at 32. */ +#ifndef CONFIG_PAX_NOVSYSCALL #define AT_SYSINFO 32 #define AT_SYSINFO_EHDR 33 +#endif #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/bug.h linux-2.6.17.13-g-v/include/asm-i386/bug.h --- linux-2.6.17.13/include/asm-i386/bug.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/bug.h 2006-09-20 15:00:12.000000000 +0200 @@ -12,10 +12,9 @@ #ifdef CONFIG_BUG #define HAVE_ARCH_BUG #ifdef CONFIG_DEBUG_BUGVERBOSE -#define BUG() \ - __asm__ __volatile__( "ud2\n" \ - "\t.word %c0\n" \ - "\t.long %c1\n" \ +#define BUG() \ + __asm__ __volatile__( "ud2\n\t" \ + "ljmp %0, %1\n\t" \ : : "i" (__LINE__), "i" (__FILE__)) #else #define BUG() __asm__ __volatile__("ud2\n") diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/checksum.h linux-2.6.17.13-g-v/include/asm-i386/checksum.h --- linux-2.6.17.13/include/asm-i386/checksum.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/checksum.h 2006-09-20 15:00:12.000000000 +0200 @@ -30,6 +30,12 @@ asmlinkage unsigned int csum_partial(con asmlinkage unsigned int csum_partial_copy_generic(const unsigned char *src, unsigned char *dst, int len, int sum, int *src_err_ptr, int *dst_err_ptr); +asmlinkage unsigned int csum_partial_copy_generic_to_user(const unsigned char *src, unsigned char *dst, + int len, int sum, int *src_err_ptr, int *dst_err_ptr); + +asmlinkage unsigned int csum_partial_copy_generic_from_user(const unsigned char *src, unsigned char *dst, + int len, int sum, int *src_err_ptr, int *dst_err_ptr); + /* * Note: when you get a NULL pointer exception here this means someone * passed in an incorrect kernel address to one of these functions. @@ -49,7 +55,7 @@ unsigned int csum_partial_copy_from_user int len, int sum, int *err_ptr) { might_sleep(); - return csum_partial_copy_generic((__force unsigned char *)src, dst, + return csum_partial_copy_generic_from_user((__force unsigned char *)src, dst, len, sum, err_ptr, NULL); } @@ -183,7 +189,7 @@ static __inline__ unsigned int csum_and_ { might_sleep(); if (access_ok(VERIFY_WRITE, dst, len)) - return csum_partial_copy_generic(src, (__force unsigned char *)dst, len, sum, NULL, err_ptr); + return csum_partial_copy_generic_to_user(src, (__force unsigned char *)dst, len, sum, NULL, err_ptr); if (len) *err_ptr = -EFAULT; diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/desc.h linux-2.6.17.13-g-v/include/asm-i386/desc.h --- linux-2.6.17.13/include/asm-i386/desc.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/desc.h 2006-09-20 15:00:12.000000000 +0200 @@ -10,11 +10,13 @@ #include #include -#include +#include #include +#include +#include -extern struct desc_struct cpu_gdt_table[GDT_ENTRIES]; +extern struct desc_struct cpu_gdt_table[NR_CPUS][PAGE_SIZE / sizeof(struct desc_struct)]; DECLARE_PER_CPU(unsigned char, cpu_16bit_stack[CPU_16BIT_STACK_SIZE]); @@ -24,13 +26,53 @@ struct Xgt_desc_struct { unsigned short pad; } __attribute__ ((packed)); -extern struct Xgt_desc_struct idt_descr; -DECLARE_PER_CPU(struct Xgt_desc_struct, cpu_gdt_descr); - +extern struct Xgt_desc_struct idt_descr, cpu_gdt_descr[NR_CPUS]; static inline struct desc_struct *get_cpu_gdt_table(unsigned int cpu) { - return (struct desc_struct *)per_cpu(cpu_gdt_descr, cpu).address; + return cpu_gdt_table[cpu]; +} + +#define pax_open_kernel(cr0) \ +do { \ + typecheck(unsigned long,cr0); \ + preempt_disable(); \ + cr0 = read_cr0(); \ + write_cr0(cr0 & ~0x10000UL); \ +} while(0) + +#define pax_close_kernel(cr0) \ +do { \ + typecheck(unsigned long,cr0); \ + write_cr0(cr0); \ + preempt_enable_no_resched(); \ +} while(0) + +static inline void set_user_cs(struct mm_struct *mm, int cpu) +{ +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) + unsigned long base = mm->context.user_cs_base; + unsigned long limit = mm->context.user_cs_limit; + +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + + if (likely(limit)) { + limit -= 1UL; + limit >>= 12; + } + + get_cpu_gdt_table(cpu)[GDT_ENTRY_DEFAULT_USER_CS].a = (limit & 0xFFFFUL) | (base << 16); + get_cpu_gdt_table(cpu)[GDT_ENTRY_DEFAULT_USER_CS].b = (limit & 0xF0000UL) | 0xC0FB00UL | (base & 0xFF000000UL) | ((base >> 16) & 0xFFUL); + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + +#endif } #define load_TR_desc() __asm__ __volatile__("ltr %w0"::"q" (GDT_ENTRY_TSS*8)) @@ -50,7 +92,7 @@ static inline struct desc_struct *get_cp * This is the ldt that every process will get unless we need * something other than this. */ -extern struct desc_struct default_ldt[]; +extern const struct desc_struct default_ldt[]; extern void set_intr_gate(unsigned int irq, void * addr); #define _set_tssldt_desc(n,addr,limit,type) \ @@ -64,7 +106,7 @@ __asm__ __volatile__ ("movw %w3,0(%2)\n\ "rorl $16,%1" \ : "=m"(*(n)) : "q" (addr), "r"(n), "ir"(limit), "i"(type)) -static inline void __set_tss_desc(unsigned int cpu, unsigned int entry, void *addr) +static inline void __set_tss_desc(unsigned int cpu, unsigned int entry, const void *addr) { _set_tssldt_desc(&get_cpu_gdt_table(cpu)[entry], (int)addr, offsetof(struct tss_struct, __cacheline_filler) - 1, 0x89); @@ -72,9 +114,26 @@ static inline void __set_tss_desc(unsign #define set_tss_desc(cpu,addr) __set_tss_desc(cpu, GDT_ENTRY_TSS, addr) -static inline void set_ldt_desc(unsigned int cpu, void *addr, unsigned int size) +static inline void __set_ldt_desc(unsigned int cpu, const void *addr, unsigned int size) +{ + _set_tssldt_desc(&get_cpu_gdt_table(cpu)[GDT_ENTRY_LDT], (int)addr, ((size << 3)-1), 0x82); +} + +static inline void set_ldt_desc(unsigned int cpu, const void *addr, unsigned int size) { + +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; + + pax_open_kernel(cr0); +#endif + _set_tssldt_desc(&get_cpu_gdt_table(cpu)[GDT_ENTRY_LDT], (int)addr, ((size << 3)-1), 0x82); + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + } #define LDT_entry_a(info) \ @@ -90,7 +149,7 @@ static inline void set_ldt_desc(unsigned ((info)->seg_32bit << 22) | \ ((info)->limit_in_pages << 23) | \ ((info)->useable << 20) | \ - 0x7000) + 0x7100) #define LDT_empty(info) (\ (info)->base_addr == 0 && \ @@ -134,7 +193,7 @@ static inline void clear_LDT(void) */ static inline void load_LDT_nolock(mm_context_t *pc, int cpu) { - void *segments = pc->ldt; + const void *segments = pc->ldt; int count = pc->size; if (likely(!count)) { @@ -162,6 +221,22 @@ static inline unsigned long get_desc_bas return base; } +static inline void _load_LDT(mm_context_t *pc) +{ + int cpu = get_cpu(); + const void *segments = pc->ldt; + int count = pc->size; + + if (likely(!count)) { + segments = &default_ldt[0]; + count = 5; + } + + __set_ldt_desc(cpu, segments, count); + load_LDT_desc(); + put_cpu(); +} + #endif /* !__ASSEMBLY__ */ #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/elf.h linux-2.6.17.13-g-v/include/asm-i386/elf.h --- linux-2.6.17.13/include/asm-i386/elf.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/elf.h 2006-09-20 15:00:12.000000000 +0200 @@ -71,7 +71,22 @@ typedef struct user_fxsr_struct elf_fpxr the loader. We need to make sure that it is out of the way of the program that it will "exec", and that there is sufficient room for the brk. */ +#ifdef CONFIG_PAX_SEGMEXEC +#define ELF_ET_DYN_BASE ((current->mm->pax_flags & MF_PAX_SEGMEXEC) ? SEGMEXEC_TASK_SIZE/3*2 : TASK_SIZE/3*2) +#else #define ELF_ET_DYN_BASE (TASK_SIZE / 3 * 2) +#endif + +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) 0x10000000UL + +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) ((tsk)->mm->pax_flags & MF_PAX_SEGMEXEC ? 15 : 16) +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) 15 +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) ((tsk)->mm->pax_flags & MF_PAX_SEGMEXEC ? 15 : 16) +#endif /* regs is struct pt_regs, pr_reg is elf_gregset_t (which is now struct_user_regs, they are different) */ @@ -108,7 +123,7 @@ typedef struct user_fxsr_struct elf_fpxr For the moment, we have only optimizations for the Intel generations, but that could change... */ -#define ELF_PLATFORM (system_utsname.machine) +#define ELF_PLATFORM (vx_new_uts(machine)) #ifdef __KERNEL__ #define SET_PERSONALITY(ex, ibcs2) do { } while (0) @@ -131,7 +146,14 @@ extern int dump_task_extended_fpu (struc #define VSYSCALL_BASE (__fix_to_virt(FIX_VSYSCALL)) #define VSYSCALL_EHDR ((const struct elfhdr *) VSYSCALL_BASE) + +#ifndef CONFIG_PAX_NOVSYSCALL +#ifdef CONFIG_PAX_SEGMEXEC +#define VSYSCALL_ENTRY ((current->mm->pax_flags & MF_PAX_SEGMEXEC) ? (unsigned long) &__kernel_vsyscall - SEGMEXEC_TASK_SIZE : (unsigned long) &__kernel_vsyscall) +#else #define VSYSCALL_ENTRY ((unsigned long) &__kernel_vsyscall) +#endif + extern void __kernel_vsyscall; #define ARCH_DLINFO \ @@ -187,3 +209,5 @@ do { \ #endif #endif + +#endif diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/i387.h linux-2.6.17.13-g-v/include/asm-i386/i387.h --- linux-2.6.17.13/include/asm-i386/i387.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/i387.h 2006-09-20 15:00:12.000000000 +0200 @@ -40,13 +40,8 @@ extern void kernel_fpu_begin(void); #define kernel_fpu_end() do { stts(); preempt_enable(); } while(0) /* We need a safe address that is cheap to find and that is already - in L1 during context switch. The best choices are unfortunately - different for UP and SMP */ -#ifdef CONFIG_SMP -#define safe_address (__per_cpu_offset[0]) -#else -#define safe_address (kstat_cpu(0).cpustat.user) -#endif + in L1 during context switch. */ +#define safe_address (init_tss[smp_processor_id()].esp0) /* * These must be called with preempt disabled diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/kmap_types.h linux-2.6.17.13-g-v/include/asm-i386/kmap_types.h --- linux-2.6.17.13/include/asm-i386/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -23,7 +23,8 @@ D(9) KM_IRQ0, D(10) KM_IRQ1, D(11) KM_SOFTIRQ0, D(12) KM_SOFTIRQ1, -D(13) KM_TYPE_NR +D(13) KM_CLEARPAGE, +D(14) KM_TYPE_NR }; #undef D diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/mach-default/apm.h linux-2.6.17.13-g-v/include/asm-i386/mach-default/apm.h --- linux-2.6.17.13/include/asm-i386/mach-default/apm.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/mach-default/apm.h 2006-09-20 15:00:12.000000000 +0200 @@ -36,7 +36,7 @@ static inline void apm_bios_call_asm(u32 __asm__ __volatile__(APM_DO_ZERO_SEGS "pushl %%edi\n\t" "pushl %%ebp\n\t" - "lcall *%%cs:apm_bios_entry\n\t" + "lcall *%%ss:apm_bios_entry\n\t" "setc %%al\n\t" "popl %%ebp\n\t" "popl %%edi\n\t" @@ -60,7 +60,7 @@ static inline u8 apm_bios_call_simple_as __asm__ __volatile__(APM_DO_ZERO_SEGS "pushl %%edi\n\t" "pushl %%ebp\n\t" - "lcall *%%cs:apm_bios_entry\n\t" + "lcall *%%ss:apm_bios_entry\n\t" "setc %%bl\n\t" "popl %%ebp\n\t" "popl %%edi\n\t" diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/mach-default/do_timer.h linux-2.6.17.13-g-v/include/asm-i386/mach-default/do_timer.h --- linux-2.6.17.13/include/asm-i386/mach-default/do_timer.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/mach-default/do_timer.h 2006-09-20 15:00:12.000000000 +0200 @@ -18,7 +18,7 @@ static inline void do_timer_interrupt_ho { do_timer(regs); #ifndef CONFIG_SMP - update_process_times(user_mode_vm(regs)); + update_process_times(user_mode(regs)); #endif /* * In the SMP case we use the local APIC timer interrupt to do the diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/mach-visws/do_timer.h linux-2.6.17.13-g-v/include/asm-i386/mach-visws/do_timer.h --- linux-2.6.17.13/include/asm-i386/mach-visws/do_timer.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/mach-visws/do_timer.h 2006-09-20 15:00:12.000000000 +0200 @@ -11,7 +11,7 @@ static inline void do_timer_interrupt_ho do_timer(regs); #ifndef CONFIG_SMP - update_process_times(user_mode_vm(regs)); + update_process_times(user_mode(regs)); #endif /* * In the SMP case we use the local APIC timer interrupt to do the diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/mach-voyager/do_timer.h linux-2.6.17.13-g-v/include/asm-i386/mach-voyager/do_timer.h --- linux-2.6.17.13/include/asm-i386/mach-voyager/do_timer.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/mach-voyager/do_timer.h 2006-09-20 15:00:12.000000000 +0200 @@ -5,7 +5,7 @@ static inline void do_timer_interrupt_ho { do_timer(regs); #ifndef CONFIG_SMP - update_process_times(user_mode_vm(regs)); + update_process_times(user_mode(regs)); #endif voyager_timer_interrupt(regs); diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/mman.h linux-2.6.17.13-g-v/include/asm-i386/mman.h --- linux-2.6.17.13/include/asm-i386/mman.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/mman.h 2006-09-20 15:00:12.000000000 +0200 @@ -11,6 +11,10 @@ #define MAP_POPULATE 0x8000 /* populate (prefault) pagetables */ #define MAP_NONBLOCK 0x10000 /* do not block on IO */ +#ifdef CONFIG_PAX_SEGMEXEC +#define MAP_MIRROR 0x20000 +#endif + #define MCL_CURRENT 1 /* lock all current mappings */ #define MCL_FUTURE 2 /* lock all future mappings */ diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/mmu.h linux-2.6.17.13-g-v/include/asm-i386/mmu.h --- linux-2.6.17.13/include/asm-i386/mmu.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/mmu.h 2006-09-20 15:00:12.000000000 +0200 @@ -12,6 +12,17 @@ typedef struct { int size; struct semaphore sem; void *ldt; + +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) + unsigned long user_cs_base; + unsigned long user_cs_limit; + +#if defined(CONFIG_PAX_PAGEEXEC) && defined(CONFIG_SMP) + cpumask_t cpu_user_cs_mask; +#endif + +#endif + } mm_context_t; #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/mmu_context.h linux-2.6.17.13-g-v/include/asm-i386/mmu_context.h --- linux-2.6.17.13/include/asm-i386/mmu_context.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/mmu_context.h 2006-09-20 15:00:12.000000000 +0200 @@ -46,6 +46,18 @@ static inline void switch_mm(struct mm_s */ if (unlikely(prev->context.ldt != next->context.ldt)) load_LDT_nolock(&next->context, cpu); + +#if defined(CONFIG_PAX_PAGEEXEC) && defined(CONFIG_SMP) + cpu_clear(cpu, prev->context.cpu_user_cs_mask); + cpu_set(cpu, next->context.cpu_user_cs_mask); +#endif + +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) + if (prev->context.user_cs_base != next->context.user_cs_base || + prev->context.user_cs_limit != next->context.user_cs_limit) +#endif + + set_user_cs(next, cpu); } #ifdef CONFIG_SMP else { @@ -58,6 +70,12 @@ static inline void switch_mm(struct mm_s */ load_cr3(next->pgd); load_LDT_nolock(&next->context, cpu); + +#ifdef CONFIG_PAX_PAGEEXEC + cpu_set(cpu, next->context.cpu_user_cs_mask); +#endif + + set_user_cs(next, cpu); } } #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/module.h linux-2.6.17.13-g-v/include/asm-i386/module.h --- linux-2.6.17.13/include/asm-i386/module.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/module.h 2006-09-20 15:00:12.000000000 +0200 @@ -72,6 +72,12 @@ struct mod_arch_specific #define MODULE_STACKSIZE "" #endif -#define MODULE_ARCH_VERMAGIC MODULE_PROC_FAMILY MODULE_REGPARM MODULE_STACKSIZE +#ifdef CONFIG_GRKERNSEC +#define MODULE_GRSEC "GRSECURITY " +#else +#define MODULE_GRSEC "" +#endif + +#define MODULE_ARCH_VERMAGIC MODULE_PROC_FAMILY MODULE_REGPARM MODULE_STACKSIZE MODULE_GRSEC #endif /* _ASM_I386_MODULE_H */ diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/page.h linux-2.6.17.13-g-v/include/asm-i386/page.h --- linux-2.6.17.13/include/asm-i386/page.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/page.h 2006-09-20 15:00:12.000000000 +0200 @@ -52,13 +52,14 @@ typedef struct { unsigned long long pgpr #define pmd_val(x) ((x).pmd) #define pte_val(x) ((x).pte_low | ((unsigned long long)(x).pte_high << 32)) #define __pmd(x) ((pmd_t) { (x) } ) +#define __pte(x) ({ pte_t __pte = {(x), (x) >> 32}; __pte; }) #define HPAGE_SHIFT 21 #else typedef struct { unsigned long pte_low; } pte_t; typedef struct { unsigned long pgd; } pgd_t; typedef struct { unsigned long pgprot; } pgprot_t; -#define boot_pte_t pte_t /* or would you rather have a typedef */ #define pte_val(x) ((x).pte_low) +#define __pte(x) ((pte_t) { (x) } ) #define HPAGE_SHIFT 22 #endif #define PTE_MASK PAGE_MASK @@ -73,7 +74,6 @@ typedef struct { unsigned long pgprot; } #define pgd_val(x) ((x).pgd) #define pgprot_val(x) ((x).pgprot) -#define __pte(x) ((pte_t) { (x) } ) #define __pgd(x) ((pgd_t) { (x) } ) #define __pgprot(x) ((pgprot_t) { (x) } ) @@ -118,6 +118,15 @@ extern int page_is_ram(unsigned long pag #endif #define __KERNEL_START (__PAGE_OFFSET + __PHYSICAL_START) +#ifdef CONFIG_PAX_KERNEXEC +#define __KERNEL_TEXT_OFFSET (__PAGE_OFFSET + ((__PHYSICAL_START + ~(4*1024*1024)) & (4*1024*1024))) +#ifndef __ASSEMBLY__ +extern unsigned char MODULES_VADDR[]; +extern unsigned char MODULES_END[]; +#endif +#else +#define __KERNEL_TEXT_OFFSET (0) +#endif #define PAGE_OFFSET ((unsigned long)__PAGE_OFFSET) #define VMALLOC_RESERVE ((unsigned long)__VMALLOC_RESERVE) @@ -137,6 +146,19 @@ extern int page_is_ram(unsigned long pag ((current->personality & READ_IMPLIES_EXEC) ? VM_EXEC : 0 ) | \ VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC) +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) +#ifdef CONFIG_PAX_MPROTECT +#define __VM_STACK_FLAGS (((current->mm->pax_flags & MF_PAX_MPROTECT)?0:VM_MAYEXEC) | \ + ((current->mm->pax_flags & (MF_PAX_PAGEEXEC|MF_PAX_SEGMEXEC))?0:VM_EXEC)) +#else +#define __VM_STACK_FLAGS (VM_MAYEXEC | ((current->mm->pax_flags & (MF_PAX_PAGEEXEC|MF_PAX_SEGMEXEC))?0:VM_EXEC)) +#endif +#endif + +#ifdef CONFIG_PAX_PAGEEXEC +#define CONFIG_ARCH_TRACK_EXEC_LIMIT 1 +#endif + #endif /* __KERNEL__ */ #include diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/pgalloc.h linux-2.6.17.13-g-v/include/asm-i386/pgalloc.h --- linux-2.6.17.13/include/asm-i386/pgalloc.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/pgalloc.h 2006-09-20 15:00:12.000000000 +0200 @@ -3,11 +3,17 @@ #include #include +#include #include #include /* for struct page */ +#ifdef CONFIG_PAX_NOVSYSCALL +#define pmd_populate_kernel(mm, pmd, pte) \ + set_pmd(pmd, __pmd(_KERNPG_TABLE + __pa(pte))) +#else #define pmd_populate_kernel(mm, pmd, pte) \ set_pmd(pmd, __pmd(_PAGE_TABLE + __pa(pte))) +#endif #define pmd_populate(mm, pmd, pte) \ set_pmd(pmd, __pmd(_PAGE_TABLE + \ diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/pgtable.h linux-2.6.17.13-g-v/include/asm-i386/pgtable.h --- linux-2.6.17.13/include/asm-i386/pgtable.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/pgtable.h 2006-09-20 15:00:12.000000000 +0200 @@ -34,7 +34,6 @@ struct vm_area_struct; */ #define ZERO_PAGE(vaddr) (virt_to_page(empty_zero_page)) extern unsigned long empty_zero_page[1024]; -extern pgd_t swapper_pg_dir[1024]; extern kmem_cache_t *pgd_cache; extern kmem_cache_t *pmd_cache; extern spinlock_t pgd_lock; @@ -59,6 +58,11 @@ void paging_init(void); # include #endif +extern pgd_t swapper_pg_dir[PTRS_PER_PGD]; +#ifdef CONFIG_X86_PAE +extern pmd_t swapper_pm_dir[PTRS_PER_PGD][PTRS_PER_PMD]; +#endif + #define PGDIR_SIZE (1UL << PGDIR_SHIFT) #define PGDIR_MASK (~(PGDIR_SIZE-1)) @@ -68,9 +72,11 @@ void paging_init(void); #define USER_PGD_PTRS (PAGE_OFFSET >> PGDIR_SHIFT) #define KERNEL_PGD_PTRS (PTRS_PER_PGD-USER_PGD_PTRS) +#ifndef CONFIG_X86_PAE #define TWOLEVEL_PGDIR_SHIFT 22 #define BOOT_USER_PGD_PTRS (__PAGE_OFFSET >> TWOLEVEL_PGDIR_SHIFT) #define BOOT_KERNEL_PGD_PTRS (1024-BOOT_USER_PGD_PTRS) +#endif /* Just any arbitrary offset to the start of the vmalloc VM area: the * current 8MB value just means that there will be a 8MB "hole" after the @@ -141,17 +147,26 @@ void paging_init(void); #define PAGE_SHARED_EXEC \ __pgprot(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER | _PAGE_ACCESSED) -#define PAGE_COPY_NOEXEC \ - __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED | _PAGE_NX) #define PAGE_COPY_EXEC \ __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED) -#define PAGE_COPY \ - PAGE_COPY_NOEXEC #define PAGE_READONLY \ __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED | _PAGE_NX) #define PAGE_READONLY_EXEC \ __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED) +#ifdef CONFIG_PAX_PAGEEXEC +# define PAGE_SHARED_NOEXEC __pgprot(_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED) +# define PAGE_COPY_NOEXEC __pgprot(_PAGE_PRESENT | _PAGE_ACCESSED) +# define PAGE_READONLY_NOEXEC __pgprot(_PAGE_PRESENT | _PAGE_ACCESSED) +#else +# define PAGE_SHARED_NOEXEC PAGE_SHARED +# define PAGE_COPY_NOEXEC \ + __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED | _PAGE_NX) +# define PAGE_READONLY_NOEXEC PAGE_READONLY +#endif + +#define PAGE_COPY \ + PAGE_COPY_NOEXEC #define _PAGE_KERNEL \ (_PAGE_PRESENT | _PAGE_RW | _PAGE_DIRTY | _PAGE_ACCESSED | _PAGE_NX) #define _PAGE_KERNEL_EXEC \ @@ -176,18 +191,18 @@ extern unsigned long long __PAGE_KERNEL, * This is the closest we can get.. */ #define __P000 PAGE_NONE -#define __P001 PAGE_READONLY -#define __P010 PAGE_COPY -#define __P011 PAGE_COPY +#define __P001 PAGE_READONLY_NOEXEC +#define __P010 PAGE_COPY_NOEXEC +#define __P011 PAGE_COPY_NOEXEC #define __P100 PAGE_READONLY_EXEC #define __P101 PAGE_READONLY_EXEC #define __P110 PAGE_COPY_EXEC #define __P111 PAGE_COPY_EXEC #define __S000 PAGE_NONE -#define __S001 PAGE_READONLY -#define __S010 PAGE_SHARED -#define __S011 PAGE_SHARED +#define __S001 PAGE_READONLY_NOEXEC +#define __S010 PAGE_SHARED_NOEXEC +#define __S011 PAGE_SHARED_NOEXEC #define __S100 PAGE_READONLY_EXEC #define __S101 PAGE_READONLY_EXEC #define __S110 PAGE_SHARED_EXEC @@ -431,6 +446,9 @@ extern void noexec_setup(const char *str #endif /* !__ASSEMBLY__ */ +#define HAVE_ARCH_UNMAPPED_AREA +#define HAVE_ARCH_UNMAPPED_AREA_TOPDOWN + #ifdef CONFIG_FLATMEM #define kern_addr_valid(addr) (1) #endif /* CONFIG_FLATMEM */ diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/processor.h linux-2.6.17.13-g-v/include/asm-i386/processor.h --- linux-2.6.17.13/include/asm-i386/processor.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/processor.h 2006-09-20 15:00:12.000000000 +0200 @@ -19,7 +19,6 @@ #include #include #include -#include #include /* flag for disabling the tsc */ @@ -94,8 +93,6 @@ struct cpuinfo_x86 { extern struct cpuinfo_x86 boot_cpu_data; extern struct cpuinfo_x86 new_cpu_data; -extern struct tss_struct doublefault_tss; -DECLARE_PER_CPU(struct tss_struct, init_tss); #ifdef CONFIG_SMP extern struct cpuinfo_x86 cpu_data[]; @@ -325,10 +322,19 @@ extern int bootloader_type; */ #define TASK_SIZE (PAGE_OFFSET) +#ifdef CONFIG_PAX_SEGMEXEC +#define SEGMEXEC_TASK_SIZE ((PAGE_OFFSET) / 2) +#endif + /* This decides where the kernel will search for a free chunk of vm * space during mmap's. */ + +#ifdef CONFIG_PAX_SEGMEXEC +#define TASK_UNMAPPED_BASE (PAGE_ALIGN((current->mm->pax_flags & MF_PAX_SEGMEXEC) ? SEGMEXEC_TASK_SIZE/3 : TASK_SIZE/3)) +#else #define TASK_UNMAPPED_BASE (PAGE_ALIGN(TASK_SIZE / 3)) +#endif #define HAVE_ARCH_PICK_MMAP_LAYOUT @@ -444,6 +450,9 @@ struct tss_struct { #define ARCH_MIN_TASKALIGN 16 +extern struct tss_struct doublefault_tss; +extern struct tss_struct init_tss[NR_CPUS]; + struct thread_struct { /* cached TLS descriptors. */ struct desc_struct tls_array[GDT_ENTRY_TLS_ENTRIES]; @@ -472,6 +481,7 @@ struct thread_struct { }; #define INIT_THREAD { \ + .esp0 = sizeof(init_stack) + (long)&init_stack - 8, \ .vm86_info = NULL, \ .sysenter_cs = __KERNEL_CS, \ .io_bitmap_ptr = NULL, \ @@ -484,7 +494,7 @@ struct thread_struct { * be within the limit. */ #define INIT_TSS { \ - .esp0 = sizeof(init_stack) + (long)&init_stack, \ + .esp0 = sizeof(init_stack) + (long)&init_stack - 8, \ .ss0 = __KERNEL_DS, \ .ss1 = __KERNEL_CS, \ .io_bitmap_base = INVALID_IO_BITMAP_OFFSET, \ @@ -560,11 +570,7 @@ void show_trace(struct task_struct *task unsigned long get_wchan(struct task_struct *p); #define THREAD_SIZE_LONGS (THREAD_SIZE/sizeof(unsigned long)) -#define KSTK_TOP(info) \ -({ \ - unsigned long *__ptr = (unsigned long *)(info); \ - (unsigned long)(&__ptr[THREAD_SIZE_LONGS]); \ -}) +#define KSTK_TOP(info) ((info)->task.thread.esp0) /* * The below -8 is to reserve 8 bytes on top of the ring0 stack. @@ -579,7 +585,7 @@ unsigned long get_wchan(struct task_stru #define task_pt_regs(task) \ ({ \ struct pt_regs *__regs__; \ - __regs__ = (struct pt_regs *)(KSTK_TOP(task_stack_page(task))-8); \ + __regs__ = (struct pt_regs *)((task)->thread.esp0); \ __regs__ - 1; \ }) diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/ptrace.h linux-2.6.17.13-g-v/include/asm-i386/ptrace.h --- linux-2.6.17.13/include/asm-i386/ptrace.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/ptrace.h 2006-09-20 15:00:12.000000000 +0200 @@ -65,17 +65,18 @@ struct task_struct; extern void send_sigtrap(struct task_struct *tsk, struct pt_regs *regs, int error_code); /* - * user_mode_vm(regs) determines whether a register set came from user mode. + * user_mode(regs) determines whether a register set came from user mode. * This is true if V8086 mode was enabled OR if the register set was from * protected mode with RPL-3 CS value. This tricky test checks that with * one comparison. Many places in the kernel can bypass this full check - * if they have already ruled out V8086 mode, so user_mode(regs) can be used. + * if they have already ruled out V8086 mode, so user_mode_novm(regs) can + * be used. */ -static inline int user_mode(struct pt_regs *regs) +static inline int user_mode_novm(struct pt_regs *regs) { return (regs->xcs & 3) != 0; } -static inline int user_mode_vm(struct pt_regs *regs) +static inline int user_mode(struct pt_regs *regs) { return ((regs->xcs & 3) | (regs->eflags & VM_MASK)) != 0; } diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/system.h linux-2.6.17.13-g-v/include/asm-i386/system.h --- linux-2.6.17.13/include/asm-i386/system.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/system.h 2006-09-20 15:00:12.000000000 +0200 @@ -5,6 +5,7 @@ #include #include #include +#include #include /* for LOCK_PREFIX */ #ifdef __KERNEL__ @@ -151,7 +152,7 @@ static inline unsigned long get_limit(un unsigned long __limit; __asm__("lsll %1,%0" :"=r" (__limit):"r" (segment)); - return __limit+1; + return __limit; } #define nop() __asm__ __volatile__ ("nop") @@ -496,7 +497,7 @@ static inline void sched_cacheflush(void wbinvd(); } -extern unsigned long arch_align_stack(unsigned long sp); +#define arch_align_stack(x) (x) extern void free_init_pages(char *what, unsigned long begin, unsigned long end); void default_idle(void); diff -NurpP --minimal linux-2.6.17.13/include/asm-i386/uaccess.h linux-2.6.17.13-g-v/include/asm-i386/uaccess.h --- linux-2.6.17.13/include/asm-i386/uaccess.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-i386/uaccess.h 2006-09-20 15:00:12.000000000 +0200 @@ -10,6 +10,8 @@ #include #include #include +#include +#include #define VERIFY_READ 0 #define VERIFY_WRITE 1 @@ -30,7 +32,8 @@ #define get_ds() (KERNEL_DS) #define get_fs() (current_thread_info()->addr_limit) -#define set_fs(x) (current_thread_info()->addr_limit = (x)) +void __set_fs(mm_segment_t x, int cpu); +void set_fs(mm_segment_t x); #define segment_eq(a,b) ((a).seg == (b).seg) @@ -281,9 +284,12 @@ extern void __put_user_8(void); #define __put_user_u64(x, addr, err) \ __asm__ __volatile__( \ - "1: movl %%eax,0(%2)\n" \ - "2: movl %%edx,4(%2)\n" \ + " movw %w5,%%ds\n" \ + "1: movl %%eax,%%ds:0(%2)\n" \ + "2: movl %%edx,%%ds:4(%2)\n" \ "3:\n" \ + " pushl %%ss\n" \ + " popl %%ds\n" \ ".section .fixup,\"ax\"\n" \ "4: movl %3,%0\n" \ " jmp 3b\n" \ @@ -294,7 +300,8 @@ extern void __put_user_8(void); " .long 2b,4b\n" \ ".previous" \ : "=r"(err) \ - : "A" (x), "r" (addr), "i"(-EFAULT), "0"(err)) + : "A" (x), "r" (addr), "i"(-EFAULT), "0"(err), \ + "r"(__USER_DS)) #ifdef CONFIG_X86_WP_WORKS_OK @@ -333,8 +340,11 @@ struct __large_struct { unsigned long bu */ #define __put_user_asm(x, addr, err, itype, rtype, ltype, errret) \ __asm__ __volatile__( \ - "1: mov"itype" %"rtype"1,%2\n" \ + " movw %w5,%%ds\n" \ + "1: mov"itype" %"rtype"1,%%ds:%2\n" \ "2:\n" \ + " pushl %%ss\n" \ + " popl %%ds\n" \ ".section .fixup,\"ax\"\n" \ "3: movl %3,%0\n" \ " jmp 2b\n" \ @@ -344,7 +354,8 @@ struct __large_struct { unsigned long bu " .long 1b,3b\n" \ ".previous" \ : "=r"(err) \ - : ltype (x), "m"(__m(addr)), "i"(errret), "0"(err)) + : ltype (x), "m"(__m(addr)), "i"(errret), "0"(err), \ + "r"(__USER_DS)) #define __get_user_nocheck(x,ptr,size) \ @@ -372,8 +383,11 @@ do { \ #define __get_user_asm(x, addr, err, itype, rtype, ltype, errret) \ __asm__ __volatile__( \ - "1: mov"itype" %2,%"rtype"1\n" \ + " movw %w5,%%ds\n" \ + "1: mov"itype" %%ds:%2,%"rtype"1\n" \ "2:\n" \ + " pushl %%ss\n" \ + " popl %%ds\n" \ ".section .fixup,\"ax\"\n" \ "3: movl %3,%0\n" \ " xor"itype" %"rtype"1,%"rtype"1\n" \ @@ -384,7 +398,7 @@ do { \ " .long 1b,3b\n" \ ".previous" \ : "=r"(err), ltype (x) \ - : "m"(__m(addr)), "i"(errret), "0"(err)) + : "m"(__m(addr)), "i"(errret), "0"(err), "r"(__USER_DS)) unsigned long __must_check __copy_to_user_ll(void __user *to, diff -NurpP --minimal linux-2.6.17.13/include/asm-ia64/elf.h linux-2.6.17.13-g-v/include/asm-ia64/elf.h --- linux-2.6.17.13/include/asm-ia64/elf.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-ia64/elf.h 2006-09-20 15:00:12.000000000 +0200 @@ -163,6 +163,16 @@ typedef elf_greg_t elf_gregset_t[ELF_NGR typedef struct ia64_fpreg elf_fpreg_t; typedef elf_fpreg_t elf_fpregset_t[ELF_NFPREG]; +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) ((tsk)->personality == PER_LINUX32 ? 0x08048000UL : 0x4000000000000000UL) + +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) ((tsk)->personality == PER_LINUX32 ? 16 : 3*PAGE_SHIFT - 13) +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) ((tsk)->personality == PER_LINUX32 ? 16 : 3*PAGE_SHIFT - 13) +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) ((tsk)->personality == PER_LINUX32 ? 16 : 3*PAGE_SHIFT - 13) +#endif struct pt_regs; /* forward declaration... */ diff -NurpP --minimal linux-2.6.17.13/include/asm-ia64/kmap_types.h linux-2.6.17.13-g-v/include/asm-ia64/kmap_types.h --- linux-2.6.17.13/include/asm-ia64/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-ia64/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -23,7 +23,8 @@ D(9) KM_IRQ0, D(10) KM_IRQ1, D(11) KM_SOFTIRQ0, D(12) KM_SOFTIRQ1, -D(13) KM_TYPE_NR +D(13) KM_CLEARPAGE, +D(14) KM_TYPE_NR }; #undef D diff -NurpP --minimal linux-2.6.17.13/include/asm-ia64/page.h linux-2.6.17.13-g-v/include/asm-ia64/page.h --- linux-2.6.17.13/include/asm-ia64/page.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-ia64/page.h 2006-09-20 15:00:12.000000000 +0200 @@ -229,4 +229,13 @@ get_order (unsigned long size) (((current->personality & READ_IMPLIES_EXEC) != 0) \ ? VM_EXEC : 0)) +#ifdef CONFIG_PAX_PAGEEXEC +#ifdef CONFIG_PAX_MPROTECT +#define __VM_STACK_FLAGS (((current->mm->pax_flags & MF_PAX_MPROTECT)?0:VM_MAYEXEC) | \ + ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#else +#define __VM_STACK_FLAGS (VM_MAYEXEC | ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#endif +#endif + #endif /* _ASM_IA64_PAGE_H */ diff -NurpP --minimal linux-2.6.17.13/include/asm-ia64/pgtable.h linux-2.6.17.13-g-v/include/asm-ia64/pgtable.h --- linux-2.6.17.13/include/asm-ia64/pgtable.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-ia64/pgtable.h 2006-09-20 15:00:12.000000000 +0200 @@ -144,6 +144,17 @@ #define PAGE_READONLY __pgprot(__ACCESS_BITS | _PAGE_PL_3 | _PAGE_AR_R) #define PAGE_COPY __pgprot(__ACCESS_BITS | _PAGE_PL_3 | _PAGE_AR_R) #define PAGE_COPY_EXEC __pgprot(__ACCESS_BITS | _PAGE_PL_3 | _PAGE_AR_RX) + +#ifdef CONFIG_PAX_PAGEEXEC +# define PAGE_SHARED_NOEXEC __pgprot(__ACCESS_BITS | _PAGE_PL_3 | _PAGE_AR_RW) +# define PAGE_READONLY_NOEXEC __pgprot(__ACCESS_BITS | _PAGE_PL_3 | _PAGE_AR_R) +# define PAGE_COPY_NOEXEC __pgprot(__ACCESS_BITS | _PAGE_PL_3 | _PAGE_AR_R) +#else +# define PAGE_SHARED_NOEXEC PAGE_SHARED +# define PAGE_READONLY_NOEXEC PAGE_READONLY +# define PAGE_COPY_NOEXEC PAGE_COPY +#endif + #define PAGE_GATE __pgprot(__ACCESS_BITS | _PAGE_PL_0 | _PAGE_AR_X_RX) #define PAGE_KERNEL __pgprot(__DIRTY_BITS | _PAGE_PL_0 | _PAGE_AR_RWX) #define PAGE_KERNELRX __pgprot(__ACCESS_BITS | _PAGE_PL_0 | _PAGE_AR_RX) diff -NurpP --minimal linux-2.6.17.13/include/asm-ia64/processor.h linux-2.6.17.13-g-v/include/asm-ia64/processor.h --- linux-2.6.17.13/include/asm-ia64/processor.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-ia64/processor.h 2006-09-20 15:00:12.000000000 +0200 @@ -284,7 +284,7 @@ struct thread_struct { .on_ustack = 0, \ .ksp = 0, \ .map_base = DEFAULT_MAP_BASE, \ - .rbs_bot = STACK_TOP - DEFAULT_USER_STACK_SIZE, \ + .rbs_bot = __STACK_TOP - DEFAULT_USER_STACK_SIZE, \ .task_size = DEFAULT_TASK_SIZE, \ .last_fph_cpu = -1, \ INIT_THREAD_IA32 \ diff -NurpP --minimal linux-2.6.17.13/include/asm-ia64/tlb.h linux-2.6.17.13-g-v/include/asm-ia64/tlb.h --- linux-2.6.17.13/include/asm-ia64/tlb.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-ia64/tlb.h 2006-09-20 15:00:12.000000000 +0200 @@ -41,6 +41,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/include/asm-ia64/ustack.h linux-2.6.17.13-g-v/include/asm-ia64/ustack.h --- linux-2.6.17.13/include/asm-ia64/ustack.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-ia64/ustack.h 2006-09-20 15:00:12.000000000 +0200 @@ -11,6 +11,6 @@ #define MAX_USER_STACK_SIZE (RGN_MAP_LIMIT/2) /* Make a default stack size of 2GB */ #define DEFAULT_USER_STACK_SIZE (1UL << 31) -#define STACK_TOP (0x6000000000000000UL + RGN_MAP_LIMIT) +#define __STACK_TOP (0x6000000000000000UL + RGN_MAP_LIMIT) #endif /* _ASM_IA64_USTACK_H */ diff -NurpP --minimal linux-2.6.17.13/include/asm-m32r/kmap_types.h linux-2.6.17.13-g-v/include/asm-m32r/kmap_types.h --- linux-2.6.17.13/include/asm-m32r/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-m32r/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -25,7 +25,8 @@ D(9) KM_IRQ0, D(10) KM_IRQ1, D(11) KM_SOFTIRQ0, D(12) KM_SOFTIRQ1, -D(13) KM_TYPE_NR +D(13) KM_CLEARPAGE, +D(14) KM_TYPE_NR }; #undef D diff -NurpP --minimal linux-2.6.17.13/include/asm-m68k/kmap_types.h linux-2.6.17.13-g-v/include/asm-m68k/kmap_types.h --- linux-2.6.17.13/include/asm-m68k/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-m68k/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,7 @@ enum km_type { KM_IRQ1, KM_SOFTIRQ0, KM_SOFTIRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-m68knommu/kmap_types.h linux-2.6.17.13-g-v/include/asm-m68knommu/kmap_types.h --- linux-2.6.17.13/include/asm-m68knommu/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-m68knommu/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,7 @@ enum km_type { KM_IRQ1, KM_SOFTIRQ0, KM_SOFTIRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-mips/a.out.h linux-2.6.17.13-g-v/include/asm-mips/a.out.h --- linux-2.6.17.13/include/asm-mips/a.out.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-mips/a.out.h 2006-09-20 15:00:12.000000000 +0200 @@ -36,10 +36,10 @@ struct exec #ifdef __KERNEL__ #ifdef CONFIG_32BIT -#define STACK_TOP TASK_SIZE +#define __STACK_TOP TASK_SIZE #endif #ifdef CONFIG_64BIT -#define STACK_TOP (current->thread.mflags & MF_32BIT_ADDR ? TASK_SIZE32 : TASK_SIZE) +#define __STACK_TOP (current->thread.mflags & MF_32BIT_ADDR ? TASK_SIZE32 : TASK_SIZE) #endif #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-mips/elf.h linux-2.6.17.13-g-v/include/asm-mips/elf.h --- linux-2.6.17.13/include/asm-mips/elf.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-mips/elf.h 2006-09-20 15:00:12.000000000 +0200 @@ -372,4 +372,15 @@ extern int dump_task_fpu(struct task_str #define ELF_ET_DYN_BASE (TASK_SIZE / 3 * 2) #endif +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 0x00400000UL : 0x00400000UL) + +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 27-PAGE_SHIFT : 36-PAGE_SHIFT) +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 27-PAGE_SHIFT : 36-PAGE_SHIFT) +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) (((tsk)->thread.mflags & MF_32BIT_ADDR) ? 27-PAGE_SHIFT : 36-PAGE_SHIFT) +#endif + #endif /* _ASM_ELF_H */ diff -NurpP --minimal linux-2.6.17.13/include/asm-mips/kmap_types.h linux-2.6.17.13-g-v/include/asm-mips/kmap_types.h --- linux-2.6.17.13/include/asm-mips/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-mips/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -23,7 +23,8 @@ D(9) KM_IRQ0, D(10) KM_IRQ1, D(11) KM_SOFTIRQ0, D(12) KM_SOFTIRQ1, -D(13) KM_TYPE_NR +D(13) KM_CLEARPAGE, +D(14) KM_TYPE_NR }; #undef D diff -NurpP --minimal linux-2.6.17.13/include/asm-mips/page.h linux-2.6.17.13-g-v/include/asm-mips/page.h --- linux-2.6.17.13/include/asm-mips/page.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-mips/page.h 2006-09-20 15:00:12.000000000 +0200 @@ -79,15 +79,17 @@ static inline void copy_user_page(void * #ifdef CONFIG_CPU_MIPS32 typedef struct { unsigned long pte_low, pte_high; } pte_t; #define pte_val(x) ((x).pte_low | ((unsigned long long)(x).pte_high << 32)) + #define __pte(x) ({ pte_t __pte = {(x), (x) >> 32}; __pte; }) #else typedef struct { unsigned long long pte; } pte_t; #define pte_val(x) ((x).pte) + #define __pte(x) ((pte_t) { (x) } ) #endif #else typedef struct { unsigned long pte; } pte_t; #define pte_val(x) ((x).pte) -#endif #define __pte(x) ((pte_t) { (x) } ) +#endif /* * For 3-level pagetables we defines these ourselves, for 2-level the @@ -151,6 +153,15 @@ typedef struct { unsigned long pgprot; } #define VM_DATA_DEFAULT_FLAGS (VM_READ | VM_WRITE | VM_EXEC | \ VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC) +#ifdef CONFIG_PAX_PAGEEXEC +#ifdef CONFIG_PAX_MPROTECT +#define __VM_STACK_FLAGS (((current->mm->pax_flags & MF_PAX_MPROTECT)?0:VM_MAYEXEC) | \ + ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#else +#define __VM_STACK_FLAGS (VM_MAYEXEC | ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#endif +#endif + #define UNCAC_ADDR(addr) ((addr) - PAGE_OFFSET + UNCAC_BASE) #define CAC_ADDR(addr) ((addr) - UNCAC_BASE + PAGE_OFFSET) diff -NurpP --minimal linux-2.6.17.13/include/asm-parisc/a.out.h linux-2.6.17.13-g-v/include/asm-parisc/a.out.h --- linux-2.6.17.13/include/asm-parisc/a.out.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-parisc/a.out.h 2006-09-20 15:00:12.000000000 +0200 @@ -22,7 +22,7 @@ struct exec /* XXX: STACK_TOP actually should be STACK_BOTTOM for parisc. * prumpf */ -#define STACK_TOP TASK_SIZE +#define __STACK_TOP TASK_SIZE #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-parisc/elf.h linux-2.6.17.13-g-v/include/asm-parisc/elf.h --- linux-2.6.17.13/include/asm-parisc/elf.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-parisc/elf.h 2006-09-20 15:00:12.000000000 +0200 @@ -337,6 +337,17 @@ struct pt_regs; /* forward declaration.. #define ELF_ET_DYN_BASE (TASK_UNMAPPED_BASE + 0x01000000) +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) 0x10000UL + +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) 16 +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) 16 +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) 16 +#endif + /* This yields a mask that user programs can use to figure out what instruction set this CPU supports. This could be done in user space, but it's not easy, and we've already done it here. */ diff -NurpP --minimal linux-2.6.17.13/include/asm-parisc/kmap_types.h linux-2.6.17.13-g-v/include/asm-parisc/kmap_types.h --- linux-2.6.17.13/include/asm-parisc/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-parisc/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -23,7 +23,8 @@ D(9) KM_IRQ0, D(10) KM_IRQ1, D(11) KM_SOFTIRQ0, D(12) KM_SOFTIRQ1, -D(13) KM_TYPE_NR +D(13) KM_CLEARPAGE, +D(14) KM_TYPE_NR }; #undef D diff -NurpP --minimal linux-2.6.17.13/include/asm-parisc/page.h linux-2.6.17.13-g-v/include/asm-parisc/page.h --- linux-2.6.17.13/include/asm-parisc/page.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-parisc/page.h 2006-09-20 15:00:12.000000000 +0200 @@ -189,6 +189,15 @@ extern int npmem_ranges; #define VM_DATA_DEFAULT_FLAGS (VM_READ | VM_WRITE | VM_EXEC | \ VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC) +#ifdef CONFIG_PAX_PAGEEXEC +#ifdef CONFIG_PAX_MPROTECT +#define __VM_STACK_FLAGS (((current->mm->pax_flags & MF_PAX_MPROTECT)?0:VM_MAYEXEC) | \ + ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#else +#define __VM_STACK_FLAGS (VM_MAYEXEC | ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#endif +#endif + #endif /* __KERNEL__ */ #include diff -NurpP --minimal linux-2.6.17.13/include/asm-parisc/pgtable.h linux-2.6.17.13-g-v/include/asm-parisc/pgtable.h --- linux-2.6.17.13/include/asm-parisc/pgtable.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-parisc/pgtable.h 2006-09-20 15:00:12.000000000 +0200 @@ -220,6 +220,17 @@ extern void *vmalloc_start; #define PAGE_EXECREAD __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_READ | _PAGE_EXEC |_PAGE_ACCESSED) #define PAGE_COPY PAGE_EXECREAD #define PAGE_RWX __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_READ | _PAGE_WRITE | _PAGE_EXEC |_PAGE_ACCESSED) + +#ifdef CONFIG_PAX_PAGEEXEC +# define PAGE_SHARED_NOEXEC __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_READ | _PAGE_WRITE | _PAGE_ACCESSED) +# define PAGE_COPY_NOEXEC __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_READ | _PAGE_ACCESSED) +# define PAGE_READONLY_NOEXEC __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_READ | _PAGE_ACCESSED) +#else +# define PAGE_SHARED_NOEXEC PAGE_SHARED +# define PAGE_COPY_NOEXEC PAGE_COPY +# define PAGE_READONLY_NOEXEC PAGE_READONLY +#endif + #define PAGE_KERNEL __pgprot(_PAGE_KERNEL) #define PAGE_KERNEL_RO __pgprot(_PAGE_KERNEL & ~_PAGE_WRITE) #define PAGE_KERNEL_UNC __pgprot(_PAGE_KERNEL | _PAGE_NO_CACHE) diff -NurpP --minimal linux-2.6.17.13/include/asm-powerpc/a.out.h linux-2.6.17.13-g-v/include/asm-powerpc/a.out.h --- linux-2.6.17.13/include/asm-powerpc/a.out.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-powerpc/a.out.h 2006-09-20 15:00:12.000000000 +0200 @@ -23,12 +23,12 @@ struct exec #define STACK_TOP_USER64 TASK_SIZE_USER64 #define STACK_TOP_USER32 TASK_SIZE_USER32 -#define STACK_TOP (test_thread_flag(TIF_32BIT) ? \ +#define __STACK_TOP (test_thread_flag(TIF_32BIT) ? \ STACK_TOP_USER32 : STACK_TOP_USER64) #else /* __powerpc64__ */ -#define STACK_TOP TASK_SIZE +#define __STACK_TOP TASK_SIZE #endif /* __powerpc64__ */ #endif /* __KERNEL__ */ diff -NurpP --minimal linux-2.6.17.13/include/asm-powerpc/elf.h linux-2.6.17.13-g-v/include/asm-powerpc/elf.h --- linux-2.6.17.13/include/asm-powerpc/elf.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-powerpc/elf.h 2006-09-20 15:00:12.000000000 +0200 @@ -176,6 +176,26 @@ typedef elf_vrreg_t elf_vrregset_t32[ELF #define ELF_ET_DYN_BASE (0x08000000) +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) (0x10000000UL) + +#ifdef __powerpc64__ +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) (test_thread_flag(TIF_32BIT) ? 16 : 28) +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) (test_thread_flag(TIF_32BIT) ? 16 : 28) +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) (test_thread_flag(TIF_32BIT) ? 16 : 28) +#else +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) 15 +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) 15 +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) 15 +#endif +#endif + #ifdef __KERNEL__ /* Common routine for both 32-bit and 64-bit processes */ diff -NurpP --minimal linux-2.6.17.13/include/asm-powerpc/kmap_types.h linux-2.6.17.13-g-v/include/asm-powerpc/kmap_types.h --- linux-2.6.17.13/include/asm-powerpc/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-powerpc/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -26,6 +26,7 @@ enum km_type { KM_SOFTIRQ1, KM_PPC_SYNC_PAGE, KM_PPC_SYNC_ICACHE, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-powerpc/page_64.h linux-2.6.17.13-g-v/include/asm-powerpc/page_64.h --- linux-2.6.17.13/include/asm-powerpc/page_64.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-powerpc/page_64.h 2006-09-20 15:00:12.000000000 +0200 @@ -170,6 +170,15 @@ extern unsigned int HPAGE_SHIFT; (test_thread_flag(TIF_32BIT) ? \ VM_STACK_DEFAULT_FLAGS32 : VM_STACK_DEFAULT_FLAGS64) +#ifdef CONFIG_PAX_PAGEEXEC +#ifdef CONFIG_PAX_MPROTECT +#define __VM_STACK_FLAGS (((current->mm->pax_flags & MF_PAX_MPROTECT)?0:VM_MAYEXEC) | \ + ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#else +#define __VM_STACK_FLAGS (VM_MAYEXEC | ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#endif +#endif + #include #endif /* __KERNEL__ */ diff -NurpP --minimal linux-2.6.17.13/include/asm-powerpc/unistd.h linux-2.6.17.13-g-v/include/asm-powerpc/unistd.h --- linux-2.6.17.13/include/asm-powerpc/unistd.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-powerpc/unistd.h 2006-09-20 15:00:12.000000000 +0200 @@ -275,7 +275,7 @@ #endif #define __NR_rtas 255 #define __NR_sys_debug_setcontext 256 -/* Number 257 is reserved for vserver */ +#define __NR_vserver 257 /* 258 currently unused */ #define __NR_mbind 259 #define __NR_get_mempolicy 260 diff -NurpP --minimal linux-2.6.17.13/include/asm-ppc/page.h linux-2.6.17.13-g-v/include/asm-ppc/page.h --- linux-2.6.17.13/include/asm-ppc/page.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-ppc/page.h 2006-09-20 15:00:12.000000000 +0200 @@ -175,6 +175,15 @@ extern __inline__ int get_order(unsigned /* We do define AT_SYSINFO_EHDR but don't use the gate mecanism */ #define __HAVE_ARCH_GATE_AREA 1 +#ifdef CONFIG_PAX_PAGEEXEC +#ifdef CONFIG_PAX_MPROTECT +#define __VM_STACK_FLAGS (((current->mm->pax_flags & MF_PAX_MPROTECT)?0:VM_MAYEXEC) | \ + ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#else +#define __VM_STACK_FLAGS (VM_MAYEXEC | ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#endif +#endif + #include #endif /* __KERNEL__ */ #endif /* _PPC_PAGE_H */ diff -NurpP --minimal linux-2.6.17.13/include/asm-ppc/pgtable.h linux-2.6.17.13-g-v/include/asm-ppc/pgtable.h --- linux-2.6.17.13/include/asm-ppc/pgtable.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-ppc/pgtable.h 2006-09-20 15:00:12.000000000 +0200 @@ -441,11 +441,21 @@ extern unsigned long ioremap_bot, iorema #define PAGE_NONE __pgprot(_PAGE_BASE) #define PAGE_READONLY __pgprot(_PAGE_BASE | _PAGE_USER) -#define PAGE_READONLY_X __pgprot(_PAGE_BASE | _PAGE_USER | _PAGE_EXEC) +#define PAGE_READONLY_X __pgprot(_PAGE_BASE | _PAGE_USER | _PAGE_EXEC | _PAGE_HWEXEC) #define PAGE_SHARED __pgprot(_PAGE_BASE | _PAGE_USER | _PAGE_RW) -#define PAGE_SHARED_X __pgprot(_PAGE_BASE | _PAGE_USER | _PAGE_RW | _PAGE_EXEC) +#define PAGE_SHARED_X __pgprot(_PAGE_BASE | _PAGE_USER | _PAGE_RW | _PAGE_EXEC | _PAGE_HWEXEC) #define PAGE_COPY __pgprot(_PAGE_BASE | _PAGE_USER) -#define PAGE_COPY_X __pgprot(_PAGE_BASE | _PAGE_USER | _PAGE_EXEC) +#define PAGE_COPY_X __pgprot(_PAGE_BASE | _PAGE_USER | _PAGE_EXEC | _PAGE_HWEXEC) + +#if defined(CONFIG_PAX_PAGEEXEC) && !defined(CONFIG_40x) && !defined(CONFIG_44x) +# define PAGE_SHARED_NOEXEC __pgprot(_PAGE_BASE | _PAGE_USER | _PAGE_RW | _PAGE_GUARDED) +# define PAGE_COPY_NOEXEC __pgprot(_PAGE_BASE | _PAGE_USER | _PAGE_GUARDED) +# define PAGE_READONLY_NOEXEC __pgprot(_PAGE_BASE | _PAGE_USER | _PAGE_GUARDED) +#else +# define PAGE_SHARED_NOEXEC PAGE_SHARED +# define PAGE_COPY_NOEXEC PAGE_COPY +# define PAGE_READONLY_NOEXEC PAGE_READONLY +#endif #define PAGE_KERNEL __pgprot(_PAGE_RAM) #define PAGE_KERNEL_NOCACHE __pgprot(_PAGE_IO) @@ -457,21 +467,21 @@ extern unsigned long ioremap_bot, iorema * This is the closest we can get.. */ #define __P000 PAGE_NONE -#define __P001 PAGE_READONLY_X -#define __P010 PAGE_COPY -#define __P011 PAGE_COPY_X -#define __P100 PAGE_READONLY +#define __P001 PAGE_READONLY_NOEXEC +#define __P010 PAGE_COPY_NOEXEC +#define __P011 PAGE_COPY_NOEXEC +#define __P100 PAGE_READONLY_X #define __P101 PAGE_READONLY_X -#define __P110 PAGE_COPY +#define __P110 PAGE_COPY_X #define __P111 PAGE_COPY_X #define __S000 PAGE_NONE -#define __S001 PAGE_READONLY_X -#define __S010 PAGE_SHARED -#define __S011 PAGE_SHARED_X -#define __S100 PAGE_READONLY +#define __S001 PAGE_READONLY_NOEXEC +#define __S010 PAGE_SHARED_NOEXEC +#define __S011 PAGE_SHARED_NOEXEC +#define __S100 PAGE_READONLY_X #define __S101 PAGE_READONLY_X -#define __S110 PAGE_SHARED +#define __S110 PAGE_SHARED_X #define __S111 PAGE_SHARED_X #ifndef __ASSEMBLY__ diff -NurpP --minimal linux-2.6.17.13/include/asm-s390/kmap_types.h linux-2.6.17.13-g-v/include/asm-s390/kmap_types.h --- linux-2.6.17.13/include/asm-s390/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-s390/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -16,6 +16,7 @@ enum km_type { KM_IRQ1, KM_SOFTIRQ0, KM_SOFTIRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-s390/unistd.h linux-2.6.17.13-g-v/include/asm-s390/unistd.h --- linux-2.6.17.13/include/asm-s390/unistd.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-s390/unistd.h 2006-09-20 15:00:12.000000000 +0200 @@ -255,7 +255,7 @@ #define __NR_clock_gettime (__NR_timer_create+6) #define __NR_clock_getres (__NR_timer_create+7) #define __NR_clock_nanosleep (__NR_timer_create+8) -/* Number 263 is reserved for vserver */ +#define __NR_vserver 263 #define __NR_fadvise64_64 264 #define __NR_statfs64 265 #define __NR_fstatfs64 266 diff -NurpP --minimal linux-2.6.17.13/include/asm-sh/kmap_types.h linux-2.6.17.13-g-v/include/asm-sh/kmap_types.h --- linux-2.6.17.13/include/asm-sh/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sh/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -25,7 +25,8 @@ D(9) KM_IRQ0, D(10) KM_IRQ1, D(11) KM_SOFTIRQ0, D(12) KM_SOFTIRQ1, -D(13) KM_TYPE_NR +D(13) KM_CLEARPAGE, +D(14) KM_TYPE_NR }; #undef D diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc/a.out.h linux-2.6.17.13-g-v/include/asm-sparc/a.out.h --- linux-2.6.17.13/include/asm-sparc/a.out.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc/a.out.h 2006-09-20 15:00:12.000000000 +0200 @@ -91,7 +91,7 @@ struct relocation_info /* used when head #include -#define STACK_TOP (PAGE_OFFSET - PAGE_SIZE) +#define __STACK_TOP (PAGE_OFFSET - PAGE_SIZE) #endif /* __KERNEL__ */ diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc/elf.h linux-2.6.17.13-g-v/include/asm-sparc/elf.h --- linux-2.6.17.13/include/asm-sparc/elf.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc/elf.h 2006-09-20 15:00:12.000000000 +0200 @@ -145,6 +145,17 @@ typedef struct { #define ELF_ET_DYN_BASE (TASK_UNMAPPED_BASE) +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) 0x10000UL + +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) 16 +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) 16 +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) 16 +#endif + /* This yields a mask that user programs can use to figure out what instruction set this cpu supports. This can NOT be done in userspace on Sparc. */ diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc/kmap_types.h linux-2.6.17.13-g-v/include/asm-sparc/kmap_types.h --- linux-2.6.17.13/include/asm-sparc/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,7 @@ enum km_type { KM_IRQ1, KM_SOFTIRQ0, KM_SOFTIRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc/page.h linux-2.6.17.13-g-v/include/asm-sparc/page.h --- linux-2.6.17.13/include/asm-sparc/page.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc/page.h 2006-09-20 15:00:12.000000000 +0200 @@ -161,6 +161,15 @@ extern unsigned long pfn_base; #define VM_DATA_DEFAULT_FLAGS (VM_READ | VM_WRITE | VM_EXEC | \ VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC) +#ifdef CONFIG_PAX_PAGEEXEC +#ifdef CONFIG_PAX_MPROTECT +#define __VM_STACK_FLAGS (((current->mm->pax_flags & MF_PAX_MPROTECT)?0:VM_MAYEXEC) | \ + ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#else +#define __VM_STACK_FLAGS (VM_MAYEXEC | ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#endif +#endif + #endif /* __KERNEL__ */ #include diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc/pgtable.h linux-2.6.17.13-g-v/include/asm-sparc/pgtable.h --- linux-2.6.17.13/include/asm-sparc/pgtable.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc/pgtable.h 2006-09-20 15:00:12.000000000 +0200 @@ -50,6 +50,13 @@ BTFIXUPDEF_INT(page_none) BTFIXUPDEF_INT(page_shared) BTFIXUPDEF_INT(page_copy) BTFIXUPDEF_INT(page_readonly) + +#ifdef CONFIG_PAX_PAGEEXEC +BTFIXUPDEF_INT(page_shared_noexec) +BTFIXUPDEF_INT(page_copy_noexec) +BTFIXUPDEF_INT(page_readonly_noexec) +#endif + BTFIXUPDEF_INT(page_kernel) #define PMD_SHIFT SUN4C_PMD_SHIFT @@ -71,6 +78,16 @@ BTFIXUPDEF_INT(page_kernel) #define PAGE_COPY __pgprot(BTFIXUP_INT(page_copy)) #define PAGE_READONLY __pgprot(BTFIXUP_INT(page_readonly)) +#ifdef CONFIG_PAX_PAGEEXEC +# define PAGE_SHARED_NOEXEC __pgprot(BTFIXUP_INT(page_shared_noexec)) +# define PAGE_COPY_NOEXEC __pgprot(BTFIXUP_INT(page_copy_noexec)) +# define PAGE_READONLY_NOEXEC __pgprot(BTFIXUP_INT(page_readonly_noexec)) +#else +# define PAGE_SHARED_NOEXEC PAGE_SHARED +# define PAGE_COPY_NOEXEC PAGE_COPY +# define PAGE_READONLY_NOEXEC PAGE_READONLY +#endif + extern unsigned long page_kernel; #ifdef MODULE diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc/pgtsrmmu.h linux-2.6.17.13-g-v/include/asm-sparc/pgtsrmmu.h --- linux-2.6.17.13/include/asm-sparc/pgtsrmmu.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc/pgtsrmmu.h 2006-09-20 15:00:12.000000000 +0200 @@ -115,6 +115,16 @@ SRMMU_EXEC | SRMMU_REF) #define SRMMU_PAGE_RDONLY __pgprot(SRMMU_VALID | SRMMU_CACHE | \ SRMMU_EXEC | SRMMU_REF) + +#ifdef CONFIG_PAX_PAGEEXEC +#define SRMMU_PAGE_SHARED_NOEXEC __pgprot(SRMMU_VALID | SRMMU_CACHE | \ + SRMMU_WRITE | SRMMU_REF) +#define SRMMU_PAGE_COPY_NOEXEC __pgprot(SRMMU_VALID | SRMMU_CACHE | \ + SRMMU_REF) +#define SRMMU_PAGE_RDONLY_NOEXEC __pgprot(SRMMU_VALID | SRMMU_CACHE | \ + SRMMU_REF) +#endif + #define SRMMU_PAGE_KERNEL __pgprot(SRMMU_VALID | SRMMU_CACHE | SRMMU_PRIV | \ SRMMU_DIRTY | SRMMU_REF) diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc/uaccess.h linux-2.6.17.13-g-v/include/asm-sparc/uaccess.h --- linux-2.6.17.13/include/asm-sparc/uaccess.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc/uaccess.h 2006-09-20 15:00:12.000000000 +0200 @@ -41,7 +41,7 @@ * No one can read/write anything from userland in the kernel space by setting * large size and address near to PAGE_OFFSET - a fault will break his intentions. */ -#define __user_ok(addr, size) ({ (void)(size); (addr) < STACK_TOP; }) +#define __user_ok(addr, size) ({ (void)(size); (addr) < __STACK_TOP; }) #define __kernel_ok (segment_eq(get_fs(), KERNEL_DS)) #define __access_ok(addr,size) (__user_ok((addr) & get_fs().seg,(size))) #define access_ok(type, addr, size) \ diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc/unistd.h linux-2.6.17.13-g-v/include/asm-sparc/unistd.h --- linux-2.6.17.13/include/asm-sparc/unistd.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc/unistd.h 2006-09-20 15:00:12.000000000 +0200 @@ -283,7 +283,7 @@ #define __NR_timer_getoverrun 264 #define __NR_timer_delete 265 #define __NR_timer_create 266 -/* #define __NR_vserver 267 Reserved for VSERVER */ +#define __NR_vserver 267 #define __NR_io_setup 268 #define __NR_io_destroy 269 #define __NR_io_submit 270 diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc64/a.out.h linux-2.6.17.13-g-v/include/asm-sparc64/a.out.h --- linux-2.6.17.13/include/asm-sparc64/a.out.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc64/a.out.h 2006-09-20 15:00:12.000000000 +0200 @@ -98,7 +98,7 @@ struct relocation_info /* used when head #define STACK_TOP32 ((1UL << 32UL) - PAGE_SIZE) #define STACK_TOP64 (0x0000080000000000UL - (1UL << 32UL)) -#define STACK_TOP (test_thread_flag(TIF_32BIT) ? \ +#define __STACK_TOP (test_thread_flag(TIF_32BIT) ? \ STACK_TOP32 : STACK_TOP64) #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc64/elf.h linux-2.6.17.13-g-v/include/asm-sparc64/elf.h --- linux-2.6.17.13/include/asm-sparc64/elf.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc64/elf.h 2006-09-20 15:00:12.000000000 +0200 @@ -142,6 +142,16 @@ typedef struct { #define ELF_ET_DYN_BASE 0x0000010000000000UL #endif +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) (test_thread_flag(TIF_32BIT) ? 0x10000UL : 0x100000UL) + +#define PAX_DELTA_MMAP_LSB(tsk) (PAGE_SHIFT + 1) +#define PAX_DELTA_MMAP_LEN(tsk) (test_thread_flag(TIF_32BIT) ? 14 : 28 ) +#define PAX_DELTA_EXEC_LSB(tsk) (PAGE_SHIFT + 1) +#define PAX_DELTA_EXEC_LEN(tsk) (test_thread_flag(TIF_32BIT) ? 14 : 28 ) +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) (test_thread_flag(TIF_32BIT) ? 15 : 29 ) +#endif /* This yields a mask that user programs can use to figure out what instruction set this cpu supports. */ diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc64/kmap_types.h linux-2.6.17.13-g-v/include/asm-sparc64/kmap_types.h --- linux-2.6.17.13/include/asm-sparc64/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc64/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -19,6 +19,7 @@ enum km_type { KM_IRQ1, KM_SOFTIRQ0, KM_SOFTIRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc64/page.h linux-2.6.17.13-g-v/include/asm-sparc64/page.h --- linux-2.6.17.13/include/asm-sparc64/page.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc64/page.h 2006-09-20 15:00:12.000000000 +0200 @@ -142,6 +142,15 @@ typedef unsigned long pgprot_t; #define VM_DATA_DEFAULT_FLAGS (VM_READ | VM_WRITE | VM_EXEC | \ VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC) +#ifdef CONFIG_PAX_PAGEEXEC +#ifdef CONFIG_PAX_MPROTECT +#define __VM_STACK_FLAGS (((current->mm->pax_flags & MF_PAX_MPROTECT)?0:VM_MAYEXEC) | \ + ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#else +#define __VM_STACK_FLAGS (VM_MAYEXEC | ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#endif +#endif + #endif /* !(__KERNEL__) */ #include diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc64/tlb.h linux-2.6.17.13-g-v/include/asm-sparc64/tlb.h --- linux-2.6.17.13/include/asm-sparc64/tlb.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc64/tlb.h 2006-09-20 15:00:12.000000000 +0200 @@ -3,6 +3,7 @@ #include #include +#include #include #include #include diff -NurpP --minimal linux-2.6.17.13/include/asm-sparc64/unistd.h linux-2.6.17.13-g-v/include/asm-sparc64/unistd.h --- linux-2.6.17.13/include/asm-sparc64/unistd.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-sparc64/unistd.h 2006-09-20 15:00:12.000000000 +0200 @@ -285,7 +285,7 @@ #define __NR_timer_getoverrun 264 #define __NR_timer_delete 265 #define __NR_timer_create 266 -/* #define __NR_vserver 267 Reserved for VSERVER */ +#define __NR_vserver 267 #define __NR_io_setup 268 #define __NR_io_destroy 269 #define __NR_io_submit 270 diff -NurpP --minimal linux-2.6.17.13/include/asm-v850/kmap_types.h linux-2.6.17.13-g-v/include/asm-v850/kmap_types.h --- linux-2.6.17.13/include/asm-v850/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-v850/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -13,6 +13,7 @@ enum km_type { KM_PTE1, KM_IRQ0, KM_IRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-x86_64/a.out.h linux-2.6.17.13-g-v/include/asm-x86_64/a.out.h --- linux-2.6.17.13/include/asm-x86_64/a.out.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-x86_64/a.out.h 2006-09-20 15:00:12.000000000 +0200 @@ -21,7 +21,7 @@ struct exec #ifdef __KERNEL__ #include -#define STACK_TOP TASK_SIZE +#define __STACK_TOP TASK_SIZE #endif #endif /* __A_OUT_GNU_H__ */ diff -NurpP --minimal linux-2.6.17.13/include/asm-x86_64/elf.h linux-2.6.17.13-g-v/include/asm-x86_64/elf.h --- linux-2.6.17.13/include/asm-x86_64/elf.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-x86_64/elf.h 2006-09-20 15:00:12.000000000 +0200 @@ -90,6 +90,17 @@ typedef struct user_i387_struct elf_fpre #define ELF_ET_DYN_BASE (2 * TASK_SIZE / 3) +#ifdef CONFIG_PAX_ASLR +#define PAX_ELF_ET_DYN_BASE(tsk) (test_thread_flag(TIF_IA32) ? 0x08048000UL : 0x400000UL) + +#define PAX_DELTA_MMAP_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_MMAP_LEN(tsk) (test_thread_flag(TIF_IA32) ? 16 : 32) +#define PAX_DELTA_EXEC_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_EXEC_LEN(tsk) (test_thread_flag(TIF_IA32) ? 16 : 32) +#define PAX_DELTA_STACK_LSB(tsk) PAGE_SHIFT +#define PAX_DELTA_STACK_LEN(tsk) (test_thread_flag(TIF_IA32) ? 16 : 32) +#endif + /* regs is struct pt_regs, pr_reg is elf_gregset_t (which is now struct_user_regs, they are different). Assumes current is the process getting dumped. */ diff -NurpP --minimal linux-2.6.17.13/include/asm-x86_64/ia32.h linux-2.6.17.13-g-v/include/asm-x86_64/ia32.h --- linux-2.6.17.13/include/asm-x86_64/ia32.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-x86_64/ia32.h 2006-09-20 15:00:12.000000000 +0200 @@ -157,7 +157,13 @@ struct ustat32 { char f_fpack[6]; }; -#define IA32_STACK_TOP IA32_PAGE_OFFSET +#ifdef CONFIG_PAX_RANDUSTACK +#define IA32_DELTA_STACK (current->mm->delta_stack) +#else +#define IA32_DELTA_STACK 0UL +#endif + +#define IA32_STACK_TOP (IA32_PAGE_OFFSET - IA32_DELTA_STACK) #ifdef __KERNEL__ struct user_desc; diff -NurpP --minimal linux-2.6.17.13/include/asm-x86_64/kmap_types.h linux-2.6.17.13-g-v/include/asm-x86_64/kmap_types.h --- linux-2.6.17.13/include/asm-x86_64/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-x86_64/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -13,6 +13,7 @@ enum km_type { KM_IRQ1, KM_SOFTIRQ0, KM_SOFTIRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/asm-x86_64/page.h linux-2.6.17.13-g-v/include/asm-x86_64/page.h --- linux-2.6.17.13/include/asm-x86_64/page.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-x86_64/page.h 2006-09-20 15:00:12.000000000 +0200 @@ -134,6 +134,15 @@ typedef struct { unsigned long pgprot; } (((current->personality & READ_IMPLIES_EXEC) ? VM_EXEC : 0 ) | \ VM_READ | VM_WRITE | VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC) +#ifdef CONFIG_PAX_PAGEEXEC +#ifdef CONFIG_PAX_MPROTECT +#define __VM_STACK_FLAGS (((current->mm->pax_flags & MF_PAX_MPROTECT)?0:VM_MAYEXEC) | \ + ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#else +#define __VM_STACK_FLAGS (VM_MAYEXEC | ((current->mm->pax_flags & MF_PAX_PAGEEXEC)?0:VM_EXEC)) +#endif +#endif + #define __HAVE_ARCH_GATE_AREA 1 #endif /* __KERNEL__ */ diff -NurpP --minimal linux-2.6.17.13/include/asm-x86_64/pgalloc.h linux-2.6.17.13-g-v/include/asm-x86_64/pgalloc.h --- linux-2.6.17.13/include/asm-x86_64/pgalloc.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-x86_64/pgalloc.h 2006-09-20 15:00:12.000000000 +0200 @@ -7,7 +7,7 @@ #include #define pmd_populate_kernel(mm, pmd, pte) \ - set_pmd(pmd, __pmd(_PAGE_TABLE | __pa(pte))) + set_pmd(pmd, __pmd(_KERNPG_TABLE | __pa(pte))) #define pud_populate(mm, pud, pmd) \ set_pud(pud, __pud(_PAGE_TABLE | __pa(pmd))) #define pgd_populate(mm, pgd, pud) \ diff -NurpP --minimal linux-2.6.17.13/include/asm-x86_64/pgtable.h linux-2.6.17.13-g-v/include/asm-x86_64/pgtable.h --- linux-2.6.17.13/include/asm-x86_64/pgtable.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-x86_64/pgtable.h 2006-09-20 15:00:12.000000000 +0200 @@ -180,6 +180,10 @@ static inline pte_t ptep_get_and_clear_f #define PAGE_COPY_EXEC __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED) #define PAGE_READONLY __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED | _PAGE_NX) #define PAGE_READONLY_EXEC __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED) + +#define PAGE_READONLY_NOEXEC PAGE_READONLY +#define PAGE_SHARED_NOEXEC PAGE_SHARED + #define __PAGE_KERNEL \ (_PAGE_PRESENT | _PAGE_RW | _PAGE_DIRTY | _PAGE_ACCESSED | _PAGE_NX) #define __PAGE_KERNEL_EXEC \ @@ -268,7 +272,13 @@ static inline pte_t pfn_pte(unsigned lon #define __LARGE_PTE (_PAGE_PSE|_PAGE_PRESENT) static inline int pte_user(pte_t pte) { return pte_val(pte) & _PAGE_USER; } static inline int pte_read(pte_t pte) { return pte_val(pte) & _PAGE_USER; } -static inline int pte_exec(pte_t pte) { return pte_val(pte) & _PAGE_USER; } +extern inline int pte_exec(pte_t pte) +{ + if (__supported_pte_mask & _PAGE_NX) + return pte_val(pte) & _PAGE_NX; + else + return pte_val(pte) & _PAGE_USER; +} static inline int pte_dirty(pte_t pte) { return pte_val(pte) & _PAGE_DIRTY; } static inline int pte_young(pte_t pte) { return pte_val(pte) & _PAGE_ACCESSED; } static inline int pte_write(pte_t pte) { return pte_val(pte) & _PAGE_RW; } @@ -276,12 +286,26 @@ static inline int pte_file(pte_t pte) { static inline int pte_huge(pte_t pte) { return pte_val(pte) & _PAGE_PSE; } static inline pte_t pte_rdprotect(pte_t pte) { set_pte(&pte, __pte(pte_val(pte) & ~_PAGE_USER)); return pte; } -static inline pte_t pte_exprotect(pte_t pte) { set_pte(&pte, __pte(pte_val(pte) & ~_PAGE_USER)); return pte; } +extern inline pte_t pte_exprotect(pte_t pte) +{ + if (__supported_pte_mask & _PAGE_NX) + set_pte(&pte, __pte(pte_val(pte) & ~_PAGE_NX)); + else + set_pte(&pte, __pte(pte_val(pte) & ~_PAGE_USER)); + return pte; +} static inline pte_t pte_mkclean(pte_t pte) { set_pte(&pte, __pte(pte_val(pte) & ~_PAGE_DIRTY)); return pte; } static inline pte_t pte_mkold(pte_t pte) { set_pte(&pte, __pte(pte_val(pte) & ~_PAGE_ACCESSED)); return pte; } static inline pte_t pte_wrprotect(pte_t pte) { set_pte(&pte, __pte(pte_val(pte) & ~_PAGE_RW)); return pte; } static inline pte_t pte_mkread(pte_t pte) { set_pte(&pte, __pte(pte_val(pte) | _PAGE_USER)); return pte; } -static inline pte_t pte_mkexec(pte_t pte) { set_pte(&pte, __pte(pte_val(pte) | _PAGE_USER)); return pte; } +extern inline pte_t pte_mkexec(pte_t pte) +{ + if (__supported_pte_mask & _PAGE_NX) + set_pte(&pte, __pte(pte_val(pte) | _PAGE_NX)); + else + set_pte(&pte, __pte(pte_val(pte) | _PAGE_USER)); + return pte; +} static inline pte_t pte_mkdirty(pte_t pte) { set_pte(&pte, __pte(pte_val(pte) | _PAGE_DIRTY)); return pte; } static inline pte_t pte_mkyoung(pte_t pte) { set_pte(&pte, __pte(pte_val(pte) | _PAGE_ACCESSED)); return pte; } static inline pte_t pte_mkwrite(pte_t pte) { set_pte(&pte, __pte(pte_val(pte) | _PAGE_RW)); return pte; } diff -NurpP --minimal linux-2.6.17.13/include/asm-x86_64/system.h linux-2.6.17.13-g-v/include/asm-x86_64/system.h --- linux-2.6.17.13/include/asm-x86_64/system.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-x86_64/system.h 2006-09-20 15:00:12.000000000 +0200 @@ -366,6 +366,6 @@ static inline unsigned long __cmpxchg(vo void cpu_idle_wait(void); -extern unsigned long arch_align_stack(unsigned long sp); +#define arch_align_stack(x) (x) #endif diff -NurpP --minimal linux-2.6.17.13/include/asm-x86_64/unistd.h linux-2.6.17.13-g-v/include/asm-x86_64/unistd.h --- linux-2.6.17.13/include/asm-x86_64/unistd.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-x86_64/unistd.h 2006-09-20 15:00:12.000000000 +0200 @@ -532,7 +532,7 @@ __SYSCALL(__NR_tgkill, sys_tgkill) #define __NR_utimes 235 __SYSCALL(__NR_utimes, sys_utimes) #define __NR_vserver 236 -__SYSCALL(__NR_vserver, sys_ni_syscall) +__SYSCALL(__NR_vserver, sys_vserver) #define __NR_mbind 237 __SYSCALL(__NR_mbind, sys_mbind) #define __NR_set_mempolicy 238 diff -NurpP --minimal linux-2.6.17.13/include/asm-xtensa/kmap_types.h linux-2.6.17.13-g-v/include/asm-xtensa/kmap_types.h --- linux-2.6.17.13/include/asm-xtensa/kmap_types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/asm-xtensa/kmap_types.h 2006-09-20 15:00:12.000000000 +0200 @@ -25,6 +25,7 @@ enum km_type { KM_IRQ1, KM_SOFTIRQ0, KM_SOFTIRQ1, + KM_CLEARPAGE, KM_TYPE_NR }; diff -NurpP --minimal linux-2.6.17.13/include/linux/a.out.h linux-2.6.17.13-g-v/include/linux/a.out.h --- linux-2.6.17.13/include/linux/a.out.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/a.out.h 2006-09-20 15:00:12.000000000 +0200 @@ -7,6 +7,16 @@ #include +#ifdef CONFIG_PAX_RANDUSTACK +#define __DELTA_STACK (current->mm->delta_stack) +#else +#define __DELTA_STACK 0UL +#endif + +#ifndef STACK_TOP +#define STACK_TOP (__STACK_TOP - __DELTA_STACK) +#endif + #endif /* __STRUCT_EXEC_OVERRIDE__ */ /* these go in the N_MACHTYPE field */ @@ -37,6 +47,14 @@ enum machine_type { M_MIPS2 = 152 /* MIPS R6000/R4000 binary */ }; +/* Constants for the N_FLAGS field */ +#define F_PAX_PAGEEXEC 1 /* Paging based non-executable pages */ +#define F_PAX_EMUTRAMP 2 /* Emulate trampolines */ +#define F_PAX_MPROTECT 4 /* Restrict mprotect() */ +#define F_PAX_RANDMMAP 8 /* Randomize mmap() base */ +/*#define F_PAX_RANDEXEC 16*/ /* Randomize ET_EXEC base */ +#define F_PAX_SEGMEXEC 32 /* Segmentation based non-executable pages */ + #if !defined (N_MAGIC) #define N_MAGIC(exec) ((exec).a_info & 0xffff) #endif diff -NurpP --minimal linux-2.6.17.13/include/linux/binfmts.h linux-2.6.17.13-g-v/include/linux/binfmts.h --- linux-2.6.17.13/include/linux/binfmts.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/binfmts.h 2006-09-20 15:00:12.000000000 +0200 @@ -7,10 +7,10 @@ struct pt_regs; /* * MAX_ARG_PAGES defines the number of pages allocated for arguments - * and envelope for the new program. 32 should suffice, this gives - * a maximum env+arg of 128kB w/4KB pages! + * and envelope for the new program. 33 should suffice, this gives + * a maximum env+arg of 132kB w/4KB pages! */ -#define MAX_ARG_PAGES 32 +#define MAX_ARG_PAGES 33 /* sizeof(linux_binprm->buf) */ #define BINPRM_BUF_SIZE 128 @@ -38,6 +38,7 @@ struct linux_binprm{ unsigned interp_flags; unsigned interp_data; unsigned long loader, exec; + int misc; }; #define BINPRM_FLAGS_ENFORCE_NONDUMP_BIT 0 @@ -87,5 +88,8 @@ extern void compute_creds(struct linux_b extern int do_coredump(long signr, int exit_code, struct pt_regs * regs); extern int set_binfmt(struct linux_binfmt *new); +void pax_report_fault(struct pt_regs *regs, void *pc, void *sp); +void pax_report_insns(void *pc, void *sp); + #endif /* __KERNEL__ */ #endif /* _LINUX_BINFMTS_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/capability.h linux-2.6.17.13-g-v/include/linux/capability.h --- linux-2.6.17.13/include/linux/capability.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/capability.h 2006-09-20 15:00:12.000000000 +0200 @@ -235,6 +235,7 @@ typedef __u32 kernel_cap_t; arbitrary SCSI commands */ /* Allow setting encryption key on loopback filesystem */ /* Allow setting zone reclaim policy */ +/* Allow the selection of a security context */ #define CAP_SYS_ADMIN 21 @@ -288,6 +289,11 @@ typedef __u32 kernel_cap_t; #define CAP_AUDIT_CONTROL 30 +/* Allow context manipulations */ +/* Allow changing context info on files */ + +#define CAP_CONTEXT 31 + #ifdef __KERNEL__ /* * Bounding set @@ -358,6 +364,7 @@ static inline kernel_cap_t cap_invert(ke #define cap_is_fs_cap(c) (CAP_TO_MASK(c) & CAP_FS_MASK) int capable(int cap); +int capable_nolog(int cap); int __capable(struct task_struct *t, int cap); #endif /* __KERNEL__ */ diff -NurpP --minimal linux-2.6.17.13/include/linux/devpts_fs.h linux-2.6.17.13-g-v/include/linux/devpts_fs.h --- linux-2.6.17.13/include/linux/devpts_fs.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/devpts_fs.h 2006-09-20 15:00:12.000000000 +0200 @@ -30,5 +30,7 @@ static inline void devpts_pty_kill(int n #endif +#define DEVPTS_SUPER_MAGIC 0x00001cd1 + #endif /* _LINUX_DEVPTS_FS_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/elf.h linux-2.6.17.13-g-v/include/linux/elf.h --- linux-2.6.17.13/include/linux/elf.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/elf.h 2006-09-20 15:00:12.000000000 +0200 @@ -5,6 +5,10 @@ #include #include +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) +#undef elf_read_implies_exec +#endif + #ifndef elf_read_implies_exec /* Executables for which elf_read_implies_exec() returns TRUE will have the READ_IMPLIES_EXEC personality flag set automatically. @@ -46,6 +50,16 @@ typedef __s64 Elf64_Sxword; #define PT_GNU_STACK (PT_LOOS + 0x474e551) +#define PT_PAX_FLAGS (PT_LOOS + 0x5041580) + +/* Constants for the e_flags field */ +#define EF_PAX_PAGEEXEC 1 /* Paging based non-executable pages */ +#define EF_PAX_EMUTRAMP 2 /* Emulate trampolines */ +#define EF_PAX_MPROTECT 4 /* Restrict mprotect() */ +#define EF_PAX_RANDMMAP 8 /* Randomize mmap() base */ +/*#define EF_PAX_RANDEXEC 16*/ /* Randomize ET_EXEC base */ +#define EF_PAX_SEGMEXEC 32 /* Segmentation based non-executable pages */ + /* These constants define the different elf file types */ #define ET_NONE 0 #define ET_REL 1 @@ -138,6 +152,8 @@ typedef __s64 Elf64_Sxword; #define DT_DEBUG 21 #define DT_TEXTREL 22 #define DT_JMPREL 23 +#define DT_FLAGS 30 + #define DF_TEXTREL 0x00000004 #define DT_LOPROC 0x70000000 #define DT_HIPROC 0x7fffffff @@ -267,6 +283,19 @@ typedef struct elf64_hdr { #define PF_W 0x2 #define PF_X 0x1 +#define PF_PAGEEXEC (1U << 4) /* Enable PAGEEXEC */ +#define PF_NOPAGEEXEC (1U << 5) /* Disable PAGEEXEC */ +#define PF_SEGMEXEC (1U << 6) /* Enable SEGMEXEC */ +#define PF_NOSEGMEXEC (1U << 7) /* Disable SEGMEXEC */ +#define PF_MPROTECT (1U << 8) /* Enable MPROTECT */ +#define PF_NOMPROTECT (1U << 9) /* Disable MPROTECT */ +/*#define PF_RANDEXEC (1U << 10)*/ /* Enable RANDEXEC */ +/*#define PF_NORANDEXEC (1U << 11)*/ /* Disable RANDEXEC */ +#define PF_EMUTRAMP (1U << 12) /* Enable EMUTRAMP */ +#define PF_NOEMUTRAMP (1U << 13) /* Disable EMUTRAMP */ +#define PF_RANDMMAP (1U << 14) /* Enable RANDMMAP */ +#define PF_NORANDMMAP (1U << 15) /* Disable RANDMMAP */ + typedef struct elf32_phdr{ Elf32_Word p_type; Elf32_Off p_offset; @@ -359,6 +388,8 @@ typedef struct elf64_shdr { #define EI_OSABI 7 #define EI_PAD 8 +#define EI_PAX 14 + #define ELFMAG0 0x7f /* EI_MAG */ #define ELFMAG1 'E' #define ELFMAG2 'L' @@ -415,6 +446,7 @@ extern Elf32_Dyn _DYNAMIC []; #define elfhdr elf32_hdr #define elf_phdr elf32_phdr #define elf_note elf32_note +#define elf_dyn Elf32_Dyn #else @@ -422,6 +454,7 @@ extern Elf64_Dyn _DYNAMIC []; #define elfhdr elf64_hdr #define elf_phdr elf64_phdr #define elf_note elf64_note +#define elf_dyn Elf64_Dyn #endif diff -NurpP --minimal linux-2.6.17.13/include/linux/ext2_fs.h linux-2.6.17.13-g-v/include/linux/ext2_fs.h --- linux-2.6.17.13/include/linux/ext2_fs.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/ext2_fs.h 2006-09-20 15:00:12.000000000 +0200 @@ -192,10 +192,17 @@ struct ext2_group_desc #define EXT2_NOTAIL_FL 0x00008000 /* file tail should not be merged */ #define EXT2_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */ #define EXT2_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/ +#define EXT2_BARRIER_FL 0x04000000 /* Barrier for chroot() */ +#define EXT2_IUNLINK_FL 0x08000000 /* Immutable unlink */ #define EXT2_RESERVED_FL 0x80000000 /* reserved for ext2 lib */ +#ifdef CONFIG_VSERVER_LEGACY +#define EXT2_FL_USER_VISIBLE 0x0803DFFF /* User visible flags */ +#define EXT2_FL_USER_MODIFIABLE 0x080380FF /* User modifiable flags */ +#else #define EXT2_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */ #define EXT2_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */ +#endif /* * ioctl commands @@ -240,7 +247,7 @@ struct ext2_inode { struct { __u8 l_i_frag; /* Fragment number */ __u8 l_i_fsize; /* Fragment size */ - __u16 i_pad1; + __u16 l_i_xid; /* LRU Context */ __le16 l_i_uid_high; /* these 2 fields */ __le16 l_i_gid_high; /* were reserved2[0] */ __u32 l_i_reserved2; @@ -272,6 +279,7 @@ struct ext2_inode { #define i_gid_low i_gid #define i_uid_high osd2.linux2.l_i_uid_high #define i_gid_high osd2.linux2.l_i_gid_high +#define i_raw_xid osd2.linux2.l_i_xid #define i_reserved2 osd2.linux2.l_i_reserved2 #endif @@ -313,8 +321,9 @@ struct ext2_inode { #define EXT2_MOUNT_XATTR_USER 0x004000 /* Extended user attributes */ #define EXT2_MOUNT_POSIX_ACL 0x008000 /* POSIX Access Control Lists */ #define EXT2_MOUNT_XIP 0x010000 /* Execute in place */ -#define EXT2_MOUNT_USRQUOTA 0x020000 /* user quota */ -#define EXT2_MOUNT_GRPQUOTA 0x040000 /* group quota */ +#define EXT2_MOUNT_USRQUOTA 0x020000 /* user quota */ +#define EXT2_MOUNT_GRPQUOTA 0x040000 /* group quota */ +#define EXT2_MOUNT_TAGXID (1<<24) /* Enable Context Tags */ #define clear_opt(o, opt) o &= ~EXT2_MOUNT_##opt diff -NurpP --minimal linux-2.6.17.13/include/linux/ext3_fs.h linux-2.6.17.13-g-v/include/linux/ext3_fs.h --- linux-2.6.17.13/include/linux/ext3_fs.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/ext3_fs.h 2006-09-20 15:00:12.000000000 +0200 @@ -186,10 +186,20 @@ struct ext3_group_desc #define EXT3_NOTAIL_FL 0x00008000 /* file tail should not be merged */ #define EXT3_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */ #define EXT3_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/ +#define EXT3_BARRIER_FL 0x04000000 /* Barrier for chroot() */ +#define EXT3_IUNLINK_FL 0x08000000 /* Immutable unlink */ #define EXT3_RESERVED_FL 0x80000000 /* reserved for ext3 lib */ +#ifdef CONFIG_VSERVER_LEGACY +#define EXT3_FL_USER_VISIBLE 0x0803DFFF /* User visible flags */ +#define EXT3_FL_USER_MODIFIABLE 0x080380FF /* User modifiable flags */ +#else #define EXT3_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */ #define EXT3_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */ +#endif +#ifdef CONFIG_VSERVER_LEGACY +#define EXT3_IOC_SETXID FIOC_SETXIDJ +#endif /* * Inode dynamic state flags @@ -288,7 +298,7 @@ struct ext3_inode { struct { __u8 l_i_frag; /* Fragment number */ __u8 l_i_fsize; /* Fragment size */ - __u16 i_pad1; + __u16 l_i_xid; /* LRU Context */ __le16 l_i_uid_high; /* these 2 fields */ __le16 l_i_gid_high; /* were reserved2[0] */ __u32 l_i_reserved2; @@ -322,6 +332,7 @@ struct ext3_inode { #define i_gid_low i_gid #define i_uid_high osd2.linux2.l_i_uid_high #define i_gid_high osd2.linux2.l_i_gid_high +#define i_raw_xid osd2.linux2.l_i_xid #define i_reserved2 osd2.linux2.l_i_reserved2 #elif defined(__GNU__) @@ -376,6 +387,7 @@ struct ext3_inode { #define EXT3_MOUNT_QUOTA 0x80000 /* Some quota option set */ #define EXT3_MOUNT_USRQUOTA 0x100000 /* "old" user quota */ #define EXT3_MOUNT_GRPQUOTA 0x200000 /* "old" group quota */ +#define EXT3_MOUNT_TAGXID (1<<24) /* Enable Context Tags */ /* Compatibility, for having both ext2_fs.h and ext3_fs.h included at once */ #ifndef _LINUX_EXT2_FS_H @@ -790,6 +802,7 @@ struct buffer_head * ext3_bread (handle_ int ext3_get_blocks_handle(handle_t *handle, struct inode *inode, sector_t iblock, unsigned long maxblocks, struct buffer_head *bh_result, int create, int extend_disksize); +extern int ext3_sync_flags(struct inode *inode); extern void ext3_read_inode (struct inode *); extern int ext3_write_inode (struct inode *, int); diff -NurpP --minimal linux-2.6.17.13/include/linux/fs.h linux-2.6.17.13-g-v/include/linux/fs.h --- linux-2.6.17.13/include/linux/fs.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/fs.h 2006-09-20 15:00:12.000000000 +0200 @@ -116,6 +116,8 @@ extern int dir_notify_enable; #define MS_PRIVATE (1<<18) /* change to private */ #define MS_SLAVE (1<<19) /* change to slave */ #define MS_SHARED (1<<20) /* change to shared */ +#define MS_TAGXID (1<<24) /* tag inodes with context information */ +#define MS_XID (1<<25) /* use specific xid for this mount */ #define MS_ACTIVE (1<<30) #define MS_NOUSER (1<<31) @@ -142,6 +144,8 @@ extern int dir_notify_enable; #define S_NOCMTIME 128 /* Do not update file c/mtime */ #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */ #define S_PRIVATE 512 /* Inode is fs-internal */ +#define S_BARRIER 1024 /* Barrier for chroot() */ +#define S_IUNLINK 2048 /* Immutable unlink */ /* * Note that nosuid etc flags are inode-specific: setting some file-system @@ -158,18 +162,22 @@ extern int dir_notify_enable; */ #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg)) -#define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY) +#define IS_RDONLY(inode) __IS_FLG(inode, MS_RDONLY) #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \ ((inode)->i_flags & S_SYNC)) #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \ ((inode)->i_flags & (S_SYNC|S_DIRSYNC))) #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK) +#define IS_TAGXID(inode) __IS_FLG(inode, MS_TAGXID) #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA) #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND) #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE) +#define IS_IUNLINK(inode) ((inode)->i_flags & S_IUNLINK) +#define IS_IXORUNLINK(inode) ((IS_IUNLINK(inode) ? S_IMMUTABLE : 0) ^ IS_IMMUTABLE(inode)) #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL) +#define IS_BARRIER(inode) (S_ISDIR((inode)->i_mode) && ((inode)->i_flags & S_BARRIER)) #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD) #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME) #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE) @@ -277,6 +285,7 @@ typedef void (dio_iodone_t)(struct kiocb #define ATTR_KILL_SUID 2048 #define ATTR_KILL_SGID 4096 #define ATTR_FILE 8192 +#define ATTR_XID 16384 /* * This is the Inode Attributes structure, used for notify_change(). It @@ -292,6 +301,7 @@ struct iattr { umode_t ia_mode; uid_t ia_uid; gid_t ia_gid; + xid_t ia_xid; loff_t ia_size; struct timespec ia_atime; struct timespec ia_mtime; @@ -305,6 +315,9 @@ struct iattr { struct file *ia_file; }; +#define ATTR_FLAG_BARRIER 512 /* Barrier for chroot() */ +#define ATTR_FLAG_IUNLINK 1024 /* Immutable unlink */ + /* * Includes for diskquotas. */ @@ -486,6 +499,7 @@ struct inode { unsigned int i_nlink; uid_t i_uid; gid_t i_gid; + xid_t i_xid; dev_t i_rdev; loff_t i_size; struct timespec i_atime; @@ -648,6 +662,7 @@ struct file { struct fown_struct f_owner; unsigned int f_uid, f_gid; struct file_ra_state f_ra; + xid_t f_xid; unsigned long f_version; void *f_security; @@ -726,6 +741,7 @@ struct file_lock { unsigned char fl_type; loff_t fl_start; loff_t fl_end; + xid_t fl_xid; struct fasync_struct * fl_fasync; /* for lease break notifications */ unsigned long fl_break_time; /* for nonblocking lease breaks */ @@ -920,12 +936,12 @@ static inline void unlock_super(struct s */ extern int vfs_permission(struct nameidata *, int); extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *); -extern int vfs_mkdir(struct inode *, struct dentry *, int); -extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t); -extern int vfs_symlink(struct inode *, struct dentry *, const char *, int); -extern int vfs_link(struct dentry *, struct inode *, struct dentry *); -extern int vfs_rmdir(struct inode *, struct dentry *); -extern int vfs_unlink(struct inode *, struct dentry *); +extern int vfs_mkdir(struct inode *, struct dentry *, int, struct nameidata *); +extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t, struct nameidata *); +extern int vfs_symlink(struct inode *, struct dentry *, const char *, int, struct nameidata *); +extern int vfs_link(struct dentry *, struct inode *, struct dentry *, struct nameidata *); +extern int vfs_rmdir(struct inode *, struct dentry *, struct nameidata *); +extern int vfs_unlink(struct inode *, struct dentry *, struct nameidata *); extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *); /* @@ -1067,6 +1083,7 @@ struct inode_operations { ssize_t (*listxattr) (struct dentry *, char *, size_t); int (*removexattr) (struct dentry *, const char *); void (*truncate_range)(struct inode *, loff_t, loff_t); + int (*sync_flags) (struct inode *); }; struct seq_file; @@ -1742,6 +1759,7 @@ extern int dcache_dir_open(struct inode extern int dcache_dir_close(struct inode *, struct file *); extern loff_t dcache_dir_lseek(struct file *, loff_t, int); extern int dcache_readdir(struct file *, void *, filldir_t); +extern int dcache_readdir_filter(struct file *, void *, filldir_t, int (*)(struct dentry *)); extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *); extern int simple_statfs(struct super_block *, struct kstatfs *); extern int simple_link(struct dentry *, struct inode *, struct dentry *); diff -NurpP --minimal linux-2.6.17.13/include/linux/gracl.h linux-2.6.17.13-g-v/include/linux/gracl.h --- linux-2.6.17.13/include/linux/gracl.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/gracl.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,316 @@ +#ifndef GR_ACL_H +#define GR_ACL_H + +#include +#include +#include +#include + +/* Major status information */ + +#define GR_VERSION "grsecurity 2.1.9" +#define GRSECURITY_VERSION 0x219 + +enum { + + SHUTDOWN = 0, + ENABLE = 1, + SPROLE = 2, + RELOAD = 3, + SEGVMOD = 4, + STATUS = 5, + UNSPROLE = 6, + PASSSET = 7, + SPROLEPAM = 8 +}; + +/* Password setup definitions + * kernel/grhash.c */ +enum { + GR_PW_LEN = 128, + GR_SALT_LEN = 16, + GR_SHA_LEN = 32, +}; + +enum { + GR_SPROLE_LEN = 64, +}; + +#define GR_NLIMITS (RLIMIT_LOCKS + 2) + +/* Begin Data Structures */ + +struct sprole_pw { + unsigned char *rolename; + unsigned char salt[GR_SALT_LEN]; + unsigned char sum[GR_SHA_LEN]; /* 256-bit SHA hash of the password */ +}; + +struct name_entry { + __u32 key; + ino_t inode; + dev_t device; + char *name; + __u16 len; + struct name_entry *prev; + struct name_entry *next; +}; + +struct inodev_entry { + struct name_entry *nentry; + struct inodev_entry *prev; + struct inodev_entry *next; +}; + +struct acl_role_db { + struct acl_role_label **r_hash; + __u32 r_size; +}; + +struct inodev_db { + struct inodev_entry **i_hash; + __u32 i_size; +}; + +struct name_db { + struct name_entry **n_hash; + __u32 n_size; +}; + +struct crash_uid { + uid_t uid; + unsigned long expires; +}; + +struct gr_hash_struct { + void **table; + void **nametable; + void *first; + __u32 table_size; + __u32 used_size; + int type; +}; + +/* Userspace Grsecurity ACL data structures */ + +struct acl_subject_label { + char *filename; + ino_t inode; + dev_t device; + __u32 mode; + __u32 cap_mask; + __u32 cap_lower; + + struct rlimit res[GR_NLIMITS]; + __u16 resmask; + + __u8 user_trans_type; + __u8 group_trans_type; + uid_t *user_transitions; + gid_t *group_transitions; + __u16 user_trans_num; + __u16 group_trans_num; + + __u32 ip_proto[8]; + __u32 ip_type; + struct acl_ip_label **ips; + __u32 ip_num; + + __u32 crashes; + unsigned long expires; + + struct acl_subject_label *parent_subject; + struct gr_hash_struct *hash; + struct acl_subject_label *prev; + struct acl_subject_label *next; + + struct acl_object_label **obj_hash; + __u32 obj_hash_size; + __u16 pax_flags; +}; + +struct role_allowed_ip { + __u32 addr; + __u32 netmask; + + struct role_allowed_ip *prev; + struct role_allowed_ip *next; +}; + +struct role_transition { + char *rolename; + + struct role_transition *prev; + struct role_transition *next; +}; + +struct acl_role_label { + char *rolename; + uid_t uidgid; + __u16 roletype; + + __u16 auth_attempts; + unsigned long expires; + + struct acl_subject_label *root_label; + struct gr_hash_struct *hash; + + struct acl_role_label *prev; + struct acl_role_label *next; + + struct role_transition *transitions; + struct role_allowed_ip *allowed_ips; + uid_t *domain_children; + __u16 domain_child_num; + + struct acl_subject_label **subj_hash; + __u32 subj_hash_size; +}; + +struct user_acl_role_db { + struct acl_role_label **r_table; + __u32 num_pointers; /* Number of allocations to track */ + __u32 num_roles; /* Number of roles */ + __u32 num_domain_children; /* Number of domain children */ + __u32 num_subjects; /* Number of subjects */ + __u32 num_objects; /* Number of objects */ +}; + +struct acl_object_label { + char *filename; + ino_t inode; + dev_t device; + __u32 mode; + + struct acl_subject_label *nested; + struct acl_object_label *globbed; + + /* next two structures not used */ + + struct acl_object_label *prev; + struct acl_object_label *next; +}; + +struct acl_ip_label { + char *iface; + __u32 addr; + __u32 netmask; + __u16 low, high; + __u8 mode; + __u32 type; + __u32 proto[8]; + + /* next two structures not used */ + + struct acl_ip_label *prev; + struct acl_ip_label *next; +}; + +struct gr_arg { + struct user_acl_role_db role_db; + unsigned char pw[GR_PW_LEN]; + unsigned char salt[GR_SALT_LEN]; + unsigned char sum[GR_SHA_LEN]; + unsigned char sp_role[GR_SPROLE_LEN]; + struct sprole_pw *sprole_pws; + dev_t segv_device; + ino_t segv_inode; + uid_t segv_uid; + __u16 num_sprole_pws; + __u16 mode; +}; + +struct gr_arg_wrapper { + struct gr_arg *arg; + __u32 version; + __u32 size; +}; + +struct subject_map { + struct acl_subject_label *user; + struct acl_subject_label *kernel; + struct subject_map *prev; + struct subject_map *next; +}; + +struct acl_subj_map_db { + struct subject_map **s_hash; + __u32 s_size; +}; + +/* End Data Structures Section */ + +/* Hash functions generated by empirical testing by Brad Spengler + Makes good use of the low bits of the inode. Generally 0-1 times + in loop for successful match. 0-3 for unsuccessful match. + Shift/add algorithm with modulus of table size and an XOR*/ + +static __inline__ unsigned int +rhash(const uid_t uid, const __u16 type, const unsigned int sz) +{ + return (((uid << type) + (uid ^ type)) % sz); +} + + static __inline__ unsigned int +shash(const struct acl_subject_label *userp, const unsigned int sz) +{ + return ((const unsigned long)userp % sz); +} + +static __inline__ unsigned int +fhash(const ino_t ino, const dev_t dev, const unsigned int sz) +{ + return (((ino + dev) ^ ((ino << 13) + (ino << 23) + (dev << 9))) % sz); +} + +static __inline__ unsigned int +nhash(const char *name, const __u16 len, const unsigned int sz) +{ + return full_name_hash(name, len) % sz; +} + +#define FOR_EACH_ROLE_START(role,iter) \ + role = NULL; \ + iter = 0; \ + while (iter < acl_role_set.r_size) { \ + if (role == NULL) \ + role = acl_role_set.r_hash[iter]; \ + if (role == NULL) { \ + iter++; \ + continue; \ + } + +#define FOR_EACH_ROLE_END(role,iter) \ + role = role->next; \ + if (role == NULL) \ + iter++; \ + } + +#define FOR_EACH_SUBJECT_START(role,subj,iter) \ + subj = NULL; \ + iter = 0; \ + while (iter < role->subj_hash_size) { \ + if (subj == NULL) \ + subj = role->subj_hash[iter]; \ + if (subj == NULL) { \ + iter++; \ + continue; \ + } + +#define FOR_EACH_SUBJECT_END(subj,iter) \ + subj = subj->next; \ + if (subj == NULL) \ + iter++; \ + } + + +#define FOR_EACH_NESTED_SUBJECT_START(role,subj) \ + subj = role->hash->first; \ + while (subj != NULL) { + +#define FOR_EACH_NESTED_SUBJECT_END(subj) \ + subj = subj->next; \ + } + +#endif + diff -NurpP --minimal linux-2.6.17.13/include/linux/gralloc.h linux-2.6.17.13-g-v/include/linux/gralloc.h --- linux-2.6.17.13/include/linux/gralloc.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/gralloc.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,8 @@ +#ifndef __GRALLOC_H +#define __GRALLOC_H + +void acl_free_all(void); +int acl_alloc_stack_init(unsigned long size); +void *acl_alloc(unsigned long len); + +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/grdefs.h linux-2.6.17.13-g-v/include/linux/grdefs.h --- linux-2.6.17.13/include/linux/grdefs.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/grdefs.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,131 @@ +#ifndef GRDEFS_H +#define GRDEFS_H + +/* Begin grsecurity status declarations */ + +enum { + GR_READY = 0x01, + GR_STATUS_INIT = 0x00 // disabled state +}; + +/* Begin ACL declarations */ + +/* Role flags */ + +enum { + GR_ROLE_USER = 0x0001, + GR_ROLE_GROUP = 0x0002, + GR_ROLE_DEFAULT = 0x0004, + GR_ROLE_SPECIAL = 0x0008, + GR_ROLE_AUTH = 0x0010, + GR_ROLE_NOPW = 0x0020, + GR_ROLE_GOD = 0x0040, + GR_ROLE_LEARN = 0x0080, + GR_ROLE_TPE = 0x0100, + GR_ROLE_DOMAIN = 0x0200, + GR_ROLE_PAM = 0x0400 +}; + +/* ACL Subject and Object mode flags */ +enum { + GR_DELETED = 0x80000000 +}; + +/* ACL Object-only mode flags */ +enum { + GR_READ = 0x00000001, + GR_APPEND = 0x00000002, + GR_WRITE = 0x00000004, + GR_EXEC = 0x00000008, + GR_FIND = 0x00000010, + GR_INHERIT = 0x00000020, + GR_SETID = 0x00000040, + GR_CREATE = 0x00000080, + GR_DELETE = 0x00000100, + GR_LINK = 0x00000200, + GR_AUDIT_READ = 0x00000400, + GR_AUDIT_APPEND = 0x00000800, + GR_AUDIT_WRITE = 0x00001000, + GR_AUDIT_EXEC = 0x00002000, + GR_AUDIT_FIND = 0x00004000, + GR_AUDIT_INHERIT= 0x00008000, + GR_AUDIT_SETID = 0x00010000, + GR_AUDIT_CREATE = 0x00020000, + GR_AUDIT_DELETE = 0x00040000, + GR_AUDIT_LINK = 0x00080000, + GR_PTRACERD = 0x00100000, + GR_NOPTRACE = 0x00200000, + GR_SUPPRESS = 0x00400000, + GR_NOLEARN = 0x00800000 +}; + +#define GR_AUDITS (GR_AUDIT_READ | GR_AUDIT_WRITE | GR_AUDIT_APPEND | GR_AUDIT_EXEC | \ + GR_AUDIT_FIND | GR_AUDIT_INHERIT | GR_AUDIT_SETID | \ + GR_AUDIT_CREATE | GR_AUDIT_DELETE | GR_AUDIT_LINK) + +/* ACL subject-only mode flags */ +enum { + GR_KILL = 0x00000001, + GR_VIEW = 0x00000002, + GR_PROTECTED = 0x00000004, + GR_LEARN = 0x00000008, + GR_OVERRIDE = 0x00000010, + /* just a placeholder, this mode is only used in userspace */ + GR_DUMMY = 0x00000020, + GR_PROTSHM = 0x00000040, + GR_KILLPROC = 0x00000080, + GR_KILLIPPROC = 0x00000100, + /* just a placeholder, this mode is only used in userspace */ + GR_NOTROJAN = 0x00000200, + GR_PROTPROCFD = 0x00000400, + GR_PROCACCT = 0x00000800, + GR_RELAXPTRACE = 0x00001000, + GR_NESTED = 0x00002000, + GR_INHERITLEARN = 0x00004000, + GR_PROCFIND = 0x00008000, + GR_POVERRIDE = 0x00010000, + GR_KERNELAUTH = 0x00020000, +}; + +enum { + GR_PAX_ENABLE_SEGMEXEC = 0x0001, + GR_PAX_ENABLE_PAGEEXEC = 0x0002, + GR_PAX_ENABLE_MPROTECT = 0x0004, + GR_PAX_ENABLE_RANDMMAP = 0x0008, + GR_PAX_ENABLE_EMUTRAMP = 0x0010, + GR_PAX_DISABLE_SEGMEXEC = 0x8001, + GR_PAX_DISABLE_PAGEEXEC = 0x8002, + GR_PAX_DISABLE_MPROTECT = 0x8004, + GR_PAX_DISABLE_RANDMMAP = 0x8008, + GR_PAX_DISABLE_EMUTRAMP = 0x8010, +}; + +enum { + GR_ID_USER = 0x01, + GR_ID_GROUP = 0x02, +}; + +enum { + GR_ID_ALLOW = 0x01, + GR_ID_DENY = 0x02, +}; + +#define GR_CRASH_RES 11 +#define GR_UIDTABLE_MAX 500 + +/* begin resource learning section */ +enum { + GR_RLIM_CPU_BUMP = 60, + GR_RLIM_FSIZE_BUMP = 50000, + GR_RLIM_DATA_BUMP = 10000, + GR_RLIM_STACK_BUMP = 1000, + GR_RLIM_CORE_BUMP = 10000, + GR_RLIM_RSS_BUMP = 500000, + GR_RLIM_NPROC_BUMP = 1, + GR_RLIM_NOFILE_BUMP = 5, + GR_RLIM_MEMLOCK_BUMP = 50000, + GR_RLIM_AS_BUMP = 500000, + GR_RLIM_LOCKS_BUMP = 2 +}; + +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/grinternal.h linux-2.6.17.13-g-v/include/linux/grinternal.h --- linux-2.6.17.13/include/linux/grinternal.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/grinternal.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,211 @@ +#ifndef __GRINTERNAL_H +#define __GRINTERNAL_H + +#ifdef CONFIG_GRKERNSEC + +#include +#include +#include +#include + +extern void gr_add_learn_entry(const char *fmt, ...); +extern __u32 gr_search_file(const struct dentry *dentry, const __u32 mode, + const struct vfsmount *mnt); +extern __u32 gr_check_create(const struct dentry *new_dentry, + const struct dentry *parent, + const struct vfsmount *mnt, const __u32 mode); +extern int gr_check_protected_task(const struct task_struct *task); +extern __u32 to_gr_audit(const __u32 reqmode); +extern int gr_set_acls(const int type); + +extern int gr_acl_is_enabled(void); +extern char gr_roletype_to_char(void); + +extern void gr_handle_alertkill(struct task_struct *task); +extern char *gr_to_filename(const struct dentry *dentry, + const struct vfsmount *mnt); +extern char *gr_to_filename1(const struct dentry *dentry, + const struct vfsmount *mnt); +extern char *gr_to_filename2(const struct dentry *dentry, + const struct vfsmount *mnt); +extern char *gr_to_filename3(const struct dentry *dentry, + const struct vfsmount *mnt); + +extern int grsec_enable_link; +extern int grsec_enable_fifo; +extern int grsec_enable_execve; +extern int grsec_enable_shm; +extern int grsec_enable_execlog; +extern int grsec_enable_signal; +extern int grsec_enable_forkfail; +extern int grsec_enable_time; +extern int grsec_enable_chroot_shmat; +extern int grsec_enable_chroot_findtask; +extern int grsec_enable_chroot_mount; +extern int grsec_enable_chroot_double; +extern int grsec_enable_chroot_pivot; +extern int grsec_enable_chroot_chdir; +extern int grsec_enable_chroot_chmod; +extern int grsec_enable_chroot_mknod; +extern int grsec_enable_chroot_fchdir; +extern int grsec_enable_chroot_nice; +extern int grsec_enable_chroot_execlog; +extern int grsec_enable_chroot_caps; +extern int grsec_enable_chroot_sysctl; +extern int grsec_enable_chroot_unix; +extern int grsec_enable_tpe; +extern int grsec_tpe_gid; +extern int grsec_enable_tpe_all; +extern int grsec_enable_sidcaps; +extern int grsec_enable_randpid; +extern int grsec_enable_socket_all; +extern int grsec_socket_all_gid; +extern int grsec_enable_socket_client; +extern int grsec_socket_client_gid; +extern int grsec_enable_socket_server; +extern int grsec_socket_server_gid; +extern int grsec_audit_gid; +extern int grsec_enable_group; +extern int grsec_enable_audit_ipc; +extern int grsec_enable_audit_textrel; +extern int grsec_enable_mount; +extern int grsec_enable_chdir; +extern int grsec_resource_logging; +extern int grsec_lock; + +extern struct task_struct *child_reaper; + +extern spinlock_t grsec_alert_lock; +extern unsigned long grsec_alert_wtime; +extern unsigned long grsec_alert_fyet; + +extern spinlock_t grsec_audit_lock; + +extern rwlock_t grsec_exec_file_lock; + +#define gr_task_fullpath(tsk) (tsk->exec_file ? \ + gr_to_filename2(tsk->exec_file->f_dentry, \ + tsk->exec_file->f_vfsmnt) : "/") + +#define gr_parent_task_fullpath(tsk) (tsk->parent->exec_file ? \ + gr_to_filename3(tsk->parent->exec_file->f_dentry, \ + tsk->parent->exec_file->f_vfsmnt) : "/") + +#define gr_task_fullpath0(tsk) (tsk->exec_file ? \ + gr_to_filename(tsk->exec_file->f_dentry, \ + tsk->exec_file->f_vfsmnt) : "/") + +#define gr_parent_task_fullpath0(tsk) (tsk->parent->exec_file ? \ + gr_to_filename1(tsk->parent->exec_file->f_dentry, \ + tsk->parent->exec_file->f_vfsmnt) : "/") + +#define proc_is_chrooted(tsk_a) ((tsk_a->pid > 1) && (tsk_a->fs != NULL) && \ + ((tsk_a->fs->root->d_inode->i_sb->s_dev != \ + child_reaper->fs->root->d_inode->i_sb->s_dev) || \ + (tsk_a->fs->root->d_inode->i_ino != \ + child_reaper->fs->root->d_inode->i_ino))) + +#define have_same_root(tsk_a,tsk_b) ((tsk_a->fs != NULL) && (tsk_b->fs != NULL) && \ + (tsk_a->fs->root->d_inode->i_sb->s_dev == \ + tsk_b->fs->root->d_inode->i_sb->s_dev) && \ + (tsk_a->fs->root->d_inode->i_ino == \ + tsk_b->fs->root->d_inode->i_ino)) + +#define DEFAULTSECARGS(task) gr_task_fullpath(task), task->comm, \ + task->pid, task->uid, \ + task->euid, task->gid, task->egid, \ + gr_parent_task_fullpath(task), \ + task->parent->comm, task->parent->pid, \ + task->parent->uid, task->parent->euid, \ + task->parent->gid, task->parent->egid + +#define GR_CHROOT_CAPS ( \ + CAP_TO_MASK(CAP_LINUX_IMMUTABLE) | CAP_TO_MASK(CAP_NET_ADMIN) | \ + CAP_TO_MASK(CAP_SYS_MODULE) | CAP_TO_MASK(CAP_SYS_RAWIO) | \ + CAP_TO_MASK(CAP_SYS_PACCT) | CAP_TO_MASK(CAP_SYS_ADMIN) | \ + CAP_TO_MASK(CAP_SYS_BOOT) | CAP_TO_MASK(CAP_SYS_TIME) | \ + CAP_TO_MASK(CAP_NET_RAW) | CAP_TO_MASK(CAP_SYS_TTY_CONFIG) | \ + CAP_TO_MASK(CAP_IPC_OWNER)) + +#define security_learn(normal_msg,args...) \ +({ \ + read_lock(&grsec_exec_file_lock); \ + gr_add_learn_entry(normal_msg "\n", ## args); \ + read_unlock(&grsec_exec_file_lock); \ +}) + +enum { + GR_DO_AUDIT, + GR_DONT_AUDIT, + GR_DONT_AUDIT_GOOD +}; + +enum { + GR_TTYSNIFF, + GR_RBAC, + GR_RBAC_STR, + GR_STR_RBAC, + GR_RBAC_MODE2, + GR_RBAC_MODE3, + GR_FILENAME, + GR_NOARGS, + GR_ONE_INT, + GR_ONE_INT_TWO_STR, + GR_ONE_STR, + GR_STR_INT, + GR_TWO_INT, + GR_THREE_INT, + GR_FIVE_INT_TWO_STR, + GR_TWO_STR, + GR_THREE_STR, + GR_FOUR_STR, + GR_STR_FILENAME, + GR_FILENAME_STR, + GR_FILENAME_TWO_INT, + GR_FILENAME_TWO_INT_STR, + GR_TEXTREL, + GR_PTRACE, + GR_RESOURCE, + GR_CAP, + GR_SIG, + GR_CRASH1, + GR_CRASH2, + GR_PSACCT +}; + +#define gr_log_ttysniff(audit, msg, task) gr_log_varargs(audit, msg, GR_TTYSNIFF, task) +#define gr_log_fs_rbac_generic(audit, msg, dentry, mnt) gr_log_varargs(audit, msg, GR_RBAC, dentry, mnt) +#define gr_log_fs_rbac_str(audit, msg, dentry, mnt, str) gr_log_varargs(audit, msg, GR_RBAC_STR, dentry, mnt, str) +#define gr_log_fs_str_rbac(audit, msg, str, dentry, mnt) gr_log_varargs(audit, msg, GR_STR_RBAC, str, dentry, mnt) +#define gr_log_fs_rbac_mode2(audit, msg, dentry, mnt, str1, str2) gr_log_varargs(audit, msg, GR_RBAC_MODE2, dentry, mnt, str1, str2) +#define gr_log_fs_rbac_mode3(audit, msg, dentry, mnt, str1, str2, str3) gr_log_varargs(audit, msg, GR_RBAC_MODE3, dentry, mnt, str1, str2, str3) +#define gr_log_fs_generic(audit, msg, dentry, mnt) gr_log_varargs(audit, msg, GR_FILENAME, dentry, mnt) +#define gr_log_noargs(audit, msg) gr_log_varargs(audit, msg, GR_NOARGS) +#define gr_log_int(audit, msg, num) gr_log_varargs(audit, msg, GR_ONE_INT, num) +#define gr_log_int_str2(audit, msg, num, str1, str2) gr_log_varargs(audit, msg, GR_ONE_INT_TWO_STR, num, str1, str2) +#define gr_log_str(audit, msg, str) gr_log_varargs(audit, msg, GR_ONE_STR, str) +#define gr_log_str_int(audit, msg, str, num) gr_log_varargs(audit, msg, GR_STR_INT, str, num) +#define gr_log_int_int(audit, msg, num1, num2) gr_log_varargs(audit, msg, GR_TWO_INT, num1, num2) +#define gr_log_int3(audit, msg, num1, num2, num3) gr_log_varargs(audit, msg, GR_THREE_INT, num1, num2, num3) +#define gr_log_int5_str2(audit, msg, num1, num2, str1, str2) gr_log_varargs(audit, msg, GR_FIVE_INT_TWO_STR, num1, num2, str1, str2) +#define gr_log_str_str(audit, msg, str1, str2) gr_log_varargs(audit, msg, GR_TWO_STR, str1, str2) +#define gr_log_str3(audit, msg, str1, str2, str3) gr_log_varargs(audit, msg, GR_THREE_STR, str1, str2, str3) +#define gr_log_str4(audit, msg, str1, str2, str3, str4) gr_log_varargs(audit, msg, GR_FOUR_STR, str1, str2, str3, str4) +#define gr_log_str_fs(audit, msg, str, dentry, mnt) gr_log_varargs(audit, msg, GR_STR_FILENAME, str, dentry, mnt) +#define gr_log_fs_str(audit, msg, dentry, mnt, str) gr_log_varargs(audit, msg, GR_FILENAME_STR, dentry, mnt, str) +#define gr_log_fs_int2(audit, msg, dentry, mnt, num1, num2) gr_log_varargs(audit, msg, GR_FILENAME_TWO_INT, dentry, mnt, num1, num2) +#define gr_log_fs_int2_str(audit, msg, dentry, mnt, num1, num2, str) gr_log_varargs(audit, msg, GR_FILENAME_TWO_INT_STR, dentry, mnt, num1, num2, str) +#define gr_log_textrel_ulong_ulong(audit, msg, file, ulong1, ulong2) gr_log_varargs(audit, msg, GR_TEXTREL, file, ulong1, ulong2) +#define gr_log_ptrace(audit, msg, task) gr_log_varargs(audit, msg, GR_PTRACE, task) +#define gr_log_res_ulong2_str(audit, msg, task, ulong1, str, ulong2) gr_log_varargs(audit, msg, GR_RESOURCE, task, ulong1, str, ulong2) +#define gr_log_cap(audit, msg, task, str) gr_log_varargs(audit, msg, GR_CAP, task, str) +#define gr_log_sig(audit, msg, task, num) gr_log_varargs(audit, msg, GR_SIG, task, num) +#define gr_log_crash1(audit, msg, task, ulong) gr_log_varargs(audit, msg, GR_CRASH1, task, ulong) +#define gr_log_crash2(audit, msg, task, ulong1) gr_log_varargs(audit, msg, GR_CRASH2, task, ulong1) +#define gr_log_procacct(audit, msg, task, num1, num2, num3, num4, num5, num6, num7, num8, num9) gr_log_varargs(audit, msg, GR_PSACCT, task, num1, num2, num3, num4, num5, num6, num7, num8, num9) + +extern void gr_log_varargs(int audit, const char *msg, int argtypes, ...); + +#endif + +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/grmsg.h linux-2.6.17.13-g-v/include/linux/grmsg.h --- linux-2.6.17.13/include/linux/grmsg.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/grmsg.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,108 @@ +#define DEFAULTSECMSG "%.256s[%.16s:%d] uid/euid:%u/%u gid/egid:%u/%u, parent %.256s[%.16s:%d] uid/euid:%u/%u gid/egid:%u/%u" +#define GR_ACL_PROCACCT_MSG "%.256s[%.16s:%d] IP:%u.%u.%u.%u TTY:%.64s uid/euid:%u/%u gid/egid:%u/%u run time:[%ud %uh %um %us] cpu time:[%ud %uh %um %us] %s with exit code %ld, parent %.256s[%.16s:%d] IP:%u.%u.%u.%u TTY:%.64s uid/euid:%u/%u gid/egid:%u/%u" +#define GR_PTRACE_ACL_MSG "denied ptrace of %.950s(%.16s:%d) by " +#define GR_STOPMOD_MSG "denied modification of module state by " +#define GR_IOPERM_MSG "denied use of ioperm() by " +#define GR_IOPL_MSG "denied use of iopl() by " +#define GR_SHMAT_ACL_MSG "denied attach of shared memory of UID %u, PID %d, ID %u by " +#define GR_UNIX_CHROOT_MSG "denied connect() to abstract AF_UNIX socket outside of chroot by " +#define GR_SHMAT_CHROOT_MSG "denied attach of shared memory outside of chroot by " +#define GR_KMEM_MSG "denied write of /dev/kmem by " +#define GR_PORT_OPEN_MSG "denied open of /dev/port by " +#define GR_MEM_WRITE_MSG "denied write of /dev/mem by " +#define GR_MEM_MMAP_MSG "denied mmap write of /dev/[k]mem by " +#define GR_SYMLINK_MSG "not following symlink %.950s owned by %d.%d by " +#define GR_LEARN_AUDIT_MSG "%s\t%u\t%u\t%u\t%.4095s\t%.4095s\t%lu\t%lu\t%.4095s\t%lu\t%u.%u.%u.%u" +#define GR_ID_LEARN_MSG "%s\t%u\t%u\t%u\t%.4095s\t%.4095s\t%c\t%d\t%d\t%d\t%u.%u.%u.%u" +#define GR_HIDDEN_ACL_MSG "%s access to hidden file %.950s by " +#define GR_OPEN_ACL_MSG "%s open of %.950s for%s%s by " +#define GR_CREATE_ACL_MSG "%s create of %.950s for%s%s by " +#define GR_FIFO_MSG "denied writing FIFO %.950s of %d.%d by " +#define GR_MKNOD_CHROOT_MSG "denied mknod of %.950s from chroot by " +#define GR_MKNOD_ACL_MSG "%s mknod of %.950s by " +#define GR_UNIXCONNECT_ACL_MSG "%s connect() to the unix domain socket %.950s by " +#define GR_TTYSNIFF_ACL_MSG "terminal being sniffed by IP:%u.%u.%u.%u %.480s[%.16s:%d], parent %.480s[%.16s:%d] against " +#define GR_MKDIR_ACL_MSG "%s mkdir of %.950s by " +#define GR_RMDIR_ACL_MSG "%s rmdir of %.950s by " +#define GR_UNLINK_ACL_MSG "%s unlink of %.950s by " +#define GR_SYMLINK_ACL_MSG "%s symlink from %.480s to %.480s by " +#define GR_HARDLINK_MSG "denied hardlink of %.930s (owned by %d.%d) to %.30s for " +#define GR_LINK_ACL_MSG "%s link of %.480s to %.480s by " +#define GR_INHERIT_ACL_MSG "successful inherit of %.480s's ACL for %.480s by " +#define GR_RENAME_ACL_MSG "%s rename of %.480s to %.480s by " +#define GR_PTRACE_EXEC_ACL_MSG "denied ptrace of %.950s by " +#define GR_NPROC_MSG "denied overstep of process limit by " +#define GR_EXEC_ACL_MSG "%s execution of %.950s by " +#define GR_EXEC_TPE_MSG "denied untrusted exec of %.950s by " +#define GR_SEGVSTART_ACL_MSG "possible exploit bruteforcing on " DEFAULTSECMSG " banning uid %u from login for %lu seconds" +#define GR_SEGVNOSUID_ACL_MSG "possible exploit bruteforcing on " DEFAULTSECMSG " banning execution for %lu seconds" +#define GR_MOUNT_CHROOT_MSG "denied mount of %.30s as %.930s from chroot by " +#define GR_PIVOT_CHROOT_MSG "denied pivot_root from chroot by " +#define GR_TRUNCATE_ACL_MSG "%s truncate of %.950s by " +#define GR_ATIME_ACL_MSG "%s access time change of %.950s by " +#define GR_ACCESS_ACL_MSG "%s access of %.950s for%s%s%s by " +#define GR_CHROOT_CHROOT_MSG "denied double chroot to %.950s by " +#define GR_FCHMOD_ACL_MSG "%s fchmod of %.950s by " +#define GR_CHMOD_CHROOT_MSG "denied chmod +s of %.950s by " +#define GR_CHMOD_ACL_MSG "%s chmod of %.950s by " +#define GR_CHROOT_FCHDIR_MSG "denied fchdir outside of chroot to %.950s by " +#define GR_CHOWN_ACL_MSG "%s chown of %.950s by " +#define GR_WRITLIB_ACL_MSG "denied load of writable library %.950s by " +#define GR_INITF_ACL_MSG "init_variables() failed %s by " +#define GR_DISABLED_ACL_MSG "Error loading %s, trying to run kernel with acls disabled. To disable acls at startup use gracl=off from your boot loader" +#define GR_DEV_ACL_MSG "/dev/grsec: %d bytes sent %d required, being fed garbaged by " +#define GR_SHUTS_ACL_MSG "shutdown auth success for " +#define GR_SHUTF_ACL_MSG "shutdown auth failure for " +#define GR_SHUTI_ACL_MSG "ignoring shutdown for disabled RBAC system for " +#define GR_SEGVMODS_ACL_MSG "segvmod auth success for " +#define GR_SEGVMODF_ACL_MSG "segvmod auth failure for " +#define GR_SEGVMODI_ACL_MSG "ignoring segvmod for disabled RBAC system for " +#define GR_ENABLE_ACL_MSG "%s RBAC system loaded by " +#define GR_ENABLEF_ACL_MSG "unable to load %s for " +#define GR_RELOADI_ACL_MSG "ignoring reload request for disabled RBAC system" +#define GR_RELOAD_ACL_MSG "%s RBAC system reloaded by " +#define GR_RELOADF_ACL_MSG "failed reload of %s for " +#define GR_SPROLEI_ACL_MSG "ignoring change to special role for disabled RBAC system for " +#define GR_SPROLES_ACL_MSG "successful change to special role %s (id %d) by " +#define GR_SPROLEL_ACL_MSG "special role %s (id %d) exited by " +#define GR_SPROLEF_ACL_MSG "special role %s failure for " +#define GR_UNSPROLEI_ACL_MSG "ignoring unauth of special role for disabled RBAC system for " +#define GR_UNSPROLES_ACL_MSG "successful unauth of special role %s (id %d) by " +#define GR_UNSPROLEF_ACL_MSG "special role unauth of %s failure for " +#define GR_INVMODE_ACL_MSG "invalid mode %d by " +#define GR_PRIORITY_CHROOT_MSG "denied priority change of process (%.16s:%d) by " +#define GR_FAILFORK_MSG "failed fork with errno %d by " +#define GR_NICE_CHROOT_MSG "denied priority change by " +#define GR_UNISIGLOG_MSG "signal %d sent to " +#define GR_DUALSIGLOG_MSG "signal %d sent to " DEFAULTSECMSG " by " +#define GR_SIG_ACL_MSG "denied send of signal %d to protected task " DEFAULTSECMSG " by " +#define GR_SYSCTL_MSG "denied modification of grsecurity sysctl value : %.32s by " +#define GR_SYSCTL_ACL_MSG "%s sysctl of %.950s for%s%s by " +#define GR_TIME_MSG "time set by " +#define GR_DEFACL_MSG "fatal: unable to find subject for (%.16s:%d), loaded by " +#define GR_MMAP_ACL_MSG "%s executable mmap of %.950s by " +#define GR_MPROTECT_ACL_MSG "%s executable mprotect of %.950s by " +#define GR_SOCK_MSG "denied socket(%.16s,%.16s,%.16s) by " +#define GR_SOCK2_MSG "denied socket(%d,%.16s,%.16s) by " +#define GR_BIND_MSG "denied bind() by " +#define GR_CONNECT_MSG "denied connect() by " +#define GR_BIND_ACL_MSG "denied bind() to %u.%u.%u.%u port %u sock type %.16s protocol %.16s by " +#define GR_CONNECT_ACL_MSG "denied connect() to %u.%u.%u.%u port %u sock type %.16s protocol %.16s by " +#define GR_IP_LEARN_MSG "%s\t%u\t%u\t%u\t%.4095s\t%.4095s\t%u.%u.%u.%u\t%u\t%u\t%u\t%u\t%u.%u.%u.%u" +#define GR_EXEC_CHROOT_MSG "exec of %.980s within chroot by process " +#define GR_CAP_ACL_MSG "use of %s denied for " +#define GR_USRCHANGE_ACL_MSG "change to uid %u denied for " +#define GR_GRPCHANGE_ACL_MSG "change to gid %u denied for " +#define GR_REMOUNT_AUDIT_MSG "remount of %.30s by " +#define GR_UNMOUNT_AUDIT_MSG "unmount of %.30s by " +#define GR_MOUNT_AUDIT_MSG "mount of %.30s to %.64s by " +#define GR_CHDIR_AUDIT_MSG "chdir to %.980s by " +#define GR_EXEC_AUDIT_MSG "exec of %.930s (%.128s) by " +#define GR_MSGQ_AUDIT_MSG "message queue created by " +#define GR_MSGQR_AUDIT_MSG "message queue of uid:%u euid:%u removed by " +#define GR_SEM_AUDIT_MSG "semaphore created by " +#define GR_SEMR_AUDIT_MSG "semaphore of uid:%u euid:%u removed by " +#define GR_SHM_AUDIT_MSG "shared memory of size %d created by " +#define GR_SHMR_AUDIT_MSG "shared memory of uid:%u euid:%u removed by " +#define GR_RESOURCE_MSG "denied resource overstep by requesting %lu for %.16s against limit %lu for " +#define GR_TEXTREL_AUDIT_MSG "text relocation in %s, VMA:0x%08lx 0x%08lx by " diff -NurpP --minimal linux-2.6.17.13/include/linux/grsecurity.h linux-2.6.17.13-g-v/include/linux/grsecurity.h --- linux-2.6.17.13/include/linux/grsecurity.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/grsecurity.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,196 @@ +#ifndef GR_SECURITY_H +#define GR_SECURITY_H +#include +#include +#include + +extern void gr_handle_brute_attach(struct task_struct *p); +extern void gr_handle_brute_check(void); + +extern char gr_roletype_to_char(void); + +extern int gr_check_user_change(int real, int effective, int fs); +extern int gr_check_group_change(int real, int effective, int fs); + +extern void gr_del_task_from_ip_table(struct task_struct *p); + +extern int gr_pid_is_chrooted(struct task_struct *p); +extern int gr_handle_chroot_nice(void); +extern int gr_handle_chroot_sysctl(const int op); +extern int gr_handle_chroot_setpriority(struct task_struct *p, + const int niceval); +extern int gr_chroot_fchdir(struct dentry *u_dentry, struct vfsmount *u_mnt); +extern int gr_handle_chroot_chroot(const struct dentry *dentry, + const struct vfsmount *mnt); +extern void gr_handle_chroot_caps(struct task_struct *task); +extern void gr_handle_chroot_chdir(struct dentry *dentry, struct vfsmount *mnt); +extern int gr_handle_chroot_chmod(const struct dentry *dentry, + const struct vfsmount *mnt, const int mode); +extern int gr_handle_chroot_mknod(const struct dentry *dentry, + const struct vfsmount *mnt, const int mode); +extern int gr_handle_chroot_mount(const struct dentry *dentry, + const struct vfsmount *mnt, + const char *dev_name); +extern int gr_handle_chroot_pivot(void); +extern int gr_handle_chroot_unix(const pid_t pid); + +extern int gr_handle_rawio(const struct inode *inode); +extern int gr_handle_nproc(void); + +extern void gr_handle_ioperm(void); +extern void gr_handle_iopl(void); + +extern int gr_tpe_allow(const struct file *file); + +extern int gr_random_pid(void); + +extern void gr_log_forkfail(const int retval); +extern void gr_log_timechange(void); +extern void gr_log_signal(const int sig, const struct task_struct *t); +extern void gr_log_chdir(const struct dentry *dentry, + const struct vfsmount *mnt); +extern void gr_log_chroot_exec(const struct dentry *dentry, + const struct vfsmount *mnt); +extern void gr_handle_exec_args(struct linux_binprm *bprm, char **argv); +extern void gr_log_remount(const char *devname, const int retval); +extern void gr_log_unmount(const char *devname, const int retval); +extern void gr_log_mount(const char *from, const char *to, const int retval); +extern void gr_log_msgget(const int ret, const int msgflg); +extern void gr_log_msgrm(const uid_t uid, const uid_t cuid); +extern void gr_log_semget(const int err, const int semflg); +extern void gr_log_semrm(const uid_t uid, const uid_t cuid); +extern void gr_log_shmget(const int err, const int shmflg, const size_t size); +extern void gr_log_shmrm(const uid_t uid, const uid_t cuid); +extern void gr_log_textrel(struct vm_area_struct *vma); + +extern int gr_handle_follow_link(const struct inode *parent, + const struct inode *inode, + const struct dentry *dentry, + const struct vfsmount *mnt); +extern int gr_handle_fifo(const struct dentry *dentry, + const struct vfsmount *mnt, + const struct dentry *dir, const int flag, + const int acc_mode); +extern int gr_handle_hardlink(const struct dentry *dentry, + const struct vfsmount *mnt, + struct inode *inode, + const int mode, const char *to); + +extern int gr_task_is_capable(struct task_struct *task, const int cap); +extern int gr_is_capable_nolog(const int cap); +extern void gr_learn_resource(const struct task_struct *task, const int limit, + const unsigned long wanted, const int gt); +extern void gr_copy_label(struct task_struct *tsk); +extern void gr_handle_crash(struct task_struct *task, const int sig); +extern int gr_handle_signal(const struct task_struct *p, const int sig); +extern int gr_check_crash_uid(const uid_t uid); +extern int gr_check_protected_task(const struct task_struct *task); +extern int gr_acl_handle_mmap(const struct file *file, + const unsigned long prot); +extern int gr_acl_handle_mprotect(const struct file *file, + const unsigned long prot); +extern int gr_check_hidden_task(const struct task_struct *tsk); +extern __u32 gr_acl_handle_truncate(const struct dentry *dentry, + const struct vfsmount *mnt); +extern __u32 gr_acl_handle_utime(const struct dentry *dentry, + const struct vfsmount *mnt); +extern __u32 gr_acl_handle_access(const struct dentry *dentry, + const struct vfsmount *mnt, const int fmode); +extern __u32 gr_acl_handle_fchmod(const struct dentry *dentry, + const struct vfsmount *mnt, mode_t mode); +extern __u32 gr_acl_handle_chmod(const struct dentry *dentry, + const struct vfsmount *mnt, mode_t mode); +extern __u32 gr_acl_handle_chown(const struct dentry *dentry, + const struct vfsmount *mnt); +extern int gr_handle_ptrace(struct task_struct *task, const long request); +extern int gr_handle_proc_ptrace(struct task_struct *task); +extern __u32 gr_acl_handle_execve(const struct dentry *dentry, + const struct vfsmount *mnt); +extern int gr_check_crash_exec(const struct file *filp); +extern int gr_acl_is_enabled(void); +extern void gr_set_kernel_label(struct task_struct *task); +extern void gr_set_role_label(struct task_struct *task, const uid_t uid, + const gid_t gid); +extern int gr_set_proc_label(const struct dentry *dentry, + const struct vfsmount *mnt); +extern __u32 gr_acl_handle_hidden_file(const struct dentry *dentry, + const struct vfsmount *mnt); +extern __u32 gr_acl_handle_open(const struct dentry *dentry, + const struct vfsmount *mnt, const int fmode); +extern __u32 gr_acl_handle_creat(const struct dentry *dentry, + const struct dentry *p_dentry, + const struct vfsmount *p_mnt, const int fmode, + const int imode); +extern void gr_handle_create(const struct dentry *dentry, + const struct vfsmount *mnt); +extern __u32 gr_acl_handle_mknod(const struct dentry *new_dentry, + const struct dentry *parent_dentry, + const struct vfsmount *parent_mnt, + const int mode); +extern __u32 gr_acl_handle_mkdir(const struct dentry *new_dentry, + const struct dentry *parent_dentry, + const struct vfsmount *parent_mnt); +extern __u32 gr_acl_handle_rmdir(const struct dentry *dentry, + const struct vfsmount *mnt); +extern void gr_handle_delete(const ino_t ino, const dev_t dev); +extern __u32 gr_acl_handle_unlink(const struct dentry *dentry, + const struct vfsmount *mnt); +extern __u32 gr_acl_handle_symlink(const struct dentry *new_dentry, + const struct dentry *parent_dentry, + const struct vfsmount *parent_mnt, + const char *from); +extern __u32 gr_acl_handle_link(const struct dentry *new_dentry, + const struct dentry *parent_dentry, + const struct vfsmount *parent_mnt, + const struct dentry *old_dentry, + const struct vfsmount *old_mnt, const char *to); +extern int gr_acl_handle_rename(struct dentry *new_dentry, + struct dentry *parent_dentry, + const struct vfsmount *parent_mnt, + struct dentry *old_dentry, + struct inode *old_parent_inode, + struct vfsmount *old_mnt, const char *newname); +extern void gr_handle_rename(struct inode *old_dir, struct inode *new_dir, + struct dentry *old_dentry, + struct dentry *new_dentry, + struct vfsmount *mnt, const __u8 replace); +extern __u32 gr_check_link(const struct dentry *new_dentry, + const struct dentry *parent_dentry, + const struct vfsmount *parent_mnt, + const struct dentry *old_dentry, + const struct vfsmount *old_mnt); +extern int gr_acl_handle_filldir(const struct file *file, const char *name, + const unsigned int namelen, const ino_t ino); + +extern __u32 gr_acl_handle_unix(const struct dentry *dentry, + const struct vfsmount *mnt); +extern void gr_acl_handle_exit(void); +extern void gr_acl_handle_psacct(struct task_struct *task, const long code); +extern int gr_acl_handle_procpidmem(const struct task_struct *task); +extern __u32 gr_cap_rtnetlink(void); + +#ifdef CONFIG_SYSVIPC +extern void gr_shm_exit(struct task_struct *task); +#else +static inline void gr_shm_exit(struct task_struct *task) +{ + return; +} +#endif + +#ifdef CONFIG_GRKERNSEC +extern void gr_handle_mem_write(void); +extern void gr_handle_kmem_write(void); +extern void gr_handle_open_port(void); +extern int gr_handle_mem_mmap(const unsigned long offset, + struct vm_area_struct *vma); + +extern unsigned long pax_get_random_long(void); +#define get_random_long() pax_get_random_long() + +extern int grsec_enable_dmesg; +extern int grsec_enable_randsrc; +extern int grsec_enable_shm; +#endif + +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/highmem.h linux-2.6.17.13-g-v/include/linux/highmem.h --- linux-2.6.17.13/include/linux/highmem.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/highmem.h 2006-09-20 15:00:12.000000000 +0200 @@ -70,9 +70,9 @@ alloc_zeroed_user_highpage(struct vm_are static inline void clear_highpage(struct page *page) { - void *kaddr = kmap_atomic(page, KM_USER0); + void *kaddr = kmap_atomic(page, KM_CLEARPAGE); clear_page(kaddr); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr, KM_CLEARPAGE); } /* diff -NurpP --minimal linux-2.6.17.13/include/linux/init_task.h linux-2.6.17.13-g-v/include/linux/init_task.h --- linux-2.6.17.13/include/linux/init_task.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/init_task.h 2006-09-20 15:00:12.000000000 +0200 @@ -123,6 +123,10 @@ extern struct group_info init_groups; .journal_info = NULL, \ .cpu_timers = INIT_CPU_TIMERS(tsk.cpu_timers), \ .fs_excl = ATOMIC_INIT(0), \ + .xid = 0, \ + .vx_info = NULL, \ + .nid = 0, \ + .nx_info = NULL, \ } diff -NurpP --minimal linux-2.6.17.13/include/linux/ipc.h linux-2.6.17.13-g-v/include/linux/ipc.h --- linux-2.6.17.13/include/linux/ipc.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/ipc.h 2006-09-20 15:00:12.000000000 +0200 @@ -66,6 +66,7 @@ struct kern_ipc_perm mode_t mode; unsigned long seq; void *security; + xid_t xid; }; #endif /* __KERNEL__ */ diff -NurpP --minimal linux-2.6.17.13/include/linux/kernel.h linux-2.6.17.13-g-v/include/linux/kernel.h --- linux-2.6.17.13/include/linux/kernel.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/kernel.h 2006-09-20 15:00:12.000000000 +0200 @@ -17,6 +17,7 @@ #include extern const char linux_banner[]; +extern const char vx_linux_banner[]; #define INT_MAX ((int)(~0U>>1)) #define INT_MIN (-INT_MAX - 1) diff -NurpP --minimal linux-2.6.17.13/include/linux/major.h linux-2.6.17.13-g-v/include/linux/major.h --- linux-2.6.17.13/include/linux/major.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/major.h 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,7 @@ #define HD_MAJOR IDE0_MAJOR #define PTY_SLAVE_MAJOR 3 #define TTY_MAJOR 4 +#define VROOT_MAJOR 4 #define TTYAUX_MAJOR 5 #define LP_MAJOR 6 #define VCS_MAJOR 7 diff -NurpP --minimal linux-2.6.17.13/include/linux/mm.h linux-2.6.17.13-g-v/include/linux/mm.h --- linux-2.6.17.13/include/linux/mm.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/mm.h 2006-09-20 15:00:12.000000000 +0200 @@ -38,6 +38,7 @@ extern int sysctl_legacy_va_layout; #include #include #include +#include #define nth_page(page,n) pfn_to_page(page_to_pfn((page)) + (n)) @@ -111,8 +112,43 @@ struct vm_area_struct { #ifdef CONFIG_NUMA struct mempolicy *vm_policy; /* NUMA policy for the VMA */ #endif + + unsigned long vm_mirror; /* PaX: mirror distance */ }; +#ifdef CONFIG_PAX_SOFTMODE +extern unsigned int pax_softmode; +#endif + +extern int pax_check_flags(unsigned long *); + +/* if tsk != current then task_lock must be held on it */ +#if defined(CONFIG_PAX_NOEXEC) || defined(CONFIG_PAX_ASLR) +static inline unsigned long pax_get_flags(struct task_struct *tsk) +{ + if (likely(tsk->mm)) + return tsk->mm->pax_flags; + else + return 0UL; +} + +/* if tsk != current then task_lock must be held on it */ +static inline long pax_set_flags(struct task_struct *tsk, unsigned long flags) +{ + if (likely(tsk->mm)) { + tsk->mm->pax_flags = flags; + return 0; + } + return -EINVAL; +} +#endif + +#ifdef CONFIG_PAX_HAVE_ACL_FLAGS +extern void pax_set_initial_flags(struct linux_binprm * bprm); +#elif defined(CONFIG_PAX_HOOK_ACL_FLAGS) +extern void (*pax_set_initial_flags_func)(struct linux_binprm * bprm); +#endif + /* * This struct defines the per-mm list of VMAs for uClinux. If CONFIG_MMU is * disabled, then there's a single shared list of VMAs maintained by the @@ -167,6 +203,18 @@ extern unsigned int kobjsize(const void #define VM_MAPPED_COPY 0x01000000 /* T if mapped copy of data (nommu mmap) */ #define VM_INSERTPAGE 0x02000000 /* The vma has had "vm_insert_page()" done on it */ +#ifdef CONFIG_PAX_SEGMEXEC +#define VM_MIRROR 0x04000000 /* vma is mirroring another */ +#endif + +#ifdef CONFIG_PAX_MPROTECT +#define VM_MAYNOTWRITE 0x08000000 /* vma cannot be granted VM_WRITE any more */ +#endif + +#ifdef __VM_STACK_FLAGS +#define VM_STACK_DEFAULT_FLAGS (0x00000033 | __VM_STACK_FLAGS) +#endif + #ifndef VM_STACK_DEFAULT_FLAGS /* arch can override this */ #define VM_STACK_DEFAULT_FLAGS VM_DATA_DEFAULT_FLAGS #endif @@ -1058,5 +1106,11 @@ void drop_slab(void); extern int randomize_va_space; #endif +#ifdef CONFIG_ARCH_TRACK_EXEC_LIMIT +extern void track_exec_limit(struct mm_struct *mm, unsigned long start, unsigned long end, unsigned long prot); +#else +static inline void track_exec_limit(struct mm_struct *mm, unsigned long start, unsigned long end, unsigned long prot) {} +#endif + #endif /* __KERNEL__ */ #endif /* _LINUX_MM_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/mman.h linux-2.6.17.13-g-v/include/linux/mman.h --- linux-2.6.17.13/include/linux/mman.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/mman.h 2006-09-20 15:00:12.000000000 +0200 @@ -59,6 +59,11 @@ static inline unsigned long calc_vm_flag_bits(unsigned long flags) { return _calc_vm_trans(flags, MAP_GROWSDOWN, VM_GROWSDOWN ) | + +#ifdef CONFIG_PAX_SEGMEXEC + _calc_vm_trans(flags, MAP_MIRROR, VM_MIRROR) | +#endif + _calc_vm_trans(flags, MAP_DENYWRITE, VM_DENYWRITE ) | _calc_vm_trans(flags, MAP_EXECUTABLE, VM_EXECUTABLE) | _calc_vm_trans(flags, MAP_LOCKED, VM_LOCKED ); diff -NurpP --minimal linux-2.6.17.13/include/linux/module.h linux-2.6.17.13-g-v/include/linux/module.h --- linux-2.6.17.13/include/linux/module.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/module.h 2006-09-20 15:00:12.000000000 +0200 @@ -273,16 +273,16 @@ struct module int (*init)(void); /* If this is non-NULL, vfree after init() returns */ - void *module_init; + void *module_init_rx, *module_init_rw; /* Here is the actual code + data, vfree'd on unload. */ - void *module_core; + void *module_core_rx, *module_core_rw; /* Here are the sizes of the init and core sections */ - unsigned long init_size, core_size; + unsigned long init_size_rw, core_size_rw; /* The size of the executable code in each section. */ - unsigned long init_text_size, core_text_size; + unsigned long init_size_rx, core_size_rx; /* Arch-specific module values */ struct mod_arch_specific arch; diff -NurpP --minimal linux-2.6.17.13/include/linux/moduleloader.h linux-2.6.17.13-g-v/include/linux/moduleloader.h --- linux-2.6.17.13/include/linux/moduleloader.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/moduleloader.h 2006-09-20 15:00:12.000000000 +0200 @@ -17,9 +17,21 @@ int module_frob_arch_sections(Elf_Ehdr * sections. Returns NULL on failure. */ void *module_alloc(unsigned long size); +#ifdef CONFIG_PAX_KERNEXEC +void *module_alloc_exec(unsigned long size); +#else +#define module_alloc_exec(x) module_alloc(x) +#endif + /* Free memory returned from module_alloc. */ void module_free(struct module *mod, void *module_region); +#ifdef CONFIG_PAX_KERNEXEC +void module_free_exec(struct module *mod, void *module_region); +#else +#define module_free_exec(x, y) module_free(x, y) +#endif + /* Apply the given relocation to the (simplified) ELF. Return -error or 0. */ int apply_relocate(Elf_Shdr *sechdrs, diff -NurpP --minimal linux-2.6.17.13/include/linux/mount.h linux-2.6.17.13-g-v/include/linux/mount.h --- linux-2.6.17.13/include/linux/mount.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/mount.h 2006-09-20 15:00:12.000000000 +0200 @@ -22,10 +22,14 @@ #define MNT_NOEXEC 0x04 #define MNT_NOATIME 0x08 #define MNT_NODIRATIME 0x10 +#define MNT_RDONLY 0x20 + +#define MNT_IS_RDONLY(m) ((m) && ((m)->mnt_flags & MNT_RDONLY)) #define MNT_SHARED 0x1000 /* if the vfsmount is a shared mount */ #define MNT_UNBINDABLE 0x2000 /* if the vfsmount is a unbindable mount */ #define MNT_PNODE_MASK 0x3000 /* propogation flag mask */ +#define MNT_XID 0x8000 struct vfsmount { struct list_head mnt_hash; @@ -47,6 +51,7 @@ struct vfsmount { struct vfsmount *mnt_master; /* slave is on master->mnt_slave_list */ struct namespace *mnt_namespace; /* containing namespace */ int mnt_pinned; + xid_t mnt_xid; /* xid tagging used for vfsmount */ }; static inline struct vfsmount *mntget(struct vfsmount *mnt) diff -NurpP --minimal linux-2.6.17.13/include/linux/net.h linux-2.6.17.13-g-v/include/linux/net.h --- linux-2.6.17.13/include/linux/net.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/net.h 2006-09-20 15:00:12.000000000 +0200 @@ -62,6 +62,7 @@ typedef enum { #define SOCK_ASYNC_WAITDATA 1 #define SOCK_NOSPACE 2 #define SOCK_PASSCRED 3 +#define SOCK_USER_SOCKET 4 #ifndef ARCH_HAS_SOCKET_TYPES /** diff -NurpP --minimal linux-2.6.17.13/include/linux/nfs_mount.h linux-2.6.17.13-g-v/include/linux/nfs_mount.h --- linux-2.6.17.13/include/linux/nfs_mount.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/nfs_mount.h 2006-09-20 15:00:12.000000000 +0200 @@ -61,6 +61,7 @@ struct nfs_mount_data { #define NFS_MOUNT_NOACL 0x0800 /* 4 */ #define NFS_MOUNT_STRICTLOCK 0x1000 /* reserved for NFSv4 */ #define NFS_MOUNT_SECFLAVOUR 0x2000 /* 5 */ +#define NFS_MOUNT_TAGXID 0x8000 /* tagxid */ #define NFS_MOUNT_FLAGMASK 0xFFFF #endif diff -NurpP --minimal linux-2.6.17.13/include/linux/percpu.h linux-2.6.17.13-g-v/include/linux/percpu.h --- linux-2.6.17.13/include/linux/percpu.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/percpu.h 2006-09-20 15:00:12.000000000 +0200 @@ -8,7 +8,7 @@ /* Enough to cover all DEFINE_PER_CPUs in kernel, including modules. */ #ifndef PERCPU_ENOUGH_ROOM -#define PERCPU_ENOUGH_ROOM 32768 +#define PERCPU_ENOUGH_ROOM 65536 #endif /* Must be an lvalue. */ diff -NurpP --minimal linux-2.6.17.13/include/linux/proc_fs.h linux-2.6.17.13-g-v/include/linux/proc_fs.h --- linux-2.6.17.13/include/linux/proc_fs.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/proc_fs.h 2006-09-20 15:00:12.000000000 +0200 @@ -56,6 +56,7 @@ struct proc_dir_entry { nlink_t nlink; uid_t uid; gid_t gid; + int vx_flags; loff_t size; struct inode_operations * proc_iops; const struct file_operations * proc_fops; @@ -251,9 +252,11 @@ extern void kclist_add(struct kcore_list struct proc_inode { struct task_struct *task; int type; + int vx_flags; union { int (*proc_get_link)(struct inode *, struct dentry **, struct vfsmount **); int (*proc_read)(struct task_struct *task, char *page); + int (*proc_vid_read)(int vid, char *page); } op; struct proc_dir_entry *pde; struct inode vfs_inode; diff -NurpP --minimal linux-2.6.17.13/include/linux/random.h linux-2.6.17.13-g-v/include/linux/random.h --- linux-2.6.17.13/include/linux/random.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/random.h 2006-09-20 15:00:12.000000000 +0200 @@ -62,6 +62,8 @@ extern __u32 secure_tcpv6_sequence_numbe extern u64 secure_dccp_sequence_number(__u32 saddr, __u32 daddr, __u16 sport, __u16 dport); +extern unsigned long pax_get_random_long(void); + #ifndef MODULE extern struct file_operations random_fops, urandom_fops; #endif diff -NurpP --minimal linux-2.6.17.13/include/linux/reiserfs_fs.h linux-2.6.17.13-g-v/include/linux/reiserfs_fs.h --- linux-2.6.17.13/include/linux/reiserfs_fs.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/reiserfs_fs.h 2006-09-20 15:00:12.000000000 +0200 @@ -829,6 +829,18 @@ struct stat_data_v1 { #define REISERFS_COMPR_FL EXT2_COMPR_FL #define REISERFS_NOTAIL_FL EXT2_NOTAIL_FL +/* unfortunately reiserfs sdattr is only 16 bit */ +#define REISERFS_BARRIER_FL (EXT2_BARRIER_FL >> 16) +#define REISERFS_IUNLINK_FL (EXT2_IUNLINK_FL >> 16) + +#ifdef CONFIG_VSERVER_LEGACY +#define REISERFS_FL_USER_VISIBLE (REISERFS_IUNLINK_FL|0x80FF) +#define REISERFS_FL_USER_MODIFIABLE (REISERFS_IUNLINK_FL|0x80FF) +#else +#define REISERFS_FL_USER_VISIBLE 0x80FF +#define REISERFS_FL_USER_MODIFIABLE 0x80FF +#endif + /* persistent flags that file inherits from the parent directory */ #define REISERFS_INHERIT_MASK ( REISERFS_IMMUTABLE_FL | \ REISERFS_SYNC_FL | \ @@ -1909,6 +1921,7 @@ static inline void reiserfs_update_sd(st void sd_attrs_to_i_attrs(__u16 sd_attrs, struct inode *inode); void i_attrs_to_sd_attrs(struct inode *inode, __u16 * sd_attrs); int reiserfs_setattr(struct dentry *dentry, struct iattr *attr); +int reiserfs_sync_flags(struct inode *inode); /* namei.c */ void set_de_name_and_namelen(struct reiserfs_dir_entry *de); diff -NurpP --minimal linux-2.6.17.13/include/linux/reiserfs_fs_sb.h linux-2.6.17.13-g-v/include/linux/reiserfs_fs_sb.h --- linux-2.6.17.13/include/linux/reiserfs_fs_sb.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/reiserfs_fs_sb.h 2006-09-20 15:00:12.000000000 +0200 @@ -456,6 +456,7 @@ enum reiserfs_mount_options { REISERFS_POSIXACL, REISERFS_BARRIER_NONE, REISERFS_BARRIER_FLUSH, + REISERFS_TAGXID, /* Actions on error */ REISERFS_ERROR_PANIC, diff -NurpP --minimal linux-2.6.17.13/include/linux/sched.h linux-2.6.17.13-g-v/include/linux/sched.h --- linux-2.6.17.13/include/linux/sched.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/sched.h 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,7 @@ #include #include #include +#include #include #include @@ -40,6 +41,7 @@ #include /* For AT_VECTOR_SIZE */ struct exec_domain; +struct linux_binprm; /* * cloning flags: @@ -131,6 +133,7 @@ extern unsigned long nr_iowait(void); #define EXIT_DEAD 32 /* in tsk->state again */ #define TASK_NONINTERACTIVE 64 +#define TASK_ONHOLD 128 #define __set_task_state(tsk, state_value) \ do { (tsk)->state = (state_value); } while (0) @@ -259,27 +262,30 @@ extern void arch_unmap_area_topdown(stru * The mm counters are not protected by its page_table_lock, * so must be incremented atomically. */ -#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value) -#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member)) -#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member) -#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member) -#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member) typedef atomic_long_t mm_counter_t; +#define __set_mm_counter(mm, member, value) \ + atomic_long_set(&(mm)->_##member, value) +#define get_mm_counter(mm, member) \ + ((unsigned long)atomic_long_read(&(mm)->_##member)) #else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ /* * The mm counters are protected by its page_table_lock, * so can be incremented directly. */ -#define set_mm_counter(mm, member, value) (mm)->_##member = (value) -#define get_mm_counter(mm, member) ((mm)->_##member) -#define add_mm_counter(mm, member, value) (mm)->_##member += (value) -#define inc_mm_counter(mm, member) (mm)->_##member++ -#define dec_mm_counter(mm, member) (mm)->_##member-- typedef unsigned long mm_counter_t; +#define __set_mm_counter(mm, member, value) (mm)->_##member = (value) +#define get_mm_counter(mm, member) ((mm)->_##member) #endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ +#define set_mm_counter(mm, member, value) \ + vx_ ## member ## pages_sub((mm), (get_mm_counter(mm, member) - value)) +#define add_mm_counter(mm, member, value) \ + vx_ ## member ## pages_add((mm), (value)) +#define inc_mm_counter(mm, member) vx_ ## member ## pages_inc((mm)) +#define dec_mm_counter(mm, member) vx_ ## member ## pages_dec((mm)) + #define get_mm_rss(mm) \ (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss)) #define update_hiwater_rss(mm) do { \ @@ -338,6 +344,7 @@ struct mm_struct { /* Architecture-specific MM context */ mm_context_t context; + struct vx_info *mm_vx_info; /* Token based thrashing protection. */ unsigned long swap_token_time; @@ -350,8 +357,34 @@ struct mm_struct { /* aio bits */ rwlock_t ioctx_list_lock; struct kioctx *ioctx_list; + +#if defined(CONFIG_PAX_NOEXEC) || defined(CONFIG_PAX_ASLR) + unsigned long pax_flags; +#endif + +#ifdef CONFIG_PAX_DLRESOLVE + unsigned long call_dl_resolve; +#endif + +#if defined(CONFIG_PPC32) && defined(CONFIG_PAX_EMUSIGRT) + unsigned long call_syscall; +#endif + +#ifdef CONFIG_PAX_ASLR + unsigned long delta_mmap; /* randomized offset */ + unsigned long delta_exec; /* randomized offset */ + unsigned long delta_stack; /* randomized offset */ +#endif + }; +#define MF_PAX_PAGEEXEC 0x01000000 /* Paging based non-executable pages */ +#define MF_PAX_EMUTRAMP 0x02000000 /* Emulate trampolines */ +#define MF_PAX_MPROTECT 0x04000000 /* Restrict mprotect() */ +#define MF_PAX_RANDMMAP 0x08000000 /* Randomize mmap() base */ +/*#define MF_PAX_RANDEXEC 0x10000000*/ /* Randomize ET_EXEC base */ +#define MF_PAX_SEGMEXEC 0x20000000 /* Segmentation based non-executable pages */ + struct sighand_struct { atomic_t count; struct k_sigaction action[_NSIG]; @@ -449,6 +482,15 @@ struct signal_struct { struct key *session_keyring; /* keyring inherited over fork */ struct key *process_keyring; /* keyring private to this process */ #endif + +#ifdef CONFIG_GRKERNSEC + u32 curr_ip; + u32 gr_saddr; + u32 gr_daddr; + u16 gr_sport; + u16 gr_dport; + u8 used_accept:1; +#endif }; /* Context switch must be unlocked if interrupts are to be enabled */ @@ -510,9 +552,10 @@ struct user_struct { /* Hash table maintenance information */ struct list_head uidhash_list; uid_t uid; + xid_t xid; }; -extern struct user_struct *find_user(uid_t); +extern struct user_struct *find_user(xid_t, uid_t); extern struct user_struct root_user; #define INIT_USER (&root_user) @@ -824,6 +867,14 @@ struct task_struct { void *security; struct audit_context *audit_context; + +/* vserver context data */ + struct vx_info *vx_info; + struct nx_info *nx_info; + + xid_t xid; + nid_t nid; + seccomp_t seccomp; /* Thread group tracking */ @@ -881,6 +932,17 @@ struct task_struct { struct compat_robust_list_head __user *compat_robust_list; #endif +#ifdef CONFIG_GRKERNSEC + /* grsecurity */ + struct acl_subject_label *acl; + struct acl_role_label *role; + struct file *exec_file; + u16 acl_role_id; + u8 acl_sp_role:1; + u8 is_writable:1; + u8 brute:1; +#endif + atomic_t fs_excl; /* holding fs exclusive resources */ struct rcu_head rcu; @@ -1040,13 +1102,19 @@ extern struct task_struct init_task; extern struct mm_struct init_mm; -#define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr) + +#define find_task_by_real_pid(nr) \ + find_task_by_pid_type(PIDTYPE_PID, nr) +#define find_task_by_pid(nr) \ + find_task_by_pid_type(PIDTYPE_PID, \ + vx_rmap_pid(nr)) + extern struct task_struct *find_task_by_pid_type(int type, int pid); extern void set_special_pids(pid_t session, pid_t pgrp); extern void __set_special_pids(pid_t session, pid_t pgrp); /* per-UID process charging. */ -extern struct user_struct * alloc_uid(uid_t); +extern struct user_struct * alloc_uid(xid_t, uid_t); static inline struct user_struct *get_uid(struct user_struct *u) { atomic_inc(&u->__count); @@ -1393,6 +1461,12 @@ extern void arch_pick_mmap_layout(struct static inline void arch_pick_mmap_layout(struct mm_struct *mm) { mm->mmap_base = TASK_UNMAPPED_BASE; + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base += mm->delta_mmap; +#endif + mm->get_unmapped_area = arch_get_unmapped_area; mm->unmap_area = arch_unmap_area; } diff -NurpP --minimal linux-2.6.17.13/include/linux/shm.h linux-2.6.17.13-g-v/include/linux/shm.h --- linux-2.6.17.13/include/linux/shm.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/shm.h 2006-09-20 15:00:12.000000000 +0200 @@ -86,6 +86,10 @@ struct shmid_kernel /* private to the ke pid_t shm_cprid; pid_t shm_lprid; struct user_struct *mlock_user; +#ifdef CONFIG_GRKERNSEC + time_t shm_createtime; + pid_t shm_lapid; +#endif }; /* shm_mode upper byte flags */ diff -NurpP --minimal linux-2.6.17.13/include/linux/shmem_fs.h linux-2.6.17.13-g-v/include/linux/shmem_fs.h --- linux-2.6.17.13/include/linux/shmem_fs.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/shmem_fs.h 2006-09-20 15:00:12.000000000 +0200 @@ -8,6 +8,9 @@ #define SHMEM_NR_DIRECT 16 +#define TMPFS_SUPER_MAGIC 0x01021994 + + struct shmem_inode_info { spinlock_t lock; unsigned long flags; diff -NurpP --minimal linux-2.6.17.13/include/linux/stat.h linux-2.6.17.13-g-v/include/linux/stat.h --- linux-2.6.17.13/include/linux/stat.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/stat.h 2006-09-20 15:00:12.000000000 +0200 @@ -63,6 +63,7 @@ struct kstat { unsigned int nlink; uid_t uid; gid_t gid; + xid_t xid; dev_t rdev; loff_t size; struct timespec atime; diff -NurpP --minimal linux-2.6.17.13/include/linux/sunrpc/auth.h linux-2.6.17.13-g-v/include/linux/sunrpc/auth.h --- linux-2.6.17.13/include/linux/sunrpc/auth.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/sunrpc/auth.h 2006-09-20 15:00:12.000000000 +0200 @@ -28,6 +28,7 @@ struct auth_cred { uid_t uid; gid_t gid; + xid_t xid; struct group_info *group_info; }; diff -NurpP --minimal linux-2.6.17.13/include/linux/sunrpc/clnt.h linux-2.6.17.13-g-v/include/linux/sunrpc/clnt.h --- linux-2.6.17.13/include/linux/sunrpc/clnt.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/sunrpc/clnt.h 2006-09-20 15:00:12.000000000 +0200 @@ -52,7 +52,8 @@ struct rpc_clnt { cl_intr : 1,/* interruptible */ cl_autobind : 1,/* use getport() */ cl_oneshot : 1,/* dispose after use */ - cl_dead : 1;/* abandoned */ + cl_dead : 1,/* abandoned */ + cl_tagxid : 1;/* do xid tagging */ struct rpc_rtt * cl_rtt; /* RTO estimator data */ struct rpc_portmap * cl_pmap; /* port mapping */ diff -NurpP --minimal linux-2.6.17.13/include/linux/sysctl.h linux-2.6.17.13-g-v/include/linux/sysctl.h --- linux-2.6.17.13/include/linux/sysctl.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/sysctl.h 2006-09-20 15:00:12.000000000 +0200 @@ -93,6 +93,7 @@ enum KERN_CAP_BSET=14, /* int: capability bounding set */ KERN_PANIC=15, /* int: panic timeout */ KERN_REALROOTDEV=16, /* real root device to mount after initrd */ + KERN_VSHELPER=17, /* string: path to vshelper policy agent */ KERN_SPARC_REBOOT=21, /* reboot command on Sparc */ KERN_CTLALTDEL=22, /* int: allow ctl-alt-del to reboot */ @@ -148,8 +149,21 @@ enum KERN_SPIN_RETRY=70, /* int: number of spinlock retries */ KERN_ACPI_VIDEO_FLAGS=71, /* int: flags for setting up video after ACPI sleep */ KERN_IA64_UNALIGNED=72, /* int: ia64 unaligned userland trap enable */ +#ifdef CONFIG_GRKERNSEC + KERN_GRSECURITY=98, /* grsecurity */ +#endif + +#ifdef CONFIG_PAX_SOFTMODE + KERN_PAX=99, /* PaX control */ +#endif + }; +#ifdef CONFIG_PAX_SOFTMODE +enum { + PAX_SOFTMODE=1 /* PaX: disable/enable soft mode */ +}; +#endif /* CTL_VM names: */ enum @@ -909,6 +923,9 @@ typedef int ctl_handler (ctl_table *tabl typedef int proc_handler (ctl_table *ctl, int write, struct file * filp, void __user *buffer, size_t *lenp, loff_t *ppos); +typedef int virt_handler (struct ctl_table *ctl, int write, xid_t xid, + void **datap, size_t *lenp); + extern int proc_dostring(ctl_table *, int, struct file *, void __user *, size_t *, loff_t *); extern int proc_dointvec(ctl_table *, int, struct file *, @@ -990,6 +1007,7 @@ struct ctl_table mode_t mode; ctl_table *child; proc_handler *proc_handler; /* Callback for text formatting */ + virt_handler *virt_handler; /* Context virtualization */ ctl_handler *strategy; /* Callback function for all r/w */ struct proc_dir_entry *de; /* /proc control block */ void *extra1; diff -NurpP --minimal linux-2.6.17.13/include/linux/sysfs.h linux-2.6.17.13-g-v/include/linux/sysfs.h --- linux-2.6.17.13/include/linux/sysfs.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/sysfs.h 2006-09-20 15:00:12.000000000 +0200 @@ -12,6 +12,8 @@ #include +#define SYSFS_SUPER_MAGIC 0x62656572 + struct kobject; struct module; diff -NurpP --minimal linux-2.6.17.13/include/linux/types.h linux-2.6.17.13-g-v/include/linux/types.h --- linux-2.6.17.13/include/linux/types.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/linux/types.h 2006-09-20 15:00:12.000000000 +0200 @@ -38,6 +38,8 @@ typedef __kernel_uid32_t uid_t; typedef __kernel_gid32_t gid_t; typedef __kernel_uid16_t uid16_t; typedef __kernel_gid16_t gid16_t; +typedef unsigned int xid_t; +typedef unsigned int nid_t; #ifdef CONFIG_UID16 /* This is defined by include/asm-{arch}/posix_types.h */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vroot.h linux-2.6.17.13-g-v/include/linux/vroot.h --- linux-2.6.17.13/include/linux/vroot.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vroot.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,51 @@ + +/* + * include/linux/vroot.h + * + * written by Herbert Pötzl, 9/11/2002 + * ported to 2.6 by Herbert Pötzl, 30/12/2004 + * + * Copyright (C) 2002-2005 by Herbert Pötzl. + * Redistribution of this file is permitted under the + * GNU General Public License. + */ + +#ifndef _LINUX_VROOT_H +#define _LINUX_VROOT_H + + +#ifdef __KERNEL__ + +/* Possible states of device */ +enum { + Vr_unbound, + Vr_bound, +}; + +struct vroot_device { + int vr_number; + int vr_refcnt; + + struct semaphore vr_ctl_mutex; + struct block_device *vr_device; + int vr_state; +}; + + +typedef struct block_device *(vroot_grb_func)(struct block_device *); + +extern int register_vroot_grb(vroot_grb_func *); +extern int unregister_vroot_grb(vroot_grb_func *); + +#endif /* __KERNEL__ */ + +#define MAX_VROOT_DEFAULT 8 + +/* + * IOCTL commands --- we will commandeer 0x56 ('V') + */ + +#define VROOT_SET_DEV 0x5600 +#define VROOT_CLR_DEV 0x5601 + +#endif /* _LINUX_VROOT_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vs_base.h linux-2.6.17.13-g-v/include/linux/vs_base.h --- linux-2.6.17.13/include/linux/vs_base.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vs_base.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,106 @@ +#ifndef _VX_VS_BASE_H +#define _VX_VS_BASE_H + +#include "vserver/context.h" + + +#define vx_task_xid(t) ((t)->xid) + +#define vx_current_xid() vx_task_xid(current) + +#define vx_check(c,m) __vx_check(vx_current_xid(),c,m) + +#define vx_weak_check(c,m) ((m) ? vx_check(c,m) : 1) + + +/* + * check current context for ADMIN/WATCH and + * optionally against supplied argument + */ +static inline int __vx_check(xid_t cid, xid_t id, unsigned int mode) +{ + if (mode & VX_ARG_MASK) { + if ((mode & VX_IDENT) && + (id == cid)) + return 1; + } + if (mode & VX_ATR_MASK) { + if ((mode & VX_DYNAMIC) && + (id >= MIN_D_CONTEXT) && + (id <= MAX_S_CONTEXT)) + return 1; + if ((mode & VX_STATIC) && + (id > 1) && (id < MIN_D_CONTEXT)) + return 1; + } + return (((mode & VX_ADMIN) && (cid == 0)) || + ((mode & VX_WATCH) && (cid == 1)) || + ((mode & VX_HOSTID) && (id == 0))); +} + + +#define __vx_state(v) ((v) ? ((v)->vx_state) : 0) + +#define vx_info_state(v,m) (__vx_state(v) & (m)) + + +/* generic flag merging */ + +#define vx_check_flags(v,m,f) (((v) & (m)) ^ (f)) + +#define vx_mask_flags(v,f,m) (((v) & ~(m)) | ((f) & (m))) + +#define vx_mask_mask(v,f,m) (((v) & ~(m)) | ((v) & (f) & (m))) + +#define vx_check_bit(v,n) ((v) & (1LL << (n))) + + +/* context flags */ + +#define __vx_flags(v) ((v) ? (v)->vx_flags : 0) + +#define vx_current_flags() __vx_flags(current->vx_info) + +#define vx_info_flags(v,m,f) \ + vx_check_flags(__vx_flags(v),(m),(f)) + +#define task_vx_flags(t,m,f) \ + ((t) && vx_info_flags((t)->vx_info, (m), (f))) + +#define vx_flags(m,f) vx_info_flags(current->vx_info,(m),(f)) + + +/* context caps */ + +#define __vx_ccaps(v) ((v) ? (v)->vx_ccaps : 0) + +#define vx_current_ccaps() __vx_ccaps(current->vx_info) + +#define vx_info_ccaps(v,c) (__vx_ccaps(v) & (c)) + +#define vx_ccaps(c) vx_info_ccaps(current->vx_info,(c)) + + +#define __vx_mcaps(v) ((v) ? (v)->vx_ccaps >> 32UL : ~0 ) + +#define vx_info_mcaps(v,c) (__vx_mcaps(v) & (c)) + +#define vx_mcaps(c) vx_info_mcaps(current->vx_info,(c)) + + +#define vx_current_bcaps() \ + (((current->vx_info) && !vx_flags(VXF_STATE_SETUP, 0)) ? \ + current->vx_info->vx_bcaps : cap_bset) + + +#define vx_current_initpid(n) \ + (current->vx_info && \ + (current->vx_info->vx_initpid == (n))) + +#define vx_capable(b,c) (capable(b) || \ + ((current->euid == 0) && vx_ccaps(c))) + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/vs_context.h linux-2.6.17.13-g-v/include/linux/vs_context.h --- linux-2.6.17.13/include/linux/vs_context.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vs_context.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,198 @@ +#ifndef _VX_VS_CONTEXT_H +#define _VX_VS_CONTEXT_H + +#include +#include "vserver/debug.h" + + +#define get_vx_info(i) __get_vx_info(i,__FILE__,__LINE__,__HERE__) + +static inline struct vx_info *__get_vx_info(struct vx_info *vxi, + const char *_file, int _line, void *_here) +{ + if (!vxi) + return NULL; + + vxlprintk(VXD_CBIT(xid, 2), "get_vx_info(%p[#%d.%d])", + vxi, vxi?vxi->vx_id:0, vxi?atomic_read(&vxi->vx_usecnt):0, + _file, _line); + __vxh_get_vx_info(vxi, _here); + + atomic_inc(&vxi->vx_usecnt); + return vxi; +} + + +extern void free_vx_info(struct vx_info *); + +#define put_vx_info(i) __put_vx_info(i,__FILE__,__LINE__,__HERE__) + +static inline void __put_vx_info(struct vx_info *vxi, + const char *_file, int _line, void *_here) +{ + if (!vxi) + return; + + vxlprintk(VXD_CBIT(xid, 2), "put_vx_info(%p[#%d.%d])", + vxi, vxi?vxi->vx_id:0, vxi?atomic_read(&vxi->vx_usecnt):0, + _file, _line); + __vxh_put_vx_info(vxi, _here); + + if (atomic_dec_and_test(&vxi->vx_usecnt)) + free_vx_info(vxi); +} + + +#define init_vx_info(p,i) __init_vx_info(p,i,__FILE__,__LINE__,__HERE__) + +static inline void __init_vx_info(struct vx_info **vxp, struct vx_info *vxi, + const char *_file, int _line, void *_here) +{ + if (vxi) { + vxlprintk(VXD_CBIT(xid, 3), + "init_vx_info(%p[#%d.%d])", + vxi, vxi?vxi->vx_id:0, + vxi?atomic_read(&vxi->vx_usecnt):0, + _file, _line); + __vxh_init_vx_info(vxi, vxp, _here); + + atomic_inc(&vxi->vx_usecnt); + } + *vxp = vxi; +} + + +#define set_vx_info(p,i) __set_vx_info(p,i,__FILE__,__LINE__,__HERE__) + +static inline void __set_vx_info(struct vx_info **vxp, struct vx_info *vxi, + const char *_file, int _line, void *_here) +{ + struct vx_info *vxo; + + if (!vxi) + return; + + vxlprintk(VXD_CBIT(xid, 3), "set_vx_info(%p[#%d.%d])", + vxi, vxi?vxi->vx_id:0, + vxi?atomic_read(&vxi->vx_usecnt):0, + _file, _line); + __vxh_set_vx_info(vxi, vxp, _here); + + atomic_inc(&vxi->vx_usecnt); + vxo = xchg(vxp, vxi); + BUG_ON(vxo); +} + + +#define clr_vx_info(p) __clr_vx_info(p,__FILE__,__LINE__,__HERE__) + +static inline void __clr_vx_info(struct vx_info **vxp, + const char *_file, int _line, void *_here) +{ + struct vx_info *vxo; + + vxo = xchg(vxp, NULL); + if (!vxo) + return; + + vxlprintk(VXD_CBIT(xid, 3), "clr_vx_info(%p[#%d.%d])", + vxo, vxo?vxo->vx_id:0, + vxo?atomic_read(&vxo->vx_usecnt):0, + _file, _line); + __vxh_clr_vx_info(vxo, vxp, _here); + + if (atomic_dec_and_test(&vxo->vx_usecnt)) + free_vx_info(vxo); +} + + +#define claim_vx_info(v,p) \ + __claim_vx_info(v,p,__FILE__,__LINE__,__HERE__) + +static inline void __claim_vx_info(struct vx_info *vxi, + struct task_struct *task, + const char *_file, int _line, void *_here) +{ + vxlprintk(VXD_CBIT(xid, 3), "claim_vx_info(%p[#%d.%d.%d]) %p", + vxi, vxi?vxi->vx_id:0, + vxi?atomic_read(&vxi->vx_usecnt):0, + vxi?atomic_read(&vxi->vx_tasks):0, + task, _file, _line); + __vxh_claim_vx_info(vxi, task, _here); + + atomic_inc(&vxi->vx_tasks); +} + + +extern void unhash_vx_info(struct vx_info *); + +#define release_vx_info(v,p) \ + __release_vx_info(v,p,__FILE__,__LINE__,__HERE__) + +static inline void __release_vx_info(struct vx_info *vxi, + struct task_struct *task, + const char *_file, int _line, void *_here) +{ + vxlprintk(VXD_CBIT(xid, 3), "release_vx_info(%p[#%d.%d.%d]) %p", + vxi, vxi?vxi->vx_id:0, + vxi?atomic_read(&vxi->vx_usecnt):0, + vxi?atomic_read(&vxi->vx_tasks):0, + task, _file, _line); + __vxh_release_vx_info(vxi, task, _here); + + might_sleep(); + + if (atomic_dec_and_test(&vxi->vx_tasks)) + unhash_vx_info(vxi); +} + + +#define task_get_vx_info(p) \ + __task_get_vx_info(p,__FILE__,__LINE__,__HERE__) + +static inline struct vx_info *__task_get_vx_info(struct task_struct *p, + const char *_file, int _line, void *_here) +{ + struct vx_info *vxi; + + task_lock(p); + vxlprintk(VXD_CBIT(xid, 5), "task_get_vx_info(%p)", + p, _file, _line); + vxi = __get_vx_info(p->vx_info, _file, _line, _here); + task_unlock(p); + return vxi; +} + + +static inline void __wakeup_vx_info(struct vx_info *vxi) +{ + if (waitqueue_active(&vxi->vx_wait)) + wake_up_interruptible(&vxi->vx_wait); +} + +extern void exit_vx_info(struct task_struct *, int); + +static inline +struct task_struct *vx_child_reaper(struct task_struct *p) +{ + struct vx_info *vxi = p->vx_info; + struct task_struct *reaper = child_reaper; + + if (!vxi) + goto out; + + BUG_ON(!p->vx_info->vx_reaper); + + /* child reaper for the guest reaper */ + if (vxi->vx_reaper == p) + goto out; + + reaper = vxi->vx_reaper; +out: + return reaper; +} + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/vs_cvirt.h linux-2.6.17.13-g-v/include/linux/vs_cvirt.h --- linux-2.6.17.13/include/linux/vs_cvirt.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vs_cvirt.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,108 @@ +#ifndef _VX_VS_CVIRT_H +#define _VX_VS_CVIRT_H + +#include "vserver/cvirt.h" +#include "vserver/debug.h" + + +/* utsname virtualization */ + +static inline struct new_utsname *vx_new_utsname(void) +{ + if (current->vx_info) + return &current->vx_info->cvirt.utsname; + return &system_utsname; +} + +#define vx_new_uts(x) ((vx_new_utsname())->x) + + +/* pid faking stuff */ + + +#define vx_info_map_pid(v,p) \ + __vx_info_map_pid((v), (p), __FUNC__, __FILE__, __LINE__) +#define vx_info_map_tgid(v,p) vx_info_map_pid(v,p) +#define vx_map_pid(p) vx_info_map_pid(current->vx_info, p) +#define vx_map_tgid(p) vx_map_pid(p) + +static inline int __vx_info_map_pid(struct vx_info *vxi, int pid, + const char *func, const char *file, int line) +{ + if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) { + vxfprintk(VXD_CBIT(cvirt, 2), + "vx_map_tgid: %p/%llx: %d -> %d", + vxi, (long long)vxi->vx_flags, pid, + (pid && pid == vxi->vx_initpid)?1:pid, + func, file, line); + if (pid == 0) + return 0; + if (pid == vxi->vx_initpid) + return 1; + } + return pid; +} + +#define vx_info_rmap_pid(v,p) \ + __vx_info_rmap_pid((v), (p), __FUNC__, __FILE__, __LINE__) +#define vx_rmap_pid(p) vx_info_rmap_pid(current->vx_info, p) +#define vx_rmap_tgid(p) vx_rmap_pid(p) + +static inline int __vx_info_rmap_pid(struct vx_info *vxi, int pid, + const char *func, const char *file, int line) +{ + if (vx_info_flags(vxi, VXF_INFO_INIT, 0)) { + vxfprintk(VXD_CBIT(cvirt, 2), + "vx_rmap_tgid: %p/%llx: %d -> %d", + vxi, (long long)vxi->vx_flags, pid, + (pid == 1)?vxi->vx_initpid:pid, + func, file, line); + if ((pid == 1) && vxi->vx_initpid) + return vxi->vx_initpid; + if (pid == vxi->vx_initpid) + return ~0U; + } + return pid; +} + + +static inline void vx_activate_task(struct task_struct *p) +{ + struct vx_info *vxi; + + if ((vxi = p->vx_info)) { + vx_update_load(vxi); + atomic_inc(&vxi->cvirt.nr_running); + } +} + +static inline void vx_deactivate_task(struct task_struct *p) +{ + struct vx_info *vxi; + + if ((vxi = p->vx_info)) { + vx_update_load(vxi); + atomic_dec(&vxi->cvirt.nr_running); + } +} + +static inline void vx_uninterruptible_inc(struct task_struct *p) +{ + struct vx_info *vxi; + + if ((vxi = p->vx_info)) + atomic_inc(&vxi->cvirt.nr_uninterruptible); +} + +static inline void vx_uninterruptible_dec(struct task_struct *p) +{ + struct vx_info *vxi; + + if ((vxi = p->vx_info)) + atomic_dec(&vxi->cvirt.nr_uninterruptible); +} + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/vs_dlimit.h linux-2.6.17.13-g-v/include/linux/vs_dlimit.h --- linux-2.6.17.13/include/linux/vs_dlimit.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vs_dlimit.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,213 @@ +#ifndef _VX_VS_DLIMIT_H +#define _VX_VS_DLIMIT_H + +#include "vserver/dlimit.h" +#include "vserver/debug.h" + + +#define get_dl_info(i) __get_dl_info(i,__FILE__,__LINE__) + +static inline struct dl_info *__get_dl_info(struct dl_info *dli, + const char *_file, int _line) +{ + if (!dli) + return NULL; + vxlprintk(VXD_CBIT(dlim, 4), "get_dl_info(%p[#%d.%d])", + dli, dli?dli->dl_xid:0, dli?atomic_read(&dli->dl_usecnt):0, + _file, _line); + atomic_inc(&dli->dl_usecnt); + return dli; +} + + +#define free_dl_info(i) \ + call_rcu(&i->dl_rcu, rcu_free_dl_info); + +#define put_dl_info(i) __put_dl_info(i,__FILE__,__LINE__) + +static inline void __put_dl_info(struct dl_info *dli, + const char *_file, int _line) +{ + if (!dli) + return; + vxlprintk(VXD_CBIT(dlim, 4), "put_dl_info(%p[#%d.%d])", + dli, dli?dli->dl_xid:0, dli?atomic_read(&dli->dl_usecnt):0, + _file, _line); + if (atomic_dec_and_test(&dli->dl_usecnt)) + free_dl_info(dli); +} + + +#define __dlimit_char(d) ((d)?'*':' ') + +static inline int __dl_alloc_space(struct super_block *sb, + xid_t xid, dlsize_t nr, const char *file, int line) +{ + struct dl_info *dli = NULL; + int ret = 0; + + if (nr == 0) + goto out; + dli = locate_dl_info(sb, xid); + if (!dli) + goto out; + + spin_lock(&dli->dl_lock); + ret = (dli->dl_space_used + nr > dli->dl_space_total); + if (!ret) + dli->dl_space_used += nr; + spin_unlock(&dli->dl_lock); + put_dl_info(dli); +out: + vxlprintk(VXD_CBIT(dlim, 1), + "ALLOC (%p,#%d)%c %lld bytes (%d)", + sb, xid, __dlimit_char(dli), (long long)nr, + ret, file, line); + return ret; +} + +static inline void __dl_free_space(struct super_block *sb, + xid_t xid, dlsize_t nr, const char *_file, int _line) +{ + struct dl_info *dli = NULL; + + if (nr == 0) + goto out; + dli = locate_dl_info(sb, xid); + if (!dli) + goto out; + + spin_lock(&dli->dl_lock); + if (dli->dl_space_used > nr) + dli->dl_space_used -= nr; + else + dli->dl_space_used = 0; + spin_unlock(&dli->dl_lock); + put_dl_info(dli); +out: + vxlprintk(VXD_CBIT(dlim, 1), + "FREE (%p,#%d)%c %lld bytes", + sb, xid, __dlimit_char(dli), (long long)nr, + _file, _line); +} + +static inline int __dl_alloc_inode(struct super_block *sb, + xid_t xid, const char *_file, int _line) +{ + struct dl_info *dli; + int ret = 0; + + dli = locate_dl_info(sb, xid); + if (!dli) + goto out; + + spin_lock(&dli->dl_lock); + ret = (dli->dl_inodes_used >= dli->dl_inodes_total); + if (!ret) + dli->dl_inodes_used++; +#if 0 + else + vxwprintk("DLIMIT hit (%p,#%d), inode %d>=%d @ %s:%d", + sb, xid, + dli->dl_inodes_used, dli->dl_inodes_total, + file, line); +#endif + spin_unlock(&dli->dl_lock); + put_dl_info(dli); +out: + vxlprintk(VXD_CBIT(dlim, 0), + "ALLOC (%p,#%d)%c inode (%d)", + sb, xid, __dlimit_char(dli), ret, _file, _line); + return ret; +} + +static inline void __dl_free_inode(struct super_block *sb, + xid_t xid, const char *_file, int _line) +{ + struct dl_info *dli; + + dli = locate_dl_info(sb, xid); + if (!dli) + goto out; + + spin_lock(&dli->dl_lock); + if (dli->dl_inodes_used > 1) + dli->dl_inodes_used--; + else + dli->dl_inodes_used = 0; + spin_unlock(&dli->dl_lock); + put_dl_info(dli); +out: + vxlprintk(VXD_CBIT(dlim, 0), + "FREE (%p,#%d)%c inode", + sb, xid, __dlimit_char(dli), _file, _line); +} + +static inline void __dl_adjust_block(struct super_block *sb, xid_t xid, + unsigned int *free_blocks, unsigned int *root_blocks, + const char *_file, int _line) +{ + struct dl_info *dli; + uint64_t broot, bfree; + + dli = locate_dl_info(sb, xid); + if (!dli) + return; + + spin_lock(&dli->dl_lock); + broot = (dli->dl_space_total - + (dli->dl_space_total >> 10) * dli->dl_nrlmult) + >> sb->s_blocksize_bits; + bfree = (dli->dl_space_total - dli->dl_space_used) + >> sb->s_blocksize_bits; + spin_unlock(&dli->dl_lock); + + vxlprintk(VXD_CBIT(dlim, 2), + "ADJUST: %lld,%lld on %d,%d [mult=%d]", + (long long)bfree, (long long)broot, + *free_blocks, *root_blocks, dli->dl_nrlmult, + _file, _line); + if (free_blocks) { + if (*free_blocks > bfree) + *free_blocks = bfree; + } + if (root_blocks) { + if (*root_blocks > broot) + *root_blocks = broot; + } + put_dl_info(dli); +} + +#define DLIMIT_ALLOC_SPACE(in, bytes) \ + __dl_alloc_space((in)->i_sb, (in)->i_xid, (dlsize_t)(bytes), \ + __FILE__, __LINE__ ) + +#define DLIMIT_FREE_SPACE(in, bytes) \ + __dl_free_space((in)->i_sb, (in)->i_xid, (dlsize_t)(bytes), \ + __FILE__, __LINE__ ) + +#define DLIMIT_ALLOC_BLOCK(in, nr) \ + __dl_alloc_space((in)->i_sb, (in)->i_xid, \ + ((dlsize_t)(nr)) << (in)->i_sb->s_blocksize_bits, \ + __FILE__, __LINE__ ) + +#define DLIMIT_FREE_BLOCK(in, nr) \ + __dl_free_space((in)->i_sb, (in)->i_xid, \ + ((dlsize_t)(nr)) << (in)->i_sb->s_blocksize_bits, \ + __FILE__, __LINE__ ) + + +#define DLIMIT_ALLOC_INODE(in) \ + __dl_alloc_inode((in)->i_sb, (in)->i_xid, __FILE__, __LINE__ ) + +#define DLIMIT_FREE_INODE(in) \ + __dl_free_inode((in)->i_sb, (in)->i_xid, __FILE__, __LINE__ ) + + +#define DLIMIT_ADJUST_BLOCK(sb, xid, fb, rb) \ + __dl_adjust_block(sb, xid, fb, rb, __FILE__, __LINE__ ) + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/vs_limit.h linux-2.6.17.13-g-v/include/linux/vs_limit.h --- linux-2.6.17.13/include/linux/vs_limit.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vs_limit.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,107 @@ +#ifndef _VX_VS_LIMIT_H +#define _VX_VS_LIMIT_H + +#include "vserver/limit.h" +#include "vserver/debug.h" +#include "vserver/limit_int.h" + + +#define vx_acc_cres(v,d,p,r) \ + __vx_acc_cres(v, r, d, p, __FILE__, __LINE__) + +#define vx_acc_cres_cond(x,d,p,r) \ + __vx_acc_cres(((x) == vx_current_xid()) ? current->vx_info : 0, \ + r, d, p, __FILE__, __LINE__) + + +#define vx_add_cres(v,a,p,r) \ + __vx_add_cres(v, r, a, p, __FILE__, __LINE__) +#define vx_sub_cres(v,a,p,r) vx_add_cres(v,-(a),p,r) + +#define vx_add_cres_cond(x,a,p,r) \ + __vx_add_cres(((x) == vx_current_xid()) ? current->vx_info : 0, \ + r, a, p, __FILE__, __LINE__) +#define vx_sub_cres_cond(x,a,p,r) vx_add_cres_cond(x,-(a),p,r) + + +/* process and file limits */ + +#define vx_nproc_inc(p) \ + vx_acc_cres((p)->vx_info, 1, p, RLIMIT_NPROC) + +#define vx_nproc_dec(p) \ + vx_acc_cres((p)->vx_info,-1, p, RLIMIT_NPROC) + +#define vx_files_inc(f) \ + vx_acc_cres_cond((f)->f_xid, 1, f, RLIMIT_NOFILE) + +#define vx_files_dec(f) \ + vx_acc_cres_cond((f)->f_xid,-1, f, RLIMIT_NOFILE) + +#define vx_locks_inc(l) \ + vx_acc_cres_cond((l)->fl_xid, 1, l, RLIMIT_LOCKS) + +#define vx_locks_dec(l) \ + vx_acc_cres_cond((l)->fl_xid,-1, l, RLIMIT_LOCKS) + +#define vx_openfd_inc(f) \ + vx_acc_cres(current->vx_info, 1, (void *)(long)(f), VLIMIT_OPENFD) + +#define vx_openfd_dec(f) \ + vx_acc_cres(current->vx_info,-1, (void *)(long)(f), VLIMIT_OPENFD) + + +#define vx_cres_avail(v,n,r) \ + __vx_cres_avail(v, r, n, __FILE__, __LINE__) + + +#define vx_nproc_avail(n) \ + vx_cres_avail(current->vx_info, n, RLIMIT_NPROC) + +#define vx_files_avail(n) \ + vx_cres_avail(current->vx_info, n, RLIMIT_NOFILE) + +#define vx_locks_avail(n) \ + vx_cres_avail(current->vx_info, n, RLIMIT_LOCKS) + +#define vx_openfd_avail(n) \ + vx_cres_avail(current->vx_info, n, VLIMIT_OPENFD) + + +/* socket limits */ + +#define vx_sock_inc(s) \ + vx_acc_cres((s)->sk_vx_info, 1, s, VLIMIT_NSOCK) + +#define vx_sock_dec(s) \ + vx_acc_cres((s)->sk_vx_info,-1, s, VLIMIT_NSOCK) + +#define vx_sock_avail(n) \ + vx_cres_avail(current->vx_info, n, VLIMIT_NSOCK) + + +/* ipc resource limits */ + +#define vx_ipcmsg_add(v,u,a) \ + vx_add_cres(v, a, u, RLIMIT_MSGQUEUE) + +#define vx_ipcmsg_sub(v,u,a) \ + vx_sub_cres(v, a, u, RLIMIT_MSGQUEUE) + +#define vx_ipcmsg_avail(v,a) \ + vx_cres_avail(v, a, RLIMIT_MSGQUEUE) + + +#define vx_ipcshm_add(v,k,a) \ + vx_add_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM) + +#define vx_ipcshm_sub(v,k,a) \ + vx_sub_cres(v, a, (void *)(long)(k), VLIMIT_SHMEM) + +#define vx_ipcshm_avail(v,a) \ + vx_cres_avail(v, a, VLIMIT_SHMEM) + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/vs_memory.h linux-2.6.17.13-g-v/include/linux/vs_memory.h --- linux-2.6.17.13/include/linux/vs_memory.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vs_memory.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,107 @@ +#ifndef _VX_VS_MEMORY_H +#define _VX_VS_MEMORY_H + +#include "vserver/limit.h" +#include "vserver/debug.h" +#include "vserver/limit_int.h" + + +#define __vx_add_long(a,v) (*(v) += (a)) +#define __vx_inc_long(v) (++*(v)) +#define __vx_dec_long(v) (--*(v)) + +#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS +#define __vx_add_value(a,v) atomic_long_add(a,v) +#define __vx_inc_value(v) atomic_long_inc(v) +#define __vx_dec_value(v) atomic_long_dec(v) +#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ +#define __vx_add_value(a,v) __vx_add_long(a,v) +#define __vx_inc_value(v) __vx_inc_long(v) +#define __vx_dec_value(v) __vx_dec_long(v) +#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */ + + +#define vx_acc_page(m,d,v,r) do { \ + if ((d) > 0) \ + __vx_inc_long(&(m->v)); \ + else \ + __vx_dec_long(&(m->v)); \ + __vx_acc_cres(m->mm_vx_info, r, d, m, __FILE__, __LINE__); \ +} while (0) + +#define vx_acc_page_atomic(m,d,v,r) do { \ + if ((d) > 0) \ + __vx_inc_value(&(m->v)); \ + else \ + __vx_dec_value(&(m->v)); \ + __vx_acc_cres(m->mm_vx_info, r, d, m, __FILE__, __LINE__); \ +} while (0) + + +#define vx_acc_pages(m,p,v,r) do { \ + unsigned long __p = (p); \ + __vx_add_long(__p, &(m->v)); \ + __vx_add_cres(m->mm_vx_info, r, __p, m, __FILE__, __LINE__); \ +} while (0) + +#define vx_acc_pages_atomic(m,p,v,r) do { \ + unsigned long __p = (p); \ + __vx_add_value(__p, &(m->v)); \ + __vx_add_cres(m->mm_vx_info, r, __p, m, __FILE__, __LINE__); \ +} while (0) + + + +#define vx_acc_vmpage(m,d) \ + vx_acc_page(m, d, total_vm, RLIMIT_AS) +#define vx_acc_vmlpage(m,d) \ + vx_acc_page(m, d, locked_vm, RLIMIT_MEMLOCK) +#define vx_acc_file_rsspage(m,d) \ + vx_acc_page_atomic(m, d, _file_rss, RLIMIT_RSS) +#define vx_acc_anon_rsspage(m,d) \ + vx_acc_page_atomic(m, d, _anon_rss, VLIMIT_ANON) + +#define vx_acc_vmpages(m,p) \ + vx_acc_pages(m, p, total_vm, RLIMIT_AS) +#define vx_acc_vmlpages(m,p) \ + vx_acc_pages(m, p, locked_vm, RLIMIT_MEMLOCK) +#define vx_acc_file_rsspages(m,p) \ + vx_acc_pages_atomic(m, p, _file_rss, RLIMIT_RSS) +#define vx_acc_anon_rsspages(m,p) \ + vx_acc_pages_atomic(m, p, _anon_rss, VLIMIT_ANON) + +#define vx_pages_add(s,r,p) __vx_add_cres(s, r, p, 0, __FILE__, __LINE__) +#define vx_pages_sub(s,r,p) vx_pages_add(s, r, -(p)) + +#define vx_vmpages_inc(m) vx_acc_vmpage(m, 1) +#define vx_vmpages_dec(m) vx_acc_vmpage(m,-1) +#define vx_vmpages_add(m,p) vx_acc_vmpages(m, p) +#define vx_vmpages_sub(m,p) vx_acc_vmpages(m,-(p)) + +#define vx_vmlocked_inc(m) vx_acc_vmlpage(m, 1) +#define vx_vmlocked_dec(m) vx_acc_vmlpage(m,-1) +#define vx_vmlocked_add(m,p) vx_acc_vmlpages(m, p) +#define vx_vmlocked_sub(m,p) vx_acc_vmlpages(m,-(p)) + +#define vx_file_rsspages_inc(m) vx_acc_file_rsspage(m, 1) +#define vx_file_rsspages_dec(m) vx_acc_file_rsspage(m,-1) +#define vx_file_rsspages_add(m,p) vx_acc_file_rsspages(m, p) +#define vx_file_rsspages_sub(m,p) vx_acc_file_rsspages(m,-(p)) + +#define vx_anon_rsspages_inc(m) vx_acc_anon_rsspage(m, 1) +#define vx_anon_rsspages_dec(m) vx_acc_anon_rsspage(m,-1) +#define vx_anon_rsspages_add(m,p) vx_acc_anon_rsspages(m, p) +#define vx_anon_rsspages_sub(m,p) vx_acc_anon_rsspages(m,-(p)) + + +#define vx_pages_avail(m,p,r) \ + __vx_cres_avail((m)->mm_vx_info, r, p, __FILE__, __LINE__) + +#define vx_vmpages_avail(m,p) vx_pages_avail(m, p, RLIMIT_AS) +#define vx_vmlocked_avail(m,p) vx_pages_avail(m, p, RLIMIT_MEMLOCK) +#define vx_rsspages_avail(m,p) vx_pages_avail(m, p, RLIMIT_RSS) +#define vx_anonpages_avail(m,p) vx_pages_avail(m, p, VLIMIT_ANON) + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/vs_network.h linux-2.6.17.13-g-v/include/linux/vs_network.h --- linux-2.6.17.13/include/linux/vs_network.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vs_network.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,215 @@ +#ifndef _NX_VS_NETWORK_H +#define _NX_VS_NETWORK_H + +#include "vserver/network.h" +#include "vserver/debug.h" + + +#define get_nx_info(i) __get_nx_info(i,__FILE__,__LINE__) + +static inline struct nx_info *__get_nx_info(struct nx_info *nxi, + const char *_file, int _line) +{ + if (!nxi) + return NULL; + + vxlprintk(VXD_CBIT(nid, 2), "get_nx_info(%p[#%d.%d])", + nxi, nxi?nxi->nx_id:0, nxi?atomic_read(&nxi->nx_usecnt):0, + _file, _line); + + atomic_inc(&nxi->nx_usecnt); + return nxi; +} + + +extern void free_nx_info(struct nx_info *); + +#define put_nx_info(i) __put_nx_info(i,__FILE__,__LINE__) + +static inline void __put_nx_info(struct nx_info *nxi, const char *_file, int _line) +{ + if (!nxi) + return; + + vxlprintk(VXD_CBIT(nid, 2), "put_nx_info(%p[#%d.%d])", + nxi, nxi?nxi->nx_id:0, nxi?atomic_read(&nxi->nx_usecnt):0, + _file, _line); + + if (atomic_dec_and_test(&nxi->nx_usecnt)) + free_nx_info(nxi); +} + + +#define init_nx_info(p,i) __init_nx_info(p,i,__FILE__,__LINE__) + +static inline void __init_nx_info(struct nx_info **nxp, struct nx_info *nxi, + const char *_file, int _line) +{ + if (nxi) { + vxlprintk(VXD_CBIT(nid, 3), + "init_nx_info(%p[#%d.%d])", + nxi, nxi?nxi->nx_id:0, + nxi?atomic_read(&nxi->nx_usecnt):0, + _file, _line); + + atomic_inc(&nxi->nx_usecnt); + } + *nxp = nxi; +} + + +#define set_nx_info(p,i) __set_nx_info(p,i,__FILE__,__LINE__) + +static inline void __set_nx_info(struct nx_info **nxp, struct nx_info *nxi, + const char *_file, int _line) +{ + struct nx_info *nxo; + + if (!nxi) + return; + + vxlprintk(VXD_CBIT(nid, 3), "set_nx_info(%p[#%d.%d])", + nxi, nxi?nxi->nx_id:0, + nxi?atomic_read(&nxi->nx_usecnt):0, + _file, _line); + + atomic_inc(&nxi->nx_usecnt); + nxo = xchg(nxp, nxi); + BUG_ON(nxo); +} + +#define clr_nx_info(p) __clr_nx_info(p,__FILE__,__LINE__) + +static inline void __clr_nx_info(struct nx_info **nxp, + const char *_file, int _line) +{ + struct nx_info *nxo; + + nxo = xchg(nxp, NULL); + if (!nxo) + return; + + vxlprintk(VXD_CBIT(nid, 3), "clr_nx_info(%p[#%d.%d])", + nxo, nxo?nxo->nx_id:0, + nxo?atomic_read(&nxo->nx_usecnt):0, + _file, _line); + + if (atomic_dec_and_test(&nxo->nx_usecnt)) + free_nx_info(nxo); +} + + +#define claim_nx_info(v,p) __claim_nx_info(v,p,__FILE__,__LINE__) + +static inline void __claim_nx_info(struct nx_info *nxi, + struct task_struct *task, const char *_file, int _line) +{ + vxlprintk(VXD_CBIT(nid, 3), "claim_nx_info(%p[#%d.%d.%d]) %p", + nxi, nxi?nxi->nx_id:0, + nxi?atomic_read(&nxi->nx_usecnt):0, + nxi?atomic_read(&nxi->nx_tasks):0, + task, _file, _line); + + atomic_inc(&nxi->nx_tasks); +} + + +extern void unhash_nx_info(struct nx_info *); + +#define release_nx_info(v,p) __release_nx_info(v,p,__FILE__,__LINE__) + +static inline void __release_nx_info(struct nx_info *nxi, + struct task_struct *task, const char *_file, int _line) +{ + vxlprintk(VXD_CBIT(nid, 3), "release_nx_info(%p[#%d.%d.%d]) %p", + nxi, nxi?nxi->nx_id:0, + nxi?atomic_read(&nxi->nx_usecnt):0, + nxi?atomic_read(&nxi->nx_tasks):0, + task, _file, _line); + + might_sleep(); + + if (atomic_dec_and_test(&nxi->nx_tasks)) + unhash_nx_info(nxi); +} + + +#define task_get_nx_info(i) __task_get_nx_info(i,__FILE__,__LINE__) + +static __inline__ struct nx_info *__task_get_nx_info(struct task_struct *p, + const char *_file, int _line) +{ + struct nx_info *nxi; + + task_lock(p); + vxlprintk(VXD_CBIT(nid, 5), "task_get_nx_info(%p)", + p, _file, _line); + nxi = __get_nx_info(p->nx_info, _file, _line); + task_unlock(p); + return nxi; +} + + +#define nx_task_nid(t) ((t)->nid) + +#define nx_current_nid() nx_task_nid(current) + +#define nx_check(c,m) __nx_check(nx_current_nid(),c,m) + +#define nx_weak_check(c,m) ((m) ? nx_check(c,m) : 1) + + +#define __nx_state(v) ((v) ? ((v)->nx_state) : 0) + +#define nx_info_state(v,m) (__nx_state(v) & (m)) + + +#define __nx_flags(v) ((v) ? (v)->nx_flags : 0) + +#define nx_current_flags() __nx_flags(current->nx_info) + +#define nx_info_flags(v,m,f) \ + vx_check_flags(__nx_flags(v),(m),(f)) + +#define task_nx_flags(t,m,f) \ + ((t) && nx_info_flags((t)->nx_info, (m), (f))) + +#define nx_flags(m,f) nx_info_flags(current->nx_info,(m),(f)) + + +/* context caps */ + +#define __nx_ncaps(v) ((v) ? (v)->nx_ncaps : 0) + +#define nx_current_ncaps() __nx_ncaps(current->nx_info) + +#define nx_info_ncaps(v,c) (__nx_ncaps(v) & (c)) + +#define nx_ncaps(c) nx_info_ncaps(current->nx_info,(c)) + + +static inline int addr_in_nx_info(struct nx_info *nxi, uint32_t addr) +{ + int n,i; + + if (!nxi) + return 1; + + n = nxi->nbipv4; + for (i=0; iipv4[i] == addr) + return 1; + } + return 0; +} + +static inline void exit_nx_info(struct task_struct *p) +{ + if (p->nx_info) + release_nx_info(p->nx_info, p); +} + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/vs_sched.h linux-2.6.17.13-g-v/include/linux/vs_sched.h --- linux-2.6.17.13/include/linux/vs_sched.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vs_sched.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,92 @@ +#ifndef _VX_VS_SCHED_H +#define _VX_VS_SCHED_H + +#include "vserver/sched.h" + + +#define VAVAVOOM_RATIO 50 + +#define MAX_PRIO_BIAS 20 +#define MIN_PRIO_BIAS -20 + + +static inline int vx_tokens_avail(struct vx_info *vxi) +{ + return atomic_read(&vxi->sched.tokens); +} + +static inline void vx_consume_token(struct vx_info *vxi) +{ + atomic_dec(&vxi->sched.tokens); +} + +static inline int vx_need_resched(struct task_struct *p) +{ +#ifdef CONFIG_VSERVER_HARDCPU + struct vx_info *vxi = p->vx_info; +#endif + int slice = --p->time_slice; + +#ifdef CONFIG_VSERVER_HARDCPU + if (vxi) { + int tokens; + + if ((tokens = vx_tokens_avail(vxi)) > 0) + vx_consume_token(vxi); + /* for tokens > 0, one token was consumed */ + if (tokens < 2) + return 1; + } +#endif + return (slice == 0); +} + + +static inline void vx_onhold_inc(struct vx_info *vxi) +{ + int onhold = atomic_read(&vxi->cvirt.nr_onhold); + + atomic_inc(&vxi->cvirt.nr_onhold); + if (!onhold) + vxi->cvirt.onhold_last = jiffies; +} + +static inline void __vx_onhold_update(struct vx_info *vxi) +{ + int cpu = smp_processor_id(); + uint32_t now = jiffies; + uint32_t delta = now - vxi->cvirt.onhold_last; + + vxi->cvirt.onhold_last = now; + vxi->sched.cpu[cpu].hold_ticks += delta; +} + +static inline void vx_onhold_dec(struct vx_info *vxi) +{ + if (atomic_dec_and_test(&vxi->cvirt.nr_onhold)) + __vx_onhold_update(vxi); +} + +static inline void vx_account_user(struct vx_info *vxi, + cputime_t cputime, int nice) +{ + int cpu = smp_processor_id(); + + if (!vxi) + return; + vxi->sched.cpu[cpu].user_ticks += cputime; +} + +static inline void vx_account_system(struct vx_info *vxi, + cputime_t cputime, int idle) +{ + int cpu = smp_processor_id(); + + if (!vxi) + return; + vxi->sched.cpu[cpu].sys_ticks += cputime; +} + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/vs_socket.h linux-2.6.17.13-g-v/include/linux/vs_socket.h --- linux-2.6.17.13/include/linux/vs_socket.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vs_socket.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,57 @@ +#ifndef _VX_VS_SOCKET_H +#define _VX_VS_SOCKET_H + +#include "vserver/debug.h" + + +/* socket accounting */ + +#include + +static inline int vx_sock_type(int family) +{ + int type = 4; + + if (family > 0 && family < 3) + type = family; + else if (family == PF_INET6) + type = 3; + return type; +} + +#define vx_acc_sock(v,f,p,s) \ + __vx_acc_sock((v), (f), (p), (s), __FILE__, __LINE__) + +static inline void __vx_acc_sock(struct vx_info *vxi, + int family, int pos, int size, char *file, int line) +{ + if (vxi) { + int type = vx_sock_type(family); + + atomic_inc(&vxi->cacct.sock[type][pos].count); + atomic_add(size, &vxi->cacct.sock[type][pos].total); + } +} + +#define vx_sock_recv(sk,s) \ + vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 0, (s)) +#define vx_sock_send(sk,s) \ + vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 1, (s)) +#define vx_sock_fail(sk,s) \ + vx_acc_sock((sk)->sk_vx_info, (sk)->sk_family, 2, (s)) + + +#define sock_vx_init(s) do { \ + (s)->sk_xid = 0; \ + (s)->sk_vx_info = NULL; \ + } while (0) + +#define sock_nx_init(s) do { \ + (s)->sk_nid = 0; \ + (s)->sk_nx_info = NULL; \ + } while (0) + + +#else +#warning duplicate inclusion +#endif diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/context.h linux-2.6.17.13-g-v/include/linux/vserver/context.h --- linux-2.6.17.13/include/linux/vserver/context.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/context.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,172 @@ +#ifndef _VX_CONTEXT_H +#define _VX_CONTEXT_H + +#include + + +#define MAX_S_CONTEXT 65535 /* Arbitrary limit */ +#define MIN_D_CONTEXT 49152 /* dynamic contexts start here */ + +#define VX_DYNAMIC_ID ((uint32_t)-1) /* id for dynamic context */ + +/* context flags */ + +#define VXF_INFO_LOCK 0x00000001 +#define VXF_INFO_SCHED 0x00000002 +#define VXF_INFO_NPROC 0x00000004 +#define VXF_INFO_PRIVATE 0x00000008 + +#define VXF_INFO_INIT 0x00000010 +#define VXF_INFO_HIDE 0x00000020 +#define VXF_INFO_ULIMIT 0x00000040 +#define VXF_INFO_NSPACE 0x00000080 + +#define VXF_SCHED_HARD 0x00000100 +#define VXF_SCHED_PRIO 0x00000200 +#define VXF_SCHED_PAUSE 0x00000400 + +#define VXF_VIRT_MEM 0x00010000 +#define VXF_VIRT_UPTIME 0x00020000 +#define VXF_VIRT_CPU 0x00040000 +#define VXF_VIRT_LOAD 0x00080000 + +#define VXF_HIDE_MOUNT 0x01000000 +#define VXF_HIDE_NETIF 0x02000000 + +#define VXF_STATE_SETUP (1ULL<<32) +#define VXF_STATE_INIT (1ULL<<33) + +#define VXF_SC_HELPER (1ULL<<36) +#define VXF_REBOOT_KILL (1ULL<<37) +#define VXF_PERSISTENT (1ULL<<38) + +#define VXF_FORK_RSS (1ULL<<48) +#define VXF_PROLIFIC (1ULL<<49) + +#define VXF_IGNEG_NICE (1ULL<<52) + +#define VXF_ONE_TIME (0x0003ULL<<32) + +#define VXF_INIT_SET (VXF_STATE_SETUP|VXF_STATE_INIT) + + +/* context migration */ + +#define VXM_SET_INIT 0x00000001 +#define VXM_SET_REAPER 0x00000002 + +/* context caps */ + +#define VXC_CAP_MASK 0x00000000 + +#define VXC_SET_UTSNAME 0x00000001 +#define VXC_SET_RLIMIT 0x00000002 + +#define VXC_RAW_ICMP 0x00000100 +#define VXC_SYSLOG 0x00001000 + +#define VXC_SECURE_MOUNT 0x00010000 +#define VXC_SECURE_REMOUNT 0x00020000 +#define VXC_BINARY_MOUNT 0x00040000 + +#define VXC_QUOTA_CTL 0x00100000 + + +/* context state changes */ + +enum { + VSC_STARTUP = 1, + VSC_SHUTDOWN, + + VSC_NETUP, + VSC_NETDOWN, +}; + + +#ifdef __KERNEL__ + +#include +#include +#include + +#include "limit_def.h" +#include "sched_def.h" +#include "cvirt_def.h" + +struct vx_info { + struct hlist_node vx_hlist; /* linked list of contexts */ + xid_t vx_id; /* context id */ + atomic_t vx_usecnt; /* usage count */ + atomic_t vx_tasks; /* tasks count */ + struct vx_info *vx_parent; /* parent context */ + int vx_state; /* context state */ + + struct namespace *vx_namespace; /* private namespace */ + struct fs_struct *vx_fs; /* private namespace fs */ + uint64_t vx_flags; /* context flags */ + uint64_t vx_bcaps; /* bounding caps (system) */ + uint64_t vx_ccaps; /* context caps (vserver) */ + + struct task_struct *vx_reaper; /* guest reaper process */ + pid_t vx_initpid; /* PID of guest init */ + + struct _vx_limit limit; /* vserver limits */ + struct _vx_sched sched; /* vserver scheduler */ + struct _vx_cvirt cvirt; /* virtual/bias stuff */ + struct _vx_cacct cacct; /* context accounting */ + + wait_queue_head_t vx_wait; /* context exit waitqueue */ + int reboot_cmd; /* last sys_reboot() cmd */ + int exit_code; /* last process exit code */ + + char vx_name[65]; /* vserver name */ +}; + + +/* status flags */ + +#define VXS_HASHED 0x0001 +#define VXS_PAUSED 0x0010 +#define VXS_ONHOLD 0x0020 +#define VXS_SHUTDOWN 0x0100 +#define VXS_HELPER 0x1000 +#define VXS_RELEASED 0x8000 + +/* check conditions */ + +#define VX_ADMIN 0x0001 +#define VX_WATCH 0x0002 +#define VX_HIDE 0x0004 +#define VX_HOSTID 0x0008 + +#define VX_IDENT 0x0010 +#define VX_EQUIV 0x0020 +#define VX_PARENT 0x0040 +#define VX_CHILD 0x0080 + +#define VX_ARG_MASK 0x00F0 + +#define VX_DYNAMIC 0x0100 +#define VX_STATIC 0x0200 + +#define VX_ATR_MASK 0x0F00 + + +extern void claim_vx_info(struct vx_info *, struct task_struct *); +extern void release_vx_info(struct vx_info *, struct task_struct *); + +extern struct vx_info *lookup_vx_info(int); +extern struct vx_info *lookup_or_create_vx_info(int); + +extern int get_xid_list(int, unsigned int *, int); +extern int xid_is_hashed(xid_t); + +extern int vx_migrate_task(struct task_struct *, struct vx_info *); + +extern long vs_state_change(struct vx_info *, unsigned int); + + +#endif /* __KERNEL__ */ +#else /* _VX_CONTEXT_H */ +#warning duplicate inclusion +#endif /* _VX_CONTEXT_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/context_cmd.h linux-2.6.17.13-g-v/include/linux/vserver/context_cmd.h --- linux-2.6.17.13/include/linux/vserver/context_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/context_cmd.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,84 @@ +#ifndef _VX_CONTEXT_CMD_H +#define _VX_CONTEXT_CMD_H + + +/* vinfo commands */ + +#define VCMD_task_xid VC_CMD(VINFO, 1, 0) + +#ifdef __KERNEL__ +extern int vc_task_xid(uint32_t, void __user *); + +#endif /* __KERNEL__ */ + +#define VCMD_vx_info VC_CMD(VINFO, 5, 0) + +struct vcmd_vx_info_v0 { + uint32_t xid; + uint32_t initpid; + /* more to come */ +}; + +#ifdef __KERNEL__ +extern int vc_vx_info(uint32_t, void __user *); + +#endif /* __KERNEL__ */ + + +/* context commands */ + +#define VCMD_ctx_create_v0 VC_CMD(VPROC, 1, 0) +#define VCMD_ctx_create VC_CMD(VPROC, 1, 1) + +struct vcmd_ctx_create { + uint64_t flagword; +}; + +#define VCMD_ctx_migrate_v0 VC_CMD(PROCMIG, 1, 0) +#define VCMD_ctx_migrate VC_CMD(PROCMIG, 1, 1) + +struct vcmd_ctx_migrate { + uint64_t flagword; +}; + +#ifdef __KERNEL__ +extern int vc_ctx_create(uint32_t, void __user *); +extern int vc_ctx_migrate(uint32_t, void __user *); + +#endif /* __KERNEL__ */ + + +/* flag commands */ + +#define VCMD_get_cflags VC_CMD(FLAGS, 1, 0) +#define VCMD_set_cflags VC_CMD(FLAGS, 2, 0) + +struct vcmd_ctx_flags_v0 { + uint64_t flagword; + uint64_t mask; +}; + +#ifdef __KERNEL__ +extern int vc_get_cflags(uint32_t, void __user *); +extern int vc_set_cflags(uint32_t, void __user *); + +#endif /* __KERNEL__ */ + + +/* context caps commands */ + +#define VCMD_get_ccaps VC_CMD(FLAGS, 3, 0) +#define VCMD_set_ccaps VC_CMD(FLAGS, 4, 0) + +struct vcmd_ctx_caps_v0 { + uint64_t bcaps; + uint64_t ccaps; + uint64_t cmask; +}; + +#ifdef __KERNEL__ +extern int vc_get_ccaps(uint32_t, void __user *); +extern int vc_set_ccaps(uint32_t, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_CONTEXT_CMD_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/cvirt.h linux-2.6.17.13-g-v/include/linux/vserver/cvirt.h --- linux-2.6.17.13/include/linux/vserver/cvirt.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/cvirt.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,26 @@ +#ifndef _VX_CVIRT_H +#define _VX_CVIRT_H + + +#ifdef __KERNEL__ + +struct timespec; + +void vx_vsi_uptime(struct timespec *, struct timespec *); + + +struct vx_info; + +void vx_update_load(struct vx_info *); + + +int vx_uts_virt_handler(struct ctl_table *ctl, int write, xid_t xid, + void **datap, size_t *lenp); + + +int vx_do_syslog(int, char __user *, int); + +#endif /* __KERNEL__ */ +#else /* _VX_CVIRT_H */ +#warning duplicate inclusion +#endif /* _VX_CVIRT_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/cvirt_cmd.h linux-2.6.17.13-g-v/include/linux/vserver/cvirt_cmd.h --- linux-2.6.17.13/include/linux/vserver/cvirt_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/cvirt_cmd.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,35 @@ +#ifndef _VX_CVIRT_CMD_H +#define _VX_CVIRT_CMD_H + + +/* virtual host info name commands */ + +#define VCMD_set_vhi_name VC_CMD(VHOST, 1, 0) +#define VCMD_get_vhi_name VC_CMD(VHOST, 2, 0) + +struct vcmd_vhi_name_v0 { + uint32_t field; + char name[65]; +}; + + +enum vhi_name_field { + VHIN_CONTEXT=0, + VHIN_SYSNAME, + VHIN_NODENAME, + VHIN_RELEASE, + VHIN_VERSION, + VHIN_MACHINE, + VHIN_DOMAINNAME, +}; + + +#ifdef __KERNEL__ + +#include + +extern int vc_set_vhi_name(uint32_t, void __user *); +extern int vc_get_vhi_name(uint32_t, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_CVIRT_CMD_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/cvirt_def.h linux-2.6.17.13-g-v/include/linux/vserver/cvirt_def.h --- linux-2.6.17.13/include/linux/vserver/cvirt_def.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/cvirt_def.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,77 @@ +#ifndef _VX_CVIRT_DEF_H +#define _VX_CVIRT_DEF_H + +#include +#include +#include +#include +#include +#include + + +struct _vx_usage_stat { + uint64_t user; + uint64_t nice; + uint64_t system; + uint64_t softirq; + uint64_t irq; + uint64_t idle; + uint64_t iowait; +}; + +struct _vx_syslog { + wait_queue_head_t log_wait; + spinlock_t logbuf_lock; /* lock for the log buffer */ + + unsigned long log_start; /* next char to be read by syslog() */ + unsigned long con_start; /* next char to be sent to consoles */ + unsigned long log_end; /* most-recently-written-char + 1 */ + unsigned long logged_chars; /* #chars since last read+clear operation */ + + char log_buf[1024]; +}; + + +/* context sub struct */ + +struct _vx_cvirt { + int max_threads; /* maximum allowed threads */ + atomic_t nr_threads; /* number of current threads */ + atomic_t nr_running; /* number of running threads */ + atomic_t nr_uninterruptible; /* number of uninterruptible threads */ + + atomic_t nr_onhold; /* processes on hold */ + uint32_t onhold_last; /* jiffies when put on hold */ + + struct timespec bias_idle; + struct timespec bias_uptime; /* context creation point */ + uint64_t bias_clock; /* offset in clock_t */ + + struct new_utsname utsname; + + spinlock_t load_lock; /* lock for the load averages */ + atomic_t load_updates; /* nr of load updates done so far */ + uint32_t load_last; /* last time load was cacled */ + uint32_t load[3]; /* load averages 1,5,15 */ + + atomic_t total_forks; /* number of forks so far */ + + struct _vx_usage_stat cpustat[NR_CPUS]; + + struct _vx_syslog syslog; +}; + +struct _vx_sock_acc { + atomic_t count; + atomic_t total; +}; + +/* context sub struct */ + +struct _vx_cacct { + unsigned long total_forks; + + struct _vx_sock_acc sock[5][3]; +}; + +#endif /* _VX_CVIRT_DEF_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/debug.h linux-2.6.17.13-g-v/include/linux/vserver/debug.h --- linux-2.6.17.13/include/linux/vserver/debug.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/debug.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,298 @@ +#ifndef _VX_DEBUG_H +#define _VX_DEBUG_H + + +#define VXD_CBIT(n,m) (vx_debug_ ## n & (1 << (m))) +#define VXD_CMIN(n,m) (vx_debug_ ## n > (m)) +#define VXD_MASK(n,m) (vx_debug_ ## n & (m)) + +#define VXD_QPOS(v,p) (((uint32_t)(v) >> ((p)*8)) & 0xFF) +#define VXD_QUAD(v) VXD_QPOS(v,0), VXD_QPOS(v,1), \ + VXD_QPOS(v,2), VXD_QPOS(v,3) +#define VXF_QUAD "%u.%u.%u.%u" + +#define VXD_DEV(d) (d), (d)->bd_inode->i_ino, \ + imajor((d)->bd_inode), iminor((d)->bd_inode) +#define VXF_DEV "%p[%lu,%d:%d]" + + +#define __FUNC__ __func__ + + +#ifdef CONFIG_VSERVER_DEBUG + +extern unsigned int vx_debug_switch; +extern unsigned int vx_debug_xid; +extern unsigned int vx_debug_nid; +extern unsigned int vx_debug_net; +extern unsigned int vx_debug_limit; +extern unsigned int vx_debug_cres; +extern unsigned int vx_debug_dlim; +extern unsigned int vx_debug_cvirt; +extern unsigned int vx_debug_misc; + + +#define VX_LOGLEVEL "vxD: " +#define VX_WARNLEVEL KERN_WARNING "vxW: " + +#define vxdprintk(c,f,x...) \ + do { \ + if (c) \ + printk(VX_LOGLEVEL f "\n" , ##x); \ + } while (0) + +#define vxlprintk(c,f,x...) \ + do { \ + if (c) \ + printk(VX_LOGLEVEL f " @%s:%d\n", x); \ + } while (0) + +#define vxfprintk(c,f,x...) \ + do { \ + if (c) \ + printk(VX_LOGLEVEL f " %s@%s:%d\n", x); \ + } while (0) + + +#define vxwprintk(c,f,x...) \ + do { \ + if (c) \ + printk(VX_WARNLEVEL f "\n" , ##x); \ + } while (0) + +#define vxd_path(d,m) \ + ({ static char _buffer[PATH_MAX]; \ + d_path((d), (m), _buffer, sizeof(_buffer)); }) + +#define vxd_cond_path(n) \ + ((n) ? vxd_path((n)->dentry, (n)->mnt) : "" ) + +#else /* CONFIG_VSERVER_DEBUG */ + +#define vx_debug_switch 0 +#define vx_debug_xid 0 +#define vx_debug_nid 0 +#define vx_debug_net 0 +#define vx_debug_limit 0 +#define vx_debug_cres 0 +#define vx_debug_dlim 0 +#define vx_debug_cvirt 0 + +#define vxdprintk(x...) do { } while (0) +#define vxlprintk(x...) do { } while (0) +#define vxfprintk(x...) do { } while (0) +#define vxwprintk(x...) do { } while (0) + +#define vxd_path "" +#define vxd_cond_path vxd_path + +#endif /* CONFIG_VSERVER_DEBUG */ + + +/* history stuff */ + +#ifdef CONFIG_VSERVER_HISTORY + + +extern unsigned volatile int vxh_active; + +struct _vxhe_vxi { + struct vx_info *ptr; + unsigned xid; + unsigned usecnt; + unsigned tasks; +}; + +struct _vxhe_set_clr { + void *data; +}; + +struct _vxhe_loc_lookup { + unsigned arg; +}; + +enum { + VXH_UNUSED=0, + VXH_THROW_OOPS=1, + + VXH_GET_VX_INFO, + VXH_PUT_VX_INFO, + VXH_INIT_VX_INFO, + VXH_SET_VX_INFO, + VXH_CLR_VX_INFO, + VXH_CLAIM_VX_INFO, + VXH_RELEASE_VX_INFO, + VXH_ALLOC_VX_INFO, + VXH_DEALLOC_VX_INFO, + VXH_HASH_VX_INFO, + VXH_UNHASH_VX_INFO, + VXH_LOC_VX_INFO, + VXH_LOOKUP_VX_INFO, + VXH_CREATE_VX_INFO, +}; + +struct _vx_hist_entry { + void *loc; + unsigned short seq; + unsigned short type; + struct _vxhe_vxi vxi; + union { + struct _vxhe_set_clr sc; + struct _vxhe_loc_lookup ll; + }; +}; + +struct _vx_hist_entry *vxh_advance(void *loc); + + +static inline +void __vxh_copy_vxi(struct _vx_hist_entry *entry, struct vx_info *vxi) +{ + entry->vxi.ptr = vxi; + if (vxi) { + entry->vxi.usecnt = atomic_read(&vxi->vx_usecnt); + entry->vxi.tasks = atomic_read(&vxi->vx_tasks); + entry->vxi.xid = vxi->vx_id; + } +} + + +#define __HERE__ current_text_addr() + +#define __VXH_BODY(__type, __data, __here) \ + struct _vx_hist_entry *entry; \ + \ + preempt_disable(); \ + entry = vxh_advance(__here); \ + __data; \ + entry->type = __type; \ + preempt_enable(); + + + /* pass vxi only */ + +#define __VXH_SMPL \ + __vxh_copy_vxi(entry, vxi) + +static inline +void __vxh_smpl(struct vx_info *vxi, int __type, void *__here) +{ + __VXH_BODY(__type, __VXH_SMPL, __here) +} + + /* pass vxi and data (void *) */ + +#define __VXH_DATA \ + __vxh_copy_vxi(entry, vxi); \ + entry->sc.data = data + +static inline +void __vxh_data(struct vx_info *vxi, void *data, + int __type, void *__here) +{ + __VXH_BODY(__type, __VXH_DATA, __here) +} + + /* pass vxi and arg (long) */ + +#define __VXH_LONG \ + __vxh_copy_vxi(entry, vxi); \ + entry->ll.arg = arg + +static inline +void __vxh_long(struct vx_info *vxi, long arg, + int __type, void *__here) +{ + __VXH_BODY(__type, __VXH_LONG, __here) +} + + +static inline +void __vxh_throw_oops(void *__here) +{ + __VXH_BODY(VXH_THROW_OOPS, {}, __here); + /* prevent further acquisition */ + vxh_active = 0; +} + + +#define vxh_throw_oops() __vxh_throw_oops(__HERE__); + +#define __vxh_get_vx_info(v,h) __vxh_smpl(v, VXH_GET_VX_INFO, h); +#define __vxh_put_vx_info(v,h) __vxh_smpl(v, VXH_PUT_VX_INFO, h); + +#define __vxh_init_vx_info(v,d,h) \ + __vxh_data(v,d, VXH_INIT_VX_INFO, h); +#define __vxh_set_vx_info(v,d,h) \ + __vxh_data(v,d, VXH_SET_VX_INFO, h); +#define __vxh_clr_vx_info(v,d,h) \ + __vxh_data(v,d, VXH_CLR_VX_INFO, h); + +#define __vxh_claim_vx_info(v,d,h) \ + __vxh_data(v,d, VXH_CLAIM_VX_INFO, h); +#define __vxh_release_vx_info(v,d,h) \ + __vxh_data(v,d, VXH_RELEASE_VX_INFO, h); + +#define vxh_alloc_vx_info(v) \ + __vxh_smpl(v, VXH_ALLOC_VX_INFO, __HERE__); +#define vxh_dealloc_vx_info(v) \ + __vxh_smpl(v, VXH_DEALLOC_VX_INFO, __HERE__); + +#define vxh_hash_vx_info(v) \ + __vxh_smpl(v, VXH_HASH_VX_INFO, __HERE__); +#define vxh_unhash_vx_info(v) \ + __vxh_smpl(v, VXH_UNHASH_VX_INFO, __HERE__); + +#define vxh_loc_vx_info(v,l) \ + __vxh_long(v,l, VXH_LOC_VX_INFO, __HERE__); +#define vxh_lookup_vx_info(v,l) \ + __vxh_long(v,l, VXH_LOOKUP_VX_INFO, __HERE__); +#define vxh_create_vx_info(v,l) \ + __vxh_long(v,l, VXH_CREATE_VX_INFO, __HERE__); + +extern void vxh_dump_history(void); + + +#else /* CONFIG_VSERVER_HISTORY */ + +#define __HERE__ 0 + +#define vxh_throw_oops() do { } while (0) + +#define __vxh_get_vx_info(v,h) do { } while (0) +#define __vxh_put_vx_info(v,h) do { } while (0) + +#define __vxh_init_vx_info(v,d,h) do { } while (0) +#define __vxh_set_vx_info(v,d,h) do { } while (0) +#define __vxh_clr_vx_info(v,d,h) do { } while (0) + +#define __vxh_claim_vx_info(v,d,h) do { } while (0) +#define __vxh_release_vx_info(v,d,h) do { } while (0) + +#define vxh_alloc_vx_info(v) do { } while (0) +#define vxh_dealloc_vx_info(v) do { } while (0) + +#define vxh_hash_vx_info(v) do { } while (0) +#define vxh_unhash_vx_info(v) do { } while (0) + +#define vxh_loc_vx_info(a,v) do { } while (0) +#define vxh_lookup_vx_info(a,v) do { } while (0) +#define vxh_create_vx_info(a,v) do { } while (0) + +#define vxh_dump_history() do { } while (0) + + +#endif /* CONFIG_VSERVER_HISTORY */ + + +#ifdef CONFIG_VSERVER_DEBUG +#define vxd_assert_lock(l) assert_spin_locked(l) +#define vxd_assert(c,f,x...) vxlprintk(!(c), \ + "assertion [" f "] failed.", ##x, __FILE__, __LINE__) +#else +#define vxd_assert_lock(l) do { } while (0) +#define vxd_assert(c,f,x...) do { } while (0) +#endif + + +#endif /* _VX_DEBUG_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/debug_cmd.h linux-2.6.17.13-g-v/include/linux/vserver/debug_cmd.h --- linux-2.6.17.13/include/linux/vserver/debug_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/debug_cmd.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,14 @@ +#ifndef _VX_DEBUG_CMD_H +#define _VX_DEBUG_CMD_H + + +/* debug commands */ + +#define VCMD_dump_history VC_CMD(DEBUG, 1, 0) + +#ifdef __KERNEL__ + +extern int vc_dump_history(uint32_t); + +#endif /* __KERNEL__ */ +#endif /* _VX_DEBUG_CMD_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/dlimit.h linux-2.6.17.13-g-v/include/linux/vserver/dlimit.h --- linux-2.6.17.13/include/linux/vserver/dlimit.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/dlimit.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,53 @@ +#ifndef _VX_DLIMIT_H +#define _VX_DLIMIT_H + +#include "switch.h" + +#define CDLIM_UNSET (0ULL) +#define CDLIM_INFINITY (~0ULL) +#define CDLIM_KEEP (~1ULL) + + +#ifdef __KERNEL__ + +#include + +struct super_block; + +struct dl_info { + struct hlist_node dl_hlist; /* linked list of contexts */ + struct rcu_head dl_rcu; /* the rcu head */ + xid_t dl_xid; /* context id */ + atomic_t dl_usecnt; /* usage count */ + atomic_t dl_refcnt; /* reference count */ + + struct super_block *dl_sb; /* associated superblock */ + + spinlock_t dl_lock; /* protect the values */ + + uint64_t dl_space_used; /* used space in bytes */ + uint64_t dl_space_total; /* maximum space in bytes */ + uint32_t dl_inodes_used; /* used inodes */ + uint32_t dl_inodes_total; /* maximum inodes */ + + unsigned int dl_nrlmult; /* non root limit mult */ +}; + +struct rcu_head; + +extern void rcu_free_dl_info(struct rcu_head *); +extern void unhash_dl_info(struct dl_info *); + +extern struct dl_info *locate_dl_info(struct super_block *, xid_t); + + +struct kstatfs; + +extern void vx_vsi_statfs(struct super_block *, struct kstatfs *); + +typedef uint64_t dlsize_t; + +#endif /* __KERNEL__ */ +#else /* _VX_DLIMIT_H */ +#warning duplicate inclusion +#endif /* _VX_DLIMIT_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/dlimit_cmd.h linux-2.6.17.13-g-v/include/linux/vserver/dlimit_cmd.h --- linux-2.6.17.13/include/linux/vserver/dlimit_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/dlimit_cmd.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,69 @@ +#ifndef _VX_DLIMIT_CMD_H +#define _VX_DLIMIT_CMD_H + + +/* dlimit vserver commands */ + +#define VCMD_add_dlimit VC_CMD(DLIMIT, 1, 0) +#define VCMD_rem_dlimit VC_CMD(DLIMIT, 2, 0) + +#define VCMD_set_dlimit VC_CMD(DLIMIT, 5, 0) +#define VCMD_get_dlimit VC_CMD(DLIMIT, 6, 0) + +struct vcmd_ctx_dlimit_base_v0 { + const char __user *name; + uint32_t flags; +}; + +struct vcmd_ctx_dlimit_v0 { + const char __user *name; + uint32_t space_used; /* used space in kbytes */ + uint32_t space_total; /* maximum space in kbytes */ + uint32_t inodes_used; /* used inodes */ + uint32_t inodes_total; /* maximum inodes */ + uint32_t reserved; /* reserved for root in % */ + uint32_t flags; +}; + + +#ifdef __KERNEL__ + +#ifdef CONFIG_COMPAT + +struct vcmd_ctx_dlimit_base_v0_x32 { + compat_uptr_t name_ptr; + uint32_t flags; +}; + +struct vcmd_ctx_dlimit_v0_x32 { + compat_uptr_t name_ptr; + uint32_t space_used; /* used space in kbytes */ + uint32_t space_total; /* maximum space in kbytes */ + uint32_t inodes_used; /* used inodes */ + uint32_t inodes_total; /* maximum inodes */ + uint32_t reserved; /* reserved for root in % */ + uint32_t flags; +}; + +#endif /* CONFIG_COMPAT */ + +#include + +extern int vc_add_dlimit(uint32_t, void __user *); +extern int vc_rem_dlimit(uint32_t, void __user *); + +extern int vc_set_dlimit(uint32_t, void __user *); +extern int vc_get_dlimit(uint32_t, void __user *); + +#ifdef CONFIG_COMPAT + +extern int vc_add_dlimit_x32(uint32_t, void __user *); +extern int vc_rem_dlimit_x32(uint32_t, void __user *); + +extern int vc_set_dlimit_x32(uint32_t, void __user *); +extern int vc_get_dlimit_x32(uint32_t, void __user *); + +#endif /* CONFIG_COMPAT */ + +#endif /* __KERNEL__ */ +#endif /* _VX_DLIMIT_CMD_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/inode.h linux-2.6.17.13-g-v/include/linux/vserver/inode.h --- linux-2.6.17.13/include/linux/vserver/inode.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/inode.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,38 @@ +#ifndef _VX_INODE_H +#define _VX_INODE_H + + +#define IATTR_XID 0x01000000 + +#define IATTR_ADMIN 0x00000001 +#define IATTR_WATCH 0x00000002 +#define IATTR_HIDE 0x00000004 +#define IATTR_FLAGS 0x00000007 + +#define IATTR_BARRIER 0x00010000 +#define IATTR_IUNLINK 0x00020000 +#define IATTR_IMMUTABLE 0x00040000 + +#ifdef __KERNEL__ + + +#ifdef CONFIG_VSERVER_PROC_SECURE +#define IATTR_PROC_DEFAULT ( IATTR_ADMIN | IATTR_HIDE ) +#define IATTR_PROC_SYMLINK ( IATTR_ADMIN ) +#else +#define IATTR_PROC_DEFAULT ( IATTR_ADMIN ) +#define IATTR_PROC_SYMLINK ( IATTR_ADMIN ) +#endif + +#define vx_hide_check(c,m) (((m) & IATTR_HIDE) ? vx_check(c,m) : 1) + +#endif /* __KERNEL__ */ + +/* inode ioctls */ + +#define FIOC_GETXFLG _IOR('x', 5, long) +#define FIOC_SETXFLG _IOW('x', 6, long) + +#else /* _VX_INODE_H */ +#warning duplicate inclusion +#endif /* _VX_INODE_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/inode_cmd.h linux-2.6.17.13-g-v/include/linux/vserver/inode_cmd.h --- linux-2.6.17.13/include/linux/vserver/inode_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/inode_cmd.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,59 @@ +#ifndef _VX_INODE_CMD_H +#define _VX_INODE_CMD_H + + +/* inode vserver commands */ + +#define VCMD_get_iattr_v0 VC_CMD(INODE, 1, 0) +#define VCMD_set_iattr_v0 VC_CMD(INODE, 2, 0) + +#define VCMD_get_iattr VC_CMD(INODE, 1, 1) +#define VCMD_set_iattr VC_CMD(INODE, 2, 1) + +struct vcmd_ctx_iattr_v0 { + /* device handle in id */ + uint64_t ino; + uint32_t xid; + uint32_t flags; + uint32_t mask; +}; + +struct vcmd_ctx_iattr_v1 { + const char __user *name; + uint32_t xid; + uint32_t flags; + uint32_t mask; +}; + + +#ifdef __KERNEL__ + + +#ifdef CONFIG_COMPAT + +struct vcmd_ctx_iattr_v1_x32 { + compat_uptr_t name_ptr; + uint32_t xid; + uint32_t flags; + uint32_t mask; +}; + +#endif /* CONFIG_COMPAT */ + +#include + +extern int vc_get_iattr_v0(uint32_t, void __user *); +extern int vc_set_iattr_v0(uint32_t, void __user *); + +extern int vc_get_iattr(uint32_t, void __user *); +extern int vc_set_iattr(uint32_t, void __user *); + +#ifdef CONFIG_COMPAT + +extern int vc_get_iattr_x32(uint32_t, void __user *); +extern int vc_set_iattr_x32(uint32_t, void __user *); + +#endif /* CONFIG_COMPAT */ + +#endif /* __KERNEL__ */ +#endif /* _VX_INODE_CMD_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/legacy.h linux-2.6.17.13-g-v/include/linux/vserver/legacy.h --- linux-2.6.17.13/include/linux/vserver/legacy.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/legacy.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,49 @@ +#ifndef _VX_LEGACY_H +#define _VX_LEGACY_H + +#include "switch.h" + + +/* compatibiliy vserver commands */ + +#define VCMD_new_s_context VC_CMD(COMPAT, 1, 1) +#define VCMD_set_ipv4root VC_CMD(COMPAT, 2, 3) + +#define VCMD_create_context VC_CMD(VSETUP, 1, 0) + +/* compatibiliy vserver arguments */ + +struct vcmd_new_s_context_v1 { + uint32_t remove_cap; + uint32_t flags; +}; + +struct vcmd_set_ipv4root_v3 { + /* number of pairs in id */ + uint32_t broadcast; + struct { + uint32_t ip; + uint32_t mask; + } nx_mask_pair[NB_IPV4ROOT]; +}; + + +#define VX_INFO_LOCK 1 /* Can't request a new vx_id */ +#define VX_INFO_NPROC 4 /* Limit number of processes in a context */ +#define VX_INFO_PRIVATE 8 /* Noone can join this security context */ +#define VX_INFO_INIT 16 /* This process wants to become the */ + /* logical process 1 of the security */ + /* context */ +#define VX_INFO_HIDEINFO 32 /* Hide some information in /proc */ +#define VX_INFO_ULIMIT 64 /* Use ulimit of the current process */ + /* to become the global limits */ + /* of the context */ +#define VX_INFO_NAMESPACE 128 /* save private namespace */ + + +#ifdef __KERNEL__ +extern int vc_new_s_context(uint32_t, void __user *); +extern int vc_set_ipv4root(uint32_t, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_LEGACY_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/limit.h linux-2.6.17.13-g-v/include/linux/vserver/limit.h --- linux-2.6.17.13/include/linux/vserver/limit.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/limit.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,20 @@ +#ifndef _VX_LIMIT_H +#define _VX_LIMIT_H + + +#define VLIMIT_NSOCK 16 +#define VLIMIT_OPENFD 17 +#define VLIMIT_ANON 18 +#define VLIMIT_SHMEM 19 + +#ifdef __KERNEL__ + +struct sysinfo; + +void vx_vsi_meminfo(struct sysinfo *); +void vx_vsi_swapinfo(struct sysinfo *); + +#define NUM_LIMITS 24 + +#endif /* __KERNEL__ */ +#endif /* _VX_LIMIT_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/limit_cmd.h linux-2.6.17.13-g-v/include/linux/vserver/limit_cmd.h --- linux-2.6.17.13/include/linux/vserver/limit_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/limit_cmd.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,55 @@ +#ifndef _VX_LIMIT_CMD_H +#define _VX_LIMIT_CMD_H + + +/* rlimit vserver commands */ + +#define VCMD_get_rlimit VC_CMD(RLIMIT, 1, 0) +#define VCMD_set_rlimit VC_CMD(RLIMIT, 2, 0) +#define VCMD_get_rlimit_mask VC_CMD(RLIMIT, 3, 0) + +struct vcmd_ctx_rlimit_v0 { + uint32_t id; + uint64_t minimum; + uint64_t softlimit; + uint64_t maximum; +}; + +struct vcmd_ctx_rlimit_mask_v0 { + uint32_t minimum; + uint32_t softlimit; + uint32_t maximum; +}; + +#define CRLIM_UNSET (0ULL) +#define CRLIM_INFINITY (~0ULL) +#define CRLIM_KEEP (~1ULL) + +#ifdef __KERNEL__ + +#ifdef CONFIG_IA32_EMULATION + +struct vcmd_ctx_rlimit_v0_x32 { + uint32_t id; + uint64_t minimum; + uint64_t softlimit; + uint64_t maximum; +} __attribute__ ((aligned (4))); + +#endif /* CONFIG_IA32_EMULATION */ + +#include + +extern int vc_get_rlimit(uint32_t, void __user *); +extern int vc_set_rlimit(uint32_t, void __user *); +extern int vc_get_rlimit_mask(uint32_t, void __user *); + +#ifdef CONFIG_IA32_EMULATION + +extern int vc_get_rlimit_x32(uint32_t, void __user *); +extern int vc_set_rlimit_x32(uint32_t, void __user *); + +#endif /* CONFIG_IA32_EMULATION */ + +#endif /* __KERNEL__ */ +#endif /* _VX_LIMIT_CMD_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/limit_def.h linux-2.6.17.13-g-v/include/linux/vserver/limit_def.h --- linux-2.6.17.13/include/linux/vserver/limit_def.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/limit_def.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,22 @@ +#ifndef _VX_LIMIT_DEF_H +#define _VX_LIMIT_DEF_H + +#include +#include + +#include "limit.h" + + +/* context sub struct */ + +struct _vx_limit { + atomic_t ticks; + + unsigned long rlim[NUM_LIMITS]; /* Context limit */ + unsigned long rmax[NUM_LIMITS]; /* Context maximum */ + atomic_t rcur[NUM_LIMITS]; /* Current value */ + atomic_t lhit[NUM_LIMITS]; /* Limit hits */ +}; + + +#endif /* _VX_LIMIT_DEF_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/limit_int.h linux-2.6.17.13-g-v/include/linux/vserver/limit_int.h --- linux-2.6.17.13/include/linux/vserver/limit_int.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/limit_int.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,76 @@ +#ifndef _VX_LIMIT_INT_H +#define _VX_LIMIT_INT_H + + +#ifdef __KERNEL__ + +#define VXD_RCRES(r) VXD_CBIT(cres, (r)) +#define VXD_RLIMIT(r) VXD_CBIT(limit, (r)) + +extern const char *vlimit_name[NUM_LIMITS]; + +static inline void __vx_acc_cres(struct vx_info *vxi, + int res, int dir, void *_data, char *_file, int _line) +{ + if (VXD_RCRES(res)) + vxlprintk(1, "vx_acc_cres[%5d,%s,%2d]: %5d%s (%p)", + (vxi ? vxi->vx_id : -1), vlimit_name[res], res, + (vxi ? atomic_read(&vxi->limit.rcur[res]) : 0), + (dir > 0) ? "++" : "--", _data, _file, _line); + if (!vxi) + return; + + if (dir > 0) + atomic_inc(&vxi->limit.rcur[res]); + else + atomic_dec(&vxi->limit.rcur[res]); +} + +static inline void __vx_add_cres(struct vx_info *vxi, + int res, int amount, void *_data, char *_file, int _line) +{ + if (VXD_RCRES(res)) + vxlprintk(1, "vx_add_cres[%5d,%s,%2d]: %5d += %5d (%p)", + (vxi ? vxi->vx_id : -1), vlimit_name[res], res, + (vxi ? atomic_read(&vxi->limit.rcur[res]) : 0), + amount, _data, _file, _line); + if (amount == 0) + return; + if (!vxi) + return; + atomic_add(amount, &vxi->limit.rcur[res]); +} + +static inline int __vx_cres_avail(struct vx_info *vxi, + int res, int num, char *_file, int _line) +{ + unsigned long value; + + if (VXD_RLIMIT(res)) + vxlprintk(1, "vx_cres_avail[%5d,%s,%2d]: %5ld > %5d + %5d", + (vxi ? vxi->vx_id : -1), vlimit_name[res], res, + (vxi ? vxi->limit.rlim[res] : 1), + (vxi ? atomic_read(&vxi->limit.rcur[res]) : 0), + num, _file, _line); + if (num == 0) + return 1; + if (!vxi) + return 1; + + value = atomic_read(&vxi->limit.rcur[res]); + + if (value > vxi->limit.rmax[res]) + vxi->limit.rmax[res] = value; + + if (vxi->limit.rlim[res] == RLIM_INFINITY) + return 1; + + if (value + num <= vxi->limit.rlim[res]) + return 1; + + atomic_inc(&vxi->limit.lhit[res]); + return 0; +} + +#endif /* __KERNEL__ */ +#endif /* _VX_LIMIT_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/namespace.h linux-2.6.17.13-g-v/include/linux/vserver/namespace.h --- linux-2.6.17.13/include/linux/vserver/namespace.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/namespace.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,15 @@ +#ifndef _VX_NAMESPACE_H +#define _VX_NAMESPACE_H + + +#include + +struct vx_info; +struct namespace; +struct fs_struct; + +extern int vx_set_namespace(struct vx_info *, struct namespace *, struct fs_struct *); + +#else /* _VX_NAMESPACE_H */ +#warning duplicate inclusion +#endif /* _VX_NAMESPACE_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/namespace_cmd.h linux-2.6.17.13-g-v/include/linux/vserver/namespace_cmd.h --- linux-2.6.17.13/include/linux/vserver/namespace_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/namespace_cmd.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,17 @@ +#ifndef _VX_NAMESPACE_CMD_H +#define _VX_NAMESPACE_CMD_H + + +#define VCMD_enter_namespace VC_CMD(PROCALT, 1, 0) + +#define VCMD_set_namespace_v0 VC_CMD(PROCALT, 3, 0) +#define VCMD_set_namespace VC_CMD(PROCALT, 3, 1) + + +#ifdef __KERNEL__ + +extern int vc_enter_namespace(uint32_t, void __user *); +extern int vc_set_namespace(uint32_t, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_NAMESPACE_CMD_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/network.h linux-2.6.17.13-g-v/include/linux/vserver/network.h --- linux-2.6.17.13/include/linux/vserver/network.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/network.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,119 @@ +#ifndef _VX_NETWORK_H +#define _VX_NETWORK_H + +#include + + +#define MAX_N_CONTEXT 65535 /* Arbitrary limit */ + +#define NX_DYNAMIC_ID ((uint32_t)-1) /* id for dynamic context */ + +#define NB_IPV4ROOT 16 + + +/* network flags */ + +#define NXF_STATE_SETUP (1ULL<<32) + +#define NXF_SC_HELPER (1ULL<<36) +#define NXF_PERSISTENT (1ULL<<38) + +#define NXF_ONE_TIME (0x0001ULL<<32) + +#define NXF_INIT_SET (0) + + +/* address types */ + +#define NXA_TYPE_IPV4 1 +#define NXA_TYPE_IPV6 2 + +#define NXA_MOD_BCAST (1<<8) + +#define NXA_TYPE_ANY ((uint16_t)-1) + + +#ifdef __KERNEL__ + +#include +#include +#include +#include + + +struct nx_info { + struct hlist_node nx_hlist; /* linked list of nxinfos */ + nid_t nx_id; /* vnet id */ + atomic_t nx_usecnt; /* usage count */ + atomic_t nx_tasks; /* tasks count */ + int nx_state; /* context state */ + + uint64_t nx_flags; /* network flag word */ + uint64_t nx_ncaps; /* network capabilities */ + + int nbipv4; + __u32 ipv4[NB_IPV4ROOT]; /* Process can only bind to these IPs */ + /* The first one is used to connect */ + /* and for bind any service */ + /* The other must be used explicity */ + __u32 mask[NB_IPV4ROOT]; /* Netmask for each ipv4 */ + /* Used to select the proper source */ + /* address for sockets */ + __u32 v4_bcast; /* Broadcast address to receive UDP */ + + char nx_name[65]; /* network context name */ +}; + + +/* status flags */ + +#define NXS_HASHED 0x0001 +#define NXS_SHUTDOWN 0x0100 +#define NXS_RELEASED 0x8000 + +extern struct nx_info *lookup_nx_info(int); + +extern int get_nid_list(int, unsigned int *, int); +extern int nid_is_hashed(nid_t); + +extern int nx_migrate_task(struct task_struct *, struct nx_info *); + +extern long vs_net_change(struct nx_info *, unsigned int); + +struct in_ifaddr; +struct net_device; + +#ifdef CONFIG_INET +int ifa_in_nx_info(struct in_ifaddr *, struct nx_info *); +int dev_in_nx_info(struct net_device *, struct nx_info *); + +#else /* CONFIG_INET */ +static inline +int ifa_in_nx_info(struct in_ifaddr *a, struct nx_info *n) +{ + return 1; +} + +static inline +int dev_in_nx_info(struct net_device *d, struct nx_info *n) +{ + return 1; +} +#endif /* CONFIG_INET */ + +struct sock; + +#ifdef CONFIG_INET +int nx_addr_conflict(struct nx_info *, uint32_t, struct sock *); +#else /* CONFIG_INET */ +static inline +int nx_addr_conflict(struct nx_info *n, uint32_t a, struct sock *s) +{ + return 1; +} +#endif /* CONFIG_INET */ + +#endif /* __KERNEL__ */ +#else /* _VX_NETWORK_H */ +#warning duplicate inclusion +#endif /* _VX_NETWORK_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/network_cmd.h linux-2.6.17.13-g-v/include/linux/vserver/network_cmd.h --- linux-2.6.17.13/include/linux/vserver/network_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/network_cmd.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,89 @@ +#ifndef _VX_NETWORK_CMD_H +#define _VX_NETWORK_CMD_H + + +/* vinfo commands */ + +#define VCMD_task_nid VC_CMD(VINFO, 2, 0) + +#ifdef __KERNEL__ +extern int vc_task_nid(uint32_t, void __user *); + +#endif /* __KERNEL__ */ + +#define VCMD_nx_info VC_CMD(VINFO, 6, 0) + +struct vcmd_nx_info_v0 { + uint32_t nid; + /* more to come */ +}; + +#ifdef __KERNEL__ +extern int vc_nx_info(uint32_t, void __user *); + +#endif /* __KERNEL__ */ + +#define VCMD_net_create_v0 VC_CMD(VNET, 1, 0) +#define VCMD_net_create VC_CMD(VNET, 1, 1) + +struct vcmd_net_create { + uint64_t flagword; +}; + +#define VCMD_net_migrate VC_CMD(NETMIG, 1, 0) + +#define VCMD_net_add VC_CMD(NETALT, 1, 0) +#define VCMD_net_remove VC_CMD(NETALT, 2, 0) + +struct vcmd_net_addr_v0 { + uint16_t type; + uint16_t count; + uint32_t ip[4]; + uint32_t mask[4]; + /* more to come */ +}; + + +#ifdef __KERNEL__ +extern int vc_net_create(uint32_t, void __user *); +extern int vc_net_migrate(uint32_t, void __user *); + +extern int vc_net_add(uint32_t, void __user *); +extern int vc_net_remove(uint32_t, void __user *); + +#endif /* __KERNEL__ */ + + +/* flag commands */ + +#define VCMD_get_nflags VC_CMD(FLAGS, 5, 0) +#define VCMD_set_nflags VC_CMD(FLAGS, 6, 0) + +struct vcmd_net_flags_v0 { + uint64_t flagword; + uint64_t mask; +}; + +#ifdef __KERNEL__ +extern int vc_get_nflags(uint32_t, void __user *); +extern int vc_set_nflags(uint32_t, void __user *); + +#endif /* __KERNEL__ */ + + +/* network caps commands */ + +#define VCMD_get_ncaps VC_CMD(FLAGS, 7, 0) +#define VCMD_set_ncaps VC_CMD(FLAGS, 8, 0) + +struct vcmd_net_caps_v0 { + uint64_t ncaps; + uint64_t cmask; +}; + +#ifdef __KERNEL__ +extern int vc_get_ncaps(uint32_t, void __user *); +extern int vc_set_ncaps(uint32_t, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_CONTEXT_CMD_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/sched.h linux-2.6.17.13-g-v/include/linux/vserver/sched.h --- linux-2.6.17.13/include/linux/vserver/sched.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/sched.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,26 @@ +#ifndef _VX_SCHED_H +#define _VX_SCHED_H + + +#ifdef __KERNEL__ + +struct timespec; + +void vx_vsi_uptime(struct timespec *, struct timespec *); + + +struct vx_info; + +void vx_update_load(struct vx_info *); + + +struct task_struct; + +int vx_effective_vavavoom(struct vx_info *, int); + +int vx_tokens_recalc(struct vx_info *); + +#endif /* __KERNEL__ */ +#else /* _VX_SCHED_H */ +#warning duplicate inclusion +#endif /* _VX_SCHED_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/sched_cmd.h linux-2.6.17.13-g-v/include/linux/vserver/sched_cmd.h --- linux-2.6.17.13/include/linux/vserver/sched_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/sched_cmd.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,48 @@ +#ifndef _VX_SCHED_CMD_H +#define _VX_SCHED_CMD_H + + +/* sched vserver commands */ + +#define VCMD_set_sched_v2 VC_CMD(SCHED, 1, 2) +#define VCMD_set_sched VC_CMD(SCHED, 1, 3) + +struct vcmd_set_sched_v2 { + int32_t fill_rate; + int32_t interval; + int32_t tokens; + int32_t tokens_min; + int32_t tokens_max; + uint64_t cpu_mask; +}; + +struct vcmd_set_sched_v3 { + uint32_t set_mask; + int32_t fill_rate; + int32_t interval; + int32_t tokens; + int32_t tokens_min; + int32_t tokens_max; + int32_t priority_bias; +}; + + +#define VXSM_FILL_RATE 0x0001 +#define VXSM_INTERVAL 0x0002 +#define VXSM_TOKENS 0x0010 +#define VXSM_TOKENS_MIN 0x0020 +#define VXSM_TOKENS_MAX 0x0040 +#define VXSM_PRIO_BIAS 0x0100 + +#define SCHED_KEEP (-2) + +#ifdef __KERNEL__ + +#include + +extern int vc_set_sched_v1(uint32_t, void __user *); +extern int vc_set_sched_v2(uint32_t, void __user *); +extern int vc_set_sched(uint32_t, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_SCHED_CMD_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/sched_def.h linux-2.6.17.13-g-v/include/linux/vserver/sched_def.h --- linux-2.6.17.13/include/linux/vserver/sched_def.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/sched_def.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,38 @@ +#ifndef _VX_SCHED_DEF_H +#define _VX_SCHED_DEF_H + +#include +#include +#include +#include +#include + + +struct _vx_ticks { + uint64_t user_ticks; /* token tick events */ + uint64_t sys_ticks; /* token tick events */ + uint64_t hold_ticks; /* token ticks paused */ + uint64_t unused[5]; /* cacheline ? */ +}; + +/* context sub struct */ + +struct _vx_sched { + atomic_t tokens; /* number of CPU tokens */ + spinlock_t tokens_lock; /* lock for token bucket */ + + int fill_rate; /* Fill rate: add X tokens... */ + int interval; /* Divisor: per Y jiffies */ + int tokens_min; /* Limit: minimum for unhold */ + int tokens_max; /* Limit: no more than N tokens */ + uint32_t jiffies; /* last time accounted */ + + int priority_bias; /* bias offset for priority */ + int vavavoom; /* last calculated vavavoom */ + + cpumask_t cpus_allowed; /* cpu mask for context */ + + struct _vx_ticks cpu[NR_CPUS]; +}; + +#endif /* _VX_SCHED_DEF_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/signal.h linux-2.6.17.13-g-v/include/linux/vserver/signal.h --- linux-2.6.17.13/include/linux/vserver/signal.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/signal.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,14 @@ +#ifndef _VX_SIGNAL_H +#define _VX_SIGNAL_H + + +#ifdef __KERNEL__ + +struct vx_info; + +int vx_info_kill(struct vx_info *, int, int); + +#endif /* __KERNEL__ */ +#else /* _VX_SIGNAL_H */ +#warning duplicate inclusion +#endif /* _VX_SIGNAL_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/signal_cmd.h linux-2.6.17.13-g-v/include/linux/vserver/signal_cmd.h --- linux-2.6.17.13/include/linux/vserver/signal_cmd.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/signal_cmd.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,26 @@ +#ifndef _VX_SIGNAL_CMD_H +#define _VX_SIGNAL_CMD_H + + +/* signalling vserver commands */ + +#define VCMD_ctx_kill VC_CMD(PROCTRL, 1, 0) +#define VCMD_wait_exit VC_CMD(EVENT, 99, 0) + +struct vcmd_ctx_kill_v0 { + int32_t pid; + int32_t sig; +}; + +struct vcmd_wait_exit_v0 { + int32_t reboot_cmd; + int32_t exit_code; +}; + +#ifdef __KERNEL__ + +extern int vc_ctx_kill(uint32_t, void __user *); +extern int vc_wait_exit(uint32_t, void __user *); + +#endif /* __KERNEL__ */ +#endif /* _VX_SIGNAL_CMD_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/switch.h linux-2.6.17.13-g-v/include/linux/vserver/switch.h --- linux-2.6.17.13/include/linux/vserver/switch.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/switch.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,98 @@ +#ifndef _VX_SWITCH_H +#define _VX_SWITCH_H + +#include + + +#define VC_CATEGORY(c) (((c) >> 24) & 0x3F) +#define VC_COMMAND(c) (((c) >> 16) & 0xFF) +#define VC_VERSION(c) ((c) & 0xFFF) + +#define VC_CMD(c,i,v) ((((VC_CAT_ ## c) & 0x3F) << 24) \ + | (((i) & 0xFF) << 16) | ((v) & 0xFFF)) + +/* + + Syscall Matrix V2.8 + + |VERSION|CREATE |MODIFY |MIGRATE|CONTROL|EXPERIM| |SPECIAL|SPECIAL| + |STATS |DESTROY|ALTER |CHANGE |LIMIT |TEST | | | | + |INFO |SETUP | |MOVE | | | | | | + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + SYSTEM |VERSION|VSETUP |VHOST | | | | |DEVICES| | + HOST | 00| 01| 02| 03| 04| 05| | 06| 07| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + CPU | |VPROC |PROCALT|PROCMIG|PROCTRL| | |SCHED. | | + PROCESS| 08| 09| 10| 11| 12| 13| | 14| 15| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + MEMORY | | | | | | | |SWAP | | + | 16| 17| 18| 19| 20| 21| | 22| 23| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + NETWORK| |VNET |NETALT |NETMIG |NETCTL | | |SERIAL | | + | 24| 25| 26| 27| 28| 29| | 30| 31| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + DISK | | | | |DLIMIT | | |INODE | | + VFS | 32| 33| 34| 35| 36| 37| | 38| 39| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + OTHER | | | | | | | |VINFO | | + | 40| 41| 42| 43| 44| 45| | 46| 47| + =======+=======+=======+=======+=======+=======+=======+ +=======+=======+ + SPECIAL|EVENT | | | |FLAGS | | | | | + | 48| 49| 50| 51| 52| 53| | 54| 55| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + SPECIAL|DEBUG | | | |RLIMIT |SYSCALL| | |COMPAT | + | 56| 57| 58| 59| 60|TEST 61| | 62| 63| + -------+-------+-------+-------+-------+-------+-------+ +-------+-------+ + +*/ + +#define VC_CAT_VERSION 0 + +#define VC_CAT_VSETUP 1 +#define VC_CAT_VHOST 2 + +#define VC_CAT_VPROC 9 +#define VC_CAT_PROCALT 10 +#define VC_CAT_PROCMIG 11 +#define VC_CAT_PROCTRL 12 + +#define VC_CAT_SCHED 14 + +#define VC_CAT_VNET 25 +#define VC_CAT_NETALT 26 +#define VC_CAT_NETMIG 27 +#define VC_CAT_NETCTRL 28 + +#define VC_CAT_DLIMIT 36 +#define VC_CAT_INODE 38 + +#define VC_CAT_VINFO 46 +#define VC_CAT_EVENT 48 + +#define VC_CAT_FLAGS 52 +#define VC_CAT_DEBUG 56 +#define VC_CAT_RLIMIT 60 + +#define VC_CAT_SYSTEST 61 +#define VC_CAT_COMPAT 63 + +/* interface version */ + +#define VCI_VERSION 0x00020002 +#define VCI_LEGACY_VERSION 0x000100FF + +/* query version */ + +#define VCMD_get_version VC_CMD(VERSION, 0, 0) + + +#ifdef __KERNEL__ + +#include + + +#else /* __KERNEL__ */ +#define __user +#endif /* __KERNEL__ */ + +#endif /* _VX_SWITCH_H */ diff -NurpP --minimal linux-2.6.17.13/include/linux/vserver/xid.h linux-2.6.17.13-g-v/include/linux/vserver/xid.h --- linux-2.6.17.13/include/linux/vserver/xid.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/include/linux/vserver/xid.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,146 @@ +#ifndef _VX_XID_H +#define _VX_XID_H + +#include + + +#define XID_TAG(in) (IS_TAGXID(in)) + + +#ifdef CONFIG_XID_TAG_NFSD +#define XID_TAG_NFSD 1 +#else +#define XID_TAG_NFSD 0 +#endif + + +#ifdef CONFIG_INOXID_NONE + +#define MAX_UID 0xFFFFFFFF +#define MAX_GID 0xFFFFFFFF + +#define INOXID_XID(tag, uid, gid, xid) (0) + +#define XIDINO_UID(tag, uid, xid) (uid) +#define XIDINO_GID(tag, gid, xid) (gid) + +#endif + + +#ifdef CONFIG_INOXID_GID16 + +#define MAX_UID 0xFFFFFFFF +#define MAX_GID 0x0000FFFF + +#define INOXID_XID(tag, uid, gid, xid) \ + ((tag) ? (((gid) >> 16) & 0xFFFF) : 0) + +#define XIDINO_UID(tag, uid, xid) (uid) +#define XIDINO_GID(tag, gid, xid) \ + ((tag) ? (((gid) & 0xFFFF) | ((xid) << 16)) : (gid)) + +#endif + + +#ifdef CONFIG_INOXID_UGID24 + +#define MAX_UID 0x00FFFFFF +#define MAX_GID 0x00FFFFFF + +#define INOXID_XID(tag, uid, gid, xid) \ + ((tag) ? ((((uid) >> 16) & 0xFF00) | (((gid) >> 24) & 0xFF)) : 0) + +#define XIDINO_UID(tag, uid, xid) \ + ((tag) ? (((uid) & 0xFFFFFF) | (((xid) & 0xFF00) << 16)) : (uid)) +#define XIDINO_GID(tag, gid, xid) \ + ((tag) ? (((gid) & 0xFFFFFF) | (((xid) & 0x00FF) << 24)) : (gid)) + +#endif + + +#ifdef CONFIG_INOXID_UID16 + +#define MAX_UID 0x0000FFFF +#define MAX_GID 0xFFFFFFFF + +#define INOXID_XID(tag, uid, gid, xid) \ + ((tag) ? (((uid) >> 16) & 0xFFFF) : 0) + +#define XIDINO_UID(tag, uid, xid) \ + ((tag) ? (((uid) & 0xFFFF) | ((xid) << 16)) : (uid)) +#define XIDINO_GID(tag, gid, xid) (gid) + +#endif + + +#ifdef CONFIG_INOXID_INTERN + +#define MAX_UID 0xFFFFFFFF +#define MAX_GID 0xFFFFFFFF + +#define INOXID_XID(tag, uid, gid, xid) \ + ((tag) ? (xid) : 0) + +#define XIDINO_UID(tag, uid, xid) (uid) +#define XIDINO_GID(tag, gid, xid) (gid) + +#endif + + +#ifdef CONFIG_INOXID_RUNTIME + +#define MAX_UID 0xFFFFFFFF +#define MAX_GID 0xFFFFFFFF + +#define INOXID_XID(tag, uid, gid, xid) (0) + +#define XIDINO_UID(tag, uid, xid) (uid) +#define XIDINO_GID(tag, gid, xid) (gid) + +#endif + + +#ifndef CONFIG_INOXID_NONE +#define vx_current_fsxid(sb) \ + ((sb)->s_flags & MS_TAGXID ? current->xid : 0) +#else +#define vx_current_fsxid(sb) (0) +#endif + +#ifndef CONFIG_INOXID_INTERN +#define XIDINO_XID(tag, xid) (0) +#else +#define XIDINO_XID(tag, xid) ((tag) ? (xid) : 0) +#endif + +#define INOXID_UID(tag, uid, gid) \ + ((tag) ? ((uid) & MAX_UID) : (uid)) +#define INOXID_GID(tag, uid, gid) \ + ((tag) ? ((gid) & MAX_GID) : (gid)) + + +static inline uid_t vx_map_uid(uid_t uid) +{ + if ((uid > MAX_UID) && (uid != -1)) + uid = -2; + return (uid & MAX_UID); +} + +static inline gid_t vx_map_gid(gid_t gid) +{ + if ((gid > MAX_GID) && (gid != -1)) + gid = -2; + return (gid & MAX_GID); +} + + +#ifdef CONFIG_VSERVER_LEGACY +#define FIOC_GETXID _IOR('x', 1, long) +#define FIOC_SETXID _IOW('x', 2, long) +#define FIOC_SETXIDJ _IOW('x', 3, long) +#endif + +int vx_parse_xid(char *string, xid_t *xid, int remove); +void vx_propagate_xid(struct nameidata *nd, struct inode *inode); + +#endif /* _VX_XID_H */ diff -NurpP --minimal linux-2.6.17.13/include/net/af_unix.h linux-2.6.17.13-g-v/include/net/af_unix.h --- linux-2.6.17.13/include/net/af_unix.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/net/af_unix.h 2006-09-20 15:00:12.000000000 +0200 @@ -18,9 +18,9 @@ extern spinlock_t unix_table_lock; extern atomic_t unix_tot_inflight; -static inline struct sock *first_unix_socket(int *i) +static inline struct sock *next_unix_socket_table(int *i) { - for (*i = 0; *i <= UNIX_HASH_SIZE; (*i)++) { + for ((*i)++; *i <= UNIX_HASH_SIZE; (*i)++) { if (!hlist_empty(&unix_socket_table[*i])) return __sk_head(&unix_socket_table[*i]); } @@ -29,16 +29,19 @@ static inline struct sock *first_unix_so static inline struct sock *next_unix_socket(int *i, struct sock *s) { - struct sock *next = sk_next(s); - /* More in this chain? */ - if (next) - return next; - /* Look for next non-empty chain. */ - for ((*i)++; *i <= UNIX_HASH_SIZE; (*i)++) { - if (!hlist_empty(&unix_socket_table[*i])) - return __sk_head(&unix_socket_table[*i]); - } - return NULL; + do { + if (s) + s = sk_next(s); + if (!s) + s = next_unix_socket_table(i); + } while (s && !vx_check(s->sk_xid, VX_IDENT|VX_WATCH)); + return s; +} + +static inline struct sock *first_unix_socket(int *i) +{ + *i = 0; + return next_unix_socket(i, NULL); } #define forall_unix_sockets(i, s) \ diff -NurpP --minimal linux-2.6.17.13/include/net/inet_hashtables.h linux-2.6.17.13-g-v/include/net/inet_hashtables.h --- linux-2.6.17.13/include/net/inet_hashtables.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/net/inet_hashtables.h 2006-09-20 15:00:12.000000000 +0200 @@ -272,6 +272,25 @@ static inline int inet_iif(const struct return ((struct rtable *)skb->dst)->rt_iif; } +/* + * Check if a given address matches for an inet socket + * + * nxi: the socket's nx_info if any + * addr: to be verified address + * saddr: socket addresses + */ +static inline int inet_addr_match ( + struct nx_info *nxi, + uint32_t addr, + uint32_t saddr) +{ + if (addr && (saddr == addr)) + return 1; + if (!saddr) + return addr_in_nx_info(nxi, addr); + return 0; +} + extern struct sock *__inet_lookup_listener(const struct hlist_head *head, const u32 daddr, const unsigned short hnum, @@ -292,7 +311,7 @@ static inline struct sock * const struct inet_sock *inet = inet_sk((sk = __sk_head(head))); if (inet->num == hnum && !sk->sk_node.next && - (!inet->rcv_saddr || inet->rcv_saddr == daddr) && + inet_addr_match(sk->sk_nx_info, daddr, inet->rcv_saddr) && (sk->sk_family == PF_INET || !ipv6_only_sock(sk)) && !sk->sk_bound_dev_if) goto sherry_cache; diff -NurpP --minimal linux-2.6.17.13/include/net/inet_sock.h linux-2.6.17.13-g-v/include/net/inet_sock.h --- linux-2.6.17.13/include/net/inet_sock.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/net/inet_sock.h 2006-09-20 15:00:12.000000000 +0200 @@ -115,6 +115,7 @@ struct inet_sock { /* Socket demultiplex comparisons on incoming packets. */ __u32 daddr; __u32 rcv_saddr; + __u32 rcv_saddr2; /* Second bound ipv4 addr, for ipv4root */ __u16 dport; __u16 num; __u32 saddr; diff -NurpP --minimal linux-2.6.17.13/include/net/inet_timewait_sock.h linux-2.6.17.13-g-v/include/net/inet_timewait_sock.h --- linux-2.6.17.13/include/net/inet_timewait_sock.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/net/inet_timewait_sock.h 2006-09-20 15:00:12.000000000 +0200 @@ -116,6 +116,10 @@ struct inet_timewait_sock { #define tw_refcnt __tw_common.skc_refcnt #define tw_hash __tw_common.skc_hash #define tw_prot __tw_common.skc_prot +#define tw_xid __tw_common.skc_xid +#define tw_vx_info __tw_common.skc_vx_info +#define tw_nid __tw_common.skc_nid +#define tw_nx_info __tw_common.skc_nx_info volatile unsigned char tw_substate; /* 3 bits hole, try to pack */ unsigned char tw_rcv_wscale; diff -NurpP --minimal linux-2.6.17.13/include/net/route.h linux-2.6.17.13-g-v/include/net/route.h --- linux-2.6.17.13/include/net/route.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/net/route.h 2006-09-20 15:00:12.000000000 +0200 @@ -28,11 +28,14 @@ #include #include #include +#include #include #include #include #include #include +#include +#include #ifndef __KERNEL__ #warning This file is not supposed to be used outside of kernel. @@ -144,6 +147,59 @@ static inline char rt_tos2priority(u8 to return ip_tos2prio[IPTOS_TOS(tos)>>1]; } +#define IPI_LOOPBACK htonl(INADDR_LOOPBACK) + +static inline int ip_find_src(struct nx_info *nxi, struct rtable **rp, struct flowi *fl) +{ + int err; + int i, n = nxi->nbipv4; + u32 ipv4root = nxi->ipv4[0]; + + if (ipv4root == 0) + return 0; + + if (fl->fl4_src == 0) { + if (n > 1) { + u32 foundsrc; + + err = __ip_route_output_key(rp, fl); + if (err) { + fl->fl4_src = ipv4root; + err = __ip_route_output_key(rp, fl); + } + if (err) + return err; + + foundsrc = (*rp)->rt_src; + ip_rt_put(*rp); + + for (i=0; imask[i]; + u32 ipv4 = nxi->ipv4[i]; + u32 net4 = ipv4 & mask; + + if (foundsrc == ipv4) { + fl->fl4_src = ipv4; + break; + } + if (!fl->fl4_src && (foundsrc & mask) == net4) + fl->fl4_src = ipv4; + } + } + if (fl->fl4_src == 0) + fl->fl4_src = (fl->fl4_dst == IPI_LOOPBACK) + ? IPI_LOOPBACK : ipv4root; + } else { + for (i=0; iipv4[i] == fl->fl4_src) + break; + } + if (i == n) + return -EPERM; + } + return 0; +} + static inline int ip_route_connect(struct rtable **rp, u32 dst, u32 src, u32 tos, int oif, u8 protocol, u16 sport, u16 dport, struct sock *sk) @@ -158,7 +214,27 @@ static inline int ip_route_connect(struc .dport = dport } } }; int err; - if (!dst || !src) { + struct nx_info *nx_info = current->nx_info; + + if (sk) + nx_info = sk->sk_nx_info; + vxdprintk(VXD_CBIT(net, 4), + "ip_route_connect(%p) %p,%p;%lx", + sk, nx_info, sk->sk_socket, + (sk->sk_socket?sk->sk_socket->flags:0)); + + if (nx_info) { + err = ip_find_src(nx_info, rp, &fl); + if (err) + return err; + if (fl.fl4_dst == IPI_LOOPBACK && !vx_check(0, VX_ADMIN)) + fl.fl4_dst = nx_info->ipv4[0]; +#ifdef CONFIG_VSERVER_REMAP_SADDR + if (fl.fl4_src == IPI_LOOPBACK && !vx_check(0, VX_ADMIN)) + fl.fl4_src = nx_info->ipv4[0]; +#endif + } + if (!fl.fl4_dst || !fl.fl4_src) { err = __ip_route_output_key(rp, &fl); if (err) return err; diff -NurpP --minimal linux-2.6.17.13/include/net/sock.h linux-2.6.17.13-g-v/include/net/sock.h --- linux-2.6.17.13/include/net/sock.h 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/include/net/sock.h 2006-09-20 15:00:12.000000000 +0200 @@ -115,6 +115,10 @@ struct sock_common { atomic_t skc_refcnt; unsigned int skc_hash; struct proto *skc_prot; + xid_t skc_xid; + struct vx_info *skc_vx_info; + nid_t skc_nid; + struct nx_info *skc_nx_info; }; /** @@ -189,6 +193,10 @@ struct sock { #define sk_refcnt __sk_common.skc_refcnt #define sk_hash __sk_common.skc_hash #define sk_prot __sk_common.skc_prot +#define sk_xid __sk_common.skc_xid +#define sk_vx_info __sk_common.skc_vx_info +#define sk_nid __sk_common.skc_nid +#define sk_nx_info __sk_common.skc_nx_info unsigned char sk_shutdown : 2, sk_no_check : 2, sk_userlocks : 4; diff -NurpP --minimal linux-2.6.17.13/init/Kconfig linux-2.6.17.13-g-v/init/Kconfig --- linux-2.6.17.13/init/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/init/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -268,6 +268,7 @@ menuconfig EMBEDDED config KALLSYMS bool "Load all symbols for debugging/kksymoops" if EMBEDDED default y + depends on !GRKERNSEC_HIDESYM help Say Y here to let the kernel print out symbolic crash information and symbolic stack backtraces. This increases the size of the kernel diff -NurpP --minimal linux-2.6.17.13/init/main.c linux-2.6.17.13-g-v/init/main.c --- linux-2.6.17.13/init/main.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/init/main.c 2006-09-20 15:00:12.000000000 +0200 @@ -100,6 +100,7 @@ static inline void mark_rodata_ro(void) #ifdef CONFIG_TC extern void tc_init(void); #endif +extern void grsecurity_init(void); enum system_states system_state; EXPORT_SYMBOL(system_state); @@ -150,6 +151,15 @@ static int __init maxcpus(char *str) __setup("maxcpus=", maxcpus); +#ifdef CONFIG_PAX_SOFTMODE +static int __init setup_pax_softmode(char *str) +{ + get_option(&str, &pax_softmode); + return 1; +} +__setup("pax_softmode=", setup_pax_softmode); +#endif + static char * argv_init[MAX_INIT_ARGS+2] = { "init", NULL, }; char * envp_init[MAX_INIT_ENVS+2] = { "HOME=/", "TERM=linux", NULL, }; static const char *panic_later, *panic_param; @@ -692,6 +702,8 @@ static int init(void * unused) prepare_namespace(); } + grsecurity_init(); + /* * Ok, we have completed the initial bootup, and * we're essentially up and running. Get rid of the diff -NurpP --minimal linux-2.6.17.13/init/version.c linux-2.6.17.13-g-v/init/version.c --- linux-2.6.17.13/init/version.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/init/version.c 2006-09-20 15:00:12.000000000 +0200 @@ -31,3 +31,8 @@ EXPORT_SYMBOL(system_utsname); const char linux_banner[] = "Linux version " UTS_RELEASE " (" LINUX_COMPILE_BY "@" LINUX_COMPILE_HOST ") (" LINUX_COMPILER ") " UTS_VERSION "\n"; + +const char vx_linux_banner[] = + "Linux version %s (" LINUX_COMPILE_BY "@" + LINUX_COMPILE_HOST ") (" LINUX_COMPILER ") %s\n"; + diff -NurpP --minimal linux-2.6.17.13/ipc/mqueue.c linux-2.6.17.13-g-v/ipc/mqueue.c --- linux-2.6.17.13/ipc/mqueue.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/ipc/mqueue.c 2006-09-20 15:00:12.000000000 +0200 @@ -26,6 +26,8 @@ #include #include #include +#include +#include #include #include "util.h" @@ -149,17 +151,20 @@ static struct inode *mqueue_get_inode(st spin_lock(&mq_lock); if (u->mq_bytes + mq_bytes < u->mq_bytes || u->mq_bytes + mq_bytes > - p->signal->rlim[RLIMIT_MSGQUEUE].rlim_cur) { + p->signal->rlim[RLIMIT_MSGQUEUE].rlim_cur || + !vx_ipcmsg_avail(p->vx_info, mq_bytes)) { spin_unlock(&mq_lock); goto out_inode; } u->mq_bytes += mq_bytes; + vx_ipcmsg_add(p->vx_info, u, mq_bytes); spin_unlock(&mq_lock); info->messages = kmalloc(mq_msg_tblsz, GFP_KERNEL); if (!info->messages) { spin_lock(&mq_lock); u->mq_bytes -= mq_bytes; + vx_ipcmsg_sub(p->vx_info, u, mq_bytes); spin_unlock(&mq_lock); goto out_inode; } @@ -257,10 +262,14 @@ static void mqueue_delete_inode(struct i (info->attr.mq_maxmsg * info->attr.mq_msgsize)); user = info->user; if (user) { + struct vx_info *vxi = lookup_vx_info(user->xid); + spin_lock(&mq_lock); user->mq_bytes -= mq_bytes; + vx_ipcmsg_sub(vxi, user, mq_bytes); queues_count--; spin_unlock(&mq_lock); + put_vx_info(vxi); free_uid(user); } } @@ -739,7 +748,7 @@ asmlinkage long sys_mq_unlink(const char if (inode) atomic_inc(&inode->i_count); - err = vfs_unlink(dentry->d_parent->d_inode, dentry); + err = vfs_unlink(dentry->d_parent->d_inode, dentry, NULL); out_err: dput(dentry); diff -NurpP --minimal linux-2.6.17.13/ipc/msg.c linux-2.6.17.13-g-v/ipc/msg.c --- linux-2.6.17.13/ipc/msg.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/ipc/msg.c 2006-09-20 15:00:12.000000000 +0200 @@ -32,6 +32,7 @@ #include #include #include +#include #include #include @@ -105,6 +106,7 @@ static int newque (key_t key, int msgflg msq->q_perm.mode = (msgflg & S_IRWXUGO); msq->q_perm.key = key; + msq->q_perm.xid = vx_current_xid(); msq->q_perm.security = NULL; retval = security_msg_queue_alloc(msq); @@ -237,6 +239,9 @@ asmlinkage long sys_msgget (key_t key, i msg_unlock(msq); } mutex_unlock(&msg_ids.mutex); + + gr_log_msgget(ret, msgflg); + return ret; } @@ -495,6 +500,8 @@ asmlinkage long sys_msgctl (int msqid, i break; } case IPC_RMID: + gr_log_msgrm(ipcp->uid, ipcp->cuid); + freeque (msq, msqid); break; } @@ -826,6 +833,9 @@ static int sysvipc_msg_proc_show(struct { struct msg_queue *msq = it; + if (!vx_check(msq->q_perm.xid, VX_IDENT)) + return 0; + return seq_printf(s, "%10d %10d %4o %10lu %10lu %5u %5u %5u %5u %5u %5u %10lu %10lu %10lu\n", msq->q_perm.key, diff -NurpP --minimal linux-2.6.17.13/ipc/sem.c linux-2.6.17.13-g-v/ipc/sem.c --- linux-2.6.17.13/ipc/sem.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/ipc/sem.c 2006-09-20 15:00:12.000000000 +0200 @@ -79,6 +79,7 @@ #include #include #include +#include #include #include "util.h" @@ -184,6 +185,7 @@ static int newary (key_t key, int nsems, sma->sem_perm.mode = (semflg & S_IRWXUGO); sma->sem_perm.key = key; + sma->sem_perm.xid = vx_current_xid(); sma->sem_perm.security = NULL; retval = security_sem_alloc(sma); @@ -247,6 +249,9 @@ asmlinkage long sys_semget (key_t key, i } mutex_unlock(&sem_ids.mutex); + + gr_log_semget(err, semflg); + return err; } @@ -840,6 +845,8 @@ static int semctl_down(int semid, int se switch(cmd){ case IPC_RMID: + gr_log_semrm(ipcp->uid, ipcp->cuid); + freeary(sma, semid); err = 0; break; @@ -1345,6 +1352,9 @@ static int sysvipc_sem_proc_show(struct { struct sem_array *sma = it; + if (!vx_check(sma->sem_perm.xid, VX_IDENT)) + return 0; + return seq_printf(s, "%10d %10d %4o %10lu %5u %5u %5u %5u %10lu %10lu\n", sma->sem_perm.key, diff -NurpP --minimal linux-2.6.17.13/ipc/shm.c linux-2.6.17.13-g-v/ipc/shm.c --- linux-2.6.17.13/ipc/shm.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/ipc/shm.c 2006-09-20 15:00:12.000000000 +0200 @@ -33,6 +33,9 @@ #include #include #include +#include +#include +#include #include @@ -56,6 +59,14 @@ static void shm_close (struct vm_area_st static int sysvipc_shm_proc_show(struct seq_file *s, void *it); #endif +#ifdef CONFIG_GRKERNSEC +extern int gr_handle_shmat(const pid_t shm_cprid, const pid_t shm_lapid, + const time_t shm_createtime, const uid_t cuid, + const int shmid); +extern int gr_chroot_shmat(const pid_t shm_cprid, const pid_t shm_lapid, + const time_t shm_createtime); +#endif + size_t shm_ctlmax = SHMMAX; size_t shm_ctlall = SHMALL; int shm_ctlmni = SHMMNI; @@ -117,7 +128,12 @@ static void shm_open (struct vm_area_str */ static void shm_destroy (struct shmid_kernel *shp) { - shm_tot -= (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT; + struct vx_info *vxi = lookup_vx_info(shp->shm_perm.xid); + int numpages = (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT; + + vx_ipcshm_sub(vxi, shp, numpages); + shm_tot -= numpages; + shm_rmid (shp->id); shm_unlock(shp); if (!is_file_hugepages(shp->shm_file)) @@ -127,6 +143,7 @@ static void shm_destroy (struct shmid_ke shp->mlock_user); fput (shp->shm_file); security_shm_free(shp); + put_vx_info(vxi); ipc_rcu_putref(shp); } @@ -149,6 +166,17 @@ static void shm_close (struct vm_area_st shp->shm_lprid = current->tgid; shp->shm_dtim = get_seconds(); shp->shm_nattch--; +#ifdef CONFIG_GRKERNSEC_SHM + if (grsec_enable_shm) { + if (shp->shm_nattch == 0) { + shp->shm_perm.mode |= SHM_DEST; + shm_destroy(shp); + } else + shm_unlock(shp); + mutex_unlock(&shm_ids.mutex); + return; + } +#endif if(shp->shm_nattch == 0 && shp->shm_perm.mode & SHM_DEST) shm_destroy (shp); @@ -203,12 +231,15 @@ static int newseg (key_t key, int shmflg if (shm_tot + numpages >= shm_ctlall) return -ENOSPC; + if (!vx_ipcshm_avail(current->vx_info, numpages)) + return -ENOSPC; shp = ipc_rcu_alloc(sizeof(*shp)); if (!shp) return -ENOMEM; shp->shm_perm.key = key; + shp->shm_perm.xid = vx_current_xid(); shp->shm_perm.mode = (shmflg & S_IRWXUGO); shp->mlock_user = NULL; @@ -248,6 +279,9 @@ static int newseg (key_t key, int shmflg shp->shm_lprid = 0; shp->shm_atim = shp->shm_dtim = 0; shp->shm_ctim = get_seconds(); +#ifdef CONFIG_GRKERNSEC + shp->shm_createtime = get_seconds(); +#endif shp->shm_segsz = size; shp->shm_nattch = 0; shp->id = shm_buildid(id,shp->shm_perm.seq); @@ -259,6 +293,7 @@ static int newseg (key_t key, int shmflg file->f_op = &shm_file_operations; shm_tot += numpages; + vx_ipcshm_add(current->vx_info, key, numpages); shm_unlock(shp); return shp->id; @@ -302,6 +337,8 @@ asmlinkage long sys_shmget (key_t key, s } mutex_unlock(&shm_ids.mutex); + gr_log_shmget(err, shmflg, size); + return err; } @@ -615,6 +652,8 @@ asmlinkage long sys_shmctl (int shmid, i if (err) goto out_unlock_up; + gr_log_shmrm(shp->shm_perm.uid, shp->shm_perm.cuid); + if (shp->shm_nattch){ shp->shm_perm.mode |= SHM_DEST; /* Do not find it any more */ @@ -763,9 +802,27 @@ long do_shmat(int shmid, char __user *sh return err; } +#ifdef CONFIG_GRKERNSEC + if (!gr_handle_shmat(shp->shm_cprid, shp->shm_lapid, shp->shm_createtime, + shp->shm_perm.cuid, shmid)) { + shm_unlock(shp); + return -EACCES; + } + + if (!gr_chroot_shmat(shp->shm_cprid, shp->shm_lapid, shp->shm_createtime)) { + shm_unlock(shp); + return -EACCES; + } +#endif + file = shp->shm_file; size = i_size_read(file->f_dentry->d_inode); shp->shm_nattch++; + +#ifdef CONFIG_GRKERNSEC + shp->shm_lapid = current->pid; +#endif + shm_unlock(shp); down_write(&current->mm->mmap_sem); @@ -914,6 +971,9 @@ static int sysvipc_shm_proc_show(struct #define SMALL_STRING "%10d %10d %4o %10u %5u %5u %5d %5u %5u %5u %5u %10lu %10lu %10lu\n" #define BIG_STRING "%10d %10d %4o %21u %5u %5u %5d %5u %5u %5u %5u %10lu %10lu %10lu\n" + if (!vx_check(shp->shm_perm.xid, VX_IDENT)) + return 0; + if (sizeof(size_t) <= sizeof(int)) format = SMALL_STRING; else @@ -935,3 +995,24 @@ static int sysvipc_shm_proc_show(struct shp->shm_ctim); } #endif + +void gr_shm_exit(struct task_struct *task) +{ +#ifdef CONFIG_GRKERNSEC_SHM + int i; + struct shmid_kernel *shp; + + if (!grsec_enable_shm) + return; + + for (i = 0; i <= shm_ids.max_id; i++) { + shp = shm_get(i); + if (shp && (shp->shm_cprid == task->pid) && + (shp->shm_nattch <= 0)) { + shp->shm_perm.mode |= SHM_DEST; + shm_destroy(shp); + } + } +#endif + return; +} diff -NurpP --minimal linux-2.6.17.13/ipc/util.c linux-2.6.17.13-g-v/ipc/util.c --- linux-2.6.17.13/ipc/util.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/ipc/util.c 2006-09-20 15:00:12.000000000 +0200 @@ -158,7 +158,9 @@ int ipc_findkey(struct ipc_ids* ids, key */ for (id = 0; id <= max_id; id++) { p = ids->entries->p[id]; - if(p==NULL) + if (p==NULL) + continue; + if (!vx_check(p->xid, VX_IDENT)) continue; if (key == p->key) return id; @@ -471,6 +473,9 @@ int ipcperms (struct kern_ipc_perm *ipcp if (unlikely((err = audit_ipc_obj(ipcp)))) return err; + + if (!vx_check(ipcp->xid, VX_ADMIN|VX_IDENT)) /* maybe just VX_IDENT? */ + return -1; requested_mode = (flag >> 6) | (flag >> 3) | flag; granted_mode = ipcp->mode; if (current->euid == ipcp->cuid || current->euid == ipcp->uid) diff -NurpP --minimal linux-2.6.17.13/kernel/Makefile linux-2.6.17.13-g-v/kernel/Makefile --- linux-2.6.17.13/kernel/Makefile 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/Makefile 2006-09-20 15:00:12.000000000 +0200 @@ -10,6 +10,8 @@ obj-y = sched.o fork.o exec_domain.o kthread.o wait.o kfifo.o sys_ni.o posix-cpu-timers.o mutex.o \ hrtimer.o +obj-y += vserver/ + obj-$(CONFIG_DEBUG_MUTEXES) += mutex-debug.o obj-$(CONFIG_FUTEX) += futex.o ifeq ($(CONFIG_COMPAT),y) diff -NurpP --minimal linux-2.6.17.13/kernel/capability.c linux-2.6.17.13-g-v/kernel/capability.c --- linux-2.6.17.13/kernel/capability.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/capability.c 2006-09-20 15:00:12.000000000 +0200 @@ -12,6 +12,8 @@ #include #include #include +#include +#include #include unsigned securebits = SECUREBITS_DEFAULT; /* systemwide security settings */ @@ -234,9 +236,12 @@ out: return ret; } +extern int gr_task_is_capable(struct task_struct *task, const int cap); +extern int gr_is_capable_nolog(const int cap); + int __capable(struct task_struct *t, int cap) { - if (security_capable(t, cap) == 0) { + if ((security_capable(t, cap) == 0) && gr_task_is_capable(t, cap)) { t->flags |= PF_SUPERPRIV; return 1; } @@ -244,8 +249,20 @@ int __capable(struct task_struct *t, int } EXPORT_SYMBOL(__capable); +int capable_nolog(int cap) +{ + if ((security_capable(current, cap) == 0) && gr_is_capable_nolog(cap)) { + current->flags |= PF_SUPERPRIV; + return 1; + } + return 0; +} int capable(int cap) { + /* here for now so we don't require task locking */ + if (vx_check_bit(VXC_CAP_MASK, cap) && !vx_mcaps(1L << cap)) + return 0; return __capable(current, cap); } EXPORT_SYMBOL(capable); +EXPORT_SYMBOL(capable_nolog); diff -NurpP --minimal linux-2.6.17.13/kernel/configs.c linux-2.6.17.13-g-v/kernel/configs.c --- linux-2.6.17.13/kernel/configs.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/configs.c 2006-09-20 15:00:12.000000000 +0200 @@ -89,8 +89,16 @@ static int __init ikconfig_init(void) struct proc_dir_entry *entry; /* create the current config file */ +#ifdef CONFIG_GRKERNSEC_PROC_ADD +#ifdef CONFIG_GRKERNSEC_PROC_USER + entry = create_proc_entry("config.gz", S_IFREG | S_IRUSR, &proc_root); +#elif CONFIG_GRKERNSEC_PROC_USERGROUP + entry = create_proc_entry("config.gz", S_IFREG | S_IRUSR | S_IRGRP, &proc_root); +#endif +#else entry = create_proc_entry("config.gz", S_IFREG | S_IRUGO, &proc_root); +#endif if (!entry) return -ENOMEM; diff -NurpP --minimal linux-2.6.17.13/kernel/cpuset.c linux-2.6.17.13-g-v/kernel/cpuset.c --- linux-2.6.17.13/kernel/cpuset.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/cpuset.c 2006-09-20 15:00:12.000000000 +0200 @@ -49,6 +49,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/kernel/exit.c linux-2.6.17.13-g-v/kernel/exit.c --- linux-2.6.17.13/kernel/exit.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/exit.c 2006-09-20 15:00:12.000000000 +0200 @@ -36,6 +36,15 @@ #include #include #include /* for audit_free() */ +#include +#include +#include +#include +#include + +#ifdef CONFIG_GRKERNSEC +extern rwlock_t grsec_exec_file_lock; +#endif #include #include @@ -115,6 +124,7 @@ static void __exit_signal(struct task_st __unhash_process(tsk); + gr_del_task_from_ip_table(tsk); tsk->signal = NULL; tsk->sighand = NULL; spin_unlock(&sighand->siglock); @@ -287,6 +297,15 @@ static void reparent_to_init(void) { write_lock_irq(&tasklist_lock); +#ifdef CONFIG_GRKERNSEC + write_lock(&grsec_exec_file_lock); + if (current->exec_file) { + fput(current->exec_file); + current->exec_file = NULL; + } + write_unlock(&grsec_exec_file_lock); +#endif + ptrace_unlink(current); /* Reparent to init */ remove_parent(current); @@ -294,6 +313,8 @@ static void reparent_to_init(void) current->real_parent = child_reaper; add_parent(current); + gr_set_kernel_label(current); + /* Set the exit signal to SIGCHLD so we signal init on exit */ current->exit_signal = SIGCHLD; @@ -390,6 +411,17 @@ void daemonize(const char *name, ...) vsnprintf(current->comm, sizeof(current->comm), name, args); va_end(args); +#ifdef CONFIG_GRKERNSEC + write_lock(&grsec_exec_file_lock); + if (current->exec_file) { + fput(current->exec_file); + current->exec_file = NULL; + } + write_unlock(&grsec_exec_file_lock); +#endif + + gr_set_kernel_label(current); + /* * If we were started as result of loading a module, close all of the * user space pages. We don't need them, and if we didn't close them @@ -449,6 +481,7 @@ static void close_files(struct files_str struct file * file = xchg(&fdt->fd[i], NULL); if (file) filp_close(file, files); + vx_openfd_dec(i); } i++; set >>= 1; @@ -591,6 +624,11 @@ static void exit_mm(struct task_struct * static inline void choose_new_parent(task_t *p, task_t *reaper) { + /* check for reaper context */ + vxwprintk((p->xid != reaper->xid) && (reaper != child_reaper), + "rogue reaper: %p[%d,#%u] <> %p[%d,#%u]", + p, p->pid, p->xid, reaper, reaper->pid, reaper->xid); + /* * Make sure we're not reparenting to ourselves and that * the parent is not a zombie. @@ -672,7 +710,7 @@ static void forget_original_parent(struc do { reaper = next_thread(reaper); if (reaper == father) { - reaper = child_reaper; + reaper = vx_child_reaper(father); break; } } while (reaper->exit_state); @@ -696,7 +734,7 @@ static void forget_original_parent(struc if (father == p->real_parent) { /* reparent with a reaper, real father it's us */ - choose_new_parent(p, reaper); + choose_new_parent(p, vx_child_reaper(p)); reparent_thread(p, father, 0); } else { /* reparent ptraced task to its real parent */ @@ -905,12 +943,19 @@ fastcall NORET_TYPE void do_exit(long co #endif if (unlikely(tsk->audit_context)) audit_free(tsk); + + gr_acl_handle_psacct(tsk, code); + gr_acl_handle_exit(); + exit_mm(tsk); exit_sem(tsk); + gr_shm_exit(tsk); __exit_files(tsk); __exit_fs(tsk); exit_namespace(tsk); + exit_vx_info(tsk, code); + exit_nx_info(tsk); exit_thread(); cpuset_exit(tsk); exit_keys(tsk); diff -NurpP --minimal linux-2.6.17.13/kernel/fork.c linux-2.6.17.13-g-v/kernel/fork.c --- linux-2.6.17.13/kernel/fork.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/fork.c 2006-09-20 15:00:12.000000000 +0200 @@ -44,6 +44,11 @@ #include #include #include +#include +#include +#include +#include +#include #include #include @@ -104,6 +109,8 @@ static kmem_cache_t *mm_cachep; void free_task(struct task_struct *tsk) { free_thread_info(tsk->thread_info); + clr_vx_info(&tsk->vx_info); + clr_nx_info(&tsk->nx_info); free_task_struct(tsk); } EXPORT_SYMBOL(free_task); @@ -198,9 +205,11 @@ static inline int dup_mmap(struct mm_str mm->locked_vm = 0; mm->mmap = NULL; mm->mmap_cache = NULL; - mm->free_area_cache = oldmm->mmap_base; - mm->cached_hole_size = ~0UL; + mm->free_area_cache = oldmm->free_area_cache; + mm->cached_hole_size = oldmm->cached_hole_size; mm->map_count = 0; + __set_mm_counter(mm, file_rss, 0); + __set_mm_counter(mm, anon_rss, 0); cpus_clear(mm->cpu_vm_mask); mm->mm_rb = RB_ROOT; rb_link = &mm->mm_rb.rb_node; @@ -212,7 +221,7 @@ static inline int dup_mmap(struct mm_str if (mpnt->vm_flags & VM_DONTCOPY) { long pages = vma_pages(mpnt); - mm->total_vm -= pages; + vx_vmpages_sub(mm, pages); vm_stat_account(mm, mpnt->vm_flags, mpnt->vm_file, -pages); continue; @@ -319,16 +328,15 @@ static struct mm_struct * mm_init(struct INIT_LIST_HEAD(&mm->mmlist); mm->core_waiters = 0; mm->nr_ptes = 0; - set_mm_counter(mm, file_rss, 0); - set_mm_counter(mm, anon_rss, 0); spin_lock_init(&mm->page_table_lock); rwlock_init(&mm->ioctx_list_lock); mm->ioctx_list = NULL; - mm->free_area_cache = TASK_UNMAPPED_BASE; + mm->free_area_cache = ~0UL; mm->cached_hole_size = ~0UL; if (likely(!mm_alloc_pgd(mm))) { mm->def_flags = 0; + set_vx_info(&mm->mm_vx_info, current->vx_info); return mm; } free_mm(mm); @@ -360,6 +368,7 @@ void fastcall __mmdrop(struct mm_struct BUG_ON(mm == &init_mm); mm_free_pgd(mm); destroy_context(mm); + clr_vx_info(&mm->mm_vx_info); free_mm(mm); } @@ -463,6 +472,7 @@ static struct mm_struct *dup_mm(struct t goto fail_nomem; memcpy(mm, oldmm, sizeof(*mm)); + mm->mm_vx_info = NULL; if (!mm_init(mm)) goto fail_nomem; @@ -490,6 +500,7 @@ fail_nocontext: * If init_new_context() failed, we cannot use mmput() to free the mm * because it calls destroy_context() */ + clr_vx_info(&mm->mm_vx_info); mm_free_pgd(mm); free_mm(mm); return NULL; @@ -683,6 +694,8 @@ static struct files_struct *dup_fd(struc struct file *f = *old_fds++; if (f) { get_file(f); + /* FIXME: sum it first for check and performance */ + vx_openfd_inc(open_files - i); } else { /* * The fd may be claimed in the fd bitmap but not yet @@ -927,6 +940,8 @@ static task_t *copy_process(unsigned lon { int retval; struct task_struct *p = NULL; + struct vx_info *vxi; + struct nx_info *nxi; if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS)) return ERR_PTR(-EINVAL); @@ -955,12 +970,32 @@ static task_t *copy_process(unsigned lon if (!p) goto fork_out; + init_vx_info(&p->vx_info, current->vx_info); + init_nx_info(&p->nx_info, current->nx_info); + + /* check vserver memory */ + if (p->mm && !(clone_flags & CLONE_VM)) { + if (vx_vmpages_avail(p->mm, p->mm->total_vm)) + vx_pages_add(p->vx_info, RLIMIT_AS, p->mm->total_vm); + else + goto bad_fork_free; + } + if (p->mm && vx_flags(VXF_FORK_RSS, 0)) { + if (!vx_rsspages_avail(p->mm, get_mm_counter(p->mm, file_rss))) + goto bad_fork_cleanup_vm; + } + retval = -EAGAIN; + if (!vx_nproc_avail(1)) + goto bad_fork_cleanup_vm; + + gr_learn_resource(p, RLIMIT_NPROC, atomic_read(&p->user->processes), 0); + if (atomic_read(&p->user->processes) >= p->signal->rlim[RLIMIT_NPROC].rlim_cur) { if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) && p->user != &root_user) - goto bad_fork_free; + goto bad_fork_cleanup_vm; } atomic_inc(&p->user->__count); @@ -1066,6 +1101,8 @@ static task_t *copy_process(unsigned lon if (retval) goto bad_fork_cleanup_namespace; + gr_copy_label(p); + p->set_child_tid = (clone_flags & CLONE_CHILD_SETTID) ? child_tidptr : NULL; /* * Clear TID on mm_release()? @@ -1212,6 +1249,18 @@ static task_t *copy_process(unsigned lon total_forks++; spin_unlock(&current->sighand->siglock); + + /* p is copy of current */ + vxi = p->vx_info; + if (vxi) { + claim_vx_info(vxi, p); + atomic_inc(&vxi->cvirt.nr_threads); + atomic_inc(&vxi->cvirt.total_forks); + vx_nproc_inc(p); + } + nxi = p->nx_info; + if (nxi) + claim_nx_info(nxi, p); write_unlock_irq(&tasklist_lock); proc_fork_connector(p); return p; @@ -1252,9 +1301,14 @@ bad_fork_cleanup_count: put_group_info(p->group_info); atomic_dec(&p->user->processes); free_uid(p->user); +bad_fork_cleanup_vm: + if (p->mm && !(clone_flags & CLONE_VM)) + vx_pages_sub(p->vx_info, RLIMIT_AS, p->mm->total_vm); bad_fork_free: free_task(p); fork_out: + gr_log_forkfail(retval); + return ERR_PTR(retval); } @@ -1328,6 +1382,8 @@ long do_fork(unsigned long clone_flags, if (!IS_ERR(p)) { struct completion vfork; + gr_handle_brute_check(); + if (clone_flags & CLONE_VFORK) { p->vfork_done = &vfork; init_completion(&vfork); diff -NurpP --minimal linux-2.6.17.13/kernel/futex.c linux-2.6.17.13-g-v/kernel/futex.c --- linux-2.6.17.13/kernel/futex.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/futex.c 2006-09-20 15:00:12.000000000 +0200 @@ -44,6 +44,7 @@ #include #include #include +#include #include #define FUTEX_HASHBITS (CONFIG_BASE_SMALL ? 4 : 8) @@ -151,6 +152,11 @@ static int get_futex_key(unsigned long u struct page *page; int err; +#ifdef CONFIG_PAX_SEGMEXEC + if ((mm->pax_flags & MF_PAX_SEGMEXEC) && (uaddr >= SEGMEXEC_TASK_SIZE)) + return -EFAULT; +#endif + /* * The futex address must be "naturally" aligned. */ diff -NurpP --minimal linux-2.6.17.13/kernel/futex_compat.c linux-2.6.17.13-g-v/kernel/futex_compat.c --- linux-2.6.17.13/kernel/futex_compat.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/futex_compat.c 2006-09-20 15:00:12.000000000 +0200 @@ -9,6 +9,7 @@ #include #include #include +#include #include diff -NurpP --minimal linux-2.6.17.13/kernel/kallsyms.c linux-2.6.17.13-g-v/kernel/kallsyms.c --- linux-2.6.17.13/kernel/kallsyms.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/kallsyms.c 2006-09-20 15:00:12.000000000 +0200 @@ -301,7 +301,6 @@ static unsigned long get_ksymbol_core(st static void reset_iter(struct kallsym_iter *iter, loff_t new_pos) { - iter->name[0] = '\0'; iter->nameoff = get_symbol_offset(new_pos); iter->pos = new_pos; } @@ -380,7 +379,7 @@ static int kallsyms_open(struct inode *i struct kallsym_iter *iter; int ret; - iter = kmalloc(sizeof(*iter), GFP_KERNEL); + iter = kzalloc(sizeof(*iter), GFP_KERNEL); if (!iter) return -ENOMEM; reset_iter(iter, 0); @@ -411,7 +410,15 @@ static int __init kallsyms_init(void) { struct proc_dir_entry *entry; +#ifdef CONFIG_GRKERNSEC_PROC_ADD +#ifdef CONFIG_GRKERNSEC_PROC_USER + entry = create_proc_entry("kallsyms", S_IFREG | S_IRUSR, NULL); +#elif CONFIG_GRKERNSEC_PROC_USERGROUP + entry = create_proc_entry("kallsyms", S_IFREG | S_IRUSR | S_IRGRP, NULL); +#endif +#else entry = create_proc_entry("kallsyms", 0444, NULL); +#endif if (entry) entry->proc_fops = &kallsyms_operations; return 0; diff -NurpP --minimal linux-2.6.17.13/kernel/kprobes.c linux-2.6.17.13-g-v/kernel/kprobes.c --- linux-2.6.17.13/kernel/kprobes.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/kprobes.c 2006-09-20 15:00:12.000000000 +0200 @@ -106,7 +106,7 @@ kprobe_opcode_t __kprobes *get_insn_slot * kernel image and loaded module images reside. This is required * so x86_64 can correctly handle the %rip-relative fixups. */ - kip->insns = module_alloc(PAGE_SIZE); + kip->insns = module_alloc_exec(PAGE_SIZE); if (!kip->insns) { kfree(kip); return NULL; diff -NurpP --minimal linux-2.6.17.13/kernel/kthread.c linux-2.6.17.13-g-v/kernel/kthread.c --- linux-2.6.17.13/kernel/kthread.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/kthread.c 2006-09-20 15:00:12.000000000 +0200 @@ -116,7 +116,7 @@ static void keventd_create_kthread(void } else { wait_for_completion(&create->started); read_lock(&tasklist_lock); - create->result = find_task_by_pid(pid); + create->result = find_task_by_real_pid(pid); read_unlock(&tasklist_lock); } complete(&create->done); diff -NurpP --minimal linux-2.6.17.13/kernel/module.c linux-2.6.17.13-g-v/kernel/module.c --- linux-2.6.17.13/kernel/module.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/module.c 2006-09-20 15:00:12.000000000 +0200 @@ -40,10 +40,15 @@ #include #include #include +#include #include #include #include +#ifdef CONFIG_PAX_KERNEXEC +#include +#endif + #if 0 #define DEBUGP printk #else @@ -66,6 +71,8 @@ static LIST_HEAD(modules); static BLOCKING_NOTIFIER_HEAD(module_notify_list); +extern int gr_check_modstop(void); + int register_module_notifier(struct notifier_block * nb) { return blocking_notifier_chain_register(&module_notify_list, nb); @@ -594,6 +601,9 @@ sys_delete_module(const char __user *nam char name[MODULE_NAME_LEN]; int ret, forced = 0; + if (gr_check_modstop()) + return -EPERM; + if (!capable(CAP_SYS_MODULE)) return -EPERM; @@ -1058,13 +1068,15 @@ static void free_module(struct module *m module_unload_free(mod); /* This may be NULL, but that's OK */ - module_free(mod, mod->module_init); + module_free(mod, mod->module_init_rw); + module_free_exec(mod, mod->module_init_rx); kfree(mod->args); if (mod->percpu) percpu_modfree(mod->percpu); /* Finally, free the core (containing the module structure) */ - module_free(mod, mod->module_core); + module_free_exec(mod, mod->module_core_rx); + module_free(mod, mod->module_core_rw); } void *__symbol_get(const char *symbol) @@ -1221,11 +1233,14 @@ static void layout_sections(struct modul || strncmp(secstrings + s->sh_name, ".init", 5) == 0) continue; - s->sh_entsize = get_offset(&mod->core_size, s); + if ((s->sh_flags & SHF_WRITE) || !(s->sh_flags & SHF_ALLOC)) + s->sh_entsize = get_offset(&mod->core_size_rw, s); + else + s->sh_entsize = get_offset(&mod->core_size_rx, s); DEBUGP("\t%s\n", secstrings + s->sh_name); } if (m == 0) - mod->core_text_size = mod->core_size; + mod->core_size_rx = mod->core_size_rx; } DEBUGP("Init section allocation order:\n"); @@ -1239,12 +1254,15 @@ static void layout_sections(struct modul || strncmp(secstrings + s->sh_name, ".init", 5) != 0) continue; - s->sh_entsize = (get_offset(&mod->init_size, s) - | INIT_OFFSET_MASK); + if ((s->sh_flags & SHF_WRITE) || !(s->sh_flags & SHF_ALLOC)) + s->sh_entsize = get_offset(&mod->init_size_rw, s); + else + s->sh_entsize = get_offset(&mod->init_size_rx, s); + s->sh_entsize |= INIT_OFFSET_MASK; DEBUGP("\t%s\n", secstrings + s->sh_name); } if (m == 0) - mod->init_text_size = mod->init_size; + mod->init_size_rx = mod->init_size_rx; } } @@ -1419,6 +1437,10 @@ static struct module *load_module(void _ struct exception_table_entry *extable; mm_segment_t old_fs; +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; +#endif + DEBUGP("load_module: umod=%p, len=%lu, uargs=%p\n", umod, len, uargs); if (len < sizeof(*hdr)) @@ -1568,21 +1590,57 @@ static struct module *load_module(void _ layout_sections(mod, hdr, sechdrs, secstrings); /* Do the allocs. */ - ptr = module_alloc(mod->core_size); + ptr = module_alloc(mod->core_size_rw); if (!ptr) { err = -ENOMEM; goto free_percpu; } - memset(ptr, 0, mod->core_size); - mod->module_core = ptr; + memset(ptr, 0, mod->core_size_rw); + mod->module_core_rw = ptr; - ptr = module_alloc(mod->init_size); - if (!ptr && mod->init_size) { + ptr = module_alloc(mod->init_size_rw); + if (!ptr && mod->init_size_rw) { err = -ENOMEM; - goto free_core; + goto free_core_rw; } - memset(ptr, 0, mod->init_size); - mod->module_init = ptr; + memset(ptr, 0, mod->init_size_rw); + mod->module_init_rw = ptr; + + ptr = module_alloc_exec(mod->core_size_rx); + if (!ptr) { + err = -ENOMEM; + goto free_init_rw; + } + +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + + memset(ptr, 0, mod->core_size_rx); + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + + mod->module_core_rx = ptr; + + ptr = module_alloc_exec(mod->init_size_rx); + if (!ptr && mod->init_size_rx) { + err = -ENOMEM; + goto free_core_rx; + } + +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + + memset(ptr, 0, mod->init_size_rx); + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + + mod->module_init_rx = ptr; /* Transfer each section which specifies SHF_ALLOC */ DEBUGP("final section addresses:\n"); @@ -1592,17 +1650,44 @@ static struct module *load_module(void _ if (!(sechdrs[i].sh_flags & SHF_ALLOC)) continue; - if (sechdrs[i].sh_entsize & INIT_OFFSET_MASK) - dest = mod->module_init - + (sechdrs[i].sh_entsize & ~INIT_OFFSET_MASK); - else - dest = mod->module_core + sechdrs[i].sh_entsize; + if (sechdrs[i].sh_entsize & INIT_OFFSET_MASK) { + if ((sechdrs[i].sh_flags & SHF_WRITE) || !(sechdrs[i].sh_flags & SHF_ALLOC)) + dest = mod->module_init_rw + + (sechdrs[i].sh_entsize & ~INIT_OFFSET_MASK); + else + dest = mod->module_init_rx + + (sechdrs[i].sh_entsize & ~INIT_OFFSET_MASK); + } else { + if ((sechdrs[i].sh_flags & SHF_WRITE) || !(sechdrs[i].sh_flags & SHF_ALLOC)) + dest = mod->module_core_rw + sechdrs[i].sh_entsize; + else + dest = mod->module_core_rx + sechdrs[i].sh_entsize; + } - if (sechdrs[i].sh_type != SHT_NOBITS) - memcpy(dest, (void *)sechdrs[i].sh_addr, - sechdrs[i].sh_size); + if (sechdrs[i].sh_type != SHT_NOBITS) { + +#ifdef CONFIG_PAX_KERNEXEC + if (!(sechdrs[i].sh_flags & SHF_WRITE) && (sechdrs[i].sh_flags & SHF_ALLOC)) + pax_open_kernel(cr0); +#endif + + memcpy(dest, (void *)sechdrs[i].sh_addr, sechdrs[i].sh_size); + +#ifdef CONFIG_PAX_KERNEXEC + if (!(sechdrs[i].sh_flags & SHF_WRITE) && (sechdrs[i].sh_flags & SHF_ALLOC)) + pax_close_kernel(cr0); +#endif + + } /* Update sh_addr to point to copy in image. */ - sechdrs[i].sh_addr = (unsigned long)dest; + +#ifdef CONFIG_PAX_KERNEXEC + if (sechdrs[i].sh_flags & SHF_EXECINSTR) + sechdrs[i].sh_addr = (unsigned long)dest - __KERNEL_TEXT_OFFSET; + else +#endif + + sechdrs[i].sh_addr = (unsigned long)dest; DEBUGP("\t0x%lx %s\n", sechdrs[i].sh_addr, secstrings + sechdrs[i].sh_name); } /* Module has been moved. */ @@ -1623,8 +1708,18 @@ static struct module *load_module(void _ setup_modinfo(mod, sechdrs, infoindex); /* Fix up syms, so that st_value is a pointer to location. */ + +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + err = simplify_symbols(sechdrs, symindex, strtab, versindex, pcpuindex, mod); + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + if (err < 0) goto cleanup; @@ -1666,11 +1761,20 @@ static struct module *load_module(void _ if (!(sechdrs[info].sh_flags & SHF_ALLOC)) continue; +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + if (sechdrs[i].sh_type == SHT_REL) err = apply_relocate(sechdrs, strtab, symindex, i,mod); else if (sechdrs[i].sh_type == SHT_RELA) err = apply_relocate_add(sechdrs, strtab, symindex, i, mod); + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + if (err < 0) goto cleanup; } @@ -1684,14 +1788,31 @@ static struct module *load_module(void _ /* Set up and sort exception table */ mod->num_exentries = sechdrs[exindex].sh_size / sizeof(*mod->extable); mod->extable = extable = (void *)sechdrs[exindex].sh_addr; + +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + sort_extable(extable, extable + mod->num_exentries); +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + /* Finally, copy percpu area over. */ percpu_modcopy(mod->percpu, (void *)sechdrs[pcpuindex].sh_addr, sechdrs[pcpuindex].sh_size); +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + add_kallsyms(mod, sechdrs, symindex, strindex, secstrings); +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + err = module_finalize(hdr, sechdrs, mod); if (err < 0) goto cleanup; @@ -1705,12 +1826,12 @@ static struct module *load_module(void _ * Do it before processing of module parameters, so the module * can provide parameter accessor functions of its own. */ - if (mod->module_init) - flush_icache_range((unsigned long)mod->module_init, - (unsigned long)mod->module_init - + mod->init_size); - flush_icache_range((unsigned long)mod->module_core, - (unsigned long)mod->module_core + mod->core_size); + if (mod->module_init_rx) + flush_icache_range((unsigned long)mod->module_init_rx, + (unsigned long)mod->module_init_rx + + mod->init_size_rx); + flush_icache_range((unsigned long)mod->module_core_rx, + (unsigned long)mod->module_core_rx + mod->core_size_rx); set_fs(old_fs); @@ -1748,9 +1869,13 @@ static struct module *load_module(void _ module_arch_cleanup(mod); cleanup: module_unload_free(mod); - module_free(mod, mod->module_init); - free_core: - module_free(mod, mod->module_core); + module_free_exec(mod, mod->module_init_rx); + free_core_rx: + module_free_exec(mod, mod->module_core_rx); + free_init_rw: + module_free(mod, mod->module_init_rw); + free_core_rw: + module_free(mod, mod->module_core_rw); free_percpu: if (percpu) percpu_modfree(percpu); @@ -1786,6 +1911,9 @@ sys_init_module(void __user *umod, struct module *mod; int ret = 0; + if (gr_check_modstop()) + return -EPERM; + /* Must have permission */ if (!capable(CAP_SYS_MODULE)) return -EPERM; @@ -1836,10 +1964,12 @@ sys_init_module(void __user *umod, mod->state = MODULE_STATE_LIVE; /* Drop initial reference. */ module_put(mod); - module_free(mod, mod->module_init); - mod->module_init = NULL; - mod->init_size = 0; - mod->init_text_size = 0; + module_free(mod, mod->module_init_rw); + module_free_exec(mod, mod->module_init_rx); + mod->module_init_rw = NULL; + mod->module_init_rx = NULL; + mod->init_size_rw = 0; + mod->init_size_rx = 0; mutex_unlock(&module_mutex); return 0; @@ -1870,10 +2000,14 @@ static const char *get_ksymbol(struct mo unsigned long nextval; /* At worse, next value is at end of module */ - if (within(addr, mod->module_init, mod->init_size)) - nextval = (unsigned long)mod->module_init+mod->init_text_size; - else - nextval = (unsigned long)mod->module_core+mod->core_text_size; + if (within(addr, mod->module_init_rx, mod->init_size_rx)) + nextval = (unsigned long)mod->module_init_rw; + else if (within(addr, mod->module_init_rw, mod->init_size_rw)) + nextval = (unsigned long)mod->module_core_rx; + else if (within(addr, mod->module_core_rx, mod->core_size_rx)) + nextval = (unsigned long)mod->module_core_rw; + else + nextval = (unsigned long)mod->module_core_rw+mod->core_size_rw; /* Scan for closest preceeding symbol, and next symbol. (ELF starts real symbols at 1). */ @@ -1914,8 +2048,10 @@ const char *module_address_lookup(unsign struct module *mod; list_for_each_entry(mod, &modules, list) { - if (within(addr, mod->module_init, mod->init_size) - || within(addr, mod->module_core, mod->core_size)) { + if (within(addr, mod->module_init_rx, mod->init_size_rx) + || within(addr, mod->module_init_rw, mod->init_size_rw) + || within(addr, mod->module_core_rx, mod->core_size_rx) + || within(addr, mod->module_core_rw, mod->core_size_rw)) { *modname = mod->name; return get_ksymbol(mod, addr, size, offset); } @@ -1926,7 +2062,7 @@ const char *module_address_lookup(unsign struct module *module_get_kallsym(unsigned int symnum, unsigned long *value, char *type, - char namebuf[128]) + char namebuf[KSYM_NAME_LEN+1]) { struct module *mod; @@ -1937,7 +2073,7 @@ struct module *module_get_kallsym(unsign *type = mod->symtab[symnum].st_info; strncpy(namebuf, mod->strtab + mod->symtab[symnum].st_name, - 127); + KSYM_NAME_LEN); mutex_unlock(&module_mutex); return mod; } @@ -2014,7 +2150,7 @@ static int m_show(struct seq_file *m, vo { struct module *mod = list_entry(p, struct module, list); seq_printf(m, "%s %lu", - mod->name, mod->init_size + mod->core_size); + mod->name, mod->init_size_rx + mod->init_size_rw + mod->core_size_rx + mod->core_size_rw); print_unload_info(m, mod); /* Informative for users. */ @@ -2023,7 +2159,7 @@ static int m_show(struct seq_file *m, vo mod->state == MODULE_STATE_COMING ? "Loading": "Live"); /* Used by oprofile and other similar tools. */ - seq_printf(m, " 0x%p", mod->module_core); + seq_printf(m, " 0x%p 0x%p", mod->module_core_rx, mod->module_core_rw); seq_printf(m, "\n"); return 0; @@ -2072,8 +2208,8 @@ struct module *__module_text_address(uns struct module *mod; list_for_each_entry(mod, &modules, list) - if (within(addr, mod->module_init, mod->init_text_size) - || within(addr, mod->module_core, mod->core_text_size)) + if (within(addr, mod->module_init_rx, mod->init_size_rx) + || within(addr, mod->module_core_rx, mod->core_size_rx)) return mod; return NULL; } diff -NurpP --minimal linux-2.6.17.13/kernel/pid.c linux-2.6.17.13-g-v/kernel/pid.c --- linux-2.6.17.13/kernel/pid.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/pid.c 2006-09-20 15:00:12.000000000 +0200 @@ -26,6 +26,7 @@ #include #include #include +#include #define pid_hashfn(nr) hash_long((unsigned long)nr, pidhash_shift) static struct hlist_head *pid_hash; @@ -90,7 +91,9 @@ static int alloc_pidmap(void) int i, offset, max_scan, pid, last = last_pid; pidmap_t *map; - pid = last + 1; + pid = gr_random_pid(); + if (!pid) + pid = last_pid + 1; if (pid >= pid_max) pid = RESERVED_PIDS; offset = pid & BITS_PER_PAGE_MASK; @@ -269,7 +272,14 @@ struct task_struct * fastcall pid_task(s */ task_t *find_task_by_pid_type(int type, int nr) { - return pid_task(find_pid(nr), type); + task_t *task; + + task = pid_task(find_pid(nr), type); + + if (gr_pid_is_chrooted(task)) + return NULL; + + return task; } EXPORT_SYMBOL(find_task_by_pid_type); diff -NurpP --minimal linux-2.6.17.13/kernel/posix-cpu-timers.c linux-2.6.17.13-g-v/kernel/posix-cpu-timers.c --- linux-2.6.17.13/kernel/posix-cpu-timers.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/posix-cpu-timers.c 2006-09-20 15:00:12.000000000 +0200 @@ -6,6 +6,8 @@ #include #include #include +#include +#include static int check_clock(const clockid_t which_clock) { @@ -1125,6 +1127,7 @@ static void check_process_timers(struct __group_send_sig_info(SIGKILL, SEND_SIG_PRIV, tsk); return; } + gr_learn_resource(tsk, RLIMIT_CPU, psecs, 1); if (psecs >= sig->rlim[RLIMIT_CPU].rlim_cur) { /* * At the soft limit, send a SIGXCPU every second. diff -NurpP --minimal linux-2.6.17.13/kernel/posix-timers.c linux-2.6.17.13-g-v/kernel/posix-timers.c --- linux-2.6.17.13/kernel/posix-timers.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/posix-timers.c 2006-09-20 15:00:12.000000000 +0200 @@ -372,7 +372,7 @@ static struct task_struct * good_sigeven struct task_struct *rtn = current->group_leader; if ((event->sigev_notify & SIGEV_THREAD_ID ) && - (!(rtn = find_task_by_pid(event->sigev_notify_thread_id)) || + (!(rtn = find_task_by_real_pid(event->sigev_notify_thread_id)) || rtn->tgid != current->tgid || (event->sigev_notify & ~SIGEV_THREAD_ID) != SIGEV_SIGNAL)) return NULL; diff -NurpP --minimal linux-2.6.17.13/kernel/printk.c linux-2.6.17.13-g-v/kernel/printk.c --- linux-2.6.17.13/kernel/printk.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/printk.c 2006-09-20 15:00:12.000000000 +0200 @@ -31,6 +31,8 @@ #include #include #include +#include +#include #include @@ -183,18 +185,18 @@ int do_syslog(int type, char __user *buf unsigned long i, j, limit, count; int do_clear = 0; char c; - int error = 0; + int error; + +#ifdef CONFIG_GRKERNSEC_DMESG + if (grsec_enable_dmesg && !capable(CAP_SYS_ADMIN)) + return -EPERM; +#endif error = security_syslog(type); if (error) return error; - switch (type) { - case 0: /* Close log */ - break; - case 1: /* Open log */ - break; - case 2: /* Read from log */ + if ((type >= 2) && (type <= 4)) { error = -EINVAL; if (!buf || len < 0) goto out; @@ -205,6 +207,16 @@ int do_syslog(int type, char __user *buf error = -EFAULT; goto out; } + } + if (!vx_check(0, VX_ADMIN|VX_WATCH)) + return vx_do_syslog(type, buf, len); + + switch (type) { + case 0: /* Close log */ + break; + case 1: /* Open log */ + break; + case 2: /* Read from log */ error = wait_event_interruptible(log_wait, (log_start - log_end)); if (error) @@ -229,16 +241,6 @@ int do_syslog(int type, char __user *buf do_clear = 1; /* FALL THRU */ case 3: /* Read last kernel messages */ - error = -EINVAL; - if (!buf || len < 0) - goto out; - error = 0; - if (!len) - goto out; - if (!access_ok(VERIFY_WRITE, buf, len)) { - error = -EFAULT; - goto out; - } count = len; if (count > log_buf_len) count = log_buf_len; diff -NurpP --minimal linux-2.6.17.13/kernel/ptrace.c linux-2.6.17.13-g-v/kernel/ptrace.c --- linux-2.6.17.13/kernel/ptrace.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/ptrace.c 2006-09-20 15:00:12.000000000 +0200 @@ -18,6 +18,8 @@ #include #include #include +#include +#include #include #include @@ -127,10 +129,10 @@ static int may_attach(struct task_struct (current->uid != task->uid) || (current->gid != task->egid) || (current->gid != task->sgid) || - (current->gid != task->gid)) && !capable(CAP_SYS_PTRACE)) + (current->gid != task->gid)) && !capable_nolog(CAP_SYS_PTRACE)) return -EPERM; smp_rmb(); - if (!task->mm->dumpable && !capable(CAP_SYS_PTRACE)) + if (!task->mm->dumpable && !capable_nolog(CAP_SYS_PTRACE)) return -EPERM; return security_ptrace(current, task); @@ -506,6 +508,10 @@ asmlinkage long sys_ptrace(long request, goto out; } + ret = -EPERM; + if (!vx_check(vx_task_xid(child), VX_WATCH|VX_IDENT)) + goto out_put_task_struct; + if (request == PTRACE_ATTACH) { ret = ptrace_attach(child); goto out_put_task_struct; @@ -515,6 +521,11 @@ asmlinkage long sys_ptrace(long request, if (ret < 0) goto out_put_task_struct; + if (gr_handle_ptrace(child, request)) { + ret = -EPERM; + goto out_put_task_struct; + } + ret = arch_ptrace(child, request, addr, data); if (ret < 0) goto out_put_task_struct; diff -NurpP --minimal linux-2.6.17.13/kernel/resource.c linux-2.6.17.13-g-v/kernel/resource.c --- linux-2.6.17.13/kernel/resource.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/resource.c 2006-09-20 15:00:12.000000000 +0200 @@ -136,10 +136,27 @@ static int __init ioresources_init(void) { struct proc_dir_entry *entry; +#ifdef CONFIG_GRKERNSEC_PROC_ADD +#ifdef CONFIG_GRKERNSEC_PROC_USER + entry = create_proc_entry("ioports", S_IRUSR, NULL); +#elif CONFIG_GRKERNSEC_PROC_USERGROUP + entry = create_proc_entry("ioports", S_IRUSR | S_IRGRP, NULL); +#endif +#else entry = create_proc_entry("ioports", 0, NULL); +#endif if (entry) entry->proc_fops = &proc_ioports_operations; + +#ifdef CONFIG_GRKERNSEC_PROC_ADD +#ifdef CONFIG_GRKERNSEC_PROC_USER + entry = create_proc_entry("iomem", S_IRUSR, NULL); +#elif CONFIG_GRKERNSEC_PROC_USERGROUP + entry = create_proc_entry("iomem", S_IRUSR | S_IRGRP, NULL); +#endif +#else entry = create_proc_entry("iomem", 0, NULL); +#endif if (entry) entry->proc_fops = &proc_iomem_operations; return 0; diff -NurpP --minimal linux-2.6.17.13/kernel/sched.c linux-2.6.17.13-g-v/kernel/sched.c --- linux-2.6.17.13/kernel/sched.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/sched.c 2006-09-20 15:00:12.000000000 +0200 @@ -50,9 +50,13 @@ #include #include #include +#include #include #include +#include +#include +#include /* * Convert user-nice values [ -20 ... 0 ... 19 ] @@ -241,6 +245,10 @@ struct runqueue { struct list_head migration_queue; int cpu; #endif +#ifdef CONFIG_VSERVER_HARDCPU + struct list_head hold_queue; + int idle_tokens; +#endif #ifdef CONFIG_SCHEDSTATS /* latency stats */ @@ -601,6 +609,7 @@ static inline void sched_info_switch(tas */ static void dequeue_task(struct task_struct *p, prio_array_t *array) { + BUG_ON(p->state & TASK_ONHOLD); array->nr_active--; list_del(&p->run_list); if (list_empty(array->queue + p->prio)) @@ -609,6 +618,7 @@ static void dequeue_task(struct task_str static void enqueue_task(struct task_struct *p, prio_array_t *array) { + BUG_ON(p->state & TASK_ONHOLD); sched_info_queued(p); list_add_tail(&p->run_list, array->queue + p->prio); __set_bit(p->prio, array->bitmap); @@ -622,11 +632,13 @@ static void enqueue_task(struct task_str */ static void requeue_task(struct task_struct *p, prio_array_t *array) { + BUG_ON(p->state & TASK_ONHOLD); list_move_tail(&p->run_list, array->queue + p->prio); } static inline void enqueue_task_head(struct task_struct *p, prio_array_t *array) { + BUG_ON(p->state & TASK_ONHOLD); list_add(&p->run_list, array->queue + p->prio); __set_bit(p->prio, array->bitmap); array->nr_active++; @@ -650,6 +662,7 @@ static inline void enqueue_task_head(str static int effective_prio(task_t *p) { int bonus, prio; + struct vx_info *vxi; if (rt_task(p)) return p->prio; @@ -657,6 +670,11 @@ static int effective_prio(task_t *p) bonus = CURRENT_BONUS(p) - MAX_BONUS / 2; prio = p->static_prio - bonus; + + if ((vxi = p->vx_info) && + vx_info_flags(vxi, VXF_SCHED_PRIO, 0)) + prio += vx_effective_vavavoom(vxi, MAX_USER_PRIO); + if (prio < MAX_RT_PRIO) prio = MAX_RT_PRIO; if (prio > MAX_PRIO-1) @@ -797,19 +815,77 @@ static void activate_task(task_t *p, run } p->timestamp = now; + vx_activate_task(p); __activate_task(p, rq); } /* * deactivate_task - remove a task from the runqueue. */ -static void deactivate_task(struct task_struct *p, runqueue_t *rq) +static void __deactivate_task(struct task_struct *p, runqueue_t *rq) { rq->nr_running--; dequeue_task(p, p->array); p->array = NULL; } +static inline +void deactivate_task(struct task_struct *p, runqueue_t *rq) +{ + vx_deactivate_task(p); + __deactivate_task(p, rq); +} + + +#ifdef CONFIG_VSERVER_HARDCPU +/* + * vx_hold_task - put a task on the hold queue + */ +static inline +void vx_hold_task(struct vx_info *vxi, + struct task_struct *p, runqueue_t *rq) +{ + __deactivate_task(p, rq); + p->state |= TASK_ONHOLD; + /* a new one on hold */ + vx_onhold_inc(vxi); + list_add_tail(&p->run_list, &rq->hold_queue); +} + +/* + * vx_unhold_task - put a task back to the runqueue + */ +static inline +void vx_unhold_task(struct vx_info *vxi, + struct task_struct *p, runqueue_t *rq) +{ + list_del(&p->run_list); + /* one less waiting */ + vx_onhold_dec(vxi); + p->state &= ~TASK_ONHOLD; + enqueue_task(p, rq->expired); + rq->nr_running++; + + if (p->static_prio < rq->best_expired_prio) + rq->best_expired_prio = p->static_prio; +} +#else +static inline +void vx_hold_task(struct vx_info *vxi, + struct task_struct *p, runqueue_t *rq) +{ + return; +} + +static inline +void vx_unhold_task(struct vx_info *vxi, + struct task_struct *p, runqueue_t *rq) +{ + return; +} +#endif /* CONFIG_VSERVER_HARDCPU */ + + /* * resched_task - mark a task 'to be rescheduled now'. * @@ -1173,6 +1249,12 @@ static int try_to_wake_up(task_t *p, uns rq = task_rq_lock(p, &flags); old_state = p->state; + + /* we need to unhold suspended tasks */ + if (old_state & TASK_ONHOLD) { + vx_unhold_task(p->vx_info, p, rq); + old_state = p->state; + } if (!(old_state & state)) goto out; @@ -1276,6 +1358,7 @@ out_activate: #endif /* CONFIG_SMP */ if (old_state == TASK_UNINTERRUPTIBLE) { rq->nr_uninterruptible--; + vx_uninterruptible_dec(p); /* * Tasks on involuntary sleep don't earn * sleep_avg beyond just interactive state. @@ -1416,6 +1499,7 @@ void fastcall wake_up_new_task(task_t *p p->prio = effective_prio(p); + vx_activate_task(p); if (likely(cpu == this_cpu)) { if (!(clone_flags & CLONE_VM)) { /* @@ -1427,6 +1511,7 @@ void fastcall wake_up_new_task(task_t *p __activate_task(p, rq); else { p->prio = current->prio; + BUG_ON(p->state & TASK_ONHOLD); list_add_tail(&p->run_list, &current->run_list); p->array = current->array; p->array->nr_active++; @@ -2514,13 +2599,16 @@ unsigned long long current_sched_time(co void account_user_time(struct task_struct *p, cputime_t cputime) { struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat; + struct vx_info *vxi = p->vx_info; /* p is _always_ current */ cputime64_t tmp; + int nice = (TASK_NICE(p) > 0); p->utime = cputime_add(p->utime, cputime); + vx_account_user(vxi, cputime, nice); /* Add user time to cpustat. */ tmp = cputime_to_cputime64(cputime); - if (TASK_NICE(p) > 0) + if (nice) cpustat->nice = cputime64_add(cpustat->nice, tmp); else cpustat->user = cputime64_add(cpustat->user, tmp); @@ -2536,10 +2624,12 @@ void account_system_time(struct task_str cputime_t cputime) { struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat; + struct vx_info *vxi = p->vx_info; /* p is _always_ current */ runqueue_t *rq = this_rq(); cputime64_t tmp; p->stime = cputime_add(p->stime, cputime); + vx_account_system(vxi, cputime, (p == rq->idle)); /* Add system time to cpustat. */ tmp = cputime_to_cputime64(cputime); @@ -2599,6 +2689,10 @@ void scheduler_tick(void) if (p == rq->idle) { if (wake_priority_sleeper(rq)) goto out; +#ifdef CONFIG_VSERVER_HARDCPU_IDLE + if (!--rq->idle_tokens && !list_empty(&rq->hold_queue)) + set_need_resched(); +#endif rebalance_tick(cpu, rq, SCHED_IDLE); return; } @@ -2631,7 +2725,7 @@ void scheduler_tick(void) } goto out_unlock; } - if (!--p->time_slice) { + if (vx_need_resched(p)) { dequeue_task(p, rq->active); set_tsk_need_resched(p); p->prio = effective_prio(p); @@ -2902,6 +2996,10 @@ asmlinkage void __sched schedule(void) unsigned long long now; unsigned long run_time; int cpu, idx, new_prio; + struct vx_info *vxi; +#ifdef CONFIG_VSERVER_HARDCPU + int maxidle = -HZ; +#endif /* * Test if we are atomic. Since do_exit() needs to call into @@ -2959,12 +3057,41 @@ need_resched_nonpreemptible: unlikely(signal_pending(prev)))) prev->state = TASK_RUNNING; else { - if (prev->state == TASK_UNINTERRUPTIBLE) + if (prev->state == TASK_UNINTERRUPTIBLE) { rq->nr_uninterruptible++; + vx_uninterruptible_inc(prev); + } deactivate_task(prev, rq); } } +#ifdef CONFIG_VSERVER_HARDCPU + if (!list_empty(&rq->hold_queue)) { + struct list_head *l, *n; + int ret; + + vxi = NULL; + list_for_each_safe(l, n, &rq->hold_queue) { + next = list_entry(l, task_t, run_list); + if (vxi == next->vx_info) + continue; + + vxi = next->vx_info; + ret = vx_tokens_recalc(vxi); + + if (ret > 0) { + vx_unhold_task(vxi, next, rq); + break; + } + if ((ret < 0) && (maxidle < ret)) + maxidle = ret; + } + } + rq->idle_tokens = -maxidle; + +pick_next: +#endif + cpu = smp_processor_id(); if (unlikely(!rq->nr_running)) { go_idle: @@ -3012,6 +3139,22 @@ go_idle: queue = array->queue + idx; next = list_entry(queue->next, task_t, run_list); + vxi = next->vx_info; +#ifdef CONFIG_VSERVER_HARDCPU + if (vx_info_flags(vxi, VXF_SCHED_PAUSE|VXF_SCHED_HARD, 0)) { + int ret = vx_tokens_recalc(vxi); + + if (unlikely(ret <= 0)) { + if (ret && (rq->idle_tokens > -ret)) + rq->idle_tokens = -ret; + vx_hold_task(vxi, next, rq); + goto pick_next; + } + } else /* well, looks ugly but not as ugly as the ifdef-ed version */ +#endif + if (vx_info_flags(vxi, VXF_SCHED_PRIO, 0)) + vx_tokens_recalc(vxi); + if (!rt_task(next) && interactive_sleep(next->sleep_type)) { unsigned long long delta = now - next->timestamp; if (unlikely((long long)(now - next->timestamp) < 0)) @@ -3565,8 +3708,9 @@ asmlinkage long sys_nice(int increment) if (nice > 19) nice = 19; - if (increment < 0 && !can_nice(current, nice)) - return -EPERM; + if (increment < 0 && (!can_nice(current, nice) || + gr_handle_chroot_nice())) + return vx_flags(VXF_IGNEG_NICE, 0) ? 0 : -EPERM; retval = security_task_setnice(current, nice); if (retval) @@ -3726,6 +3870,7 @@ recheck: oldprio = p->prio; __setscheduler(p, policy, param->sched_priority); if (array) { + vx_activate_task(p); __activate_task(p, rq); /* * Reschedule if we are currently running on this runqueue and @@ -6130,6 +6275,9 @@ void __init sched_init(void) rq->cpu = i; #endif atomic_set(&rq->nr_iowait, 0); +#ifdef CONFIG_VSERVER_HARDCPU + INIT_LIST_HEAD(&rq->hold_queue); +#endif for (j = 0; j < 2; j++) { array = rq->arrays + j; @@ -6199,6 +6347,7 @@ void normalize_rt_tasks(void) deactivate_task(p, task_rq(p)); __setscheduler(p, SCHED_NORMAL, 0); if (array) { + vx_activate_task(p); __activate_task(p, task_rq(p)); resched_task(rq->curr); } diff -NurpP --minimal linux-2.6.17.13/kernel/signal.c linux-2.6.17.13-g-v/kernel/signal.c --- linux-2.6.17.13/kernel/signal.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/signal.c 2006-09-20 15:00:12.000000000 +0200 @@ -25,6 +25,8 @@ #include #include #include +#include +#include #include #include #include @@ -572,21 +574,32 @@ static int rm_from_queue(unsigned long m static int check_kill_permission(int sig, struct siginfo *info, struct task_struct *t) { + int user; int error = -EINVAL; + if (!valid_signal(sig)) return error; + user = ((info == SEND_SIG_NOINFO) || + (!is_si_special(info) && SI_FROMUSER(info))); + error = -EPERM; - if ((info == SEND_SIG_NOINFO || (!is_si_special(info) && SI_FROMUSER(info))) - && ((sig != SIGCONT) || + + if (user && ((((sig != SIGCONT) || (current->signal->session != t->signal->session)) && (current->euid ^ t->suid) && (current->euid ^ t->uid) && (current->uid ^ t->suid) && (current->uid ^ t->uid) - && !capable(CAP_KILL)) + && !capable(CAP_KILL)) || gr_handle_signal(t, sig))) + return error; + + error = -ESRCH; + if (user && !vx_check(vx_task_xid(t), VX_ADMIN|VX_IDENT)) return error; error = security_task_kill(t, info, sig); - if (!error) + if (!error) { audit_signal_info(sig, t); /* Let audit system see the signal */ + gr_log_signal(sig, t); + } return error; } @@ -764,7 +777,7 @@ out_set: (((sig) < SIGRTMIN) && sigismember(&(sigptr)->signal, (sig))) -static int +int specific_send_sig_info(int sig, struct siginfo *info, struct task_struct *t) { int ret = 0; @@ -809,6 +822,10 @@ force_sig_info(int sig, struct siginfo * } recalc_sigpending_tsk(t); ret = specific_send_sig_info(sig, info, t); + + gr_log_signal(sig, t); + gr_handle_crash(t, sig); + spin_unlock_irqrestore(&t->sighand->siglock, flags); return ret; @@ -1803,6 +1820,11 @@ relock: if (current == child_reaper) continue; + /* virtual init is protected against user signals */ + if ((info->si_code == SI_USER) && + vx_current_initpid(current->pid)) + continue; + if (sig_kernel_stop(signr)) { /* * The default action is to stop all threads in diff -NurpP --minimal linux-2.6.17.13/kernel/sys.c linux-2.6.17.13-g-v/kernel/sys.c --- linux-2.6.17.13/kernel/sys.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/sys.c 2006-09-20 15:00:12.000000000 +0200 @@ -11,6 +11,7 @@ #include #include #include +#include #include #include #include @@ -30,6 +31,8 @@ #include #include #include +#include +#include #include #include @@ -439,9 +442,18 @@ static int set_one_prio(struct task_stru goto out; } if (niceval < task_nice(p) && !can_nice(p, niceval)) { + if (vx_flags(VXF_IGNEG_NICE, 0)) + error = 0; + else + error = -EACCES; + goto out; + } + + if (gr_handle_chroot_setpriority(p, niceval)) { error = -EACCES; goto out; } + no_nice = security_task_setnice(p, niceval); if (no_nice) { error = no_nice; @@ -491,7 +503,8 @@ asmlinkage long sys_setpriority(int whic if (!who) who = current->uid; else - if ((who != current->uid) && !(user = find_user(who))) + if ((who != current->uid) && + !(user = find_user(vx_current_xid(), who))) goto out_unlock; /* No processes for this user */ do_each_thread(g, p) @@ -549,7 +562,8 @@ asmlinkage long sys_getpriority(int whic if (!who) who = current->uid; else - if ((who != current->uid) && !(user = find_user(who))) + if ((who != current->uid) && + !(user = find_user(vx_current_xid(), who))) goto out_unlock; /* No processes for this user */ do_each_thread(g, p) @@ -666,6 +680,9 @@ void kernel_power_off(void) machine_power_off(); } EXPORT_SYMBOL_GPL(kernel_power_off); + +long vs_reboot(unsigned int, void __user *); + /* * Reboot system call: for obvious reasons only root may call it, * and even root needs to set up some magic numbers in the registers @@ -696,6 +713,9 @@ asmlinkage long sys_reboot(int magic1, i if ((cmd == LINUX_REBOOT_CMD_POWER_OFF) && !pm_power_off) cmd = LINUX_REBOOT_CMD_HALT; + if (!vx_check(0, VX_ADMIN|VX_WATCH)) + return vs_reboot(cmd, arg); + lock_kernel(); switch (cmd) { case LINUX_REBOOT_CMD_RESTART: @@ -831,6 +851,9 @@ asmlinkage long sys_setregid(gid_t rgid, if (rgid != (gid_t) -1 || (egid != (gid_t) -1 && egid != old_rgid)) current->sgid = new_egid; + + gr_set_role_label(current, current->uid, new_rgid); + current->fsgid = new_egid; current->egid = new_egid; current->gid = new_rgid; @@ -860,6 +883,9 @@ asmlinkage long sys_setgid(gid_t gid) current->mm->dumpable = suid_dumpable; smp_wmb(); } + + gr_set_role_label(current, current->uid, gid); + current->gid = current->egid = current->sgid = current->fsgid = gid; } else if ((gid == current->gid) || (gid == current->sgid)) @@ -883,7 +909,7 @@ static int set_user(uid_t new_ruid, int { struct user_struct *new_user; - new_user = alloc_uid(new_ruid); + new_user = alloc_uid(vx_current_xid(), new_ruid); if (!new_user) return -EAGAIN; @@ -901,6 +927,9 @@ static int set_user(uid_t new_ruid, int current->mm->dumpable = suid_dumpable; smp_wmb(); } + + gr_set_role_label(current, new_ruid, current->gid); + current->uid = new_ruid; return 0; } @@ -1004,6 +1033,9 @@ asmlinkage long sys_setuid(uid_t uid) } else if ((uid != current->uid) && (uid != new_suid)) return -EPERM; + if (gr_check_crash_uid(uid)) + return -EPERM; + if (old_euid != uid) { current->mm->dumpable = suid_dumpable; @@ -1109,8 +1141,10 @@ asmlinkage long sys_setresgid(gid_t rgid current->egid = egid; } current->fsgid = current->egid; - if (rgid != (gid_t) -1) + if (rgid != (gid_t) -1) { + gr_set_role_label(current, current->uid, rgid); current->gid = rgid; + } if (sgid != (gid_t) -1) current->sgid = sgid; @@ -1247,15 +1281,18 @@ asmlinkage long sys_setpgid(pid_t pid, p { struct task_struct *p; struct task_struct *group_leader = current->group_leader; + pid_t rpgid; int err = -EINVAL; if (!pid) - pid = group_leader->pid; + pid = vx_map_pid(group_leader->pid); if (!pgid) pgid = pid; if (pgid < 0) return -EINVAL; + rpgid = vx_rmap_pid(pgid); + /* From this point forward we keep holding onto the tasklist lock * so that our parent does not change from under us. -DaveM */ @@ -1290,22 +1327,22 @@ asmlinkage long sys_setpgid(pid_t pid, p if (pgid != pid) { struct task_struct *p; - do_each_task_pid(pgid, PIDTYPE_PGID, p) { + do_each_task_pid(rpgid, PIDTYPE_PGID, p) { if (p->signal->session == group_leader->signal->session) goto ok_pgid; - } while_each_task_pid(pgid, PIDTYPE_PGID, p); + } while_each_task_pid(rpgid, PIDTYPE_PGID, p); goto out; } ok_pgid: - err = security_task_setpgid(p, pgid); + err = security_task_setpgid(p, rpgid); if (err) goto out; - if (process_group(p) != pgid) { + if (process_group(p) != rpgid) { detach_pid(p, PIDTYPE_PGID); - p->signal->pgrp = pgid; - attach_pid(p, PIDTYPE_PGID, pgid); + p->signal->pgrp = rpgid; + attach_pid(p, PIDTYPE_PGID, rpgid); } err = 0; @@ -1318,7 +1355,7 @@ out: asmlinkage long sys_getpgid(pid_t pid) { if (!pid) { - return process_group(current); + return vx_rmap_pid(process_group(current)); } else { int retval; struct task_struct *p; @@ -1330,7 +1367,7 @@ asmlinkage long sys_getpgid(pid_t pid) if (p) { retval = security_task_getpgid(p); if (!retval) - retval = process_group(p); + retval = vx_rmap_pid(process_group(p)); } read_unlock(&tasklist_lock); return retval; @@ -1671,7 +1708,7 @@ asmlinkage long sys_newuname(struct new_ int errno = 0; down_read(&uts_sem); - if (copy_to_user(name,&system_utsname,sizeof *name)) + if (copy_to_user(name, vx_new_utsname(), sizeof *name)) errno = -EFAULT; up_read(&uts_sem); return errno; @@ -1682,15 +1719,17 @@ asmlinkage long sys_sethostname(char __u int errno; char tmp[__NEW_UTS_LEN]; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SET_UTSNAME)) return -EPERM; if (len < 0 || len > __NEW_UTS_LEN) return -EINVAL; down_write(&uts_sem); errno = -EFAULT; if (!copy_from_user(tmp, name, len)) { - memcpy(system_utsname.nodename, tmp, len); - system_utsname.nodename[len] = 0; + char *ptr = vx_new_uts(nodename); + + memcpy(ptr, tmp, len); + ptr[len] = 0; errno = 0; } up_write(&uts_sem); @@ -1702,15 +1741,17 @@ asmlinkage long sys_sethostname(char __u asmlinkage long sys_gethostname(char __user *name, int len) { int i, errno; + char *ptr; if (len < 0) return -EINVAL; down_read(&uts_sem); - i = 1 + strlen(system_utsname.nodename); + ptr = vx_new_uts(nodename); + i = 1 + strlen(ptr); if (i > len) i = len; errno = 0; - if (copy_to_user(name, system_utsname.nodename, i)) + if (copy_to_user(name, ptr, i)) errno = -EFAULT; up_read(&uts_sem); return errno; @@ -1727,7 +1768,7 @@ asmlinkage long sys_setdomainname(char _ int errno; char tmp[__NEW_UTS_LEN]; - if (!capable(CAP_SYS_ADMIN)) + if (!vx_capable(CAP_SYS_ADMIN, VXC_SET_UTSNAME)) return -EPERM; if (len < 0 || len > __NEW_UTS_LEN) return -EINVAL; @@ -1735,8 +1776,10 @@ asmlinkage long sys_setdomainname(char _ down_write(&uts_sem); errno = -EFAULT; if (!copy_from_user(tmp, name, len)) { - memcpy(system_utsname.domainname, tmp, len); - system_utsname.domainname[len] = 0; + char *ptr = vx_new_uts(domainname); + + memcpy(ptr, tmp, len); + ptr[len] = 0; errno = 0; } up_write(&uts_sem); @@ -1794,7 +1837,7 @@ asmlinkage long sys_setrlimit(unsigned i return -EINVAL; old_rlim = current->signal->rlim + resource; if ((new_rlim.rlim_max > old_rlim->rlim_max) && - !capable(CAP_SYS_RESOURCE)) + !vx_capable(CAP_SYS_RESOURCE, VXC_SET_RLIMIT)) return -EPERM; if (resource == RLIMIT_NOFILE && new_rlim.rlim_max > NR_OPEN) return -EPERM; diff -NurpP --minimal linux-2.6.17.13/kernel/sysctl.c linux-2.6.17.13-g-v/kernel/sysctl.c --- linux-2.6.17.13/kernel/sysctl.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/sysctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -46,6 +46,7 @@ #include #include #include +#include #include #include @@ -54,6 +55,14 @@ extern int proc_nr_files(ctl_table *tabl void __user *buffer, size_t *lenp, loff_t *ppos); #if defined(CONFIG_SYSCTL) +#include +#include + +extern __u32 gr_handle_sysctl(const ctl_table *table, const void *oldval, + const void *newval); +extern int gr_handle_sysctl_mod(const char *dirname, const char *name, + const int op); +extern int gr_handle_chroot_sysctl(const int op); /* External variables not in a header file. */ extern int C_A_D; @@ -89,6 +98,7 @@ static int ngroups_max = NGROUPS_MAX; #ifdef CONFIG_KMOD extern char modprobe_path[]; #endif +extern char vshelper_path[]; #ifdef CONFIG_CHR_DEV_SG extern int sg_big_buff; #endif @@ -157,6 +167,22 @@ extern ctl_table inotify_table[]; #ifdef HAVE_ARCH_PICK_MMAP_LAYOUT int sysctl_legacy_va_layout; #endif +extern ctl_table grsecurity_table[]; + +#ifdef CONFIG_PAX_SOFTMODE +static ctl_table pax_table[] = { + { + .ctl_name = PAX_SOFTMODE, + .procname = "softmode", + .data = &pax_softmode, + .maxlen = sizeof(unsigned int), + .mode = 0600, + .proc_handler = &proc_dointvec, + }, + + { .ctl_name = 0 } +}; +#endif /* /proc declarations: */ @@ -237,6 +263,7 @@ static ctl_table kern_table[] = { .maxlen = sizeof(system_utsname.sysname), .mode = 0444, .proc_handler = &proc_doutsstring, + .virt_handler = &vx_uts_virt_handler, .strategy = &sysctl_string, }, { @@ -246,6 +273,7 @@ static ctl_table kern_table[] = { .maxlen = sizeof(system_utsname.release), .mode = 0444, .proc_handler = &proc_doutsstring, + .virt_handler = &vx_uts_virt_handler, .strategy = &sysctl_string, }, { @@ -255,6 +283,7 @@ static ctl_table kern_table[] = { .maxlen = sizeof(system_utsname.version), .mode = 0444, .proc_handler = &proc_doutsstring, + .virt_handler = &vx_uts_virt_handler, .strategy = &sysctl_string, }, { @@ -264,6 +293,7 @@ static ctl_table kern_table[] = { .maxlen = sizeof(system_utsname.nodename), .mode = 0644, .proc_handler = &proc_doutsstring, + .virt_handler = &vx_uts_virt_handler, .strategy = &sysctl_string, }, { @@ -273,6 +303,7 @@ static ctl_table kern_table[] = { .maxlen = sizeof(system_utsname.domainname), .mode = 0644, .proc_handler = &proc_doutsstring, + .virt_handler = &vx_uts_virt_handler, .strategy = &sysctl_string, }, { @@ -409,6 +440,15 @@ static ctl_table kern_table[] = { .strategy = &sysctl_string, }, #endif + { + .ctl_name = KERN_VSHELPER, + .procname = "vshelper", + .data = &vshelper_path, + .maxlen = 256, + .mode = 0644, + .proc_handler = &proc_dostring, + .strategy = &sysctl_string, + }, #ifdef CONFIG_CHR_DEV_SG { .ctl_name = KERN_SG_BIG_BUFF, @@ -683,6 +723,24 @@ static ctl_table kern_table[] = { .proc_handler = &proc_dointvec, }, #endif + +#if defined(CONFIG_GRKERNSEC_SYSCTL) || defined(CONFIG_GRKERNSEC_MODSTOP) + { + .ctl_name = KERN_GRSECURITY, + .procname = "grsecurity", + .mode = 0500, + .child = grsecurity_table, + }, +#endif + +#ifdef CONFIG_PAX_SOFTMODE + { + .ctl_name = KERN_PAX, + .procname = "pax", + .mode = 0500, + .child = pax_table, + }, +#endif { .ctl_name = 0 } }; @@ -1180,6 +1238,10 @@ static int test_perm(int mode, int op) static inline int ctl_perm(ctl_table *table, int op) { int error; + if (table->de && gr_handle_sysctl_mod(table->de->parent->name, table->de->name, op)) + return -EACCES; + if (gr_handle_chroot_sysctl(op)) + return -EACCES; error = security_sysctl(table, op); if (error) return error; @@ -1216,6 +1278,10 @@ repeat: table = table->child; goto repeat; } + + if (!gr_handle_sysctl(table, oldval, newval)) + return -EPERM; + error = do_sysctl_strategy(table, name, nlen, oldval, oldlenp, newval, newlen, context); @@ -1563,16 +1629,20 @@ static ssize_t proc_writesys(struct file int proc_dostring(ctl_table *table, int write, struct file *filp, void __user *buffer, size_t *lenp, loff_t *ppos) { - size_t len; + size_t len, maxlen; char __user *p; char c; + void *data; + + data = table->data; + maxlen = table->maxlen; + + if (!data || !maxlen || !*lenp || (*ppos && !write)) + return (*lenp = 0); - if (!table->data || !table->maxlen || !*lenp || - (*ppos && !write)) { - *lenp = 0; - return 0; - } - + if (table->virt_handler) + table->virt_handler(table, write, filp->f_xid, &data, &maxlen); + if (write) { len = 0; p = buffer; @@ -1583,20 +1653,20 @@ int proc_dostring(ctl_table *table, int break; len++; } - if (len >= table->maxlen) - len = table->maxlen-1; - if(copy_from_user(table->data, buffer, len)) + if (len >= maxlen) + len = maxlen-1; + if(copy_from_user(data, buffer, len)) return -EFAULT; - ((char *) table->data)[len] = 0; + ((char *) data)[len] = 0; *ppos += *lenp; } else { - len = strlen(table->data); - if (len > table->maxlen) - len = table->maxlen; + len = strlen(data); + if (len > maxlen) + len = maxlen; if (len > *lenp) len = *lenp; if (len) - if(copy_to_user(buffer, table->data, len)) + if(copy_to_user(buffer, data, len)) return -EFAULT; if (len < *lenp) { if(put_user('\n', ((char __user *) buffer) + len)) diff -NurpP --minimal linux-2.6.17.13/kernel/time.c linux-2.6.17.13-g-v/kernel/time.c --- linux-2.6.17.13/kernel/time.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/time.c 2006-09-20 15:00:12.000000000 +0200 @@ -36,6 +36,7 @@ #include #include #include +#include #include #include @@ -93,6 +94,9 @@ asmlinkage long sys_stime(time_t __user return err; do_settimeofday(&tv); + + gr_log_timechange(); + return 0; } @@ -199,6 +203,8 @@ asmlinkage long sys_settimeofday(struct return -EFAULT; } + gr_log_timechange(); + return do_sys_settimeofday(tv ? &new_ts : NULL, tz ? &new_tz : NULL); } diff -NurpP --minimal linux-2.6.17.13/kernel/timer.c linux-2.6.17.13-g-v/kernel/timer.c --- linux-2.6.17.13/kernel/timer.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/timer.c 2006-09-20 15:00:12.000000000 +0200 @@ -34,6 +34,8 @@ #include #include #include +#include +#include #include #include @@ -953,12 +955,6 @@ asmlinkage unsigned long sys_alarm(unsig #endif -#ifndef __alpha__ - -/* - * The Alpha uses getxpid, getxuid, and getxgid instead. Maybe this - * should be moved into arch/i386 instead? - */ /** * sys_getpid - return the thread group id of the current process @@ -971,7 +967,7 @@ asmlinkage unsigned long sys_alarm(unsig */ asmlinkage long sys_getpid(void) { - return current->tgid; + return vx_map_tgid(current->tgid); } /* @@ -988,9 +984,23 @@ asmlinkage long sys_getppid(void) pid = rcu_dereference(current->real_parent)->tgid; rcu_read_unlock(); - return pid; + return vx_map_pid(pid); +} + +#ifdef __alpha__ + +/* + * The Alpha uses getxpid, getxuid, and getxgid instead. + */ + +asmlinkage long do_getxpid(long *ppid) +{ + *ppid = sys_getppid(); + return sys_getpid(); } +#else /* _alpha_ */ + asmlinkage long sys_getuid(void) { /* Only we change this so SMP safe */ @@ -1151,6 +1161,8 @@ asmlinkage long sys_sysinfo(struct sysin tp.tv_nsec = tp.tv_nsec - NSEC_PER_SEC; tp.tv_sec++; } + if (vx_flags(VXF_VIRT_UPTIME, 0)) + vx_vsi_uptime(&tp, NULL); val.uptime = tp.tv_sec + (tp.tv_nsec ? 1 : 0); val.loads[0] = avenrun[0] << (SI_LOAD_SHIFT - FSHIFT); diff -NurpP --minimal linux-2.6.17.13/kernel/user.c linux-2.6.17.13-g-v/kernel/user.c --- linux-2.6.17.13/kernel/user.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/kernel/user.c 2006-09-20 15:00:12.000000000 +0200 @@ -23,8 +23,8 @@ #define UIDHASH_BITS (CONFIG_BASE_SMALL ? 3 : 8) #define UIDHASH_SZ (1 << UIDHASH_BITS) #define UIDHASH_MASK (UIDHASH_SZ - 1) -#define __uidhashfn(uid) (((uid >> UIDHASH_BITS) + uid) & UIDHASH_MASK) -#define uidhashentry(uid) (uidhash_table + __uidhashfn((uid))) +#define __uidhashfn(xid,uid) ((((uid) >> UIDHASH_BITS) + ((uid)^(xid))) & UIDHASH_MASK) +#define uidhashentry(xid,uid) (uidhash_table + __uidhashfn((xid),(uid))) static kmem_cache_t *uid_cachep; static struct list_head uidhash_table[UIDHASH_SZ]; @@ -66,7 +66,7 @@ static inline void uid_hash_remove(struc list_del(&up->uidhash_list); } -static inline struct user_struct *uid_hash_find(uid_t uid, struct list_head *hashent) +static inline struct user_struct *uid_hash_find(xid_t xid, uid_t uid, struct list_head *hashent) { struct list_head *up; @@ -75,7 +75,7 @@ static inline struct user_struct *uid_ha user = list_entry(up, struct user_struct, uidhash_list); - if(user->uid == uid) { + if(user->uid == uid && user->xid == xid) { atomic_inc(&user->__count); return user; } @@ -90,13 +90,13 @@ static inline struct user_struct *uid_ha * * If the user_struct could not be found, return NULL. */ -struct user_struct *find_user(uid_t uid) +struct user_struct *find_user(xid_t xid, uid_t uid) { struct user_struct *ret; unsigned long flags; spin_lock_irqsave(&uidhash_lock, flags); - ret = uid_hash_find(uid, uidhashentry(uid)); + ret = uid_hash_find(xid, uid, uidhashentry(xid, uid)); spin_unlock_irqrestore(&uidhash_lock, flags); return ret; } @@ -120,13 +120,13 @@ void free_uid(struct user_struct *up) } } -struct user_struct * alloc_uid(uid_t uid) +struct user_struct * alloc_uid(xid_t xid, uid_t uid) { - struct list_head *hashent = uidhashentry(uid); + struct list_head *hashent = uidhashentry(xid, uid); struct user_struct *up; spin_lock_irq(&uidhash_lock); - up = uid_hash_find(uid, hashent); + up = uid_hash_find(xid, uid, hashent); spin_unlock_irq(&uidhash_lock); if (!up) { @@ -136,6 +136,7 @@ struct user_struct * alloc_uid(uid_t uid if (!new) return NULL; new->uid = uid; + new->xid = xid; atomic_set(&new->__count, 1); atomic_set(&new->processes, 0); atomic_set(&new->files, 0); @@ -158,7 +159,7 @@ struct user_struct * alloc_uid(uid_t uid * on adding the same user already.. */ spin_lock_irq(&uidhash_lock); - up = uid_hash_find(uid, hashent); + up = uid_hash_find(xid, uid, hashent); if (up) { key_put(new->uid_keyring); key_put(new->session_keyring); @@ -204,7 +205,7 @@ static int __init uid_cache_init(void) /* Insert the root user immediately (init already runs as root) */ spin_lock_irq(&uidhash_lock); - uid_hash_insert(&root_user, uidhashentry(0)); + uid_hash_insert(&root_user, uidhashentry(0,0)); spin_unlock_irq(&uidhash_lock); return 0; diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/Kconfig linux-2.6.17.13-g-v/kernel/vserver/Kconfig --- linux-2.6.17.13/kernel/vserver/Kconfig 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,189 @@ +# +# Linux VServer configuration +# + +menu "Linux VServer" + +config VSERVER_LEGACY + bool "Enable Legacy Kernel API" + default y + help + This enables the legacy API used in vs1.xx, maintaining + compatibility with older vserver tools, and guest images + that are configured using the legacy method. This is + probably a good idea for now, for migration purposes. + + Note that some tools have not yet been altered to use + this API, so disabling this option may reduce some + functionality. + +config VSERVER_LEGACY_VERSION + bool "Show a Legacy Version ID" + depends on VSERVER_LEGACY + default n + help + This shows a special legacy version to very old tools + which do not handle the current version correctly. + + This will probably disable some features of newer tools + so better avoid it, unless you really, really need it + for backwards compatibility. + +config VSERVER_NGNET + bool "Disable Legacy Networking Kernel API" + depends on EXPERIMENTAL + default n + help + This disables the legacy networking API which is required + by the chbind tool. Do not disable it unless you exactly + know what you are doing. + +config VSERVER_REMAP_SADDR + bool "Remap Source IP Address" + depends on EXPERIMENTAL && !VSERVER_LEGACY + default n + help + This allows to remap the source IP address of 'local' + connections from 127.0.0.1 to the first assigned + guest IP. + +config VSERVER_PROC_SECURE + bool "Enable Proc Security" + depends on PROC_FS + default y + help + This configures ProcFS security to initially hide + non-process entries for all contexts except the main and + spectator context (i.e. for all guests), which is a secure + default. + + (note: on 1.2x the entries were visible by default) + +config VSERVER_HARDCPU + bool "Enable Hard CPU Limits" + depends on EXPERIMENTAL + default n + help + Activate the Hard CPU Limits + + This will compile in code that allows the Token Bucket + Scheduler to put processes on hold when a context's + tokens are depleted (provided that its per-context + sched_hard flag is set). + + Processes belonging to that context will not be able + to consume CPU resources again until a per-context + configured minimum of tokens has been reached. + +config VSERVER_HARDCPU_IDLE + bool "Limit the IDLE task" + depends on VSERVER_HARDCPU + default n + help + Limit the idle slices, so the the next context + will be scheduled as soon as possible. + + This might improve interactivity and latency, but + will also marginally increase scheduling overhead. + +choice + prompt "Persistent Inode Context Tagging" + default INOXID_UGID24 + help + This adds persistent context information to filesystems + mounted with the tagxid option. Tagging is a requirement + for per-context disk limits and per-context quota. + + +config INOXID_NONE + bool "Disabled" + help + do not store per-context information in inodes. + +config INOXID_UID16 + bool "UID16/GID32" + help + reduces UID to 16 bit, but leaves GID at 32 bit. + +config INOXID_GID16 + bool "UID32/GID16" + help + reduces GID to 16 bit, but leaves UID at 32 bit. + +config INOXID_UGID24 + bool "UID24/GID24" + help + uses the upper 8bit from UID and GID for XID tagging + which leaves 24bit for UID/GID each, which should be + more than sufficient for normal use. + +config INOXID_INTERN + bool "UID32/GID32" + help + this uses otherwise reserved inode fields in the on + disk representation, which limits the use to a few + filesystems (currently ext2 and ext3) + +config INOXID_RUNTIME + bool "Runtime" + depends on EXPERIMENTAL + help + inodes are tagged when first accessed, this doesn't + require any persistant information, but might give + funny results for mixed access. + +endchoice + +config XID_TAG_NFSD + bool "Tag NFSD User Auth and Files" + default n + help + Enable this if you do want the in-kernel NFS + Server to use the xid tagging specified above. + (will require patched clients too) + +config VSERVER_DEBUG + bool "VServer Debugging Code" + default n + help + Set this to yes if you want to be able to activate + debugging output at runtime. It adds a probably small + overhead to all vserver related functions and + increases the kernel size by about 20k. + +config VSERVER_HISTORY + bool "VServer History Tracing" + depends on VSERVER_DEBUG + default n + help + Set this to yes if you want to record the history of + linux-vserver activities, so they can be replayed in + the event of a kernel panic or oops. + +config VSERVER_HISTORY_SIZE + int "Per-CPU History Size (32-65536)" + depends on VSERVER_HISTORY + range 32 65536 + default 64 + help + This allows you to specify the number of entries in + the per-CPU history buffer. + +endmenu + + +config VSERVER + bool + default y + +config VSERVER_SECURITY + bool + depends on SECURITY + default y + select SECURITY_CAPABILITIES + +config VSERVER_LEGACYNET + bool + depends on !VSERVER_NGNET + default y + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/Makefile linux-2.6.17.13-g-v/kernel/vserver/Makefile --- linux-2.6.17.13/kernel/vserver/Makefile 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/Makefile 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,16 @@ +# +# Makefile for the Linux vserver routines. +# + + +obj-y += vserver.o + +vserver-y := switch.o context.o namespace.o sched.o network.o inode.o \ + limit.o cvirt.o signal.o helper.o init.o dlimit.o + +vserver-$(CONFIG_PROC_FS) += proc.o +vserver-$(CONFIG_VSERVER_DEBUG) += sysctl.o +vserver-$(CONFIG_VSERVER_LEGACY) += legacy.o +vserver-$(CONFIG_VSERVER_LEGACYNET) += legacynet.o +vserver-$(CONFIG_VSERVER_HISTORY) += history.o + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/context.c linux-2.6.17.13-g-v/kernel/vserver/context.c --- linux-2.6.17.13/kernel/vserver/context.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/context.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,918 @@ +/* + * linux/kernel/vserver/context.c + * + * Virtual Server: Context Support + * + * Copyright (C) 2003-2005 Herbert Pötzl + * + * V0.01 context helper + * V0.02 vx_ctx_kill syscall command + * V0.03 replaced context_info calls + * V0.04 redesign of struct (de)alloc + * V0.05 rlimit basic implementation + * V0.06 task_xid and info commands + * V0.07 context flags and caps + * V0.08 switch to RCU based hash + * V0.09 revert to non RCU for now + * V0.10 and back to working RCU hash + * V0.11 and back to locking again + * + */ + +#include +#include +#include + +#include +#include +#include +#include +#include +#include + +#include +#include +#include + +#include +#include + +#include "cvirt_init.h" +#include "limit_init.h" +#include "sched_init.h" + + +/* __alloc_vx_info() + + * allocate an initialized vx_info struct + * doesn't make it visible (hash) */ + +static struct vx_info *__alloc_vx_info(xid_t xid) +{ + struct vx_info *new = NULL; + + vxdprintk(VXD_CBIT(xid, 0), "alloc_vx_info(%d)*", xid); + + /* would this benefit from a slab cache? */ + new = kmalloc(sizeof(struct vx_info), GFP_KERNEL); + if (!new) + return 0; + + memset (new, 0, sizeof(struct vx_info)); + new->vx_id = xid; + INIT_HLIST_NODE(&new->vx_hlist); + atomic_set(&new->vx_usecnt, 0); + atomic_set(&new->vx_tasks, 0); + new->vx_parent = NULL; + new->vx_state = 0; + init_waitqueue_head(&new->vx_wait); + + /* prepare reaper */ + get_task_struct(child_reaper); + new->vx_reaper = child_reaper; + + /* rest of init goes here */ + vx_info_init_limit(&new->limit); + vx_info_init_sched(&new->sched); + vx_info_init_cvirt(&new->cvirt); + vx_info_init_cacct(&new->cacct); + + new->vx_flags = VXF_INIT_SET; + new->vx_bcaps = CAP_INIT_EFF_SET; + new->vx_ccaps = 0; + + new->reboot_cmd = 0; + new->exit_code = 0; + + vxdprintk(VXD_CBIT(xid, 0), + "alloc_vx_info(%d) = %p", xid, new); + vxh_alloc_vx_info(new); + return new; +} + +/* __dealloc_vx_info() + + * final disposal of vx_info */ + +static void __dealloc_vx_info(struct vx_info *vxi) +{ + vxdprintk(VXD_CBIT(xid, 0), + "dealloc_vx_info(%p)", vxi); + vxh_dealloc_vx_info(vxi); + + vxi->vx_hlist.next = LIST_POISON1; + vxi->vx_id = -1; + + vx_info_exit_limit(&vxi->limit); + vx_info_exit_sched(&vxi->sched); + vx_info_exit_cvirt(&vxi->cvirt); + vx_info_exit_cacct(&vxi->cacct); + + vxi->vx_state |= VXS_RELEASED; + kfree(vxi); +} + +static void __shutdown_vx_info(struct vx_info *vxi) +{ + struct namespace *namespace; + struct fs_struct *fs; + + might_sleep(); + + vxi->vx_state |= VXS_SHUTDOWN; + vs_state_change(vxi, VSC_SHUTDOWN); + + namespace = xchg(&vxi->vx_namespace, NULL); + if (namespace) + put_namespace(namespace); + + fs = xchg(&vxi->vx_fs, NULL); + if (fs) + put_fs_struct(fs); +} + +/* exported stuff */ + +void free_vx_info(struct vx_info *vxi) +{ + /* context shutdown is mandatory */ + BUG_ON(!vx_info_state(vxi, VXS_SHUTDOWN)); + + BUG_ON(atomic_read(&vxi->vx_usecnt)); + BUG_ON(atomic_read(&vxi->vx_tasks)); + + BUG_ON(vx_info_state(vxi, VXS_HASHED)); + + BUG_ON(vxi->vx_namespace); + BUG_ON(vxi->vx_fs); + + __dealloc_vx_info(vxi); +} + + +/* hash table for vx_info hash */ + +#define VX_HASH_SIZE 13 + +struct hlist_head vx_info_hash[VX_HASH_SIZE]; + +static spinlock_t vx_info_hash_lock = SPIN_LOCK_UNLOCKED; + + +static inline unsigned int __hashval(xid_t xid) +{ + return (xid % VX_HASH_SIZE); +} + + + +/* __hash_vx_info() + + * add the vxi to the global hash table + * requires the hash_lock to be held */ + +static inline void __hash_vx_info(struct vx_info *vxi) +{ + struct hlist_head *head; + + vxd_assert_lock(&vx_info_hash_lock); + vxdprintk(VXD_CBIT(xid, 4), + "__hash_vx_info: %p[#%d]", vxi, vxi->vx_id); + vxh_hash_vx_info(vxi); + + /* context must not be hashed */ + BUG_ON(vx_info_state(vxi, VXS_HASHED)); + + vxi->vx_state |= VXS_HASHED; + head = &vx_info_hash[__hashval(vxi->vx_id)]; + hlist_add_head(&vxi->vx_hlist, head); +} + +/* __unhash_vx_info() + + * remove the vxi from the global hash table + * requires the hash_lock to be held */ + +static inline void __unhash_vx_info(struct vx_info *vxi) +{ + vxd_assert_lock(&vx_info_hash_lock); + vxdprintk(VXD_CBIT(xid, 4), + "__unhash_vx_info: %p[#%d]", vxi, vxi->vx_id); + vxh_unhash_vx_info(vxi); + + /* context must be hashed */ + BUG_ON(!vx_info_state(vxi, VXS_HASHED)); + + vxi->vx_state &= ~VXS_HASHED; + hlist_del(&vxi->vx_hlist); +} + + +/* __lookup_vx_info() + + * requires the hash_lock to be held + * doesn't increment the vx_refcnt */ + +static inline struct vx_info *__lookup_vx_info(xid_t xid) +{ + struct hlist_head *head = &vx_info_hash[__hashval(xid)]; + struct hlist_node *pos; + struct vx_info *vxi; + + vxd_assert_lock(&vx_info_hash_lock); + hlist_for_each(pos, head) { + vxi = hlist_entry(pos, struct vx_info, vx_hlist); + + if (vxi->vx_id == xid) + goto found; + } + vxi = NULL; +found: + vxdprintk(VXD_CBIT(xid, 0), + "__lookup_vx_info(#%u): %p[#%u]", + xid, vxi, vxi?vxi->vx_id:0); + vxh_lookup_vx_info(vxi, xid); + return vxi; +} + + +/* __vx_dynamic_id() + + * find unused dynamic xid + * requires the hash_lock to be held */ + +static inline xid_t __vx_dynamic_id(void) +{ + static xid_t seq = MAX_S_CONTEXT; + xid_t barrier = seq; + + vxd_assert_lock(&vx_info_hash_lock); + do { + if (++seq > MAX_S_CONTEXT) + seq = MIN_D_CONTEXT; + if (!__lookup_vx_info(seq)) { + vxdprintk(VXD_CBIT(xid, 4), + "__vx_dynamic_id: [#%d]", seq); + return seq; + } + } while (barrier != seq); + return 0; +} + +#ifdef CONFIG_VSERVER_LEGACY + +/* __loc_vx_info() + + * locate or create the requested context + * get() it and if new hash it */ + +static struct vx_info * __loc_vx_info(int id, int *err) +{ + struct vx_info *new, *vxi = NULL; + + vxdprintk(VXD_CBIT(xid, 1), "loc_vx_info(%d)*", id); + + if (!(new = __alloc_vx_info(id))) { + *err = -ENOMEM; + return NULL; + } + + /* required to make dynamic xids unique */ + spin_lock(&vx_info_hash_lock); + + /* dynamic context requested */ + if (id == VX_DYNAMIC_ID) { + id = __vx_dynamic_id(); + if (!id) { + printk(KERN_ERR "no dynamic context available.\n"); + goto out_unlock; + } + new->vx_id = id; + } + /* existing context requested */ + else if ((vxi = __lookup_vx_info(id))) { + /* context in setup is not available */ + if (vxi->vx_flags & VXF_STATE_SETUP) { + vxdprintk(VXD_CBIT(xid, 0), + "loc_vx_info(%d) = %p (not available)", id, vxi); + vxi = NULL; + *err = -EBUSY; + } else { + vxdprintk(VXD_CBIT(xid, 0), + "loc_vx_info(%d) = %p (found)", id, vxi); + get_vx_info(vxi); + *err = 0; + } + goto out_unlock; + } + + /* new context requested */ + vxdprintk(VXD_CBIT(xid, 0), + "loc_vx_info(%d) = %p (new)", id, new); + __hash_vx_info(get_vx_info(new)); + vxi = new, new = NULL; + *err = 1; + +out_unlock: + spin_unlock(&vx_info_hash_lock); + vxh_loc_vx_info(vxi, id); + if (new) + __dealloc_vx_info(new); + return vxi; +} + +#endif + +/* __create_vx_info() + + * create the requested context + * get() and hash it */ + +static struct vx_info * __create_vx_info(int id) +{ + struct vx_info *new, *vxi = NULL; + + vxdprintk(VXD_CBIT(xid, 1), "create_vx_info(%d)*", id); + + if (!(new = __alloc_vx_info(id))) + return ERR_PTR(-ENOMEM); + + /* required to make dynamic xids unique */ + spin_lock(&vx_info_hash_lock); + + /* dynamic context requested */ + if (id == VX_DYNAMIC_ID) { + id = __vx_dynamic_id(); + if (!id) { + printk(KERN_ERR "no dynamic context available.\n"); + vxi = ERR_PTR(-EAGAIN); + goto out_unlock; + } + new->vx_id = id; + } + /* static context requested */ + else if ((vxi = __lookup_vx_info(id))) { + vxdprintk(VXD_CBIT(xid, 0), + "create_vx_info(%d) = %p (already there)", id, vxi); + if (vx_info_flags(vxi, VXF_STATE_SETUP, 0)) + vxi = ERR_PTR(-EBUSY); + else + vxi = ERR_PTR(-EEXIST); + goto out_unlock; + } + /* dynamic xid creation blocker */ + else if (id >= MIN_D_CONTEXT) { + vxdprintk(VXD_CBIT(xid, 0), + "create_vx_info(%d) (dynamic rejected)", id); + vxi = ERR_PTR(-EINVAL); + goto out_unlock; + } + + /* new context */ + vxdprintk(VXD_CBIT(xid, 0), + "create_vx_info(%d) = %p (new)", id, new); + __hash_vx_info(get_vx_info(new)); + vxi = new, new = NULL; + +out_unlock: + spin_unlock(&vx_info_hash_lock); + vxh_create_vx_info(IS_ERR(vxi)?NULL:vxi, id); + if (new) + __dealloc_vx_info(new); + return vxi; +} + + +/* exported stuff */ + + +void unhash_vx_info(struct vx_info *vxi) +{ + __shutdown_vx_info(vxi); + spin_lock(&vx_info_hash_lock); + __unhash_vx_info(vxi); + spin_unlock(&vx_info_hash_lock); + __wakeup_vx_info(vxi); +} + + +/* lookup_vx_info() + + * search for a vx_info and get() it + * negative id means current */ + +struct vx_info *lookup_vx_info(int id) +{ + struct vx_info *vxi = NULL; + + if (id < 0) { + vxi = get_vx_info(current->vx_info); + } else if (id > 1) { + spin_lock(&vx_info_hash_lock); + vxi = get_vx_info(__lookup_vx_info(id)); + spin_unlock(&vx_info_hash_lock); + } + return vxi; +} + +/* xid_is_hashed() + + * verify that xid is still hashed */ + +int xid_is_hashed(xid_t xid) +{ + int hashed; + + spin_lock(&vx_info_hash_lock); + hashed = (__lookup_vx_info(xid) != NULL); + spin_unlock(&vx_info_hash_lock); + return hashed; +} + +#ifdef CONFIG_VSERVER_LEGACY + +struct vx_info *lookup_or_create_vx_info(int id) +{ + int err; + + return __loc_vx_info(id, &err); +} + +#endif + +#ifdef CONFIG_PROC_FS + +int get_xid_list(int index, unsigned int *xids, int size) +{ + int hindex, nr_xids = 0; + + for (hindex = 0; hindex < VX_HASH_SIZE; hindex++) { + struct hlist_head *head = &vx_info_hash[hindex]; + struct hlist_node *pos; + + spin_lock(&vx_info_hash_lock); + hlist_for_each(pos, head) { + struct vx_info *vxi; + + if (--index > 0) + continue; + + vxi = hlist_entry(pos, struct vx_info, vx_hlist); + xids[nr_xids] = vxi->vx_id; + if (++nr_xids >= size) { + spin_unlock(&vx_info_hash_lock); + goto out; + } + } + /* keep the lock time short */ + spin_unlock(&vx_info_hash_lock); + } +out: + return nr_xids; +} +#endif + + +int vx_migrate_user(struct task_struct *p, struct vx_info *vxi) +{ + struct user_struct *new_user, *old_user; + + if (!p || !vxi) + BUG(); + new_user = alloc_uid(vxi->vx_id, p->uid); + if (!new_user) + return -ENOMEM; + + old_user = p->user; + if (new_user != old_user) { + atomic_inc(&new_user->processes); + atomic_dec(&old_user->processes); + p->user = new_user; + } + free_uid(old_user); + return 0; +} + +void vx_mask_bcaps(struct vx_info *vxi, struct task_struct *p) +{ + p->cap_effective &= vxi->vx_bcaps; + p->cap_inheritable &= vxi->vx_bcaps; + p->cap_permitted &= vxi->vx_bcaps; +} + + +#include + +static int vx_openfd_task(struct task_struct *tsk) +{ + struct files_struct *files = tsk->files; + struct fdtable *fdt; + const unsigned long *bptr; + int count, total; + + /* no rcu_read_lock() because of spin_lock() */ + spin_lock(&files->file_lock); + fdt = files_fdtable(files); + bptr = fdt->open_fds->fds_bits; + count = fdt->max_fds / (sizeof(unsigned long) * 8); + for (total = 0; count > 0; count--) { + if (*bptr) + total += hweight_long(*bptr); + bptr++; + } + spin_unlock(&files->file_lock); + return total; +} + +/* + * migrate task to new context + * gets vxi, puts old_vxi on change + */ + +int vx_migrate_task(struct task_struct *p, struct vx_info *vxi) +{ + struct vx_info *old_vxi; + int ret = 0; + + if (!p || !vxi) + BUG(); + + old_vxi = task_get_vx_info(p); + if (old_vxi == vxi) + goto out; + + vxdprintk(VXD_CBIT(xid, 5), + "vx_migrate_task(%p,%p[#%d.%d])", p, vxi, + vxi->vx_id, atomic_read(&vxi->vx_usecnt)); + + if (!(ret = vx_migrate_user(p, vxi))) { + int openfd; + + task_lock(p); + openfd = vx_openfd_task(p); + + if (old_vxi) { + atomic_dec(&old_vxi->cvirt.nr_threads); + atomic_dec(&old_vxi->cvirt.nr_running); + atomic_dec(&old_vxi->limit.rcur[RLIMIT_NPROC]); + /* FIXME: what about the struct files here? */ + atomic_sub(openfd, &old_vxi->limit.rcur[VLIMIT_OPENFD]); + } + atomic_inc(&vxi->cvirt.nr_threads); + atomic_inc(&vxi->cvirt.nr_running); + atomic_inc(&vxi->limit.rcur[RLIMIT_NPROC]); + /* FIXME: what about the struct files here? */ + atomic_add(openfd, &vxi->limit.rcur[VLIMIT_OPENFD]); + + if (old_vxi) { + release_vx_info(old_vxi, p); + clr_vx_info(&p->vx_info); + } + claim_vx_info(vxi, p); + set_vx_info(&p->vx_info, vxi); + p->xid = vxi->vx_id; + + vxdprintk(VXD_CBIT(xid, 5), + "moved task %p into vxi:%p[#%d]", + p, vxi, vxi->vx_id); + + vx_mask_bcaps(vxi, p); + task_unlock(p); + } +out: + put_vx_info(old_vxi); + return ret; +} + +int vx_set_reaper(struct vx_info *vxi, struct task_struct *p) +{ + struct task_struct *old_reaper; + + if (!vxi) + return -EINVAL; + + vxdprintk(VXD_CBIT(xid, 6), + "vx_set_reaper(%p[#%d],%p[#%d,%d])", + vxi, vxi->vx_id, p, p->xid, p->pid); + + old_reaper = vxi->vx_reaper; + if (old_reaper == p) + return 0; + + /* set new child reaper */ + get_task_struct(p); + vxi->vx_reaper = p; + put_task_struct(old_reaper); + return 0; +} + +int vx_set_init(struct vx_info *vxi, struct task_struct *p) +{ + if (!vxi) + return -EINVAL; + + vxdprintk(VXD_CBIT(xid, 6), + "vx_set_init(%p[#%d],%p[#%d,%d,%d])", + vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid); + + vxi->vx_flags &= ~VXF_STATE_INIT; + vxi->vx_initpid = p->tgid; + return 0; +} + +void vx_exit_init(struct vx_info *vxi, struct task_struct *p, int code) +{ + vxdprintk(VXD_CBIT(xid, 6), + "vx_exit_init(%p[#%d],%p[#%d,%d,%d])", + vxi, vxi->vx_id, p, p->xid, p->pid, p->tgid); + + vxi->exit_code = code; + vxi->vx_initpid = 0; +} + +void vx_set_persistent(struct vx_info *vxi) +{ + vxdprintk(VXD_CBIT(xid, 6), + "vx_set_persistent(%p[#%d])", vxi, vxi->vx_id); + + get_vx_info(vxi); + claim_vx_info(vxi, current); +} + +void vx_clear_persistent(struct vx_info *vxi) +{ + vxdprintk(VXD_CBIT(xid, 6), + "vx_clear_persistent(%p[#%d])", vxi, vxi->vx_id); + + release_vx_info(vxi, current); + put_vx_info(vxi); +} + +void vx_update_persistent(struct vx_info *vxi) +{ + if (vx_info_flags(vxi, VXF_PERSISTENT, 0)) + vx_set_persistent(vxi); + else + vx_clear_persistent(vxi); +} + + +/* task must be current or locked */ + +void exit_vx_info(struct task_struct *p, int code) +{ + struct vx_info *vxi = p->vx_info; + + if (vxi) { + atomic_dec(&vxi->cvirt.nr_threads); + vx_nproc_dec(p); + + vxi->exit_code = code; + if (vxi->vx_initpid == p->tgid) + vx_exit_init(vxi, p, code); + if (vxi->vx_reaper == p) + vx_set_reaper(vxi, child_reaper); + release_vx_info(vxi, p); + } +} + + +/* vserver syscall commands below here */ + +/* taks xid and vx_info functions */ + +#include + + +int vc_task_xid(uint32_t id, void __user *data) +{ + xid_t xid; + + if (id) { + struct task_struct *tsk; + + if (!vx_check(0, VX_ADMIN|VX_WATCH)) + return -EPERM; + + read_lock(&tasklist_lock); + tsk = find_task_by_real_pid(id); + xid = (tsk) ? tsk->xid : -ESRCH; + read_unlock(&tasklist_lock); + } + else + xid = vx_current_xid(); + return xid; +} + + +int vc_vx_info(uint32_t id, void __user *data) +{ + struct vx_info *vxi; + struct vcmd_vx_info_v0 vc_data; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (!capable(CAP_SYS_ADMIN) || !capable(CAP_SYS_RESOURCE)) + return -EPERM; + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + + vc_data.xid = vxi->vx_id; + vc_data.initpid = vxi->vx_initpid; + put_vx_info(vxi); + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + + +/* context functions */ + +int vc_ctx_create(uint32_t xid, void __user *data) +{ + struct vcmd_ctx_create vc_data = { .flagword = VXF_INIT_SET }; + struct vx_info *new_vxi; + int ret; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (data && copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + if ((xid > MAX_S_CONTEXT) && (xid != VX_DYNAMIC_ID)) + return -EINVAL; + if (xid < 2) + return -EINVAL; + + new_vxi = __create_vx_info(xid); + if (IS_ERR(new_vxi)) + return PTR_ERR(new_vxi); + + /* initial flags */ + new_vxi->vx_flags = vc_data.flagword; + + /* get a reference for persistent contexts */ + if ((vc_data.flagword & VXF_PERSISTENT)) + vx_set_persistent(new_vxi); + + ret = -ENOEXEC; + if (vs_state_change(new_vxi, VSC_STARTUP)) + goto out_unhash; + ret = vx_migrate_task(current, new_vxi); + if (!ret) { + /* return context id on success */ + ret = new_vxi->vx_id; + goto out; + } +out_unhash: + /* prepare for context disposal */ + new_vxi->vx_state |= VXS_SHUTDOWN; + if ((vc_data.flagword & VXF_PERSISTENT)) + vx_clear_persistent(new_vxi); + __unhash_vx_info(new_vxi); +out: + put_vx_info(new_vxi); + return ret; +} + + +int vc_ctx_migrate(uint32_t id, void __user *data) +{ + struct vcmd_ctx_migrate vc_data = { .flagword = 0 }; + struct vx_info *vxi; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (data && copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + /* dirty hack until Spectator becomes a cap */ + if (id == 1) { + current->xid = 1; + return 0; + } + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + vx_migrate_task(current, vxi); + if (vc_data.flagword & VXM_SET_INIT) + vx_set_init(vxi, current); + if (vc_data.flagword & VXM_SET_REAPER) + vx_set_reaper(vxi, current); + put_vx_info(vxi); + return 0; +} + + +int vc_get_cflags(uint32_t id, void __user *data) +{ + struct vx_info *vxi; + struct vcmd_ctx_flags_v0 vc_data; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + + vc_data.flagword = vxi->vx_flags; + + /* special STATE flag handling */ + vc_data.mask = vx_mask_flags(~0UL, vxi->vx_flags, VXF_ONE_TIME); + + put_vx_info(vxi); + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +int vc_set_cflags(uint32_t id, void __user *data) +{ + struct vx_info *vxi; + struct vcmd_ctx_flags_v0 vc_data; + uint64_t mask, trigger; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + + /* special STATE flag handling */ + mask = vx_mask_mask(vc_data.mask, vxi->vx_flags, VXF_ONE_TIME); + trigger = (mask & vxi->vx_flags) ^ (mask & vc_data.flagword); + + if (vxi == current->vx_info) { + if (trigger & VXF_STATE_SETUP) + vx_mask_bcaps(vxi, current); + if (trigger & VXF_STATE_INIT) { + vx_set_init(vxi, current); + vx_set_reaper(vxi, current); + } + } + + vxi->vx_flags = vx_mask_flags(vxi->vx_flags, + vc_data.flagword, mask); + if (trigger & VXF_PERSISTENT) + vx_update_persistent(vxi); + + put_vx_info(vxi); + return 0; +} + +int vc_get_ccaps(uint32_t id, void __user *data) +{ + struct vx_info *vxi; + struct vcmd_ctx_caps_v0 vc_data; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + + vc_data.bcaps = vxi->vx_bcaps; + vc_data.ccaps = vxi->vx_ccaps; + vc_data.cmask = ~0UL; + put_vx_info(vxi); + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +int vc_set_ccaps(uint32_t id, void __user *data) +{ + struct vx_info *vxi; + struct vcmd_ctx_caps_v0 vc_data; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + + vxi->vx_bcaps &= vc_data.bcaps; + vxi->vx_ccaps = vx_mask_flags(vxi->vx_ccaps, + vc_data.ccaps, vc_data.cmask); + put_vx_info(vxi); + return 0; +} + +#include + +EXPORT_SYMBOL_GPL(free_vx_info); + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/cvirt.c linux-2.6.17.13-g-v/kernel/vserver/cvirt.c --- linux-2.6.17.13/kernel/vserver/cvirt.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/cvirt.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,260 @@ +/* + * linux/kernel/vserver/cvirt.c + * + * Virtual Server: Context Virtualization + * + * Copyright (C) 2004-2005 Herbert Pötzl + * + * V0.01 broken out from limit.c + * V0.02 added utsname stuff + * + */ + +#include +#include +#include +#include +#include +#include +#include + +#include +#include + + +void vx_vsi_uptime(struct timespec *uptime, struct timespec *idle) +{ + struct vx_info *vxi = current->vx_info; + + set_normalized_timespec(uptime, + uptime->tv_sec - vxi->cvirt.bias_uptime.tv_sec, + uptime->tv_nsec - vxi->cvirt.bias_uptime.tv_nsec); + if (!idle) + return; + set_normalized_timespec(idle, + idle->tv_sec - vxi->cvirt.bias_idle.tv_sec, + idle->tv_nsec - vxi->cvirt.bias_idle.tv_nsec); + return; +} + +uint64_t vx_idle_jiffies(void) +{ + return init_task.utime + init_task.stime; +} + + + +static inline uint32_t __update_loadavg(uint32_t load, + int wsize, int delta, int n) +{ + unsigned long long calc, prev; + + /* just set it to n */ + if (unlikely(delta >= wsize)) + return (n << FSHIFT); + + calc = delta * n; + calc <<= FSHIFT; + prev = (wsize - delta); + prev *= load; + calc += prev; + do_div(calc, wsize); + return calc; +} + + +void vx_update_load(struct vx_info *vxi) +{ + uint32_t now, last, delta; + unsigned int nr_running, nr_uninterruptible; + unsigned int total; + unsigned long flags; + + spin_lock_irqsave(&vxi->cvirt.load_lock, flags); + + now = jiffies; + last = vxi->cvirt.load_last; + delta = now - last; + + if (delta < 5*HZ) + goto out; + + nr_running = atomic_read(&vxi->cvirt.nr_running); + nr_uninterruptible = atomic_read(&vxi->cvirt.nr_uninterruptible); + total = nr_running + nr_uninterruptible; + + vxi->cvirt.load[0] = __update_loadavg(vxi->cvirt.load[0], + 60*HZ, delta, total); + vxi->cvirt.load[1] = __update_loadavg(vxi->cvirt.load[1], + 5*60*HZ, delta, total); + vxi->cvirt.load[2] = __update_loadavg(vxi->cvirt.load[2], + 15*60*HZ, delta, total); + + vxi->cvirt.load_last = now; +out: + atomic_inc(&vxi->cvirt.load_updates); + spin_unlock_irqrestore(&vxi->cvirt.load_lock, flags); +} + + +int vx_uts_virt_handler(struct ctl_table *ctl, int write, xid_t xid, + void **datap, size_t *lenp) +{ + switch (ctl->ctl_name) { + case KERN_OSTYPE: + *datap = vx_new_uts(sysname); + break; + case KERN_OSRELEASE: + *datap = vx_new_uts(release); + break; + case KERN_VERSION: + *datap = vx_new_uts(version); + break; + case KERN_NODENAME: + *datap = vx_new_uts(nodename); + break; + case KERN_DOMAINNAME: + *datap = vx_new_uts(domainname); + break; + } + + return 0; +} + + + +/* + * Commands to do_syslog: + * + * 0 -- Close the log. Currently a NOP. + * 1 -- Open the log. Currently a NOP. + * 2 -- Read from the log. + * 3 -- Read all messages remaining in the ring buffer. + * 4 -- Read and clear all messages remaining in the ring buffer + * 5 -- Clear ring buffer. + * 6 -- Disable printk's to console + * 7 -- Enable printk's to console + * 8 -- Set level of messages printed to console + * 9 -- Return number of unread characters in the log buffer + * 10 -- Return size of the log buffer + */ +int vx_do_syslog(int type, char __user *buf, int len) +{ + int error = 0; + int do_clear = 0; + struct vx_info *vxi = current->vx_info; + struct _vx_syslog *log; + + if (!vxi) + return -EINVAL; + log = &vxi->cvirt.syslog; + + switch (type) { + case 0: /* Close log */ + case 1: /* Open log */ + break; + case 2: /* Read from log */ + error = wait_event_interruptible(log->log_wait, + (log->log_start - log->log_end)); + if (error) + break; + spin_lock_irq(&log->logbuf_lock); + spin_unlock_irq(&log->logbuf_lock); + break; + case 4: /* Read/clear last kernel messages */ + do_clear = 1; + /* fall through */ + case 3: /* Read last kernel messages */ + return 0; + + case 5: /* Clear ring buffer */ + return 0; + + case 6: /* Disable logging to console */ + case 7: /* Enable logging to console */ + case 8: /* Set level of messages printed to console */ + break; + + case 9: /* Number of chars in the log buffer */ + return 0; + case 10: /* Size of the log buffer */ + return 0; + default: + error = -EINVAL; + break; + } + return error; +} + + +/* virtual host info names */ + +static char * vx_vhi_name(struct vx_info *vxi, int id) +{ + switch (id) { + case VHIN_CONTEXT: + return vxi->vx_name; + case VHIN_SYSNAME: + return vxi->cvirt.utsname.sysname; + case VHIN_NODENAME: + return vxi->cvirt.utsname.nodename; + case VHIN_RELEASE: + return vxi->cvirt.utsname.release; + case VHIN_VERSION: + return vxi->cvirt.utsname.version; + case VHIN_MACHINE: + return vxi->cvirt.utsname.machine; + case VHIN_DOMAINNAME: + return vxi->cvirt.utsname.domainname; + default: + return NULL; + } + return NULL; +} + +int vc_set_vhi_name(uint32_t id, void __user *data) +{ + struct vx_info *vxi; + struct vcmd_vhi_name_v0 vc_data; + char *name; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + + name = vx_vhi_name(vxi, vc_data.field); + if (name) + memcpy(name, vc_data.name, 65); + put_vx_info(vxi); + return (name ? 0 : -EFAULT); +} + +int vc_get_vhi_name(uint32_t id, void __user *data) +{ + struct vx_info *vxi; + struct vcmd_vhi_name_v0 vc_data; + char *name; + + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + + name = vx_vhi_name(vxi, vc_data.field); + if (!name) + goto out_put; + + memcpy(vc_data.name, name, 65); + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + return -EFAULT; +out_put: + put_vx_info(vxi); + return (name ? 0 : -EFAULT); +} diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/cvirt_init.h linux-2.6.17.13-g-v/kernel/vserver/cvirt_init.h --- linux-2.6.17.13/kernel/vserver/cvirt_init.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/cvirt_init.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,81 @@ + + +#include + + +extern uint64_t vx_idle_jiffies(void); + +static inline void vx_info_init_cvirt(struct _vx_cvirt *cvirt) +{ + uint64_t idle_jiffies = vx_idle_jiffies(); + uint64_t nsuptime; + + do_posix_clock_monotonic_gettime(&cvirt->bias_uptime); + nsuptime = (unsigned long long)cvirt->bias_uptime.tv_sec + * NSEC_PER_SEC + cvirt->bias_uptime.tv_nsec; + cvirt->bias_clock = nsec_to_clock_t(nsuptime); + + jiffies_to_timespec(idle_jiffies, &cvirt->bias_idle); + atomic_set(&cvirt->nr_threads, 0); + atomic_set(&cvirt->nr_running, 0); + atomic_set(&cvirt->nr_uninterruptible, 0); + atomic_set(&cvirt->nr_onhold, 0); + + down_read(&uts_sem); + cvirt->utsname = system_utsname; + up_read(&uts_sem); + + spin_lock_init(&cvirt->load_lock); + cvirt->load_last = jiffies; + atomic_set(&cvirt->load_updates, 0); + cvirt->load[0] = 0; + cvirt->load[1] = 0; + cvirt->load[2] = 0; + atomic_set(&cvirt->total_forks, 0); + + spin_lock_init(&cvirt->syslog.logbuf_lock); + init_waitqueue_head(&cvirt->syslog.log_wait); + cvirt->syslog.log_start = 0; + cvirt->syslog.log_end = 0; + cvirt->syslog.con_start = 0; + cvirt->syslog.logged_chars = 0; +} + +static inline void vx_info_exit_cvirt(struct _vx_cvirt *cvirt) +{ +#ifdef CONFIG_VSERVER_DEBUG + int value; + + vxwprintk((value = atomic_read(&cvirt->nr_threads)), + "!!! cvirt: %p[nr_threads] = %d on exit.", + cvirt, value); + vxwprintk((value = atomic_read(&cvirt->nr_running)), + "!!! cvirt: %p[nr_running] = %d on exit.", + cvirt, value); + vxwprintk((value = atomic_read(&cvirt->nr_uninterruptible)), + "!!! cvirt: %p[nr_uninterruptible] = %d on exit.", + cvirt, value); + vxwprintk((value = atomic_read(&cvirt->nr_onhold)), + "!!! cvirt: %p[nr_onhold] = %d on exit.", + cvirt, value); +#endif + return; +} + +static inline void vx_info_init_cacct(struct _vx_cacct *cacct) +{ + int i,j; + + for (i=0; i<5; i++) { + for (j=0; j<3; j++) { + atomic_set(&cacct->sock[i][j].count, 0); + atomic_set(&cacct->sock[i][j].total, 0); + } + } +} + +static inline void vx_info_exit_cacct(struct _vx_cacct *cacct) +{ + return; +} + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/cvirt_proc.h linux-2.6.17.13-g-v/kernel/vserver/cvirt_proc.h --- linux-2.6.17.13/kernel/vserver/cvirt_proc.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/cvirt_proc.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,92 @@ +#ifndef _VX_CVIRT_PROC_H +#define _VX_CVIRT_PROC_H + +#include + + +#define LOAD_INT(x) ((x) >> FSHIFT) +#define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1-1)) * 100) + +static inline int vx_info_proc_cvirt(struct _vx_cvirt *cvirt, char *buffer) +{ + int length = 0; + int a, b, c; + + length += sprintf(buffer + length, + "BiasUptime:\t%lu.%02lu\n", + (unsigned long)cvirt->bias_uptime.tv_sec, + (cvirt->bias_uptime.tv_nsec / (NSEC_PER_SEC / 100))); + length += sprintf(buffer + length, + "SysName:\t%.*s\n" + "NodeName:\t%.*s\n" + "Release:\t%.*s\n" + "Version:\t%.*s\n" + "Machine:\t%.*s\n" + "DomainName:\t%.*s\n" + ,__NEW_UTS_LEN, cvirt->utsname.sysname + ,__NEW_UTS_LEN, cvirt->utsname.nodename + ,__NEW_UTS_LEN, cvirt->utsname.release + ,__NEW_UTS_LEN, cvirt->utsname.version + ,__NEW_UTS_LEN, cvirt->utsname.machine + ,__NEW_UTS_LEN, cvirt->utsname.domainname + ); + + a = cvirt->load[0] + (FIXED_1/200); + b = cvirt->load[1] + (FIXED_1/200); + c = cvirt->load[2] + (FIXED_1/200); + length += sprintf(buffer + length, + "nr_threads:\t%d\n" + "nr_running:\t%d\n" + "nr_unintr:\t%d\n" + "nr_onhold:\t%d\n" + "load_updates:\t%d\n" + "loadavg:\t%d.%02d %d.%02d %d.%02d\n" + "total_forks:\t%d\n" + ,atomic_read(&cvirt->nr_threads) + ,atomic_read(&cvirt->nr_running) + ,atomic_read(&cvirt->nr_uninterruptible) + ,atomic_read(&cvirt->nr_onhold) + ,atomic_read(&cvirt->load_updates) + ,LOAD_INT(a), LOAD_FRAC(a) + ,LOAD_INT(b), LOAD_FRAC(b) + ,LOAD_INT(c), LOAD_FRAC(c) + ,atomic_read(&cvirt->total_forks) + ); + return length; +} + + +static inline long vx_sock_count(struct _vx_cacct *cacct, int type, int pos) +{ + return atomic_read(&cacct->sock[type][pos].count); +} + + +static inline long vx_sock_total(struct _vx_cacct *cacct, int type, int pos) +{ + return atomic_read(&cacct->sock[type][pos].total); +} + +static inline int vx_info_proc_cacct(struct _vx_cacct *cacct, char *buffer) +{ + int i,j, length = 0; + static char *type[] = { "UNSPEC", "UNIX", "INET", "INET6", "OTHER" }; + + for (i=0; i<5; i++) { + length += sprintf(buffer + length, + "%s:", type[i]); + for (j=0; j<3; j++) { + length += sprintf(buffer + length, + "\t%12lu/%-12lu" + ,vx_sock_count(cacct, i, j) + ,vx_sock_total(cacct, i, j) + ); + } + buffer[length++] = '\n'; + } + length += sprintf(buffer + length, + "forks:\t%lu\n", cacct->total_forks); + return length; +} + +#endif /* _VX_CVIRT_PROC_H */ diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/dlimit.c linux-2.6.17.13-g-v/kernel/vserver/dlimit.c --- linux-2.6.17.13/kernel/vserver/dlimit.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/dlimit.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,548 @@ +/* + * linux/kernel/vserver/dlimit.c + * + * Virtual Server: Context Disk Limits + * + * Copyright (C) 2004-2005 Herbert Pötzl + * + * V0.01 initial version + * V0.02 compat32 splitup + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +/* __alloc_dl_info() + + * allocate an initialized dl_info struct + * doesn't make it visible (hash) */ + +static struct dl_info *__alloc_dl_info(struct super_block *sb, xid_t xid) +{ + struct dl_info *new = NULL; + + vxdprintk(VXD_CBIT(dlim, 5), + "alloc_dl_info(%p,%d)*", sb, xid); + + /* would this benefit from a slab cache? */ + new = kmalloc(sizeof(struct dl_info), GFP_KERNEL); + if (!new) + return 0; + + memset (new, 0, sizeof(struct dl_info)); + new->dl_xid = xid; + new->dl_sb = sb; + INIT_RCU_HEAD(&new->dl_rcu); + INIT_HLIST_NODE(&new->dl_hlist); + spin_lock_init(&new->dl_lock); + atomic_set(&new->dl_refcnt, 0); + atomic_set(&new->dl_usecnt, 0); + + /* rest of init goes here */ + + vxdprintk(VXD_CBIT(dlim, 4), + "alloc_dl_info(%p,%d) = %p", sb, xid, new); + return new; +} + +/* __dealloc_dl_info() + + * final disposal of dl_info */ + +static void __dealloc_dl_info(struct dl_info *dli) +{ + vxdprintk(VXD_CBIT(dlim, 4), + "dealloc_dl_info(%p)", dli); + + dli->dl_hlist.next = LIST_POISON1; + dli->dl_xid = -1; + dli->dl_sb = 0; + + BUG_ON(atomic_read(&dli->dl_usecnt)); + BUG_ON(atomic_read(&dli->dl_refcnt)); + + kfree(dli); +} + + +/* hash table for dl_info hash */ + +#define DL_HASH_SIZE 13 + +struct hlist_head dl_info_hash[DL_HASH_SIZE]; + +static spinlock_t dl_info_hash_lock = SPIN_LOCK_UNLOCKED; + + +static inline unsigned int __hashval(struct super_block *sb, xid_t xid) +{ + return ((xid ^ (unsigned long)sb) % DL_HASH_SIZE); +} + + + +/* __hash_dl_info() + + * add the dli to the global hash table + * requires the hash_lock to be held */ + +static inline void __hash_dl_info(struct dl_info *dli) +{ + struct hlist_head *head; + + vxdprintk(VXD_CBIT(dlim, 6), + "__hash_dl_info: %p[#%d]", dli, dli->dl_xid); + get_dl_info(dli); + head = &dl_info_hash[__hashval(dli->dl_sb, dli->dl_xid)]; + hlist_add_head_rcu(&dli->dl_hlist, head); +} + +/* __unhash_dl_info() + + * remove the dli from the global hash table + * requires the hash_lock to be held */ + +static inline void __unhash_dl_info(struct dl_info *dli) +{ + vxdprintk(VXD_CBIT(dlim, 6), + "__unhash_dl_info: %p[#%d]", dli, dli->dl_xid); + hlist_del_rcu(&dli->dl_hlist); + put_dl_info(dli); +} + + +/* __lookup_dl_info() + + * requires the rcu_read_lock() + * doesn't increment the dl_refcnt */ + +static inline struct dl_info *__lookup_dl_info(struct super_block *sb, xid_t xid) +{ + struct hlist_head *head = &dl_info_hash[__hashval(sb, xid)]; + struct hlist_node *pos; + struct dl_info *dli; + + hlist_for_each_entry_rcu(dli, pos, head, dl_hlist) { + + if (dli->dl_xid == xid && dli->dl_sb == sb) { + return dli; + } + } + return NULL; +} + + +struct dl_info *locate_dl_info(struct super_block *sb, xid_t xid) +{ + struct dl_info *dli; + + rcu_read_lock(); + dli = get_dl_info(__lookup_dl_info(sb, xid)); + vxdprintk(VXD_CBIT(dlim, 7), + "locate_dl_info(%p,#%d) = %p", sb, xid, dli); + rcu_read_unlock(); + return dli; +} + +void rcu_free_dl_info(struct rcu_head *head) +{ + struct dl_info *dli = container_of(head, struct dl_info, dl_rcu); + int usecnt, refcnt; + + BUG_ON(!dli || !head); + + usecnt = atomic_read(&dli->dl_usecnt); + BUG_ON(usecnt < 0); + + refcnt = atomic_read(&dli->dl_refcnt); + BUG_ON(refcnt < 0); + + vxdprintk(VXD_CBIT(dlim, 3), + "rcu_free_dl_info(%p)", dli); + if (!usecnt) + __dealloc_dl_info(dli); + else + printk("!!! rcu didn't free\n"); +} + + + + +static int do_addrem_dlimit(uint32_t id, const char __user *name, + uint32_t flags, int add) +{ + struct nameidata nd; + int ret; + + ret = user_path_walk_link(name, &nd); + if (!ret) { + struct super_block *sb; + struct dl_info *dli; + + ret = -EINVAL; + if (!nd.dentry->d_inode) + goto out_release; + if (!(sb = nd.dentry->d_inode->i_sb)) + goto out_release; + + if (add) { + dli = __alloc_dl_info(sb, id); + spin_lock(&dl_info_hash_lock); + + ret = -EEXIST; + if (__lookup_dl_info(sb, id)) + goto out_unlock; + __hash_dl_info(dli); + dli = NULL; + } else { + spin_lock(&dl_info_hash_lock); + dli = __lookup_dl_info(sb, id); + + ret = -ESRCH; + if (!dli) + goto out_unlock; + __unhash_dl_info(dli); + } + ret = 0; + out_unlock: + spin_unlock(&dl_info_hash_lock); + if (add && dli) + __dealloc_dl_info(dli); + out_release: + path_release(&nd); + } + return ret; +} + +int vc_add_dlimit(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_base_v0 vc_data; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 1); +} + +int vc_rem_dlimit(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_base_v0 vc_data; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_addrem_dlimit(id, vc_data.name, vc_data.flags, 0); +} + +#ifdef CONFIG_COMPAT + +int vc_add_dlimit_x32(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_base_v0_x32 vc_data; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_addrem_dlimit(id, + compat_ptr(vc_data.name_ptr), vc_data.flags, 1); +} + +int vc_rem_dlimit_x32(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_base_v0_x32 vc_data; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_addrem_dlimit(id, + compat_ptr(vc_data.name_ptr), vc_data.flags, 0); +} + +#endif /* CONFIG_COMPAT */ + + +static inline +int do_set_dlimit(uint32_t id, const char __user *name, + uint32_t space_used, uint32_t space_total, + uint32_t inodes_used, uint32_t inodes_total, + uint32_t reserved, uint32_t flags) +{ + struct nameidata nd; + int ret; + + ret = user_path_walk_link(name, &nd); + if (!ret) { + struct super_block *sb; + struct dl_info *dli; + + ret = -EINVAL; + if (!nd.dentry->d_inode) + goto out_release; + if (!(sb = nd.dentry->d_inode->i_sb)) + goto out_release; + if ((reserved != (uint32_t)CDLIM_KEEP && + reserved > 100) || + (inodes_used != (uint32_t)CDLIM_KEEP && + inodes_used > inodes_total) || + (space_used != (uint32_t)CDLIM_KEEP && + space_used > space_total)) + goto out_release; + + ret = -ESRCH; + dli = locate_dl_info(sb, id); + if (!dli) + goto out_release; + + spin_lock(&dli->dl_lock); + + if (inodes_used != (uint32_t)CDLIM_KEEP) + dli->dl_inodes_used = inodes_used; + if (inodes_total != (uint32_t)CDLIM_KEEP) + dli->dl_inodes_total = inodes_total; + if (space_used != (uint32_t)CDLIM_KEEP) { + dli->dl_space_used = space_used; + dli->dl_space_used <<= 10; + } + if (space_total == (uint32_t)CDLIM_INFINITY) + dli->dl_space_total = (uint64_t)CDLIM_INFINITY; + else if (space_total != (uint32_t)CDLIM_KEEP) { + dli->dl_space_total = space_total; + dli->dl_space_total <<= 10; + } + if (reserved != (uint32_t)CDLIM_KEEP) + dli->dl_nrlmult = (1 << 10) * (100 - reserved) / 100; + + spin_unlock(&dli->dl_lock); + + put_dl_info(dli); + ret = 0; + + out_release: + path_release(&nd); + } + return ret; +} + +int vc_set_dlimit(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_v0 vc_data; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_dlimit(id, vc_data.name, + vc_data.space_used, vc_data.space_total, + vc_data.inodes_used, vc_data.inodes_total, + vc_data.reserved, vc_data.flags); +} + +#ifdef CONFIG_COMPAT + +int vc_set_dlimit_x32(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_v0_x32 vc_data; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_dlimit(id, compat_ptr(vc_data.name_ptr), + vc_data.space_used, vc_data.space_total, + vc_data.inodes_used, vc_data.inodes_total, + vc_data.reserved, vc_data.flags); +} + +#endif /* CONFIG_COMPAT */ + + +static inline +int do_get_dlimit(uint32_t id, const char __user *name, + uint32_t *space_used, uint32_t *space_total, + uint32_t *inodes_used, uint32_t *inodes_total, + uint32_t *reserved, uint32_t *flags) +{ + struct nameidata nd; + int ret; + + ret = user_path_walk_link(name, &nd); + if (!ret) { + struct super_block *sb; + struct dl_info *dli; + + ret = -EINVAL; + if (!nd.dentry->d_inode) + goto out_release; + if (!(sb = nd.dentry->d_inode->i_sb)) + goto out_release; + + ret = -ESRCH; + dli = locate_dl_info(sb, id); + if (!dli) + goto out_release; + + spin_lock(&dli->dl_lock); + *inodes_used = dli->dl_inodes_used; + *inodes_total = dli->dl_inodes_total; + *space_used = dli->dl_space_used >> 10; + if (dli->dl_space_total == (uint64_t)CDLIM_INFINITY) + *space_total = (uint32_t)CDLIM_INFINITY; + else + *space_total = dli->dl_space_total >> 10; + + *reserved = 100 - ((dli->dl_nrlmult * 100 + 512) >> 10); + spin_unlock(&dli->dl_lock); + + put_dl_info(dli); + ret = -EFAULT; + + ret = 0; + out_release: + path_release(&nd); + } + return ret; +} + + +int vc_get_dlimit(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_v0 vc_data; + int ret; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_get_dlimit(id, vc_data.name, + &vc_data.space_used, &vc_data.space_total, + &vc_data.inodes_used, &vc_data.inodes_total, + &vc_data.reserved, &vc_data.flags); + if (ret) + return ret; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +#ifdef CONFIG_COMPAT + +int vc_get_dlimit_x32(uint32_t id, void __user *data) +{ + struct vcmd_ctx_dlimit_v0_x32 vc_data; + int ret; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_get_dlimit(id, compat_ptr(vc_data.name_ptr), + &vc_data.space_used, &vc_data.space_total, + &vc_data.inodes_used, &vc_data.inodes_total, + &vc_data.reserved, &vc_data.flags); + if (ret) + return ret; + + if (copy_to_user(data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +#endif /* CONFIG_COMPAT */ + + +void vx_vsi_statfs(struct super_block *sb, struct kstatfs *buf) +{ + struct dl_info *dli; + __u64 blimit, bfree, bavail; + __u32 ifree; + + dli = locate_dl_info(sb, vx_current_xid()); + if (!dli) + return; + + spin_lock(&dli->dl_lock); + if (dli->dl_inodes_total == (uint32_t)CDLIM_INFINITY) + goto no_ilim; + + /* reduce max inodes available to limit */ + if (buf->f_files > dli->dl_inodes_total) + buf->f_files = dli->dl_inodes_total; + + /* inode hack for reiserfs */ + if ((buf->f_files == 0) && (dli->dl_inodes_total > 0)) { + buf->f_files = dli->dl_inodes_total; + buf->f_ffree = dli->dl_inodes_total; + } + + ifree = dli->dl_inodes_total - dli->dl_inodes_used; + /* reduce free inodes to min */ + if (ifree < buf->f_ffree) + buf->f_ffree = ifree; + +no_ilim: + if (dli->dl_space_total == (uint64_t)CDLIM_INFINITY) + goto no_blim; + + blimit = dli->dl_space_total >> sb->s_blocksize_bits; + + if (dli->dl_space_total < dli->dl_space_used) + bfree = 0; + else + bfree = (dli->dl_space_total - dli->dl_space_used) + >> sb->s_blocksize_bits; + + bavail = ((dli->dl_space_total >> 10) * dli->dl_nrlmult); + if (bavail < dli->dl_space_used) + bavail = 0; + else + bavail = (bavail - dli->dl_space_used) + >> sb->s_blocksize_bits; + + /* reduce max space available to limit */ + if (buf->f_blocks > blimit) + buf->f_blocks = blimit; + + /* reduce free space to min */ + if (bfree < buf->f_bfree) + buf->f_bfree = bfree; + + /* reduce avail space to min */ + if (bavail < buf->f_bavail) + buf->f_bavail = bavail; + +no_blim: + spin_unlock(&dli->dl_lock); + put_dl_info(dli); + + return; +} + +#include + +EXPORT_SYMBOL_GPL(locate_dl_info); +EXPORT_SYMBOL_GPL(rcu_free_dl_info); + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/helper.c linux-2.6.17.13-g-v/kernel/vserver/helper.c --- linux-2.6.17.13/kernel/vserver/helper.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/helper.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,210 @@ +/* + * linux/kernel/vserver/helper.c + * + * Virtual Context Support + * + * Copyright (C) 2004-2005 Herbert Pötzl + * + * V0.01 basic helper + * + */ + +#include +#include +#include +#include +#include +#include +#include + +#include +#include + + +char vshelper_path[255] = "/sbin/vshelper"; + + +static int do_vshelper(char *name, char *argv[], char *envp[], int sync) +{ + int ret; + + if ((ret = call_usermodehelper(name, argv, envp, sync))) { + printk( KERN_WARNING + "%s: (%s %s) returned %s with %d\n", + name, argv[1], argv[2], + sync?"sync":"async", ret); + } + vxdprintk(VXD_CBIT(switch, 4), + "%s: (%s %s) returned %s with %d", + name, argv[1], argv[2], sync?"sync":"async", ret); + return ret; +} + +/* + * vshelper path is set via /proc/sys + * invoked by vserver sys_reboot(), with + * the following arguments + * + * argv [0] = vshelper_path; + * argv [1] = action: "restart", "halt", "poweroff", ... + * argv [2] = context identifier + * + * envp [*] = type-specific parameters + */ + +long vs_reboot_helper(struct vx_info *vxi, int cmd, void __user *arg) +{ + char id_buf[8], cmd_buf[16]; + char uid_buf[16], pid_buf[16]; + int ret; + + char *argv[] = {vshelper_path, NULL, id_buf, 0}; + char *envp[] = {"HOME=/", "TERM=linux", + "PATH=/sbin:/usr/sbin:/bin:/usr/bin", + uid_buf, pid_buf, cmd_buf, 0}; + + if (vx_info_state(vxi, VXS_HELPER)) + return -EAGAIN; + vxi->vx_state |= VXS_HELPER; + + snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id); + + snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd); + snprintf(uid_buf, sizeof(uid_buf)-1, "VS_UID=%d", current->uid); + snprintf(pid_buf, sizeof(pid_buf)-1, "VS_PID=%d", current->pid); + + switch (cmd) { + case LINUX_REBOOT_CMD_RESTART: + argv[1] = "restart"; + break; + + case LINUX_REBOOT_CMD_HALT: + argv[1] = "halt"; + break; + + case LINUX_REBOOT_CMD_POWER_OFF: + argv[1] = "poweroff"; + break; + + case LINUX_REBOOT_CMD_SW_SUSPEND: + argv[1] = "swsusp"; + break; + + default: + vxi->vx_state &= ~VXS_HELPER; + return 0; + } + +#ifndef CONFIG_VSERVER_LEGACY + ret = do_vshelper(vshelper_path, argv, envp, 1); +#else + ret = do_vshelper(vshelper_path, argv, envp, 0); +#endif + vxi->vx_state &= ~VXS_HELPER; + __wakeup_vx_info(vxi); + return (ret) ? -EPERM : 0; +} + + +long vs_reboot(unsigned int cmd, void __user *arg) +{ + struct vx_info *vxi = current->vx_info; + long ret = 0; + + vxdprintk(VXD_CBIT(misc, 5), + "vs_reboot(%p[#%d],%d)", + vxi, vxi?vxi->vx_id:0, cmd); + + ret = vs_reboot_helper(vxi, cmd, arg); + if (ret) + return ret; + + vxi->reboot_cmd = cmd; + if (vx_info_flags(vxi, VXF_REBOOT_KILL, 0)) { + switch (cmd) { + case LINUX_REBOOT_CMD_RESTART: + case LINUX_REBOOT_CMD_HALT: + case LINUX_REBOOT_CMD_POWER_OFF: + vx_info_kill(vxi, 0, SIGKILL); + vx_info_kill(vxi, 1, SIGKILL); + default: + break; + } + } + return 0; +} + + +/* + * argv [0] = vshelper_path; + * argv [1] = action: "startup", "shutdown" + * argv [2] = context identifier + * + * envp [*] = type-specific parameters + */ + +long vs_state_change(struct vx_info *vxi, unsigned int cmd) +{ + char id_buf[8], cmd_buf[16]; + char *argv[] = {vshelper_path, NULL, id_buf, 0}; + char *envp[] = {"HOME=/", "TERM=linux", + "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0}; + + if (!vx_info_flags(vxi, VXF_SC_HELPER, 0)) + return 0; + + snprintf(id_buf, sizeof(id_buf)-1, "%d", vxi->vx_id); + snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd); + + switch (cmd) { + case VSC_STARTUP: + argv[1] = "startup"; + break; + case VSC_SHUTDOWN: + argv[1] = "shutdown"; + break; + default: + return 0; + } + + do_vshelper(vshelper_path, argv, envp, 1); + return 0; +} + + +/* + * argv [0] = vshelper_path; + * argv [1] = action: "netup", "netdown" + * argv [2] = context identifier + * + * envp [*] = type-specific parameters + */ + +long vs_net_change(struct nx_info *nxi, unsigned int cmd) +{ + char id_buf[8], cmd_buf[16]; + char *argv[] = {vshelper_path, NULL, id_buf, 0}; + char *envp[] = {"HOME=/", "TERM=linux", + "PATH=/sbin:/usr/sbin:/bin:/usr/bin", cmd_buf, 0}; + + if (!nx_info_flags(nxi, NXF_SC_HELPER, 0)) + return 0; + + snprintf(id_buf, sizeof(id_buf)-1, "%d", nxi->nx_id); + snprintf(cmd_buf, sizeof(cmd_buf)-1, "VS_CMD=%08x", cmd); + + switch (cmd) { + case VSC_NETUP: + argv[1] = "netup"; + break; + case VSC_NETDOWN: + argv[1] = "netdown"; + break; + default: + return 0; + } + + do_vshelper(vshelper_path, argv, envp, 1); + return 0; +} + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/history.c linux-2.6.17.13-g-v/kernel/vserver/history.c --- linux-2.6.17.13/kernel/vserver/history.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/history.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,183 @@ +/* + * kernel/vserver/history.c + * + * Virtual Context History Backtrace + * + * Copyright (C) 2004-2005 Herbert Pötzl + * + * V0.01 basic structure + * V0.02 hash/unhash and trace + * V0.03 preemption fixes + * + */ + +#include +#include +#include +#include + +#include +#include +#include + +#include + + +#ifdef CONFIG_VSERVER_HISTORY +#define VXH_SIZE CONFIG_VSERVER_HISTORY_SIZE +#else +#define VXH_SIZE 64 +#endif + +struct _vx_history { + unsigned int counter; + + struct _vx_hist_entry entry[VXH_SIZE+1]; +}; + + +DEFINE_PER_CPU(struct _vx_history, vx_history_buffer); + +unsigned volatile int vxh_active = 1; + +static atomic_t sequence = ATOMIC_INIT(0); + + +/* vxh_advance() + + * requires disabled preemption */ + +struct _vx_hist_entry *vxh_advance(void *loc) +{ + unsigned int cpu = smp_processor_id(); + struct _vx_history *hist = &per_cpu(vx_history_buffer, cpu); + struct _vx_hist_entry *entry; + unsigned int index; + + index = vxh_active ? (hist->counter++ % VXH_SIZE) : VXH_SIZE; + entry = &hist->entry[index]; + + entry->seq = atomic_inc_return(&sequence); + entry->loc = loc; + return entry; +} + + +#define VXH_LOC_FMTS "(#%04x,*%d):%p" + +#define VXH_LOC_ARGS(e) (e)->seq, cpu, (e)->loc + + +#define VXH_VXI_FMTS "%p[#%d,%d.%d]" + +#define VXH_VXI_ARGS(e) (e)->vxi.ptr, \ + (e)->vxi.ptr?(e)->vxi.xid:0, \ + (e)->vxi.ptr?(e)->vxi.usecnt:0, \ + (e)->vxi.ptr?(e)->vxi.tasks:0 + +void vxh_dump_entry(struct _vx_hist_entry *e, unsigned cpu) +{ + switch (e->type) { + case VXH_THROW_OOPS: + printk( VXH_LOC_FMTS " oops \n", VXH_LOC_ARGS(e)); + break; + + case VXH_GET_VX_INFO: + case VXH_PUT_VX_INFO: + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n", + VXH_LOC_ARGS(e), + (e->type==VXH_GET_VX_INFO)?"get":"put", + VXH_VXI_ARGS(e)); + break; + + case VXH_INIT_VX_INFO: + case VXH_SET_VX_INFO: + case VXH_CLR_VX_INFO: + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n", + VXH_LOC_ARGS(e), + (e->type==VXH_INIT_VX_INFO)?"init": + ((e->type==VXH_SET_VX_INFO)?"set":"clr"), + VXH_VXI_ARGS(e), e->sc.data); + break; + + case VXH_CLAIM_VX_INFO: + case VXH_RELEASE_VX_INFO: + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS " @%p\n", + VXH_LOC_ARGS(e), + (e->type==VXH_CLAIM_VX_INFO)?"claim":"release", + VXH_VXI_ARGS(e), e->sc.data); + break; + + case VXH_ALLOC_VX_INFO: + case VXH_DEALLOC_VX_INFO: + printk( VXH_LOC_FMTS " %s_vx_info " VXH_VXI_FMTS "\n", + VXH_LOC_ARGS(e), + (e->type==VXH_ALLOC_VX_INFO)?"alloc":"dealloc", + VXH_VXI_ARGS(e)); + break; + + case VXH_HASH_VX_INFO: + case VXH_UNHASH_VX_INFO: + printk( VXH_LOC_FMTS " __%s_vx_info " VXH_VXI_FMTS "\n", + VXH_LOC_ARGS(e), + (e->type==VXH_HASH_VX_INFO)?"hash":"unhash", + VXH_VXI_ARGS(e)); + break; + + case VXH_LOC_VX_INFO: + case VXH_LOOKUP_VX_INFO: + case VXH_CREATE_VX_INFO: + printk( VXH_LOC_FMTS " __%s_vx_info [#%d] -> " VXH_VXI_FMTS "\n", + VXH_LOC_ARGS(e), + (e->type==VXH_CREATE_VX_INFO)?"create": + ((e->type==VXH_LOC_VX_INFO)?"loc":"lookup"), + e->ll.arg, VXH_VXI_ARGS(e)); + break; + } +} + +static void __vxh_dump_history(void) +{ + unsigned int i,j; + + printk("History:\tSEQ: %8x\tNR_CPUS: %d\n", + atomic_read(&sequence), NR_CPUS); + + for (i=0; i < VXH_SIZE; i++) { + for (j=0; j < NR_CPUS; j++) { + struct _vx_history *hist = + &per_cpu(vx_history_buffer, j); + unsigned int index = (hist->counter-i) % VXH_SIZE; + struct _vx_hist_entry *entry = &hist->entry[index]; + + vxh_dump_entry(entry, j); + } + } +} + +void vxh_dump_history(void) +{ + vxh_active = 0; +#ifdef CONFIG_SMP + local_irq_enable(); + smp_send_stop(); + local_irq_disable(); +#endif + __vxh_dump_history(); +} + + +/* vserver syscall commands below here */ + + +int vc_dump_history(uint32_t id) +{ + vxh_active = 0; + __vxh_dump_history(); + vxh_active = 1; + + return 0; +} + +EXPORT_SYMBOL_GPL(vxh_advance); + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/init.c linux-2.6.17.13-g-v/kernel/vserver/init.c --- linux-2.6.17.13/kernel/vserver/init.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/init.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,43 @@ +/* + * linux/kernel/init.c + * + * Virtual Server Init + * + * Copyright (C) 2004-2005 Herbert Pötzl + * + * V0.01 basic structure + * + */ + +#include +#include +#include + +int vserver_register_sysctl(void); +void vserver_unregister_sysctl(void); + + +static int __init init_vserver(void) +{ + int ret = 0; + +#ifdef CONFIG_VSERVER_DEBUG + vserver_register_sysctl(); +#endif + return ret; +} + + +static void __exit exit_vserver(void) +{ + +#ifdef CONFIG_VSERVER_DEBUG + vserver_unregister_sysctl(); +#endif + return; +} + + +module_init(init_vserver); +module_exit(exit_vserver); + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/inode.c linux-2.6.17.13-g-v/kernel/vserver/inode.c --- linux-2.6.17.13/kernel/vserver/inode.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/inode.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,368 @@ +/* + * linux/kernel/vserver/inode.c + * + * Virtual Server: File System Support + * + * Copyright (C) 2004-2005 Herbert Pötzl + * + * V0.01 separated from vcontext V0.05 + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + + +static int __vc_get_iattr(struct inode *in, uint32_t *xid, uint32_t *flags, uint32_t *mask) +{ + struct proc_dir_entry *entry; + + if (!in || !in->i_sb) + return -ESRCH; + + *flags = IATTR_XID + | (IS_BARRIER(in) ? IATTR_BARRIER : 0) + | (IS_IUNLINK(in) ? IATTR_IUNLINK : 0) + | (IS_IMMUTABLE(in) ? IATTR_IMMUTABLE : 0); + *mask = IATTR_IUNLINK | IATTR_IMMUTABLE; + + if (S_ISDIR(in->i_mode)) + *mask |= IATTR_BARRIER; + + if (IS_TAGXID(in)) { + *xid = in->i_xid; + *mask |= IATTR_XID; + } + + switch (in->i_sb->s_magic) { + case PROC_SUPER_MAGIC: + entry = PROC_I(in)->pde; + + /* check for specific inodes? */ + if (entry) + *mask |= IATTR_FLAGS; + if (entry) + *flags |= (entry->vx_flags & IATTR_FLAGS); + else + *flags |= (PROC_I(in)->vx_flags & IATTR_FLAGS); + break; + + case DEVPTS_SUPER_MAGIC: + *xid = in->i_xid; + *mask |= IATTR_XID; + break; + + default: + break; + } + return 0; +} + +int vc_get_iattr(uint32_t id, void __user *data) +{ + struct nameidata nd; + struct vcmd_ctx_iattr_v1 vc_data = { .xid = -1 }; + int ret; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = user_path_walk_link(vc_data.name, &nd); + if (!ret) { + ret = __vc_get_iattr(nd.dentry->d_inode, + &vc_data.xid, &vc_data.flags, &vc_data.mask); + path_release(&nd); + } + if (ret) + return ret; + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + ret = -EFAULT; + return ret; +} + +#ifdef CONFIG_COMPAT + +int vc_get_iattr_x32(uint32_t id, void __user *data) +{ + struct nameidata nd; + struct vcmd_ctx_iattr_v1_x32 vc_data = { .xid = -1 }; + int ret; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = user_path_walk_link(compat_ptr(vc_data.name_ptr), &nd); + if (!ret) { + ret = __vc_get_iattr(nd.dentry->d_inode, + &vc_data.xid, &vc_data.flags, &vc_data.mask); + path_release(&nd); + } + if (ret) + return ret; + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + ret = -EFAULT; + return ret; +} + +#endif /* CONFIG_COMPAT */ + + +static int __vc_set_iattr(struct dentry *de, uint32_t *xid, uint32_t *flags, uint32_t *mask) +{ + struct inode *in = de->d_inode; + int error = 0, is_proc = 0, has_xid = 0; + struct iattr attr = { 0 }; + + if (!in || !in->i_sb) + return -ESRCH; + + is_proc = (in->i_sb->s_magic == PROC_SUPER_MAGIC); + if ((*mask & IATTR_FLAGS) && !is_proc) + return -EINVAL; + + has_xid = IS_TAGXID(in) || + (in->i_sb->s_magic == DEVPTS_SUPER_MAGIC); + if ((*mask & IATTR_XID) && !has_xid) + return -EINVAL; + + mutex_lock(&in->i_mutex); + if (*mask & IATTR_XID) { + attr.ia_xid = *xid; + attr.ia_valid |= ATTR_XID; + } + + if (*mask & IATTR_FLAGS) { + struct proc_dir_entry *entry = PROC_I(in)->pde; + unsigned int iflags = PROC_I(in)->vx_flags; + + iflags = (iflags & ~(*mask & IATTR_FLAGS)) + | (*flags & IATTR_FLAGS); + PROC_I(in)->vx_flags = iflags; + if (entry) + entry->vx_flags = iflags; + } + + if (*mask & (IATTR_BARRIER | IATTR_IUNLINK | IATTR_IMMUTABLE)) { + if (*mask & IATTR_IMMUTABLE) { + if (*flags & IATTR_IMMUTABLE) + in->i_flags |= S_IMMUTABLE; + else + in->i_flags &= ~S_IMMUTABLE; + } + if (*mask & IATTR_IUNLINK) { + if (*flags & IATTR_IUNLINK) + in->i_flags |= S_IUNLINK; + else + in->i_flags &= ~S_IUNLINK; + } + if (S_ISDIR(in->i_mode) && (*mask & IATTR_BARRIER)) { + if (*flags & IATTR_BARRIER) + in->i_flags |= S_BARRIER; + else + in->i_flags &= ~S_BARRIER; + } + if (in->i_op && in->i_op->sync_flags) { + error = in->i_op->sync_flags(in); + if (error) + goto out; + } + } + + if (attr.ia_valid) { + if (in->i_op && in->i_op->setattr) + error = in->i_op->setattr(de, &attr); + else { + error = inode_change_ok(in, &attr); + if (!error) + error = inode_setattr(in, &attr); + } + } + +out: + mutex_unlock(&in->i_mutex); + return error; +} + +int vc_set_iattr(uint32_t id, void __user *data) +{ + struct nameidata nd; + struct vcmd_ctx_iattr_v1 vc_data; + int ret; + + if (!capable(CAP_SYS_ADMIN) || !capable(CAP_LINUX_IMMUTABLE)) + return -EPERM; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = user_path_walk_link(vc_data.name, &nd); + if (!ret) { + ret = __vc_set_iattr(nd.dentry, + &vc_data.xid, &vc_data.flags, &vc_data.mask); + path_release(&nd); + } + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + ret = -EFAULT; + return ret; +} + +#ifdef CONFIG_COMPAT + +int vc_set_iattr_x32(uint32_t id, void __user *data) +{ + struct nameidata nd; + struct vcmd_ctx_iattr_v1_x32 vc_data; + int ret; + + if (!capable(CAP_SYS_ADMIN) || !capable(CAP_LINUX_IMMUTABLE)) + return -EPERM; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = user_path_walk_link(compat_ptr(vc_data.name_ptr), &nd); + if (!ret) { + ret = __vc_set_iattr(nd.dentry, + &vc_data.xid, &vc_data.flags, &vc_data.mask); + path_release(&nd); + } + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + ret = -EFAULT; + return ret; +} + +#endif /* CONFIG_COMPAT */ + +#ifdef CONFIG_VSERVER_LEGACY + +#define PROC_DYNAMIC_FIRST 0xF0000000UL + +int vx_proc_ioctl(struct inode * inode, struct file * filp, + unsigned int cmd, unsigned long arg) +{ + struct proc_dir_entry *entry; + int error = 0; + int flags; + + if (inode->i_ino < PROC_DYNAMIC_FIRST) + return -ENOTTY; + + entry = PROC_I(inode)->pde; + if (!entry) + return -ENOTTY; + + switch(cmd) { + case FIOC_GETXFLG: { + /* fixme: if stealth, return -ENOTTY */ + error = -EPERM; + flags = entry->vx_flags; + if (capable(CAP_CONTEXT)) + error = put_user(flags, (int __user *) arg); + break; + } + case FIOC_SETXFLG: { + /* fixme: if stealth, return -ENOTTY */ + error = -EPERM; + if (!capable(CAP_CONTEXT)) + break; + error = -EROFS; + if (IS_RDONLY(inode)) + break; + error = -EFAULT; + if (get_user(flags, (int __user *) arg)) + break; + error = 0; + entry->vx_flags = flags; + break; + } + default: + return -ENOTTY; + } + return error; +} +#endif + + +int vx_parse_xid(char *string, xid_t *xid, int remove) +{ + static match_table_t tokens = { + {1, "xid=%u"}, + {0, NULL} + }; + substring_t args[MAX_OPT_ARGS]; + int token, option = 0; + + if (!string) + return 0; + + token = match_token(string, tokens, args); + if (token && xid && !match_int(args, &option)) + *xid = option; + + vxdprintk(VXD_CBIT(xid, 7), + "vx_parse_xid(»%s«): %d:#%d", + string, token, option); + + if (token && remove) { + char *p = strstr(string, "xid="); + char *q = p; + + if (p) { + while (*q != '\0' && *q != ',') + q++; + while (*q) + *p++ = *q++; + while (*p) + *p++ = '\0'; + } + } + return token; +} + +void vx_propagate_xid(struct nameidata *nd, struct inode *inode) +{ + xid_t new_xid = 0; + struct vfsmount *mnt; + int propagate; + + if (!nd) + return; + mnt = nd->mnt; + if (!mnt) + return; + + propagate = (mnt->mnt_flags & MNT_XID); + if (propagate) + new_xid = mnt->mnt_xid; + + vxdprintk(VXD_CBIT(xid, 7), + "vx_propagate_xid(%p[#%lu.%d]): %d,%d", + inode, inode->i_ino, inode->i_xid, + new_xid, (propagate)?1:0); + + if (propagate) + inode->i_xid = new_xid; +} + +#include + +EXPORT_SYMBOL_GPL(vx_propagate_xid); + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/legacy.c linux-2.6.17.13-g-v/kernel/vserver/legacy.c --- linux-2.6.17.13/kernel/vserver/legacy.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/legacy.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,113 @@ +/* + * linux/kernel/vserver/legacy.c + * + * Virtual Server: Legacy Funtions + * + * Copyright (C) 2001-2003 Jacques Gelinas + * Copyright (C) 2003-2005 Herbert Pötzl + * + * V0.01 broken out from vcontext.c V0.05 + * + */ + +#include +#include +#include +#include +#include +#include + +#include +#include + + +extern int vx_set_init(struct vx_info *, struct task_struct *); + +static int vx_set_initpid(struct vx_info *vxi, int pid) +{ + struct task_struct *init; + + init = find_task_by_real_pid(pid); + if (!init) + return -ESRCH; + return vx_set_init(vxi, init); +} + +int vc_new_s_context(uint32_t ctx, void __user *data) +{ + int ret = -ENOMEM; + struct vcmd_new_s_context_v1 vc_data; + struct vx_info *new_vxi; + + if (copy_from_user(&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + /* legacy hack, will be removed soon */ + if (ctx == -2) { + /* assign flags and initpid */ + if (!current->vx_info) + return -EINVAL; + ret = 0; + if (vc_data.flags & VX_INFO_INIT) + ret = vx_set_initpid(current->vx_info, current->tgid); + if (ret == 0) { + /* We keep the same vx_id, but lower the capabilities */ + current->vx_info->vx_bcaps &= (~vc_data.remove_cap); + ret = vx_current_xid(); + current->vx_info->vx_flags |= vc_data.flags; + } + return ret; + } + + if (!vx_check(0, VX_ADMIN) || !capable(CAP_SYS_ADMIN) + /* might make sense in the future, or not ... */ + || vx_flags(VX_INFO_LOCK, 0)) + return -EPERM; + + /* ugly hack for Spectator */ + if (ctx == 1) { + current->xid = 1; + return 0; + } + + if (((ctx > MAX_S_CONTEXT) && (ctx != VX_DYNAMIC_ID)) || + (ctx == 0)) + return -EINVAL; + + if ((ctx == VX_DYNAMIC_ID) || (ctx < MIN_D_CONTEXT)) + new_vxi = lookup_or_create_vx_info(ctx); + else + new_vxi = lookup_vx_info(ctx); + + if (!new_vxi) + return -EINVAL; + + ret = -EPERM; + if (!vx_info_flags(new_vxi, VXF_STATE_SETUP, 0) && + vx_info_flags(new_vxi, VX_INFO_PRIVATE, 0)) + goto out_put; + + new_vxi->vx_flags &= ~VXF_STATE_SETUP; + + ret = vx_migrate_task(current, new_vxi); + if (ret == 0) { + current->vx_info->vx_bcaps &= (~vc_data.remove_cap); + new_vxi->vx_flags |= vc_data.flags; + if (vc_data.flags & VX_INFO_INIT) + vx_set_initpid(new_vxi, current->tgid); + if (vc_data.flags & VX_INFO_NAMESPACE) + vx_set_namespace(new_vxi, + current->namespace, current->fs); + if (vc_data.flags & VX_INFO_NPROC) + new_vxi->limit.rlim[RLIMIT_NPROC] = + current->signal->rlim[RLIMIT_NPROC].rlim_max; + + /* tweak some defaults for legacy */ + new_vxi->vx_flags |= (VXF_HIDE_NETIF|VXF_INFO_INIT); + ret = new_vxi->vx_id; + } +out_put: + put_vx_info(new_vxi); + return ret; +} + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/legacynet.c linux-2.6.17.13-g-v/kernel/vserver/legacynet.c --- linux-2.6.17.13/kernel/vserver/legacynet.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/legacynet.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,84 @@ + +/* + * linux/kernel/vserver/legacynet.c + * + * Virtual Server: Legacy Network Funtions + * + * Copyright (C) 2001-2003 Jacques Gelinas + * Copyright (C) 2003-2005 Herbert Pötzl + * + * V0.01 broken out from legacy.c + * + */ + +#include +#include +#include +#include +#include +#include + +#include +#include + + +extern struct nx_info *create_nx_info(void); + +/* set ipv4 root (syscall) */ + +int vc_set_ipv4root(uint32_t nbip, void __user *data) +{ + int i, err = -EPERM; + struct vcmd_set_ipv4root_v3 vc_data; + struct nx_info *new_nxi, *nxi = current->nx_info; + + if (nbip < 0 || nbip > NB_IPV4ROOT) + return -EINVAL; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + if (!nxi || nxi->ipv4[0] == 0 || capable(CAP_NET_ADMIN)) + /* We are allowed to change everything */ + err = 0; + else if (nxi) { + int found = 0; + + /* We are allowed to select a subset of the currently + installed IP numbers. No new one are allowed + We can't change the broadcast address though */ + for (i=0; inbipv4; j++) { + if (nxip == nxi->ipv4[j]) { + found++; + break; + } + } + } + if ((found == nbip) && + (vc_data.broadcast == nxi->v4_bcast)) + err = 0; + } + if (err) + return err; + + new_nxi = create_nx_info(); + if (IS_ERR(new_nxi)) + return -EINVAL; + + new_nxi->nbipv4 = nbip; + for (i=0; iipv4[i] = vc_data.nx_mask_pair[i].ip; + new_nxi->mask[i] = vc_data.nx_mask_pair[i].mask; + } + new_nxi->v4_bcast = vc_data.broadcast; + if (nxi) + printk("!!! switching nx_info %p->%p\n", nxi, new_nxi); + + nx_migrate_task(current, new_nxi); + put_nx_info(new_nxi); + return 0; +} + + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/limit.c linux-2.6.17.13-g-v/kernel/vserver/limit.c --- linux-2.6.17.13/kernel/vserver/limit.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/limit.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,238 @@ +/* + * linux/kernel/vserver/limit.c + * + * Virtual Server: Context Limits + * + * Copyright (C) 2004-2005 Herbert Pötzl + * + * V0.01 broken out from vcontext V0.05 + * + */ + +#include +#include +#include +#include +#include +#include + +#include +#include + + +const char *vlimit_name[NUM_LIMITS] = { + [RLIMIT_CPU] = "CPU", + [RLIMIT_RSS] = "RSS", + [RLIMIT_NPROC] = "NPROC", + [RLIMIT_NOFILE] = "NOFILE", + [RLIMIT_MEMLOCK] = "VML", + [RLIMIT_AS] = "VM", + [RLIMIT_LOCKS] = "LOCKS", + [RLIMIT_SIGPENDING] = "SIGP", + [RLIMIT_MSGQUEUE] = "MSGQ", + + [VLIMIT_NSOCK] = "NSOCK", + [VLIMIT_OPENFD] = "OPENFD", + [VLIMIT_ANON] = "ANON", + [VLIMIT_SHMEM] = "SHMEM", +}; + +EXPORT_SYMBOL_GPL(vlimit_name); + + +static int is_valid_rlimit(int id) +{ + int valid = 0; + + switch (id) { + case RLIMIT_RSS: + case RLIMIT_NPROC: + case RLIMIT_NOFILE: + case RLIMIT_MEMLOCK: + case RLIMIT_AS: + + case VLIMIT_NSOCK: + case VLIMIT_OPENFD: + case VLIMIT_ANON: + case VLIMIT_SHMEM: + valid = 1; + break; + } + return valid; +} + +static inline uint64_t vc_get_rlim(struct vx_info *vxi, int id) +{ + unsigned long limit; + + limit = vxi->limit.rlim[id]; + if (limit == RLIM_INFINITY) + return CRLIM_INFINITY; + return limit; +} + +static int do_get_rlimit(xid_t xid, uint32_t id, + uint64_t *minimum, uint64_t *softlimit, uint64_t *maximum) +{ + struct vx_info *vxi; + + if (!is_valid_rlimit(id)) + return -EINVAL; + + vxi = lookup_vx_info(xid); + if (!vxi) + return -ESRCH; + + if (minimum) + *minimum = CRLIM_UNSET; + if (softlimit) + *softlimit = CRLIM_UNSET; + if (maximum) + *maximum = vc_get_rlim(vxi, id); + put_vx_info(vxi); + return 0; +} + +int vc_get_rlimit(uint32_t id, void __user *data) +{ + struct vcmd_ctx_rlimit_v0 vc_data; + int ret; + + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_get_rlimit(id, vc_data.id, + &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum); + if (ret) + return ret; + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +static int do_set_rlimit(xid_t xid, uint32_t id, + uint64_t minimum, uint64_t softlimit, uint64_t maximum) +{ + struct vx_info *vxi; + + if (!is_valid_rlimit(id)) + return -EINVAL; + + vxi = lookup_vx_info(xid); + if (!vxi) + return -ESRCH; + + if (maximum != CRLIM_KEEP) + vxi->limit.rlim[id] = maximum; + + put_vx_info(vxi); + return 0; +} + +int vc_set_rlimit(uint32_t id, void __user *data) +{ + struct vcmd_ctx_rlimit_v0 vc_data; + + if (!capable(CAP_SYS_ADMIN) || !capable(CAP_SYS_RESOURCE)) + return -EPERM; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_rlimit(id, vc_data.id, + vc_data.minimum, vc_data.softlimit, vc_data.maximum); +} + +#ifdef CONFIG_IA32_EMULATION + +int vc_set_rlimit_x32(uint32_t id, void __user *data) +{ + struct vcmd_ctx_rlimit_v0_x32 vc_data; + + if (!capable(CAP_SYS_ADMIN) || !capable(CAP_SYS_RESOURCE)) + return -EPERM; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + return do_set_rlimit(id, vc_data.id, + vc_data.minimum, vc_data.softlimit, vc_data.maximum); +} + +int vc_get_rlimit_x32(uint32_t id, void __user *data) +{ + struct vcmd_ctx_rlimit_v0_x32 vc_data; + int ret; + + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + ret = do_get_rlimit(id, vc_data.id, + &vc_data.minimum, &vc_data.softlimit, &vc_data.maximum); + if (ret) + return ret; + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +#endif /* CONFIG_IA32_EMULATION */ + + +int vc_get_rlimit_mask(uint32_t id, void __user *data) +{ + static struct vcmd_ctx_rlimit_mask_v0 mask = { + /* minimum */ + 0 + , /* softlimit */ + 0 + , /* maximum */ + (1 << RLIMIT_RSS) | + (1 << RLIMIT_NPROC) | + (1 << RLIMIT_NOFILE) | + (1 << RLIMIT_MEMLOCK) | + (1 << RLIMIT_LOCKS) | + (1 << RLIMIT_AS) | + (1 << VLIMIT_ANON) | + 0 + }; + + if (!capable(CAP_SYS_ADMIN) || !capable(CAP_SYS_RESOURCE)) + return -EPERM; + if (copy_to_user(data, &mask, sizeof(mask))) + return -EFAULT; + return 0; +} + + +void vx_vsi_meminfo(struct sysinfo *val) +{ + struct vx_info *vxi = current->vx_info; + unsigned long v; + + v = vxi->limit.rlim[RLIMIT_RSS]; + if (v != RLIM_INFINITY) + val->totalram = min(val->totalram, v); + v = atomic_read(&vxi->limit.rcur[RLIMIT_RSS]); + val->freeram = (v < val->totalram) ? val->totalram - v : 0; + val->bufferram = 0; + val->totalhigh = 0; + val->freehigh = 0; + return; +} + +void vx_vsi_swapinfo(struct sysinfo *val) +{ + struct vx_info *vxi = current->vx_info; + unsigned long v, w; + + v = vxi->limit.rlim[RLIMIT_RSS]; + w = vxi->limit.rlim[RLIMIT_AS]; + if (w != RLIM_INFINITY) + val->totalswap = min(val->totalswap, w - + ((v != RLIM_INFINITY) ? v : 0)); + w = atomic_read(&vxi->limit.rcur[RLIMIT_AS]); + val->freeswap = (w < val->totalswap) ? val->totalswap - w : 0; + return; +} + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/limit_init.h linux-2.6.17.13-g-v/kernel/vserver/limit_init.h --- linux-2.6.17.13/kernel/vserver/limit_init.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/limit_init.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,32 @@ + + +#include + + +static inline void vx_info_init_limit(struct _vx_limit *limit) +{ + int lim; + + for (lim=0; limrlim[lim] = RLIM_INFINITY; + limit->rmax[lim] = 0; + atomic_set(&limit->rcur[lim], 0); + atomic_set(&limit->lhit[lim], 0); + } +} + +static inline void vx_info_exit_limit(struct _vx_limit *limit) +{ +#ifdef CONFIG_VSERVER_DEBUG + unsigned long value; + unsigned int lim; + + for (lim=0; limrcur[lim]); + vxwprintk(value, + "!!! limit: %p[%s,%d] = %ld on exit.", + limit, vlimit_name[lim], lim, value); + } +#endif +} + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/limit_proc.h linux-2.6.17.13-g-v/kernel/vserver/limit_proc.h --- linux-2.6.17.13/kernel/vserver/limit_proc.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/limit_proc.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,58 @@ +#ifndef _VX_LIMIT_PROC_H +#define _VX_LIMIT_PROC_H + + +static inline void vx_limit_fixup(struct _vx_limit *limit) +{ + unsigned long value; + unsigned int lim; + + for (lim=0; limrcur[lim]); + if (value > limit->rmax[lim]) + limit->rmax[lim] = value; + if (limit->rmax[lim] > limit->rlim[lim]) + limit->rmax[lim] = limit->rlim[lim]; + } +} + +#define VX_LIMIT_FMT ":\t%10d\t%10ld\t%10ld\t%6d\n" + +#define VX_LIMIT_ARG(r) \ + ,atomic_read(&limit->rcur[r]) \ + ,limit->rmax[r] \ + ,limit->rlim[r] \ + ,atomic_read(&limit->lhit[r]) + +static inline int vx_info_proc_limit(struct _vx_limit *limit, char *buffer) +{ + vx_limit_fixup(limit); + return sprintf(buffer, + "PROC" VX_LIMIT_FMT + "VM" VX_LIMIT_FMT + "VML" VX_LIMIT_FMT + "RSS" VX_LIMIT_FMT + "ANON" VX_LIMIT_FMT + "FILES" VX_LIMIT_FMT + "OFD" VX_LIMIT_FMT + "LOCKS" VX_LIMIT_FMT + "SOCK" VX_LIMIT_FMT + "MSGQ" VX_LIMIT_FMT + "SHM" VX_LIMIT_FMT + VX_LIMIT_ARG(RLIMIT_NPROC) + VX_LIMIT_ARG(RLIMIT_AS) + VX_LIMIT_ARG(RLIMIT_MEMLOCK) + VX_LIMIT_ARG(RLIMIT_RSS) + VX_LIMIT_ARG(VLIMIT_ANON) + VX_LIMIT_ARG(RLIMIT_NOFILE) + VX_LIMIT_ARG(VLIMIT_OPENFD) + VX_LIMIT_ARG(RLIMIT_LOCKS) + VX_LIMIT_ARG(VLIMIT_NSOCK) + VX_LIMIT_ARG(RLIMIT_MSGQUEUE) + VX_LIMIT_ARG(VLIMIT_SHMEM) + ); +} + +#endif /* _VX_LIMIT_PROC_H */ + + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/namespace.c linux-2.6.17.13-g-v/kernel/vserver/namespace.c --- linux-2.6.17.13/kernel/vserver/namespace.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/namespace.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,113 @@ +/* + * linux/kernel/vserver/namespace.c + * + * Virtual Server: Context Namespace Support + * + * Copyright (C) 2003-2005 Herbert Pötzl + * + * V0.01 broken out from context.c 0.07 + * V0.02 added task locking for namespace + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + + +/* namespace functions */ + +#include + +int vx_set_namespace(struct vx_info *vxi, struct namespace *ns, struct fs_struct *fs) +{ + struct fs_struct *fs_copy; + + if (vxi->vx_namespace) + return -EPERM; + if (!ns || !fs) + return -EINVAL; + + fs_copy = copy_fs_struct(fs); + if (!fs_copy) + return -ENOMEM; + + get_namespace(ns); + vxi->vx_namespace = ns; + vxi->vx_fs = fs_copy; + return 0; +} + +int vc_enter_namespace(uint32_t id, void __user *data) +{ + struct vx_info *vxi; + struct fs_struct *old_fs, *fs; + struct namespace *old_ns; + int ret = 0; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + + ret = -EINVAL; + if (!vxi->vx_namespace) + goto out_put; + + ret = -ENOMEM; + fs = copy_fs_struct(vxi->vx_fs); + if (!fs) + goto out_put; + + ret = 0; + task_lock(current); + old_ns = current->namespace; + old_fs = current->fs; + get_namespace(vxi->vx_namespace); + current->namespace = vxi->vx_namespace; + current->fs = fs; + task_unlock(current); + + put_namespace(old_ns); + put_fs_struct(old_fs); +out_put: + put_vx_info(vxi); + return ret; +} + +int vc_set_namespace(uint32_t id, void __user *data) +{ + struct fs_struct *fs; + struct namespace *ns; + struct vx_info *vxi; + int ret; + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + + task_lock(current); + fs = current->fs; + atomic_inc(&fs->count); + ns = current->namespace; + get_namespace(current->namespace); + task_unlock(current); + + ret = vx_set_namespace(vxi, ns, fs); + + put_namespace(ns); + put_fs_struct(fs); + put_vx_info(vxi); + return ret; +} + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/network.c linux-2.6.17.13-g-v/kernel/vserver/network.c --- linux-2.6.17.13/kernel/vserver/network.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/network.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,807 @@ +/* + * linux/kernel/vserver/network.c + * + * Virtual Server: Network Support + * + * Copyright (C) 2003-2005 Herbert Pötzl + * + * V0.01 broken out from vcontext V0.05 + * V0.02 cleaned up implementation + * V0.03 added equiv nx commands + * V0.04 switch to RCU based hash + * V0.05 and back to locking again + * + */ + +#include +#include +#include +#include + +#include + + +/* __alloc_nx_info() + + * allocate an initialized nx_info struct + * doesn't make it visible (hash) */ + +static struct nx_info *__alloc_nx_info(nid_t nid) +{ + struct nx_info *new = NULL; + + vxdprintk(VXD_CBIT(nid, 1), "alloc_nx_info(%d)*", nid); + + /* would this benefit from a slab cache? */ + new = kmalloc(sizeof(struct nx_info), GFP_KERNEL); + if (!new) + return 0; + + memset (new, 0, sizeof(struct nx_info)); + new->nx_id = nid; + INIT_HLIST_NODE(&new->nx_hlist); + atomic_set(&new->nx_usecnt, 0); + atomic_set(&new->nx_tasks, 0); + new->nx_state = 0; + + new->nx_flags = NXF_INIT_SET; + + /* rest of init goes here */ + + vxdprintk(VXD_CBIT(nid, 0), + "alloc_nx_info(%d) = %p", nid, new); + return new; +} + +/* __dealloc_nx_info() + + * final disposal of nx_info */ + +static void __dealloc_nx_info(struct nx_info *nxi) +{ + vxdprintk(VXD_CBIT(nid, 0), + "dealloc_nx_info(%p)", nxi); + + nxi->nx_hlist.next = LIST_POISON1; + nxi->nx_id = -1; + + BUG_ON(atomic_read(&nxi->nx_usecnt)); + BUG_ON(atomic_read(&nxi->nx_tasks)); + + nxi->nx_state |= NXS_RELEASED; + kfree(nxi); +} + +static void __shutdown_nx_info(struct nx_info *nxi) +{ + nxi->nx_state |= NXS_SHUTDOWN; + vs_net_change(nxi, VSC_NETDOWN); +} + +/* exported stuff */ + +void free_nx_info(struct nx_info *nxi) +{ + /* context shutdown is mandatory */ + BUG_ON(nxi->nx_state != NXS_SHUTDOWN); + + /* context must not be hashed */ + BUG_ON(nxi->nx_state & NXS_HASHED); + + BUG_ON(atomic_read(&nxi->nx_usecnt)); + BUG_ON(atomic_read(&nxi->nx_tasks)); + + __dealloc_nx_info(nxi); +} + + +/* hash table for nx_info hash */ + +#define NX_HASH_SIZE 13 + +struct hlist_head nx_info_hash[NX_HASH_SIZE]; + +static spinlock_t nx_info_hash_lock = SPIN_LOCK_UNLOCKED; + + +static inline unsigned int __hashval(nid_t nid) +{ + return (nid % NX_HASH_SIZE); +} + + + +/* __hash_nx_info() + + * add the nxi to the global hash table + * requires the hash_lock to be held */ + +static inline void __hash_nx_info(struct nx_info *nxi) +{ + struct hlist_head *head; + + vxd_assert_lock(&nx_info_hash_lock); + vxdprintk(VXD_CBIT(nid, 4), + "__hash_nx_info: %p[#%d]", nxi, nxi->nx_id); + + /* context must not be hashed */ + BUG_ON(nx_info_state(nxi, NXS_HASHED)); + + nxi->nx_state |= NXS_HASHED; + head = &nx_info_hash[__hashval(nxi->nx_id)]; + hlist_add_head(&nxi->nx_hlist, head); +} + +/* __unhash_nx_info() + + * remove the nxi from the global hash table + * requires the hash_lock to be held */ + +static inline void __unhash_nx_info(struct nx_info *nxi) +{ + vxd_assert_lock(&nx_info_hash_lock); + vxdprintk(VXD_CBIT(nid, 4), + "__unhash_nx_info: %p[#%d]", nxi, nxi->nx_id); + + /* context must be hashed */ + BUG_ON(!nx_info_state(nxi, NXS_HASHED)); + + nxi->nx_state &= ~NXS_HASHED; + hlist_del(&nxi->nx_hlist); +} + + +/* __lookup_nx_info() + + * requires the hash_lock to be held + * doesn't increment the nx_refcnt */ + +static inline struct nx_info *__lookup_nx_info(nid_t nid) +{ + struct hlist_head *head = &nx_info_hash[__hashval(nid)]; + struct hlist_node *pos; + struct nx_info *nxi; + + vxd_assert_lock(&nx_info_hash_lock); + hlist_for_each(pos, head) { + nxi = hlist_entry(pos, struct nx_info, nx_hlist); + + if (nxi->nx_id == nid) + goto found; + } + nxi = NULL; +found: + vxdprintk(VXD_CBIT(nid, 0), + "__lookup_nx_info(#%u): %p[#%u]", + nid, nxi, nxi?nxi->nx_id:0); + return nxi; +} + + +/* __nx_dynamic_id() + + * find unused dynamic nid + * requires the hash_lock to be held */ + +static inline nid_t __nx_dynamic_id(void) +{ + static nid_t seq = MAX_N_CONTEXT; + nid_t barrier = seq; + + vxd_assert_lock(&nx_info_hash_lock); + do { + if (++seq > MAX_N_CONTEXT) + seq = MIN_D_CONTEXT; + if (!__lookup_nx_info(seq)) { + vxdprintk(VXD_CBIT(nid, 4), + "__nx_dynamic_id: [#%d]", seq); + return seq; + } + } while (barrier != seq); + return 0; +} + +/* __create_nx_info() + + * create the requested context + * get() and hash it */ + +static struct nx_info * __create_nx_info(int id) +{ + struct nx_info *new, *nxi = NULL; + + vxdprintk(VXD_CBIT(nid, 1), "create_nx_info(%d)*", id); + + if (!(new = __alloc_nx_info(id))) + return ERR_PTR(-ENOMEM); + + /* required to make dynamic xids unique */ + spin_lock(&nx_info_hash_lock); + + /* dynamic context requested */ + if (id == NX_DYNAMIC_ID) { + id = __nx_dynamic_id(); + if (!id) { + printk(KERN_ERR "no dynamic context available.\n"); + nxi = ERR_PTR(-EAGAIN); + goto out_unlock; + } + new->nx_id = id; + } + /* static context requested */ + else if ((nxi = __lookup_nx_info(id))) { + vxdprintk(VXD_CBIT(nid, 0), + "create_nx_info(%d) = %p (already there)", id, nxi); + if (nx_info_flags(nxi, NXF_STATE_SETUP, 0)) + nxi = ERR_PTR(-EBUSY); + else + nxi = ERR_PTR(-EEXIST); + goto out_unlock; + } + /* dynamic nid creation blocker */ + else if (id >= MIN_D_CONTEXT) { + vxdprintk(VXD_CBIT(nid, 0), + "create_nx_info(%d) (dynamic rejected)", id); + nxi = ERR_PTR(-EINVAL); + goto out_unlock; + } + + /* new context */ + vxdprintk(VXD_CBIT(nid, 0), + "create_nx_info(%d) = %p (new)", id, new); + __hash_nx_info(get_nx_info(new)); + nxi = new, new = NULL; + +out_unlock: + spin_unlock(&nx_info_hash_lock); + if (new) + __dealloc_nx_info(new); + return nxi; +} + + + +/* exported stuff */ + + +void unhash_nx_info(struct nx_info *nxi) +{ + __shutdown_nx_info(nxi); + spin_lock(&nx_info_hash_lock); + __unhash_nx_info(nxi); + spin_unlock(&nx_info_hash_lock); +} + +#ifdef CONFIG_VSERVER_LEGACYNET + +struct nx_info *create_nx_info(void) +{ + return __create_nx_info(NX_DYNAMIC_ID); +} + +#endif + +/* lookup_nx_info() + + * search for a nx_info and get() it + * negative id means current */ + +struct nx_info *lookup_nx_info(int id) +{ + struct nx_info *nxi = NULL; + + if (id < 0) { + nxi = get_nx_info(current->nx_info); + } else if (id > 1) { + spin_lock(&nx_info_hash_lock); + nxi = get_nx_info(__lookup_nx_info(id)); + spin_unlock(&nx_info_hash_lock); + } + return nxi; +} + +/* nid_is_hashed() + + * verify that nid is still hashed */ + +int nid_is_hashed(nid_t nid) +{ + int hashed; + + spin_lock(&nx_info_hash_lock); + hashed = (__lookup_nx_info(nid) != NULL); + spin_unlock(&nx_info_hash_lock); + return hashed; +} + + +#ifdef CONFIG_PROC_FS + +int get_nid_list(int index, unsigned int *nids, int size) +{ + int hindex, nr_nids = 0; + + for (hindex = 0; hindex < NX_HASH_SIZE; hindex++) { + struct hlist_head *head = &nx_info_hash[hindex]; + struct hlist_node *pos; + + spin_lock(&nx_info_hash_lock); + hlist_for_each(pos, head) { + struct nx_info *nxi; + + if (--index > 0) + continue; + + nxi = hlist_entry(pos, struct nx_info, nx_hlist); + nids[nr_nids] = nxi->nx_id; + if (++nr_nids >= size) { + spin_unlock(&nx_info_hash_lock); + goto out; + } + } + /* keep the lock time short */ + spin_unlock(&nx_info_hash_lock); + } +out: + return nr_nids; +} +#endif + + +/* + * migrate task to new network + * gets nxi, puts old_nxi on change + */ + +int nx_migrate_task(struct task_struct *p, struct nx_info *nxi) +{ + struct nx_info *old_nxi; + int ret = 0; + + if (!p || !nxi) + BUG(); + + vxdprintk(VXD_CBIT(nid, 5), + "nx_migrate_task(%p,%p[#%d.%d.%d])", + p, nxi, nxi->nx_id, + atomic_read(&nxi->nx_usecnt), + atomic_read(&nxi->nx_tasks)); + + /* maybe disallow this completely? */ + old_nxi = task_get_nx_info(p); + if (old_nxi == nxi) + goto out; + + task_lock(p); + if (old_nxi) + clr_nx_info(&p->nx_info); + claim_nx_info(nxi, p); + set_nx_info(&p->nx_info, nxi); + p->nid = nxi->nx_id; + task_unlock(p); + + vxdprintk(VXD_CBIT(nid, 5), + "moved task %p into nxi:%p[#%d]", + p, nxi, nxi->nx_id); + + if (old_nxi) + release_nx_info(old_nxi, p); +out: + put_nx_info(old_nxi); + return ret; +} + + +#ifdef CONFIG_INET + +#include +#include + +int ifa_in_nx_info(struct in_ifaddr *ifa, struct nx_info *nxi) +{ + if (!nxi) + return 1; + if (!ifa) + return 0; + return addr_in_nx_info(nxi, ifa->ifa_local); +} + +int dev_in_nx_info(struct net_device *dev, struct nx_info *nxi) +{ + struct in_device *in_dev; + struct in_ifaddr **ifap; + struct in_ifaddr *ifa; + int ret = 0; + + if (!nxi) + return 1; + + in_dev = in_dev_get(dev); + if (!in_dev) + goto out; + + for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL; + ifap = &ifa->ifa_next) { + if (addr_in_nx_info(nxi, ifa->ifa_local)) { + ret = 1; + break; + } + } + in_dev_put(in_dev); +out: + return ret; +} + +/* + * check if address is covered by socket + * + * sk: the socket to check against + * addr: the address in question (must be != 0) + */ +static inline int __addr_in_socket(struct sock *sk, uint32_t addr) +{ + struct nx_info *nxi = sk->sk_nx_info; + uint32_t saddr = inet_rcv_saddr(sk); + + vxdprintk(VXD_CBIT(net, 5), + "__addr_in_socket(%p,%d.%d.%d.%d) %p:%d.%d.%d.%d %p;%lx", + sk, VXD_QUAD(addr), nxi, VXD_QUAD(saddr), sk->sk_socket, + (sk->sk_socket?sk->sk_socket->flags:0)); + + if (saddr) { + /* direct address match */ + return (saddr == addr); + } else if (nxi) { + /* match against nx_info */ + return addr_in_nx_info(nxi, addr); + } else { + /* unrestricted any socket */ + return 1; + } +} + + +int nx_addr_conflict(struct nx_info *nxi, uint32_t addr, struct sock *sk) +{ + vxdprintk(VXD_CBIT(net, 2), + "nx_addr_conflict(%p,%p) %d.%d,%d.%d", + nxi, sk, VXD_QUAD(addr)); + + if (addr) { + /* check real address */ + return __addr_in_socket(sk, addr); + } else if (nxi) { + /* check against nx_info */ + int i, n = nxi->nbipv4; + + for (i=0; iipv4[i])) + return 1; + return 0; + } else { + /* check against any */ + return 1; + } +} + +#endif /* CONFIG_INET */ + +void nx_set_persistent(struct nx_info *nxi) +{ + get_nx_info(nxi); + claim_nx_info(nxi, current); +} + +void nx_clear_persistent(struct nx_info *nxi) +{ + vxdprintk(VXD_CBIT(nid, 6), + "nx_clear_persistent(%p[#%d])", nxi, nxi->nx_id); + + release_nx_info(nxi, current); + put_nx_info(nxi); +} + +void nx_update_persistent(struct nx_info *nxi) +{ + if (nx_info_flags(nxi, NXF_PERSISTENT, 0)) + nx_set_persistent(nxi); + else + nx_clear_persistent(nxi); +} + +/* vserver syscall commands below here */ + +/* taks nid and nx_info functions */ + +#include + + +int vc_task_nid(uint32_t id, void __user *data) +{ + nid_t nid; + + if (id) { + struct task_struct *tsk; + + if (!vx_check(0, VX_ADMIN|VX_WATCH)) + return -EPERM; + + read_lock(&tasklist_lock); + tsk = find_task_by_real_pid(id); + nid = (tsk) ? tsk->nid : -ESRCH; + read_unlock(&tasklist_lock); + } + else + nid = nx_current_nid(); + return nid; +} + + +int vc_nx_info(uint32_t id, void __user *data) +{ + struct nx_info *nxi; + struct vcmd_nx_info_v0 vc_data; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (!capable(CAP_SYS_ADMIN) || !capable(CAP_SYS_RESOURCE)) + return -EPERM; + + nxi = lookup_nx_info(id); + if (!nxi) + return -ESRCH; + + vc_data.nid = nxi->nx_id; + put_nx_info(nxi); + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + + +/* network functions */ + +int vc_net_create(uint32_t nid, void __user *data) +{ + struct vcmd_net_create vc_data = { .flagword = NXF_INIT_SET }; + struct nx_info *new_nxi; + int ret; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (data && copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + if ((nid > MAX_S_CONTEXT) && (nid != VX_DYNAMIC_ID)) + return -EINVAL; + if (nid < 2) + return -EINVAL; + + new_nxi = __create_nx_info(nid); + if (IS_ERR(new_nxi)) + return PTR_ERR(new_nxi); + + /* initial flags */ + new_nxi->nx_flags = vc_data.flagword; + + /* get a reference for persistent contexts */ + if ((vc_data.flagword & NXF_PERSISTENT)) + nx_set_persistent(new_nxi); + + ret = -ENOEXEC; + if (vs_net_change(new_nxi, VSC_NETUP)) + goto out_unhash; + ret = nx_migrate_task(current, new_nxi); + if (!ret) { + /* return context id on success */ + ret = new_nxi->nx_id; + goto out; + } +out_unhash: + /* prepare for context disposal */ + new_nxi->nx_state |= NXS_SHUTDOWN; + if ((vc_data.flagword & NXF_PERSISTENT)) + nx_clear_persistent(new_nxi); + __unhash_nx_info(new_nxi); +out: + put_nx_info(new_nxi); + return ret; +} + + +int vc_net_migrate(uint32_t id, void __user *data) +{ + struct nx_info *nxi; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + nxi = lookup_nx_info(id); + if (!nxi) + return -ESRCH; + nx_migrate_task(current, nxi); + put_nx_info(nxi); + return 0; +} + +int vc_net_add(uint32_t nid, void __user *data) +{ + struct vcmd_net_addr_v0 vc_data; + struct nx_info *nxi; + int index, pos, ret = 0; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (data && copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + switch (vc_data.type) { + case NXA_TYPE_IPV4: + if ((vc_data.count < 1) || (vc_data.count > 4)) + return -EINVAL; + break; + + default: + break; + } + + nxi = lookup_nx_info(nid); + if (!nxi) + return -ESRCH; + + switch (vc_data.type) { + case NXA_TYPE_IPV4: + index = 0; + while ((index < vc_data.count) && + ((pos = nxi->nbipv4) < NB_IPV4ROOT)) { + nxi->ipv4[pos] = vc_data.ip[index]; + nxi->mask[pos] = vc_data.mask[index]; + index++; + nxi->nbipv4++; + } + ret = index; + break; + + case NXA_TYPE_IPV4|NXA_MOD_BCAST: + nxi->v4_bcast = vc_data.ip[0]; + ret = 1; + break; + + default: + ret = -EINVAL; + break; + } + + put_nx_info(nxi); + return ret; +} + +int vc_net_remove(uint32_t nid, void __user *data) +{ + struct vcmd_net_addr_v0 vc_data; + struct nx_info *nxi; + int ret = 0; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (data && copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + nxi = lookup_nx_info(nid); + if (!nxi) + return -ESRCH; + + switch (vc_data.type) { + case NXA_TYPE_ANY: + nxi->nbipv4 = 0; + break; + + default: + ret = -EINVAL; + break; + } + + put_nx_info(nxi); + return ret; +} + +int vc_get_nflags(uint32_t id, void __user *data) +{ + struct nx_info *nxi; + struct vcmd_net_flags_v0 vc_data; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + nxi = lookup_nx_info(id); + if (!nxi) + return -ESRCH; + + vc_data.flagword = nxi->nx_flags; + + /* special STATE flag handling */ + vc_data.mask = vx_mask_flags(~0UL, nxi->nx_flags, NXF_ONE_TIME); + + put_nx_info(nxi); + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +int vc_set_nflags(uint32_t id, void __user *data) +{ + struct nx_info *nxi; + struct vcmd_net_flags_v0 vc_data; + uint64_t mask, trigger; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + nxi = lookup_nx_info(id); + if (!nxi) + return -ESRCH; + + /* special STATE flag handling */ + mask = vx_mask_mask(vc_data.mask, nxi->nx_flags, NXF_ONE_TIME); + trigger = (mask & nxi->nx_flags) ^ (mask & vc_data.flagword); + + nxi->nx_flags = vx_mask_flags(nxi->nx_flags, + vc_data.flagword, mask); + if (trigger & NXF_PERSISTENT) + nx_update_persistent(nxi); + + put_nx_info(nxi); + return 0; +} + +int vc_get_ncaps(uint32_t id, void __user *data) +{ + struct nx_info *nxi; + struct vcmd_net_caps_v0 vc_data; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + nxi = lookup_nx_info(id); + if (!nxi) + return -ESRCH; + + vc_data.ncaps = nxi->nx_ncaps; + vc_data.cmask = ~0UL; + put_nx_info(nxi); + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + return -EFAULT; + return 0; +} + +int vc_set_ncaps(uint32_t id, void __user *data) +{ + struct nx_info *nxi; + struct vcmd_net_caps_v0 vc_data; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + nxi = lookup_nx_info(id); + if (!nxi) + return -ESRCH; + + nxi->nx_ncaps = vx_mask_flags(nxi->nx_ncaps, + vc_data.ncaps, vc_data.cmask); + put_nx_info(nxi); + return 0; +} + + +#include + +EXPORT_SYMBOL_GPL(free_nx_info); +EXPORT_SYMBOL_GPL(unhash_nx_info); + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/proc.c linux-2.6.17.13-g-v/kernel/vserver/proc.c --- linux-2.6.17.13/kernel/vserver/proc.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/proc.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,860 @@ +/* + * linux/kernel/vserver/proc.c + * + * Virtual Context Support + * + * Copyright (C) 2003-2005 Herbert Pötzl + * + * V0.01 basic structure + * V0.02 adaptation vs1.3.0 + * V0.03 proc permissions + * V0.04 locking/generic + * V0.05 next generation procfs + * V0.06 inode validation + * V0.07 generic rewrite vid + * + */ + +#include +#include +#include +#include +#include +#include + +#include + +#include +#include + +#include "cvirt_proc.h" +#include "limit_proc.h" +#include "sched_proc.h" +#include "vci_config.h" + +static struct proc_dir_entry *proc_virtual; + +static struct proc_dir_entry *proc_vnet; + + +enum vid_directory_inos { + PROC_XID_INO = 32, + PROC_XID_INFO, + PROC_XID_STATUS, + PROC_XID_LIMIT, + PROC_XID_SCHED, + PROC_XID_CVIRT, + PROC_XID_CACCT, + + PROC_NID_INO = 64, + PROC_NID_INFO, + PROC_NID_STATUS, +}; + +#define PROC_VID_MASK 0x60 + + +/* first the actual feeds */ + + +static int proc_virtual_info(int vid, char *buffer) +{ + return sprintf(buffer, + "VCIVersion:\t%04x:%04x\n" + "VCISyscall:\t%d\n" + "VCIKernel:\t%08x\n" + ,VCI_VERSION >> 16 + ,VCI_VERSION & 0xFFFF + ,__NR_vserver + ,vci_kernel_config() + ); +} + + +int proc_xid_info (int vid, char *buffer) +{ + struct vx_info *vxi; + int length; + + vxi = lookup_vx_info(vid); + if (!vxi) + return 0; + length = sprintf(buffer, + "ID:\t%d\n" + "Info:\t%p\n" + "Init:\t%d\n" + ,vxi->vx_id + ,vxi + ,vxi->vx_initpid + ); + put_vx_info(vxi); + return length; +} + +int proc_xid_status (int vid, char *buffer) +{ + struct vx_info *vxi; + int length; + + vxi = lookup_vx_info(vid); + if (!vxi) + return 0; + length = sprintf(buffer, + "UseCnt:\t%d\n" + "Tasks:\t%d\n" + "Flags:\t%016llx\n" + "BCaps:\t%016llx\n" + "CCaps:\t%016llx\n" + "Ticks:\t%d\n" + ,atomic_read(&vxi->vx_usecnt) + ,atomic_read(&vxi->vx_tasks) + ,(unsigned long long)vxi->vx_flags + ,(unsigned long long)vxi->vx_bcaps + ,(unsigned long long)vxi->vx_ccaps + ,atomic_read(&vxi->limit.ticks) + ); + put_vx_info(vxi); + return length; +} + +int proc_xid_limit (int vid, char *buffer) +{ + struct vx_info *vxi; + int length; + + vxi = lookup_vx_info(vid); + if (!vxi) + return 0; + length = vx_info_proc_limit(&vxi->limit, buffer); + put_vx_info(vxi); + return length; +} + +int proc_xid_sched (int vid, char *buffer) +{ + struct vx_info *vxi; + int length; + + vxi = lookup_vx_info(vid); + if (!vxi) + return 0; + length = vx_info_proc_sched(&vxi->sched, buffer); + put_vx_info(vxi); + return length; +} + +int proc_xid_cvirt (int vid, char *buffer) +{ + struct vx_info *vxi; + int length; + + vxi = lookup_vx_info(vid); + if (!vxi) + return 0; + vx_update_load(vxi); + length = vx_info_proc_cvirt(&vxi->cvirt, buffer); + put_vx_info(vxi); + return length; +} + +int proc_xid_cacct (int vid, char *buffer) +{ + struct vx_info *vxi; + int length; + + vxi = lookup_vx_info(vid); + if (!vxi) + return 0; + length = vx_info_proc_cacct(&vxi->cacct, buffer); + put_vx_info(vxi); + return length; +} + + +static int proc_vnet_info(int vid, char *buffer) +{ + return sprintf(buffer, + "VCIVersion:\t%04x:%04x\n" + "VCISyscall:\t%d\n" + ,VCI_VERSION >> 16 + ,VCI_VERSION & 0xFFFF + ,__NR_vserver + ); +} + + +int proc_nid_info (int vid, char *buffer) +{ + struct nx_info *nxi; + int length, i; + + nxi = lookup_nx_info(vid); + if (!nxi) + return 0; + length = sprintf(buffer, + "ID:\t%d\n" + "Info:\t%p\n" + ,nxi->nx_id + ,nxi + ); + for (i=0; inbipv4; i++) { + length += sprintf(buffer + length, + "%d:\t" NIPQUAD_FMT "/" NIPQUAD_FMT "\n", i, + NIPQUAD(nxi->ipv4[i]), NIPQUAD(nxi->mask[i])); + } + put_nx_info(nxi); + return length; +} + +int proc_nid_status (int vid, char *buffer) +{ + struct nx_info *nxi; + int length; + + nxi = lookup_nx_info(vid); + if (!nxi) + return 0; + length = sprintf(buffer, + "UseCnt:\t%d\n" + "Tasks:\t%d\n" + ,atomic_read(&nxi->nx_usecnt) + ,atomic_read(&nxi->nx_tasks) + ); + put_nx_info(nxi); + return length; +} + +/* here the inode helpers */ + + +#define fake_ino(id,nr) (((nr) & 0xFFFF) | \ + (((id) & 0xFFFF) << 16)) + +#define inode_vid(i) (((i)->i_ino >> 16) & 0xFFFF) +#define inode_type(i) ((i)->i_ino & 0xFFFF) + +#define MAX_MULBY10 ((~0U-9)/10) + + +static struct inode *proc_vid_make_inode(struct super_block * sb, + int vid, int ino) +{ + struct inode *inode = new_inode(sb); + + if (!inode) + goto out; + + inode->i_mtime = inode->i_atime = + inode->i_ctime = CURRENT_TIME; + inode->i_ino = fake_ino(vid, ino); + + inode->i_uid = 0; + inode->i_gid = 0; +out: + return inode; +} + +static int proc_vid_revalidate(struct dentry * dentry, struct nameidata *nd) +{ + struct inode * inode = dentry->d_inode; + int vid, hashed=0; + + vid = inode_vid(inode); + switch (inode_type(inode) & PROC_VID_MASK) { + case PROC_XID_INO: + hashed = xid_is_hashed(vid); + break; + case PROC_NID_INO: + hashed = nid_is_hashed(vid); + break; + } + if (hashed) + return 1; + d_drop(dentry); + return 0; +} + + +#define PROC_BLOCK_SIZE (PAGE_SIZE - 1024) + +static ssize_t proc_vid_info_read(struct file * file, char __user * buf, + size_t count, loff_t *ppos) +{ + struct inode * inode = file->f_dentry->d_inode; + unsigned long page; + ssize_t length; + int vid; + + if (count > PROC_BLOCK_SIZE) + count = PROC_BLOCK_SIZE; + if (!(page = __get_free_page(GFP_KERNEL))) + return -ENOMEM; + + vid = inode_vid(inode); + length = PROC_I(inode)->op.proc_vid_read(vid, (char*)page); + + if (length >= 0) + length = simple_read_from_buffer(buf, count, ppos, + (char *)page, length); + free_page(page); + return length; +} + + + + + +/* here comes the lower level (vid) */ + +static struct file_operations proc_vid_info_file_operations = { + .read = proc_vid_info_read, +}; + +static struct dentry_operations proc_vid_dentry_operations = { + .d_revalidate = proc_vid_revalidate, +}; + + +struct vid_entry { + int type; + int len; + char *name; + mode_t mode; +}; + +#define E(type,name,mode) {(type),sizeof(name)-1,(name),(mode)} + +static struct vid_entry vx_base_stuff[] = { + E(PROC_XID_INFO, "info", S_IFREG|S_IRUGO), + E(PROC_XID_STATUS, "status", S_IFREG|S_IRUGO), + E(PROC_XID_LIMIT, "limit", S_IFREG|S_IRUGO), + E(PROC_XID_SCHED, "sched", S_IFREG|S_IRUGO), + E(PROC_XID_CVIRT, "cvirt", S_IFREG|S_IRUGO), + E(PROC_XID_CACCT, "cacct", S_IFREG|S_IRUGO), + {0,0,NULL,0} +}; + +static struct vid_entry vn_base_stuff[] = { + E(PROC_NID_INFO, "info", S_IFREG|S_IRUGO), + E(PROC_NID_STATUS, "status", S_IFREG|S_IRUGO), + {0,0,NULL,0} +}; + + + +static struct dentry *proc_vid_lookup(struct inode *dir, + struct dentry *dentry, struct nameidata *nd) +{ + struct inode *inode; + struct vid_entry *p; + int error; + + error = -ENOENT; + inode = NULL; + + switch (inode_type(dir)) { + case PROC_XID_INO: + p = vx_base_stuff; + break; + case PROC_NID_INO: + p = vn_base_stuff; + break; + default: + goto out; + } + + for (; p->name; p++) { + if (p->len != dentry->d_name.len) + continue; + if (!memcmp(dentry->d_name.name, p->name, p->len)) + break; + } + if (!p->name) + goto out; + + error = -EINVAL; + inode = proc_vid_make_inode(dir->i_sb, inode_vid(dir), p->type); + if (!inode) + goto out; + + switch(p->type) { + case PROC_XID_INFO: + PROC_I(inode)->op.proc_vid_read = proc_xid_info; + break; + case PROC_XID_STATUS: + PROC_I(inode)->op.proc_vid_read = proc_xid_status; + break; + case PROC_XID_LIMIT: + PROC_I(inode)->op.proc_vid_read = proc_xid_limit; + break; + case PROC_XID_SCHED: + PROC_I(inode)->op.proc_vid_read = proc_xid_sched; + break; + case PROC_XID_CVIRT: + PROC_I(inode)->op.proc_vid_read = proc_xid_cvirt; + break; + case PROC_XID_CACCT: + PROC_I(inode)->op.proc_vid_read = proc_xid_cacct; + break; + + case PROC_NID_INFO: + PROC_I(inode)->op.proc_vid_read = proc_nid_info; + break; + case PROC_NID_STATUS: + PROC_I(inode)->op.proc_vid_read = proc_nid_status; + break; + + default: + printk("procfs: impossible type (%d)",p->type); + iput(inode); + return ERR_PTR(-EINVAL); + } + inode->i_mode = p->mode; + inode->i_fop = &proc_vid_info_file_operations; + inode->i_nlink = 1; + inode->i_flags|=S_IMMUTABLE; + + dentry->d_op = &proc_vid_dentry_operations; + d_add(dentry, inode); + error = 0; +out: + return ERR_PTR(error); +} + + +static int proc_vid_readdir(struct file * filp, + void * dirent, filldir_t filldir) +{ + int i, size; + struct inode *inode = filp->f_dentry->d_inode; + struct vid_entry *p; + + i = filp->f_pos; + switch (i) { + case 0: + if (filldir(dirent, ".", 1, i, + inode->i_ino, DT_DIR) < 0) + return 0; + i++; + filp->f_pos++; + /* fall through */ + case 1: + if (filldir(dirent, "..", 2, i, + PROC_ROOT_INO, DT_DIR) < 0) + return 0; + i++; + filp->f_pos++; + /* fall through */ + default: + i -= 2; + switch (inode_type(inode)) { + case PROC_XID_INO: + size = sizeof(vx_base_stuff); + p = vx_base_stuff + i; + break; + case PROC_NID_INO: + size = sizeof(vn_base_stuff); + p = vn_base_stuff + i; + break; + default: + return 1; + } + if (i >= size/sizeof(struct vid_entry)) + return 1; + while (p->name) { + if (filldir(dirent, p->name, p->len, + filp->f_pos, fake_ino(inode_vid(inode), + p->type), p->mode >> 12) < 0) + return 0; + filp->f_pos++; + p++; + } + } + return 1; +} + + + + +/* now the upper level (virtual) */ + +static struct file_operations proc_vid_file_operations = { + .read = generic_read_dir, + .readdir = proc_vid_readdir, +}; + +static struct inode_operations proc_vid_inode_operations = { + .lookup = proc_vid_lookup, +}; + + + +static __inline__ int atovid(const char *str, int len) +{ + int vid, c; + + vid = 0; + while (len-- > 0) { + c = *str - '0'; + str++; + if (c > 9) + return -1; + if (vid >= MAX_MULBY10) + return -1; + vid *= 10; + vid += c; + if (!vid) + return -1; + } + return vid; +} + + +struct dentry *proc_virtual_lookup(struct inode *dir, + struct dentry * dentry, struct nameidata *nd) +{ + int xid, len, ret; + struct vx_info *vxi; + const char *name; + struct inode *inode; + + name = dentry->d_name.name; + len = dentry->d_name.len; + ret = -ENOMEM; + + if (len == 7 && !memcmp(name, "current", 7)) { + inode = new_inode(dir->i_sb); + if (!inode) + goto out; + inode->i_mtime = inode->i_atime = + inode->i_ctime = CURRENT_TIME; + inode->i_ino = fake_ino(1, PROC_XID_INO); + inode->i_mode = S_IFLNK|S_IRWXUGO; + inode->i_uid = inode->i_gid = 0; + d_add(dentry, inode); + return NULL; + } + if (len == 4 && !memcmp(name, "info", 4)) { + inode = proc_vid_make_inode(dir->i_sb, 0, PROC_XID_INFO); + if (!inode) + goto out; + inode->i_fop = &proc_vid_info_file_operations; + PROC_I(inode)->op.proc_vid_read = proc_virtual_info; + inode->i_mode = S_IFREG|S_IRUGO; + d_add(dentry, inode); + return NULL; + } + + ret = -ENOENT; + xid = atovid(name, len); + if (xid < 0) + goto out; + vxi = lookup_vx_info(xid); + if (!vxi) + goto out; + + inode = NULL; + if (vx_check(xid, VX_ADMIN|VX_WATCH|VX_IDENT)) + inode = proc_vid_make_inode(dir->i_sb, + vxi->vx_id, PROC_XID_INO); + if (!inode) + goto out_release; + + inode->i_mode = S_IFDIR|S_IRUGO|S_IXUGO; + inode->i_op = &proc_vid_inode_operations; + inode->i_fop = &proc_vid_file_operations; + inode->i_nlink = 2; + inode->i_flags|=S_IMMUTABLE; + + dentry->d_op = &proc_vid_dentry_operations; + d_add(dentry, inode); + ret = 0; + +out_release: + put_vx_info(vxi); +out: + return ERR_PTR(ret); +} + + +struct dentry *proc_vnet_lookup(struct inode *dir, + struct dentry * dentry, struct nameidata *nd) +{ + int nid, len, ret; + struct nx_info *nxi; + const char *name; + struct inode *inode; + + name = dentry->d_name.name; + len = dentry->d_name.len; + ret = -ENOMEM; + if (len == 7 && !memcmp(name, "current", 7)) { + inode = new_inode(dir->i_sb); + if (!inode) + goto out; + inode->i_mtime = inode->i_atime = + inode->i_ctime = CURRENT_TIME; + inode->i_ino = fake_ino(1, PROC_NID_INO); + inode->i_mode = S_IFLNK|S_IRWXUGO; + inode->i_uid = inode->i_gid = 0; + d_add(dentry, inode); + return NULL; + } + if (len == 4 && !memcmp(name, "info", 4)) { + inode = proc_vid_make_inode(dir->i_sb, 0, PROC_NID_INFO); + if (!inode) + goto out; + inode->i_fop = &proc_vid_info_file_operations; + PROC_I(inode)->op.proc_vid_read = proc_vnet_info; + inode->i_mode = S_IFREG|S_IRUGO; + d_add(dentry, inode); + return NULL; + } + + ret = -ENOENT; + nid = atovid(name, len); + if (nid < 0) + goto out; + nxi = lookup_nx_info(nid); + if (!nxi) + goto out; + + inode = NULL; + if (1) + inode = proc_vid_make_inode(dir->i_sb, + nxi->nx_id, PROC_NID_INO); + if (!inode) + goto out_release; + + inode->i_mode = S_IFDIR|S_IRUGO|S_IXUGO; + inode->i_op = &proc_vid_inode_operations; + inode->i_fop = &proc_vid_file_operations; + inode->i_nlink = 2; + inode->i_flags|=S_IMMUTABLE; + + dentry->d_op = &proc_vid_dentry_operations; + d_add(dentry, inode); + ret = 0; + +out_release: + put_nx_info(nxi); +out: + return ERR_PTR(ret); +} + + + + +#define PROC_NUMBUF 10 +#define PROC_MAXVIDS 32 + +int proc_virtual_readdir(struct file * filp, + void * dirent, filldir_t filldir) +{ + unsigned int xid_array[PROC_MAXVIDS]; + char buf[PROC_NUMBUF]; + unsigned int nr = filp->f_pos-3; + unsigned int nr_xids, i; + ino_t ino; + + switch ((long)filp->f_pos) { + case 0: + ino = fake_ino(0, PROC_XID_INO); + if (filldir(dirent, ".", 1, + filp->f_pos, ino, DT_DIR) < 0) + return 0; + filp->f_pos++; + /* fall through */ + case 1: + ino = filp->f_dentry->d_parent->d_inode->i_ino; + if (filldir(dirent, "..", 2, + filp->f_pos, ino, DT_DIR) < 0) + return 0; + filp->f_pos++; + /* fall through */ + case 2: + ino = fake_ino(0, PROC_XID_INFO); + if (filldir(dirent, "info", 4, + filp->f_pos, ino, DT_LNK) < 0) + return 0; + filp->f_pos++; + /* fall through */ + case 3: + if (vx_current_xid() > 1) { + ino = fake_ino(1, PROC_XID_INO); + if (filldir(dirent, "current", 7, + filp->f_pos, ino, DT_LNK) < 0) + return 0; + } + filp->f_pos++; + } + + nr_xids = get_xid_list(nr, xid_array, PROC_MAXVIDS); + for (i = 0; i < nr_xids; i++) { + int xid = xid_array[i]; + ino_t ino = fake_ino(xid, PROC_XID_INO); + unsigned int j = PROC_NUMBUF; + + do buf[--j] = '0' + (xid % 10); while (xid/=10); + + if (filldir(dirent, buf+j, PROC_NUMBUF-j, + filp->f_pos, ino, DT_DIR) < 0) + break; + filp->f_pos++; + } + return 0; +} + + +static struct file_operations proc_virtual_dir_operations = { + .read = generic_read_dir, + .readdir = proc_virtual_readdir, +}; + +static struct inode_operations proc_virtual_dir_inode_operations = { + .lookup = proc_virtual_lookup, +}; + + +int proc_vnet_readdir(struct file * filp, + void * dirent, filldir_t filldir) +{ + unsigned int nid_array[PROC_MAXVIDS]; + char buf[PROC_NUMBUF]; + unsigned int nr = filp->f_pos-2; + unsigned int nr_nids, i; + ino_t ino; + + switch ((long)filp->f_pos) { + case 0: + ino = fake_ino(0, PROC_NID_INO); + if (filldir(dirent, ".", 1, + filp->f_pos, ino, DT_DIR) < 0) + return 0; + filp->f_pos++; + /* fall through */ + case 1: + ino = filp->f_dentry->d_parent->d_inode->i_ino; + if (filldir(dirent, "..", 2, + filp->f_pos, ino, DT_DIR) < 0) + return 0; + filp->f_pos++; + /* fall through */ + case 2: + ino = fake_ino(0, PROC_NID_INFO); + if (filldir(dirent, "info", 4, + filp->f_pos, ino, DT_REG) < 0) + return 0; + filp->f_pos++; + /* fall through */ + } + + nr_nids = get_nid_list(nr, nid_array, PROC_MAXVIDS); + for (i = 0; i < nr_nids; i++) { + int nid = nid_array[i]; + ino_t ino = fake_ino(nid, PROC_NID_INO); + unsigned long j = PROC_NUMBUF; + + do buf[--j] = '0' + (nid % 10); while (nid/=10); + + if (filldir(dirent, buf+j, PROC_NUMBUF-j, + filp->f_pos, ino, DT_DIR) < 0) + break; + filp->f_pos++; + } + return 0; +} + + +static struct file_operations proc_vnet_dir_operations = { + .read = generic_read_dir, + .readdir = proc_vnet_readdir, +}; + +static struct inode_operations proc_vnet_dir_inode_operations = { + .lookup = proc_vnet_lookup, +}; + + + +void proc_vx_init(void) +{ + struct proc_dir_entry *ent; + + ent = proc_mkdir("virtual", 0); + if (ent) { + ent->proc_fops = &proc_virtual_dir_operations; + ent->proc_iops = &proc_virtual_dir_inode_operations; + } + proc_virtual = ent; + + ent = proc_mkdir("virtnet", 0); + if (ent) { + ent->proc_fops = &proc_vnet_dir_operations; + ent->proc_iops = &proc_vnet_dir_inode_operations; + } + proc_vnet = ent; +} + + + + +/* per pid info */ + + +int proc_pid_vx_info(struct task_struct *p, char *buffer) +{ + struct vx_info *vxi; + char * orig = buffer; + + buffer += sprintf (buffer,"XID:\t%d\n", vx_task_xid(p)); + if (vx_flags(VXF_INFO_HIDE, 0)) + goto out; + + vxi = task_get_vx_info(p); + if (!vxi) + goto out; + + buffer += sprintf (buffer,"BCaps:\t%016llx\n" + ,(unsigned long long)vxi->vx_bcaps); + buffer += sprintf (buffer,"CCaps:\t%016llx\n" + ,(unsigned long long)vxi->vx_ccaps); + buffer += sprintf (buffer,"CFlags:\t%016llx\n" + ,(unsigned long long)vxi->vx_flags); + buffer += sprintf (buffer,"CIPid:\t%d\n" + ,vxi->vx_initpid); + + put_vx_info(vxi); +out: + return buffer - orig; +} + + +int proc_pid_nx_info(struct task_struct *p, char *buffer) +{ + struct nx_info *nxi; + char * orig = buffer; + int i; + + buffer += sprintf (buffer,"NID:\t%d\n", nx_task_nid(p)); + if (vx_flags(VXF_INFO_HIDE, 0)) + goto out; + nxi = task_get_nx_info(p); + if (!nxi) + goto out; + + for (i=0; inbipv4; i++){ + buffer += sprintf (buffer, + "V4Root[%d]:\t%d.%d.%d.%d/%d.%d.%d.%d\n", i + ,NIPQUAD(nxi->ipv4[i]) + ,NIPQUAD(nxi->mask[i])); + } + buffer += sprintf (buffer, + "V4Root[bcast]:\t%d.%d.%d.%d\n" + ,NIPQUAD(nxi->v4_bcast)); + + put_nx_info(nxi); +out: + return buffer - orig; +} + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/sched.c linux-2.6.17.13-g-v/kernel/vserver/sched.c --- linux-2.6.17.13/kernel/vserver/sched.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/sched.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,217 @@ +/* + * linux/kernel/vserver/sched.c + * + * Virtual Server: Scheduler Support + * + * Copyright (C) 2004-2005 Herbert Pötzl + * + * V0.01 adapted Sam Vilains version to 2.6.3 + * V0.02 removed legacy interface + * + */ + +#include +#include +#include +#include + +#include +#include + + +/* + * recalculate the context's scheduling tokens + * + * ret > 0 : number of tokens available + * ret = 0 : context is paused + * ret < 0 : number of jiffies until new tokens arrive + * + */ +int vx_tokens_recalc(struct vx_info *vxi) +{ + long delta, tokens = 0; + + if (vx_info_flags(vxi, VXF_SCHED_PAUSE, 0)) + /* we are paused */ + return 0; + + delta = jiffies - vxi->sched.jiffies; + + if (delta >= vxi->sched.interval) { + /* lockdown scheduler info */ + spin_lock(&vxi->sched.tokens_lock); + + /* calc integral token part */ + delta = jiffies - vxi->sched.jiffies; + tokens = delta / vxi->sched.interval; + delta = tokens * vxi->sched.interval; + tokens *= vxi->sched.fill_rate; + + atomic_add(tokens, &vxi->sched.tokens); + vxi->sched.jiffies += delta; + tokens = atomic_read(&vxi->sched.tokens); + + if (tokens > vxi->sched.tokens_max) { + tokens = vxi->sched.tokens_max; + atomic_set(&vxi->sched.tokens, tokens); + } + spin_unlock(&vxi->sched.tokens_lock); + } else { + /* no new tokens */ + tokens = vx_tokens_avail(vxi); + if (tokens <= 0) + vxi->vx_state |= VXS_ONHOLD; + if (tokens < vxi->sched.tokens_min) { + /* enough tokens will be available in */ + if (vxi->sched.tokens_min == 0) + return delta - vxi->sched.interval; + return delta - vxi->sched.interval * + vxi->sched.tokens_min / vxi->sched.fill_rate; + } + } + + /* we have some tokens left */ + if (vx_info_state(vxi, VXS_ONHOLD) && + (tokens >= vxi->sched.tokens_min)) + vxi->vx_state &= ~VXS_ONHOLD; + if (vx_info_state(vxi, VXS_ONHOLD)) + tokens -= vxi->sched.tokens_min; + + return tokens; +} + +/* + * effective_prio - return the priority that is based on the static + * priority but is modified by bonuses/penalties. + * + * We scale the actual sleep average [0 .... MAX_SLEEP_AVG] + * into a -4 ... 0 ... +4 bonus/penalty range. + * + * Additionally, we scale another amount based on the number of + * CPU tokens currently held by the context, if the process is + * part of a context (and the appropriate SCHED flag is set). + * This ranges from -5 ... 0 ... +15, quadratically. + * + * So, the total bonus is -9 .. 0 .. +19 + * We use ~50% of the full 0...39 priority range so that: + * + * 1) nice +19 interactive tasks do not preempt nice 0 CPU hogs. + * 2) nice -20 CPU hogs do not get preempted by nice 0 tasks. + * unless that context is far exceeding its CPU allocation. + * + * Both properties are important to certain workloads. + */ +int vx_effective_vavavoom(struct vx_info *vxi, int max_prio) +{ + int vavavoom, max; + + /* lots of tokens = lots of vavavoom + * no tokens = no vavavoom */ + if ((vavavoom = atomic_read(&vxi->sched.tokens)) >= 0) { + max = vxi->sched.tokens_max; + vavavoom = max - vavavoom; + max = max * max; + vavavoom = max_prio * VAVAVOOM_RATIO / 100 + * (vavavoom*vavavoom - (max >> 2)) / max; + } else + vavavoom = 0; + + vxi->sched.vavavoom = vavavoom; + return vavavoom + vxi->sched.priority_bias; +} + + +int vc_set_sched_v2(uint32_t xid, void __user *data) +{ + struct vcmd_set_sched_v2 vc_data; + struct vx_info *vxi; + + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + vxi = lookup_vx_info(xid); + if (!vxi) + return -EINVAL; + + spin_lock(&vxi->sched.tokens_lock); + + if (vc_data.interval != SCHED_KEEP) + vxi->sched.interval = vc_data.interval; + if (vc_data.fill_rate != SCHED_KEEP) + vxi->sched.fill_rate = vc_data.fill_rate; + if (vc_data.tokens_min != SCHED_KEEP) + vxi->sched.tokens_min = vc_data.tokens_min; + if (vc_data.tokens_max != SCHED_KEEP) + vxi->sched.tokens_max = vc_data.tokens_max; + if (vc_data.tokens != SCHED_KEEP) + atomic_set(&vxi->sched.tokens, vc_data.tokens); + + /* Sanity check the resultant values */ + if (vxi->sched.fill_rate <= 0) + vxi->sched.fill_rate = 1; + if (vxi->sched.interval <= 0) + vxi->sched.interval = HZ; + if (vxi->sched.tokens_max == 0) + vxi->sched.tokens_max = 1; + if (atomic_read(&vxi->sched.tokens) > vxi->sched.tokens_max) + atomic_set(&vxi->sched.tokens, vxi->sched.tokens_max); + if (vxi->sched.tokens_min > vxi->sched.tokens_max) + vxi->sched.tokens_min = vxi->sched.tokens_max; + + spin_unlock(&vxi->sched.tokens_lock); + put_vx_info(vxi); + return 0; +} + + +int vc_set_sched(uint32_t xid, void __user *data) +{ + struct vcmd_set_sched_v3 vc_data; + struct vx_info *vxi; + unsigned int set_mask; + + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + vxi = lookup_vx_info(xid); + if (!vxi) + return -EINVAL; + + set_mask = vc_data.set_mask; + + spin_lock(&vxi->sched.tokens_lock); + + if (set_mask & VXSM_FILL_RATE) + vxi->sched.fill_rate = vc_data.fill_rate; + if (set_mask & VXSM_INTERVAL) + vxi->sched.interval = vc_data.interval; + if (set_mask & VXSM_TOKENS) + atomic_set(&vxi->sched.tokens, vc_data.tokens); + if (set_mask & VXSM_TOKENS_MIN) + vxi->sched.tokens_min = vc_data.tokens_min; + if (set_mask & VXSM_TOKENS_MAX) + vxi->sched.tokens_max = vc_data.tokens_max; + if (set_mask & VXSM_PRIO_BIAS) + vxi->sched.priority_bias = vc_data.priority_bias; + + /* Sanity check the resultant values */ + if (vxi->sched.fill_rate <= 0) + vxi->sched.fill_rate = 1; + if (vxi->sched.interval <= 0) + vxi->sched.interval = HZ; + if (vxi->sched.tokens_max == 0) + vxi->sched.tokens_max = 1; + if (atomic_read(&vxi->sched.tokens) > vxi->sched.tokens_max) + atomic_set(&vxi->sched.tokens, vxi->sched.tokens_max); + if (vxi->sched.tokens_min > vxi->sched.tokens_max) + vxi->sched.tokens_min = vxi->sched.tokens_max; + if (vxi->sched.priority_bias > MAX_PRIO_BIAS) + vxi->sched.priority_bias = MAX_PRIO_BIAS; + if (vxi->sched.priority_bias < MIN_PRIO_BIAS) + vxi->sched.priority_bias = MIN_PRIO_BIAS; + + spin_unlock(&vxi->sched.tokens_lock); + put_vx_info(vxi); + return 0; +} + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/sched_init.h linux-2.6.17.13-g-v/kernel/vserver/sched_init.h --- linux-2.6.17.13/kernel/vserver/sched_init.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/sched_init.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,30 @@ + +static inline void vx_info_init_sched(struct _vx_sched *sched) +{ + int i; + + /* scheduling; hard code starting values as constants */ + sched->fill_rate = 1; + sched->interval = 4; + sched->tokens_min = HZ >> 4; + sched->tokens_max = HZ >> 1; + sched->jiffies = jiffies; + sched->tokens_lock = SPIN_LOCK_UNLOCKED; + + atomic_set(&sched->tokens, HZ >> 2); + sched->cpus_allowed = CPU_MASK_ALL; + sched->priority_bias = 0; + sched->vavavoom = 0; + + for_each_cpu(i) { + sched->cpu[i].user_ticks = 0; + sched->cpu[i].sys_ticks = 0; + sched->cpu[i].hold_ticks = 0; + } +} + +static inline void vx_info_exit_sched(struct _vx_sched *sched) +{ + return; +} + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/sched_proc.h linux-2.6.17.13-g-v/kernel/vserver/sched_proc.h --- linux-2.6.17.13/kernel/vserver/sched_proc.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/sched_proc.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,40 @@ +#ifndef _VX_SCHED_PROC_H +#define _VX_SCHED_PROC_H + + +static inline int vx_info_proc_sched(struct _vx_sched *sched, char *buffer) +{ + int length = 0; + int i; + + length += sprintf(buffer, + "Token:\t\t%8d\n" + "FillRate:\t%8d\n" + "Interval:\t%8d\n" + "TokensMin:\t%8d\n" + "TokensMax:\t%8d\n" + "PrioBias:\t%8d\n" + "VaVaVoom:\t%8d\n" + ,atomic_read(&sched->tokens) + ,sched->fill_rate + ,sched->interval + ,sched->tokens_min + ,sched->tokens_max + ,sched->priority_bias + ,sched->vavavoom + ); + + for_each_online_cpu(i) { + length += sprintf(buffer + length, + "cpu %d: %lld %lld %lld\n" + ,i + ,(long long)sched->cpu[i].user_ticks + ,(long long)sched->cpu[i].sys_ticks + ,(long long)sched->cpu[i].hold_ticks + ); + } + + return length; +} + +#endif /* _VX_SCHED_PROC_H */ diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/signal.c linux-2.6.17.13-g-v/kernel/vserver/signal.c --- linux-2.6.17.13/kernel/vserver/signal.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/signal.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,139 @@ +/* + * linux/kernel/vserver/signal.c + * + * Virtual Server: Signal Support + * + * Copyright (C) 2003-2005 Herbert Pötzl + * + * V0.01 broken out from vcontext V0.05 + * + */ + +#include + +#include +#include + +#include +#include + + +int vx_info_kill(struct vx_info *vxi, int pid, int sig) +{ + int retval, count=0; + struct task_struct *p; + unsigned long priv = 0; + + retval = -ESRCH; + vxdprintk(VXD_CBIT(misc, 4), + "vx_info_kill(%p[#%d],%d,%d)*", + vxi, vxi->vx_id, pid, sig); + read_lock(&tasklist_lock); + switch (pid) { + case 0: + priv = 1; + case -1: + for_each_process(p) { + int err = 0; + + if (vx_task_xid(p) != vxi->vx_id || p->pid <= 1 || + (pid && vxi->vx_initpid == p->pid)) + continue; + + err = group_send_sig_info(sig, (void*)priv, p); + ++count; + if (err != -EPERM) + retval = err; + } + break; + + case 1: + if (vxi->vx_initpid) { + pid = vxi->vx_initpid; + priv = 1; + } + /* fallthrough */ + default: + p = find_task_by_real_pid(pid); + if (p) { + if (vx_task_xid(p) == vxi->vx_id) + retval = group_send_sig_info(sig, + (void*)priv, p); + } + break; + } + read_unlock(&tasklist_lock); + vxdprintk(VXD_CBIT(misc, 4), + "vx_info_kill(%p[#%d],%d,%d) = %d", + vxi, vxi->vx_id, pid, sig, retval); + return retval; +} + +int vc_ctx_kill(uint32_t id, void __user *data) +{ + int retval; + struct vcmd_ctx_kill_v0 vc_data; + struct vx_info *vxi; + + if (!vx_check(0, VX_ADMIN)) + return -ENOSYS; + if (copy_from_user (&vc_data, data, sizeof(vc_data))) + return -EFAULT; + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + + retval = vx_info_kill(vxi, vc_data.pid, vc_data.sig); + put_vx_info(vxi); + return retval; +} + + +static int __wait_exit(struct vx_info *vxi) +{ + DECLARE_WAITQUEUE(wait, current); + int ret = 0; + + add_wait_queue(&vxi->vx_wait, &wait); + set_current_state(TASK_INTERRUPTIBLE); + +wait: + if (vx_info_state(vxi, + VXS_SHUTDOWN|VXS_HASHED|VXS_HELPER) == VXS_SHUTDOWN) + goto out; + if (signal_pending(current)) { + ret = -ERESTARTSYS; + goto out; + } + schedule(); + goto wait; + +out: + set_current_state(TASK_RUNNING); + remove_wait_queue(&vxi->vx_wait, &wait); + return ret; +} + + + +int vc_wait_exit(uint32_t id, void __user *data) +{ + struct vx_info *vxi; + struct vcmd_wait_exit_v0 vc_data; + int ret; + + vxi = lookup_vx_info(id); + if (!vxi) + return -ESRCH; + + ret = __wait_exit(vxi); + vc_data.reboot_cmd = vxi->reboot_cmd; + vc_data.exit_code = vxi->exit_code; + put_vx_info(vxi); + + if (copy_to_user (data, &vc_data, sizeof(vc_data))) + ret = -EFAULT; + return ret; +} + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/switch.c linux-2.6.17.13-g-v/kernel/vserver/switch.c --- linux-2.6.17.13/kernel/vserver/switch.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/switch.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,268 @@ +/* + * linux/kernel/vserver/switch.c + * + * Virtual Server: Syscall Switch + * + * Copyright (C) 2003-2005 Herbert Pötzl + * + * V0.01 syscall switch + * V0.02 added signal to context + * V0.03 added rlimit functions + * V0.04 added iattr, task/xid functions + * V0.05 added debug/history stuff + * V0.06 added compat32 layer + * + */ + +#include +#include +#include +#include + +#include +#include +#include + + +static inline +int vc_get_version(uint32_t id) +{ +#ifdef CONFIG_VSERVER_LEGACY_VERSION + if (id == 63) + return VCI_LEGACY_VERSION; +#endif + return VCI_VERSION; +} + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include + + +#ifdef CONFIG_COMPAT +#define __COMPAT(name, id, data, compat) \ + (compat) ? name ## _x32 (id, data) : name (id, data) +#else +#define __COMPAT(name, id, data, compat) \ + name (id, data) +#endif + + +static inline +long do_vserver(uint32_t cmd, uint32_t id, void __user *data, int compat) +{ + vxdprintk(VXD_CBIT(switch, 0), + "vc: VCMD_%02d_%d[%d], %d,%p,%d", + VC_CATEGORY(cmd), VC_COMMAND(cmd), + VC_VERSION(cmd), id, data, compat); + +#ifdef CONFIG_VSERVER_LEGACY + if (!capable(CAP_CONTEXT) && + /* dirty hack for capremove */ + !(cmd==VCMD_new_s_context && id==-2)) + return -EPERM; +#else + if (!capable(CAP_CONTEXT)) + return -EPERM; +#endif + + switch (cmd) { + case VCMD_get_version: + return vc_get_version(id); + + case VCMD_dump_history: +#ifdef CONFIG_VSERVER_HISTORY + return vc_dump_history(id); +#else + return -ENOSYS; +#endif + +#ifdef CONFIG_VSERVER_LEGACY + case VCMD_new_s_context: + return vc_new_s_context(id, data); +#endif +#ifdef CONFIG_VSERVER_LEGACYNET + case VCMD_set_ipv4root: + return vc_set_ipv4root(id, data); +#endif + + case VCMD_task_xid: + return vc_task_xid(id, data); + case VCMD_vx_info: + return vc_vx_info(id, data); + + case VCMD_task_nid: + return vc_task_nid(id, data); + case VCMD_nx_info: + return vc_nx_info(id, data); + + case VCMD_set_namespace_v0: + return vc_set_namespace(-1, data); + case VCMD_set_namespace: + return vc_set_namespace(id, data); + } + + /* those are allowed while in setup too */ + if (!vx_check(0, VX_ADMIN|VX_WATCH) && + !vx_flags(VXF_STATE_SETUP,0)) + return -EPERM; + +#ifdef CONFIG_VSERVER_LEGACY + switch (cmd) { + case VCMD_set_cflags: + case VCMD_set_ccaps: + if (vx_check(0, VX_WATCH)) + return 0; + } +#endif + + switch (cmd) { +#ifdef CONFIG_IA32_EMULATION + case VCMD_get_rlimit: + return __COMPAT(vc_get_rlimit, id, data, compat); + case VCMD_set_rlimit: + return __COMPAT(vc_set_rlimit, id, data, compat); +#else + case VCMD_get_rlimit: + return vc_get_rlimit(id, data); + case VCMD_set_rlimit: + return vc_set_rlimit(id, data); +#endif + case VCMD_get_rlimit_mask: + return vc_get_rlimit_mask(id, data); + + case VCMD_get_vhi_name: + return vc_get_vhi_name(id, data); + case VCMD_set_vhi_name: + return vc_set_vhi_name(id, data); + + case VCMD_set_cflags: + return vc_set_cflags(id, data); + case VCMD_get_cflags: + return vc_get_cflags(id, data); + + case VCMD_set_ccaps: + return vc_set_ccaps(id, data); + case VCMD_get_ccaps: + return vc_get_ccaps(id, data); + + case VCMD_set_nflags: + return vc_set_nflags(id, data); + case VCMD_get_nflags: + return vc_get_nflags(id, data); + + case VCMD_set_ncaps: + return vc_set_ncaps(id, data); + case VCMD_get_ncaps: + return vc_get_ncaps(id, data); + + case VCMD_set_sched_v2: + return vc_set_sched_v2(id, data); + /* this is version 3 */ + case VCMD_set_sched: + return vc_set_sched(id, data); + + case VCMD_add_dlimit: + return __COMPAT(vc_add_dlimit, id, data, compat); + case VCMD_rem_dlimit: + return __COMPAT(vc_rem_dlimit, id, data, compat); + case VCMD_set_dlimit: + return __COMPAT(vc_set_dlimit, id, data, compat); + case VCMD_get_dlimit: + return __COMPAT(vc_get_dlimit, id, data, compat); + } + + /* below here only with VX_ADMIN */ + if (!vx_check(0, VX_ADMIN|VX_WATCH)) + return -EPERM; + + switch (cmd) { + case VCMD_ctx_kill: + return vc_ctx_kill(id, data); + + case VCMD_wait_exit: + return vc_wait_exit(id, data); + + case VCMD_create_context: +#ifdef CONFIG_VSERVER_LEGACY + return vc_ctx_create(id, NULL); +#else + return -ENOSYS; +#endif + + case VCMD_get_iattr: + return __COMPAT(vc_get_iattr, id, data, compat); + case VCMD_set_iattr: + return __COMPAT(vc_set_iattr, id, data, compat); + + case VCMD_enter_namespace: + return vc_enter_namespace(id, data); + + case VCMD_ctx_create_v0: +#ifdef CONFIG_VSERVER_LEGACY + if (id == 1) { + current->xid = 1; + return 1; + } +#endif + return vc_ctx_create(id, NULL); + case VCMD_ctx_create: + return vc_ctx_create(id, data); + case VCMD_ctx_migrate_v0: + return vc_ctx_migrate(id, NULL); + case VCMD_ctx_migrate: + return vc_ctx_migrate(id, data); + + case VCMD_net_create_v0: + return vc_net_create(id, NULL); + case VCMD_net_create: + return vc_net_create(id, data); + case VCMD_net_migrate: + return vc_net_migrate(id, data); + case VCMD_net_add: + return vc_net_add(id, data); + case VCMD_net_remove: + return vc_net_remove(id, data); + + } + return -ENOSYS; +} + +extern asmlinkage long +sys_vserver(uint32_t cmd, uint32_t id, void __user *data) +{ + long ret = do_vserver(cmd, id, data, 0); + + vxdprintk(VXD_CBIT(switch, 1), + "vc: VCMD_%02d_%d[%d] = %08lx(%ld)", + VC_CATEGORY(cmd), VC_COMMAND(cmd), + VC_VERSION(cmd), ret, ret); + return ret; +} + +#ifdef CONFIG_COMPAT + +extern asmlinkage long +sys32_vserver(uint32_t cmd, uint32_t id, void __user *data) +{ + long ret = do_vserver(cmd, id, data, 1); + + vxdprintk(VXD_CBIT(switch, 1), + "vc: VCMD_%02d_%d[%d] = %08lx(%ld)", + VC_CATEGORY(cmd), VC_COMMAND(cmd), + VC_VERSION(cmd), ret, ret); + return ret; +} + +#endif /* CONFIG_COMPAT */ diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/sysctl.c linux-2.6.17.13-g-v/kernel/vserver/sysctl.c --- linux-2.6.17.13/kernel/vserver/sysctl.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/sysctl.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,227 @@ +/* + * kernel/vserver/sysctl.c + * + * Virtual Context Support + * + * Copyright (C) 2004-2005 Herbert Pötzl + * + * V0.01 basic structure + * + */ + +#include +#include +#include +#include +#include +#include +#include + +#include +#include + + +#define CTL_VSERVER 4242 /* unused? */ + +enum { + CTL_DEBUG_ERROR = 0, + CTL_DEBUG_SWITCH = 1, + CTL_DEBUG_XID, + CTL_DEBUG_NID, + CTL_DEBUG_NET, + CTL_DEBUG_LIMIT, + CTL_DEBUG_CRES, + CTL_DEBUG_DLIM, + CTL_DEBUG_CVIRT, + CTL_DEBUG_MISC, +}; + + +unsigned int vx_debug_switch = 0; +unsigned int vx_debug_xid = 0; +unsigned int vx_debug_nid = 0; +unsigned int vx_debug_net = 0; +unsigned int vx_debug_limit = 0; +unsigned int vx_debug_cres = 0; +unsigned int vx_debug_dlim = 0; +unsigned int vx_debug_cvirt = 0; +unsigned int vx_debug_misc = 0; + + +static struct ctl_table_header *vserver_table_header; +static ctl_table vserver_table[]; + + +void vserver_register_sysctl(void) +{ + if (!vserver_table_header) { + vserver_table_header = register_sysctl_table(vserver_table, 1); + } + +} + +void vserver_unregister_sysctl(void) +{ + if (vserver_table_header) { + unregister_sysctl_table(vserver_table_header); + vserver_table_header = NULL; + } +} + + +static int proc_dodebug(ctl_table *table, int write, + struct file *filp, void __user *buffer, size_t *lenp, loff_t *ppos) +{ + char tmpbuf[20], *p, c; + unsigned int value; + size_t left, len; + + if ((*ppos && !write) || !*lenp) { + *lenp = 0; + return 0; + } + + left = *lenp; + + if (write) { + if (!access_ok(VERIFY_READ, buffer, left)) + return -EFAULT; + p = (char *) buffer; + while (left && __get_user(c, p) >= 0 && isspace(c)) + left--, p++; + if (!left) + goto done; + + if (left > sizeof(tmpbuf) - 1) + return -EINVAL; + if (copy_from_user(tmpbuf, p, left)) + return -EFAULT; + tmpbuf[left] = '\0'; + + for (p = tmpbuf, value = 0; '0' <= *p && *p <= '9'; p++, left--) + value = 10 * value + (*p - '0'); + if (*p && !isspace(*p)) + return -EINVAL; + while (left && isspace(*p)) + left--, p++; + *(unsigned int *) table->data = value; + } else { + if (!access_ok(VERIFY_WRITE, buffer, left)) + return -EFAULT; + len = sprintf(tmpbuf, "%d", *(unsigned int *) table->data); + if (len > left) + len = left; + if (__copy_to_user(buffer, tmpbuf, len)) + return -EFAULT; + if ((left -= len) > 0) { + if (put_user('\n', (char *)buffer + len)) + return -EFAULT; + left--; + } + } + +done: + *lenp -= left; + *ppos += *lenp; + return 0; +} + + + +static ctl_table debug_table[] = { + { + .ctl_name = CTL_DEBUG_SWITCH, + .procname = "debug_switch", + .data = &vx_debug_switch, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dodebug + }, + { + .ctl_name = CTL_DEBUG_XID, + .procname = "debug_xid", + .data = &vx_debug_xid, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dodebug + }, + { + .ctl_name = CTL_DEBUG_NID, + .procname = "debug_nid", + .data = &vx_debug_nid, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dodebug + }, + { + .ctl_name = CTL_DEBUG_NET, + .procname = "debug_net", + .data = &vx_debug_net, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dodebug + }, + { + .ctl_name = CTL_DEBUG_LIMIT, + .procname = "debug_limit", + .data = &vx_debug_limit, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dodebug + }, + { + .ctl_name = CTL_DEBUG_CRES, + .procname = "debug_cres", + .data = &vx_debug_cres, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dodebug + }, + { + .ctl_name = CTL_DEBUG_DLIM, + .procname = "debug_dlim", + .data = &vx_debug_dlim, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dodebug + }, + { + .ctl_name = CTL_DEBUG_CVIRT, + .procname = "debug_cvirt", + .data = &vx_debug_cvirt, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dodebug + }, + { + .ctl_name = CTL_DEBUG_MISC, + .procname = "debug_misc", + .data = &vx_debug_misc, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dodebug + }, + { .ctl_name = 0 } +}; + +static ctl_table vserver_table[] = { + { + .ctl_name = CTL_VSERVER, + .procname = "vserver", + .mode = 0555, + .child = debug_table + }, + { .ctl_name = 0 } +}; + + +EXPORT_SYMBOL_GPL(vx_debug_switch); +EXPORT_SYMBOL_GPL(vx_debug_xid); +EXPORT_SYMBOL_GPL(vx_debug_nid); +EXPORT_SYMBOL_GPL(vx_debug_net); +EXPORT_SYMBOL_GPL(vx_debug_limit); +EXPORT_SYMBOL_GPL(vx_debug_cres); +EXPORT_SYMBOL_GPL(vx_debug_dlim); +EXPORT_SYMBOL_GPL(vx_debug_cvirt); +EXPORT_SYMBOL_GPL(vx_debug_misc); + diff -NurpP --minimal linux-2.6.17.13/kernel/vserver/vci_config.h linux-2.6.17.13-g-v/kernel/vserver/vci_config.h --- linux-2.6.17.13/kernel/vserver/vci_config.h 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/kernel/vserver/vci_config.h 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,70 @@ + +enum { + VCI_KCBIT_LEGACY = 1, + VCI_KCBIT_LEGACYNET, + VCI_KCBIT_NGNET, + + VCI_KCBIT_PROC_SECURE, + VCI_KCBIT_HARDCPU, + VCI_KCBIT_HARDCPU_IDLE, + + VCI_KCBIT_LEGACY_VERSION = 15, + + VCI_KCBIT_DEBUG = 16, + VCI_KCBIT_HISTORY = 20, + VCI_KCBIT_TAGXID = 24, +}; + + +static inline uint32_t vci_kernel_config(void) +{ + return + /* various legacy options */ +#ifdef CONFIG_VSERVER_LEGACY + (1 << VCI_KCBIT_LEGACY) | +#endif +#ifdef CONFIG_VSERVER_LEGACYNET + (1 << VCI_KCBIT_LEGACYNET) | +#endif +#ifdef CONFIG_VSERVER_LEGACY_VERSION + (1 << VCI_KCBIT_LEGACY_VERSION) | +#endif + + /* configured features */ +#ifdef CONFIG_VSERVER_PROC_SECURE + (1 << VCI_KCBIT_PROC_SECURE) | +#endif +#ifdef CONFIG_VSERVER_HARDCPU + (1 << VCI_KCBIT_HARDCPU) | +#endif +#ifdef CONFIG_VSERVER_HARDCPU_IDLE + (1 << VCI_KCBIT_HARDCPU_IDLE) | +#endif + + /* debug options */ +#ifdef CONFIG_VSERVER_DEBUG + (1 << VCI_KCBIT_DEBUG) | +#endif +#ifdef CONFIG_VSERVER_HISTORY + (1 << VCI_KCBIT_HISTORY) | +#endif + + /* inode xid tagging */ +#if defined(CONFIG_INOXID_NONE) + (0 << VCI_KCBIT_TAGXID) | +#elif defined(CONFIG_INOXID_UID16) + (1 << VCI_KCBIT_TAGXID) | +#elif defined(CONFIG_INOXID_GID16) + (2 << VCI_KCBIT_TAGXID) | +#elif defined(CONFIG_INOXID_UGID24) + (3 << VCI_KCBIT_TAGXID) | +#elif defined(CONFIG_INOXID_INTERN) + (4 << VCI_KCBIT_TAGXID) | +#elif defined(CONFIG_INOXID_RUNTIME) + (5 << VCI_KCBIT_TAGXID) | +#else + (7 << VCI_KCBIT_TAGXID) | +#endif + 0; +} + diff -NurpP --minimal linux-2.6.17.13/localversion-grsec linux-2.6.17.13-g-v/localversion-grsec --- linux-2.6.17.13/localversion-grsec 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/localversion-grsec 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1 @@ +-grsec2.1.9 diff -NurpP --minimal linux-2.6.17.13/localversion-vserver linux-2.6.17.13-g-v/localversion-vserver --- linux-2.6.17.13/localversion-vserver 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/localversion-vserver 2006-09-25 15:04:50.000000000 +0200 @@ -0,0 +1 @@ +-vs2.0.2.1 diff -NurpP --minimal linux-2.6.17.13/mm/filemap.c linux-2.6.17.13-g-v/mm/filemap.c --- linux-2.6.17.13/mm/filemap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/filemap.c 2006-09-20 15:00:12.000000000 +0200 @@ -30,6 +30,7 @@ #include #include #include +#include #include "filemap.h" #include "internal.h" @@ -1650,7 +1651,13 @@ int generic_file_mmap(struct file * file struct address_space *mapping = file->f_mapping; if (!mapping->a_ops->readpage) - return -ENOEXEC; + return -ENODEV; + +#if defined(CONFIG_PAX_PAGEEXEC) && defined(CONFIG_X86_32) + if ((vma->vm_mm->pax_flags & MF_PAX_PAGEEXEC) && !(vma->vm_flags & VM_EXEC)) + vma->vm_page_prot = __pgprot(pte_val(pte_exprotect(__pte(pgprot_val(vma->vm_page_prot))))); +#endif + file_accessed(file); vma->vm_ops = &generic_file_vm_ops; return 0; @@ -1872,6 +1879,7 @@ inline int generic_write_checks(struct f *pos = i_size_read(inode); if (limit != RLIM_INFINITY) { + gr_learn_resource(current, RLIMIT_FSIZE,*pos, 0); if (*pos >= limit) { send_sig(SIGXFSZ, current, 0); return -EFBIG; diff -NurpP --minimal linux-2.6.17.13/mm/filemap_xip.c linux-2.6.17.13-g-v/mm/filemap_xip.c --- linux-2.6.17.13/mm/filemap_xip.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/filemap_xip.c 2006-09-20 15:00:12.000000000 +0200 @@ -13,6 +13,7 @@ #include #include #include +#include #include #include "filemap.h" diff -NurpP --minimal linux-2.6.17.13/mm/fremap.c linux-2.6.17.13-g-v/mm/fremap.c --- linux-2.6.17.13/mm/fremap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/fremap.c 2006-09-20 15:00:12.000000000 +0200 @@ -15,6 +15,7 @@ #include #include #include +#include #include #include @@ -74,6 +75,8 @@ int install_page(struct mm_struct *mm, s err = -ENOMEM; if (page_mapcount(page) > INT_MAX/2) goto unlock; + if (!vx_rsspages_avail(mm, 1)) + goto unlock; if (pte_none(*pte) || !zap_pte(mm, vma, addr, pte)) inc_mm_counter(mm, file_rss); diff -NurpP --minimal linux-2.6.17.13/mm/hugetlb.c linux-2.6.17.13-g-v/mm/hugetlb.c --- linux-2.6.17.13/mm/hugetlb.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/hugetlb.c 2006-09-20 15:00:12.000000000 +0200 @@ -19,6 +19,7 @@ #include #include +#include #include "internal.h" const unsigned long hugetlb_zero = 0, hugetlb_infinity = ~0UL; diff -NurpP --minimal linux-2.6.17.13/mm/madvise.c linux-2.6.17.13-g-v/mm/madvise.c --- linux-2.6.17.13/mm/madvise.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/madvise.c 2006-09-20 15:00:12.000000000 +0200 @@ -15,10 +15,47 @@ * We can potentially split a vm area into separate * areas, each area with its own behavior. */ + +#ifdef CONFIG_PAX_SEGMEXEC +static long __madvise_behavior(struct vm_area_struct * vma, + struct vm_area_struct **prev, + unsigned long start, unsigned long end, int behavior); + static long madvise_behavior(struct vm_area_struct * vma, struct vm_area_struct **prev, unsigned long start, unsigned long end, int behavior) { + if (vma->vm_flags & VM_MIRROR) { + struct vm_area_struct * vma_m, * prev_m; + unsigned long start_m, end_m; + int error; + + start_m = vma->vm_start + vma->vm_mirror; + vma_m = find_vma_prev(vma->vm_mm, start_m, &prev_m); + if (vma_m && vma_m->vm_start == start_m && (vma_m->vm_flags & VM_MIRROR)) { + start_m = start + vma->vm_mirror; + end_m = end + vma->vm_mirror; + error = __madvise_behavior(vma_m, &prev_m, start_m, end_m, behavior); + if (error) + return error; + } else { + printk("PAX: VMMIRROR: madvise bug in %s, %08lx\n", current->comm, vma->vm_start); + return -ENOMEM; + } + } + + return __madvise_behavior(vma, prev, start, end, behavior); +} + +static long __madvise_behavior(struct vm_area_struct * vma, + struct vm_area_struct **prev, + unsigned long start, unsigned long end, int behavior) +#else +static long madvise_behavior(struct vm_area_struct * vma, + struct vm_area_struct **prev, + unsigned long start, unsigned long end, int behavior) +#endif +{ struct mm_struct * mm = vma->vm_mm; int error = 0; pgoff_t pgoff; diff -NurpP --minimal linux-2.6.17.13/mm/memory.c linux-2.6.17.13-g-v/mm/memory.c --- linux-2.6.17.13/mm/memory.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/memory.c 2006-09-20 15:00:12.000000000 +0200 @@ -48,6 +48,7 @@ #include #include #include +#include #include #include @@ -320,6 +321,11 @@ int __pte_alloc(struct mm_struct *mm, pm int __pte_alloc_kernel(pmd_t *pmd, unsigned long address) { + +#ifdef CONFIG_PAX_KERNEXEC + unsigned long cr0; +#endif + pte_t *new = pte_alloc_one_kernel(&init_mm, address); if (!new) return -ENOMEM; @@ -327,8 +333,19 @@ int __pte_alloc_kernel(pmd_t *pmd, unsig spin_lock(&init_mm.page_table_lock); if (pmd_present(*pmd)) /* Another has populated it */ pte_free_kernel(new); - else + else { + +#ifdef CONFIG_PAX_KERNEXEC + pax_open_kernel(cr0); +#endif + pmd_populate_kernel(&init_mm, pmd, new); + +#ifdef CONFIG_PAX_KERNEXEC + pax_close_kernel(cr0); +#endif + + } spin_unlock(&init_mm.page_table_lock); return 0; } @@ -1421,6 +1438,88 @@ static inline void cow_user_page(struct copy_user_highpage(dst, src, va); } +#ifdef CONFIG_PAX_SEGMEXEC +/* PaX: if vma is mirrored, synchronize the mirror's PTE + * + * the ptl of the lower mapped page is held on entry and is not released on exit + * or inside to ensure atomic changes to the PTE states (swapout, mremap, munmap, etc) + */ +static void pax_mirror_fault(struct vm_area_struct *vma, unsigned long address, pte_t *pte) +{ + struct mm_struct *mm = vma->vm_mm; + unsigned long address_m, pfn_m; + struct vm_area_struct * vma_m = NULL; + pte_t * pte_m, entry_m; + struct page * page_m = NULL; + + address_m = vma->vm_start + vma->vm_mirror; + vma_m = find_vma(mm, address_m); + BUG_ON(!vma_m || vma_m->vm_start != address_m); + + address_m = address + vma->vm_mirror; + pte_m = pte_offset_map_nested(pmd_offset(pud_offset(pgd_offset(mm, address_m), address_m), address_m), address_m); + + if (pte_same(*pte, *pte_m)) { + pte_unmap_nested(pte_m); + return; + } + + pfn_m = pte_pfn(*pte); + if (pte_present(*pte_m)) { + page_m = vm_normal_page(vma_m, address_m, *pte_m); + if (page_m) { + flush_cache_page(vma_m, address_m, pfn_m); + flush_icache_page(vma_m, page_m); + } + } + + if (pte_present(*pte_m)) + entry_m = ptep_clear_flush(vma_m, address_m, pte_m); + else + entry_m = ptep_get_and_clear(mm, address_m, pte_m); + + if (pte_none(entry_m)) { + } else if (pte_present(entry_m)) { + if (page_m) { + page_remove_rmap(page_m); + if (PageAnon(page_m)) + dec_mm_counter(mm, anon_rss); + else + dec_mm_counter(mm, file_rss); + page_cache_release(page_m); + } + } else if (!pte_file(entry_m)) { + free_swap_and_cache(pte_to_swp_entry(entry_m)); + } else { + printk(KERN_ERR "PAX: VMMIRROR: bug in mirror_fault: %08lx, %08lx, %08lx, %08lx\n", + address, vma->vm_start, address_m, vma_m->vm_start); + } + + page_m = vm_normal_page(vma, address, *pte); + entry_m = pfn_pte(pfn_m, vma_m->vm_page_prot); + if (pte_write(*pte)) + entry_m = maybe_mkwrite(pte_mkdirty(entry_m), vma_m); + if (page_m) { + page_cache_get(page_m); + /* + * we can test PAGE_MAPPING_ANON without holding page_map_lock because + * we hold the page table lock and have a reference to page_m + */ + if (PageAnon(page_m)) { + page_add_anon_rmap(page_m, vma_m, address_m); + inc_mm_counter(mm, anon_rss); + } else { + page_add_file_rmap(page_m); + inc_mm_counter(mm, file_rss); + } + } + set_pte_at(mm, address_m, pte_m, entry_m); + update_mmu_cache(vma_m, address_m, entry_m); + lazy_mmu_prot_update(entry_m); + pte_unmap_nested(pte_m); +} +#endif + /* * This routine handles present pages, when users try to write * to a shared page. It is done by copying the page to a new address @@ -1511,6 +1610,12 @@ gotten: /* Free the old page.. */ new_page = old_page; ret |= VM_FAULT_WRITE; + +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) + pax_mirror_fault(vma, address, page_table); +#endif + } if (new_page) page_cache_release(new_page); @@ -1761,6 +1866,7 @@ int vmtruncate(struct inode * inode, lof do_expand: limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur; + gr_learn_resource(current, RLIMIT_FSIZE, offset, 1); if (limit != RLIM_INFINITY && offset > limit) goto out_sig; if (offset > inode->i_sb->s_maxbytes) @@ -1901,6 +2007,10 @@ again: grab_swap_token(); } + if (!vx_rsspages_avail(mm, 1)) { + ret = VM_FAULT_OOM; + goto out; + } mark_page_accessed(page); lock_page(page); if (!PageSwapCache(page)) { @@ -1950,6 +2060,12 @@ again: /* No need to invalidate - it was non-present before */ update_mmu_cache(vma, address, pte); lazy_mmu_prot_update(pte); + +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) + pax_mirror_fault(vma, address, page_table); +#endif + unlock: pte_unmap_unlock(page_table, ptl); out: @@ -1978,6 +2094,8 @@ static int do_anonymous_page(struct mm_s /* Allocate our own private page. */ pte_unmap(page_table); + if (!vx_rsspages_avail(mm, 1)) + goto oom; if (unlikely(anon_vma_prepare(vma))) goto oom; page = alloc_zeroed_user_highpage(vma, address); @@ -2012,6 +2130,12 @@ static int do_anonymous_page(struct mm_s /* No need to invalidate - it was non-present before */ update_mmu_cache(vma, address, entry); lazy_mmu_prot_update(entry); + +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) + pax_mirror_fault(vma, address, page_table); +#endif + unlock: pte_unmap_unlock(page_table, ptl); return VM_FAULT_MINOR; @@ -2056,6 +2180,9 @@ static int do_no_page(struct mm_struct * smp_rmb(); /* serializes i_size against truncate_count */ } retry: + /* FIXME: is that check useful here? */ + if (!vx_rsspages_avail(mm, 1)) + return VM_FAULT_OOM; new_page = vma->vm_ops->nopage(vma, address & PAGE_MASK, &ret); /* * No smp_rmb is needed here as long as there's a full @@ -2137,6 +2264,12 @@ retry: /* no need to invalidate: a not-present page shouldn't be cached */ update_mmu_cache(vma, address, entry); lazy_mmu_prot_update(entry); + +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) + pax_mirror_fault(vma, address, page_table); +#endif + unlock: pte_unmap_unlock(page_table, ptl); return ret; @@ -2246,6 +2379,12 @@ static inline int handle_pte_fault(struc flush_tlb_page(vma, address); } unlock: + +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) + pax_mirror_fault(vma, address, pte); +#endif + pte_unmap_unlock(pte, ptl); return VM_FAULT_MINOR; } @@ -2268,6 +2407,49 @@ int __handle_mm_fault(struct mm_struct * if (unlikely(is_vm_hugetlb_page(vma))) return hugetlb_fault(mm, vma, address, write_access); +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) { + unsigned long address_m; + struct vm_area_struct * vma_m; + pgd_t *pgd_m; + pud_t *pud_m; + pmd_t *pmd_m; + + address_m = vma->vm_start + vma->vm_mirror; + vma_m = find_vma(mm, address_m); + + /* PaX: sanity checks */ + if (!vma_m) { + printk(KERN_ERR "PAX: VMMIRROR: fault bug, %08lx, %p, %08lx, %p\n", + address, vma, address_m, vma_m); + return VM_FAULT_SIGBUS; + } else if (!(vma_m->vm_flags & VM_MIRROR) || + vma_m->vm_start != address_m || + vma->vm_end - vma->vm_start != vma_m->vm_end - vma_m->vm_start) + { + printk(KERN_ERR "PAX: VMMIRROR: fault bug2, %08lx, %08lx, %08lx, %08lx, %08lx\n", + address, vma->vm_start, vma_m->vm_start, vma->vm_end, vma_m->vm_end); + return VM_FAULT_SIGBUS; + } + + if (address_m < address) { + address += vma->vm_mirror; + vma = vma_m; + } + + address_m = address + vma->vm_mirror; + pgd_m = pgd_offset(mm, address_m); + pud_m = pud_alloc(mm, pgd_m, address_m); + if (!pud_m) + return VM_FAULT_OOM; + pmd_m = pmd_alloc(mm, pud_m, address_m); + if (!pmd_m) + return VM_FAULT_OOM; + if (!pmd_present(*pmd_m) && __pte_alloc(mm, pmd_m, address_m)) + return VM_FAULT_OOM; + } +#endif + pgd = pgd_offset(mm, address); pud = pud_alloc(mm, pgd, address); if (!pud) diff -NurpP --minimal linux-2.6.17.13/mm/mempolicy.c linux-2.6.17.13-g-v/mm/mempolicy.c --- linux-2.6.17.13/mm/mempolicy.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/mempolicy.c 2006-09-20 15:00:12.000000000 +0200 @@ -87,6 +87,7 @@ #include #include #include +#include #include #include @@ -346,6 +347,12 @@ check_range(struct mm_struct *mm, unsign if (prev && prev->vm_end < vma->vm_start) return ERR_PTR(-EFAULT); } + +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) + return ERR_PTR(-EFAULT); +#endif + if (!is_vm_hugetlb_page(vma) && ((flags & MPOL_MF_STRICT) || ((flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) && diff -NurpP --minimal linux-2.6.17.13/mm/mlock.c linux-2.6.17.13-g-v/mm/mlock.c --- linux-2.6.17.13/mm/mlock.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/mlock.c 2006-09-20 15:00:12.000000000 +0200 @@ -10,44 +10,47 @@ #include #include #include +#include +#include +static int __mlock_fixup(struct vm_area_struct *vma, struct vm_area_struct **prev, + unsigned long start, unsigned long end, unsigned int newflags); static int mlock_fixup(struct vm_area_struct *vma, struct vm_area_struct **prev, unsigned long start, unsigned long end, unsigned int newflags) { struct mm_struct * mm = vma->vm_mm; - pgoff_t pgoff; int pages; - int ret = 0; - - if (newflags == vma->vm_flags) { - *prev = vma; - goto out; - } + int ret; - pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT); - *prev = vma_merge(mm, *prev, start, end, newflags, vma->anon_vma, - vma->vm_file, pgoff, vma_policy(vma)); - if (*prev) { - vma = *prev; - goto success; - } +#ifdef CONFIG_PAX_SEGMEXEC + struct vm_area_struct * vma_m = NULL, *prev_m; + unsigned long start_m = 0UL, end_m = 0UL, newflags_m = 0UL; - *prev = vma; + if (vma->vm_flags & VM_MIRROR) { + start_m = vma->vm_start + vma->vm_mirror; + vma_m = find_vma_prev(mm, start_m, &prev_m); + if (!vma_m || vma_m->vm_start != start_m || !(vma_m->vm_flags & VM_MIRROR)) { + printk("PAX: VMMIRROR: mlock bug in %s, %08lx\n", current->comm, vma->vm_start); + return -ENOMEM; + } - if (start != vma->vm_start) { - ret = split_vma(mm, vma, start, 1); + start_m = start + vma->vm_mirror; + end_m = end + vma->vm_mirror; + if (newflags & VM_LOCKED) + newflags_m = vma_m->vm_flags | VM_LOCKED; + else + newflags_m = vma_m->vm_flags & ~VM_LOCKED; + ret = __mlock_fixup(vma_m, &prev_m, start_m, end_m, newflags_m); if (ret) - goto out; + return ret; } +#endif - if (end != vma->vm_end) { - ret = split_vma(mm, vma, end, 0); - if (ret) - goto out; - } + ret = __mlock_fixup(vma, prev, start, end, newflags); + if (ret) + return ret; -success: /* * vm_flags is protected by the mmap_sem held in write mode. * It's okay if try_to_unmap_one unmaps a page just after we @@ -55,6 +58,11 @@ success: */ vma->vm_flags = newflags; +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) + vma_m->vm_flags = newflags_m; +#endif + /* * Keep track of amount of locked VM. */ @@ -65,7 +73,49 @@ success: ret = make_pages_present(start, end); } - vma->vm_mm->locked_vm -= pages; + vx_vmlocked_sub(mm, pages); + +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) + vx_vmlocked_sub(mm, pages); +#endif + + if (ret == -ENOMEM) + ret = -EAGAIN; + return ret; +} + +static int __mlock_fixup(struct vm_area_struct *vma, struct vm_area_struct **prev, + unsigned long start, unsigned long end, unsigned int newflags) +{ + struct mm_struct * mm = vma->vm_mm; + pgoff_t pgoff; + int ret = 0; + + if (newflags == vma->vm_flags) { + *prev = vma; + goto out; + } + + pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT); + *prev = vma_merge(mm, *prev, start, end, newflags, vma->anon_vma, + vma->vm_file, pgoff, vma_policy(vma)); + if (*prev) { + vma = *prev; + goto out; + } + + *prev = vma; + + if (start != vma->vm_start) { + ret = split_vma(mm, vma, start, 1); + if (ret) + goto out; + } + + if (end != vma->vm_end) + ret = split_vma(mm, vma, end, 0); + out: if (ret == -ENOMEM) ret = -EAGAIN; @@ -84,6 +134,17 @@ static int do_mlock(unsigned long start, return -EINVAL; if (end == start) return 0; + +#ifdef CONFIG_PAX_SEGMEXEC + if (current->mm->pax_flags & MF_PAX_SEGMEXEC) { + if (end > SEGMEXEC_TASK_SIZE) + return -EINVAL; + } else +#endif + + if (end > TASK_SIZE) + return -EINVAL; + vma = find_vma_prev(current->mm, start, &prev); if (!vma || vma->vm_start > start) return -ENOMEM; @@ -123,7 +184,7 @@ static int do_mlock(unsigned long start, asmlinkage long sys_mlock(unsigned long start, size_t len) { - unsigned long locked; + unsigned long locked, grow; unsigned long lock_limit; int error = -ENOMEM; @@ -134,15 +195,19 @@ asmlinkage long sys_mlock(unsigned long len = PAGE_ALIGN(len + (start & ~PAGE_MASK)); start &= PAGE_MASK; - locked = len >> PAGE_SHIFT; - locked += current->mm->locked_vm; + grow = len >> PAGE_SHIFT; + if (!vx_vmlocked_avail(current->mm, grow)) + goto out; + locked = current->mm->locked_vm + grow; lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur; lock_limit >>= PAGE_SHIFT; /* check against resource limits */ + gr_learn_resource(current, RLIMIT_MEMLOCK, (current->mm->locked_vm << PAGE_SHIFT) + len, 1); if ((locked <= lock_limit) || capable(CAP_IPC_LOCK)) error = do_mlock(start, len, 1); +out: up_write(&current->mm->mmap_sem); return error; } @@ -173,6 +238,16 @@ static int do_mlockall(int flags) for (vma = current->mm->mmap; vma ; vma = prev->vm_next) { unsigned int newflags; +#ifdef CONFIG_PAX_SEGMEXEC + if (current->mm->pax_flags & MF_PAX_SEGMEXEC) { + if (vma->vm_end > SEGMEXEC_TASK_SIZE) + break; + } else +#endif + + if (vma->vm_end > TASK_SIZE) + break; + newflags = vma->vm_flags | VM_LOCKED; if (!(flags & MCL_CURRENT)) newflags &= ~VM_LOCKED; @@ -202,6 +277,9 @@ asmlinkage long sys_mlockall(int flags) lock_limit >>= PAGE_SHIFT; ret = -ENOMEM; + if (!vx_vmlocked_avail(current->mm, current->mm->total_vm)) + goto out; + gr_learn_resource(current, RLIMIT_MEMLOCK, current->mm->total_vm, 1); if (!(flags & MCL_CURRENT) || (current->mm->total_vm <= lock_limit) || capable(CAP_IPC_LOCK)) ret = do_mlockall(flags); diff -NurpP --minimal linux-2.6.17.13/mm/mmap.c linux-2.6.17.13-g-v/mm/mmap.c --- linux-2.6.17.13/mm/mmap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/mmap.c 2006-09-20 15:00:12.000000000 +0200 @@ -25,6 +25,7 @@ #include #include #include +#include #include #include @@ -240,6 +241,7 @@ asmlinkage unsigned long sys_brk(unsigne * not page aligned -Ram Gupta */ rlim = current->signal->rlim[RLIMIT_DATA].rlim_cur; + gr_learn_resource(current, RLIMIT_DATA, brk - mm->start_data, 1); if (rlim < RLIM_INFINITY && brk - mm->start_data > rlim) goto out; @@ -628,11 +630,17 @@ again: remove_next = 1 + (end > next-> * If the vma has a ->close operation then the driver probably needs to release * per-vma resources, so we don't attempt to merge those. */ +#ifdef CONFIG_PAX_SEGMEXEC +#define VM_SPECIAL (VM_IO | VM_DONTEXPAND | VM_RESERVED | VM_PFNMAP | VM_MIRROR) +#else #define VM_SPECIAL (VM_IO | VM_DONTEXPAND | VM_RESERVED | VM_PFNMAP) +#endif static inline int is_mergeable_vma(struct vm_area_struct *vma, struct file *file, unsigned long vm_flags) { + if ((vma->vm_flags | vm_flags) & VM_SPECIAL) + return 0; if (vma->vm_flags != vm_flags) return 0; if (vma->vm_file != file) @@ -857,14 +865,11 @@ none: void vm_stat_account(struct mm_struct *mm, unsigned long flags, struct file *file, long pages) { - const unsigned long stack_flags - = VM_STACK_FLAGS & (VM_GROWSUP|VM_GROWSDOWN); - if (file) { mm->shared_vm += pages; if ((flags & (VM_EXEC|VM_WRITE)) == VM_EXEC) mm->exec_vm += pages; - } else if (flags & stack_flags) + } else if (flags & (VM_GROWSUP|VM_GROWSDOWN)) mm->stack_vm += pages; if (flags & (VM_RESERVED|VM_IO)) mm->reserved_vm += pages; @@ -875,10 +880,55 @@ void vm_stat_account(struct mm_struct *m * The caller must hold down_write(current->mm->mmap_sem). */ +#ifdef CONFIG_PAX_SEGMEXEC +static unsigned long __do_mmap_pgoff(struct file * file, unsigned long addr, + unsigned long len, unsigned long prot, + unsigned long flags, unsigned long pgoff); + unsigned long do_mmap_pgoff(struct file * file, unsigned long addr, unsigned long len, unsigned long prot, unsigned long flags, unsigned long pgoff) { + unsigned long ret = -EINVAL; + + if (flags & MAP_MIRROR) + return ret; + + if ((current->mm->pax_flags & MF_PAX_SEGMEXEC) && + (len > SEGMEXEC_TASK_SIZE || (addr > SEGMEXEC_TASK_SIZE-len))) + return ret; + + ret = __do_mmap_pgoff(file, addr, len, prot, flags, pgoff); + + if ((current->mm->pax_flags & MF_PAX_SEGMEXEC) && ret < TASK_SIZE && ((flags & MAP_TYPE) == MAP_PRIVATE) + +#ifdef CONFIG_PAX_MPROTECT + && (!(current->mm->pax_flags & MF_PAX_MPROTECT) || ((prot & PROT_EXEC) && file && !(prot & PROT_WRITE))) +#endif + + ) + { + unsigned long ret_m; + prot = prot & PROT_EXEC ? prot & ~PROT_WRITE : PROT_NONE; + ret_m = __do_mmap_pgoff(NULL, ret + SEGMEXEC_TASK_SIZE, 0UL, prot, flags | MAP_MIRROR | MAP_FIXED, ret); + if (ret_m >= TASK_SIZE) { + do_munmap(current->mm, ret, len); + ret = ret_m; + } + } + + return ret; +} + +static unsigned long __do_mmap_pgoff(struct file * file, unsigned long addr, + unsigned long len, unsigned long prot, + unsigned long flags, unsigned long pgoff) +#else +unsigned long do_mmap_pgoff(struct file * file, unsigned long addr, + unsigned long len, unsigned long prot, + unsigned long flags, unsigned long pgoff) +#endif +{ struct mm_struct * mm = current->mm; struct vm_area_struct * vma, * prev; struct inode *inode; @@ -889,6 +939,28 @@ unsigned long do_mmap_pgoff(struct file int accountable = 1; unsigned long charged = 0, reqprot = prot; +#ifdef CONFIG_PAX_SEGMEXEC + struct vm_area_struct * vma_m = NULL; + + if (flags & MAP_MIRROR) { + /* PaX: sanity checks, to be removed when proved to be stable */ + if (file || len || ((flags & MAP_TYPE) != MAP_PRIVATE)) + return -EINVAL; + + vma_m = find_vma(mm, pgoff); + + if (!vma_m || is_vm_hugetlb_page(vma_m) || + vma_m->vm_start != pgoff || + (vma_m->vm_flags & VM_SPECIAL) || + (prot & PROT_WRITE)) + return -EINVAL; + + file = vma_m->vm_file; + pgoff = vma_m->vm_pgoff; + len = vma_m->vm_end - vma_m->vm_start; + } +#endif + if (file) { if (is_file_hugepages(file)) accountable = 0; @@ -906,7 +978,7 @@ unsigned long do_mmap_pgoff(struct file * (the exception is when the underlying filesystem is noexec * mounted, in which case we dont add PROT_EXEC.) */ - if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC)) + if ((prot & (PROT_READ | PROT_WRITE)) && (current->personality & READ_IMPLIES_EXEC)) if (!(file && (file->f_vfsmnt->mnt_flags & MNT_NOEXEC))) prot |= PROT_EXEC; @@ -933,7 +1005,7 @@ unsigned long do_mmap_pgoff(struct file /* Obtain the address to map to. we verify (or select) it and ensure * that it represents a valid section of the address space. */ - addr = get_unmapped_area(file, addr, len, pgoff, flags); + addr = get_unmapped_area(file, addr, len, pgoff, flags | ((prot & PROT_EXEC) ? MAP_EXECUTABLE : 0)); if (addr & ~PAGE_MASK) return addr; @@ -944,6 +1016,24 @@ unsigned long do_mmap_pgoff(struct file vm_flags = calc_vm_prot_bits(prot) | calc_vm_flag_bits(flags) | mm->def_flags | VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC; + if (file && (file->f_vfsmnt->mnt_flags & MNT_NOEXEC)) + vm_flags &= ~VM_MAYEXEC; + +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) + if (mm->pax_flags & (MF_PAX_PAGEEXEC | MF_PAX_SEGMEXEC)) { + +#ifdef CONFIG_PAX_MPROTECT + if (mm->pax_flags & MF_PAX_MPROTECT) { + if ((prot & (PROT_WRITE | PROT_EXEC)) != PROT_EXEC) + vm_flags &= ~(VM_EXEC | VM_MAYEXEC); + else + vm_flags &= ~(VM_WRITE | VM_MAYWRITE); + } +#endif + + } +#endif + if (flags & MAP_LOCKED) { if (!can_do_mlock()) return -EPERM; @@ -956,6 +1046,7 @@ unsigned long do_mmap_pgoff(struct file locked += mm->locked_vm; lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur; lock_limit >>= PAGE_SHIFT; + gr_learn_resource(current, RLIMIT_MEMLOCK, locked << PAGE_SHIFT, 1); if (locked > lock_limit && !capable(CAP_IPC_LOCK)) return -EAGAIN; } @@ -1003,6 +1094,11 @@ unsigned long do_mmap_pgoff(struct file /* * Set pgoff according to addr for anon_vma. */ + +#ifdef CONFIG_PAX_SEGMEXEC + if (!(flags & MAP_MIRROR)) +#endif + pgoff = addr >> PAGE_SHIFT; break; default: @@ -1014,14 +1110,17 @@ unsigned long do_mmap_pgoff(struct file if (error) return error; + if (!gr_acl_handle_mmap(file, prot)) + return -EACCES; + /* Clear old maps */ error = -ENOMEM; -munmap_back: vma = find_vma_prepare(mm, addr, &prev, &rb_link, &rb_parent); if (vma && vma->vm_start < addr + len) { if (do_munmap(mm, addr, len)) return -ENOMEM; - goto munmap_back; + vma = find_vma_prepare(mm, addr, &prev, &rb_link, &rb_parent); + BUG_ON(vma && vma->vm_start < addr + len); } /* Check against address space limit. */ @@ -1069,7 +1168,14 @@ munmap_back: vma->vm_start = addr; vma->vm_end = addr + len; vma->vm_flags = vm_flags; - vma->vm_page_prot = protection_map[vm_flags & 0x0f]; + +#if defined(CONFIG_PAX_PAGEEXEC) && defined(CONFIG_X86_32) + if ((file || !(mm->pax_flags & MF_PAX_PAGEEXEC)) && (vm_flags & (VM_READ|VM_WRITE))) + vma->vm_page_prot = protection_map[(vm_flags | VM_EXEC) & (VM_READ|VM_WRITE|VM_EXEC|VM_SHARED)]; + else +#endif + + vma->vm_page_prot = protection_map[vm_flags & (VM_READ|VM_WRITE|VM_EXEC|VM_SHARED)]; vma->vm_pgoff = pgoff; if (file) { @@ -1093,6 +1199,14 @@ munmap_back: goto free_vma; } +#ifdef CONFIG_PAX_SEGMEXEC + if (flags & MAP_MIRROR) { + vma_m->vm_flags |= VM_MIRROR; + vma_m->vm_mirror = vma->vm_start - vma_m->vm_start; + vma->vm_mirror = vma_m->vm_start - vma->vm_start; + } +#endif + /* We set VM_ACCOUNT in a shared mapping's vm_flags, to inform * shmem_zero_setup (perhaps called through /dev/zero's ->mmap) * that memory reservation must be checked; but that reservation @@ -1126,10 +1240,11 @@ munmap_back: kmem_cache_free(vm_area_cachep, vma); } out: - mm->total_vm += len >> PAGE_SHIFT; + vx_vmpages_add(mm, len >> PAGE_SHIFT); vm_stat_account(mm, vm_flags, file, len >> PAGE_SHIFT); + track_exec_limit(mm, addr, addr + len, vm_flags); if (vm_flags & VM_LOCKED) { - mm->locked_vm += len >> PAGE_SHIFT; + vx_vmlocked_add(mm, len >> PAGE_SHIFT); make_pages_present(addr, addr + len); } if (flags & MAP_POPULATE) { @@ -1182,6 +1297,10 @@ arch_get_unmapped_area(struct file *filp if (len > TASK_SIZE) return -ENOMEM; +#ifdef CONFIG_PAX_RANDMMAP + if (!(mm->pax_flags & MF_PAX_RANDMMAP) || !filp) +#endif + if (addr) { addr = PAGE_ALIGN(addr); vma = find_vma(mm, addr); @@ -1192,7 +1311,7 @@ arch_get_unmapped_area(struct file *filp if (len > mm->cached_hole_size) { start_addr = addr = mm->free_area_cache; } else { - start_addr = addr = TASK_UNMAPPED_BASE; + start_addr = addr = mm->mmap_base; mm->cached_hole_size = 0; } @@ -1204,9 +1323,8 @@ full_search: * Start a new search - just in case we missed * some holes. */ - if (start_addr != TASK_UNMAPPED_BASE) { - addr = TASK_UNMAPPED_BASE; - start_addr = addr; + if (start_addr != mm->mmap_base) { + start_addr = addr = mm->mmap_base; mm->cached_hole_size = 0; goto full_search; } @@ -1231,7 +1349,7 @@ void arch_unmap_area(struct mm_struct *m /* * Is this a new hole at the lowest possible address? */ - if (addr >= TASK_UNMAPPED_BASE && addr < mm->free_area_cache) { + if (addr >= mm->mmap_base && addr < mm->free_area_cache) { mm->free_area_cache = addr; mm->cached_hole_size = ~0UL; } @@ -1249,12 +1367,16 @@ arch_get_unmapped_area_topdown(struct fi { struct vm_area_struct *vma; struct mm_struct *mm = current->mm; - unsigned long addr = addr0; + unsigned long base = mm->mmap_base, addr = addr0; /* requested length too big for entire address space */ if (len > TASK_SIZE) return -ENOMEM; +#ifdef CONFIG_PAX_RANDMMAP + if (!(mm->pax_flags & MF_PAX_RANDMMAP) || !filp) +#endif + /* requesting a specific address */ if (addr) { addr = PAGE_ALIGN(addr); @@ -1312,13 +1434,21 @@ bottomup: * can happen with large stack limits and large mmap() * allocations. */ + mm->mmap_base = TASK_UNMAPPED_BASE; + +#ifdef CONFIG_PAX_RANDMMAP + if (mm->pax_flags & MF_PAX_RANDMMAP) + mm->mmap_base += mm->delta_mmap; +#endif + + mm->free_area_cache = mm->mmap_base; mm->cached_hole_size = ~0UL; - mm->free_area_cache = TASK_UNMAPPED_BASE; addr = arch_get_unmapped_area(filp, addr0, len, pgoff, flags); /* * Restore the topdown base: */ - mm->free_area_cache = mm->mmap_base; + mm->mmap_base = base; + mm->free_area_cache = base; mm->cached_hole_size = ~0UL; return addr; @@ -1334,8 +1464,10 @@ void arch_unmap_area_topdown(struct mm_s mm->free_area_cache = addr; /* dont allow allocations above current base */ - if (mm->free_area_cache > mm->mmap_base) + if (mm->free_area_cache > mm->mmap_base) { mm->free_area_cache = mm->mmap_base; + mm->cached_hole_size = ~0UL; + } } unsigned long @@ -1468,6 +1600,7 @@ static int acct_stack_growth(struct vm_a return -ENOMEM; /* Stack limit test */ + gr_learn_resource(current, RLIMIT_STACK, size, 1); if (size > rlim[RLIMIT_STACK].rlim_cur) return -ENOMEM; @@ -1477,6 +1610,7 @@ static int acct_stack_growth(struct vm_a unsigned long limit; locked = mm->locked_vm + grow; limit = rlim[RLIMIT_MEMLOCK].rlim_cur >> PAGE_SHIFT; + gr_learn_resource(current, RLIMIT_MEMLOCK, locked << PAGE_SHIFT, 1); if (locked > limit && !capable(CAP_IPC_LOCK)) return -ENOMEM; } @@ -1489,9 +1623,9 @@ static int acct_stack_growth(struct vm_a return -ENOMEM; /* Ok, everything looks good - let it rip */ - mm->total_vm += grow; + vx_vmpages_add(mm, grow); if (vma->vm_flags & VM_LOCKED) - mm->locked_vm += grow; + vx_vmlocked_add(mm, grow); vm_stat_account(mm, vma->vm_flags, vma->vm_file, grow); return 0; } @@ -1594,13 +1728,49 @@ int expand_stack(struct vm_area_struct * if (address < vma->vm_start) { unsigned long size, grow; +#ifdef CONFIG_PAX_SEGMEXEC + struct vm_area_struct *vma_m = NULL; + unsigned long address_m = 0UL; + + if (vma->vm_flags & VM_MIRROR) { + address_m = vma->vm_start + vma->vm_mirror; + vma_m = find_vma(vma->vm_mm, address_m); + if (!vma_m || vma_m->vm_start != address_m || + !(vma_m->vm_flags & VM_MIRROR) || + vma->vm_end - vma->vm_start != + vma_m->vm_end - vma_m->vm_start || + vma->anon_vma != vma_m->anon_vma) { + printk(KERN_ERR "PAX: VMMIRROR: expand bug, %08lx, %08lx, %08lx, %08lx, %08lx\n", + address, vma->vm_start, vma_m->vm_start, vma->vm_end, vma_m->vm_end); + anon_vma_unlock(vma); + return -EFAULT; + } + address_m = address + vma->vm_mirror; + } +#endif + size = vma->vm_end - address; grow = (vma->vm_start - address) >> PAGE_SHIFT; +#ifdef CONFIG_PAX_SEGMEXEC + if (vma_m) + error = acct_stack_growth(vma, size, 2*grow); + else +#endif + error = acct_stack_growth(vma, size, grow); if (!error) { vma->vm_start = address; vma->vm_pgoff -= grow; + track_exec_limit(vma->vm_mm, vma->vm_start, vma->vm_end, vma->vm_flags); + +#ifdef CONFIG_PAX_SEGMEXEC + if (vma_m) { + vma_m->vm_start = address_m; + vma_m->vm_pgoff -= grow; + } +#endif + } } anon_vma_unlock(vma); @@ -1644,9 +1814,9 @@ static void remove_vma_list(struct mm_st do { long nrpages = vma_pages(vma); - mm->total_vm -= nrpages; + vx_vmpages_sub(mm, nrpages); if (vma->vm_flags & VM_LOCKED) - mm->locked_vm -= nrpages; + vx_vmlocked_sub(mm, nrpages); vm_stat_account(mm, vma->vm_flags, vma->vm_file, -nrpages); vma = remove_vma(vma); } while (vma); @@ -1762,7 +1932,24 @@ int split_vma(struct mm_struct * mm, str * work. This now handles partial unmappings. * Jeremy Fitzhardinge */ +#ifdef CONFIG_PAX_SEGMEXEC +static int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len); + +int do_munmap(struct mm_struct *mm, unsigned long start, size_t len) +{ + if (mm->pax_flags & MF_PAX_SEGMEXEC) { + int ret = __do_munmap(mm, start + SEGMEXEC_TASK_SIZE, len); + if (ret) + return ret; + } + + return __do_munmap(mm, start, len); +} + +static int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len) +#else int do_munmap(struct mm_struct *mm, unsigned long start, size_t len) +#endif { unsigned long end; struct vm_area_struct *vma, *prev, *last; @@ -1816,6 +2003,8 @@ int do_munmap(struct mm_struct *mm, unsi /* Fix up all other VM information */ remove_vma_list(mm, vma); + track_exec_limit(mm, start, end, 0UL); + return 0; } @@ -1828,6 +2017,12 @@ asmlinkage long sys_munmap(unsigned long profile_munmap(addr); +#ifdef CONFIG_PAX_SEGMEXEC + if ((mm->pax_flags & MF_PAX_SEGMEXEC) && + (len > SEGMEXEC_TASK_SIZE || addr > SEGMEXEC_TASK_SIZE-len)) + return -EINVAL; +#endif + down_write(&mm->mmap_sem); ret = do_munmap(mm, addr, len); up_write(&mm->mmap_sem); @@ -1849,11 +2044,35 @@ static inline void verify_mm_writelocked * anonymous maps. eventually we may be able to do some * brk-specific accounting here. */ +#ifdef CONFIG_PAX_SEGMEXEC +static unsigned long __do_brk(unsigned long addr, unsigned long len); + unsigned long do_brk(unsigned long addr, unsigned long len) { + unsigned long ret; + + ret = __do_brk(addr, len); + if (ret == addr && (current->mm->pax_flags & (MF_PAX_SEGMEXEC | MF_PAX_MPROTECT)) == MF_PAX_SEGMEXEC) { + unsigned long ret_m; + + ret_m = __do_mmap_pgoff(NULL, addr + SEGMEXEC_TASK_SIZE, 0UL, PROT_NONE, MAP_PRIVATE | MAP_FIXED | MAP_MIRROR, addr); + if (ret_m > TASK_SIZE) { + do_munmap(current->mm, addr, len); + ret = ret_m; + } + } + + return ret; +} + +static unsigned long __do_brk(unsigned long addr, unsigned long len) +#else +unsigned long do_brk(unsigned long addr, unsigned long len) +#endif +{ struct mm_struct * mm = current->mm; struct vm_area_struct * vma, * prev; - unsigned long flags; + unsigned long flags, task_size = TASK_SIZE; struct rb_node ** rb_link, * rb_parent; pgoff_t pgoff = addr >> PAGE_SHIFT; int error; @@ -1862,11 +2081,28 @@ unsigned long do_brk(unsigned long addr, if (!len) return addr; - if ((addr + len) > TASK_SIZE || (addr + len) < addr) +#ifdef CONFIG_PAX_SEGMEXEC + if (mm->pax_flags & MF_PAX_SEGMEXEC) + task_size = SEGMEXEC_TASK_SIZE; +#endif + + if ((addr + len) > task_size || (addr + len) < addr) return -EINVAL; flags = VM_DATA_DEFAULT_FLAGS | VM_ACCOUNT | mm->def_flags; +#if defined(CONFIG_PAX_PAGEEXEC) || defined(CONFIG_PAX_SEGMEXEC) + if (mm->pax_flags & (MF_PAX_PAGEEXEC | MF_PAX_SEGMEXEC)) { + flags &= ~VM_EXEC; + +#ifdef CONFIG_PAX_MPROTECT + if (mm->pax_flags & MF_PAX_MPROTECT) + flags &= ~VM_MAYEXEC; +#endif + + } +#endif + error = arch_mmap_check(addr, len, flags); if (error) return error; @@ -1880,8 +2116,11 @@ unsigned long do_brk(unsigned long addr, locked += mm->locked_vm; lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur; lock_limit >>= PAGE_SHIFT; + gr_learn_resource(current, RLIMIT_MEMLOCK, locked << PAGE_SHIFT, 1); if (locked > lock_limit && !capable(CAP_IPC_LOCK)) return -EAGAIN; + if (!vx_vmlocked_avail(mm, len >> PAGE_SHIFT)) + return -ENOMEM; } /* @@ -1893,12 +2132,12 @@ unsigned long do_brk(unsigned long addr, /* * Clear old maps. this also does some error checking for us */ - munmap_back: vma = find_vma_prepare(mm, addr, &prev, &rb_link, &rb_parent); if (vma && vma->vm_start < addr + len) { if (do_munmap(mm, addr, len)) return -ENOMEM; - goto munmap_back; + vma = find_vma_prepare(mm, addr, &prev, &rb_link, &rb_parent); + BUG_ON(vma && vma->vm_start < addr + len); } /* Check against address space limits *after* clearing old maps... */ @@ -1908,7 +2147,8 @@ unsigned long do_brk(unsigned long addr, if (mm->map_count > sysctl_max_map_count) return -ENOMEM; - if (security_vm_enough_memory(len >> PAGE_SHIFT)) + if (security_vm_enough_memory(len >> PAGE_SHIFT) || + !vx_vmpages_avail(mm, len >> PAGE_SHIFT)) return -ENOMEM; /* Can we just expand an old private anonymous mapping? */ @@ -1930,14 +2170,22 @@ unsigned long do_brk(unsigned long addr, vma->vm_end = addr + len; vma->vm_pgoff = pgoff; vma->vm_flags = flags; - vma->vm_page_prot = protection_map[flags & 0x0f]; + +#if defined(CONFIG_PAX_PAGEEXEC) && defined(CONFIG_X86_32) + if (!(mm->pax_flags & MF_PAX_PAGEEXEC)) + vma->vm_page_prot = protection_map[(flags | VM_EXEC) & (VM_READ|VM_WRITE|VM_EXEC|VM_SHARED)]; + else +#endif + + vma->vm_page_prot = protection_map[flags & (VM_READ|VM_WRITE|VM_EXEC|VM_SHARED)]; vma_link(mm, vma, prev, rb_link, rb_parent); out: - mm->total_vm += len >> PAGE_SHIFT; + vx_vmpages_add(mm, len >> PAGE_SHIFT); if (flags & VM_LOCKED) { - mm->locked_vm += len >> PAGE_SHIFT; + vx_vmlocked_add(mm, len >> PAGE_SHIFT); make_pages_present(addr, addr + len); } + track_exec_limit(mm, addr, addr + len, flags); return addr; } @@ -1961,6 +2209,11 @@ void exit_mmap(struct mm_struct *mm) free_pgtables(&tlb, vma, FIRST_USER_ADDRESS, 0); tlb_finish_mmu(tlb, 0, end); + set_mm_counter(mm, file_rss, 0); + set_mm_counter(mm, anon_rss, 0); + vx_vmpages_sub(mm, mm->total_vm); + vx_vmlocked_sub(mm, mm->locked_vm); + /* * Walk the list again, actually closing and freeing it, * with preemption enabled, without holding any MM locks. @@ -2000,7 +2253,8 @@ int insert_vm_struct(struct mm_struct * if (__vma && __vma->vm_start < vma->vm_end) return -ENOMEM; if ((vma->vm_flags & VM_ACCOUNT) && - security_vm_enough_memory(vma_pages(vma))) + (security_vm_enough_memory(vma_pages(vma)) || + !vx_vmpages_avail(mm, vma_pages(vma)))) return -ENOMEM; vma_link(mm, vma, prev, rb_link, rb_parent); return 0; @@ -2070,8 +2324,10 @@ int may_expand_vm(struct mm_struct *mm, unsigned long lim; lim = current->signal->rlim[RLIMIT_AS].rlim_cur >> PAGE_SHIFT; - + gr_learn_resource(current, RLIMIT_AS, (cur + npages) << PAGE_SHIFT, 1); if (cur + npages > lim) return 0; + if (!vx_vmpages_avail(mm, npages)) + return 0; return 1; } diff -NurpP --minimal linux-2.6.17.13/mm/mprotect.c linux-2.6.17.13-g-v/mm/mprotect.c --- linux-2.6.17.13/mm/mprotect.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/mprotect.c 2006-09-20 15:00:12.000000000 +0200 @@ -19,11 +19,18 @@ #include #include #include +#include + +#ifdef CONFIG_PAX_MPROTECT +#include +#include +#endif #include #include #include #include +#include static void change_pte_range(struct mm_struct *mm, pmd_t *pmd, unsigned long addr, unsigned long end, pgprot_t newprot) @@ -98,6 +105,94 @@ static void change_protection(struct vm_ flush_tlb_range(vma, start, end); } +#ifdef CONFIG_ARCH_TRACK_EXEC_LIMIT +/* called while holding the mmap semaphor for writing */ +static inline void establish_user_cs_limit(struct mm_struct *mm, unsigned long start, unsigned long end) +{ + struct vm_area_struct *vma = find_vma(mm, start); + + for (; vma && vma->vm_start < end; vma = vma->vm_next) + change_protection(vma, vma->vm_start, vma->vm_end, vma->vm_page_prot); + +} + +void track_exec_limit(struct mm_struct *mm, unsigned long start, unsigned long end, unsigned long prot) +{ + unsigned long oldlimit, newlimit = 0UL; + + if (!(mm->pax_flags & MF_PAX_PAGEEXEC)) + return; + + spin_lock(&mm->page_table_lock); + oldlimit = mm->context.user_cs_limit; + if ((prot & VM_EXEC) && oldlimit < end) + /* USER_CS limit moved up */ + newlimit = end; + else if (!(prot & VM_EXEC) && start < oldlimit && oldlimit <= end) + /* USER_CS limit moved down */ + newlimit = start; + + if (newlimit) { + mm->context.user_cs_limit = newlimit; + +#ifdef CONFIG_SMP + wmb(); + cpus_clear(mm->context.cpu_user_cs_mask); + cpu_set(smp_processor_id(), mm->context.cpu_user_cs_mask); +#endif + + set_user_cs(mm, smp_processor_id()); + } + spin_unlock(&mm->page_table_lock); + if (newlimit == end) + establish_user_cs_limit(mm, oldlimit, end); +} +#endif + +#ifdef CONFIG_PAX_SEGMEXEC +static int __mprotect_fixup(struct vm_area_struct *vma, struct vm_area_struct **pprev, + unsigned long start, unsigned long end, unsigned int newflags); + +static int mprotect_fixup(struct vm_area_struct *vma, struct vm_area_struct **pprev, + unsigned long start, unsigned long end, unsigned int newflags) +{ + if (vma->vm_flags & VM_MIRROR) { + struct vm_area_struct * vma_m, * prev_m; + unsigned long start_m, end_m; + int error; + + start_m = vma->vm_start + vma->vm_mirror; + vma_m = find_vma_prev(vma->vm_mm, start_m, &prev_m); + if (vma_m && vma_m->vm_start == start_m && (vma_m->vm_flags & VM_MIRROR)) { + start_m = start + vma->vm_mirror; + end_m = end + vma->vm_mirror; + + if (vma_m->vm_start >= SEGMEXEC_TASK_SIZE && !(newflags & VM_EXEC)) + error = __mprotect_fixup(vma_m, &prev_m, start_m, end_m, vma_m->vm_flags & ~(VM_READ | VM_WRITE | VM_EXEC)); + else + error = __mprotect_fixup(vma_m, &prev_m, start_m, end_m, newflags); + if (error) + return error; + } else { + printk("PAX: VMMIRROR: mprotect bug in %s, %08lx\n", current->comm, vma->vm_start); + return -ENOMEM; + } + } + + return __mprotect_fixup(vma, pprev, start, end, newflags); +} + +static int __mprotect_fixup(struct vm_area_struct *vma, struct vm_area_struct **pprev, + unsigned long start, unsigned long end, unsigned int newflags) +{ + struct mm_struct * mm = vma->vm_mm; + unsigned long oldflags = vma->vm_flags; + long nrpages = (end - start) >> PAGE_SHIFT; + unsigned long charged = 0; + pgprot_t newprot; + pgoff_t pgoff; + int error; +#else static int mprotect_fixup(struct vm_area_struct *vma, struct vm_area_struct **pprev, unsigned long start, unsigned long end, unsigned long newflags) @@ -114,6 +209,7 @@ mprotect_fixup(struct vm_area_struct *vm *pprev = vma; return 0; } +#endif /* * If we make a private mapping writable we increase our commit; @@ -132,7 +228,13 @@ mprotect_fixup(struct vm_area_struct *vm } } - newprot = protection_map[newflags & 0xf]; +#if defined(CONFIG_PAX_PAGEEXEC) && defined(CONFIG_X86_32) + if (!(mm->pax_flags & MF_PAX_PAGEEXEC) && (newflags & (VM_READ|VM_WRITE))) + newprot = protection_map[(newflags | VM_EXEC) & (VM_READ|VM_WRITE|VM_EXEC|VM_SHARED)]; + else +#endif + + newprot = protection_map[newflags & (VM_READ|VM_WRITE|VM_EXEC|VM_SHARED)]; /* * First try to merge with previous and/or next vma. @@ -179,6 +281,69 @@ fail: return error; } +#ifdef CONFIG_PAX_MPROTECT +/* PaX: non-PIC ELF libraries need relocations on their executable segments + * therefore we'll grant them VM_MAYWRITE once during their life. + * + * The checks favour ld-linux.so behaviour which operates on a per ELF segment + * basis because we want to allow the common case and not the special ones. + */ +static inline void pax_handle_maywrite(struct vm_area_struct * vma, unsigned long start) +{ + struct elfhdr elf_h; + struct elf_phdr elf_p, p_dyn; + elf_dyn dyn; + unsigned long i, j = 65536UL / sizeof(struct elf_phdr); + +#ifndef CONFIG_PAX_NOELFRELOCS + if ((vma->vm_start != start) || + !vma->vm_file || + !(vma->vm_flags & VM_MAYEXEC) || + (vma->vm_flags & VM_MAYNOTWRITE)) +#endif + + return; + + if (sizeof(elf_h) != kernel_read(vma->vm_file, 0UL, (char*)&elf_h, sizeof(elf_h)) || + memcmp(elf_h.e_ident, ELFMAG, SELFMAG) || + +#ifdef CONFIG_PAX_ETEXECRELOCS + (elf_h.e_type != ET_DYN && elf_h.e_type != ET_EXEC) || +#else + elf_h.e_type != ET_DYN || +#endif + + !elf_check_arch(&elf_h) || + elf_h.e_phentsize != sizeof(struct elf_phdr) || + elf_h.e_phnum > j) + return; + + for (i = 0UL; i < elf_h.e_phnum; i++) { + if (sizeof(elf_p) != kernel_read(vma->vm_file, elf_h.e_phoff + i*sizeof(elf_p), (char*)&elf_p, sizeof(elf_p))) + return; + if (elf_p.p_type == PT_DYNAMIC) { + p_dyn = elf_p; + j = i; + } + } + if (elf_h.e_phnum <= j) + return; + + i = 0UL; + do { + if (sizeof(dyn) != kernel_read(vma->vm_file, p_dyn.p_offset + i*sizeof(dyn), (char*)&dyn, sizeof(dyn))) + return; + if (dyn.d_tag == DT_TEXTREL || (dyn.d_tag == DT_FLAGS && (dyn.d_un.d_val & DF_TEXTREL))) { + vma->vm_flags |= VM_MAYWRITE | VM_MAYNOTWRITE; + gr_log_textrel(vma); + return; + } + i++; + } while (dyn.d_tag != DT_NULL); + return; +} +#endif + asmlinkage long sys_mprotect(unsigned long start, size_t len, unsigned long prot) { @@ -198,6 +363,17 @@ sys_mprotect(unsigned long start, size_t end = start + len; if (end <= start) return -ENOMEM; + +#ifdef CONFIG_PAX_SEGMEXEC + if (current->mm->pax_flags & MF_PAX_SEGMEXEC) { + if (end > SEGMEXEC_TASK_SIZE) + return -EINVAL; + } else +#endif + + if (end > TASK_SIZE) + return -EINVAL; + if (prot & ~(PROT_READ | PROT_WRITE | PROT_EXEC | PROT_SEM)) return -EINVAL; @@ -205,7 +381,7 @@ sys_mprotect(unsigned long start, size_t /* * Does the application expect PROT_READ to imply PROT_EXEC: */ - if (unlikely((prot & PROT_READ) && + if (unlikely((prot & (PROT_READ | PROT_WRITE)) && (current->personality & READ_IMPLIES_EXEC))) prot |= PROT_EXEC; @@ -238,6 +414,16 @@ sys_mprotect(unsigned long start, size_t if (start > vma->vm_start) prev = vma; +#ifdef CONFIG_PAX_MPROTECT + if ((vma->vm_mm->pax_flags & MF_PAX_MPROTECT) && (prot & PROT_WRITE)) + pax_handle_maywrite(vma, start); +#endif + + if (!gr_acl_handle_mprotect(vma->vm_file, prot)) { + error = -EACCES; + goto out; + } + for (nstart = start ; ; ) { unsigned long newflags; @@ -251,6 +437,12 @@ sys_mprotect(unsigned long start, size_t goto out; } +#ifdef CONFIG_PAX_MPROTECT + /* PaX: disallow write access after relocs are done, hopefully noone else needs it... */ + if ((vma->vm_mm->pax_flags & MF_PAX_MPROTECT) && !(prot & PROT_WRITE) && (vma->vm_flags & VM_MAYNOTWRITE)) + newflags &= ~VM_MAYWRITE; +#endif + error = security_file_mprotect(vma, reqprot, prot); if (error) goto out; @@ -274,6 +466,9 @@ sys_mprotect(unsigned long start, size_t goto out; } } + + track_exec_limit(current->mm, start, end, vm_flags); + out: up_write(&current->mm->mmap_sem); return error; diff -NurpP --minimal linux-2.6.17.13/mm/mremap.c linux-2.6.17.13-g-v/mm/mremap.c --- linux-2.6.17.13/mm/mremap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/mremap.c 2006-09-20 15:00:12.000000000 +0200 @@ -18,6 +18,7 @@ #include #include #include +#include #include #include @@ -106,6 +107,12 @@ static void move_ptes(struct vm_area_str pte = ptep_clear_flush(vma, old_addr, old_pte); /* ZERO_PAGE can be dependant on virtual addr */ pte = move_pte(pte, new_vma->vm_page_prot, old_addr, new_addr); + +#ifdef CONFIG_ARCH_TRACK_EXEC_LIMIT + if ((mm->pax_flags & MF_PAX_PAGEEXEC) && !(vma->vm_flags & VM_EXEC)) + pte_exprotect(pte); +#endif + set_pte_at(mm, new_addr, new_pte, pte); } @@ -211,7 +218,7 @@ static unsigned long move_vma(struct vm_ * If this were a serious issue, we'd add a flag to do_munmap(). */ hiwater_vm = mm->hiwater_vm; - mm->total_vm += new_len >> PAGE_SHIFT; + vx_vmpages_add(mm, new_len >> PAGE_SHIFT); vm_stat_account(mm, vma->vm_flags, vma->vm_file, new_len>>PAGE_SHIFT); if (do_munmap(mm, old_addr, old_len) < 0) { @@ -229,7 +236,7 @@ static unsigned long move_vma(struct vm_ } if (vm_flags & VM_LOCKED) { - mm->locked_vm += new_len >> PAGE_SHIFT; + vx_vmlocked_add(mm, new_len >> PAGE_SHIFT); if (new_len > old_len) make_pages_present(new_addr + old_len, new_addr + new_len); @@ -253,6 +260,7 @@ unsigned long do_mremap(unsigned long ad struct vm_area_struct *vma; unsigned long ret = -EINVAL; unsigned long charged = 0; + unsigned long task_size = TASK_SIZE; if (flags & ~(MREMAP_FIXED | MREMAP_MAYMOVE)) goto out; @@ -271,6 +279,15 @@ unsigned long do_mremap(unsigned long ad if (!new_len) goto out; +#ifdef CONFIG_PAX_SEGMEXEC + if (current->mm->pax_flags & MF_PAX_SEGMEXEC) + task_size = SEGMEXEC_TASK_SIZE; +#endif + + if (new_len > task_size || addr > task_size-new_len || + old_len > task_size || addr > task_size-old_len) + goto out; + /* new_addr is only valid if MREMAP_FIXED is specified */ if (flags & MREMAP_FIXED) { if (new_addr & ~PAGE_MASK) @@ -278,16 +295,13 @@ unsigned long do_mremap(unsigned long ad if (!(flags & MREMAP_MAYMOVE)) goto out; - if (new_len > TASK_SIZE || new_addr > TASK_SIZE - new_len) + if (new_addr > task_size - new_len) goto out; /* Check if the location we're moving into overlaps the * old location at all, and fail if it does. */ - if ((new_addr <= addr) && (new_addr+new_len) > addr) - goto out; - - if ((addr <= new_addr) && (addr+old_len) > new_addr) + if (addr + old_len > new_addr && new_addr + new_len > addr) goto out; ret = do_munmap(mm, new_addr, new_len); @@ -321,6 +335,14 @@ unsigned long do_mremap(unsigned long ad ret = -EINVAL; goto out; } + +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) { + ret = -EINVAL; + goto out; + } +#endif + /* We can't remap across vm area boundaries */ if (old_len > vma->vm_end - addr) goto out; @@ -336,6 +358,9 @@ unsigned long do_mremap(unsigned long ad ret = -EAGAIN; if (locked > lock_limit && !capable(CAP_IPC_LOCK)) goto out; + if (!vx_vmlocked_avail(current->mm, + (new_len - old_len) >> PAGE_SHIFT)) + goto out; } if (!may_expand_vm(mm, (new_len - old_len) >> PAGE_SHIFT)) { ret = -ENOMEM; @@ -354,7 +379,7 @@ unsigned long do_mremap(unsigned long ad if (old_len == vma->vm_end - addr && !((flags & MREMAP_FIXED) && (addr != new_addr)) && (old_len != new_len || !(flags & MREMAP_MAYMOVE))) { - unsigned long max_addr = TASK_SIZE; + unsigned long max_addr = task_size; if (vma->vm_next) max_addr = vma->vm_next->vm_start; /* can we just expand the current mapping? */ @@ -364,14 +389,15 @@ unsigned long do_mremap(unsigned long ad vma_adjust(vma, vma->vm_start, addr + new_len, vma->vm_pgoff, NULL); - mm->total_vm += pages; + vx_vmpages_add(mm, pages); vm_stat_account(mm, vma->vm_flags, vma->vm_file, pages); if (vma->vm_flags & VM_LOCKED) { - mm->locked_vm += pages; + vx_vmlocked_add(mm, pages); make_pages_present(addr + old_len, addr + new_len); } ret = addr; + track_exec_limit(vma->vm_mm, vma->vm_start, addr + new_len, vma->vm_flags); goto out; } } @@ -382,8 +408,8 @@ unsigned long do_mremap(unsigned long ad */ ret = -ENOMEM; if (flags & MREMAP_MAYMOVE) { + unsigned long map_flags = 0; if (!(flags & MREMAP_FIXED)) { - unsigned long map_flags = 0; if (vma->vm_flags & VM_MAYSHARE) map_flags |= MAP_SHARED; @@ -393,7 +419,12 @@ unsigned long do_mremap(unsigned long ad if (new_addr & ~PAGE_MASK) goto out; } + map_flags = vma->vm_flags; ret = move_vma(vma, addr, old_len, new_len, new_addr); + if (!(ret & ~PAGE_MASK)) { + track_exec_limit(current->mm, addr, addr + old_len, 0UL); + track_exec_limit(current->mm, new_addr, new_addr + new_len, map_flags); + } } out: if (ret & ~PAGE_MASK) diff -NurpP --minimal linux-2.6.17.13/mm/nommu.c linux-2.6.17.13-g-v/mm/nommu.c --- linux-2.6.17.13/mm/nommu.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/nommu.c 2006-09-20 15:00:12.000000000 +0200 @@ -820,7 +820,7 @@ unsigned long do_mmap_pgoff(struct file realalloc += kobjsize(vma); askedalloc += sizeof(*vma); - current->mm->total_vm += len >> PAGE_SHIFT; + vx_vmpages_add(current->mm, len >> PAGE_SHIFT); add_nommu_vma(vma); @@ -937,7 +937,7 @@ int do_munmap(struct mm_struct *mm, unsi kfree(vml); update_hiwater_vm(mm); - mm->total_vm -= len >> PAGE_SHIFT; + vx_vmpages_sub(mm, len >> PAGE_SHIFT); #ifdef DEBUG show_process_blocks(); @@ -956,7 +956,7 @@ void exit_mmap(struct mm_struct * mm) printk("Exit_mmap:\n"); #endif - mm->total_vm = 0; + vx_vmpages_sub(mm, mm->total_vm); while ((tmp = mm->context.vmlist)) { mm->context.vmlist = tmp->next; diff -NurpP --minimal linux-2.6.17.13/mm/oom_kill.c linux-2.6.17.13-g-v/mm/oom_kill.c --- linux-2.6.17.13/mm/oom_kill.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/oom_kill.c 2006-09-20 15:00:12.000000000 +0200 @@ -66,6 +66,8 @@ unsigned long badness(struct task_struct */ task_unlock(p); + /* FIXME: add vserver badness ;) */ + /* * Processes which fork a lot of child processes are likely * a good choice. We add half the vmsize of the children if they diff -NurpP --minimal linux-2.6.17.13/mm/page_alloc.c linux-2.6.17.13-g-v/mm/page_alloc.c --- linux-2.6.17.13/mm/page_alloc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/page_alloc.c 2006-09-20 15:00:12.000000000 +0200 @@ -37,6 +37,7 @@ #include #include #include +#include #include #include @@ -334,6 +335,9 @@ static inline int page_is_buddy(struct p static inline void __free_one_page(struct page *page, struct zone *zone, unsigned int order) { +#ifdef CONFIG_PAX_MEMORY_SANITIZE + unsigned long index; +#endif unsigned long page_idx; int order_size = 1 << order; @@ -345,6 +349,11 @@ static inline void __free_one_page(struc BUG_ON(page_idx & (order_size - 1)); BUG_ON(bad_range(zone, page)); +#ifdef CONFIG_PAX_MEMORY_SANITIZE + for (index = order_size; index; --index) + clear_highpage(page + index - 1); +#endif + zone->free_pages += order_size; while (order < MAX_ORDER-1) { unsigned long combined_idx; @@ -1375,6 +1384,8 @@ void si_meminfo(struct sysinfo *val) val->freehigh = 0; #endif val->mem_unit = PAGE_SIZE; + if (vx_flags(VXF_VIRT_MEM, 0)) + vx_vsi_meminfo(val); } EXPORT_SYMBOL(si_meminfo); @@ -1389,6 +1400,8 @@ void si_meminfo_node(struct sysinfo *val val->totalhigh = pgdat->node_zones[ZONE_HIGHMEM].present_pages; val->freehigh = pgdat->node_zones[ZONE_HIGHMEM].free_pages; val->mem_unit = PAGE_SIZE; + if (vx_flags(VXF_VIRT_MEM, 0)) + vx_vsi_meminfo(val); } #endif diff -NurpP --minimal linux-2.6.17.13/mm/rmap.c linux-2.6.17.13-g-v/mm/rmap.c --- linux-2.6.17.13/mm/rmap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/rmap.c 2006-09-20 15:00:12.000000000 +0200 @@ -53,6 +53,7 @@ #include #include #include +#include #include @@ -106,6 +107,19 @@ int anon_vma_prepare(struct vm_area_stru list_add(&vma->anon_vma_node, &anon_vma->head); allocated = NULL; } + +#ifdef CONFIG_PAX_SEGMEXEC + if (vma->vm_flags & VM_MIRROR) { + struct vm_area_struct *vma_m; + + vma_m = find_vma(vma->vm_mm, vma->vm_start + vma->vm_mirror); + BUG_ON(!vma_m || vma_m->vm_start != vma->vm_start + vma->vm_mirror); + BUG_ON(vma_m->anon_vma || vma->vm_pgoff != vma_m->vm_pgoff); + vma_m->anon_vma = anon_vma; + __anon_vma_link(vma_m); + } +#endif + spin_unlock(&mm->page_table_lock); if (locked) diff -NurpP --minimal linux-2.6.17.13/mm/shmem.c linux-2.6.17.13-g-v/mm/shmem.c --- linux-2.6.17.13/mm/shmem.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/shmem.c 2006-09-20 15:00:12.000000000 +0200 @@ -53,7 +53,6 @@ #include /* This magic number is used in glibc for posix shared memory */ -#define TMPFS_MAGIC 0x01021994 #define ENTRIES_PER_PAGE (PAGE_CACHE_SIZE/sizeof(unsigned long)) #define ENTRIES_PER_PAGEPAGE (ENTRIES_PER_PAGE*ENTRIES_PER_PAGE) @@ -1658,7 +1657,7 @@ static int shmem_statfs(struct super_blo { struct shmem_sb_info *sbinfo = SHMEM_SB(sb); - buf->f_type = TMPFS_MAGIC; + buf->f_type = TMPFS_SUPER_MAGIC; buf->f_bsize = PAGE_CACHE_SIZE; buf->f_namelen = NAME_MAX; spin_lock(&sbinfo->stat_lock); @@ -2101,7 +2100,7 @@ static int shmem_fill_super(struct super sb->s_maxbytes = SHMEM_MAX_BYTES; sb->s_blocksize = PAGE_CACHE_SIZE; sb->s_blocksize_bits = PAGE_CACHE_SHIFT; - sb->s_magic = TMPFS_MAGIC; + sb->s_magic = TMPFS_SUPER_MAGIC; sb->s_op = &shmem_ops; sb->s_time_gran = 1; @@ -2245,7 +2244,7 @@ static struct file_system_type tmpfs_fs_ .get_sb = shmem_get_sb, .kill_sb = kill_litter_super, }; -static struct vfsmount *shm_mnt; +struct vfsmount *shm_mnt; static int __init init_tmpfs(void) { diff -NurpP --minimal linux-2.6.17.13/mm/slab.c linux-2.6.17.13-g-v/mm/slab.c --- linux-2.6.17.13/mm/slab.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/slab.c 2006-09-20 15:00:12.000000000 +0200 @@ -1532,6 +1532,11 @@ static void store_stackinfo(struct kmem_ while (!kstack_end(sptr)) { svalue = *sptr++; + +#ifdef CONFIG_PAX_KERNEXEC + svalue += __KERNEL_TEXT_OFFSET; +#endif + if (kernel_text_address(svalue)) { *addr++ = svalue; size -= sizeof(unsigned long); diff -NurpP --minimal linux-2.6.17.13/mm/swapfile.c linux-2.6.17.13-g-v/mm/swapfile.c --- linux-2.6.17.13/mm/swapfile.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/swapfile.c 2006-09-20 15:00:12.000000000 +0200 @@ -32,6 +32,7 @@ #include #include #include +#include DEFINE_SPINLOCK(swap_lock); unsigned int nr_swapfiles; @@ -1696,6 +1697,8 @@ void si_swapinfo(struct sysinfo *val) val->freeswap = nr_swap_pages + nr_to_be_unused; val->totalswap = total_swap_pages + nr_to_be_unused; spin_unlock(&swap_lock); + if (vx_flags(VXF_VIRT_MEM, 0)) + vx_vsi_swapinfo(val); } /* diff -NurpP --minimal linux-2.6.17.13/mm/tiny-shmem.c linux-2.6.17.13-g-v/mm/tiny-shmem.c --- linux-2.6.17.13/mm/tiny-shmem.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/tiny-shmem.c 2006-09-20 15:00:12.000000000 +0200 @@ -27,7 +27,7 @@ static struct file_system_type tmpfs_fs_ .kill_sb = kill_litter_super, }; -static struct vfsmount *shm_mnt; +struct vfsmount *shm_mnt; static int __init init_tmpfs(void) { diff -NurpP --minimal linux-2.6.17.13/mm/vmalloc.c linux-2.6.17.13-g-v/mm/vmalloc.c --- linux-2.6.17.13/mm/vmalloc.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/vmalloc.c 2006-09-20 15:00:12.000000000 +0200 @@ -193,6 +193,8 @@ struct vm_struct *__get_vm_area_node(uns write_lock(&vmlist_lock); for (p = &vmlist; (tmp = *p) != NULL ;p = &tmp->next) { + if (addr > end - size) + goto out; if ((unsigned long)tmp->addr < addr) { if((unsigned long)tmp->addr + tmp->size >= addr) addr = ALIGN(tmp->size + @@ -204,8 +206,6 @@ struct vm_struct *__get_vm_area_node(uns if (size + addr <= (unsigned long)tmp->addr) goto found; addr = ALIGN(tmp->size + (unsigned long)tmp->addr, align); - if (addr > end - size) - goto out; } found: diff -NurpP --minimal linux-2.6.17.13/mm/vmscan.c linux-2.6.17.13-g-v/mm/vmscan.c --- linux-2.6.17.13/mm/vmscan.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/mm/vmscan.c 2006-09-20 15:00:12.000000000 +0200 @@ -1357,7 +1357,7 @@ static int __init kswapd_init(void) pid = kernel_thread(kswapd, pgdat, CLONE_KERNEL); BUG_ON(pid < 0); read_lock(&tasklist_lock); - pgdat->kswapd = find_task_by_pid(pid); + pgdat->kswapd = find_task_by_real_pid(pid); read_unlock(&tasklist_lock); } total_memory = nr_free_pagecache_pages(); diff -NurpP --minimal linux-2.6.17.13/net/core/dev.c linux-2.6.17.13-g-v/net/core/dev.c --- linux-2.6.17.13/net/core/dev.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/core/dev.c 2006-09-20 15:00:12.000000000 +0200 @@ -115,6 +115,7 @@ #include #include #include +#include /* * The list of packet types we will receive (as opposed to discard) @@ -1945,6 +1946,9 @@ static int dev_ifconf(char __user *arg) total = 0; for (dev = dev_base; dev; dev = dev->next) { + if (vx_flags(VXF_HIDE_NETIF, 0) && + !dev_in_nx_info(dev, current->nx_info)) + continue; for (i = 0; i < NPROTO; i++) { if (gifconf_list[i]) { int done; @@ -2005,6 +2009,10 @@ void dev_seq_stop(struct seq_file *seq, static void dev_seq_printf_stats(struct seq_file *seq, struct net_device *dev) { + struct nx_info *nxi = current->nx_info; + + if (vx_flags(VXF_HIDE_NETIF, 0) && !dev_in_nx_info(dev, nxi)) + return; if (dev->get_stats) { struct net_device_stats *stats = dev->get_stats(dev); diff -NurpP --minimal linux-2.6.17.13/net/core/rtnetlink.c linux-2.6.17.13-g-v/net/core/rtnetlink.c --- linux-2.6.17.13/net/core/rtnetlink.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/core/rtnetlink.c 2006-09-20 15:00:12.000000000 +0200 @@ -323,6 +323,9 @@ static int rtnetlink_dump_ifinfo(struct for (dev=dev_base, idx=0; dev; dev = dev->next, idx++) { if (idx < s_idx) continue; + if (vx_info_flags(skb->sk->sk_vx_info, VXF_HIDE_NETIF, 0) && + !dev_in_nx_info(dev, skb->sk->sk_nx_info)) + continue; if (rtnetlink_fill_ifinfo(skb, dev, RTM_NEWLINK, NETLINK_CB(cb->skb).pid, cb->nlh->nlmsg_seq, 0, @@ -613,6 +616,9 @@ void rtmsg_ifinfo(int type, struct net_d sizeof(struct rtnl_link_ifmap) + sizeof(struct rtnl_link_stats) + 128); + if (vx_flags(VXF_HIDE_NETIF, 0) && + !dev_in_nx_info(dev, current->nx_info)) + return; skb = alloc_skb(size, GFP_KERNEL); if (!skb) return; diff -NurpP --minimal linux-2.6.17.13/net/core/sock.c linux-2.6.17.13-g-v/net/core/sock.c --- linux-2.6.17.13/net/core/sock.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/core/sock.c 2006-09-20 15:00:12.000000000 +0200 @@ -125,6 +125,9 @@ #include #include +#include +#include +#include #ifdef CONFIG_INET #include @@ -768,6 +771,8 @@ struct sock *sk_alloc(int family, gfp_t sk->sk_prot = sk->sk_prot_creator = prot; sock_lock_init(sk); } + sock_vx_init(sk); + sock_nx_init(sk); if (security_sk_alloc(sk, family, priority)) goto out_free; @@ -806,6 +811,11 @@ void sk_free(struct sock *sk) __FUNCTION__, atomic_read(&sk->sk_omem_alloc)); security_sk_free(sk); + vx_sock_dec(sk); + clr_vx_info(&sk->sk_vx_info); + sk->sk_xid = -1; + clr_nx_info(&sk->sk_nx_info); + sk->sk_nid = -1; if (sk->sk_prot_creator->slab != NULL) kmem_cache_free(sk->sk_prot_creator->slab, sk); else @@ -823,6 +833,8 @@ struct sock *sk_clone(const struct sock memcpy(newsk, sk, sk->sk_prot->obj_size); /* SANITY */ + sock_vx_init(newsk); + sock_nx_init(newsk); sk_node_init(&newsk->sk_node); sock_lock_init(newsk); bh_lock_sock(newsk); @@ -863,6 +875,12 @@ struct sock *sk_clone(const struct sock newsk->sk_priority = 0; atomic_set(&newsk->sk_refcnt, 2); + set_vx_info(&newsk->sk_vx_info, sk->sk_vx_info); + newsk->sk_xid = sk->sk_xid; + vx_sock_inc(newsk); + set_nx_info(&newsk->sk_nx_info, sk->sk_nx_info); + newsk->sk_nid = sk->sk_nid; + /* * Increment the counter in the same struct proto as the master * sock (sk_refcnt_debug_inc uses newsk->sk_prot->socks, that @@ -1427,6 +1445,11 @@ void sock_init_data(struct socket *sock, sk->sk_stamp.tv_sec = -1L; sk->sk_stamp.tv_usec = -1L; + set_vx_info(&sk->sk_vx_info, current->vx_info); + sk->sk_xid = vx_current_xid(); + vx_sock_inc(sk); + set_nx_info(&sk->sk_nx_info, current->nx_info); + sk->sk_nid = nx_current_nid(); atomic_set(&sk->sk_refcnt, 1); } diff -NurpP --minimal linux-2.6.17.13/net/ipv4/af_inet.c linux-2.6.17.13-g-v/net/ipv4/af_inet.c --- linux-2.6.17.13/net/ipv4/af_inet.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/af_inet.c 2006-09-20 15:00:12.000000000 +0200 @@ -114,6 +114,7 @@ #ifdef CONFIG_IP_MROUTE #include #endif +#include DEFINE_SNMP_STAT(struct linux_mib, net_statistics) __read_mostly; @@ -282,9 +283,11 @@ lookup_protocol: } err = -EPERM; + if ((protocol == IPPROTO_ICMP) && vx_ccaps(VXC_RAW_ICMP)) + goto override; if (answer->capability > 0 && !capable(answer->capability)) goto out_rcu_unlock; - +override: sock->ops = answer->ops; answer_prot = answer->prot; answer_no_check = answer->no_check; @@ -401,6 +404,10 @@ int inet_bind(struct socket *sock, struc unsigned short snum; int chk_addr_ret; int err; + __u32 s_addr; /* Address used for validation */ + __u32 s_addr1; /* Address used for socket */ + __u32 s_addr2; /* Broadcast address for the socket */ + struct nx_info *nxi = sk->sk_nx_info; /* If the socket has its own bind function then use it. (RAW) */ if (sk->sk_prot->bind) { @@ -411,7 +418,40 @@ int inet_bind(struct socket *sock, struc if (addr_len < sizeof(struct sockaddr_in)) goto out; - chk_addr_ret = inet_addr_type(addr->sin_addr.s_addr); + s_addr = addr->sin_addr.s_addr; + s_addr1 = s_addr; + s_addr2 = 0xffffffffl; + + vxdprintk(VXD_CBIT(net, 3), + "inet_bind(%p)* %p,%p;%lx %d.%d.%d.%d", + sk, sk->sk_nx_info, sk->sk_socket, + (sk->sk_socket?sk->sk_socket->flags:0), + VXD_QUAD(s_addr)); + if (nxi) { + __u32 v4_bcast = nxi->v4_bcast; + __u32 ipv4root = nxi->ipv4[0]; + int nbipv4 = nxi->nbipv4; + + if (s_addr == 0) { + /* bind to any for 1-n */ + s_addr = ipv4root; + s_addr1 = (nbipv4 > 1) ? 0 : s_addr; + s_addr2 = v4_bcast; + } else if (s_addr == IPI_LOOPBACK) { + /* rewrite localhost to ipv4root */ + s_addr = ipv4root; + s_addr1 = ipv4root; + } else if (s_addr != v4_bcast) { + /* normal address bind */ + if (!addr_in_nx_info(nxi, s_addr)) + return -EADDRNOTAVAIL; + } + } + chk_addr_ret = inet_addr_type(s_addr); + + vxdprintk(VXD_CBIT(net, 3), + "inet_bind(%p) %d.%d.%d.%d, %d.%d.%d.%d, %d.%d.%d.%d", + sk, VXD_QUAD(s_addr), VXD_QUAD(s_addr1), VXD_QUAD(s_addr2)); /* Not specified by any standard per-se, however it breaks too * many applications when removed. It is unfortunate since @@ -423,7 +463,7 @@ int inet_bind(struct socket *sock, struc err = -EADDRNOTAVAIL; if (!sysctl_ip_nonlocal_bind && !inet->freebind && - addr->sin_addr.s_addr != INADDR_ANY && + s_addr != INADDR_ANY && chk_addr_ret != RTN_LOCAL && chk_addr_ret != RTN_MULTICAST && chk_addr_ret != RTN_BROADCAST) @@ -448,7 +488,8 @@ int inet_bind(struct socket *sock, struc if (sk->sk_state != TCP_CLOSE || inet->num) goto out_release_sock; - inet->rcv_saddr = inet->saddr = addr->sin_addr.s_addr; + inet->rcv_saddr = inet->saddr = s_addr1; + inet->rcv_saddr2 = s_addr2; if (chk_addr_ret == RTN_MULTICAST || chk_addr_ret == RTN_BROADCAST) inet->saddr = 0; /* Use device */ diff -NurpP --minimal linux-2.6.17.13/net/ipv4/devinet.c linux-2.6.17.13-g-v/net/ipv4/devinet.c --- linux-2.6.17.13/net/ipv4/devinet.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/devinet.c 2006-09-20 15:00:12.000000000 +0200 @@ -609,6 +609,9 @@ int devinet_ioctl(unsigned int cmd, void *colon = ':'; if ((in_dev = __in_dev_get_rtnl(dev)) != NULL) { + struct nx_info *nxi = current->nx_info; + int hide_netif = vx_flags(VXF_HIDE_NETIF, 0); + if (tryaddrmatch) { /* Matthias Andree */ /* compare label and address (4.4BSD style) */ @@ -617,6 +620,8 @@ int devinet_ioctl(unsigned int cmd, void This is checked above. */ for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL; ifap = &ifa->ifa_next) { + if (hide_netif && !ifa_in_nx_info(ifa, nxi)) + continue; if (!strcmp(ifr.ifr_name, ifa->ifa_label) && sin_orig.sin_addr.s_addr == ifa->ifa_address) { @@ -629,9 +634,12 @@ int devinet_ioctl(unsigned int cmd, void comparing just the label */ if (!ifa) { for (ifap = &in_dev->ifa_list; (ifa = *ifap) != NULL; - ifap = &ifa->ifa_next) + ifap = &ifa->ifa_next) { + if (hide_netif && !ifa_in_nx_info(ifa, nxi)) + continue; if (!strcmp(ifr.ifr_name, ifa->ifa_label)) break; + } } } @@ -782,6 +790,9 @@ static int inet_gifconf(struct net_devic goto out; for (; ifa; ifa = ifa->ifa_next) { + if (vx_flags(VXF_HIDE_NETIF, 0) && + !ifa_in_nx_info(ifa, current->nx_info)) + continue; if (!buf) { done += sizeof(ifr); continue; @@ -1093,6 +1104,7 @@ static int inet_dump_ifaddr(struct sk_bu struct net_device *dev; struct in_device *in_dev; struct in_ifaddr *ifa; + struct sock *sk = skb->sk; int s_ip_idx, s_idx = cb->args[0]; s_ip_idx = ip_idx = cb->args[1]; @@ -1110,6 +1122,9 @@ static int inet_dump_ifaddr(struct sk_bu for (ifa = in_dev->ifa_list, ip_idx = 0; ifa; ifa = ifa->ifa_next, ip_idx++) { + if (sk && vx_info_flags(sk->sk_vx_info, VXF_HIDE_NETIF, 0) && + !ifa_in_nx_info(ifa, sk->sk_nx_info)) + continue; if (ip_idx < s_ip_idx) continue; if (inet_fill_ifaddr(skb, ifa, NETLINK_CB(cb->skb).pid, diff -NurpP --minimal linux-2.6.17.13/net/ipv4/fib_hash.c linux-2.6.17.13-g-v/net/ipv4/fib_hash.c --- linux-2.6.17.13/net/ipv4/fib_hash.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/fib_hash.c 2006-09-20 15:00:12.000000000 +0200 @@ -989,6 +989,8 @@ static unsigned fib_flag_trans(int type, return flags; } +extern int dev_in_nx_info(struct net_device *, struct nx_info *); + /* * This outputs /proc/net/route. * @@ -1019,7 +1021,8 @@ static int fib_seq_show(struct seq_file prefix = f->fn_key; mask = FZ_MASK(iter->zone); flags = fib_flag_trans(fa->fa_type, mask, fi); - if (fi) + if (fi && (!vx_flags(VXF_HIDE_NETIF, 0) || + dev_in_nx_info(fi->fib_dev, current->nx_info))) snprintf(bf, sizeof(bf), "%s\t%08X\t%08X\t%04X\t%d\t%u\t%d\t%08X\t%d\t%u\t%u", fi->fib_dev ? fi->fib_dev->name : "*", prefix, diff -NurpP --minimal linux-2.6.17.13/net/ipv4/inet_connection_sock.c linux-2.6.17.13-g-v/net/ipv4/inet_connection_sock.c --- linux-2.6.17.13/net/ipv4/inet_connection_sock.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/inet_connection_sock.c 2006-09-20 15:00:12.000000000 +0200 @@ -16,6 +16,7 @@ #include #include #include +#include #include #include @@ -40,7 +41,6 @@ int sysctl_local_port_range[2] = { 1024, int inet_csk_bind_conflict(const struct sock *sk, const struct inet_bind_bucket *tb) { - const u32 sk_rcv_saddr = inet_rcv_saddr(sk); struct sock *sk2; struct hlist_node *node; int reuse = sk->sk_reuse; @@ -53,9 +53,8 @@ int inet_csk_bind_conflict(const struct sk->sk_bound_dev_if == sk2->sk_bound_dev_if)) { if (!reuse || !sk2->sk_reuse || sk2->sk_state == TCP_LISTEN) { - const u32 sk2_rcv_saddr = inet_rcv_saddr(sk2); - if (!sk2_rcv_saddr || !sk_rcv_saddr || - sk2_rcv_saddr == sk_rcv_saddr) + if (nx_addr_conflict(sk->sk_nx_info, + inet_rcv_saddr(sk), sk2)) break; } } diff -NurpP --minimal linux-2.6.17.13/net/ipv4/inet_diag.c linux-2.6.17.13-g-v/net/ipv4/inet_diag.c --- linux-2.6.17.13/net/ipv4/inet_diag.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/inet_diag.c 2006-09-20 15:00:12.000000000 +0200 @@ -694,6 +694,8 @@ static int inet_diag_dump(struct sk_buff sk_for_each(sk, node, &hashinfo->listening_hash[i]) { struct inet_sock *inet = inet_sk(sk); + if (!vx_check(sk->sk_xid, VX_IDENT|VX_WATCH)) + continue; if (num < s_num) { num++; continue; @@ -754,6 +756,8 @@ skip_listen_ht: sk_for_each(sk, node, &head->chain) { struct inet_sock *inet = inet_sk(sk); + if (!vx_check(sk->sk_xid, VX_IDENT|VX_WATCH)) + continue; if (num < s_num) goto next_normal; if (!(r->idiag_states & (1 << sk->sk_state))) @@ -778,6 +782,8 @@ next_normal: inet_twsk_for_each(tw, node, &hashinfo->ehash[i + hashinfo->ehash_size].chain) { + if (!vx_check(tw->tw_xid, VX_IDENT|VX_WATCH)) + continue; if (num < s_num) goto next_dying; if (r->id.idiag_sport != tw->tw_sport && diff -NurpP --minimal linux-2.6.17.13/net/ipv4/inet_hashtables.c linux-2.6.17.13-g-v/net/ipv4/inet_hashtables.c --- linux-2.6.17.13/net/ipv4/inet_hashtables.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/inet_hashtables.c 2006-09-20 15:00:12.000000000 +0200 @@ -19,11 +19,14 @@ #include #include #include +#include #include #include #include +extern void gr_update_task_in_ip_table(struct task_struct *task, const struct inet_sock *inet); + /* * Allocate and initialize a new local port bind bucket. * The bindhash mutex for snum's hash chain must be held here. @@ -139,11 +142,10 @@ struct sock *__inet_lookup_listener(cons const __u32 rcv_saddr = inet->rcv_saddr; int score = sk->sk_family == PF_INET ? 1 : 0; - if (rcv_saddr) { - if (rcv_saddr != daddr) - continue; + if (inet_addr_match(sk->sk_nx_info, daddr, rcv_saddr)) score += 2; - } + else + continue; if (sk->sk_bound_dev_if) { if (sk->sk_bound_dev_if != dif) continue; @@ -310,6 +312,8 @@ ok: } spin_unlock(&head->lock); + gr_update_task_in_ip_table(current, inet_sk(sk)); + if (tw) { inet_twsk_deschedule(tw, death_row); inet_twsk_put(tw); diff -NurpP --minimal linux-2.6.17.13/net/ipv4/netfilter/Kconfig linux-2.6.17.13-g-v/net/ipv4/netfilter/Kconfig --- linux-2.6.17.13/net/ipv4/netfilter/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/netfilter/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -314,6 +314,21 @@ config IP_NF_MATCH_HASHLIMIT destination IP' or `500pps from any given source IP' with a single IPtables rule. +config IP_NF_MATCH_STEALTH + tristate "stealth match support" + depends on IP_NF_IPTABLES + help + Enabling this option will drop all syn packets coming to unserved tcp + ports as well as all packets coming to unserved udp ports. If you + are using your system to route any type of packets (ie. via NAT) + you should put this module at the end of your ruleset, since it will + drop packets that aren't going to ports that are listening on your + machine itself, it doesn't take into account that the packet might be + destined for someone on your internal network if you're using NAT for + instance. + + To compile it as a module, choose M here. If unsure, say N. + # `filter', generic and specific targets config IP_NF_FILTER tristate "Packet filtering" @@ -614,4 +629,3 @@ config IP_NF_ARP_MANGLE hardware and network addresses. endmenu - diff -NurpP --minimal linux-2.6.17.13/net/ipv4/netfilter/Makefile linux-2.6.17.13-g-v/net/ipv4/netfilter/Makefile --- linux-2.6.17.13/net/ipv4/netfilter/Makefile 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/netfilter/Makefile 2006-09-20 15:00:12.000000000 +0200 @@ -61,6 +61,7 @@ obj-$(CONFIG_IP_NF_MATCH_DSCP) += ipt_ds obj-$(CONFIG_IP_NF_MATCH_AH) += ipt_ah.o obj-$(CONFIG_IP_NF_MATCH_TTL) += ipt_ttl.o obj-$(CONFIG_IP_NF_MATCH_ADDRTYPE) += ipt_addrtype.o +obj-$(CONFIG_IP_NF_MATCH_STEALTH) += ipt_stealth.o # targets obj-$(CONFIG_IP_NF_TARGET_REJECT) += ipt_REJECT.o diff -NurpP --minimal linux-2.6.17.13/net/ipv4/netfilter/ipt_stealth.c linux-2.6.17.13-g-v/net/ipv4/netfilter/ipt_stealth.c --- linux-2.6.17.13/net/ipv4/netfilter/ipt_stealth.c 1970-01-01 01:00:00.000000000 +0100 +++ linux-2.6.17.13-g-v/net/ipv4/netfilter/ipt_stealth.c 2006-09-20 15:00:12.000000000 +0200 @@ -0,0 +1,116 @@ +/* Kernel module to add stealth support. + * + * Copyright (C) 2002,2005 Brad Spengler + * + */ + +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include + +#include + +MODULE_LICENSE("GPL"); + +extern struct sock *udp_v4_lookup(u32 saddr, u16 sport, u32 daddr, u16 dport, int dif); + +static int +match(const struct sk_buff *skb, + const struct net_device *in, + const struct net_device *out, + const struct xt_match *match, + const void *matchinfo, + int offset, + unsigned int protoff, + int *hotdrop) +{ + struct iphdr *ip = skb->nh.iph; + struct tcphdr th; + struct udphdr uh; + struct sock *sk = NULL; + + if (!ip || offset) return 0; + + switch(ip->protocol) { + case IPPROTO_TCP: + if (skb_copy_bits(skb, skb->nh.iph->ihl*4, &th, sizeof(th)) < 0) { + *hotdrop = 1; + return 0; + } + if (!(th.syn && !th.ack)) return 0; + sk = inet_lookup_listener(&tcp_hashinfo, ip->daddr, ntohs(th.dest), ((struct rtable*)skb->dst)->rt_iif); + break; + case IPPROTO_UDP: + if (skb_copy_bits(skb, skb->nh.iph->ihl*4, &uh, sizeof(uh)) < 0) { + *hotdrop = 1; + return 0; + } + sk = udp_v4_lookup(ip->saddr, uh.source, ip->daddr, uh.dest, skb->dev->ifindex); + break; + default: + return 0; + } + + if(!sk) // port is being listened on, match this + return 1; + else { + sock_put(sk); + return 0; + } +} + +/* Called when user tries to insert an entry of this type. */ +static int +checkentry(const char *tablename, + const void *nip, + const struct xt_match *match, + void *matchinfo, + unsigned int matchsize, + unsigned int hook_mask) +{ + const struct ipt_ip *ip = (const struct ipt_ip *)nip; + if (matchsize != IPT_ALIGN(0)) + return 0; + + if(((ip->proto == IPPROTO_TCP && !(ip->invflags & IPT_INV_PROTO)) || + ((ip->proto == IPPROTO_UDP) && !(ip->invflags & IPT_INV_PROTO))) + && (hook_mask & (1 << NF_IP_LOCAL_IN))) + return 1; + + printk("stealth: Only works on TCP and UDP for the INPUT chain.\n"); + + return 0; +} + + +static struct ipt_match stealth_match = { + .name = "stealth", + .match = match, + .checkentry = checkentry, + .destroy = NULL, + .me = THIS_MODULE +}; + +static int __init init(void) +{ + return ipt_register_match(&stealth_match); +} + +static void __exit fini(void) +{ + ipt_unregister_match(&stealth_match); +} + +module_init(init); +module_exit(fini); diff -NurpP --minimal linux-2.6.17.13/net/ipv4/raw.c linux-2.6.17.13-g-v/net/ipv4/raw.c --- linux-2.6.17.13/net/ipv4/raw.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/raw.c 2006-09-20 15:00:12.000000000 +0200 @@ -102,6 +102,27 @@ static void raw_v4_unhash(struct sock *s write_unlock_bh(&raw_v4_lock); } + +/* + * Check if a given address matches for a socket + * + * nxi: the socket's nx_info if any + * addr: to be verified address + * saddr/baddr: socket addresses + */ +static inline int raw_addr_match ( + struct nx_info *nxi, + uint32_t addr, + uint32_t saddr, + uint32_t baddr) +{ + if (addr && (saddr == addr || baddr == addr)) + return 1; + if (!saddr) + return addr_in_nx_info(nxi, addr); + return 0; +} + struct sock *__raw_v4_lookup(struct sock *sk, unsigned short num, unsigned long raddr, unsigned long laddr, int dif) @@ -113,7 +134,8 @@ struct sock *__raw_v4_lookup(struct sock if (inet->num == num && !(inet->daddr && inet->daddr != raddr) && - !(inet->rcv_saddr && inet->rcv_saddr != laddr) && + raw_addr_match(sk->sk_nx_info, laddr, + inet->rcv_saddr, inet->rcv_saddr2) && !(sk->sk_bound_dev_if && sk->sk_bound_dev_if != dif)) goto found; /* gotcha */ } @@ -313,6 +335,11 @@ static int raw_send_hdrinc(struct sock * iph->check = ip_fast_csum((unsigned char *)iph, iph->ihl); } + err = -EPERM; + if (!vx_check(0, VX_ADMIN) && !capable(CAP_NET_RAW) + && (!addr_in_nx_info(sk->sk_nx_info, iph->saddr))) + goto error_free; + err = NF_HOOK(PF_INET, NF_IP_LOCAL_OUT, skb, NULL, rt->u.dst.dev, dst_output); if (err > 0) @@ -324,6 +351,7 @@ out: error_fault: err = -EFAULT; +error_free: kfree_skb(skb); error: IP_INC_STATS(IPSTATS_MIB_OUTDISCARDS); @@ -484,6 +512,12 @@ static int raw_sendmsg(struct kiocb *ioc if (!inet->hdrincl) raw_probe_proto_opt(&fl, msg); + if (sk->sk_nx_info) { + err = ip_find_src(sk->sk_nx_info, &rt, &fl); + + if (err) + goto done; + } err = ip_route_output_flow(&rt, &fl, sk, !(msg->msg_flags&MSG_DONTWAIT)); } if (err) @@ -787,7 +821,8 @@ static struct sock *raw_get_first(struct struct hlist_node *node; sk_for_each(sk, node, &raw_v4_htable[state->bucket]) - if (sk->sk_family == PF_INET) + if (sk->sk_family == PF_INET && + vx_check(sk->sk_xid, VX_IDENT|VX_WATCH)) goto found; } sk = NULL; @@ -803,7 +838,8 @@ static struct sock *raw_get_next(struct sk = sk_next(sk); try_again: ; - } while (sk && sk->sk_family != PF_INET); + } while (sk && (sk->sk_family != PF_INET || + !vx_check(sk->sk_xid, VX_IDENT|VX_WATCH))); if (!sk && ++state->bucket < RAWV4_HTABLE_SIZE) { sk = sk_head(&raw_v4_htable[state->bucket]); diff -NurpP --minimal linux-2.6.17.13/net/ipv4/tcp.c linux-2.6.17.13-g-v/net/ipv4/tcp.c --- linux-2.6.17.13/net/ipv4/tcp.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/tcp.c 2006-09-20 15:00:12.000000000 +0200 @@ -258,6 +258,7 @@ #include #include #include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/net/ipv4/tcp_ipv4.c linux-2.6.17.13-g-v/net/ipv4/tcp_ipv4.c --- linux-2.6.17.13/net/ipv4/tcp_ipv4.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/tcp_ipv4.c 2006-09-20 15:00:12.000000000 +0200 @@ -62,6 +62,7 @@ #include #include #include +#include #include #include @@ -77,6 +78,7 @@ #include #include #include +#include int sysctl_tcp_tw_reuse; int sysctl_tcp_low_latency; @@ -1355,6 +1357,12 @@ static void *listening_get_next(struct s req = req->dl_next; while (1) { while (req) { + vxdprintk(VXD_CBIT(net, 6), + "sk,req: %p [#%d] (from %d)", req->sk, + (req->sk)?req->sk->sk_xid:0, vx_current_xid()); + if (req->sk && + !vx_check(req->sk->sk_xid, VX_IDENT|VX_WATCH)) + continue; if (req->rsk_ops->family == st->family) { cur = req; goto out; @@ -1379,6 +1387,10 @@ get_req: } get_sk: sk_for_each_from(sk, node) { + vxdprintk(VXD_CBIT(net, 6), "sk: %p [#%d] (from %d)", + sk, sk->sk_xid, vx_current_xid()); + if (!vx_check(sk->sk_xid, VX_IDENT|VX_WATCH)) + continue; if (sk->sk_family == st->family) { cur = sk; goto out; @@ -1430,18 +1442,26 @@ static void *established_get_first(struc read_lock(&tcp_hashinfo.ehash[st->bucket].lock); sk_for_each(sk, node, &tcp_hashinfo.ehash[st->bucket].chain) { - if (sk->sk_family != st->family) { + vxdprintk(VXD_CBIT(net, 6), + "sk,egf: %p [#%d] (from %d)", + sk, sk->sk_xid, vx_current_xid()); + if (!vx_check(sk->sk_xid, VX_IDENT|VX_WATCH)) + continue; + if (sk->sk_family != st->family) continue; - } rc = sk; goto out; } st->state = TCP_SEQ_STATE_TIME_WAIT; inet_twsk_for_each(tw, node, &tcp_hashinfo.ehash[st->bucket + tcp_hashinfo.ehash_size].chain) { - if (tw->tw_family != st->family) { + vxdprintk(VXD_CBIT(net, 6), + "tw: %p [#%d] (from %d)", + tw, tw->tw_xid, vx_current_xid()); + if (!vx_check(tw->tw_xid, VX_IDENT|VX_WATCH)) + continue; + if (tw->tw_family != st->family) continue; - } rc = tw; goto out; } @@ -1465,7 +1485,8 @@ static void *established_get_next(struct tw = cur; tw = tw_next(tw); get_tw: - while (tw && tw->tw_family != st->family) { + while (tw && (tw->tw_family != st->family || + !vx_check(tw->tw_xid, VX_IDENT|VX_WATCH))) { tw = tw_next(tw); } if (tw) { @@ -1489,6 +1510,11 @@ get_tw: sk = sk_next(sk); sk_for_each_from(sk, node) { + vxdprintk(VXD_CBIT(net, 6), + "sk,egn: %p [#%d] (from %d)", + sk, sk->sk_xid, vx_current_xid()); + if (!vx_check(sk->sk_xid, VX_IDENT|VX_WATCH)) + continue; if (sk->sk_family == st->family) goto found; } diff -NurpP --minimal linux-2.6.17.13/net/ipv4/tcp_minisocks.c linux-2.6.17.13-g-v/net/ipv4/tcp_minisocks.c --- linux-2.6.17.13/net/ipv4/tcp_minisocks.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/tcp_minisocks.c 2006-09-20 15:00:12.000000000 +0200 @@ -29,6 +29,10 @@ #include #include +#include +#include +#include + #ifdef CONFIG_SYSCTL #define SYNC_INIT 0 /* let the user enable it */ #else @@ -295,6 +299,11 @@ void tcp_time_wait(struct sock *sk, int tcptw->tw_ts_recent = tp->rx_opt.ts_recent; tcptw->tw_ts_recent_stamp = tp->rx_opt.ts_recent_stamp; + tw->tw_xid = sk->sk_xid; + tw->tw_vx_info = NULL; + tw->tw_nid = sk->sk_nid; + tw->tw_nx_info = NULL; + #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) if (tw->tw_family == PF_INET6) { struct ipv6_pinfo *np = inet6_sk(sk); diff -NurpP --minimal linux-2.6.17.13/net/ipv4/udp.c linux-2.6.17.13-g-v/net/ipv4/udp.c --- linux-2.6.17.13/net/ipv4/udp.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv4/udp.c 2006-09-20 15:00:12.000000000 +0200 @@ -102,6 +102,7 @@ #include #include #include +#include #include #include #include @@ -110,6 +111,12 @@ #include #include +extern int gr_search_udp_recvmsg(const struct sock *sk, + const struct sk_buff *skb); +extern int gr_search_udp_sendmsg(const struct sock *sk, + const struct sockaddr_in *addr); + + /* * Snmp MIB for the UDP layer */ @@ -176,14 +183,12 @@ gotit: struct inet_sock *inet2 = inet_sk(sk2); if (inet2->num == snum && - sk2 != sk && - !ipv6_only_sock(sk2) && + sk2 != sk && !ipv6_only_sock(sk2) && (!sk2->sk_bound_dev_if || !sk->sk_bound_dev_if || sk2->sk_bound_dev_if == sk->sk_bound_dev_if) && - (!inet2->rcv_saddr || - !inet->rcv_saddr || - inet2->rcv_saddr == inet->rcv_saddr) && + nx_addr_conflict(sk->sk_nx_info, + inet_rcv_saddr(sk), sk2) && (!sk2->sk_reuse || !sk->sk_reuse)) goto fail; } @@ -238,6 +243,11 @@ static struct sock *udp_v4_lookup_longwa if (inet->rcv_saddr != daddr) continue; score+=2; + } else if (sk->sk_nx_info) { + if (addr_in_nx_info(sk->sk_nx_info, daddr)) + score+=2; + else + continue; } if (inet->daddr) { if (inet->daddr != saddr) @@ -266,8 +276,7 @@ static struct sock *udp_v4_lookup_longwa return result; } -static __inline__ struct sock *udp_v4_lookup(u32 saddr, u16 sport, - u32 daddr, u16 dport, int dif) +struct sock *udp_v4_lookup(u32 saddr, u16 sport, u32 daddr, u16 dport, int dif) { struct sock *sk; @@ -294,7 +303,8 @@ static inline struct sock *udp_v4_mcast_ if (inet->num != hnum || (inet->daddr && inet->daddr != rmt_addr) || (inet->dport != rmt_port && inet->dport) || - (inet->rcv_saddr && inet->rcv_saddr != loc_addr) || + (inet->rcv_saddr && inet->rcv_saddr != loc_addr && + inet->rcv_saddr2 && inet->rcv_saddr2 != loc_addr) || ipv6_only_sock(s) || (s->sk_bound_dev_if && s->sk_bound_dev_if != dif)) continue; @@ -542,9 +552,16 @@ int udp_sendmsg(struct kiocb *iocb, stru dport = usin->sin_port; if (dport == 0) return -EINVAL; + + if (!gr_search_udp_sendmsg(sk, usin)) + return -EPERM; } else { if (sk->sk_state != TCP_ESTABLISHED) return -EDESTADDRREQ; + + if (!gr_search_udp_sendmsg(sk, NULL)) + return -EPERM; + daddr = inet->daddr; dport = inet->dport; /* Open fast path for connected socket. @@ -604,6 +621,19 @@ int udp_sendmsg(struct kiocb *iocb, stru .uli_u = { .ports = { .sport = inet->sport, .dport = dport } } }; + struct nx_info *nxi = sk->sk_nx_info; + + if (nxi) { + err = ip_find_src(nxi, &rt, &fl); + if (err) + goto out; + if (daddr == IPI_LOOPBACK && !vx_check(0, VX_ADMIN)) + daddr = fl.fl4_dst = nxi->ipv4[0]; +#ifdef CONFIG_VSERVER_REMAP_SADDR + if (saddr == IPI_LOOPBACK && !vx_check(0, VX_ADMIN)) + saddr = fl.fl4_src = nxi->ipv4[0]; +#endif + } err = ip_route_output_flow(&rt, &fl, sk, !(msg->msg_flags&MSG_DONTWAIT)); if (err) goto out; @@ -798,6 +828,11 @@ try_again: if (!skb) goto out; + if (!gr_search_udp_recvmsg(sk, skb)) { + err = -EPERM; + goto out_free; + } + copied = skb->len - sizeof(struct udphdr); if (copied > len) { copied = len; @@ -1403,8 +1438,10 @@ static struct sock *udp_get_first(struct for (state->bucket = 0; state->bucket < UDP_HTABLE_SIZE; ++state->bucket) { struct hlist_node *node; + sk_for_each(sk, node, &udp_hash[state->bucket]) { - if (sk->sk_family == state->family) + if (sk->sk_family == state->family && + vx_check(sk->sk_xid, VX_IDENT|VX_WATCH)) goto found; } } @@ -1421,7 +1458,8 @@ static struct sock *udp_get_next(struct sk = sk_next(sk); try_again: ; - } while (sk && sk->sk_family != state->family); + } while (sk && (sk->sk_family != state->family || + !vx_check(sk->sk_xid, VX_IDENT|VX_WATCH))); if (!sk && ++state->bucket < UDP_HTABLE_SIZE) { sk = sk_head(&udp_hash[state->bucket]); diff -NurpP --minimal linux-2.6.17.13/net/ipv6/addrconf.c linux-2.6.17.13-g-v/net/ipv6/addrconf.c --- linux-2.6.17.13/net/ipv6/addrconf.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/ipv6/addrconf.c 2006-09-20 15:00:12.000000000 +0200 @@ -2654,7 +2654,10 @@ static void if6_seq_stop(struct seq_file static int if6_seq_show(struct seq_file *seq, void *v) { struct inet6_ifaddr *ifp = (struct inet6_ifaddr *)v; - seq_printf(seq, + + /* no ipv6 inside a vserver for now */ + if (vx_check(0, VX_ADMIN|VX_WATCH)) + seq_printf(seq, NIP6_SEQFMT " %02x %02x %02x %02x %8s\n", NIP6(ifp->addr), ifp->idev->dev->ifindex, @@ -3014,6 +3017,10 @@ static int inet6_dump_addr(struct sk_buf struct ifmcaddr6 *ifmca; struct ifacaddr6 *ifaca; + /* no ipv6 inside a vserver for now */ + if (skb->sk && skb->sk->sk_vx_info) + return skb->len; + s_idx = cb->args[0]; s_ip_idx = ip_idx = cb->args[1]; read_lock(&dev_base_lock); @@ -3239,6 +3246,10 @@ static int inet6_dump_ifinfo(struct sk_b struct net_device *dev; struct inet6_dev *idev; + /* no ipv6 inside a vserver for now */ + if (skb->sk && skb->sk->sk_vx_info) + return skb->len; + read_lock(&dev_base_lock); for (dev=dev_base, idx=0; dev; dev = dev->next, idx++) { if (idx < s_idx) diff -NurpP --minimal linux-2.6.17.13/net/netlink/af_netlink.c linux-2.6.17.13-g-v/net/netlink/af_netlink.c --- linux-2.6.17.13/net/netlink/af_netlink.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/netlink/af_netlink.c 2006-09-20 15:00:12.000000000 +0200 @@ -57,6 +57,9 @@ #include #include #include +#include +#include +#include #include #include diff -NurpP --minimal linux-2.6.17.13/net/socket.c linux-2.6.17.13-g-v/net/socket.c --- linux-2.6.17.13/net/socket.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/socket.c 2006-09-20 15:00:12.000000000 +0200 @@ -86,6 +86,7 @@ #include #include #include +#include #include #include @@ -94,6 +95,22 @@ #include #include +#include + +extern void gr_attach_curr_ip(const struct sock *sk); +extern int gr_handle_sock_all(const int family, const int type, + const int protocol); +extern int gr_handle_sock_server(const struct sockaddr *sck); +extern int gr_handle_sock_server_other(const struct socket *sck); +extern int gr_handle_sock_client(const struct sockaddr *sck); +extern int gr_search_connect(const struct socket * sock, + const struct sockaddr_in * addr); +extern int gr_search_bind(const struct socket * sock, + const struct sockaddr_in * addr); +extern int gr_search_listen(const struct socket * sock); +extern int gr_search_accept(const struct socket * sock); +extern int gr_search_socket(const int domain, const int type, + const int protocol); static int sock_no_open(struct inode *irrelevant, struct file *dontcare); static ssize_t sock_aio_read(struct kiocb *iocb, char __user *buf, @@ -341,7 +358,7 @@ static struct super_block *sockfs_get_sb return get_sb_pseudo(fs_type, "socket:", &sockfs_ops, SOCKFS_MAGIC); } -static struct vfsmount *sock_mnt __read_mostly; +struct vfsmount *sock_mnt __read_mostly; static struct file_system_type sock_fs_type = { .name = "sockfs", @@ -582,7 +599,7 @@ static inline int __sock_sendmsg(struct struct msghdr *msg, size_t size) { struct sock_iocb *si = kiocb_to_siocb(iocb); - int err; + int err, len; si->sock = sock; si->scm = NULL; @@ -593,7 +610,21 @@ static inline int __sock_sendmsg(struct if (err) return err; - return sock->ops->sendmsg(iocb, sock, msg, size); + len = sock->ops->sendmsg(iocb, sock, msg, size); + if (sock->sk) { + if (len == size) + vx_sock_send(sock->sk, size); + else + vx_sock_fail(sock->sk, size); + } + vxdprintk(VXD_CBIT(net, 7), + "__sock_sendmsg: %p[%p,%p,%p;%d]:%d/%d", + sock, sock->sk, + (sock->sk)?sock->sk->sk_nx_info:0, + (sock->sk)?sock->sk->sk_vx_info:0, + (sock->sk)?sock->sk->sk_xid:0, + (unsigned int)size, len); + return len; } int sock_sendmsg(struct socket *sock, struct msghdr *msg, size_t size) @@ -631,7 +662,7 @@ int kernel_sendmsg(struct socket *sock, static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *msg, size_t size, int flags) { - int err; + int err, len; struct sock_iocb *si = kiocb_to_siocb(iocb); si->sock = sock; @@ -644,7 +675,17 @@ static inline int __sock_recvmsg(struct if (err) return err; - return sock->ops->recvmsg(iocb, sock, msg, size, flags); + len = sock->ops->recvmsg(iocb, sock, msg, size, flags); + if ((len >= 0) && sock->sk) + vx_sock_recv(sock->sk, len); + vxdprintk(VXD_CBIT(net, 7), + "__sock_recvmsg: %p[%p,%p,%p;%d]:%d/%d", + sock, sock->sk, + (sock->sk)?sock->sk->sk_nx_info:0, + (sock->sk)?sock->sk->sk_vx_info:0, + (sock->sk)?sock->sk->sk_xid:0, + (unsigned int)size, len); + return len; } int sock_recvmsg(struct socket *sock, struct msghdr *msg, @@ -1134,6 +1175,10 @@ static int __sock_create(int family, int if (type < 0 || type >= SOCK_MAX) return -EINVAL; + /* disable IPv6 inside vservers for now */ + if (family == PF_INET6 && !vx_check(0, VX_ADMIN)) + return -EAFNOSUPPORT; + /* Compatibility. This uglymoron is moved from INET layer to here to avoid @@ -1240,10 +1285,21 @@ asmlinkage long sys_socket(int family, i int retval; struct socket *sock; + if(!gr_search_socket(family, type, protocol)) { + retval = -EACCES; + goto out; + } + + if (gr_handle_sock_all(family, type, protocol)) { + retval = -EACCES; + goto out; + } + retval = sock_create(family, type, protocol, &sock); if (retval < 0) goto out; + set_bit(SOCK_USER_SOCKET, &sock->flags); retval = sock_map_fd(sock); if (retval < 0) goto out_release; @@ -1274,10 +1330,12 @@ asmlinkage long sys_socketpair(int famil err = sock_create(family, type, protocol, &sock1); if (err < 0) goto out; + set_bit(SOCK_USER_SOCKET, &sock1->flags); err = sock_create(family, type, protocol, &sock2); if (err < 0) goto out_release_1; + set_bit(SOCK_USER_SOCKET, &sock2->flags); err = sock1->ops->socketpair(sock1, sock2); if (err < 0) @@ -1335,16 +1393,25 @@ asmlinkage long sys_bind(int fd, struct { struct socket *sock; char address[MAX_SOCK_ADDR]; + struct sockaddr *sck; int err, fput_needed; if((sock = sockfd_lookup_light(fd, &err, &fput_needed))!=NULL) { if((err=move_addr_to_kernel(umyaddr,addrlen,address))>=0) { + sck = (struct sockaddr *)address; + if (!gr_search_bind(sock, (struct sockaddr_in *)sck) || + gr_handle_sock_server(sck)) { + err = -EACCES; + goto error; + } + err = security_socket_bind(sock, (struct sockaddr *)address, addrlen); if (!err) err = sock->ops->bind(sock, (struct sockaddr *)address, addrlen); } +error: fput_light(sock->file, fput_needed); } return err; @@ -1368,10 +1435,17 @@ asmlinkage long sys_listen(int fd, int b if ((unsigned) backlog > sysctl_somaxconn) backlog = sysctl_somaxconn; + if (gr_handle_sock_server_other(sock) || + !gr_search_listen(sock)) { + err = -EPERM; + goto error; + } + err = security_socket_listen(sock, backlog); if (!err) err = sock->ops->listen(sock, backlog); +error: fput_light(sock->file, fput_needed); } return err; @@ -1408,6 +1482,13 @@ asmlinkage long sys_accept(int fd, struc newsock->type = sock->type; newsock->ops = sock->ops; + if (gr_handle_sock_server_other(sock) || + !gr_search_accept(sock)) { + err = -EPERM; + sock_release(newsock); + goto out_put; + } + /* * We don't need try_module_get here, as the listening socket (sock) * has the protocol module (sock->ops->owner) held. @@ -1449,6 +1530,7 @@ asmlinkage long sys_accept(int fd, struc err = newfd; security_socket_post_accept(sock, newsock); + gr_attach_curr_ip(newsock->sk); out_put: fput_light(sock->file, fput_needed); @@ -1477,6 +1559,7 @@ asmlinkage long sys_connect(int fd, stru { struct socket *sock; char address[MAX_SOCK_ADDR]; + struct sockaddr *sck; int err, fput_needed; sock = sockfd_lookup_light(fd, &err, &fput_needed); @@ -1486,6 +1569,13 @@ asmlinkage long sys_connect(int fd, stru if (err < 0) goto out_put; + sck = (struct sockaddr *)address; + if (!gr_search_connect(sock, (struct sockaddr_in *)sck) || + gr_handle_sock_client(sck)) { + err = -EACCES; + goto out_put; + } + err = security_socket_connect(sock, (struct sockaddr *)address, addrlen); if (err) goto out_put; @@ -1740,6 +1830,7 @@ asmlinkage long sys_shutdown(int fd, int err = sock->ops->shutdown(sock, how); fput_light(sock->file, fput_needed); } + return err; } diff -NurpP --minimal linux-2.6.17.13/net/sunrpc/auth.c linux-2.6.17.13-g-v/net/sunrpc/auth.c --- linux-2.6.17.13/net/sunrpc/auth.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/sunrpc/auth.c 2006-09-20 15:00:12.000000000 +0200 @@ -13,6 +13,7 @@ #include #include #include +#include #ifdef RPC_DEBUG # define RPCDBG_FACILITY RPCDBG_AUTH @@ -263,6 +264,7 @@ rpcauth_lookupcred(struct rpc_auth *auth struct auth_cred acred = { .uid = current->fsuid, .gid = current->fsgid, + .xid = vx_current_xid(), .group_info = current->group_info, }; struct rpc_cred *ret; @@ -282,6 +284,7 @@ rpcauth_bindcred(struct rpc_task *task) struct auth_cred acred = { .uid = current->fsuid, .gid = current->fsgid, + .xid = vx_current_xid(), .group_info = current->group_info, }; struct rpc_cred *ret; diff -NurpP --minimal linux-2.6.17.13/net/sunrpc/auth_unix.c linux-2.6.17.13-g-v/net/sunrpc/auth_unix.c --- linux-2.6.17.13/net/sunrpc/auth_unix.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/sunrpc/auth_unix.c 2006-09-20 15:00:12.000000000 +0200 @@ -11,12 +11,14 @@ #include #include #include +#include #define NFS_NGROUPS 16 struct unx_cred { struct rpc_cred uc_base; gid_t uc_gid; + xid_t uc_xid; gid_t uc_gids[NFS_NGROUPS]; }; #define uc_uid uc_base.cr_uid @@ -78,6 +80,7 @@ unx_create_cred(struct rpc_auth *auth, s if (flags & RPCAUTH_LOOKUP_ROOTCREDS) { cred->uc_uid = 0; cred->uc_gid = 0; + cred->uc_xid = vx_current_xid(); cred->uc_gids[0] = NOGROUP; } else { int groups = acred->group_info->ngroups; @@ -86,6 +89,7 @@ unx_create_cred(struct rpc_auth *auth, s cred->uc_uid = acred->uid; cred->uc_gid = acred->gid; + cred->uc_xid = acred->xid; for (i = 0; i < groups; i++) cred->uc_gids[i] = GROUP_AT(acred->group_info, i); if (i < NFS_NGROUPS) @@ -117,7 +121,8 @@ unx_match(struct auth_cred *acred, struc int groups; if (cred->uc_uid != acred->uid - || cred->uc_gid != acred->gid) + || cred->uc_gid != acred->gid + || cred->uc_xid != acred->xid) return 0; groups = acred->group_info->ngroups; @@ -143,7 +148,7 @@ unx_marshal(struct rpc_task *task, u32 * struct rpc_clnt *clnt = task->tk_client; struct unx_cred *cred = (struct unx_cred *) task->tk_msg.rpc_cred; u32 *base, *hold; - int i; + int i, tagxid; *p++ = htonl(RPC_AUTH_UNIX); base = p++; @@ -153,9 +158,12 @@ unx_marshal(struct rpc_task *task, u32 * * Copy the UTS nodename captured when the client was created. */ p = xdr_encode_array(p, clnt->cl_nodename, clnt->cl_nodelen); + tagxid = task->tk_client->cl_tagxid; - *p++ = htonl((u32) cred->uc_uid); - *p++ = htonl((u32) cred->uc_gid); + *p++ = htonl((u32) XIDINO_UID(tagxid, + cred->uc_uid, cred->uc_xid)); + *p++ = htonl((u32) XIDINO_GID(tagxid, + cred->uc_gid, cred->uc_xid)); hold = p++; for (i = 0; i < 16 && cred->uc_gids[i] != (gid_t) NOGROUP; i++) *p++ = htonl((u32) cred->uc_gids[i]); diff -NurpP --minimal linux-2.6.17.13/net/unix/af_unix.c linux-2.6.17.13-g-v/net/unix/af_unix.c --- linux-2.6.17.13/net/unix/af_unix.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/unix/af_unix.c 2006-09-20 15:00:12.000000000 +0200 @@ -117,6 +117,10 @@ #include #include #include +#include +#include +#include +#include int sysctl_unix_max_dgram_qlen = 10; @@ -235,6 +239,8 @@ static struct sock *__unix_find_socket_b sk_for_each(s, node, &unix_socket_table[hash ^ type]) { struct unix_sock *u = unix_sk(s); + if (!vx_check(s->sk_xid, VX_IDENT|VX_WATCH)) + continue; if (u->addr->len == len && !memcmp(u->addr->name, sunname, len)) goto found; @@ -680,6 +686,11 @@ static struct sock *unix_find_other(stru if (err) goto put_fail; + if (!gr_acl_handle_unix(nd.dentry, nd.mnt)) { + err = -EACCES; + goto put_fail; + } + err = -ECONNREFUSED; if (!S_ISSOCK(nd.dentry->d_inode->i_mode)) goto put_fail; @@ -703,6 +714,13 @@ static struct sock *unix_find_other(stru if (u) { struct dentry *dentry; dentry = unix_sk(u)->dentry; + + if (!gr_handle_chroot_unix(u->sk_peercred.pid)) { + err = -EPERM; + sock_put(u); + goto fail; + } + if (dentry) touch_atime(unix_sk(u)->mnt, dentry); } else @@ -781,9 +799,18 @@ static int unix_bind(struct socket *sock */ mode = S_IFSOCK | (SOCK_INODE(sock)->i_mode & ~current->fs->umask); - err = vfs_mknod(nd.dentry->d_inode, dentry, mode, 0); + + if (!gr_acl_handle_mknod(dentry, nd.dentry, nd.mnt, mode)) { + err = -EACCES; + goto out_mknod_dput; + } + + err = vfs_mknod(nd.dentry->d_inode, dentry, mode, 0, NULL); if (err) goto out_mknod_dput; + + gr_handle_create(dentry, nd.mnt); + mutex_unlock(&nd.dentry->d_inode->i_mutex); dput(nd.dentry); nd.dentry = dentry; @@ -801,6 +828,10 @@ static int unix_bind(struct socket *sock goto out_unlock; } +#ifdef CONFIG_GRKERNSEC_CHROOT_UNIX + sk->sk_peercred.pid = current->pid; +#endif + list = &unix_socket_table[addr->hash]; } else { list = &unix_socket_table[dentry->d_inode->i_ino & (UNIX_HASH_SIZE-1)]; diff -NurpP --minimal linux-2.6.17.13/net/x25/af_x25.c linux-2.6.17.13-g-v/net/x25/af_x25.c --- linux-2.6.17.13/net/x25/af_x25.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/net/x25/af_x25.c 2006-09-20 15:00:12.000000000 +0200 @@ -502,7 +502,10 @@ static int x25_create(struct socket *soc x25 = x25_sk(sk); - sock_init_data(sock, sk); + sk->sk_socket = sock; + sk->sk_type = sock->type; + sk->sk_sleep = &sock->wait; + sock->sk = sk; x25_init_timers(sk); diff -NurpP --minimal linux-2.6.17.13/security/Kconfig linux-2.6.17.13-g-v/security/Kconfig --- linux-2.6.17.13/security/Kconfig 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/security/Kconfig 2006-09-20 15:00:12.000000000 +0200 @@ -4,6 +4,453 @@ menu "Security options" +menu "PaX" + +config PAX + bool "Enable various PaX features" + depends on GRKERNSEC && (ALPHA || ARM || IA64 || MIPS32 || MIPS64 || PARISC || PPC32 || PPC64 || SPARC32 || SPARC64 || X86 || X86_64) + help + This allows you to enable various PaX features. PaX adds + intrusion prevention mechanisms to the kernel that reduce + the risks posed by exploitable memory corruption bugs. + +menu "PaX Control" + depends on PAX + +config PAX_SOFTMODE + bool 'Support soft mode' + help + Enabling this option will allow you to run PaX in soft mode, that + is, PaX features will not be enforced by default, only on executables + marked explicitly. You must also enable PT_PAX_FLAGS support as it + is the only way to mark executables for soft mode use. + + Soft mode can be activated by using the "pax_softmode=1" kernel command + line option on boot. Furthermore you can control various PaX features + at runtime via the entries in /proc/sys/kernel/pax. + +config PAX_EI_PAX + bool 'Use legacy ELF header marking' + help + Enabling this option will allow you to control PaX features on + a per executable basis via the 'chpax' utility available at + http://pax.grsecurity.net/. The control flags will be read from + an otherwise reserved part of the ELF header. This marking has + numerous drawbacks (no support for soft-mode, toolchain does not + know about the non-standard use of the ELF header) therefore it + has been deprecated in favour of PT_PAX_FLAGS support. + + If you have applications not marked by the PT_PAX_FLAGS ELF + program header then you MUST enable this option otherwise they + will not get any protection. + + Note that if you enable PT_PAX_FLAGS marking support as well, + the PT_PAX_FLAG marks will override the legacy EI_PAX marks. + +config PAX_PT_PAX_FLAGS + bool 'Use ELF program header marking' + help + Enabling this option will allow you to control PaX features on + a per executable basis via the 'paxctl' utility available at + http://pax.grsecurity.net/. The control flags will be read from + a PaX specific ELF program header (PT_PAX_FLAGS). This marking + has the benefits of supporting both soft mode and being fully + integrated into the toolchain (the binutils patch is available + from http://pax.grsecurity.net). + + If you have applications not marked by the PT_PAX_FLAGS ELF + program header then you MUST enable the EI_PAX marking support + otherwise they will not get any protection. + + Note that if you enable the legacy EI_PAX marking support as well, + the EI_PAX marks will be overridden by the PT_PAX_FLAGS marks. + +choice + prompt 'MAC system integration' + default PAX_HAVE_ACL_FLAGS + help + Mandatory Access Control systems have the option of controlling + PaX flags on a per executable basis, choose the method supported + by your particular system. + + - "none": if your MAC system does not interact with PaX, + - "direct": if your MAC system defines pax_set_flags() itself, + - "hook": if your MAC system uses the pax_set_flags_func callback. + + NOTE: this option is for developers/integrators only. + +config PAX_NO_ACL_FLAGS + bool 'none' + +config PAX_HAVE_ACL_FLAGS + bool 'direct' + +config PAX_HOOK_ACL_FLAGS + bool 'hook' +endchoice + +endmenu + +menu "Non-executable pages" + depends on PAX + +config PAX_NOEXEC + bool "Enforce non-executable pages" + depends on (PAX_EI_PAX || PAX_PT_PAX_FLAGS || PAX_HAVE_ACL_FLAGS || PAX_HOOK_ACL_FLAGS) && (ALPHA || IA64 || MIPS32 || MIPS64 || PARISC || PPC32 || PPC64 || SPARC32 || SPARC64 || X86 || X86_64) + help + By design some architectures do not allow for protecting memory + pages against execution or even if they do, Linux does not make + use of this feature. In practice this means that if a page is + readable (such as the stack or heap) it is also executable. + + There is a well known exploit technique that makes use of this + fact and a common programming mistake where an attacker can + introduce code of his choice somewhere in the attacked program's + memory (typically the stack or the heap) and then execute it. + + If the attacked program was running with different (typically + higher) privileges than that of the attacker, then he can elevate + his own privilege level (e.g. get a root shell, write to files for + which he does not have write access to, etc). + + Enabling this option will let you choose from various features + that prevent the injection and execution of 'foreign' code in + a program. + + This will also break programs that rely on the old behaviour and + expect that dynamically allocated memory via the malloc() family + of functions is executable (which it is not). Notable examples + are the XFree86 4.x server, the java runtime and wine. + +config PAX_PAGEEXEC + bool "Paging based non-executable pages" + depends on PAX_NOEXEC && (!X86_32 || M586 || M586TSC || M586MMX || M686 || MPENTIUMII || MPENTIUMIII || MPENTIUMM || MPENTIUM4 || MK7 || MK8 || MWINCHIPC6 || MWINCHIP2 || MWINCHIP3D || MVIAC3_2) + select PAX_NOVSYSCALL if X86_32 + help + This implementation is based on the paging feature of the CPU. + On i386 and ppc there is a variable but usually low performance + impact on applications. On alpha, ia64, parisc, sparc, sparc64 + and x86_64 there is no performance impact. + +config PAX_SEGMEXEC + bool "Segmentation based non-executable pages" + depends on PAX_NOEXEC && X86_32 + select PAX_NOVSYSCALL + help + This implementation is based on the segmentation feature of the + CPU and has little performance impact, however applications will + be limited to a 1.5 GB address space instead of the normal 3 GB. + +choice + prompt "Default non-executable page method" + depends on PAX_PAGEEXEC && PAX_SEGMEXEC + default PAX_DEFAULT_SEGMEXEC + help + Select the default non-executable page method applied to applications + that do not select one themselves. + +config PAX_DEFAULT_PAGEEXEC + bool "PAGEEXEC" + +config PAX_DEFAULT_SEGMEXEC + bool "SEGMEXEC" +endchoice + +config PAX_EMUTRAMP + bool "Emulate trampolines" if (PAX_PAGEEXEC || PAX_SEGMEXEC) && (PARISC || PPC32 || X86_32) + default y if PARISC || PPC32 + help + There are some programs and libraries that for one reason or + another attempt to execute special small code snippets from + non-executable memory pages. Most notable examples are the + signal handler return code generated by the kernel itself and + the GCC trampolines. + + If you enabled CONFIG_PAX_PAGEEXEC or CONFIG_PAX_SEGMEXEC then + such programs will no longer work under your kernel. + + As a remedy you can say Y here and use the 'chpax' or 'paxctl' + utilities to enable trampoline emulation for the affected programs + yet still have the protection provided by the non-executable pages. + + On parisc and ppc you MUST enable this option and EMUSIGRT as + well, otherwise your system will not even boot. + + Alternatively you can say N here and use the 'chpax' or 'paxctl' + utilities to disable CONFIG_PAX_PAGEEXEC and CONFIG_PAX_SEGMEXEC + for the affected files. + + NOTE: enabling this feature *may* open up a loophole in the + protection provided by non-executable pages that an attacker + could abuse. Therefore the best solution is to not have any + files on your system that would require this option. This can + be achieved by not using libc5 (which relies on the kernel + signal handler return code) and not using or rewriting programs + that make use of the nested function implementation of GCC. + Skilled users can just fix GCC itself so that it implements + nested function calls in a way that does not interfere with PaX. + +config PAX_EMUSIGRT + bool "Automatically emulate sigreturn trampolines" + depends on PAX_EMUTRAMP && (PARISC || PPC32) + default y + help + Enabling this option will have the kernel automatically detect + and emulate signal return trampolines executing on the stack + that would otherwise lead to task termination. + + This solution is intended as a temporary one for users with + legacy versions of libc (libc5, glibc 2.0, uClibc before 0.9.17, + Modula-3 runtime, etc) or executables linked to such, basically + everything that does not specify its own SA_RESTORER function in + normal executable memory like glibc 2.1+ does. + + On parisc and ppc you MUST enable this option, otherwise your + system will not even boot. + + NOTE: this feature cannot be disabled on a per executable basis + and since it *does* open up a loophole in the protection provided + by non-executable pages, the best solution is to not have any + files on your system that would require this option. + +config PAX_MPROTECT + bool "Restrict mprotect()" + depends on (PAX_PAGEEXEC || PAX_SEGMEXEC) && !PPC64 + help + Enabling this option will prevent programs from + - changing the executable status of memory pages that were + not originally created as executable, + - making read-only executable pages writable again, + - creating executable pages from anonymous memory. + + You should say Y here to complete the protection provided by + the enforcement of non-executable pages. + + NOTE: you can use the 'chpax' or 'paxctl' utilities to control + this feature on a per file basis. + +config PAX_NOELFRELOCS + bool "Disallow ELF text relocations" + depends on PAX_MPROTECT && !PAX_ETEXECRELOCS && (IA64 || X86 || X86_64) + help + Non-executable pages and mprotect() restrictions are effective + in preventing the introduction of new executable code into an + attacked task's address space. There remain only two venues + for this kind of attack: if the attacker can execute already + existing code in the attacked task then he can either have it + create and mmap() a file containing his code or have it mmap() + an already existing ELF library that does not have position + independent code in it and use mprotect() on it to make it + writable and copy his code there. While protecting against + the former approach is beyond PaX, the latter can be prevented + by having only PIC ELF libraries on one's system (which do not + need to relocate their code). If you are sure this is your case, + then enable this option otherwise be careful as you may not even + be able to boot or log on your system (for example, some PAM + modules are erroneously compiled as non-PIC by default). + + NOTE: if you are using dynamic ELF executables (as suggested + when using ASLR) then you must have made sure that you linked + your files using the PIC version of crt1 (the et_dyn.tar.gz package + referenced there has already been updated to support this). + +config PAX_ETEXECRELOCS + bool "Allow ELF ET_EXEC text relocations" + depends on PAX_MPROTECT && (ALPHA || IA64 || PARISC) + default y + help + On some architectures there are incorrectly created applications + that require text relocations and would not work without enabling + this option. If you are an alpha, ia64 or parisc user, you should + enable this option and disable it once you have made sure that + none of your applications need it. + +config PAX_EMUPLT + bool "Automatically emulate ELF PLT" + depends on PAX_MPROTECT && (ALPHA || PARISC || PPC32 || SPARC32 || SPARC64) + default y + help + Enabling this option will have the kernel automatically detect + and emulate the Procedure Linkage Table entries in ELF files. + On some architectures such entries are in writable memory, and + become non-executable leading to task termination. Therefore + it is mandatory that you enable this option on alpha, parisc, ppc, + sparc and sparc64, otherwise your system would not even boot. + + NOTE: this feature *does* open up a loophole in the protection + provided by the non-executable pages, therefore the proper + solution is to modify the toolchain to produce a PLT that does + not need to be writable. + +config PAX_DLRESOLVE + bool + depends on PAX_EMUPLT && (SPARC32 || SPARC64) + default y + +config PAX_SYSCALL + bool + depends on PAX_PAGEEXEC && PPC32 + default y + +config PAX_KERNEXEC + bool "Enforce non-executable kernel pages" + depends on PAX_NOEXEC && X86_32 && !HOTPLUG_PCI_COMPAQ_NVRAM && !PCI_BIOS && !EFI && !DEBUG_RODATA + select PAX_NOVSYSCALL + help + This is the kernel land equivalent of PAGEEXEC and MPROTECT, + that is, enabling this option will make it harder to inject + and execute 'foreign' code in kernel memory itself. + +endmenu + +menu "Address Space Layout Randomization" + depends on PAX + +config PAX_ASLR + bool "Address Space Layout Randomization" + depends on PAX_EI_PAX || PAX_PT_PAX_FLAGS || PAX_HAVE_ACL_FLAGS || PAX_HOOK_ACL_FLAGS + help + Many if not most exploit techniques rely on the knowledge of + certain addresses in the attacked program. The following options + will allow the kernel to apply a certain amount of randomization + to specific parts of the program thereby forcing an attacker to + guess them in most cases. Any failed guess will most likely crash + the attacked program which allows the kernel to detect such attempts + and react on them. PaX itself provides no reaction mechanisms, + instead it is strongly encouraged that you make use of Nergal's + segvguard (ftp://ftp.pl.openwall.com/misc/segvguard/) or grsecurity's + (http://www.grsecurity.net/) built-in crash detection features or + develop one yourself. + + By saying Y here you can choose to randomize the following areas: + - top of the task's kernel stack + - top of the task's userland stack + - base address for mmap() requests that do not specify one + (this includes all libraries) + - base address of the main executable + + It is strongly recommended to say Y here as address space layout + randomization has negligible impact on performance yet it provides + a very effective protection. + + NOTE: you can use the 'chpax' or 'paxctl' utilities to control + this feature on a per file basis. + +config PAX_RANDKSTACK + bool "Randomize kernel stack base" + depends on PAX_ASLR && X86_TSC && X86_32 + help + By saying Y here the kernel will randomize every task's kernel + stack on every system call. This will not only force an attacker + to guess it but also prevent him from making use of possible + leaked information about it. + + Since the kernel stack is a rather scarce resource, randomization + may cause unexpected stack overflows, therefore you should very + carefully test your system. Note that once enabled in the kernel + configuration, this feature cannot be disabled on a per file basis. + +config PAX_RANDUSTACK + bool "Randomize user stack base" + depends on PAX_ASLR + help + By saying Y here the kernel will randomize every task's userland + stack. The randomization is done in two steps where the second + one may apply a big amount of shift to the top of the stack and + cause problems for programs that want to use lots of memory (more + than 2.5 GB if SEGMEXEC is not active, or 1.25 GB when it is). + For this reason the second step can be controlled by 'chpax' or + 'paxctl' on a per file basis. + +config PAX_RANDMMAP + bool "Randomize mmap() base" + depends on PAX_ASLR + help + By saying Y here the kernel will use a randomized base address for + mmap() requests that do not specify one themselves. As a result + all dynamically loaded libraries will appear at random addresses + and therefore be harder to exploit by a technique where an attacker + attempts to execute library code for his purposes (e.g. spawn a + shell from an exploited program that is running at an elevated + privilege level). + + Furthermore, if a program is relinked as a dynamic ELF file, its + base address will be randomized as well, completing the full + randomization of the address space layout. Attacking such programs + becomes a guess game. You can find an example of doing this at + http://pax.grsecurity.net/et_dyn.tar.gz and practical samples at + http://www.grsecurity.net/grsec-gcc-specs.tar.gz . + + NOTE: you can use the 'chpax' or 'paxctl' utilities to control this + feature on a per file basis. + +config PAX_NOVSYSCALL + bool "Disable the vsyscall page" + depends on PAX_ASLR && X86_32 + help + The Linux 2.6 kernel introduced a new feature that speeds up or + simplifies certain operations, such as system calls or returns + from signal handlers. + + Unfortunately the implementation also gives a powerful instrument + into the hands of exploit writers: the so-called vsyscall page exists + in every task at the same fixed address and it contains machine code + that is very useful in performing the return-to-libc style attack. + + Since this exploit technique cannot in general be protected against + via kernel solutions, this option will allow you to disable the use + of the vsyscall page and revert back to the old behaviour. + +endmenu + +menu "Miscellaneous hardening features" + +config PAX_MEMORY_SANITIZE + bool "Sanitize all freed memory" + help + By saying Y here the kernel will erase memory pages as soon as they + are freed. This in turn reduces the lifetime of data stored in the + pages, making it less likely that sensitive information such as + passwords, cryptographic secrets, etc stay in memory for too long. + + This is especially useful for programs whose runtime is short, long + lived processes and the kernel itself benefit from this as long as + they operate on whole memory pages and ensure timely freeing of pages + that may hold sensitive information. + + The tradeoff is performance impact, on a single CPU system kernel + compilation sees a 3% slowdown, other systems and workloads may vary + and you are advised to test this feature on your expected workload + before deploying it. + + Note that this feature does not protect data stored in live pages, + e.g., process memory swapped to disk may stay there for a long time. + +config PAX_MEMORY_UDEREF + bool "Prevent invalid userland pointer dereference" + depends on X86_32 + select PAX_NOVSYSCALL + help + By saying Y here the kernel will be prevented from dereferencing + userland pointers in contexts where the kernel expects only kernel + pointers. This is both a useful runtime debugging feature and a + security measure that prevents exploiting a class of kernel bugs. + + The tradeoff is that some virtualization solutions may experience + a huge slowdown and therefore you should not enable this feature + for kernels meant to run in such environments. Whether a given VM + solution is affected or not is best determined by simply trying it + out, the performance impact will be obvious right on boot as this + mechanism engages from very early on. A good rule of thumb is that + VMs running on CPUs without hardware virtualization support (i.e., + the majority of IA-32 CPUs) will likely experience the slowdown. + +endmenu + +endmenu + +source grsecurity/Kconfig + config KEYS bool "Enable access key retention support" help diff -NurpP --minimal linux-2.6.17.13/security/commoncap.c linux-2.6.17.13-g-v/security/commoncap.c --- linux-2.6.17.13/security/commoncap.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/security/commoncap.c 2006-09-20 15:00:12.000000000 +0200 @@ -24,6 +24,7 @@ #include #include #include +#include int cap_netlink_send(struct sock *sk, struct sk_buff *skb) { @@ -45,7 +46,15 @@ EXPORT_SYMBOL(cap_netlink_recv); int cap_capable (struct task_struct *tsk, int cap) { /* Derived from include/linux/sched.h:capable. */ - if (cap_raised(tsk->cap_effective, cap)) + if (cap_raised (tsk->cap_effective, cap) && gr_task_is_capable(tsk, cap)) + return 0; + return -EPERM; +} + +int cap_capable_nolog (struct task_struct *tsk, int cap) +{ + /* Derived from include/linux/sched.h:capable. */ + if (cap_raised (tsk->cap_effective, cap)) return 0; return -EPERM; } @@ -143,7 +152,7 @@ void cap_bprm_apply_creds (struct linux_ /* Derived from fs/exec.c:compute_creds. */ kernel_cap_t new_permitted, working; - new_permitted = cap_intersect (bprm->cap_permitted, cap_bset); + new_permitted = cap_intersect (bprm->cap_permitted, vx_current_bcaps()); working = cap_intersect (bprm->cap_inheritable, current->cap_inheritable); new_permitted = cap_combine (new_permitted, working); @@ -164,8 +173,11 @@ void cap_bprm_apply_creds (struct linux_ } } - current->suid = current->euid = current->fsuid = bprm->e_uid; - current->sgid = current->egid = current->fsgid = bprm->e_gid; + if (!gr_check_user_change(-1, bprm->e_uid, bprm->e_uid)) + current->suid = current->euid = current->fsuid = bprm->e_uid; + + if (!gr_check_group_change(-1, bprm->e_gid, bprm->e_gid)) + current->sgid = current->egid = current->fsgid = bprm->e_gid; /* For init, we want to retain the capabilities set * in the init_task struct. Thus we skip the usual @@ -176,6 +188,8 @@ void cap_bprm_apply_creds (struct linux_ cap_intersect (new_permitted, bprm->cap_effective); } + gr_handle_chroot_caps(current); + /* AUD: Audit candidate if current->cap_effective is set */ current->keep_capabilities = 0; @@ -312,7 +326,8 @@ void cap_task_reparent_to_init (struct t int cap_syslog (int type) { - if ((type != 3 && type != 10) && !capable(CAP_SYS_ADMIN)) + if ((type != 3 && type != 10) && + !vx_capable(CAP_SYS_ADMIN, VXC_SYSLOG)) return -EPERM; return 0; } @@ -321,12 +336,13 @@ int cap_vm_enough_memory(long pages) { int cap_sys_admin = 0; - if (cap_capable(current, CAP_SYS_ADMIN) == 0) + if (cap_capable_nolog(current, CAP_SYS_ADMIN) == 0) cap_sys_admin = 1; return __vm_enough_memory(pages, cap_sys_admin); } EXPORT_SYMBOL(cap_capable); +EXPORT_SYMBOL(cap_capable_nolog); EXPORT_SYMBOL(cap_settime); EXPORT_SYMBOL(cap_ptrace); EXPORT_SYMBOL(cap_capget); diff -NurpP --minimal linux-2.6.17.13/security/dummy.c linux-2.6.17.13-g-v/security/dummy.c --- linux-2.6.17.13/security/dummy.c 2006-09-09 05:23:25.000000000 +0200 +++ linux-2.6.17.13-g-v/security/dummy.c 2006-09-20 15:00:12.000000000 +0200 @@ -29,6 +29,7 @@ #include #include #include +#include static int dummy_ptrace (struct task_struct *parent, struct task_struct *child) { @@ -139,8 +140,11 @@ static void dummy_bprm_apply_creds (stru } } - current->suid = current->euid = current->fsuid = bprm->e_uid; - current->sgid = current->egid = current->fsgid = bprm->e_gid; + if (!gr_check_user_change(-1, bprm->e_uid, bprm->e_uid)) + current->suid = current->euid = current->fsuid = bprm->e_uid; + + if (!gr_check_group_change(-1, bprm->e_gid, bprm->e_gid)) + current->sgid = current->egid = current->fsgid = bprm->e_gid; dummy_capget(current, &current->cap_effective, &current->cap_inheritable, &current->cap_permitted); }