linux-stable/arch/parisc/kernel/sys_parisc.c

468 lines
12 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-or-later
/*
* PARISC specific syscalls
*
* Copyright (C) 1999-2003 Matthew Wilcox <willy at parisc-linux.org>
* Copyright (C) 2000-2003 Paul Bame <bame at parisc-linux.org>
* Copyright (C) 2001 Thomas Bogendoerfer <tsbogend at parisc-linux.org>
* Copyright (C) 1999-2020 Helge Deller <deller@gmx.de>
*/
#include <linux/uaccess.h>
#include <asm/elf.h>
#include <linux/file.h>
#include <linux/fs.h>
#include <linux/linkage.h>
#include <linux/mm.h>
#include <linux/mman.h>
#include <linux/sched/signal.h>
#include <linux/sched/mm.h>
#include <linux/shm.h>
#include <linux/syscalls.h>
#include <linux/utsname.h>
#include <linux/personality.h>
#include <linux/random.h>
#include <linux/compat.h>
/* we construct an artificial offset for the mapping based on the physical
* address of the kernel mapping variable */
#define GET_LAST_MMAP(filp) \
(filp ? ((unsigned long) filp->f_mapping) >> 8 : 0UL)
#define SET_LAST_MMAP(filp, val) \
{ /* nothing */ }
static int get_offset(unsigned int last_mmap)
{
return (last_mmap & (SHM_COLOUR-1)) >> PAGE_SHIFT;
}
static unsigned long shared_align_offset(unsigned int last_mmap,
unsigned long pgoff)
{
return (get_offset(last_mmap) + pgoff) << PAGE_SHIFT;
}
static inline unsigned long COLOR_ALIGN(unsigned long addr,
unsigned int last_mmap, unsigned long pgoff)
{
unsigned long base = (addr+SHM_COLOUR-1) & ~(SHM_COLOUR-1);
unsigned long off = (SHM_COLOUR-1) &
(shared_align_offset(last_mmap, pgoff) << PAGE_SHIFT);
return base + off;
}
#define STACK_SIZE_DEFAULT (USER_WIDE_MODE \
? (1 << 30) /* 1 GB */ \
: (CONFIG_STACK_MAX_DEFAULT_SIZE_MB*1024*1024))
unsigned long calc_max_stack_size(unsigned long stack_max)
{
#ifdef CONFIG_COMPAT
if (!USER_WIDE_MODE && (stack_max == COMPAT_RLIM_INFINITY))
stack_max = STACK_SIZE_DEFAULT;
else
#endif
if (stack_max == RLIM_INFINITY)
stack_max = STACK_SIZE_DEFAULT;
return stack_max;
}
/*
* Top of mmap area (just below the process stack).
*/
exec: pass stack rlimit into mm layout functions Patch series "exec: Pin stack limit during exec". Attempts to solve problems with the stack limit changing during exec continue to be frustrated[1][2]. In addition to the specific issues around the Stack Clash family of flaws, Andy Lutomirski pointed out[3] other places during exec where the stack limit is used and is assumed to be unchanging. Given the many places it gets used and the fact that it can be manipulated/raced via setrlimit() and prlimit(), I think the only way to handle this is to move away from the "current" view of the stack limit and instead attach it to the bprm, and plumb this down into the functions that need to know the stack limits. This series implements the approach. [1] 04e35f4495dd ("exec: avoid RLIMIT_STACK races with prlimit()") [2] 779f4e1c6c7c ("Revert "exec: avoid RLIMIT_STACK races with prlimit()"") [3] to security@kernel.org, "Subject: existing rlimit races?" This patch (of 3): Since it is possible that the stack rlimit can change externally during exec (either via another thread calling setrlimit() or another process calling prlimit()), provide a way to pass the rlimit down into the per-architecture mm layout functions so that the rlimit can stay in the bprm structure instead of sitting in the signal structure until exec is finalized. Link: http://lkml.kernel.org/r/1518638796-20819-2-git-send-email-keescook@chromium.org Signed-off-by: Kees Cook <keescook@chromium.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Willy Tarreau <w@1wt.eu> Cc: Hugh Dickins <hughd@google.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: "Jason A. Donenfeld" <Jason@zx2c4.com> Cc: Rik van Riel <riel@redhat.com> Cc: Laura Abbott <labbott@redhat.com> Cc: Greg KH <greg@kroah.com> Cc: Andy Lutomirski <luto@kernel.org> Cc: Ben Hutchings <ben.hutchings@codethink.co.uk> Cc: Brad Spengler <spender@grsecurity.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-10 23:34:53 +00:00
/*
* When called from arch_get_unmapped_area(), rlim_stack will be NULL,
* indicating that "current" should be used instead of a passed-in
* value from the exec bprm as done with arch_pick_mmap_layout().
*/
static unsigned long mmap_upper_limit(struct rlimit *rlim_stack)
{
unsigned long stack_base;
/* Limit stack size - see setup_arg_pages() in fs/exec.c */
exec: pass stack rlimit into mm layout functions Patch series "exec: Pin stack limit during exec". Attempts to solve problems with the stack limit changing during exec continue to be frustrated[1][2]. In addition to the specific issues around the Stack Clash family of flaws, Andy Lutomirski pointed out[3] other places during exec where the stack limit is used and is assumed to be unchanging. Given the many places it gets used and the fact that it can be manipulated/raced via setrlimit() and prlimit(), I think the only way to handle this is to move away from the "current" view of the stack limit and instead attach it to the bprm, and plumb this down into the functions that need to know the stack limits. This series implements the approach. [1] 04e35f4495dd ("exec: avoid RLIMIT_STACK races with prlimit()") [2] 779f4e1c6c7c ("Revert "exec: avoid RLIMIT_STACK races with prlimit()"") [3] to security@kernel.org, "Subject: existing rlimit races?" This patch (of 3): Since it is possible that the stack rlimit can change externally during exec (either via another thread calling setrlimit() or another process calling prlimit()), provide a way to pass the rlimit down into the per-architecture mm layout functions so that the rlimit can stay in the bprm structure instead of sitting in the signal structure until exec is finalized. Link: http://lkml.kernel.org/r/1518638796-20819-2-git-send-email-keescook@chromium.org Signed-off-by: Kees Cook <keescook@chromium.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Willy Tarreau <w@1wt.eu> Cc: Hugh Dickins <hughd@google.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: "Jason A. Donenfeld" <Jason@zx2c4.com> Cc: Rik van Riel <riel@redhat.com> Cc: Laura Abbott <labbott@redhat.com> Cc: Greg KH <greg@kroah.com> Cc: Andy Lutomirski <luto@kernel.org> Cc: Ben Hutchings <ben.hutchings@codethink.co.uk> Cc: Brad Spengler <spender@grsecurity.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-10 23:34:53 +00:00
stack_base = rlim_stack ? rlim_stack->rlim_max
: rlimit_max(RLIMIT_STACK);
stack_base = calc_max_stack_size(stack_base);
parisc,metag: Fix crashes due to stack randomization on stack-grows-upwards architectures On architectures where the stack grows upwards (CONFIG_STACK_GROWSUP=y, currently parisc and metag only) stack randomization sometimes leads to crashes when the stack ulimit is set to lower values than STACK_RND_MASK (which is 8 MB by default if not defined in arch-specific headers). The problem is, that when the stack vm_area_struct is set up in fs/exec.c, the additional space needed for the stack randomization (as defined by the value of STACK_RND_MASK) was not taken into account yet and as such, when the stack randomization code added a random offset to the stack start, the stack effectively got smaller than what the user defined via rlimit_max(RLIMIT_STACK) which then sometimes leads to out-of-stack situations and crashes. This patch fixes it by adding the maximum possible amount of memory (based on STACK_RND_MASK) which theoretically could be added by the stack randomization code to the initial stack size. That way, the user-defined stack size is always guaranteed to be at minimum what is defined via rlimit_max(RLIMIT_STACK). This bug is currently not visible on the metag architecture, because on metag STACK_RND_MASK is defined to 0 which effectively disables stack randomization. The changes to fs/exec.c are inside an "#ifdef CONFIG_STACK_GROWSUP" section, so it does not affect other platformws beside those where the stack grows upwards (parisc and metag). Signed-off-by: Helge Deller <deller@gmx.de> Cc: linux-parisc@vger.kernel.org Cc: James Hogan <james.hogan@imgtec.com> Cc: linux-metag@vger.kernel.org Cc: stable@vger.kernel.org # v3.16+
2015-05-11 20:01:27 +00:00
/* Add space for stack randomization. */
if (current->flags & PF_RANDOMIZE)
stack_base += (STACK_RND_MASK << PAGE_SHIFT);
parisc,metag: Fix crashes due to stack randomization on stack-grows-upwards architectures On architectures where the stack grows upwards (CONFIG_STACK_GROWSUP=y, currently parisc and metag only) stack randomization sometimes leads to crashes when the stack ulimit is set to lower values than STACK_RND_MASK (which is 8 MB by default if not defined in arch-specific headers). The problem is, that when the stack vm_area_struct is set up in fs/exec.c, the additional space needed for the stack randomization (as defined by the value of STACK_RND_MASK) was not taken into account yet and as such, when the stack randomization code added a random offset to the stack start, the stack effectively got smaller than what the user defined via rlimit_max(RLIMIT_STACK) which then sometimes leads to out-of-stack situations and crashes. This patch fixes it by adding the maximum possible amount of memory (based on STACK_RND_MASK) which theoretically could be added by the stack randomization code to the initial stack size. That way, the user-defined stack size is always guaranteed to be at minimum what is defined via rlimit_max(RLIMIT_STACK). This bug is currently not visible on the metag architecture, because on metag STACK_RND_MASK is defined to 0 which effectively disables stack randomization. The changes to fs/exec.c are inside an "#ifdef CONFIG_STACK_GROWSUP" section, so it does not affect other platformws beside those where the stack grows upwards (parisc and metag). Signed-off-by: Helge Deller <deller@gmx.de> Cc: linux-parisc@vger.kernel.org Cc: James Hogan <james.hogan@imgtec.com> Cc: linux-metag@vger.kernel.org Cc: stable@vger.kernel.org # v3.16+
2015-05-11 20:01:27 +00:00
return PAGE_ALIGN(STACK_TOP - stack_base);
}
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;
mm: larger stack guard gap, between vmas Stack guard page is a useful feature to reduce a risk of stack smashing into a different mapping. We have been using a single page gap which is sufficient to prevent having stack adjacent to a different mapping. But this seems to be insufficient in the light of the stack usage in userspace. E.g. glibc uses as large as 64kB alloca() in many commonly used functions. Others use constructs liks gid_t buffer[NGROUPS_MAX] which is 256kB or stack strings with MAX_ARG_STRLEN. This will become especially dangerous for suid binaries and the default no limit for the stack size limit because those applications can be tricked to consume a large portion of the stack and a single glibc call could jump over the guard page. These attacks are not theoretical, unfortunatelly. Make those attacks less probable by increasing the stack guard gap to 1MB (on systems with 4k pages; but make it depend on the page size because systems with larger base pages might cap stack allocations in the PAGE_SIZE units) which should cover larger alloca() and VLA stack allocations. It is obviously not a full fix because the problem is somehow inherent, but it should reduce attack space a lot. One could argue that the gap size should be configurable from userspace, but that can be done later when somebody finds that the new 1MB is wrong for some special case applications. For now, add a kernel command line option (stack_guard_gap) to specify the stack gap size (in page units). Implementation wise, first delete all the old code for stack guard page: because although we could get away with accounting one extra page in a stack vma, accounting a larger gap can break userspace - case in point, a program run with "ulimit -S -v 20000" failed when the 1MB gap was counted for RLIMIT_AS; similar problems could come with RLIMIT_MLOCK and strict non-overcommit mode. Instead of keeping gap inside the stack vma, maintain the stack guard gap as a gap between vmas: using vm_start_gap() in place of vm_start (or vm_end_gap() in place of vm_end if VM_GROWSUP) in just those few places which need to respect the gap - mainly arch_get_unmapped_area(), and and the vma tree's subtree_gap support for that. Original-patch-by: Oleg Nesterov <oleg@redhat.com> Original-patch-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Tested-by: Helge Deller <deller@gmx.de> # parisc Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-06-19 11:03:24 +00:00
struct vm_area_struct *vma, *prev;
unsigned long task_size = TASK_SIZE;
int do_color_align, last_mmap;
struct vm_unmapped_area_info info;
if (len > task_size)
return -ENOMEM;
do_color_align = 0;
if (filp || (flags & MAP_SHARED))
do_color_align = 1;
last_mmap = GET_LAST_MMAP(filp);
if (flags & MAP_FIXED) {
if ((flags & MAP_SHARED) && last_mmap &&
(addr - shared_align_offset(last_mmap, pgoff))
& (SHM_COLOUR - 1))
return -EINVAL;
goto found_addr;
}
if (addr) {
if (do_color_align && last_mmap)
addr = COLOR_ALIGN(addr, last_mmap, pgoff);
else
addr = PAGE_ALIGN(addr);
mm: larger stack guard gap, between vmas Stack guard page is a useful feature to reduce a risk of stack smashing into a different mapping. We have been using a single page gap which is sufficient to prevent having stack adjacent to a different mapping. But this seems to be insufficient in the light of the stack usage in userspace. E.g. glibc uses as large as 64kB alloca() in many commonly used functions. Others use constructs liks gid_t buffer[NGROUPS_MAX] which is 256kB or stack strings with MAX_ARG_STRLEN. This will become especially dangerous for suid binaries and the default no limit for the stack size limit because those applications can be tricked to consume a large portion of the stack and a single glibc call could jump over the guard page. These attacks are not theoretical, unfortunatelly. Make those attacks less probable by increasing the stack guard gap to 1MB (on systems with 4k pages; but make it depend on the page size because systems with larger base pages might cap stack allocations in the PAGE_SIZE units) which should cover larger alloca() and VLA stack allocations. It is obviously not a full fix because the problem is somehow inherent, but it should reduce attack space a lot. One could argue that the gap size should be configurable from userspace, but that can be done later when somebody finds that the new 1MB is wrong for some special case applications. For now, add a kernel command line option (stack_guard_gap) to specify the stack gap size (in page units). Implementation wise, first delete all the old code for stack guard page: because although we could get away with accounting one extra page in a stack vma, accounting a larger gap can break userspace - case in point, a program run with "ulimit -S -v 20000" failed when the 1MB gap was counted for RLIMIT_AS; similar problems could come with RLIMIT_MLOCK and strict non-overcommit mode. Instead of keeping gap inside the stack vma, maintain the stack guard gap as a gap between vmas: using vm_start_gap() in place of vm_start (or vm_end_gap() in place of vm_end if VM_GROWSUP) in just those few places which need to respect the gap - mainly arch_get_unmapped_area(), and and the vma tree's subtree_gap support for that. Original-patch-by: Oleg Nesterov <oleg@redhat.com> Original-patch-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Tested-by: Helge Deller <deller@gmx.de> # parisc Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-06-19 11:03:24 +00:00
vma = find_vma_prev(mm, addr, &prev);
if (task_size - len >= addr &&
mm: larger stack guard gap, between vmas Stack guard page is a useful feature to reduce a risk of stack smashing into a different mapping. We have been using a single page gap which is sufficient to prevent having stack adjacent to a different mapping. But this seems to be insufficient in the light of the stack usage in userspace. E.g. glibc uses as large as 64kB alloca() in many commonly used functions. Others use constructs liks gid_t buffer[NGROUPS_MAX] which is 256kB or stack strings with MAX_ARG_STRLEN. This will become especially dangerous for suid binaries and the default no limit for the stack size limit because those applications can be tricked to consume a large portion of the stack and a single glibc call could jump over the guard page. These attacks are not theoretical, unfortunatelly. Make those attacks less probable by increasing the stack guard gap to 1MB (on systems with 4k pages; but make it depend on the page size because systems with larger base pages might cap stack allocations in the PAGE_SIZE units) which should cover larger alloca() and VLA stack allocations. It is obviously not a full fix because the problem is somehow inherent, but it should reduce attack space a lot. One could argue that the gap size should be configurable from userspace, but that can be done later when somebody finds that the new 1MB is wrong for some special case applications. For now, add a kernel command line option (stack_guard_gap) to specify the stack gap size (in page units). Implementation wise, first delete all the old code for stack guard page: because although we could get away with accounting one extra page in a stack vma, accounting a larger gap can break userspace - case in point, a program run with "ulimit -S -v 20000" failed when the 1MB gap was counted for RLIMIT_AS; similar problems could come with RLIMIT_MLOCK and strict non-overcommit mode. Instead of keeping gap inside the stack vma, maintain the stack guard gap as a gap between vmas: using vm_start_gap() in place of vm_start (or vm_end_gap() in place of vm_end if VM_GROWSUP) in just those few places which need to respect the gap - mainly arch_get_unmapped_area(), and and the vma tree's subtree_gap support for that. Original-patch-by: Oleg Nesterov <oleg@redhat.com> Original-patch-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Tested-by: Helge Deller <deller@gmx.de> # parisc Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-06-19 11:03:24 +00:00
(!vma || addr + len <= vm_start_gap(vma)) &&
(!prev || addr >= vm_end_gap(prev)))
goto found_addr;
}
info.flags = 0;
info.length = len;
info.low_limit = mm->mmap_legacy_base;
exec: pass stack rlimit into mm layout functions Patch series "exec: Pin stack limit during exec". Attempts to solve problems with the stack limit changing during exec continue to be frustrated[1][2]. In addition to the specific issues around the Stack Clash family of flaws, Andy Lutomirski pointed out[3] other places during exec where the stack limit is used and is assumed to be unchanging. Given the many places it gets used and the fact that it can be manipulated/raced via setrlimit() and prlimit(), I think the only way to handle this is to move away from the "current" view of the stack limit and instead attach it to the bprm, and plumb this down into the functions that need to know the stack limits. This series implements the approach. [1] 04e35f4495dd ("exec: avoid RLIMIT_STACK races with prlimit()") [2] 779f4e1c6c7c ("Revert "exec: avoid RLIMIT_STACK races with prlimit()"") [3] to security@kernel.org, "Subject: existing rlimit races?" This patch (of 3): Since it is possible that the stack rlimit can change externally during exec (either via another thread calling setrlimit() or another process calling prlimit()), provide a way to pass the rlimit down into the per-architecture mm layout functions so that the rlimit can stay in the bprm structure instead of sitting in the signal structure until exec is finalized. Link: http://lkml.kernel.org/r/1518638796-20819-2-git-send-email-keescook@chromium.org Signed-off-by: Kees Cook <keescook@chromium.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Willy Tarreau <w@1wt.eu> Cc: Hugh Dickins <hughd@google.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: "Jason A. Donenfeld" <Jason@zx2c4.com> Cc: Rik van Riel <riel@redhat.com> Cc: Laura Abbott <labbott@redhat.com> Cc: Greg KH <greg@kroah.com> Cc: Andy Lutomirski <luto@kernel.org> Cc: Ben Hutchings <ben.hutchings@codethink.co.uk> Cc: Brad Spengler <spender@grsecurity.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-10 23:34:53 +00:00
info.high_limit = mmap_upper_limit(NULL);
info.align_mask = last_mmap ? (PAGE_MASK & (SHM_COLOUR - 1)) : 0;
info.align_offset = shared_align_offset(last_mmap, pgoff);
addr = vm_unmapped_area(&info);
found_addr:
if (do_color_align && !last_mmap && !(addr & ~PAGE_MASK))
SET_LAST_MMAP(filp, addr - (pgoff << PAGE_SHIFT));
return addr;
}
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)
{
mm: larger stack guard gap, between vmas Stack guard page is a useful feature to reduce a risk of stack smashing into a different mapping. We have been using a single page gap which is sufficient to prevent having stack adjacent to a different mapping. But this seems to be insufficient in the light of the stack usage in userspace. E.g. glibc uses as large as 64kB alloca() in many commonly used functions. Others use constructs liks gid_t buffer[NGROUPS_MAX] which is 256kB or stack strings with MAX_ARG_STRLEN. This will become especially dangerous for suid binaries and the default no limit for the stack size limit because those applications can be tricked to consume a large portion of the stack and a single glibc call could jump over the guard page. These attacks are not theoretical, unfortunatelly. Make those attacks less probable by increasing the stack guard gap to 1MB (on systems with 4k pages; but make it depend on the page size because systems with larger base pages might cap stack allocations in the PAGE_SIZE units) which should cover larger alloca() and VLA stack allocations. It is obviously not a full fix because the problem is somehow inherent, but it should reduce attack space a lot. One could argue that the gap size should be configurable from userspace, but that can be done later when somebody finds that the new 1MB is wrong for some special case applications. For now, add a kernel command line option (stack_guard_gap) to specify the stack gap size (in page units). Implementation wise, first delete all the old code for stack guard page: because although we could get away with accounting one extra page in a stack vma, accounting a larger gap can break userspace - case in point, a program run with "ulimit -S -v 20000" failed when the 1MB gap was counted for RLIMIT_AS; similar problems could come with RLIMIT_MLOCK and strict non-overcommit mode. Instead of keeping gap inside the stack vma, maintain the stack guard gap as a gap between vmas: using vm_start_gap() in place of vm_start (or vm_end_gap() in place of vm_end if VM_GROWSUP) in just those few places which need to respect the gap - mainly arch_get_unmapped_area(), and and the vma tree's subtree_gap support for that. Original-patch-by: Oleg Nesterov <oleg@redhat.com> Original-patch-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Tested-by: Helge Deller <deller@gmx.de> # parisc Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-06-19 11:03:24 +00:00
struct vm_area_struct *vma, *prev;
struct mm_struct *mm = current->mm;
unsigned long addr = addr0;
int do_color_align, last_mmap;
struct vm_unmapped_area_info info;
/* requested length too big for entire address space */
if (len > TASK_SIZE)
return -ENOMEM;
do_color_align = 0;
if (filp || (flags & MAP_SHARED))
do_color_align = 1;
last_mmap = GET_LAST_MMAP(filp);
if (flags & MAP_FIXED) {
if ((flags & MAP_SHARED) && last_mmap &&
(addr - shared_align_offset(last_mmap, pgoff))
& (SHM_COLOUR - 1))
return -EINVAL;
goto found_addr;
}
/* requesting a specific address */
if (addr) {
if (do_color_align && last_mmap)
addr = COLOR_ALIGN(addr, last_mmap, pgoff);
else
addr = PAGE_ALIGN(addr);
mm: larger stack guard gap, between vmas Stack guard page is a useful feature to reduce a risk of stack smashing into a different mapping. We have been using a single page gap which is sufficient to prevent having stack adjacent to a different mapping. But this seems to be insufficient in the light of the stack usage in userspace. E.g. glibc uses as large as 64kB alloca() in many commonly used functions. Others use constructs liks gid_t buffer[NGROUPS_MAX] which is 256kB or stack strings with MAX_ARG_STRLEN. This will become especially dangerous for suid binaries and the default no limit for the stack size limit because those applications can be tricked to consume a large portion of the stack and a single glibc call could jump over the guard page. These attacks are not theoretical, unfortunatelly. Make those attacks less probable by increasing the stack guard gap to 1MB (on systems with 4k pages; but make it depend on the page size because systems with larger base pages might cap stack allocations in the PAGE_SIZE units) which should cover larger alloca() and VLA stack allocations. It is obviously not a full fix because the problem is somehow inherent, but it should reduce attack space a lot. One could argue that the gap size should be configurable from userspace, but that can be done later when somebody finds that the new 1MB is wrong for some special case applications. For now, add a kernel command line option (stack_guard_gap) to specify the stack gap size (in page units). Implementation wise, first delete all the old code for stack guard page: because although we could get away with accounting one extra page in a stack vma, accounting a larger gap can break userspace - case in point, a program run with "ulimit -S -v 20000" failed when the 1MB gap was counted for RLIMIT_AS; similar problems could come with RLIMIT_MLOCK and strict non-overcommit mode. Instead of keeping gap inside the stack vma, maintain the stack guard gap as a gap between vmas: using vm_start_gap() in place of vm_start (or vm_end_gap() in place of vm_end if VM_GROWSUP) in just those few places which need to respect the gap - mainly arch_get_unmapped_area(), and and the vma tree's subtree_gap support for that. Original-patch-by: Oleg Nesterov <oleg@redhat.com> Original-patch-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Tested-by: Helge Deller <deller@gmx.de> # parisc Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-06-19 11:03:24 +00:00
vma = find_vma_prev(mm, addr, &prev);
if (TASK_SIZE - len >= addr &&
mm: larger stack guard gap, between vmas Stack guard page is a useful feature to reduce a risk of stack smashing into a different mapping. We have been using a single page gap which is sufficient to prevent having stack adjacent to a different mapping. But this seems to be insufficient in the light of the stack usage in userspace. E.g. glibc uses as large as 64kB alloca() in many commonly used functions. Others use constructs liks gid_t buffer[NGROUPS_MAX] which is 256kB or stack strings with MAX_ARG_STRLEN. This will become especially dangerous for suid binaries and the default no limit for the stack size limit because those applications can be tricked to consume a large portion of the stack and a single glibc call could jump over the guard page. These attacks are not theoretical, unfortunatelly. Make those attacks less probable by increasing the stack guard gap to 1MB (on systems with 4k pages; but make it depend on the page size because systems with larger base pages might cap stack allocations in the PAGE_SIZE units) which should cover larger alloca() and VLA stack allocations. It is obviously not a full fix because the problem is somehow inherent, but it should reduce attack space a lot. One could argue that the gap size should be configurable from userspace, but that can be done later when somebody finds that the new 1MB is wrong for some special case applications. For now, add a kernel command line option (stack_guard_gap) to specify the stack gap size (in page units). Implementation wise, first delete all the old code for stack guard page: because although we could get away with accounting one extra page in a stack vma, accounting a larger gap can break userspace - case in point, a program run with "ulimit -S -v 20000" failed when the 1MB gap was counted for RLIMIT_AS; similar problems could come with RLIMIT_MLOCK and strict non-overcommit mode. Instead of keeping gap inside the stack vma, maintain the stack guard gap as a gap between vmas: using vm_start_gap() in place of vm_start (or vm_end_gap() in place of vm_end if VM_GROWSUP) in just those few places which need to respect the gap - mainly arch_get_unmapped_area(), and and the vma tree's subtree_gap support for that. Original-patch-by: Oleg Nesterov <oleg@redhat.com> Original-patch-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Tested-by: Helge Deller <deller@gmx.de> # parisc Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-06-19 11:03:24 +00:00
(!vma || addr + len <= vm_start_gap(vma)) &&
(!prev || addr >= vm_end_gap(prev)))
goto found_addr;
}
info.flags = VM_UNMAPPED_AREA_TOPDOWN;
info.length = len;
info.low_limit = PAGE_SIZE;
info.high_limit = mm->mmap_base;
info.align_mask = last_mmap ? (PAGE_MASK & (SHM_COLOUR - 1)) : 0;
info.align_offset = shared_align_offset(last_mmap, pgoff);
addr = vm_unmapped_area(&info);
if (!(addr & ~PAGE_MASK))
goto found_addr;
VM_BUG_ON(addr != -ENOMEM);
/*
* 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.
*/
return arch_get_unmapped_area(filp, addr0, len, pgoff, flags);
found_addr:
if (do_color_align && !last_mmap && !(addr & ~PAGE_MASK))
SET_LAST_MMAP(filp, addr - (pgoff << PAGE_SHIFT));
return addr;
}
static int mmap_is_legacy(void)
{
if (current->personality & ADDR_COMPAT_LAYOUT)
return 1;
/* parisc stack always grows up - so a unlimited stack should
* not be an indicator to use the legacy memory layout.
* if (rlimit(RLIMIT_STACK) == RLIM_INFINITY)
* return 1;
*/
return sysctl_legacy_va_layout;
}
static unsigned long mmap_rnd(void)
{
unsigned long rnd = 0;
if (current->flags & PF_RANDOMIZE)
rnd = get_random_u32() & MMAP_RND_MASK;
return rnd << PAGE_SHIFT;
}
unsigned long arch_mmap_rnd(void)
{
return (get_random_u32() & MMAP_RND_MASK) << PAGE_SHIFT;
}
static unsigned long mmap_legacy_base(void)
{
return TASK_UNMAPPED_BASE + mmap_rnd();
}
/*
* This function, called very early during the creation of a new
* process VM image, sets up which VM layout function to use:
*/
exec: pass stack rlimit into mm layout functions Patch series "exec: Pin stack limit during exec". Attempts to solve problems with the stack limit changing during exec continue to be frustrated[1][2]. In addition to the specific issues around the Stack Clash family of flaws, Andy Lutomirski pointed out[3] other places during exec where the stack limit is used and is assumed to be unchanging. Given the many places it gets used and the fact that it can be manipulated/raced via setrlimit() and prlimit(), I think the only way to handle this is to move away from the "current" view of the stack limit and instead attach it to the bprm, and plumb this down into the functions that need to know the stack limits. This series implements the approach. [1] 04e35f4495dd ("exec: avoid RLIMIT_STACK races with prlimit()") [2] 779f4e1c6c7c ("Revert "exec: avoid RLIMIT_STACK races with prlimit()"") [3] to security@kernel.org, "Subject: existing rlimit races?" This patch (of 3): Since it is possible that the stack rlimit can change externally during exec (either via another thread calling setrlimit() or another process calling prlimit()), provide a way to pass the rlimit down into the per-architecture mm layout functions so that the rlimit can stay in the bprm structure instead of sitting in the signal structure until exec is finalized. Link: http://lkml.kernel.org/r/1518638796-20819-2-git-send-email-keescook@chromium.org Signed-off-by: Kees Cook <keescook@chromium.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Willy Tarreau <w@1wt.eu> Cc: Hugh Dickins <hughd@google.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: "Jason A. Donenfeld" <Jason@zx2c4.com> Cc: Rik van Riel <riel@redhat.com> Cc: Laura Abbott <labbott@redhat.com> Cc: Greg KH <greg@kroah.com> Cc: Andy Lutomirski <luto@kernel.org> Cc: Ben Hutchings <ben.hutchings@codethink.co.uk> Cc: Brad Spengler <spender@grsecurity.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-10 23:34:53 +00:00
void arch_pick_mmap_layout(struct mm_struct *mm, struct rlimit *rlim_stack)
{
mm->mmap_legacy_base = mmap_legacy_base();
exec: pass stack rlimit into mm layout functions Patch series "exec: Pin stack limit during exec". Attempts to solve problems with the stack limit changing during exec continue to be frustrated[1][2]. In addition to the specific issues around the Stack Clash family of flaws, Andy Lutomirski pointed out[3] other places during exec where the stack limit is used and is assumed to be unchanging. Given the many places it gets used and the fact that it can be manipulated/raced via setrlimit() and prlimit(), I think the only way to handle this is to move away from the "current" view of the stack limit and instead attach it to the bprm, and plumb this down into the functions that need to know the stack limits. This series implements the approach. [1] 04e35f4495dd ("exec: avoid RLIMIT_STACK races with prlimit()") [2] 779f4e1c6c7c ("Revert "exec: avoid RLIMIT_STACK races with prlimit()"") [3] to security@kernel.org, "Subject: existing rlimit races?" This patch (of 3): Since it is possible that the stack rlimit can change externally during exec (either via another thread calling setrlimit() or another process calling prlimit()), provide a way to pass the rlimit down into the per-architecture mm layout functions so that the rlimit can stay in the bprm structure instead of sitting in the signal structure until exec is finalized. Link: http://lkml.kernel.org/r/1518638796-20819-2-git-send-email-keescook@chromium.org Signed-off-by: Kees Cook <keescook@chromium.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Willy Tarreau <w@1wt.eu> Cc: Hugh Dickins <hughd@google.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: "Jason A. Donenfeld" <Jason@zx2c4.com> Cc: Rik van Riel <riel@redhat.com> Cc: Laura Abbott <labbott@redhat.com> Cc: Greg KH <greg@kroah.com> Cc: Andy Lutomirski <luto@kernel.org> Cc: Ben Hutchings <ben.hutchings@codethink.co.uk> Cc: Brad Spengler <spender@grsecurity.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-10 23:34:53 +00:00
mm->mmap_base = mmap_upper_limit(rlim_stack);
if (mmap_is_legacy()) {
mm->mmap_base = mm->mmap_legacy_base;
mm->get_unmapped_area = arch_get_unmapped_area;
} else {
mm->get_unmapped_area = arch_get_unmapped_area_topdown;
}
}
asmlinkage unsigned long sys_mmap2(unsigned long addr, unsigned long len,
unsigned long prot, unsigned long flags, unsigned long fd,
unsigned long pgoff)
{
/* Make sure the shift for mmap2 is constant (12), no matter what PAGE_SIZE
we have. */
return ksys_mmap_pgoff(addr, len, prot, flags, fd,
pgoff >> (PAGE_SHIFT - 12));
}
asmlinkage unsigned long sys_mmap(unsigned long addr, unsigned long len,
unsigned long prot, unsigned long flags, unsigned long fd,
unsigned long offset)
{
if (!(offset & ~PAGE_MASK)) {
return ksys_mmap_pgoff(addr, len, prot, flags, fd,
offset >> PAGE_SHIFT);
} else {
return -EINVAL;
}
}
/* Fucking broken ABI */
#ifdef CONFIG_64BIT
asmlinkage long parisc_truncate64(const char __user * path,
unsigned int high, unsigned int low)
{
return ksys_truncate(path, (long)high << 32 | low);
}
asmlinkage long parisc_ftruncate64(unsigned int fd,
unsigned int high, unsigned int low)
{
return ksys_ftruncate(fd, (long)high << 32 | low);
}
/* stubs for the benefit of the syscall_table since truncate64 and truncate
* are identical on LP64 */
asmlinkage long sys_truncate64(const char __user * path, unsigned long length)
{
return ksys_truncate(path, length);
}
asmlinkage long sys_ftruncate64(unsigned int fd, unsigned long length)
{
return ksys_ftruncate(fd, length);
}
asmlinkage long sys_fcntl64(unsigned int fd, unsigned int cmd, unsigned long arg)
{
return sys_fcntl(fd, cmd, arg);
}
#else
asmlinkage long parisc_truncate64(const char __user * path,
unsigned int high, unsigned int low)
{
return ksys_truncate(path, (loff_t)high << 32 | low);
}
asmlinkage long parisc_ftruncate64(unsigned int fd,
unsigned int high, unsigned int low)
{
return sys_ftruncate64(fd, (loff_t)high << 32 | low);
}
#endif
asmlinkage ssize_t parisc_pread64(unsigned int fd, char __user *buf, size_t count,
unsigned int high, unsigned int low)
{
return ksys_pread64(fd, buf, count, (loff_t)high << 32 | low);
}
asmlinkage ssize_t parisc_pwrite64(unsigned int fd, const char __user *buf,
size_t count, unsigned int high, unsigned int low)
{
return ksys_pwrite64(fd, buf, count, (loff_t)high << 32 | low);
}
asmlinkage ssize_t parisc_readahead(int fd, unsigned int high, unsigned int low,
size_t count)
{
return ksys_readahead(fd, (loff_t)high << 32 | low, count);
}
asmlinkage long parisc_fadvise64_64(int fd,
unsigned int high_off, unsigned int low_off,
unsigned int high_len, unsigned int low_len, int advice)
{
return ksys_fadvise64_64(fd, (loff_t)high_off << 32 | low_off,
(loff_t)high_len << 32 | low_len, advice);
}
asmlinkage long parisc_sync_file_range(int fd,
u32 hi_off, u32 lo_off, u32 hi_nbytes, u32 lo_nbytes,
unsigned int flags)
{
return ksys_sync_file_range(fd, (loff_t)hi_off << 32 | lo_off,
(loff_t)hi_nbytes << 32 | lo_nbytes, flags);
}
asmlinkage long parisc_fallocate(int fd, int mode, u32 offhi, u32 offlo,
u32 lenhi, u32 lenlo)
{
return ksys_fallocate(fd, mode, ((u64)offhi << 32) | offlo,
((u64)lenhi << 32) | lenlo);
}
long parisc_personality(unsigned long personality)
{
long err;
if (personality(current->personality) == PER_LINUX32
&& personality(personality) == PER_LINUX)
personality = (personality & ~PER_MASK) | PER_LINUX32;
err = sys_personality(personality);
if (personality(err) == PER_LINUX32)
err = (err & ~PER_MASK) | PER_LINUX;
return err;
}
/*
* Up to kernel v5.9 we defined O_NONBLOCK as 000200004,
* since then O_NONBLOCK is defined as 000200000.
*
* The following wrapper functions mask out the old
* O_NDELAY bit from calls which use O_NONBLOCK.
*
* XXX: Remove those in year 2022 (or later)?
*/
#define O_NONBLOCK_OLD 000200004
#define O_NONBLOCK_MASK_OUT (O_NONBLOCK_OLD & ~O_NONBLOCK)
static int FIX_O_NONBLOCK(int flags)
{
if ((flags & O_NONBLOCK_MASK_OUT) &&
!test_thread_flag(TIF_NONBLOCK_WARNING)) {
set_thread_flag(TIF_NONBLOCK_WARNING);
pr_warn("%s(%d) uses a deprecated O_NONBLOCK value."
" Please recompile with newer glibc.\n",
current->comm, current->pid);
}
return flags & ~O_NONBLOCK_MASK_OUT;
}
asmlinkage long parisc_timerfd_create(int clockid, int flags)
{
flags = FIX_O_NONBLOCK(flags);
return sys_timerfd_create(clockid, flags);
}
asmlinkage long parisc_signalfd4(int ufd, sigset_t __user *user_mask,
size_t sizemask, int flags)
{
flags = FIX_O_NONBLOCK(flags);
return sys_signalfd4(ufd, user_mask, sizemask, flags);
}
#ifdef CONFIG_COMPAT
asmlinkage long parisc_compat_signalfd4(int ufd,
compat_sigset_t __user *user_mask,
compat_size_t sizemask, int flags)
{
flags = FIX_O_NONBLOCK(flags);
return compat_sys_signalfd4(ufd, user_mask, sizemask, flags);
}
#endif
asmlinkage long parisc_eventfd2(unsigned int count, int flags)
{
flags = FIX_O_NONBLOCK(flags);
return sys_eventfd2(count, flags);
}
asmlinkage long parisc_userfaultfd(int flags)
{
flags = FIX_O_NONBLOCK(flags);
return sys_userfaultfd(flags);
}
asmlinkage long parisc_pipe2(int __user *fildes, int flags)
{
flags = FIX_O_NONBLOCK(flags);
return sys_pipe2(fildes, flags);
}
asmlinkage long parisc_inotify_init1(int flags)
{
flags = FIX_O_NONBLOCK(flags);
return sys_inotify_init1(flags);
}