mm: larger stack guard gap, between vmas

commit 1be7107fbe18eed3e319a6c3e83c78254b693acb upstream. 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> [wt: backport to 4.11: adjust context] [wt: backport to 4.9: adjust context ; kernel doc was not in admin-guide] Signed-off-by: Willy Tarreau <w@1wt.eu> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mm: larger stack guard gap, between vmas
commit 1be7107fbe18eed3e319a6c3e83c78254b693acb upstream. 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> [wt: backport to 4.11: adjust context] [wt: backport to 4.9: adjust context ; kernel doc was not in admin-guide] Signed-off-by: Willy Tarreau <w@1wt.eu> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Hugh Dickins · Greg Kroah-Hartman
1 parent 04651048c7
Showing 23 changed files with 152 additions and 163 deletions Side-by-side Diff
Documentation/kernel-parameters.txt
arch/arc/mm/mmap.c
arch/arm/mm/mmap.c
arch/frv/mm/elf-fdpic.c
arch/mips/mm/mmap.c
arch/parisc/kernel/sys_parisc.c
arch/powerpc/mm/hugetlbpage-radix.c
arch/powerpc/mm/mmap.c
arch/powerpc/mm/slice.c
arch/s390/mm/mmap.c
arch/sh/mm/mmap.c
arch/sparc/kernel/sys_sparc_64.c
arch/sparc/mm/hugetlbpage.c
arch/tile/mm/hugetlbpage.c
arch/x86/kernel/sys_x86_64.c
arch/x86/mm/hugetlbpage.c
arch/xtensa/kernel/syscall.c
fs/hugetlbfs/inode.c
fs/proc/task_mmu.c
include/linux/mm.h
@@ -3932,6 +3932,13 @@
 	spia_pedr=
 	spia_peddr=
  
+	stack_guard_gap=	[MM]
+			override the default stack gap protection. The value
+			is in page units and it defines how many pages prior
+			to (for stacks growing down) resp. after (for stacks
+			growing up) the main stack are reserved for no other
+			mapping. Default value is 256 pages.
+
 	stacktrace	[FTRACE]
 			Enabled the stack tracer on boot up.
  
@@ -64,7 +64,7 @@
  
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -89,7 +89,7 @@
  
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -140,7 +140,7 @@
 			addr = PAGE_ALIGN(addr);
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr &&
-				(!vma || addr + len <= vma->vm_start))
+				(!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -74,7 +74,7 @@
 		addr = PAGE_ALIGN(addr);
 		vma = find_vma(current->mm, addr);
 		if (TASK_SIZE - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			goto success;
 	}
  
@@ -92,7 +92,7 @@
  
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -88,7 +88,7 @@
 		unsigned long len, unsigned long pgoff, unsigned long flags)
 {
 	struct mm_struct *mm = current->mm;
-	struct vm_area_struct *vma;
+	struct vm_area_struct *vma, *prev;
 	unsigned long task_size = TASK_SIZE;
 	int do_color_align, last_mmap;
 	struct vm_unmapped_area_info info;
  
@@ -115,9 +115,10 @@
 		else
 			addr = PAGE_ALIGN(addr);
  
-		vma = find_vma(mm, addr);
+		vma = find_vma_prev(mm, addr, &prev);
 		if (task_size - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)) &&
+		    (!prev || addr >= vm_end_gap(prev)))
 			goto found_addr;
 	}
  
@@ -141,7 +142,7 @@
 			  const unsigned long len, const unsigned long pgoff,
 			  const unsigned long flags)
 {
-	struct vm_area_struct *vma;
+	struct vm_area_struct *vma, *prev;
 	struct mm_struct *mm = current->mm;
 	unsigned long addr = addr0;
 	int do_color_align, last_mmap;
  
@@ -175,9 +176,11 @@
 			addr = COLOR_ALIGN(addr, last_mmap, pgoff);
 		else
 			addr = PAGE_ALIGN(addr);
-		vma = find_vma(mm, addr);
+
+		vma = find_vma_prev(mm, addr, &prev);
 		if (TASK_SIZE - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)) &&
+		    (!prev || addr >= vm_end_gap(prev)))
 			goto found_addr;
 	}
  
@@ -65,7 +65,7 @@
 		addr = ALIGN(addr, huge_page_size(h));
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
 	/*
@@ -106,7 +106,7 @@
 		addr = PAGE_ALIGN(addr);
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -142,7 +142,7 @@
 		addr = PAGE_ALIGN(addr);
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
-				(!vma || addr + len <= vma->vm_start))
+				(!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -105,7 +105,7 @@
 	if ((mm->task_size - len) < addr)
 		return 0;
 	vma = find_vma(mm, addr);
-	return (!vma || (addr + len) <= vma->vm_start);
+	return (!vma || (addr + len) <= vm_start_gap(vma));
 }
  
 static int slice_low_has_vma(struct mm_struct *mm, unsigned long slice)
@@ -98,7 +98,7 @@
 		addr = PAGE_ALIGN(addr);
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -136,7 +136,7 @@
 		addr = PAGE_ALIGN(addr);
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
-				(!vma || addr + len <= vma->vm_start))
+				(!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -63,7 +63,7 @@
  
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -113,7 +113,7 @@
  
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -118,7 +118,7 @@
  
 		vma = find_vma(mm, addr);
 		if (task_size - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -181,7 +181,7 @@
  
 		vma = find_vma(mm, addr);
 		if (task_size - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -116,7 +116,7 @@
 		addr = ALIGN(addr, HPAGE_SIZE);
 		vma = find_vma(mm, addr);
 		if (task_size - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
 	if (mm->get_unmapped_area == arch_get_unmapped_area)
@@ -232,7 +232,7 @@
 		addr = ALIGN(addr, huge_page_size(h));
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
 	if (current->mm->get_unmapped_area == arch_get_unmapped_area)
@@ -140,7 +140,7 @@
 		addr = PAGE_ALIGN(addr);
 		vma = find_vma(mm, addr);
 		if (end - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -183,7 +183,7 @@
 		addr = PAGE_ALIGN(addr);
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr &&
-				(!vma || addr + len <= vma->vm_start))
+				(!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -144,7 +144,7 @@
 		addr = ALIGN(addr, huge_page_size(h));
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
 	if (mm->get_unmapped_area == arch_get_unmapped_area)
@@ -87,7 +87,7 @@
 		/* At this point:  (!vmm || addr < vmm->vm_end). */
 		if (TASK_SIZE - len < addr)
 			return -ENOMEM;
-		if (!vmm || addr + len <= vmm->vm_start)
+		if (!vmm || addr + len <= vm_start_gap(vmm))
 			return addr;
 		addr = vmm->vm_end;
 		if (flags & MAP_SHARED)
@@ -191,7 +191,7 @@
 		addr = ALIGN(addr, huge_page_size(h));
 		vma = find_vma(mm, addr);
 		if (TASK_SIZE - len >= addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
  
@@ -299,11 +299,7 @@
  
 	/* We don't show the stack guard page in /proc/maps */
 	start = vma->vm_start;
-	if (stack_guard_page_start(vma, start))
-		start += PAGE_SIZE;
 	end = vma->vm_end;
-	if (stack_guard_page_end(vma, end))
-		end -= PAGE_SIZE;
  
 	seq_setwidth(m, 25 + sizeof(void *) * 6 - 1);
 	seq_printf(m, "%08lx-%08lx %c%c%c%c %08llx %02x:%02x %lu ",
@@ -1356,39 +1356,11 @@
  
 int get_cmdline(struct task_struct *task, char *buffer, int buflen);
  
-/* Is the vma a continuation of the stack vma above it? */
-static inline int vma_growsdown(struct vm_area_struct *vma, unsigned long addr)
-{
-	return vma && (vma->vm_end == addr) && (vma->vm_flags & VM_GROWSDOWN);
-}
-
 static inline bool vma_is_anonymous(struct vm_area_struct *vma)
 {
 	return !vma->vm_ops;
 }
  
-static inline int stack_guard_page_start(struct vm_area_struct *vma,
-					     unsigned long addr)
-{
-	return (vma->vm_flags & VM_GROWSDOWN) &&
-		(vma->vm_start == addr) &&
-		!vma_growsdown(vma->vm_prev, addr);
-}
-
-/* Is the vma a continuation of the stack vma below it? */
-static inline int vma_growsup(struct vm_area_struct *vma, unsigned long addr)
-{
-	return vma && (vma->vm_start == addr) && (vma->vm_flags & VM_GROWSUP);
-}
-
-static inline int stack_guard_page_end(struct vm_area_struct *vma,
-					   unsigned long addr)
-{
-	return (vma->vm_flags & VM_GROWSUP) &&
-		(vma->vm_end == addr) &&
-		!vma_growsup(vma->vm_next, addr);
-}
-
 int vma_is_stack_for_current(struct vm_area_struct *vma);
  
 extern unsigned long move_page_tables(struct vm_area_struct *vma,
@@ -2127,6 +2099,7 @@
 				pgoff_t offset,
 				unsigned long size);
  
+extern unsigned long stack_guard_gap;
 /* Generic expand stack which grows the stack according to GROWS{UP,DOWN} */
 extern int expand_stack(struct vm_area_struct *vma, unsigned long address);
  
@@ -2153,6 +2126,30 @@
 	if (vma && end_addr <= vma->vm_start)
 		vma = NULL;
 	return vma;
+}
+
+static inline unsigned long vm_start_gap(struct vm_area_struct *vma)
+{
+	unsigned long vm_start = vma->vm_start;
+
+	if (vma->vm_flags & VM_GROWSDOWN) {
+		vm_start -= stack_guard_gap;
+		if (vm_start > vma->vm_start)
+			vm_start = 0;
+	}
+	return vm_start;
+}
+
+static inline unsigned long vm_end_gap(struct vm_area_struct *vma)
+{
+	unsigned long vm_end = vma->vm_end;
+
+	if (vma->vm_flags & VM_GROWSUP) {
+		vm_end += stack_guard_gap;
+		if (vm_end < vma->vm_end)
+			vm_end = -PAGE_SIZE;
+	}
+	return vm_end;
 }
  
 static inline unsigned long vma_pages(struct vm_area_struct *vma)
@@ -370,11 +370,6 @@
 	/* mlock all present pages, but do not fault in new pages */
 	if ((*flags & (FOLL_POPULATE | FOLL_MLOCK)) == FOLL_MLOCK)
 		return -ENOENT;
-	/* For mm_populate(), just skip the stack guard page. */
-	if ((*flags & FOLL_POPULATE) &&
-			(stack_guard_page_start(vma, address) ||
-			 stack_guard_page_end(vma, address + PAGE_SIZE)))
-		return -ENOENT;
 	if (*flags & FOLL_WRITE)
 		fault_flags |= FAULT_FLAG_WRITE;
 	if (*flags & FOLL_REMOTE)
@@ -2699,40 +2699,6 @@
 }
  
 /*
- * This is like a special single-page "expand_{down|up}wards()",
- * except we must first make sure that 'address{-|+}PAGE_SIZE'
- * doesn't hit another vma.
- */
-static inline int check_stack_guard_page(struct vm_area_struct *vma, unsigned long address)
-{
-	address &= PAGE_MASK;
-	if ((vma->vm_flags & VM_GROWSDOWN) && address == vma->vm_start) {
-		struct vm_area_struct *prev = vma->vm_prev;
-
-		/*
-		 * Is there a mapping abutting this one below?
-		 *
-		 * That's only ok if it's the same stack mapping
-		 * that has gotten split..
-		 */
-		if (prev && prev->vm_end == address)
-			return prev->vm_flags & VM_GROWSDOWN ? 0 : -ENOMEM;
-
-		return expand_downwards(vma, address - PAGE_SIZE);
-	}
-	if ((vma->vm_flags & VM_GROWSUP) && address + PAGE_SIZE == vma->vm_end) {
-		struct vm_area_struct *next = vma->vm_next;
-
-		/* As VM_GROWSDOWN but s/below/above/ */
-		if (next && next->vm_start == address + PAGE_SIZE)
-			return next->vm_flags & VM_GROWSUP ? 0 : -ENOMEM;
-
-		return expand_upwards(vma, address + PAGE_SIZE);
-	}
-	return 0;
-}
-
-/*
  * We enter with non-exclusive mmap_sem (to exclude vma changes,
  * but allow concurrent faults), and pte mapped but not yet locked.
  * We return with mmap_sem still held, but pte unmapped and unlocked.
@@ -2747,10 +2713,6 @@
 	/* File mapping without ->vm_ops ? */
 	if (vma->vm_flags & VM_SHARED)
 		return VM_FAULT_SIGBUS;
-
-	/* Check if we need to add a guard page to the stack */
-	if (check_stack_guard_page(vma, fe->address) < 0)
-		return VM_FAULT_SIGSEGV;
  
 	/*
 	 * Use pte_alloc() instead of pte_alloc_map().  We can't run
@@ -183,6 +183,7 @@
 	unsigned long retval;
 	unsigned long newbrk, oldbrk;
 	struct mm_struct *mm = current->mm;
+	struct vm_area_struct *next;
 	unsigned long min_brk;
 	bool populate;
  
@@ -228,7 +229,8 @@
 	}
  
 	/* Check against existing mmap mappings. */
-	if (find_vma_intersection(mm, oldbrk, newbrk+PAGE_SIZE))
+	next = find_vma(mm, oldbrk);
+	if (next && newbrk + PAGE_SIZE > vm_start_gap(next))
 		goto out;
  
 	/* Ok, looks good - let it rip. */
@@ -251,10 +253,22 @@
  
 static long vma_compute_subtree_gap(struct vm_area_struct *vma)
 {
-	unsigned long max, subtree_gap;
-	max = vma->vm_start;
-	if (vma->vm_prev)
-		max -= vma->vm_prev->vm_end;
+	unsigned long max, prev_end, subtree_gap;
+
+	/*
+	 * Note: in the rare case of a VM_GROWSDOWN above a VM_GROWSUP, we
+	 * allow two stack_guard_gaps between them here, and when choosing
+	 * an unmapped area; whereas when expanding we only require one.
+	 * That's a little inconsistent, but keeps the code here simpler.
+	 */
+	max = vm_start_gap(vma);
+	if (vma->vm_prev) {
+		prev_end = vm_end_gap(vma->vm_prev);
+		if (max > prev_end)
+			max -= prev_end;
+		else
+			max = 0;
+	}
 	if (vma->vm_rb.rb_left) {
 		subtree_gap = rb_entry(vma->vm_rb.rb_left,
 				struct vm_area_struct, vm_rb)->rb_subtree_gap;
@@ -350,7 +364,7 @@
 			anon_vma_unlock_read(anon_vma);
 		}
  
-		highest_address = vma->vm_end;
+		highest_address = vm_end_gap(vma);
 		vma = vma->vm_next;
 		i++;
 	}
@@ -539,7 +553,7 @@
 	if (vma->vm_next)
 		vma_gap_update(vma->vm_next);
 	else
-		mm->highest_vm_end = vma->vm_end;
+		mm->highest_vm_end = vm_end_gap(vma);
  
 	/*
 	 * vma->vm_prev wasn't known when we followed the rbtree to find the
@@ -854,7 +868,7 @@
 			vma_gap_update(vma);
 		if (end_changed) {
 			if (!next)
-				mm->highest_vm_end = end;
+				mm->highest_vm_end = vm_end_gap(vma);
 			else if (!adjust_next)
 				vma_gap_update(next);
 		}
@@ -939,7 +953,7 @@
 			 * mm->highest_vm_end doesn't need any update
 			 * in remove_next == 1 case.
 			 */
-			VM_WARN_ON(mm->highest_vm_end != end);
+			VM_WARN_ON(mm->highest_vm_end != vm_end_gap(vma));
 		}
 	}
 	if (insert && file)
@@ -1783,7 +1797,7 @@
  
 	while (true) {
 		/* Visit left subtree if it looks promising */
-		gap_end = vma->vm_start;
+		gap_end = vm_start_gap(vma);
 		if (gap_end >= low_limit && vma->vm_rb.rb_left) {
 			struct vm_area_struct *left =
 				rb_entry(vma->vm_rb.rb_left,
@@ -1794,7 +1808,7 @@
 			}
 		}
  
-		gap_start = vma->vm_prev ? vma->vm_prev->vm_end : 0;
+		gap_start = vma->vm_prev ? vm_end_gap(vma->vm_prev) : 0;
 check_current:
 		/* Check if current node has a suitable gap */
 		if (gap_start > high_limit)
@@ -1821,8 +1835,8 @@
 			vma = rb_entry(rb_parent(prev),
 				       struct vm_area_struct, vm_rb);
 			if (prev == vma->vm_rb.rb_left) {
-				gap_start = vma->vm_prev->vm_end;
-				gap_end = vma->vm_start;
+				gap_start = vm_end_gap(vma->vm_prev);
+				gap_end = vm_start_gap(vma);
 				goto check_current;
 			}
 		}
@@ -1886,7 +1900,7 @@
  
 	while (true) {
 		/* Visit right subtree if it looks promising */
-		gap_start = vma->vm_prev ? vma->vm_prev->vm_end : 0;
+		gap_start = vma->vm_prev ? vm_end_gap(vma->vm_prev) : 0;
 		if (gap_start <= high_limit && vma->vm_rb.rb_right) {
 			struct vm_area_struct *right =
 				rb_entry(vma->vm_rb.rb_right,
@@ -1899,7 +1913,7 @@
  
 check_current:
 		/* Check if current node has a suitable gap */
-		gap_end = vma->vm_start;
+		gap_end = vm_start_gap(vma);
 		if (gap_end < low_limit)
 			return -ENOMEM;
 		if (gap_start <= high_limit && gap_end - gap_start >= length)
@@ -1925,7 +1939,7 @@
 				       struct vm_area_struct, vm_rb);
 			if (prev == vma->vm_rb.rb_right) {
 				gap_start = vma->vm_prev ?
-					vma->vm_prev->vm_end : 0;
+					vm_end_gap(vma->vm_prev) : 0;
 				goto check_current;
 			}
 		}
@@ -1963,7 +1977,7 @@
 		unsigned long len, unsigned long pgoff, unsigned long flags)
 {
 	struct mm_struct *mm = current->mm;
-	struct vm_area_struct *vma;
+	struct vm_area_struct *vma, *prev;
 	struct vm_unmapped_area_info info;
  
 	if (len > TASK_SIZE - mmap_min_addr)
  
@@ -1974,9 +1988,10 @@
  
 	if (addr) {
 		addr = PAGE_ALIGN(addr);
-		vma = find_vma(mm, addr);
+		vma = find_vma_prev(mm, addr, &prev);
 		if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
-		    (!vma || addr + len <= vma->vm_start))
+		    (!vma || addr + len <= vm_start_gap(vma)) &&
+		    (!prev || addr >= vm_end_gap(prev)))
 			return addr;
 	}
  
@@ -1999,7 +2014,7 @@
 			  const unsigned long len, const unsigned long pgoff,
 			  const unsigned long flags)
 {
-	struct vm_area_struct *vma;
+	struct vm_area_struct *vma, *prev;
 	struct mm_struct *mm = current->mm;
 	unsigned long addr = addr0;
 	struct vm_unmapped_area_info info;
  
@@ -2014,9 +2029,10 @@
 	/* requesting a specific address */
 	if (addr) {
 		addr = PAGE_ALIGN(addr);
-		vma = find_vma(mm, addr);
+		vma = find_vma_prev(mm, addr, &prev);
 		if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
-				(!vma || addr + len <= vma->vm_start))
+				(!vma || addr + len <= vm_start_gap(vma)) &&
+				(!prev || addr >= vm_end_gap(prev)))
 			return addr;
 	}
  
  
  
@@ -2151,21 +2167,19 @@
  * update accounting. This is shared with both the
  * grow-up and grow-down cases.
  */
-static int acct_stack_growth(struct vm_area_struct *vma, unsigned long size, unsigned long grow)
+static int acct_stack_growth(struct vm_area_struct *vma,
+			     unsigned long size, unsigned long grow)
 {
 	struct mm_struct *mm = vma->vm_mm;
 	struct rlimit *rlim = current->signal->rlim;
-	unsigned long new_start, actual_size;
+	unsigned long new_start;
  
 	/* address space limit tests */
 	if (!may_expand_vm(mm, vma->vm_flags, grow))
 		return -ENOMEM;
  
 	/* Stack limit test */
-	actual_size = size;
-	if (size && (vma->vm_flags & (VM_GROWSUP | VM_GROWSDOWN)))
-		actual_size -= PAGE_SIZE;
-	if (actual_size > READ_ONCE(rlim[RLIMIT_STACK].rlim_cur))
+	if (size > READ_ONCE(rlim[RLIMIT_STACK].rlim_cur))
 		return -ENOMEM;
  
 	/* mlock limit tests */
  
  
@@ -2203,17 +2217,30 @@
 int expand_upwards(struct vm_area_struct *vma, unsigned long address)
 {
 	struct mm_struct *mm = vma->vm_mm;
+	struct vm_area_struct *next;
+	unsigned long gap_addr;
 	int error = 0;
  
 	if (!(vma->vm_flags & VM_GROWSUP))
 		return -EFAULT;
  
 	/* Guard against wrapping around to address 0. */
-	if (address < PAGE_ALIGN(address+4))
-		address = PAGE_ALIGN(address+4);
-	else
+	address &= PAGE_MASK;
+	address += PAGE_SIZE;
+	if (!address)
 		return -ENOMEM;
  
+	/* Enforce stack_guard_gap */
+	gap_addr = address + stack_guard_gap;
+	if (gap_addr < address)
+		return -ENOMEM;
+	next = vma->vm_next;
+	if (next && next->vm_start < gap_addr) {
+		if (!(next->vm_flags & VM_GROWSUP))
+			return -ENOMEM;
+		/* Check that both stack segments have the same anon_vma? */
+	}
+
 	/* We must make sure the anon_vma is allocated. */
 	if (unlikely(anon_vma_prepare(vma)))
 		return -ENOMEM;
@@ -2257,7 +2284,7 @@
 				if (vma->vm_next)
 					vma_gap_update(vma->vm_next);
 				else
-					mm->highest_vm_end = address;
+					mm->highest_vm_end = vm_end_gap(vma);
 				spin_unlock(&mm->page_table_lock);
  
 				perf_event_mmap(vma);
@@ -2278,6 +2305,8 @@
 				   unsigned long address)
 {
 	struct mm_struct *mm = vma->vm_mm;
+	struct vm_area_struct *prev;
+	unsigned long gap_addr;
 	int error;
  
 	address &= PAGE_MASK;
@@ -2285,6 +2314,17 @@
 	if (error)
 		return error;
  
+	/* Enforce stack_guard_gap */
+	gap_addr = address - stack_guard_gap;
+	if (gap_addr > address)
+		return -ENOMEM;
+	prev = vma->vm_prev;
+	if (prev && prev->vm_end > gap_addr) {
+		if (!(prev->vm_flags & VM_GROWSDOWN))
+			return -ENOMEM;
+		/* Check that both stack segments have the same anon_vma? */
+	}
+
 	/* We must make sure the anon_vma is allocated. */
 	if (unlikely(anon_vma_prepare(vma)))
 		return -ENOMEM;
  
@@ -2339,28 +2379,25 @@
 	return error;
 }
  
-/*
- * Note how expand_stack() refuses to expand the stack all the way to
- * abut the next virtual mapping, *unless* that mapping itself is also
- * a stack mapping. We want to leave room for a guard page, after all
- * (the guard page itself is not added here, that is done by the
- * actual page faulting logic)
- *
- * This matches the behavior of the guard page logic (see mm/memory.c:
- * check_stack_guard_page()), which only allows the guard page to be
- * removed under these circumstances.
- */
+/* enforced gap between the expanding stack and other mappings. */
+unsigned long stack_guard_gap = 256UL<<PAGE_SHIFT;
+
+static int __init cmdline_parse_stack_guard_gap(char *p)
+{
+	unsigned long val;
+	char *endptr;
+
+	val = simple_strtoul(p, &endptr, 10);
+	if (!*endptr)
+		stack_guard_gap = val << PAGE_SHIFT;
+
+	return 0;
+}
+__setup("stack_guard_gap=", cmdline_parse_stack_guard_gap);
+
 #ifdef CONFIG_STACK_GROWSUP
 int expand_stack(struct vm_area_struct *vma, unsigned long address)
 {
-	struct vm_area_struct *next;
-
-	address &= PAGE_MASK;
-	next = vma->vm_next;
-	if (next && next->vm_start == address + PAGE_SIZE) {
-		if (!(next->vm_flags & VM_GROWSUP))
-			return -ENOMEM;
-	}
 	return expand_upwards(vma, address);
 }
  
@@ -2382,14 +2419,6 @@
 #else
 int expand_stack(struct vm_area_struct *vma, unsigned long address)
 {
-	struct vm_area_struct *prev;
-
-	address &= PAGE_MASK;
-	prev = vma->vm_prev;
-	if (prev && prev->vm_end == address) {
-		if (!(prev->vm_flags & VM_GROWSDOWN))
-			return -ENOMEM;
-	}
 	return expand_downwards(vma, address);
 }
  
@@ -2487,7 +2516,7 @@
 		vma->vm_prev = prev;
 		vma_gap_update(vma);
 	} else
-		mm->highest_vm_end = prev ? prev->vm_end : 0;
+		mm->highest_vm_end = prev ? vm_end_gap(prev) : 0;
 	tail_vma->vm_next = NULL;
  
 	/* Kill the cache */
...	...	@@ -3932,6 +3932,13 @@
3932	3932	spia_pedr=
3933	3933	spia_peddr=
3934	3934
	3935	+ stack_guard_gap= [MM]
	3936	+ override the default stack gap protection. The value
	3937	+ is in page units and it defines how many pages prior
	3938	+ to (for stacks growing down) resp. after (for stacks
	3939	+ growing up) the main stack are reserved for no other
	3940	+ mapping. Default value is 256 pages.
	3941	+
3935	3942	stacktrace [FTRACE]
3936	3943	Enabled the stack tracer on boot up.
3937	3944
...	...	@@ -64,7 +64,7 @@
64	64
65	65	vma = find_vma(mm, addr);
66	66	if (TASK_SIZE - len >= addr &&
67		- (!vma \|\| addr + len <= vma->vm_start))
	67	+ (!vma \|\| addr + len <= vm_start_gap(vma)))
68	68	return addr;
69	69	}
70	70
...	...	@@ -89,7 +89,7 @@
89	89
90	90	vma = find_vma(mm, addr);
91	91	if (TASK_SIZE - len >= addr &&
92		- (!vma \|\| addr + len <= vma->vm_start))
	92	+ (!vma \|\| addr + len <= vm_start_gap(vma)))
93	93	return addr;
94	94	}
95	95
...	...	@@ -140,7 +140,7 @@
140	140	addr = PAGE_ALIGN(addr);
141	141	vma = find_vma(mm, addr);
142	142	if (TASK_SIZE - len >= addr &&
143		- (!vma \|\| addr + len <= vma->vm_start))
	143	+ (!vma \|\| addr + len <= vm_start_gap(vma)))
144	144	return addr;
145	145	}
146	146
...	...	@@ -74,7 +74,7 @@
74	74	addr = PAGE_ALIGN(addr);
75	75	vma = find_vma(current->mm, addr);
76	76	if (TASK_SIZE - len >= addr &&
77		- (!vma \|\| addr + len <= vma->vm_start))
	77	+ (!vma \|\| addr + len <= vm_start_gap(vma)))
78	78	goto success;
79	79	}
80	80
...	...	@@ -92,7 +92,7 @@
92	92
93	93	vma = find_vma(mm, addr);
94	94	if (TASK_SIZE - len >= addr &&
95		- (!vma \|\| addr + len <= vma->vm_start))
	95	+ (!vma \|\| addr + len <= vm_start_gap(vma)))
96	96	return addr;
97	97	}
98	98
...	...	@@ -88,7 +88,7 @@
88	88	unsigned long len, unsigned long pgoff, unsigned long flags)
89	89	{
90	90	struct mm_struct *mm = current->mm;
91		- struct vm_area_struct *vma;
	91	+ struct vm_area_struct vma, prev;
92	92	unsigned long task_size = TASK_SIZE;
93	93	int do_color_align, last_mmap;
94	94	struct vm_unmapped_area_info info;
95	95
...	...	@@ -115,9 +115,10 @@
115	115	else
116	116	addr = PAGE_ALIGN(addr);
117	117
118		- vma = find_vma(mm, addr);
	118	+ vma = find_vma_prev(mm, addr, &prev);
119	119	if (task_size - len >= addr &&
120		- (!vma \|\| addr + len <= vma->vm_start))
	120	+ (!vma \|\| addr + len <= vm_start_gap(vma)) &&
	121	+ (!prev \|\| addr >= vm_end_gap(prev)))
121	122	goto found_addr;
122	123	}
123	124
...	...	@@ -141,7 +142,7 @@
141	142	const unsigned long len, const unsigned long pgoff,
142	143	const unsigned long flags)
143	144	{
144		- struct vm_area_struct *vma;
	145	+ struct vm_area_struct vma, prev;
145	146	struct mm_struct *mm = current->mm;
146	147	unsigned long addr = addr0;
147	148	int do_color_align, last_mmap;
148	149
...	...	@@ -175,9 +176,11 @@
175	176	addr = COLOR_ALIGN(addr, last_mmap, pgoff);
176	177	else
177	178	addr = PAGE_ALIGN(addr);
178		- vma = find_vma(mm, addr);
	179	+
	180	+ vma = find_vma_prev(mm, addr, &prev);
179	181	if (TASK_SIZE - len >= addr &&
180		- (!vma \|\| addr + len <= vma->vm_start))
	182	+ (!vma \|\| addr + len <= vm_start_gap(vma)) &&
	183	+ (!prev \|\| addr >= vm_end_gap(prev)))
181	184	goto found_addr;
182	185	}
183	186
...	...	@@ -65,7 +65,7 @@
65	65	addr = ALIGN(addr, huge_page_size(h));
66	66	vma = find_vma(mm, addr);
67	67	if (TASK_SIZE - len >= addr &&
68		- (!vma \|\| addr + len <= vma->vm_start))
	68	+ (!vma \|\| addr + len <= vm_start_gap(vma)))
69	69	return addr;
70	70	}
71	71	/*
...	...	@@ -106,7 +106,7 @@
106	106	addr = PAGE_ALIGN(addr);
107	107	vma = find_vma(mm, addr);
108	108	if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
109		- (!vma \|\| addr + len <= vma->vm_start))
	109	+ (!vma \|\| addr + len <= vm_start_gap(vma)))
110	110	return addr;
111	111	}
112	112
...	...	@@ -142,7 +142,7 @@
142	142	addr = PAGE_ALIGN(addr);
143	143	vma = find_vma(mm, addr);
144	144	if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
145		- (!vma \|\| addr + len <= vma->vm_start))
	145	+ (!vma \|\| addr + len <= vm_start_gap(vma)))
146	146	return addr;
147	147	}
148	148
...	...	@@ -105,7 +105,7 @@
105	105	if ((mm->task_size - len) < addr)
106	106	return 0;
107	107	vma = find_vma(mm, addr);
108		- return (!vma \|\| (addr + len) <= vma->vm_start);
	108	+ return (!vma \|\| (addr + len) <= vm_start_gap(vma));
109	109	}
110	110
111	111	static int slice_low_has_vma(struct mm_struct *mm, unsigned long slice)
...	...	@@ -98,7 +98,7 @@
98	98	addr = PAGE_ALIGN(addr);
99	99	vma = find_vma(mm, addr);
100	100	if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
101		- (!vma \|\| addr + len <= vma->vm_start))
	101	+ (!vma \|\| addr + len <= vm_start_gap(vma)))
102	102	return addr;
103	103	}
104	104
...	...	@@ -136,7 +136,7 @@
136	136	addr = PAGE_ALIGN(addr);
137	137	vma = find_vma(mm, addr);
138	138	if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
139		- (!vma \|\| addr + len <= vma->vm_start))
	139	+ (!vma \|\| addr + len <= vm_start_gap(vma)))
140	140	return addr;
141	141	}
142	142