Skip to contentSkip to conversation

mm: Fix hiding the stack guard page in proc

PATCH In review Stefan Bader 2 participants 3 messages Aug 31, 2010

Subsystems (4)
Memory management MM core Proc filesystem Memory mapping
lore
Share & tools
lore

Stefan Bader

author
#0 lore raw

PATCH 1/1 mm: Fix hiding the stack guard page in proc

Looking at the other changes I believe that the logic for proc also needs to look at the previous vma to decide whether to hide the first page of it or not. The following patch against current head should fix that (only compile tested here . I only made the change to task_mmu.c here. But basically its the same check as in mlock.c. Maybe its would be worth moving the inlines to mm.h.

-Stefan

Current output before:
...
b778e000-b7790000 rw-p 00000000 00:00 0
bfdc8000-bfddd000 rw-p 00000000 00:00 0 [stack]

After mlocking parts of the stack:
...
b778d000-b7790000 rw-p 00000000 00:00 0
bfdc8000-bfdda000 rw-p 00000000 00:00 0
bfddb000-bfddb000 rw-p 00000000 00:00 0 [stack]
bfddc000-bfddd000 rw-p 00000000 00:00 0

Quoted 1 line

From baa30fb2027096d6604c9f0ba2bad04a30edc482 Mon Sep 17 00:00:00 2001

From: Stefan Bader <[email protected]>
Date: Tue, 31 Aug 2010 13:59:58 +0200
Subject: [PATCH] mm: Fix hiding the stack guard page in proc

Commit d7824370e26325c881b665350ce64fb0a4fde24a tried to hide the stack guard page from user-space by adapting mlock and hiding it in /proc. Commit 7798330ac8114c731cfab83e634c6ecedaa233d7 fixed the mlock part, using the newly added double linked vma lists.
Still the proc change would hide the first page of each stack vma which would be wrong for those vmas that are just continuations of the stack after its vma has been split.

Signed-off-by Stefan Bader <[email protected]>
Cc [email protected]
---
 fs/proc/task_mmu.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)
fs/proc/task_mmu.c +3 -1
index 439fc1f..2e302b4 100644
@@ -224,7 +224,9 @@ static void show_map_vma(struct seq_file *m, struct vm_area_struct *vma)
 	/* We don't show the stack guard page in /proc/maps */
 	start = vma->vm_start;
 	if (vma->vm_flags & VM_GROWSDOWN)
-		start += PAGE_SIZE;
+		if (!vma->vm_prev || vma->vm_prev->vm_flags != VM_GROWSDOWN ||
+		    vma->vm_start != vma->vm_prev->vm_end)
+			start += PAGE_SIZE;
 
 	seq_printf(m, "%08lx-%08lx %c%c%c%c %08llx %02x:%02x %lu %n",
 			start,
Signature · 1 line
1.7.0.4
#1 lore raw
Quoted 2 lines · Stefan Bader

Maybe its would be worth moving the inlines to mm.h.

Yes, I think it would be a good idea to try to abstract it out a bit...

Linus

author
#2 lore raw
Quoted 7 lines · Linus Torvalds
Quoted 2 lines · Stefan Bader

Maybe its would be worth moving the inlines to mm.h.

Yes, I think it would be a good idea to try to abstract it out a bit...

Linus

Quick and dirty (and very untested)

Stefan

Quoted 1 line

From 33f7a4a8d572628dd82c8953d434b2806a1c448e Mon Sep 17 00:00:00 2001

From: Stefan Bader <[email protected]>
Date: Tue, 31 Aug 2010 15:52:27 +0200
Subject: [PATCH] mm: Move vma_stack_continue into mm.h

So it can be used by all that need to check for that.

Signed-off-by Stefan Bader <[email protected]>
---
 fs/proc/task_mmu.c |    3 +--
 include/linux/mm.h |    7 +++++++
 mm/mlock.c         |    6 ------
 3 files changed, 8 insertions(+), 8 deletions(-)
fs/proc/task_mmu.c +1 -2
index 2e302b4..ac0101a 100644
@@ -224,8 +224,7 @@ static void show_map_vma(struct seq_file *m, struct vm_area_struct *vma)
 	/* We don't show the stack guard page in /proc/maps */
 	start = vma->vm_start;
 	if (vma->vm_flags & VM_GROWSDOWN)
-		if (!vma->vm_prev || vma->vm_prev->vm_flags != VM_GROWSDOWN ||
-		    vma->vm_start != vma->vm_prev->vm_end)
+		if(!vma_stack_continue(vma->vm_prev, vma->vm_start))
 			start += PAGE_SIZE;
 
 	seq_printf(m, "%08lx-%08lx %c%c%c%c %08llx %02x:%02x %lu %n",
include/linux/mm.h +7 -0
index e6b1210..4f4773a 100644
@@ -864,6 +864,13 @@ int set_page_dirty(struct page *page);
 int set_page_dirty_lock(struct page *page);
 int clear_page_dirty_for_io(struct page *page);
 
+/* Is the vma a continuation of the stack vma above it? */
+static inline int vma_stack_continue(struct vm_area_struct *vma,
+				     unsigned long addr)
+{
+	return vma && (vma->vm_end == addr) && (vma->vm_flags & VM_GROWSDOWN);
+}
+
 extern unsigned long move_page_tables(struct vm_area_struct *vma,
 		unsigned long old_addr, struct vm_area_struct *new_vma,
 		unsigned long new_addr, unsigned long len);
mm/mlock.c +0 -6
index cbae7c5..b70919c 100644
@@ -135,12 +135,6 @@ void munlock_vma_page(struct page *page)
 	}
 }
 
-/* Is the vma a continuation of the stack vma above it? */
-static inline int vma_stack_continue(struct vm_area_struct *vma, unsigned long addr)
-{
-	return vma && (vma->vm_end == addr) && (vma->vm_flags & VM_GROWSDOWN);
-}
-
 static inline int stack_guard_page(struct vm_area_struct *vma, unsigned long addr)
 {
 	return (vma->vm_flags & VM_GROWSDOWN) &&
Signature · 1 line
1.7.0.4