If the mmap_sem is not held while we walk_page_range(), then
it is possible for find_vma() to race with a remove_vma_list()
caused by do_munmap() (or others).
Unable to handle kernel paging request at virtual address 6b6b6b5b
Internal error: Oops: 5 [#1] PREEMPT
CPU: 0 Not tainted (2.6.32.9-27154-ge3e6e27 #1)
PC is at find_vma+0x40/0x7c
LR is at walk_page_range+0x70/0x230
pc : [<c00aa3ac>] lr : [<c00b298c>] psr: 20000013
sp : c6aa9eb8 ip : 6b6b6b53 fp : c6a58f60
r10: c7e1d1b8 r9 : 0001bca0 r8 : 47000000
r7 : c6aa9f80 r6 : c6aa8000 r5 : 46fbd000 r4 : 6b6b6b6b
r3 : c7ca4820 r2 : 6b6b6b6b r1 : 46fbd000 r0 : c70e3e40
Flags: nzCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment user
Control: 10c5787d Table: 26574019 DAC: 00000015
[<c00aa3ac>] (find_vma+0x40/0x7c) from [<c00b298c>] (walk_page_range+0x70/0x230)
[<c00b298c>] (walk_page_range+0x70/0x230) from [<c00f5d3c>] (pagemap_read+0x1a4/0x278)
[<c00f5d3c>] (pagemap_read+0x1a4/0x278) from [<c00bac40>] (vfs_read+0xa8/0x150)
[<c00bac40>] (vfs_read+0xa8/0x150) from [<c00bad94>] (sys_read+0x3c/0x68)
[<c00bad94>] (sys_read+0x3c/0x68) from [<c0026f00>] (ret_fast_syscall+0x0/0x2c)
Code: 98bd8010 e5932004 e3a00000 ea000008 (e5124010)
Signed-off-by: San Mehat <san@google.com>
Cc: Brian Swetland <swetland@google.com>
Cc: Matt Mackall <mpm@selenic.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
---
fs/proc/task_mmu.c | 3 +--
1 files changed, 1 insertions(+), 2 deletions(-)
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 2a1bef9..3f300c1 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -726,8 +726,6 @@ static ssize_t pagemap_read(struct file *file, char __user *buf,
down_read(¤t->mm->mmap_sem);
ret = get_user_pages(current, current->mm, uaddr, pagecount,
1, 0, pages, NULL);
- up_read(¤t->mm->mmap_sem);
-
if (ret < 0)
goto out_free;
@@ -776,6 +774,7 @@ out_pages:
page_cache_release(page);
}
out_free:
+ up_read(¤t->mm->mmap_sem);
kfree(pages);
out_mm:
mmput(mm);
--
1.7.0.1
--
