lkml.org 
[lkml]   [2011]   [Oct]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
Subject[possible deadlock][3.1.0-g138c4ae] possible circular locking dependency detected
Hi folks:

My dmesg said that:

======================================================
[ INFO: possible circular locking dependency detected ]
3.1.0-138c4ae #2
-------------------------------------------------------
hugemmap05/18198 is trying to acquire lock:
(&mm->mmap_sem){++++++}, at: [<ffffffff8114d85c>] might_fault+0x5c/0xb0
but task is already holding lock:
(&sb->s_type->i_mutex_key#21){+.+.+.}, at: [<ffffffff811a10f6>] vfs_readdir+0x86/0xe0
which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #1 (&sb->s_type->i_mutex_key#21){+.+.+.}:
[<ffffffff810afd34>] validate_chain+0x704/0x860
[<ffffffff810b018c>] __lock_acquire+0x2fc/0x500
[<ffffffff810b0b01>] lock_acquire+0xb1/0x1a0
[<ffffffff815464f2>] __mutex_lock_common+0x62/0x420
[<ffffffff81546a1a>] mutex_lock_nested+0x4a/0x60
[<ffffffff8120b4ba>] hugetlbfs_file_mmap+0xaa/0x160
[<ffffffff81158071>] mmap_region+0x3e1/0x590
[<ffffffff81158584>] do_mmap_pgoff+0x364/0x3b0
[<ffffffff811587d9>] sys_mmap_pgoff+0x209/0x240
[<ffffffff8101aac9>] sys_mmap+0x29/0x30
[<ffffffff81551542>] system_call_fastpath+0x16/0x1b
-> #0 (&mm->mmap_sem){++++++}:
[<ffffffff810af607>] check_prev_add+0x537/0x560
[<ffffffff810afd34>] validate_chain+0x704/0x860
[<ffffffff810b018c>] __lock_acquire+0x2fc/0x500
[<ffffffff810b0b01>] lock_acquire+0xb1/0x1a0
[<ffffffff8114d889>] might_fault+0x89/0xb0
[<ffffffff811a0f2e>] filldir+0x7e/0xe0
[<ffffffff811b445e>] dcache_readdir+0x5e/0x230
[<ffffffff811a1130>] vfs_readdir+0xc0/0xe0
[<ffffffff811a12c9>] sys_getdents+0x89/0x100
[<ffffffff81551542>] system_call_fastpath+0x16/0x1b
other info that might help us debug this:

Possible unsafe locking scenario:

CPU0 CPU1
---- ----
lock(&sb->s_type->i_mutex_key);
lock(&mm->mmap_sem);
lock(&sb->s_type->i_mutex_key);
lock(&mm->mmap_sem);
*** DEADLOCK ***

1 lock held by hugemmap05/18198:
#0: (&sb->s_type->i_mutex_key#21){+.+.+.}, at: [<ffffffff811a10f6>] vfs_readdir+0x86/0xe0
stack backtrace:
Pid: 18198, comm: hugemmap05 Not tainted 3.1.0-138c4ae #2
Call Trace:
[<ffffffff810ad469>] print_circular_bug+0x109/0x110
[<ffffffff810af607>] check_prev_add+0x537/0x560
[<ffffffff8114e112>] ? do_anonymous_page+0xf2/0x2d0
[<ffffffff810afd34>] validate_chain+0x704/0x860
[<ffffffff810b018c>] __lock_acquire+0x2fc/0x500
[<ffffffff810b0b01>] lock_acquire+0xb1/0x1a0
[<ffffffff8114d85c>] ? might_fault+0x5c/0xb0
[<ffffffff8114d889>] might_fault+0x89/0xb0
[<ffffffff8114d85c>] ? might_fault+0x5c/0xb0
[<ffffffff81546763>] ? __mutex_lock_common+0x2d3/0x420
[<ffffffff811a10f6>] ? vfs_readdir+0x86/0xe0
[<ffffffff811a0f2e>] filldir+0x7e/0xe0
[<ffffffff811b445e>] dcache_readdir+0x5e/0x230
[<ffffffff811a0eb0>] ? filldir64+0xf0/0xf0
[<ffffffff811a0eb0>] ? filldir64+0xf0/0xf0
[<ffffffff811a0eb0>] ? filldir64+0xf0/0xf0
[<ffffffff811a1130>] vfs_readdir+0xc0/0xe0
[<ffffffff8118e9be>] ? fget+0xee/0x220
[<ffffffff8118e8d0>] ? fget_raw+0x220/0x220
[<ffffffff811a12c9>] sys_getdents+0x89/0x100
[<ffffffff81551542>] system_call_fastpath+0x16/0x1b


Wile hugemmap05 is a test case from LTP.
http://ltp.git.sourceforge.net/git/gitweb.cgi?p=ltp/ltp.git;a=blob;f=testcases/kernel/mem/hugetlb/hugemmap/hugemmap05.c;h=50bb8ca23ae9686662740f9ea5d7187affff8b60;hb=HEAD
But I don't know how to reproduce this.


Thanks
-Wanlong Gao


\
 
 \ /
  Last update: 2011-10-28 05:45    [from the cache]
©2003-2011 Jasper Spaans