lkml.org 
[lkml]   [2010]   [Jun]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: 2.6.35-rc3 - include/linux/fdtable.h:85 invoked rcu_dereference_check() without protection!
On Sun, Jun 13, 2010 at 02:36:03AM -0400, Miles Lane wrote:
> [ 2074.377176] [ INFO: suspicious rcu_dereference_check() usage. ]
> [ 2074.377180] ---------------------------------------------------
> [ 2074.377185] include/linux/fdtable.h:85 invoked
> rcu_dereference_check() without protection!
> [ 2074.377189]
> [ 2074.377190] other info that might help us debug this:
> [ 2074.377191]
> [ 2074.377195]
> [ 2074.377196] rcu_scheduler_active = 1, debug_locks = 1
> [ 2074.377200] no locks held by gvfsd/4402.
> [ 2074.377203]
> [ 2074.377204] stack backtrace:
> [ 2074.377209] Pid: 4402, comm: gvfsd Not tainted 2.6.35-rc3 #1
> [ 2074.377213] Call Trace:
> [ 2074.377226] [<ffffffff8106f2c5>] lockdep_rcu_dereference+0x9d/0xa5
> [ 2074.377235] [<ffffffff810e9104>] fcheck_files+0x5c/0xc9
> [ 2074.377242] [<ffffffff810e91c9>] fget_light+0x35/0x96
> [ 2074.377250] [<ffffffff810f77b0>] do_sys_poll+0x1ff/0x3e5
> [ 2074.377257] [<ffffffff810f6eb2>] ? __pollwait+0x0/0xc7
> [ 2074.377264] [<ffffffff810f6f79>] ? pollwake+0x0/0x4f
> [ 2074.377271] [<ffffffff810f6f79>] ? pollwake+0x0/0x4f
> [ 2074.377277] [<ffffffff810f6f79>] ? pollwake+0x0/0x4f
> [ 2074.377285] [<ffffffff812d2f18>] ? __sock_recvmsg_nosec+0x65/0x6e
> [ 2074.377292] [<ffffffff812d4401>] ? __sock_recvmsg+0x49/0x54
> [ 2074.377299] [<ffffffff812d6842>] ? sock_aio_read+0xc2/0xd6
> [ 2074.377307] [<ffffffff810e7faf>] ? do_sync_read+0xc6/0x103
> [ 2074.377316] [<ffffffff8119ec0c>] ? security_file_permission+0x11/0x13
> [ 2074.377324] [<ffffffff8100877a>] ? sysret_check+0x2e/0x69
> [ 2074.377331] [<ffffffff810f7b2a>] sys_poll+0x50/0xbb
> [ 2074.377338] [<ffffffff81008742>] system_call_fastpath+0x16/0x1b

Hmmm... This was supposed to have been fixed by the following:

http://lkml.org/lkml/2010/5/11/489

The rcu_my_thread_group_empty() function that it depends on is in
mainline, but this is not. Checking the -vfs tree... Not seeing it
there, either.

I will repost the patch.

Thank you for following up on this!!!

Thanx, Paul


\
 
 \ /
  Last update: 2010-06-17 02:15    [from the cache]
©2003-2011 Jasper Spaans