[PATCH] xfs: use INIT_WORK_ONSTACK in xfs_alloc_vextent()

[Eric Sandeen]

On 6/8/12 6:12 PM, Dave Chinner wrote:
> On Fri, Jun 08, 2012 at 02:32:42PM -0500, Eric Sandeen wrote:
>> w/o this change I was getting:
>>
>> [  742.376495] ODEBUG: object is on stack, but not annotated
>> [  742.377007] ------------[ cut here ]------------
>> [  742.377007] WARNING: at lib/debugobjects.c:302 __debug_object_init+0x1c0/0x3d0()
>> ...
>> [  742.377007] Call Trace:
>> [  742.377007]  [<ffffffff81060dbf>] warn_slowpath_common+0x7f/0xc0
>> [  742.377007]  [<ffffffff81060e1a>] warn_slowpath_null+0x1a/0x20
>> [  742.377007]  [<ffffffff81336640>] __debug_object_init+0x1c0/0x3d0
>> [  742.377007]  [<ffffffff8133689f>] debug_object_init+0x1f/0x30
>> [  742.377007]  [<ffffffff81081cd9>] __init_work+0x19/0x30
>> [  742.377007]  [<ffffffffa0299f81>] xfs_alloc_vextent+0x101/0x190 [xfs]
>>
>> on a debug 3.4 kernel.
>>
>> Signed-off-by: Eric Sandeen <sandeen at redhat.com>
> 
> Already posted by Jeff twice - I picked this up and started testing
> it a couple of days ago:
> 
> http://marc.info/?l=linux-fsdevel&m=133877118608852&w=2
> http://oss.sgi.com/archives/xfs/2012-06/msg00081.html
> 
> Sorry to burst you bubble, Eric, but I think Jeff gets this one. ;)

sigh.  Can't keep up.  But I try!

-Eric

> Cheers,
> 
> Dave.