RE: [PATCH 2/3] xfs: Calling destroy_work_on_stack() to pair with INIT_WORK_ONSTACK()

["Liu, Chuansheng" <chuansheng.liu@xxxxxxxxx>]

Hello Jeff,

> -----Original Message-----
> From: Jeff Liu [mailto:jeff.liu@xxxxxxxxxx]
> Sent: Tuesday, January 07, 2014 7:21 PM
> To: Liu, Chuansheng; dchinner@xxxxxxxxxxxxx; bpm@xxxxxxx
> Cc: linux-kernel@xxxxxxxxxxxxxxx; xfs@xxxxxxxxxxx
> Subject: Re: [PATCH 2/3] xfs: Calling destroy_work_on_stack() to pair with
> INIT_WORK_ONSTACK()
> 
> Hi Chuansheng,
> 
> On 01/07 2014 16:53 PM, Chuansheng Liu wrote:
> >
> > In case CONFIG_DEBUG_OBJECTS_WORK is defined, it is needed to
> > call destroy_work_on_stack() which frees the debug object to pair
> > with INIT_WORK_ONSTACK().
> >
> > Signed-off-by: Liu, Chuansheng <chuansheng.liu@xxxxxxxxx>
> > ---
> Thanks for your patch and it work fine for my testing.  I missed this in an
> old commit: [ 3b876c8f2a xfs: fix debug_object WARN at xfs_alloc_vextent() ]
Thanks your testing.

> 
> Just out of curious, do you notice memory leaks or other hints which help you
> finding out this problem?
I am trying to use the INIT_WORK_ONSTACK() in my code, then I found in some 
places
Calling destroy_work_on_stack() is missed.