[PATCH 1/2] drm/xe: Fix missing workqueue destroy in xe_gt_pagefault

Summers, Stuart stuart.summers at intel.com
Tue Jul 9 22:54:36 UTC 2024


On Tue, 2024-07-09 at 20:48 +0000, Matthew Brost wrote:
> On Tue, Jul 09, 2024 at 02:59:41PM -0400, Rodrigo Vivi wrote:
> > On Tue, Jul 09, 2024 at 06:40:59PM +0000, Stuart Summers wrote:
> > > On driver reload we never free up the memory for the pagefault
> > > and
> > > access counter workqueues. Add those destroy calls here.
> > > 
> > > Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for
> > > Intel GPUs")
> > > Signed-off-by: Stuart Summers <stuart.summers at intel.com>
> > > ---
> > >  drivers/gpu/drm/xe/xe_gt_pagefault.c | 14 +++++++++++++-
> > >  1 file changed, 13 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/drivers/gpu/drm/xe/xe_gt_pagefault.c
> > > b/drivers/gpu/drm/xe/xe_gt_pagefault.c
> > > index e18659201208..ded8e3c33a82 100644
> > > --- a/drivers/gpu/drm/xe/xe_gt_pagefault.c
> > > +++ b/drivers/gpu/drm/xe/xe_gt_pagefault.c
> > > @@ -398,6 +398,18 @@ static void pf_queue_work_func(struct
> > > work_struct *w)
> > >  
> > >  static void acc_queue_work_func(struct work_struct *w);
> > >  
> > > +static void pagefault_fini(void *arg)
> > > +{
> > > +       struct xe_gt *gt = arg;
> > > +       struct xe_device *xe = gt_to_xe(gt);
> > > +
> > > +       if (!xe->info.has_usm)
> > > +               return;
> > > +
> > > +       destroy_workqueue(gt->usm.acc_wq);
> > > +       destroy_workqueue(gt->usm.pf_wq);
> > > +}
> > > +
> > >  int xe_gt_pagefault_init(struct xe_gt *gt)
> > >  {
> > >         struct xe_device *xe = gt_to_xe(gt);
> > > @@ -428,7 +440,7 @@ int xe_gt_pagefault_init(struct xe_gt *gt)
> > >         if (!gt->usm.acc_wq)
> > >                 return -ENOMEM;
> > 
> > I believe we should destroy the  pf_wq in this return, no?!
> > 
> 
> Yes, I was going to say the same thing in my review - we need to
> destroy
> pf_wq here.

Yep that's right :( I'll fix it up and resend.

> 
> Matt
> 
> > with that fixed or clarified,
> > 
> > Reviewed-by: Rodrigo Vivi <rodrigo.vivi at intel.com>

Thanks Rodrigo!

-Stuart

> > 
> > >  
> > > -       return 0;
> > > +       return devm_add_action_or_reset(xe->drm.dev,
> > > pagefault_fini, gt);
> > >  }
> > >  
> > >  void xe_gt_pagefault_reset(struct xe_gt *gt)
> > > -- 
> > > 2.34.1
> > > 



More information about the Intel-xe mailing list