[Intel-gfx] [PATCH 2/3] drm/i915: Detect page faults during hangcheck

[Chris Wilson]

On Mon, Mar 16, 2015 at 06:53:40PM +0100, Daniel Vetter wrote:
> On Mon, Mar 16, 2015 at 03:46:35PM +0200, Mika Kuoppala wrote:
> > From: Chris Wilson <chris at chris-wilson.co.uk>
> > 
> > On Sandybridge+, the GPU provides the ERROR register for detecting page
> > faults. Hook this up to our hangcheck so that we can dump the error
> > state soon after such an event occurs. This would be better inside an
> > interrupt handler, but it serves a purpose here as it detects that our
> > initial context setup is invalid...
> > 
> > Reviewed-by: Mika Kuoppala <mika.kuoppala at intel.com>
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > ---
> >  drivers/gpu/drm/i915/i915_irq.c     | 5 +++++
> >  drivers/gpu/drm/i915/intel_uncore.c | 2 ++
> >  2 files changed, 7 insertions(+)
> > 
> > diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
> > index 49ad5fb..ea668fc 100644
> > --- a/drivers/gpu/drm/i915/i915_irq.c
> > +++ b/drivers/gpu/drm/i915/i915_irq.c
> > @@ -2929,6 +2929,11 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
> >  	if (!i915.enable_hangcheck)
> >  		return;
> >  
> > +	if (INTEL_INFO(dev_priv)->gen >= 6 && I915_READ(ERROR_GEN6)) {
> > +		i915_handle_error(dev, false, "GPU reported a page fault");
> > +		I915_WRITE(ERROR_GEN6, 0);
> 
> Shouldn't we also at least report the bits from the ERROR register
> somewhere? Or are they supremely useless?

It's recorded in the hangcheck already - that's how I knew we were
generating pagefaults during module init in the first place.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre