[Intel-gfx] [PATCH 6/6] drm/i915: make gpu hangman more resilient

[Daniel Vetter]

On Wed, Apr 25, 2012 at 11:30:24PM -0300, Eugeni Dodonov wrote:
> On Apr 25, 2012 9:01 AM, "Daniel Vetter" <daniel.vetter at ffwll.ch> wrote:
> > +       /* Also reset the gpu hangman. */
> > +       if (dev_priv->stop_rings) {
> > +               DRM_DEBUG("Simulated gpu hang, resetting stop_rings\n");
> 
> Maybe also print the stop_ring value here?
> 
> > +               dev_priv->stop_rings = 0;
> > +               if (ret) {
> > +                       DRM_ERROR("Ignoring reset failure in simulated
> gpu hang\n");
> 
> ...and error value here as well?
> 
> If we are simulating hangs, I'd guess that we are interested in those
> values..

Well, I think I'll check for -ENODEV only instead because I simply don't
want to kill the gpu and force a module reload just because we don't have
the reset code implemented.
-Daniel
-- 
Daniel Vetter
Mail: daniel at ffwll.ch
Mobile: +41 (0)79 365 57 48