[Intel-gfx] [PATCH v2] drm/i915: Skip HW reinitialisation on resume if still wedged

[Chris Wilson]

Quoting Mika Kuoppala (2017-10-16 15:24:33)
> Chris Wilson <chris at chris-wilson.co.uk> writes:
> 
> > If we fail to recover the HW state upon resume (i.e. our attempt to
> > clear the wedged bit and reset during i915_gem_sanitize() fails), then
> > skip the HW restart inside i915_gem_init_hw(). We will ultimately do the
> > HW restart when successfully unwedging and resetting the HW later,
> > but attempting to restore a wedged device upon resume is risky as the HW
> > is in an unknown state.
> >
> > v2: Suppress the error message when detecting the already wedged HW.
> >
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > Cc: Mika Kuoppala <mika.kuoppala at linux.intel.com>
> > ---
> >  drivers/gpu/drm/i915/i915_gem.c | 10 ++++++++--
> >  1 file changed, 8 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> > index d9d39b309ce8..449f8c3788b1 100644
> > --- a/drivers/gpu/drm/i915/i915_gem.c
> > +++ b/drivers/gpu/drm/i915/i915_gem.c
> > @@ -4835,6 +4835,10 @@ int i915_gem_init_hw(struct drm_i915_private *dev_priv)
> >       init_unused_rings(dev_priv);
> >  
> >       BUG_ON(!dev_priv->kernel_context);
> > +     if (i915_terminally_wedged(&dev_priv->gpu_error)) {
> > +             ret = -EIO;
> > +             goto out;
> > +     }
> >
> 
> You have done some hw initialization already before this point.
> Is there a reason for not moving this right before acquiring
> forcewake?

init_unused_rings() is part of the sanitisation I wanted to keep. The
other mmio writes we need to sort out in the right w/a category; if they
are display related we need to keep them. Hence, being chicken and
sticking the escape clause here, right before we commit to restarting
the engines.
-Chris