Quoting Michal Wajdeczko (2018-04-09 13:23:26)
> By calling in i915_reset only i915_gem_init_hw without previous
> i915_gem_fini_hw we introduced asymmetry. Let's fix that.
>
> Signed-off-by: Michal Wajdeczko <michal.wajdeczko@xxxxxxxxx>
> Cc: Sagar Arun Kamble <sagar.a.kamble@xxxxxxxxx>
> Cc: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
Reviewed-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> ---
> drivers/gpu/drm/i915/i915_drv.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
> index 854b26c..a0a5af0 100644
> --- a/drivers/gpu/drm/i915/i915_drv.c
> +++ b/drivers/gpu/drm/i915/i915_drv.c
> @@ -1902,6 +1902,8 @@ void i915_reset(struct drm_i915_private *i915,
> goto error;
> }
>
> + i915_gem_fini_hw(i915);
> +
> for (i = 0; i < 3; i++) {
> ret = intel_gpu_reset(i915, ALL_ENGINES);
> if (ret == 0)
I still have a feeling that i915_gem_reset() will cause trouble. Hmm,
the wedged -> recovery path should be triggering the submission from
inside i915_gem_reset. So it should be exploding already...
I think where we use GEM_BUG_ON(!gt.awake) in execlists, we want a
GEM_BUG_ON(!irq_pinned) in guc_submission_tasklet().
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
