On Wed, Jan 11, 2017 at 03:13:29PM +0100, Michal Wajdeczko wrote:
> > + vma = i915_gem_object_ggtt_pin(huc_fw->obj, NULL, 0, 0, 0);
> > + if (IS_ERR(vma)) {
> > + DRM_DEBUG_DRIVER("pin failed %d\n", (int)PTR_ERR(vma));
> > + return PTR_ERR(vma);
> > + }
Just asking a stupid question: Does the HuC have the same limitation as
the GuC on not being able to map certain ranges of the GuC? From the
earlier discussion on the failures, I got the impression the HuC had the
same limitations.
> > +
> > + /* Invalidate GuC TLB to let GuC take the latest updates to GTT. */
> > + I915_WRITE(GEN8_GTCR, GEN8_GTCR_INVALIDATE);
> > +
> > + intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL);
> > +
> > + /* init WOPCM */
> > + I915_WRITE(GUC_WOPCM_SIZE, intel_guc_wopcm_size(dev_priv));
> > + I915_WRITE(DMA_GUC_WOPCM_OFFSET, GUC_WOPCM_OFFSET_VALUE |
> > + HUC_LOADING_AGENT_GUC);
> > +
> > + /* Set the source address for the uCode */
> > + offset = i915_ggtt_offset(vma) + huc_fw->header_offset;
If huc does have the same limits as the guc, please use guc_ggtt_offset()
for the extra verification on the address before use.
-Chris
--
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
