Quoting Patchwork (2017-10-07 06:19:37) > For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_5938/shards.html No surprises this time, so pushed. Need to sit down and think about the context issue. It's a bad interaction between context-pinning and evict_for_something, I knew it existed (as it was explicitly avoided in the original context switching code which used active references to avoid pin-until-retirement), just hadn't expected to ever see it. I think it may just be as simple as always waiting-for-idle on ENOSPC if ggtt. Hmm, which iirc we should already be doing... -Chris > _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
