These two patches fix/mitigate an issue I was seeing during suspend/resume. I would see a "blt ring initialization failed" error in the logs, followed by a NULL dereference in gen6_render_ring_flush shortly after. The first patch adds a BUG_ON to catch the NULL, and a bread crumb. The second patch fixes the init error by waiting for a little bit before failing the read (I found the register values were correct after waiting a bit). Sean Paul (2): drm/i915: Add BUG_ON when ring->private is NULL drm/i915: Add wait_for in init_ring_common drivers/gpu/drm/i915/intel_ringbuffer.c | 16 ++++++++++++---- 1 files changed, 12 insertions(+), 4 deletions(-) -- 1.7.7.3
