drm/i915: read HEAD register back in init_ring_common() to enforce ordering
authorJiri Kosina <jkosina@suse.cz>
Thu, 7 Aug 2014 14:29:53 +0000 (16:29 +0200)
committerDaniel Vetter <daniel.vetter@ffwll.ch>
Fri, 8 Aug 2014 14:22:18 +0000 (16:22 +0200)
Withtout this, ring initialization fails reliabily during resume with

[drm:init_ring_common] *ERROR* render ring initialization failed ctl 0001f001 head ffffff8804 tail 00000000 start 000e4000

This is not a complete fix, but it is verified to make the ring
initialization failures during resume much less likely.

We were not able to root-cause this bug (likely HW-specific to Gen4 chips)
yet. This is therefore used as a ducttape before problem is fully
understood and proper fix created, so that people don't suffer from
completely unusable systems in the meantime.

The discussion and debugging is happening at

https://bugs.freedesktop.org/show_bug.cgi?id=76554

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
drivers/gpu/drm/i915/intel_ringbuffer.c

index 05969f03c0c1bd03188f9c26a13a8395382ec708..16371a444426d190d73b4baddd38cd8abeae66b2 100644 (file)
@@ -532,6 +532,9 @@ static int init_ring_common(struct intel_engine_cs *ring)
        else
                ring_setup_phys_status_page(ring);
 
+       /* Enforce ordering by reading HEAD register back */
+       I915_READ_HEAD(ring);
+
        /* Initialize the ring. This must happen _after_ we've cleared the ring
         * registers with the above sequence (the readback of the HEAD registers
         * also enforces ordering), otherwise the hw might lose the new ring