drm/i915: Force the CS stall for invalidate flushes
authorChris Wilson <chris@chris-wilson.co.uk>
Tue, 16 Dec 2014 08:44:32 +0000 (08:44 +0000)
committerJani Nikula <jani.nikula@intel.com>
Tue, 16 Dec 2014 13:06:48 +0000 (15:06 +0200)
In order to act as a full command barrier by itself, we need to tell the
pipecontrol to actually stall the command streamer while the flush runs.
We require the full command barrier before operations like
MI_SET_CONTEXT, which currently rely on a prior invalidate flush.

References: https://bugs.freedesktop.org/show_bug.cgi?id=83677
Cc: Simon Farnsworth <simon@farnz.org.uk>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@vger.kernel.org
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
drivers/gpu/drm/i915/intel_ringbuffer.c

index 3d6bc8d56e675f6ea02c30044d4c3b570fc2c22c..c7bc93d28d84ec4356c0c5f5c4e4cd67296df709 100644 (file)
@@ -369,6 +369,8 @@ gen7_render_ring_flush(struct intel_engine_cs *ring,
                flags |= PIPE_CONTROL_QW_WRITE;
                flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
 
+               flags |= PIPE_CONTROL_STALL_AT_SCOREBOARD;
+
                /* Workaround: we must issue a pipe_control with CS-stall bit
                 * set before a pipe_control command that has the state cache
                 * invalidate bit set. */
This page took 0.026069 seconds and 5 git commands to generate.