drm/i915/ringbuffer: Remove irq-seqno w/a for gen6/7 rcs
authorChris Wilson <chris@chris-wilson.co.uk>
Fri, 28 Dec 2018 17:16:37 +0000 (17:16 +0000)
committerChris Wilson <chris@chris-wilson.co.uk>
Mon, 31 Dec 2018 15:35:45 +0000 (15:35 +0000)
Having transitioned to using PIPECONTROL to combine the flush with the
breadcrumb write using their post-sync functions, assume that this will
resolve the serialisation with the subsequent MI_USER_INTERRUPT. That is
when inspecting the breadcrumb after an interrupt we can rely on the write
being posted (i.e. the HWSP will be coherent).

Testing using gem_sync shows that the PIPECONTROL + CS stall does
serialise the command streamer sufficient that the breadcrumb lands
before the MI_USER_INTERRUPT. The same is not true for MI_FLUSH_DW.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20181228171641.16531-2-chris@chris-wilson.co.uk
drivers/gpu/drm/i915/intel_ringbuffer.c

index d773f7d..1b92648 100644 (file)
@@ -2218,13 +2218,11 @@ int intel_init_render_ring_buffer(struct intel_engine_cs *engine)
                engine->emit_flush = gen7_render_ring_flush;
                engine->emit_breadcrumb = gen7_rcs_emit_breadcrumb;
                engine->emit_breadcrumb_sz = gen7_rcs_emit_breadcrumb_sz;
-               engine->irq_seqno_barrier = gen6_seqno_barrier;
        } else if (IS_GEN(dev_priv, 6)) {
                engine->init_context = intel_rcs_ctx_init;
                engine->emit_flush = gen6_render_ring_flush;
                engine->emit_breadcrumb = gen6_rcs_emit_breadcrumb;
                engine->emit_breadcrumb_sz = gen6_rcs_emit_breadcrumb_sz;
-               engine->irq_seqno_barrier = gen6_seqno_barrier;
        } else if (IS_GEN(dev_priv, 5)) {
                engine->emit_flush = gen4_render_ring_flush;
        } else {