gen4_render_ring_flush(struct i915_request *rq, u32 mode)
{
u32 cmd, *cs;
+ int i;
/*
* read/write caches:
cmd |= MI_INVALIDATE_ISP;
}
- cs = intel_ring_begin(rq, 2);
+ i = 2;
+ if (mode & EMIT_INVALIDATE)
+ i += 20;
+
+ cs = intel_ring_begin(rq, i);
if (IS_ERR(cs))
return PTR_ERR(cs);
*cs++ = cmd;
- *cs++ = MI_NOOP;
+
+ /*
+ * A random delay to let the CS invalidate take effect? Without this
+ * delay, the GPU relocation path fails as the CS does not see
+ * the updated contents. Just as important, if we apply the flushes
+ * to the EMIT_FLUSH branch (i.e. immediately after the relocation
+ * write and before the invalidate on the next batch), the relocations
+ * still fail. This implies that is a delay following invalidation
+ * that is required to reset the caches as opposed to a delay to
+ * ensure the memory is written.
+ */
+ if (mode & EMIT_INVALIDATE) {
+ *cs++ = GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE;
+ *cs++ = i915_ggtt_offset(rq->engine->scratch) |
+ PIPE_CONTROL_GLOBAL_GTT;
+ *cs++ = 0;
+ *cs++ = 0;
+
+ for (i = 0; i < 12; i++)
+ *cs++ = MI_FLUSH;
+
+ *cs++ = GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE;
+ *cs++ = i915_ggtt_offset(rq->engine->scratch) |
+ PIPE_CONTROL_GLOBAL_GTT;
+ *cs++ = 0;
+ *cs++ = 0;
+ }
+
+ *cs++ = cmd;
+
intel_ring_advance(rq, cs);
return 0;
intel_engine_reset_breadcrumbs(engine);
+ if (HAS_LEGACY_SEMAPHORES(engine->i915)) {
+ I915_WRITE(RING_SYNC_0(engine->mmio_base), 0);
+ I915_WRITE(RING_SYNC_1(engine->mmio_base), 0);
+ if (HAS_VEBOX(dev_priv))
+ I915_WRITE(RING_SYNC_2(engine->mmio_base), 0);
+ }
+
/* Enforce ordering by reading HEAD register back */
I915_READ_HEAD(engine);
/* Check that the ring offsets point within the ring! */
GEM_BUG_ON(!intel_ring_offset_valid(ring, ring->head));
GEM_BUG_ON(!intel_ring_offset_valid(ring, ring->tail));
-
intel_ring_update_space(ring);
+
+ /* First wake the ring up to an empty/idle ring */
I915_WRITE_HEAD(engine, ring->head);
- I915_WRITE_TAIL(engine, ring->tail);
+ I915_WRITE_TAIL(engine, ring->head);
(void)I915_READ_TAIL(engine);
I915_WRITE_CTL(engine, RING_CTL_SIZE(ring->size) | RING_VALID);
if (INTEL_GEN(dev_priv) > 2)
I915_WRITE_MODE(engine, _MASKED_BIT_DISABLE(STOP_RING));
+ /* Now awake, let it get started */
+ if (ring->tail != ring->head) {
+ I915_WRITE_TAIL(engine, ring->tail);
+ (void)I915_READ_TAIL(engine);
+ }
+
/* Papering over lost _interrupts_ immediately following the restart */
intel_engine_wakeup(engine);
out:
static void reset_ring(struct intel_engine_cs *engine, struct i915_request *rq)
{
- GEM_TRACE("%s seqno=%x\n", engine->name, rq ? rq->global_seqno : 0);
+ GEM_TRACE("%s request global=%d, current=%d\n",
+ engine->name, rq ? rq->global_seqno : 0,
+ intel_engine_get_seqno(engine));
/*
* Try to restore the logical GPU state to match the continuation
if (ret)
return ret;
- intel_whitelist_workarounds_apply(engine);
-
/* WaTimedSingleVertexDispatch:cl,bw,ctg,elk,ilk,snb */
if (IS_GEN(dev_priv, 4, 6))
I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(VS_TIMER_DISPATCH));
int intel_ring_pin(struct intel_ring *ring)
{
struct i915_vma *vma = ring->vma;
- enum i915_map_type map =
- HAS_LLC(vma->vm->i915) ? I915_MAP_WB : I915_MAP_WC;
+ enum i915_map_type map = i915_coherent_map_type(vma->vm->i915);
unsigned int flags;
void *addr;
int ret;