drm/radeon: Sync ME and PFP after CP semaphore waits v4

Fixes lockups due to CP read GPUVM faults when running piglit on Cape
Verde.

v2 (chk): apply the fix to R600+ as well, on CIK only the GFX CP has
	  a PFP, add more comments to R600 code, enable flushing again
v3: (agd5f): only apply to 7xx+.  r6xx does not have the packet.
v4: (agd5f): split flush change into a separate patch, fix formatting

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c
index d61bc18..af2cbc6 100644
--- a/drivers/gpu/drm/radeon/cik.c
+++ b/drivers/gpu/drm/radeon/cik.c
@@ -3920,6 +3920,17 @@
 	radeon_ring_write(ring, 0);
 }
 
+/**
+ * cik_semaphore_ring_emit - emit a semaphore on the CP ring
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon ring buffer object
+ * @semaphore: radeon semaphore object
+ * @emit_wait: Is this a sempahore wait?
+ *
+ * Emits a semaphore signal/wait packet to the CP ring and prevents the PFP
+ * from running ahead of semaphore waits.
+ */
 bool cik_semaphore_ring_emit(struct radeon_device *rdev,
 			     struct radeon_ring *ring,
 			     struct radeon_semaphore *semaphore,
@@ -3932,6 +3943,12 @@
 	radeon_ring_write(ring, lower_32_bits(addr));
 	radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) | sel);
 
+	if (emit_wait && ring->idx == RADEON_RING_TYPE_GFX_INDEX) {
+		/* Prevent the PFP from running ahead of the semaphore wait */
+		radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
+		radeon_ring_write(ring, 0x0);
+	}
+
 	return true;
 }