drm/amdgpu:impl vgt_flush for VI(V5)

when shadowing enabled, tesselation app will trigger
vm fault because below three tesselation registers:

VGT_TF_RING_SIZE__CI__VI,
VGT_HS_OFFCHIP_PARAM__CI__VI,
VGT_TF_MEMORY_BASE__CI__VI,

need to be programed after vgt-flush.

Tesselation picture vm fault disappeared after vgt-flush
introduced.

v2:implement vgt-flush for CI & SI.
v3:move vgt flush inside of cntx_cntrl
v4:count vgt flush in frame_size
v5:squash in typo fix
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c

@@ -1463,6 +1463,13 @@ static void gfx_v6_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, 0x1);
 }
 
+static void gfx_v6_0_ring_emit_vgt_flush(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE, 0));
+	amdgpu_ring_write(ring, EVENT_TYPE(VGT_FLUSH) |
+		EVENT_INDEX(0));
+}
+
 /**
  * gfx_v6_0_ring_emit_hdp_invalidate - emit an hdp invalidate on the cp
  *
@@ -2802,6 +2809,8 @@ static uint64_t gfx_v6_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 
 static void gfx_v6_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
 {
+	if (flags & AMDGPU_HAVE_CTX_SWITCH)
+		gfx_v6_0_ring_emit_vgt_flush(ring);
 	amdgpu_ring_write(ring, PACKET3(PACKET3_CONTEXT_CONTROL, 1));
 	amdgpu_ring_write(ring, 0x80000000);
 	amdgpu_ring_write(ring, 0);
@@ -3265,7 +3274,7 @@ static const struct amdgpu_ring_funcs gfx_v6_0_ring_funcs_gfx = {
 		14 + 14 + 14 + /* gfx_v6_0_ring_emit_fence x3 for user fence, vm fence */
 		7 + 4 + /* gfx_v6_0_ring_emit_pipeline_sync */
 		17 + 6 + /* gfx_v6_0_ring_emit_vm_flush */
-		3, /* gfx_v6_ring_emit_cntxcntl */
+		3 + 2, /* gfx_v6_ring_emit_cntxcntl including vgt flush */
 	.emit_ib_size = 6, /* gfx_v6_0_ring_emit_ib */
 	.emit_ib = gfx_v6_0_ring_emit_ib,
 	.emit_fence = gfx_v6_0_ring_emit_fence,

drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c

@@ -2105,6 +2105,18 @@ static void gfx_v7_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, 0x20); /* poll interval */
 }
 
+static void gfx_v7_0_ring_emit_vgt_flush(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE, 0));
+	amdgpu_ring_write(ring, EVENT_TYPE(VS_PARTIAL_FLUSH) |
+		EVENT_INDEX(4));
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE, 0));
+	amdgpu_ring_write(ring, EVENT_TYPE(VGT_FLUSH) |
+		EVENT_INDEX(0));
+}
+
+
 /**
  * gfx_v7_0_ring_emit_hdp_invalidate - emit an hdp invalidate on the cp
  *
@@ -2260,6 +2272,7 @@ static void gfx_v7_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
 
 	dw2 |= 0x80000000; /* set load_enable otherwise this package is just NOPs */
 	if (flags & AMDGPU_HAVE_CTX_SWITCH) {
+		gfx_v7_0_ring_emit_vgt_flush(ring);
 		/* set load_global_config & load_global_uconfig */
 		dw2 |= 0x8001;
 		/* set load_cs_sh_regs */
@@ -5153,7 +5166,7 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_gfx = {
 		12 + 12 + 12 + /* gfx_v7_0_ring_emit_fence_gfx x3 for user fence, vm fence */
 		7 + 4 + /* gfx_v7_0_ring_emit_pipeline_sync */
 		17 + 6 + /* gfx_v7_0_ring_emit_vm_flush */
-		3, /* gfx_v7_ring_emit_cntxcntl */
+		3 + 4, /* gfx_v7_ring_emit_cntxcntl including vgt flush*/
 	.emit_ib_size = 4, /* gfx_v7_0_ring_emit_ib_gfx */
 	.emit_ib = gfx_v7_0_ring_emit_ib_gfx,
 	.emit_fence = gfx_v7_0_ring_emit_fence_gfx,

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c

@@ -6186,6 +6186,18 @@ static void gfx_v8_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
 	amdgpu_ring_write(ring, 0x20); /* poll interval */
 }
 
+static void gfx_v8_0_ring_emit_vgt_flush(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE, 0));
+	amdgpu_ring_write(ring, EVENT_TYPE(VS_PARTIAL_FLUSH) |
+		EVENT_INDEX(4));
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE, 0));
+	amdgpu_ring_write(ring, EVENT_TYPE(VGT_FLUSH) |
+		EVENT_INDEX(0));
+}
+
+
 static void gfx_v8_0_ring_emit_hdp_invalidate(struct amdgpu_ring *ring)
 {
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
@@ -6371,6 +6383,7 @@ static void gfx_v8_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
 
 	dw2 |= 0x80000000; /* set load_enable otherwise this package is just NOPs */
 	if (flags & AMDGPU_HAVE_CTX_SWITCH) {
+		gfx_v8_0_ring_emit_vgt_flush(ring);
 		/* set load_global_config & load_global_uconfig */
 		dw2 |= 0x8001;
 		/* set load_cs_sh_regs */
@@ -6574,7 +6587,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = {
 		7 + /* gfx_v8_0_ring_emit_pipeline_sync */
 		128 + 19 + /* gfx_v8_0_ring_emit_vm_flush */
 		2 + /* gfx_v8_ring_emit_sb */
-		3, /* gfx_v8_ring_emit_cntxcntl */
+		3 + 4, /* gfx_v8_ring_emit_cntxcntl including vgt flush */
 	.emit_ib_size =	4, /* gfx_v8_0_ring_emit_ib_gfx */
 	.emit_ib = gfx_v8_0_ring_emit_ib_gfx,
 	.emit_fence = gfx_v8_0_ring_emit_fence_gfx,