[Mesa-stable] [PATCH] anv: Implement a depth stall restriction on gen7

[Jason Ekstrand]

Fixes 61 Vulkan CTS tests on Haswell

Cc: "13.0" <mesa-stable at lists.freedesktop.org>
---
 src/intel/vulkan/anv_genX.h        |  2 ++
 src/intel/vulkan/genX_blorp_exec.c |  2 ++
 src/intel/vulkan/genX_cmd_buffer.c | 31 +++++++++++++++++++++++++++++++
 3 files changed, 35 insertions(+)

diff --git a/src/intel/vulkan/anv_genX.h b/src/intel/vulkan/anv_genX.h
index 7937170..296b1ae 100644
--- a/src/intel/vulkan/anv_genX.h
+++ b/src/intel/vulkan/anv_genX.h
@@ -42,6 +42,8 @@ void genX(cmd_buffer_emit_state_base_address)(struct anv_cmd_buffer *cmd_buffer)
 
 void genX(cmd_buffer_apply_pipe_flushes)(struct anv_cmd_buffer *cmd_buffer);
 
+void genX(cmd_buffer_emit_gen7_depth_flush)(struct anv_cmd_buffer *cmd_buffer);
+
 void genX(flush_pipeline_select_3d)(struct anv_cmd_buffer *cmd_buffer);
 void genX(flush_pipeline_select_gpgpu)(struct anv_cmd_buffer *cmd_buffer);
 
diff --git a/src/intel/vulkan/genX_blorp_exec.c b/src/intel/vulkan/genX_blorp_exec.c
index a705de0..721b6e5 100644
--- a/src/intel/vulkan/genX_blorp_exec.c
+++ b/src/intel/vulkan/genX_blorp_exec.c
@@ -150,6 +150,8 @@ genX(blorp_exec)(struct blorp_batch *batch,
 
    genX(flush_pipeline_select_3d)(cmd_buffer);
 
+   genX(cmd_buffer_emit_gen7_depth_flush)(cmd_buffer);
+
    blorp_exec(batch, params);
 
    cmd_buffer->state.vb_dirty = ~0;
diff --git a/src/intel/vulkan/genX_cmd_buffer.c b/src/intel/vulkan/genX_cmd_buffer.c
index 985bae0..b873a69 100644
--- a/src/intel/vulkan/genX_cmd_buffer.c
+++ b/src/intel/vulkan/genX_cmd_buffer.c
@@ -1978,6 +1978,35 @@ genX(flush_pipeline_select_gpgpu)(struct anv_cmd_buffer *cmd_buffer)
    }
 }
 
+void
+genX(cmd_buffer_emit_gen7_depth_flush)(struct anv_cmd_buffer *cmd_buffer)
+{
+   if (GEN_GEN > 7)
+      return;
+
+   /* From the Haswell PRM, documentation for 3DSTATE_DEPTH_BUFFER:
+    *
+    *    "Restriction: Prior to changing Depth/Stencil Buffer state (i.e., any
+    *    combination of 3DSTATE_DEPTH_BUFFER, 3DSTATE_CLEAR_PARAMS,
+    *    3DSTATE_STENCIL_BUFFER, 3DSTATE_HIER_DEPTH_BUFFER) SW must first
+    *    issue a pipelined depth stall (PIPE_CONTROL with Depth Stall bit
+    *    set), followed by a pipelined depth cache flush (PIPE_CONTROL with
+    *    Depth Flush Bit set, followed by another pipelined depth stall
+    *    (PIPE_CONTROL with Depth Stall Bit set), unless SW can otherwise
+    *    guarantee that the pipeline from WM onwards is already flushed (e.g.,
+    *    via a preceding MI_FLUSH)."
+    */
+   anv_batch_emit(&cmd_buffer->batch, GENX(PIPE_CONTROL), pipe) {
+      pipe.DepthStallEnable = true;
+   }
+   anv_batch_emit(&cmd_buffer->batch, GENX(PIPE_CONTROL), pipe) {
+      pipe.DepthCacheFlushEnable = true;
+   }
+   anv_batch_emit(&cmd_buffer->batch, GENX(PIPE_CONTROL), pipe) {
+      pipe.DepthStallEnable = true;
+   }
+}
+
 static void
 cmd_buffer_emit_depth_stencil(struct anv_cmd_buffer *cmd_buffer)
 {
@@ -1994,6 +2023,8 @@ cmd_buffer_emit_depth_stencil(struct anv_cmd_buffer *cmd_buffer)
    /* FIXME: Implement the PMA stall W/A */
    /* FIXME: Width and Height are wrong */
 
+   genX(cmd_buffer_emit_gen7_depth_flush)(cmd_buffer);
+
    /* Emit 3DSTATE_DEPTH_BUFFER */
    if (has_depth) {
       anv_batch_emit(&cmd_buffer->batch, GENX(3DSTATE_DEPTH_BUFFER), db) {
-- 
2.5.0.400.gff86faf