radv: move calculating primgroup_size to pipeline.

This moves this out of the draw paths.

Reviewed-by: Bas Nieuwenhuizen <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>

3 files changed, 11 insertions, 9 deletions

diff --git a/src/amd/vulkan/radv_pipeline.c b/src/amd/vulkan/radv_pipeline.c
index acc955f32b2..590dd67c462 100644
--- a/src/amd/vulkan/radv_pipeline.c
+++ b/src/amd/vulkan/radv_pipeline.c
@@ -2007,6 +2007,13 @@ radv_pipeline_init(struct radv_pipeline *pipeline,
 		calculate_tess_state(pipeline, pCreateInfo);
 	}
 
+	if (radv_pipeline_has_tess(pipeline))
+		pipeline->graphics.primgroup_size = pipeline->graphics.tess.num_patches;
+	else if (radv_pipeline_has_gs(pipeline))
+		pipeline->graphics.primgroup_size = 64;
+	else
+		pipeline->graphics.primgroup_size = 128; /* recommended without a GS */
+
 	const VkPipelineVertexInputStateCreateInfo *vi_info =
 		pCreateInfo->pVertexInputState;
 	struct radv_vertex_elements_info *velems = &pipeline->vertex_elements;
diff --git a/src/amd/vulkan/radv_private.h b/src/amd/vulkan/radv_private.h
index dd99d7faa78..c647efdcbcc 100644
--- a/src/amd/vulkan/radv_private.h
+++ b/src/amd/vulkan/radv_private.h
@@ -1075,6 +1075,7 @@ struct radv_pipeline {
 			uint32_t vgt_gs_mode;
 			bool vgt_primitiveid_en;
 			bool prim_restart_enable;
+			uint8_t primgroup_size;
 			unsigned esgs_ring_size;
 			unsigned gsvs_ring_size;
 			uint32_t ps_input_cntl[32];
diff --git a/src/amd/vulkan/si_cmd_buffer.c b/src/amd/vulkan/si_cmd_buffer.c
index 57823904930..9abd05583c5 100644
--- a/src/amd/vulkan/si_cmd_buffer.c
+++ b/src/amd/vulkan/si_cmd_buffer.c
@@ -683,7 +683,6 @@ si_get_ia_multi_vgt_param(struct radv_cmd_buffer *cmd_buffer,
 	enum radeon_family family = cmd_buffer->device->physical_device->rad_info.family;
 	struct radeon_info *info = &cmd_buffer->device->physical_device->rad_info;
 	unsigned prim = cmd_buffer->state.pipeline->graphics.prim;
-	unsigned primgroup_size = 128; /* recommended without a GS */
 	unsigned max_primgroup_in_wave = 2;
 	/* SWITCH_ON_EOP(0) is always preferable. */
 	bool wd_switch_on_eop = false;
@@ -693,15 +692,10 @@ si_get_ia_multi_vgt_param(struct radv_cmd_buffer *cmd_buffer,
 	bool partial_es_wave = false;
 	bool multi_instances_smaller_than_primgroup;
 
-	if (radv_pipeline_has_tess(cmd_buffer->state.pipeline))
-		primgroup_size = cmd_buffer->state.pipeline->graphics.tess.num_patches;
-	else if (radv_pipeline_has_gs(cmd_buffer->state.pipeline))
-		primgroup_size = 64;  /* recommended with a GS */
-
 	multi_instances_smaller_than_primgroup = indirect_draw;
 	if (!multi_instances_smaller_than_primgroup && instanced_draw) {
 		uint32_t num_prims = radv_prims_for_vertices(&cmd_buffer->state.pipeline->graphics.prim_vertex_count, draw_vertex_count);
-		if (num_prims < primgroup_size)
+		if (num_prims < cmd_buffer->state.pipeline->graphics.primgroup_size)
 			multi_instances_smaller_than_primgroup = true;
 	}
 
@@ -803,7 +797,7 @@ si_get_ia_multi_vgt_param(struct radv_cmd_buffer *cmd_buffer,
 			ia_switch_on_eoi = true;
 
 		/* GS requirement. */
-		if (SI_GS_PER_ES / primgroup_size >= cmd_buffer->device->gs_table_depth - 3)
+		if (SI_GS_PER_ES / cmd_buffer->state.pipeline->graphics.primgroup_size >= cmd_buffer->device->gs_table_depth - 3)
 			partial_es_wave = true;
 
 		/* GS hw bug with single-primitive instances and SWITCH_ON_EOI.
@@ -826,7 +820,7 @@ si_get_ia_multi_vgt_param(struct radv_cmd_buffer *cmd_buffer,
 		S_028AA8_SWITCH_ON_EOI(ia_switch_on_eoi) |
 		S_028AA8_PARTIAL_VS_WAVE_ON(partial_vs_wave) |
 		S_028AA8_PARTIAL_ES_WAVE_ON(partial_es_wave) |
-		S_028AA8_PRIMGROUP_SIZE(primgroup_size - 1) |
+		S_028AA8_PRIMGROUP_SIZE(cmd_buffer->state.pipeline->graphics.primgroup_size - 1) |
 		S_028AA8_WD_SWITCH_ON_EOP(chip_class >= CIK ? wd_switch_on_eop : 0) |
 		/* The following field was moved to VGT_SHADER_STAGES_EN in GFX9. */
 		S_028AA8_MAX_PRIMGRP_IN_WAVE(chip_class == VI ?