intel/blorp: Use wm_prog_data instead of hand-rolling our own

Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98012
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Cc: "13.0" <mesa-stable@lists.freedesktop.org>

src/intel/blorp/blorp.c

@@ -169,7 +169,7 @@ blorp_compile_fs(struct blorp_context *blorp, void *mem_ctx,
                  struct nir_shader *nir,
                  const struct brw_wm_prog_key *wm_key,
                  bool use_repclear,
-                 struct brw_blorp_prog_data *prog_data,
+                 struct brw_wm_prog_data *wm_prog_data,
                  unsigned *program_size)
 {
    const struct brw_compiler *compiler = blorp->compiler;
@@ -177,15 +177,14 @@ blorp_compile_fs(struct blorp_context *blorp, void *mem_ctx,
    nir->options =
       compiler->glsl_compiler_options[MESA_SHADER_FRAGMENT].NirOptions;
 
-   struct brw_wm_prog_data wm_prog_data;
+   memset(wm_prog_data, 0, sizeof(*wm_prog_data));
-   memset(&wm_prog_data, 0, sizeof(wm_prog_data));
 
-   wm_prog_data.base.nr_params = 0;
+   wm_prog_data->base.nr_params = 0;
-   wm_prog_data.base.param = NULL;
+   wm_prog_data->base.param = NULL;
 
    /* BLORP always just uses the first two binding table entries */
-   wm_prog_data.binding_table.render_target_start = BLORP_RENDERBUFFER_BT_INDEX;
+   wm_prog_data->binding_table.render_target_start = BLORP_RENDERBUFFER_BT_INDEX;
-   wm_prog_data.base.binding_table.texture_start = BLORP_TEXTURE_BT_INDEX;
+   wm_prog_data->base.binding_table.texture_start = BLORP_TEXTURE_BT_INDEX;
 
    nir = brw_preprocess_nir(compiler, nir);
    nir_remove_dead_variables(nir, nir_var_shader_in);
@@ -206,22 +205,9 @@ blorp_compile_fs(struct blorp_context *blorp, void *mem_ctx,
 
    const unsigned *program =
       brw_compile_fs(compiler, blorp->driver_ctx, mem_ctx, wm_key,
-                     &wm_prog_data, nir, NULL, -1, -1, false, use_repclear,
+                     wm_prog_data, nir, NULL, -1, -1, false, use_repclear,
                      NULL, program_size, NULL);
 
-   /* Copy the relavent bits of wm_prog_data over into the blorp prog data */
-   prog_data->dispatch_8 = wm_prog_data.dispatch_8;
-   prog_data->dispatch_16 = wm_prog_data.dispatch_16;
-   prog_data->first_curbe_grf_0 = wm_prog_data.base.dispatch_grf_start_reg;
-   prog_data->first_curbe_grf_2 = wm_prog_data.dispatch_grf_start_reg_2;
-   prog_data->ksp_offset_2 = wm_prog_data.prog_offset_2;
-   prog_data->persample_msaa_dispatch = wm_prog_data.persample_dispatch;
-   prog_data->flat_inputs = wm_prog_data.flat_inputs;
-   prog_data->num_varying_inputs = wm_prog_data.num_varying_inputs;
-   prog_data->inputs_read = nir->info->inputs_read;
-
-   assert(wm_prog_data.base.nr_params == 0);
-
    return program;
 }
 

src/intel/blorp/blorp_blit.c

@@ -1237,7 +1237,7 @@ brw_blorp_get_blit_kernel(struct blorp_context *blorp,
 
    const unsigned *program;
    unsigned program_size;
-   struct brw_blorp_prog_data prog_data;
+   struct brw_wm_prog_data prog_data;
 
    nir_shader *nir = brw_blorp_build_nir_shader(blorp, mem_ctx, prog_key);
    struct brw_wm_prog_key wm_key;

src/intel/blorp/blorp_clear.c

@@ -73,7 +73,7 @@ blorp_params_get_clear_kernel(struct blorp_context *blorp,
    struct brw_wm_prog_key wm_key;
    brw_blorp_init_wm_prog_key(&wm_key);
 
-   struct brw_blorp_prog_data prog_data;
+   struct brw_wm_prog_data prog_data;
    unsigned program_size;
    const unsigned *program =
       blorp_compile_fs(blorp, mem_ctx, b.shader, &wm_key, use_replicated_data,

src/intel/blorp/blorp_genX_exec.h

@@ -207,7 +207,8 @@ blorp_emit_input_varying_data(struct blorp_batch *batch,
    for (unsigned i = 0; i < max_num_varyings; i++) {
       const gl_varying_slot attr = VARYING_SLOT_VAR0 + i;
 
-      if (!(params->wm_prog_data->inputs_read & (1ull << attr)))
+      const int input_index = params->wm_prog_data->urb_setup[attr];
+      if (input_index < 0)
          continue;
 
       memcpy(inputs, inputs_src + i * 4, vec4_size_in_bytes);
@@ -401,7 +402,7 @@ static void
 blorp_emit_sf_config(struct blorp_batch *batch,
                      const struct blorp_params *params)
 {
-   const struct brw_blorp_prog_data *prog_data = params->wm_prog_data;
+   const struct brw_wm_prog_data *prog_data = params->wm_prog_data;
 
    /* 3DSTATE_SF
     *
@@ -502,7 +503,7 @@ static void
 blorp_emit_ps_config(struct blorp_batch *batch,
                      const struct blorp_params *params)
 {
-   const struct brw_blorp_prog_data *prog_data = params->wm_prog_data;
+   const struct brw_wm_prog_data *prog_data = params->wm_prog_data;
 
    /* Even when thread dispatch is disabled, max threads (dw5.25:31) must be
     * nonzero to prevent the GPU from hanging.  While the documentation doesn't
@@ -527,16 +528,16 @@ blorp_emit_ps_config(struct blorp_batch *batch,
 
       if (prog_data) {
          ps.DispatchGRFStartRegisterForConstantSetupData0 =
-            prog_data->first_curbe_grf_0;
+            prog_data->base.dispatch_grf_start_reg;
          ps.DispatchGRFStartRegisterForConstantSetupData2 =
-            prog_data->first_curbe_grf_2;
+            prog_data->dispatch_grf_start_reg_2;
 
          ps._8PixelDispatchEnable = prog_data->dispatch_8;
          ps._16PixelDispatchEnable = prog_data->dispatch_16;
 
          ps.KernelStartPointer0 = params->wm_prog_kernel;
          ps.KernelStartPointer2 =
-            params->wm_prog_kernel + prog_data->ksp_offset_2;
+            params->wm_prog_kernel + prog_data->prog_offset_2;
       }
 
       /* 3DSTATE_PS expects the number of threads per PSD, which is always 64;
@@ -577,7 +578,7 @@ blorp_emit_ps_config(struct blorp_batch *batch,
       if (prog_data) {
          psx.PixelShaderValid = true;
          psx.AttributeEnable = prog_data->num_varying_inputs > 0;
-         psx.PixelShaderIsPerSample = prog_data->persample_msaa_dispatch;
+         psx.PixelShaderIsPerSample = prog_data->persample_dispatch;
       }
 
       if (params->src.enabled)
@@ -612,7 +613,7 @@ blorp_emit_ps_config(struct blorp_batch *batch,
       if (params->dst.surf.samples > 1) {
          wm.MultisampleRasterizationMode = MSRASTMODE_ON_PATTERN;
          wm.MultisampleDispatchMode =
-            (prog_data && prog_data->persample_msaa_dispatch) ?
+            (prog_data && prog_data->persample_dispatch) ?
             MSDISPMODE_PERSAMPLE : MSDISPMODE_PERPIXEL;
       } else {
          wm.MultisampleRasterizationMode = MSRASTMODE_OFF_PIXEL;
@@ -630,13 +631,13 @@ blorp_emit_ps_config(struct blorp_batch *batch,
 
       if (prog_data) {
          ps.DispatchGRFStartRegisterforConstantSetupData0 =
-            prog_data->first_curbe_grf_0;
+            prog_data->base.dispatch_grf_start_reg;
          ps.DispatchGRFStartRegisterforConstantSetupData2 =
-            prog_data->first_curbe_grf_2;
+            prog_data->dispatch_grf_start_reg_2;
 
          ps.KernelStartPointer0 = params->wm_prog_kernel;
          ps.KernelStartPointer2 =
-            params->wm_prog_kernel + prog_data->ksp_offset_2;
+            params->wm_prog_kernel + prog_data->prog_offset_2;
 
          ps._8PixelDispatchEnable = prog_data->dispatch_8;
          ps._16PixelDispatchEnable = prog_data->dispatch_16;
@@ -692,13 +693,13 @@ blorp_emit_ps_config(struct blorp_batch *batch,
          wm.ThreadDispatchEnable = true;
 
          wm.DispatchGRFStartRegisterforConstantSetupData0 =
-            prog_data->first_curbe_grf_0;
+            prog_data->base.dispatch_grf_start_reg;
          wm.DispatchGRFStartRegisterforConstantSetupData2 =
-            prog_data->first_curbe_grf_2;
+            prog_data->dispatch_grf_start_reg_2;
 
          wm.KernelStartPointer0 = params->wm_prog_kernel;
          wm.KernelStartPointer2 =
-            params->wm_prog_kernel + prog_data->ksp_offset_2;
+            params->wm_prog_kernel + prog_data->prog_offset_2;
 
          wm._8PixelDispatchEnable = prog_data->dispatch_8;
          wm._16PixelDispatchEnable = prog_data->dispatch_16;
@@ -714,7 +715,7 @@ blorp_emit_ps_config(struct blorp_batch *batch,
       if (params->dst.surf.samples > 1) {
          wm.MultisampleRasterizationMode = MSRASTMODE_ON_PATTERN;
          wm.MultisampleDispatchMode =
-            (prog_data && prog_data->persample_msaa_dispatch) ?
+            (prog_data && prog_data->persample_dispatch) ?
             MSDISPMODE_PERSAMPLE : MSDISPMODE_PERPIXEL;
       } else {
          wm.MultisampleRasterizationMode = MSRASTMODE_OFF_PIXEL;

src/intel/blorp/blorp_priv.h

@@ -138,33 +138,8 @@ struct brw_blorp_wm_inputs
    uint32_t pad[1];
 };
 
-struct brw_blorp_prog_data
-{
-   bool dispatch_8;
-   bool dispatch_16;
-
-   uint8_t first_curbe_grf_0;
-   uint8_t first_curbe_grf_2;
-
-   uint32_t ksp_offset_2;
-
-   /**
-    * True if the WM program should be run in MSDISPMODE_PERSAMPLE with more
-    * than one sample per pixel.
-    */
-   bool persample_msaa_dispatch;
-
-   /**
-    * Mask of which FS inputs are marked flat by the shader source.  This is
-    * needed for setting up 3DSTATE_SF/SBE.
-    */
-   uint32_t flat_inputs;
-   unsigned num_varying_inputs;
-   uint64_t inputs_read;
-};
-
 static inline unsigned
-brw_blorp_get_urb_length(const struct brw_blorp_prog_data *prog_data)
+brw_blorp_get_urb_length(const struct brw_wm_prog_data *prog_data)
 {
    if (prog_data == NULL)
       return 1;
@@ -197,7 +172,7 @@ struct blorp_params
    unsigned num_draw_buffers;
    unsigned num_layers;
    uint32_t wm_prog_kernel;
-   struct brw_blorp_prog_data *wm_prog_data;
+   struct brw_wm_prog_data *wm_prog_data;
 };
 
 void blorp_params_init(struct blorp_params *params);
@@ -314,7 +289,7 @@ blorp_compile_fs(struct blorp_context *blorp, void *mem_ctx,
                  struct nir_shader *nir,
                  const struct brw_wm_prog_key *wm_key,
                  bool use_repclear,
-                 struct brw_blorp_prog_data *prog_data,
+                 struct brw_wm_prog_data *wm_prog_data,
                  unsigned *program_size);
 
 /** \} */