mesa,driconf: Add WA to initialize vertex program outputs to vec4(0,0,0,1)

Per ARB_vertex_program spec result registers are 4-component and initially
undefined, and the FF fragment program expects its intputs to be
4-component too. So, if the client's vertex program does not write the
whole vector it will cause misrenderings unless the same client also
supplies fragment program that expects less than 4 componens.

This commit adds a workaround that initializes results to vec4(0, 0, 0, 1)
which seems to be an expected behavior for such clients.

Cc: mesa-stable
Signed-off-by: Sviatoslav Peleshko <sviatoslav.peleshko@globallogic.com>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
(cherry picked from commit f03432c81a)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38432>

.pick_status.json

@@ -994,7 +994,7 @@
         "description": "mesa,driconf: Add WA to initialize vertex program outputs to vec4(0,0,0,1)",
         "nominated": true,
         "nomination_type": 1,
-        "resolution": 0,
+        "resolution": 1,
         "main_sha": null,
         "because_sha": null,
         "notes": null

src/gallium/auxiliary/pipe-loader/driinfo_gallium.h

@@ -64,6 +64,7 @@ DRI_CONF_SECTION_END
 DRI_CONF_SECTION_MISCELLANEOUS
    DRI_CONF_ALWAYS_HAVE_DEPTH_BUFFER(false)
    DRI_CONF_GLSL_ZERO_INIT(false)
+   DRI_CONF_VERTEX_PROGRAM_DEFAULT_OUT(false)
    DRI_CONF_VS_POSITION_ALWAYS_INVARIANT(false)
    DRI_CONF_VS_POSITION_ALWAYS_PRECISE(false)
    DRI_CONF_ALLOW_RGB10_CONFIGS(true)

src/gallium/auxiliary/util/u_driconf.c

@@ -76,6 +76,7 @@ u_driconf_fill_st_options(struct st_config_options *options,
    query_string_option(force_gl_renderer);
    query_string_option(mesa_extension_override);
    query_bool_option(allow_multisampled_copyteximage);
+   query_bool_option(vertex_program_default_out);
 
    driComputeOptionsSha1(optionCache, options->config_options_sha1);
 }

src/gallium/include/frontend/api.h

@@ -205,6 +205,7 @@ struct st_config_options
    char *force_gl_renderer;
    char *mesa_extension_override;
    bool allow_multisampled_copyteximage;
+   bool vertex_program_default_out;
 
    unsigned char config_options_sha1[20];
 };

src/mesa/main/consts_exts.h

@@ -944,6 +944,9 @@ struct gl_constants
     */
    bool GLThreadNopCheckFramebufferStatus;
 
+   /** (driconf) Initialize outputs of vertex program to a default value vec4(0, 0, 0, 1) */
+   GLboolean VertexProgramDefaultOut;
+
    /** GL_ARB_sparse_texture */
    GLuint MaxSparseTextureSize;
    GLuint MaxSparse3DTextureSize;

src/mesa/program/prog_to_nir.c

@@ -784,6 +784,13 @@ setup_registers_and_variables(struct ptn_compile *c)
        * the shader.
        */
       c->output_regs[i] = nir_decl_reg(b, 4, 32, 0);
+
+      /* Initialize output registers with default value vec4(0, 0, 0, 1) */
+      if (c->ctx->Const.VertexProgramDefaultOut &&
+          c->prog->info.stage == MESA_SHADER_VERTEX &&
+          i != VARYING_SLOT_FOGC && i <= VARYING_SLOT_TEX7) {
+         nir_store_reg(b, nir_imm_vec4(b, 0, 0, 0, 1), c->output_regs[i]);
+      }
    }
 
    /* Create temporary registers. */

src/mesa/state_tracker/st_extensions.c

@@ -1331,6 +1331,8 @@ void st_init_extensions(struct pipe_screen *screen,
       consts->GLSLZeroInit = screen->caps.glsl_zero_init;
    }
 
+   consts->VertexProgramDefaultOut = options->vertex_program_default_out;
+
    if (extensions->EXT_semaphore) {
       consts->MaxTimelineSemaphoreValueDifference = screen->caps.max_timeline_semaphore_difference;
       extensions->NV_timeline_semaphore = consts->MaxTimelineSemaphoreValueDifference > 0;

src/util/driconf.h

@@ -520,6 +520,10 @@
    DRI_CONF_OPT_B(allow_multisampled_copyteximage, def, \
                   "Allow CopyTexSubImage and other to copy sampled framebuffer")
 
+#define DRI_CONF_VERTEX_PROGRAM_DEFAULT_OUT(def) \
+   DRI_CONF_OPT_B(vertex_program_default_out, def, \
+                  "Initialize outputs of vertex program to a default value vec4(0, 0, 0, 1)")
+
 #define DRI_CONF_CUSTOM_BORDER_COLORS_WITHOUT_FORMAT(def) \
    DRI_CONF_OPT_B(custom_border_colors_without_format, def, \
                   "Enable custom border colors without format")