i965/fs: Add support for translating ir_triop_fma into MAD.

[mesa.git] / src / mesa / drivers / dri / i965 / brw_wm_surface_state.c

diff --git a/src/mesa/drivers/dri/i965/brw_wm_surface_state.c b/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
index 3b67bd2884eab89abf9a871ebccc34474a819f7b..e2c7b77472d86ed7cdf50eb51caf1d2d2bff1b31 100644 (file)
--- a/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_surface_state.c
@@ -347,9 +347,9 @@ brw_create_constant_surface(struct brw_context *brw,
    surf[4] = 0;
    surf[5] = 0;
 
-   /* Emit relocation to surface contents.  Section 5.1.1 of the gen4
-    * bspec ("Data Cache") says that the data cache does not exist as
-    * a separate cache and is just the sampler cache.
+   /* Emit relocation to surface contents.  The 965 PRM, Volume 4, section
+    * 5.1.2 "Data Cache" says: "the data cache does not exist as a separate
+    * physical cache.  It is mapped in hardware to the sampler cache."
     */
    drm_intel_bo_emit_reloc(brw->batch.bo,
                           *out_offset + 4,
@@ -451,7 +451,7 @@ brw_update_sol_surface(struct brw_context *brw,
 static void
 brw_upload_wm_pull_constants(struct brw_context *brw)
 {
-   struct gl_context *ctx = &brw->intel.ctx;
+   struct gl_context *ctx = &brw->ctx;
    /* BRW_NEW_FRAGMENT_PROGRAM */
    struct brw_fragment_program *fp =
       (struct brw_fragment_program *) brw->fragment_program;
@@ -523,7 +523,7 @@ brw_update_null_renderbuffer_surface(struct brw_context *brw, unsigned int unit)
     *
     *     - Surface Format must be R8G8B8A8_UNORM.
     */
-   struct gl_context *ctx = &brw->intel.ctx;
+   struct gl_context *ctx = &brw->ctx;
    uint32_t *surf;
    unsigned surface_type = BRW_SURFACE_NULL;
    drm_intel_bo *bo = NULL;
@@ -533,8 +533,8 @@ brw_update_null_renderbuffer_surface(struct brw_context *brw, unsigned int unit)
    /* _NEW_BUFFERS */
    const struct gl_framebuffer *fb = ctx->DrawBuffer;
 
-   surf = brw_state_batch(brw, AUB_TRACE_SURFACE_STATE,
-                         6 * 4, 32, &brw->wm.surf_offset[unit]);
+   surf = brw_state_batch(brw, AUB_TRACE_SURFACE_STATE, 6 * 4, 32,
+                          &brw->wm.surf_offset[SURF_INDEX_DRAW(unit)]);
 
    if (fb->Visual.samples > 1) {
       /* On Gen6, null render targets seem to cause GPU hangs when
@@ -587,7 +587,7 @@ brw_update_null_renderbuffer_surface(struct brw_context *brw, unsigned int unit)
 
    if (bo) {
       drm_intel_bo_emit_reloc(brw->batch.bo,
-                              brw->wm.surf_offset[unit] + 4,
+                              brw->wm.surf_offset[SURF_INDEX_DRAW(unit)] + 4,
                               bo, 0,
                               I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER);
    }
@@ -604,7 +604,7 @@ brw_update_renderbuffer_surface(struct brw_context *brw,
                                bool layered,
                                unsigned int unit)
 {
-   struct gl_context *ctx = &brw->intel.ctx;
+   struct gl_context *ctx = &brw->ctx;
    struct intel_renderbuffer *irb = intel_renderbuffer(rb);
    struct intel_mipmap_tree *mt = irb->mt;
    struct intel_region *region;
@@ -635,8 +635,8 @@ brw_update_renderbuffer_surface(struct brw_context *brw,
 
    region = irb->mt->region;
 
-   surf = brw_state_batch(brw, AUB_TRACE_SURFACE_STATE,
-                         6 * 4, 32, &brw->wm.surf_offset[unit]);
+   surf = brw_state_batch(brw, AUB_TRACE_SURFACE_STATE, 6 * 4, 32,
+                          &brw->wm.surf_offset[SURF_INDEX_DRAW(unit)]);
 
    format = brw->render_target_format[rb_format];
    if (unlikely(!brw->format_supported_as_render_target[rb_format])) {
@@ -692,7 +692,7 @@ brw_update_renderbuffer_surface(struct brw_context *brw,
    }
 
    drm_intel_bo_emit_reloc(brw->batch.bo,
-                          brw->wm.surf_offset[unit] + 4,
+                          brw->wm.surf_offset[SURF_INDEX_DRAW(unit)] + 4,
                           region->bo,
                           surf[1] - region->bo->offset,
                           I915_GEM_DOMAIN_RENDER,
@@ -705,7 +705,7 @@ brw_update_renderbuffer_surface(struct brw_context *brw,
 static void
 brw_update_renderbuffer_surfaces(struct brw_context *brw)
 {
-   struct gl_context *ctx = &brw->intel.ctx;
+   struct gl_context *ctx = &brw->ctx;
    GLuint i;
 
    /* _NEW_BUFFERS | _NEW_COLOR */
@@ -750,7 +750,7 @@ const struct brw_tracked_state gen6_renderbuffer_surfaces = {
 static void
 brw_update_texture_surfaces(struct brw_context *brw)
 {
-   struct gl_context *ctx = &brw->intel.ctx;
+   struct gl_context *ctx = &brw->ctx;
 
    /* BRW_NEW_VERTEX_PROGRAM and BRW_NEW_FRAGMENT_PROGRAM:
     * Unfortunately, we're stuck using the gl_program structs until the
@@ -809,7 +809,7 @@ brw_upload_ubo_surfaces(struct brw_context *brw,
                        struct gl_shader *shader,
                        uint32_t *surf_offsets)
 {
-   struct gl_context *ctx = &brw->intel.ctx;
+   struct gl_context *ctx = &brw->ctx;
 
    if (!shader)
       return;
@@ -839,7 +839,7 @@ brw_upload_ubo_surfaces(struct brw_context *brw,
 static void
 brw_upload_wm_ubo_surfaces(struct brw_context *brw)
 {
-   struct gl_context *ctx = &brw->intel.ctx;
+   struct gl_context *ctx = &brw->ctx;
    /* _NEW_PROGRAM */
    struct gl_shader_program *prog = ctx->Shader._CurrentFragmentProgram;
 
@@ -873,15 +873,14 @@ brw_upload_wm_binding_table(struct brw_context *brw)
       gen7_create_shader_time_surface(brw, &brw->wm.surf_offset[SURF_INDEX_WM_SHADER_TIME]);
    }
 
-   /* Might want to calculate nr_surfaces first, to avoid taking up so much
-    * space for the binding table.
-    */
+   /* CACHE_NEW_WM_PROG */
+   unsigned entries = brw->wm.prog_data->binding_table_size;
    bind = brw_state_batch(brw, AUB_TRACE_BINDING_TABLE,
-                         sizeof(uint32_t) * BRW_MAX_WM_SURFACES,
+                         sizeof(uint32_t) * entries,
                          32, &brw->wm.bind_bo_offset);
 
    /* BRW_NEW_SURFACES */
-   for (i = 0; i < BRW_MAX_WM_SURFACES; i++) {
+   for (i = 0; i < entries; i++) {
       bind[i] = brw->wm.surf_offset[i];
    }
 
@@ -893,7 +892,7 @@ const struct brw_tracked_state brw_wm_binding_table = {
       .mesa = 0,
       .brw = (BRW_NEW_BATCH |
              BRW_NEW_SURFACES),
-      .cache = 0
+      .cache = CACHE_NEW_WM_PROG
    },
    .emit = brw_upload_wm_binding_table,
 };