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

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

diff --git a/src/mesa/drivers/dri/i965/brw_draw.c b/src/mesa/drivers/dri/i965/brw_draw.c
index 01e6b2443a78d0c251df61e0feff80a587f689e8..c7164acfb02a22bc601076db29aabf0d02ae1936 100644 (file)
--- a/src/mesa/drivers/dri/i965/brw_draw.c
+++ b/src/mesa/drivers/dri/i965/brw_draw.c
@@ -54,7 +54,7 @@
 
 #define FILE_DEBUG_FLAG DEBUG_PRIMS
 
-static GLuint prim_to_hw_prim[GL_POLYGON+1] = {
+const GLuint prim_to_hw_prim[GL_POLYGON+1] = {
    _3DPRIM_POINTLIST,
    _3DPRIM_LINELIST,
    _3DPRIM_LINELOOP,
@@ -90,7 +90,7 @@ static const GLenum reduced_prim[GL_POLYGON+1] = {
 static void brw_set_prim(struct brw_context *brw,
                          const struct _mesa_prim *prim)
 {
-   struct gl_context *ctx = &brw->intel.ctx;
+   struct gl_context *ctx = &brw->ctx;
    uint32_t hw_prim = prim_to_hw_prim[prim->mode];
 
    DBG("PRIM: %s\n", _mesa_lookup_enum_by_nr(prim->mode));
@@ -160,7 +160,6 @@ static void brw_emit_prim(struct brw_context *brw,
                          const struct _mesa_prim *prim,
                          uint32_t hw_prim)
 {
-   struct intel_context *intel = &brw->intel;
    int verts_per_instance;
    int vertex_access_type;
    int start_vertex_location;
@@ -181,7 +180,7 @@ static void brw_emit_prim(struct brw_context *brw,
    }
 
    /* We only need to trim the primitive count on pre-Gen6. */
-   if (intel->gen < 6)
+   if (brw->gen < 6)
       verts_per_instance = trim(prim->mode, prim->count);
    else
       verts_per_instance = prim->count;
@@ -210,7 +209,7 @@ static void brw_emit_prim(struct brw_context *brw,
    OUT_BATCH(base_vertex_location);
    ADVANCE_BATCH();
 
-   intel->batch.need_workaround_flush = true;
+   brw->batch.need_workaround_flush = true;
 
    if (brw->always_flush_cache) {
       intel_batchbuffer_emit_mi_flush(brw);
@@ -300,7 +299,7 @@ static void brw_merge_inputs( struct brw_context *brw,
 static void
 brw_predraw_resolve_buffers(struct brw_context *brw)
 {
-   struct gl_context *ctx = &brw->intel.ctx;
+   struct gl_context *ctx = &brw->ctx;
    struct intel_renderbuffer *depth_irb;
    struct intel_texture_object *tex_obj;
 
@@ -334,12 +333,13 @@ brw_predraw_resolve_buffers(struct brw_context *brw)
  */
 static void brw_postdraw_set_buffers_need_resolve(struct brw_context *brw)
 {
-   struct gl_context *ctx = &brw->intel.ctx;
+   struct gl_context *ctx = &brw->ctx;
    struct gl_framebuffer *fb = ctx->DrawBuffer;
 
    struct intel_renderbuffer *front_irb = NULL;
    struct intel_renderbuffer *back_irb = intel_get_renderbuffer(fb, BUFFER_BACK_LEFT);
    struct intel_renderbuffer *depth_irb = intel_get_renderbuffer(fb, BUFFER_DEPTH);
+   struct gl_renderbuffer_attachment *depth_att = &fb->Attachment[BUFFER_DEPTH];
 
    if (brw->is_front_buffer_rendering)
       front_irb = intel_get_renderbuffer(fb, BUFFER_FRONT_LEFT);
@@ -349,7 +349,7 @@ static void brw_postdraw_set_buffers_need_resolve(struct brw_context *brw)
    if (back_irb)
       intel_renderbuffer_set_needs_downsample(back_irb);
    if (depth_irb && ctx->Depth.Mask)
-      intel_renderbuffer_set_needs_depth_resolve(depth_irb);
+      intel_renderbuffer_att_set_needs_depth_resolve(depth_att);
 }
 
 /* May fail if out of video memory for texture or vbo upload, or on
@@ -363,7 +363,6 @@ static bool brw_try_draw_prims( struct gl_context *ctx,
                                     GLuint min_index,
                                     GLuint max_index )
 {
-   struct intel_context *intel = intel_context(ctx);
    struct brw_context *brw = brw_context(ctx);
    bool retval = true;
    GLuint i;
@@ -372,6 +371,13 @@ static bool brw_try_draw_prims( struct gl_context *ctx,
    if (ctx->NewState)
       _mesa_update_state( ctx );
 
+   /* Find the highest sampler unit used by each shader program.  A bit-count
+    * won't work since ARB programs use the texture unit number as the sampler
+    * index.
+    */
+   brw->wm.sampler_count = _mesa_fls(ctx->FragmentProgram._Current->Base.SamplersUsed);
+   brw->vs.sampler_count = _mesa_fls(ctx->VertexProgram._Current->Base.SamplersUsed);
+
    /* We have to validate the textures *before* checking for fallbacks;
     * otherwise, the software fallback won't be able to rely on the
     * texture state, the firstLevel and lastLevel fields won't be
@@ -431,7 +437,7 @@ static bool brw_try_draw_prims( struct gl_context *ctx,
          brw->basevertex = prim->basevertex;
          brw->state.dirty.brw |= BRW_NEW_VERTICES;
       }
-      if (intel->gen < 6)
+      if (brw->gen < 6)
         brw_set_prim(brw, &prim[i]);
       else
         gen6_set_prim(brw, &prim[i]);
@@ -443,18 +449,18 @@ retry:
        * brw->state.dirty.brw.
        */
       if (brw->state.dirty.brw) {
-        intel->no_batch_wrap = true;
+        brw->no_batch_wrap = true;
         brw_upload_state(brw);
       }
 
-      if (intel->gen >= 7)
+      if (brw->gen >= 7)
         gen7_emit_prim(brw, &prim[i], brw->primitive);
       else
         brw_emit_prim(brw, &prim[i], brw->primitive);
 
-      intel->no_batch_wrap = false;
+      brw->no_batch_wrap = false;
 
-      if (dri_bufmgr_check_aperture_space(&intel->batch.bo, 1)) {
+      if (dri_bufmgr_check_aperture_space(&brw->batch.bo, 1)) {
         if (!fail_next) {
            intel_batchbuffer_reset_to_saved(brw);
            intel_batchbuffer_flush(brw);
@@ -494,7 +500,7 @@ void brw_draw_prims( struct gl_context *ctx,
                     GLuint max_index,
                     struct gl_transform_feedback_object *tfb_vertcount )
 {
-   struct intel_context *intel = intel_context(ctx);
+   struct brw_context *brw = brw_context(ctx);
    const struct gl_client_array **arrays = ctx->Array._DrawArrays;
 
    if (!_mesa_check_conditional_render(ctx))
@@ -534,7 +540,7 @@ void brw_draw_prims( struct gl_context *ctx,
 
 void brw_draw_init( struct brw_context *brw )
 {
-   struct gl_context *ctx = &brw->intel.ctx;
+   struct gl_context *ctx = &brw->ctx;
    struct vbo_context *vbo = vbo_context(ctx);
    int i;