Merge branch 'gallium-vertex-linear' into gallium-tex-surfaces

3 files changed, 25 insertions, 37 deletions

diff --git a/src/gallium/auxiliary/translate/translate.h b/src/gallium/auxiliary/translate/translate.h
index b8210af50c..c3b754a902 100644
--- a/src/gallium/auxiliary/translate/translate.h
+++ b/src/gallium/auxiliary/translate/translate.h
@@ -71,15 +71,15 @@ struct translate {
 		       const void *ptr,
 		       unsigned stride );
 
-   void (*run_elts)( struct translate *,
-		     const unsigned *elts,
-		     unsigned count,
-		     void *output_buffer);
-
-   void (*run)( struct translate *,
-		unsigned start,
-		unsigned count,
-		void *output_buffer);
+   void (PIPE_CDECL *run_elts)( struct translate *,
+                                const unsigned *elts,
+                                unsigned count,
+                                void *output_buffer);
+
+   void (PIPE_CDECL *run)( struct translate *,
+                           unsigned start,
+                           unsigned count,
+                           void *output_buffer);
 };
 
 
diff --git a/src/gallium/auxiliary/translate/translate_generic.c b/src/gallium/auxiliary/translate/translate_generic.c
index 402780ee53..a25d94f2ca 100644
--- a/src/gallium/auxiliary/translate/translate_generic.c
+++ b/src/gallium/auxiliary/translate/translate_generic.c
@@ -541,10 +541,10 @@ static emit_func get_emit_func( enum pipe_format format )
 /**
  * Fetch vertex attributes for 'count' vertices.
  */
-static void generic_run_elts( struct translate *translate,
-			      const unsigned *elts,
-			      unsigned count,
-			      void *output_buffer )
+static void PIPE_CDECL generic_run_elts( struct translate *translate,
+                                         const unsigned *elts,
+                                         unsigned count,
+                                         void *output_buffer )
 {
    struct translate_generic *tg = translate_generic(translate);
    char *vert = output_buffer;
@@ -580,10 +580,10 @@ static void generic_run_elts( struct translate *translate,
 
 
 
-static void generic_run( struct translate *translate,
-			 unsigned start,
-			 unsigned count,
-			 void *output_buffer )
+static void PIPE_CDECL generic_run( struct translate *translate,
+                                    unsigned start,
+                                    unsigned count,
+                                    void *output_buffer )
 {
    struct translate_generic *tg = translate_generic(translate);
    char *vert = output_buffer;
diff --git a/src/gallium/auxiliary/translate/translate_sse.c b/src/gallium/auxiliary/translate/translate_sse.c
index a54ac5a82f..2fc8b9d3d0 100644
--- a/src/gallium/auxiliary/translate/translate_sse.c
+++ b/src/gallium/auxiliary/translate/translate_sse.c
@@ -45,22 +45,16 @@
 #define W    3
 
 
-#ifdef WIN32
-#define RTASM __cdecl
-#else
-#define RTASM
-#endif
-
-typedef void (RTASM *run_func)( struct translate *translate,
-                                unsigned start,
-                                unsigned count,
-                                void *output_buffer );
-
-typedef void (RTASM *run_elts_func)( struct translate *translate,
-                                     const unsigned *elts,
+typedef void (PIPE_CDECL *run_func)( struct translate *translate,
+                                     unsigned start,
                                      unsigned count,
                                      void *output_buffer );
 
+typedef void (PIPE_CDECL *run_elts_func)( struct translate *translate,
+                                          const unsigned *elts,
+                                          unsigned count,
+                                          void *output_buffer );
+
 
 
 struct translate_sse {
@@ -472,13 +466,7 @@ static boolean build_vertex_emit( struct translate_sse *p,
    x86_lea(p->func, vertexECX, x86_make_disp(vertexECX, p->translate.key.output_stride));
 
    /* Incr index
-    */   /* Emit code for each of the attributes.  Currently routes
-    * everything through SSE registers, even when it might be more
-    * efficient to stick with regular old x86.  No optimization or
-    * other tricks - enough new ground to cover here just getting
-    * things working.
-    */
-
+    */ 
    if (linear) {
       x86_inc(p->func, idxEBX);
    }