vbo: use MapBufferRange where available

Previously would have to allocate a new VBO after firing a draw command
as subsequent call to Map() on old VBO might block if the driver had
submitted the commands to hardware.

1 files changed, 101 insertions, 17 deletions

diff --git a/src/mesa/vbo/vbo_exec_draw.c b/src/mesa/vbo/vbo_exec_draw.c
index ae43857c8ae..a89f1506544 100644
--- a/src/mesa/vbo/vbo_exec_draw.c
+++ b/src/mesa/vbo/vbo_exec_draw.c
@@ -70,7 +70,7 @@ static GLuint vbo_copy_vertices( struct vbo_exec_context *exec )
    GLuint ovf, i;
    GLuint sz = exec->vtx.vertex_size;
    GLfloat *dst = exec->vtx.copied.buffer;
-   GLfloat *src = ((GLfloat *)exec->vtx.buffer_map + 
+   GLfloat *src = (exec->vtx.buffer_map + 
 		   exec->vtx.prim[exec->vtx.prim_count-1].start * 
 		   exec->vtx.vertex_size);
 
@@ -147,7 +147,7 @@ static void vbo_exec_bind_arrays( GLcontext *ctx )
    struct vbo_exec_context *exec = &vbo->exec;
    struct gl_client_array *arrays = exec->vtx.arrays;
    GLuint count = exec->vtx.vert_count;
-   GLubyte *data = exec->vtx.buffer_map;
+   GLubyte *data = (GLubyte *)exec->vtx.buffer_map;
    const GLuint *map;
    GLuint attr;
    GLbitfield varying_inputs = 0x0;
@@ -220,10 +220,78 @@ static void vbo_exec_bind_arrays( GLcontext *ctx )
 }
 
 
+static void vbo_exec_vtx_unmap( struct vbo_exec_context *exec )
+{
+   GLenum target = GL_ARRAY_BUFFER_ARB;
+
+   if (exec->vtx.bufferobj->Name) {
+      GLcontext *ctx = exec->ctx;
+
+      exec->vtx.buffer_used += (exec->vtx.buffer_ptr -
+                                exec->vtx.buffer_map) * sizeof(float);
+            
+      ctx->Driver.UnmapBuffer(ctx, target, exec->vtx.bufferobj);
+      exec->vtx.buffer_map = NULL;
+      exec->vtx.buffer_ptr = NULL;
+      exec->vtx.max_vert = 0;
+   }
+}
+
+void vbo_exec_vtx_map( struct vbo_exec_context *exec )
+{
+   GLcontext *ctx = exec->ctx;
+   GLenum target = GL_ARRAY_BUFFER_ARB;
+   GLenum access = GL_READ_WRITE_ARB;
+   GLenum usage = GL_STREAM_DRAW_ARB;
+   
+   if (exec->vtx.bufferobj->Name == 0)
+      return;
+
+   if (exec->vtx.buffer_map != NULL) {
+      assert(0);
+      exec->vtx.buffer_map = NULL;
+   }
+
+   if (VBO_VERT_BUFFER_SIZE > exec->vtx.buffer_used + 1024 &&
+       ctx->Driver.MapBufferRange)
+   {
+      exec->vtx.buffer_map = 
+         (GLfloat *)ctx->Driver.MapBufferRange(ctx, 
+                                               target, 
+                                               exec->vtx.buffer_used,
+                                               (VBO_VERT_BUFFER_SIZE - 
+                                                exec->vtx.buffer_used),
+                                               (GL_MAP_WRITE_BIT |
+                                                GL_MAP_INVALIDATE_RANGE_BIT | 
+                                                GL_MAP_UNSYNCHRONIZED_BIT | 
+                                                MESA_MAP_NOWAIT_BIT),
+                                               exec->vtx.bufferobj);
+   }
+
+   if (exec->vtx.buffer_map) {
+      exec->vtx.buffer_map += exec->vtx.buffer_used / sizeof(float);
+   }
+   else {
+      exec->vtx.buffer_used = 0;
+
+      ctx->Driver.BufferData(ctx, target, 
+                             VBO_VERT_BUFFER_SIZE, 
+                             NULL, usage, exec->vtx.bufferobj);
+
+      exec->vtx.buffer_map = 
+         (GLfloat *)ctx->Driver.MapBuffer(ctx, target, access, exec->vtx.bufferobj);
+   }
+
+   if (0) _mesa_printf("map %d..\n", exec->vtx.buffer_used);
+}
+
+
+
 /**
  * Execute the buffer and save copied verts.
  */
-void vbo_exec_vtx_flush( struct vbo_exec_context *exec )
+void vbo_exec_vtx_flush( struct vbo_exec_context *exec,
+                         GLboolean unmap )
 {
    if (0)
       vbo_exec_debug_verts( exec );
@@ -236,13 +304,9 @@ void vbo_exec_vtx_flush( struct vbo_exec_context *exec )
 
       if (exec->vtx.copied.nr != exec->vtx.vert_count) {
 	 GLcontext *ctx = exec->ctx;
-
-	 GLenum target = GL_ARRAY_BUFFER_ARB;
-	 GLenum access = GL_READ_WRITE_ARB;
-	 GLenum usage = GL_STREAM_DRAW_ARB;
-	 GLsizei size = VBO_VERT_BUFFER_SIZE * sizeof(GLfloat);
 	 
-	 /* Before the unmap (why?)
+	 /* Before the update_state() as this may raise _NEW_ARRAY
+          * from _mesa_set_varying_vp_inputs().
 	  */
 	 vbo_exec_bind_arrays( ctx );
 
@@ -250,8 +314,12 @@ void vbo_exec_vtx_flush( struct vbo_exec_context *exec )
             _mesa_update_state( ctx );
 
 
-	 ctx->Driver.UnmapBuffer(ctx, target, exec->vtx.bufferobj);
-	 exec->vtx.buffer_map = NULL;
+         if (exec->vtx.bufferobj->Name) {
+            vbo_exec_vtx_unmap( exec );
+         }
+
+         if (0) _mesa_printf("%s %d %d\n", __FUNCTION__, exec->vtx.prim_count,
+                      exec->vtx.vert_count);
 
 	 vbo_context(ctx)->draw_prims( ctx, 
 				       exec->vtx.inputs, 
@@ -261,15 +329,31 @@ void vbo_exec_vtx_flush( struct vbo_exec_context *exec )
 				       0,
 				       exec->vtx.vert_count - 1);
 
-	 /* Get new data:
-	  */
-	 ctx->Driver.BufferData(ctx, target, size, NULL, usage, exec->vtx.bufferobj);
-	 exec->vtx.buffer_map
-	    = ctx->Driver.MapBuffer(ctx, target, access, exec->vtx.bufferobj);
+	 /* If using a real VBO, get new storage -- unless asked not to.
+          */
+         if (exec->vtx.bufferobj->Name && !unmap) {
+            vbo_exec_vtx_map( exec );
+          }
       }
    }
 
+   /* May have to unmap explicitly if we didn't draw:
+    */
+   if (unmap && 
+       exec->vtx.bufferobj->Name &&
+       exec->vtx.buffer_map) {
+      vbo_exec_vtx_unmap( exec );
+   }
+
+
+   if (unmap) 
+      exec->vtx.max_vert = 0;
+   else
+      exec->vtx.max_vert = ((VBO_VERT_BUFFER_SIZE - exec->vtx.buffer_used) / 
+                            (exec->vtx.vertex_size * sizeof(GLfloat)));
+
+
+   exec->vtx.buffer_ptr = exec->vtx.buffer_map;
    exec->vtx.prim_count = 0;
    exec->vtx.vert_count = 0;
-   exec->vtx.vbptr = (GLfloat *)exec->vtx.buffer_map;
 }