vbo: fix VBO optimization regression

The optimization in change 8e4efdc895 ("vbo: optimize some display
list drawing") missed the loopback case.  This is used when the
glBegin/End primitive doesn't have a uniform set of vertex attributes.
The new Piglit gl-1.0-dlist-materials test hits this.

So check the aligned_vertex_buffer_offset(list) value and adjust the
buffer offset accordingly.

We also need to remove the 'start == 0' assertion in the loopback
code since it no longer applies.

Reviewed-by: Roland Scheidegger <sroland@vmware.com>

src/mesa/vbo/vbo_save_draw.c

@@ -240,8 +240,11 @@ loopback_vertex_list(struct gl_context *ctx,
                                  list->vertex_store->bufferobj,
                                  MAP_INTERNAL);
 
+   unsigned buffer_offset =
+      aligned_vertex_buffer_offset(list) ? 0 : list->buffer_offset;
+
    vbo_loopback_vertex_list(ctx,
-                            (const GLfloat *)(buffer + list->buffer_offset),
+                            (const GLfloat *) (buffer + buffer_offset),
                             list->attrsz,
                             list->prims,
                             list->prim_count,

src/mesa/vbo/vbo_save_loopback.c

@@ -107,17 +107,17 @@ loopback_prim(struct gl_context *ctx,
    GLuint k;
 
    if (0)
-      printf("loopback prim %s(%s,%s) verts %d..%d\n",
+      printf("loopback prim %s(%s,%s) verts %d..%d  vsize %d\n",
              _mesa_lookup_prim_by_nr(prim->mode),
              prim->begin ? "begin" : "..",
              prim->end ? "end" : "..",
-             start, end);
+             start, end,
+             vertex_size);
 
    if (prim->begin) {
       CALL_Begin(GET_DISPATCH(), (prim->mode));
    }
    else {
-      assert(start == 0);
       start += wrap_count;
    }