Mesa (master): mesa: use _mesa_memcpy for COPY_4FV macro
Roland Scheidegger
sroland at kemper.freedesktop.org
Tue Dec 8 18:28:44 UTC 2009
Module: Mesa
Branch: master
Commit: 86c8f70db10a584aa78e4d5f397ad3543fdb77d2
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=86c8f70db10a584aa78e4d5f397ad3543fdb77d2
Author: Roland Scheidegger <sroland at vmware.com>
Date: Thu Dec 3 23:26:13 2009 +0100
mesa: use _mesa_memcpy for COPY_4FV macro
Gets rid of one of the worst strict-aliasing offenders, and actually
produces faster code (at least in some cases, when compiler can use
for instance 64bit moves for memcpy).
(note _mesa_memcpy should get inlined)
---
src/mesa/main/macros.h | 11 +++--------
1 files changed, 3 insertions(+), 8 deletions(-)
diff --git a/src/mesa/main/macros.h b/src/mesa/main/macros.h
index 3d9a1ab..f0ea463 100644
--- a/src/mesa/main/macros.h
+++ b/src/mesa/main/macros.h
@@ -202,17 +202,12 @@ do { \
#endif
/**
- * Copy a 4-element float vector (avoid using FPU registers)
- * XXX Could use two 64-bit moves on 64-bit systems
+ * Copy a 4-element float vector
+ * memcpy seems to be most efficient
*/
#define COPY_4FV( DST, SRC ) \
do { \
- const GLuint *_s = (const GLuint *) (SRC); \
- GLuint *_d = (GLuint *) (DST); \
- _d[0] = _s[0]; \
- _d[1] = _s[1]; \
- _d[2] = _s[2]; \
- _d[3] = _s[3]; \
+ _mesa_memcpy(DST, SRC, sizeof(GLfloat) * 4); \
} while (0)
/** Copy \p SZ elements into a 4-element vector */
More information about the mesa-commit
mailing list