"fast unaligned" bytestream functions

patch by Ramiro Polla ramiro lisha ufsc br
original thread:
date: 03/11/2007 03:06 AM
subject: [Ffmpeg-devel] [PATCH] Machine endian bytestream functions

Originally committed as revision 8803 to svn://svn.ffmpeg.org/ffmpeg/trunk

diff --git a/configure b/configure
index 1c6efb4..55c89a6 100755 (executable)
--- a/configure
+++ b/configure
@@ -628,6 +628,7 @@ HAVE_LIST="
     ebx_available
     fast_64bit
     fast_cmov
+    fast_unaligned
     freetype2
     imlib2
     inet_aton
@@ -765,6 +766,7 @@ powerpc_perf="no"
 mmx="default"
 cmov="no"
 fast_cmov="no"
+fast_unaligned="no"
 armv5te="default"
 armv6="default"
 iwmmxt="default"
@@ -975,9 +977,11 @@ done
 case "$arch" in
   i386|i486|i586|i686|i86pc|BePC)
     arch="x86_32"
+    enable fast_unaligned
   ;;
   x86_64|amd64)
     arch="x86_32"
+    enable fast_unaligned
     canon_arch="`$cc -dumpmachine | sed -e 's,\([^-]*\)-.*,\1,'`"
     if [ x"$canon_arch" = x"x86_64" -o x"$canon_arch" = x"amd64" ]; then
       if [ -z "`echo $CFLAGS | grep -- -m32`"  ]; then

diff --git a/libavutil/intreadwrite.h b/libavutil/intreadwrite.h
index b77a92a..b59c481 100644 (file)
--- a/libavutil/intreadwrite.h
+++ b/libavutil/intreadwrite.h
@@ -50,6 +50,21 @@ struct unaligned_16 { uint16_t l; } __attribute__((packed));
 #define AV_RL8(x)     AV_RB8(x)
 #define AV_WL8(p, d)  AV_WB8(p, d)
 
+#ifdef HAVE_FAST_UNALIGNED
+# ifdef WORDS_BIGENDIAN
+#  define AV_RB16(x)    LD16(x)
+#  define AV_WB16(p, d) ST16(p, d)
+
+#  define AV_RL16(x)    bswap_16(LD16(x))
+#  define AV_WL16(p, d) ST16(p, bswap_16(d))
+# else /* WORDS_BIGENDIAN */
+#  define AV_RB16(x)    bswap_16(LD16(x))
+#  define AV_WB16(p, d) ST16(p, bswap_16(d))
+
+#  define AV_RL16(x)    LD16(x)
+#  define AV_WL16(p, d) ST16(p, d)
+# endif
+#else /* HAVE_FAST_UNALIGNED */
 #define AV_RB16(x)  ((((uint8_t*)(x))[0] << 8) | ((uint8_t*)(x))[1])
 #define AV_WB16(p, d) { \
                     ((uint8_t*)(p))[1] = (d); \
@@ -60,6 +75,7 @@ struct unaligned_16 { uint16_t l; } __attribute__((packed));
 #define AV_WL16(p, d) { \
                     ((uint8_t*)(p))[0] = (d); \
                     ((uint8_t*)(p))[1] = (d)>>8; }
+#endif
 
 #define AV_RB24(x)  ((((uint8_t*)(x))[0] << 16) | \
                      (((uint8_t*)(x))[1] <<  8) | \
@@ -77,6 +93,21 @@ struct unaligned_16 { uint16_t l; } __attribute__((packed));
                     ((uint8_t*)(p))[1] = (d)>>8; \
                     ((uint8_t*)(p))[2] = (d)>>16; }
 
+#ifdef HAVE_FAST_UNALIGNED
+# ifdef WORDS_BIGENDIAN
+#  define AV_RB32(x)    LD32(x)
+#  define AV_WB32(p, d) ST32(p, d)
+
+#  define AV_RL32(x)    bswap_32(LD32(x))
+#  define AV_WL32(p, d) ST32(p, bswap_32(d))
+# else /* WORDS_BIGENDIAN */
+#  define AV_RB32(x)    bswap_32(LD32(x))
+#  define AV_WB32(p, d) ST32(p, bswap_32(d))
+
+#  define AV_RL32(x)    LD32(x)
+#  define AV_WL32(p, d) ST32(p, d)
+# endif
+#else /* HAVE_FAST_UNALIGNED */
 #define AV_RB32(x)  ((((uint8_t*)(x))[0] << 24) | \
                      (((uint8_t*)(x))[1] << 16) | \
                      (((uint8_t*)(x))[2] <<  8) | \
@@ -96,5 +127,6 @@ struct unaligned_16 { uint16_t l; } __attribute__((packed));
                     ((uint8_t*)(p))[1] = (d)>>8; \
                     ((uint8_t*)(p))[2] = (d)>>16; \
                     ((uint8_t*)(p))[3] = (d)>>24; }
+#endif
 
 #endif /* INTREADWRITE_H */