PPC asm for AV_RL*()

PPC is normally big endian but has special little endian load/store
instructions.  Using these avoids a separate byteswap.  This makes the
vorbis decoder about 5% faster.  Not much else uses little-endian
read/write extensively.

GCC generates horrible PPC code for the default AV_[RW]B64 (which uses
a packed struct), so we override it with a plain pointer cast.

Originally committed as revision 18602 to svn://svn.ffmpeg.org/ffmpeg/trunk

1 files changed, 104 insertions, 0 deletions

diff --git a/libavutil/ppc/intreadwrite.h b/libavutil/ppc/intreadwrite.h
new file mode 100644
index 0000000000..f9898172c5
--- /dev/null
+++ b/libavutil/ppc/intreadwrite.h
@@ -0,0 +1,104 @@
+/*
+ * Copyright (c) 2008 Mans Rullgard <mans@mansr.com>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef AVUTIL_PPC_INTREADWRITE_H
+#define AVUTIL_PPC_INTREADWRITE_H
+
+#include <stdint.h>
+#include "config.h"
+
+#define AV_RL16 AV_RL16
+static inline uint16_t AV_RL16(const void *p)
+{
+    uint16_t v;
+    __asm__ ("lhbrx   %0, %y1" : "=r"(v) : "Z"(*(const uint16_t*)p));
+    return v;
+}
+
+#define AV_WL16 AV_WL16
+static inline void AV_WL16(void *p, uint16_t v)
+{
+    __asm__ ("sthbrx  %1, %y0" : "=Z"(*(uint16_t*)p) : "r"(v));
+}
+
+#define AV_RL32 AV_RL32
+static inline uint32_t AV_RL32(const void *p)
+{
+    uint32_t v;
+    __asm__ ("lwbrx   %0, %y1" : "=r"(v) : "Z"(*(const uint32_t*)p));
+    return v;
+}
+
+#define AV_WL32 AV_WL32
+static inline void AV_WL32(void *p, uint32_t v)
+{
+    __asm__ ("stwbrx  %1, %y0" : "=Z"(*(uint32_t*)p) : "r"(v));
+}
+
+#if HAVE_LDBRX
+
+#define AV_RL64 AV_RL64
+static inline uint64_t AV_RL64(const void *p)
+{
+    uint64_t v;
+    __asm__ ("ldbrx   %0, %y1" : "=r"(v) : "Z"(*(const uint64_t*)p));
+    return v;
+}
+
+#define AV_WL64 AV_WL64
+static inline void AV_WL64(void *p, uint64_t v)
+{
+    __asm__ ("stdbrx  %1, %y0" : "=Z"(*(uint64_t*)p) : "r"(v));
+}
+
+#else
+
+#define AV_RL64 AV_RL64
+static inline uint64_t AV_RL64(const void *p)
+{
+    union { uint64_t v; uint32_t hl[2]; } v;
+    __asm__ ("lwbrx   %0, %y2  \n\t"
+             "lwbrx   %1, %y3  \n\t"
+             : "=r"(v.hl[1]), "=r"(v.hl[0])
+             : "Z"(*(const uint32_t*)p), "Z"(*((const uint32_t*)p+1)));
+    return v.v;
+}
+
+#define AV_WL64 AV_WL64
+static inline void AV_WL64(void *p, uint64_t v)
+{
+    union { uint64_t v; uint32_t hl[2]; } vv = { v };
+    __asm__ ("stwbrx  %2, %y0  \n\t"
+             "stwbrx  %3, %y1  \n\t"
+             : "=Z"(*(uint32_t*)p), "=Z"(*((uint32_t*)p+1))
+             : "r"(vv.hl[1]), "r"(vv.hl[0]));
+}
+
+#endif /* HAVE_LDBRX */
+
+/*
+ * GCC fails miserably on the packed struct version which is used by
+ * default, so we override it here.
+ */
+
+#define AV_RB64(p) (*(const uint64_t *)(p))
+#define AV_WB64(p, v) (*(uint64_t *)(p) = (v))
+
+#endif /* AVUTIL_PPC_INTREADWRITE_H */