PPC: 32-bit asm for MAC64 and MLS64

[frescor/ffmpeg.git] / libavcodec / ppc / mathops.h

diff --git a/libavcodec/ppc/mathops.h b/libavcodec/ppc/mathops.h
index 2259f2af09c5d9708e46b6208bdd5d465e70d088..a74fcfa792aa96de26ba1b64731de6e5af2428ca 100644 (file)
--- a/libavcodec/ppc/mathops.h
+++ b/libavcodec/ppc/mathops.h
@@ -1,6 +1,6 @@
 /*
  * simple math operations
- * Copyright (c) 2001, 2002 Fabrice Bellard.
+ * Copyright (c) 2001, 2002 Fabrice Bellard
  * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> et al
  *
  * This file is part of FFmpeg.
@@ -20,19 +20,58 @@
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
-#ifndef FFMPEG_PPC_MATHOPS_H
-#define FFMPEG_PPC_MATHOPS_H
+#ifndef AVCODEC_PPC_MATHOPS_H
+#define AVCODEC_PPC_MATHOPS_H
 
-#if defined(ARCH_POWERPC_405)
+#include "config.h"
+
+#if HAVE_PPC4XX
 /* signed 16x16 -> 32 multiply add accumulate */
 #define MAC16(rt, ra, rb) \
-    asm ("maclhw %0, %2, %3" : "=r" (rt) : "0" (rt), "r" (ra), "r" (rb));
+    __asm__ ("maclhw %0, %2, %3" : "=r" (rt) : "0" (rt), "r" (ra), "r" (rb));
 
 /* signed 16x16 -> 32 multiply */
 #define MUL16(ra, rb) \
     ({ int __rt; \
-    asm ("mullhw %0, %1, %2" : "=r" (__rt) : "r" (ra), "r" (rb)); \
+    __asm__ ("mullhw %0, %1, %2" : "=r" (__rt) : "r" (ra), "r" (rb)); \
     __rt; })
 #endif
 
-#endif /* FFMPEG_PPC_MATHOPS_H */
+#define MULH MULH
+static inline av_const int MULH(int a, int b){
+    int r;
+    __asm__ ("mulhw %0, %1, %2" : "=r"(r) : "r"(a), "r"(b));
+    return r;
+}
+
+#if !HAVE_PPC64
+static inline av_const int64_t MAC64(int64_t d, int a, int b)
+{
+    union { uint64_t x; unsigned hl[2]; } x = { d };
+    int h, l;
+    __asm__ ("mullw %3, %4, %5   \n\t"
+             "mulhw %2, %4, %5   \n\t"
+             "addc  %1, %1, %3   \n\t"
+             "adde  %0, %0, %2   \n\t"
+             : "+r"(x.hl[0]), "+r"(x.hl[1]), "=&r"(h), "=&r"(l)
+             : "r"(a), "r"(b));
+    return x.x;
+}
+#define MAC64(d, a, b) ((d) = MAC64(d, a, b))
+
+static inline av_const int64_t MLS64(int64_t d, int a, int b)
+{
+    union { uint64_t x; unsigned hl[2]; } x = { d };
+    int h, l;
+    __asm__ ("mullw %3, %4, %5   \n\t"
+             "mulhw %2, %4, %5   \n\t"
+             "subfc %1, %3, %1   \n\t"
+             "subfe %0, %2, %0   \n\t"
+             : "+r"(x.hl[0]), "+r"(x.hl[1]), "=&r"(h), "=&r"(l)
+             : "r"(a), "r"(b));
+    return x.x;
+}
+#define MLS64(d, a, b) ((d) = MLS64(d, a, b))
+#endif
+
+#endif /* AVCODEC_PPC_MATHOPS_H */