From 8a03d428f1759a7d3af505ab747662747360ea60 Mon Sep 17 00:00:00 2001
From: Rong Yan <rongyan236@gmail.com>
Date: Mon, 27 Apr 2015 03:29:51 +0000
Subject: [PATCH] avcodec/ppc/fdctdsp: POWER LE support in ff_fdct_altivec()
 delete macros VEC_FMERGEH() VEC_FMERGEL(), they where wrong

GCC tool had a bug of PPC intrinsic interpret, which has been fixed in GCC 4.9.1. This bug lead to
errors in two of our previous patches. We found this when we update our GCC tools to 4.9.1 and by
reading the related info on GCC website. We fix our previous error in two separate commits

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
---
 libavcodec/ppc/fdctdsp.c | 86 ++++++++++++++++++----------------------
 1 file changed, 39 insertions(+), 47 deletions(-)

diff --git a/libavcodec/ppc/fdctdsp.c b/libavcodec/ppc/fdctdsp.c
index 80f1366c65..40f4c6c967 100644
--- a/libavcodec/ppc/fdctdsp.c
+++ b/libavcodec/ppc/fdctdsp.c
@@ -78,14 +78,6 @@ static const vector float fdctconsts[3] = {
 #define LD_WA vec_splat(cnsts2, 2)
 #define LD_WB vec_splat(cnsts2, 3)
 
-#if HAVE_BIGENDIAN
-#define VEC_FMERGEH(a, b) vec_mergeh(a, b)
-#define VEC_FMERGEL(a, b) vec_mergel(a, b)
-#else
-#define VEC_FMERGEH(a, b) vec_mergel(b, a)
-#define VEC_FMERGEL(a, b) vec_mergeh(b, a)
-#endif
-
 #define FDCTROW(b0, b1, b2, b3, b4, b5, b6, b7) /* {{{ */           \
     x0 = vec_add(b0, b7);               /* x0 = b0 + b7; */         \
     x7 = vec_sub(b0, b7);               /* x7 = b0 - b7; */         \
@@ -393,45 +385,45 @@ void ff_fdct_altivec(int16_t *block)
     /* }}} */
 
     /* 8x8 matrix transpose (vector float[8][2]) {{{ */
-    x0 = VEC_FMERGEL(b00, b20);
-    x1 = VEC_FMERGEH(b00, b20);
-    x2 = VEC_FMERGEL(b10, b30);
-    x3 = VEC_FMERGEH(b10, b30);
-
-    b00 = VEC_FMERGEH(x1, x3);
-    b10 = VEC_FMERGEL(x1, x3);
-    b20 = VEC_FMERGEH(x0, x2);
-    b30 = VEC_FMERGEL(x0, x2);
-
-    x4 = VEC_FMERGEL(b41, b61);
-    x5 = VEC_FMERGEH(b41, b61);
-    x6 = VEC_FMERGEL(b51, b71);
-    x7 = VEC_FMERGEH(b51, b71);
-
-    b41 = VEC_FMERGEH(x5, x7);
-    b51 = VEC_FMERGEL(x5, x7);
-    b61 = VEC_FMERGEH(x4, x6);
-    b71 = VEC_FMERGEL(x4, x6);
-
-    x0 = VEC_FMERGEL(b01, b21);
-    x1 = VEC_FMERGEH(b01, b21);
-    x2 = VEC_FMERGEL(b11, b31);
-    x3 = VEC_FMERGEH(b11, b31);
-
-    x4 = VEC_FMERGEL(b40, b60);
-    x5 = VEC_FMERGEH(b40, b60);
-    x6 = VEC_FMERGEL(b50, b70);
-    x7 = VEC_FMERGEH(b50, b70);
-
-    b40 = VEC_FMERGEH(x1, x3);
-    b50 = VEC_FMERGEL(x1, x3);
-    b60 = VEC_FMERGEH(x0, x2);
-    b70 = VEC_FMERGEL(x0, x2);
-
-    b01 = VEC_FMERGEH(x5, x7);
-    b11 = VEC_FMERGEL(x5, x7);
-    b21 = VEC_FMERGEH(x4, x6);
-    b31 = VEC_FMERGEL(x4, x6);
+    x0 = vec_mergel(b00, b20);
+    x1 = vec_mergeh(b00, b20);
+    x2 = vec_mergel(b10, b30);
+    x3 = vec_mergeh(b10, b30);
+
+    b00 = vec_mergeh(x1, x3);
+    b10 = vec_mergel(x1, x3);
+    b20 = vec_mergeh(x0, x2);
+    b30 = vec_mergel(x0, x2);
+
+    x4 = vec_mergel(b41, b61);
+    x5 = vec_mergeh(b41, b61);
+    x6 = vec_mergel(b51, b71);
+    x7 = vec_mergeh(b51, b71);
+
+    b41 = vec_mergeh(x5, x7);
+    b51 = vec_mergel(x5, x7);
+    b61 = vec_mergeh(x4, x6);
+    b71 = vec_mergel(x4, x6);
+
+    x0 = vec_mergel(b01, b21);
+    x1 = vec_mergeh(b01, b21);
+    x2 = vec_mergel(b11, b31);
+    x3 = vec_mergeh(b11, b31);
+
+    x4 = vec_mergel(b40, b60);
+    x5 = vec_mergeh(b40, b60);
+    x6 = vec_mergel(b50, b70);
+    x7 = vec_mergeh(b50, b70);
+
+    b40 = vec_mergeh(x1, x3);
+    b50 = vec_mergel(x1, x3);
+    b60 = vec_mergeh(x0, x2);
+    b70 = vec_mergel(x0, x2);
+
+    b01 = vec_mergeh(x5, x7);
+    b11 = vec_mergel(x5, x7);
+    b21 = vec_mergeh(x4, x6);
+    b31 = vec_mergel(x4, x6);
     /* }}} */
 
     FDCTCOL(b00, b10, b20, b30, b40, b50, b60, b70);