[FFmpeg-devel] [PATCH 5/5] x86: sbrdsp/fft: reuse ps_neg constant

[Christophe Gisquet]

---
 libavcodec/x86/constants.c | 2 ++
 libavcodec/x86/constants.h | 2 ++
 libavcodec/x86/fft.asm     | 7 ++++---
 libavcodec/x86/sbrdsp.asm  | 2 +-
 4 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/libavcodec/x86/constants.c b/libavcodec/x86/constants.c
index bfb0ff3..5d2c237 100644
--- a/libavcodec/x86/constants.c
+++ b/libavcodec/x86/constants.c
@@ -56,3 +56,5 @@ DECLARE_ALIGNED(32, const ymm_reg,  ff_pb_3)    = { 0x0303030303030303ULL, 0x030
                                                     0x0303030303030303ULL, 0x0303030303030303ULL };
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pb_80)   = { 0x8080808080808080ULL, 0x8080808080808080ULL };
 DECLARE_ALIGNED(8,  const uint64_t, ff_pb_FC)   =   0xFCFCFCFCFCFCFCFCULL;
+
+DECLARE_ALIGNED(16, const xmm_reg,  ff_ps_neg)  = { 0x8000000080000000ULL, 0x8000000080000000ULL };
diff --git a/libavcodec/x86/constants.h b/libavcodec/x86/constants.h
index f73a9f2..e75fff9 100644
--- a/libavcodec/x86/constants.h
+++ b/libavcodec/x86/constants.h
@@ -55,4 +55,6 @@ extern const xmm_reg  ff_pb_80;
 extern const xmm_reg  ff_pb_F8;
 extern const uint64_t ff_pb_FC;
 
+extern const xmm_reg  ff_ps_neg;
+
 #endif /* AVCODEC_X86_CONSTANTS_H */
diff --git a/libavcodec/x86/fft.asm b/libavcodec/x86/fft.asm
index cae404c..877997e 100644
--- a/libavcodec/x86/fft.asm
+++ b/libavcodec/x86/fft.asm
@@ -68,9 +68,10 @@ perm1: dd 0x00, 0x02, 0x03, 0x01, 0x03, 0x00, 0x02, 0x01
 perm2: dd 0x00, 0x01, 0x02, 0x03, 0x01, 0x00, 0x02, 0x03
 ps_p1p1m1p1root2: dd 1.0, 1.0, -1.0, 1.0, M_SQRT1_2, M_SQRT1_2, M_SQRT1_2, M_SQRT1_2
 ps_m1m1p1m1p1m1m1m1: dd 1<<31, 1<<31, 0, 1<<31, 0, 1<<31, 1<<31, 1<<31
-ps_m1m1m1m1: times 4 dd 1<<31
 ps_m1p1: dd 1<<31, 0
 
+cextern ps_neg
+
 %assign i 16
 %rep 13
 cextern cos_ %+ i
@@ -685,7 +686,7 @@ cglobal imdct_calc, 3,5,3
     mov     r2, r3
     sub     r3, mmsize
     neg     r2
-    mova    m2, [ps_m1m1m1m1]
+    mova    m2, [ps_neg]
 .loop:
 %if mmsize == 8
     PSWAPD  m0, [r1 + r3]
@@ -998,7 +999,7 @@ cglobal imdct_half, 3,12,8; FFTContext *s, FFTSample *output, const FFTSample *i
     sub   r4, r3
 %endif
 %if notcpuflag(3dnowext) && mmsize == 8
-    movd  m7, [ps_m1m1m1m1]
+    movd  m7, [ps_neg]
 %endif
 .pre:
 %if ARCH_X86_64 == 0
diff --git a/libavcodec/x86/sbrdsp.asm b/libavcodec/x86/sbrdsp.asm
index d556f27..6f2e4f4 100644
--- a/libavcodec/x86/sbrdsp.asm
+++ b/libavcodec/x86/sbrdsp.asm
@@ -25,13 +25,13 @@ SECTION_RODATA
 ; mask equivalent for multiply by -1.0 1.0
 ps_mask         times 2 dd 1<<31, 0
 ps_mask2        times 2 dd 0, 1<<31
-ps_neg          times 4 dd 1<<31
 ps_noise0       times 2 dd  1.0,  0.0,
 ps_noise2       times 2 dd -1.0,  0.0
 ps_noise13      dd  0.0,  1.0, 0.0, -1.0
                 dd  0.0, -1.0, 0.0,  1.0
                 dd  0.0,  1.0, 0.0, -1.0
 cextern         sbr_noise_table
+cextern         ps_neg
 
 SECTION_TEXT
 
-- 
1.9.2.msysgit.0