Now RET checks whether it immediately follows a branch, so the programmer
dosen't have to keep track of that condition.
REP_RET is still needed manually when it's a branch target, but that's much
rarer.
The implementation involves lots of spurious labels, so get rid of them with
`strip`.
---
configure | 3 +++
library.mak | 1 +
libavutil/x86/x86inc.asm | 36 +++++++++++++++++++++++----
libavcodec/x86/ac3dsp.asm | 6 ++---
libavcodec/x86/deinterlace.asm | 2 +-
libavcodec/x86/dsputil.asm | 22 ++++++++---------
libavcodec/x86/fft.asm | 8 +++---
libavcodec/x86/fmtconvert.asm | 12 ++++-----
libavcodec/x86/h264_chromamc.asm | 18 +++++++-------
libavcodec/x86/h264_chromamc_10bit.asm | 10 ++++----
libavcodec/x86/h264_deblock.asm | 2 +-
libavcodec/x86/h264_deblock_10bit.asm | 10 ++++----
libavcodec/x86/h264_idct.asm | 16 ++++++------
libavcodec/x86/h264_intrapred.asm | 28 ++++++++++-----------
libavcodec/x86/h264_intrapred_10bit.asm | 16 ++++++------
libavcodec/x86/h264_qpel_10bit.asm | 2 +-
libavcodec/x86/h264_qpel_8bit.asm | 20 +++++++--------
libavcodec/x86/h264_weight.asm | 16 ++++++------
libavcodec/x86/h264_weight_10bit.asm | 12 ++++-----
libavcodec/x86/pngdsp.asm | 2 +-
libavcodec/x86/rv34dsp.asm | 2 +-
libavcodec/x86/rv40dsp.asm | 10 ++++----
libavcodec/x86/sbrdsp.asm | 2 +-
libavcodec/x86/vp8dsp.asm | 40 +++++++++++++++---------------
libavfilter/x86/af_volume.asm | 6 ++---
libavfilter/x86/hqdn3d.asm | 2 +-
libavresample/x86/audio_convert.asm | 44 ++++++++++++++++-----------------
libavresample/x86/audio_mix.asm | 10 ++++----
libavutil/x86/float_dsp.asm | 8 +++---
libswscale/x86/input.asm | 14 +++++------
libswscale/x86/output.asm | 8 +++---
libswscale/x86/scale.asm | 2 +-
32 files changed, 210 insertions(+), 180 deletions(-)
diff --git a/configure b/configure
index 08f1d82..ac056b3 100755
--- a/configure
+++ b/configure
@@ -1802,6 +1802,7 @@ nm_default="nm -g"
objformat="elf"
pkg_config_default=pkg-config
ranlib="ranlib"
+strip="strip"
yasmexe="yasm"
nogas=":"
@@ -2051,6 +2052,7 @@ cc_default="${cross_prefix}${cc_default}"
nm_default="${cross_prefix}${nm_default}"
pkg_config_default="${cross_prefix}${pkg_config_default}"
ranlib="${cross_prefix}${ranlib}"
+strip="${cross_prefix}${strip}"
sysinclude_default="${sysroot}/usr/include"
@@ -3842,6 +3844,7 @@ ARCH=$arch
CC=$cc
AS=$as
LD=$ld
+STRIP=$strip
DEPCC=$dep_cc
DEPCCFLAGS=$DEPCCFLAGS \$(CPPFLAGS)
DEPAS=$as
diff --git a/library.mak b/library.mak
index 3b4bd2d..23ee5fa 100644
--- a/library.mak
+++ b/library.mak
@@ -26,6 +26,7 @@ $(SUBDIR)%-test.i: $(SUBDIR)%.c
$(SUBDIR)x86/%.o: $(SUBDIR)x86/%.asm
$(DEPYASM) $(YASMFLAGS) -I $(<D)/ -M -o $@ $< > $(@:.o=.d)
$(YASM) $(YASMFLAGS) -I $(<D)/ -o $@ $<
+ -@ $(if $(STRIP), $(STRIP) -wN '..@*' $@)
LIBOBJS := $(OBJS) $(SUBDIR)%.h.o $(TESTOBJS)
$(LIBOBJS) $(LIBOBJS:.o=.i): CPPFLAGS += -DHAVE_AV_CONFIG_H
diff --git a/libavutil/x86/x86inc.asm b/libavutil/x86/x86inc.asm
index 2617cdf..5594db4 100644
--- a/libavutil/x86/x86inc.asm
+++ b/libavutil/x86/x86inc.asm
@@ -131,8 +131,7 @@ CPUNOP amdnop
; Pops anything that was pushed by PROLOGUE, and returns.
; REP_RET:
-; Same, but if it doesn't pop anything it becomes a 2-byte ret, for athlons
-; which are slow when a normal ret follows a branch.
+; Use this instead of RET if it's a branch target.
; registers:
; rN and rNq are the native-size register holding function argument N
@@ -480,7 +479,7 @@ DECLARE_REG 14, R15, 120
%if mmsize == 32
vzeroupper
%endif
- ret
+ AUTO_REP_RET
%endmacro
%elif ARCH_X86_64 ; *nix x64 ;=============================================
@@ -527,7 +526,7 @@ DECLARE_REG 14, R15, 72
%if mmsize == 32
vzeroupper
%endif
- ret
+ AUTO_REP_RET
%endmacro
%else ; X86_32 ;==============================================================
@@ -583,7 +582,7 @@ DECLARE_ARG 7, 8, 9, 10, 11, 12, 13, 14
%if mmsize == 32
vzeroupper
%endif
- ret
+ AUTO_REP_RET
%endmacro
%endif ;======================================================================
@@ -597,6 +596,10 @@ DECLARE_ARG 7, 8, 9, 10, 11, 12, 13, 14
%endmacro
%endif
+; On AMD cpus <=K10, an ordinary ret is slow if it immediately follows either
+; a branch or a branch target. So switch to a 2-byte form of ret in that case.
+; We can automatically detect "follows a branch", but not a branch target.
+; (SSSE3 is a sufficient condition to know that your cpu doesn't have this
problem.)
%macro REP_RET 0
%if has_epilogue
RET
@@ -605,6 +608,29 @@ DECLARE_ARG 7, 8, 9, 10, 11, 12, 13, 14
%endif
%endmacro
+%define last_branch_adr $$
+%macro AUTO_REP_RET 0
+ %ifndef cpuflags
+ times ((last_branch_adr-$)>>31)+1 rep ; times 1 iff $ !=
last_branch_adr.
+ %elif notcpuflag(ssse3)
+ times ((last_branch_adr-$)>>31)+1 rep
+ %endif
+ ret
+%endmacro
+
+%macro BRANCH_INSTR 0-*
+ %rep %0
+ %macro %1 1-2 %1
+ %2 %1
+ %%branch_instr:
+ %xdefine last_branch_adr %%branch_instr
+ %endmacro
+ %rotate 1
+ %endrep
+%endmacro
+
+BRANCH_INSTR jz, je, jnz, jne, jl, jle, jnl, jnle, jg, jge, jng, jnge, ja,
jae, jna, jnae, jb, jbe, jnb, jnbe, jc, jnc, js, jns, jo, jno, jp, jnp
+
%macro TAIL_CALL 2 ; callee, is_nonadjacent
%if has_epilogue
call %1
diff --git a/libavcodec/x86/ac3dsp.asm b/libavcodec/x86/ac3dsp.asm
index a5d9458..ca5b9fe 100644
--- a/libavcodec/x86/ac3dsp.asm
+++ b/libavcodec/x86/ac3dsp.asm
@@ -317,7 +317,7 @@ cglobal float_to_fixed24, 3, 3, 9, dst, src, len
sub lenq, 16
%endif
ja .loop
- REP_RET
+ RET
;------------------------------------------------------------------------------
; int ff_ac3_compute_mantissa_size(uint16_t mant_cnt[6][16])
@@ -415,7 +415,7 @@ cglobal ac3_extract_exponents, 3, 3, 0, exp, coef, len
movd [expq+lenq], m0
add lenq, 4
jl .loop
- REP_RET
+ RET
%endif
%macro AC3_EXTRACT_EXPONENTS 0
@@ -447,7 +447,7 @@ cglobal ac3_extract_exponents, 3, 3, 4, exp, coef, len
add lenq, 4
jl .loop
- REP_RET
+ RET
%endmacro
%if HAVE_SSE2_EXTERNAL
diff --git a/libavcodec/x86/deinterlace.asm b/libavcodec/x86/deinterlace.asm
index b2828f3..ba087de 100644
--- a/libavcodec/x86/deinterlace.asm
+++ b/libavcodec/x86/deinterlace.asm
@@ -74,7 +74,7 @@ cglobal deinterlace_line_mmx, 7,7,7, dst, lum_m4,
lum_m3, lum_m2, lum_m1
add lumq, 4
sub sized, 4
jg .nextrow
- REP_RET
+ RET
%endmacro
DEINTERLACE ""
diff --git a/libavcodec/x86/dsputil.asm b/libavcodec/x86/dsputil.asm
index dd0c795..e613f15 100644
--- a/libavcodec/x86/dsputil.asm
+++ b/libavcodec/x86/dsputil.asm
@@ -313,7 +313,7 @@ cglobal apply_window_int16_round, 4,5,6, output, input,
window, offset, offset2
add offsetd, mmsize
sub offset2d, mmsize
jae .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -596,7 +596,7 @@ cglobal emu_edge_core, 2, 7, 0
%endif
call r1
.h_extend_end:
- RET
+ REP_RET
%if ARCH_X86_64
%define vall al
@@ -1120,7 +1120,7 @@ cglobal vector_clip_int32%5, 5,5,%1, dst, src, min, max,
len
add dstq, mmsize*4*(%2+%3)
sub lend, mmsize*(%2+%3)
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmx
@@ -1167,7 +1167,7 @@ ALIGN 16
add src1q, 2*mmsize
sub lenq, 2*mmsize
jge .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse
@@ -1195,7 +1195,7 @@ ALIGN 16
sub lenq, 2*mmsize
jge .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse
@@ -1346,7 +1346,7 @@ cglobal bswap32_buf, 3,4,5
jnz .loop2
%endif
.end:
- RET
+ REP_RET
%endmacro
INIT_XMM sse2
@@ -1410,7 +1410,7 @@ cglobal %1_pixels4_l2, 6,6
add r2, 16
sub r5d, 4
jne .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -1453,7 +1453,7 @@ cglobal %1_pixels8_l2, 6,6
add r2, 32
sub r5d, 4
jne .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -1498,7 +1498,7 @@ cglobal %1_pixels16_l2, 6,6
add r2, 32
sub r5d, 2
jne .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -1561,7 +1561,7 @@ cglobal put_pixels16, 4,5,4
sub r3d, 4
lea r0, [r0+r2*4]
jnz .loop
- REP_RET
+ RET
; void avg_pixels16_sse2(uint8_t *block, const uint8_t *pixels, int line_size,
int h)
cglobal avg_pixels16, 4,5,4
@@ -1584,4 +1584,4 @@ cglobal avg_pixels16, 4,5,4
sub r3d, 4
lea r0, [r0+r2*4]
jnz .loop
- REP_RET
+ RET
diff --git a/libavcodec/x86/fft.asm b/libavcodec/x86/fft.asm
index c87752b..e0a6a09 100644
--- a/libavcodec/x86/fft.asm
+++ b/libavcodec/x86/fft.asm
@@ -556,7 +556,7 @@ cglobal fft_calc, 2,5,8
mov r0, r1
mov r1, r3
FFT_DISPATCH _interleave %+ SUFFIX, r1
- REP_RET
+ RET
INIT_XMM sse
@@ -609,7 +609,7 @@ cglobal fft_calc, 2,5,8
femms
RET
%else
- REP_RET
+ RET
%endif
%endmacro
@@ -653,7 +653,7 @@ cglobal fft_permute, 2,7,1
movaps [r1 + r2 + 16], xmm1
add r2, 32
jl .loopcopy
- REP_RET
+ RET
%macro IMDCT_CALC_FUNC 0
cglobal imdct_calc, 3,5,3
@@ -703,7 +703,7 @@ cglobal imdct_calc, 3,5,3
femms
RET
%else
- REP_RET
+ RET
%endif
%endmacro
diff --git a/libavcodec/x86/fmtconvert.asm b/libavcodec/x86/fmtconvert.asm
index 8267bd4..2e0e6ac 100644
--- a/libavcodec/x86/fmtconvert.asm
+++ b/libavcodec/x86/fmtconvert.asm
@@ -68,7 +68,7 @@ cglobal int32_to_float_fmul_scalar, 4, 4, %1, dst, src, mul,
len
mova [dstq+lenq+16], m2
add lenq, 32
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse
@@ -107,7 +107,7 @@ cglobal float_to_int16, 3, 3, %1, dst, src, len
%if mmsize == 8
emms
%endif
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -185,7 +185,7 @@ cglobal float_to_int16_step, 4, 7, %1, dst, src, len, step,
step3, v1, v2
%if mmsize == 8
emms
%endif
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -233,7 +233,7 @@ cglobal float_to_int16_interleave2, 3, 4, 2, dst, src0,
src1, len
%if mmsize == 8
emms
%endif
- REP_RET
+ RET
%endmacro
INIT_MMX 3dnow
@@ -373,7 +373,7 @@ cglobal float_interleave6, 2, 8, %1, dst, src, src1, src2,
src3, src4, src5, len
%if mmsize == 8
emms
%endif
- REP_RET
+ RET
%endmacro
INIT_MMX mmx
@@ -416,7 +416,7 @@ cglobal float_interleave2, 3, 4, %1, dst, src, len, src1
%if mmsize == 8
emms
%endif
- REP_RET
+ RET
%endmacro
INIT_MMX mmx
diff --git a/libavcodec/x86/h264_chromamc.asm b/libavcodec/x86/h264_chromamc.asm
index 1be37de..c50b337 100644
--- a/libavcodec/x86/h264_chromamc.asm
+++ b/libavcodec/x86/h264_chromamc.asm
@@ -114,7 +114,7 @@ cglobal %1_%2_chroma_mc8%3, 6, 7 + extra_regs, 0
jne .at_least_one_non_zero
; mx == 0 AND my == 0 - no filter needed
mv0_pixels_mc8
- REP_RET
+ RET
.at_least_one_non_zero:
%ifidn %2, rv40
@@ -194,7 +194,7 @@ cglobal %1_%2_chroma_mc8%3, 6, 7 + extra_regs, 0
add r1, r2
dec r3d
jne .next1drow
- REP_RET
+ RET
.both_non_zero: ; general case, bilinear
movd m4, r4d ; x
@@ -370,7 +370,7 @@ cglobal %1_%2_chroma_mc4, 6, 6 + extra_regs, 0
add r0, r2
sub r3d, 2
jnz .next2rows
- REP_RET
+ RET
%endmacro
%macro chroma_mc2_mmx_func 2
@@ -416,7 +416,7 @@ cglobal %1_%2_chroma_mc2, 6, 7, 0
add r0, r2
sub r3d, 1
jnz .nextrow
- REP_RET
+ RET
%endmacro
%define rnd_1d_h264 pw_4
@@ -474,7 +474,7 @@ cglobal %1_%2_chroma_mc8%3, 6, 7, 8
jne .at_least_one_non_zero
; mx == 0 AND my == 0 - no filter needed
mv0_pixels_mc8
- REP_RET
+ RET
.at_least_one_non_zero:
test r5d, r5d
@@ -535,7 +535,7 @@ cglobal %1_%2_chroma_mc8%3, 6, 7, 8
sub r3d, 2
lea r0, [r0+r2*2]
jg .next2rows
- REP_RET
+ RET
.my_is_zero:
mov r5d, r4d
@@ -572,7 +572,7 @@ cglobal %1_%2_chroma_mc8%3, 6, 7, 8
lea r0, [r0+r2*2]
lea r1, [r1+r2*2]
jg .next2xrows
- REP_RET
+ RET
.mx_is_zero:
mov r4d, r5d
@@ -609,7 +609,7 @@ cglobal %1_%2_chroma_mc8%3, 6, 7, 8
sub r3d, 2
lea r0, [r0+r2*2]
jg .next2yrows
- REP_RET
+ RET
%endmacro
%macro chroma_mc4_ssse3_func 2
@@ -662,7 +662,7 @@ cglobal %1_%2_chroma_mc4, 6, 7, 0
sub r3d, 2
lea r0, [r0+r2*2]
jg .next2rows
- REP_RET
+ RET
%endmacro
%define CHROMAMC_AVG NOTHING
diff --git a/libavcodec/x86/h264_chromamc_10bit.asm
b/libavcodec/x86/h264_chromamc_10bit.asm
index aec7678..cf87027 100644
--- a/libavcodec/x86/h264_chromamc_10bit.asm
+++ b/libavcodec/x86/h264_chromamc_10bit.asm
@@ -69,7 +69,7 @@ cglobal %1_h264_chroma_mc8_10, 6,7,8
jne .at_least_one_non_zero
; mx == 0 AND my == 0 - no filter needed
MV0_PIXELS_MC8
- REP_RET
+ RET
.at_least_one_non_zero:
mov r6d, 2
@@ -104,7 +104,7 @@ cglobal %1_h264_chroma_mc8_10, 6,7,8
add r1, r2
dec r3d
jne .next1drow
- REP_RET
+ RET
.xy_interpolation: ; general case, bilinear
movd m4, r4m ; x
@@ -146,7 +146,7 @@ cglobal %1_h264_chroma_mc8_10, 6,7,8
add r0, r2
dec r3d
jne .next2drow
- REP_RET
+ RET
%endmacro
;-----------------------------------------------------------------------------
@@ -196,7 +196,7 @@ cglobal %1_h264_chroma_mc4_10, 6,6,7
MC4_OP m6, m0
sub r3d, 2
jnz .next2rows
- REP_RET
+ RET
%endmacro
;-----------------------------------------------------------------------------
@@ -236,7 +236,7 @@ cglobal %1_h264_chroma_mc2_10, 6,7
add r0, r2
dec r3d
jnz .nextrow
- REP_RET
+ RET
%endmacro
%macro NOTHING 2-3
diff --git a/libavcodec/x86/h264_deblock.asm b/libavcodec/x86/h264_deblock.asm
index 1f1dbc6..02d14d7 100644
--- a/libavcodec/x86/h264_deblock.asm
+++ b/libavcodec/x86/h264_deblock.asm
@@ -689,7 +689,7 @@ cglobal deblock_%1_luma_intra_8,
4,6,16,ARCH_X86_64*0x50-0x50
LUMA_INTRA_SWAP_PQ
LUMA_INTRA_P012 [r0], [r0+r1], [r0+2*r1], [r0+r5]
.end:
- RET
+ REP_RET
INIT_MMX cpuname
%if ARCH_X86_64
diff --git a/libavcodec/x86/h264_deblock_10bit.asm
b/libavcodec/x86/h264_deblock_10bit.asm
index 3b81ef6..1f5bc82 100644
--- a/libavcodec/x86/h264_deblock_10bit.asm
+++ b/libavcodec/x86/h264_deblock_10bit.asm
@@ -373,7 +373,7 @@ cglobal deblock_v_luma_10, 5,5,15
add r4, 2
dec r3
jg .loop
- REP_RET
+ RET
cglobal deblock_h_luma_10, 5,7,15
shl r2d, 2
@@ -412,7 +412,7 @@ cglobal deblock_h_luma_10, 5,7,15
lea r5, [r5+r1*8]
dec r6
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -646,7 +646,7 @@ cglobal deblock_v_luma_intra_10, 4,7,16
add r4, mmsize
dec r6
jg .loop
- REP_RET
+ RET
;-----------------------------------------------------------------------------
; void deblock_h_luma_intra( uint16_t *pix, int stride, int alpha, int beta )
@@ -871,7 +871,7 @@ cglobal deblock_v_chroma_10, 5,7-(mmsize/16),8*(mmsize/16)
add r4, mmsize/4
dec r6
jg .loop
- REP_RET
+ RET
%else
RET
%endif
@@ -899,7 +899,7 @@ cglobal deblock_v_chroma_intra_10,
4,6-(mmsize/16),8*(mmsize/16)
add r4, mmsize
dec r5
jg .loop
- REP_RET
+ RET
%else
RET
%endif
diff --git a/libavcodec/x86/h264_idct.asm b/libavcodec/x86/h264_idct.asm
index 30cecd9..c123be3 100644
--- a/libavcodec/x86/h264_idct.asm
+++ b/libavcodec/x86/h264_idct.asm
@@ -321,7 +321,7 @@ cglobal h264_idct_add16_8, 5, 7 + npicregs, 0, dst,
block_offset, block, stride,
add r2, 32
cmp r5, 16
jl .nextblock
- REP_RET
+ RET
; ff_h264_idct8_add4_mmx(uint8_t *dst, const int *block_offset,
; DCTELEM *block, int stride, const uint8_t nnzc[6*8])
@@ -388,7 +388,7 @@ cglobal h264_idct_add16_8, 5, 8 + npicregs, 0, dst1,
block_offset, block, stride
add r2, 32
cmp r5, 16
jl .nextblock
- REP_RET
+ RET
.no_dc:
mov r6d, dword [r1+r5*4]
add r6, r0
@@ -398,7 +398,7 @@ cglobal h264_idct_add16_8, 5, 8 + npicregs, 0, dst1,
block_offset, block, stride
add r2, 32
cmp r5, 16
jl .nextblock
- REP_RET
+ RET
INIT_MMX mmx
; ff_h264_idct_add16intra_mmx(uint8_t *dst, const int *block_offset,
@@ -422,7 +422,7 @@ cglobal h264_idct_add16intra_8, 5, 7 + npicregs, 0, dst,
block_offset, block, st
add r2, 32
cmp r5, 16
jl .nextblock
- REP_RET
+ RET
INIT_MMX mmxext
; ff_h264_idct_add16intra_mmxext(uint8_t *dst, const int *block_offset,
@@ -445,7 +445,7 @@ cglobal h264_idct_add16intra_8, 5, 8 + npicregs, 0, dst1,
block_offset, block, s
add r2, 32
cmp r5, 16
jl .nextblock
- REP_RET
+ RET
.try_dc:
movsx r6, word [r2]
test r6, r6
@@ -466,7 +466,7 @@ cglobal h264_idct_add16intra_8, 5, 8 + npicregs, 0, dst1,
block_offset, block, s
add r2, 32
cmp r5, 16
jl .nextblock
- REP_RET
+ RET
; ff_h264_idct8_add4_mmxext(uint8_t *dst, const int *block_offset,
; DCTELEM *block, int stride,
@@ -564,7 +564,7 @@ INIT_MMX cpuname
add r2, 128
cmp r5, 16
jl .nextblock
- REP_RET
+ RET
.no_dc:
INIT_XMM cpuname
mov dst2d, dword [r1+r5*4]
@@ -578,7 +578,7 @@ INIT_XMM cpuname
add r2, 128
cmp r5, 16
jl .nextblock
- REP_RET
+ RET
INIT_MMX mmx
h264_idct_add8_mmx_plane:
diff --git a/libavcodec/x86/h264_intrapred.asm
b/libavcodec/x86/h264_intrapred.asm
index b9db3f4..a1b71c8 100644
--- a/libavcodec/x86/h264_intrapred.asm
+++ b/libavcodec/x86/h264_intrapred.asm
@@ -66,7 +66,7 @@ cglobal pred16x16_vertical_8, 2,3
lea r0, [r0+r1*2]
dec r2
jg .loop
- REP_RET
+ RET
INIT_XMM sse
cglobal pred16x16_vertical_8, 2,3
@@ -82,7 +82,7 @@ cglobal pred16x16_vertical_8, 2,3
lea r0, [r0+r1*2]
dec r2
jg .loop
- REP_RET
+ RET
;-----------------------------------------------------------------------------
; void pred16x16_horizontal_8(uint8_t *src, int stride)
@@ -115,7 +115,7 @@ cglobal pred16x16_horizontal_8, 2,3
lea r0, [r0+r1*2]
dec r2
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmx
@@ -177,7 +177,7 @@ cglobal pred16x16_dc_8, 2,7
lea r4, [r4+r1*2]
dec r3d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -224,7 +224,7 @@ cglobal pred16x16_tm_vp8_8, 2,5
add r0, r1
dec r4d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmx
@@ -266,7 +266,7 @@ cglobal pred16x16_tm_vp8_8, 2,6,6
lea r0, [r0+r1*2]
dec r5d
jg .loop
- REP_RET
+ RET
;-----------------------------------------------------------------------------
; void pred16x16_plane_*_8(uint8_t *src, int stride)
@@ -529,7 +529,7 @@ cglobal pred16x16_plane_%1_8, 2,9,7
lea r0, [r0+r2*2]
dec r4
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmx
@@ -711,7 +711,7 @@ ALIGN 16
lea r0, [r0+r2*2]
dec r4
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmx
@@ -758,7 +758,7 @@ cglobal pred8x8_horizontal_8, 2,3
lea r0, [r0+r1*2]
dec r2
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmx
@@ -898,7 +898,7 @@ cglobal pred8x8_dc_rv40_8, 2,7
lea r4, [r4+r1*2]
dec r3d
jg .loop
- REP_RET
+ RET
;-----------------------------------------------------------------------------
; void pred8x8_tm_vp8_8(uint8_t *src, int stride)
@@ -936,7 +936,7 @@ cglobal pred8x8_tm_vp8_8, 2,6
lea r0, [r0+r1*2]
dec r5d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmx
@@ -971,7 +971,7 @@ cglobal pred8x8_tm_vp8_8, 2,6,4
lea r0, [r0+r1*2]
dec r5d
jg .loop
- REP_RET
+ RET
INIT_XMM ssse3
cglobal pred8x8_tm_vp8_8, 2,3,6
@@ -998,7 +998,7 @@ cglobal pred8x8_tm_vp8_8, 2,3,6
lea r0, [r0+r1*2]
dec r2d
jg .loop
- REP_RET
+ RET
; dest, left, right, src, tmp
; output: %1 = (t[n-1] + t[n]*2 + t[n+1] + 2) >> 2
@@ -2460,7 +2460,7 @@ cglobal pred4x4_tm_vp8_8, 3,6
lea r0, [r0+r2*2]
dec r5d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmx
diff --git a/libavcodec/x86/h264_intrapred_10bit.asm
b/libavcodec/x86/h264_intrapred_10bit.asm
index 1b7974b..43cef59 100644
--- a/libavcodec/x86/h264_intrapred_10bit.asm
+++ b/libavcodec/x86/h264_intrapred_10bit.asm
@@ -326,7 +326,7 @@ cglobal pred8x8_horizontal_10, 2, 3
lea r0, [r0+r1*2]
dec r2d
jg .loop
- REP_RET
+ RET
;-----------------------------------------------------------------------------
; void predict_8x8_dc(pixel *src, int stride)
@@ -502,7 +502,7 @@ cglobal pred8x8_plane_10, 2, 7, 7
add r0, r1
dec r2d
jg .loop
- REP_RET
+ RET
;-----------------------------------------------------------------------------
@@ -1004,7 +1004,7 @@ cglobal pred16x16_vertical_10, 2, 3
lea r0, [r0+r1*2]
dec r2d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -1028,7 +1028,7 @@ cglobal pred16x16_horizontal_10, 2, 3
lea r0, [r0+r1*2]
dec r2d
jg .vloop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -1074,7 +1074,7 @@ cglobal pred16x16_dc_10, 2, 6
lea r5, [r5+r1*2]
dec r3d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -1106,7 +1106,7 @@ cglobal pred16x16_top_dc_10, 2, 3
lea r0, [r0+r1*2]
dec r2d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -1143,7 +1143,7 @@ cglobal pred16x16_left_dc_10, 2, 6
lea r5, [r5+r1*2]
dec r3d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -1164,7 +1164,7 @@ cglobal pred16x16_128_dc_10, 2,3
lea r0, [r0+r1*2]
dec r2d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
diff --git a/libavcodec/x86/h264_qpel_10bit.asm
b/libavcodec/x86/h264_qpel_10bit.asm
index e14df84..122c6dc 100644
--- a/libavcodec/x86/h264_qpel_10bit.asm
+++ b/libavcodec/x86/h264_qpel_10bit.asm
@@ -211,7 +211,7 @@ cglobal %1_h264_qpel16_mc00_10, 3,4
lea r1, [r1+r2*2]
dec r3d
jg .loop
- REP_RET
+ RET
%endmacro
%define OP_MOV mova
diff --git a/libavcodec/x86/h264_qpel_8bit.asm
b/libavcodec/x86/h264_qpel_8bit.asm
index bc6c725..1c2f4f4 100644
--- a/libavcodec/x86/h264_qpel_8bit.asm
+++ b/libavcodec/x86/h264_qpel_8bit.asm
@@ -89,7 +89,7 @@ cglobal %1_h264_qpel4_h_lowpass, 4,5 ; dst, src, dstStride,
srcStride
add r1, r3
dec r4d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -149,7 +149,7 @@ cglobal %1_h264_qpel8_h_lowpass, 4,5 ; dst, src, dstStride,
srcStride
add r1, r3
dec r4d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -192,7 +192,7 @@ cglobal %1_h264_qpel8_h_lowpass, 4,5,8 ; dst, src,
dstStride, srcStride
add r0, r2
dec r4d
jne .loop
- REP_RET
+ RET
%endmacro
INIT_XMM ssse3
@@ -239,7 +239,7 @@ cglobal %1_h264_qpel4_h_lowpass_l2, 5,6 ; dst, src, src2,
dstStride, srcStride
add r2, r4
dec r5d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -303,7 +303,7 @@ cglobal %1_h264_qpel8_h_lowpass_l2, 5,6 ; dst, src, src2,
dstStride, srcStride
add r2, r4
dec r5d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -350,7 +350,7 @@ cglobal %1_h264_qpel8_h_lowpass_l2, 5,6,8 ; dst, src, src2,
dstStride, src2Strid
add r2, r4
dec r5d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM ssse3
@@ -535,7 +535,7 @@ cglobal %1_h264_qpel4_hv_lowpass_h, 3,4 ; tmp, dst,
dstStride
add r1, r2
dec r3d
jnz .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -627,7 +627,7 @@ cglobal %1_h264_qpel8or16_hv2_lowpass_op, 5,5 ; dst, tmp,
dstStride, unused, h
add r0, r2
dec r4d
jne .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -784,7 +784,7 @@ cglobal %1_pixels8_l2_shift5, 6, 6 ; dst, src16, src8,
dstStride, src8Stride, h
lea r0, [r0+2*r3]
sub r5d, 2
jne .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -853,7 +853,7 @@ cglobal %1_h264_qpel16_h_lowpass_l2, 5, 6, 16 ; dst, src,
src2, dstStride, src2S
add r2, r4
dec r5d
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM ssse3
diff --git a/libavcodec/x86/h264_weight.asm b/libavcodec/x86/h264_weight.asm
index 646acdf..11937e6 100644
--- a/libavcodec/x86/h264_weight.asm
+++ b/libavcodec/x86/h264_weight.asm
@@ -81,7 +81,7 @@ cglobal h264_weight_16, 6, 6, 0
add r0, r1
dec r2d
jnz .nextrow
- REP_RET
+ RET
%macro WEIGHT_FUNC_MM 2
cglobal h264_weight_%1, 6, 6, %2
@@ -92,7 +92,7 @@ cglobal h264_weight_%1, 6, 6, %2
add r0, r1
dec r2d
jnz .nextrow
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -117,7 +117,7 @@ cglobal h264_weight_%1, 6, 6, %2
add r0, r3
dec r2d
jnz .nextrow
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -205,7 +205,7 @@ cglobal h264_biweight_16, 7, 8, 0
add r1, r2
dec r3d
jnz .nextrow
- REP_RET
+ RET
%macro BIWEIGHT_FUNC_MM 2
cglobal h264_biweight_%1, 7, 8, %2
@@ -220,7 +220,7 @@ cglobal h264_biweight_%1, 7, 8, %2
add r1, r2
dec r3d
jnz .nextrow
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -249,7 +249,7 @@ cglobal h264_biweight_%1, 7, 8, %2
add r1, r4
dec r3d
jnz .nextrow
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -284,7 +284,7 @@ cglobal h264_biweight_16, 7, 8, 8
add r1, r2
dec r3d
jnz .nextrow
- REP_RET
+ RET
INIT_XMM ssse3
cglobal h264_biweight_8, 7, 8, 8
@@ -307,4 +307,4 @@ cglobal h264_biweight_8, 7, 8, 8
add r1, r4
dec r3d
jnz .nextrow
- REP_RET
+ RET
diff --git a/libavcodec/x86/h264_weight_10bit.asm
b/libavcodec/x86/h264_weight_10bit.asm
index 3b09e42..962fb7b 100644
--- a/libavcodec/x86/h264_weight_10bit.asm
+++ b/libavcodec/x86/h264_weight_10bit.asm
@@ -100,7 +100,7 @@ cglobal h264_weight_16_10
add r0, r1
dec r2d
jnz .nextrow
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -119,7 +119,7 @@ cglobal h264_weight_8_10
add r0, r1
dec r2d
jnz .nextrow
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -141,7 +141,7 @@ cglobal h264_weight_4_10
add r0, r3
dec r2d
jnz .nextrow
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -232,7 +232,7 @@ cglobal h264_biweight_16_10
add r1, r2
dec r3d
jnz .nextrow
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -251,7 +251,7 @@ cglobal h264_biweight_8_10
add r1, r2
dec r3d
jnz .nextrow
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -273,7 +273,7 @@ cglobal h264_biweight_4_10
add r1, r4
dec r3d
jnz .nextrow
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
diff --git a/libavcodec/x86/pngdsp.asm b/libavcodec/x86/pngdsp.asm
index c05f3da..aa17a5f 100644
--- a/libavcodec/x86/pngdsp.asm
+++ b/libavcodec/x86/pngdsp.asm
@@ -78,7 +78,7 @@ cglobal add_bytes_l2, 4, 6, %1, dst, src1, src2, wa, w, i
.end_s:
cmp iq, wq
jl .loop_s
- REP_RET
+ RET
%endmacro
%if ARCH_X86_32
diff --git a/libavcodec/x86/rv34dsp.asm b/libavcodec/x86/rv34dsp.asm
index c099ac5..4a36d05 100644
--- a/libavcodec/x86/rv34dsp.asm
+++ b/libavcodec/x86/rv34dsp.asm
@@ -54,7 +54,7 @@ cglobal rv34_idct_%1, 1, 2, 0
movq [r0+ 8], m0
movq [r0+16], m0
movq [r0+24], m0
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
diff --git a/libavcodec/x86/rv40dsp.asm b/libavcodec/x86/rv40dsp.asm
index 7ec72be..e058294 100644
--- a/libavcodec/x86/rv40dsp.asm
+++ b/libavcodec/x86/rv40dsp.asm
@@ -175,7 +175,7 @@ cglobal %1_rv40_qpel_v, 6,6+npicregs,12, dst, dststride,
src, srcstride, height,
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
%endmacro
%macro FILTER_H 1
@@ -232,7 +232,7 @@ cglobal %1_rv40_qpel_h, 6, 6+npicregs, 12, dst, dststride,
src, srcstride, heigh
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
%endmacro
%if ARCH_X86_32
@@ -299,7 +299,7 @@ cglobal %1_rv40_qpel_v, 6,6+npicregs,8, dst, dststride,
src, srcstride, height,
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
cglobal %1_rv40_qpel_h, 6,6+npicregs,8, dst, dststride, src, srcstride,
height, mx, picreg
%ifdef PIC
@@ -332,7 +332,7 @@ cglobal %1_rv40_qpel_h, 6,6+npicregs,8, dst, dststride,
src, srcstride, height,
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
%endmacro
INIT_XMM ssse3
@@ -483,7 +483,7 @@ cglobal rv40_weight_func_%1_%2, 6, 7, 8
.loop:
MAIN_LOOP %2, RND
jnz .loop
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
diff --git a/libavcodec/x86/sbrdsp.asm b/libavcodec/x86/sbrdsp.asm
index b87da4a..93ba7c5 100644
--- a/libavcodec/x86/sbrdsp.asm
+++ b/libavcodec/x86/sbrdsp.asm
@@ -114,7 +114,7 @@ cglobal sbr_hf_g_filt, 5, 6, 5
dec r5
jnz .loop1
.end:
- RET
+ REP_RET
; static void sbr_hf_gen_c(float (*X_high)[2], const float (*X_low)[2],
; const float alpha0[2], const float alpha1[2],
diff --git a/libavcodec/x86/vp8dsp.asm b/libavcodec/x86/vp8dsp.asm
index af8403e..3e64cab 100644
--- a/libavcodec/x86/vp8dsp.asm
+++ b/libavcodec/x86/vp8dsp.asm
@@ -215,7 +215,7 @@ cglobal put_vp8_epel%1_h6, 6, 6 + npicregs, 8, dst,
dststride, src, srcstride, h
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
cglobal put_vp8_epel%1_h4, 6, 6 + npicregs, 7, dst, dststride, src, srcstride,
height, mx, picreg
shl mxd, 4
@@ -245,7 +245,7 @@ cglobal put_vp8_epel%1_h4, 6, 6 + npicregs, 7, dst,
dststride, src, srcstride, h
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
cglobal put_vp8_epel%1_v4, 7, 7, 8, dst, dststride, src, srcstride, height,
picreg, my
shl myd, 4
@@ -283,7 +283,7 @@ cglobal put_vp8_epel%1_v4, 7, 7, 8, dst, dststride, src,
srcstride, height, picr
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
cglobal put_vp8_epel%1_v6, 7, 7, 8, dst, dststride, src, srcstride, height,
picreg, my
lea myd, [myq*3]
@@ -329,7 +329,7 @@ cglobal put_vp8_epel%1_v6, 7, 7, 8, dst, dststride, src,
srcstride, height, picr
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
%endmacro
INIT_MMX ssse3
@@ -383,7 +383,7 @@ cglobal put_vp8_epel4_h4, 6, 6 + npicregs, 0, dst,
dststride, src, srcstride, he
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
; 4x4 block, H-only 6-tap filter
INIT_MMX mmxext
@@ -441,7 +441,7 @@ cglobal put_vp8_epel4_h6, 6, 6 + npicregs, 0, dst,
dststride, src, srcstride, he
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
INIT_XMM sse2
cglobal put_vp8_epel8_h4, 6, 6 + npicregs, 10, dst, dststride, src, srcstride,
height, mx, picreg
@@ -489,7 +489,7 @@ cglobal put_vp8_epel8_h4, 6, 6 + npicregs, 10, dst,
dststride, src, srcstride, h
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
INIT_XMM sse2
cglobal put_vp8_epel8_h6, 6, 6 + npicregs, 14, dst, dststride, src, srcstride,
height, mx, picreg
@@ -552,7 +552,7 @@ cglobal put_vp8_epel8_h6, 6, 6 + npicregs, 14, dst,
dststride, src, srcstride, h
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
%macro FILTER_V 1
; 4x4 block, V-only 4-tap filter
@@ -605,7 +605,7 @@ cglobal put_vp8_epel%1_v4, 7, 7, 8, dst, dststride, src,
srcstride, height, picr
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
; 4x4 block, V-only 6-tap filter
@@ -670,7 +670,7 @@ cglobal put_vp8_epel%1_v6, 7, 7, 8, dst, dststride, src,
srcstride, height, picr
add srcq, srcstrideq
dec heightd ; next row
jg .nextrow
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -721,7 +721,7 @@ cglobal put_vp8_bilinear%1_v, 7, 7, 7, dst, dststride, src,
srcstride, height, p
lea srcq, [srcq+srcstrideq*2]
sub heightd, 2
jg .nextrow
- REP_RET
+ RET
cglobal put_vp8_bilinear%1_h, 6, 6 + npicregs, 7, dst, dststride, src,
srcstride, height, mx, picreg
shl mxd, 4
@@ -766,7 +766,7 @@ cglobal put_vp8_bilinear%1_h, 6, 6 + npicregs, 7, dst,
dststride, src, srcstride
lea srcq, [srcq+srcstrideq*2]
sub heightd, 2
jg .nextrow
- REP_RET
+ RET
%endmacro
INIT_MMX mmxext
@@ -809,7 +809,7 @@ cglobal put_vp8_bilinear%1_v, 7, 7, 5, dst, dststride, src,
srcstride, height, p
lea srcq, [srcq+srcstrideq*2]
sub heightd, 2
jg .nextrow
- REP_RET
+ RET
cglobal put_vp8_bilinear%1_h, 6, 6 + npicregs, 5, dst, dststride, src,
srcstride, height, mx, picreg
shl mxd, 4
@@ -845,7 +845,7 @@ cglobal put_vp8_bilinear%1_h, 6, 6 + npicregs, 5, dst,
dststride, src, srcstride
lea srcq, [srcq+srcstrideq*2]
sub heightd, 2
jg .nextrow
- REP_RET
+ RET
%endmacro
INIT_MMX ssse3
@@ -864,7 +864,7 @@ cglobal put_vp8_pixels8, 5, 5, 0, dst, dststride, src,
srcstride, height
lea dstq, [dstq+dststrideq*2]
sub heightd, 2
jg .nextrow
- REP_RET
+ RET
%if ARCH_X86_32
INIT_MMX mmx
@@ -882,7 +882,7 @@ cglobal put_vp8_pixels16, 5, 5, 0, dst, dststride, src,
srcstride, height
lea dstq, [dstq+dststrideq*2]
sub heightd, 2
jg .nextrow
- REP_RET
+ RET
%endif
INIT_XMM sse
@@ -896,7 +896,7 @@ cglobal put_vp8_pixels16, 5, 5, 2, dst, dststride, src,
srcstride, height
lea dstq, [dstq+dststrideq*2]
sub heightd, 2
jg .nextrow
- REP_RET
+ RET
;-----------------------------------------------------------------------------
; void vp8_idct_dc_add_<opt>(uint8_t *dst, DCTELEM block[16], int stride);
@@ -1601,7 +1601,7 @@ cglobal vp8_%1_loop_filter_simple, 3, %2, 8, dst, stride,
flim, cntr
%endif
dec cntrq
jg .next8px
- REP_RET
+ RET
%else ; sse2
RET
%endif
@@ -2085,7 +2085,7 @@ cglobal vp8_%1_loop_filter16y_inner, 5, 5, 13,
stack_size, dst, stride, flimE, f
dec cntrq
jg .next8px
%endif
- REP_RET
+ RET
%else ; mmsize == 16
RET
%endif
@@ -2744,7 +2744,7 @@ cglobal vp8_%1_loop_filter16y_mbedge, 5, 5, 15,
stack_size, dst1, stride, flimE,
dec cntrq
jg .next8px
%endif
- REP_RET
+ RET
%else ; mmsize == 16
RET
%endif
diff --git a/libavfilter/x86/af_volume.asm b/libavfilter/x86/af_volume.asm
index 4e5ad22..a031bd8 100644
--- a/libavfilter/x86/af_volume.asm
+++ b/libavfilter/x86/af_volume.asm
@@ -56,7 +56,7 @@ cglobal scale_samples_s16, 4,4,4, dst, src, len, volume
mova [dstq+lenq], m3
sub lenq, mmsize
jge .loop
- REP_RET
+ RET
;------------------------------------------------------------------------------
; void ff_scale_samples_s32(uint8_t *dst, const uint8_t *src, int len,
@@ -93,7 +93,7 @@ cglobal scale_samples_s32, 4,4,4, dst, src, len, volume
%endif
sub lenq, mmsize
jge .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -135,4 +135,4 @@ cglobal scale_samples_s32, 4,4,8, dst, src, len, volume
mova [dstq+lenq], m0
sub lenq, mmsize
jge .loop
- REP_RET
+ RET
diff --git a/libavfilter/x86/hqdn3d.asm b/libavfilter/x86/hqdn3d.asm
index dee2c96..10e503b 100644
--- a/libavfilter/x86/hqdn3d.asm
+++ b/libavfilter/x86/hqdn3d.asm
@@ -97,7 +97,7 @@ ALIGN 16
inc xq
jl .loop
je .loop2
- REP_RET
+ RET
%endmacro ; HQDN3D_ROW
HQDN3D_ROW 8
diff --git a/libavresample/x86/audio_convert.asm
b/libavresample/x86/audio_convert.asm
index 1af1429..2ce970b 100644
--- a/libavresample/x86/audio_convert.asm
+++ b/libavresample/x86/audio_convert.asm
@@ -58,7 +58,7 @@ cglobal conv_s16_to_s32, 3,3,3, dst, src, len
mova [dstq+2*lenq+mmsize], m1
add lenq, mmsize
jl .loop
- REP_RET
+ RET
;------------------------------------------------------------------------------
; void ff_conv_s16_to_flt(float *dst, const int16_t *src, int len);
@@ -83,7 +83,7 @@ cglobal conv_s16_to_flt, 3,3,3, dst, src, len
mova [dstq+2*lenq+mmsize], m1
add lenq, mmsize
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -120,7 +120,7 @@ cglobal conv_s32_to_s16, 3,3,4, dst, src, len
emms
RET
%else
- REP_RET
+ RET
%endif
%endmacro
@@ -150,7 +150,7 @@ cglobal conv_s32_to_flt, 3,3,3, dst, src, len
mova [dstq+lenq+mmsize], m2
add lenq, mmsize*2
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -188,7 +188,7 @@ cglobal conv_flt_to_s16, 3,3,5, dst, src, len
mova [dstq+lenq+mmsize], m2
add lenq, mmsize*2
jl .loop
- REP_RET
+ RET
;------------------------------------------------------------------------------
; void ff_conv_flt_to_s32(int32_t *dst, const float *src, int len);
@@ -221,7 +221,7 @@ cglobal conv_flt_to_s32, 3,3,6, dst, src, len
mova [dstq+lenq+3*mmsize], m3
add lenq, mmsize*4
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -256,7 +256,7 @@ cglobal conv_s16p_to_s16_2ch, 3,4,5, dst, src0, len, src1
mova [dstq+2*lenq+3*mmsize], m3
add lenq, 2*mmsize
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -375,7 +375,7 @@ cglobal conv_s16p_to_s16_6ch, 2,7,7, dst, src0, src1, src2,
src3, src4, src5
sub lend, mmsize/2
%endif
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -424,7 +424,7 @@ cglobal conv_s16p_to_flt_2ch, 3,4,6, dst, src0, len, src1
mova [dstq+4*lenq+3*mmsize], m3
add lenq, mmsize
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -524,7 +524,7 @@ cglobal conv_s16p_to_flt_6ch, 2,7,8, dst, src, src1, src2,
src3, src4, src5
add dstq, mmsize*6
sub lend, mmsize/4
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -568,7 +568,7 @@ cglobal conv_fltp_to_s16_2ch, 3,4,3, dst, src0, len, src1
mova [dstq+lenq], m0
add lenq, mmsize
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -677,7 +677,7 @@ cglobal conv_fltp_to_s16_6ch, 2,7,7, dst, src, src1, src2,
src3, src4, src5
emms
RET
%else
- REP_RET
+ RET
%endif
%endmacro
@@ -715,7 +715,7 @@ cglobal conv_fltp_to_flt_2ch, 3,4,5, dst, src0, len, src1
mova [dstq+2*lenq+3*mmsize], m3
add lenq, 2*mmsize
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse
@@ -791,7 +791,7 @@ cglobal conv_fltp_to_flt_6ch, 2,8,7, dst, src, src1, src2,
src3, src4, src5, len
emms
RET
%else
- REP_RET
+ RET
%endif
%endmacro
@@ -839,7 +839,7 @@ cglobal conv_s16_to_s16p_2ch, 3,4,4, dst0, src, len, dst1
mova [dst1q+lenq], m1
add lenq, mmsize
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -896,7 +896,7 @@ cglobal conv_s16_to_s16p_6ch, 2,7,5, dst, src, dst1, dst2,
dst3, dst4, dst5
add dstq, mmsize/2
sub lend, mmsize/4
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -934,7 +934,7 @@ cglobal conv_s16_to_fltp_2ch, 3,4,5, dst0, src, len, dst1
mova [dst1q+lenq], m1
add lenq, mmsize
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -1009,7 +1009,7 @@ cglobal conv_s16_to_fltp_6ch, 2,7,7, dst, src, dst1,
dst2, dst3, dst4, dst5
add dstq, mmsize
sub lend, mmsize/4
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -1057,7 +1057,7 @@ cglobal conv_flt_to_s16p_2ch, 3,4,6, dst0, src, len, dst1
mova [dst1q+lenq], m1
add lenq, mmsize
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -1125,7 +1125,7 @@ cglobal conv_flt_to_s16p_6ch, 2,7,7, dst, src, dst1,
dst2, dst3, dst4, dst5
add dstq, mmsize/2
sub lend, mmsize/4
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -1157,7 +1157,7 @@ cglobal conv_flt_to_fltp_2ch, 3,4,3, dst0, src, len, dst1
mova [dst1q+lenq], m1
add lenq, mmsize
jl .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse
@@ -1218,7 +1218,7 @@ cglobal conv_flt_to_fltp_6ch, 2,7,7, dst, src, dst1,
dst2, dst3, dst4, dst5
add dstq, mmsize
sub lend, mmsize/4
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
diff --git a/libavresample/x86/audio_mix.asm b/libavresample/x86/audio_mix.asm
index daf0a7b..a25c893 100644
--- a/libavresample/x86/audio_mix.asm
+++ b/libavresample/x86/audio_mix.asm
@@ -50,7 +50,7 @@ cglobal mix_2_to_1_fltp_flt, 3,4,6, src, matrix, len, src1
add srcq, mmsize*2
sub lend, mmsize*2/4
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse
@@ -94,7 +94,7 @@ cglobal mix_2_to_1_s16p_flt, 3,4,6, src, matrix, len, src1
add srcq, mmsize
sub lend, mmsize/2
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
@@ -141,7 +141,7 @@ cglobal mix_2_to_1_s16p_q8, 3,4,6, src, matrix, len, src1
add srcq, mmsize
sub lend, mmsize/2
jg .loop
- REP_RET
+ RET
;-----------------------------------------------------------------------------
; void ff_mix_1_to_2_fltp_flt(float **src, float **matrix, int len,
@@ -167,7 +167,7 @@ cglobal mix_1_to_2_fltp_flt, 3,5,4, src0, matrix0, len,
src1, matrix1
add src0q, mmsize
sub lend, mmsize/4
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse
@@ -210,7 +210,7 @@ cglobal mix_1_to_2_s16p_flt, 3,5,6, src0, matrix0, len,
src1, matrix1
add src0q, mmsize
sub lend, mmsize/2
jg .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
diff --git a/libavutil/x86/float_dsp.asm b/libavutil/x86/float_dsp.asm
index 4113fd9..fd22d7f 100644
--- a/libavutil/x86/float_dsp.asm
+++ b/libavutil/x86/float_dsp.asm
@@ -39,7 +39,7 @@ ALIGN 16
sub lenq, 2*mmsize
jge .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse
@@ -78,7 +78,7 @@ cglobal vector_fmac_scalar, 4,4,3, dst, src, mul, len
mova [dstq+lenq+mmsize], m2
sub lenq, 2*mmsize
jge .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse
@@ -109,7 +109,7 @@ cglobal vector_fmul_scalar, 4,4,3, dst, src, mul, len
mova [dstq+lenq], m1
sub lenq, mmsize
jge .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse
@@ -153,7 +153,7 @@ cglobal vector_dmul_scalar, 4,4,3, dst, src, mul, len
mova [dstq+lenq+mmsize], m2
sub lenq, 2*mmsize
jge .loop
- REP_RET
+ RET
%endmacro
INIT_XMM sse2
diff --git a/libswscale/x86/input.asm b/libswscale/x86/input.asm
index a8d5a5a..4db74df 100644
--- a/libswscale/x86/input.asm
+++ b/libswscale/x86/input.asm
@@ -164,7 +164,7 @@ cglobal %2 %+ 24ToY, 3, 3, %1, dst, src, w
movh [dstq+wq], m0
add wq, mmsize / 2
jl .loop
- REP_RET
+ RET
%endif ; (ARCH_X86_64 && %0 == 3) || mmsize == 8
%endmacro
@@ -281,7 +281,7 @@ cglobal %2 %+ 24ToUV, 3, 4, %1, dstU, dstV, src, w
%endif ; mmsize == 8/16
add wq, mmsize / 2
jl .loop
- REP_RET
+ RET
%endif ; ARCH_X86_64 && %0 == 3
%endmacro
@@ -347,7 +347,7 @@ cglobal %2%3%4%5 %+ ToY, 3, 3, %1, dst, src, w
movh [dstq+wq], m0
add wq, mmsize / 2
jl .loop
- REP_RET
+ RET
%endif ; %0 == 3
%endmacro
@@ -425,7 +425,7 @@ cglobal %2%3%4%5 %+ ToUV, 3, 4, %1, dstU, dstV, src, w
%endif ; mmsize == 8/16
add wq, mmsize / 2
jl .loop
- REP_RET
+ RET
%endif ; ARCH_X86_64 && %0 == 3
%endmacro
@@ -480,7 +480,7 @@ RGB32_FUNCS 8, 12
mova [dstq+wq], m0
add wq, mmsize
jl .loop_%1
- REP_RET
+ RET
%endmacro
; %1 = nr. of XMM registers
@@ -550,7 +550,7 @@ cglobal %2ToY, 3, 3, %1, dst, src, w
%endif ; mmsize == 8/16
add wq, mmsize / 2
jl .loop_%1
- REP_RET
+ RET
%endmacro
; %1 = nr. of XMM registers
@@ -608,7 +608,7 @@ cglobal %2ToUV, 3, 4, %1, dstU, dstV, src, w
%endif ; nv12/21
add wq, mmsize
jl .loop_%1
- REP_RET
+ RET
%endmacro
; %1 = nr. of XMM registers
diff --git a/libswscale/x86/output.asm b/libswscale/x86/output.asm
index e1ceded..478f27c 100644
--- a/libswscale/x86/output.asm
+++ b/libswscale/x86/output.asm
@@ -239,10 +239,10 @@ cglobal yuv2planeX_%1, %3, 8, %2, filter, fltsize, src,
dst, w, dither, offset
ADD rsp, pad
RET
%else ; x86-64
- REP_RET
+ RET
%endif ; x86-32/64
%else ; %1 == 9/10/16
- REP_RET
+ RET
%endif ; %1 == 8/9/10/16
%endmacro
@@ -376,11 +376,11 @@ cglobal yuv2plane1_%1, %3, %3, %2, src, dst, w, dither,
offset
test dstq, 15
jnz .unaligned
yuv2plane1_mainloop %1, a
- REP_RET
+ RET
.unaligned:
yuv2plane1_mainloop %1, u
%endif ; mmsize == 8/16
- REP_RET
+ RET
%endmacro
%if ARCH_X86_32
diff --git a/libswscale/x86/scale.asm b/libswscale/x86/scale.asm
index 440a27b..1f0c7c9 100644
--- a/libswscale/x86/scale.asm
+++ b/libswscale/x86/scale.asm
@@ -386,7 +386,7 @@ cglobal hscale%1to%2_%4, %5, 10, %6, pos0, dst, w, srcmem,
filter, fltpos, fltsi
add wq, 2
%endif ; %3 ==/!= X
jl .loop
- REP_RET
+ RET
%endmacro
; SCALE_FUNCS source_width, intermediate_nbits, n_xmm
--
1.7.12
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
