Skip to content

F
ffmpeg.wasm-core
Commit 1e9c5bf4 authored by Diego Biurrun's avatar Diego Biurrun
Browse files
Options

asm: FF_-prefix internal macros used in inline assembly 

These warnings conflict with system macros on Solaris, producing
truckloads of warnings about macro redefinition.
parent dc40a70c
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 677 additions and 677 deletions
libavcodec/x86/blockdsp.c
View file @ 1e9c5bf4
...@@ -33,18 +33,18 @@ ...@@ -33,18 +33,18 @@
static void name(int16_t *blocks) \ static void name(int16_t *blocks) \
{ \ { \
__asm__ volatile ( \ __asm__ volatile ( \
"pxor %%mm7, %%mm7 \n\t" \ "pxor %%mm7, %%mm7 \n\t" \
"mov %1, %%"REG_a" \n\t" \ "mov %1, %%"FF_REG_a" \n\t" \
"1: \n\t" \ "1: \n\t" \
"movq %%mm7, (%0, %%"REG_a") \n\t" \ "movq %%mm7, (%0, %%"FF_REG_a") \n\t" \
"movq %%mm7, 8(%0, %%"REG_a") \n\t" \ "movq %%mm7, 8(%0, %%"FF_REG_a") \n\t" \
"movq %%mm7, 16(%0, %%"REG_a") \n\t" \ "movq %%mm7, 16(%0, %%"FF_REG_a") \n\t" \
"movq %%mm7, 24(%0, %%"REG_a") \n\t" \ "movq %%mm7, 24(%0, %%"FF_REG_a") \n\t" \
"add $32, %%"REG_a" \n\t" \ "add $32, %%"FF_REG_a" \n\t" \
"js 1b \n\t" \ "js 1b \n\t" \
:: "r"(((uint8_t *) blocks) + 128 * n), \ :: "r"(((uint8_t *) blocks) + 128 * n), \
"i"(-128 * n) \ "i"(-128 * n) \
: "%"REG_a); \ : "%"FF_REG_a); \
} }
CLEAR_BLOCKS(clear_blocks_mmx, 6) CLEAR_BLOCKS(clear_blocks_mmx, 6)
CLEAR_BLOCKS(clear_block_mmx, 1) CLEAR_BLOCKS(clear_block_mmx, 1)
...@@ -68,21 +68,21 @@ static void clear_block_sse(int16_t *block) ...@@ -68,21 +68,21 @@ static void clear_block_sse(int16_t *block)
static void clear_blocks_sse(int16_t *blocks) static void clear_blocks_sse(int16_t *blocks)
{ {
__asm__ volatile ( __asm__ volatile (
"xorps %%xmm0, %%xmm0 \n" "xorps %%xmm0, %%xmm0 \n"
"mov %1, %%"REG_a" \n" "mov %1, %%"FF_REG_a" \n"
"1: \n" "1: \n"
"movaps %%xmm0, (%0, %%"REG_a") \n" "movaps %%xmm0, (%0, %%"FF_REG_a") \n"
"movaps %%xmm0, 16(%0, %%"REG_a") \n" "movaps %%xmm0, 16(%0, %%"FF_REG_a") \n"
"movaps %%xmm0, 32(%0, %%"REG_a") \n" "movaps %%xmm0, 32(%0, %%"FF_REG_a") \n"
"movaps %%xmm0, 48(%0, %%"REG_a") \n" "movaps %%xmm0, 48(%0, %%"FF_REG_a") \n"
"movaps %%xmm0, 64(%0, %%"REG_a") \n" "movaps %%xmm0, 64(%0, %%"FF_REG_a") \n"
"movaps %%xmm0, 80(%0, %%"REG_a") \n" "movaps %%xmm0, 80(%0, %%"FF_REG_a") \n"
"movaps %%xmm0, 96(%0, %%"REG_a") \n" "movaps %%xmm0, 96(%0, %%"FF_REG_a") \n"
"movaps %%xmm0, 112(%0, %%"REG_a") \n" "movaps %%xmm0, 112(%0, %%"FF_REG_a") \n"
"add $128, %%"REG_a" \n" "add $128, %%"FF_REG_a" \n"
"js 1b \n" "js 1b \n"
:: "r"(((uint8_t *) blocks) + 128 * 6), "i"(-128 * 6) :: "r"(((uint8_t *) blocks) + 128 * 6), "i"(-128 * 6)
: "%"REG_a); : "%"FF_REG_a);
} }
#endif /* HAVE_INLINE_ASM */ #endif /* HAVE_INLINE_ASM */
......
libavcodec/x86/cabac.h
View file @ 1e9c5bf4
...@@ -72,12 +72,12 @@ ...@@ -72,12 +72,12 @@
"mov "tmpbyte" , "statep" \n\t"\ "mov "tmpbyte" , "statep" \n\t"\
"test "lowword" , "lowword" \n\t"\ "test "lowword" , "lowword" \n\t"\
"jnz 2f \n\t"\ "jnz 2f \n\t"\
"mov "byte" , %%"REG_c" \n\t"\ "mov "byte" , %%"FF_REG_c" \n\t"\
"cmp "end" , %%"REG_c" \n\t"\ "cmp "end" , %%"FF_REG_c" \n\t"\
"jge 1f \n\t"\ "jge 1f \n\t"\
"add"OPSIZE" $2 , "byte" \n\t"\ "add"FF_OPSIZE" $2 , "byte" \n\t"\
"1: \n\t"\ "1: \n\t"\
"movzwl (%%"REG_c") , "tmp" \n\t"\ "movzwl (%%"FF_REG_c"), "tmp" \n\t"\
"lea -1("low") , %%ecx \n\t"\ "lea -1("low") , %%ecx \n\t"\
"xor "low" , %%ecx \n\t"\ "xor "low" , %%ecx \n\t"\
"shr $15 , %%ecx \n\t"\ "shr $15 , %%ecx \n\t"\
...@@ -133,12 +133,12 @@ ...@@ -133,12 +133,12 @@
"mov "tmpbyte" , "statep" \n\t"\ "mov "tmpbyte" , "statep" \n\t"\
"test "lowword" , "lowword" \n\t"\ "test "lowword" , "lowword" \n\t"\
" jnz 2f \n\t"\ " jnz 2f \n\t"\
"mov "byte" , %%"REG_c" \n\t"\ "mov "byte" , %%"FF_REG_c" \n\t"\
"cmp "end" , %%"REG_c" \n\t"\ "cmp "end" , %%"FF_REG_c" \n\t"\
"jge 1f \n\t"\ "jge 1f \n\t"\
"add"OPSIZE" $2 , "byte" \n\t"\ "add"FF_OPSIZE" $2 , "byte" \n\t"\
"1: \n\t"\ "1: \n\t"\
"movzwl (%%"REG_c") , "tmp" \n\t"\ "movzwl (%%"FF_REG_c") , "tmp" \n\t"\
"lea -1("low") , %%ecx \n\t"\ "lea -1("low") , %%ecx \n\t"\
"xor "low" , %%ecx \n\t"\ "xor "low" , %%ecx \n\t"\
"shr $15 , %%ecx \n\t"\ "shr $15 , %%ecx \n\t"\
...@@ -183,7 +183,7 @@ static av_always_inline int get_cabac_inline_x86(CABACContext *c, ...@@ -183,7 +183,7 @@ static av_always_inline int get_cabac_inline_x86(CABACContext *c,
"i"(offsetof(CABACContext, bytestream)), "i"(offsetof(CABACContext, bytestream)),
"i"(offsetof(CABACContext, bytestream_end)) "i"(offsetof(CABACContext, bytestream_end))
TABLES_ARG TABLES_ARG
: "%"REG_c, "memory" : "%"FF_REG_c, "memory"
); );
return bit & 1; return bit & 1;
} }
...@@ -214,7 +214,7 @@ static av_always_inline int get_cabac_bypass_sign_x86(CABACContext *c, int val) ...@@ -214,7 +214,7 @@ static av_always_inline int get_cabac_bypass_sign_x86(CABACContext *c, int val)
"addl %%edx, %%eax \n\t" "addl %%edx, %%eax \n\t"
"cmp %c5(%2), %1 \n\t" "cmp %c5(%2), %1 \n\t"
"jge 1f \n\t" "jge 1f \n\t"
"add"OPSIZE" $2, %c4(%2) \n\t" "add"FF_OPSIZE" $2, %c4(%2) \n\t"
"1: \n\t" "1: \n\t"
"movl %%eax, %c3(%2) \n\t" "movl %%eax, %c3(%2) \n\t"
...@@ -254,7 +254,7 @@ static av_always_inline int get_cabac_bypass_x86(CABACContext *c) ...@@ -254,7 +254,7 @@ static av_always_inline int get_cabac_bypass_x86(CABACContext *c)
"addl %%ecx, %%eax \n\t" "addl %%ecx, %%eax \n\t"
"cmp %c5(%2), %1 \n\t" "cmp %c5(%2), %1 \n\t"
"jge 1f \n\t" "jge 1f \n\t"
"add"OPSIZE" $2, %c4(%2) \n\t" "add"FF_OPSIZE" $2, %c4(%2) \n\t"
"1: \n\t" "1: \n\t"
"movl %%eax, %c3(%2) \n\t" "movl %%eax, %c3(%2) \n\t"
......
libavcodec/x86/fpel_mmx.c
View file @ 1e9c5bf4
...@@ -79,26 +79,26 @@ void ff_put_pixels8_mmx(uint8_t *block, const uint8_t *pixels, ...@@ -79,26 +79,26 @@ void ff_put_pixels8_mmx(uint8_t *block, const uint8_t *pixels,
ptrdiff_t line_size, int h) ptrdiff_t line_size, int h)
{ {
__asm__ volatile ( __asm__ volatile (
"lea (%3, %3), %%"REG_a" \n\t" "lea (%3, %3), %%"FF_REG_a" \n\t"
".p2align 3 \n\t" ".p2align 3 \n\t"
"1: \n\t" "1: \n\t"
"movq (%1 ), %%mm0 \n\t" "movq (%1 ), %%mm0 \n\t"
"movq (%1, %3), %%mm1 \n\t" "movq (%1, %3), %%mm1 \n\t"
"movq %%mm0, (%2) \n\t" "movq %%mm0, (%2) \n\t"
"movq %%mm1, (%2, %3) \n\t" "movq %%mm1, (%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"movq (%1 ), %%mm0 \n\t" "movq (%1 ), %%mm0 \n\t"
"movq (%1, %3), %%mm1 \n\t" "movq (%1, %3), %%mm1 \n\t"
"movq %%mm0, (%2) \n\t" "movq %%mm0, (%2) \n\t"
"movq %%mm1, (%2, %3) \n\t" "movq %%mm1, (%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"subl $4, %0 \n\t" "subl $4, %0 \n\t"
"jnz 1b \n\t" "jnz 1b \n\t"
: "+g"(h), "+r"(pixels), "+r"(block) : "+g"(h), "+r"(pixels), "+r"(block)
: "r"((x86_reg)line_size) : "r"((x86_reg)line_size)
: "%"REG_a, "memory" : "%"FF_REG_a, "memory"
); );
} }
...@@ -106,7 +106,7 @@ void ff_put_pixels16_mmx(uint8_t *block, const uint8_t *pixels, ...@@ -106,7 +106,7 @@ void ff_put_pixels16_mmx(uint8_t *block, const uint8_t *pixels,
ptrdiff_t line_size, int h) ptrdiff_t line_size, int h)
{ {
__asm__ volatile ( __asm__ volatile (
"lea (%3, %3), %%"REG_a" \n\t" "lea (%3, %3), %%"FF_REG_a" \n\t"
".p2align 3 \n\t" ".p2align 3 \n\t"
"1: \n\t" "1: \n\t"
"movq (%1 ), %%mm0 \n\t" "movq (%1 ), %%mm0 \n\t"
...@@ -117,8 +117,8 @@ void ff_put_pixels16_mmx(uint8_t *block, const uint8_t *pixels, ...@@ -117,8 +117,8 @@ void ff_put_pixels16_mmx(uint8_t *block, const uint8_t *pixels,
"movq %%mm4, 8(%2) \n\t" "movq %%mm4, 8(%2) \n\t"
"movq %%mm1, (%2, %3) \n\t" "movq %%mm1, (%2, %3) \n\t"
"movq %%mm5, 8(%2, %3) \n\t" "movq %%mm5, 8(%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"movq (%1 ), %%mm0 \n\t" "movq (%1 ), %%mm0 \n\t"
"movq 8(%1 ), %%mm4 \n\t" "movq 8(%1 ), %%mm4 \n\t"
"movq (%1, %3), %%mm1 \n\t" "movq (%1, %3), %%mm1 \n\t"
...@@ -127,13 +127,13 @@ void ff_put_pixels16_mmx(uint8_t *block, const uint8_t *pixels, ...@@ -127,13 +127,13 @@ void ff_put_pixels16_mmx(uint8_t *block, const uint8_t *pixels,
"movq %%mm4, 8(%2) \n\t" "movq %%mm4, 8(%2) \n\t"
"movq %%mm1, (%2, %3) \n\t" "movq %%mm1, (%2, %3) \n\t"
"movq %%mm5, 8(%2, %3) \n\t" "movq %%mm5, 8(%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"subl $4, %0 \n\t" "subl $4, %0 \n\t"
"jnz 1b \n\t" "jnz 1b \n\t"
: "+g"(h), "+r"(pixels), "+r"(block) : "+g"(h), "+r"(pixels), "+r"(block)
: "r"((x86_reg)line_size) : "r"((x86_reg)line_size)
: "%"REG_a, "memory" : "%"FF_REG_a, "memory"
); );
} }
......
libavcodec/x86/h264_i386.h
View file @ 1e9c5bf4
...@@ -84,13 +84,13 @@ static int decode_significance_x86(CABACContext *c, int max_coeff, ...@@ -84,13 +84,13 @@ static int decode_significance_x86(CABACContext *c, int max_coeff,
"sub %10, %1 \n\t" "sub %10, %1 \n\t"
"mov %2, %0 \n\t" "mov %2, %0 \n\t"
"movl %7, %%ecx \n\t" "movl %7, %%ecx \n\t"
"add %1, %%"REG_c" \n\t" "add %1, %%"FF_REG_c" \n\t"
"movl %%ecx, (%0) \n\t" "movl %%ecx, (%0) \n\t"
"test $1, %4 \n\t" "test $1, %4 \n\t"
" jnz 5f \n\t" " jnz 5f \n\t"
"add"OPSIZE" $4, %2 \n\t" "add"FF_OPSIZE" $4, %2 \n\t"
"4: \n\t" "4: \n\t"
"add $1, %1 \n\t" "add $1, %1 \n\t"
...@@ -98,7 +98,7 @@ static int decode_significance_x86(CABACContext *c, int max_coeff, ...@@ -98,7 +98,7 @@ static int decode_significance_x86(CABACContext *c, int max_coeff,
" jb 3b \n\t" " jb 3b \n\t"
"mov %2, %0 \n\t" "mov %2, %0 \n\t"
"movl %7, %%ecx \n\t" "movl %7, %%ecx \n\t"
"add %1, %%"REG_c" \n\t" "add %1, %%"FF_REG_c" \n\t"
"movl %%ecx, (%0) \n\t" "movl %%ecx, (%0) \n\t"
"5: \n\t" "5: \n\t"
"add %9, %k0 \n\t" "add %9, %k0 \n\t"
...@@ -109,7 +109,7 @@ static int decode_significance_x86(CABACContext *c, int max_coeff, ...@@ -109,7 +109,7 @@ static int decode_significance_x86(CABACContext *c, int max_coeff,
"i"(offsetof(CABACContext, bytestream)), "i"(offsetof(CABACContext, bytestream)),
"i"(offsetof(CABACContext, bytestream_end)) "i"(offsetof(CABACContext, bytestream_end))
TABLES_ARG TABLES_ARG
: "%"REG_c, "memory" : "%"FF_REG_c, "memory"
); );
return coeff_count; return coeff_count;
} }
...@@ -175,7 +175,7 @@ static int decode_significance_8x8_x86(CABACContext *c, ...@@ -175,7 +175,7 @@ static int decode_significance_8x8_x86(CABACContext *c,
"test $1, %4 \n\t" "test $1, %4 \n\t"
" jnz 5f \n\t" " jnz 5f \n\t"
"add"OPSIZE" $4, %2 \n\t" "add"FF_OPSIZE" $4, %2 \n\t"
"4: \n\t" "4: \n\t"
"addl $1, %k6 \n\t" "addl $1, %k6 \n\t"
...@@ -194,7 +194,7 @@ static int decode_significance_8x8_x86(CABACContext *c, ...@@ -194,7 +194,7 @@ static int decode_significance_8x8_x86(CABACContext *c,
"i"(offsetof(CABACContext, bytestream)), "i"(offsetof(CABACContext, bytestream)),
"i"(offsetof(CABACContext, bytestream_end)), "i"(offsetof(CABACContext, bytestream_end)),
"i"(H264_LAST_COEFF_FLAG_OFFSET_8x8_OFFSET) TABLES_ARG "i"(H264_LAST_COEFF_FLAG_OFFSET_8x8_OFFSET) TABLES_ARG
: "%"REG_c, "memory" : "%"FF_REG_c, "memory"
); );
return coeff_count; return coeff_count;
} }
......
libavcodec/x86/hpeldsp_rnd_template.c
View file @ 1e9c5bf4
...@@ -32,7 +32,7 @@ static void DEF(put, pixels8_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff_ ...@@ -32,7 +32,7 @@ static void DEF(put, pixels8_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff_
{ {
MOVQ_BFE(mm6); MOVQ_BFE(mm6);
__asm__ volatile( __asm__ volatile(
"lea (%3, %3), %%"REG_a" \n\t" "lea (%3, %3), %%"FF_REG_a" \n\t"
".p2align 3 \n\t" ".p2align 3 \n\t"
"1: \n\t" "1: \n\t"
"movq (%1), %%mm0 \n\t" "movq (%1), %%mm0 \n\t"
...@@ -42,8 +42,8 @@ static void DEF(put, pixels8_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff_ ...@@ -42,8 +42,8 @@ static void DEF(put, pixels8_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff_
PAVGBP(%%mm0, %%mm1, %%mm4, %%mm2, %%mm3, %%mm5) PAVGBP(%%mm0, %%mm1, %%mm4, %%mm2, %%mm3, %%mm5)
"movq %%mm4, (%2) \n\t" "movq %%mm4, (%2) \n\t"
"movq %%mm5, (%2, %3) \n\t" "movq %%mm5, (%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"movq (%1), %%mm0 \n\t" "movq (%1), %%mm0 \n\t"
"movq 1(%1), %%mm1 \n\t" "movq 1(%1), %%mm1 \n\t"
"movq (%1, %3), %%mm2 \n\t" "movq (%1, %3), %%mm2 \n\t"
...@@ -51,20 +51,20 @@ static void DEF(put, pixels8_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff_ ...@@ -51,20 +51,20 @@ static void DEF(put, pixels8_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff_
PAVGBP(%%mm0, %%mm1, %%mm4, %%mm2, %%mm3, %%mm5) PAVGBP(%%mm0, %%mm1, %%mm4, %%mm2, %%mm3, %%mm5)
"movq %%mm4, (%2) \n\t" "movq %%mm4, (%2) \n\t"
"movq %%mm5, (%2, %3) \n\t" "movq %%mm5, (%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"subl $4, %0 \n\t" "subl $4, %0 \n\t"
"jnz 1b \n\t" "jnz 1b \n\t"
:"+g"(h), "+S"(pixels), "+D"(block) :"+g"(h), "+S"(pixels), "+D"(block)
:"r"((x86_reg)line_size) :"r"((x86_reg)line_size)
:REG_a, "memory"); :FF_REG_a, "memory");
} }
static void DEF(put, pixels16_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h) static void DEF(put, pixels16_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
{ {
MOVQ_BFE(mm6); MOVQ_BFE(mm6);
__asm__ volatile( __asm__ volatile(
"lea (%3, %3), %%"REG_a" \n\t" "lea (%3, %3), %%"FF_REG_a" \n\t"
".p2align 3 \n\t" ".p2align 3 \n\t"
"1: \n\t" "1: \n\t"
"movq (%1), %%mm0 \n\t" "movq (%1), %%mm0 \n\t"
...@@ -81,8 +81,8 @@ static void DEF(put, pixels16_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff ...@@ -81,8 +81,8 @@ static void DEF(put, pixels16_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff
PAVGBP(%%mm0, %%mm1, %%mm4, %%mm2, %%mm3, %%mm5) PAVGBP(%%mm0, %%mm1, %%mm4, %%mm2, %%mm3, %%mm5)
"movq %%mm4, 8(%2) \n\t" "movq %%mm4, 8(%2) \n\t"
"movq %%mm5, 8(%2, %3) \n\t" "movq %%mm5, 8(%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"movq (%1), %%mm0 \n\t" "movq (%1), %%mm0 \n\t"
"movq 1(%1), %%mm1 \n\t" "movq 1(%1), %%mm1 \n\t"
"movq (%1, %3), %%mm2 \n\t" "movq (%1, %3), %%mm2 \n\t"
...@@ -97,42 +97,42 @@ static void DEF(put, pixels16_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff ...@@ -97,42 +97,42 @@ static void DEF(put, pixels16_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff
PAVGBP(%%mm0, %%mm1, %%mm4, %%mm2, %%mm3, %%mm5) PAVGBP(%%mm0, %%mm1, %%mm4, %%mm2, %%mm3, %%mm5)
"movq %%mm4, 8(%2) \n\t" "movq %%mm4, 8(%2) \n\t"
"movq %%mm5, 8(%2, %3) \n\t" "movq %%mm5, 8(%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"subl $4, %0 \n\t" "subl $4, %0 \n\t"
"jnz 1b \n\t" "jnz 1b \n\t"
:"+g"(h), "+S"(pixels), "+D"(block) :"+g"(h), "+S"(pixels), "+D"(block)
:"r"((x86_reg)line_size) :"r"((x86_reg)line_size)
:REG_a, "memory"); :FF_REG_a, "memory");
} }
static void DEF(put, pixels8_y2)(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h) static void DEF(put, pixels8_y2)(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
{ {
MOVQ_BFE(mm6); MOVQ_BFE(mm6);
__asm__ volatile( __asm__ volatile(
"lea (%3, %3), %%"REG_a" \n\t" "lea (%3, %3), %%"FF_REG_a" \n\t"
"movq (%1), %%mm0 \n\t" "movq (%1), %%mm0 \n\t"
".p2align 3 \n\t" ".p2align 3 \n\t"
"1: \n\t" "1: \n\t"
"movq (%1, %3), %%mm1 \n\t" "movq (%1, %3), %%mm1 \n\t"
"movq (%1, %%"REG_a"),%%mm2 \n\t" "movq (%1, %%"FF_REG_a"),%%mm2 \n\t"
PAVGBP(%%mm1, %%mm0, %%mm4, %%mm2, %%mm1, %%mm5) PAVGBP(%%mm1, %%mm0, %%mm4, %%mm2, %%mm1, %%mm5)
"movq %%mm4, (%2) \n\t" "movq %%mm4, (%2) \n\t"
"movq %%mm5, (%2, %3) \n\t" "movq %%mm5, (%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"movq (%1, %3), %%mm1 \n\t" "movq (%1, %3), %%mm1 \n\t"
"movq (%1, %%"REG_a"),%%mm0 \n\t" "movq (%1, %%"FF_REG_a"),%%mm0 \n\t"
PAVGBP(%%mm1, %%mm2, %%mm4, %%mm0, %%mm1, %%mm5) PAVGBP(%%mm1, %%mm2, %%mm4, %%mm0, %%mm1, %%mm5)
"movq %%mm4, (%2) \n\t" "movq %%mm4, (%2) \n\t"
"movq %%mm5, (%2, %3) \n\t" "movq %%mm5, (%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"subl $4, %0 \n\t" "subl $4, %0 \n\t"
"jnz 1b \n\t" "jnz 1b \n\t"
:"+g"(h), "+S"(pixels), "+D"(block) :"+g"(h), "+S"(pixels), "+D"(block)
:"r"((x86_reg)line_size) :"r"((x86_reg)line_size)
:REG_a, "memory"); :FF_REG_a, "memory");
} }
static void DEF(avg, pixels16_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h) static void DEF(avg, pixels16_x2)(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
...@@ -165,12 +165,12 @@ static void DEF(avg, pixels8_y2)(uint8_t *block, const uint8_t *pixels, ptrdiff_ ...@@ -165,12 +165,12 @@ static void DEF(avg, pixels8_y2)(uint8_t *block, const uint8_t *pixels, ptrdiff_
{ {
MOVQ_BFE(mm6); MOVQ_BFE(mm6);
__asm__ volatile( __asm__ volatile(
"lea (%3, %3), %%"REG_a" \n\t" "lea (%3, %3), %%"FF_REG_a" \n\t"
"movq (%1), %%mm0 \n\t" "movq (%1), %%mm0 \n\t"
".p2align 3 \n\t" ".p2align 3 \n\t"
"1: \n\t" "1: \n\t"
"movq (%1, %3), %%mm1 \n\t" "movq (%1, %3), %%mm1 \n\t"
"movq (%1, %%"REG_a"), %%mm2 \n\t" "movq (%1, %%"FF_REG_a"), %%mm2 \n\t"
PAVGBP(%%mm1, %%mm0, %%mm4, %%mm2, %%mm1, %%mm5) PAVGBP(%%mm1, %%mm0, %%mm4, %%mm2, %%mm1, %%mm5)
"movq (%2), %%mm3 \n\t" "movq (%2), %%mm3 \n\t"
PAVGB_MMX(%%mm3, %%mm4, %%mm0, %%mm6) PAVGB_MMX(%%mm3, %%mm4, %%mm0, %%mm6)
...@@ -178,11 +178,11 @@ static void DEF(avg, pixels8_y2)(uint8_t *block, const uint8_t *pixels, ptrdiff_ ...@@ -178,11 +178,11 @@ static void DEF(avg, pixels8_y2)(uint8_t *block, const uint8_t *pixels, ptrdiff_
PAVGB_MMX(%%mm3, %%mm5, %%mm1, %%mm6) PAVGB_MMX(%%mm3, %%mm5, %%mm1, %%mm6)
"movq %%mm0, (%2) \n\t" "movq %%mm0, (%2) \n\t"
"movq %%mm1, (%2, %3) \n\t" "movq %%mm1, (%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"movq (%1, %3), %%mm1 \n\t" "movq (%1, %3), %%mm1 \n\t"
"movq (%1, %%"REG_a"), %%mm0 \n\t" "movq (%1, %%"FF_REG_a"), %%mm0 \n\t"
PAVGBP(%%mm1, %%mm2, %%mm4, %%mm0, %%mm1, %%mm5) PAVGBP(%%mm1, %%mm2, %%mm4, %%mm0, %%mm1, %%mm5)
"movq (%2), %%mm3 \n\t" "movq (%2), %%mm3 \n\t"
PAVGB_MMX(%%mm3, %%mm4, %%mm2, %%mm6) PAVGB_MMX(%%mm3, %%mm4, %%mm2, %%mm6)
...@@ -190,12 +190,12 @@ static void DEF(avg, pixels8_y2)(uint8_t *block, const uint8_t *pixels, ptrdiff_ ...@@ -190,12 +190,12 @@ static void DEF(avg, pixels8_y2)(uint8_t *block, const uint8_t *pixels, ptrdiff_
PAVGB_MMX(%%mm3, %%mm5, %%mm1, %%mm6) PAVGB_MMX(%%mm3, %%mm5, %%mm1, %%mm6)
"movq %%mm2, (%2) \n\t" "movq %%mm2, (%2) \n\t"
"movq %%mm1, (%2, %3) \n\t" "movq %%mm1, (%2, %3) \n\t"
"add %%"REG_a", %1 \n\t" "add %%"FF_REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t" "add %%"FF_REG_a", %2 \n\t"
"subl $4, %0 \n\t" "subl $4, %0 \n\t"
"jnz 1b \n\t" "jnz 1b \n\t"
:"+g"(h), "+S"(pixels), "+D"(block) :"+g"(h), "+S"(pixels), "+D"(block)
:"r"((x86_reg)line_size) :"r"((x86_reg)line_size)
:REG_a, "memory"); :FF_REG_a, "memory");
} }
libavcodec/x86/me_cmp_init.c
View file @ 1e9c5bf4
...@@ -812,15 +812,15 @@ static inline void sad8_1_mmx(uint8_t *blk1, uint8_t *blk2, ...@@ -812,15 +812,15 @@ static inline void sad8_1_mmx(uint8_t *blk1, uint8_t *blk2,
__asm__ volatile ( __asm__ volatile (
".p2align 4 \n\t" ".p2align 4 \n\t"
"1: \n\t" "1: \n\t"
"movq (%1, %%"REG_a"), %%mm0 \n\t" "movq (%1, %%"FF_REG_a"), %%mm0 \n\t"
"movq (%2, %%"REG_a"), %%mm2 \n\t" "movq (%2, %%"FF_REG_a"), %%mm2 \n\t"
"movq (%2, %%"REG_a"), %%mm4 \n\t" "movq (%2, %%"FF_REG_a"), %%mm4 \n\t"
"add %3, %%"REG_a" \n\t" "add %3, %%"FF_REG_a" \n\t"
"psubusb %%mm0, %%mm2 \n\t" "psubusb %%mm0, %%mm2 \n\t"
"psubusb %%mm4, %%mm0 \n\t" "psubusb %%mm4, %%mm0 \n\t"
"movq (%1, %%"REG_a"), %%mm1 \n\t" "movq (%1, %%"FF_REG_a"), %%mm1 \n\t"
"movq (%2, %%"REG_a"), %%mm3 \n\t" "movq (%2, %%"FF_REG_a"), %%mm3 \n\t"
"movq (%2, %%"REG_a"), %%mm5 \n\t" "movq (%2, %%"FF_REG_a"), %%mm5 \n\t"
"psubusb %%mm1, %%mm3 \n\t" "psubusb %%mm1, %%mm3 \n\t"
"psubusb %%mm5, %%mm1 \n\t" "psubusb %%mm5, %%mm1 \n\t"
"por %%mm2, %%mm0 \n\t" "por %%mm2, %%mm0 \n\t"
...@@ -835,7 +835,7 @@ static inline void sad8_1_mmx(uint8_t *blk1, uint8_t *blk2, ...@@ -835,7 +835,7 @@ static inline void sad8_1_mmx(uint8_t *blk1, uint8_t *blk2,
"paddw %%mm3, %%mm2 \n\t" "paddw %%mm3, %%mm2 \n\t"
"paddw %%mm2, %%mm0 \n\t" "paddw %%mm2, %%mm0 \n\t"
"paddw %%mm0, %%mm6 \n\t" "paddw %%mm0, %%mm6 \n\t"
"add %3, %%"REG_a" \n\t" "add %3, %%"FF_REG_a" \n\t"
" js 1b \n\t" " js 1b \n\t"
: "+a" (len) : "+a" (len)
: "r" (blk1 - len), "r" (blk2 - len), "r" (stride)); : "r" (blk1 - len), "r" (blk2 - len), "r" (stride));
...@@ -971,18 +971,18 @@ static inline void sad8_2_mmx(uint8_t *blk1a, uint8_t *blk1b, uint8_t *blk2, ...@@ -971,18 +971,18 @@ static inline void sad8_2_mmx(uint8_t *blk1a, uint8_t *blk1b, uint8_t *blk2,
__asm__ volatile ( __asm__ volatile (
".p2align 4 \n\t" ".p2align 4 \n\t"
"1: \n\t" "1: \n\t"
"movq (%1, %%"REG_a"), %%mm0 \n\t" "movq (%1, %%"FF_REG_a"), %%mm0 \n\t"
"movq (%2, %%"REG_a"), %%mm1 \n\t" "movq (%2, %%"FF_REG_a"), %%mm1 \n\t"
"movq (%1, %%"REG_a"), %%mm2 \n\t" "movq (%1, %%"FF_REG_a"), %%mm2 \n\t"
"movq (%2, %%"REG_a"), %%mm3 \n\t" "movq (%2, %%"FF_REG_a"), %%mm3 \n\t"
"punpcklbw %%mm7, %%mm0 \n\t" "punpcklbw %%mm7, %%mm0 \n\t"
"punpcklbw %%mm7, %%mm1 \n\t" "punpcklbw %%mm7, %%mm1 \n\t"
"punpckhbw %%mm7, %%mm2 \n\t" "punpckhbw %%mm7, %%mm2 \n\t"
"punpckhbw %%mm7, %%mm3 \n\t" "punpckhbw %%mm7, %%mm3 \n\t"
"paddw %%mm0, %%mm1 \n\t" "paddw %%mm0, %%mm1 \n\t"
"paddw %%mm2, %%mm3 \n\t" "paddw %%mm2, %%mm3 \n\t"
"movq (%3, %%"REG_a"), %%mm4 \n\t" "movq (%3, %%"FF_REG_a"), %%mm4 \n\t"
"movq (%3, %%"REG_a"), %%mm2 \n\t" "movq (%3, %%"FF_REG_a"), %%mm2 \n\t"
"paddw %%mm5, %%mm1 \n\t" "paddw %%mm5, %%mm1 \n\t"
"paddw %%mm5, %%mm3 \n\t" "paddw %%mm5, %%mm3 \n\t"
"psrlw $1, %%mm1 \n\t" "psrlw $1, %%mm1 \n\t"
...@@ -996,7 +996,7 @@ static inline void sad8_2_mmx(uint8_t *blk1a, uint8_t *blk1b, uint8_t *blk2, ...@@ -996,7 +996,7 @@ static inline void sad8_2_mmx(uint8_t *blk1a, uint8_t *blk1b, uint8_t *blk2,
"punpckhbw %%mm7, %%mm1 \n\t" "punpckhbw %%mm7, %%mm1 \n\t"
"paddw %%mm1, %%mm0 \n\t" "paddw %%mm1, %%mm0 \n\t"
"paddw %%mm0, %%mm6 \n\t" "paddw %%mm0, %%mm6 \n\t"
"add %4, %%"REG_a" \n\t" "add %4, %%"FF_REG_a" \n\t"
" js 1b \n\t" " js 1b \n\t"
: "+a" (len) : "+a" (len)
: "r" (blk1a - len), "r" (blk1b - len), "r" (blk2 - len), : "r" (blk1a - len), "r" (blk1b - len), "r" (blk2 - len),
...@@ -1008,8 +1008,8 @@ static inline void sad8_4_mmx(uint8_t *blk1, uint8_t *blk2, ...@@ -1008,8 +1008,8 @@ static inline void sad8_4_mmx(uint8_t *blk1, uint8_t *blk2,
{ {
x86_reg len = -(stride * h); x86_reg len = -(stride * h);
__asm__ volatile ( __asm__ volatile (
"movq (%1, %%"REG_a"), %%mm0 \n\t" "movq (%1, %%"FF_REG_a"), %%mm0\n\t"
"movq 1(%1, %%"REG_a"), %%mm2 \n\t" "movq 1(%1, %%"FF_REG_a"), %%mm2\n\t"
"movq %%mm0, %%mm1 \n\t" "movq %%mm0, %%mm1 \n\t"
"movq %%mm2, %%mm3 \n\t" "movq %%mm2, %%mm3 \n\t"
"punpcklbw %%mm7, %%mm0 \n\t" "punpcklbw %%mm7, %%mm0 \n\t"
...@@ -1020,8 +1020,8 @@ static inline void sad8_4_mmx(uint8_t *blk1, uint8_t *blk2, ...@@ -1020,8 +1020,8 @@ static inline void sad8_4_mmx(uint8_t *blk1, uint8_t *blk2,
"paddw %%mm3, %%mm1 \n\t" "paddw %%mm3, %%mm1 \n\t"
".p2align 4 \n\t" ".p2align 4 \n\t"
"1: \n\t" "1: \n\t"
"movq (%2, %%"REG_a"), %%mm2 \n\t" "movq (%2, %%"FF_REG_a"), %%mm2\n\t"
"movq 1(%2, %%"REG_a"), %%mm4 \n\t" "movq 1(%2, %%"FF_REG_a"), %%mm4\n\t"
"movq %%mm2, %%mm3 \n\t" "movq %%mm2, %%mm3 \n\t"
"movq %%mm4, %%mm5 \n\t" "movq %%mm4, %%mm5 \n\t"
"punpcklbw %%mm7, %%mm2 \n\t" "punpcklbw %%mm7, %%mm2 \n\t"
...@@ -1035,8 +1035,8 @@ static inline void sad8_4_mmx(uint8_t *blk1, uint8_t *blk2, ...@@ -1035,8 +1035,8 @@ static inline void sad8_4_mmx(uint8_t *blk1, uint8_t *blk2,
"paddw %%mm3, %%mm1 \n\t" "paddw %%mm3, %%mm1 \n\t"
"paddw %%mm5, %%mm0 \n\t" "paddw %%mm5, %%mm0 \n\t"
"paddw %%mm5, %%mm1 \n\t" "paddw %%mm5, %%mm1 \n\t"
"movq (%3, %%"REG_a"), %%mm4 \n\t" "movq (%3, %%"FF_REG_a"), %%mm4 \n\t"
"movq (%3, %%"REG_a"), %%mm5 \n\t" "movq (%3, %%"FF_REG_a"), %%mm5 \n\t"
"psrlw $2, %%mm0 \n\t" "psrlw $2, %%mm0 \n\t"
"psrlw $2, %%mm1 \n\t" "psrlw $2, %%mm1 \n\t"
"packuswb %%mm1, %%mm0 \n\t" "packuswb %%mm1, %%mm0 \n\t"
...@@ -1050,7 +1050,7 @@ static inline void sad8_4_mmx(uint8_t *blk1, uint8_t *blk2, ...@@ -1050,7 +1050,7 @@ static inline void sad8_4_mmx(uint8_t *blk1, uint8_t *blk2,
"paddw %%mm4, %%mm6 \n\t" "paddw %%mm4, %%mm6 \n\t"
"movq %%mm2, %%mm0 \n\t" "movq %%mm2, %%mm0 \n\t"
"movq %%mm3, %%mm1 \n\t" "movq %%mm3, %%mm1 \n\t"
"add %4, %%"REG_a" \n\t" "add %4, %%"FF_REG_a" \n\t"
" js 1b \n\t" " js 1b \n\t"
: "+a" (len) : "+a" (len)
: "r" (blk1 - len), "r" (blk1 - len + stride), "r" (blk2 - len), : "r" (blk1 - len), "r" (blk1 - len + stride), "r" (blk2 - len),
......
libavcodec/x86/mpegvideo.c
View file @ 1e9c5bf4
This diff is collapsed.
libavcodec/x86/mpegvideoenc_template.c
View file @ 1e9c5bf4
...@@ -147,33 +147,33 @@ static int RENAME(dct_quantize)(MpegEncContext *s, ...@@ -147,33 +147,33 @@ static int RENAME(dct_quantize)(MpegEncContext *s,
if((s->out_format == FMT_H263 || s->out_format == FMT_H261) && s->mpeg_quant==0){ if((s->out_format == FMT_H263 || s->out_format == FMT_H261) && s->mpeg_quant==0){
__asm__ volatile( __asm__ volatile(
"movd %%"REG_a", "MM"3 \n\t" // last_non_zero_p1 "movd %%"FF_REG_a", "MM"3 \n\t" // last_non_zero_p1
SPREADW(MM"3") SPREADW(MM"3")
"pxor "MM"7, "MM"7 \n\t" // 0 "pxor "MM"7, "MM"7 \n\t" // 0
"pxor "MM"4, "MM"4 \n\t" // 0 "pxor "MM"4, "MM"4 \n\t" // 0
MOVQ" (%2), "MM"5 \n\t" // qmat[0] MOVQ" (%2), "MM"5 \n\t" // qmat[0]
"pxor "MM"6, "MM"6 \n\t" "pxor "MM"6, "MM"6 \n\t"
"psubw (%3), "MM"6 \n\t" // -bias[0] "psubw (%3), "MM"6 \n\t" // -bias[0]
"mov $-128, %%"REG_a" \n\t" "mov $-128, %%"FF_REG_a" \n\t"
".p2align 4 \n\t" ".p2align 4 \n\t"
"1: \n\t" "1: \n\t"
MOVQ" (%1, %%"REG_a"), "MM"0 \n\t" // block[i] MOVQ" (%1, %%"FF_REG_a"), "MM"0 \n\t" // block[i]
SAVE_SIGN(MM"1", MM"0") // ABS(block[i]) SAVE_SIGN(MM"1", MM"0") // ABS(block[i])
"psubusw "MM"6, "MM"0 \n\t" // ABS(block[i]) + bias[0] "psubusw "MM"6, "MM"0 \n\t" // ABS(block[i]) + bias[0]
"pmulhw "MM"5, "MM"0 \n\t" // (ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16 "pmulhw "MM"5, "MM"0 \n\t" // (ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16
"por "MM"0, "MM"4 \n\t" "por "MM"0, "MM"4 \n\t"
RESTORE_SIGN(MM"1", MM"0") // out=((ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16)*sign(block[i]) RESTORE_SIGN(MM"1", MM"0") // out=((ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16)*sign(block[i])
MOVQ" "MM"0, (%5, %%"REG_a") \n\t" MOVQ" "MM"0, (%5, %%"FF_REG_a") \n\t"
"pcmpeqw "MM"7, "MM"0 \n\t" // out==0 ? 0xFF : 0x00 "pcmpeqw "MM"7, "MM"0 \n\t" // out==0 ? 0xFF : 0x00
MOVQ" (%4, %%"REG_a"), "MM"1 \n\t" MOVQ" (%4, %%"FF_REG_a"), "MM"1 \n\t"
MOVQ" "MM"7, (%1, %%"REG_a") \n\t" // 0 MOVQ" "MM"7, (%1, %%"FF_REG_a") \n\t" // 0
"pandn "MM"1, "MM"0 \n\t" "pandn "MM"1, "MM"0 \n\t"
PMAXW(MM"0", MM"3") PMAXW(MM"0", MM"3")
"add $"MMREG_WIDTH", %%"REG_a" \n\t" "add $"MMREG_WIDTH", %%"FF_REG_a" \n\t"
" js 1b \n\t" " js 1b \n\t"
PMAX(MM"3", MM"0") PMAX(MM"3", MM"0")
"movd "MM"3, %%"REG_a" \n\t" "movd "MM"3, %%"FF_REG_a" \n\t"
"movzb %%al, %%"REG_a" \n\t" // last_non_zero_p1 "movzb %%al, %%"FF_REG_a" \n\t" // last_non_zero_p1
: "+a" (last_non_zero_p1) : "+a" (last_non_zero_p1)
: "r" (block+64), "r" (qmat), "r" (bias), : "r" (block+64), "r" (qmat), "r" (bias),
"r" (inv_zigzag_direct16 + 64), "r" (temp_block + 64) "r" (inv_zigzag_direct16 + 64), "r" (temp_block + 64)
...@@ -182,32 +182,32 @@ static int RENAME(dct_quantize)(MpegEncContext *s, ...@@ -182,32 +182,32 @@ static int RENAME(dct_quantize)(MpegEncContext *s,
); );
}else{ // FMT_H263 }else{ // FMT_H263
__asm__ volatile( __asm__ volatile(
"movd %%"REG_a", "MM"3 \n\t" // last_non_zero_p1 "movd %%"FF_REG_a", "MM"3 \n\t" // last_non_zero_p1
SPREADW(MM"3") SPREADW(MM"3")
"pxor "MM"7, "MM"7 \n\t" // 0 "pxor "MM"7, "MM"7 \n\t" // 0
"pxor "MM"4, "MM"4 \n\t" // 0 "pxor "MM"4, "MM"4 \n\t" // 0
"mov $-128, %%"REG_a" \n\t" "mov $-128, %%"FF_REG_a" \n\t"
".p2align 4 \n\t" ".p2align 4 \n\t"
"1: \n\t" "1: \n\t"
MOVQ" (%1, %%"REG_a"), "MM"0 \n\t" // block[i] MOVQ" (%1, %%"FF_REG_a"), "MM"0 \n\t" // block[i]
SAVE_SIGN(MM"1", MM"0") // ABS(block[i]) SAVE_SIGN(MM"1", MM"0") // ABS(block[i])
MOVQ" (%3, %%"REG_a"), "MM"6 \n\t" // bias[0] MOVQ" (%3, %%"FF_REG_a"), "MM"6 \n\t" // bias[0]
"paddusw "MM"6, "MM"0 \n\t" // ABS(block[i]) + bias[0] "paddusw "MM"6, "MM"0 \n\t" // ABS(block[i]) + bias[0]
MOVQ" (%2, %%"REG_a"), "MM"5 \n\t" // qmat[i] MOVQ" (%2, %%"FF_REG_a"), "MM"5 \n\t" // qmat[i]
"pmulhw "MM"5, "MM"0 \n\t" // (ABS(block[i])*qmat[0] + bias[0]*qmat[0])>>16 "pmulhw "MM"5, "MM"0 \n\t" // (ABS(block[i])*qmat[0] + bias[0]*qmat[0])>>16
"por "MM"0, "MM"4 \n\t" "por "MM"0, "MM"4 \n\t"
RESTORE_SIGN(MM"1", MM"0") // out=((ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16)*sign(block[i]) RESTORE_SIGN(MM"1", MM"0") // out=((ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16)*sign(block[i])
MOVQ" "MM"0, (%5, %%"REG_a") \n\t" MOVQ" "MM"0, (%5, %%"FF_REG_a") \n\t"
"pcmpeqw "MM"7, "MM"0 \n\t" // out==0 ? 0xFF : 0x00 "pcmpeqw "MM"7, "MM"0 \n\t" // out==0 ? 0xFF : 0x00
MOVQ" (%4, %%"REG_a"), "MM"1 \n\t" MOVQ" (%4, %%"FF_REG_a"), "MM"1 \n\t"
MOVQ" "MM"7, (%1, %%"REG_a") \n\t" // 0 MOVQ" "MM"7, (%1, %%"FF_REG_a") \n\t" // 0
"pandn "MM"1, "MM"0 \n\t" "pandn "MM"1, "MM"0 \n\t"
PMAXW(MM"0", MM"3") PMAXW(MM"0", MM"3")
"add $"MMREG_WIDTH", %%"REG_a" \n\t" "add $"MMREG_WIDTH", %%"FF_REG_a" \n\t"
" js 1b \n\t" " js 1b \n\t"
PMAX(MM"3", MM"0") PMAX(MM"3", MM"0")
"movd "MM"3, %%"REG_a" \n\t" "movd "MM"3, %%"FF_REG_a" \n\t"
"movzb %%al, %%"REG_a" \n\t" // last_non_zero_p1 "movzb %%al, %%"FF_REG_a" \n\t" // last_non_zero_p1
: "+a" (last_non_zero_p1) : "+a" (last_non_zero_p1)
: "r" (block+64), "r" (qmat+64), "r" (bias+64), : "r" (block+64), "r" (qmat+64), "r" (bias+64),
"r" (inv_zigzag_direct16 + 64), "r" (temp_block + 64) "r" (inv_zigzag_direct16 + 64), "r" (temp_block + 64)
......
libavcodec/x86/rnd_template.c
View file @ 1e9c5bf4
...@@ -46,12 +46,12 @@ STATIC void DEF(put, pixels8_xy2)(uint8_t *block, const uint8_t *pixels, ...@@ -46,12 +46,12 @@ STATIC void DEF(put, pixels8_xy2)(uint8_t *block, const uint8_t *pixels,
"punpckhbw %%mm7, %%mm5 \n\t" "punpckhbw %%mm7, %%mm5 \n\t"
"paddusw %%mm0, %%mm4 \n\t" "paddusw %%mm0, %%mm4 \n\t"
"paddusw %%mm1, %%mm5 \n\t" "paddusw %%mm1, %%mm5 \n\t"
"xor %%"REG_a", %%"REG_a" \n\t" "xor %%"FF_REG_a", %%"FF_REG_a" \n\t"
"add %3, %1 \n\t" "add %3, %1 \n\t"
".p2align 3 \n\t" ".p2align 3 \n\t"
"1: \n\t" "1: \n\t"
"movq (%1, %%"REG_a"), %%mm0 \n\t" "movq (%1, %%"FF_REG_a"), %%mm0\n\t"
"movq 1(%1, %%"REG_a"), %%mm2 \n\t" "movq 1(%1, %%"FF_REG_a"), %%mm2\n\t"
"movq %%mm0, %%mm1 \n\t" "movq %%mm0, %%mm1 \n\t"
"movq %%mm2, %%mm3 \n\t" "movq %%mm2, %%mm3 \n\t"
"punpcklbw %%mm7, %%mm0 \n\t" "punpcklbw %%mm7, %%mm0 \n\t"
...@@ -67,11 +67,11 @@ STATIC void DEF(put, pixels8_xy2)(uint8_t *block, const uint8_t *pixels, ...@@ -67,11 +67,11 @@ STATIC void DEF(put, pixels8_xy2)(uint8_t *block, const uint8_t *pixels,
"psrlw $2, %%mm4 \n\t" "psrlw $2, %%mm4 \n\t"
"psrlw $2, %%mm5 \n\t" "psrlw $2, %%mm5 \n\t"
"packuswb %%mm5, %%mm4 \n\t" "packuswb %%mm5, %%mm4 \n\t"
"movq %%mm4, (%2, %%"REG_a") \n\t" "movq %%mm4, (%2, %%"FF_REG_a") \n\t"
"add %3, %%"REG_a" \n\t" "add %3, %%"FF_REG_a" \n\t"
"movq (%1, %%"REG_a"), %%mm2 \n\t" // 0 <-> 2 1 <-> 3 "movq (%1, %%"FF_REG_a"), %%mm2\n\t" // 0 <-> 2 1 <-> 3
"movq 1(%1, %%"REG_a"), %%mm4 \n\t" "movq 1(%1, %%"FF_REG_a"), %%mm4\n\t"
"movq %%mm2, %%mm3 \n\t" "movq %%mm2, %%mm3 \n\t"
"movq %%mm4, %%mm5 \n\t" "movq %%mm4, %%mm5 \n\t"
"punpcklbw %%mm7, %%mm2 \n\t" "punpcklbw %%mm7, %%mm2 \n\t"
...@@ -87,14 +87,14 @@ STATIC void DEF(put, pixels8_xy2)(uint8_t *block, const uint8_t *pixels, ...@@ -87,14 +87,14 @@ STATIC void DEF(put, pixels8_xy2)(uint8_t *block, const uint8_t *pixels,
"psrlw $2, %%mm0 \n\t" "psrlw $2, %%mm0 \n\t"
"psrlw $2, %%mm1 \n\t" "psrlw $2, %%mm1 \n\t"
"packuswb %%mm1, %%mm0 \n\t" "packuswb %%mm1, %%mm0 \n\t"
"movq %%mm0, (%2, %%"REG_a") \n\t" "movq %%mm0, (%2, %%"FF_REG_a") \n\t"
"add %3, %%"REG_a" \n\t" "add %3, %%"FF_REG_a" \n\t"
"subl $2, %0 \n\t" "subl $2, %0 \n\t"
"jnz 1b \n\t" "jnz 1b \n\t"
:"+g"(h), "+S"(pixels) :"+g"(h), "+S"(pixels)
:"D"(block), "r"((x86_reg)line_size) :"D"(block), "r"((x86_reg)line_size)
:REG_a, "memory"); :FF_REG_a, "memory");
} }
// avg_pixels // avg_pixels
...@@ -115,12 +115,12 @@ STATIC void DEF(avg, pixels8_xy2)(uint8_t *block, const uint8_t *pixels, ...@@ -115,12 +115,12 @@ STATIC void DEF(avg, pixels8_xy2)(uint8_t *block, const uint8_t *pixels,
"punpckhbw %%mm7, %%mm5 \n\t" "punpckhbw %%mm7, %%mm5 \n\t"
"paddusw %%mm0, %%mm4 \n\t" "paddusw %%mm0, %%mm4 \n\t"
"paddusw %%mm1, %%mm5 \n\t" "paddusw %%mm1, %%mm5 \n\t"
"xor %%"REG_a", %%"REG_a" \n\t" "xor %%"FF_REG_a", %%"FF_REG_a" \n\t"
"add %3, %1 \n\t" "add %3, %1 \n\t"
".p2align 3 \n\t" ".p2align 3 \n\t"
"1: \n\t" "1: \n\t"
"movq (%1, %%"REG_a"), %%mm0 \n\t" "movq (%1, %%"FF_REG_a"), %%mm0\n\t"
"movq 1(%1, %%"REG_a"), %%mm2 \n\t" "movq 1(%1, %%"FF_REG_a"), %%mm2\n\t"
"movq %%mm0, %%mm1 \n\t" "movq %%mm0, %%mm1 \n\t"
"movq %%mm2, %%mm3 \n\t" "movq %%mm2, %%mm3 \n\t"
"punpcklbw %%mm7, %%mm0 \n\t" "punpcklbw %%mm7, %%mm0 \n\t"
...@@ -135,16 +135,16 @@ STATIC void DEF(avg, pixels8_xy2)(uint8_t *block, const uint8_t *pixels, ...@@ -135,16 +135,16 @@ STATIC void DEF(avg, pixels8_xy2)(uint8_t *block, const uint8_t *pixels,
"paddusw %%mm1, %%mm5 \n\t" "paddusw %%mm1, %%mm5 \n\t"
"psrlw $2, %%mm4 \n\t" "psrlw $2, %%mm4 \n\t"
"psrlw $2, %%mm5 \n\t" "psrlw $2, %%mm5 \n\t"
"movq (%2, %%"REG_a"), %%mm3 \n\t" "movq (%2, %%"FF_REG_a"), %%mm3 \n\t"
"packuswb %%mm5, %%mm4 \n\t" "packuswb %%mm5, %%mm4 \n\t"
"pcmpeqd %%mm2, %%mm2 \n\t" "pcmpeqd %%mm2, %%mm2 \n\t"
"paddb %%mm2, %%mm2 \n\t" "paddb %%mm2, %%mm2 \n\t"
PAVGB_MMX(%%mm3, %%mm4, %%mm5, %%mm2) PAVGB_MMX(%%mm3, %%mm4, %%mm5, %%mm2)
"movq %%mm5, (%2, %%"REG_a") \n\t" "movq %%mm5, (%2, %%"FF_REG_a") \n\t"
"add %3, %%"REG_a" \n\t" "add %3, %%"FF_REG_a" \n\t"
"movq (%1, %%"REG_a"), %%mm2 \n\t" // 0 <-> 2 1 <-> 3 "movq (%1, %%"FF_REG_a"), %%mm2\n\t" // 0 <-> 2 1 <-> 3
"movq 1(%1, %%"REG_a"), %%mm4 \n\t" "movq 1(%1, %%"FF_REG_a"), %%mm4\n\t"
"movq %%mm2, %%mm3 \n\t" "movq %%mm2, %%mm3 \n\t"
"movq %%mm4, %%mm5 \n\t" "movq %%mm4, %%mm5 \n\t"
"punpcklbw %%mm7, %%mm2 \n\t" "punpcklbw %%mm7, %%mm2 \n\t"
...@@ -159,17 +159,17 @@ STATIC void DEF(avg, pixels8_xy2)(uint8_t *block, const uint8_t *pixels, ...@@ -159,17 +159,17 @@ STATIC void DEF(avg, pixels8_xy2)(uint8_t *block, const uint8_t *pixels,
"paddusw %%mm5, %%mm1 \n\t" "paddusw %%mm5, %%mm1 \n\t"
"psrlw $2, %%mm0 \n\t" "psrlw $2, %%mm0 \n\t"
"psrlw $2, %%mm1 \n\t" "psrlw $2, %%mm1 \n\t"
"movq (%2, %%"REG_a"), %%mm3 \n\t" "movq (%2, %%"FF_REG_a"), %%mm3 \n\t"
"packuswb %%mm1, %%mm0 \n\t" "packuswb %%mm1, %%mm0 \n\t"
"pcmpeqd %%mm2, %%mm2 \n\t" "pcmpeqd %%mm2, %%mm2 \n\t"
"paddb %%mm2, %%mm2 \n\t" "paddb %%mm2, %%mm2 \n\t"
PAVGB_MMX(%%mm3, %%mm0, %%mm1, %%mm2) PAVGB_MMX(%%mm3, %%mm0, %%mm1, %%mm2)
"movq %%mm1, (%2, %%"REG_a") \n\t" "movq %%mm1, (%2, %%"FF_REG_a") \n\t"
"add %3, %%"REG_a" \n\t" "add %3, %%"FF_REG_a" \n\t"
"subl $2, %0 \n\t" "subl $2, %0 \n\t"
"jnz 1b \n\t" "jnz 1b \n\t"
:"+g"(h), "+S"(pixels) :"+g"(h), "+S"(pixels)
:"D"(block), "r"((x86_reg)line_size) :"D"(block), "r"((x86_reg)line_size)
:REG_a, "memory"); :FF_REG_a, "memory");
} }
libavcodec/x86/vc1dsp_mmx.c
View file @ 1e9c5bf4
...@@ -87,7 +87,7 @@ static void vc1_put_ver_16b_shift2_mmx(int16_t *dst, ...@@ -87,7 +87,7 @@ static void vc1_put_ver_16b_shift2_mmx(int16_t *dst,
int rnd, int64_t shift) int rnd, int64_t shift)
{ {
__asm__ volatile( __asm__ volatile(
"mov $3, %%"REG_c" \n\t" "mov $3, %%"FF_REG_c" \n\t"
LOAD_ROUNDER_MMX("%5") LOAD_ROUNDER_MMX("%5")
"movq "MANGLE(ff_pw_9)", %%mm6 \n\t" "movq "MANGLE(ff_pw_9)", %%mm6 \n\t"
"1: \n\t" "1: \n\t"
...@@ -106,12 +106,12 @@ static void vc1_put_ver_16b_shift2_mmx(int16_t *dst, ...@@ -106,12 +106,12 @@ static void vc1_put_ver_16b_shift2_mmx(int16_t *dst,
SHIFT2_LINE(168, 4, 1, 2, 3) SHIFT2_LINE(168, 4, 1, 2, 3)
"sub %6, %0 \n\t" "sub %6, %0 \n\t"
"add $8, %1 \n\t" "add $8, %1 \n\t"
"dec %%"REG_c" \n\t" "dec %%"FF_REG_c" \n\t"
"jnz 1b \n\t" "jnz 1b \n\t"
: "+r"(src), "+r"(dst) : "+r"(src), "+r"(dst)
: "r"(stride), "r"(-2*stride), : "r"(stride), "r"(-2*stride),
"m"(shift), "m"(rnd), "r"(9*stride-4) "m"(shift), "m"(rnd), "r"(9*stride-4)
: "%"REG_c, "memory" : "%"FF_REG_c, "memory"
); );
} }
...@@ -173,7 +173,7 @@ static void OPNAME ## vc1_shift2_mmx(uint8_t *dst, const uint8_t *src,\ ...@@ -173,7 +173,7 @@ static void OPNAME ## vc1_shift2_mmx(uint8_t *dst, const uint8_t *src,\
{\ {\
rnd = 8-rnd;\ rnd = 8-rnd;\
__asm__ volatile(\ __asm__ volatile(\
"mov $8, %%"REG_c" \n\t"\ "mov $8, %%"FF_REG_c" \n\t"\
LOAD_ROUNDER_MMX("%5")\ LOAD_ROUNDER_MMX("%5")\
"movq "MANGLE(ff_pw_9)", %%mm6\n\t"\ "movq "MANGLE(ff_pw_9)", %%mm6\n\t"\
"1: \n\t"\ "1: \n\t"\
...@@ -208,12 +208,12 @@ static void OPNAME ## vc1_shift2_mmx(uint8_t *dst, const uint8_t *src,\ ...@@ -208,12 +208,12 @@ static void OPNAME ## vc1_shift2_mmx(uint8_t *dst, const uint8_t *src,\
"movq %%mm3, (%1) \n\t"\ "movq %%mm3, (%1) \n\t"\
"add %6, %0 \n\t"\ "add %6, %0 \n\t"\
"add %4, %1 \n\t"\ "add %4, %1 \n\t"\
"dec %%"REG_c" \n\t"\ "dec %%"FF_REG_c" \n\t"\
"jnz 1b \n\t"\ "jnz 1b \n\t"\
: "+r"(src), "+r"(dst)\ : "+r"(src), "+r"(dst)\
: "r"(offset), "r"(-2*offset), "g"(stride), "m"(rnd),\ : "r"(offset), "r"(-2*offset), "g"(stride), "m"(rnd),\
"g"(stride-offset)\ "g"(stride-offset)\
: "%"REG_c, "memory"\ : "%"FF_REG_c, "memory"\
);\ );\
} }
......
libavutil/x86/asm.h
View file @ 1e9c5bf4
...@@ -28,45 +28,45 @@ typedef struct xmm_reg { uint64_t a, b; } xmm_reg; ...@@ -28,45 +28,45 @@ typedef struct xmm_reg { uint64_t a, b; } xmm_reg;
typedef struct ymm_reg { uint64_t a, b, c, d; } ymm_reg; typedef struct ymm_reg { uint64_t a, b, c, d; } ymm_reg;
#if ARCH_X86_64 #if ARCH_X86_64
# define OPSIZE "q" # define FF_OPSIZE "q"
# define REG_a "rax" # define FF_REG_a "rax"
# define REG_b "rbx" # define FF_REG_b "rbx"
# define REG_c "rcx" # define FF_REG_c "rcx"
# define REG_d "rdx" # define FF_REG_d "rdx"
# define REG_D "rdi" # define FF_REG_D "rdi"
# define REG_S "rsi" # define FF_REG_S "rsi"
# define PTR_SIZE "8" # define FF_PTR_SIZE "8"
typedef int64_t x86_reg; typedef int64_t x86_reg;
# define REG_SP "rsp" # define FF_REG_SP "rsp"
# define REG_BP "rbp" # define FF_REG_BP "rbp"
# define REGBP rbp # define FF_REGBP rbp
# define REGa rax # define FF_REGa rax
# define REGb rbx # define FF_REGb rbx
# define REGc rcx # define FF_REGc rcx
# define REGd rdx # define FF_REGd rdx
# define REGSP rsp # define FF_REGSP rsp
#elif ARCH_X86_32 #elif ARCH_X86_32
# define OPSIZE "l" # define FF_OPSIZE "l"
# define REG_a "eax" # define FF_REG_a "eax"
# define REG_b "ebx" # define FF_REG_b "ebx"
# define REG_c "ecx" # define FF_REG_c "ecx"
# define REG_d "edx" # define FF_REG_d "edx"
# define REG_D "edi" # define FF_REG_D "edi"
# define REG_S "esi" # define FF_REG_S "esi"
# define PTR_SIZE "4" # define FF_PTR_SIZE "4"
typedef int32_t x86_reg; typedef int32_t x86_reg;
# define REG_SP "esp" # define FF_REG_SP "esp"
# define REG_BP "ebp" # define FF_REG_BP "ebp"
# define REGBP ebp # define FF_REGBP ebp
# define REGa eax # define FF_REGa eax
# define REGb ebx # define FF_REGb ebx
# define REGc ecx # define FF_REGc ecx
# define REGd edx # define FF_REGd edx
# define REGSP esp # define FF_REGSP esp
#else #else
typedef int x86_reg; typedef int x86_reg;
#endif #endif
......
libavutil/x86/cpu.c
View file @ 1e9c5bf4
...@@ -41,9 +41,9 @@ ...@@ -41,9 +41,9 @@
/* ebx saving is necessary for PIC. gcc seems unable to see it alone */ /* ebx saving is necessary for PIC. gcc seems unable to see it alone */
#define cpuid(index, eax, ebx, ecx, edx) \ #define cpuid(index, eax, ebx, ecx, edx) \
__asm__ volatile ( \ __asm__ volatile ( \
"mov %%"REG_b", %%"REG_S" \n\t" \ "mov %%"FF_REG_b", %%"FF_REG_S" \n\t" \
"cpuid \n\t" \ "cpuid \n\t" \
"xchg %%"REG_b", %%"REG_S \ "xchg %%"FF_REG_b", %%"FF_REG_S \
: "=a" (eax), "=S" (ebx), "=c" (ecx), "=d" (edx) \ : "=a" (eax), "=S" (ebx), "=c" (ecx), "=d" (edx) \
: "0" (index)) : "0" (index))
......
libswscale/utils.c
View file @ 1e9c5bf4
...@@ -649,9 +649,9 @@ static av_cold int init_hscaler_mmxext(int dstW, int xInc, uint8_t *filterCode, ...@@ -649,9 +649,9 @@ static av_cold int init_hscaler_mmxext(int dstW, int xInc, uint8_t *filterCode,
"jmp 9f \n\t" "jmp 9f \n\t"
// Begin // Begin
"0: \n\t" "0: \n\t"
"movq (%%"REG_d", %%"REG_a"), %%mm3 \n\t" "movq (%%"FF_REG_d", %%"FF_REG_a"), %%mm3 \n\t"
"movd (%%"REG_c", %%"REG_S"), %%mm0 \n\t" "movd (%%"FF_REG_c", %%"FF_REG_S"), %%mm0 \n\t"
"movd 1(%%"REG_c", %%"REG_S"), %%mm1 \n\t" "movd 1(%%"FF_REG_c", %%"FF_REG_S"), %%mm1 \n\t"
"punpcklbw %%mm7, %%mm1 \n\t" "punpcklbw %%mm7, %%mm1 \n\t"
"punpcklbw %%mm7, %%mm0 \n\t" "punpcklbw %%mm7, %%mm0 \n\t"
"pshufw $0xFF, %%mm1, %%mm1 \n\t" "pshufw $0xFF, %%mm1, %%mm1 \n\t"
...@@ -659,14 +659,14 @@ static av_cold int init_hscaler_mmxext(int dstW, int xInc, uint8_t *filterCode, ...@@ -659,14 +659,14 @@ static av_cold int init_hscaler_mmxext(int dstW, int xInc, uint8_t *filterCode,
"pshufw $0xFF, %%mm0, %%mm0 \n\t" "pshufw $0xFF, %%mm0, %%mm0 \n\t"
"2: \n\t" "2: \n\t"
"psubw %%mm1, %%mm0 \n\t" "psubw %%mm1, %%mm0 \n\t"
"movl 8(%%"REG_b", %%"REG_a"), %%esi \n\t" "movl 8(%%"FF_REG_b", %%"FF_REG_a"), %%esi \n\t"
"pmullw %%mm3, %%mm0 \n\t" "pmullw %%mm3, %%mm0 \n\t"
"psllw $7, %%mm1 \n\t" "psllw $7, %%mm1 \n\t"
"paddw %%mm1, %%mm0 \n\t" "paddw %%mm1, %%mm0 \n\t"
"movq %%mm0, (%%"REG_D", %%"REG_a") \n\t" "movq %%mm0, (%%"FF_REG_D", %%"FF_REG_a") \n\t"
"add $8, %%"REG_a" \n\t" "add $8, %%"FF_REG_a" \n\t"
// End // End
"9: \n\t" "9: \n\t"
// "int $3 \n\t" // "int $3 \n\t"
...@@ -689,22 +689,22 @@ static av_cold int init_hscaler_mmxext(int dstW, int xInc, uint8_t *filterCode, ...@@ -689,22 +689,22 @@ static av_cold int init_hscaler_mmxext(int dstW, int xInc, uint8_t *filterCode,
"jmp 9f \n\t" "jmp 9f \n\t"
// Begin // Begin
"0: \n\t" "0: \n\t"
"movq (%%"REG_d", %%"REG_a"), %%mm3 \n\t" "movq (%%"FF_REG_d", %%"FF_REG_a"), %%mm3 \n\t"
"movd (%%"REG_c", %%"REG_S"), %%mm0 \n\t" "movd (%%"FF_REG_c", %%"FF_REG_S"), %%mm0 \n\t"
"punpcklbw %%mm7, %%mm0 \n\t" "punpcklbw %%mm7, %%mm0 \n\t"
"pshufw $0xFF, %%mm0, %%mm1 \n\t" "pshufw $0xFF, %%mm0, %%mm1 \n\t"
"1: \n\t" "1: \n\t"
"pshufw $0xFF, %%mm0, %%mm0 \n\t" "pshufw $0xFF, %%mm0, %%mm0 \n\t"
"2: \n\t" "2: \n\t"
"psubw %%mm1, %%mm0 \n\t" "psubw %%mm1, %%mm0 \n\t"
"movl 8(%%"REG_b", %%"REG_a"), %%esi \n\t" "movl 8(%%"FF_REG_b", %%"FF_REG_a"), %%esi \n\t"
"pmullw %%mm3, %%mm0 \n\t" "pmullw %%mm3, %%mm0 \n\t"
"psllw $7, %%mm1 \n\t" "psllw $7, %%mm1 \n\t"
"paddw %%mm1, %%mm0 \n\t" "paddw %%mm1, %%mm0 \n\t"
"movq %%mm0, (%%"REG_D", %%"REG_a") \n\t" "movq %%mm0, (%%"FF_REG_D", %%"FF_REG_a") \n\t"
"add $8, %%"REG_a" \n\t" "add $8, %%"FF_REG_a" \n\t"
// End // End
"9: \n\t" "9: \n\t"
// "int $3 \n\t" // "int $3 \n\t"
......
libswscale/x86/rgb2rgb_template.c
View file @ 1e9c5bf4
This diff is collapsed.
libswscale/x86/swscale_template.c
View file @ 1e9c5bf4
This diff is collapsed.
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment