Commit 5cdc3a4c authored by Jan Kratochvil's avatar Jan Kratochvil Committed by John Koleszar

nasm: address labels 'rel label' vice 'wrt rip'

nasm does not support `label wrt rip', it requires `rel label'. It is
still fully compatible with yasm.

Provide nasm compatibility. No binary change by this patch with yasm on
{x86_64,i686}-fedora13-linux-gnu. Few longer opcodes with nasm on
{x86_64,i686}-fedora13-linux-gnu have been checked as safe.

Change-Id: I488773a4e930a56e43b0cc72d867ee5291215f50
parent e114f699
......@@ -58,11 +58,11 @@ sym(vp8_short_idct4x4llm_mmx):
movq mm5, mm1
paddw mm2, mm0 ; a1 =0+2
pmulhw mm5, [x_s1sqr2 GLOBAL] ;
pmulhw mm5, [GLOBAL(x_s1sqr2)] ;
paddw mm5, mm1 ; ip1 * sin(pi/8) * sqrt(2)
movq mm7, mm3 ;
pmulhw mm7, [x_c1sqr2less1 GLOBAL] ;
pmulhw mm7, [GLOBAL(x_c1sqr2less1)] ;
paddw mm7, mm3 ; ip3 * cos(pi/8) * sqrt(2)
psubw mm7, mm5 ; c1
......@@ -70,10 +70,10 @@ sym(vp8_short_idct4x4llm_mmx):
movq mm5, mm1
movq mm4, mm3
pmulhw mm5, [x_c1sqr2less1 GLOBAL]
pmulhw mm5, [GLOBAL(x_c1sqr2less1)]
paddw mm5, mm1
pmulhw mm3, [x_s1sqr2 GLOBAL]
pmulhw mm3, [GLOBAL(x_s1sqr2)]
paddw mm3, mm4
paddw mm3, mm5 ; d1
......@@ -113,11 +113,11 @@ sym(vp8_short_idct4x4llm_mmx):
movq mm5, mm1
paddw mm2, mm0 ; a1 =0+2
pmulhw mm5, [x_s1sqr2 GLOBAL] ;
pmulhw mm5, [GLOBAL(x_s1sqr2)] ;
paddw mm5, mm1 ; ip1 * sin(pi/8) * sqrt(2)
movq mm7, mm3 ;
pmulhw mm7, [x_c1sqr2less1 GLOBAL] ;
pmulhw mm7, [GLOBAL(x_c1sqr2less1)] ;
paddw mm7, mm3 ; ip3 * cos(pi/8) * sqrt(2)
psubw mm7, mm5 ; c1
......@@ -125,16 +125,16 @@ sym(vp8_short_idct4x4llm_mmx):
movq mm5, mm1
movq mm4, mm3
pmulhw mm5, [x_c1sqr2less1 GLOBAL]
pmulhw mm5, [GLOBAL(x_c1sqr2less1)]
paddw mm5, mm1
pmulhw mm3, [x_s1sqr2 GLOBAL]
pmulhw mm3, [GLOBAL(x_s1sqr2)]
paddw mm3, mm4
paddw mm3, mm5 ; d1
paddw mm0, [fours GLOBAL]
paddw mm0, [GLOBAL(fours)]
paddw mm2, [fours GLOBAL]
paddw mm2, [GLOBAL(fours)]
movq mm6, mm2 ; a1
movq mm4, mm0 ; b1
......@@ -196,7 +196,7 @@ sym(vp8_short_idct4x4llm_1_mmx):
mov rax, arg(0) ;input
movd mm0, [rax]
paddw mm0, [fours GLOBAL]
paddw mm0, [GLOBAL(fours)]
mov rdx, arg(1) ;output
psraw mm0, 3
......@@ -239,7 +239,7 @@ sym(vp8_dc_only_idct_add_mmx):
movd mm5, arg(0) ;input_dc
paddw mm5, [fours GLOBAL]
paddw mm5, [GLOBAL(fours)]
psraw mm5, 3
......
......@@ -51,7 +51,7 @@ sym(idct_dequant_0_2x_sse2):
pshufhw xmm4, xmm4, 00000000b
mov rax, arg(2) ; pre
paddw xmm4, [fours GLOBAL]
paddw xmm4, [GLOBAL(fours)]
movsxd rcx, dword ptr arg(5) ; blk_stride
psraw xmm4, 3
......@@ -160,11 +160,11 @@ sym(idct_dequant_full_2x_sse2):
movdqa xmm5, xmm1
paddw xmm2, xmm0 ; a1 = 0+2
pmulhw xmm5, [x_s1sqr2 GLOBAL]
pmulhw xmm5, [GLOBAL(x_s1sqr2)]
paddw xmm5, xmm1 ; ip1 * sin(pi/8) * sqrt(2)
movdqa xmm7, xmm3
pmulhw xmm7, [x_c1sqr2less1 GLOBAL]
pmulhw xmm7, [GLOBAL(x_c1sqr2less1)]
paddw xmm7, xmm3 ; ip3 * cos(pi/8) * sqrt(2)
psubw xmm7, xmm5 ; c1
......@@ -172,10 +172,10 @@ sym(idct_dequant_full_2x_sse2):
movdqa xmm5, xmm1
movdqa xmm4, xmm3
pmulhw xmm5, [x_c1sqr2less1 GLOBAL]
pmulhw xmm5, [GLOBAL(x_c1sqr2less1)]
paddw xmm5, xmm1
pmulhw xmm3, [x_s1sqr2 GLOBAL]
pmulhw xmm3, [GLOBAL(x_s1sqr2)]
paddw xmm3, xmm4
paddw xmm3, xmm5 ; d1
......@@ -229,11 +229,11 @@ sym(idct_dequant_full_2x_sse2):
movdqa xmm5, xmm1
paddw xmm2, xmm0 ; a1 = 0+2
pmulhw xmm5, [x_s1sqr2 GLOBAL]
pmulhw xmm5, [GLOBAL(x_s1sqr2)]
paddw xmm5, xmm1 ; ip1 * sin(pi/8) * sqrt(2)
movdqa xmm7, xmm3
pmulhw xmm7, [x_c1sqr2less1 GLOBAL]
pmulhw xmm7, [GLOBAL(x_c1sqr2less1)]
paddw xmm7, xmm3 ; ip3 * cos(pi/8) * sqrt(2)
psubw xmm7, xmm5 ; c1
......@@ -241,16 +241,16 @@ sym(idct_dequant_full_2x_sse2):
movdqa xmm5, xmm1
movdqa xmm4, xmm3
pmulhw xmm5, [x_c1sqr2less1 GLOBAL]
pmulhw xmm5, [GLOBAL(x_c1sqr2less1)]
paddw xmm5, xmm1
pmulhw xmm3, [x_s1sqr2 GLOBAL]
pmulhw xmm3, [GLOBAL(x_s1sqr2)]
paddw xmm3, xmm4
paddw xmm3, xmm5 ; d1
paddw xmm0, [fours GLOBAL]
paddw xmm0, [GLOBAL(fours)]
paddw xmm2, [fours GLOBAL]
paddw xmm2, [GLOBAL(fours)]
movdqa xmm6, xmm2 ; a1
movdqa xmm4, xmm0 ; b1
......@@ -394,7 +394,7 @@ sym(idct_dequant_dc_0_2x_sse2):
punpckldq xmm4, xmm4
; Rounding to dequant and downshift
paddw xmm4, [fours GLOBAL]
paddw xmm4, [GLOBAL(fours)]
psraw xmm4, 3
; Predict buffer needs to be expanded from bytes to words
......@@ -505,11 +505,11 @@ sym(idct_dequant_dc_full_2x_sse2):
movdqa xmm5, xmm1
paddw xmm2, xmm0 ; a1 = 0+2
pmulhw xmm5, [x_s1sqr2 GLOBAL]
pmulhw xmm5, [GLOBAL(x_s1sqr2)]
paddw xmm5, xmm1 ; ip1 * sin(pi/8) * sqrt(2)
movdqa xmm7, xmm3
pmulhw xmm7, [x_c1sqr2less1 GLOBAL]
pmulhw xmm7, [GLOBAL(x_c1sqr2less1)]
paddw xmm7, xmm3 ; ip3 * cos(pi/8) * sqrt(2)
psubw xmm7, xmm5 ; c1
......@@ -517,10 +517,10 @@ sym(idct_dequant_dc_full_2x_sse2):
movdqa xmm5, xmm1
movdqa xmm4, xmm3
pmulhw xmm5, [x_c1sqr2less1 GLOBAL]
pmulhw xmm5, [GLOBAL(x_c1sqr2less1)]
paddw xmm5, xmm1
pmulhw xmm3, [x_s1sqr2 GLOBAL]
pmulhw xmm3, [GLOBAL(x_s1sqr2)]
paddw xmm3, xmm4
paddw xmm3, xmm5 ; d1
......@@ -574,11 +574,11 @@ sym(idct_dequant_dc_full_2x_sse2):
movdqa xmm5, xmm1
paddw xmm2, xmm0 ; a1 = 0+2
pmulhw xmm5, [x_s1sqr2 GLOBAL]
pmulhw xmm5, [GLOBAL(x_s1sqr2)]
paddw xmm5, xmm1 ; ip1 * sin(pi/8) * sqrt(2)
movdqa xmm7, xmm3
pmulhw xmm7, [x_c1sqr2less1 GLOBAL]
pmulhw xmm7, [GLOBAL(x_c1sqr2less1)]
paddw xmm7, xmm3 ; ip3 * cos(pi/8) * sqrt(2)
psubw xmm7, xmm5 ; c1
......@@ -586,16 +586,16 @@ sym(idct_dequant_dc_full_2x_sse2):
movdqa xmm5, xmm1
movdqa xmm4, xmm3
pmulhw xmm5, [x_c1sqr2less1 GLOBAL]
pmulhw xmm5, [GLOBAL(x_c1sqr2less1)]
paddw xmm5, xmm1
pmulhw xmm3, [x_s1sqr2 GLOBAL]
pmulhw xmm3, [GLOBAL(x_s1sqr2)]
paddw xmm3, xmm4
paddw xmm3, xmm5 ; d1
paddw xmm0, [fours GLOBAL]
paddw xmm0, [GLOBAL(fours)]
paddw xmm2, [fours GLOBAL]
paddw xmm2, [GLOBAL(fours)]
movdqa xmm6, xmm2 ; a1
movdqa xmm4, xmm0 ; b1
......
This diff is collapsed.
This diff is collapsed.
......@@ -37,16 +37,16 @@ sym(vp8_post_proc_down_and_across_mmx):
%if ABI_IS_32BIT=1 && CONFIG_PIC=1
; move the global rd onto the stack, since we don't have enough registers
; to do PIC addressing
movq mm0, [rd GLOBAL]
movq mm0, [GLOBAL(rd)]
sub rsp, 8
movq [rsp], mm0
%define RD [rsp]
%else
%define RD [rd GLOBAL]
%define RD [GLOBAL(rd)]
%endif
push rbx
lea rbx, [Blur GLOBAL]
lea rbx, [GLOBAL(Blur)]
movd mm2, dword ptr arg(6) ;flimit
punpcklwd mm2, mm2
punpckldq mm2, mm2
......@@ -286,7 +286,7 @@ sym(vp8_mbpost_proc_down_mmx):
%define flimit2 [rsp+128]
%if ABI_IS_32BIT=0
lea r8, [sym(vp8_rv) GLOBAL]
lea r8, [GLOBAL(sym(vp8_rv))]
%endif
;rows +=8;
......@@ -404,7 +404,7 @@ loop_row:
and rcx, 127
%if ABI_IS_32BIT=1 && CONFIG_PIC=1
push rax
lea rax, [sym(vp8_rv) GLOBAL]
lea rax, [GLOBAL(sym(vp8_rv))]
movq mm4, [rax + rcx*2] ;vp8_rv[rcx*2]
pop rax
%elif ABI_IS_32BIT=0
......
......@@ -36,12 +36,12 @@ sym(vp8_post_proc_down_and_across_xmm):
ALIGN_STACK 16, rax
; move the global rd onto the stack, since we don't have enough registers
; to do PIC addressing
movdqa xmm0, [rd42 GLOBAL]
movdqa xmm0, [GLOBAL(rd42)]
sub rsp, 16
movdqa [rsp], xmm0
%define RD42 [rsp]
%else
%define RD42 [rd42 GLOBAL]
%define RD42 [GLOBAL(rd42)]
%endif
......@@ -275,7 +275,7 @@ sym(vp8_mbpost_proc_down_xmm):
%define flimit4 [rsp+128]
%if ABI_IS_32BIT=0
lea r8, [sym(vp8_rv) GLOBAL]
lea r8, [GLOBAL(sym(vp8_rv))]
%endif
;rows +=8;
......@@ -393,7 +393,7 @@ loop_row:
and rcx, 127
%if ABI_IS_32BIT=1 && CONFIG_PIC=1
push rax
lea rax, [sym(vp8_rv) GLOBAL]
lea rax, [GLOBAL(sym(vp8_rv))]
movdqu xmm4, [rax + rcx*2] ;vp8_rv[rcx*2]
pop rax
%elif ABI_IS_32BIT=0
......@@ -579,7 +579,7 @@ nextcol4:
punpcklwd xmm1, xmm0
paddd xmm1, xmm6
paddd xmm1, [four8s GLOBAL]
paddd xmm1, [GLOBAL(four8s)]
psrad xmm1, 4
packssdw xmm1, xmm0
......
......@@ -84,7 +84,7 @@ nextrow:
pmullw mm5, [rdx] ; mm5 *= kernel 5 modifiers
paddsw mm3, mm5 ; mm3 += mm5
paddsw mm3, [rd GLOBAL] ; mm3 += round value
paddsw mm3, [GLOBAL(rd)] ; mm3 += round value
psraw mm3, VP8_FILTER_SHIFT ; mm3 /= 128
packuswb mm3, mm0 ; pack and unpack to saturate
punpcklbw mm3, mm0 ;
......@@ -136,7 +136,7 @@ sym(vp8_filter_block1d_v6_mmx):
push rdi
; end prolog
movq mm5, [rd GLOBAL]
movq mm5, [GLOBAL(rd)]
push rbx
mov rbx, arg(6) ;vp8_filter
movq mm1, [rbx + 16] ; do both the negative taps first!!!
......@@ -225,7 +225,7 @@ sym(vp8_filter_block1dc_v6_mmx):
push rdi
; end prolog
movq mm5, [rd GLOBAL]
movq mm5, [GLOBAL(rd)]
push rbx
mov rbx, arg(7) ;vp8_filter
movq mm1, [rbx + 16] ; do both the negative taps first!!!
......@@ -320,7 +320,7 @@ sym(vp8_bilinear_predict8x8_mmx):
mov rdi, arg(4) ;dst_ptr ;
shl rax, 5 ; offset * 32
lea rcx, [sym(vp8_bilinear_filters_mmx) GLOBAL]
lea rcx, [GLOBAL(sym(vp8_bilinear_filters_mmx))]
add rax, rcx ; HFilter
mov rsi, arg(0) ;src_ptr ;
......@@ -363,10 +363,10 @@ sym(vp8_bilinear_predict8x8_mmx):
paddw mm3, mm5 ;
paddw mm4, mm6 ;
paddw mm3, [rd GLOBAL] ; xmm3 += round value
paddw mm3, [GLOBAL(rd)] ; xmm3 += round value
psraw mm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw mm4, [rd GLOBAL] ;
paddw mm4, [GLOBAL(rd)] ;
psraw mm4, VP8_FILTER_SHIFT ;
movq mm7, mm3 ;
......@@ -404,10 +404,10 @@ next_row_8x8:
pmullw mm5, [rax] ;
pmullw mm6, [rax] ;
paddw mm3, [rd GLOBAL] ; xmm3 += round value
paddw mm3, [GLOBAL(rd)] ; xmm3 += round value
psraw mm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw mm4, [rd GLOBAL] ;
paddw mm4, [GLOBAL(rd)] ;
psraw mm4, VP8_FILTER_SHIFT ;
movq mm7, mm3 ;
......@@ -421,10 +421,10 @@ next_row_8x8:
paddw mm4, mm6 ;
paddw mm3, [rd GLOBAL] ; xmm3 += round value
paddw mm3, [GLOBAL(rd)] ; xmm3 += round value
psraw mm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw mm4, [rd GLOBAL] ;
paddw mm4, [GLOBAL(rd)] ;
psraw mm4, VP8_FILTER_SHIFT ;
packuswb mm3, mm4
......@@ -476,7 +476,7 @@ sym(vp8_bilinear_predict8x4_mmx):
movsxd rax, dword ptr arg(2) ;xoffset
mov rdi, arg(4) ;dst_ptr ;
lea rcx, [sym(vp8_bilinear_filters_mmx) GLOBAL]
lea rcx, [GLOBAL(sym(vp8_bilinear_filters_mmx))]
shl rax, 5
mov rsi, arg(0) ;src_ptr ;
......@@ -518,10 +518,10 @@ sym(vp8_bilinear_predict8x4_mmx):
paddw mm3, mm5 ;
paddw mm4, mm6 ;
paddw mm3, [rd GLOBAL] ; xmm3 += round value
paddw mm3, [GLOBAL(rd)] ; xmm3 += round value
psraw mm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw mm4, [rd GLOBAL] ;
paddw mm4, [GLOBAL(rd)] ;
psraw mm4, VP8_FILTER_SHIFT ;
movq mm7, mm3 ;
......@@ -559,10 +559,10 @@ next_row_8x4:
pmullw mm5, [rax] ;
pmullw mm6, [rax] ;
paddw mm3, [rd GLOBAL] ; xmm3 += round value
paddw mm3, [GLOBAL(rd)] ; xmm3 += round value
psraw mm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw mm4, [rd GLOBAL] ;
paddw mm4, [GLOBAL(rd)] ;
psraw mm4, VP8_FILTER_SHIFT ;
movq mm7, mm3 ;
......@@ -576,10 +576,10 @@ next_row_8x4:
paddw mm4, mm6 ;
paddw mm3, [rd GLOBAL] ; xmm3 += round value
paddw mm3, [GLOBAL(rd)] ; xmm3 += round value
psraw mm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw mm4, [rd GLOBAL] ;
paddw mm4, [GLOBAL(rd)] ;
psraw mm4, VP8_FILTER_SHIFT ;
packuswb mm3, mm4
......@@ -631,7 +631,7 @@ sym(vp8_bilinear_predict4x4_mmx):
movsxd rax, dword ptr arg(2) ;xoffset
mov rdi, arg(4) ;dst_ptr ;
lea rcx, [sym(vp8_bilinear_filters_mmx) GLOBAL]
lea rcx, [GLOBAL(sym(vp8_bilinear_filters_mmx))]
shl rax, 5
add rax, rcx ; HFilter
......@@ -662,7 +662,7 @@ sym(vp8_bilinear_predict4x4_mmx):
pmullw mm5, mm2 ;
paddw mm3, mm5 ;
paddw mm3, [rd GLOBAL] ; xmm3 += round value
paddw mm3, [GLOBAL(rd)] ; xmm3 += round value
psraw mm3, VP8_FILTER_SHIFT ; xmm3 /= 128
......@@ -686,7 +686,7 @@ next_row_4x4:
punpcklbw mm5, mm0 ;
pmullw mm5, [rax] ;
paddw mm3, [rd GLOBAL] ; xmm3 += round value
paddw mm3, [GLOBAL(rd)] ; xmm3 += round value
psraw mm3, VP8_FILTER_SHIFT ; xmm3 /= 128
movq mm7, mm3 ;
......@@ -697,7 +697,7 @@ next_row_4x4:
paddw mm3, mm5 ;
paddw mm3, [rd GLOBAL] ; xmm3 += round value
paddw mm3, [GLOBAL(rd)] ; xmm3 += round value
psraw mm3, VP8_FILTER_SHIFT ; xmm3 /= 128
packuswb mm3, mm0
......
......@@ -107,7 +107,7 @@ filter_block1d8_h6_rowloop:
paddsw xmm4, xmm6
paddsw xmm4, xmm1
paddsw xmm4, [rd GLOBAL]
paddsw xmm4, [GLOBAL(rd)]
psraw xmm4, 7
......@@ -231,7 +231,7 @@ filter_block1d16_h6_sse2_rowloop:
paddsw xmm4, xmm6
paddsw xmm4, xmm1
paddsw xmm4, [rd GLOBAL]
paddsw xmm4, [GLOBAL(rd)]
psraw xmm4, 7
......@@ -284,7 +284,7 @@ filter_block1d16_h6_sse2_rowloop:
paddsw xmm4, xmm6
paddsw xmm4, xmm2
paddsw xmm4, [rd GLOBAL]
paddsw xmm4, [GLOBAL(rd)]
psraw xmm4, 7
......@@ -351,7 +351,7 @@ sym(vp8_filter_block1d8_v6_sse2):
movsxd rcx, DWORD PTR arg(5) ;[output_height]
pxor xmm0, xmm0 ; clear xmm0
movdqa xmm7, XMMWORD PTR [rd GLOBAL]
movdqa xmm7, XMMWORD PTR [GLOBAL(rd)]
%if ABI_IS_32BIT=0
movsxd r8, dword ptr arg(2) ; dst_ptich
%endif
......@@ -489,7 +489,7 @@ vp8_filter_block1d16_v6_sse2_loop:
pmullw xmm5, [rax + 80]
pmullw xmm6, [rax + 80]
movdqa xmm7, XMMWORD PTR [rd GLOBAL]
movdqa xmm7, XMMWORD PTR [GLOBAL(rd)]
pxor xmm0, xmm0 ; clear xmm0
paddsw xmm1, xmm3
......@@ -608,7 +608,7 @@ filter_block1d8_h6_only_rowloop:
paddsw xmm4, xmm6
paddsw xmm4, xmm1
paddsw xmm4, [rd GLOBAL]
paddsw xmm4, [GLOBAL(rd)]
psraw xmm4, 7
......@@ -723,7 +723,7 @@ filter_block1d16_h6_only_sse2_rowloop:
paddsw xmm4, xmm6
paddsw xmm4, xmm1
paddsw xmm4, [rd GLOBAL]
paddsw xmm4, [GLOBAL(rd)]
psraw xmm4, 7
......@@ -773,7 +773,7 @@ filter_block1d16_h6_only_sse2_rowloop:
paddsw xmm4, xmm6
paddsw xmm4, xmm2
paddsw xmm4, [rd GLOBAL]
paddsw xmm4, [GLOBAL(rd)]
psraw xmm4, 7
......@@ -832,7 +832,7 @@ sym(vp8_filter_block1d8_v6_only_sse2):
pxor xmm0, xmm0 ; clear xmm0
movdqa xmm7, XMMWORD PTR [rd GLOBAL]
movdqa xmm7, XMMWORD PTR [GLOBAL(rd)]
%if ABI_IS_32BIT=0
movsxd r8, dword ptr arg(3) ; dst_ptich
%endif
......@@ -978,7 +978,7 @@ sym(vp8_bilinear_predict16x16_sse2):
;const short *HFilter = bilinear_filters_mmx[xoffset]
;const short *VFilter = bilinear_filters_mmx[yoffset]
lea rcx, [sym(vp8_bilinear_filters_mmx) GLOBAL]
lea rcx, [GLOBAL(sym(vp8_bilinear_filters_mmx))]
movsxd rax, dword ptr arg(2) ;xoffset
cmp rax, 0 ;skip first_pass filter if xoffset=0
......@@ -1033,10 +1033,10 @@ sym(vp8_bilinear_predict16x16_sse2):
paddw xmm3, xmm5
paddw xmm4, xmm6
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
paddw xmm3, [GLOBAL(rd)] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw xmm4, [rd GLOBAL]
paddw xmm4, [GLOBAL(rd)]
psraw xmm4, VP8_FILTER_SHIFT
movdqa xmm7, xmm3
......@@ -1074,10 +1074,10 @@ next_row:
pmullw xmm5, [rax]
pmullw xmm6, [rax]
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
paddw xmm3, [GLOBAL(rd)] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw xmm4, [rd GLOBAL]
paddw xmm4, [GLOBAL(rd)]
psraw xmm4, VP8_FILTER_SHIFT
movdqa xmm7, xmm3
......@@ -1089,10 +1089,10 @@ next_row:
paddw xmm3, xmm5
paddw xmm4, xmm6
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
paddw xmm3, [GLOBAL(rd)] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw xmm4, [rd GLOBAL]
paddw xmm4, [GLOBAL(rd)]
psraw xmm4, VP8_FILTER_SHIFT
packuswb xmm3, xmm4
......@@ -1154,10 +1154,10 @@ next_row_spo:
paddw xmm3, xmm5
paddw xmm4, xmm6
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
paddw xmm3, [GLOBAL(rd)] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw xmm4, [rd GLOBAL]
paddw xmm4, [GLOBAL(rd)]
psraw xmm4, VP8_FILTER_SHIFT
packuswb xmm3, xmm4
......@@ -1198,10 +1198,10 @@ next_row_fpo:
paddw xmm3, xmm5
paddw xmm4, xmm6
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
paddw xmm3, [GLOBAL(rd)] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
paddw xmm4, [rd GLOBAL]
paddw xmm4, [GLOBAL(rd)]
psraw xmm4, VP8_FILTER_SHIFT
packuswb xmm3, xmm4
......@@ -1249,7 +1249,7 @@ sym(vp8_bilinear_predict8x8_sse2):
;const short *HFilter = bilinear_filters_mmx[xoffset]
;const short *VFilter = bilinear_filters_mmx[yoffset]
lea rcx, [sym(vp8_bilinear_filters_mmx) GLOBAL]
lea rcx, [GLOBAL(sym(vp8_bilinear_filters_mmx))]
mov rsi, arg(0) ;src_ptr
movsxd rdx, dword ptr arg(1) ;src_pixels_per_line
......@@ -1315,7 +1315,7 @@ sym(vp8_bilinear_predict8x8_sse2):
paddw xmm3, xmm4
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
paddw xmm3, [GLOBAL(rd)] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
movdqa xmm7, xmm3
......@@ -1334,7 +1334,7 @@ next_row8x8:
paddw xmm3, xmm4
pmullw xmm7, xmm5
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
paddw xmm3, [GLOBAL(rd)] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
movdqa xmm4, xmm3
......@@ -1344,7 +1344,7 @@ next_row8x8:
movdqa xmm7, xmm4
paddw xmm3, [rd GLOBAL] ; xmm3 += round value
paddw xmm3, [GLOBAL(rd)] ; xmm3 += round value
psraw xmm3, VP8_FILTER_SHIFT ; xmm3 /= 128
packuswb xmm3, xmm0
......
This diff is collapsed.
......@@ -98,11 +98,11 @@ sym(vp8_dequant_idct_add_mmx):
movq mm5, mm1
paddw mm2, mm0 ; a1 =0+2
pmulhw mm5, [x_s1sqr2 GLOBAL];
pmulhw mm5, [GLOBAL(x_s1sqr2)];
paddw mm5, mm1 ; ip1 * sin(pi/8) * sqrt(2)
movq mm7, mm3 ;
pmulhw mm7, [x_c1sqr2less1 GLOBAL];
pmulhw mm7, [GLOBAL(x_c1sqr2less1)];
paddw mm7, mm3 ; ip3 * cos(pi/8) * sqrt(2)
psubw mm7, mm5 ; c1
......@@ -110,10 +110,10 @@ sym(vp8_dequant_idct_add_mmx):
movq mm5, mm1
movq mm4, mm3
pmulhw mm5, [x_c1sqr2less1 GLOBAL]
pmulhw mm5, [GLOBAL(x_c1sqr2less1)]
paddw mm5, mm1
pmulhw mm3, [x_s1sqr2 GLOBAL]
pmulhw mm3, [GLOBAL(x_s1sqr2)]
paddw mm3, mm4
paddw mm3, mm5 ; d1
......@@ -153,11 +153,11 @@ sym(vp8_dequant_idct_add_mmx):
movq mm5, mm1
paddw mm2, mm0 ; a1 =0+2
pmulhw mm5, [x_s1sqr2 GLOBAL];
pmulhw mm5, [GLOBAL(x_s1sqr2)];
paddw mm5, mm1 ; ip1 * sin(pi/8) * sqrt(2)
movq mm7, mm3 ;
pmulhw mm7, [x_c1sqr2less1 GLOBAL];
pmulhw mm7, [GLOBAL(x_c1sqr2less1)];