|
 |
0a4abc |
https://gerrit.chromium.org/gerrit/#/c/71007/
|
|
|
0a4abc |
|
|
|
0a4abc |
[modified]
|
|
|
0a4abc |
|
|
|
0a4abc |
commit 2695f18211e9b1017647af608a64d72a688ffbe7
|
|
|
0a4abc |
Author: Jan Kratochvil <jan.kratochvil@redhat.com>
|
|
|
0a4abc |
Date: Mon Jul 28 20:07:38 2014 +0200
|
|
|
0a4abc |
|
|
|
0a4abc |
Fix --as=nasm compatibility for new asm code.
|
|
|
0a4abc |
|
|
|
0a4abc |
s/movd/movq/
|
|
|
0a4abc |
s/pmovmskb rX,/pmovmskb rXd,/
|
|
|
0a4abc |
Add end-of-line ':' label markers.
|
|
|
0a4abc |
|
|
|
0a4abc |
Change-Id: Icc2c8b66af3cf72598361021699e099739f813d7
|
|
|
0a4abc |
|
|
|
0a4abc |
diff --git a/vp9/common/x86/vp9_subpixel_8t_ssse3.asm b/vp9/common/x86/vp9_subpixel_8t_ssse3.asm
|
|
|
0a4abc |
index fd781d4..57eee70 100644
|
|
|
0a4abc |
--- a/vp9/common/x86/vp9_subpixel_8t_ssse3.asm
|
|
|
0a4abc |
+++ b/vp9/common/x86/vp9_subpixel_8t_ssse3.asm
|
|
|
0a4abc |
@@ -18,7 +18,7 @@
|
|
|
0a4abc |
mov rcx, 0x0400040
|
|
|
0a4abc |
|
|
|
0a4abc |
movdqa xmm4, [rdx] ;load filters
|
|
|
0a4abc |
- movd xmm5, rcx
|
|
|
0a4abc |
+ movq xmm5, rcx
|
|
|
0a4abc |
packsswb xmm4, xmm4
|
|
|
0a4abc |
pshuflw xmm0, xmm4, 0b ;k0_k1
|
|
|
0a4abc |
pshuflw xmm1, xmm4, 01010101b ;k2_k3
|
|
|
0a4abc |
@@ -624,7 +624,7 @@ sym(vp9_filter_block1d16_v8_avg_ssse3):
|
|
|
0a4abc |
pavgb xmm0, xmm1
|
|
|
0a4abc |
%endif
|
|
|
0a4abc |
movd [rdi], xmm0
|
|
|
0a4abc |
-.done
|
|
|
0a4abc |
+.done:
|
|
|
0a4abc |
%endm
|
|
|
0a4abc |
|
|
|
0a4abc |
%macro HORIZx8_ROW 4
|
|
|
0a4abc |
@@ -661,7 +661,7 @@ sym(vp9_filter_block1d16_v8_avg_ssse3):
|
|
|
0a4abc |
mov rcx, 0x0400040
|
|
|
0a4abc |
|
|
|
0a4abc |
movdqa xmm4, [rdx] ;load filters
|
|
|
0a4abc |
- movd xmm5, rcx
|
|
|
0a4abc |
+ movq xmm5, rcx
|
|
|
0a4abc |
packsswb xmm4, xmm4
|
|
|
0a4abc |
pshuflw xmm0, xmm4, 0b ;k0_k1
|
|
|
0a4abc |
pshuflw xmm1, xmm4, 01010101b ;k2_k3
|
|
|
0a4abc |
@@ -727,7 +727,7 @@ sym(vp9_filter_block1d16_v8_avg_ssse3):
|
|
|
0a4abc |
pavgb xmm0, xmm1
|
|
|
0a4abc |
%endif
|
|
|
0a4abc |
movq [rdi], xmm0
|
|
|
0a4abc |
-.done
|
|
|
0a4abc |
+.done:
|
|
|
0a4abc |
%endm
|
|
|
0a4abc |
|
|
|
0a4abc |
%macro HORIZx16 1
|
|
|
0a4abc |
diff --git a/vp9/encoder/x86/vp9_quantize_ssse3_x86_64.asm b/vp9/encoder/x86/vp9_quantize_ssse3_x86_64.asm
|
|
|
0a4abc |
index 508e1d4..1c97342 100644
|
|
|
0a4abc |
--- a/vp9/encoder/x86/vp9_quantize_ssse3.asm
|
|
|
0a4abc |
+++ b/vp9/encoder/x86/vp9_quantize_ssse3.asm
|
|
|
0a4abc |
@@ -122,8 +122,8 @@ cglobal quantize_%1, 0, %2, 15, coeff, ncoeff, skip, zbin, round, quant, \
|
|
|
0a4abc |
pcmpgtw m7, m6, m0 ; m7 = c[i] >= zbin
|
|
|
0a4abc |
pcmpgtw m12, m11, m0 ; m12 = c[i] >= zbin
|
|
|
0a4abc |
%ifidn %1, b_32x32
|
|
|
0a4abc |
- pmovmskb r6, m7
|
|
|
0a4abc |
- pmovmskb r2, m12
|
|
|
0a4abc |
+ pmovmskb r6d, m7
|
|
|
0a4abc |
+ pmovmskb r2d, m12
|
|
|
0a4abc |
or r6, r2
|
|
|
0a4abc |
jz .skip_iter
|
|
|
0a4abc |
%endif
|
|
|
0a4abc |
diff --git a/vp9/encoder/x86/vp9_subpel_variance.asm b/vp9/encoder/x86/vp9_subpel_variance.asm
|
|
|
0a4abc |
index 1a9e4e8..e06e220 100644
|
|
|
0a4abc |
--- a/vp9/encoder/x86/vp9_subpel_variance.asm
|
|
|
0a4abc |
+++ b/vp9/encoder/x86/vp9_subpel_variance.asm
|
|
|
0a4abc |
@@ -101,7 +101,6 @@ SECTION .text
|
|
|
0a4abc |
pshufd m4, m6, 0x1
|
|
|
0a4abc |
movd [r1], m7 ; store sse
|
|
|
0a4abc |
paddd m6, m4
|
|
|
0a4abc |
- movd rax, m6 ; store sum as return value
|
|
|
0a4abc |
%else ; mmsize == 8
|
|
|
0a4abc |
pshufw m4, m6, 0xe
|
|
|
0a4abc |
pshufw m3, m7, 0xe
|
|
|
0a4abc |
@@ -113,7 +112,11 @@ SECTION .text
|
|
|
0a4abc |
movd [r1], m7 ; store sse
|
|
|
0a4abc |
pshufw m4, m6, 0xe
|
|
|
0a4abc |
paddd m6, m4
|
|
|
0a4abc |
- movd rax, m6 ; store sum as return value
|
|
|
0a4abc |
+%endif
|
|
|
0a4abc |
+%if ARCH_X86_64
|
|
|
0a4abc |
+ movq rax, m6 ; store sum as return value
|
|
|
0a4abc |
+%else
|
|
|
0a4abc |
+ movd eax, m6 ; store sum as return value
|
|
|
0a4abc |
%endif
|
|
|
0a4abc |
RET
|
|
|
0a4abc |
%endmacro
|