diff options
Diffstat (limited to 'vp9-nasm.patch')
| -rw-r--r-- | vp9-nasm.patch | 96 |
1 files changed, 96 insertions, 0 deletions
diff --git a/vp9-nasm.patch b/vp9-nasm.patch new file mode 100644 index 0000000..5e80cb7 --- /dev/null +++ b/vp9-nasm.patch @@ -0,0 +1,96 @@ +https://gerrit.chromium.org/gerrit/#/c/71007/ + +[modified] + +commit 2695f18211e9b1017647af608a64d72a688ffbe7 +Author: Jan Kratochvil <jan.kratochvil@redhat.com> +Date: Mon Jul 28 20:07:38 2014 +0200 + + Fix --as=nasm compatibility for new asm code. + + s/movd/movq/ + s/pmovmskb rX,/pmovmskb rXd,/ + Add end-of-line ':' label markers. + + Change-Id: Icc2c8b66af3cf72598361021699e099739f813d7 + +diff --git a/vp9/common/x86/vp9_subpixel_8t_ssse3.asm b/vp9/common/x86/vp9_subpixel_8t_ssse3.asm +index fd781d4..57eee70 100644 +--- a/vp9/common/x86/vp9_subpixel_8t_ssse3.asm ++++ b/vp9/common/x86/vp9_subpixel_8t_ssse3.asm +@@ -18,7 +18,7 @@ + mov rcx, 0x0400040 + + movdqa xmm4, [rdx] ;load filters +- movd xmm5, rcx ++ movq xmm5, rcx + packsswb xmm4, xmm4 + pshuflw xmm0, xmm4, 0b ;k0_k1 + pshuflw xmm1, xmm4, 01010101b ;k2_k3 +@@ -624,7 +624,7 @@ sym(vp9_filter_block1d16_v8_avg_ssse3): + pavgb xmm0, xmm1 + %endif + movd [rdi], xmm0 +-.done ++.done: + %endm + + %macro HORIZx8_ROW 4 +@@ -661,7 +661,7 @@ sym(vp9_filter_block1d16_v8_avg_ssse3): + mov rcx, 0x0400040 + + movdqa xmm4, [rdx] ;load filters +- movd xmm5, rcx ++ movq xmm5, rcx + packsswb xmm4, xmm4 + pshuflw xmm0, xmm4, 0b ;k0_k1 + pshuflw xmm1, xmm4, 01010101b ;k2_k3 +@@ -727,7 +727,7 @@ sym(vp9_filter_block1d16_v8_avg_ssse3): + pavgb xmm0, xmm1 + %endif + movq [rdi], xmm0 +-.done ++.done: + %endm + + %macro HORIZx16 1 +diff --git a/vp9/encoder/x86/vp9_quantize_ssse3_x86_64.asm b/vp9/encoder/x86/vp9_quantize_ssse3_x86_64.asm +index 508e1d4..1c97342 100644 +--- a/vp9/encoder/x86/vp9_quantize_ssse3.asm ++++ b/vp9/encoder/x86/vp9_quantize_ssse3.asm +@@ -122,8 +122,8 @@ cglobal quantize_%1, 0, %2, 15, coeff, ncoeff, skip, zbin, round, quant, \ + pcmpgtw m7, m6, m0 ; m7 = c[i] >= zbin + pcmpgtw m12, m11, m0 ; m12 = c[i] >= zbin + %ifidn %1, b_32x32 +- pmovmskb r6, m7 +- pmovmskb r2, m12 ++ pmovmskb r6d, m7 ++ pmovmskb r2d, m12 + or r6, r2 + jz .skip_iter + %endif +diff --git a/vp9/encoder/x86/vp9_subpel_variance.asm b/vp9/encoder/x86/vp9_subpel_variance.asm +index 1a9e4e8..e06e220 100644 +--- a/vp9/encoder/x86/vp9_subpel_variance.asm ++++ b/vp9/encoder/x86/vp9_subpel_variance.asm +@@ -101,7 +101,6 @@ SECTION .text + pshufd m4, m6, 0x1 + movd [r1], m7 ; store sse + paddd m6, m4 +- movd rax, m6 ; store sum as return value + %else ; mmsize == 8 + pshufw m4, m6, 0xe + pshufw m3, m7, 0xe +@@ -113,7 +112,11 @@ SECTION .text + movd [r1], m7 ; store sse + pshufw m4, m6, 0xe + paddd m6, m4 +- movd rax, m6 ; store sum as return value ++%endif ++%if ARCH_X86_64 ++ movq rax, m6 ; store sum as return value ++%else ++ movd eax, m6 ; store sum as return value + %endif + RET + %endmacro |