Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(7)

Issues Search
    My Issues | Starred     Open | Closed | All

Unified Diff: source/libvpx/vp8/common/x86/loopfilter_sse2.asm

Issue 7671004: Update libvpx snapshot to v0.9.7-p1 (Cayuga). (Closed) Base URL: svn://chrome-svn/chrome/trunk/deps/third_party/libvpx/
Patch Set: '' Created 9 years, 4 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View side-by-side diff with in-line comments
Download patch
« no previous file with comments | « source/libvpx/vp8/common/x86/loopfilter_mmx.asm ('k') | source/libvpx/vp8/common/x86/loopfilter_x86.h » ('j') | no next file with comments »
Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
Index: source/libvpx/vp8/common/x86/loopfilter_sse2.asm
===================================================================
--- source/libvpx/vp8/common/x86/loopfilter_sse2.asm (revision 96967)
+++ source/libvpx/vp8/common/x86/loopfilter_sse2.asm (working copy)
@@ -110,7 +110,7 @@
psubusb xmm6, xmm5 ; p1-=p0
por xmm6, xmm4 ; abs(p1 - p0)
- mov rdx, arg(2) ; get flimit
+ mov rdx, arg(2) ; get blimit
movdqa t1, xmm6 ; save to t1
@@ -123,7 +123,7 @@
psubusb xmm1, xmm7
por xmm2, xmm3 ; abs(p1-q1)
- movdqa xmm4, XMMWORD PTR [rdx] ; flimit
+ movdqa xmm7, XMMWORD PTR [rdx] ; blimit
movdqa xmm3, xmm0 ; q0
pand xmm2, [GLOBAL(tfe)] ; set lsb of each byte to zero
@@ -134,13 +134,11 @@
psrlw xmm2, 1 ; abs(p1-q1)/2
psubusb xmm5, xmm3 ; p0-=q0
- paddb xmm4, xmm4 ; flimit*2 (less than 255)
psubusb xmm3, xmm6 ; q0-=p0
por xmm5, xmm3 ; abs(p0 - q0)
paddusb xmm5, xmm5 ; abs(p0-q0)*2
- paddb xmm7, xmm4 ; flimit * 2 + limit (less than 255)
movdqa xmm4, t0 ; hev get abs (q1 - q0)
@@ -150,7 +148,7 @@
movdqa xmm2, XMMWORD PTR [rdx] ; hev
- psubusb xmm5, xmm7 ; abs (p0 - q0) *2 + abs(p1-q1)/2 > flimit * 2 + limit
+ psubusb xmm5, xmm7 ; abs (p0 - q0) *2 + abs(p1-q1)/2 > blimit
psubusb xmm4, xmm2 ; hev
psubusb xmm3, xmm2 ; hev
@@ -278,7 +276,7 @@
;(
; unsigned char *src_ptr,
; int src_pixel_step,
-; const char *flimit,
+; const char *blimit,
; const char *limit,
; const char *thresh,
; int count
@@ -288,7 +286,7 @@
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
- SAVE_XMM
+ SAVE_XMM 7
GET_GOT rbx
push rsi
push rdi
@@ -328,7 +326,7 @@
;(
; unsigned char *src_ptr,
; int src_pixel_step,
-; const char *flimit,
+; const char *blimit,
; const char *limit,
; const char *thresh,
; int count
@@ -338,7 +336,7 @@
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
- SAVE_XMM
+ SAVE_XMM 7
GET_GOT rbx
push rsi
push rdi
@@ -574,7 +572,7 @@
;(
; unsigned char *src_ptr,
; int src_pixel_step,
-; const char *flimit,
+; const char *blimit,
; const char *limit,
; const char *thresh,
; int count
@@ -584,7 +582,7 @@
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
- SAVE_XMM
+ SAVE_XMM 7
GET_GOT rbx
push rsi
push rdi
@@ -624,7 +622,7 @@
;(
; unsigned char *u,
; int src_pixel_step,
-; const char *flimit,
+; const char *blimit,
; const char *limit,
; const char *thresh,
; unsigned char *v
@@ -634,7 +632,7 @@
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
- SAVE_XMM
+ SAVE_XMM 7
GET_GOT rbx
push rsi
push rdi
@@ -904,7 +902,7 @@
movdqa xmm4, XMMWORD PTR [rdx]; limit
pmaxub xmm0, xmm7
- mov rdx, arg(2) ; flimit
+ mov rdx, arg(2) ; blimit
psubusb xmm0, xmm4
movdqa xmm5, xmm2 ; q1
@@ -921,12 +919,11 @@
psrlw xmm5, 1 ; abs(p1-q1)/2
psubusb xmm6, xmm3 ; q0-p0
- movdqa xmm2, XMMWORD PTR [rdx]; flimit
+ movdqa xmm4, XMMWORD PTR [rdx]; blimit
mov rdx, arg(4) ; get thresh
por xmm1, xmm6 ; abs(q0-p0)
- paddb xmm2, xmm2 ; flimit*2 (less than 255)
movdqa xmm6, t0 ; get abs (q1 - q0)
@@ -939,10 +936,9 @@
paddusb xmm1, xmm5 ; abs (p0 - q0) *2 + abs(p1-q1)/2
psubusb xmm6, xmm7 ; abs(q1 - q0) > thresh
- paddb xmm4, xmm2 ; flimit * 2 + limit (less than 255)
psubusb xmm3, xmm7 ; abs(p1 - p0)> thresh
- psubusb xmm1, xmm4 ; abs (p0 - q0) *2 + abs(p1-q1)/2 > flimit * 2 + limit
+ psubusb xmm1, xmm4 ; abs (p0 - q0) *2 + abs(p1-q1)/2 > blimit
por xmm6, xmm3 ; abs(q1 - q0) > thresh || abs(p1 - p0) > thresh
por xmm1, xmm0 ; mask
@@ -1014,7 +1010,7 @@
;(
; unsigned char *src_ptr,
; int src_pixel_step,
-; const char *flimit,
+; const char *blimit,
; const char *limit,
; const char *thresh,
; int count
@@ -1024,7 +1020,7 @@
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
- SAVE_XMM
+ SAVE_XMM 7
GET_GOT rbx
push rsi
push rdi
@@ -1081,7 +1077,7 @@
;(
; unsigned char *u,
; int src_pixel_step,
-; const char *flimit,
+; const char *blimit,
; const char *limit,
; const char *thresh,
; unsigned char *v
@@ -1091,7 +1087,7 @@
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
- SAVE_XMM
+ SAVE_XMM 7
GET_GOT rbx
push rsi
push rdi
@@ -1239,7 +1235,7 @@
;(
; unsigned char *src_ptr,
; int src_pixel_step,
-; const char *flimit,
+; const char *blimit,
; const char *limit,
; const char *thresh,
; int count
@@ -1249,7 +1245,7 @@
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
- SAVE_XMM
+ SAVE_XMM 7
GET_GOT rbx
push rsi
push rdi
@@ -1308,7 +1304,7 @@
;(
; unsigned char *u,
; int src_pixel_step,
-; const char *flimit,
+; const char *blimit,
; const char *limit,
; const char *thresh,
; unsigned char *v
@@ -1318,7 +1314,7 @@
push rbp
mov rbp, rsp
SHADOW_ARGS_TO_STACK 6
- SAVE_XMM
+ SAVE_XMM 7
GET_GOT rbx
push rsi
push rdi
@@ -1376,17 +1372,14 @@
;(
; unsigned char *src_ptr,
; int src_pixel_step,
-; const char *flimit,
-; const char *limit,
-; const char *thresh,
-; int count
+; const char *blimit,
;)
global sym(vp8_loop_filter_simple_horizontal_edge_sse2)
sym(vp8_loop_filter_simple_horizontal_edge_sse2):
push rbp
mov rbp, rsp
- SHADOW_ARGS_TO_STACK 6
- SAVE_XMM
+ SHADOW_ARGS_TO_STACK 3
+ SAVE_XMM 7
GET_GOT rbx
push rsi
push rdi
@@ -1394,14 +1387,9 @@
mov rsi, arg(0) ;src_ptr
movsxd rax, dword ptr arg(1) ;src_pixel_step ; destination pitch?
- mov rdx, arg(2) ;flimit ; get flimit
+ mov rdx, arg(2) ;blimit
movdqa xmm3, XMMWORD PTR [rdx]
- mov rdx, arg(3) ;limit
- movdqa xmm7, XMMWORD PTR [rdx]
- paddb xmm3, xmm3 ; flimit*2 (less than 255)
- paddb xmm3, xmm7 ; flimit * 2 + limit (less than 255)
-
mov rdi, rsi ; rdi points to row +1 for indirect addressing
add rdi, rax
neg rax
@@ -1428,7 +1416,7 @@
paddusb xmm5, xmm5 ; abs(p0-q0)*2
paddusb xmm5, xmm1 ; abs (p0 - q0) *2 + abs(p1-q1)/2
- psubusb xmm5, xmm3 ; abs(p0 - q0) *2 + abs(p1-q1)/2 > flimit * 2 + limit
+ psubusb xmm5, xmm3 ; abs(p0 - q0) *2 + abs(p1-q1)/2 > blimit
pxor xmm3, xmm3
pcmpeqb xmm5, xmm3
@@ -1493,17 +1481,14 @@
;(
; unsigned char *src_ptr,
; int src_pixel_step,
-; const char *flimit,
-; const char *limit,
-; const char *thresh,
-; int count
+; const char *blimit,
;)
global sym(vp8_loop_filter_simple_vertical_edge_sse2)
sym(vp8_loop_filter_simple_vertical_edge_sse2):
push rbp ; save old base pointer value.
mov rbp, rsp ; set new base pointer value.
- SHADOW_ARGS_TO_STACK 6
- SAVE_XMM
+ SHADOW_ARGS_TO_STACK 3
+ SAVE_XMM 7
GET_GOT rbx ; save callee-saved reg
push rsi
push rdi
@@ -1607,14 +1592,10 @@
paddusb xmm5, xmm5 ; abs(p0-q0)*2
paddusb xmm5, xmm6 ; abs (p0 - q0) *2 + abs(p1-q1)/2
- mov rdx, arg(2) ;flimit
+ mov rdx, arg(2) ;blimit
movdqa xmm7, XMMWORD PTR [rdx]
- mov rdx, arg(3) ; get limit
- movdqa xmm6, XMMWORD PTR [rdx]
- paddb xmm7, xmm7 ; flimit*2 (less than 255)
- paddb xmm7, xmm6 ; flimit * 2 + limit (less than 255)
- psubusb xmm5, xmm7 ; abs(p0 - q0) *2 + abs(p1-q1)/2 > flimit * 2 + limit
+ psubusb xmm5, xmm7 ; abs(p0 - q0) *2 + abs(p1-q1)/2 > blimit
pxor xmm7, xmm7
pcmpeqb xmm5, xmm7 ; mm5 = mask
« no previous file with comments | « source/libvpx/vp8/common/x86/loopfilter_mmx.asm ('k') | source/libvpx/vp8/common/x86/loopfilter_x86.h » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld 408576698