RepoMirrors/ffmpeg
1
0
Fork 0
mirror of https://git.ffmpeg.org/ffmpeg.git synced 2025-02-09 16:19:35 +00:00
Code Issues Projects Releases Wiki Activity

x86: hevc_mc: fix register count usage

Browse Source 
A macro was using a fixed register, causing too many GPRs to be
declared as used.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This commit is contained in:
Christophe Gisquet 2014-07-28 11:48:21 +02:00 committed by Michael Niedermayer
parent d38edeee9b
commit a507623bad
1 changed files with 12 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
libavcodec/x86/hevc_mc.asm
View File

@ -270,12 +270,12 @@ QPEL_TABLE 12, 4, w, sse4
%endif %endif
%endmacro %endmacro
%macro QPEL_V_LOAD 4 %macro QPEL_V_LOAD 5
lea r12q, [%2] lea %5q, [%2]
sub r12q, r3srcq sub %5q, r3srcq
movdqu m0, [r12 ] ;load x- 3*srcstride movdqu m0, [%5q ] ;load x- 3*srcstride
movdqu m1, [r12+ %3q ] ;load x- 2*srcstride movdqu m1, [%5q+ %3q ] ;load x- 2*srcstride
movdqu m2, [r12+ 2*%3q ] ;load x-srcstride movdqu m2, [%5q+ 2*%3q ] ;load x-srcstride
movdqu m3, [%2 ] ;load x movdqu m3, [%2 ] ;load x
movdqu m4, [%2+ %3q] ;load x+stride movdqu m4, [%2+ %3q] ;load x+stride
movdqu m5, [%2+ 2*%3q] ;load x+2*stride movdqu m5, [%2+ 2*%3q] ;load x+2*stride
@ -869,11 +869,11 @@ cglobal hevc_put_hevc_bi_qpel_h%1_%2, 8, 9, 16 , dst, dststride, src, srcstride,
; int width, int height, int mx, int my) ; int width, int height, int mx, int my)
; ****************************** ; ******************************
cglobal hevc_put_hevc_qpel_v%1_%2, 7, 14, 15 , dst, dststride, src, srcstride, height, r3src, my, rfilter cglobal hevc_put_hevc_qpel_v%1_%2, 7, 9, 15, dst, dststride, src, srcstride, height, r3src, my, rfilter
lea r3srcq, [srcstrideq*3] lea r3srcq, [srcstrideq*3]
QPEL_FILTER %2, my QPEL_FILTER %2, my
.loop .loop
QPEL_V_LOAD %2, srcq, srcstride, %1 QPEL_V_LOAD %2, srcq, srcstride, %1, r8
QPEL_COMPUTE %1, %2 QPEL_COMPUTE %1, %2
%if %2 > 8 %if %2 > 8
packssdw m0, m1 packssdw m0, m1
@ -882,12 +882,12 @@ cglobal hevc_put_hevc_qpel_v%1_%2, 7, 14, 15 , dst, dststride, src, srcstride, h
LOOP_END dst, dststride, src, srcstride LOOP_END dst, dststride, src, srcstride
RET RET
cglobal hevc_put_hevc_uni_qpel_v%1_%2, 7, 14, 15 , dst, dststride, src, srcstride, height, r3src, my, rfilter cglobal hevc_put_hevc_uni_qpel_v%1_%2, 7, 9, 15, dst, dststride, src, srcstride, height, r3src, my, rfilter
movdqa m9, [pw_%2] movdqa m9, [pw_%2]
lea r3srcq, [srcstrideq*3] lea r3srcq, [srcstrideq*3]
QPEL_FILTER %2, my QPEL_FILTER %2, my
.loop .loop
QPEL_V_LOAD %2, srcq, srcstride, %1 QPEL_V_LOAD %2, srcq, srcstride, %1, r8
QPEL_COMPUTE %1, %2 QPEL_COMPUTE %1, %2
%if %2 > 8 %if %2 > 8
packusdw m0, m1 packusdw m0, m1
@ -900,13 +900,13 @@ cglobal hevc_put_hevc_uni_qpel_v%1_%2, 7, 14, 15 , dst, dststride, src, srcstrid
jnz .loop ; height loop jnz .loop ; height loop
RET RET
cglobal hevc_put_hevc_bi_qpel_v%1_%2, 9, 14, 16 , dst, dststride, src, srcstride, src2, src2stride, height, r3src, my, rfilter cglobal hevc_put_hevc_bi_qpel_v%1_%2, 9, 11, 16, dst, dststride, src, srcstride, src2, src2stride, height, r3src, my, rfilter
movdqa m9, [pw_bi_%2] movdqa m9, [pw_bi_%2]
lea r3srcq, [srcstrideq*3] lea r3srcq, [srcstrideq*3]
QPEL_FILTER %2, my QPEL_FILTER %2, my
.loop .loop
SIMPLE_BILOAD %1, src2q, m10, m11 SIMPLE_BILOAD %1, src2q, m10, m11
QPEL_V_LOAD %2, srcq, srcstride, %1 QPEL_V_LOAD %2, srcq, srcstride, %1, r10
QPEL_COMPUTE %1, %2 QPEL_COMPUTE %1, %2
%if %2 > 8 %if %2 > 8
packssdw m0, m1 packssdw m0, m1