[FFmpeg-cvslog] avcodec/x86/hevc_mc: fix sse register counts
Michael Niedermayer
git at videolan.org
Thu Dec 11 13:49:34 CET 2014
ffmpeg | branch: master | Michael Niedermayer <michaelni at gmx.at> | Wed Dec 10 07:48:52 2014 +0100| [a03f72e7447b98139e37480218ab5d512ac2c683] | committer: Michael Niedermayer
avcodec/x86/hevc_mc: fix sse register counts
These fix failures of --enable-xmm-clobber-test
It would be better to change the code to use fewer registers, but until
someone does the used register count must not be too small
Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=a03f72e7447b98139e37480218ab5d512ac2c683
---
libavcodec/x86/hevc_mc.asm | 28 ++++++++++++++--------------
1 file changed, 14 insertions(+), 14 deletions(-)
diff --git a/libavcodec/x86/hevc_mc.asm b/libavcodec/x86/hevc_mc.asm
index ca444e4..8f9f939 100644
--- a/libavcodec/x86/hevc_mc.asm
+++ b/libavcodec/x86/hevc_mc.asm
@@ -586,7 +586,7 @@ cglobal hevc_put_hevc_bi_pel_pixels%1_%2, 6, 6, 6, dst, dststride, src, srcstrid
%macro HEVC_PUT_HEVC_EPEL 2
-cglobal hevc_put_hevc_epel_h%1_%2, 5, 6, 6, dst, src, srcstride, height, mx, rfilter
+cglobal hevc_put_hevc_epel_h%1_%2, 5, 6, 11, dst, src, srcstride, height, mx, rfilter
%assign %%stride ((%2 + 7)/8)
EPEL_FILTER %2, mx, m4, m5
.loop
@@ -596,7 +596,7 @@ cglobal hevc_put_hevc_epel_h%1_%2, 5, 6, 6, dst, src, srcstride, height, mx, rfi
LOOP_END dst, src, srcstride
RET
-cglobal hevc_put_hevc_uni_epel_h%1_%2, 6, 7, 7, dst, dststride, src, srcstride, height, mx, rfilter
+cglobal hevc_put_hevc_uni_epel_h%1_%2, 6, 7, 11, dst, dststride, src, srcstride, height, mx, rfilter
%assign %%stride ((%2 + 7)/8)
movdqa m6, [pw_%2]
EPEL_FILTER %2, mx, m4, m5
@@ -611,7 +611,7 @@ cglobal hevc_put_hevc_uni_epel_h%1_%2, 6, 7, 7, dst, dststride, src, srcstride,
jnz .loop ; height loop
RET
-cglobal hevc_put_hevc_bi_epel_h%1_%2, 7, 8, 7, dst, dststride, src, srcstride, src2, height, mx, rfilter
+cglobal hevc_put_hevc_bi_epel_h%1_%2, 7, 8, 11, dst, dststride, src, srcstride, src2, height, mx, rfilter
movdqa m6, [pw_bi_%2]
EPEL_FILTER %2, mx, m4, m5
.loop
@@ -634,7 +634,7 @@ cglobal hevc_put_hevc_bi_epel_h%1_%2, 7, 8, 7, dst, dststride, src, srcstride, s
; int16_t* mcbuffer)
; ******************************
-cglobal hevc_put_hevc_epel_v%1_%2, 6, 7, 6, dst, src, srcstride, height, r3src, my, rfilter
+cglobal hevc_put_hevc_epel_v%1_%2, 6, 7, 11, dst, src, srcstride, height, r3src, my, rfilter
lea r3srcq, [srcstrideq*3]
sub srcq, srcstrideq
EPEL_FILTER %2, my, m4, m5
@@ -645,7 +645,7 @@ cglobal hevc_put_hevc_epel_v%1_%2, 6, 7, 6, dst, src, srcstride, height, r3src,
LOOP_END dst, src, srcstride
RET
-cglobal hevc_put_hevc_uni_epel_v%1_%2, 7, 8, 7, dst, dststride, src, srcstride, height, r3src, my, rfilter
+cglobal hevc_put_hevc_uni_epel_v%1_%2, 7, 8, 11, dst, dststride, src, srcstride, height, r3src, my, rfilter
lea r3srcq, [srcstrideq*3]
movdqa m6, [pw_%2]
sub srcq, srcstrideq
@@ -662,7 +662,7 @@ cglobal hevc_put_hevc_uni_epel_v%1_%2, 7, 8, 7, dst, dststride, src, srcstride,
RET
-cglobal hevc_put_hevc_bi_epel_v%1_%2, 8, 9, 7, dst, dststride, src, srcstride, src2, height, r3src, my, rfilter
+cglobal hevc_put_hevc_bi_epel_v%1_%2, 8, 9, 11, dst, dststride, src, srcstride, src2, height, r3src, my, rfilter
lea r3srcq, [srcstrideq*3]
movdqa m6, [pw_bi_%2]
sub srcq, srcstrideq
@@ -689,7 +689,7 @@ cglobal hevc_put_hevc_bi_epel_v%1_%2, 8, 9, 7, dst, dststride, src, srcstride, s
; ******************************
%macro HEVC_PUT_HEVC_EPEL_HV 2
-cglobal hevc_put_hevc_epel_hv%1_%2, 6, 8, 12 , dst, src, srcstride, height, mx, my, r3src, rfilter
+cglobal hevc_put_hevc_epel_hv%1_%2, 6, 8, 16 , dst, src, srcstride, height, mx, my, r3src, rfilter
%assign %%stride ((%2 + 7)/8)
sub srcq, srcstrideq
EPEL_HV_FILTER %2
@@ -723,7 +723,7 @@ cglobal hevc_put_hevc_epel_hv%1_%2, 6, 8, 12 , dst, src, srcstride, height, mx,
LOOP_END dst, src, srcstride
RET
-cglobal hevc_put_hevc_uni_epel_hv%1_%2, 7, 9, 12 , dst, dststride, src, srcstride, height, mx, my, r3src, rfilter
+cglobal hevc_put_hevc_uni_epel_hv%1_%2, 7, 9, 16 , dst, dststride, src, srcstride, height, mx, my, r3src, rfilter
%assign %%stride ((%2 + 7)/8)
sub srcq, srcstrideq
EPEL_HV_FILTER %2
@@ -810,7 +810,7 @@ cglobal hevc_put_hevc_bi_epel_hv%1_%2, 8, 10, 16, dst, dststride, src, srcstride
; ******************************
%macro HEVC_PUT_HEVC_QPEL 2
-cglobal hevc_put_hevc_qpel_h%1_%2, 5, 6, 15, dst, src, srcstride, height, mx, rfilter
+cglobal hevc_put_hevc_qpel_h%1_%2, 5, 6, 16, dst, src, srcstride, height, mx, rfilter
QPEL_FILTER %2, mx
.loop
QPEL_H_LOAD %2, srcq, %1, 10
@@ -822,7 +822,7 @@ cglobal hevc_put_hevc_qpel_h%1_%2, 5, 6, 15, dst, src, srcstride, height, mx, rf
LOOP_END dst, src, srcstride
RET
-cglobal hevc_put_hevc_uni_qpel_h%1_%2, 6, 7, 15 , dst, dststride, src, srcstride, height, mx, rfilter
+cglobal hevc_put_hevc_uni_qpel_h%1_%2, 6, 7, 16 , dst, dststride, src, srcstride, height, mx, rfilter
movdqa m9, [pw_%2]
QPEL_FILTER %2, mx
.loop
@@ -865,7 +865,7 @@ cglobal hevc_put_hevc_bi_qpel_h%1_%2, 7, 8, 16 , dst, dststride, src, srcstride,
; int width, int height, int mx, int my)
; ******************************
-cglobal hevc_put_hevc_qpel_v%1_%2, 6, 8, 15, dst, src, srcstride, height, r3src, my, rfilter
+cglobal hevc_put_hevc_qpel_v%1_%2, 6, 8, 16, dst, src, srcstride, height, r3src, my, rfilter
lea r3srcq, [srcstrideq*3]
QPEL_FILTER %2, my
.loop
@@ -878,7 +878,7 @@ cglobal hevc_put_hevc_qpel_v%1_%2, 6, 8, 15, dst, src, srcstride, height, r3src,
LOOP_END dst, src, srcstride
RET
-cglobal hevc_put_hevc_uni_qpel_v%1_%2, 7, 9, 15, dst, dststride, src, srcstride, height, r3src, my, rfilter
+cglobal hevc_put_hevc_uni_qpel_v%1_%2, 7, 9, 16, dst, dststride, src, srcstride, height, r3src, my, rfilter
movdqa m9, [pw_%2]
lea r3srcq, [srcstrideq*3]
QPEL_FILTER %2, my
@@ -924,7 +924,7 @@ cglobal hevc_put_hevc_bi_qpel_v%1_%2, 8, 10, 16, dst, dststride, src, srcstride,
; int height, int mx, int my)
; ******************************
%macro HEVC_PUT_HEVC_QPEL_HV 2
-cglobal hevc_put_hevc_qpel_hv%1_%2, 6, 8, 12, dst, src, srcstride, height, mx, my, r3src, rfilter
+cglobal hevc_put_hevc_qpel_hv%1_%2, 6, 8, 16, dst, src, srcstride, height, mx, my, r3src, rfilter
lea mxq, [mxq*8-8]
lea myq, [myq*8-8]
lea r3srcq, [srcstrideq*3]
@@ -993,7 +993,7 @@ cglobal hevc_put_hevc_qpel_hv%1_%2, 6, 8, 12, dst, src, srcstride, height, mx, m
LOOP_END dst, src, srcstride
RET
-cglobal hevc_put_hevc_uni_qpel_hv%1_%2, 7, 9, 12 , dst, dststride, src, srcstride, height, mx, my, r3src, rfilter
+cglobal hevc_put_hevc_uni_qpel_hv%1_%2, 7, 9, 16 , dst, dststride, src, srcstride, height, mx, my, r3src, rfilter
lea mxq, [mxq*8-8]
lea myq, [myq*8-8]
lea r3srcq, [srcstrideq*3]
More information about the ffmpeg-cvslog
mailing list