virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2025-08-15 14:13:16 +02:00
Code Issues Releases Activity

x86/sbrdsp: sign extend start and end gprs in ff_sbr_hf_gen_sse

Browse Source 
Tested-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Almer <jamrial@gmail.com>
This commit is contained in:
James Almer
2017-06-29 22:51:04 -03:00
parent de54a37c1d
commit ac8ad8d098
1 changed files with 13 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
libavcodec/x86/sbrdsp.asm
View File

@@ -149,19 +149,19 @@ cglobal sbr_hf_gen, 4,4,8, X_high, X_low, alpha0, alpha1, BW, S, E
; start and end 6th and 7th args on stack ; start and end 6th and 7th args on stack
mov r2d, Sm mov r2d, Sm
mov r3d, Em mov r3d, Em
%define start r2q DEFINE_ARGS X_high, X_low, start, end
%define end r3q
%else %else
; BW does not actually occupy a register, so shift by 1 ; BW does not actually occupy a register, so shift by 1
%define start BWq DEFINE_ARGS X_high, X_low, alpha0, alpha1, start, end
%define end Sq movsxd startq, startd
movsxd endq, endd
%endif %endif
sub start, end ; neg num of loops sub startq, endq ; neg num of loops
lea X_highq, [X_highq + end*2*4] lea X_highq, [X_highq + endq*2*4]
lea X_lowq, [X_lowq + end*2*4 - 2*2*4] lea X_lowq, [X_lowq + endq*2*4 - 2*2*4]
shl start, 3 ; offset from num loops shl startq, 3 ; offset from num loops
mova m0, [X_lowq + start] mova m0, [X_lowq + startq]
shufps m3, m3, q1111 shufps m3, m3, q1111
shufps m4, m4, q1111 shufps m4, m4, q1111
xorps m3, [ps_mask] xorps m3, [ps_mask]
@@ -169,7 +169,7 @@ cglobal sbr_hf_gen, 4,4,8, X_high, X_low, alpha0, alpha1, BW, S, E
shufps m2, m2, q0000 shufps m2, m2, q0000
xorps m4, [ps_mask] xorps m4, [ps_mask]
.loop2: .loop2:
movu m7, [X_lowq + start + 8] ; BbCc movu m7, [X_lowq + startq + 8] ; BbCc
mova m6, m0 mova m6, m0
mova m5, m7 mova m5, m7
shufps m0, m0, q2301 ; aAbB shufps m0, m0, q2301 ; aAbB
@@ -179,12 +179,12 @@ cglobal sbr_hf_gen, 4,4,8, X_high, X_low, alpha0, alpha1, BW, S, E
mulps m6, m2 mulps m6, m2
mulps m5, m1 mulps m5, m1
addps m7, m0 addps m7, m0
mova m0, [X_lowq + start +16] ; CcDd mova m0, [X_lowq + startq + 16] ; CcDd
addps m7, m0 addps m7, m0
addps m6, m5 addps m6, m5
addps m7, m6 addps m7, m6
mova [X_highq + start], m7 mova [X_highq + startq], m7
add start, 16 add startq, 16
jnz .loop2 jnz .loop2
RET RET