archive/Cemu
1
0
Fork 0
mirror of https://github.com/cemu-project/Cemu.git synced 2025-07-06 23:11:18 +12:00
Code Issues Projects Releases Packages Wiki Activity Actions

ih264d: Process arguments correctly on Apple Silicon

Browse source 
Apple seemed to not have offsets for arguments on the stack
Either that or the offsets were just wrong, I'll test on a Linux VM and remove the conditonal if this still happens
This commit is contained in:
Exverge 2024-07-18 22:59:56 -04:00
parent 10bed1abc8
commit 804aa4f345
No known key found for this signature in database
GPG key ID: 19AAFC0AC6A9B35A
1 changed files with 23 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

31
dependencies/ih264d/common/armv8/ih264_weighted_bi_pred_av8.s vendored
View file

@ -142,14 +142,22 @@ ih264_weighted_bi_pred_luma_av8:
sxtw x4, w4
sxtw x5, w5
stp x19, x20, [sp, #-16]!
#ifndef __APPLE__
ldr w8, [sp, #80] //Load wt2 in w8
ldr w9, [sp, #88] //Load ofst1 in w9
add w6, w6, #1 //w6 = log_WD + 1
neg w10, w6 //w10 = -(log_WD + 1)
dup v0.8h, w10 //Q0 = -(log_WD + 1) (32-bit)
ldr w10, [sp, #96] //Load ofst2 in w10
ldr w11, [sp, #104] //Load ht in w11
ldr w12, [sp, #112] //Load wd in w12
#else
ldr w8, [sp, #80] //Load wd in w12
ldr w9, [sp, #84] //Load ht in w11
ldr w10, [sp, #88] //Load offst2 in w10
ldr w11, [sp, #92] //Load offst1 in w9
ldr w12, [sp, #96] //Load offst1 in w9
#endif
add w6, w6, #1 //w6 = log_WD + 1
neg w10, w6 //w10 = -(log_WD + 1)
dup v0.8h, w10 //Q0 = -(log_WD + 1) (32-bit)
add w9, w9, #1 //w9 = ofst1 + 1
add w9, w9, w10 //w9 = ofst1 + ofst2 + 1
mov v2.s[0], w7
@ -424,17 +432,24 @@ ih264_weighted_bi_pred_chroma_av8:
sxtw x5, w5
stp x19, x20, [sp, #-16]!
#ifndef __APPLE__
ldr w8, [sp, #80] //Load wt2 in w8
ldr w9, [sp, #88] //Load ofst1 in w9
ldr w10, [sp, #96] //Load ofst2 in w10
ldr w11, [sp, #104] //Load ht in w11
ldr w12, [sp, #112] //Load wd in w12
#else
ldr w8, [sp, #80] //Load wd in w12
ldr w9, [sp, #84] //Load ht in w11
ldr w10, [sp, #88] //Load offst2 in w10
ldr w11, [sp, #92] //Load offst1 in w9
ldr w12, [sp, #96] //Load offst1 in w9
#endif
dup v4.4s, w8 //Q2 = (wt2_u, wt2_v) (32-bit)
dup v2.4s, w7 //Q1 = (wt1_u, wt1_v) (32-bit)
add w6, w6, #1 //w6 = log_WD + 1
ldr w9, [sp, #88] //Load ofst1 in w9
ldr w10, [sp, #96] //Load ofst2 in w10
neg w20, w6 //w20 = -(log_WD + 1)
dup v0.8h, w20 //Q0 = -(log_WD + 1) (16-bit)
ldr w11, [sp, #104] //Load ht in x11
ldr w12, [sp, #112] //Load wd in x12
dup v20.8h, w9 //0ffset1
dup v21.8h, w10 //0ffset2
srhadd v6.8b, v20.8b, v21.8b