Skip to content

Commit

Permalink
ih264d: Process arguments correctly on Apple Silicon
Browse files Browse the repository at this point in the history
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
  • Loading branch information
exverge-0 committed Jul 20, 2024
1 parent 250e77e commit ebb249c
Showing 1 changed file with 23 additions and 8 deletions.
31 changes: 23 additions & 8 deletions dependencies/ih264d/common/armv8/ih264_weighted_bi_pred_av8.s
Original file line number Diff line number Diff line change
Expand Up @@ -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
Expand Down Expand Up @@ -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
Expand Down

0 comments on commit ebb249c

Please sign in to comment.