History

lavc/h264idct: fix RISC-V group multiplier

After the branch, the expected SEW/LMUL ratio is 1 byte/vector. So we have to set the same ratio before branching (QEMU does not care, but real hardware does).

Rémi Denis-Courmont committed 1 year ago

Commit bbb0fde1 parent 8a5b74f1 file · +4 −4Browse files

Unified · Split

libavcodec/riscv/h264idct_rvv.S+4−4
@@ -644,7 +644,7 @@ func ff_h264_idct4_add\type\()_\depth\()_rvv, zve32x · 643 unmodified lines
644644 sd a4, 24(sp)
645645 ld a0, 0(a0) # dest[0]
646646 addi a1, a1, 16 * 4 # &block_offset[16]
647− vsetivli zero, 4, e8, mf4, ta, ma
647+ vsetivli zero, 4, e8, m1, ta, ma
648648 jal .Lidct4_add4_\depth\()_rvv
649649
650650 ld a4, 24(sp) # nnzc
@@ -655,7 +655,7 @@ func ff_h264_idct4_add\type\()_\depth\()_rvv, zve32x · 4 unmodified lines
655655 ld a0, 8(a0) # dest[1]
656656 lla t0, ff_h264_scan8 + 32
657657 .ifc \type, 8_422
658− vsetivli zero, 4, e8, mf4, ta, ma
658+ vsetivli zero, 4, e8, m1, ta, ma
659659 jal .Lidct4_add4_\depth\()_rvv
660660
661661 ld a4, 24(sp) # nnzc
@@ -665,7 +665,7 @@ func ff_h264_idct4_add\type\()_\depth\()_rvv, zve32x · 3 unmodified lines
665665 addi a1, t5, (-8 - 4) * 4 # &block_offset[24]
666666 ld a0, 0(a0) # dest[0]
667667 lla t0, ff_h264_scan8 + 24
668− vsetivli zero, 4, e8, mf4, ta, ma
668+ vsetivli zero, 4, e8, m1, ta, ma
669669 jal .Lidct4_add4_\depth\()_rvv
670670
671671 ld a4, 24(sp) # nnzc
@@ -679,7 +679,7 @@ func ff_h264_idct4_add\type\()_\depth\()_rvv, zve32x · 7 unmodified lines
679679 ld ra, 8(sp)
680680 ld s0, 0(sp)
681681 addi sp, sp, 32
682− vsetivli zero, 4, e8, mf4, ta, ma
682+ vsetivli zero, 4, e8, m1, ta, ma
683683 j .Lidct4_add4_\depth\()_rvv
684684 endfunc
685685 .endm

Keyboard shortcuts

General

Show keyboard shortcuts
?
Focus the current search
/

Issues and pull requests

Open filters
f

Issues, pull requests and Inbox

Select next item
j↓
Select previous item
k↑
Open selected item
Enter
Dismiss filters or clear selection
Esc

Repository code

Go to file
t