History

aarch64: Fix ff_hevc_put_hevc_epel_h48_8_neon_i8mm

The first 32 elements of each row were correct, while the last 16 were scrambled. This hasn't been noticed, because the checkasm test erroneously only checked half of the output (for 8 bit functions), and apparently none of the samples as part of "fate-hevc" seem to trigger this specific function.

Martin Storsjö committed 2 years ago

Commit 0c5da7b1 parent 2129d661 file · +9 −5Browse files

Unified · Split

libavcodec/aarch64/hevcdsp_epel_neon.S+9−5
@@ -1572,6 +1572,7 @@ function ff_hevc_put_hevc_epel_h48_8_neon_i8mm, export=1 · 1571 unmodified lines
15721572 xtn2 v22.8h, v26.4s
15731573 xtn v23.4h, v23.4s
15741574 xtn2 v23.8h, v27.4s
1575+ add x7, x0, #64
15751576 st4 {v20.8h, v21.8h, v22.8h, v23.8h}, [x0], x10
15761577 ext v4.16b, v2.16b, v3.16b, #1
15771578 ext v5.16b, v2.16b, v3.16b, #2
@@ -1584,11 +1585,14 @@ function ff_hevc_put_hevc_epel_h48_8_neon_i8mm, export=1 · 6 unmodified lines
15841585 usdot v21.4s, v4.16b, v30.16b
15851586 usdot v22.4s, v5.16b, v30.16b
15861587 usdot v23.4s, v6.16b, v30.16b
1587− xtn v20.4h, v20.4s
1588− xtn2 v20.8h, v22.4s
1589− xtn v21.4h, v21.4s
1590− xtn2 v21.8h, v23.4s
1591− add x7, x0, #64
1588+ zip1 v24.4s, v20.4s, v22.4s
1589+ zip2 v25.4s, v20.4s, v22.4s
1590+ zip1 v26.4s, v21.4s, v23.4s
1591+ zip2 v27.4s, v21.4s, v23.4s
1592+ xtn v20.4h, v24.4s
1593+ xtn2 v20.8h, v25.4s
1594+ xtn v21.4h, v26.4s
1595+ xtn2 v21.8h, v27.4s
15921596 st2 {v20.8h, v21.8h}, [x7]
15931597 b.ne 1b
15941598 ret

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