diff options
| author | Bin Peng <pengbin@visionular.com> | 2024-12-16 10:31:23 +0800 |
|---|---|---|
| committer | Martin Storsjö <martin@martin.st> | 2024-12-17 14:50:29 +0200 |
| commit | decc9e643cc3ac5537f42b465e2637fbefbf41cc (patch) | |
| tree | fa80d2ad358cb9b9c4d024ce1b418e6ff165f0c9 | |
| parent | 7b0bd6c4a7bcb53eacf082e472ff32a81b14edba (diff) | |
| download | ffmpeg-decc9e643cc3ac5537f42b465e2637fbefbf41cc.tar.gz | |
lavc/aarch64: Fix ff_pred8x8_plane_neon_10
Fix test failure on aarch64:
./tests/checkasm/checkasm --test=h264pred 479612
The mismatch between neon and C functions can also be reproduced using the following bitstream and command line.
wget https://streams.videolan.org/ffmpeg/incoming/intra8x8pred_10bit.264
./ffmpeg -cpuflags 0 -threads 1 -i intra8x8pred_10bit.264 -f framemd5 -y md5_ref
./ffmpeg -threads 1 -i intra8x8pred_10bit.264 -f framemd5 -y md5_neon
Signed-off-by: Bin Peng <pengbin@visionular.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
| -rw-r--r-- | libavcodec/aarch64/h264pred_neon.S | 5 |
1 files changed, 2 insertions, 3 deletions
diff --git a/libavcodec/aarch64/h264pred_neon.S b/libavcodec/aarch64/h264pred_neon.S index ea37689f34..168f8191ad 100644 --- a/libavcodec/aarch64/h264pred_neon.S +++ b/libavcodec/aarch64/h264pred_neon.S @@ -595,12 +595,11 @@ function ff_pred8x8_plane_neon_10, export=1 ssubl v2.4s, v2.4h, v3.4h ext v0.16b, v0.16b, v0.16b, #14 mov v0.h[0], wzr - mul v0.8h, v0.8h, v5.h[0] dup v1.4s, v2.s[0] dup v2.4s, v2.s[0] dup v3.8h, v5.h[1] - saddw v1.4s, v1.4s, v0.4h - saddw2 v2.4s, v2.4s, v0.8h + smlal v1.4s, v0.4h, v5.h[0] + smlal2 v2.4s, v0.8h, v5.h[0] mov w3, #8 mvni v4.8h, #0xFC, lsl #8 // 1023 for clipping 1: |