[FFmpeg-devel] [PATCH v2] avcodec/mips: Fix MMI macro replaces in HEVC Decoder

[Michael Niedermayer]

On Tue, Sep 06, 2022 at 07:34:14PM +0800, Shiyou Yin wrote:
> 
> 
> > 2022年8月18日 20:29，Shiyou Yin <yinshiyou-hf at loongson.cn> 写道：
> > 
> > 
> > 
> >> 2022年8月18日 19:44，戚铁铮 <qitiezheng at 360.cn <mailto:qitiezheng at 360.cn>> 写道：
> >> 
> >> 
> >> At 2022/8/18 PM 7:01, "Qi Tiezheng" <qitiezheng at 360.cn <mailto:qitiezheng at 360.cn>> wrote:
> >> 
> >>> The latest commit of Loongson MMI macro replaces were incorrect.
> >>> It makes a mass of green tints on HEVC videos when playing. I've
> >>> compared it with the older MMI implementation, and found out that
> >>> several lines have been replaced by wrong macros.
> >>> 
> >>> Signed-off-by: Qi Tiezheng <qitiezheng at 360.cn <mailto:qitiezheng at 360.cn>>
> >>> ---
> >>> libavcodec/mips/hevcdsp_mmi.c | 16 ++++++++--------
> >>> 1 file changed, 8 insertions(+), 8 deletions(-)
> >>> 
> >>> diff --git a/libavcodec/mips/hevcdsp_mmi.c b/libavcodec/mips/hevcdsp_mmi.c
> >>> index 0ea88a7c08..1da56d3d87 100644
> >>> --- a/libavcodec/mips/hevcdsp_mmi.c
> >>> +++ b/libavcodec/mips/hevcdsp_mmi.c
> >>> @@ -80,7 +80,7 @@ void ff_hevc_put_hevc_qpel_h##w##_8_mmi(int16_t *dst, const uint8_t *_src, \
> >>> "paddh %[ftmp3], %[ftmp3], %[ftmp4] \n\t" \
> >>> "paddh %[ftmp5], %[ftmp5], %[ftmp6] \n\t" \
> >>> "paddh %[ftmp3], %[ftmp3], %[ftmp5] \n\t" \
> >>> - MMI_ULDC1(%[ftmp3], %[dst], 0x00) \
> >>> + MMI_USDC1(%[ftmp3], %[dst], 0x00) \
> >>> \
> >>> "daddi %[x], %[x], -0x01 \n\t" \
> >>> PTR_ADDIU "%[src], %[src], 0x04 \n\t" \
> >>> @@ -178,7 +178,7 @@ void ff_hevc_put_hevc_qpel_hv##w##_8_mmi(int16_t *dst, const uint8_t *_src,\
> >>> "paddh %[ftmp3], %[ftmp3], %[ftmp4] \n\t" \
> >>> "paddh %[ftmp5], %[ftmp5], %[ftmp6] \n\t" \
> >>> "paddh %[ftmp3], %[ftmp3], %[ftmp5] \n\t" \
> >>> - MMI_ULDC1(%[ftmp3], %[tmp], 0x00) \
> >>> + MMI_USDC1(%[ftmp3], %[tmp], 0x00) \
> >>> \
> >>> "daddi %[x], %[x], -0x01 \n\t" \
> >>> PTR_ADDIU "%[src], %[src], 0x04 \n\t" \
> >>> @@ -690,10 +690,10 @@ void ff_hevc_put_hevc_epel_bi_hv##w##_8_mmi(uint8_t *_dst, \
> >>> \
> >>> "1: \n\t" \
> >>> "2: \n\t" \
> >>> - MMI_ULDC1(%[ftmp3], %[src], 0x00) \
> >>> - MMI_ULDC1(%[ftmp4], %[src], 0x01) \
> >>> - MMI_ULDC1(%[ftmp5], %[src], 0x02) \
> >>> - MMI_ULDC1(%[ftmp6], %[src], 0x03) \
> >>> + MMI_ULWC1(%[ftmp2], %[src], 0x00) \
> >>> + MMI_ULWC1(%[ftmp3], %[src], 0x01) \
> >>> + MMI_ULWC1(%[ftmp4], %[src], 0x02) \
> >>> + MMI_ULWC1(%[ftmp5], %[src], 0x03) \
> >>> "punpcklbh %[ftmp2], %[ftmp2], %[ftmp0] \n\t" \
> >>> "pmullh %[ftmp2], %[ftmp2], %[ftmp1] \n\t" \
> >>> "punpcklbh %[ftmp3], %[ftmp3], %[ftmp0] \n\t" \
> >>> @@ -707,7 +707,7 @@ void ff_hevc_put_hevc_epel_bi_hv##w##_8_mmi(uint8_t *_dst, \
> >>> "paddh %[ftmp2], %[ftmp2], %[ftmp3] \n\t" \
> >>> "paddh %[ftmp4], %[ftmp4], %[ftmp5] \n\t" \
> >>> "paddh %[ftmp2], %[ftmp2], %[ftmp4] \n\t" \
> >>> - MMI_ULDC1(%[ftmp2], %[tmp], 0x00) \
> >>> + MMI_USDC1(%[ftmp2], %[tmp], 0x00) \
> >>> \
> >>> "daddi %[x], %[x], -0x01 \n\t" \
> >>> PTR_ADDIU "%[src], %[src], 0x04 \n\t" \
> >>> @@ -773,7 +773,7 @@ void ff_hevc_put_hevc_epel_bi_hv##w##_8_mmi(uint8_t *_dst, \
> >>> "paddw %[ftmp5], %[ftmp5], %[ftmp6] \n\t" \
> >>> "psraw %[ftmp5], %[ftmp5], %[ftmp0] \n\t" \
> >>> "packsswh %[ftmp3], %[ftmp3], %[ftmp5] \n\t" \
> >>> - MMI_ULDC1(%[ftmp4], %[tmp], 0x02) \
> >>> + MMI_ULDC1(%[ftmp4], %[src2], 0x00) \
> >>> "li %[rtmp0], 0x10 \n\t" \
> >>> "dmtc1 %[rtmp0], %[ftmp8] \n\t" \
> >>> "punpcklhw %[ftmp5], %[ftmp2], %[ftmp3] \n\t" \
> >>> -- 
> >>> 2.25.1
> >> 
> >> Sorry, I must use e-mail client because our e-mail server is Exchange not SMTP.
> >> The patch system seems cannot process UTF-8 Chinese characters correctly.
> >> I try sending it as attachment again.
> >> 
> > Thank you for fixing this bug.
> > LGTM.
> > 
> 
> Hi, Michael
> 
> Could you please help to merge this Fix.

will apply

thx

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Everything should be made as simple as possible, but not simpler.
-- Albert Einstein
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20220908/a64c6645/attachment.sig>