diff mbox series

[FFmpeg-devel] avcodec/mips: Support old style mmi instruction mnemonics

Message ID 20210729063235.20602-1-jiaxun.yang@flygoat.com
State New
Headers show
Series [FFmpeg-devel] avcodec/mips: Support old style mmi instruction mnemonics | expand

Checks

Context Check Description
andriy/x86_make success Make finished
andriy/x86_make_fate success Make fate finished
andriy/PPC64_make success Make finished
andriy/PPC64_make_fate success Make fate finished

Commit Message

Jiaxun Yang July 29, 2021, 6:32 a.m. UTC 
Loongson had renamed serval instruction mnemonics to distinguish
integral and simd opreations. In ebedd26 ("libavcodec/mips: Fix
specification of instruction name"), all old style mnemonics are
replaced by the new one.
However, upstream GCC doesn't support new style mnemonics, it breaks
build.
As stated by Loongson, their GCC toolchain support both style while
clang only support the new one, we'll use new style for clang only.

Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
---
 libavcodec/mips/blockdsp_mmi.c    |   8 +-
 libavcodec/mips/h264chroma_mmi.c  |  20 +--
 libavcodec/mips/h264dsp_mmi.c     | 288 +++++++++++++++---------------
 libavcodec/mips/h264pred_mmi.c    |  18 +-
 libavcodec/mips/h264qpel_mmi.c    |  26 +--
 libavcodec/mips/hevcdsp_mmi.c     |  32 ++--
 libavcodec/mips/hpeldsp_mmi.c     |  26 +--
 libavcodec/mips/idctdsp_mmi.c     |   2 +-
 libavcodec/mips/mpegvideo_mmi.c   |  94 +++++-----
 libavcodec/mips/pixblockdsp_mmi.c |   8 +-
 libavcodec/mips/simple_idct_mmi.c |  14 +-
 libavcodec/mips/vc1dsp_mmi.c      |  34 ++--
 libavcodec/mips/vp3dsp_idct_mmi.c | 132 +++++++-------
 libavcodec/mips/vp8dsp_mmi.c      |  80 ++++-----
 libavcodec/mips/vp9_mc_mmi.c      |  10 +-
 libavcodec/mips/wmv2dsp_mmi.c     |   2 +-
 libavutil/mips/mmiutils.h         |  22 +++
 17 files changed, 419 insertions(+), 397 deletions(-)

Comments

Shiyou Yin July 29, 2021, 9:29 a.m. UTC | #1 
> -----原始邮件-----
> 发件人: "Jiaxun Yang" <jiaxun.yang@flygoat.com>
> 发送时间: 2021-07-29 14:32:35 (星期四)
> 收件人: ffmpeg-devel@ffmpeg.org
> 抄送: yinshiyou-hf@loongson.cn, "Jiaxun Yang" <jiaxun.yang@flygoat.com>
> 主题: [PATCH] avcodec/mips: Support old style mmi instruction mnemonics
> 
> Loongson had renamed serval instruction mnemonics to distinguish
> integral and simd opreations. In ebedd26 ("libavcodec/mips: Fix
> specification of instruction name"), all old style mnemonics are
> replaced by the new one.
> However, upstream GCC doesn't support new style mnemonics, it breaks
> build.
> As stated by Loongson, their GCC toolchain support both style while
> clang only support the new one, we'll use new style for clang only.
> 
> Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
> ---
>  libavcodec/mips/blockdsp_mmi.c    |   8 +-
>  libavcodec/mips/h264chroma_mmi.c  |  20 +--
>  libavcodec/mips/h264dsp_mmi.c     | 288 +++++++++++++++---------------
>  libavcodec/mips/h264pred_mmi.c    |  18 +-
>  libavcodec/mips/h264qpel_mmi.c    |  26 +--
>  libavcodec/mips/hevcdsp_mmi.c     |  32 ++--
>  libavcodec/mips/hpeldsp_mmi.c     |  26 +--
>  libavcodec/mips/idctdsp_mmi.c     |   2 +-
>  libavcodec/mips/mpegvideo_mmi.c   |  94 +++++-----
>  libavcodec/mips/pixblockdsp_mmi.c |   8 +-
>  libavcodec/mips/simple_idct_mmi.c |  14 +-
>  libavcodec/mips/vc1dsp_mmi.c      |  34 ++--
>  libavcodec/mips/vp3dsp_idct_mmi.c | 132 +++++++-------
>  libavcodec/mips/vp8dsp_mmi.c      |  80 ++++-----
>  libavcodec/mips/vp9_mc_mmi.c      |  10 +-
>  libavcodec/mips/wmv2dsp_mmi.c     |   2 +-
>  libavutil/mips/mmiutils.h         |  22 +++
>  17 files changed, 419 insertions(+), 397 deletions(-)


> 

I suggest adding 'pxor' check in configure

and adding 'pxor' support in upstream gcc/clang.
Jiaxun Yang July 29, 2021, 12:26 p.m. UTC | #2 
在2021年7月29日七月 下午5:29，yinshiyou-hf@loongson.cn写道：
> > -----原始邮件-----
> > 发件人: "Jiaxun Yang" <jiaxun.yang@flygoat.com>
> > 发送时间: 2021-07-29 14:32:35 (星期四)
> > 收件人: ffmpeg-devel@ffmpeg.org
> > 抄送: yinshiyou-hf@loongson.cn, "Jiaxun Yang" <jiaxun.yang@flygoat.com>
> > 主题: [PATCH] avcodec/mips: Support old style mmi instruction mnemonics
> > 
> > Loongson had renamed serval instruction mnemonics to distinguish
> > integral and simd opreations. In ebedd26 ("libavcodec/mips: Fix
> > specification of instruction name"), all old style mnemonics are
> > replaced by the new one.
> > However, upstream GCC doesn't support new style mnemonics, it breaks
> > build.
> > As stated by Loongson, their GCC toolchain support both style while
> > clang only support the new one, we'll use new style for clang only.
> > 
> > Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
> > ---
> >  libavcodec/mips/blockdsp_mmi.c    |   8 +-
> >  libavcodec/mips/h264chroma_mmi.c  |  20 +--
> >  libavcodec/mips/h264dsp_mmi.c     | 288 +++++++++++++++---------------
> >  libavcodec/mips/h264pred_mmi.c    |  18 +-
> >  libavcodec/mips/h264qpel_mmi.c    |  26 +--
> >  libavcodec/mips/hevcdsp_mmi.c     |  32 ++--
> >  libavcodec/mips/hpeldsp_mmi.c     |  26 +--
> >  libavcodec/mips/idctdsp_mmi.c     |   2 +-
> >  libavcodec/mips/mpegvideo_mmi.c   |  94 +++++-----
> >  libavcodec/mips/pixblockdsp_mmi.c |   8 +-
> >  libavcodec/mips/simple_idct_mmi.c |  14 +-
> >  libavcodec/mips/vc1dsp_mmi.c      |  34 ++--
> >  libavcodec/mips/vp3dsp_idct_mmi.c | 132 +++++++-------
> >  libavcodec/mips/vp8dsp_mmi.c      |  80 ++++-----
> >  libavcodec/mips/vp9_mc_mmi.c      |  10 +-
> >  libavcodec/mips/wmv2dsp_mmi.c     |   2 +-
> >  libavutil/mips/mmiutils.h         |  22 +++
> >  17 files changed, 419 insertions(+), 397 deletions(-)
> 
> 
> > 
> 
> I suggest adding 'pxor' check in configure

Hmm, I don't think it's worthy to hava a configure option for this.

> 
> and adding 'pxor' support in upstream gcc/clang.

For clang I don't think it's possible as we don't have loongson mmi support in upstream codebase.

For GCC I'd like to have a try but I need a full list of instructions needs to be converted as I asked in another thread.

Also given that we've already have many users with older toolchain and it's impossible to let them upgrade at once, this method could address the issue without trouble present users.

Thanks.

> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> 
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
>
Jiaxun Yang Aug. 2, 2021, 5:40 a.m. UTC | #3 
在 2021/7/29 20:26, Jiaxun Yang 写道:
>
> 在2021年7月29日七月 下午5:29，yinshiyou-hf@loongson.cn写道：
>>> -----原始邮件-----
>>> 发件人: "Jiaxun Yang" <jiaxun.yang@flygoat.com>
>>> 发送时间: 2021-07-29 14:32:35 (星期四)
>>> 收件人: ffmpeg-devel@ffmpeg.org
>>> 抄送: yinshiyou-hf@loongson.cn, "Jiaxun Yang" <jiaxun.yang@flygoat.com>
>>> 主题: [PATCH] avcodec/mips: Support old style mmi instruction mnemonics
>>>
>>> Loongson had renamed serval instruction mnemonics to distinguish
>>> integral and simd opreations. In ebedd26 ("libavcodec/mips: Fix
>>> specification of instruction name"), all old style mnemonics are
>>> replaced by the new one.
>>> However, upstream GCC doesn't support new style mnemonics, it breaks
>>> build.
>>> As stated by Loongson, their GCC toolchain support both style while
>>> clang only support the new one, we'll use new style for clang only.
>>>
>>> Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
>>> ---
>>>   libavcodec/mips/blockdsp_mmi.c    |   8 +-
>>>   libavcodec/mips/h264chroma_mmi.c  |  20 +--
>>>   libavcodec/mips/h264dsp_mmi.c     | 288 +++++++++++++++---------------
>>>   libavcodec/mips/h264pred_mmi.c    |  18 +-
>>>   libavcodec/mips/h264qpel_mmi.c    |  26 +--
>>>   libavcodec/mips/hevcdsp_mmi.c     |  32 ++--
>>>   libavcodec/mips/hpeldsp_mmi.c     |  26 +--
>>>   libavcodec/mips/idctdsp_mmi.c     |   2 +-
>>>   libavcodec/mips/mpegvideo_mmi.c   |  94 +++++-----
>>>   libavcodec/mips/pixblockdsp_mmi.c |   8 +-
>>>   libavcodec/mips/simple_idct_mmi.c |  14 +-
>>>   libavcodec/mips/vc1dsp_mmi.c      |  34 ++--
>>>   libavcodec/mips/vp3dsp_idct_mmi.c | 132 +++++++-------
>>>   libavcodec/mips/vp8dsp_mmi.c      |  80 ++++-----
>>>   libavcodec/mips/vp9_mc_mmi.c      |  10 +-
>>>   libavcodec/mips/wmv2dsp_mmi.c     |   2 +-
>>>   libavutil/mips/mmiutils.h         |  22 +++
>>>   17 files changed, 419 insertions(+), 397 deletions(-)
>>
>> I suggest adding 'pxor' check in configure
> Hmm, I don't think it's worthy to hava a configure option for this.
>
>> and adding 'pxor' support in upstream gcc/clang.


Are you going to input anything?

If not I may request TC comment next week.

Downstream users are complaining about the breakage.

Thanks.

- Jiaxun


> For clang I don't think it's possible as we don't have loongson mmi support in upstream codebase.
>
> For GCC I'd like to have a try but I need a full list of instructions needs to be converted as I asked in another thread.
>
> Also given that we've already have many users with older toolchain and it's impossible to let them upgrade at once, this method could address the issue without trouble present users.
>
> Thanks.
>
>> _______________________________________________
>> ffmpeg-devel mailing list
>> ffmpeg-devel@ffmpeg.org
>> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>>
>> To unsubscribe, visit link above, or email
>> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
>>
>
Shiyou Yin Aug. 2, 2021, 8:33 a.m. UTC | #4 
> -----原始邮件-----
> 发件人: "Jiaxun Yang" <jiaxun.yang@flygoat.com>
> 发送时间: 2021-08-02 13:40:54 (星期一)
> 收件人: "FFmpeg development discussions and patches" <ffmpeg-devel@ffmpeg.org>, yinshiyou-hf@loongson.cn
> 抄送: 
> 主题: Re: [FFmpeg-devel] [PATCH] avcodec/mips: Support old style mmi instruction mnemonics
> 
> 
> 在 2021/7/29 20:26, Jiaxun Yang 写道:
> >
> > 在2021年7月29日七月 下午5:29，yinshiyou-hf@loongson.cn写道：
> >>> -----原始邮件-----
> >>> 发件人: "Jiaxun Yang" <jiaxun.yang@flygoat.com>
> >>> 发送时间: 2021-07-29 14:32:35 (星期四)
> >>> 收件人: ffmpeg-devel@ffmpeg.org
> >>> 抄送: yinshiyou-hf@loongson.cn, "Jiaxun Yang" <jiaxun.yang@flygoat.com>
> >>> 主题: [PATCH] avcodec/mips: Support old style mmi instruction mnemonics
> >>>
> >>> Loongson had renamed serval instruction mnemonics to distinguish
> >>> integral and simd opreations. In ebedd26 ("libavcodec/mips: Fix
> >>> specification of instruction name"), all old style mnemonics are
> >>> replaced by the new one.
> >>> However, upstream GCC doesn't support new style mnemonics, it breaks
> >>> build.
> >>> As stated by Loongson, their GCC toolchain support both style while
> >>> clang only support the new one, we'll use new style for clang only.
> >>>
> >>> Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
> >>> ---
> >>>   libavcodec/mips/blockdsp_mmi.c    |   8 +-
> >>>   libavcodec/mips/h264chroma_mmi.c  |  20 +--
> >>>   libavcodec/mips/h264dsp_mmi.c     | 288 +++++++++++++++---------------
> >>>   libavcodec/mips/h264pred_mmi.c    |  18 +-
> >>>   libavcodec/mips/h264qpel_mmi.c    |  26 +--
> >>>   libavcodec/mips/hevcdsp_mmi.c     |  32 ++--
> >>>   libavcodec/mips/hpeldsp_mmi.c     |  26 +--
> >>>   libavcodec/mips/idctdsp_mmi.c     |   2 +-
> >>>   libavcodec/mips/mpegvideo_mmi.c   |  94 +++++-----
> >>>   libavcodec/mips/pixblockdsp_mmi.c |   8 +-
> >>>   libavcodec/mips/simple_idct_mmi.c |  14 +-
> >>>   libavcodec/mips/vc1dsp_mmi.c      |  34 ++--
> >>>   libavcodec/mips/vp3dsp_idct_mmi.c | 132 +++++++-------
> >>>   libavcodec/mips/vp8dsp_mmi.c      |  80 ++++-----
> >>>   libavcodec/mips/vp9_mc_mmi.c      |  10 +-
> >>>   libavcodec/mips/wmv2dsp_mmi.c     |   2 +-
> >>>   libavutil/mips/mmiutils.h         |  22 +++
> >>>   17 files changed, 419 insertions(+), 397 deletions(-)
> >>
> >> I suggest adding 'pxor' check in configure
> > Hmm, I don't think it's worthy to hava a configure option for this.
> >
> >> and adding 'pxor' support in upstream gcc/clang.
> 
> 
> Are you going to input anything?
> 
> If not I may request TC comment next week.
> 
> Downstream users are complaining about the breakage.
> 
> Thanks.
> 
> - Jiaxun
> 
> 
> > For clang I don't think it's possible as we don't have loongson mmi support in upstream codebase.
> >
> > For GCC I'd like to have a try but I need a full list of instructions needs to be converted as I asked in another thread.
> >
> > Also given that we've already have many users with older toolchain and it's impossible to let them upgrade at once, this method could address the issue without trouble present users.
> >
> > Thanks.
> >


1. What I mean adding 'pxor' check in configure is just replace 'punpcklhw' with 'pxor'. Refer to jinbo's patch，it can workaround the build failed.

2. It's great that you will try to add support in gcc, it's helpful for MIPS code to become concise.

3. About the full list of instruction, you can CC to me in the thread you have mentioned, I will try to help.
Jiaxun Yang Aug. 7, 2021, 3:01 a.m. UTC | #5 
在 2021/8/2 下午4:33, yinshiyou-hf@loongson.cn 写道:
>> -----原始邮件-----
>> 发件人: "Jiaxun Yang" <jiaxun.yang@flygoat.com>
>> 发送时间: 2021-08-02 13:40:54 (星期一)
>> 收件人: "FFmpeg development discussions and patches" <ffmpeg-devel@ffmpeg.org>, yinshiyou-hf@loongson.cn
>> 抄送:
>> 主题: Re: [FFmpeg-devel] [PATCH] avcodec/mips: Support old style mmi instruction mnemonics
>>
>>
>> 在 2021/7/29 20:26, Jiaxun Yang 写道:
>>> 在2021年7月29日七月 下午5:29，yinshiyou-hf@loongson.cn写道：
>>>>> -----原始邮件-----
>>>>> 发件人: "Jiaxun Yang" <jiaxun.yang@flygoat.com>
>>>>> 发送时间: 2021-07-29 14:32:35 (星期四)
>>>>> 收件人: ffmpeg-devel@ffmpeg.org
>>>>> 抄送: yinshiyou-hf@loongson.cn, "Jiaxun Yang" <jiaxun.yang@flygoat.com>
>>>>> 主题: [PATCH] avcodec/mips: Support old style mmi instruction mnemonics
>>>>>
>>>>> Loongson had renamed serval instruction mnemonics to distinguish
>>>>> integral and simd opreations. In ebedd26 ("libavcodec/mips: Fix
>>>>> specification of instruction name"), all old style mnemonics are
>>>>> replaced by the new one.
>>>>> However, upstream GCC doesn't support new style mnemonics, it breaks
>>>>> build.
>>>>> As stated by Loongson, their GCC toolchain support both style while
>>>>> clang only support the new one, we'll use new style for clang only.
>>>>>
>>>>> Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
>>>>> ---
>>>>>    libavcodec/mips/blockdsp_mmi.c    |   8 +-
>>>>>    libavcodec/mips/h264chroma_mmi.c  |  20 +--
>>>>>    libavcodec/mips/h264dsp_mmi.c     | 288 +++++++++++++++---------------
>>>>>    libavcodec/mips/h264pred_mmi.c    |  18 +-
>>>>>    libavcodec/mips/h264qpel_mmi.c    |  26 +--
>>>>>    libavcodec/mips/hevcdsp_mmi.c     |  32 ++--
>>>>>    libavcodec/mips/hpeldsp_mmi.c     |  26 +--
>>>>>    libavcodec/mips/idctdsp_mmi.c     |   2 +-
>>>>>    libavcodec/mips/mpegvideo_mmi.c   |  94 +++++-----
>>>>>    libavcodec/mips/pixblockdsp_mmi.c |   8 +-
>>>>>    libavcodec/mips/simple_idct_mmi.c |  14 +-
>>>>>    libavcodec/mips/vc1dsp_mmi.c      |  34 ++--
>>>>>    libavcodec/mips/vp3dsp_idct_mmi.c | 132 +++++++-------
>>>>>    libavcodec/mips/vp8dsp_mmi.c      |  80 ++++-----
>>>>>    libavcodec/mips/vp9_mc_mmi.c      |  10 +-
>>>>>    libavcodec/mips/wmv2dsp_mmi.c     |   2 +-
>>>>>    libavutil/mips/mmiutils.h         |  22 +++
>>>>>    17 files changed, 419 insertions(+), 397 deletions(-)
>>>> I suggest adding 'pxor' check in configure
>>> Hmm, I don't think it's worthy to hava a configure option for this.
>>>
>>>> and adding 'pxor' support in upstream gcc/clang.
>>
>> Are you going to input anything?
>>
>> If not I may request TC comment next week.
>>
>> Downstream users are complaining about the breakage.
>>
>> Thanks.
>>
>> - Jiaxun
>>
>>
>>> For clang I don't think it's possible as we don't have loongson mmi support in upstream codebase.
>>>
>>> For GCC I'd like to have a try but I need a full list of instructions needs to be converted as I asked in another thread.
>>>
>>> Also given that we've already have many users with older toolchain and it's impossible to let them upgrade at once, this method could address the issue without trouble present users.
>>>
>>> Thanks.
>>>
>
> 1. What I mean adding 'pxor' check in configure is just replace 'punpcklhw' with 'pxor'. Refer to jinbo's patch，it can workaround the build failed.
It means distros using upstream toolchain lost their right to enjoy 
Loongson MMI?
Why we're going to make it happen?

> 2. It's great that you will try to add support in gcc, it's helpful for MIPS code to become concise.
>
> 3. About the full list of instruction, you can CC to me in the thread you have mentioned, I will try to help.

I'm pasting the previous post here:

I'm trying to get that part fixed in upstream toolchain, I wonder that 
beside these four instructions, do we have other instructions renamed?

I have a list of MMI instructions sharing name with regular instructions.

     add
     addu
     dadd
     sub
     subu
     dsub
     or
     sll
     dsll
     xor
     nor
     and
     srl
     dsrl
     sra
     dsra
     sequ
     sltu
     sleu
     seq
     slt
     sle

Could you please help me check your manual?

Thanks.

- Jiaxun

> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
Jiaxun Yang Jan. 28, 2022, 12:55 p.m. UTC | #6 
在2021年8月2日八月 上午9:33，yinshiyou-hf@loongson.cn写道：
>> -----原始邮件-----
>> 发件人: "Jiaxun Yang" <jiaxun.yang@flygoat.com>
>> 发送时间: 2021-08-02 13:40:54 (星期一)
>> 收件人: "FFmpeg development discussions and patches" <ffmpeg-devel@ffmpeg.org>, yinshiyou-hf@loongson.cn
>> 抄送: 
>> 主题: Re: [FFmpeg-devel] [PATCH] avcodec/mips: Support old style mmi instruction mnemonics
>> 
>> 
>> 在 2021/7/29 20:26, Jiaxun Yang 写道:
>> >
>> > 在2021年7月29日七月 下午5:29，yinshiyou-hf@loongson.cn写道：
>> >>> -----原始邮件-----
>> >>> 发件人: "Jiaxun Yang" <jiaxun.yang@flygoat.com>
>> >>> 发送时间: 2021-07-29 14:32:35 (星期四)
>> >>> 收件人: ffmpeg-devel@ffmpeg.org
>> >>> 抄送: yinshiyou-hf@loongson.cn, "Jiaxun Yang" <jiaxun.yang@flygoat.com>
>> >>> 主题: [PATCH] avcodec/mips: Support old style mmi instruction mnemonics
>> >>>
>> >>> Loongson had renamed serval instruction mnemonics to distinguish
>> >>> integral and simd opreations. In ebedd26 ("libavcodec/mips: Fix
>> >>> specification of instruction name"), all old style mnemonics are
>> >>> replaced by the new one.
>> >>> However, upstream GCC doesn't support new style mnemonics, it breaks
>> >>> build.
>> >>> As stated by Loongson, their GCC toolchain support both style while
>> >>> clang only support the new one, we'll use new style for clang only.
>> >>>
>> >>> Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
>> >>> ---
>> >>>   libavcodec/mips/blockdsp_mmi.c    |   8 +-
>> >>>   libavcodec/mips/h264chroma_mmi.c  |  20 +--
>> >>>   libavcodec/mips/h264dsp_mmi.c     | 288 +++++++++++++++---------------
>> >>>   libavcodec/mips/h264pred_mmi.c    |  18 +-
>> >>>   libavcodec/mips/h264qpel_mmi.c    |  26 +--
>> >>>   libavcodec/mips/hevcdsp_mmi.c     |  32 ++--
>> >>>   libavcodec/mips/hpeldsp_mmi.c     |  26 +--
>> >>>   libavcodec/mips/idctdsp_mmi.c     |   2 +-
>> >>>   libavcodec/mips/mpegvideo_mmi.c   |  94 +++++-----
>> >>>   libavcodec/mips/pixblockdsp_mmi.c |   8 +-
>> >>>   libavcodec/mips/simple_idct_mmi.c |  14 +-
>> >>>   libavcodec/mips/vc1dsp_mmi.c      |  34 ++--
>> >>>   libavcodec/mips/vp3dsp_idct_mmi.c | 132 +++++++-------
>> >>>   libavcodec/mips/vp8dsp_mmi.c      |  80 ++++-----
>> >>>   libavcodec/mips/vp9_mc_mmi.c      |  10 +-
>> >>>   libavcodec/mips/wmv2dsp_mmi.c     |   2 +-
>> >>>   libavutil/mips/mmiutils.h         |  22 +++
>> >>>   17 files changed, 419 insertions(+), 397 deletions(-)
>> >>
>> >> I suggest adding 'pxor' check in configure
>> > Hmm, I don't think it's worthy to hava a configure option for this.
>> >
>> >> and adding 'pxor' support in upstream gcc/clang.
>> 
>> 
>> Are you going to input anything?
>> 
>> If not I may request TC comment next week.
>> 
>> Downstream users are complaining about the breakage.
>> 
>> Thanks.
>> 
>> - Jiaxun
>> 
>> 
>> > For clang I don't think it's possible as we don't have loongson mmi support in upstream codebase.
>> >
>> > For GCC I'd like to have a try but I need a full list of instructions needs to be converted as I asked in another thread.
>> >
>> > Also given that we've already have many users with older toolchain and it's impossible to let them upgrade at once, this method could address the issue without trouble present users.
>> >
>> > Thanks.
>> >
>
>
> 1. What I mean adding 'pxor' check in configure is just replace 
> 'punpcklhw' with 'pxor'. Refer to jinbo's patch，it can workaround the 
> build failed.

Sorry but I must revisit this patch, as I recieved much complains from
both downsream distros and users.

Debian complains about that, Loongson-2K users using old Loongnix complained
that, NeoKylin users complained that.

I'll revisit and send v2 soonish.

Thanks.
diff mbox series

Patch

diff --git a/libavcodec/mips/blockdsp_mmi.c b/libavcodec/mips/blockdsp_mmi.c
index 8b5c7e955c..ce3049ae26 100644
--- a/libavcodec/mips/blockdsp_mmi.c
+++ b/libavcodec/mips/blockdsp_mmi.c
@@ -76,8 +76,8 @@  void ff_clear_block_mmi(int16_t *block)
     double ftmp[2];
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp1]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp1]                \n\t"
         MMI_SQC1(%[ftmp0], %[ftmp1], %[block], 0x00)
         MMI_SQC1(%[ftmp0], %[ftmp1], %[block], 0x10)
         MMI_SQC1(%[ftmp0], %[ftmp1], %[block], 0x20)
@@ -97,8 +97,8 @@  void ff_clear_blocks_mmi(int16_t *block)
     double ftmp[2];
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp1]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp1]                \n\t"
         MMI_SQC1(%[ftmp0], %[ftmp1], %[block], 0x00)
         MMI_SQC1(%[ftmp0], %[ftmp1], %[block], 0x10)
         MMI_SQC1(%[ftmp0], %[ftmp1], %[block], 0x20)
diff --git a/libavcodec/mips/h264chroma_mmi.c b/libavcodec/mips/h264chroma_mmi.c
index ec35c5a72e..c68de686b2 100644
--- a/libavcodec/mips/h264chroma_mmi.c
+++ b/libavcodec/mips/h264chroma_mmi.c
@@ -75,7 +75,7 @@  void ff_put_h264_chroma_mc8_mmi(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
         A.i = 64 - D.i - B.i - C.i;
 
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]           \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]           \n\t"
             "pshufh     %[A],       %[A],           %[ftmp0]           \n\t"
             "pshufh     %[B],       %[B],           %[ftmp0]           \n\t"
             "mtc1       %[tmp0],    %[ftmp9]                           \n\t"
@@ -175,7 +175,7 @@  void ff_put_h264_chroma_mc8_mmi(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
         A.i = 64 - E.i;
 
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]           \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]           \n\t"
             "pshufh     %[A],       %[A],           %[ftmp0]           \n\t"
             "pshufh     %[E],       %[E],           %[ftmp0]           \n\t"
             "mtc1       %[tmp0],    %[ftmp7]                           \n\t"
@@ -223,7 +223,7 @@  void ff_put_h264_chroma_mc8_mmi(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
         A.i = 64 - E.i;
 
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]           \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]           \n\t"
             "pshufh     %[A],       %[A],           %[ftmp0]           \n\t"
             "pshufh     %[E],       %[E],           %[ftmp0]           \n\t"
             "mtc1       %[tmp0],    %[ftmp7]                           \n\t"
@@ -333,7 +333,7 @@  void ff_avg_h264_chroma_mc8_mmi(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
         C.i = (y << 3) - D.i;
         A.i = 64 - D.i - B.i - C.i;
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]       \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]       \n\t"
             "pshufh     %[A],       %[A],           %[ftmp0]       \n\t"
             "pshufh     %[B],       %[B],           %[ftmp0]       \n\t"
             "mtc1       %[tmp0],    %[ftmp9]                       \n\t"
@@ -401,7 +401,7 @@  void ff_avg_h264_chroma_mc8_mmi(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
         E.i = x << 3;
         A.i = 64 - E.i;
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]       \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]       \n\t"
             "pshufh     %[A],       %[A],           %[ftmp0]       \n\t"
             "pshufh     %[E],       %[E],           %[ftmp0]       \n\t"
             "mtc1       %[tmp0],    %[ftmp7]                       \n\t"
@@ -450,7 +450,7 @@  void ff_avg_h264_chroma_mc8_mmi(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
         E.i = y << 3;
         A.i = 64 - E.i;
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]       \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]       \n\t"
             "pshufh     %[A],       %[A],           %[ftmp0]       \n\t"
             "pshufh     %[E],       %[E],           %[ftmp0]       \n\t"
             "mtc1       %[tmp0],    %[ftmp7]                       \n\t"
@@ -512,7 +512,7 @@  void ff_put_h264_chroma_mc4_mmi(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
 
     if (D.i) {
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
             "pshufh     %[A],       %[A],           %[ftmp0]            \n\t"
             "pshufh     %[B],       %[B],           %[ftmp0]            \n\t"
             "mtc1       %[tmp0],    %[ftmp7]                            \n\t"
@@ -561,7 +561,7 @@  void ff_put_h264_chroma_mc4_mmi(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
     } else if (E.i) {
         const int step = C.i ? stride : 1;
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
             "pshufh     %[A],       %[A],           %[ftmp0]            \n\t"
             "pshufh     %[E],       %[E],           %[ftmp0]            \n\t"
             "mtc1       %[tmp0],    %[ftmp5]                            \n\t"
@@ -633,7 +633,7 @@  void ff_avg_h264_chroma_mc4_mmi(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
 
     if (D.i) {
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
             "pshufh     %[A],       %[A],           %[ftmp0]            \n\t"
             "pshufh     %[B],       %[B],           %[ftmp0]            \n\t"
             "mtc1       %[tmp0],    %[ftmp7]                            \n\t"
@@ -684,7 +684,7 @@  void ff_avg_h264_chroma_mc4_mmi(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
     } else if (E.i) {
         const int step = C.i ? stride : 1;
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
             "pshufh     %[A],       %[A],           %[ftmp0]            \n\t"
             "pshufh     %[E],       %[E],           %[ftmp0]            \n\t"
             "mtc1       %[tmp0],    %[ftmp5]                            \n\t"
diff --git a/libavcodec/mips/h264dsp_mmi.c b/libavcodec/mips/h264dsp_mmi.c
index b5ab07c863..8d5dae021d 100644
--- a/libavcodec/mips/h264dsp_mmi.c
+++ b/libavcodec/mips/h264dsp_mmi.c
@@ -34,7 +34,7 @@  void ff_h264_add_pixels4_8_mmi(uint8_t *dst, int16_t *src, int stride)
     DECLARE_VAR_LOW32;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         MMI_LDC1(%[ftmp1], %[src], 0x00)
         MMI_LDC1(%[ftmp2], %[src], 0x08)
         MMI_LDC1(%[ftmp3], %[src], 0x10)
@@ -89,7 +89,7 @@  void ff_h264_idct_add_8_mmi(uint8_t *dst, int16_t *block, int stride)
         MMI_LDC1(%[ftmp2], %[block], 0x10)
         MMI_LDC1(%[ftmp3], %[block], 0x18)
         /* memset(block, 0, 32) */
-        "pxor       %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
         MMI_SQC1(%[ftmp4], %[ftmp4], %[block], 0x00)
         MMI_SQC1(%[ftmp4], %[ftmp4], %[block], 0x10)
         "dli        %[tmp0],    0x01                                    \n\t"
@@ -127,7 +127,7 @@  void ff_h264_idct_add_8_mmi(uint8_t *dst, int16_t *block, int stride)
         "psubh      %[ftmp5],   %[ftmp5],       %[ftmp4]                \n\t"
         MMI_ULWC1(%[ftmp2], %[dst], 0x00)
         MMI_LWXC1(%[ftmp0], %[dst], %[stride], 0x00)
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
         "psrah      %[ftmp3],   %[ftmp10],      %[ftmp9]                \n\t"
         "psrah      %[ftmp4],   %[ftmp11],      %[ftmp9]                \n\t"
         "punpcklbh  %[ftmp2],   %[ftmp2],       %[ftmp7]                \n\t"
@@ -419,7 +419,7 @@  void ff_h264_idct8_add_8_mmi(uint8_t *dst, int16_t *block, int stride)
         "psubh      %[ftmp7],   %[ftmp7],       %[ftmp1]                \n\t"
         MMI_SDC1(%[ftmp0], $sp, 0x10)
         "dmfc1      %[tmp1],    %[ftmp2]                                \n\t"
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp2]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp2]                \n\t"
         MMI_SDC1(%[ftmp2], %[block], 0x00)
         MMI_SDC1(%[ftmp2], %[block], 0x08)
         MMI_SDC1(%[ftmp2], %[block], 0x10)
@@ -555,7 +555,7 @@  void ff_h264_idct8_add_8_mmi(uint8_t *dst, int16_t *block, int stride)
         "psubh      %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
         MMI_SDC1(%[ftmp7], $sp, 0x18)
         "dmfc1      %[tmp2],    %[ftmp0]                                \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         MMI_ULWC1(%[ftmp6], %[addr0], 0x00)
         MMI_LWXC1(%[ftmp7], %[addr0], %[stride], 0x00)
         "psrah      %[ftmp2],   %[ftmp2],       %[ftmp10]               \n\t"
@@ -646,7 +646,7 @@  void ff_h264_idct_dc_add_8_mmi(uint8_t *dst, int16_t *block, int stride)
 
     __asm__ volatile (
         "mtc1       %[dc],      %[ftmp5]                                \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "pshufh     %[ftmp5],   %[ftmp5],       %[ftmp0]                \n\t"
         MMI_ULWC1(%[ftmp1], %[dst0], 0x00)
         MMI_ULWC1(%[ftmp2], %[dst1], 0x00)
@@ -690,7 +690,7 @@  void ff_h264_idct8_dc_add_8_mmi(uint8_t *dst, int16_t *block, int stride)
 
     __asm__ volatile (
         "mtc1       %[dc],      %[ftmp5]                                \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "pshufh     %[ftmp5],   %[ftmp5],       %[ftmp0]                \n\t"
         MMI_LDC1(%[ftmp1], %[dst0], 0x00)
         MMI_LDC1(%[ftmp2], %[dst1], 0x00)
@@ -929,7 +929,7 @@  void ff_h264_luma_dc_dequant_idct_8_mmi(int16_t *output, int16_t *input,
         "packsswh   %[ftmp0],   %[ftmp0],       %[ftmp1]                \n\t"
         "packsswh   %[ftmp2],   %[ftmp2],       %[ftmp5]                \n\t"
         "dmfc1      %[tmp1],    %[ftmp0]                                \n\t"
-        "ssrld      %[ftmp0],   %[ftmp0],       %[ftmp9]                \n\t"
+        SSRLD "     %[ftmp0],   %[ftmp0],       %[ftmp9]                \n\t"
         "mfc1       %[input],   %[ftmp0]                                \n\t"
         "sh         %[tmp1],    0x00(%[output])                         \n\t"
         "sh         %[input],   0x80(%[output])                         \n\t"
@@ -938,7 +938,7 @@  void ff_h264_luma_dc_dequant_idct_8_mmi(int16_t *output, int16_t *input,
         "sh         %[tmp1],    0x20(%[output])                         \n\t"
         "sh         %[input],   0xa0(%[output])                         \n\t"
         "dmfc1      %[tmp1],    %[ftmp2]                                \n\t"
-        "ssrld      %[ftmp2],   %[ftmp2],       %[ftmp9]                \n\t"
+        SSRLD "     %[ftmp2],   %[ftmp2],       %[ftmp9]                \n\t"
         "mfc1       %[input],   %[ftmp2]                                \n\t"
         "sh         %[tmp1],    0x40(%[output])                         \n\t"
         "sh         %[input],   0xc0(%[output])                         \n\t"
@@ -963,7 +963,7 @@  void ff_h264_luma_dc_dequant_idct_8_mmi(int16_t *output, int16_t *input,
         "packsswh   %[ftmp3],   %[ftmp3],       %[ftmp1]                \n\t"
         "packsswh   %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
         "dmfc1      %[tmp1],    %[ftmp3]                                \n\t"
-        "ssrld      %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
+        SSRLD "     %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
         "mfc1       %[input],   %[ftmp3]                                \n\t"
         "sh         %[tmp1],    0x100(%[output])                        \n\t"
         "sh         %[input],   0x180(%[output])                        \n\t"
@@ -972,7 +972,7 @@  void ff_h264_luma_dc_dequant_idct_8_mmi(int16_t *output, int16_t *input,
         "sh         %[tmp1],    0x120(%[output])                        \n\t"
         "sh         %[input],   0x1a0(%[output])                        \n\t"
         "dmfc1      %[tmp1],    %[ftmp4]                                \n\t"
-        "ssrld      %[ftmp4],   %[ftmp4],       %[ftmp9]                \n\t"
+        SSRLD "     %[ftmp4],   %[ftmp4],       %[ftmp9]                \n\t"
         "mfc1       %[input],   %[ftmp4]                                \n\t"
         "sh         %[tmp1],    0x140(%[output])                        \n\t"
         "sh         %[input],   0x1c0(%[output])                        \n\t"
@@ -1016,7 +1016,7 @@  void ff_h264_luma_dc_dequant_idct_8_mmi(int16_t *output, int16_t *input,
         "packsswh   %[ftmp0],   %[ftmp0],       %[ftmp1]                \n\t"
         "packsswh   %[ftmp2],   %[ftmp2],       %[ftmp5]                \n\t"
         "dmfc1      %[tmp1],    %[ftmp0]                                \n\t"
-        "ssrld      %[ftmp0],   %[ftmp0],       %[ftmp9]                \n\t"
+        SSRLD "     %[ftmp0],   %[ftmp0],       %[ftmp9]                \n\t"
         "sh         %[tmp1],    0x00(%[output])                         \n\t"
         "mfc1       %[input],   %[ftmp0]                                \n\t"
         "dsrl       %[tmp1],    %[tmp1],        0x10                    \n\t"
@@ -1025,7 +1025,7 @@  void ff_h264_luma_dc_dequant_idct_8_mmi(int16_t *output, int16_t *input,
         PTR_SRL    "%[input],   %[input],       0x10                    \n\t"
         "dmfc1      %[tmp1],    %[ftmp2]                                \n\t"
         "sh         %[input],   0xa0(%[output])                         \n\t"
-        "ssrld      %[ftmp2],   %[ftmp2],       %[ftmp9]                \n\t"
+        SSRLD "     %[ftmp2],   %[ftmp2],       %[ftmp9]                \n\t"
         "sh         %[tmp1],    0x40(%[output])                         \n\t"
         "mfc1       %[input],   %[ftmp2]                                \n\t"
         "dsrl       %[tmp1],    %[tmp1],        0x10                    \n\t"
@@ -1050,7 +1050,7 @@  void ff_h264_luma_dc_dequant_idct_8_mmi(int16_t *output, int16_t *input,
         "packsswh   %[ftmp3],   %[ftmp3],       %[ftmp1]                \n\t"
         "packsswh   %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
         "dmfc1      %[tmp1],    %[ftmp3]                                \n\t"
-        "ssrld      %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
+        SSRLD "     %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
         "mfc1       %[input],   %[ftmp3]                                \n\t"
         "sh         %[tmp1],    0x100(%[output])                        \n\t"
         "sh         %[input],   0x180(%[output])                        \n\t"
@@ -1059,7 +1059,7 @@  void ff_h264_luma_dc_dequant_idct_8_mmi(int16_t *output, int16_t *input,
         "sh         %[tmp1],    0x120(%[output])                        \n\t"
         "sh         %[input],   0x1a0(%[output])                        \n\t"
         "dmfc1      %[tmp1],    %[ftmp4]                                \n\t"
-        "ssrld      %[ftmp4],   %[ftmp4],       %[ftmp9]                \n\t"
+        SSRLD "     %[ftmp4],   %[ftmp4],       %[ftmp9]                \n\t"
         "mfc1       %[input],   %[ftmp4]                                \n\t"
         "sh         %[tmp1],    0x140(%[output])                        \n\t"
         "sh         %[input],   0x1c0(%[output])                        \n\t"
@@ -1144,7 +1144,7 @@  void ff_h264_weight_pixels16_8_mmi(uint8_t *block, ptrdiff_t stride, int height,
 
     for (y=0; y<height; y++, block+=stride) {
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
             MMI_LDC1(%[ftmp1], %[block0], 0x00)
             MMI_LDC1(%[ftmp2], %[block1], 0x00)
             "mtc1       %[weight],  %[ftmp3]                            \n\t"
@@ -1198,7 +1198,7 @@  void ff_h264_biweight_pixels16_8_mmi(uint8_t *dst, uint8_t *src,
 
     for (y=0; y<height; y++, dst+=stride, src+=stride) {
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
             MMI_LDC1(%[ftmp1], %[src0], 0x00)
             MMI_LDC1(%[ftmp2], %[dst0], 0x00)
             "mtc1       %[weights], %[ftmp3]                            \n\t"
@@ -1271,7 +1271,7 @@  void ff_h264_weight_pixels8_8_mmi(uint8_t *block, ptrdiff_t stride, int height,
 
     for (y=0; y<height; y++, block+=stride) {
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
             MMI_LDC1(%[ftmp1], %[block], 0x00)
             "mtc1       %[weight],  %[ftmp2]                            \n\t"
             "mtc1       %[offset],  %[ftmp3]                            \n\t"
@@ -1312,7 +1312,7 @@  void ff_h264_biweight_pixels8_8_mmi(uint8_t *dst, uint8_t *src,
 
     for (y=0; y<height; y++, dst+=stride, src+=stride) {
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
             MMI_LDC1(%[ftmp1], %[src], 0x00)
             MMI_LDC1(%[ftmp2], %[dst], 0x00)
             "mtc1       %[weights], %[ftmp3]                            \n\t"
@@ -1366,7 +1366,7 @@  void ff_h264_weight_pixels4_8_mmi(uint8_t *block, ptrdiff_t stride, int height,
 
     for (y=0; y<height; y++, block+=stride) {
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
             MMI_ULWC1(%[ftmp1], %[block], 0x00)
             "mtc1       %[weight],  %[ftmp2]                            \n\t"
             "mtc1       %[offset],  %[ftmp3]                            \n\t"
@@ -1402,7 +1402,7 @@  void ff_h264_biweight_pixels4_8_mmi(uint8_t *dst, uint8_t *src,
 
     for (y=0; y<height; y++, dst+=stride, src+=stride) {
         __asm__ volatile (
-            "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+            PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
             MMI_ULWC1(%[ftmp1], %[src], 0x00)
             MMI_ULWC1(%[ftmp2], %[dst], 0x00)
             "mtc1       %[weight],  %[ftmp3]                            \n\t"
@@ -1445,7 +1445,7 @@  void ff_deblock_v8_luma_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha, int bet
 
     __asm__ volatile (
         PTR_ADDU   "%[addr0],   %[stride],      %[stride]               \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         PTR_ADDU   "%[addr1],   %[stride],      %[addr0]                \n\t"
         "addi       %[alpha],   %[alpha],       -0x01                   \n\t"
         PTR_SUBU   "%[addr1],   $0,             %[addr1]                \n\t"
@@ -1463,18 +1463,18 @@  void ff_deblock_v8_luma_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha, int bet
         "packushb   %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp3],       %[ftmp2]                \n\t"
         "psubusb    %[ftmp8],   %[ftmp2],       %[ftmp3]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp2],       %[ftmp1]                \n\t"
         "psubusb    %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp1],       %[ftmp2]                \n\t"
-        "por        %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
+        POR "       %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp3],       %[ftmp4]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp5],       %[ftmp6]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp4],       %[ftmp3]                \n\t"
-        "por        %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
+        POR "       %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp5],       %[ftmp6]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
         "pcmpeqb    %[ftmp8],   %[ftmp8],       %[ftmp0]                \n\t"
         "pcmpeqb    %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
         MMI_ULWC1(%[ftmp5], %[tc0], 0x00)
@@ -1482,21 +1482,21 @@  void ff_deblock_v8_luma_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha, int bet
         "punpcklbh  %[ftmp9],   %[ftmp5],       %[ftmp5]                \n\t"
         "pcmpgtb    %[ftmp5],   %[ftmp9],       %[ftmp4]                \n\t"
         MMI_LDC1(%[ftmp4], %[addr1], 0x00)
-        "pand       %[ftmp10],  %[ftmp5],       %[ftmp8]                \n\t"
+        PAND "      %[ftmp10],  %[ftmp5],       %[ftmp8]                \n\t"
         "psubusb    %[ftmp8],   %[ftmp4],       %[ftmp2]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp2],       %[ftmp4]                \n\t"
         "psubusb    %[ftmp8],   %[ftmp8],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
         "pcmpeqb    %[ftmp7],   %[ftmp7],       %[ftmp8]                \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
-        "pand       %[ftmp5],   %[ftmp10],      %[ftmp9]                \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
+        PAND "      %[ftmp5],   %[ftmp10],      %[ftmp9]                \n\t"
         "psubb      %[ftmp8],   %[ftmp5],       %[ftmp7]                \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp5]                \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp5]                \n\t"
         "pavgb      %[ftmp5],   %[ftmp2],       %[ftmp3]                \n\t"
         MMI_LDC1(%[ftmp11], %[addr1], 0x00)
         "pavgb      %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
-        "pxor       %[ftmp5],   %[ftmp5],       %[ftmp11]               \n\t"
-        "pand       %[ftmp5],   %[ftmp5],       %[ff_pb_1]              \n\t"
+        PXOR "      %[ftmp5],   %[ftmp5],       %[ftmp11]               \n\t"
+        PAND "      %[ftmp5],   %[ftmp5],       %[ff_pb_1]              \n\t"
         "psubusb    %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp1],       %[ftmp7]                \n\t"
         "paddusb    %[ftmp7],   %[ftmp7],       %[ftmp1]                \n\t"
@@ -1509,26 +1509,26 @@  void ff_deblock_v8_luma_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha, int bet
         "psubusb    %[ftmp4],   %[ftmp4],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
         "pcmpeqb    %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
         "psubb      %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
-        "pand       %[ftmp6],   %[ftmp9],       %[ftmp7]                \n\t"
+        PAND "      %[ftmp6],   %[ftmp9],       %[ftmp7]                \n\t"
         MMI_LDXC1(%[ftmp4], %[pix], %[stride], 0x00)
         "pavgb      %[ftmp7],   %[ftmp2],       %[ftmp3]                \n\t"
         MMI_LDXC1(%[ftmp11], %[pix], %[addr0], 0x00)
         "pavgb      %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp11]               \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ff_pb_1]              \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp11]               \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ff_pb_1]              \n\t"
         "psubusb    %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp4],       %[ftmp6]                \n\t"
         "paddusb    %[ftmp6],   %[ftmp6],       %[ftmp4]                \n\t"
         "pmaxub     %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "pminub     %[ftmp5],   %[ftmp5],       %[ftmp6]                \n\t"
         MMI_SDXC1(%[ftmp5], %[pix], %[stride], 0x00)
-        "pxor       %[ftmp6],   %[ftmp2],       %[ftmp3]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp2],       %[ftmp3]                \n\t"
         "pcmpeqb    %[ftmp5],   %[ftmp5],       %[ftmp5]                \n\t"
-        "pand       %[ftmp6],   %[ftmp6],       %[ff_pb_1]              \n\t"
-        "pxor       %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
-        "pxor       %[ftmp5],   %[ftmp5],       %[ftmp2]                \n\t"
+        PAND "      %[ftmp6],   %[ftmp6],       %[ff_pb_1]              \n\t"
+        PXOR "      %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
+        PXOR "      %[ftmp5],   %[ftmp5],       %[ftmp2]                \n\t"
         "pavgb      %[ftmp4],   %[ftmp4],       %[ftmp1]                \n\t"
         "pavgb      %[ftmp4],   %[ftmp4],       %[ff_pb_3]              \n\t"
         "pavgb      %[ftmp5],   %[ftmp5],       %[ftmp3]                \n\t"
@@ -1574,12 +1574,12 @@  static void deblock_v8_luma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alph
 
     __asm__ volatile (
         "ori        %[tmp0],    $0,             0x01                    \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "mtc1       %[tmp0],    %[ftmp9]                                \n\t"
         PTR_SLL    "%[addr0],   %[stride],      0x02                    \n\t"
         PTR_ADDU   "%[addr2],   %[stride],      %[stride]               \n\t"
         PTR_ADDIU  "%[alpha],   %[alpha],       -0x01                   \n\t"
-        "sslld      %[ftmp11],  %[ftmp9],       %[ftmp9]                \n\t"
+        SSLLD "     %[ftmp11],  %[ftmp9],       %[ftmp9]                \n\t"
         "bltz       %[alpha],   1f                                      \n\t"
         PTR_ADDU   "%[addr1],   %[addr2],       %[stride]               \n\t"
         PTR_ADDIU  "%[beta],    %[beta],        -0x01                   \n\t"
@@ -1598,20 +1598,20 @@  static void deblock_v8_luma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alph
         "psubusb    %[ftmp7],   %[ftmp3],       %[ftmp2]                \n\t"
         "psubusb    %[ftmp8],   %[ftmp2],       %[ftmp3]                \n\t"
         "packushb   %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
         MMI_SDC1(%[ftmp5], %[stack], 0x10)
         "psubusb    %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp2],       %[ftmp1]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp1],       %[ftmp2]                \n\t"
-        "por        %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
+        POR "       %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp5],       %[ftmp6]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp3],       %[ftmp4]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp4],       %[ftmp3]                \n\t"
-        "por        %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
+        POR "       %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp5],       %[ftmp6]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
         MMI_LDC1(%[ftmp5], %[stack], 0x10)
         "pcmpeqb    %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
         "ldc1       %[ftmp10],  %[ff_pb_1]                              \n\t"
@@ -1624,14 +1624,14 @@  static void deblock_v8_luma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alph
         "psubusb    %[ftmp7],   %[ftmp7],       %[ftmp5]                \n\t"
         MMI_LDC1(%[ftmp15], %[stack], 0x20)
         "pcmpeqb    %[ftmp7],   %[ftmp7],       %[ftmp8]                \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp15]               \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp15]               \n\t"
         MMI_LDXC1(%[ftmp15], %[addr0], %[stride], 0x00)
         "psubusb    %[ftmp8],   %[ftmp15],      %[ftmp2]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp2],       %[ftmp15]               \n\t"
         "psubusb    %[ftmp8],   %[ftmp8],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp5],       %[ftmp6]                \n\t"
         "pcmpeqb    %[ftmp5],   %[ftmp5],       %[ftmp8]                \n\t"
-        "pand       %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
+        PAND "      %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         MMI_LDXC1(%[ftmp14], %[pix], %[addr2], 0x00)
         MMI_SDC1(%[ftmp5], %[stack], 0x30)
         "psubusb    %[ftmp8],   %[ftmp14],      %[ftmp3]                \n\t"
@@ -1639,7 +1639,7 @@  static void deblock_v8_luma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alph
         "psubusb    %[ftmp8],   %[ftmp8],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp5],       %[ftmp6]                \n\t"
         "pcmpeqb    %[ftmp5],   %[ftmp5],       %[ftmp8]                \n\t"
-        "pand       %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
+        PAND "      %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         MMI_SDC1(%[ftmp5], %[stack], 0x40)
         "pavgb      %[ftmp5],   %[ftmp15],      %[ftmp1]                \n\t"
         "pavgb      %[ftmp6],   %[ftmp2],       %[ftmp3]                \n\t"
@@ -1652,36 +1652,36 @@  static void deblock_v8_luma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alph
         MMI_SDC1(%[ftmp7], %[stack], 0x00)
         "psrlh      %[ftmp7],   %[ftmp7],       %[ftmp9]                \n\t"
         "pavgb      %[ftmp7],   %[ftmp7],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp5]                \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp5]                \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
         "psubb      %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "pavgb      %[ftmp6],   %[ftmp15],      %[ftmp4]                \n\t"
         "psubb      %[ftmp7],   %[ftmp15],      %[ftmp4]                \n\t"
         "paddb      %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
         "psubb      %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
         "psubb      %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
         MMI_LDC1(%[ftmp13], %[stack], 0x10)
         "pavgb      %[ftmp6],   %[ftmp6],       %[ftmp1]                \n\t"
         "psrlh      %[ftmp8],   %[ftmp8],       %[ftmp11]               \n\t"
         "pavgb      %[ftmp6],   %[ftmp6],       %[ftmp13]               \n\t"
         "pavgb      %[ftmp8],   %[ftmp8],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp8],   %[ftmp8],       %[ftmp6]                \n\t"
-        "pand       %[ftmp8],   %[ftmp8],       %[ftmp10]               \n\t"
+        PXOR "      %[ftmp8],   %[ftmp8],       %[ftmp6]                \n\t"
+        PAND "      %[ftmp8],   %[ftmp8],       %[ftmp10]               \n\t"
         "psubb      %[ftmp6],   %[ftmp6],       %[ftmp8]                \n\t"
-        "pxor       %[ftmp8],   %[ftmp2],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp8],   %[ftmp2],       %[ftmp4]                \n\t"
         "pavgb      %[ftmp7],   %[ftmp2],       %[ftmp4]                \n\t"
-        "pand       %[ftmp8],   %[ftmp8],       %[ftmp10]               \n\t"
+        PAND "      %[ftmp8],   %[ftmp8],       %[ftmp10]               \n\t"
         "psubb      %[ftmp7],   %[ftmp7],       %[ftmp8]                \n\t"
         MMI_LDC1(%[ftmp13], %[stack], 0x30)
         "pavgb      %[ftmp7],   %[ftmp7],       %[ftmp1]                \n\t"
         MMI_LDC1(%[ftmp12], %[stack], 0x20)
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp2]                \n\t"
-        "pand       %[ftmp6],   %[ftmp6],       %[ftmp13]               \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp12]               \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp2]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp2]                \n\t"
+        PAND "      %[ftmp6],   %[ftmp6],       %[ftmp13]               \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp12]               \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp2]                \n\t"
         MMI_SDXC1(%[ftmp6], %[addr0], %[addr1], 0x00)
         MMI_LDC1(%[ftmp6], %[addr0], 0x00)
         "paddb      %[ftmp7],   %[ftmp15],      %[ftmp6]                \n\t"
@@ -1692,16 +1692,16 @@  static void deblock_v8_luma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alph
         "paddb      %[ftmp7],   %[ftmp7],       %[ftmp12]               \n\t"
         "psrlh      %[ftmp7],   %[ftmp7],       %[ftmp11]               \n\t"
         "pavgb      %[ftmp7],   %[ftmp7],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
         MMI_LDC1(%[ftmp12], %[stack], 0x30)
         "psubb      %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp5],   %[ftmp5],       %[ftmp1]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp15]               \n\t"
-        "pand       %[ftmp5],   %[ftmp5],       %[ftmp12]               \n\t"
-        "pand       %[ftmp6],   %[ftmp6],       %[ftmp12]               \n\t"
-        "pxor       %[ftmp5],   %[ftmp5],       %[ftmp1]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp15]               \n\t"
+        PXOR "      %[ftmp5],   %[ftmp5],       %[ftmp1]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp15]               \n\t"
+        PAND "      %[ftmp5],   %[ftmp5],       %[ftmp12]               \n\t"
+        PAND "      %[ftmp6],   %[ftmp6],       %[ftmp12]               \n\t"
+        PXOR "      %[ftmp5],   %[ftmp5],       %[ftmp1]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp15]               \n\t"
         MMI_SDXC1(%[ftmp5], %[addr0], %[addr2], 0x00)
         MMI_SDXC1(%[ftmp6], %[addr0], %[stride], 0x00)
         "pavgb      %[ftmp5],   %[ftmp14],      %[ftmp4]                \n\t"
@@ -1715,36 +1715,36 @@  static void deblock_v8_luma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alph
         MMI_SDC1(%[ftmp7], %[stack], 0x00)
         "psrlh      %[ftmp7],   %[ftmp7],       %[ftmp9]                \n\t"
         "pavgb      %[ftmp7],   %[ftmp7],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp5]                \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp5]                \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
         "psubb      %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "pavgb      %[ftmp6],   %[ftmp14],      %[ftmp1]                \n\t"
         "paddb      %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
         "psubb      %[ftmp7],   %[ftmp14],      %[ftmp1]                \n\t"
         "psubb      %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
         "psubb      %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
         MMI_LDC1(%[ftmp12], %[stack], 0x10)
         "pavgb      %[ftmp6],   %[ftmp6],       %[ftmp4]                \n\t"
         "pavgb      %[ftmp6],   %[ftmp6],       %[ftmp12]               \n\t"
         "psrlh      %[ftmp8],   %[ftmp8],       %[ftmp11]               \n\t"
         "pavgb      %[ftmp8],   %[ftmp8],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp8],   %[ftmp8],       %[ftmp6]                \n\t"
-        "pand       %[ftmp8],   %[ftmp8],       %[ftmp10]               \n\t"
+        PXOR "      %[ftmp8],   %[ftmp8],       %[ftmp6]                \n\t"
+        PAND "      %[ftmp8],   %[ftmp8],       %[ftmp10]               \n\t"
         "psubb      %[ftmp6],   %[ftmp6],       %[ftmp8]                \n\t"
-        "pxor       %[ftmp8],   %[ftmp3],       %[ftmp1]                \n\t"
+        PXOR "      %[ftmp8],   %[ftmp3],       %[ftmp1]                \n\t"
         "pavgb      %[ftmp7],   %[ftmp3],       %[ftmp1]                \n\t"
-        "pand       %[ftmp8],   %[ftmp8],       %[ftmp10]               \n\t"
+        PAND "      %[ftmp8],   %[ftmp8],       %[ftmp10]               \n\t"
         MMI_LDC1(%[ftmp12], %[stack], 0x40)
         "psubb      %[ftmp7],   %[ftmp7],       %[ftmp8]                \n\t"
         MMI_LDC1(%[ftmp13], %[stack], 0x20)
         "pavgb      %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp3]                \n\t"
-        "pand       %[ftmp6],   %[ftmp6],       %[ftmp12]               \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp13]               \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp3]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp3]                \n\t"
+        PAND "      %[ftmp6],   %[ftmp6],       %[ftmp12]               \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp13]               \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp3]                \n\t"
         MMI_SDC1(%[ftmp6], %[pix], 0x00)
         MMI_LDXC1(%[ftmp6], %[pix], %[addr1], 0x00)
         "paddb      %[ftmp7],   %[ftmp14],      %[ftmp6]                \n\t"
@@ -1755,16 +1755,16 @@  static void deblock_v8_luma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alph
         "paddb      %[ftmp7],   %[ftmp7],       %[ftmp12]               \n\t"
         "psrlh      %[ftmp7],   %[ftmp7],       %[ftmp11]               \n\t"
         "pavgb      %[ftmp7],   %[ftmp7],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp10]               \n\t"
         MMI_LDC1(%[ftmp12], %[stack], 0x40)
         "psubb      %[ftmp6],   %[ftmp6],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp5],   %[ftmp5],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp14]               \n\t"
-        "pand       %[ftmp5],   %[ftmp5],       %[ftmp12]               \n\t"
-        "pand       %[ftmp6],   %[ftmp6],       %[ftmp12]               \n\t"
-        "pxor       %[ftmp5],   %[ftmp5],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp14]               \n\t"
+        PXOR "      %[ftmp5],   %[ftmp5],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp14]               \n\t"
+        PAND "      %[ftmp5],   %[ftmp5],       %[ftmp12]               \n\t"
+        PAND "      %[ftmp6],   %[ftmp6],       %[ftmp12]               \n\t"
+        PXOR "      %[ftmp5],   %[ftmp5],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp14]               \n\t"
         MMI_SDXC1(%[ftmp5], %[pix], %[stride], 0x00)
         MMI_SDXC1(%[ftmp6], %[pix], %[addr2], 0x00)
         "1:                                                             \n\t"
@@ -1808,7 +1808,7 @@  void ff_deblock_v_chroma_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha,
         MMI_LDC1(%[ftmp3], %[pix], 0x00)
         MMI_LDXC1(%[ftmp4], %[pix], %[stride], 0x00)
 
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "mtc1       %[alpha],   %[ftmp5]                                \n\t"
         "mtc1       %[beta],    %[ftmp6]                                \n\t"
         "pshufh     %[ftmp5],   %[ftmp5],       %[ftmp0]                \n\t"
@@ -1817,29 +1817,29 @@  void ff_deblock_v_chroma_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha,
         "packushb   %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp3],       %[ftmp2]                \n\t"
         "psubusb    %[ftmp8],   %[ftmp2],       %[ftmp3]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp2],       %[ftmp1]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp1],       %[ftmp2]                \n\t"
-        "por        %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
+        POR "       %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp5],       %[ftmp6]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp3],       %[ftmp4]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp4],       %[ftmp3]                \n\t"
-        "por        %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
+        POR "       %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp5],       %[ftmp6]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
         "pcmpeqb    %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
         MMI_ULWC1(%[ftmp7], %[tc0], 0x00)
         "punpcklbh  %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
-        "pand       %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
+        PAND "      %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
         "pcmpeqb    %[ftmp5],   %[ftmp5],       %[ftmp5]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp2],       %[ftmp3]                \n\t"
-        "pxor       %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
-        "pand       %[ftmp6],   %[ftmp6],       %[ff_pb_1]              \n\t"
+        PXOR "      %[ftmp6],   %[ftmp2],       %[ftmp3]                \n\t"
+        PXOR "      %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
+        PAND "      %[ftmp6],   %[ftmp6],       %[ff_pb_1]              \n\t"
         "pavgb      %[ftmp4],   %[ftmp4],       %[ftmp1]                \n\t"
-        "pxor       %[ftmp5],   %[ftmp5],       %[ftmp2]                \n\t"
+        PXOR "      %[ftmp5],   %[ftmp5],       %[ftmp2]                \n\t"
         "pavgb      %[ftmp4],   %[ftmp4],       %[ff_pb_3]              \n\t"
         "pavgb      %[ftmp5],   %[ftmp5],       %[ftmp3]                \n\t"
         "pavgb      %[ftmp4],   %[ftmp4],       %[ftmp6]                \n\t"
@@ -1891,7 +1891,7 @@  void ff_deblock_v_chroma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha,
         MMI_LDC1(%[ftmp3], %[pix], 0x00)
         MMI_LDXC1(%[ftmp4], %[pix], %[stride], 0x00)
 
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "mtc1       %[alpha],   %[ftmp5]                                \n\t"
         "mtc1       %[beta],    %[ftmp6]                                \n\t"
         "pshufh     %[ftmp5],   %[ftmp5],       %[ftmp0]                \n\t"
@@ -1900,36 +1900,36 @@  void ff_deblock_v_chroma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha,
         "packushb   %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp3],       %[ftmp2]                \n\t"
         "psubusb    %[ftmp8],   %[ftmp2],       %[ftmp3]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp2],       %[ftmp1]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp1],       %[ftmp2]                \n\t"
-        "por        %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
+        POR "       %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp5],       %[ftmp6]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp3],       %[ftmp4]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp4],       %[ftmp3]                \n\t"
-        "por        %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
+        POR "       %[ftmp5],   %[ftmp5],       %[ftmp7]                \n\t"
         "psubusb    %[ftmp5],   %[ftmp5],       %[ftmp6]                \n\t"
-        "por        %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
+        POR "       %[ftmp8],   %[ftmp8],       %[ftmp5]                \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
         "pcmpeqb    %[ftmp8],   %[ftmp8],       %[ftmp7]                \n\t"
         "mov.d      %[ftmp6],   %[ftmp2]                                \n\t"
         "mov.d      %[ftmp7],   %[ftmp3]                                \n\t"
-        "pxor       %[ftmp5],   %[ftmp2],       %[ftmp4]                \n\t"
-        "pand       %[ftmp5],   %[ftmp5],       %[ff_pb_1]              \n\t"
+        PXOR "      %[ftmp5],   %[ftmp2],       %[ftmp4]                \n\t"
+        PAND "      %[ftmp5],   %[ftmp5],       %[ff_pb_1]              \n\t"
         "pavgb      %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
         "psubusb    %[ftmp2],   %[ftmp2],       %[ftmp5]                \n\t"
         "pavgb      %[ftmp2],   %[ftmp2],       %[ftmp1]                \n\t"
-        "pxor       %[ftmp5],   %[ftmp3],       %[ftmp1]                \n\t"
-        "pand       %[ftmp5],   %[ftmp5],       %[ff_pb_1]              \n\t"
+        PXOR "      %[ftmp5],   %[ftmp3],       %[ftmp1]                \n\t"
+        PAND "      %[ftmp5],   %[ftmp5],       %[ff_pb_1]              \n\t"
         "pavgb      %[ftmp3],   %[ftmp3],       %[ftmp1]                \n\t"
         "psubusb    %[ftmp3],   %[ftmp3],       %[ftmp5]                \n\t"
         "pavgb      %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
         "psubb      %[ftmp2],   %[ftmp2],       %[ftmp6]                \n\t"
         "psubb      %[ftmp3],   %[ftmp3],       %[ftmp7]                \n\t"
-        "pand       %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
-        "pand       %[ftmp3],   %[ftmp3],       %[ftmp8]                \n\t"
+        PAND "      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
+        PAND "      %[ftmp3],   %[ftmp3],       %[ftmp8]                \n\t"
         "paddb      %[ftmp2],   %[ftmp2],       %[ftmp6]                \n\t"
         "paddb      %[ftmp3],   %[ftmp3],       %[ftmp7]                \n\t"
 
@@ -1996,7 +1996,7 @@  void ff_deblock_h_chroma_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha, int be
         "mov.d      %[ftmp9],   %[ftmp0]                                \n\t"
         "mov.d      %[ftmp10],  %[ftmp3]                                \n\t"
 
-        "pxor       %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
+        PXOR "      %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
         "mtc1       %[alpha],   %[ftmp4]                                \n\t"
         "mtc1       %[beta],    %[ftmp5]                                \n\t"
         "pshufh     %[ftmp4],   %[ftmp4],       %[ftmp8]                \n\t"
@@ -2005,29 +2005,29 @@  void ff_deblock_h_chroma_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha, int be
         "packushb   %[ftmp5],   %[ftmp5],       %[ftmp5]                \n\t"
         "psubusb    %[ftmp6],   %[ftmp2],       %[ftmp1]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp1],       %[ftmp2]                \n\t"
-        "por        %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
+        POR "       %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
         "psubusb    %[ftmp6],   %[ftmp1],       %[ftmp0]                \n\t"
         "psubusb    %[ftmp4],   %[ftmp0],       %[ftmp1]                \n\t"
-        "por        %[ftmp4],   %[ftmp4],       %[ftmp6]                \n\t"
+        POR "       %[ftmp4],   %[ftmp4],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
-        "por        %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
+        POR "       %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
         "psubusb    %[ftmp6],   %[ftmp2],       %[ftmp3]                \n\t"
         "psubusb    %[ftmp4],   %[ftmp3],       %[ftmp2]                \n\t"
-        "por        %[ftmp4],   %[ftmp4],       %[ftmp6]                \n\t"
+        POR "       %[ftmp4],   %[ftmp4],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
-        "por        %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
+        POR "       %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
         "pcmpeqb    %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
         MMI_ULWC1(%[ftmp6], %[tc0], 0x00)
         "punpcklbh  %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
-        "pand       %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
+        PAND "      %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
         "pcmpeqb    %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp5],   %[ftmp1],       %[ftmp2]                \n\t"
-        "pxor       %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
-        "pand       %[ftmp5],   %[ftmp5],       %[ff_pb_1]              \n\t"
+        PXOR "      %[ftmp5],   %[ftmp1],       %[ftmp2]                \n\t"
+        PXOR "      %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
+        PAND "      %[ftmp5],   %[ftmp5],       %[ff_pb_1]              \n\t"
         "pavgb      %[ftmp3],   %[ftmp3],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp4],   %[ftmp4],       %[ftmp1]                \n\t"
+        PXOR "      %[ftmp4],   %[ftmp4],       %[ftmp1]                \n\t"
         "pavgb      %[ftmp3],   %[ftmp3],       %[ff_pb_3]              \n\t"
         "pavgb      %[ftmp4],   %[ftmp4],       %[ftmp2]                \n\t"
         "pavgb      %[ftmp3],   %[ftmp3],       %[ftmp5]                \n\t"
@@ -2134,7 +2134,7 @@  void ff_deblock_h_chroma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha,
         "punpcklwd  %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
         "punpcklwd  %[ftmp2],   %[ftmp2],       %[ftmp6]                \n\t"
 
-        "pxor       %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
+        PXOR "      %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
         "mtc1       %[alpha],   %[ftmp4]                                \n\t"
         "mtc1       %[beta],    %[ftmp5]                                \n\t"
         "pshufh     %[ftmp4],   %[ftmp4],       %[ftmp8]                \n\t"
@@ -2143,36 +2143,36 @@  void ff_deblock_h_chroma_intra_8_mmi(uint8_t *pix, ptrdiff_t stride, int alpha,
         "packushb   %[ftmp5],   %[ftmp5],       %[ftmp5]                \n\t"
         "psubusb    %[ftmp6],   %[ftmp2],       %[ftmp1]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp1],       %[ftmp2]                \n\t"
-        "por        %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
+        POR "       %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
         "psubusb    %[ftmp6],   %[ftmp1],       %[ftmp0]                \n\t"
         "psubusb    %[ftmp4],   %[ftmp0],       %[ftmp1]                \n\t"
-        "por        %[ftmp4],   %[ftmp4],       %[ftmp6]                \n\t"
+        POR "       %[ftmp4],   %[ftmp4],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
-        "por        %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
+        POR "       %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
         "psubusb    %[ftmp6],   %[ftmp2],       %[ftmp3]                \n\t"
         "psubusb    %[ftmp4],   %[ftmp3],       %[ftmp2]                \n\t"
-        "por        %[ftmp4],   %[ftmp4],       %[ftmp6]                \n\t"
+        POR "       %[ftmp4],   %[ftmp4],       %[ftmp6]                \n\t"
         "psubusb    %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
-        "por        %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
+        POR "       %[ftmp7],   %[ftmp7],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
         "pcmpeqb    %[ftmp7],   %[ftmp7],       %[ftmp6]                \n\t"
         "mov.d      %[ftmp5],   %[ftmp1]                                \n\t"
         "mov.d      %[ftmp6],   %[ftmp2]                                \n\t"
-        "pxor       %[ftmp4],   %[ftmp1],       %[ftmp3]                \n\t"
-        "pand       %[ftmp4],   %[ftmp4],       %[ff_pb_1]              \n\t"
+        PXOR "      %[ftmp4],   %[ftmp1],       %[ftmp3]                \n\t"
+        PAND "      %[ftmp4],   %[ftmp4],       %[ff_pb_1]              \n\t"
         "pavgb      %[ftmp1],   %[ftmp1],       %[ftmp3]                \n\t"
         "psubusb    %[ftmp1],   %[ftmp1],       %[ftmp4]                \n\t"
         "pavgb      %[ftmp1],   %[ftmp1],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp4],   %[ftmp2],       %[ftmp0]                \n\t"
-        "pand       %[ftmp4],   %[ftmp4],       %[ff_pb_1]              \n\t"
+        PXOR "      %[ftmp4],   %[ftmp2],       %[ftmp0]                \n\t"
+        PAND "      %[ftmp4],   %[ftmp4],       %[ff_pb_1]              \n\t"
         "pavgb      %[ftmp2],   %[ftmp2],       %[ftmp0]                \n\t"
         "psubusb    %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
         "pavgb      %[ftmp2],   %[ftmp2],       %[ftmp3]                \n\t"
         "psubb      %[ftmp1],   %[ftmp1],       %[ftmp5]                \n\t"
         "psubb      %[ftmp2],   %[ftmp2],       %[ftmp6]                \n\t"
-        "pand       %[ftmp1],   %[ftmp1],       %[ftmp7]                \n\t"
-        "pand       %[ftmp2],   %[ftmp2],       %[ftmp7]                \n\t"
+        PAND "      %[ftmp1],   %[ftmp1],       %[ftmp7]                \n\t"
+        PAND "      %[ftmp2],   %[ftmp2],       %[ftmp7]                \n\t"
         "paddb      %[ftmp1],   %[ftmp1],       %[ftmp5]                \n\t"
         "paddb      %[ftmp2],   %[ftmp2],       %[ftmp6]                \n\t"
 
diff --git a/libavcodec/mips/h264pred_mmi.c b/libavcodec/mips/h264pred_mmi.c
index 480411f5b5..ee0f730379 100644
--- a/libavcodec/mips/h264pred_mmi.c
+++ b/libavcodec/mips/h264pred_mmi.c
@@ -162,7 +162,7 @@  void ff_pred8x8l_top_dc_8_mmi(uint8_t *src, int has_topleft,
     DECLARE_VAR_ADDRT;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         MMI_ULDC1(%[ftmp10], %[srcA], 0x00)
         MMI_ULDC1(%[ftmp9], %[src0], 0x00)
         MMI_ULDC1(%[ftmp8], %[src1], 0x00)
@@ -267,7 +267,7 @@  void ff_pred8x8l_dc_8_mmi(uint8_t *src, int has_topleft, int has_topright,
         MMI_ULDC1(%[ftmp4], %[srcA], 0x00)
         MMI_ULDC1(%[ftmp5], %[src0], 0x00)
         MMI_ULDC1(%[ftmp6], %[src1], 0x00)
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "dli        %[tmp0],    0x03                                    \n\t"
         "punpcklbh  %[ftmp7],   %[ftmp4],       %[ftmp0]                \n\t"
         "punpckhbh  %[ftmp8],   %[ftmp4],       %[ftmp0]                \n\t"
@@ -358,7 +358,7 @@  void ff_pred8x8l_vertical_8_mmi(uint8_t *src, int has_topleft,
     DECLARE_VAR_ALL64;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         MMI_LDC1(%[ftmp3], %[srcA], 0x00)
         MMI_LDC1(%[ftmp4], %[src0], 0x00)
         MMI_LDC1(%[ftmp5], %[src1], 0x00)
@@ -531,7 +531,7 @@  void ff_pred8x8_top_dc_8_mmi(uint8_t *src, ptrdiff_t stride)
 
     __asm__ volatile (
         "dli        %[tmp0],    0x02                                    \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         PTR_SUBU   "%[addr0],   %[src],         %[stride]               \n\t"
         MMI_LDC1(%[ftmp1], %[addr0], 0x00)
         "punpcklbh  %[ftmp2],   %[ftmp1],       %[ftmp0]                \n\t"
@@ -641,7 +641,7 @@  void ff_pred8x8_dc_8_mmi(uint8_t *src, ptrdiff_t stride)
         PTR_SRL    "%[addr4],   0x02                                    \n\t"
         PTR_SRL    "%[addr1],   0x02                                    \n\t"
         PTR_SRL    "%[addr2],   0x03                                    \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "dmtc1      %[addr3],   %[ftmp1]                                \n\t"
         "pshufh     %[ftmp1],   %[ftmp1],       %[ftmp0]                \n\t"
         "dmtc1      %[addr4],   %[ftmp2]                                \n\t"
@@ -758,9 +758,9 @@  static inline void pred16x16_plane_compat_mmi(uint8_t *src, int stride,
         "dmtc1      %[tmp0],    %[ftmp4]                                \n\t"
         MMI_ULDC1(%[ftmp0], %[addr0], -0x01)
         MMI_ULDC1(%[ftmp2], %[addr0],  0x08)
-        "ssrld      %[ftmp1],   %[ftmp0],       %[ftmp4]                \n\t"
-        "ssrld      %[ftmp3],   %[ftmp2],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
+        SSRLD "     %[ftmp1],   %[ftmp0],       %[ftmp4]                \n\t"
+        SSRLD "     %[ftmp3],   %[ftmp2],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
         "punpcklbh  %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
         "punpcklbh  %[ftmp1],   %[ftmp1],       %[ftmp4]                \n\t"
         "punpcklbh  %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
@@ -916,7 +916,7 @@  static inline void pred16x16_plane_compat_mmi(uint8_t *src, int stride,
         "dmul       %[tmp3],    %[tmp3],        %[tmp2]                 \n\t"
         "dsubu      %[tmp5],    %[tmp5],        %[tmp3]                 \n\t"
 
-        "pxor       %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
         "dmtc1      %[tmp0],    %[ftmp0]                                \n\t"
         "pshufh     %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
         "dmtc1      %[tmp1],    %[ftmp5]                                \n\t"
diff --git a/libavcodec/mips/h264qpel_mmi.c b/libavcodec/mips/h264qpel_mmi.c
index 3482956e13..e1d5e5c43a 100644
--- a/libavcodec/mips/h264qpel_mmi.c
+++ b/libavcodec/mips/h264qpel_mmi.c
@@ -114,7 +114,7 @@  static void put_h264_qpel4_h_lowpass_mmi(uint8_t *dst, const uint8_t *src,
     DECLARE_VAR_LOW32;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "dli        %[tmp0],    0x04                                    \n\t"
         "1:                                                             \n\t"
         MMI_ULWC1(%[ftmp1], %[src], -0x02)
@@ -169,7 +169,7 @@  static void put_h264_qpel8_h_lowpass_mmi(uint8_t *dst, const uint8_t *src,
     DECLARE_VAR_ALL64;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "dli        %[tmp0],    0x08                                    \n\t"
         "1:                                                             \n\t"
         MMI_ULDC1(%[ftmp1], %[src], -0x02)
@@ -250,7 +250,7 @@  static void avg_h264_qpel4_h_lowpass_mmi(uint8_t *dst, const uint8_t *src,
     DECLARE_VAR_LOW32;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "dli        %[tmp0],    0x04                                    \n\t"
         "1:                                                             \n\t"
         MMI_ULWC1(%[ftmp1], %[src], -0x02)
@@ -307,7 +307,7 @@  static void avg_h264_qpel8_h_lowpass_mmi(uint8_t *dst, const uint8_t *src,
     DECLARE_VAR_ALL64;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "dli        %[tmp0],    0x08                                    \n\t"
         "1:                                                             \n\t"
         MMI_ULDC1(%[ftmp1], %[src], -0x02)
@@ -394,7 +394,7 @@  static void put_h264_qpel4_v_lowpass_mmi(uint8_t *dst, const uint8_t *src,
     __asm__ volatile (
         ".set       push                                                \n\t"
         ".set       noreorder                                           \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "dli        %[tmp0],    0x02                                    \n\t"
         MMI_LWC1(%[ftmp1], %[src], 0x00)
         "mtc1       %[tmp0],    %[ftmp10]                               \n\t"
@@ -516,7 +516,7 @@  static void put_h264_qpel8_v_lowpass_mmi(uint8_t *dst, const uint8_t *src,
             PTR_ADDU   "%[src],     %[src],         %[srcStride]        \n\t"
             MMI_LWC1(%[ftmp2], %[src], 0x00)
             PTR_ADDU   "%[src],     %[src],         %[srcStride]        \n\t"
-            "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]            \n\t"
+            PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]            \n\t"
             MMI_LWC1(%[ftmp3], %[src], 0x00)
             PTR_ADDU   "%[src],     %[src],         %[srcStride]        \n\t"
             MMI_LWC1(%[ftmp4], %[src], 0x00)
@@ -812,7 +812,7 @@  static void avg_h264_qpel4_v_lowpass_mmi(uint8_t *dst, const uint8_t *src,
         ".set       push                                                \n\t"
         ".set       noreorder                                           \n\t"
         "dli        %[tmp0],    0x02                                    \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
         "mtc1       %[tmp0],    %[ftmp9]                                \n\t"
         "dli        %[tmp0],    0x05                                    \n\t"
         MMI_LWC1(%[ftmp0], %[src], 0x00)
@@ -930,7 +930,7 @@  static void avg_h264_qpel8_v_lowpass_mmi(uint8_t *dst, const uint8_t *src,
             ".set       push                                            \n\t"
             ".set       noreorder                                       \n\t"
             "dli        %[tmp0],    0x02                                \n\t"
-            "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]            \n\t"
+            PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]            \n\t"
             "mtc1       %[tmp0],    %[ftmp9]                            \n\t"
             "dli        %[tmp0],    0x05                                \n\t"
             MMI_LWC1(%[ftmp0], %[src], 0x00)
@@ -1269,7 +1269,7 @@  static void put_h264_qpel4_hv_lowpass_mmi(uint8_t *dst, const uint8_t *src,
     src -= 2*srcStride;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "dli        %[tmp0],    0x09                                    \n\t"
         "1:                                                             \n\t"
         MMI_ULWC1(%[ftmp1], %[src], -0x02)
@@ -1347,7 +1347,7 @@  static void put_h264_qpel8or16_hv1_lowpass_mmi(int16_t *tmp,
             MMI_ULWC1(%[ftmp0], %[src], 0x00)
             "mtc1       %[tmp0],    %[ftmp10]                           \n\t"
             PTR_ADDU   "%[src],     %[src],         %[srcStride]        \n\t"
-            "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]            \n\t"
+            PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]            \n\t"
             MMI_ULWC1(%[ftmp1], %[src], 0x00)
             PTR_ADDU   "%[src],     %[src],         %[srcStride]        \n\t"
             MMI_ULWC1(%[ftmp2], %[src], 0x00)
@@ -1684,7 +1684,7 @@  static void put_h264_qpel8_h_lowpass_l2_mmi(uint8_t *dst, const uint8_t *src,
         "dli        %[tmp0],    0x02                                    \n\t"
         "mtc1       %[tmp0],    %[ftmp7]                                \n\t"
         "dli        %[tmp0],    0x05                                    \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "mtc1       %[tmp0],    %[ftmp8]                                \n\t"
         "1:                                                             \n\t"
         MMI_ULDC1(%[ftmp1], %[src], 0x00)
@@ -1833,7 +1833,7 @@  static void avg_h264_qpel4_hv_lowpass_mmi(uint8_t *dst, const uint8_t *src,
     src -= 2*srcStride;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "dli        %[tmp0],    0x09                                    \n\t"
         "1:                                                             \n\t"
         MMI_ULWC1(%[ftmp1], %[src], -0x02)
@@ -2005,7 +2005,7 @@  static void avg_h264_qpel8_h_lowpass_l2_mmi(uint8_t *dst, const uint8_t *src,
         "ori        %[tmp0],    $0,             0x8                     \n\t"
         "mtc1       %[tmp1],    %[ftmp7]                                \n\t"
         "dli        %[tmp1],    0x05                                    \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "mtc1       %[tmp1],    %[ftmp8]                                \n\t"
         "1:                                                             \n\t"
         MMI_ULDC1(%[ftmp1], %[src], 0x00)
diff --git a/libavcodec/mips/hevcdsp_mmi.c b/libavcodec/mips/hevcdsp_mmi.c
index 6583bef5da..5d2def9caa 100644
--- a/libavcodec/mips/hevcdsp_mmi.c
+++ b/libavcodec/mips/hevcdsp_mmi.c
@@ -47,7 +47,7 @@  void ff_hevc_put_hevc_qpel_h##w##_8_mmi(int16_t *dst, uint8_t *_src,     \
         "punpcklbh    %[ftmp1],      %[ftmp0],      %[ftmp1]    \n\t"    \
         "psrah        %[ftmp1],      %[ftmp1],      %[ftmp0]    \n\t"    \
         "psrah        %[ftmp2],      %[ftmp2],      %[ftmp0]    \n\t"    \
-        "pxor         %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"    \
+        PXOR "        %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"    \
                                                                          \
         "1:                                                     \n\t"    \
         "2:                                                     \n\t"    \
@@ -145,7 +145,7 @@  void ff_hevc_put_hevc_qpel_hv##w##_8_mmi(int16_t *dst, uint8_t *_src,    \
         "punpcklbh    %[ftmp1],      %[ftmp0],      %[ftmp1]    \n\t"    \
         "psrah        %[ftmp1],      %[ftmp1],      %[ftmp0]    \n\t"    \
         "psrah        %[ftmp2],      %[ftmp2],      %[ftmp0]    \n\t"    \
-        "pxor         %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"    \
+        PXOR "        %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"    \
                                                                          \
         "1:                                                     \n\t"    \
         "2:                                                     \n\t"    \
@@ -334,7 +334,7 @@  void ff_hevc_put_hevc_qpel_bi_h##w##_8_mmi(uint8_t *_dst,               \
         "punpcklbh    %[ftmp1],      %[ftmp0],      %[ftmp1]    \n\t"   \
         "psrah        %[ftmp1],      %[ftmp1],      %[ftmp0]    \n\t"   \
         "psrah        %[ftmp2],      %[ftmp2],      %[ftmp0]    \n\t"   \
-        "pxor         %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"   \
+        PXOR "        %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"   \
         "punpcklhw    %[offset],     %[offset],     %[offset]   \n\t"   \
         "punpcklwd    %[offset],     %[offset],     %[offset]   \n\t"   \
                                                                         \
@@ -388,7 +388,7 @@  void ff_hevc_put_hevc_qpel_bi_h##w##_8_mmi(uint8_t *_dst,               \
         "psraw        %[ftmp6],      %[ftmp6],      %[shift]    \n\t"   \
         "packsswh     %[ftmp5],      %[ftmp5],      %[ftmp6]    \n\t"   \
         "pcmpgth      %[ftmp7],      %[ftmp5],      %[ftmp0]    \n\t"   \
-        "pand         %[ftmp3],      %[ftmp5],      %[ftmp7]    \n\t"   \
+        PAND "        %[ftmp3],      %[ftmp5],      %[ftmp7]    \n\t"   \
         "packushb     %[ftmp3],      %[ftmp3],      %[ftmp3]    \n\t"   \
         MMI_USWC1(%[ftmp3], %[dst], 0x00)                               \
                                                                         \
@@ -469,7 +469,7 @@  void ff_hevc_put_hevc_qpel_bi_hv##w##_8_mmi(uint8_t *_dst,              \
         "punpcklbh    %[ftmp1],      %[ftmp0],      %[ftmp1]    \n\t"   \
         "psrah        %[ftmp1],      %[ftmp1],      %[ftmp0]    \n\t"   \
         "psrah        %[ftmp2],      %[ftmp2],      %[ftmp0]    \n\t"   \
-        "pxor         %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"   \
+        PXOR "        %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"   \
                                                                         \
         "1:                                                     \n\t"   \
         "2:                                                     \n\t"   \
@@ -588,7 +588,7 @@  void ff_hevc_put_hevc_qpel_bi_hv##w##_8_mmi(uint8_t *_dst,              \
         "psraw        %[ftmp5],      %[ftmp5],      %[ftmp0]    \n\t"   \
         "packsswh     %[ftmp3],      %[ftmp3],      %[ftmp5]    \n\t"   \
         MMI_ULDC1(%[ftmp4], %[src2], 0x00)                              \
-        "pxor         %[ftmp7],      %[ftmp7],      %[ftmp7]    \n\t"   \
+        PXOR "        %[ftmp7],      %[ftmp7],      %[ftmp7]    \n\t"   \
         "li           %[rtmp0],      0x10                       \n\t"   \
         "dmtc1        %[rtmp0],      %[ftmp8]                   \n\t"   \
         "punpcklhw    %[ftmp5],      %[ftmp7],      %[ftmp3]    \n\t"   \
@@ -607,7 +607,7 @@  void ff_hevc_put_hevc_qpel_bi_hv##w##_8_mmi(uint8_t *_dst,              \
         "psraw        %[ftmp6],      %[ftmp6],      %[shift]    \n\t"   \
         "packsswh     %[ftmp5],      %[ftmp5],      %[ftmp6]    \n\t"   \
         "pcmpgth      %[ftmp7],      %[ftmp5],      %[ftmp7]    \n\t"   \
-        "pand         %[ftmp3],      %[ftmp5],      %[ftmp7]    \n\t"   \
+        PAND "        %[ftmp3],      %[ftmp5],      %[ftmp7]    \n\t"   \
         "packushb     %[ftmp3],      %[ftmp3],      %[ftmp3]    \n\t"   \
         MMI_USWC1(%[ftmp3], %[dst], 0x00)                               \
                                                                         \
@@ -686,7 +686,7 @@  void ff_hevc_put_hevc_epel_bi_hv##w##_8_mmi(uint8_t *_dst,              \
         "dmtc1        %[rtmp0],      %[ftmp0]                   \n\t"   \
         "punpcklbh    %[ftmp1],      %[ftmp0],      %[ftmp1]    \n\t"   \
         "psrah        %[ftmp1],      %[ftmp1],      %[ftmp0]    \n\t"   \
-        "pxor         %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"   \
+        PXOR "        %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"   \
                                                                         \
         "1:                                                     \n\t"   \
         "2:                                                     \n\t"   \
@@ -747,7 +747,7 @@  void ff_hevc_put_hevc_epel_bi_hv##w##_8_mmi(uint8_t *_dst,              \
         "li           %[rtmp0],      0x06                       \n\t"   \
         "dmtc1        %[rtmp0],      %[ftmp0]                   \n\t"   \
         "punpcklwd    %[offset],     %[offset],     %[offset]   \n\t"   \
-        "pxor         %[ftmp2],      %[ftmp2],      %[ftmp2]    \n\t"   \
+        PXOR "        %[ftmp2],      %[ftmp2],      %[ftmp2]    \n\t"   \
                                                                         \
         "1:                                                     \n\t"   \
         "li           %[x],        " #x_step "                  \n\t"   \
@@ -792,7 +792,7 @@  void ff_hevc_put_hevc_epel_bi_hv##w##_8_mmi(uint8_t *_dst,              \
         "psraw        %[ftmp6],      %[ftmp6],      %[shift]    \n\t"   \
         "packsswh     %[ftmp5],      %[ftmp5],      %[ftmp6]    \n\t"   \
         "pcmpgth      %[ftmp7],      %[ftmp5],      %[ftmp2]    \n\t"   \
-        "pand         %[ftmp3],      %[ftmp5],      %[ftmp7]    \n\t"   \
+        PAND "        %[ftmp3],      %[ftmp5],      %[ftmp7]    \n\t"   \
         "packushb     %[ftmp3],      %[ftmp3],      %[ftmp3]    \n\t"   \
         MMI_USWC1(%[ftmp3], %[dst], 0x0)                                \
                                                                         \
@@ -855,7 +855,7 @@  void ff_hevc_put_hevc_pel_bi_pixels##w##_8_mmi(uint8_t *_dst,             \
     y = height;                                                           \
     x = width >> 3;                                                       \
     __asm__ volatile(                                                     \
-        "pxor         %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"     \
+        PXOR "        %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"     \
         "li           %[rtmp0],      0x06                       \n\t"     \
         "dmtc1        %[rtmp0],      %[ftmp1]                   \n\t"     \
         "li           %[rtmp0],      0x10                       \n\t"     \
@@ -900,8 +900,8 @@  void ff_hevc_put_hevc_pel_bi_pixels##w##_8_mmi(uint8_t *_dst,             \
         "packsswh     %[ftmp4],      %[ftmp4],      %[ftmp5]    \n\t"     \
         "pcmpgth      %[ftmp3],      %[ftmp2],      %[ftmp0]    \n\t"     \
         "pcmpgth      %[ftmp5],      %[ftmp4],      %[ftmp0]    \n\t"     \
-        "pand         %[ftmp2],      %[ftmp2],      %[ftmp3]    \n\t"     \
-        "pand         %[ftmp4],      %[ftmp4],      %[ftmp5]    \n\t"     \
+        PAND "        %[ftmp2],      %[ftmp2],      %[ftmp3]    \n\t"     \
+        PAND "        %[ftmp4],      %[ftmp4],      %[ftmp5]    \n\t"     \
         "packushb     %[ftmp2],      %[ftmp2],      %[ftmp4]    \n\t"     \
         MMI_USDC1(%[ftmp2], %[dst], 0x0)                                  \
                                                                           \
@@ -980,7 +980,7 @@  void ff_hevc_put_hevc_qpel_uni_hv##w##_8_mmi(uint8_t *_dst,             \
         "punpcklbh    %[ftmp1],      %[ftmp0],      %[ftmp1]    \n\t"   \
         "psrah        %[ftmp1],      %[ftmp1],      %[ftmp0]    \n\t"   \
         "psrah        %[ftmp2],      %[ftmp2],      %[ftmp0]    \n\t"   \
-        "pxor         %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"   \
+        PXOR "        %[ftmp0],      %[ftmp0],      %[ftmp0]    \n\t"   \
                                                                         \
         "1:                                                     \n\t"   \
         "2:                                                     \n\t"   \
@@ -1101,9 +1101,9 @@  void ff_hevc_put_hevc_qpel_uni_hv##w##_8_mmi(uint8_t *_dst,             \
         "packsswh     %[ftmp3],      %[ftmp3],      %[ftmp5]    \n\t"   \
         "paddh        %[ftmp3],      %[ftmp3],      %[offset]   \n\t"   \
         "psrah        %[ftmp3],      %[ftmp3],      %[shift]    \n\t"   \
-        "pxor         %[ftmp7],      %[ftmp7],      %[ftmp7]    \n\t"   \
+        PXOR "        %[ftmp7],      %[ftmp7],      %[ftmp7]    \n\t"   \
         "pcmpgth      %[ftmp7],      %[ftmp3],      %[ftmp7]    \n\t"   \
-        "pand         %[ftmp3],      %[ftmp3],      %[ftmp7]    \n\t"   \
+        PAND "        %[ftmp3],      %[ftmp3],      %[ftmp7]    \n\t"   \
         "packushb     %[ftmp3],      %[ftmp3],      %[ftmp3]    \n\t"   \
         MMI_USWC1(%[ftmp3], %[dst], 0x00)                               \
                                                                         \
diff --git a/libavcodec/mips/hpeldsp_mmi.c b/libavcodec/mips/hpeldsp_mmi.c
index 8e9c0fa821..03a817b7dd 100644
--- a/libavcodec/mips/hpeldsp_mmi.c
+++ b/libavcodec/mips/hpeldsp_mmi.c
@@ -677,14 +677,14 @@  inline void ff_put_no_rnd_pixels8_l2_8_mmi(uint8_t *dst, const uint8_t *src1,
         PTR_ADDU   "%[addr1],   %[src2],        %[src_stride2]          \n\t"
         MMI_ULDC1(%[ftmp3], %[addr1], 0x00)
         PTR_ADDU   "%[src1],    %[src1],        %[addr2]                \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
         "pavgb      %[ftmp0],   %[ftmp0],       %[ftmp2]                \n\t"
         "pavgb      %[ftmp1],   %[ftmp1],       %[ftmp3]                \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp4]                \n\t"
         MMI_SDC1(%[ftmp0], %[dst], 0x00)
         MMI_SDXC1(%[ftmp1], %[dst], %[dst_stride], 0x00)
         PTR_ADDU   "%[src2],    %[src2],        %[addr3]                \n\t"
@@ -697,14 +697,14 @@  inline void ff_put_no_rnd_pixels8_l2_8_mmi(uint8_t *dst, const uint8_t *src1,
         PTR_ADDU   "%[addr1],   %[src2],        %[src_stride2]          \n\t"
         MMI_ULDC1(%[ftmp3], %[addr1], 0x00)
         PTR_ADDU   "%[src1],    %[src1],        %[addr2]                \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
         "pavgb      %[ftmp0],   %[ftmp0],       %[ftmp2]                \n\t"
         "pavgb      %[ftmp1],   %[ftmp1],       %[ftmp3]                \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp4]                \n\t"
         MMI_SDC1(%[ftmp0], %[dst], 0x00)
         MMI_SDXC1(%[ftmp1], %[dst], %[dst_stride], 0x00)
         PTR_ADDU   "%[src2],    %[src2],        %[addr3]                \n\t"
@@ -847,7 +847,7 @@  void ff_put_pixels8_xy2_8_mmi(uint8_t *block, const uint8_t *pixels,
     DECLARE_VAR_ADDRT;
 
     __asm__ volatile (
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
         "dli        %[addr0],   0x0f                                    \n\t"
         "pcmpeqw    %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
         "dmtc1      %[addr0],   %[ftmp8]                                \n\t"
diff --git a/libavcodec/mips/idctdsp_mmi.c b/libavcodec/mips/idctdsp_mmi.c
index d22e5eedd7..3de0c9f39c 100644
--- a/libavcodec/mips/idctdsp_mmi.c
+++ b/libavcodec/mips/idctdsp_mmi.c
@@ -154,7 +154,7 @@  void ff_add_pixels_clamped_mmi(const int16_t *block,
     uint64_t tmp[1];
     __asm__ volatile (
         "li         %[tmp0],    0x04                           \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],   %[ftmp0]           \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],   %[ftmp0]           \n\t"
         "1:                                                    \n\t"
         MMI_LDC1(%[ftmp5], %[pixels], 0x00)
         PTR_ADDU   "%[pixels],  %[pixels],  %[line_size]       \n\t"
diff --git a/libavcodec/mips/mpegvideo_mmi.c b/libavcodec/mips/mpegvideo_mmi.c
index 3d5b5e20ab..a7f9da9b65 100644
--- a/libavcodec/mips/mpegvideo_mmi.c
+++ b/libavcodec/mips/mpegvideo_mmi.c
@@ -54,13 +54,13 @@  void ff_dct_unquantize_h263_intra_mmi(MpegEncContext *s, int16_t *block,
         nCoeffs = s->inter_scantable.raster_end[s->block_last_index[n]];
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "packsswh   %[qmul],    %[qmul],        %[qmul]                 \n\t"
         "packsswh   %[qmul],    %[qmul],        %[qmul]                 \n\t"
         "packsswh   %[qadd],    %[qadd],        %[qadd]                 \n\t"
         "packsswh   %[qadd],    %[qadd],        %[qadd]                 \n\t"
         "psubh      %[ftmp0],   %[ftmp0],       %[qadd]                 \n\t"
-        "pxor       %[ftmp5],   %[ftmp5],       %[ftmp5]                \n\t"
+        PXOR "      %[ftmp5],   %[ftmp5],       %[ftmp5]                \n\t"
         ".p2align   4                                                   \n\t"
 
         "1:                                                             \n\t"
@@ -73,12 +73,12 @@  void ff_dct_unquantize_h263_intra_mmi(MpegEncContext *s, int16_t *block,
         "pmullh     %[ftmp2],   %[ftmp2],       %[qmul]                 \n\t"
         "pcmpgth    %[ftmp3],   %[ftmp3],       %[ftmp5]                \n\t"
         "pcmpgth    %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp3]                \n\t"
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp3]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
         "paddh      %[ftmp1],   %[ftmp1],       %[ftmp0]                \n\t"
         "paddh      %[ftmp2],   %[ftmp2],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp3],   %[ftmp3],       %[ftmp1]                \n\t"
-        "pxor       %[ftmp4],   %[ftmp4],       %[ftmp2]                \n\t"
+        PXOR "      %[ftmp3],   %[ftmp3],       %[ftmp1]                \n\t"
+        PXOR "      %[ftmp4],   %[ftmp4],       %[ftmp2]                \n\t"
         "pcmpeqh    %[ftmp1],   %[ftmp1],       %[ftmp0]                \n\t"
         "pcmpeqh    %[ftmp2],   %[ftmp2],       %[ftmp0]                \n\t"
         "pandn      %[ftmp1],   %[ftmp1],       %[ftmp3]                \n\t"
@@ -118,11 +118,11 @@  void ff_dct_unquantize_h263_inter_mmi(MpegEncContext *s, int16_t *block,
     __asm__ volatile (
         "packsswh   %[qmul],    %[qmul],        %[qmul]                 \n\t"
         "packsswh   %[qmul],    %[qmul],        %[qmul]                 \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "packsswh   %[qadd],    %[qadd],        %[qadd]                 \n\t"
         "packsswh   %[qadd],    %[qadd],        %[qadd]                 \n\t"
         "psubh      %[ftmp0],   %[ftmp0],       %[qadd]                 \n\t"
-        "pxor       %[ftmp5],   %[ftmp5],       %[ftmp5]                \n\t"
+        PXOR "      %[ftmp5],   %[ftmp5],       %[ftmp5]                \n\t"
         ".p2align   4                                                   \n\t"
         "1:                                                             \n\t"
         PTR_ADDU   "%[addr0],   %[block],       %[nCoeffs]              \n\t"
@@ -134,12 +134,12 @@  void ff_dct_unquantize_h263_inter_mmi(MpegEncContext *s, int16_t *block,
         "pmullh     %[ftmp2],   %[ftmp2],       %[qmul]                 \n\t"
         "pcmpgth    %[ftmp3],   %[ftmp3],       %[ftmp5]                \n\t"
         "pcmpgth    %[ftmp4],   %[ftmp4],       %[ftmp5]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp3]                \n\t"
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp3]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
         "paddh      %[ftmp1],   %[ftmp1],       %[ftmp0]                \n\t"
         "paddh      %[ftmp2],   %[ftmp2],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp3],   %[ftmp3],       %[ftmp1]                \n\t"
-        "pxor       %[ftmp4],   %[ftmp4],       %[ftmp2]                \n\t"
+        PXOR "      %[ftmp3],   %[ftmp3],       %[ftmp1]                \n\t"
+        PXOR "      %[ftmp4],   %[ftmp4],       %[ftmp2]                \n\t"
         "pcmpeqh    %[ftmp1],   %[ftmp1],       %[ftmp0]                \n\t"
         "pcmpeqh    %[ftmp2],   %[ftmp2],       %[ftmp0]                \n\t"
         "pandn      %[ftmp1],   %[ftmp1],       %[ftmp3]                \n\t"
@@ -203,18 +203,18 @@  void ff_dct_unquantize_mpeg1_intra_mmi(MpegEncContext *s, int16_t *block,
         MMI_LDXC1(%[ftmp7], %[addr0], %[quant], 0x08)
         "pmullh     %[ftmp6],   %[ftmp6],       %[ftmp1]                \n\t"
         "pmullh     %[ftmp7],   %[ftmp7],       %[ftmp1]                \n\t"
-        "pxor       %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
-        "pxor       %[ftmp9],   %[ftmp9],       %[ftmp9]                \n\t"
+        PXOR "      %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
+        PXOR "      %[ftmp9],   %[ftmp9],       %[ftmp9]                \n\t"
         "pcmpgth    %[ftmp8],   %[ftmp8],       %[ftmp2]                \n\t"
         "pcmpgth    %[ftmp9],   %[ftmp9],       %[ftmp3]                \n\t"
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
-        "pxor       %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
+        PXOR "      %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
         "psubh      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
         "psubh      %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
         "pmullh     %[ftmp2],   %[ftmp2],       %[ftmp6]                \n\t"
         "pmullh     %[ftmp3],   %[ftmp3],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
         "pcmpeqh    %[ftmp6],   %[ftmp6],       %[ftmp4]                \n\t"
         "dli        %[tmp0],    0x03                                    \n\t"
         "pcmpeqh    %[ftmp7],   %[ftmp7],       %[ftmp5]                \n\t"
@@ -223,10 +223,10 @@  void ff_dct_unquantize_mpeg1_intra_mmi(MpegEncContext *s, int16_t *block,
         "psrah      %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
         "psubh      %[ftmp2],   %[ftmp2],       %[ftmp0]                \n\t"
         "psubh      %[ftmp3],   %[ftmp3],       %[ftmp0]                \n\t"
-        "por        %[ftmp2],   %[ftmp2],       %[ftmp0]                \n\t"
-        "por        %[ftmp3],   %[ftmp3],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
-        "pxor       %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
+        POR "       %[ftmp2],   %[ftmp2],       %[ftmp0]                \n\t"
+        POR "       %[ftmp3],   %[ftmp3],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
+        PXOR "      %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
         "psubh      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
         "psubh      %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
         "pandn      %[ftmp6],   %[ftmp6],       %[ftmp2]                \n\t"
@@ -289,12 +289,12 @@  void ff_dct_unquantize_mpeg1_inter_mmi(MpegEncContext *s, int16_t *block,
         MMI_LDXC1(%[ftmp7], %[addr0], %[quant], 0x08)
         "pmullh     %[ftmp6],   %[ftmp6],       %[ftmp1]                \n\t"
         "pmullh     %[ftmp7],   %[ftmp7],       %[ftmp1]                \n\t"
-        "pxor       %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
-        "pxor       %[ftmp9],   %[ftmp9],       %[ftmp9]                \n\t"
+        PXOR "      %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
+        PXOR "      %[ftmp9],   %[ftmp9],       %[ftmp9]                \n\t"
         "pcmpgth    %[ftmp8],   %[ftmp8],       %[ftmp2]                \n\t"
         "pcmpgth    %[ftmp9],   %[ftmp9],       %[ftmp3]                \n\t"
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
-        "pxor       %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
+        PXOR "      %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
         "psubh      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
         "psubh      %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
         "paddh      %[ftmp2],   %[ftmp2],       %[ftmp2]                \n\t"
@@ -303,8 +303,8 @@  void ff_dct_unquantize_mpeg1_inter_mmi(MpegEncContext *s, int16_t *block,
         "paddh      %[ftmp3],   %[ftmp3],       %[ftmp0]                \n\t"
         "pmullh     %[ftmp2],   %[ftmp2],       %[ftmp6]                \n\t"
         "pmullh     %[ftmp3],   %[ftmp3],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
         "pcmpeqh    %[ftmp6],   %[ftmp6],       %[ftmp4]                \n\t"
         "dli        %[tmp0],    0x04                                    \n\t"
         "pcmpeqh    %[ftmp7],   %[ftmp7],       %[ftmp5]                \n\t"
@@ -313,10 +313,10 @@  void ff_dct_unquantize_mpeg1_inter_mmi(MpegEncContext *s, int16_t *block,
         "psrah      %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
         "psubh      %[ftmp2],   %[ftmp2],       %[ftmp0]                \n\t"
         "psubh      %[ftmp3],   %[ftmp3],       %[ftmp0]                \n\t"
-        "por        %[ftmp2],   %[ftmp2],       %[ftmp0]                \n\t"
-        "por        %[ftmp3],   %[ftmp3],       %[ftmp0]                \n\t"
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
-        "pxor       %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
+        POR "       %[ftmp2],   %[ftmp2],       %[ftmp0]                \n\t"
+        POR "       %[ftmp3],   %[ftmp3],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
+        PXOR "      %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
         "psubh      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
         "psubh      %[ftmp3],   %[ftmp3],       %[ftmp9]                \n\t"
         "pandn      %[ftmp6],   %[ftmp6],       %[ftmp2]                \n\t"
@@ -388,26 +388,26 @@  void ff_dct_unquantize_mpeg2_intra_mmi(MpegEncContext *s, int16_t *block,
         MMI_LDXC1(%[ftmp6], %[addr0], %[quant], 0x08)
         "pmullh     %[ftmp5],   %[ftmp5],       %[ftmp9]                \n\t"
         "pmullh     %[ftmp6],   %[ftmp6],       %[ftmp9]                \n\t"
-        "pxor       %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
+        PXOR "      %[ftmp7],   %[ftmp7],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp8],   %[ftmp8],       %[ftmp8]                \n\t"
         "pcmpgth    %[ftmp7],   %[ftmp7],       %[ftmp1]                \n\t"
         "pcmpgth    %[ftmp8],   %[ftmp8],       %[ftmp2]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
         "psubh      %[ftmp1],   %[ftmp1],       %[ftmp7]                \n\t"
         "psubh      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
         "pmullh     %[ftmp1],   %[ftmp1],       %[ftmp5]                \n\t"
         "pmullh     %[ftmp2],   %[ftmp2],       %[ftmp6]                \n\t"
-        "pxor       %[ftmp5],   %[ftmp5],       %[ftmp5]                \n\t"
-        "pxor       %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
+        PXOR "      %[ftmp5],   %[ftmp5],       %[ftmp5]                \n\t"
+        PXOR "      %[ftmp6],   %[ftmp6],       %[ftmp6]                \n\t"
         "pcmpeqh    %[ftmp5],   %[ftmp5],       %[ftmp3]                \n\t"
         "dli        %[tmp0],    0x03                                    \n\t"
         "pcmpeqh    %[ftmp6] ,  %[ftmp6],       %[ftmp4]                \n\t"
         "mtc1       %[tmp0],    %[ftmp3]                                \n\t"
         "psrah      %[ftmp1],   %[ftmp1],       %[ftmp3]                \n\t"
         "psrah      %[ftmp2],   %[ftmp2],       %[ftmp3]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp7]                \n\t"
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp7]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
         "psubh      %[ftmp1],   %[ftmp1],       %[ftmp7]                \n\t"
         "psubh      %[ftmp2],   %[ftmp2],       %[ftmp8]                \n\t"
         "pandn      %[ftmp5],   %[ftmp5],       %[ftmp1]                \n\t"
@@ -447,16 +447,16 @@  void ff_denoise_dct_mmi(MpegEncContext *s, int16_t *block)
     s->dct_count[intra]++;
 
     __asm__ volatile(
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "1:                                                             \n\t"
         MMI_LDC1(%[ftmp1], %[block], 0x00)
-        "pxor       %[ftmp2],   %[ftmp2],       %[ftmp2]                \n\t"
+        PXOR "      %[ftmp2],   %[ftmp2],       %[ftmp2]                \n\t"
         MMI_LDC1(%[ftmp3], %[block], 0x08)
-        "pxor       %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
         "pcmpgth    %[ftmp2],   %[ftmp2],       %[ftmp1]                \n\t"
         "pcmpgth    %[ftmp4],   %[ftmp4],       %[ftmp3]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp2]                \n\t"
-        "pxor       %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp2]                \n\t"
+        PXOR "      %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
         "psubh      %[ftmp1],   %[ftmp1],       %[ftmp2]                \n\t"
         "psubh      %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
         MMI_LDC1(%[ftmp6], %[offset], 0x00)
@@ -465,8 +465,8 @@  void ff_denoise_dct_mmi(MpegEncContext *s, int16_t *block)
         MMI_LDC1(%[ftmp6], %[offset], 0x08)
         "mov.d      %[ftmp7],   %[ftmp3]                                \n\t"
         "psubush    %[ftmp3],   %[ftmp3],       %[ftmp6]                \n\t"
-        "pxor       %[ftmp1],   %[ftmp1],       %[ftmp2]                \n\t"
-        "pxor       %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp1],   %[ftmp1],       %[ftmp2]                \n\t"
+        PXOR "      %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
         "psubh      %[ftmp1],   %[ftmp1],       %[ftmp2]                \n\t"
         "psubh      %[ftmp3],   %[ftmp3],       %[ftmp4]                \n\t"
         MMI_SDC1(%[ftmp1], %[block], 0x00)
diff --git a/libavcodec/mips/pixblockdsp_mmi.c b/libavcodec/mips/pixblockdsp_mmi.c
index 1230f5de88..6f18637f69 100644
--- a/libavcodec/mips/pixblockdsp_mmi.c
+++ b/libavcodec/mips/pixblockdsp_mmi.c
@@ -33,7 +33,7 @@  void ff_get_pixels_8_mmi(int16_t *av_restrict block, const uint8_t *pixels,
     DECLARE_VAR_ADDRT;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
 
         MMI_LDC1(%[ftmp1], %[pixels], 0x00)
         MMI_LDXC1(%[ftmp2], %[pixels], %[stride], 0x00)
@@ -103,12 +103,12 @@  void ff_diff_pixels_mmi(int16_t *av_restrict block, const uint8_t *src1,
 
     __asm__ volatile (
         "li         %[tmp0],    0x08                                    \n\t"
-        "pxor       %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
+        PXOR "      %[ftmp4],   %[ftmp4],       %[ftmp4]                \n\t"
         "1:                                                             \n\t"
         MMI_LDC1(%[ftmp0], %[src1], 0x00)
-        "por        %[ftmp1],   %[ftmp0],       %[ftmp0]                \n\t"
+        POR "       %[ftmp1],   %[ftmp0],       %[ftmp0]                \n\t"
         MMI_LDC1(%[ftmp2], %[src2], 0x00)
-        "por        %[ftmp3],   %[ftmp2],       %[ftmp2]                \n\t"
+        POR "       %[ftmp3],   %[ftmp2],       %[ftmp2]                \n\t"
         "punpcklbh  %[ftmp0],   %[ftmp0],       %[ftmp4]                \n\t"
         "punpckhbh  %[ftmp1],   %[ftmp1],       %[ftmp4]                \n\t"
         "punpcklbh  %[ftmp2],   %[ftmp2],       %[ftmp4]                \n\t"
diff --git a/libavcodec/mips/simple_idct_mmi.c b/libavcodec/mips/simple_idct_mmi.c
index 4680520edc..ab7df64300 100644
--- a/libavcodec/mips/simple_idct_mmi.c
+++ b/libavcodec/mips/simple_idct_mmi.c
@@ -135,7 +135,7 @@  void ff_simple_idct_8_mmi(int16_t *block)
         "psllh        $f28,     "#src1",    $f30                \n\t" \
         "dmtc1        $9,        $f31                           \n\t" \
         "punpcklhw    $f29,      $f28,      $f28                \n\t" \
-        "pand         $f29,      $f29,      $f31                \n\t" \
+        PAND "        $f29,      $f29,      $f31                \n\t" \
         "paddw        $f28,      $f28,      $f29                \n\t" \
         "punpcklwd   "#src1",    $f28,      $f28                \n\t" \
         "punpcklwd   "#src2",    $f28,      $f28                \n\t" \
@@ -269,9 +269,9 @@  void ff_simple_idct_8_mmi(int16_t *block)
         "punpcklwd    $f8,       $f27,      $f29                \n\t"
         "punpckhwd    $f12,      $f27,      $f29                \n\t"
 
-        "por          $f26,      $f2,       $f6                 \n\t"
-        "por          $f26,      $f26,      $f10                \n\t"
-        "por          $f26,      $f26,      $f14                \n\t"
+        POR "         $f26,      $f2,       $f6                 \n\t"
+        POR "         $f26,      $f26,      $f10                \n\t"
+        POR "         $f26,      $f26,      $f14                \n\t"
         "dmfc1        $10,       $f26                           \n\t"
         "bnez         $10,       1f                             \n\t"
         /* case1: In this case, row[1,3,5,7] are all zero */
@@ -339,9 +339,9 @@  void ff_simple_idct_8_mmi(int16_t *block)
         "punpcklwd    $f9,       $f27,      $f29                \n\t"
         "punpckhwd    $f13,      $f27,      $f29                \n\t"
 
-        "por          $f26,      $f3,       $f7                 \n\t"
-        "por          $f26,      $f26,      $f11                \n\t"
-        "por          $f26,      $f26,      $f15                \n\t"
+        POR "         $f26,      $f3,       $f7                 \n\t"
+        POR "         $f26,      $f26,      $f11                \n\t"
+        POR "         $f26,      $f26,      $f15                \n\t"
         "dmfc1        $10,       $f26                           \n\t"
         "bnez         $10,       1f                             \n\t"
         /* case1: In this case, row[1,3,5,7] are all zero */
diff --git a/libavcodec/mips/vc1dsp_mmi.c b/libavcodec/mips/vc1dsp_mmi.c
index 27a3c813da..aa5290b051 100644
--- a/libavcodec/mips/vc1dsp_mmi.c
+++ b/libavcodec/mips/vc1dsp_mmi.c
@@ -136,7 +136,7 @@  void ff_vc1_inv_trans_8x8_dc_mmi(uint8_t *dest, ptrdiff_t linesize, int16_t *blo
     dc_u.i = dc;
 
     __asm__ volatile(
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "pshufh     %[dc],      %[dc],          %[ftmp0]                \n\t"
         "li         %[count],   0x02                                    \n\t"
 
@@ -426,7 +426,7 @@  void ff_vc1_inv_trans_8x4_dc_mmi(uint8_t *dest, ptrdiff_t linesize, int16_t *blo
     dc_u.i = dc;
 
     __asm__ volatile(
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "pshufh     %[dc],      %[dc],          %[ftmp0]                \n\t"
 
         MMI_LDC1(%[ftmp1], %[dest0], 0x00)
@@ -704,7 +704,7 @@  void ff_vc1_inv_trans_8x4_mmi(uint8_t *dest, ptrdiff_t linesize, int16_t *block)
         MMI_LWC1(%[ftmp3], %[tmp0], 0x00)
         PTR_ADDU    "%[tmp0],   %[tmp0],    %[linesize]                 \n\t"
         MMI_LWC1(%[ftmp4], %[tmp0], 0x00)
-        "pxor       %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp1],   %[ftmp1],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp2],   %[ftmp2],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp3],   %[ftmp3],   %[ftmp0]                    \n\t"
@@ -828,7 +828,7 @@  void ff_vc1_inv_trans_8x4_mmi(uint8_t *dest, ptrdiff_t linesize, int16_t *block)
         MMI_LWC1(%[ftmp3], %[tmp0], 0x04)
         PTR_ADDU    "%[tmp0],   %[tmp0],    %[linesize]                 \n\t"
         MMI_LWC1(%[ftmp4], %[tmp0], 0x04)
-        "pxor       %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp1],   %[ftmp1],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp2],   %[ftmp2],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp3],   %[ftmp3],   %[ftmp0]                    \n\t"
@@ -878,7 +878,7 @@  void ff_vc1_inv_trans_4x8_dc_mmi(uint8_t *dest, ptrdiff_t linesize, int16_t *blo
     dc_u.i = dc;
 
     __asm__ volatile(
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "pshufh     %[dc],      %[dc],          %[ftmp0]                \n\t"
 
         MMI_LWC1(%[ftmp1], %[dest0], 0x00)
@@ -1056,7 +1056,7 @@  void ff_vc1_inv_trans_4x8_mmi(uint8_t *dest, ptrdiff_t linesize, int16_t *block)
         MMI_LWC1(%[ftmp7], %[tmp0], 0x00)
         PTR_ADDU  "%[tmp0],   %[tmp0],    %[linesize]                 \n\t"
         MMI_LWC1(%[ftmp8], %[tmp0], 0x00)
-        "pxor       %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp1],   %[ftmp1],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp2],   %[ftmp2],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp3],   %[ftmp3],   %[ftmp0]                    \n\t"
@@ -1133,7 +1133,7 @@  void ff_vc1_inv_trans_4x4_dc_mmi(uint8_t *dest, ptrdiff_t linesize, int16_t *blo
     dc_u.i = dc;
 
     __asm__ volatile(
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]                \n\t"
         "pshufh     %[dc],      %[dc],          %[ftmp0]                \n\t"
 
         MMI_LWC1(%[ftmp1], %[dest0], 0x00)
@@ -1337,7 +1337,7 @@  void ff_vc1_inv_trans_4x4_mmi(uint8_t *dest, ptrdiff_t linesize, int16_t *block)
         MMI_LWC1(%[ftmp3], %[tmp0], 0x00)
         PTR_ADDU    "%[tmp0],   %[tmp0],    %[linesize]                 \n\t"
         MMI_LWC1(%[ftmp4], %[tmp0], 0x00)
-        "pxor       %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp1],   %[ftmp1],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp2],   %[ftmp2],   %[ftmp0]                    \n\t"
         "punpcklbh  %[ftmp3],   %[ftmp3],   %[ftmp0]                    \n\t"
@@ -1664,7 +1664,7 @@  static void vc1_put_ver_16b_shift2_mmi(int16_t *dst,
     shift_u.i = shift;
 
     __asm__ volatile(
-        "pxor       $f0,    $f0,    $f0             \n\t"
+        PXOR "      $f0,    $f0,    $f0             \n\t"
         "li         $8,     0x03                    \n\t"
         LOAD_ROUNDER_MMI("%[rnd]")
         "1:                                         \n\t"
@@ -1768,7 +1768,7 @@  static void OPNAME ## vc1_shift2_mmi(uint8_t *dst, const uint8_t *src,      \
     rnd = 8 - rnd;                                                          \
                                                                             \
     __asm__ volatile(                                                       \
-        "pxor       $f0,    $f0,    $f0             \n\t"                   \
+        PXOR "      $f0,    $f0,    $f0             \n\t"                   \
         "li         $10,    0x08                    \n\t"                   \
         LOAD_ROUNDER_MMI("%[rnd]")                                          \
         "1:                                         \n\t"                   \
@@ -1896,7 +1896,7 @@  vc1_put_ver_16b_ ## NAME ## _mmi(int16_t *dst, const uint8_t *src,          \
     src -= src_stride;                                                      \
                                                                             \
     __asm__ volatile(                                                       \
-        "pxor       $f0,    $f0,    $f0             \n\t"                   \
+        PXOR "      $f0,    $f0,    $f0             \n\t"                   \
         LOAD_ROUNDER_MMI("%[rnd]")                                          \
         ".p2align 3                                 \n\t"                   \
         "1:                                         \n\t"                   \
@@ -1969,7 +1969,7 @@  OPNAME ## vc1_hor_16b_ ## NAME ## _mmi(uint8_t *dst, mips_reg stride,       \
     rnd -= (-4+58+13-3)*256; /* Add -256 bias */                            \
                                                                             \
     __asm__ volatile(                                                       \
-        "pxor       $f0,    $f0,    $f0             \n\t"                   \
+        PXOR "      $f0,    $f0,    $f0             \n\t"                   \
         LOAD_ROUNDER_MMI("%[rnd]")                                          \
         ".p2align 3                                 \n\t"                   \
         "1:                                         \n\t"                   \
@@ -2017,7 +2017,7 @@  OPNAME ## vc1_## NAME ## _mmi(uint8_t *dst, const uint8_t *src,             \
     rnd = 32-rnd;                                                           \
                                                                             \
     __asm__ volatile (                                                      \
-        "pxor       $f0,    $f0,    $f0             \n\t"                   \
+        PXOR "      $f0,    $f0,    $f0             \n\t"                   \
         LOAD_ROUNDER_MMI("%[rnd]")                                          \
         ".p2align 3                                 \n\t"                   \
         "1:                                         \n\t"                   \
@@ -2252,7 +2252,7 @@  void ff_put_no_rnd_vc1_chroma_mc8_mmi(uint8_t *dst /* align 8 */,
 
     __asm__ volatile(
         "li         %[tmp0],    0x06                                    \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
         "mtc1       %[tmp0],    %[ftmp9]                                \n\t"
         "pshufh     %[A],       %[A],       %[ftmp0]                    \n\t"
         "pshufh     %[B],       %[B],       %[ftmp0]                    \n\t"
@@ -2308,7 +2308,7 @@  void ff_put_no_rnd_vc1_chroma_mc4_mmi(uint8_t *dst /* align 8 */,
 
     __asm__ volatile(
         "li         %[tmp0],    0x06                                    \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
         "mtc1       %[tmp0],    %[ftmp5]                                \n\t"
         "pshufh     %[A],       %[A],       %[ftmp0]                    \n\t"
         "pshufh     %[B],       %[B],       %[ftmp0]                    \n\t"
@@ -2362,7 +2362,7 @@  void ff_avg_no_rnd_vc1_chroma_mc8_mmi(uint8_t *dst /* align 8 */,
 
     __asm__ volatile(
         "li         %[tmp0],    0x06                                    \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
         "mtc1       %[tmp0],    %[ftmp9]                                \n\t"
         "pshufh     %[A],       %[A],       %[ftmp0]                    \n\t"
         "pshufh     %[B],       %[B],       %[ftmp0]                    \n\t"
@@ -2421,7 +2421,7 @@  void ff_avg_no_rnd_vc1_chroma_mc4_mmi(uint8_t *dst /* align 8 */,
 
     __asm__ volatile(
         "li         %[tmp0],    0x06                                    \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
         "mtc1       %[tmp0],    %[ftmp5]                                \n\t"
         "pshufh     %[A],       %[A],       %[ftmp0]                    \n\t"
         "pshufh     %[B],       %[B],       %[ftmp0]                    \n\t"
diff --git a/libavcodec/mips/vp3dsp_idct_mmi.c b/libavcodec/mips/vp3dsp_idct_mmi.c
index d505fcb765..e27cfc3494 100644
--- a/libavcodec/mips/vp3dsp_idct_mmi.c
+++ b/libavcodec/mips/vp3dsp_idct_mmi.c
@@ -34,7 +34,7 @@  static void idct_row_mmi(int16_t *input)
     double ftmp[23];
     uint64_t tmp[2];
     __asm__ volatile (
-        "pxor       %[ftmp10],      %[ftmp10],        %[ftmp10] \n\t"
+        PXOR "      %[ftmp10],      %[ftmp10],        %[ftmp10] \n\t"
         LOAD_CONST(%[csth_1], 1)
         "li         %[tmp0],        0x02                        \n\t"
         "1:                                                     \n\t"
@@ -51,14 +51,14 @@  static void idct_row_mmi(int16_t *input)
         LOAD_CONST(%[ftmp9], 12785)
         "pmulhh     %[A],           %[ftmp9],         %[ftmp7]  \n\t"
         "pcmpgth    %[C],           %[ftmp10],        %[ftmp1]  \n\t"
-        "por        %[mask],        %[C],             %[csth_1] \n\t"
+        POR "       %[mask],        %[C],             %[csth_1] \n\t"
         "pmullh     %[B],           %[ftmp1],         %[mask]   \n\t"
         "pmulhuh    %[B],           %[ftmp8],         %[B]      \n\t"
         "pmullh     %[B],           %[B],             %[mask]   \n\t"
         "paddh      %[A],           %[A],             %[B]      \n\t"
         "paddh      %[A],           %[A],             %[C]      \n\t"
         "pcmpgth    %[D],           %[ftmp10],        %[ftmp7]  \n\t"
-        "por        %[mask],        %[D],             %[csth_1] \n\t"
+        POR "       %[mask],        %[D],             %[csth_1] \n\t"
         "pmullh     %[ftmp7],       %[ftmp7],         %[mask]   \n\t"
         "pmulhuh    %[B],           %[ftmp8],         %[ftmp7]  \n\t"
         "pmullh     %[B],           %[B],             %[mask]   \n\t"
@@ -69,12 +69,12 @@  static void idct_row_mmi(int16_t *input)
         LOAD_CONST(%[ftmp8], 54491)
         LOAD_CONST(%[ftmp9], 36410)
         "pcmpgth    %[Ad],          %[ftmp10],        %[ftmp5]  \n\t"
-        "por        %[mask],        %[Ad],            %[csth_1] \n\t"
+        POR "       %[mask],        %[Ad],            %[csth_1] \n\t"
         "pmullh     %[ftmp1],       %[ftmp5],         %[mask]   \n\t"
         "pmulhuh    %[C],           %[ftmp9],         %[ftmp1]  \n\t"
         "pmullh     %[C],           %[C],             %[mask]   \n\t"
         "pcmpgth    %[Bd],          %[ftmp10],        %[ftmp3]  \n\t"
-        "por        %[mask],        %[Bd],            %[csth_1] \n\t"
+        POR "       %[mask],        %[Bd],            %[csth_1] \n\t"
         "pmullh     %[D],           %[ftmp3],         %[mask]   \n\t"
         "pmulhuh    %[D],           %[ftmp8],         %[D]      \n\t"
         "pmullh     %[D],           %[D],             %[mask]   \n\t"
@@ -82,12 +82,12 @@  static void idct_row_mmi(int16_t *input)
         "paddh      %[C],           %[C],             %[Ad]     \n\t"
         "paddh      %[C],           %[C],             %[Bd]     \n\t"
         "pcmpgth    %[Bd],          %[ftmp10],        %[ftmp3]  \n\t"
-        "por        %[mask],        %[Bd],            %[csth_1] \n\t"
+        POR "       %[mask],        %[Bd],            %[csth_1] \n\t"
         "pmullh     %[ftmp1],       %[ftmp3],         %[mask]   \n\t"
         "pmulhuh    %[D],           %[ftmp9],         %[ftmp1]  \n\t"
         "pmullh     %[D],           %[D],             %[mask]   \n\t"
         "pcmpgth    %[Ed],          %[ftmp10],        %[ftmp5]  \n\t"
-        "por        %[mask],        %[Ed],            %[csth_1] \n\t"
+        POR "       %[mask],        %[Ed],            %[csth_1] \n\t"
         "pmullh     %[Ad],          %[ftmp5],         %[mask]   \n\t"
         "pmulhuh    %[Ad],          %[ftmp8],         %[Ad]     \n\t"
         "pmullh     %[Ad],          %[Ad],            %[mask]   \n\t"
@@ -98,14 +98,14 @@  static void idct_row_mmi(int16_t *input)
         LOAD_CONST(%[ftmp8], 46341)
         "psubh      %[Ad],          %[A],             %[C]      \n\t"
         "pcmpgth    %[Bd],          %[ftmp10],        %[Ad]     \n\t"
-        "por        %[mask],        %[Bd],            %[csth_1] \n\t"
+        POR "       %[mask],        %[Bd],            %[csth_1] \n\t"
         "pmullh     %[Ad],          %[Ad],            %[mask]   \n\t"
         "pmulhuh    %[Ad],          %[ftmp8],         %[Ad]     \n\t"
         "pmullh     %[Ad],          %[Ad],            %[mask]   \n\t"
         "paddh      %[Ad],          %[Ad],            %[Bd]     \n\t"
         "psubh      %[Bd],          %[B],             %[D]      \n\t"
         "pcmpgth    %[Cd],          %[ftmp10],        %[Bd]     \n\t"
-        "por        %[mask],        %[Cd],            %[csth_1] \n\t"
+        POR "       %[mask],        %[Cd],            %[csth_1] \n\t"
         "pmullh     %[Bd],          %[Bd],            %[mask]   \n\t"
         "pmulhuh    %[Bd],          %[ftmp8],         %[Bd]     \n\t"
         "pmullh     %[Bd],          %[Bd],            %[mask]   \n\t"
@@ -114,14 +114,14 @@  static void idct_row_mmi(int16_t *input)
         "paddh      %[Dd],          %[B],             %[D]      \n\t"
         "paddh      %[A],           %[ftmp0],         %[ftmp4]  \n\t"
         "pcmpgth    %[B],           %[ftmp10],        %[A]      \n\t"
-        "por        %[mask],        %[B],             %[csth_1] \n\t"
+        POR "       %[mask],        %[B],             %[csth_1] \n\t"
         "pmullh     %[A],           %[A],             %[mask]   \n\t"
         "pmulhuh    %[A],           %[ftmp8],         %[A]      \n\t"
         "pmullh     %[A],           %[A],             %[mask]   \n\t"
         "paddh      %[A],           %[A],             %[B]      \n\t"
         "psubh      %[B],           %[ftmp0],         %[ftmp4]  \n\t"
         "pcmpgth    %[C],           %[ftmp10],        %[B]      \n\t"
-        "por        %[mask],        %[C],             %[csth_1] \n\t"
+        POR "       %[mask],        %[C],             %[csth_1] \n\t"
         "pmullh     %[B],           %[B],             %[mask]   \n\t"
         "pmulhuh    %[B],           %[ftmp8],         %[B]      \n\t"
         "pmullh     %[B],           %[B],             %[mask]   \n\t"
@@ -131,14 +131,14 @@  static void idct_row_mmi(int16_t *input)
         LOAD_CONST(%[ftmp9], 25080)
         "pmulhh     %[C],           %[ftmp9],         %[ftmp6]  \n\t"
         "pcmpgth    %[D],           %[ftmp10],        %[ftmp2]  \n\t"
-        "por        %[mask],        %[D],             %[csth_1] \n\t"
+        POR "       %[mask],        %[D],             %[csth_1] \n\t"
         "pmullh     %[Ed],          %[ftmp2],         %[mask]   \n\t"
         "pmulhuh    %[Ed],          %[ftmp8],         %[Ed]     \n\t"
         "pmullh     %[Ed],          %[Ed],            %[mask]   \n\t"
         "paddh      %[C],           %[C],             %[Ed]     \n\t"
         "paddh      %[C],           %[C],             %[D]      \n\t"
         "pcmpgth    %[Ed],          %[ftmp10],        %[ftmp6]  \n\t"
-        "por        %[mask],        %[Ed],            %[csth_1] \n\t"
+        POR "       %[mask],        %[Ed],            %[csth_1] \n\t"
         "pmullh     %[ftmp6],       %[ftmp6],         %[mask]   \n\t"
         "pmulhuh    %[D],           %[ftmp8],         %[ftmp6]  \n\t"
         "pmullh     %[D],           %[D],             %[mask]   \n\t"
@@ -193,7 +193,7 @@  static void idct_column_true_mmi(uint8_t *dst, int stride, int16_t *input)
     for (int i = 0; i < 8; ++i)
         temp_value[i] = av_clip_uint8(128 + ((46341 * input[i << 3] + (8 << 16)) >> 20));
     __asm__ volatile (
-        "pxor       %[ftmp10],      %[ftmp10],          %[ftmp10] \n\t"
+        PXOR "      %[ftmp10],      %[ftmp10],          %[ftmp10] \n\t"
         "li         %[tmp0],        0x02                          \n\t"
         "1:                                                       \n\t"
         "ldc1       %[ftmp0],       0x00(%[input])                \n\t"
@@ -213,14 +213,14 @@  static void idct_column_true_mmi(uint8_t *dst, int stride, int16_t *input)
         LOAD_CONST(%[Gd], 1)
         "pmulhh     %[A],           %[ftmp9],           %[ftmp7]  \n\t"
         "pcmpgth    %[C],           %[ftmp10],          %[ftmp1]  \n\t"
-        "por        %[mask],        %[C],               %[Gd]     \n\t"
+        POR "       %[mask],        %[C],               %[Gd]     \n\t"
         "pmullh     %[B],           %[ftmp1],           %[mask]   \n\t"
         "pmulhuh    %[B],           %[ftmp8],           %[B]      \n\t"
         "pmullh     %[B],           %[B],               %[mask]   \n\t"
         "paddh      %[A],           %[A],               %[B]      \n\t"
         "paddh      %[A],           %[A],               %[C]      \n\t"
         "pcmpgth    %[D],           %[ftmp10],          %[ftmp7]  \n\t"
-        "por        %[mask],        %[D],               %[Gd]     \n\t"
+        POR "       %[mask],        %[D],               %[Gd]     \n\t"
         "pmullh     %[Ad],          %[ftmp7],           %[mask]   \n\t"
         "pmulhuh    %[B],           %[ftmp8],           %[Ad]     \n\t"
         "pmullh     %[B],           %[B],               %[mask]   \n\t"
@@ -231,12 +231,12 @@  static void idct_column_true_mmi(uint8_t *dst, int stride, int16_t *input)
         LOAD_CONST(%[ftmp8], 54491)
         LOAD_CONST(%[ftmp9], 36410)
         "pcmpgth    %[Ad],          %[ftmp10],          %[ftmp5]  \n\t"
-        "por        %[mask],        %[Ad],              %[Gd]     \n\t"
+        POR "       %[mask],        %[Ad],              %[Gd]     \n\t"
         "pmullh     %[Cd],          %[ftmp5],           %[mask]   \n\t"
         "pmulhuh    %[C],           %[ftmp9],           %[Cd]     \n\t"
         "pmullh     %[C],           %[C],               %[mask]   \n\t"
         "pcmpgth    %[Bd],          %[ftmp10],          %[ftmp3]  \n\t"
-        "por        %[mask],        %[Bd],              %[Gd]     \n\t"
+        POR "       %[mask],        %[Bd],              %[Gd]     \n\t"
         "pmullh     %[D],           %[ftmp3],           %[mask]   \n\t"
         "pmulhuh    %[D],           %[ftmp8],           %[D]      \n\t"
         "pmullh     %[D],           %[D],               %[mask]   \n\t"
@@ -244,12 +244,12 @@  static void idct_column_true_mmi(uint8_t *dst, int stride, int16_t *input)
         "paddh      %[C],           %[C],               %[Ad]     \n\t"
         "paddh      %[C],           %[C],               %[Bd]     \n\t"
         "pcmpgth    %[Bd],          %[ftmp10],          %[ftmp3]  \n\t"
-        "por        %[mask],        %[Bd],              %[Gd]     \n\t"
+        POR "       %[mask],        %[Bd],              %[Gd]     \n\t"
         "pmullh     %[Cd],          %[ftmp3],           %[mask]   \n\t"
         "pmulhuh    %[D],           %[ftmp9],           %[Cd]     \n\t"
         "pmullh     %[D],           %[D],               %[mask]   \n\t"
         "pcmpgth    %[Ed],          %[ftmp10],          %[ftmp5]  \n\t"
-        "por        %[mask],        %[Ed],              %[Gd]     \n\t"
+        POR "       %[mask],        %[Ed],              %[Gd]     \n\t"
         "pmullh     %[Ad],          %[ftmp5],           %[mask]   \n\t"
         "pmulhuh    %[Ad],          %[ftmp8],           %[Ad]     \n\t"
         "pmullh     %[Ad],          %[Ad],              %[mask]   \n\t"
@@ -260,14 +260,14 @@  static void idct_column_true_mmi(uint8_t *dst, int stride, int16_t *input)
         LOAD_CONST(%[ftmp8], 46341)
         "psubh      %[Ad],          %[A],             %[C]        \n\t"
         "pcmpgth    %[Bd],          %[ftmp10],        %[Ad]       \n\t"
-        "por        %[mask],        %[Bd],            %[Gd]       \n\t"
+        POR "       %[mask],        %[Bd],            %[Gd]       \n\t"
         "pmullh     %[Ad],          %[Ad],            %[mask]     \n\t"
         "pmulhuh    %[Ad],          %[ftmp8],         %[Ad]       \n\t"
         "pmullh     %[Ad],          %[Ad],            %[mask]     \n\t"
         "paddh      %[Ad],          %[Ad],            %[Bd]       \n\t"
         "psubh      %[Bd],          %[B],             %[D]        \n\t"
         "pcmpgth    %[Cd],          %[ftmp10],        %[Bd]       \n\t"
-        "por        %[mask],        %[Cd],            %[Gd]       \n\t"
+        POR "       %[mask],        %[Cd],            %[Gd]       \n\t"
         "pmullh     %[Bd],          %[Bd],            %[mask]     \n\t"
         "pmulhuh    %[Bd],          %[ftmp8],         %[Bd]       \n\t"
         "pmullh     %[Bd],          %[Bd],            %[mask]     \n\t"
@@ -278,7 +278,7 @@  static void idct_column_true_mmi(uint8_t *dst, int stride, int16_t *input)
         LOAD_CONST(%[Ed], 2056)
         "paddh      %[A],           %[ftmp0],         %[ftmp4]    \n\t"
         "pcmpgth    %[B],           %[ftmp10],        %[A]        \n\t"
-        "por        %[mask],        %[B],             %[Gd]       \n\t"
+        POR "       %[mask],        %[B],             %[Gd]       \n\t"
         "pmullh     %[A],           %[A],             %[mask]     \n\t"
         "pmulhuh    %[A],           %[ftmp8],         %[A]        \n\t"
         "pmullh     %[A],           %[A],             %[mask]     \n\t"
@@ -286,7 +286,7 @@  static void idct_column_true_mmi(uint8_t *dst, int stride, int16_t *input)
         "paddh      %[A],           %[A],             %[Ed]       \n\t"
         "psubh      %[B],           %[ftmp0],         %[ftmp4]    \n\t"
         "pcmpgth    %[C],           %[ftmp10],        %[B]        \n\t"
-        "por        %[mask],        %[C],             %[Gd]       \n\t"
+        POR "       %[mask],        %[C],             %[Gd]       \n\t"
         "pmullh     %[B],           %[B],             %[mask]     \n\t"
         "pmulhuh    %[B],           %[ftmp8],         %[B]        \n\t"
         "pmullh     %[B],           %[B],             %[mask]     \n\t"
@@ -297,14 +297,14 @@  static void idct_column_true_mmi(uint8_t *dst, int stride, int16_t *input)
         LOAD_CONST(%[ftmp9], 25080)
         "pmulhh     %[C],           %[ftmp9],         %[ftmp6]    \n\t"
         "pcmpgth    %[D],           %[ftmp10],        %[ftmp2]    \n\t"
-        "por        %[mask],        %[D],             %[Gd]       \n\t"
+        POR "       %[mask],        %[D],             %[Gd]       \n\t"
         "pmullh     %[Ed],          %[ftmp2],         %[mask]     \n\t"
         "pmulhuh    %[Ed],          %[ftmp8],         %[Ed]       \n\t"
         "pmullh     %[Ed],          %[Ed],            %[mask]     \n\t"
         "paddh      %[C],           %[C],             %[Ed]       \n\t"
         "paddh      %[C],           %[C],             %[D]        \n\t"
         "pcmpgth    %[Ed],          %[ftmp10],        %[ftmp6]    \n\t"
-        "por        %[mask],        %[Ed],            %[Gd]       \n\t"
+        POR "       %[mask],        %[Ed],            %[Gd]       \n\t"
         "pmullh     %[D],           %[ftmp6],         %[mask]     \n\t"
         "pmulhuh    %[D],           %[ftmp8],         %[D]        \n\t"
         "pmullh     %[D],           %[D],             %[mask]     \n\t"
@@ -317,12 +317,12 @@  static void idct_column_true_mmi(uint8_t *dst, int stride, int16_t *input)
         "psubh      %[C],           %[B],             %[Ad]       \n\t"
         "psubh      %[B],           %[Bd],            %[D]        \n\t"
         "paddh      %[D],           %[Bd],            %[D]        \n\t"
-        "por        %[mask],        %[ftmp1],         %[ftmp2]    \n\t"
-        "por        %[mask],        %[mask],          %[ftmp3]    \n\t"
-        "por        %[mask],        %[mask],          %[ftmp4]    \n\t"
-        "por        %[mask],        %[mask],          %[ftmp5]    \n\t"
-        "por        %[mask],        %[mask],          %[ftmp6]    \n\t"
-        "por        %[mask],        %[mask],          %[ftmp7]    \n\t"
+        POR "       %[mask],        %[ftmp1],         %[ftmp2]    \n\t"
+        POR "       %[mask],        %[mask],          %[ftmp3]    \n\t"
+        POR "       %[mask],        %[mask],          %[ftmp4]    \n\t"
+        POR "       %[mask],        %[mask],          %[ftmp5]    \n\t"
+        POR "       %[mask],        %[mask],          %[ftmp6]    \n\t"
+        POR "       %[mask],        %[mask],          %[ftmp7]    \n\t"
         "pcmpeqh    %[mask],        %[mask],          %[ftmp10]   \n\t"
         "packushb   %[mask],        %[mask],          %[ftmp10]   \n\t"
         "li         %[tmp1],        0x04                          \n\t"
@@ -361,7 +361,7 @@  static void idct_column_true_mmi(uint8_t *dst, int stride, int16_t *input)
         "packushb   %[ftmp7],       %[ftmp7],         %[ftmp10]   \n\t"
 
         "lwc1       %[Ed],          0x00(%[temp_value])           \n\t"
-        "pand       %[Ed],          %[Ed],            %[mask]     \n\t"
+        PAND "      %[Ed],          %[Ed],            %[mask]     \n\t"
         "paddb      %[ftmp0],       %[ftmp0],         %[Ed]       \n\t"
         "paddb      %[ftmp1],       %[ftmp1],         %[Ed]       \n\t"
         "paddb      %[ftmp2],       %[ftmp2],         %[Ed]       \n\t"
@@ -412,7 +412,7 @@  static void idct_column_false_mmi(uint8_t *dst, int stride, int16_t *input)
     for (int i = 0; i < 8; ++i)
         temp_value[i] = (46341 * input[i << 3] + (8 << 16)) >> 20;
     __asm__ volatile (
-        "pxor       %[ftmp10],      %[ftmp10],          %[ftmp10] \n\t"
+        PXOR "      %[ftmp10],      %[ftmp10],          %[ftmp10] \n\t"
         "li         %[tmp0],        0x02                          \n\t"
         "1:                                                       \n\t"
         "ldc1       %[ftmp0],       0x00(%[input])                \n\t"
@@ -432,14 +432,14 @@  static void idct_column_false_mmi(uint8_t *dst, int stride, int16_t *input)
         LOAD_CONST(%[Gd], 1)
         "pmulhh     %[A],           %[ftmp9],           %[ftmp7]  \n\t"
         "pcmpgth    %[C],           %[ftmp10],          %[ftmp1]  \n\t"
-        "por        %[mask],        %[C],               %[Gd]     \n\t"
+        POR "       %[mask],        %[C],               %[Gd]     \n\t"
         "pmullh     %[B],           %[ftmp1],           %[mask]   \n\t"
         "pmulhuh    %[B],           %[ftmp8],           %[B]      \n\t"
         "pmullh     %[B],           %[B],               %[mask]   \n\t"
         "paddh      %[A],           %[A],               %[B]      \n\t"
         "paddh      %[A],           %[A],               %[C]      \n\t"
         "pcmpgth    %[D],           %[ftmp10],          %[ftmp7]  \n\t"
-        "por        %[mask],        %[D],               %[Gd]     \n\t"
+        POR "       %[mask],        %[D],               %[Gd]     \n\t"
         "pmullh     %[Ad],          %[ftmp7],           %[mask]   \n\t"
         "pmulhuh    %[B],           %[ftmp8],           %[Ad]     \n\t"
         "pmullh     %[B],           %[B],               %[mask]   \n\t"
@@ -450,12 +450,12 @@  static void idct_column_false_mmi(uint8_t *dst, int stride, int16_t *input)
         LOAD_CONST(%[ftmp8], 54491)
         LOAD_CONST(%[ftmp9], 36410)
         "pcmpgth    %[Ad],          %[ftmp10],          %[ftmp5]  \n\t"
-        "por        %[mask],        %[Ad],              %[Gd]     \n\t"
+        POR "       %[mask],        %[Ad],              %[Gd]     \n\t"
         "pmullh     %[Cd],          %[ftmp5],           %[mask]   \n\t"
         "pmulhuh    %[C],           %[ftmp9],           %[Cd]     \n\t"
         "pmullh     %[C],           %[C],               %[mask]   \n\t"
         "pcmpgth    %[Bd],          %[ftmp10],          %[ftmp3]  \n\t"
-        "por        %[mask],        %[Bd],              %[Gd]     \n\t"
+        POR "       %[mask],        %[Bd],              %[Gd]     \n\t"
         "pmullh     %[D],           %[ftmp3],           %[mask]   \n\t"
         "pmulhuh    %[D],           %[ftmp8],           %[D]      \n\t"
         "pmullh     %[D],           %[D],               %[mask]   \n\t"
@@ -463,12 +463,12 @@  static void idct_column_false_mmi(uint8_t *dst, int stride, int16_t *input)
         "paddh      %[C],           %[C],               %[Ad]     \n\t"
         "paddh      %[C],           %[C],               %[Bd]     \n\t"
         "pcmpgth    %[Bd],          %[ftmp10],          %[ftmp3]  \n\t"
-        "por        %[mask],        %[Bd],              %[Gd]     \n\t"
+        POR "       %[mask],        %[Bd],              %[Gd]     \n\t"
         "pmullh     %[Cd],          %[ftmp3],           %[mask]   \n\t"
         "pmulhuh    %[D],           %[ftmp9],           %[Cd]     \n\t"
         "pmullh     %[D],           %[D],               %[mask]   \n\t"
         "pcmpgth    %[Ed],          %[ftmp10],          %[ftmp5]  \n\t"
-        "por        %[mask],        %[Ed],              %[Gd]     \n\t"
+        POR "       %[mask],        %[Ed],              %[Gd]     \n\t"
         "pmullh     %[Ad],          %[ftmp5],           %[mask]   \n\t"
         "pmulhuh    %[Ad],          %[ftmp8],           %[Ad]     \n\t"
         "pmullh     %[Ad],          %[Ad],              %[mask]   \n\t"
@@ -479,14 +479,14 @@  static void idct_column_false_mmi(uint8_t *dst, int stride, int16_t *input)
         LOAD_CONST(%[ftmp8], 46341)
         "psubh      %[Ad],          %[A],             %[C]        \n\t"
         "pcmpgth    %[Bd],          %[ftmp10],        %[Ad]       \n\t"
-        "por        %[mask],        %[Bd],            %[Gd]       \n\t"
+        POR "       %[mask],        %[Bd],            %[Gd]       \n\t"
         "pmullh     %[Ad],          %[Ad],            %[mask]     \n\t"
         "pmulhuh    %[Ad],          %[ftmp8],         %[Ad]       \n\t"
         "pmullh     %[Ad],          %[Ad],            %[mask]     \n\t"
         "paddh      %[Ad],          %[Ad],            %[Bd]       \n\t"
         "psubh      %[Bd],          %[B],             %[D]        \n\t"
         "pcmpgth    %[Cd],          %[ftmp10],        %[Bd]       \n\t"
-        "por        %[mask],        %[Cd],            %[Gd]       \n\t"
+        POR "       %[mask],        %[Cd],            %[Gd]       \n\t"
         "pmullh     %[Bd],          %[Bd],            %[mask]     \n\t"
         "pmulhuh    %[Bd],          %[ftmp8],         %[Bd]       \n\t"
         "pmullh     %[Bd],          %[Bd],            %[mask]     \n\t"
@@ -497,7 +497,7 @@  static void idct_column_false_mmi(uint8_t *dst, int stride, int16_t *input)
         LOAD_CONST(%[Ed], 8)
         "paddh      %[A],           %[ftmp0],         %[ftmp4]    \n\t"
         "pcmpgth    %[B],           %[ftmp10],        %[A]        \n\t"
-        "por        %[mask],        %[B],             %[Gd]       \n\t"
+        POR "       %[mask],        %[B],             %[Gd]       \n\t"
         "pmullh     %[A],           %[A],             %[mask]     \n\t"
         "pmulhuh    %[A],           %[ftmp8],         %[A]        \n\t"
         "pmullh     %[A],           %[A],             %[mask]     \n\t"
@@ -505,7 +505,7 @@  static void idct_column_false_mmi(uint8_t *dst, int stride, int16_t *input)
         "paddh      %[A],           %[A],             %[Ed]       \n\t"
         "psubh      %[B],           %[ftmp0],         %[ftmp4]    \n\t"
         "pcmpgth    %[C],           %[ftmp10],        %[B]        \n\t"
-        "por        %[mask],        %[C],             %[Gd]       \n\t"
+        POR "       %[mask],        %[C],             %[Gd]       \n\t"
         "pmullh     %[B],           %[B],             %[mask]     \n\t"
         "pmulhuh    %[B],           %[ftmp8],         %[B]        \n\t"
         "pmullh     %[B],           %[B],             %[mask]     \n\t"
@@ -516,14 +516,14 @@  static void idct_column_false_mmi(uint8_t *dst, int stride, int16_t *input)
         LOAD_CONST(%[ftmp9], 25080)
         "pmulhh     %[C],           %[ftmp9],         %[ftmp6]    \n\t"
         "pcmpgth    %[D],           %[ftmp10],        %[ftmp2]    \n\t"
-        "por        %[mask],        %[D],             %[Gd]       \n\t"
+        POR "       %[mask],        %[D],             %[Gd]       \n\t"
         "pmullh     %[Ed],          %[ftmp2],         %[mask]     \n\t"
         "pmulhuh    %[Ed],          %[ftmp8],         %[Ed]       \n\t"
         "pmullh     %[Ed],          %[Ed],            %[mask]     \n\t"
         "paddh      %[C],           %[C],             %[Ed]       \n\t"
         "paddh      %[C],           %[C],             %[D]        \n\t"
         "pcmpgth    %[Ed],          %[ftmp10],        %[ftmp6]    \n\t"
-        "por        %[mask],        %[Ed],            %[Gd]       \n\t"
+        POR "       %[mask],        %[Ed],            %[Gd]       \n\t"
         "pmullh     %[D],           %[ftmp6],         %[mask]     \n\t"
         "pmulhuh    %[D],           %[ftmp8],         %[D]        \n\t"
         "pmullh     %[D],           %[D],             %[mask]     \n\t"
@@ -536,12 +536,12 @@  static void idct_column_false_mmi(uint8_t *dst, int stride, int16_t *input)
         "psubh      %[C],           %[B],             %[Ad]       \n\t"
         "psubh      %[B],           %[Bd],            %[D]        \n\t"
         "paddh      %[D],           %[Bd],            %[D]        \n\t"
-        "por        %[mask],        %[ftmp1],         %[ftmp2]    \n\t"
-        "por        %[mask],        %[mask],          %[ftmp3]    \n\t"
-        "por        %[mask],        %[mask],          %[ftmp4]    \n\t"
-        "por        %[mask],        %[mask],          %[ftmp5]    \n\t"
-        "por        %[mask],        %[mask],          %[ftmp6]    \n\t"
-        "por        %[mask],        %[mask],          %[ftmp7]    \n\t"
+        POR "       %[mask],        %[ftmp1],         %[ftmp2]    \n\t"
+        POR "       %[mask],        %[mask],          %[ftmp3]    \n\t"
+        POR "       %[mask],        %[mask],          %[ftmp4]    \n\t"
+        POR "       %[mask],        %[mask],          %[ftmp5]    \n\t"
+        POR "       %[mask],        %[mask],          %[ftmp6]    \n\t"
+        POR "       %[mask],        %[mask],          %[ftmp7]    \n\t"
         "pcmpeqh    %[mask],        %[mask],          %[ftmp10]   \n\t"
         "li         %[tmp1],        0x04                          \n\t"
         "dmtc1      %[tmp1],        %[ftmp8]                      \n\t"
@@ -587,16 +587,16 @@  static void idct_column_false_mmi(uint8_t *dst, int stride, int16_t *input)
         "punpcklbh  %[Cd],          %[Cd],            %[ftmp10]   \n\t"
         "punpcklbh  %[Dd],          %[Dd],            %[ftmp10]   \n\t"
         "ldc1       %[Ed],          0x00(%[temp_value])           \n\t"
-        "pand       %[Ed],          %[Ed],            %[mask]     \n\t"
-        "pnor       %[mask],        %[mask],          %[mask]     \n\t"
-        "pand       %[ftmp0],       %[ftmp0],         %[mask]     \n\t"
-        "pand       %[ftmp1],       %[ftmp1],         %[mask]     \n\t"
-        "pand       %[ftmp2],       %[ftmp2],         %[mask]     \n\t"
-        "pand       %[ftmp3],       %[ftmp3],         %[mask]     \n\t"
-        "pand       %[ftmp4],       %[ftmp4],         %[mask]     \n\t"
-        "pand       %[ftmp5],       %[ftmp5],         %[mask]     \n\t"
-        "pand       %[ftmp6],       %[ftmp6],         %[mask]     \n\t"
-        "pand       %[ftmp7],       %[ftmp7],         %[mask]     \n\t"
+        PAND "      %[Ed],          %[Ed],            %[mask]     \n\t"
+        PNOR "      %[mask],        %[mask],          %[mask]     \n\t"
+        PAND "      %[ftmp0],       %[ftmp0],         %[mask]     \n\t"
+        PAND "      %[ftmp1],       %[ftmp1],         %[mask]     \n\t"
+        PAND "      %[ftmp2],       %[ftmp2],         %[mask]     \n\t"
+        PAND "      %[ftmp3],       %[ftmp3],         %[mask]     \n\t"
+        PAND "      %[ftmp4],       %[ftmp4],         %[mask]     \n\t"
+        PAND "      %[ftmp5],       %[ftmp5],         %[mask]     \n\t"
+        PAND "      %[ftmp6],       %[ftmp6],         %[mask]     \n\t"
+        PAND "      %[ftmp7],       %[ftmp7],         %[mask]     \n\t"
         "paddh      %[ftmp0],       %[ftmp0],         %[A]        \n\t"
         "paddh      %[ftmp1],       %[ftmp1],         %[B]        \n\t"
         "paddh      %[ftmp2],       %[ftmp2],         %[C]        \n\t"
@@ -689,7 +689,7 @@  void ff_vp3_idct_dc_add_mmi(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
     double ftmp[7];
     uint64_t tmp;
     __asm__ volatile (
-        "pxor       %[ftmp0],     %[ftmp0],           %[ftmp0]      \n\t"
+        PXOR "      %[ftmp0],     %[ftmp0],           %[ftmp0]      \n\t"
         "mtc1       %[dc],        %[ftmp5]                          \n\t"
         "pshufh     %[ftmp5],     %[ftmp5],           %[ftmp0]      \n\t"
         "li         %[tmp0],      0x08                              \n\t"
@@ -734,10 +734,10 @@  void ff_put_no_rnd_pixels_l2_mmi(uint8_t *dst, const uint8_t *src1,
             "1:                                                          \n\t"
             MMI_ULDC1(%[ftmp1], %[src1], 0x0)
             MMI_ULDC1(%[ftmp2], %[src2], 0x0)
-            "pxor        %[ftmp3],       %[ftmp1],             %[ftmp2]  \n\t"
-            "pand        %[ftmp3],       %[ftmp3],             %[ftmp4]  \n\t"
+            PXOR "       %[ftmp3],       %[ftmp1],             %[ftmp2]  \n\t"
+            PAND "       %[ftmp3],       %[ftmp3],             %[ftmp4]  \n\t"
             "psrlw       %[ftmp3],       %[ftmp3],             %[ftmp5]  \n\t"
-            "pand        %[ftmp6],       %[ftmp1],             %[ftmp2]  \n\t"
+            PAND "       %[ftmp6],       %[ftmp1],             %[ftmp2]  \n\t"
             "paddw       %[ftmp3],       %[ftmp3],             %[ftmp6]  \n\t"
             "sdc1        %[ftmp3],       0x00(%[dst])                    \n\t"
             PTR_ADDU    "%[src1],        %[src1],              %[stride] \n\t"
diff --git a/libavcodec/mips/vp8dsp_mmi.c b/libavcodec/mips/vp8dsp_mmi.c
index bc774aa365..e12990abc7 100644
--- a/libavcodec/mips/vp8dsp_mmi.c
+++ b/libavcodec/mips/vp8dsp_mmi.c
@@ -38,10 +38,10 @@ 
         "pcmpeqb    %[db_1],    "#src1",        "#src2"             \n\t"   \
         "pmaxub     %[db_2],    "#src1",        "#src2"             \n\t"   \
         "pcmpeqb    %[db_2],    %[db_2],        "#src1"             \n\t"   \
-        "pxor       "#dst",     %[db_2],        %[db_1]             \n\t"
+        PXOR "      "#dst",     %[db_2],        %[db_1]             \n\t"
 
 #define MMI_BTOH(dst_l, dst_r, src)                                         \
-        "pxor       %[db_1],    %[db_1],        %[db_1]             \n\t"   \
+        PXOR "      %[db_1],    %[db_1],        %[db_1]             \n\t"   \
         "pcmpgtb    %[db_2],    %[db_1],        "#src"              \n\t"   \
         "punpcklbh  "#dst_r",   "#src",         %[db_2]             \n\t"   \
         "punpckhbh  "#dst_l",   "#src",         %[db_2]             \n\t"
@@ -84,17 +84,17 @@ 
         "punpcklwd  %[ftmp3],   %[ftmp3],       %[ftmp3]            \n\t"   \
         MMI_PCMPGTUB(%[mask], %[mask], %[ftmp3])                            \
         "pcmpeqw    %[ftmp3],   %[ftmp3],       %[ftmp3]            \n\t"   \
-        "pxor       %[mask],    %[mask],        %[ftmp3]            \n\t"   \
+        PXOR "      %[mask],    %[mask],        %[ftmp3]            \n\t"   \
         /* VP8_MBFILTER */                                                  \
         "li         %[tmp0],    0x80808080                          \n\t"   \
         "dmtc1      %[tmp0],    %[ftmp7]                            \n\t"   \
         "punpcklwd  %[ftmp7],   %[ftmp7],       %[ftmp7]            \n\t"   \
-        "pxor       %[p2],      %[p2],          %[ftmp7]            \n\t"   \
-        "pxor       %[p1],      %[p1],          %[ftmp7]            \n\t"   \
-        "pxor       %[p0],      %[p0],          %[ftmp7]            \n\t"   \
-        "pxor       %[q0],      %[q0],          %[ftmp7]            \n\t"   \
-        "pxor       %[q1],      %[q1],          %[ftmp7]            \n\t"   \
-        "pxor       %[q2],      %[q2],          %[ftmp7]            \n\t"   \
+        PXOR "      %[p2],      %[p2],          %[ftmp7]            \n\t"   \
+        PXOR "      %[p1],      %[p1],          %[ftmp7]            \n\t"   \
+        PXOR "      %[p0],      %[p0],          %[ftmp7]            \n\t"   \
+        PXOR "      %[q0],      %[q0],          %[ftmp7]            \n\t"   \
+        PXOR "      %[q1],      %[q1],          %[ftmp7]            \n\t"   \
+        PXOR "      %[q2],      %[q2],          %[ftmp7]            \n\t"   \
         "psubsb     %[ftmp4],   %[p1],          %[q1]               \n\t"   \
         "psubb      %[ftmp5],   %[q0],          %[p0]               \n\t"   \
         MMI_BTOH(%[ftmp1],  %[ftmp0],  %[ftmp5])                            \
@@ -109,8 +109,8 @@ 
         "paddh      %[ftmp1],   %[ftmp3],       %[ftmp1]            \n\t"   \
         /* Combine left and right part */                                   \
         "packsshb   %[ftmp1],   %[ftmp0],       %[ftmp1]            \n\t"   \
-        "pand       %[ftmp1],   %[ftmp1],       %[mask]             \n\t"   \
-        "pand       %[ftmp2],   %[ftmp1],       %[hev]              \n\t"   \
+        PAND "      %[ftmp1],   %[ftmp1],       %[mask]             \n\t"   \
+        PAND "      %[ftmp2],   %[ftmp1],       %[hev]              \n\t"   \
         "li         %[tmp0],    0x04040404                          \n\t"   \
         "dmtc1      %[tmp0],    %[ftmp0]                            \n\t"   \
         "punpcklwd  %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"   \
@@ -129,8 +129,8 @@ 
         "paddsb     %[p0],      %[p0],          %[ftmp4]            \n\t"   \
         /* filt_val &= ~hev */                                              \
         "pcmpeqw    %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"   \
-        "pxor       %[hev],     %[hev],         %[ftmp0]            \n\t"   \
-        "pand       %[ftmp1],   %[ftmp1],       %[hev]              \n\t"   \
+        PXOR "      %[hev],     %[hev],         %[ftmp0]            \n\t"   \
+        PAND "      %[ftmp1],   %[ftmp1],       %[hev]              \n\t"   \
         MMI_BTOH(%[ftmp5],  %[ftmp6],  %[ftmp1])                            \
         "li         %[tmp0],    0x07                                \n\t"   \
         "dmtc1      %[tmp0],    %[ftmp2]                            \n\t"   \
@@ -151,9 +151,9 @@ 
         /* Combine left and right part */                                   \
         "packsshb   %[ftmp4],   %[ftmp3],       %[ftmp4]            \n\t"   \
         "psubsb     %[q0],      %[q0],          %[ftmp4]            \n\t"   \
-        "pxor       %[q0],      %[q0],          %[ftmp7]            \n\t"   \
+        PXOR "      %[q0],      %[q0],          %[ftmp7]            \n\t"   \
         "paddsb     %[p0],      %[p0],          %[ftmp4]            \n\t"   \
-        "pxor       %[p0],      %[p0],          %[ftmp7]            \n\t"   \
+        PXOR "      %[p0],      %[p0],          %[ftmp7]            \n\t"   \
         "li         %[tmp0],    0x00120012                          \n\t"   \
         "dmtc1      %[tmp0],    %[ftmp1]                            \n\t"   \
         "punpcklwd  %[ftmp1],   %[ftmp1],       %[ftmp1]            \n\t"   \
@@ -168,9 +168,9 @@ 
         /* Combine left and right part */                                   \
         "packsshb   %[ftmp4],   %[ftmp3],       %[ftmp4]            \n\t"   \
         "psubsb     %[q1],      %[q1],          %[ftmp4]            \n\t"   \
-        "pxor       %[q1],      %[q1],          %[ftmp7]            \n\t"   \
+        PXOR "      %[q1],      %[q1],          %[ftmp7]            \n\t"   \
         "paddsb     %[p1],      %[p1],          %[ftmp4]            \n\t"   \
-        "pxor       %[p1],      %[p1],          %[ftmp7]            \n\t"   \
+        PXOR "      %[p1],      %[p1],          %[ftmp7]            \n\t"   \
         "li         %[tmp0],    0x03                                \n\t"   \
         "dmtc1      %[tmp0],    %[ftmp1]                            \n\t"   \
         /* Right part */                                                    \
@@ -186,9 +186,9 @@ 
         /* Combine left and right part */                                   \
         "packsshb   %[ftmp4],   %[ftmp3],       %[ftmp4]            \n\t"   \
         "psubsb     %[q2],      %[q2],          %[ftmp4]            \n\t"   \
-        "pxor       %[q2],      %[q2],          %[ftmp7]            \n\t"   \
+        PXOR "      %[q2],      %[q2],          %[ftmp7]            \n\t"   \
         "paddsb     %[p2],      %[p2],          %[ftmp4]            \n\t"   \
-        "pxor       %[p2],      %[p2],          %[ftmp7]            \n\t"
+        PXOR "      %[p2],      %[p2],          %[ftmp7]            \n\t"
 
 #define PUT_VP8_EPEL4_H6_MMI(src, dst)                                      \
         MMI_ULWC1(%[ftmp1], src, 0x00)                                      \
@@ -997,7 +997,7 @@  void ff_vp8_luma_dc_wht_mmi(int16_t block[4][4][16], int16_t dc[16])
     block[3][3][0] = (dc[12] - dc[15] + 3 - dc[13] + dc[14]) >> 3;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         MMI_SDC1(%[ftmp0], %[dc], 0x00)
         MMI_SDC1(%[ftmp0], %[dc], 0x08)
         MMI_SDC1(%[ftmp0], %[dc], 0x10)
@@ -1114,7 +1114,7 @@  void ff_vp8_idct_add_mmi(uint8_t *dst, int16_t block[16], ptrdiff_t stride)
     ff_ph_22a3_u.i = 0x22a322a322a322a3ULL;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         MMI_LDC1(%[ftmp1], %[block], 0x00)
         MMI_LDC1(%[ftmp2], %[block], 0x08)
         MMI_LDC1(%[ftmp3], %[block], 0x10)
@@ -1280,7 +1280,7 @@  void ff_vp8_idct_dc_add_mmi(uint8_t *dst, int16_t block[16], ptrdiff_t stride)
     block[0] = 0;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "mtc1       %[dc],      %[ftmp5]                            \n\t"
         MMI_LWC1(%[ftmp1], %[dst0], 0x00)
         MMI_LWC1(%[ftmp2], %[dst1], 0x00)
@@ -1604,7 +1604,7 @@  void ff_put_vp8_epel16_h4_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[15] = cm[(filter[2] * src[15] - filter[1] * src[14] + filter[3] * src[16] - filter[4] * src[17] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -1680,7 +1680,7 @@  void ff_put_vp8_epel8_h4_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[7] = cm[(filter[2] * src[7] - filter[1] * src[ 6] + filter[3] * src[8] - filter[4] * src[9] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -1745,7 +1745,7 @@  void ff_put_vp8_epel4_h4_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[3] = cm[(filter[2] * src[3] - filter[1] * src[ 2] + filter[3] * src[4] - filter[4] * src[5] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -1826,7 +1826,7 @@  void ff_put_vp8_epel16_h6_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[15] = cm[(filter[2]*src[15] - filter[1]*src[14] + filter[0]*src[13] + filter[3]*src[16] - filter[4]*src[17] + filter[5]*src[18] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -1906,7 +1906,7 @@  void ff_put_vp8_epel8_h6_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[7] = cm[(filter[2]*src[7] - filter[1]*src[ 6] + filter[0]*src[ 5] + filter[3]*src[8] - filter[4]*src[9] + filter[5]*src[10] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -1976,7 +1976,7 @@  void ff_put_vp8_epel4_h6_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[3] = cm[(filter[2]*src[3] - filter[1]*src[ 2] + filter[0]*src[ 1] + filter[3]*src[4] - filter[4]*src[5] + filter[5]*src[ 6] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -2054,7 +2054,7 @@  void ff_put_vp8_epel16_v4_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[15] = cm[(filter[2] * src[15] - filter[1] * src[15-srcstride] + filter[3] * src[15+srcstride] - filter[4] * src[15+2*srcstride] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -2131,7 +2131,7 @@  void ff_put_vp8_epel8_v4_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[7] = cm[(filter[2] * src[7] - filter[1] * src[7-srcstride] + filter[3] * src[7+srcstride] - filter[4] * src[7+2*srcstride] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -2198,7 +2198,7 @@  void ff_put_vp8_epel4_v4_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[3] = cm[(filter[2] * src[3] - filter[1] * src[3-srcstride] + filter[3] * src[3+srcstride] - filter[4] * src[3+2*srcstride] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -2280,7 +2280,7 @@  void ff_put_vp8_epel16_v6_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[15] = cm[(filter[2]*src[15] - filter[1]*src[15-srcstride] + filter[0]*src[15-2*srcstride] + filter[3]*src[15+srcstride] - filter[4]*src[15+2*srcstride] + filter[5]*src[15+3*srcstride] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -2362,7 +2362,7 @@  void ff_put_vp8_epel8_v6_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[7] = cm[(filter[2]*src[7] - filter[1]*src[7-srcstride] + filter[0]*src[7-2*srcstride] + filter[3]*src[7+srcstride] - filter[4]*src[7+2*srcstride] + filter[5]*src[7+3*srcstride] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -2434,7 +2434,7 @@  void ff_put_vp8_epel4_v6_mmi(uint8_t *dst, ptrdiff_t dststride, uint8_t *src,
     dst[3] = cm[(filter[2]*src[3] - filter[1]*src[3-srcstride] + filter[0]*src[3-2*srcstride] + filter[3]*src[3+srcstride] - filter[4]*src[3+2*srcstride] + filter[5]*src[3+3*srcstride] + 64) >> 7];
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x07                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
 
@@ -2974,7 +2974,7 @@  void ff_put_vp8_bilinear16_h_mmi(uint8_t *dst, ptrdiff_t dstride, uint8_t *src,
     dst[15] = (a * src[15] + b * src[16] + 4) >> 3;
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x03                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
         "pshufh     %[a],       %[a],           %[ftmp0]            \n\t"
@@ -3043,7 +3043,7 @@  void ff_put_vp8_bilinear16_v_mmi(uint8_t *dst, ptrdiff_t dstride, uint8_t *src,
     dst[7] = (c * src[7] + d * src[7 + sstride] + 4) >> 3;
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x03                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
         "pshufh     %[c],       %[c],           %[ftmp0]            \n\t"
@@ -3146,7 +3146,7 @@  void ff_put_vp8_bilinear8_h_mmi(uint8_t *dst, ptrdiff_t dstride, uint8_t *src,
     dst[7] = (a * src[7] + b * src[8] + 4) >> 3;
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x03                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
         "pshufh     %[a],       %[a],           %[ftmp0]            \n\t"
@@ -3209,7 +3209,7 @@  void ff_put_vp8_bilinear8_v_mmi(uint8_t *dst, ptrdiff_t dstride, uint8_t *src,
     dst[7] = (c * src[7] + d * src[7 + sstride] + 4) >> 3;
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x03                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
         "pshufh     %[c],       %[c],           %[ftmp0]            \n\t"
@@ -3303,7 +3303,7 @@  void ff_put_vp8_bilinear4_h_mmi(uint8_t *dst, ptrdiff_t dstride, uint8_t *src,
     dst[3] = (a * src[3] + b * src[4] + 4) >> 3;
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x03                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
         "pshufh     %[a],       %[a],           %[ftmp0]            \n\t"
@@ -3363,7 +3363,7 @@  void ff_put_vp8_bilinear4_v_mmi(uint8_t *dst, ptrdiff_t dstride, uint8_t *src,
     dst[3] = (c * src[3] + d * src[3 + sstride] + 4) >> 3;
     */
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],       %[ftmp0]            \n\t"
         "li         %[tmp0],    0x03                                \n\t"
         "mtc1       %[tmp0],    %[ftmp4]                            \n\t"
         "pshufh     %[c],       %[c],           %[ftmp0]            \n\t"
diff --git a/libavcodec/mips/vp9_mc_mmi.c b/libavcodec/mips/vp9_mc_mmi.c
index 495cac3d0b..3e9df4d504 100644
--- a/libavcodec/mips/vp9_mc_mmi.c
+++ b/libavcodec/mips/vp9_mc_mmi.c
@@ -83,7 +83,7 @@  static void convolve_horiz_mmi(const uint8_t *src, int32_t src_stride,
     dst_stride -= w;
     __asm__ volatile (
         "move       %[tmp1],    %[width]                   \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],    %[ftmp0]      \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],    %[ftmp0]      \n\t"
         MMI_ULDC1(%[filter1], %[filter], 0x00)
         MMI_ULDC1(%[filter2], %[filter], 0x08)
         "li         %[tmp0],    0x07                       \n\t"
@@ -154,7 +154,7 @@  static void convolve_vert_mmi(const uint8_t *src, int32_t src_stride,
     dst_stride -= w;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],    %[ftmp0],   %[ftmp0]      \n\t"
+        PXOR "      %[ftmp0],    %[ftmp0],   %[ftmp0]      \n\t"
         MMI_ULDC1(%[ftmp4], %[filter], 0x00)
         MMI_ULDC1(%[ftmp5], %[filter], 0x08)
         "punpcklwd  %[filter10], %[ftmp4],   %[ftmp4]      \n\t"
@@ -242,7 +242,7 @@  static void convolve_avg_horiz_mmi(const uint8_t *src, int32_t src_stride,
 
     __asm__ volatile (
         "move       %[tmp1],    %[width]                   \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],    %[ftmp0]      \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],    %[ftmp0]      \n\t"
         MMI_ULDC1(%[filter1], %[filter], 0x00)
         MMI_ULDC1(%[filter2], %[filter], 0x08)
         "li         %[tmp0],    0x07                       \n\t"
@@ -323,7 +323,7 @@  static void convolve_avg_vert_mmi(const uint8_t *src, int32_t src_stride,
     dst_stride -= w;
 
     __asm__ volatile (
-        "pxor       %[ftmp0],    %[ftmp0],   %[ftmp0]      \n\t"
+        PXOR "      %[ftmp0],    %[ftmp0],   %[ftmp0]      \n\t"
         MMI_ULDC1(%[ftmp4], %[filter], 0x00)
         MMI_ULDC1(%[ftmp5], %[filter], 0x08)
         "punpcklwd  %[filter10], %[ftmp4],   %[ftmp4]      \n\t"
@@ -419,7 +419,7 @@  static void convolve_avg_mmi(const uint8_t *src, int32_t src_stride,
 
     __asm__ volatile (
         "move       %[tmp1],    %[width]                  \n\t"
-        "pxor       %[ftmp0],   %[ftmp0],   %[ftmp0]      \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],   %[ftmp0]      \n\t"
         "li         %[tmp0],    0x10001                   \n\t"
         "dmtc1      %[tmp0],    %[ftmp3]                  \n\t"
         "punpcklhw  %[ftmp3],   %[ftmp3],   %[ftmp3]      \n\t"
diff --git a/libavcodec/mips/wmv2dsp_mmi.c b/libavcodec/mips/wmv2dsp_mmi.c
index 1a6781ae77..c1bb94eeef 100644
--- a/libavcodec/mips/wmv2dsp_mmi.c
+++ b/libavcodec/mips/wmv2dsp_mmi.c
@@ -106,7 +106,7 @@  void ff_wmv2_idct_add_mmi(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
         wmv2_idct_col_mmi(block + i);
 
     __asm__ volatile (
-        "pxor       %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
+        PXOR "      %[ftmp0],   %[ftmp0],   %[ftmp0]                    \n\t"
 
         // low 4 loop
         MMI_LDC1(%[ftmp1], %[block], 0x00)
diff --git a/libavutil/mips/mmiutils.h b/libavutil/mips/mmiutils.h
index 7991e14e84..a030aa475f 100644
--- a/libavutil/mips/mmiutils.h
+++ b/libavutil/mips/mmiutils.h
@@ -29,6 +29,28 @@ 
 #include "libavutil/mem_internal.h"
 #include "libavutil/mips/asmdefs.h"
 
+
+/* Alias for instruction mnemonics, clang only support new one */
+#if defined(__clang__)
+#define PAND    "pand"
+#define POR     "por"
+#define PNOR    "pnor"
+#define PXOR    "pxor"
+#define SSRLD   "ssrld"
+#define SSRLW   "ssrlw"
+#define SSLLD   "sslld"
+#define SSLLW   "ssllw"
+#else
+#define POR     "or"
+#define PNOR    "nor"
+#define PXOR    "xor"
+#define PAND    "and"
+#define SSRLD   "dsrl"
+#define SSRLW   "srl"
+#define SSLLD   "dsll"
+#define SSLLW   "sll"
+#endif
+
 /*
  * These were used to define temporary registers for MMI marcos
  * however now we're using $at. They're theoretically unnecessary