[FFmpeg-devel] [PATCH] lavc/mediacodec: add hwaccel support
Benoit Fouet
benoit.fouet at free.fr
Fri Jul 8 17:37:30 EEST 2016
On 08/07/2016 15:40, Matthieu Bouron wrote:
> On Fri, Jul 08, 2016 at 01:21:48PM +0200, Benoit Fouet wrote:
>> Hi,
>>
>> On 07/07/2016 17:43, Matthieu Bouron wrote:
>>
>> [...]
>>
>>> 0001-lavc-add-mediacodec-hwaccel-support.patch
>>>
>>>
>>> From 9bb86990f0f7a26d25878a771f5977ae83d14769 Mon Sep 17 00:00:00 2001
>>> From: Matthieu Bouron<matthieu.bouron at stupeflix.com>
>>> Date: Fri, 11 Mar 2016 17:21:04 +0100
>>> Subject: [PATCH] lavc: add mediacodec hwaccel support
>>>
>>> ---
>>> configure | 1 +
>>> libavcodec/Makefile | 6 +-
>>> libavcodec/allcodecs.c | 1 +
>>> libavcodec/mediacodec.c | 133 ++++++++++++++++++++
>>> libavcodec/mediacodec.h | 88 +++++++++++++
>>> libavcodec/mediacodec_surface.c | 66 ++++++++++
>>> libavcodec/mediacodec_surface.h | 31 +++++
>>> libavcodec/mediacodec_wrapper.c | 5 +-
>>> libavcodec/mediacodecdec.c | 270 +++++++++++++++++++++++++++++++++-------
>>> libavcodec/mediacodecdec.h | 17 +++
>>> libavcodec/mediacodecdec_h264.c | 44 ++++++-
>>> libavcodec/version.h | 2 +-
>>> libavutil/pixdesc.c | 4 +
>>> libavutil/pixfmt.h | 2 +
>>> 14 files changed, 611 insertions(+), 59 deletions(-)
>>> create mode 100644 libavcodec/mediacodec.c
>>> create mode 100644 libavcodec/mediacodec.h
>>> create mode 100644 libavcodec/mediacodec_surface.c
>>> create mode 100644 libavcodec/mediacodec_surface.h
[...]
>>> diff --git a/libavcodec/mediacodecdec.h b/libavcodec/mediacodecdec.h
>>> index 646b628..8613352 100644
>>> --- a/libavcodec/mediacodecdec.h
>>> +++ b/libavcodec/mediacodecdec.h
>>> @@ -34,12 +34,17 @@
>>> typedef struct MediaCodecDecContext {
>>> + volatile int refcount;
>>> +
>> I don't think this needs to be marked volatile
> The avpriv_atomic_[...] functions take a volatile int *ptr. Also there are
> examples in our code base that declare the atomic as volatile int (see
> libavutil/buffer_internal.h, libavcodec/mmaldec.c for example). I can
> be missing something though.
as the atomic api is working on addresses, it should not make any
difference.
I don't really mind, though, keep it this way if you prefer
I'm not a volatile fan boy, so I try to refrain from using them when
it's not strictly needed :-)
[...]
> If you are OK with my comments, I will push the patch.
Yes, sure.
Cheers,
--
Ben
More information about the ffmpeg-devel
mailing list