Re: [PATCH v2, 0/2] This patchset add Read-only(Ro) request for capture queue

From: yunfei.dong <yunfei.dong@mediatek.com>
To: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Cc: Mark Rutland <mark.rutland@arm.com>,
	devicetree <devicetree@vger.kernel.org>,
	Tiffany Lin <tiffany.lin@mediatek.com>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Tomasz Figa <tfiga@chromium.org>,
	Hans Verkuil <hverkuil@xs4all.nl>,
	Rob Herring <robh+dt@kernel.org>,
	"moderated list:ARM/Mediatek SoC support"
	<linux-mediatek@lists.infradead.org>,
	Matthias Brugger <matthias.bgg@gmail.com>,
	Mauro Carvalho Chehab <mchehab@kernel.org>,
	linux-arm-kernel <linux-arm-kernel@lists.infradead.org>,
	linux-media <linux-media@vger.kernel.org>
Subject: Re: [PATCH v2, 0/2] This patchset add Read-only(Ro) request for capture queue
Date: Mon, 29 Jun 2020 14:47:29 +0800	[thread overview]
Message-ID: <1593413249.19566.23.camel@mhfsdcap03> (raw)
In-Reply-To: <CAAEAJfBtuZUn-LxiwCZ75rwT_oEuM-_QEXCH4-vFhV6X+8=4SA@mail.gmail.com>

Hi Ezequiel Garcia,

Thanks for your advice.

1: An upstream driver using the feature would be important, but it would
be nice to also have: an open-source userspace application,
>>> Ro request is one feature based on media request. In userspace,
application can use it the same as with request. Ro request and request
are separated in kernel space with different ctrl handler.

2:and a proper explanation in the stateless decoder interface
specification.
>>> Whether add the explanation in cover-letter is enough ? In fact,
these patches just add one new function v4l2_check_ro_ext_ctrls(), other
changes separate Ro request and request.

Thanks again.

Best Regards,
Yunfei Dong

On Tue, 2020-06-23 at 15:43 -0300, Ezequiel Garcia wrote:
> Hi Yunfei,
> 
> Thanks for the patch.
> 
> On Sun, 21 Jun 2020 at 22:55, Yunfei Dong <yunfei.dong@mediatek.com> wrote:
> >
> > User driver need to get HDR10+ information for each capture buffer;
> > For some encoder cases, user driver need to get encoded message for
> > each frame. So add support read-only(Ro) request for capture queue.
> >
> > Ro request mean that user driver just can get ext ctrls, set ext ctrls
> > is not not allowed. Ro Request also can be used in output queue.
> >
> > There is not upstream driver to use this feature at now, but we are
> > developing internal driver to use it. If it is ready, we will try to
> > upstream vdec/venc driver based on this feature.
> >
> 
> An upstream driver using the feature would be important, but it would
> be nice to also have: an open-source userspace application,
> and a proper explanation in the stateless decoder interface specification.
> 
> Thanks,
> Ezequiel
> 
> > Change compared to v1:
> > -change commit message of patch 01/02
> > -change commit message of patch 02/02
> >
> > Yunfei Dong (2):
> >   media: v4l UAPI: add V4L2_BUF_CAP_SUPPORTS_RO_REQUESTS
> >   media: v4l: Add Ro request api for capture queue
> >
> >  .../media/v4l/vidioc-reqbufs.rst              |   4 +
> >  .../media/common/videobuf2/videobuf2-v4l2.c   |   7 ++
> >  drivers/media/mc/mc-request.c                 |  10 +-
> >  drivers/media/v4l2-core/v4l2-ctrls.c          | 107 +++++++++++++++---
> >  drivers/media/v4l2-core/v4l2-ioctl.c          |  22 ++++
> >  drivers/media/v4l2-core/v4l2-mem2mem.c        |  19 ++--
> >  include/media/v4l2-ctrls.h                    |  22 +++-
> >  include/media/v4l2-fh.h                       |   2 +
> >  include/media/videobuf2-core.h                |   2 +
> >  include/uapi/linux/videodev2.h                |   1 +
> >  10 files changed, 158 insertions(+), 38 deletions(-)
> >

_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek