[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1567522916.5229.11.camel@pengutronix.de>
Date: Tue, 03 Sep 2019 17:01:56 +0200
From: Philipp Zabel <p.zabel@...gutronix.de>
To: Jonas Karlman <jonas@...boo.se>,
Ezequiel Garcia <ezequiel@...labora.com>
Cc: Mauro Carvalho Chehab <mchehab@...nel.org>,
Hans Verkuil <hverkuil@...all.nl>,
Boris Brezillon <boris.brezillon@...labora.com>,
Paul Kocialkowski <paul.kocialkowski@...tlin.com>,
"linux-media@...r.kernel.org" <linux-media@...r.kernel.org>,
"linux-rockchip@...ts.infradead.org"
<linux-rockchip@...ts.infradead.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [RFC 08/12] media: hantro: Fix H264 decoding of field encoded
content
On Tue, 2019-09-03 at 14:02 +0000, Jonas Karlman wrote:
> On 2019-09-03 15:21, Philipp Zabel wrote:
> > On Sun, 2019-09-01 at 12:45 +0000, Jonas Karlman wrote:
> > > This need code cleanup and formatting
> > >
> > > Signed-off-by: Jonas Karlman <jonas@...boo.se>
> >
> > The previous patches all work, but this patch breaks decoding of
> > progressive content for me (i.MX8MQ with FFmpeg based on Ezequiel's
> > branch).
>
> Please try with ffmpeg based on my v4l2-request-hwaccel-4.0.4-hantro branch at [1],
> up to and including the commit "HACK: add dpb flags for reference usage and field picture".
> That commit adds code to set reference flags needed by the changes in this patch.
>
> There is probably also some other minor difference between our two ffmpeg branches.
> I have not observed any issues with progressive content with this patch and my ffmpeg branch (on a RK3288 device).
> Some H264 reference samples do have visual issues after this patch, however all my real world samples does seem to work.
>
> My branch use libudev to probe media/video devices and needs to be configured with:
> --enable-v4l2-request --enable-libudev --enable-libdrm
>
> [1] https://github.com/Kwiboo/FFmpeg/commits/v4l2-request-hwaccel-4.0.4-hantro
I hadn't realized that this is a backwards incompatible change. With
your branch rebased onto n4.2, and this patch applied, decoding seems to
work.
regards
Philipp
Powered by blists - more mailing lists