[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <e0e144be-5cf9-4a79-a602-2ab2b7cd9aa1@timsurber.de>
Date: Wed, 5 Mar 2025 11:41:56 +0100
From: Tim Surber <me@...surber.de>
To: Dmitry Osipenko <dmitry.osipenko@...labora.com>,
Hans Verkuil <hverkuil@...all.nl>,
Shreeya Patel <shreeya.patel@...labora.com>, Heiko Stuebner
<heiko@...ech.de>, Mauro Carvalho Chehab <mchehab@...nel.org>,
Rob Herring <robh@...nel.org>, Krzysztof Kozlowski <krzk+dt@...nel.org>,
Conor Dooley <conor+dt@...nel.org>, jose.abreu@...opsys.com,
nelson.costa@...opsys.com, shawn.wen@...k-chips.com,
nicolas.dufresne@...labora.com,
Sebastian Reichel <sebastian.reichel@...labora.com>
Cc: kernel@...labora.com, linux-media@...r.kernel.org,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-rockchip@...ts.infradead.org,
Christophe JAILLET <christophe.jaillet@...adoo.fr>,
Diederik de Haas <didi.debian@...ow.org>
Subject: Re: [PATCH v13 0/6] Add Synopsys DesignWare HDMI RX Controller
Hi,
so the 4:4:4 issue was just a gstreamer bug and it worked when I applied
an experimental fix [1].
So everything works for me using the default EDID now.
Tested-by: Tim Surber <me@...surber.de>
[1]:
https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/8534
On 3/5/25 10:09, Dmitry Osipenko wrote:
> On 3/5/25 12:03, Dmitry Osipenko wrote:
>> On 3/5/25 11:10, Hans Verkuil wrote:
>>> On 05/03/2025 01:59, Tim Surber wrote:
>>>> Hi Dmitry,
>>>>
>>>> I did some more testing. That the Apple TV did not work was a bit
>>>> misleading.
>>>>
>>>> It was just, that the Apple TV defaulted to 4:4:4 Chroma which does not
>>>> work at all for me. (The same happens using the vendor driver).
>>>>
>>>> When I changed the EDID to match the vendor driver the HDMI handshake
>>>> happened with 4:2:0 chroma, where I could verify even 4k60fps using your
>>>> driver, nice!
>>>>
>>>> So the remaining problems I see are:
>>>> - 4:4:4 chroma not working in any resolution
>>>> - 4:2:2 and RGB not working in 4k60fps (is this a hardware limitation?)
>>>>
>>>> A possible workaround could be to disable these non supported formats in
>>>> the default EDID.
>>> I would like to merge this driver this week, since otherwise it will likely
>>> slip to v6.16. So if there is a working EDID, perhaps it can be used for now,
>>> and later on it can be patched if there is a better EDID.
>>>
>>> Would this EDID work? Tim, can you try this?
>>>
>>> v4l2-ctl --set-edid type=hdmi-4k-600mhz,ycbcr444,ycbcr422
>>>
>>> Alternatively, if there is indeed a HW limitation that prevents 4kp60 to work,
>>> try this:
>>>
>>> v4l2-ctl --set-edid type=hdmi-4k-300mhz,ycbcr444,ycbcr422
>>>
>>> Whichever of the two works is what we can use as default EDID.
>>
>> Disabling 444 and 422 is an option. Though, they work on my setup at
>> 4k@....
>>
>> In general, it often a challenge to get 4k@60p properly with any of
>> these small board devices. 4k@60p works only using a short HDMI cable
>> for me. Also, not everyone aware that the micro HDMI adapter needs to be
>> compliant with HDMI 2.0 for 4k@60, that's why 300MHz is the default.
>>
>> Will be nice to have the good EDID enabled by default in the defconfig.
>> Dealing with problems like that will be a headache for majority of
>> people, IMO.
>
> BTW, I don't see it as a blocker. Driver works in general, new issues
> can be resolved with additional patches.
>
Powered by blists - more mailing lists