Linux-Rockchip Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Piotr Oniszczuk <piotr.oniszczuk@gmail.com>
To: Andy Yan <andy.yan@rock-chips.com>
Cc: Linux-rockchip@lists.infradead.org
Subject: Re: vop2: 3566 vs 3568: green screen on 3566; on 3568 all ok...
Date: Tue, 26 Dec 2023 12:22:47 +0100	[thread overview]
Message-ID: <DEF50A27-6DF2-4B83-8F30-A47AA3878BC6@gmail.com> (raw)
In-Reply-To: <ef6186aa-f600-4d3a-be0d-155ff3336a27@rock-chips.com>

Andy,

pls see inline

> Wiadomość napisana przez Andy Yan <andy.yan@rock-chips.com> w dniu 26.12.2023, o godz. 02:41:
> 
> Hi Piotr,
> 
> On 12/25/23 20:14, Piotr Oniszczuk wrote:
>> Andy,
>> Thx for looking on this!
>> Before passing dri state logs - let me give some context why i’m asking here for help.
>> I’m developing 2 distros with goal to offer user convenient run of multimedia players (Kodi and MythTV): appliance and general distro.
>> Goal is out of box run of Kodi/MythTV in 2 rendering modes: EGL DAMBuf and DRM direct planes.
>> Currently i have out-of-box (no any SoC specific configs in Kodi/MythTV) for: rk3328/rk3399/H6/H313/H616/rpi3/rpi4.
>> Issues ia have are with: rk3566 and rk3568
>>  
> 
> Yes, it is a cool project, I run it on rk3566-box-demo for a while last year.

Many thx for kind words! 

> 
>> Vanilla 6.6.8 kernel on 3568 not offers me well working out-of-box Kodi/MythTV in DRM direct planes - but i was able to patch vop2 to get such out-of-box operation (pls see patch below).
>> I have also issue with 3566 - where i also developed patch (also see below) - but i’m not able to get 3566 fully working (EGL DMAbuf is ok; in DRM Planes GUI is ok, but video playback gives part screen green part black; green/black border is vibrantly changing at playback)
>> Logs:
> 
> 
> The dri/state dump already indicates the root case of the problem:
> 
> For rk3566/rk3568, the Cluster windows only support AFBC based RGB or YUV format,
> linear formats(such as NV12 you used for Cluster0-win0 on rk3566) are not supported, these
> linear formats are only support by Esmart(RGB/YUV),smart(RGB only) windows.
> 
> Compared with rk3568 you use, the NV12 is committed to Esmart1, and XRGB UI is
> committed to Esmart0.
> But on rk3566, there are only three windows/plane you can use, if you can only decoder
> linear NV12, it should be committed to Esmart0, and use Smart0 for UI.
> 
> 
> But when the vop2 drivers first lands, it mistakenly add linear formats(NV12/NV16) to
> Cluster windows format lists, this will mislead user space.
> 
> I have a patch series to fix this that already be merged in linux 6.7[0]
> 
> Feel free to let me know if you still have problem.
> 
> [0] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=01d5a75370a60c3a8d691347ae6ebb2a9f8dc44a

Perfect.

Above patch + vop2_win_data rk3568_vop_win_data[] reordering to:

Esmart1-win0
Esmart0-win0
Smart0-win0
Smart1-win0
Cluster0-win0
Cluster1-win0

makes Kodi and MythTV working out-of-box with EGL DMAbuf and DRM planes HW decoded video rendering :-)

So with your help we have single 6.6.8 kernel offering EGL DMAbuf and DRM planes HW decoded video rendering on rk3328/rk3399/rk3566/rk3568/H6/H313/H616/H618/rpi3/rpi4

once again: many thx

(p.s: it looks like only thing missing in 356x to be perfect open-source multimedia soc is: lack of rkvdec2 for hevc/vp9 hw decoding. Are there any plans from rock-chips to help foss to bring-up mainline kernel support for rkvdec2?)
   


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

  reply	other threads:[~2023-12-26 11:23 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-12-24 12:37 vop2: 3566 vs 3568: green screen on 3566; on 3568 all ok Piotr Oniszczuk
2023-12-25  1:40 ` Andy Yan
2023-12-25 12:14   ` Piotr Oniszczuk
2023-12-26  1:41     ` Andy Yan
2023-12-26 11:22       ` Piotr Oniszczuk [this message]
2023-12-27  7:25         ` Kever Yang
2023-12-27  7:36           ` Dragan Simic

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=DEF50A27-6DF2-4B83-8F30-A47AA3878BC6@gmail.com \
    --to=piotr.oniszczuk@gmail.com \
    --cc=Linux-rockchip@lists.infradead.org \
    --cc=andy.yan@rock-chips.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox