linux-kernel - Issues with Radeon DP audio and Dell U2711 monitor

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <2717558.zfCjFtmnXs@feathers>
Date:	Fri, 15 May 2015 19:09:17 +0200
From:	Stefan Gödel <mail@...fan-goedel.de>
To:	linux-kernel@...r.kernel.org
Subject: Issues with Radeon DP audio and Dell U2711 monitor

Hello,

I run a dual monitor setup as my main workstation with Debian sid (unstable) 
on it. My primary monitor is a Dell U2711 and the secondary a Dell 2209WA. The 
GPU is a Radeon HD 5450 with 1xDisplayPort, 1x DVI, 1xHDMI. 

Two days ago I updated my system and Linux kernel 4.0.2 was installed. When I 
rebooted, everything went smooth and I was happy. In the evening I suspended 
the system and when I resumed it yesterday morning, the primary monitor (which 
is attached to the Display Port) didn't come up. The secondary monitor showed 
the right part of my desktop as if the primary monitor was still connected. 
After a reboot, everything was OK again but after suspend/resume the Dell 
U2711 again stayed dark. I fiddled around with xset but soon realized that the 
problem was not limited to X. I disabled kdm and suspended/resumed from the 
console with the same result. I also noticed that the power button on the 
U2711 switched from "breathing" yellow (idle state) to blue on resume, but 
switched back to idle after the usual timeout. Then I looked at the entries in  
/sys/class/drm/card0-DP-1/ and realized that the monitor power button switched 
to active whenever I did a "# cat /sys/class/drm/card0-DP-1/status" and then 
timed out again.

I studied the release notes for 4.0 and stumbled over the new Radeon support 
for audio over the DisplayPort. Several mails on LKML and patches for 4.0.2 
indicated that there had already been some issues with audio over DP and so I 
rebooted with parameter "radeon.audio=0" - and that solved my issue! I can now 
suspend/resume without loosing the primary monitor.

Thinking about these experiences I remembered that there is an optional 
"soundbar" available for the Dell U2711, which I haven't installed. Maybe the 
monitor tells the GPU that it has audio capabilities and then something 
stumbles over the absent soundbar. I also found forums where MS-Windows users 
complained about similar issues with this monitor. I guess that the Windows 
Radeon drivers supported audio over DP from the start?

I don't know if this is a specific problem with the Radeon / Dell U 2711 combo 
or if other combinations might trigger this too. I won't recommend a patch to 
solve this. It might be possible to include a check for this specific monitor 
before enabling audio over DP, but then there will be people with the soundbar 
installed, who will complain. My solution with the "radeon.audio=0" kernel 
parameter is a bit radical, since it also shuts off the HDMI sound.

My favorite solution would be an additional parameter for selectively turning 
off audio over DP (dp-audio=0 ???). This way people with issues like me (but 
maybe with other GPU / monitor combos) could revert this functionality to pre 
4.0 times.

I hope this might help someone :-)

Cheers,

Stefan

p.s.: I am not subscribed to LKML - so please CC me if you want to respond...
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/