Bug 83505 - AMD A4-5300 APU : radeon.dpm=1 get random reboots with 3.16.1 kernel.
Summary: AMD A4-5300 APU : radeon.dpm=1 get random reboots with 3.16.1 kernel.
Status: RESOLVED WORKSFORME
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Radeon (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium minor
Assignee: Default DRI bug account
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-09-04 16:54 UTC by Rpnpif
Modified: 2014-09-22 06:39 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
dmesg with radeon.dpm=1 (61.63 KB, text/plain)
2014-09-04 16:54 UTC, Rpnpif
no flags Details
Dmesg from machine based on ASRock K7S41GX without radeon.dpm (35.87 KB, text/plain)
2014-09-11 15:08 UTC, Rpnpif
no flags Details

Description Rpnpif 2014-09-04 16:54:42 UTC
Created attachment 105760 [details]
dmesg with radeon.dpm=1

I have read that for this AMD A4-5300 APU (Aruba), radeon.dpm=1 on the grub command line of the kernel booting is useless because this option is already per default in the kernel.

But if I don't use this option, Debian 7.6 reboots randomly after some minutes, hours or days of work when I use Iceweasel/Firefox and some flash animation, Libreoffice or other heavy application.

With this option, no reboots. All work fine after several days.

This issue seems about the same as https://bugs.freedesktop.org/show_bug.cgi?id=82201#c26 but I am not sure.

I report this issue because dpm=1 should be useless.
Comment 1 Rpnpif 2014-09-06 10:40:34 UTC
I have now two consecutive reboot : I describe.

1. I worked on Iceweasel 24.8.0 with two opened windows. One editing a SPIP site.
The second with editing a Wikipedia article, more one other tab with Wikipedia.
Claws-mail was also opened.

2. A reboot occurred without any log.

3. The system reboots on 3.16.1 kernel with always radeon.dpm=1.

4. I reopened a session on Xfce.

5. I rerun Iceweasel only. All the tab and windows are restored.

6. I opened this bug report and begun this new comments.

7. After writing two or three lines in the text area, the system rebooted again!

8. I replaced radeon.dpm=1 with radeon.dpm=0, rerun update-grub.I rebooted.

9. I rerun again Iceweasel with restored tabs and I can now fill this comment !

To conclude. The reboot problem is not fixed by a radeon.dpm=1 in grub. So the summary of this bug is not exact. I expect that no new reboot happened again. I'll follow with dpm=0 to see.

Note that about 2 two seconds before a reboot, xorg freezes (a watchdog effect ?).
I have also a little doubt with a kernel network side-effect.

My question : how get a debug log when an immediate reboot happened?
Comment 2 Rpnpif 2014-09-11 14:51:23 UTC
I can confirm now that radeon.dpm=1 or none option get random immediate reboots.

Only radeon.dpm=0 prevents this reboots as tested during one week.

So I change the summary.
Comment 3 Rpnpif 2014-09-11 15:03:47 UTC
Same issue on another machine with ASRock K7S41GX motherboard and Radeon RV280 with 3.14 Kernel from Debian Wheezy-Backport.
Comment 4 Rpnpif 2014-09-11 15:08:22 UTC
Created attachment 106141 [details]
Dmesg from machine based on ASRock K7S41GX without radeon.dpm
Comment 5 Alex Deucher 2014-09-11 15:14:31 UTC
(In reply to comment #3)
> Same issue on another machine with ASRock K7S41GX motherboard and Radeon
> RV280 with 3.14 Kernel from Debian Wheezy-Backport.

RV280 does not support dpm so this parameter is useless for this asic.
Comment 6 Rpnpif 2014-09-17 08:01:15 UTC
No random reboots occurred during five days of use of 3.16.2 kernel.
I am waiting some days yet until I will say that this bug is resolved.
Comment 7 Rpnpif 2014-09-19 14:02:20 UTC
After one week without random reboots, I think that this bug is fixed in 3.16.2.
Thank you.
Comment 8 Michel Dänzer 2014-09-22 06:39:34 UTC
(In reply to comment #7)
> After one week without random reboots, I think that this bug is fixed in
> 3.16.2.

That's weird though, as there are no directly related changes between 3.16.1 and 3.16.2. Can you bisect which change fixed it for you? (Note that you'll have to reverse 'good' and 'bad' for the bisection, as git bisect can only handle good -> bad)


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.