Bug 76837

Summary: sound glitches with radeon.dpm
Product: xorg Reporter: francesco <markurujapan>
Component: Driver/RadeonAssignee: xf86-video-ati maintainers <xorg-driver-ati>
Status: RESOLVED MOVED QA Contact: Xorg Project Team <xorg-team>
Severity: major    
Priority: highest CC: f.gorzialka
Version: git   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
log of xorg.log
none
dmesg output
none
Gigabyte F2A88XM dmesg.
none
Xorg.0.log for AMD A6-6400K.
none
dmesg
none
Xorg.0.log none

Description francesco 2014-03-31 14:11:24 UTC
Since kernel 3.12 , 3.13 and 3.14, and the latest radeon drivers, I have serious problems with the pulseaudio sound and alsa sound, glitches, skips, so annoying. This problem is not present when I use this kernels with catalyst drivers, are present only when I use opensource drivers.

I have an apu amd a8 5600k.
Comment 1 Alex Deucher 2014-03-31 14:22:54 UTC
Please attach your dmesg output and xorg log.  For clarity, are you only having problems with dpm enabled?  Is this a regression?  When was the last working kernel?  How are you connecting to your speakers (hdmi, audio jack, etc.)?
Comment 2 francesco 2014-03-31 18:34:28 UTC
(In reply to comment #1)
> Please attach your dmesg output and xorg log.  For clarity, are you only
> having problems with dpm enabled?  Is this a regression?  When was the last
> working kernel?  How are you connecting to your speakers (hdmi, audio jack,
> etc.)?

the latest kernel that work perfectly was the 3.10 and 3.11 without dpm enabled. the problem is present on 3.12/3.13 and 3.14. I tried it with fedora, ubuntu and archlinux.
I use headphones with jack connection.
Comment 3 francesco 2014-03-31 18:34:57 UTC
Created attachment 96677 [details]
log of xorg.log
Comment 4 francesco 2014-03-31 18:35:10 UTC
Created attachment 96678 [details]
dmesg output
Comment 5 mumart 2014-05-02 23:30:31 UTC
Hello, I've just installed a similar FM2 motherboard to this one and am experiencing the same problem with my PCI sound card (Terratec DMX 6Fire), which I bisected to commit 3a118989 "drm/radeon: enable DPM by default in TN asics".

Tried playing with the DPM settings, but to no avail. Issuing an "echo high > power_dpm_force_performance_level" gives me "write error: Invalid argument".

Kernel messages and Xorg logs attached.

Cheers,
Martin
Comment 6 mumart 2014-05-02 23:33:38 UTC
Created attachment 98355 [details]
Gigabyte F2A88XM dmesg.
Comment 7 mumart 2014-05-02 23:34:27 UTC
Created attachment 98356 [details]
Xorg.0.log for AMD A6-6400K.
Comment 8 mumart 2014-05-02 23:43:22 UTC
Audio problem does not occur when booting with "radeon.dpm=0".

Cheers,
Martin
Comment 9 francesco 2014-08-06 13:39:32 UTC
the problem still present on kernel 3.16, I need to disable radeon dpm (radeon.dpm=0).
Comment 10 francesco 2014-10-07 14:00:13 UTC
problem continue with kernel 3.17, no changes
Comment 11 mumart 2014-11-30 22:30:25 UTC
Problem does not occur for me in 3.17 if I enable dpm and set "radeon.bapm=1".

Cheers,
Martin
Comment 12 francesco 2014-11-30 22:33:43 UTC
radeon.bapm=1 or radeon.dpm?
Comment 13 mumart 2014-12-01 21:46:21 UTC
I only need to set "radeon.bapm=1" to make it work for me. DPM is on by default with my hardware.

Cheers,
Martin
Comment 14 bgunteriv 2015-01-22 17:03:41 UTC
Created attachment 112670 [details]
dmesg

my dmesg output
Comment 15 bgunteriv 2015-01-22 17:07:39 UTC
Created attachment 112671 [details]
Xorg.0.log

here's my Xorg.0.log
Comment 16 bgunteriv 2015-01-22 17:10:53 UTC
I believe that i'm affected too by this issue.

I have attached my logs above. 
switching radeon.dpm does not have any effect for me.

trying radeon.bapm to see if this helps.

let me know if there is anything else i can try.
Comment 17 bgunteriv 2015-01-22 17:24:35 UTC
I use HDMI, and pure ALSA.

i used to be affected by this bug -- https://bugs.freedesktop.org/show_bug.cgi?id=77002

but I now think i'm affected by this issue.
but for me, I didn't start noticing it until kernel 3.18.
I too have been utilizing Mesa-git-devel on my system.
but even when I go back to Mesa 10.3.2, i still have this issue.
Comment 18 Alex Deucher 2015-01-22 17:25:54 UTC
(In reply to bgunteriv from comment #16)
> I believe that i'm affected too by this issue.
> 
> I have attached my logs above. 
> switching radeon.dpm does not have any effect for me.
> 
> trying radeon.bapm to see if this helps.
> 
> let me know if there is anything else i can try.

Does increasing the CPU load help?  See:
https://bugzilla.kernel.org/show_bug.cgi?id=86351
Comment 19 bgunteriv 2015-01-22 18:24:01 UTC
this seems to be working for me:
radeon.bapm=0 radeon.dpm=0
kernel 3.18.3

i have not had any interruptions -- to date.

now testing with kernel 3.19rc5 -- i will report back to let you know.
Comment 20 bgunteriv 2015-01-22 18:28:45 UTC
(In reply to Alex Deucher from comment #18)
> (In reply to bgunteriv from comment #16)
> > I believe that i'm affected too by this issue.
> > 
> > I have attached my logs above. 
> > switching radeon.dpm does not have any effect for me.
> > 
> > trying radeon.bapm to see if this helps.
> > 
> > let me know if there is anything else i can try.
> 
> Does increasing the CPU load help?  See:
> https://bugzilla.kernel.org/show_bug.cgi?id=86351

i get permission denied when trying to run:
echo high > /sys/class/drm/card0/device/power_dpm_force_performance_level

yes, I am root.
am i doing something wrong?
Comment 21 bgunteriv 2015-01-22 18:38:47 UTC
i have the file > echo high > /sys/class/drm/card0/device/power_method
which has just the word:
default

in it.

should i try changing: default to high
?
Comment 22 bgunteriv 2015-01-23 19:50:02 UTC
when I try to edit this file with:
echo high > /sys/class/drm/card0/device/power_method

i get:
-bash: echo: write error: Invalid argument
Comment 23 Alex Deucher 2015-01-23 19:56:33 UTC
(In reply to bgunteriv from comment #22)
> when I try to edit this file with:
> echo high > /sys/class/drm/card0/device/power_method
> 
> i get:
> -bash: echo: write error: Invalid argument

Forcing high is not possible on trinity hw.  Can you try forcing to low or creating CPU load?
Comment 24 bgunteriv 2015-01-23 21:44:42 UTC
it's actually Richland.

If I set:
radeon.bapm=0 radeon.dpm=0

then,
/sys/class/drm/card0/device/power_method
is set to default

and, I can set it to High or Low, or any setting:
echo high > /sys/class/drm/card0/device/power_profile 

consequently, if I set: 
radeon.bapm=1 radeon.dpm=1

or completely leave it out of grub (default), then,
/sys/class/drm/card0/device/power_method
is set to dpm

and I can not change the profile to anything -- High or Low

hope this makes sense.

not really sure about creating CPU load...?
Comment 25 bgunteriv 2015-02-09 03:34:57 UTC
so..... I've been using the latest drm-next kernel -- http://lists.freedesktop.org/archives/dri-devel/2015-January/076003.html

and so far so good.
I have commented out radeon.dpm and radeon.bapm from grub, and I haven't experienced any sound glitches.
Comment 26 Martin Peres 2019-11-19 07:46:14 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/xorg/driver/xf86-video-ati/issues/98.

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.