Bug 97850 - Suspending AIGLX clients for VT switch when Radeon is configured
Summary: Suspending AIGLX clients for VT switch when Radeon is configured
Status: RESOLVED NOTOURBUG
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/Radeon (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: xf86-video-ati maintainers
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-09-18 18:41 UTC by Yu Tzu Wu
Modified: 2019-09-20 15:35 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
var/log/Xorg.0.log (1.08 KB, text/plain)
2016-09-19 15:42 UTC, Yu Tzu Wu
no flags Details
/var/log/sddm.log (11.02 KB, text/plain)
2016-09-19 15:43 UTC, Yu Tzu Wu
no flags Details
dmesg (61.27 KB, text/plain)
2016-09-19 15:44 UTC, Yu Tzu Wu
no flags Details
lsmod (2.84 KB, text/plain)
2016-09-19 15:45 UTC, Yu Tzu Wu
no flags Details
lspci | grep VGA (264 bytes, text/plain)
2016-09-19 15:46 UTC, Yu Tzu Wu
no flags Details
/usr/src/linux/.config (143.60 KB, text/plain)
2016-09-19 15:53 UTC, Yu Tzu Wu
no flags Details
dmesg (57.25 KB, text/plain)
2016-09-19 15:59 UTC, Yu Tzu Wu
no flags Details
lsmod (2.62 KB, text/plain)
2016-09-19 15:59 UTC, Yu Tzu Wu
no flags Details
dmesg without fglrx (63.42 KB, text/plain)
2016-09-20 10:34 UTC, Yu Tzu Wu
no flags Details
xorg.0.log without fglrx (3.95 KB, text/plain)
2016-09-20 10:35 UTC, Yu Tzu Wu
no flags Details
sddm.log without fglrx (59.18 KB, text/plain)
2016-09-20 10:36 UTC, Yu Tzu Wu
no flags Details
/var/log/Xorg.0.log after emerge xf86-video-ati (35.55 KB, text/plain)
2016-09-21 07:29 UTC, Yu Tzu Wu
no flags Details
/var/log/sddm.log when /var/Xorg.0.log no Errors (79.83 KB, text/plain)
2016-09-21 11:02 UTC, Yu Tzu Wu
no flags Details

Description Yu Tzu Wu 2016-09-18 18:41:39 UTC
uname -a
Linux 3820TG 4.4.6-gentoo #4 SMP Mon Sep 19 03:38:26 CST 2016 x86_64 Intel(R) Core(TM) i3 CPU M 370 @ 2.40GHz GenuineIntel GNU/Linux

equery l xorg-server
[IP-] [  ] x11-base/xorg-server-1.17.4:0/1.17.4

 equery l ati-drivers
[IP-] [  ] x11-drivers/ati-drivers-15.9-r1:1

equery l xorg-drivers                                                                                                                                                                                                                                                                                                                      
[IP-] [  ] x11-base/xorg-drivers-1.17:0   

It's reproducible that open a black screen with a flickering underscore when /etc/init.d/xdm starts using radeon driver.

Here is /var/log/Xorg.0.log
https://bpaste.net/show/3d79333cd7a9

Here is /var/log/sddm.log
https://bpaste.net/show/a5603d7ba3ac


Here is /etc/X11/xorg.conf

Section "ServerLayout" 
   Identifier   "Default Layout" 
   Option   "ignoreABI"   "1"    
EndSection 

Section   "Monitor" 
   Identifier   "Default Monitor" 
EndSection 

Section "Device" 
   Identifier   "Radeon" 
   Driver   "radeon" 
   #Option   "AccelMethod"   "glamor" 
   #Option   "DRI" "2" 
   #Option   "TearFree" "on" 
   BusID   "PCI:02:00:0" 
EndSection 

Section "Screen" 
   Identifier "Default Screen" 
   Monitor   "Default Monitor" 
   Device "Radeon" 
EndSection

Here is dmesg
https://bpaste.net/show/f4f499696294

Here is lsmod
https://bpaste.net/show/dd851eee28b5

Here is lspci | grep VGA
https://bpaste.net/show/2c203259714e

Here is /usr/src/linux/.config
https://bpaste.net/show/81e880c35864
Comment 1 Alex Deucher 2016-09-19 15:19:07 UTC
Please attach your dmesg, etc. rather than providing links to be sure the links don't break, etc.  Please also give a clear description of what problem you are seeing.
Comment 2 Yu Tzu Wu 2016-09-19 15:42:17 UTC
Created attachment 126618 [details]
var/log/Xorg.0.log
Comment 3 Yu Tzu Wu 2016-09-19 15:43:53 UTC
Created attachment 126619 [details]
/var/log/sddm.log
Comment 4 Yu Tzu Wu 2016-09-19 15:44:48 UTC
Created attachment 126620 [details]
dmesg
Comment 5 Yu Tzu Wu 2016-09-19 15:45:23 UTC
Created attachment 126621 [details]
lsmod
Comment 6 Yu Tzu Wu 2016-09-19 15:46:31 UTC
Created attachment 126622 [details]
lspci | grep VGA
Comment 7 Yu Tzu Wu 2016-09-19 15:53:32 UTC
Created attachment 126623 [details]
/usr/src/linux/.config
Comment 8 Yu Tzu Wu 2016-09-19 15:59:03 UTC
Created attachment 126624 [details]
dmesg
Comment 9 Yu Tzu Wu 2016-09-19 15:59:36 UTC
Created attachment 126625 [details]
lsmod
Comment 10 Yu Tzu Wu 2016-09-19 16:02:23 UTC
After /etc/init.d/xdm start, it shows 'sddm setting' at tty1, a black screen shows, and back to tty1.
Comment 11 Michel Dänzer 2016-09-20 03:21:08 UTC
dmesg shows that the fglrx kernel module gets loaded. If you want to use the open source radeon drivers, you have to completely remove all traces of fglrx.

If that doesn't help, please attach a full Xorg log file corresponding to the failure. You may need to look at Xorg.*.log.old .
Comment 12 Yu Tzu Wu 2016-09-20 10:34:54 UTC
Created attachment 126650 [details]
dmesg without fglrx
Comment 13 Yu Tzu Wu 2016-09-20 10:35:38 UTC
Created attachment 126651 [details]
xorg.0.log without fglrx
Comment 14 Yu Tzu Wu 2016-09-20 10:36:36 UTC
Created attachment 126652 [details]
sddm.log without fglrx
Comment 15 Yu Tzu Wu 2016-09-20 10:38:16 UTC
(In reply to Michel Dänzer from comment #11)
> dmesg shows that the fglrx kernel module gets loaded. If you want to use the
> open source radeon drivers, you have to completely remove all traces of
> fglrx.
> 
> If that doesn't help, please attach a full Xorg log file corresponding to
> the failure. You may need to look at Xorg.*.log.old .

I already remove all fglrx modules, but situation is the same.
log files are attached.
Comment 16 Michel Dänzer 2016-09-21 01:07:58 UTC
From the Xorg log file:

[   129.365] (WW) Warning, couldn't open module radeon
[   129.365] (II) UnloadModule: "radeon"
[   129.365] (II) Unloading radeon
[   129.365] (EE) Failed to load module "radeon" (module does not exist, 0)
[   129.365] (EE) No drivers available.

It looks like /etc/X11/xorg.conf forces the radeon driver (xf86-video-ati), but that isn't installed.
Comment 17 Yu Tzu Wu 2016-09-21 07:29:27 UTC
Created attachment 126691 [details]
/var/log/Xorg.0.log after emerge xf86-video-ati
Comment 18 Yu Tzu Wu 2016-09-21 07:41:49 UTC
(In reply to Michel Dänzer from comment #16)
> From the Xorg log file:
> 
> [   129.365] (WW) Warning, couldn't open module radeon
> [   129.365] (II) UnloadModule: "radeon"
> [   129.365] (II) Unloading radeon
> [   129.365] (EE) Failed to load module "radeon" (module does not exist, 0)
> [   129.365] (EE) No drivers available.
> 
> It looks like /etc/X11/xorg.conf forces the radeon driver (xf86-video-ati),
> but that isn't installed.

Yes, it isn't there, so I emerge xf86-video-ati again.
After /etc/init.d/xdm start, screen shows whole black and no jump back to tty1.
(I have to switch it to console manually.)

Another thing that warned during emerge xf86-video-ati process was
portage tree suggested that I should not set the below
CONFIG_DRM_RADEON_UMS=y
CONFIG_FB_RADEON=m
Does this matter?
Comment 19 Michel Dänzer 2016-09-21 08:31:19 UTC
(In reply to Yu Tzu Wu from comment #17)
> /var/log/Xorg.0.log after emerge xf86-video-ati

Assuming the Xorg process still existed when you grabbed the log file, it looks like Xorg starts up normally and that you could successfully switch away from its VT to console.

Note that a black screen is Xorg's default startup behaviour, so unless the problem doesn't happen e.g. using the modesetting driver instead of radeon, it looks like the problem is more likely with sddm or something it uses.


(In reply to Yu Tzu Wu from comment #18)
> Another thing that warned during emerge xf86-video-ati process was
> portage tree suggested that I should not set the below
> CONFIG_DRM_RADEON_UMS=y
> CONFIG_FB_RADEON=m
> Does this matter?

No.
Comment 20 Yu Tzu Wu 2016-09-21 11:02:12 UTC
Created attachment 126699 [details]
/var/log/sddm.log when /var/Xorg.0.log no Errors
Comment 21 Yu Tzu Wu 2016-09-21 11:06:13 UTC
(In reply to Michel Dänzer from comment #19)
> Note that a black screen is Xorg's default startup behaviour, so unless the
> problem doesn't happen e.g. using the modesetting driver instead of radeon,
> it looks like the problem is more likely with sddm or something it uses.

Yes. There is no error in Xorg.0.log. But errors appear in sddm.log, 

[23:35:44.594] (II) HELPER: [PAM] Starting...
[23:35:44.595] (II) HELPER: [PAM] Authenticating...
[23:35:44.595] (II) HELPER: [PAM] returning.
[23:35:44.595] (EE) HELPER:  Auth: SafeDataStream: Could not write all stored data
[23:35:44.595] (WW) HELPER: QIODevice::read (QTcpSocket): device not open
[23:35:44.595] (EE) HELPER: Received a wrong opcode instead of AUTHENTICATED: 0
[23:35:44.595] (EE) HELPER: Unable to run user session: unknown session type
[23:35:44.595] (EE) HELPER:  Auth: SafeDataStream: Could not write all stored data
[23:35:44.595] (WW) HELPER: QIODevice::read (QTcpSocket): device not open
[23:35:44.595] (EE) HELPER: Received a wrong opcode instead of SESSION_STATUS: 0
[23:35:44.595] (II) HELPER: [PAM] Ended.


I should go to report a bug to sddm. Thank you very much!
Comment 22 Michel Dänzer 2019-09-20 15:35:56 UTC
Assuming this was an SDDM issue.


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.