Recent kernel upgrade & Radeon RX 5700 XT - Xorg crashes when
evolution is invoked
by ja
Hi
I have recently put together a new machine using
an RX 5700 XT graphics card & X570 motherboard.
I have been using it for about a month.
My understanding is that kernel 5.3 provides "fuller" support for RX 5700 graphics
With the 5.2 kernel (eg 5.2.18-200.fc30.x86_64) it has been working fine.
Only one resolution appeared available but it seemed to work OK
xrandr shows
Screen 0: minimum 1920 x 1080, current 1920 x 1080, maximum 1920 x 1080
default connected 1920x1080+0+0 0mm x 0mm
1920x1080 77.00*
Today's update to 5.3 (5.3.5-200.fc30.x86_64) causes Xorg to crash
when evolution is used.
Xorg restarts and returns the user to the login screen.
Other applications seem to work OK google-chrome, google-earth, brave-browser,
pcmanfm,
xrandr now shows loads of different resolutions (but not that given above)
Screen 0: minimum 320 x 200, current 1920 x 1080, maximum 16384 x 16384
DP-1 disconnected (normal left inverted right x axis y axis)
DP-2 disconnected (normal left inverted right x axis y axis)
DP-3 disconnected (normal left inverted right x axis y axis)
HDMI-1 connected 1920x1080+0+0 (normal left inverted right x axis y axis) 477mm x 268mm
1920x1080 60.00*+ 59.96 60.00 50.00 59.94 59.93
...
Further tests show that
Logging in as "fred" and "su - ted" & running evolution as ted avoids the crash.
Logging in as "fred" and "su - fred" & running evolution as fred crashes Xorg
Any thoughts/advice?
John
//-------------------------------------------------------------------------------------------------
/var/log/Xorg.0.log.old shows
[ 530.028] (II) modeset(0): Modeline "1152x864"x0.0 108.00 1152 1216 1344 1600 864 865 868 900 +hsync +vsync (67.5 kHz e)
[ 530.028] (II) modeset(0): Modeline "1440x900"x0.0 136.75 1440 1536 1688 1936 900 903 909 942 -hsync +vsync (70.6 kHz e)
[ 1793.339] (EE)
[ 1793.339] (EE) Backtrace:
[ 1793.340] (EE) 0: /usr/libexec/Xorg (OsLookupColor+0x13c) [0x564a610dbd2c]
[ 1793.340] (EE) 1: /lib64/libpthread.so.0 (funlockfile+0x50) [0x7fe5322c6caf]
[ 1793.340] (EE) 2: /lib64/libc.so.6 (gsignal+0x145) [0x7fe532125e35]
[ 1793.341] (EE) 3: /lib64/libc.so.6 (abort+0x127) [0x7fe532110895]
[ 1793.341] (EE) 4: /lib64/libc.so.6 (__assert_fail_base.cold+0xf) [0x7fe532110769]
[ 1793.341] (EE) 5: /lib64/libc.so.6 (__assert_fail+0x46) [0x7fe53211e566]
[ 1793.341] (EE) 6: /usr/libexec/Xorg (DRIMoveBuffersHelper+0xc15) [0x564a610a6e15]
[ 1793.341] (EE) 7: /usr/libexec/Xorg (DRI2Authenticate+0xa2) [0x564a610a8462]
[ 1793.341] (EE) 8: /usr/libexec/Xorg (DRI2GetParam+0x944) [0x564a610a9814]
[ 1793.341] (EE) 9: /usr/libexec/Xorg (SendErrorToClient+0x354) [0x564a60f7cd94]
[ 1793.341] (EE) 10: /usr/libexec/Xorg (InitFonts+0x3b4) [0x564a60f80e34]
[ 1793.342] (EE) 11: /lib64/libc.so.6 (__libc_start_main+0xf3) [0x7fe532111f43]
[ 1793.342] (EE) 12: /usr/libexec/Xorg (_start+0x2e) [0x564a60f6a86e]
[ 1793.342] (EE)
[ 1793.342] (EE)
Fatal server error:
[ 1793.342] (EE) Caught signal 6 (Aborted). Server aborting
[ 1793.342] (EE)
[ 1793.342] (EE)
Please consult the Fedora Project support
at http://wiki.x.org
for help.
[ 1793.342] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[ 1793.342] (EE)
[ 1793.383] (EE) Server terminated with error (1). Closing log file.
//-------------------------------------------------------------------------------------------------
journalctl shows
Stack trace of thread 4693:
#0 0x00007fe532125e35 raise (libc.so.6)
#1 0x00007fe532110895 abort (libc.so.6)
#2 0x0000564a610de940 OsAbort (Xorg)
#3 0x0000564a610e4539 AbortServer (Xorg)
#4 0x0000564a610e539a FatalError (Xorg)
#5 0x0000564a610dbc89 OsSigHandler (Xorg)
#6 0x00007fe5322c6c60 __restore_rt (libpthread.so.0)
#7 0x00007fe532125e35 raise (libc.so.6)
#8 0x00007fe532110895 abort (libc.so.6)
#9 0x00007fe532110769 __assert_fail_base.cold (libc.so.6)
#10 0x00007fe53211e566 __assert_fail (libc.so.6)
#11 0x0000564a610a6225 dixGetPrivateAddr.isra.0.part.0 (Xorg)
#12 0x0000564a610a8462 DRI2Authenticate (Xorg)
#13 0x0000564a610a9484 ProcDRI2Dispatch (Xorg)
#14 0x0000564a60f7cd34 Dispatch (Xorg)
#15 0x0000564a60f80df4 dix_main (Xorg)
#16 0x00007fe532111f43 __libc_start_main (libc.so.6)
#17 0x0000564a60f6a86e _start (Xorg)
Stack trace of thread 4722:
#0 0x00007fe5322c23c5 pthread_cond_wait@(a)GLIBC_2.3.2 (libpthread.so.0)
#1 0x00007fe52b109c13 cnd_wait (swrast_dri.so)
#2 0x00007fe52b109a7b impl_thrd_routine (swrast_dri.so)
#3 0x00007fe5322bc4c0 start_thread (libpthread.so.0)
#4 0x00007fe5321ea553 __clone (libc.so.6)
...
#0 0x00007fe5322c23c5 pthread_cond_wait@(a)GLIBC_2.3.2 (libpthread.so.0)
#1 0x00007fe52b109c13 cnd_wait (swrast_dri.so)
#2 0x00007fe52b109a7b impl_thrd_routine (swrast_dri.so)
#3 0x00007fe5322bc4c0 start_thread (libpthread.so.0)
#4 0x00007fe5321ea553 __clone (libc.so.6)
3 hours, 35 minutes
F30: latest kernel no external monitor
by Gianluca Cecchi
Hello,
executed "dnf update" after about 15 days of inactivity.
Updates has brought in kernel 5.2.8-200.fc30.x86_64
But after reboot the system is not able to detect (in the sense of no
monitor connection at all seen) when connecting a 24" Dell monitor using an
usb-c to hdmi adapter on my laptop and then connecting an HDMI-to-dvi cable
to the monitor, that I normally used before.
Tried to reboot two times without effect.
As soon as I restart with previous kernel 5.1.20-300.fc30.x86_64, external
monitor detection and usage is ok again.
This is in default Gnome environment and Wayland.
Eg in gnome terminal I have:
$ loginctl show-session 2 -p Type
Type=wayland
and
$ echo $WAYLAND_DISPLAY
wayland-0
Anyone else with similar problems?
If you point me on how to get related log lines in journal in both working
and not working configurations, I can get more info.
Seen on several threads that with Xwayland and gnome-shell not so easy to
see startup log files...
Thanks in advance,
Gianluca
13 hours, 56 minutes
too-nearly-full filesystem '/'. (was upgrade problem: space on '/'
filesystem.)
by home user
(Fedora-30; x86-64; personal workstation)
This problem goes back to last Thursday, so a brief review might be good.
brief background
================
Last Thursday, I tried to upgrade from Fedora-29 to Fedora-30. dnf
complained it didn't have enough space in the '/' filesystem. List
members successfully got me through the upgrade. But the root
problem(s) probably still remains.
some data (all commands run this (Tuesday) afternoon, Oct. 15)
=========
-bash.11[~]: df -hP
Filesystem Size Used Avail Use% Mounted on
devtmpfs 7.8G 0 7.8G 0% /dev
tmpfs 7.9G 76M 7.8G 1% /dev/shm
tmpfs 7.9G 1.7M 7.9G 1% /run
tmpfs 7.9G 0 7.9G 0% /sys/fs/cgroup
/dev/sda6 50G 41G 6.0G 88% /
tmpfs 7.9G 68K 7.9G 1% /tmp
/dev/sda3 477M 236M 213M 53% /boot
/dev/sda7 904G 7.7G 851G 1% /home
tmpfs 1.6G 84K 1.6G 1% /run/user/1001
-bash.12[~]:
-bash.13[/]: du -sh | grep G
du: cannot access './run/user/1001/gvfs': Permission denied
du: cannot access './proc/1776/task/1776/fd/4': No such file or directory
du: cannot access './proc/1776/task/1776/fdinfo/4': No such file or
directory
du: cannot access './proc/1776/fd/3': No such file or directory
du: cannot access './proc/1776/fdinfo/3': No such file or directory
49G .
-bash.14[/]:
-bash.14[/]: du -hx -d1 /
23G /var
4.0K /system-upgrade-root
4.0K /srv
17G /usr
16K /lost+found
51M /etc
4.0K /sysroot
4.0K /mnt
1.4G /root
4.0K /media
238M /opt
4.0K /.cache
41G /
-bash.15[/]:
starting questions
==================
1. What else do you want to know to effectively help?
2. A good specific first step?
Please: one step at a time!
thanks,
Bill.
15 hours, 9 minutes
Suspend problems on Dell XPS 13
by Marcel Oliver
I am running Fedora 30 (Cinnamon Spin) on a Dell XPS 13 (9380) laptop.
Until about early August, everything was working perfectly, but then,
with some update, suspend problems developed which I have difficulties
debugging.
Initially, I thought this is a kernel issue. Kernels up to
5.1.18-300.fc30.x86_64 are working, everything starting from
5.2.6-200.fc30.x86_64 (I did not try anything in between) is failing in
the following way:
When I press the the suspend button, the screen goes blank, after a few
seconds, it comes back to live, goes blank again, I hear a lot of coil
whine. Then nothing happens, the power button light stays on and the
laptop is doing something in a high-power state (it's getting warmer
than when it's on and idle). I think the same is happening on lid
close, definitely the final state is the same which is a real problem
because when I absentmindedly close the lid and stow the laptop away (as
I am used to doing), I have a very hot laptop in the sleeve, which is
definitely not very good. When the system is in this state, all I can do
is a hard reboot by holding the power button for half a minute. When
the system reboots, it actually comes up as if it was hibernated, i.e.,
I am still logged in with all windows open.
Looking into the problem more, I noticed two things: (i) running
"systemctl suspend" from the command line works perfectly, on all
kernels, and (ii) even with a 5.1 kernel, the system is doing something
different than simply suspending when pressing the suspend button: the
command line suspend will simply turn off the screen, then the power
button light goes off after another second or so. Suspend-button
suspend will always do the dance where the display comes back on briefly
- it was doing this since I had installed Fedora for the first time, I
just didn't think anything of it since it did not cause any problems.
So there is a definitive difference.
Googling further, I found an LWN article
(https://lwn.net/Articles/764841/) discussing that Fedora is
experimenting with hybrid suspend/hibernate modes, but no definitive
conclusion; the article made it seem that it is essentially a
Gnome-not-showing-configuration-options issue, but it's not clear
whether this problem ever made it into a regular Fedora release and it's
also not clear whether it would also affect Cinnamon via some shared code.
Some mailing list post I found somewhere suggests that I look at
/etc/systemd/logind.conf, which in my case says
HandleLidSwitch=suspend
HandleLidSwitchExternalPower=suspend
with all other entries commented out. This seems to indicated that on
closing my lid, "systemctl suspend" should be run. Only, for all I can
tell, it isn't. So who controls what is run on lid close? Is there
some Gnome override (not explicitly installed here), Cinnamon override,
or kernel override for what the systemd configuration suggests?
Is there any documentation on suspend and its hybrid variants with a
clear debug protocol? (I have been thinking about filing a bug report,
but I have no clue which component is responsible and what information
to include.)
Strangely, why is this hitting me and seemingly not everybody else, on
one of the arguably best-supported hardware for Linux, on, for all I can
tell, a vanilla install from earlier this year?
I would appreciate any hints.
Regards,
Marcel
15 hours, 26 minutes
Wifi systens -
by Bob Goodwin
.
Does anyone have experience with the so called "mesh wifi" systems. My
daughter suggested that I need one after reading a review. All I can
find is a lot of hype devoid of the things I would like to know, mainly
how are the units interconnected? It occurs to me that they might use
Ethernet via the ac power line.
I would appreciate anything anyone might have to share on the subject.
Bob
--
Bob Goodwin - Zuni, Virginia, USA
17 hours, 4 minutes
upgrade problem: space on '/' filesystem.
by home user
I'm trying to upgrade from Fedora-29 to Fedora-30. Everything seems to
have downloaded fine. The transaction check succeeded. But then I get
this:
====================
Running transaction test
The downloaded packages were saved in cache until the next successful
transaction.
You can remove cached packages by executing 'dnf clean packages'.
Error: Transaction check error:
installing package ibus-1.5.20-5.fc30.x86_64 needs 7MB on the /
filesystem
[... snip; there are a huge number of similar lines snipped out here ...]
installing package libcap-2.26-5.fc30.i686 needs 1137MB on the /
filesystem
Error Summary
-------------
Disk Requirements:
At least 1274MB more space needed on the / filesystem.
bash.6[~]:
====================
How do I safely get the needed filespace?
Am I correct in assuming that deleting files from
/home/[whatever]
would not help?
thanks,
Bill.
1 day, 5 hours
Two way trust IPA <--> AD
by Michal Zacek
Hello,
it is possible to use an IPA accounts to login to the AD domain
(cross-forest trust)? I would need create a file share on Windows, which
should be accessible by IPA users. I know the opposite (AD --> IPA) work
well, but after i read all docs, I came to conclusion that this so
called "two way trust" is in reality only one way trust.
Thanks for any help.
Best regards,
Michal
2 days, 10 hours
Time to refresh my hardware
by Sam Varshavchik
My C++ compiles are getting longer. It's time to get new hardware, but I'm
having some difficulty finding Fedora-friendly hardware, that's slightly
above average grade, such as dual CPU and spinning rust (I haven't gotten
quite aboard the SSD train, with its built-in expiration date).
Tips and links appreciated.
2 days, 16 hours
How to shut off org.freedesktop.trackers
by Javier Perez
HI
How do I shut off the Freedesktop trackers?
Right now they are filling up my logs with this message.
Oct 13 10:12:38 pepewin.localdomain dbus-broker-launch[9415]: Activation
request for 'org.freedesktop.Tracker1' failed.
Oct 13 10:13:49 pepewin.localdomain dbus-broker-launch[9415]: Activation
request for 'org.freedesktop.Tracker1' failed.
Every few seconds I get one like that. I masked all the trackers because
they were keeping the hard disk almost continuously in use (hdd light lit
on all the time).
Thanks
--
------------------------------
/\_/\
|O O| pepebuho(a)gmail.com
~~~~ Javier Perez
~~~~ While the night runs
~~~~ toward the day...
m m Pepebuho watches
from his high perch.
3 days, 14 hours
