View previous topic :: View next topic |
Author |
Message |
DawgG l33t
Joined: 17 Sep 2003 Posts: 866
|
Posted: Sat Jul 14, 2018 9:55 am Post subject: strange behavior of xfce-compositor on nouveau/nvidia-binary |
|
|
i run xfce (among others) on an ~amd64-system that uses different kernels for the same system to run X with either the nouveau or the nvidia-binary-drivers (as described here https://wiki.gentoo.org/wiki/Nouveau_%26_nvidia-drivers_switching#Switching_using_two_kernels)
in about 95% of the time on this box i use fluxbox with the nouveau-kernel, so i do not know exactly when this error started occurring: xfce (and X) would not start (.xinitrc and startx) and xfce4-session produced a segfault.
i found out that when compositing is activated i a running xfce-session on the nouveau-kernel, it runs perfectly (with compositing) until the session is closed; when the next session is started the segfault occurs.
to start a session after that i had to boot with the nvidia-binary-kernel and turn off compositing inside that session if i wanted to use xfce with the nouveau-kernel again (where i could use compositing again once the session was up).
it's not a really pressing problem but i'd like to know if is this strange or expected behavior? _________________ DUMM KLICKT GUT. |
|
Back to top |
|
|
krinn Watchman
Joined: 02 May 2003 Posts: 7470
|
Posted: Sat Jul 14, 2018 10:27 am Post subject: |
|
|
i would say strange myself.
it looks like composing stuck it, trying to removing it and reloading it would reset it and may point who is keeping nouveau stuck ; so did you try reloading the nouveau module in between? |
|
Back to top |
|
|
Hu Moderator
Joined: 06 Mar 2007 Posts: 21633
|
Posted: Sat Jul 14, 2018 3:32 pm Post subject: |
|
|
Rather than booting into the nVidia proprietary mode, can you boot the Nouveau mode, disable compositing from the console, and then start Xfce directly? |
|
Back to top |
|
|
DawgG l33t
Joined: 17 Sep 2003 Posts: 866
|
Posted: Tue Jul 17, 2018 10:20 am Post subject: |
|
|
THX for your replies!
Quote: | so did you try reloading the nouveau module in between? |
i only noticed this behavior after a reboot, so, yes, the nouveau-module was re-loaded in between.
Quote: | disable compositing from the console, |
i would just set the value in xfwm4.xml - will check that soon and post back (right now some uninterruptible work is done under fluxbox). _________________ DUMM KLICKT GUT. |
|
Back to top |
|
|
krinn Watchman
Joined: 02 May 2003 Posts: 7470
|
Posted: Tue Jul 17, 2018 11:47 am Post subject: |
|
|
DawgG wrote: | Quote: | so did you try reloading the nouveau module in between? |
i only noticed this behavior after a reboot, so, yes, the nouveau-module was re-loaded in between. |
I was more having in mind: modprobe -r follow by modprobe the module again |
|
Back to top |
|
|
DawgG l33t
Joined: 17 Sep 2003 Posts: 866
|
Posted: Tue Jul 17, 2018 2:26 pm Post subject: |
|
|
Quote: | was more having in mind: modprobe -r follow by modprobe the module again |
i had, too, but it's built-in.
the thing is, when compositing is turned on while xfce4 is running, it works, but if it's turned on before, xfce will not start.
this is what dmesg told me after a failed start:
Code: | [10160.202164] traps: xfce4-session[24042] general protection ip:7f76f1f0fe3d sp:7ffec4327fb0 error:0 in libgobject-2.0.so.0.5400.3[7f76f1eda000+53000] |
(xfce4-session and glib have been recently recompiled)
and this was in Xorg.0.log (EE and WW): Code: |
[ 10202.140] Current Operating System: Linux gen2 4.17.5-gentoo #1 SMP PREEMPT Fri Jul 13 15:03:50 CEST 2018 x86_64
(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[ 10202.163] (EE) Failed to load module "nv" (module does not exist, 0)
[ 10202.163] (EE) Failed to load module "fbdev" (module does not exist, 0)
[ 10202.163] (EE) Failed to load module "vesa" (module does not exist, 0)
[ 10202.431] (II) Initializing extension MIT-SCREEN-SAVER
[ 10203.539] (EE)
[ 10203.539] (EE) Backtrace:
[ 10203.539] (EE) 0: /usr/bin/X (xorg_backtrace+0x4d) [0x564ac39c1edd]
[ 10203.539] (EE) 1: /usr/bin/X (0x564ac381c000+0x1a9b59) [0x564ac39c5b59]
[ 10203.540] (EE) 2: /lib64/libpthread.so.0 (0x7f1195f2f000+0x14090) [0x7f1195f43090]
[ 10203.540] (EE) 3: /usr/bin/X (miRenderColorToPixel+0xe) [0x564ac3937e7e]
[ 10203.540] (EE) 4: /usr/lib64/xorg/modules/libexa.so (0x7f1191688000+0xf23b) [0x7f119169723b]
[ 10203.540] (EE) 5: /usr/bin/X (0x564ac381c000+0x12eb66) [0x564ac394ab66]
[ 10203.540] (EE) 6: /usr/bin/X (0x564ac381c000+0x122ecc) [0x564ac393eecc]
[ 10203.540] (EE) 7: /usr/bin/X (0x564ac381c000+0x56c98) [0x564ac3872c98]
[ 10203.540] (EE) 8: /usr/bin/X (0x564ac381c000+0x5ad08) [0x564ac3876d08]
[ 10203.540] (EE) 9: /lib64/libc.so.6 (__libc_start_main+0xe7) [0x7f1195b889f7]
[ 10203.540] (EE) 10: /usr/bin/X (_start+0x2a) [0x564ac386093a]
[ 10203.540] (EE)
[ 10203.540] (EE) Segmentation fault at address 0x8
[ 10203.540] (EE)
[ 10203.540] (EE) Caught signal 11 (Segmentation fault). Server aborting
[ 10203.540] (EE)
[ 10203.540] (EE)
[ 10203.540] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[ 10203.540] (EE)
[ 10203.659] (EE) Server terminated with error (1). Closing log file. |
it's not really a pressing problem but i think it's a little odd. _________________ DUMM KLICKT GUT. |
|
Back to top |
|
|
krinn Watchman
Joined: 02 May 2003 Posts: 7470
|
Posted: Tue Jul 17, 2018 3:44 pm Post subject: |
|
|
you should strace fail attempt and see where exactly it segfault ; quiet a nasty one |
|
Back to top |
|
|
Hu Moderator
Joined: 06 Mar 2007 Posts: 21633
|
Posted: Wed Jul 18, 2018 1:41 am Post subject: |
|
|
That looks like you have at least two problems. Pick which one you want to work first: the X server crash or the xfce-session crash. Either way, you probably need a symbolized backtrace of the crash. Currently, at least for X, you do not have debug symbols. That will make solving this much harder. I suggest you rebuild the X server with debug symbols, then crash it again and check the resulting backtrace. |
|
Back to top |
|
|
|