View previous topic :: View next topic |
Author |
Message |
phobos13013 Apprentice
Joined: 22 Sep 2004 Posts: 277 Location: PNW
|
Posted: Mon Jan 26, 2015 4:25 am Post subject: Nouveau Causes Segfault |
|
|
OK, so for the longest time I was using the nvidia driver to run my video on a very aging box; however, within the last few months... when running the nvidia driver my computer hung on boot with the "waiting for uevents to be processed" deal. I tried falling back to older and older versions of nvidia, but to no avail. Soooo, I switched to nouveau; now, i get constant segfaults whenever video is running whether it be streamed thru a browser or running on a media player.
Here is the end of output I am getting from Xorg on crash:
Quote: |
...............
[ 911.854] (EE) 12: /usr/bin/X (0x400000+0x112a38) [0x512a38]
[ 911.854] (EE) 13: /usr/bin/X (0x400000+0x108b36) [0x508b36]
[ 911.854] (EE) 14: /usr/bin/X (0x400000+0x37166) [0x437166]
[ 911.854] (EE) 15: /usr/bin/X (0x400000+0x3afba) [0x43afba]
[ 911.854] (EE) 16: /lib64/libc.so.6 (__libc_start_main+0xf5) [0x7f130ee8ad85]
[ 911.854] (EE) 17: /usr/bin/X (0x400000+0x25f71) [0x425f71]
[ 911.854] (EE)
[ 911.854] (EE) Segmentation fault at address 0x8
[ 911.854] (EE)
Fatal server error:
[ 911.854] (EE) Caught signal 11 (Segmentation fault). Server aborting
[ 911.854] (EE)
[ 911.855] (EE)
Please consult the The X.Org Foundation support
at http://wiki.x.org
for help.
[ 911.855] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[ 911.855] (EE)
[ 911.855] (II) AIGLX: Suspending AIGLX clients for VT switch
[ 911.855] (II) NOUVEAU(0): NVLeaveVT is called.
[ 911.942] (EE) Server terminated with error (1). Closing log file. |
Any ideas what this referring to? I dont really know how to interpret this log. Any other pertinent info needed to diagnosis? Going to nvidia is not possible so i have to make it work with nouveau.... _________________ ...and it should be known by now |
|
Back to top |
|
|
chithanh Developer
Joined: 05 Aug 2006 Posts: 2158 Location: Berlin, Germany
|
Posted: Mon Jan 26, 2015 2:54 pm Post subject: |
|
|
This could be a VDPAU or other acceleration problem.
First thing you could try is upgrading to unstable mesa-10.4.2, xorg-server-1.16.3 and x86-video-nouveau-1.0.11
If the problem still happens, try disabling USE="vdpau" for mesa. As a last resort, boot with nouveau.noaccel=1 kernel parameter (but things will be slow then).
To get further help, you will need to obtain a proper backtrace from the X server to isolate the problem more. Build xorg-server with debug symbols (CFLAGS=-g FEATURES=splitdebug). Connect from another computer via ssh and attach gdb to your X server.
More instructions at http://www.x.org/wiki/Development/Documentation/ServerDebugging/ |
|
Back to top |
|
|
phobos13013 Apprentice
Joined: 22 Sep 2004 Posts: 277 Location: PNW
|
Posted: Mon Jan 26, 2015 7:13 pm Post subject: |
|
|
Chit!
Great info thanks, will implement these suggestions and see what happens/post further progress or hiccups! _________________ ...and it should be known by now |
|
Back to top |
|
|
phobos13013 Apprentice
Joined: 22 Sep 2004 Posts: 277 Location: PNW
|
Posted: Tue Jan 27, 2015 2:52 am Post subject: |
|
|
OK, well turns out not so great news. At least no progress:
I tried to upgrade mesa/xorg-server/nouveau but there were just too many conflicts from dependicies to move forward with. Many of them are in the x11 base or drivers and I don't know their current unstables let alone how to resolve them all.
I did re-emerge mesa with -vdpau but that didnt do the trick. I wont set a kernel parameter thats just useless and kernel debugging is something I just dont have the time for as I have never tried it and thats a good week long progress. Hoping a kernel or nouveau update solves this.
Willing to accept other suggestions, so bumping _________________ ...and it should be known by now |
|
Back to top |
|
|
chithanh Developer
Joined: 05 Aug 2006 Posts: 2158 Location: Berlin, Germany
|
Posted: Tue Jan 27, 2015 10:19 am Post subject: |
|
|
The following should be enough to allow an upgrade:
/etc/portage/package.accept_keywords: | =app-admin/eselect-opengl-1.3.1-r1
=media-libs/libepoxy-1.2
=media-libs/mesa-10.4.2
=x11-libs/libXfont-1.5.0
=x11-libs/libdrm-2.4.58
=x11-base/xorg-drivers-1.16
=x11-base/xorg-server-1.16.3-r1
=x11-drivers/xf86-video-nouveau-1.0.11
=x11-proto/xproto-7.0.26
=x11-proto/glproto-1.4.17-r1
=x11-proto/fontsproto-2.1.3
|
Then run "emerge --oneshot --ask mesa xorg-server libXfont"
The kernel parameter I suggested you to try is just for isolating the issue. |
|
Back to top |
|
|
phobos13013 Apprentice
Joined: 22 Sep 2004 Posts: 277 Location: PNW
|
Posted: Tue Jan 27, 2015 3:12 pm Post subject: |
|
|
Sweet, libXfont was the sticking point; will try thst one later, gotta run for the day. I did try the --oneshot option for mesa alone but not the libXfont.
Thanks again and will let all know. _________________ ...and it should be known by now |
|
Back to top |
|
|
phobos13013 Apprentice
Joined: 22 Sep 2004 Posts: 277 Location: PNW
|
Posted: Wed Jan 28, 2015 1:21 am Post subject: |
|
|
Wheeew,
So I tried to run the blanket command with the newly included accept_keywords file (btw chit you didnt write ~amd64 next to all those packages but I got the drift ) and i still had a bunch of collisions and blocks, so I just went down that list one at a time. When I finally got xorg-server installed it told me that you should/had to rebuild all x11-modules if coming from <1.16 which at that point I couldnt remember if i had or not but I think I did, so made that interlude "emerge @x11-modules-rebuild (although I think it was redundant because they had probably been rebuilt in that initial command).
ANYWAY, after that I got to mesa... now I have a problem. It will not upgrade mesa to unstable without upgrading eselect-opengl to 1.3.1 BUT even the latest xorg-server and others require less than that (1.3.0 which is currently installed). So I cannot upgrade mesa because of critical collision with eselect-opengl. EDIT - WAS going to give it a go with just this upgraded setup, but it segfaulted within 5 mins. If mesa is the key, any idea how to get through the eselect-opengl limitation?
Oh, and one other package that is blocking is emul-linux-x86-opengl.... I tried to activate the r1 version in accept_keywords but it wasnt enough to get rid of the block. So its xorg-server and emul-opengl that is holding back mesa...
So, still stuck with a failed X _________________ ...and it should be known by now |
|
Back to top |
|
|
phobos13013 Apprentice
Joined: 22 Sep 2004 Posts: 277 Location: PNW
|
Posted: Fri Jan 30, 2015 2:51 am Post subject: |
|
|
Bump.... Still crashing requiring hard reboot every five mins whenever video is played.
See above _________________ ...and it should be known by now |
|
Back to top |
|
|
phobos13013 Apprentice
Joined: 22 Sep 2004 Posts: 277 Location: PNW
|
Posted: Sun Feb 01, 2015 8:05 am Post subject: |
|
|
OK, so this one is not solved but my suspicion is no longer centered on nouveau or vicarously mesa, xorg, etc. I am on kernel 3.17 and I am fairly certain this is the dreaded lockup bug that plagued 3.17/3.18. Why did this get let into the repositories if there is known/unresolved/unlocatable issues with this kernel?! My life was fine until this damned thing entered my system. I have tried backing down kernels but to no avail. This is unresolved, still have the issue and until the kernel issue is sorted it seems i am stuck with an unusable gentoo. _________________ ...and it should be known by now |
|
Back to top |
|
|
|