You are not logged in.
See topic. When updating to xf86-input-evdev-2.8.2-2 xf86-input-joystick-1.6.2-3 xf86-input-keyboard-1.8.0-2 xf86-input-mouse-1.9.0-2 and xorg-server-1.15.0-1 X won't start and 'systemctl status kdm.service' states that the X server has died. According to Phoronix this should not happen.
Harvey
Last edited by Harey (2014-01-14 11:06:53)
Linux is like a wigwam: No Gates, no Windows and an Apache inside
Offline
... error messages? logs? something?
"UNIX is simple and coherent" - Dennis Ritchie; "GNU's Not Unix" - Richard Stallman
Offline
Including `nvidia` in your initramfs could do the trick… – perhaps.
Offline
Well, Xorg.0.log states there is an ABI mismatch:
[ 37.122] (II) Loading /usr/lib/xorg/modules/drivers/nvidia_drv.so
[ 37.151] (II) Module nvidia: vendor="NVIDIA Corporation"
[ 37.151] compiled for 4.0.2, module version = 1.0.0
[ 37.151] Module class: X.Org Video Driver
[ 37.151] ================ WARNING WARNING WARNING WARNING ================
[ 37.151] This server has a video driver ABI version of 15.0 that is not
supported by this NVIDIA driver. Please check
http://www.nvidia.com/ for driver updates or downgrade to an X
server with a supported driver ABI.
[ 37.151] =================================================================
[ 37.151] (EE) NVIDIA: Use the -ignoreABI option to override this check.
[ 37.151] (II) UnloadModule: "nvidia"
[ 37.151] (II) Unloading nvidia
[ 37.151] (EE) Failed to load module "nvidia" (module requirement mismatch, 0)
But the phoronix guys tell us this should not happen. Unfortunately they do not mention the driver version number...
Harvey
Linux is like a wigwam: No Gates, no Windows and an Apache inside
Offline
I know that there is a switch that you can add to nvidia to "-ignoreABI" where it won't do the check for working with specific xorg versions, but damnit I'm forget where to append it.
edit: oh, hey look at Harey's org log, it says it right there. Does anyone know where to append this?
Offline
Offline
The ABI check is there for a purpose:
[ 38.018] Loading extension NV-GLX
[ 38.027] (EE)
[ 38.027] (EE) Backtrace:
[ 38.029] (EE) 0: /usr/bin/X (xorg_backtrace+0x48) [0x5834c8]
[ 38.029] (EE) 1: /usr/bin/X (0x400000+0x1871a9) [0x5871a9]
[ 38.029] (EE) 2: /usr/lib/libpthread.so.0 (0x7ff70edbf000+0xf870) [0x7ff70edce870]
[ 38.029] (EE) 3: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7ff708c51000+0x4e427f) [0x7ff70913527f]
[ 38.029] (EE) 4: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7ff708c51000+0x4eb501) [0x7ff70913c501]
[ 38.029] (EE) 5: /usr/bin/X (AddScreen+0x71) [0x435c81]
[ 38.029] (EE) 6: /usr/bin/X (InitOutput+0x3df) [0x4769df]
[ 38.029] (EE) 7: /usr/bin/X (0x400000+0x3961b) [0x43961b]
[ 38.029] (EE) 8: /usr/lib/libc.so.6 (__libc_start_main+0xf5) [0x7ff70dc35b05]
[ 38.029] (EE) 9: /usr/bin/X (0x400000+0x24dae) [0x424dae]
[ 38.029] (EE)
[ 38.029] (EE) Segmentation fault at address 0x18
[ 38.029] (EE)
Fatal server error:
[ 38.029] (EE) Caught signal 11 (Segmentation fault). Server aborting
Seems we have to wait for nvidia to include xorg-server 1.15 into their driver
Harvey
Linux is like a wigwam: No Gates, no Windows and an Apache inside
Offline
Downgrade nvidia or try nouveau.
Offline
Does that mean that xorg 1.15 won't hit extra until nvidia finally supports it or how does Arch deal w/ situations like these?
Offline
Often Arch follows upstream and goes ahead with e.g. the new X server, because devs say it's not their problem if a closed source package can't keep up.
Offline
I totally get your point, but in this case, I guess this would affect at least a quarter(?) of all users; so is there some sort of written policy for issues like this or is it just up to the dev's assessments (which would mean in your described case no X for a large number of users for an unknown period of time)?
Of course, you can say "then switch over to nouveau", but it isn't a secret that many NVIDIA cards simply don't run or at least run quite poorly w/ nouveau…
Edit: I'm not even sure if opening a bug report then would be the right measure for this case.
Last edited by cju (2013-12-29 16:58:12)
Offline
Downgrade nvidia or try nouveau.
downgrading is no option as the driver mentioned by Gusar (319.76) is not available for Arch Linux. As a workaround I am using nouveau for now, but 3D performance is poor and my favourite game is not running anymore through wine although I installed lib32-nouveau-dri
Harvey
Linux is like a wigwam: No Gates, no Windows and an Apache inside
Offline
Unless your card is really really new, it's supported by the 304xx branch (the point of this branch is Geforce 6/7 support, but it can handle all cards that were available when the branch was created). And Arch does package 304xx. So it can serve as a simple workaround for many users.
Then there's the not quite so simple but still not hard method of rolling your own 319.76 packages.
Offline
Got exactly the same error (ABI doesn't match etc.) on a vbox(!), so perhaps this isn't just nvidias fault? Or is it even NVIDIA's and Oracle's fault?
Btw.: Anyone testet yet on an ATI w/ fglrx?
Offline
karol wrote:Downgrade nvidia or try nouveau.
downgrading is no option as the driver mentioned by Gusar (319.76) is not available for Arch Linux.
https://aur.archlinux.org/packages.php?K=319
Maybe you can use A.R.M. to downgrade to previous X server version and wait. This may mean downgrading some other packages that depend on X server, but, if everything else fails, it is worth a try.
Offline
This is 319.32, not 319.76. NVidia seems to have incorporated the xorg-server 1.15 support into a new incarnation of an older driver.
Maybe you can use A.R.M. to downgrade to previous X server version and wait.
I'll stay with nouveau for now and hope for a fast solution from NVidia Just kidding...
Harvey
Linux is like a wigwam: No Gates, no Windows and an Apache inside
Offline
karol wrote:This is 319.32, not 319.76.
I know.
Have you tried bumping the package version? That's another thing you can try.
Offline
Have you tried bumping the package version?
319.76 is not on the ftp server, so bumping the package version does not work. Sigh.
And with nouveau suspend does not work
Harvey
Linux is like a wigwam: No Gates, no Windows and an Apache inside
Offline
Download it manually then. I gave you the links.
Offline
Download it manually then. I gave you the links.
Of course I tried this as well. But since I am on X86_64 and there is no no-compat version guess where this led me... I downgraded to the 304xx-branch for now after verifying that my card is supported there. Nouveau is not working well with my card (A 8600GT, by the way). I'll stay there for the time being.
Thanks for all your help
Harvey
Linux is like a wigwam: No Gates, no Windows and an Apache inside
Offline
I downgrade to nvidia-304.17 and run perfect. My card is 560ti and use ck kernel.
Last edited by botika (2013-12-30 17:45:28)
Offline
Same problem here... after a couple of failed attempts at downgrading xorg-server (to many conflicting interdependancies) I gave up and removed the nvidia driver I had in there in favour of the nouveau driver. Now fan-control for my GPU doesn't work and my pc sounds like it's ready for take-off LOL
I might try later just going back to a previous version of the nividia driver that does support xorg-server 1.15 as stated above.
Offline
The nvidia driver 304 support xorg-server 1.15. Nvidia driver and do not give problems with fans, etc.
I don't understand your problem. If works me with xorg-server 1.15 and nvidia 304.17.
Please if your problem has solved edit post to [solved].
Last edited by botika (2014-01-02 20:15:19)
Offline
Btw.: Anyone testet yet on an ATI w/ fglrx?
Dream on, we don't even have 1.14 support in the [stable] branch of AMD Catalyst. It's probably going to be another month before 1.15 support hits the AMD beta.
Claire is fine.
Problems? I have dysgraphia, so clear and concise please.
My public GPG key for package signing
My x86_64 package repository
Offline