Re: [arch-general] [arch-dev-public] [extra signoff] nvidia/nvidia-utils 169.07
On Tue, Jan 08, 2008 at 09:11:34 +0100, Thomas Bächler wrote:
I added new nvidia drivers 169.07 for i686 and x86_64 to testing. My nvidia machine is out of order for now, so please sign off both architectures and as many GPUs as possible.
I am also interested in signoffs from users, so please don't hesitate to reply to this message to the arch-general list!
everything looks fine on i686, except for this error on upgrade, I don't know how pacman lost track of it, unless I was installinng my own nvidia* a long time ago. nvidia-utils: /usr/lib/libXvMCNVIDIA_dynamic.so.1 exists in filesystem lspci: 01:00.0 VGA compatible controller: nVidia Corporation NV34 [GeForce FX 5200] (rev a1)
Jesse Young schrieb:
everything looks fine on i686, except for this error on upgrade, I don't know how pacman lost track of it, unless I was installinng my own nvidia* a long time ago. nvidia-utils: /usr/lib/libXvMCNVIDIA_dynamic.so.1 exists in filesystem
This is a symlink that was created by ldconfig locally on your system as it was missing in the package in the past. I don't know how to work around this problem, as file-conflicts are checked before I have a chance to delete it, I guess everyone has to -Sf the package.
lspci: 01:00.0 VGA compatible controller: nVidia Corporation NV34 [GeForce FX 5200] (rev a1)
Thanks, that is the one I could have tested as well.
I added new nvidia drivers 169.07 for i686 and x86_64 to testing. My nvidia machine is out of order for now, so please sign off both architectures and as many GPUs as possible.
I am also interested in signoffs from users, so please don't hesitate to reply to this message to the arch-general list! I can try it today.. but with new nvidias I have always a simple
Hello out there On Tue, Jan 08, 2008 at 09:11:34 +0100, Thomas Bächler wrote: problem with libswf (or sth. similar). I do the symlink by hand than.. But for testing it's nothing more than pacman -Sy testing/nvidia ? -- Gruß, Johannes Täglich http://blog.hehejo.de und du fühlst dich gut. http://cryptocd.eduforge.org/online_version
Same conflict error for me brain0, (x86_64) error: could not prepare transaction error: failed to commit transaction (conflicting files) nvidia-utils: /usr/lib/libXvMCNVIDIA_dynamic.so.1 exists in filesystem But more importantly, X fails to start up... here are the last few lines from Xorg.0.log (II) Initializing built-in extension COMPOSITE (II) Initializing built-in extension DAMAGE (II) Initializing built-in extension XEVIE (II) Initializing extension GLX Backtrace: 0: /usr/bin/X(xf86SigHandler+0x6a) [0x49580a] 1: /lib/libc.so.6 [0x2ba54b0aa120] 2: /usr/lib64/tls/libnvidia-tls.so.1 [0x2ba54cddd65e] Fatal server error: Caught signal 11. Server aborting I'm just faced with a black screen. There's no hard lock, and I can still switch tty. So obviously, no signoff from me. Varun
Varun Acharya schrieb:
Same conflict error for me brain0, (x86_64)
Of course, it was a mistake in the old package version .
2: /usr/lib64/tls/libnvidia-tls.so.1 [0x2ba54cddd65e]
This file is not contained in the nvidia-utils package, which means you previously used the nvidia installer to install the driver. This almost always fails and leads to this error when you switch back to the Arch package. We neither use the lib64 directory, nor the tls directory. Instead we install the tls version of the library directly to /usr/lib instead of /usr/lib/tls (as the non-tls version is useless for us).
On Jan 8, 2008 5:35 PM, Thomas Bächler <thomas@archlinux.org> wrote:
This file is not contained in the nvidia-utils package, which means you previously used the nvidia installer to install the driver.
I've been using the arch package without any problems for a while now, and didn't have any problems while upgrading previously. Anyways, I cleared the old files and re ran the upgrade, and it works fine now. A small side note..fps values in glxgears seem to have taken a hit.. previously I was getting ~ 15500fps, now with the new drivers I get ~ 14000fps See http://pastebin.archlinux.org/23421 If this fps drop is expected behavior (since glxgears is not a true index of performance), then a signoff for x86_64 from me. Nvidia 7800 GTX for me btw. Varun
participants (4)
-
Jesse Young
-
Johannes Held
-
Thomas Bächler
-
Varun Acharya