Not able to run Xorg 1.5.3 on a GeForce 9300M (in an EMS64T) using driver ver 180.51

[SIl0x0000]

Hello all,

I recently reinstalled my laptop (sony vaio VGN-CS21S/R) with gentoo (amd64) was (x86). Since then i'm not able to run Xorg with the nvidia binary drivers for linux. After looking at all kinds of log files (dmesg, /var/log/messages, Xorg.0.log) i could not find much help, Xorg only says it can't find devices. There are no NVIDIA error messages in the log files. I read the complete readme file (from the nvidia package) and usin the parts i understood (not everything), i tried several things. The nvidia kernel driver is succesfully loaded by the kernel without errors and it reports my vidieo card in the proc file system. The Xorg server seems to load the nvidia_drv.so library and does not report any errors.

I dont know what to do next.

PS. Xorg starts with the vesa driver enabled

[AaronP]

(II) Primary Device is: ISA

Your X server is very confused about which device is primary. Please uncomment the BusID line you have in xorg.conf and generate a new log. Please also attach the log from the vesa driver.

Typically the X server will do a PCI bus scan and list all of the devices it finds right before the line that reads "(II) System resource ranges:". It looks like your X server didn't do that. I'm not sure what could cause that problem, but it sounds like a bug in your X server or kernel. Could your X server have been built without PCI support somehow?

[SIl0x0000]

Hi AaronP,

Originally Posted by AaronP

Your X server is very confused about which device is primary. Please uncomment the BusID line you have in xorg.conf and generate a new log. Please also attach the log from the vesa driver.

I tried to start te x server with both vesa and nvidia drivers enabled see the attachments for the logging. The results with the nvidia driver is exactly the same. The vesa driver still works.

Originally Posted by AaronP

Typically the X server will do a PCI bus scan and list all of the devices it finds right before the line that reads "(II) System resource ranges:". It looks like your X server didn't do that. I'm not sure what could cause that problem, but it sounds like a bug in your X server or kernel. Could your X server have been built without PCI support somehow?

I built the X server as i usually do on gentoo... I will look into what it exactly compiles... I will post it when i have that clear.

thanx in advance!

PS. In the Vesa log there are resources ranges, so i guess that works or at least X tries too. above the first resource scans there is the same line...

[AaronP]

Okay, it looks like what's happening here is that the new PCI probe routine is failing to find any PCI devices, which causes it to fall back to the old-style probe routine for the vesa driver, which always works because it bypasses PCI and uses the legacy VGA resources directly.

On X servers with the PCI rework, the NVIDIA driver doesn't plug in an old-style probe routine. The X server is failing to detect any of your PCI devices, so it doesn't call the NVIDIA PCI probe routine at all. Since there's no old-style probe routine, it just fails with "No devices detected."

The fundamental problem here is that your X server cannot find any PCI devices. Please run "ldd /usr/bin/Xorg" and verify that it's linked against libpciaccess. If it's not, I suspect it somehow got compiled with PCI support disabled and I would recommend recompiling it and making sure that it finds libpciaccess correctly.

[SIl0x0000]

Originally Posted by AaronP

The fundamental problem here is that your X server cannot find any PCI devices. Please run "ldd /usr/bin/Xorg" and verify that it's linked against libpciaccess. If it's not, I suspect it somehow got compiled with PCI support disabled and I would recommend recompiling it and making sure that it finds libpciaccess correctly.

Its seems that Xorg uses the libpciaccess library with version 0.10.5. I'm going to search on the Xorg site for more information on this. I've to work now...

ldd output printed below

# ldd /usr/bin/Xorg
/usr/bin/ldd: line 161: /lib/ld-linux.so.2: cannot execute binary file
        linux-vdso.so.1 =>  (0x00007fff259ff000)
        libpciaccess.so.0 => /usr/lib/libpciaccess.so.0 (0x00007f721d5b4000)
        libdl.so.2 => /lib/libdl.so.2 (0x00007f721d3b0000)
        libpthread.so.0 => /lib/libpthread.so.0 (0x00007f721d195000)
        libXfont.so.1 => /usr/lib/libXfont.so.1 (0x00007f721cf59000)
        libfreetype.so.6 => /usr/lib/libfreetype.so.6 (0x00007f721ccc8000)
        libbz2.so.1 => /lib/libbz2.so.1 (0x00007f721cab8000)
        libXau.so.6 => /usr/lib/libXau.so.6 (0x00007f721c8b5000)
        libfontenc.so.1 => /usr/lib/libfontenc.so.1 (0x00007f721c6ae000)
        libz.so.1 => /lib/libz.so.1 (0x00007f721c499000)
        libpixman-1.so.0 => /usr/lib/libpixman-1.so.0 (0x00007f721c256000)
        libhal.so.1 => /usr/lib/libhal.so.1 (0x00007f721c045000)
        libdbus-1.so.3 => /usr/lib/libdbus-1.so.3 (0x00007f721be0a000)
        libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0x00007f721bc04000)
        libcrypto.so.0.9.8 => /usr/lib/libcrypto.so.0.9.8 (0x00007f721b88a000)
        libm.so.6 => /lib/libm.so.6 (0x00007f721b609000)
        librt.so.1 => /lib/librt.so.1 (0x00007f721b400000)
        libc.so.6 => /lib/libc.so.6 (0x00007f721b0b7000)
        /lib64/ld-linux-x86-64.so.2 (0x00007f721d7bb000)

I also recompiled libpciacces, no difference (i included the build log in an attachment) for the sake of being complete /lib64 is accessible through a symbolic link /lib

thanx again

[SIl0x0000]

As always the answer to difficult propblems was fairly simple. I did not have a sysfs mounted on /sys. Where it depends on this mechanism. When i mounted the pseudo filesystem it worked very nice... going tweaking now.

It might be handy that this is reported somehow, maybe to let libpciaccess depend on the kernel sysfs configuration option? and maybe an easy warning or debug message when sysfs is not mounted.

thinking on this.

for now YaY!

thanx for helping

[AaronP]

Aha, of course! I forgot that libpciaccess uses sysfs for that. I'll keep that in mind in case anyone else has similar problems.