On 8/8/07, Colleen Beamer <[EMAIL PROTECTED]> wrote: > Jesús Guerrero wrote: > > On Wed, 08 Aug 2007 17:57:53 -0500 > > Colleen Beamer <[EMAIL PROTECTED]> wrote: > > > > Hello > >> 'API mismatch. This nvidia component has version 100.14.09, but the > >> nvidia's kernel modules version does not match'
This sounds like you didn't reboot after emerging nvidia-drivers. What I believe has happened (my hypothesis, at least) is that you compiled/installed the new kernel, rebooted, logged into X, then compiled/installed nvidia-drivers, and tried to restart X. The problem comes from the fact that the previous version of the nvidia module is still loaded into the kernel. X looks to see if it needs to load the nvidia module, sees one currently loaded, so doesn't load the new one. Try this: 1. Log out of X (if you are logged in) 2. Switch to a virtual terminal (<CTRL>-<ALT>-<F2>) 3. /etc/init.d/xdm stop 4. rmmod nvidia 5. /etc/init.d/xdm start And see if that fixes it. If it doesn't, then do steps 1-4 again, then do: modprobe nvidia dmesg | tail -n10 uname -a modinfo nvidia | grep vermagic And e-mail back any error messages the modprobe gives at that point, and the output of the dmesg command and the modinfo command. Hope that helps! -James > >> > > This is not relate to nvidia. It would have happened the same with your ATI > > board. > > > > This means that the kernel modules version and the X driver versions do not > > match. This can have an easy explanation: probably your /usr/src/linux > > symlink is not pointing to the correct directory, so, the module was > > compile for any other kernel. And your current kernel is still using the > > old piece, while Xorg is using the new one. > > I'm not stupid - ls -al linux points to linux-2.6.21-gentoo-r4 the > latest version of gentoo-sources installed and this is what is compiled > on my laptop. I rebooted after compiling the kernel and reinstalled the > nvidia driver. I followed the nvidia How To to the letter. And I still > get the same message. > > > > You can use 'uname -a' to check your kernel version. Then use > > 'eselect kernel list' to find the that same version in the list, and then > > you can use 'eselect kernel set <x>' to set the correct one. > > I know all this. I also know how to check what the linux symlink in > /usr/src is pointing to. The last time I tried this (before writing > this message) I had run genkernel --menuconfig all to check to make sure > all my settings agreed with what was in the nvidia How To. > > > > > > > Emerge nvidia drivers again, rmmod nvidia, and modprobe nvidia. Look > > at dmesg's output to check there's no error loading nvidia. Now launch > > X. If it starts, all is ok. If not, probably your kernel was compiled > > with a different version of glibc/gcc than the nvidia-driver. So, you > > will need to recompile and reinstall your kernel, and then emerge > > nvidia-drivers again. That is the big picture. If you have any doubt > > just ask. > > I've done all this and as I stated, the previous version of > nvidia-drivers that was installed before I upgraded worked fine. And as > previously stated, I just recompiled the kernel today, so it was *not* > compiled with a different version of gcc. > > And I still think nvidia cards are crap! > > > > Regards, > > Colleen > -- > > Registered Linux User #411143 with the Linux Counter, http://counter.li.org > -- > [EMAIL PROTECTED] mailing list > > -- [EMAIL PROTECTED] mailing list
