[driverloader] kernel oops on X11 shutdown with stk16 kernel
richibb at rochester.rr.com
Mon Oct 4 21:07:56 EDT 2004
That kernel works fine, no strange X11 behavior. Driverloader works
great, too. So I guess it's something in one of the 45 patches (yikes,
45 patches!) included in the FC2 kernel. I'm a little hesitant about
not using any of those patches, since I'm not sure why they're being
applied. I can try reversing some of the patches on the FC2 kernel
sources, to see if I can find the problematic one. Any suggestions on
where to start? Or maybe a different suggestion altogether?
thanks very much for the help,
Linuxant support (Jonathan) wrote:
> please try with a fresh 18.104.22.168 kernel from 'http://www.kernel.org'.
> Before you compile this kernel, please apply this patch :
> You should make sure that the following configuration option is disabled
> and you should enable the following option
> Since you will now use a custom compiled Linux kernel, you will need to
> use the generic RPM package (if you do not already use it) that we offer
> on the DriverLoader download page at :
> Technical specialist / Linuxant
> support at linuxant.com
> Rich Ibbotson wrote:
>> Thanks, Jonathan. It looks like that doesn't help for my case (intel
>> centrino with XP drivers).
>> I still get messages indicating that I need to use a kernel with >4K
>> stack size.
>> Jonathan Baron wrote:
>>> On 10/03/04 08:00, Rich Ibbotson wrote:
>>>> Hi all,
>>>> I'm having some difficulties getting driverloader running on my Fedora
>>>> Core 2 system. Actually, it's not the driverloader code that is the
>>>> problem, it's the "stk16" kernel that I'm using to get around the 4k
>>>> stack restriction in the FC2 kernel.
>>>> The problem:
>>>> When I boot using the 2.6.8-1.521.stk16.i686 replacement kernel,
>>>> in runlevel 5, all is well until I log out of a kde session. At that
>>>> point, instead of getting a graphical login screen, the X server
>>>> and I see a kernel oops. I've included the relevant bits of
>>>> /var/log/messages and /var/log/Xorg.0.log below.
>>>> Note that I am not sure whether or not this is due to the 4k/16k patch.
>>>> I've built a kernel from the sources for 2.6.8-1.521 (from Fedora Core
>>>> 2) with only the change to disable 4 GB kernel-space/user-space vm
>>>> support (no 16k stack patch), and I see slightly different problems -
>>>> sometimes an oops at logout, sometimes a system freeze when I log in
>>>> graphically, sometimes it works fine.
>>>> Any ideas? I've tried building a custom kernel starting from the stk16
>>>> source but re-enabling 4G/4G virtual memory - the resulting kernel
>>>> doesn't even boot. Last thing I see is "Uncompressing Linux... OK,
>>>> booting the kernel".
>>> With Fedora Core 2 and the 2.6.8-1.521 kernel (the latest),
>>> driverloader works for me WITHOUT using a modified kernel (either
>>> custom or the stk16 from Linuxant). You might try that. I did
>>> do the following:
>>> dldrconfig --enable-workaround=stack
>> driverloader mailing list
>> driverloader at lists.linuxant.com
More information about the driverloader