[lug] Unresolved Symbols
D. Stimits
stimits at idcomm.com
Sat Jun 30 15:36:43 MDT 2001
Justin wrote:
>
> I did not backup the old modules, I removed the kernel packages so
> the /lib/modules/2.4.2/ was gone. I did move the new System.map over
> after I compiled the 2.4.5 kernel. Right now I'm just back to the stock
> setup with all the 2.4.2-2 kernel packages and no more 2.4.5 stuff
> anywhere. Here is what my /boot and /lib/modules looks like if it makes
> any difference:
>
> total 3924
> drwxr-xr-x 2 root root 4096 Jun 29 10:44 ./
> drwxr-xr-x 32 root root 4096 Jun 29 10:44 ../
> -rw-r--r-- 1 root root 512 Apr 20 08:07 boot.0800
> -rw-r--r-- 1 root root 5824 Feb 23 12:33 boot.b
> -rw-r--r-- 1 root root 612 Feb 23 12:33 chain.b
> -rw-r--r-- 1 root root 379943 Jun 29 09:51 initrd-2.4.2.img
On and older 7.1 install my initrd is "initrd2.4.2-2.img", not what you
have (very close). Looks like the "EXTRAVERSION" is not considered there
(search the kernel source top level Makefile for EXTRAVERSION). But that
probably would not cause the symbol problem; if there was a required
module during bootup, it would probably mean lilo would fail, depending
on what you did (does /etc/lilo.conf refer to "initrd-2.4.2-2.img" or to
"initrd-2.4.2.img"?).
When you run "uname -r", the Redhat kernel should use the "-2". FYI, I
think there is a newer RH rpm kernel release with some fixes, based on
2.4.3.
If you have created all of this with the "-2" EXTRAVERSION, which the
stock RH kernel is, but the modules were compiled with out it (or the
other way around, but in any case, if the EXTRAVERSION is inconsistent
between kernel, System.map, and modules), then you'll likely have
unresolved module symbols.
If you recompiled and did not do a "make mrproper" first, that too can
cause strange problems.
D. Stimits, stimits at idcomm.com
> lrwxrwxrwx 1 root root 16 Jun 29 09:51 initrd.img ->
> initrd-2.4.2.img
> lrwxrwxrwx 1 root root 14 Jun 28 17:00 kernel.h ->
> kernel.h-2.4.2
> -rw-r--r-- 1 root root 405 Jun 28 17:27 kernel.h-2.4.2
> -rw------- 1 root root 15872 Jun 29 10:40 map
> -rw-r--r-- 1 root root 23108 Feb 23 12:33 message
> lrwxrwxrwx 1 root root 19 Jun 28 16:59 module-info ->
> module-info-2.4.2-2
> -rw-r--r-- 1 root root 13598 Apr 8 17:03 module-info-
> 2.4.2-2
> -rw-r--r-- 1 root root 640 Feb 23 12:33 os2_d.b
> lrwxrwxrwx 1 root root 18 Jun 29 10:44 System.map ->
> System.map-2.4.2-2
> -rw-r--r-- 1 root root 419935 Apr 8 17:03 System.map-
> 2.4.2-2
> -rwxr-xr-x 1 root root 2310234 Apr 8 17:03 vmlinux-2.4.2-2*
> lrwxrwxrwx 1 root root 15 Jun 28 16:59 vmlinuz ->
> vmlinuz-2.4.2-2
> -rw-r--r-- 1 root root 793965 Apr 8 17:03 vmlinuz-2.4.2-2
> [jm at hogan modules]$ pwd
> /lib/modules
> [jm at hogan modules]$ ls
> 2.4.2-2/
>
> Justin
>
> > Justin wrote:
> > >
> > > This all started when I was uprading my RH 7.1 box to kernel 2.4.5
> > > yesterday. I compiled fine and got everything setup and relinked in
> > > /boot. I also (this is where I boneheaded it up) removed the
> > > /boot/initrd.img thinking my system here at work was ide and not
> scsi.
> > > Alas, it is a scsi machine and now it wouldn't boot up, kernel panic
> > > because the boot device wasn't mounted. I tried making a new initrd
> > > with 'mkinitrd /boot/initrd-2.4.5.img 2.4.5' but I got the
> error "All
> > > of your loopback devices are in use!". I tried loading the module
> for
> > > loop with insmod loop, but this is where I started running into all
> > > the unresolved symbol errors. At this point I removed all the 2.4.5
> > > stuff and reinstalled the stock RH 7.1 kernel packages (kernel,
> > > -headers, and -source). Even now though whenever I reboot to the
> > > original kernel I still get the unresolved symbol errors, but now I
> > > get them for everything, ugh. Any idea where I should start?
> > >
> > > Justin
> > >
> > > -----
> > > glow at jackmoves.com
> > > www.jackmoves.com
> >
> > Did you move your old modules to a backup location prior to
> doing "make
> > modules_install"? Did you copy your new System.map to
> > /boot/System.map-someversion (where "someversion" basically matches
> the
> > format of "uname -r")?
> >
> > D. Stimits, stimits at idcomm.com
> > _______________________________________________
> > Web Page: http://lug.boulder.co.us
> > Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> >
> >
>
> -----
> glow at jackmoves.com
> www.jackmoves.com
> _______________________________________________
> Web Page: http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
More information about the LUG
mailing list