lilo segfaults

Automatically imported from Debian bug report #271593 http://bugs.debian.org/271593

Message-Id: <email address hidden>
Date: Tue, 14 Sep 2004 10:39:50 +1000
From: Peter Chubb <email address hidden>
To: <email address hidden>
Subject: lilo segfaults

Package: lilo
Version: 22.5.9-6
Severity: grave

I see:
sudo lilo
/sbin/lilo: line 23: 21978 Segmentation fault $0.real "$@"

I rebuilt from source, and ran under GDB:

(gdb) r
Program received signal SIGSEGV, Segmentation fault.
0xb7f16c78 in strncmp () from /lib/tls/libc.so.6
(gdb) bt
#0 0xb7f16c78 in strncmp () from /lib/tls/libc.so.6
#1 0xb7fd51b0 in timezone () from /lib/tls/libc.so.6
#2 0xbffffb00 in ?? ()
#3 0x08052b3e in boot_image (spec=0x807d420 "/boot/vmlinuz-2.6.9-rc2",
    descr=0xbffff8d0) at boot.c:70
#4 0x0805c6b7 in do_image () at bsect.c:1456
#5 0x0805d83f in cfg_do_set (table=0x80781e0, item=0x807d2e0 "image",
    value=0x807d420 "/boot/vmlinuz-2.6.9-rc2", copy=0, context=0x80781e0)
    at cfg.c:475
#6 0x0805da40 in cfg_parse (table=0x80781e0) at cfg.c:533
#7 0x0804c73c in main (argc=0, argv=0xbffffb74) at lilo.c:825
(gdb)

Libc6 is version 2.3.2.ds1-16; platform is Pentium-4.

I suspect that frames #1 and #2 are artefacts; boot_image()
calls strncmp() directly.

/etc/lilo.conf is (without the comments):

---
vga=791
lba32
boot=/dev/hda
root=/dev/hda3
install=menu
map=/boot/map
delay=20
append="hdc=cdrom acpi=force elevator=deadline"
default=2.6.9-rc2

image=/boot/vmlinuz-2.6.9-rc2
        label=2.6.9-rc2
        root=/dev/hda3
        read-only

image=/boot/vmlinuz-2.6.9-rc1
        label=2.6.9-rc1
        root=/dev/hda3
        read-only

image=/boot/vmlinuz-2.6.8-rc4
        label=2.6.8-rc4
        root=/dev/hda3
        read-only

image=/boot/vmlinuz-2.6.8-rc3
        label=2.6.8-rc3
        root=/dev/hda3
        read-only

image=/boot/vmlinuz-2.4.20-xfs
        append="hda=scsi hdb=scsi hdc=scsi hdd=scsi hde=scsi hdf=scsi hdg=scsi h
dh=scsi apm=power-off nomce"
        label=Linux
        initrd=/boot/initrd.gz
        read-only

image=/vmlinuz.old
        label=LinuxOLD
        read-only
        optional

Message-ID: <email address hidden>
Date: Tue, 14 Sep 2004 10:52:53 +1000
From: Peter Chubb <email address hidden>
To: <email address hidden>
Subject: Worked out the problem...

OK, I tried dropping back to lilo/stable and had the same problem,
which then indicated that the problem wasn't with LILO, but somewhere
else.

Turns out that it's a kernel problem, not a LILO problem: dropping
back to 2.6.8-rc3 meant I could intsall the current 2.6.9-rc2 kernel,
and lilo works again!

--
Dr Peter Chubb http://www.gelato.unsw.edu.au <email address hidden>
You are lost in a maze of BitKeeper repositories, all slightly different.

Message-ID: <email address hidden>
Date: Tue, 14 Sep 2004 10:53:43 +1000
From: Peter Chubb <email address hidden>
To: <email address hidden>
Subject: Close 271593

package lilo
severity 271593 minor
close 271593

Not a bug in lilo. It doesn't affect warty