Red Hat Bugzilla – Bug 437573
loader segv after DHCP when using 'ip=dhcp ipv6=auto method=http://...'
Last modified: 2008-08-02 19:40:36 EDT
loader in today's rawhide (20080314, anaconda-126.96.36.199-1) will crash when
booted with the following commandline:
ip=dhcp ipv6=auto method=http://[url]
It gets a SIGSEGV right after attempting to fetch stage2.img. It works fine if
you're booting from boot.iso or other media, because then you have no need to
It still fails if you use 'noipv6'. Manually configuring in the TUI works fine.
There is no backtrace on tty1; tty4 has:
loader: segfault at 2fe5b101 ip 00123853 sp bfc93bf4 error 4 in
Interestingly I'm not able to reproduce this crash on x86_64.
I just tried that here and X started up fine (x86, inside of qemu)... nothing
else on your command line?
Nope, nothing else. See also bug 437909 - Jesse seems to have hit the same thing in qemu.
I'm seeing a loader backtrace on tty1 when I try to use boot.iso from 2008-03-18
w/ http method and http updates.img. Nothing too interesting on tty3 or tty4.
Backtrace points to httpGetFileDesc in loader, then down into libc. This is bare
FWIW, my cmdline is 'ip=dhcp ksdevice=xx:xx:xx:xx:xx:xx noipv6 lang=en_US.UTF-8
clumens mentions that there was a fix for bug 437909 (ip=dhcp + updates) in
today's rawhide. Can you retest and confirm that bug?
This works fine for me with rawhide-20080319. But I'm not using updates.img.
Yeah, it works for me today. Sorry for the delay.