Re: [Fastboot] Re: Kdump Testing

From: Randy.Dunlap
Date: Mon Apr 25 2005 - 18:11:22 EST


On Mon, 25 Apr 2005 17:45:43 +0530
Nagesh Sharyathi <sharyathi@xxxxxxxxxx> wrote:

> vgoyal@xxxxxxxxxxxxxxxxxxxxxxx wrote on 23/04/2005 09:00:03:
>
> > Quoting "Eric W. Biederman" <ebiederm@xxxxxxxxxxxx>:
>
> > > Nagesh Sharyathi <sharyathi@xxxxxxxxxx> writes:
> > >
> > > > Here is the console boot log, before the machine jumps to BIOS
> > > > after hang during panic kerenl boot
> > >
> > > Ok thanks. So this is manually triggered with SysRq
> > > and the kexec part works but the recover kernel simply fails
> > > to boot.
> > >
> > > It looks like that hunk of the ACPI code that messes up maxcpus=1
> > > needs to be looked at.
>
> > It works well with Uniporcessor capture kernel. For the time being
> sufficient
> > to capture the dump but it is always good idea to be able to boot
> > and SMP kernel
> > as well.
> >
> > Vivek
> I verified on my machine where earlier kdump used to fail and after
> disabling CONFIG_SMP(ie CONFIG_SMP=n) crash kernel boots properly and I am
> able to take the memory dump


Thanks for those hints. However, my testing didn't go quite
as well as that.


2.6.12-rc2-mm3 reboots vmlinux-recover-UP on panic.
(vmlinux-recover-SMP hangs during [early] reboot, but -UP
goes further....)

(BTW, how does I do serial console from the second
kernel...? It has the drivers, but not the command
line info? TBD.)

vmlinux-recover-UP gets to this point, hand-written,
several lines missing:

kfree_debugcheck: bad ptr c3dbffb0h. ( == %esi)
kernel BUG at <bad filename>:23128!
invalid operand: 0000 [#1]
DEBUG_PAGEALLOC
EIP is at kfree_debugcheck+0x45/0x50

Stack dump shows lots of ext3 cache and inode functions...

On a dual-proc P4 with 1 GB RAM.
--
~Randy
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/