Subject: Re: Disklabel problems.
From: David Laight <firstname.lastname@example.org>
Date: 11/11/2003 16:24:29
> It is interruptable. I can <ctrl>-c out of it and my machine is no worse
> for the wear.
Off hand I can't think where the loop would be. The only likely one
is in the kernel trying to follow the extended partition chain.
(But I don't think that is interruptable, and should have a check that
the sector number is increasing_.
Some freebsd disklabels have been known to cause disklable to core dump.
Maybe there is something odd in sector 2...
> > > Can anyone offer any suggestions on how to fix this or how to futher
> > > diagnose the problem?
> > What does NetBSDs fdisk say about wd0?
> > Depending on the version (I've forgotten when 1.6U was) one or more -v
> > will give more info.
> fdisk displayed the partition table correctly. I can't get to the machine
> at the moment, so I can't paste it in here, but it did show the correct
You need 'fdisk -vvv sd0' to get all the info displayed.
> > If you run current (probably later than 1.6U) the kernel ought to have
> > code to generate a label from the extended partition list.
> Hmmm... Well, I've grabbed -current source. I'll update tonight. Will
> the system automatically see this label, or do I have to write it to the
The kernel will try to read a label from a variety of places:
sector 1 of a type 169 (NetBSD) partition
sector 1 of a type 165 (FreeBSD) partition
sector 1 of the disk
failing that it will create one containing the partitions from the MBR
and extended partition list.
With current the netbsd partition can be an extended partition - although
you might fail to boot that with grub.
David Laight: email@example.com