Brian Kress wrote:
> LVM has to run a userspace program vgscan to load the location
> of the lvm partitions into kernelspace. It does this by scanning
> EVERY device in /dev for an LVM signature. While I really like LVM,
^^^
Nope. Only block devices.
> this part strikes me as rather unclean. Devfs does help with this
> a bit, since then at least its not scanning non-existant devices. An
/proc/partitions is used. So again only known block devices are
scanned.
And this needs to be done in order to detect pv signature and
construct the volume groups.
> MD style kernel-based autodetect for LVM would be really nice, but I
> think Heinz is really busy (since LVM 0.8 for 2.2.x still hasn't
> shown up).
IMHO there is no need to rush and bloat the kernel.
It is not too hard to place vgscan/vgchange in the kernel but
there are other options (like a temporary root block device)
which should be considered.
IMHO the initrd aproach should work reasonable for now.
Michael
-- Michael Marxmeier Marxmeier Software AG E-Mail: mike@msede.com Besenbruchstrasse 9 Phone : +49 202 2431440 42285 Wuppertal, Germany Fax : +49 202 2431420 http://www.msede.com/- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun May 07 2000 - 21:00:18 EST