On Tue, 22 May 2001, Steve Lord wrote:
> I would really encourage you to get the 2.4.4 xfs kernel up and running,
> I am much happier with the code base we have there now, and we cannot yet
> replicate the problem with the development tree. Hmm, maybe fiber is the
> difference, the qlogic driver has a patchy history in linux.
I'd like to check out 2.4.4 as well -- installed 2.4.2 XFS 1.0 a couple of
days ago, and I'd like to say "good work, give yourself a collective pat
on the back" ;-)
Unfortunately, my system has a VIA Apollo Pro chip set, which isn't 100%
happy with 2.4.2 (none of the UDMA modes above 2 seem to work, e.g, and
you get fairly poor perfomance by the looks of it).
Gather 2.4.4 has some fixes for Via Crapollo... how would I go about
downloading 2.4.4 XFS?
I installed 2.4.2 XFS on a 20GB Seagate IDE disk (5,400 rpm, 512KB cache,
ATA-66) as well as a pair of IBM SCSI disks (7,200 rpm, 512KB cache, FW
SCSI 40MBps).
This is probably a Linux issue, but when I boot up from the IDE drive, and
do hdparm -tT, I see 78/15-16MBps results; however, if I boot up from the
SCSI drives, I see 88/9MBps for both the SCSI and the IDE drives. Can't
explain this, unless hdparm is giving some completely bogus results.
--
Regards,
Juha
PGP fingerprint:
B7E1 CC52 5FCA 9756 B502 10C8 4CD8 B066 12F3 9544
|