Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 20 Sep 2000 18:29:21 -0700
From:      Kent Stewart <kstewart@urx.com>
To:        David Uhring <duhring@charter.net>
Cc:        Charlie Root <root@oblomow.demon.nl>, freebsd-stable@FreeBSD.ORG
Subject:   Re: UDMA write errors on stable and release
Message-ID:  <39C96471.4A031E4D@urx.com>
References:  <200009202229.e8KMThw12433@oblomow.demon.nl> <39C94161.19E4D26B@urx.com> <00092018390000.00177@dave>

next in thread | previous in thread | raw e-mail | index | archive | help


David Uhring wrote:
> 
> On Wed, 20 Sep 2000, Kent Stewart wrote:
> > Charlie Root wrote:
> > > Hi,
> > >
> > > I am not sure if this is the right forum to ask, but since it concerns
> > > 4.1 stable I might as well try.
> > >
> > > Since switching to a new motherboard I get the following errors:
> > >
> > > Sep 14 22:37:27 oblomow /kernel.GENERIC: ad0: UDMA ICRC WRITE ERROR blk#
> > > 2523344 retrying Sep 14 22:37:27 oblomow /kernel.GENERIC: ad0: UDMA ICRC
> > > WRITE ERROR blk# 2535904 retrying
> > >
> > > Not only with this installation ( 4.1 Release ) but also with 4.1 stable
> > > (see attached messages file)
> > >
> > > I changed from an Abit AX5 motherboard with a K6 to an Abit KT7 with a
> > > Duron 600 (Via chipset).
> > > Eventually these errors disappear because the system decides to switch to
> > > PIO4 mode for disk access. I have gone through the mailing list archive
> > > and found several more reports about this, but no solution there.
> > >
> > > The cable used to attach these UDMA33/66 capable drives is new and not
> > > too long and should be capable of handling these speeds. The disks that
> > > worked are 2 *  WDC AC34300L.  I also tried a new IBM GXP 30 GB with the
> > > same result.
> > >
> > > Suggestions anyone?
> >
> > Do a "man ata" and learn how to use sysctl to turn off UDMA33+ on your
> > motherboard. The Via chips and the ata driver seem to be very
> > sensitive to some UDMA drive combinations. You can corrupt your data
> > in really bad situations, which occured at 4.0-Release, and degrading
> > to PIO 4 or using SCSI HD's seems to be the only solution at this
> > point. The sysctl commands go into /etc/sysctl.conf
> >
> > Kent
> >
> > > Marc Albers
> > >
> > > /var/log/messages:
>  - - snipped - -
> 
> I had the identical problem with 4.0-R on a FIC-VA503A MB using the VIA MVP4
> chipset.  Do you mean that the VIA KT133 chipset for the Athlon/Duron in
> UDMA66 mode is unsupported by FreeBSD, also?  If so, this is really a shame
> because the UDMA66 chipset improves Bonnie benchmarks by 25-30% with OpenBSD
> and Linux.  BTW, on a FIC PA2013 MB with the MVP3 chipset, FreeBSD fully
> supports UDMA33 mode, at least for the HD.

If you go back to the archives and look at the trouble, most of the
UDMA problems have been associated with the Via chips (in general) and
UDMA. Some combinations of HD's and MB's are worse than others. There
are people with UDMA drives that aren't having problems. There are
people with UDMA drives that were being trashed during the boot. There
are people that replaced the 80-pin UDMA cables with 40-pin cables to
get around the problem. I have 4 UDMA66 drives hooked up to 440bx
boards that don't have problems unless I try to boot them dirty and I
get the message on one system. As soon as I fsck the drives, the
problem went away. I don't think that 4.1-Stable has as many problems
as 4.0-Release did. I don't know any more than this. They transfer
data significantly faster in UDMA mode.

Kent

> 
> Dave

-- 
Kent Stewart
Richland, WA

mailto:kbstew99@hotmail.com
http://kstewart.urx.com/kstewart/index.html
FreeBSD News http://daily.daemonnews.org/


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?39C96471.4A031E4D>