6.3 And VIA 8237S Controller

I just bought a new MSI P4M900M2 mobo.  It works just fine
with both
Windoze and SUSE Linux.  When I tried booting 6.2 on it, it
refused to
set the drive (ad0 - I tried several different drives) into
the
higher speed UDMA modes.  So, I downloaded 6.3, and it
*seemed* to
be fine.  The drives come up as UDMA 100 or UDMA 133.

But ... under long disk operations - say untaring a 2G
tarball
stored on a USB drive - I start to see this:

ad0: WARNING WRITE_DMA UDMA ICRC ERROR

And eventually:

ad0: FAILURE WRITE_DMA Status=51 .... Error=84

g_vfs_done():ad0s1f ...


What's going on here? Is there a known driver problem with
the VIA
chipsets?  I took the two drives I tried this with, and
stuck them
in another machine - no problem, so I kind of doubt this is
a drive
problem.  I have replaced the IDE cables as well.  Again,
this same
mobo and drive combo worked flawlessly doing the same thing
under SUSE
Linux, so I'm thinking this is a software problem.

Any help much appreciated...
-- 
------------------------------------------------------------
----------------
Tim Daneliuk     tundratundraware.com
PGP Key:         http://www.tundraware.
com/PGP/

_______________________________________________
freebsd-questionsfreebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-que
stions
To unsubscribe, send any mail to
"freebsd-questions-unsubscribefreebsd.org"

Tim Daneliuk wrote:

One point of clarification I neglected to mention in the 
description below.  I have not actually installed FreeBSD
on the disk.  I paritioned/labeled the disk with the
install
disk, then rebooted the install disk, went into the Fixit
environment and manually mounted ad0x under the various 
/mnt directories.  I then insert the USB drive into the
system that has a full image of FreeBSD from another
machine
on it, stored in a tarball, and mount it under /mnt/mnt.
I then start to untar it (to load that image onto my
newly labeled disk), and that's when I see the errors.
The OS running at that time is the FreeBSD 6.3 Fixit
environment.


> I just bought a new MSI P4M900M2 mobo.  It works just
fine with both
> Windoze and SUSE Linux.  When I tried booting 6.2 on
it, it refused to
> set the drive (ad0 - I tried several different drives)
into the
> higher speed UDMA modes.  So, I downloaded 6.3, and it
*seemed* to
> be fine.  The drives come up as UDMA 100 or UDMA 133.
> 
> But ... under long disk operations - say untaring a 2G
tarball
> stored on a USB drive - I start to see this:
> 
> ad0: WARNING WRITE_DMA UDMA ICRC ERROR
> 
> And eventually:
> 
> ad0: FAILURE WRITE_DMA Status=51 .... Error=84
> 
> g_vfs_done():ad0s1f ...
> 
> 
> What's going on here? Is there a known driver problem
with the VIA
> chipsets?  I took the two drives I tried this with, and
stuck them
> in another machine - no problem, so I kind of doubt
this is a drive
> problem.  I have replaced the IDE cables as well. 
Again, this same
> mobo and drive combo worked flawlessly doing the same
thing under SUSE
> Linux, so I'm thinking this is a software problem.
> 
> Any help much appreciated...


-- 
------------------------------------------------------------
----------------
Tim Daneliuk     tundratundraware.com
PGP Key:         http://www.tundraware.
com/PGP/

_______________________________________________
freebsd-questionsfreebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-que
stions
To unsubscribe, send any mail to
"freebsd-questions-unsubscribefreebsd.org"

Tim Daneliuk wrote:
> I just bought a new MSI P4M900M2 mobo.  It works just
fine with both
> Windoze and SUSE Linux.  When I tried booting 6.2 on
it, it refused to
> set the drive (ad0 - I tried several different drives)
into the
> higher speed UDMA modes.  So, I downloaded 6.3, and it
*seemed* to
> be fine.  The drives come up as UDMA 100 or UDMA 133.
> 
> But ... under long disk operations - say untaring a 2G
tarball
> stored on a USB drive - I start to see this:
> 
> ad0: WARNING WRITE_DMA UDMA ICRC ERROR
>

I have resolved this and thought I'd share with the class in
case
anyone else runs into the problem.

It occurred to me that this chipset has been around long
enough that
it was very likely not a driver problem. I went back and
replaced the
IDE cable with another one known to be good and, voila',
problem
solved.

What's weird about this is that the "bad" cable is
a more-or-less new
low profile round IDE cable I got from Tiger Direct a while
back. It
is the 20" variety which may be contributing noise to
the problem.
Weirder still is that neither Linux nor Windows seemed to
have
problems with it, though I did not test as thoroughly with
those OSs.
I'd guess that the FBSD driver is perhaps trying to squeeze
the last
bit of optimization out of the controller and thus drives
the IDE bus
to its limits, hence the problem shows up there.  Either
that,
or I just didn't pound on the machine hard enough with
Linux
especially to see the problem.

I should have guessed "cable problem" right away,
but given the
relative newness of the cable, that seemed unlikely.

In a related note: I also discovered that the FreeBSD
install CD Fixit
environment does flakey things when you try to untar a large
file from
a USB drive plugged in through an external hub. Plugging the
drive
directly into one of the mobo ports made that problem go
away.




-- 
------------------------------------------------------------
----------------
Tim Daneliuk     tundratundraware.com
PGP Key:         http://www.tundraware.
com/PGP/

_______________________________________________
freebsd-questionsfreebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-que
stions
To unsubscribe, send any mail to
"freebsd-questions-unsubscribefreebsd.org"