Subject: Re: UltraDMA problems under 1.6F (probably hardware)
To: None <>
From: gabriel rosenkoetter <>
List: current-users
Date: 08/16/2002 22:37:03
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Aug 16, 2002 at 11:50:31AM +0100, Richard Earnshaw wrote:
> So I messed around with this last night (it is a M5229 BTW).  Changing
> 	drvp->n_dmaerrs =3D NERRS_MAX-1;
> in ata.c to
> 	drvp->n_dmaerrs =3D NERRS_MAX-2;
> And my disk runs quite happily in UDMA-2 mode.

Hrm. That same change doesn't fix anything for me; my disk still
degrades all the way to PIO mode 4 (and nothing else). But that's
not too surprising since it was kind of silly for me to try, being
as my wd?s are rooted not at a wdc, but at a pciide. D'oh. Isn't it
kind of redundant to have this stuff in separate places? Isn't a
pciide also a wdc? (Color me uneducated, but if it controls IDE

Also, it seems a bit shady... I mean, n_dmaerrs is getting set to
that because we saw NERRS_MAX while just scanning the bus, if I'm
interpolating what little I just read correctly. I mean, why should
we be seeing errors at all? The disk's still spinning up? (Should we
be letting wdc/pciide settle like we do scsibus?) The BIOS init'ed
it wrong? Maybe those are dumb questions: I don't know how IDE
busses work. (Suggested reading?)

In any case, I'm playing with sys/dev/pci/pciide.c (especially
/^acer_/), but I've done as much as I feel like doing on a Friday
evening without getting anything worth mentioning done. Tomorrow...

gabriel rosenkoetter

Content-Type: application/pgp-signature
Content-Disposition: inline

Version: GnuPG v1.0.7 (NetBSD)