Date: Sat, 17 Oct 1998 21:49:33 +0200 (CEST) From: Søren Schmidt <sos@FreeBSD.ORG> To: brett@lariat.org (Brett Glass) Cc: sos@FreeBSD.ORG, bugs@FreeBSD.ORG Subject: Re: Crash workaround: ATAPI and ATAPI_STATIC required in 2.2.7 wd driver Message-ID: <199810171949.VAA11529@sos.freebsd.dk> In-Reply-To: <4.1.19981017130810.06f04590@mail.lariat.org> from Brett Glass at "Oct 17, 98 01:23:16 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
In reply to Brett Glass who wrote:
> At 09:05 PM 10/17/98 +0200, Søren Schmidt wrote:
>
> >Hmm, I doing a rewrite (slowly :( ) of the ata/atapi system, but the
> >problem you see I've never heard of before. What kind of drives do
> >you have
>
> We have several, but the machine that acted up first had a Western Digital
> Caviar 32500. Here are the messages displayed during a boot with the -v
> (verbose) boot option:
>
> wdc0 at 0x1f0-0x1f7 irq 14 on isa
> wdc0: unit 0 (wd0): <WDC AC32500H>
> wd0: 2441MB (4999680 sectors), 4960 cyls, 16 heads, 63 S/T, 512 B/S
Well, some WD drives are a bit wierd, the 1.2G i had said it could
handle DMA, but fu*ed up real bad when told to,...
> >and what flags do you have on the wd driver ??
>
> The defaults.
Ok.
> >The only thing that I can come up with is that the ATAPI options
> >alters the timing a bit...
>
> Are you sure? It seems to me that there might be an error in buffering
> (hence the page faults) or in IRQ handling. I don't know the code
> well enough to tell, though. (I'd love to become a FreeBSD kernel
> hacker, but can find no good reference on register use conventions,
> timing conventions, kernel data structures, etc. -- and I know that
> one can't always glean these subtleties from the code. The one time
> I hacked the wd driver and published some changes, I was told that
> I'd broken a bunch of rules, though the code somehow worked. But
> anyway, this is a side issue.)
I just looked again (in -current though), and I cannot see how adding
the ATAPI* options should heal some disk problem. You didn't use
any atapi drives right ??
> There's got to be some reason why the code is trying to access
> invalid addresses in one case but not the other....
Well, yes :)
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Søren Schmidt (sos@FreeBSD.org) FreeBSD Core Team
Even more code to hack -- will it ever end?
..
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199810171949.VAA11529>