Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 20 May 2002 13:05:04 +0200 (CEST)
From:      Søren Schmidt <sos@freebsd.dk>
To:        Matthias Andree <ma@dt.e-technik.uni-dortmund.de>
Cc:        freebsd-stable@FreeBSD.ORG, sos@FreeBSD.ORG, freebsd-bugs@FreeBSD.ORG
Subject:   Re: ATA observations in FreeBSD 4.6-RC
Message-ID:  <200205201105.g4KB54iY080077@freebsd.dk>
In-Reply-To: <m3lmaf5ewv.fsf@merlin.emma.line.org>

next in thread | previous in thread | raw e-mail | index | archive | help
It seems Matthias Andree wrote:
> Hi,
> 
> it seems that the ATA subsystem has some bugs in recent FreeBSD-STABLE
> versions, and it looks as though FreeBSD 4.6 is going to ship with these
> bugs. Whom do I seat in front of these to prevent this?
> 
> 1. PR kern/37060: kernel panic with hw.ata.tags=1 in ata-disk.c:710
>    filed April 14 (five weeks ago)
> 
>    The kernel does not boot at all on my machine.
>    This PR has never been replied to, but it's marked as critical with
>    high priority because it's a show-stopper.
> 
>    This MUST be fixed before 4.6 RELEASE. If the fix is: disable ATA
>    tagged queueing altogether not only by default, but also ignore
>    hw.ata.tags, so be it.

This is a known problem (for some) but until now I havn't been able
to reproduce it here at all, and hence have not been able to 
work on a fix. The solution would be to make a note in the docs
somewhere that there is potential problems on some HW wtih this..

> 2. to be filed shortly: at initial boot, all my three ATA drives
>    attached to my VIA KT133 are used in UDMA66 mode. After wakeup from
>    APM standby, all ATA interfaces are reset and all primary drives are
>    in UDMA33 mode henceforth.

Hmm, I've seen on eother report of this, no solution for it yet though.

>    Affects performance of the system, should also be fixed before
>    4.6-RELEASE.

I need help on this as well, I does work on all my VIA based systems :(

> 3. not yet filed, because it's documented: FreeBSD doesn't try Tagged
>    Queueing on WDC 420400D or IBM DJNA drives. Linux 2.5.15 uses tagged
>    queueing on these drives, it seems, it works.

I have some local patches for some of the WDC drives, the DNJA is 
buggy and cant do tags reliably (documented by IBM and easily
verified), the old DTTA series can do tags if the transfer length 
is limitted (also fixed here locally)..

-Søren

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?200205201105.g4KB54iY080077>