Date: Tue, 02 Jun 1998 17:13:05 -0400 (EDT) From: Simon Shapiro <shimon@simon-shapiro.org> To: Mike Smith <mike@smith.net.au> Cc: "freebsd-scsi@freebsd.org" <freebsd-scsi@FreeBSD.ORG>, "freebsd-current@freebsd.org" <freebsd-current@FreeBSD.ORG> Subject: Re: DPT Redux Message-ID: <XFMail.980602171305.shimon@simon-shapiro.org> In-Reply-To: <199806021707.KAA00985@antipodes.cdrom.com>
next in thread | previous in thread | raw e-mail | index | archive | help
I am deleting the cross-post to current....
On 02-Jun-98 Mike Smith wrote:
...
> Then could you characterise "returning a successful completion code for
> an incomplete/failed transfer"? The SCSI stack has to assume at this
> point that the transaction is complete, even though you're admitting
> that it's not.
There was one specific failure mode, yes. To get there, especially on a
RAID array, you had to have so many ducks lined up:
* DPT free enough to accept commands
* DPT so busy it will not reply to INQUIRY on a RAID array (there is no
SCSI bus involved here)
* DPT not setting the hardware registers as busy, while oh, so busy.
This condition had to persist for several minutes. The patch checked in
against current fixes that. the same patch will/should be checked in
against 2.2 any moment now.
BTW, for this to exist, the DPT firmware has to be rather sick. I can
rplicate this scsnario with some broken, unpublished versions of the
firmware.
Simon
---
Sincerely Yours,
Simon Shapiro Shimon@Simon-Shapiro.ORG
770.265.7340
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?XFMail.980602171305.shimon>