Date: Tue, 15 Dec 1998 14:17:40 -0700 (MST) From: "Justin T. Gibbs" <gibbs@narnia.plutotech.com> To: Richard Smith <richard@jezebel.demon.co.uk> Cc: scsi@FreeBSD.ORG Subject: Re: Update: ncr0: timeout problem Message-ID: <199812152117.OAA59717@narnia.plutotech.com> In-Reply-To: <36753F8E.D2A105FE@jezebel.demon.co.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
> Over a month ago, I posted a timeout problem I was having with my
> symbios controller and fujitsu drive, on 2.2.7R
>
> I have now upgraded to 3.0R and the problem although different, still
> remains. I was advised at the time that the problem would be easier to
> diagnose under 3.0, so here goes...
My guess would be that the Fujitsu doesn't like getting lots of tags.
CAM will actually send more transactions to the drive in parallel than
the old SCSI code which could be making the problem worse than before.
Unfortunately, my knowledge of the underlying NCR hardware is pretty
limitted, so I can't say for sure that this is the cause of the problem.
Try reducing the number of transactions that CAM sends to the drive
and see if this helps. You can do this by adding a quirk entry
for your drive in sys/cam/cam_xpt.c. Something like this should get
you started:
{
/* Broken tagged queuing drive?? */
{ T_DIRECT, SIP_MEDIA_FIXED, "FUJITSU", "M2949E-512", "0134" },
/*quirks*/0, /*mintags*/2, /*maxtags*/2
},
If this works, increment the maxtags value until it stops working.
--
Justin
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199812152117.OAA59717>