Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 14 Feb 2000 07:37:47 -0700
From:      "Justin T. Gibbs" <gibbs@FreeBSD.org>
To:        Chris Johnson <cjohnson@palomine.net>
Cc:        "Justin T. Gibbs" <gibbs@plutotech.com>, scsi@FreeBSD.org, lamaster@nren.nasa.gov
Subject:   Re: BT-948 timeouts 
Message-ID:  <200002141437.HAA00367@caspian.plutotech.com>
In-Reply-To: Your message of "Thu, 10 Feb 2000 04:55:43 EST." <20000210045543.A57016@palomine.net> 

next in thread | previous in thread | raw e-mail | index | archive | help
>Thanks very much for your response.
>
>Here's what I did:
>
>su-2.03# camcontrol tags da0 -N 8
>
>And here's the result:
>
>su-2.03# camcontrol tags da0 -v
>(pass0:bt0:0:6:0): dev_openings  8
>(pass0:bt0:0:6:0): dev_active    0
>(pass0:bt0:0:6:0): devq_openings 8
>(pass0:bt0:0:6:0): devq_queued   0
>(pass0:bt0:0:6:0): held          0
>(pass0:bt0:0:6:0): mintags       2
>(pass0:bt0:0:6:0): maxtags       255
>
>This seems to have done the trick. I built a world without any timeouts, which
>I hadn't been able to do before I made the change.
>
>Is there a fancy way to make this change survive across boots, or should
>I just shove the camcontrol command in a script that runs at boot time?

You can add a quirk to sys/cam/cam_xpt.c for your drive that will limit
the number of tags to use.

>How would I go about learning what the tags setting should be for a particular
>drive?

CAM can determine the maximum number if you attach the drive to a controller
that will report the QUEUE FULL status (ahc, sym, ncr, aic).  When the system
stops reducing the number of tags (while the device is under a heavy
transaction load), you know that you've found the device's limit.

>Chris
>

--
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?200002141437.HAA00367>