[8245] in linux-scsi channel archive
Re: scatter gather or not to scatter gather
daemon@ATHENA.MIT.EDU (Matthew Jacob)
Fri Feb 25 14:28:24 2000
Date: Fri, 25 Feb 2000 11:00:30 -0800 (PST)
From: Matthew Jacob <mjacob@feral.com>
Reply-To: mjacob@feral.com
To: "Justin T. Gibbs" <gibbs@FreeBSD.org>
Cc: Kurt Garloff <garloff@suse.de>,
Douglas Gilbert <dgilbert@interlog.com>,
The Linux SCSI list <linux-scsi@vger.rutgers.edu>
In-Reply-To: <200002251811.LAA00525@caspian.plutotech.com>
Message-ID: <Pine.BSF.4.05.10002251053200.9763-100000@semuta.feral.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
> >On Thu, Feb 24, 2000 at 12:53:28PM -0700, Justin T. Gibbs wrote:
> >> >Kurt,
> >> >This url is to a white paper by IBM about "packetized" scsi.
> >>
> >> Information Unit Transfers (what you call "packetized scsi") is
> >> already defined in SPI3. The Adaptec U160 controllers can be made
> >> to support it with suitable firmware modifications. The QLogic
> >> U160 controllers may already support it, but I'm not sure.
> >
> >Well, but the firmware will not affect Linux operation, will it?
> >AFAIK, the Linux drivers don't rely on firmware but use their
> >own routines and scripts.
>
> All QLogic drivers rely on QLogic's firmware. I would expect any
> IU support in the QLogic to work transparently to the device driver.
I just got Ultra3 Qlogic cards in the last month- finished the linux version
updates last week. I'll have to admit that I don't pay as much attention to
SPI as I should- I've mostly been concentrating in fibre channel.
As far as I can tell, the only SPI3 type 'features' are the enabling or
disabling of PPR. One of the problems/features of the Qlogic approach is that
you're mostly shielded from low level stuff. This can be good, but it can be a
pain.
For example, in Fibre Channel, I'd love to have the ability to run an
arbitrary ELS (extended link request), say for when the newer T10 ACL approach
gets implemented- but the Qlogic folks are very against this and have been
unwilling to supply a feature for that. I suspect that much the same reasoning
may apply for various SPI3 things.
Note that what you've said about f/w modification also applies to the Qlogic-
I mean, with some more time and energy I *could* rewrite the f/w to do
anything, but it's not something I'd really like to spend my time on.
-matt
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
