Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 29 Jan 1996 23:55:42 +0100
From:      se@zpr.uni-koeln.de (Stefan Esser)
To:        Julian Elischer <julian@ref.tfs.com>
Cc:        freebsd-scsi@freebsd.org
Subject:   Re: scsi cleanup
Message-ID:  <199601292255.AA04501@Sysiphos>
In-Reply-To: Julian Elischer <julian@ref.tfs.com> "Re: scsi cleanup" (Jan 28, 14:02)

next in thread | previous in thread | raw e-mail | index | archive | help
On Jan 28, 14:02, Julian Elischer wrote:
} Subject: Re: scsi cleanup

} > Get the SCSI target interface working on something other than just
} > the AHA-1542B working - specifically on the two drivers we have control
} > of the firmware (NCR and ahc).  Add interfaces to support TCP-IP over
} > SCSI.
} now we're getting more long-term :)

Yes, I'd be interested in implementing this for the NCR.

} > Reduce the size of the NCR driver so that it is less than 7000 lines.

Definitely true, and I've already been working on that, too.
But note the amount of comments in the file, it could easily
be reduced to half the number of lines by stripping off all
the comments :)

I've already completely designed a new "SCRIPT" language and 
will implement it as soon as possible. It is a much compressed
form of the NCR code to be executed (a copy is required anyway
for execution) and will allow to add support for the advanced 
features of the 53c810A, 53c860 and 53c875 (the last one offers
an on-chip SRAM, which can be used for microcode and variable
storage).

It will require some testing (I won't release it before it hasn't
successfully run for at least a week and one make world per night :)

} Add:
} Justin would like a call-down from the generic code to the adapter at the time of allocating a scsi_transfer struct, so that he can also allocate a matching
} adaptrer level struct and pin them together....

The NCR driver could take advantage of a per controller (instead
of per device) limit of outstanding commands.

I'd also like to see more support for CHECK CONDITION/REQUEST
SENSE by the generic code: The NCR work queue is a circular 
buffer and it takes extra code to start the failed command 
another time, after a CHECK CONDITION situation. If the generic
code re-issued the complete command, the start queue management
could be simplified. (I know there are some callback hooks in 
the generic code, but on first glance it wasn't obvious whether
it is already possible with the current code).

Regards, STefan
-- 
 Stefan Esser, Zentrum fuer Paralleles Rechnen		Tel:	+49 221 4706021
 Universitaet zu Koeln, Weyertal 80, 50931 Koeln	FAX:	+49 221 4705160
 ==============================================================================
 http://www.zpr.uni-koeln.de/~se			  <se@ZPR.Uni-Koeln.DE>



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199601292255.AA04501>