Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 22 Dec 1998 18:15:35 +0000 (GMT)
From:      Doug Rabson <dfr@nlsystems.com>
To:        Mike Smith <mike@smith.net.au>
Cc:        Andrew Gallatin <gallatin@cs.duke.edu>, "Kaleb S. KEITHLEY" <kaleb@ics.com>, freebsd-alpha@FreeBSD.ORG
Subject:   Re: Boot kern.flp wedges -- now what? 
Message-ID:  <Pine.BSF.4.01.9812221814460.25052-100000@herring.nlsystems.com>
In-Reply-To: <199812202047.MAA46687@dingo.cdrom.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 20 Dec 1998, Mike Smith wrote:

> > On Fri, 18 Dec 1998, Mike Smith wrote:
> > > > Last I checked, isa devices requiring DMA didn't work yet, and the
> > > > new, pci-only driver wasn't yet done.  (at least I though somebody was 
> > > > working on one...) I'd be happy to be wrong about that ;-)
> > > 
> > > This took a hit when Soren's system and his backup tapes were stolen.  
> > > I'm not sure how much momentum we've lost, but I'd say quite a lot.
> > 
> > Ouch, that hurts.  I was daydreaming about writing a CAM driver for atapi
> > which would support CD and Zip drives.  I spent an hour comparing the
> > atapicd and scsi cd drivers and figuring out the differences in the
> > protocol and I'm sure that the translation needed for the (few) different
> > commands would be minimal.
> 
> ... you could also look at the way that Linux and NetBSD do it.  I 
> would certainly consider hanging ATAPI devices out to the CAM stack a 
> reasonably good idea.
> 
> > Unfortunately fixed disks don't fit into the
> > scsi-like protocol afaik.
> 
> You'd be up for a little more work in the translation process, but not 
> *really* that much; you could probably get away with implementing:
> 
>  Test Unit Ready
>  Inquiry
>  Read10
>  Write10
> 
> and a couple of modepages.  The actual translation wouldn't be very 
> expensive. 
> 
> I was throwing the whole ATA architecture thing around the other day 
> (while kidding myself that I had time to work on it); I guess I see 
> that if we were to integrate it fully with CAM you'd have a layering 
> arrangement with chipset-specific drivers providing an 'atabus' with 
> I/O methods etc. suited to the chipset/bus in question.  Then you hang 
> the generic 'ata' client code off the atabus, which offers generic ATA 
> command services, does device probing, etc.  From that hangs 'atadisk' 
> and 'atapi' drivers to suit, which offer themselves to the CAM layer
> as though they were host adapters.
> 
> This is all moderately complex, and adds the extra worry that for an 
> all-IDE system the code bulk for CAM might be excessive.  I'm not sure, 
> but I do know that I don't have time to implement it, so it's all a bit 
> too much like fantasy.  8(

Unfortunately thats more or less the same place I ended up in after my
daydreaming.


--
Doug Rabson				Mail:  dfr@nlsystems.com
Nonlinear Systems Ltd.			Phone: +44 181 442 9037



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-alpha" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.01.9812221814460.25052-100000>