Subject: Re: 160MB/s (was Re: scsibus capabilities.)
To: NetBSD Kernel Technical Discussion List <>
From: Greg A. Woods <>
List: tech-kern
Date: 05/22/2002 17:31:04
[ On Wednesday, May 22, 2002 at 22:15:29 (+0200), Manuel Bouyer wrote: ]
> Subject: Re: 160MB/s (was Re: scsibus capabilities.)
> On Mon, May 20, 2002 at 06:35:09PM -0400, Greg A. Woods wrote:
> > [ On Monday, May 20, 2002 at 19:17:56 (+0200), Manuel Bouyer wrote: ]
> > > Subject: Re: 160MB/s (was Re: scsibus capabilities.)
> > >
> > > There are CAM bits in the core portion. It's possible that the linux
> > > version includes a CAM emulation, but as the linux version is GPL'd we
> > > can't use it.
> > 
> > Then there must be CAM emulation bits in the NetBSD version too, no?
> > 
> > Or were the CAM bits removed from the core portion during the last port?
> They were removed, or weren't there.

If they were removed then they can be removed again in much the same way....  :-)

> The driver which was ported, and
> the current freebsd driver are very, very different in the interfaces they
> use.

It must have changed a lot since the last time I looked at it.....  I
know there's been now over a year's worth of evolution...

> > Either way the port of the FreeBSD AHC driver to NetBSD has already been
> > done once and the guy who did it says it'll take another SCSI-familiar
> > driver author about a week to re-do again from the current FreeBSD
> > sources -- without chip documentation.....
> What's silly is that with docs it would take less time to port it properly
> integrated in scsipi.

Are we suddenly talking about the CAM documentation?  I don't understand
why you'd need chip docs to re-port a working driver that has been
ported to not just NetBSD but other platforms as well, and at least one
of those additional ports is freely available for examination.

(of course if something breaks then understanding the details of the
hardware could help a lot to fix it, though still with a very good
working example, or possibly even two, already running to compare

Even if there has been some "interface creep" in the mean time, the
driver's author has stated that other ports have been done quite
successfully with no chip documentation....

I just wish I had at least a solid two weeks I could dedicated to the
task, _and_ the hardware to test with of course......  :-)  Then I could
move some machines back to NetBSD (and maybe even do a head-to-head
comparison between platforms too! :-)

								Greg A. Woods

+1 416 218-0098;  <>;  <>;  <>
Planix, Inc. <>; VE3TCP; Secrets of the Weird <>