[1243] in linux-scsi channel archive
Re: AHA2742T - switch channels (Linux) ?
daemon@ATHENA.MIT.EDU (deischen@iworks.InterWorks.org)
Thu Jan 9 06:16:54 1997
Date: Thu, 9 Jan 1997 05:17:52 -0600
From: <deischen@iworks.InterWorks.org>
To: steffen@gfz-potsdam.de
Cc: aic7xxx@FreeBSD.ORG, linux-scsi@vger.rutgers.edu
> Well, that's o.k.as long as there are no other hosts in the system, the
> order of which also would be reversed. IMHO, the necessary changes should
> be done in the low-level code (as there we can determine whether the BIOS
> swapping of the channels is in effect) - perhaps by re-ordering the make
> of the entries in the host list ?
No, the low-level code should only inform the mid-level SCSI
code which channel is primary (or better yet, the order that
the channels be probed). It doesn't make sense to have all
the low-level drivers duplicate the same code in order to
re-order devices. The mid-level SCSI code is currently responsible
for probing adapters and channels/busses.
This has already been talked about, and future enhancements
/changes to the mid-level SCSI code will try to incorporate
this feature. The low-level drivers will have to be changed
to inform the mid-level SCSI code in which order the channels
are to be probed when this feature is added.
Dan Eischen
deischen@iworks.InterWorks.org
