[833] in linux-scsi channel archive
Re: Problems with Adaptec 2940 as a module
daemon@ATHENA.MIT.EDU (Leonard N. Zubkoff)
Tue Oct 22 12:59:42 1996
Date: Tue, 22 Oct 1996 09:47:47 -0700
From: "Leonard N. Zubkoff" <lnz@dandelion.com>
To: neuffer@nomis.i-connect.net
CC: johnsonm@redhat.com, linux-kernel@vger.rutgers.edu,
linux-scsi@vger.rutgers.edu, rcoleman@cc.gatech.edu
In-reply-to: <Pine.LNX.3.95.961022030112.5869B-100000@nomis.i-Connect.Net>
(message from Michael Neuffer on Tue, 22 Oct 1996 03:49:08 -0700
(PDT))
By the way, the patch to hosts.c is definitely not acceptable for 2.0.x.
Moving the eata_dma probe earlier in the sequence will quite possibly break
some working configurations by causing the devices to be recognized in a
different order.
For 2.1, we really should come up with a better method of recognizing
controllers in the same order the BIOS will. Unfortunately, as yet I've not
heard of any universally applicable algorithm for doing so. Just in testing
with BusLogic MultiMaster and FlashPoint, I've seen just about every
conceivable ordering in different motherboards.
Anyone have any better ideas?
Leonard
