Re: Problem with 2.0.24 SCSI change

Leonard N. Zubkoff (
Wed, 30 Oct 1996 11:27:47 -0800

From: Chris Adams <>
Date: Wed, 30 Oct 1996 13:10:39 -0600 (CST)

I am running 2.0.24 (with the no_atime patch) on a news server. I have
two SCSI adapters in it: an Adaptec 2940 and a DPT 3224. I have the OS
& swap on sda and the history file on sdb attached to the Adaptec and a
big RAID array for news spool attached to the DPT. When I boot up with
2.0.24, I get "Unable to open initial console" and everything stops. I
finally realized that the DPT was suddenly being detected first. I had
to change drivers/scsi/hosts.c to get the Adaptec back first. I have 2

1) Why was the DPT moved above the Adaptec in 2.0.24?

Michael Neuffer sent a patch that made updates to a number of EATA files, as
well as the common file hosts.c. Despite my having stated publicly that the
reordering in hosts.c was inappropriate, Linus applied the patch without
noticing this. Changes to any common SCSI files are supposed to be approved by
me before they go in, but sometimes screwups happen, as in this case.

I sent Linus email last night asking why he applied it, and he replied that the
1500 email messages awaiting him after his trip had something to do with it.
This reordering will be reverted when 2.0.25 eventually comes out.

2) Is there not a better way to get the order of multiple host adapters?

Not at the present time. There's no consistency to the order in which a
motherboard's BIOS will install the BIOS extensions from the SCSI controllers.
I think we will eventually need to allow for explicitly specifying the probe