Primary channel B?

[Mike Bilow]

Reinhold J. Gerharz wrote in a message to Mike Bilow:

 RJG> Now wait a minute. It sounds like Eugene Blanchard is
 RJG> referring to the old problem of Linux scanning and assigning
 RJG> /dev/sdX entries from the physical "A" bus before the
 RJG> physical "B" bus of the AHA-2742T, ignoring the EISA setup
 RJG> parameter setting the "B" bus to be treated as the primary
 RJG> bus. The last I heard about this was that it may be a SCSI
 RJG> subsystem problem in Linux, rather than the AIC?xxx driver.
 RJG> (A search of the mailing list archives should provide better
 RJG> information.) 

 RJG> I still have one of these cards, and hope to put it back
 RJG> into service later this year. I doubt that I will set B =
 RJG> primary, though that is how I used to use it, which is why I
 RJG> originally subscribed to this list. 

It is my opinion that this should not make any difference, and that Lilo should
support booting from any device for which BIOS support is provided.  Obviously,
the controller and BIOS must assign devices in a consistent manner from one
boot to the next, but other than that I see no reason why additional
requirements should be imposed.  What should happen is that BIOS should load
the boot sector of the first partition marked "active" in the scanned order,
and then Lilo should be able to load from there.  The scan order should not
play a role unless you are foolish enough to mark more than one partition
"active" simultaneously, in which case you should always get the fist one.

If I want to put Lilo on the master boot record of /dev/sdc and have it boot an
ext2fs partition at /dev/sdc3, Lilo should be able to do that as long as BIOS
knows how to read that device and assigns a consistent device number.  The
issue as concerns Lilo is one of consistency, not absolute device order.

There is no necessary connection between BIOS device numbers and Unix device
names.  Among other things, Unix should be able to support controllers and
devices which are entirely unsupported by BIOS, and therefore from which boot
is impossible.  The AHA154x controllers have this limitation, supporting boot
from PUN 0 or 1 only, and then only if there are less than two IDE (or ST506)
devices on the system.
 
-- Mike



To Unsubscribe: send mail to majordomo at FreeBSD.org
with "unsubscribe aic7xxx" in the body of the message