Subject: PCI device mapping method changes.
To: None <current-users@NetBSD.ORG>
From: Chris G. Demetriou <cgd@cs.cmu.edu>
List: current-users
Date: 04/13/1997 18:08:29
FYI:

Today i've checked in a bunch of changes to the way PCI devices are
mapped.

These changes have been fairly well tested on some devices (ahc, ncr,
de, fpa, isp), but have been less well tested on others.  If you run
into any new problems starting with tomorrow's -current with your PCI
device autoconfiguration, you should probably send me mail about it.

The result of these changes is that most of the drivers are now more
correct and robust than they were, in the face of some systems' desire
to avoid enabling memory or I/O spaces on some devices (especially
devices behind PCI-PCI bridges).


chris