[tulip] Interesting problem with 2.2.19pre16.

Donald Becker becker@scyld.com
Wed, 7 Mar 2001 01:47:37 -0500 (EST)


On Tue, 6 Mar 2001, Ben Greear wrote:

> I just downloaded the development tulip driver and pci-scan code...
> 
> Compiled with kgcc against the 2.2.19 headers...and was able to
> insmod both modules w/out problems...  (Running RH-7, btw)
> 
> The interesting thing came when I ran a script, that just happens to
> insmod the de4x5 driver...  So, before I ran it, I had eth0-4
> (4 of those are 21143 (ZNYX), and one is a built-in via-rhine).
> After the insmod de4x5 I have 4 new ethX (eth5,6,7,8)
> 
> I assume that I shouldn't be allowed to get into this kind
> of situation...or is that true?

The Tulip driver may be compiled to use either PCI I/O space or memory
space.  It can only check if the region has been claimed when compiled
for I/O space, thus cannot block other drivers (or even itself) from
multiple detections when compiled for memory space accesses.

Donald Becker				becker@scyld.com
Scyld Computing Corporation		http://www.scyld.com
410 Severn Ave. Suite 210		Second Generation Beowulf Clusters
Annapolis MD 21403			410-990-9993