Subject: RE: ym driver
To: 'M.Drochner@fz-juelich.de' <M.Drochner@fz-juelich.de>
From: Adam Glass (Exchange) <adamg@Exchange.Microsoft.com>
Date: 09/29/1999 08:22:06
I'm confused here about your comment about the PNP0C01 node having too many
ranges. The limitation on the number of memory ranges is specific to PNPISA
because it only has 4 "BARs" for memory descriptors. The prohibition on
mixing the 24 bit and 32 bit memory descriptors is also specific to ISAPNP.
From: Matthias Drochner [mailto:firstname.lastname@example.org]
Sent: Wednesday, September 22, 1999 5:07 AM
Cc: email@example.com; firstname.lastname@example.org
Subject: Re: ym driver
> EISA ID: PNP0C01
> mem: 0-9ffff flags b mem: f0000-fffff flags a mem: 100000-400ffff
> flags b\
> mem: 4010000-403ffff flags a mem: fff00000-ffffffff flags a\
> mem: fef80000-feffffff flags a
Another board which doesn't follow the specs. Normally,
only 4 memory ranges per device are allowed.
Soren obviously increased the limits...
(Did you use a fixed limit or something dynamic?)
> When will the pnpbios support be committed? :)
The basic framework is almost ready for the tree.
There could be some code shared with isapnp, but
dev/isapnp has so many problems that I'd prefer to
not use it as sample.
There are also some ideas how to improve ressource
management which would need some work.
The device attachments are hacks partly - mostly
because the driver backends are quite ISA specific.
This might improve once the stuff is in the tree.
I'm quite busy until Thursday next week. After that,
we can think about it.