Subject: probe/attach and msize
To: None <tech-kern@NetBSD.ORG>
From: Michael C. Richardson <mcr@sandelman.ottawa.on.ca>
List: tech-kern
Date: 10/02/1997 23:01:27
-----BEGIN PGP SIGNED MESSAGE-----


  1. can someone explain the "broken indirect" stuff to me? Does
- -current still have this problem?

  2. I understand that I'm supposed to unmap my IO and memory at the
end of my probe routine and remap it in the attach routine. 

  Now, if my card did not have a memory address in the config file,
but since there is only four desirable mappings (all software
selectable), I can will scan and pick an available one. 
  I map the memory in the probe to make sure I can reset the
card. (which is the real way to find out if the card is there)
  If I have multiple cards, then should I be leaving each card mapped?
If I don't, then I may not discover during probe that I don't have
enough memory spaces to map all the cards (since all can use the first
mapping). [I don't wish to compensate in software]

  Should my probe leave the memory space mapped, mark it in
ia->maddr/ia->msize, or should it find a space again. What does the
attach routine do if this fails?

  Yes, I have the 4.4 Daemon book... but 4.4. doesn't have bus space
macros.

   :!mcr!:            |  Network security programming, currently
   Michael Richardson | on contract with SSH IPSEC (http://www.ssh.fi/)
 WWW: mcr@sandelman.ottawa.on.ca. PGP key available.



-----BEGIN PGP SIGNATURE-----
Version: 2.6.3ia
Charset: latin1
Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface

iQB1AwUBNDRgBaZpLyXYhL+BAQG/AQL+KQ5sve84hZo5jCIL65KaY2P+6JY3UjFr
2NUE9w3ZjtPMhYSoDofBqHpD2tecYeEAFRn296xThsM1Ujagqq4b6u6hCvmAb80v
i9dzSMqP+mOsdblmItIqMhnaAlXV0Boa
=xIVH
-----END PGP SIGNATURE-----