PCI Support && CVS updated
From: Smári P. McCarthy (email@example.com)
Date: Tue 07 May 2002 - 08:13:26 UTC
Next message: Smári P. McCarthy: "CVS updated"
-----BEGIN PGP SIGNED MESSAGE-----
Start with formalities:
Checking in drivers/pci/pci.c;
Checking in drivers/pci/pci_list.h;
PCI bus scanning works perfectly. Our next task on the PCI front would be to
add the PCI devices to a table in kernel memory. Then we should start loading
modules, but that's a later task - first thing's first: For the PCI table to
work properly, we want a linked list. But linked lists imply memory
management, and hence we need that first. That's why all further PCI
development is *on hold* until we've finished the other few minor things we
need for Yaxic 0.0.1 (Brjánn). Ahh.. yes... I can see it clearly now..
The pci_list.h file really sucks. I mean, it sucks more than anything that has
ever sucked before. It's bloated, it's cruddy, and it's standard. However,
the standardisation part can be fixed.
- if anybody is in the mood for a little Perl-o-suprimo then he may please
take http://www.yourvote.com/pci/pcidev.csv and create something that
converts it into a file that looks almost identical to pci_list.h - but skips
the crud. Please?
- if anybody is in the mood for a little malloc() super-fun (after we
implement malloc()), he may please create a system for updating the tables in
pci_list.h during runtime (loading it from a file?) *this is not important*
Now.. next up: i8253 watchdogs.. úúú!
Smári P. McCarthy
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org
-----END PGP SIGNATURE-----