Hi,
On Tue, 19 Jul 2005 01:58:43 +0900, Lee Braiden wrote:
...
> It's a fairly obvious idea, and I'm sure it would have already been done if
> it
> was easy, but what /is/ the thinking on NetBSD drivers in (L4-)HURD?
I would consider more some (HURD-friendly) generic drivers layer and several
adapting layers on top of it for specific OS drivers.
[HURD]
| generic drivers API
/---------------------+------------------------\
| | |
[Linux-kernel-iface] [NetBSD-kernel-iface] [MS-Windows-kernel-iface]
| Linux API | NetBSD API | MS-Windows ABI
[sata_via.c] [???.c] [ntfs.sys]
I wrote such Linux-kernel-filesystem-API <-> MS-Windows-kernel-filesystem-ABI
before
http://www.jankratochvil.net/project/captive/
and the borken monolithic Linux kernel just avoid any reasonable drivers
adaptation, compatibility and security safety layers there.
Sure there should be some proper IDL separation there, maybe like
modules: [HURD modules]
| | | IDL
module: [Linux-kernel-iface layer] ; maybe not needed
| IDL
module: [sata_via.c + iface-stub]
Sorry for design flaws I am not much aware of L4/HURD details.
Regards,
Lace
_______________________________________________
L4-hurd mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/l4-hurd
