> When Linux comes along, storage devices will
> be ready for use on, for example, the PCI bus.
Yes, and we need to shut them down cleanly, reinitialize them and then
restart the drivers. #2 is difficult from userspace after #1.
> I'm talking about the initial boot process.
> Once you are post-boot (as you would be before
> you'd go into a sleep state) you have a chance
> to initialize devices correctly with ACPI.
Yes, but you have to do legacy init again then and have the driver save
state before that so it can restart properly afterwards. Also, how do you
want to handle conditionals in the AML code? You disagreed with the idea
of a compiler that would build modules and insert them and with the idea
of a static VM. I cannot see how this should work otherwise.
> I believe the chances for kernel inclusion of
> the thick model *right now* are fairly close to
> zero and for me that's pretty much a trump
> for the thin model.
Do you really think that the chance for inclusion into the kernel should
be the unit of measure here? I like a clean solution that I have to apply
manually more than a dirty one that comes with the tarball.
Anyways, there is much to be done before, for example the shutdown/restart
code in the native drivers. Assuming they maintain state is nice, but a
clean shutdown is better.
Simon
PGP public key available from ftp://phobos.fs.tum.de/pub/pgp/geier.asc
Fingerprint: 10 62 F6 F5 C0 5D 9E D8 47 05 1B 8A 22 E5 4E C1
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/