Re: Cyrix Detection -- NO SMP, please ?????

Albert D. Cahalan (acahalan@cs.uml.edu)
Sun, 18 Oct 1998 19:48:21 -0400 (EDT)

>> We ought to use self-modifying code anyway, for a generic 386/486
>> kernel. There is a 6-byte sequence that would handle a cache issue.
>> (something Linus came up with a year or two ago) Some of the user
>> access code could be ripped out on the fly too.
>
> It isnt worth the overhead. A 486 wants totally different memory
> alignment before you even get into the differences in cache,
> write protect on supervisor and instruction sets.

Last I heard, Red Hat didn't expect people to recompile the kernel.
That was the whole reason they had you do the modular sound hacking.
The old system with 100 kernel images has been abandoned too.
The only other solution is to drop 386 support. Not that I'd object
to that, but I can't imagine Red Hat doing that any time soon.
(it was dropped for NT 4.0 though, and I've not seen a 386 since 1995)

> Power down 'ought to work' ignores a lot of design aspects of
> SMP motherboards (eg what happens if we halt the CPU
> that gets SMI interrupts and try and power down on the other).

Use the boot CPU. In any case, doesn't the power down signal just
go to the power supply? The system will crash anyway as soon as the
power supply turns off; that is the intent.

> More productive would be for you to download the ACPI spec
> and get programming.

That was reported as buggy beyond all hope.

-
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/