Re: [PATCH 2/2] x86_64: move apic init in init_IRQs

[ebiederm@xmission.com (Eric W. Biederman)]

Andi Kleen <ak@suse.de> writes:

> On Tuesday 04 October 2005 17:13, Eric W. Biederman wrote:
>> All kinds of ugliness exists because we don't initialize
>> the apics during init_IRQs.
>> - We calibrate jiffies in non apic mode even when we are using apics.
>> - We have to have special code to initialize the apics when non-smp.
>> - The legacy i8259 must exist and be setup correctly, even
>>   we won't use it past initialization.
>
> Actually some setups use its timer even in ACPI mode. Do you take this into 
> account?

The patch is basically code motion and duplication removal with just a
few tweaks thrown in to account for the change in location in the
kernel boot process when all of occurs.

So everything that was working should continue to work.

> In theory it looks reasonable, but I guess it won't work without the bogus
> watchdog change. That would need to be resolved first.

Sure.  That should be straight forward.  The only part that breaks is
the watchdog initialization/testing.  As I recall there was no obvious
place to put it.  The call has to happen after everything is running
so it is a bit tricky to find a good place to put it.

Eric