RE: [RFC PATCH 1/2] x86/mce: Handle AMD threshold interrupt storms

[Luck, Tony]

> It looks to me most of the code can be shared except in few places
> where AMD and Intel use different registers to set error thresholds.

Hopefully easy to abstract.

> And the fact that AMD's threshold interrupts just handles corrected
> errors unlike CMCI.

That makes your life much simpler than mine :-)

> I'm thinking of coming up with a shared code between both by keeping
> the Intel's new storm handling code as base and incorporating AMD
> changes on them and send for review.
>
> Let me know if thats okay?

My new Intel code hasn't had Boris look through it yet to point
out all the bits where I have bugs, or just make things more complex
than they need to be.

So it would be helpful if Boris could do at least a quick scan to
say my code is a worthy base. I'd hate to see you waste time
building a merged version and then have Boris say "Nack".

-Tony