Re: [patch 3/4] genirq: add a quick check handler

[Christoph Hellwig <hch@infradead.org>]

On Thu, Mar 05, 2009 at 11:59:09AM -0800, Sven-Thorsten Dietrich wrote:
> Most of the IRQ handler, whether run in a thread or IRQ context, will be
> the same code - so what you are proposing would have to eliminate code
> duplication as well as heavy runtime branching overhead.
> 
> Ultimately, no matter how its done, the concept of disabling IRQ assert
> at the device level, rather than the apic level, is the optimal
> "correct" implementation.
> 
> Formulating that into the code, as Thomas proposed with the quickcheck,
> supplies structural demarcation for semi as well as software design.


Umm, the code will be look more or less the same either way.  I just
think overloading the current handler to mean two different things is
a bad idea.  For a driver using a quick disable handler and a long slow
threaded one the only difference is naming the two functions
differently.

I wonder if you're still thinking in the way of a -RT like setup where
threaded interrupts can be enabled and disabled globally?  I don't think
we should ever do that for mainline.