[cfe-dev] RFC: Support x86 interrupt and exception handlers

[Joachim Durchholz via cfe-dev]

Am 22.09.2015 um 09:56 schrieb Hal Finkel:
> The overall cost of doing the (hardware-level) mode switch can be
> only low 10s of cycles on modern hardware,

I see. That's still quite expensive given that modern CPUs would like to 
execute multiple instructions per cycle. More importantly, it's more 
than what I'd expect a typical interrupt stub to take but not by 
multiple orders of magnitude, so this could be just relevant enough to 
make a difference on an interrupt-heavy system.
I guess it's back to "this needs to be measured to decide anything" :-)

Regards,
Jo