Re: [PATCH] I3 sched tweaks...

[Paul Mackerras]

Ingo Molnar writes:

> It's
> slightly cheaper to pass an already existing (calculated) 'current'
> pointer over to another function, instead of calculating it once more in
> that function.

On x86 that's true; many other architectures - alpha, ia64, m68k,
mips, mips64, parisc, ppc, ppc64, sparc, sparc64 - keep current in a
register already and thus it is slightly more expensive to pass it as
a parameter instead of just using current in the function.

Either way surely the cost is tiny, and the maintainability
considerations should prevail.  Having a function which takes a
task_struct * parameter which _has_ to be current sounds to me like an
invitation for somebody to get it wrong down the track.

Paul.