Re: MI non-interlocked atomic ops?

[Mindaugas Rasiukevicius <rmind%netbsd.org@localhost>]

yamt%mwd.biglobe.ne.jp@localhost (YAMAMOTO Takashi) wrote:
> >> which spends (on x86) a small but not immeasurable amount of time in
> >> splx(s), so I was tempted to rewrite it like this:
> >> 
> >>         void
> >>         mbstat_type_add(int type, int diff)
> >>         {
> >>                 struct mbstat_cpu *mb;
> >> 
> >>                 mb = percpu_getref(mbstat_percpu);
> >>                 atomic_add_uint_ni(&mb->m_mtypes[type], diff);
> >>                 percpu_putref(mbstat_percpu);
> >>         }
> >> 
> >> There is no such routine as atomic_add_uint_ni(), though, and I don't
> >> know if every NetBSD architecture can supply that routine.
> > 
> > Right, spl*()/splx() have a cost.  However, code like this (networking)
> > should not be in the hard-interrupt handler.  As we move towards a wider
> > use of software interrupts, SPL calls will disappear and kernel
> > preemption provides the necessary protection (although software
> > interrupts are not yet preemptable).
> 
> iirc kpreempt_disable() doesn't provide a protection from a softint.
> 

You are right, sorry.  I was thinking of soft-interrupt internals, where
more or less the same mechanism as for kernel preemption is used..

> YAMAMOTO Takashi
> 

-- 
Mindaugas