8 Apr
2005
8 Apr
'05
6:49 p.m.
On Fri, 2005-04-08 at 15:33 -0500, Jack O'Quin wrote:
Lee Revell <rlrevell(a)joe-job.com> writes:
Yup, and the whole thing started with Paul's post to LKML last year to
the effect that "2.6 latency sucks".
Basically the RT preempt kernel achieves
determinism by making every
code path in the kernel preemptible, except for a few like the scheduler
(and timer ISR, for now) that fundamentally can't be made preemptible,
and those few code paths can be analyzed to ensure that they execute in
constant time. This is achieved by turning all spinlocks into priority
inheriting mutexes.
That's interesting, I didn't know that. So, it seems Ingo's patchset
really *is* hard-RT. I'll be surprised to see full hard-RT integrated
into the base kernel.
But, maybe as a Kconfig option some day.
Linus has not said much about it, but I think this is the plan.
Lee