On Wed, Mar 29, 2000 at 12:07:29PM +0200, Ingo Molnar wrote:
>
> Victor, the lowlatency patch takes care of all code paths that keep the
> CPU for more than ~200-400 usecs without looking at current->need_resched.
> While such work (the lowlatency patch) is always inexact because
> source-level latency validation of 2 million lines of code is pretty much
> impossible (although work on that front is advancing and some of the most
> obvious abusers were fixed in 2.3), 'practical measurements' (Senno's
> measurement tool + a loaded system) show good results. RTLinux is so
> lowlevel (because it provides it's own small & atomic API orothogonal to
> the base kernel) that source-level validation is much easier.
My reservation about the low-latency patch is
in the method of fixing these long sections. If inode reclaim is too
slow, we can either introduce more checks for needs resched or try to
find a smart data structure/algorithm that does the work faster.
If read/write is too slow, figuring out some way to force it into
mmap or do something else smart is an alternative to checking needs
resched.
I wish that one of the vendors would put together a big test system
to evaluate patches on a range of systems and conditions.
-- --------------------------------------------------------- Victor Yodaiken FSMLabs: www.fsmlabs.com www.rtlinux.com FSMLabs is a servicemark and a service of VJY Associates L.L.C, New Mexico.- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Fri Mar 31 2000 - 21:00:25 EST