lkml.org 
[lkml]   [2008]   [Apr]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [git pull] scheduler changes for v2.6.26
From
Date

On Tue, 2008-04-22 at 14:25 +0200, Mike Galbraith wrote:
> On Tue, 2008-04-22 at 10:59 +0200, Ingo Molnar wrote:
> > * Mike Galbraith <efault@gmx.de> wrote:
> >
> > > > the 700-800 msecs of delays you see are very "brutal" so there must
> > > > be something fundamentally wrong going on here.
> > >
> > > I'm seeing latency hits with 26.git, whereas 25 is hit free.
> >
> > ok. I think what happens is that your broken sched-clock hid the real
> > breakage. Lets try fix the real breakage now.
> >
> > I've uploaded a new sched-devel.git that is against very latest -git,
> > could you try ftrace (with a sufficiently large
> > /debug/tracing/tracing_max_entries value) - perhaps the
> > worst-case-wakeup-latency tracer shows large latencies? If not, then
> > maybe the sched_switch tracer gives a better insight into what's
> > happening?
>
> Well, looks like this is going to have to wait. sched-devel is locking
> hard on me, both sched-devel and .git are losing my darn keyboard right
> in the middle of things, and my serial ports are (still) so screwed up I
> can't get minicom working right. Arg.

I dug my old p3/500 out and hooked it up since I still can't get minicom
functional plugging P4->Q6600, but no joy. Whatever the heck is going
wrong in sched-devel is so catastrophic that not one character hits the
console when it locks, and nmi_watchdog=1 does nada.

FWIW, all I am doing is listening to music in amarok while a make -j4
glibc build is running, watching latencytop and watch-rq-clock.sh (nice
-15). Latency hits are smaller, but still present without NO_HZ btw.
The largest I ever saw in latencytop before death arrived was < 300ms.
(still size mondo)

config attached.

-Mike
[unhandled content-type:application/x-gzip]
\
 
 \ /
  Last update: 2008-04-23 10:15    [W:0.079 / U:0.024 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site