lkml.org 
[lkml]   [2001]   [Sep]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: ksoftirqd? (Was: Re: [PATCH] Preemption Latency Measurement Tool)
    On Sat, Sep 22, 2001 at 10:51:16PM +0200, Roger Larsson wrote:
    > On Saturday 22 September 2001 15.14, Andrea Arcangeli wrote:
    > > On Sat, Sep 22, 2001 at 02:56:58PM +0200, Roger Larsson wrote:
    > > > Hi,
    > > >
    > > > We have a new kid on the block since we started thinking of a preemptive
    > > > kernel.
    > > >
    > > > ksoftirqd...
    > > >
    > > > Running with nice 19 (shouldn't it really be -19?)
    >
    > I repeat this question - should it be nice 19 and not nice -19 (not nice)?

    Being it +19 it is light with respect to the other tasks in the system,
    mainly considering there's one for each cpu.

    > Sep 22 22:37:38 jeloin logger: *** ./stress_dbench begin ***
    > Sep 22 22:37:42 jeloin kernel: ksoftirqd: do_softirq
    > Sep 22 22:37:44 jeloin last message repeated 2 times
    > Sep 22 22:37:48 jeloin kernel: Latency 8ms PID 7 %% kupdated
    > Sep 22 22:37:49 jeloin kernel: ksoftirqd: do_softirq
    > Sep 22 22:38:19 jeloin last message repeated 18 times
    > Sep 22 22:38:20 jeloin su: (to root) roger on /dev/pts/2
    > Sep 22 22:38:20 jeloin PAM-unix2[988]: session started for user root, service
    > su
    > Sep 22 22:38:23 jeloin kernel: ksoftirqd: do_softirq
    > Sep 22 22:38:24 jeloin modprobe: modprobe: Can't locate module net-pf-10
    > Sep 22 22:38:24 jeloin kernel: ksoftirqd: do_softirq
    > Sep 22 22:38:58 jeloin last message repeated 20 times
    > Sep 22 22:39:05 jeloin last message repeated 4 times
    > Sep 22 22:42:03 jeloin kernel: ksoftirqd: do_softirq
    > Sep 22 22:43:58 jeloin logger: *** ./stress_dbench end ***

    one reschedule every few seconds sounds ok.

    > Possibly but the other case is quite common too...

    if you try to put a printk also in do_softirq you'll see why it is
    not the common case :)

    also you may want to put the printk in the schedule() path, ksoftirqd
    may just be rescheduled once, and run many of the do_softirq().

    Andrea
    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

    \
     
     \ /
      Last update: 2005-03-22 13:03    [W:0.023 / U:60.440 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site