lkml.org 
[lkml]   [2004]   [Dec]   [7]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: [patch] Real-Time Preemption, -RT-2.6.10-rc2-mm2-V0.7.32-2

    * Gene Heskett <gene.heskett@verizon.net> wrote:

    > I'd like to report a slight improvement in the reports from tvtime
    > while running 32-6 with full preemtion turned on, cfq scheduler:
    >
    > ------
    > Dec 7 13:29:39 coyote kernel: wow! That was a 15 millisec bump
    > Dec 7 13:29:39 coyote kernel: `IRQ 8'[838] is being piggy.
    > need_resched=0, cpu=0
    > Dec 7 13:29:39 coyote kernel: Read missed before next interrupt
    > -----
    > Each second of the log contains about 27 of these, so its pretty
    > steady. Picture looks pretty good though.
    >
    > Although, that particular snip was taken from the log while I was
    > *not* on the same screen as tvtime. Switching back it its screen
    > returned the slip times to the 22 millisecond area. [...]

    could enable WAKEUP_TIMING and LATENCY_TRACING, boot into the new kernel
    and do:

    echo 0 > /proc/sys/kernel/preempt_max_latency

    then you'll get the worst-case trace in /proc/latency_trace. Does it
    show any millisecs-range latencies?

    also, do this:

    chrt -f 90 -p `pidof 'IRQ 8'`
    chrt -f 91 -p `pidof 'IRQ 0'`

    to make sure IRQ8 doesnt get preempted by other stuff. How often do you
    get the rtc histogram in the syslog? Does tvtime use /dev/rtc for normal
    operations perhaps? If yes then you might want to disable RTC_HISTOGRAM.

    Ingo
    -
    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 14:08    [W:0.023 / U:1.504 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site