[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#330366: marked as done (kded activity disrupts video grabbing despite realtime sched)

Your message dated Tue, 27 Mar 2007 10:22:31 -0400
with message-id <9f694b820703270722y174ee097o18e1656d27f04bb5@mail.gmail.com>
and subject line Old bug, do you still have the problem ?
has caused the attached Bug report to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what I am
talking about this indicates a serious mail system misconfiguration
somewhere.  Please contact me immediately.)

Debian bug tracking system administrator
(administrator, Debian Bugs database)
--- Begin Message --- 
Package: kdelibs-bin
Version: 4:3.4.2-3

Hi,

I've been having huge trouble doing frame grabbing from my video camera via
firewire recently. Every ~ 20 seconds a frame got dropped.
And this all despite running a -ck kernel with dvgrab running in SCHED_ISO,
which should normally avoid any and all frame drops quite easily.

Environment: P3/700 with 2.6.13-ck* kernels, CONFIG_PREEMPT and CONFIG_PREEMPT_BKL.

I then changed to IceWM which immediately worked.
So it had to be something within the KDE environment.
A kill -STOP kded then fixed it.

Possibilities:
a) kded is doing way too much and thus stealing too much CPU. Since this should still
   never cause frame drops since dvgrab is SCHED_ISO, it would indicate a scheduling
   weakness. Highly unlikely.

b) kded calling one huge non-preemptive system call, causing huge latency during
   grabbing. --> kernel issue.

c) kded calling non-preemptive system call with too high latency too often.
    kded issue / kernel issue.

Note that top shows kded sometimes having 3% CPU spikes, and strace illustrates
this activity.

dvgrab has been verified to use SCHED_ISO, via schedtool.

Is there a way to improve kded to have less activity within one burst?

Other ideas?

Thanks!

Andreas Mohr

--- End Message ---
--- Begin Message --- 
Submitter was asked to provide more information about this bug
( http://bugs.debian.org/330366 ) four weeks ago.

Since more info was not provided, we are closing the bug.
Feel free to reopen this bug if you are still experiencing this issue
and you have extra information on how reproduce the issue.

Thanks,
Olivier

--- End Message ---
Reply to: