Unix VM still coughing sockets...

Andreas Raab Andreas.Raab at gmx.de
Tue May 14 15:25:58 UTC 2002


> I think this one is best left to Ian - he has been digging 
> this through A LOT and did rewrite most of the aio in the
> process... :-)

In which case you should cc: him directly since otherwise he might just
skip this bit of information (as he has said before he's filtering the
list due to overwhelming traffic).

Cheers,
  - Andreas

> -----Original Message-----
> From: squeak-dev-admin at lists.squeakfoundation.org 
> [mailto:squeak-dev-admin at lists.squeakfoundation.org] On 
> Behalf Of goran.hultgren at bluefish.se
> Sent: Tuesday, May 14, 2002 6:24 PM
> To: squeak-dev at lists.squeakfoundation.org
> Subject: Re: Unix VM still coughing sockets...
> 
> 
> Ned Konz <ned at bike-nomad.com> wrote:
> > On Tuesday 14 May 2002 05:51 am, goran.hultgren at bluefish.se wrote:
> > 
> > > I still end up sometimes with Processes in #waitForDataUntil:
> > > running wild on me and chewing up all the CPU. I attached a silly
> > > GIF to show you that I am not hallucinating. (the reason for so
> > > many of them is that the CPU watcher suspends the naughty guy and
> > > asks me what to do and then another one pops up - I guess it is
> > > Comanche doing it)
> > 
> > So what's the timeout? It doesn't seem that waiting on a 
> semaphore for 
> > a timeout (presumably in the multi-second range) could eat 
> up 71% of 
> > the CPU. Waiting should eat precisely 0 until the timeout.
> 
> That is the problem - if you haven't followed this thread the thing is
> (probably, if it's still the same problem) that the semaphore ends up
> with an excessive signal so the "wait" just goes through. Voila - a
> busyloop.
> 
> > Perhaps there's a problem with the CPUWatcher. I'll take a 
> look at it 
> > today.
> 
> I don't think the CPUWatcher is the problem. This is exactly 
> the problem
> Ian "fixed" and sure - it doesn't happen that often now - but it still
> happens (in his newest 3.1beta that is, funny enough not in one of the
> interim VMs).
> 
> > Do you see the system (OS) CPU usage being high as well?
> 
> Haven't checked but I think so - oh, yes - KDEs "CPU guard" 
> kicked in so
> it did notice it too.
> 
> I think this one is best left to Ian - he has been digging 
> this through
> A LOT and did rewrite most of the aio in the process... :-)
> 
> regards, Göran
> 




More information about the Squeak-dev mailing list