tech-kern archive

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

Re: SCHED_M2 livelock



On Fri, Jan 04, 2008 at 06:08:26PM +0100, Joerg Sonnenberger wrote:
> On Fri, Jan 04, 2008 at 04:31:48PM +1100, Daniel Carosone wrote:
> > I'll experiment further (say, with more or less MAKE_JOBS) to see how
> > the problem moves around.  With 3 of the 4 gmakes running and one
> > stopped, i have 3 makes taking 50% of a cpu, and the rest of the cpu
> > used by cc and others making some progress.
> 
> I haven't seen any such issues without MAKE_JOBS, so it is not unlikely
> to be a gmake bug as well.

Very possibly - gmake is certainly behaving anti-socially, whether the
root cause is purely internal or made worse with some outside
assistance from (say) poll timeouts not running to term or something.

My concern at this stage was to highlight the scheduler issue that
allows 4 such miscreants to effectively kill the system -- even "pkill
-STOP gmake" regularly took 5-10 mins to get enough time in to take
effect. 

Fixing gmake to stop unintended behaviour is one thing; preventing
intentional miscreants from reproducing this result is another.

--
Dan.

Attachment: pgpu9LjpgvL9W.pgp
Description: PGP signature



Home | Main Index | Thread Index | Old Index