Re: [chrony-dev] crash when overloading refclock_sock

[Miroslav Lichvar <mlichvar@xxxxxxxxxx>]

On Wed, Oct 05, 2016 at 10:56:04AM +0200, Nuno Gonçalves wrote:
> I was testing a refclock_sock provider and when the rate of messages
> is really high (over 50k/s), I manage to crash chronyd.
> 
> Yes, this is not the standard use case, usually I publish 1/s, but
> anyway chrony ideally should resist to this abuse.
> 
> 
> 2016-10-05T07:58:01Z refclock.c:707:(filter_add_sample) filter sample
> 52 t=1475654280.876999978 offset=0.122999928 dispersion=0.000000001
> 2016-10-05T07:58:01Z refclock.c:707:(filter_add_sample) filter sample
> 53 t=1475654280.876999978 offset=0.122999928 dispersion=0.000000001
> 2016-10-05T07:58:01Z refclock.c:707:(filter_add_sample) filter sample
> 54 t=1475654280.876999978 offset=0.122999928 dispersion=0.000000001
> 2016-10-05T07:58:01Z refclock.c:707:(filter_add_sample) filter sample
> 55 t=1475654280.876999978 offset=0.122999928 dispersion=0.000000001
> 2016-10-05T07:58:01Z sources.c:360:(SRC_AccumulateSample) ip=[SOCK]
> t=1475654280.876999978 ofs=-0.123000 del=0.000000 disp=-nan str=0
> chronyd: sourcestats.c:388: find_best_sample_index: Assertion
> `best_index >= 0' failed.
> Aborted (core dumped)

Thanks for the bug report. It looks like the problem is that there are
multiple samples with the same local timestamp, which breaks the
calculation of the dispersion.

Can you please pull from git and see if it's fixed?

-- 
Miroslav Lichvar

-- 
To unsubscribe email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "unsubscribe" in the subject.
For help email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "help" in the subject.
Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.