I've just resolved a problem that's been reported to me off-list. It
turns out that
1) This is a bug introduced in 2.88
2) If it's triggered, it's bad: dnsmasq logs about cache internal error
and the DNS subsystem becomes broken.
3) Actually triggering the bug is pretty unlikely: I've only seen one
report of it, and it took a weeks work to be able to reproduce it.
4) The bug is more likely to occur when DNSSEC validation is in use, but
still quite unlikely: I have several production instances with DNSSEC
validation and have not seen it.
Since the bug is quite bad, and since there's no workaround that
guarantees to suppress it, even though the chances of seeing the problem
are very low, I think it makes sense to make a new, fixed, release
fairly rapidly. I'll work though the uncontroversial patches in my
queue, with the aim of making a 2.89 release towards the end of the week.
https://thekelleys.org.uk/gitweb/?p=dnsmasq.git;a=commit;h=f172fdbb77c422e27d3b7530f3fe95b98d1608e7
has the fix and the explanation.
Cheers,
Simon.
_______________________________________________
Dnsmasq-discuss mailing list
Dnsmasq-discuss@lists.thekelleys.org.uk
https://lists.thekelleys.org.uk/cgi-bin/mailman/listinfo/dnsmasq-discuss
