> On Jun 26, 2024, at 2:24 PM, Marek Zarychta <zarych...@plan-b.pwste.edu.pl> > wrote: > > W dniu 26.06.2024 o 06:55, Zhenlei Huang pisze: >> >> The MFC is Done ! >> >> Marek has extensive testing with them. I did stress testing on my dev >> machine. No known issues so far. > Thank you very much for the MFC of the commits set I suggested. I have been > testing these changes for several weeks on 10+ various hosts, including > servers and routers, and I only see positive changes. I hope that no > important commit was omitted in selecting the set for this MFC and there will > be no complaints. > > The only thing I would suggest is to change in future the default value of > net.inet6.icmp6.errppslimit to 200. It will be the same value as defaults for > net.inet.icmp.icmplim. In my case it was required to silence spurious > warnings on two web servers and BGP router. > The value 100 of tunable `net.inet6.icmp6.errppslimit` dates back to 2000, the first time it was introduced [1].
I do not find any RFCs that has recommendation value for it. RFC 4443 section 2.4 (f) [2] has an example default of 10 that is much smaller ( for small/mid-size devices ). I expect the default value of net.inet6.icmp6.errppslimit will stay 100 for much longer time ;) 1. https://cgit.freebsd.org/src/commit/sys/netinet6/in6_proto.c?h=stable/7&id=686cdd19b1b182f2257bc158116e78c5fef84980 2. https://www.rfc-editor.org/rfc/rfc4443.html#section-2.4 Best regards, Zhenlei > With best regards, > > -- > Marek Zarychta