At Sat, 21 Mar 2026 22:16:29 -0000 (UTC), mlelstv%serpens.de@localhost (Michael van Elst) wrote:
Subject: Re: network outages
>
> woods%planix.ca@localhost ("Greg A. Woods") writes:
>
> >Isn't the simplest to just make kern.sbmax the right size for the
> >largest given {send,recv}buf_max value?
>
> It's a limit for all socket buffers, not just TCP. And other protocols
> don't have their own "max" values to look at.
I figured as much, but
Surely it could be auto-tuned up if a net.inet.tcp.sendbuf_max value
would bump into it before reaching its own max, and definitely if a
sendspace value exceeds it?
It seems horrible for the network stack to fail to open a TCP connection
if the tuning is wrong, and surely other protocols won't mind if it is
bigger than they need.
--
Greg A. Woods <gwoods%acm.org@localhost>
Kelowna, BC +1 250 762-7675 RoboHack <woods%robohack.ca@localhost>
Planix, Inc. <woods%planix.com@localhost> Avoncote Farms <woods%avoncote.ca@localhost>
Attachment:
pgprhJ_IpEB0W.pgp
Description: OpenPGP Digital Signature