On Fri, Oct 30, 2026 at 06:58:36AM -0400, Martin Mirchev wrote: > Hi Amaury, > Thanks for merging the first one, and for the rewrite. Your version suits > me, and the single check is cleaner. You are right about the allocation > failure, and about the backport going further back than 3.1. AF_NETLINK was > only how I reproduced it, so it is fine to leave it out of the message. > I applied your patch to current master (995295edc) and ran the same test > again. For the inherited netlink socket, "debug dev fd" now prints "laddr=? > raddr=?", and the process keeps running.
Great, thanks for your feedback, I'm merging it as well. -- Amaury Denoyelle
