On Mon, 22 Sep 2025 02:26:12 +0200, Guido Falsi wrote: > > On 9/22/25 01:00, Gleb Smirnoff wrote: > > On Sun, Sep 21, 2025 at 09:17:15PM +0200, Guido Falsi wrote: > > G> > Is struct nd_ifinfo something that is used as argument to ioctl(2)? > > G> > If so, then adding counter_u64_t into it is not correct. > > G> > > > G> > I would suggest to not revert the change as a whole, but remove the > > G> > counter only and bring struct nd_ifinfo back to original. The counter > > G> > should probably go into struct in6_ifextra that is pointed by > > G> > if->if_afdata[AF_INET6]. > > G> > > G> Thanks for the feedback and suggestion! > > G> > > G> You are definitely right! I'm in the process to create a review for such > > a > > G> change shortly, I'll post it as soon as I'm able to have the code tested. > > > > As long as Herbert confirms that the patch fixes the issue for him, please > > consider the change as approved by an src committer and please push it. We > > want to enter the stabweek that starts tomorrow without a ABI breakage. > > > > @Herbert I'm attaching a patch against base sources for you to test. > > It moves the new counter to in6_ifextra structure, removing the ABI > breakage I introduced. Everything should be back to working with this. > > Can you confirm it fixes the issue for you?
Thanks a lot, this patch fixes the issue.
