Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx> wrote: > > Yes, since 6c959fd5e17387201dba3619b2e6af213939a0a7 > > the legacy symbol is user visible so next step is to replace > > various "select ...TABLES_LEGACY" with "depends on" clauses. > > Okay. So I would repost the series fixing what the bot complained in > 2/3. The action in case people complain about slow insertion would be: > - Use iptables-legacy-restore if mass insertion is performance critical. > - Use iptables-nft which does not have this problem. > - If both option don't work, copy the counters immediately risking to > miss in-flight updates, free the memory after a grace period. Seems like a good plan, thanks Sebastian. > Any objections? Not from my side.
