Hi Lukas, On Thu, Apr 23, 2020 at 06:05:42PM +0200, Lukas Wunner wrote: [...] > Daniel submitted a revert of this series but didn't cc me: > > https://lore.kernel.org/netdev/bbdee6355234e730ef686f9321bd072bcf4bb232.1584523237.git.daniel@xxxxxxxxxxxxx/ > > In the ensuing discussion it turned out that the performance argument > may be addressed by a rearrangement of sch_handle_egress() and > nf_egress() invocations. I could look into amending the series > accordingly and resubmitting, though I'm currently swamped with > other work. > > The question is whether that's going to be sufficient because Daniel > mentioned having an in-tree user as a prerequisite for accepting this > feature, to which Pablo responded with NAT64/NAT46. I don't have > intentions of implementing those, but maybe someone else has. I'd love to work on integrating that feature, there are a few implementations outthere that might be useful for this. However, I'm terribly biased, I'm the Netfilter maintainer. Even though, I really think this hook is going to be very useful for the Linux community from day one.
