> > I root caused the reason for the regression to be the > netmem_is_net_iov() check in the fast path. I removed this regression > in v7 (see the change log) by conditionally compiling the check in > that function. > > In v8, Pavel/Jens/David pushed back on the ifdef check. See this > entire thread, but in particular this response from Jens: It seemed the main objection is about how to enable this feature for the io_uring? And it seemed that you had added the CONFIG_NET_DEVMEM for this devmem thing, why not use it for that? > > https://lore.kernel.org/lkml/11f52113-7b67-4b45-ba1d-29b070050cec@xxxxxxxxx/ > > Seems consensus that it's 'not really worth it in this scenario'. I was only reading through the above thread, it didn't seemed to reach to consensus as Jesper pointed out the performance impact for the XDP DROP case in the same thread. https://lore.kernel.org/lkml/779b9542-4170-483a-af54-ca0dd471f774@xxxxxxxxxx/ >
