On Thu, 9 Jan 2025 at 01:35, Stanislav Fomichev <sdf@xxxxxxxxxxx> wrote:
>
> Commit 86e25f40aa1e ("net: napi: Add napi_config") moved napi->napi_id
> assignment to a later point in time (napi_hash_add_with_id). This breaks
> __xdp_rxq_info_reg which copies napi_id at an earlier time and now
> stores 0 napi_id. It also makes sk_mark_napi_id_once_xdp and
> __sk_mark_napi_id_once useless because they now work against 0 napi_id.
> Since sk_busy_loop requires valid napi_id to busy-poll on, there is no way
> to busy-poll AF_XDP sockets anymore.
>
> Bring back the ability to busy-poll on XSK by resolving socket's napi_id
> at bind time. This relies on relatively recent netif_queue_set_napi,
> but (assume) at this point most popular drivers should have been converted.
> This also removes per-tx/rx cycles which used to check and/or set
> the napi_id value.
>
> Confirmed by running a busy-polling AF_XDP socket
> (github.com/fomichev/xskrtt) on mlx5 and looking at BusyPollRxPackets
> from /proc/net/netstat.
Thanks Stanislav for finding and fixing this. As a bonus, the
resulting code is much nicer too.
I just took a look at the Intel drivers and some of our drivers have
not been converted to use netif_queue_set_napi() yet. Just ice, e1000,
and e1000e use it. But that is on us to fix.
