Re: [PATCH net-next] net: inline eth_type_trans() fast path

[Alexander Lobakin <aleksander.lobakin@intel.com>]

From: Edumazet@kernel.org <edumazet@kernel.org>
Date: Wed, 30 Sep 2026 19:29:57 +0200

> eth_type_trans() is called once per received packet by most
> Ethernet drivers, and by core helpers (napi_gro_frags(),
> xdp_build_skb_from_*(), veth, tun, tunnels, loopback...).
> 
> With CONFIG_MITIGATION_RETHUNK / SRSO, the call and return
> are not free anymore.
> 
> Add eth_type_trans_inline(), handling the common case inline:
> unicast frame sent to dev->dev_addr, with a real ethertype,
> on a device which is not a DSA conduit. All other frames
> (multicast, broadcast, otherhost, 802.2, runts, DSA)
> are handled by the out-of-line eth_type_trans().
> 
> eth_type_trans() is now a macro calling eth_type_trans_inline(),
> so that all existing callers get the fast path. The out-of-line
> version remains exported, and can be called with
> (eth_type_trans)(skb, dev). bpf_prog_test_run_skb() uses this

Shouldn't it get renamed to e.g. eth_type_trans_slow() to avoid this
confusion?

> to keep tools/testing/selftests/bpf fentry/fexit tests working.
> 
> The inline part is about 25 instructions on x86_64, and the whole

Nice!

> function would be almost twice as big when fully inlined.
> 
> Signed-off-by: Eric Dumazet <edumazet@kernel.org>
> ---
>  include/linux/etherdevice.h | 43 +++++++++++++++++++++++++++++++++++++
>  net/bpf/test_run.c          |  3 ++-
>  net/ethernet/eth.c          |  8 +++++++
>  3 files changed, 53 insertions(+), 1 deletion(-)
Thanks,
Olek