Re: [PATCH bpf-next v3 3/4] xdp: recycle Page Pool backed skbs built from XDP frames

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

From: Jesper Dangaard Brouer <jbrouer@redhat.com>
Date: Thu, 16 Mar 2023 18:10:26 +0100

> 
> On 15/03/2023 15.58, Alexander Lobakin wrote:
>> From: Jesper Dangaard Brouer <jbrouer@redhat.com>
>> Date: Wed, 15 Mar 2023 15:55:44 +0100

[...]

> Thanks for signing up for fixing these issues down-the-road :-)

At some point, I wasn't sure which commit tags to put to Fixes:. Like,
from one PoV, it's not my patch which introduced them. From the other
side, there was no chance to have 0x42 overwritten in the metadata
during the selftest before that switch and no one could even predict it
(I didn't expect XDP_PASS frames from the test_run to reach neigh xmit
at all), so the original code is not buggy itself as well ._.

> 
> For what is it worth, I've rebased to include this patchset on my
> testlab.
> 
> For now, I've tested mlx5 with cpumap redirect and net stack processing,
> everything seems to be working nicely. When disabling GRO/GRO, then the
> cpumap get same and sometimes better TCP throughput performance,
> even-though checksum have to be done in software. (Hopefully we can soon
> close the missing HW checksum gap with XDP-hints).

Yeah I'm also looking forward to having some hints being passed to
cpumap/veth, so that __xdp_build_skb_from_frame() could consume it. So
that I could pick a bunch of patches from my RFC back to switch cpumap
to GRO finally :D

> 
> --Jesper
> 

Thanks,
Olek