Richard Gobert wrote:
Eric Dumazet wrote:
I do not understand this patch 4/4 then.
Why bother moving stuff in net/ipv4/tcp_offload.c if we plan to move it back to where it belongs ?
Willem also pointed that out, and I agree. I'll post a v5 and move this functionality to gro.c. Currently, gro_network_flush will be called from tcp_gro_receive and in a separate series I'll fix the bug by calling gro_network_flush in skb_gro_receive or adding it to udp_gro_receive_segment - whichever is better.
This patch is meaningful by itself - removing checks against non-relevant packets and making the flush/flush_id checks in a single place.
One issue: if the do this move in net/next, then a later fix that relies on it cannot be backporter to older stable kernels.