On 5/20/25 12:32 PM, Eric Dumazet wrote:
On Tue, May 20, 2025 at 2:31 AM Paolo Abeni pabeni@redhat.com wrote:
On 5/14/25 3:56 PM, chia-yu.chang@nokia-bell-labs.com wrote:
This patch uses the existing 1-byte holes in the tcp_sock_write_txrx group for new u8 members, but adds a 4-byte hole in tcp_sock_write_rx group after the new u32 delivered_ecn_bytes[3] member. Therefore, the group size of tcp_sock_write_rx is increased from 96 to 112.
Note that I'm still concerned by the relevant increase of the cacheline groups size. My fear is that this change could defeat some/most of the benefist from the cacheline reorg for all tcp users.
Some additional feedback from Eric and/or Neal more than welcome!
I have been trapped lately with production issues, sorry for the delay.
I am still working on an idpf bug, hopefully done today.
Then, I am OOO tomorrow, and can have a look at the whole series on Thursday.
Thanks Eric!
@Chia-Yu: please consider the above timeline before posting a new revision. i.e. it would be likely wise to wait for the additional review.
/P