Re: [RFC net-next 4/4] tcp: defer regular ACK while processing socket backlog
flat view
From: Stephen Hemminger <stephen@networkplumber.org>
Date: 2023-09-06 23:07:33
On Wed, 6 Sep 2023 13:31:17 -0700 Yuchung Cheng [off-list ref] wrote:
quoted
Signed-off-by: Eric Dumazet <redacted>Signed-off-by: Yuchung Cheng <redacted> Looks great. Any reason any workload may want to switch this feature off? I can't think of one...
Never underestimate the likelihood that some TCP stack or middlebox is broken and needs more ACK's to work.