Le ven. 9 oct. 2026 à 19:15, Neal Cardwell <[email protected]> a écrit : > > From: Neal Cardwell <[email protected]> > > On Fri, Oct 9, 2026 at 1:47#AM Eric Dumazet <[email protected]> wrote: > > > Let me come back to tp->tsorted_lost_queue [1]. I was too terse there > > ("often") and it did not get discussed. > > Thanks, Eric. Given that you are OK with the space overhead for a new > tp->tsorted_lost_queue field, I think it would be quite nice to use your > suggested approach, since it should give us better loss recovery behavior. > > FWIW, I asked an LLM to code up your idea, and below is what it came up with, > in case that is useful for discussion or using. It also proposed 2 packetdrill > tests (one that claims "This is the case that Eric suggested for segments with > TCPCB_EVER_RETRANS."), which I can post if there is interest. >
Excellent, thanks Neal. I will take a deeper look on Monday.
