[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[lwip-devel] [bug #50614] TCP: new ACK processing, unsent queue check an

From: Joel Cunningham
Subject: [lwip-devel] [bug #50614] TCP: new ACK processing, unsent queue check and pcb->sndbuf update in wrong spot
Date: Wed, 29 Mar 2017 15:55:04 -0400 (EDT)
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_4) AppleWebKit/603.1.30 (KHTML, like Gecko) Version/10.1 Safari/603.1.30

Follow-up Comment #4, bug #50614 (project lwip):

I agree that ACKs <= pcb->lastack would no longer invoke the check of
pcb->unsent.  Checking pcb->unsent for that case should NOT be unnecessary
because we would have already performed the check when lastack was updated. 
The only way a segment could remain is if the ACK didn't fully ack the
segment, at which-point an ACK greater than lastack would be required to ACK
it.  Further, if a check was really needed in that case, it should also be
performed on unacked as well (which hasn't been the case)

Good point on combining the two loops, they do almost identical logic on the
queues.  For right now, I'd like to minimize the change and commit this patch.
 Another patch can generalize the logic to operate on both loops


Reply to this item at:


  Message sent via/by Savannah

reply via email to

[Prev in Thread] Current Thread [Next in Thread]