* Kieran Mansley <[EMAIL PROTECTED]> [071220 13:55]: > I'm working on merging this now. The code to split the pbuf and unacked > head is pretty complex though. I wonder if a simpler solution would be > to just make a new pbuf when you want to send a persist timer probe, > copy in the first byte from the unacknowledged queue, and send that > (without adding it to the unacknowledged queue).
What I didn't tell you (but I guess you already found out) is that using the split function you also get the possibility to fill up the other sides receive window. Without spit, lwIP relies on the that the other side eventually will open up its receive window and that we then can transfer the data in the unsent queue. In our application we pump data between two hosts running lwIP where one of the hosts rust echoed back the data it received. In that setup/application we managed to end up in a deadlock situation if we didn't fill up the receivers window. That is why the split function exists. As you say, it is not really necessary for the persist probes. /PH -- Per-Henrik Lundblom epost: [EMAIL PROTECTED] telefon: 0733-20 71 26 hemsida: www.whatever.nu _______________________________________________ lwip-users mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/lwip-users
