[Ns-bugs] [Bug 424] TCP FIN notification callback needed

[code@nsnam.ece.gatech.edu]

http://www.nsnam.org/bugzilla/show_bug.cgi?id=424





--- Comment #11 from Tom Henderson <tomh at tomh.org>  2009-10-07 00:16:32 EDT ---
(In reply to comment #10)
> Created an attachment (id=610)
 --> (http://www.nsnam.org/bugzilla/attachment.cgi?id=610) [details]
> Correct broken close semantics/implementation
> 
> Fixes several problems with TCP close semantics.  Appears to fix bugs
> 326, 559, 663, and 664 also.
> 

I think that the newly proposed socket API is necessary.  Because it isn't
plumbed in for NSC yet, and because it is too close to code freeze, I suggest
to split this patch into two patches; one dealing with TCP internal closing
fixes (that also fixes other TCP bugs) and one dealing with the two new socket
callbacks, and keep this open and at P2, and that we resolve this in the next
month including for NSC.

One question I have concerns when NotifyNormalClose should be called.  The
proposed patch calls it when the TCP receiver has received the FIN in sequence
(or plugs all remaining sequence gaps) but is not tied to whether the
application has read through to EOF.  Another possible implementation would
call it only after the application has read through to EOF.  


-- 
Configure bugmail: http://www.nsnam.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.