Re: [OpenTransport] accepting on a different TCP endpoint than listening
Re: [OpenTransport] accepting on a different TCP endpoint than listening
- Subject: Re: [OpenTransport] accepting on a different TCP endpoint than listening
- From: "Thomas E. Knowlton, Jr." <email@hidden>
- Date: Thu, 4 Dec 2003 11:12:50 -0700
Quinn, thanks for your reply. See my responses inline below.
--
Tk!
On Dec 3, 2003, at 11:00 PM, Quinn wrote:
>
At 15:00 -0700 1/12/03, Thomas E. Knowlton, Jr. wrote:
>
> Can someone confirm my assumption that a TCP endpoint (conceptually,
>
> a listening endpoint, with an OTQLen > 1 on the bind) that "hands
>
> off" to an unbound TCP endpoint when calling OTAccept() transitions
>
> back to the T_IDLE state, not (as documented) to the T_DATAXFER
>
> state?
>
>
It definitely transitions to T_IDLE. This is clearly documented in
>
the TPI specification.
>
>
<http://developer.apple.com/macos/opentransport//docs/dev/tpi.pdf>
Excellent, another source of reference. I wonder whether you can tell
me what I should have searched for in order to have found that
document? I have been looking near
http://developer.apple.com/macos/opentransport/
for just such documentation. BTW, there is a broken link on that page,
about 2/3 of the way down, labelled "Inside Macintosh: Networking with
Open Transport". The href should say "documentation" where it now says
"techpubs".
Also, I notice that the directory that contains the document you
referred to is not browseable, and I wonder whether there is more "must
have" documentation near there?
FYI, I had been relying almost exclusively on "Networking with
OpenTransport"
http://developer.apple.com/documentation/mac/pdf/NetworkingOT.pdf
which does not clearly document what happens when you handoff to
another endpoint when T_LISTEN arrives, although it does mention this
approach. Additionally, I have been relying on "OTVirtualServer" sample
to help me along. (I have a copy of Stevens Sockets and XTI but Stevens
lacks details in many cases).
>
>
Specifically, look at Tables 4-2 and 4-7. Table 4-2 describes event
>
"accept2" as "successful return of t_accept() with ocnt == 1, fd !=
>
resfd". Table 4-7 shows that if the state is T_INCON and event
>
accept2 happens, the provide transitions to state T_IDLE.
>
>
If the Apple documentation contracts this, it's simply a bug in the
>
documentation. For my information, I'd like to know where you saw
>
this bug.
Sure, I was reading the text and diagram near Figure 4-2, p93 of the
above mentioned "Networking with OpenTransport". Nowhere in that
document, which seemed to me to be the most comprehensive OpenTransport
document available, can I find a clear definition of what happens when
you handoff an incoming connection to a second endpoint. I will
carefully study this new document you've provided.
--
Tk!
_______________________________________________
macnetworkprog mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/macnetworkprog
Do not post admin requests to the list. They will be ignored.