[anonsec] I-D Action:draft-ietf-btns-connection-latching-06.txt
Nicolas Williams
Nicolas.Williams at sun.com
Thu Apr 10 08:19:42 PDT 2008
On Thu, Apr 10, 2008 at 04:22:02PM +0200, Daniel Migault wrote:
> Maybe there is a solution to drop the SUSPEND state and merge it with the
> BROKEN state by considering different transition condition from ESTABLISHED
> to BROKEN.
That's exactly what I did:
<CREATE_LISTENER_LATCH(3-tuple, ...)>
:
v <CREATE_CONNECTION_LATCH(5-tuple, ...)>
/--------\ : :
+------|LISTENER|...... : :
| \--------/ : : : +--------------------+
| : : : : |Legend: |
| : : : : | dotted lines denote|
| <conn. trigger event> : : | latch creation |
| (e.g., TCP SYN : : : | |
| received, : : : | solid lines denote |
| connect() : : : | state transition|
| called, ...) v v : | |
| : /-----------\ : | semi-solid lines |
| : |ESTABLISHED| : | denote async |
| <conflict> \-----------/ : | notification |
| : ^ | : +--------------------+
| : | <conflict>
| : <conflict | :
| : cleared> | :
| : (OPTIONAL) | :
| : | v v
| : /----------------\
| :.....>| BROKEN |.-.-.-.-.-> <ALERT()>
| \----------------/
| |
<RELEASE_LATCH()> <RELEASE_LATCH()>
| |
| v
| /------\
+------------------->|CLOSED|
\------/
> I don't think we have too many states, and I eventually would add CONECTION
> larval state for LC objects. On the other hand, if I really had to drop one
> state I would rather drop larval state like the LISTEN state.
Check out the above diagram. I think it's simple enough now.
Nico
--
More information about the ANONSEC
mailing list