On Wed, 2015-11-18 at 12:19 +0100, Hannes Frederic Sowa wrote: > On Wed, Nov 18, 2015, at 11:47, Lorenzo Colitti wrote: > > On Wed, Nov 18, 2015 at 7:19 PM, Hannes Frederic Sowa > > <han...@stressinduktion.org> wrote: > > > I bet there will soon be a timewaitd which handles the not configurable > > > (David has rejected all those patches so far) timeout of TIME_WAIT > > > sockets. And I bet it will be used. :/ > > > > No, SOCK_DESTROY has no effect on TCP_TIME_WAIT sockets or any other > > non-full socket. > > > > When called on any socket where sk_fullsock(sk) is false, it returns > > EOPNOTSUPP because there is nothing to do. Its purpose is to interrupt > > blocked userspace socket calls, not to release resources. > > Okay, thanks for clarification! Still I don't see how you enter > TIME_WAIT in case you kill an active socket. At least my start-up idea > timewaitd seems to be not implementable. ;) > > I was wondering why you didn't use tcp_close function, because still we > could have the address and we would like to do a proper shutdown of the > connection. While this patchset wants to tear down sockets for addresses > no longer alive, it still can be used with full sockets.
I guess this could work, but we would have to also propose a mechanism where no FIN/RST message is sent. This might be the time for upstreaming our TCP_SILENT_CLOSE implementation. /* on close(): free sock, no FIN/RST */ It seems there is a lot of confusion on this thread about these patches. Proposal is _not_ changing TCP stack behavior, as David Laight seems to fear. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
