Rgoer, I'm suggesting that ECN could be turned off in *retransmitted* SYN packets, i.e. those where no response was received, so Linux with ECN would interoperate with sites that *drop ECN packets on the floor*. Of course you're right that it would be a bad idea to retry without ECN in the original poster's case, when the connection was refused (i.e. a RST response was received). I'll change the Subject: to avoid further confusion. Jim Rogier Wolff wrote: > On Wed, Oct 30, 2002 at 06:26:10PM -0700, Jim Castleberry wrote: > >>Maybe a silly question: Couldn't Linux turn off ECN in retransmitted >>SYN packets, so it would still work with sites that drop ECN packets >>on the floor? > > > If they were dropping the ECN-enabled packets, yes. However instead > of dropping the packets, they are REJECTING the connection with > a fully valid "I dont' want to talk to you". Retrying the connection > without ECN would cause at least double traffic on connections that > are legitimately refused. > > Roger. > - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
