Date: Mon, 4 Jun 2007 15:37:54 +0400 From: Gleb Smirnoff <glebius@FreeBSD.org> To: Nikos Vassiliadis <nvass@teledomenet.gr> Cc: freebsd-net@FreeBSD.org, Alexander Motin <mav@FreeBSD.org> Subject: Re: ng_pptpgre + netmask_change results to lock Message-ID: <20070604113754.GA89017@glebius.int.ru> In-Reply-To: <200706040957.38906.nvass@teledomenet.gr> References: <1179829392.00743801.1179818401@10.7.7.3> <465D60C5.50503@freebsd.org> <20070601111457.GY89017@FreeBSD.org> <200706040957.38906.nvass@teledomenet.gr>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Jun 04, 2007 at 09:57:36AM +0300, Nikos Vassiliadis wrote: N> On Friday 01 June 2007 14:14, Gleb Smirnoff wrote: N> > On Wed, May 30, 2007 at 02:32:21PM +0300, Alexander Motin wrote: N> > A> Nikos Vassiliadis wrote: N> > A> >> In my case doing a few times: N> > A> >> ifconfig fxp0 192.168.1.71/16 N> > A> >> ifconfig fxp0 192.168.1.71/24 N> > A> >> locks the box. N> > A> N> > A> Doing this you are dropping all routes going via that network. N> > Probably you A> are dropping route to your pptp peer which can lead to N> > wrapping tunnel A> inside himself. N> > N> > Yes. This is a known problem with netgraph tunnels. N> N> I see. Would it be possible/desirable to control this kind N> of situations using a LINK[012] flag? I mean having code/kludge N> doing loop detection? I think that in this case loop detection can be implemented only using mbuf tags. -- Totus tuus, Glebius. GLEBIUS-RIPN GLEB-RIPE
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070604113754.GA89017>