Date: Thu, 10 Jan 2008 13:23:52 +0200 From: Nikos Vassiliadis <nvass@teledomenet.gr> To: cvs-all@freebsd.org Cc: Maxime Henrion <mux@freebsd.org>, cvs-src@freebsd.org, src-committers@freebsd.org, Mike Tancsa <mike@sentex.net> Subject: Re: cvs commit: src/sys/net route.c route.h Message-ID: <200801101323.53589.nvass@teledomenet.gr> In-Reply-To: <200801091554.m09FsnUC049542@lava.sentex.ca> References: <200801091538.m09Fcn0w050430@repoman.freebsd.org> <200801091554.m09FsnUC049542@lava.sentex.ca>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday 09 January 2008 17:54:48 Mike Tancsa wrote: > At 10:38 AM 1/9/2008, Maxime Henrion wrote: > >mux 2008-01-09 15:38:49 UTC > > > > FreeBSD src repository > > > > Modified files: (Branch: RELENG_6) > > sys/net route.c route.h > > Log: > > MFC route.c@1.125 and route.h@1.68: > > Add a workaround for a deadlock between the rt_setgate() and > > rt_check() functions. It is easily triggered by running routed, and, > > I expect, by running any other daemon that uses routing sockets. > > Hi, > I have a number of boxes running full bgp tables with > Quagga. How does the deadlock manifest itself ? I havent seen the > boxes lockup ? I was using routed over a UDP tunnel (/usr/share/examples/netgraph/udp.tunnel) to distribute routing information over some DSL lines. This stopped working when I upgraded one of the routers to FreeBSD-6. I never reported it, feeling that the setup(PPPoE + netgraph UDP tunnel + routed) to reproduce it, is somehow complicated. It seems that after Maxime's changes, the deadlock I was seeing is gone. Testing a similar setup, seems to work, as it did in the pre-FBSD-6 days. Nikos
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200801101323.53589.nvass>