Date: Tue, 7 Dec 2004 12:35:04 +0300 From: Gleb Smirnoff <glebius@freebsd.org> To: SUZUKI Shinsuke <suz@crl.hitachi.co.jp>, ru@freebsd.org Cc: cvs-all@freebsd.org Subject: Re: cvs commit: src/sys/netinet in_gif.c Message-ID: <20041207093504.GA75987@cell.sick.ru> In-Reply-To: <x7fz2i7la1.wl%suz@crl.hitachi.co.jp> References: <200412061902.iB6J2hJ6000543@repoman.freebsd.org> <x7fz2i7la1.wl%suz@crl.hitachi.co.jp>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Dec 07, 2004 at 02:03:02PM +0900, SUZUKI Shinsuke wrote: S> - you need the same fix to sys/netinet6/in6_gif.c (for IPv4 over IPv6 S> tunnel) S> S> - it seems like sys/net/if_stf.c also needs the similar fix S> (I haven't confirmed it though, I'll check it later) S> S> - Here's a patch to update route cache periodically, to prevent S> unexpected gif failure due to a stale route cache. Is it okay to S> commit it together with your change? The only reasonable value for in_gif_rtcachettl would be 1 second. I don't think a 300 second lag in network is acceptable for us. Even 5 second lag is not. I admit that my fix is dirty and should be reworked. I like Ruslan's idea. We should reallocate route every time when routing table age changes. This does not happen too often, and this will give us ability to switch to a new route as fast as possible with zero packet loss. As soon as Ruslan commits route table aging, I will fix in_gif and in6_gif.c and look at other tunnels. I'll send both of you patches for review. P.S. I've written to net@ a week ago. Why do I get review only after change is commited? -- Totus tuus, Glebius. GLEBIUS-RIPN GLEB-RIPE
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20041207093504.GA75987>