From owner-freebsd-current Fri Apr 30 1:29: 4 1999 Delivered-To: freebsd-current@freebsd.org Received: from ewok.creative.net.au (ewok.creative.net.au [203.30.44.41]) by hub.freebsd.org (Postfix) with SMTP id F3EB0158EA for ; Fri, 30 Apr 1999 01:28:59 -0700 (PDT) (envelope-from adrian@freebsd.org) Received: (qmail 14596 invoked by uid 1008); 30 Apr 1999 08:28:56 -0000 Message-ID: <19990430082856.14594.qmail@ewok.creative.net.au> From: adrian@freebsd.org To: John Polstra Cc: freebsd-current@freebsd.org, mike@sentex.net, andrewb@demon.net Subject: Re: Any action on PR 10570 ? getting closer to 65K :-( In-reply-to: Your message of "Thu, 29 Apr 1999 22:55:43 MST." <199904300555.WAA14843@vashon.polstra.com> Date: Fri, 30 Apr 1999 16:28:55 +0800 Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG John Polstra writes: >In article <19990429212555.12573.qmail@ewok.creative.net.au>, > wrote: >> Mike Tancsa writes: >> > >> >I am currently using a couple of FreeBSD boxes for my border routers, and >> >it seems that the global routing table is getting closer and closer to 65K+ >> >routes (i.e. sooner than later I am going to bump into PR 10570). I am not >> >a systems programmer by any stretch of the imagination, so I would not know >> >how to go about debugging anything and everything that would be effected by >> >changing the size from a short to a long. I guess this PR only effects >> >very few people, so I can understand it not being a priority. But are >> >there any plans to look at it soon ? >> > >> >wat-border# netstat -nr | wc >> > 59118 355941 4200959 >> > >> > >> >> Hey, that is an interesting pr. >> >> I'd like to tackle this one if noone minds, I have a similar environment >> to debug this one in, but it might take a little time. > >Of course nobody minds! :-) > >I posted a follow-up, but it went into the black hole of cvs-all. >Briefly, I grepped the source tree and I believe that if you change >the field to an "int" and then do a full make world and kernel >build, it should work just fine. I recommend using an int rather >than a long because on the Alpha, a long is 64 bits. (int32_t would >be OK, too.) Yup. I checked the source tree for ifa_refcnt and IFAFREE, and including a couple of man page references, all the code plays with ifa_refcnt directly without assigning it to variables and vice versa. I'll submit a followup to the pr, and a patch after I've verified it doesn't panic the system, but that will be sometime early next week (I can't setup a BGP connection to flood routes in and out before then..) (I've CC:ed the original pr poster, although I'm not entirely sure if the same holds for OpenBSDs source tree..) Adrian Adrian To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message