From owner-freebsd-net@FreeBSD.ORG Tue Jul 15 17:51:17 2003 Return-Path: Delivered-To: freebsd-net@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id C0B3137B401; Tue, 15 Jul 2003 17:51:17 -0700 (PDT) Received: from mail.sandvine.com (sandvine.com [199.243.201.138]) by mx1.FreeBSD.org (Postfix) with ESMTP id DCEF843FA3; Tue, 15 Jul 2003 17:51:16 -0700 (PDT) (envelope-from sloach@sandvine.com) Received: by mail.sandvine.com with Internet Mail Service (5.5.2653.19) id <305LFX67>; Tue, 15 Jul 2003 20:51:13 -0400 Message-ID: From: Scot Loach To: 'Don Lewis' , silby@silby.com Date: Tue, 15 Jul 2003 20:51:12 -0400 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" cc: freebsd-net@FreeBSD.org Subject: RE: Kernel tuning for large maxsockets X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 00:51:18 -0000 True, I can add a tunable for each of tcp, udp, raw, divert. What will happen when the system runs out of pcbs? -----Original Message----- From: Don Lewis [mailto:truckman@FreeBSD.org] Sent: Tuesday, July 15, 2003 8:48 PM To: silby@silby.com Cc: sloach@sandvine.com; freebsd-net@FreeBSD.org Subject: Re: Kernel tuning for large maxsockets On 15 Jul, Mike Silbersack wrote: > > On Tue, 15 Jul 2003, Scot Loach wrote: > >> Is there any reason I should not modify the kernel code to only let a small, >> fixed number of raw and divert pcbs be preallocated instead of having them >> scale with maxsockets? > > Your idea is sound. > >> Next, does this seem like a generally useful thing that could be rolled back >> into the source tree? I could make this a kernel option or a tunable sysctl >> variable. >> >> thanks >> >> Scot Loach > > A tunable maximum for each of those settings sounds good, that should fit > well in subr_param.c. Send me your patch when it's done, and I'll look > into incorporating it. I'd prefer separate tunables. I suspect that it is also common to have vastly different requirements for the numbers of TCP and UDP sockets.