From owner-freebsd-net@FreeBSD.ORG Thu Oct 22 09:37:33 2009 Return-Path: Delivered-To: freebsd-net@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 0AB86106568F for ; Thu, 22 Oct 2009 09:37:33 +0000 (UTC) (envelope-from brendan.kennedy@gmail.com) Received: from mail-ew0-f209.google.com (mail-ew0-f209.google.com [209.85.219.209]) by mx1.freebsd.org (Postfix) with ESMTP id 8FED08FC1A for ; Thu, 22 Oct 2009 09:37:32 +0000 (UTC) Received: by ewy5 with SMTP id 5so4308675ewy.36 for ; Thu, 22 Oct 2009 02:37:31 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:date:message-id:subject :from:to:content-type; bh=h25qBafJPlqn4sIiKFkFNkV/5cr26eSoAsOfk7cZ3sk=; b=wqlirJVB7TmcqrXODHkKV7ysgTV+l5JghxW/7W34srGCmo7BlC8ueRwbyJiLwXgJCs va73tFOiCppvRC/FqY7wEX+Uh8iidhJt/c+UJ7LUd2CieoasuTho8KUr+31wtbFx+87H VZHMEZWirrpeORM/yQh8a179albQqXwhLPDRc= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:date:message-id:subject:from:to:content-type; b=OgTQpqfcf7ZC7XoQRGsnLQ0cI3FAe2n9FFBu15Ze3rUm1qXFt+o03h36Wua+Sf0DbS AJWzl0YGJpFa2EDpNvgyqJ3l7qpK4xfz4YxNS7bXcbyzvhDzn678m4SOqiPDOTL1LQkW AFQCh09EmC4nFHO9MS05omJTutMP1iuY0bU78= MIME-Version: 1.0 Received: by 10.216.90.141 with SMTP id e13mr3055622wef.166.1256202408056; Thu, 22 Oct 2009 02:06:48 -0700 (PDT) Date: Thu, 22 Oct 2009 10:06:48 +0100 Message-ID: From: Brendan Kennedy To: freebsd-net@freebsd.org Content-Type: text/plain; charset=ISO-8859-1 Subject: FreeBSD 7.1 mbuf_cluster memory leak X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 22 Oct 2009 09:37:33 -0000 Hi all, I'm having a problem with a crypto driver and netipsec/racoon. At high data rates (1Gig of traffic, 64bit packets) we seem to be leaking 'mbuf_cluster's. These are allocated/freed by either netipsec I think. At low data rates, 'mbuf_cluster's are not used, so we don't have this problem. I just have a few questions around this issue: 1) What is the difference between mbuf and mbuf_packet? 1) Why am I getting mbuf_cluster memory being used when the packets are so small? 2) Is netipsec grouping the 64bit IP packets into a single ipsec packet? 3) Should my crypto driver be dealing with the mbuf_cluster any differently than a linked list of data (e.g. should a digest be added at the bottom of each fragment)? Regards, Brendan