From owner-freebsd-net Mon Jul 16 8:52: 0 2001 Delivered-To: freebsd-net@freebsd.org Received: from ringworld.nanolink.com (ringworld.nanolink.com [195.24.48.39]) by hub.freebsd.org (Postfix) with SMTP id D497A37B409 for ; Mon, 16 Jul 2001 08:51:07 -0700 (PDT) (envelope-from roam@orbitel.bg) Received: (qmail 58036 invoked by uid 1000); 16 Jul 2001 15:55:16 -0000 Date: Mon, 16 Jul 2001 18:55:16 +0300 From: Peter Pentchev To: Joseph Gleason Cc: net@FreeBSD.ORG Subject: Re: ICMP source quench - deprecated? Message-ID: <20010716185516.F56285@ringworld.oblivion.bg> References: <20010716152638.B52566@ringworld.oblivion.bg> <00d201c10e0e$4720fdc0$0b2d2d0a@battleship> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <00d201c10e0e$4720fdc0$0b2d2d0a@battleship>; from clash@tasam.com on Mon, Jul 16, 2001 at 11:45:02AM -0400 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org I know; those were my reasons for lifting an eyebrow at the 'deprecated' comment in the Linux kernel source.. G'luck, Peter -- "yields falsehood, when appended to its quotation." yields falsehood, when appended to its quotation. On Mon, Jul 16, 2001 at 11:45:02AM -0400, Joseph Gleason wrote: > Unless I am mistaken, ICMP source quench applies to any IP traffic, not just > TCP. > > Also, ICMP source quench can be sent by intermediate routers. > > > > ----- Original Message ----- > From: "Peter Pentchev" > To: > Sent: Monday, July 16, 2001 08:26 > Subject: ICMP source quench - deprecated? > > > > Hi, > > > > I just had a friend ask me a weird question about ICMP source quench > > and its handling in various OS's. He proceeded to show me a part > > of some version of the Linux kernel source, which processed a source > > quench request properly, yet had a 'This is deprecated' comment > > at the top. > > > > I had a quick look at src/sys/netinet/tcp_subr.c, and on FreeBSD 4.x > > and -current, an ICMP source quench seems to provoke a proper downsizing > > of the connection window. > > > > Is there any reason for ICMP source quench to be deprecated? > > If so, what is it to be replaced with? Or are the TCP window scaling > > algorithms smart enough to downsize the window quick enough in case of > > a noticeable packet loss? (I guess they should be..) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message