From owner-freebsd-hackers Wed Nov 20 9:41:58 2002 Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 0079A37B401 for ; Wed, 20 Nov 2002 09:41:57 -0800 (PST) Received: from mail.sandvine.com (sandvine.com [199.243.201.138]) by mx1.FreeBSD.org (Postfix) with ESMTP id 3CDA043E6E for ; Wed, 20 Nov 2002 09:41:56 -0800 (PST) (envelope-from don@sandvine.com) Received: by mail.sandvine.com with Internet Mail Service (5.5.2653.19) id <42S9WNS0>; Wed, 20 Nov 2002 12:41:54 -0500 Message-ID: From: Don Bowman To: 'Terry Lambert' , hiten@unixdaemons.com Cc: tony@ubik.demon.co.uk, freebsd-hackers@freebsd.org Subject: RE: if_ti DEVICE_POLLING patch (Was: Re: [hackers] Re: Netgraphc ouldbe a router also) Date: Wed, 20 Nov 2002 12:41:53 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG > From: Terry Lambert [mailto:tlambert2@mindspring.com] > Hiten Pandya wrote: > > Hehe. I was not being sarcastic. I just acknoledged your > statement. The > > reason I asked about committing the patch in the first > place, is because I > > thought that the patch had been around on the -net@ list, > and had been reviewed. > > > > My apologies. > > It's been around on the list for a while, but if it's working for > the people involved, they are not saying anything. > > Last I heard, the person who had the problem had tried the Intel > Gigabit card, but not the Tigon III, but was looking for a Gigabit > card recommendation. > > I suppose I could make the same patch for the Intel card. > > The problem is that DEVICE_POLLING is an all-or-nothing thing, > you can't control it on a card-by-card basis, even with a > sysctl, because it fills out an entry point that's statically, > rather than procedurally initialized. Is there any point to using device polling with the tigon 3 (broadcom 570x etc)? It has a pretty good interrupt reducer in it by itself. Just tune the 2 rx and the 2 tx parameters and you get a constant interrupt rate with good latency for any packet rate. I changed (if_bge.c): /* RX Interrupt no more than every ~500 us */ sc->bge_rx_coal_ticks = 512; /* TX Interrupt no more than every ~500 us */ sc->bge_tx_coal_ticks = 512; /* RX Interrupt no more than every ~120 packets */ sc->bge_rx_max_coal_bds = 128; /* TX Interrupt no more than every ~120 packets */ sc->bge_tx_max_coal_bds = 128; and this gives no more than 2000 interrupts/s, and no more than 500us of latency. The if_em has similar functionality AFAIK. --don To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message