From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-wi0-f181.google.com (mail-wi0-f181.google.com [209.85.212.181]) (using TLSv1 with cipher RC4-MD5 (128/128 bits)) (Client CN "smtp.gmail.com", Issuer "Google Internet Authority" (verified OK)) by huchra.bufferbloat.net (Postfix) with ESMTPS id 82F24201251 for ; Thu, 14 Jun 2012 11:37:57 -0700 (PDT) Received: by wibhn14 with SMTP id hn14so1780422wib.10 for ; Thu, 14 Jun 2012 11:37:55 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type:content-transfer-encoding; bh=ylJRCV423ppRtuvUhMxjitB6qoetkeb7Z2kNJfbDbDc=; b=gSK4hLBhSPaK+tpzsFAOSVFNeN+PR1mceDG6YSew1d20i/BaTeL27FmZRNHuqGCg5S UVyjBZU6J7je19fJ0o9D78jPCZTEvZ82VXA4lSDKR4nZwG7YxKD1aHGDbxyQ/iOWI6bY nQi/PPFtlI5fPmG/P4LONlcp69qlJkLmFfqUcBwolJdVIH/S/ldE3w0PARgSR4dUgfzt KGGkWk8eoTKFDV9f+vVdFnN8Wb6Gt3xA2zqR5Lii8SAXMowf2vx+ENDTj+763ukjgbuy 5j8FVG/HQ/UIwUuB/UVTwyEJJis5pdFp5ZguCQ0i/c0ccQo0PA95ZJuMyde/Dl1IyCbJ qyjA== MIME-Version: 1.0 Received: by 10.216.198.14 with SMTP id u14mr1605879wen.12.1339699074887; Thu, 14 Jun 2012 11:37:54 -0700 (PDT) Received: by 10.223.103.199 with HTTP; Thu, 14 Jun 2012 11:37:54 -0700 (PDT) In-Reply-To: References: Date: Thu, 14 Jun 2012 14:37:54 -0400 Message-ID: From: Dave Taht To: Maciej Soltysiak Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Cc: cerowrt-devel@lists.bufferbloat.net Subject: Re: [Cerowrt-devel] cerowrt 3.3.8-1 and related X-BeenThere: cerowrt-devel@lists.bufferbloat.net X-Mailman-Version: 2.1.13 Precedence: list List-Id: Development issues regarding the cerowrt test router project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 14 Jun 2012 18:38:01 -0000 3.3.8-3 is the first major resync between openwrt head and cerowrt in nearly a month... and needed a lot of love just to get that far. I'm very happy that it's "final" versions of the codel codel and getting-close-to-openwrt's-freeze versions of openwrt.... But: Some more polish and testing is needed, so I would be reluctant to have you just slam it into your default setup. Notably: ethtool's syntax changed on me, so fq_codel is not being enabled properly on the ethernet interfaces. The new version of quagga isn't quite baked yet, so babeld is the default s= till. There is a bug in the new firewall rules in the last stanza - it should read "guest" and have a different name than the previous stanza. I did fix the "lan" rule mentioned in the previous email. Mo bettah rules desired... Felix has made some major improvements on the ath9k front, which I haven't tested against bug #379 and related yet. I haven't addressed the long-standing upnp issue. There was a switch to netifd also. My concern here is that ahcpd will break= . gpsd is now from openwrt head, not ceropackages. dnsmasq isn't fully enabled with ipv6 dhcp support yet. jabberd is installed on the router with access limited to the ip address ranges on the secure network (see /etc/xinetd.d/xmpp-client for details) There are two default jabber users there, with default passwords. I'm not happy with this setup, but it is kind of neat to be able to leave notes to the other person's that use the router, on the router.... Transmission (bittorrent) is included but disabled by default. I have some patches for it for both ipv6 classification and uTP diffserv, too, that didn't make 3.3.8-3. And so on. By all means, please test, but my usual concerns about installing something alpha-quality on your primary gateway to the universe in the presence of wives, parents, or children, apply. On Thu, Jun 14, 2012 at 1:36 PM, Maciej Soltysiak wr= ote: > Hi Dave, > > Cool, just flashed my router with the 3.3.8-3, going to use it at home. Jeeze, I have barely had that built for an hour... > About codel, does it mean that I shouldn't enable anything in AQM and Please note that codel is the base substrate for a new class of network queuing algorithms, of which the first is fq_codel. I would recommend you run fq_codel in preference to codel in nearly all situations. I keep trying to make these distinctions clear, because many other things besides fq_codel are possible on top of codel. I hope others will clearly distinguish between "codel": the underlying theory and basic drop strategy, and stuff layered on top of it like fq_codel, qfq+codel, and stuff like the openwrt qos scripts and the currently under test simple_qos script. (which use hfsc and htb, respectively, for soft-rate limiting) > QOS tabs and just edit > /usr/sbin/simple_qos.sh to tune my down/up bandwidth? I re-enabled the QoS tab to see how good the std openwrt + fq_codel implementation now was, so as to compare it against simple_qos, so you can try either and let us know how it works. :) The AQM tab remains entirely disfunctional. > > Regards, > Maciej > > On Fri, Jun 8, 2012 at 4:54 PM, Dave Taht wrote: >> there is at least one mistake in the current firewall rules. the "lan" r= ule >> for se00, sw10, sw00 should have ACCEPT on input,output,and forward. >> >> Plenty more fixes to come by the end of the weekend. >> >> On Fri, Jun 8, 2012 at 5:19 AM, Dave Taht wrote: >>> I have updated cerowrt 3.3 to Linux version 3.3.8-1 and merged with >>> openwrt head. >>> >>> My goal for this release was simply to get caught up with openwrt, then= push >>> out the remaining useful cerowrt stuff into openwrt. After taking 3 wee= ks >>> off it's amazing how difficult it was to get back in sync. >>> >>> Changes in this release: >>> >>> + Linux 3.3.8 >>> + netifd support (more dynamic application startup/shutdown management) >>> + Bind 9.1.1-P1 (Fix for a CVE) >>> + transmission and transmission web server (uTP) added >>> + Jabber chat server added >>> >>> - I accidentally took a step back on dnsmasq (2.61 to 2.59) >>> - Firewall rules still need work (transmission support, block port 81, = too) >>> - High performance hang (#379 and friends) untested (but possibly impro= ved) >>> - UPnP work needs to be folded in >>> - quagga and gpsd didn't build >>> >>> There are exciting things happening on the quagga front with babel - >>> there is a security architecture being prototyped as I write that >>> looks promising. >>> As for the gpsd issue, there was a flurry of controversy over how best >>> to stick it in openwrt, and obviously whatever one I picked lost. >>> >>> And, of course, fq_codel is running on all interfaces. >>> >>> I will try to resolve these issues by sunday and get out a 3.3.8-2 >>> >>> Let me know if any new ones have cropped up. >>> >>> -- >>> Dave T=E4ht >>> SKYPE: davetaht >>> http://ronsravings.blogspot.com/ >> >> >> >> -- >> Dave T=E4ht >> SKYPE: davetaht >> http://ronsravings.blogspot.com/ >> _______________________________________________ >> Cerowrt-devel mailing list >> Cerowrt-devel@lists.bufferbloat.net >> https://lists.bufferbloat.net/listinfo/cerowrt-devel --=20 Dave T=E4ht SKYPE: davetaht http://ronsravings.blogspot.com/