**** BEGIN LOGGING AT Sat Mar 08 10:59:56 2008 Mar 08 13:15:34 matteo * r10569 /packages/net/wshaper/ (. Makefile files/ files/wshaper.htb): added the wshaper script Mar 08 17:00:36 matteo * r10570 /packages/net/wshaper/files/wshaper.htb: added module (un)load to wshaper Mar 08 19:10:43 matteo * r10571 /trunk/package/ar7-atm/patches/100-compile_fix.patch: [AR7]: Fix two of the three bugs in #3122 Mar 08 21:54:05 iptables-mod-extra is flawed in HEAD and in r10465 Mar 08 21:54:38 it claims to toggle libipt_physdev, but this is not correct Mar 08 21:55:00 I checked netfilter.mk, but it seems to be fine to that point Mar 08 21:55:48 I am not sure how to verify where the build process breaks down, I am working on r10465 and 2.6 kernel for broadcom Mar 08 21:57:30 by breaks down, I don't have a build failure per say, but at the same time, physdev is not built and included into the resulting kmod-ipt-extra...ipk Mar 09 00:28:20 I opened a ticket https://dev.openwrt.org/ticket/3218 **** BEGIN LOGGING AT Sun Mar 09 04:41:09 2008 Mar 09 10:07:57 olli * r10572 /packages/admin/gkrellmd/ (Makefile files/gkrellmd.init patches/100-conf.patch): [Packages] admin/gkrellmd: Update to 2.3.1 (http://lists.openwrt.org/pipermail/openwrt-devel/2008-March/001753.html) Mar 09 10:55:23 florian * r10573 /trunk/package/rt2x00/ (27 files in 2 dirs): Fix rt2x00 compilation and upgrade to the current mainline version (2.6.24) Mar 09 11:08:40 florian * r10574 /trunk/target/linux/rdc/image/Makefile: Fix bzImage copy with 2.6.24 kernels Mar 09 11:21:20 florian * r10575 /trunk/ (4 files in 3 dirs): Upgrade x86 to 2.6.24 Mar 09 17:58:46 matteo * r10576 /trunk/package/ar7-atm/patches/160-module-params.patch: ar7-atm: add an useful patch from #2561, tnx Axel Gembe Mar 09 20:29:36 ping ejka Mar 09 21:22:25 nbd * r10577 /trunk/tools/Makefile: only build dtc for the target that uses it Mar 09 21:22:32 nbd * r10578 /trunk/toolchain/uClibc/Makefile: fix uclibc build on osx Mar 09 21:42:13 nbd * r10579 /trunk/tools/Makefile: build dtc for all powerpc targets (suggested by Kaloz) Mar 09 22:40:29 ryd * r10580 /packages/libs/libvorbisidec/Makefile: Adding libtool fixup to libvorbisidec Mar 09 22:44:27 ryd * r10581 /packages/sound/mpd/Makefile: Mar 09 22:44:27 Fix alsa to mpd Mar 09 22:44:27 Patch by Mirko Vogt Mar 09 22:44:27 fix [3025] Mar 09 22:49:21 argl fix in #3025 Mar 09 22:50:00 hehe. Mar 10 02:07:21 *grmbl* Mar 10 02:07:45 why does ifupdown, debian and openwrt still use ifconfig *sigh* Mar 10 02:08:39 familiarity? Mar 10 02:09:03 i assume you favor "ip link set" Mar 10 02:09:15 i do Mar 10 02:09:28 especially since i need 2 ips because of dnscache/tinydns Mar 10 02:09:47 i 'just' wanted a router for the flat *sighs* Mar 10 02:10:09 br-lan:0 looks ugly! ;) Mar 10 02:10:42 well, you're welcome to install the ip stack Mar 10 02:10:48 i did. Mar 10 02:11:17 still it does not help me getting rid of the ugly ifconfig aslong as uci/ifupdown use it Mar 10 02:12:08 uninstall those Mar 10 02:12:42 well... i didnt want to reinvent the wheel completely by myself Mar 10 02:13:37 then just edit the initscripts Mar 10 02:13:49 btw.. i did Mar 10 02:14:34 currently i use uci and a 'lan1' which generates br-lan:0 to set the second ip and call it from /etc/init.d/network as last thing in boot() Mar 10 02:19:39 roh: rewrite the scripts, submit a patch, we'll strongly consider it. Mar 10 02:19:52 Bartman007 sounds good Mar 10 02:20:38 i mean.. i wouldn do it only to have a patch for myself and see it not being mergeable upstream.. but i think since busybox also supports the ip command that could be done, right? Mar 10 02:21:03 busybox's ip is buggy/gimped/minimalist. Mar 10 02:21:31 i know. but for the minimal stuff uci does it should surfice.. what do you think? Mar 10 02:21:59 .oO(busybox being buggy? no way!!1) Mar 10 02:22:02 ;) Mar 10 02:22:46 possibly, it is something that would have to be discussed in more depth. I don't consider it too extreme include tc/ip in the default images (I don't have a single install without them) Mar 10 02:24:04 well.. i dunno much about the policies in openwrt. whats the minimum needed flash, thus the maximum allowed base-set? Mar 10 02:24:31 i only use 'big' systems it seems... nothing <8mb flash and <16mb ram Mar 10 02:24:51 .oO(and 2.6.x kernels) Mar 10 02:24:54 4MB flash is the minimum, which is pretty tight if people want anything that uses libopenssl, libpcap, etc. Mar 10 02:25:25 my WRTs have 4MB flash, but the rest of my boards all have 8-16MB flash. Mar 10 02:26:27 4MB flash broadcom users are going to be in for quite a suprise when wireless becomes fully functional on 2.6. Broadcom's wl and nas are siginificantly smaller than OSS equivalents. Mar 10 02:26:54 what i am still kinda missing is a alternative to isc-dhcpd which doesnt suck or has too limited featureset (not that isc software wouldnt suck) Mar 10 02:27:32 isc-dhcpd is offered but personally I find all my needs (except for one fringe case) are met by dnsmasq. Mar 10 02:28:30 that thing stiÃll scares me somehow Mar 10 02:29:09 the fringe case is that I have a few users that want the same IP when connecting via wired or wireless but most of them don't remember to kill wireless when connected via a wired connection anyway so things get messy anyway w/ what they want. Mar 10 02:29:20 does too much and too less.. i sometimes need the ability to fix ips to macs, and do kinda strage option-setups Mar 10 02:29:38 mostly when hacking other embed. devices Mar 10 02:29:43 roh: mac affinity is possible w/ dnsmasq Mar 10 02:30:44 cat one set arbitrary dhcp-options? also overwriting globals like dns and allow bootp requests? Mar 10 02:32:10 dnsmasq's dhcp is pretty versatile, but sometimes you really need the isc dhcpd Mar 10 02:32:50 yes you can set arbitrary dhcp options, even create classes of devices for custom settings. Mar 10 02:35:55 sounds good.. i should take a second look at dnsmasq Mar 10 02:44:43 hm.. can one use it only for dhcp, and disable the dns component? Mar 10 02:45:58 why not just dump tinydns? or do you use it for an authoritative server? Mar 10 02:46:04 both. Mar 10 02:46:25 i mean.. authoritative for the local network and a instance of dnscache Mar 10 02:48:27 dnsmasq can cache just fine, but it won't manage a domain Mar 10 02:48:53 well.. can i disable that without patching? Mar 10 02:50:40 i don't remember offhand, but you'll probably end up looking at the source either way Mar 10 03:10:03 hm.. one can disable dns in dnsmasq by setting the port to 0 Mar 10 03:41:07 or just make it give a different dns server address to it's dhcp clients ... Mar 10 07:32:36 any ar7 people around? Mar 10 08:18:53 well? Mar 10 08:29:31 *gnnhh* Mar 10 08:30:20 i-hate-these-webif-stuff *sigh* .. just had to find out how to force this 'thingie' to be play a dsl modem Mar 10 08:30:44 finally terminating the pppoe on a selfcompiled kernel Mar 10 08:31:10 heh Mar 10 08:31:23 btw.. are there platforms where openwrt supports a adsl22Ã+ modem? Mar 10 08:31:36 ar7 Mar 10 08:31:43 oh Mar 10 08:31:57 how fast are these? Mar 10 08:32:03 adsl2+ Mar 10 08:32:48 i mean the cpu.. i need some headroom for pppoe and openvpn ;) Mar 10 08:33:22 but for what its worth.. uci rocks compared to all these webifs Mar 10 08:33:23 roh: the Infineon Danube platform will come in time :-) Mar 10 08:33:48 Bartman007 i see... is there a list of hw which uses these already? Mar 10 08:34:22 there isn't any hardware on the market as far as I know. We support the platform, just waiting for the boards to come to market :-) Mar 10 08:34:23 http://wiki.openwrt.org/CategoryAR7Device Mar 10 08:34:25 have a second box here which has infineon for all support chips, sadly the soc has a glued-on cooler Mar 10 08:35:11 eeeks.. ADAM2 is still in use *scream* Mar 10 08:35:12 Infineon has been working with several of our devs, it has been a real joy compared to the normal efforts required for supporting a new platform from what I've heard. Mar 10 08:35:55 nice.. ive listened to the chaosradio-express podcast with nbd earlier.. he mentioned that Mar 10 08:36:28 Bartman007: have you done much with cpmac? Mar 10 08:36:59 btw.. will there be a SOC2008 for openwrt? Mar 10 08:37:08 not that i want to apply, but would be cool Mar 10 08:37:38 sn9: nope Mar 10 08:37:39 i'm sure google would jump at the chance Mar 10 08:38:38 roh: it has been mentioned, but not much more than that. Mar 10 08:38:50 Bartman007: hmm, i was wondering why it's working on the netgear but on the linksys Mar 10 08:39:15 Bartman007 i see.. hopefully somebody from the coredevs takes care about it. i think the application-deadline is near Mar 10 08:40:04 near? i would have thought there was maybe a month or two left Mar 10 08:40:45 "We are now accepting applications from open source projects who would like to act as mentoring organizations for 2008, and will be accepting these applications through March 12, 2008. We will begin accepting student applications on Monday, March 24th." Mar 10 08:40:56 means... *yikes* Mar 10 08:40:57 wow Mar 10 08:41:08 two days Mar 10 08:43:01 bah, they don't accept documentation projects Mar 10 08:43:23 .oO(i have an project, iproute2 support for uci, and documentation for uci) Mar 10 08:43:40 Bartman007: that's dumb Mar 10 08:43:55 sn9: yeah, I could have easily spent an entire summer working on that. Mar 10 08:44:58 do you have enough ideas for projects or should i try do dump what i am missing? Mar 10 08:45:09 in the land of the blind, the one-eyed man is king Mar 10 08:45:28 in the land of coders, the documenter is king Mar 10 08:46:34 roh: as far as I know, only one proposal so far. Mar 10 08:46:53 Bartman007 whats it about? Mar 10 08:48:38 getting a working driver for the bcm53xx crypto core (IPsec accel) that can be included in the mainline kernel Mar 10 08:49:01 53xx? Mar 10 08:49:15 sounds cool.. nice Mar 10 08:49:19 i know 43xx and 63xx Mar 10 08:50:01 have a 6358 here.. should have that unit i think Mar 10 08:50:27 many devices supported by brcm-2.4 and brcm47xx have the crypto core. Mar 10 08:50:29 roh: 6358 is 63xx -- adsl2+ with usb Mar 10 08:51:18 the openwrt port to the bcm63xx is in a sad state Mar 10 08:52:33 yeah, it won't have much of a future unless some decent network drivers surface. Mar 10 08:53:11 one has to check gpl tarballs from 63xx vendors regularly Mar 10 08:53:27 sn9 uh... Mar 10 08:53:48 i've looked a little, and i think there is something in some of them Mar 10 08:53:53 sn9 would be cool to have that box supported.. because of the legerity stuff Mar 10 08:54:07 roh: which box? Mar 10 08:54:32 tcom sells one named 'speedport W 502V' Mar 10 08:54:45 is that a 63xx? Mar 10 08:55:10 the 'type C' is a 6358 with 4318 wifi Mar 10 08:55:51 it features 2 pots fxs via the legerity stuff and one which can be fxo or isdn te Mar 10 08:56:07 isdn is done with some infineon chip Mar 10 08:56:27 the table of hardware lists the W500V as a 6358 with 4318 Mar 10 08:56:41 but the W501V is ar7 Mar 10 08:57:01 its all oem stuff. the 6358 version is from arcadyan afaik Mar 10 08:57:10 thats the 'type C' Mar 10 08:57:23 type A is from hitachi supposedly and i dunno the soc. Mar 10 08:57:52 the rest in there is infineon. but the case looks the same, just a bit bigger. even the webif is styled exactly the same Mar 10 08:58:16 there are ar7 devices from linksys with the same legerity chip for fxs Mar 10 08:58:51 gotnone talked to zarlink, which bought legerity Mar 10 08:59:09 they will rerelease the drivers under gpl this week Mar 10 08:59:21 is there some 'embedded appliance disassembling/documentation' in some wiki where one can add stuff? Mar 10 09:00:08 you mean hardware take-apart, or machine code? Mar 10 09:01:31 documentation about the hardware, means where to get, how to identify, known revisions/differences, hi-res photos... Mar 10 09:02:18 oh, the openwrt wiki is a good source of that. so is the seattlewireless wiki Mar 10 09:02:18 one could make a list of used components and then make wikipages for all of these with links to known facts/drivers/docs etc Mar 10 09:02:42 sounds like the openwrt wiki to me Mar 10 09:02:44 also listing oem-versions and similarities Mar 10 09:03:14 you've seen the table of hardware, right? Mar 10 09:03:21 well.. the openwrt wiki is, with respect in a even worse state than what our wiki at openmoko is.. and thats already bad Mar 10 09:03:31 jap. and thats quite 'basic' Mar 10 09:04:55 it's still a central source of information Mar 10 09:06:41 are there any plans to move the moin-moin to something else? Mar 10 09:06:56 dunno Mar 10 09:07:09 i don't see why Mar 10 09:07:27 well.. its a pain in the ass to use for one and it looks bad for another Mar 10 09:07:51 also with something like trac wiki one could hack a nice macro helping documenting stuff nicer Mar 10 09:08:16 most wikis are harder to use than moin, and the look is changeable Mar 10 09:08:41 a) tags b) macro which is used to show the status of chip-support in the page of the box for all chips Mar 10 09:09:01 sn9 trac wiki is the same basic syntax afaik. Mar 10 09:09:08 the dev wiki already has that Mar 10 09:09:16 means you could even import it easy Mar 10 09:09:33 i know. thats why i wonder the regular wiki is still moinmoin Mar 10 09:11:09 because the openwrt documentation effort is severely shorthanded Mar 10 09:13:47 roh: yes, we will be moving to mediawiki Mar 10 09:14:11 oh, no! Mar 10 09:14:47 that can only multiply the problems Mar 10 09:16:20 mediawiki is to wikis as redhat was to linux distros a decade ago Mar 10 09:17:12 before ubuntu Mar 10 09:17:59 does it have a git backend? :) Mar 10 09:18:04 sn9: personally I like and use moinmoin, but I'm not the openwrt webmaster and our wiki has been plagued with issues that have been difficult to track down Mar 10 09:18:20 loswillios: patches are welcomed ;-) Mar 10 09:18:32 heh Mar 10 09:22:16 Bartman007 urghs.. thats bad Mar 10 09:22:39 we have mediawiki at OM and its a pure pain Mar 10 09:23:01 i saw some very strange issues with the table of hardware Mar 10 09:23:06 .oO(and its php) Mar 10 09:26:30 voice these concerns to Kaloz he's the man in charge **** ENDING LOGGING AT Mon Mar 10 09:59:57 2008