**** BEGIN LOGGING AT Wed May 05 02:59:56 2010 May 05 03:05:32 build #67 of ppc44x is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ppc44x/builds/67 May 05 03:53:26 cshore * r21368 /packages/libs/savedynamic/ (. files/ files/savedynamic.sh): May 05 03:53:26 packages: libs/savedynamic: Add package that uses new modular firewall to save May 05 03:53:26 chains with dynamically generates rules from packages such as miniupnpd and May 05 03:53:26 strongswan in order that they might be preserved across a firewall restart (but May 05 03:53:26 not a router reboot). May 05 03:54:17 Before I modify miniupnpd, if I depend on savedynamic, will it break backfire builds since I don't think they use the new firewall? May 05 03:56:54 And if it does, how do I deal with the need for different versions for trunk vs. backfire? May 05 03:58:19 can I have a conditionaly dependency, "if not backfire depend on savedynamic"? May 05 03:58:27 *conditional May 05 04:35:47 build #52 of ep93xx is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ep93xx/builds/52 May 05 04:36:28 hi May 05 04:38:07 I'm a student in Computing and currently studying OpenWrt as part of my course project work May 05 04:39:00 I tried to get as much information as i could from openwrt wiki but i think i would get a more detail information about this project here May 05 04:39:41 I have compile some short question i would like to know about OpenWrt project May 05 04:43:10 How development decisions are made? eg. most of the time mbm making the final decision? May 05 04:43:14 Legal decisions? eg. the time when dealing with Sveasoft GPL violation issue May 05 04:43:18 Tools that are used for making decisions? May 05 04:43:23 what are the OpenWrt project aims? May 05 04:46:33 You need to explore the website (not just wiki), including the forums the Trac wiki (not the doc wiki), and review the ailing list archives May 05 04:46:42 ailing=mailing May 05 04:49:39 yep I found heaps of information about it but its was abit here and there and its not very clear. but i'll definitely check out the mailing list May 05 04:50:35 That's why it's call research...not snap and grab May 05 04:51:13 *called May 05 04:52:03 is trac part of decision making tool used by OpenWrt? May 05 04:52:36 decision making tool? you like software? afaik there isn't such a thing May 05 04:53:00 it's consensus-driven May 05 04:53:30 IRC and mailing list, and bug reports on Trac and the forums are the 'tools' May 05 04:55:03 There's not a big formal structure and such....it's basically open communication amongst developers and community collaborating May 05 04:55:55 ok, so if most people agree on something then the project would go ahead in that direction? May 05 04:56:51 or maybe is there like a forum voting on a particular path to take for the project decision? May 05 04:58:16 build #53 of octeon is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/53 May 05 04:58:56 I'm not aware of forum voting....the developer's basically have agreed to ask other developers before making any major changes, and other than that the project goes where individual developers go (with a *lot* of patches etc input from the community). May 05 05:01:50 It's not a formal structure (though there are some formal rules, for the case of handling monies and such, but I don't know about that; mostly the donations that come in are of hardware for doing testing and development, and they're handled I think by either glp, or the contact person with the donor depend on what the donor wants) May 05 05:02:01 so where does usually major discussion take place? I would assume mailing list and forums? May 05 05:02:27 Yes...but I think there is a private developer mailing list as well May 05 05:03:14 Or at least person-to-person discussion not on the public lists May 05 05:03:58 oh yes I joined the LuCI developer mailing list May 05 05:04:00 It depends on the developer and nature of the discussion May 05 05:04:12 not, I mean for devs only May 05 05:04:29 i.e. you can't sign up unless you're a developer May 05 05:05:18 do i need to be a developer to submit a patch? May 05 05:05:22 no May 05 05:05:50 build #55 of sibyte is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/sibyte/builds/55 May 05 05:06:22 developers are the ones who can commit changes May 05 05:06:33 and package maintainers can commit to packages May 05 05:09:07 i notice from the wiki that shows all the developers involved in the project and notice mbm's title is co-founder instead of founder? May 05 05:09:15 Mostly, while I've been here there hasn't been any major policy discussion, nor any major change in policy May 05 05:09:21 is he not the project leader anymore? May 05 05:09:30 there were outer founders May 05 05:09:34 other May 05 05:09:57 and not all are still involved much anymore May 05 05:10:09 mbm for example isn't very active May 05 05:10:50 and something i notice as well is that there is no project leader from the developer list May 05 05:10:53 and there is no single leader May 05 05:11:02 ohh no wonder May 05 05:13:10 so is OpenWrt getting any funding from any source? May 05 05:14:25 maybe donations by people who used and like OpenWrt or some hardware company like Linksys, Buffalo, etc provide them funding to encourage OpenWrt development? May 05 05:14:45 I don't know; you'd have to ask glp May 05 05:16:22 he's really the guy to ask questions about overall organization etc....I don't really pay that much attention to it May 05 05:17:01 but u seems to know alot about the project, are you part of the development team? May 05 05:17:29 I'm helping out in doing translation for LuCI May 05 05:17:33 I've been around a while and have recently been given full commit access, so I guess I am...not formally yet May 05 05:17:34 cool May 05 05:17:53 build #50 of cobalt is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/cobalt/builds/50 May 05 05:22:31 So how does OpenWrt started? May 05 05:23:21 as far as i know Sveasoft and DD-WRT comes later and uses OpenWrt code May 05 05:26:24 Sorry I don't know about that. Florian would know, he helped start IIRC May 05 05:27:56 What did you guys handle the Sveason case? May 05 05:29:18 when was that? I hadn't heard about it, but I've only started participating about a year and a half ago May 05 05:29:42 and not that much at first May 05 05:29:58 well Sveasoft took OpenWrt code and violate the GPL license May 05 05:30:18 when was the case? May 05 05:30:34 2006 May 05 05:30:40 it was ugly lol May 05 05:30:40 ah, long before my time May 05 05:31:09 u can just google Sveason OpenWrt May 05 05:31:19 it is quite interesting to read May 05 05:31:40 I've seen mention of it....I though it was ancient history which is why I asked May 05 05:34:30 yeah and it appears DD-WRT is going that way too May 05 05:34:48 yeah, so I've gathered May 05 05:34:52 What they doing now seems questionable May 05 05:35:12 I agree....that's why I work on OpenWRT not DD-WRT May 05 05:35:39 same here I almost choose DD-WRT as part of my study May 05 05:38:21 Any news if someone in OpenWrt is gonna any legal action on DD-WRT? May 05 05:43:50 No clue May 05 05:45:02 are there currently any aims or goal that you are aware of? May 05 06:00:07 isn't dd-wrt just a fork off of the linksys fw for wrt54g gpl release May 05 06:11:40 from my understanding DD-WRT is a fork of Sveasoft Alchemy release May 05 06:12:28 as far as i know OpenWrt > Sveasoft > DD-WRT May 05 06:13:31 "D-WRT versions up to v22 were based on the Alchemy firmware from Sveasoft, which in turn is based on the original Linksys firmware." May 05 06:13:39 -- wikipedia May 05 06:14:34 nico * r21369 /trunk/include/image.mk: image/ext2: fix rootfs partition size after [21362] May 05 06:16:05 there was a big dispute between OpenWrt and sveasoft using OpenWrt code but violate the GPL license it inherit from the code May 05 06:18:38 yeah, can't remember what it was about exactly May 05 06:19:33 it was a long time ago May 05 06:20:15 history as I recall it (probably a bit wrong): May 05 06:20:31 openwrt and sveasoft both came into existence. May 05 06:20:48 sveasoft sought to extend the stock linksys firmware with more functionality May 05 06:21:12 openwrt sought to replace it with a more functional system build from scratch May 05 06:21:13 dd-wrt project started in response to the sveasoft issue but what they doing now is almost like what sveasoft did May 05 06:22:04 sveasoft dev accuses openwrt of stealing lzma code, but there is a record of who contributed the code to openwrt May 05 06:23:24 so the sveasoft developer was claiming out of the blue that code with a clear history was stolen from them despite being around before the time sveasoft claimed they wrote it May 05 06:24:02 later on, dd-wrt comes about using the sveasoft alchemy sources May 05 06:24:23 eventually it is switched to use a lot of openwrt white russian underpinnings May 05 06:24:29 the openwrt industrial espionage unit eh? May 05 06:24:42 I hear they wear ninja outfits May 05 06:24:47 sveasoft fades into non-existence after they go completely commercial with talisman May 05 06:25:22 maligor: with time machines so that can grab code from the future. May 05 06:25:31 s/that/they/ May 05 06:26:28 as time goes on, dd-wrt pulls a fair bit of code from openwrt, as is allowed under gpl May 05 06:26:44 sveasoftt has lost practically all their possible arguments now that gpl's even been court tested May 05 06:27:16 I remember them having some totally weird arguments too May 05 06:27:17 maligor: no need to go to court anyway, sveasoft lost developers one by one until just the founder was left May 05 06:27:58 was sveasoft even an opensource project? May 05 06:29:11 it was open source but not free May 05 06:29:23 well, kinda. May 05 06:29:38 if you bought a subscription, you got access to the source May 05 06:29:56 but if you gave that source to anyone else you got b& for evar May 05 06:30:03 but if you released the source (or subscriber images to others,) your subscription was terminated and forum account baned. May 05 06:30:36 once each new release left beta, the source and images were freely available May 05 06:30:44 hmm so the developers are paid? May 05 06:31:40 but GPL requires that the source be made available by request to anyone who receives binaries, and (IIRC) no restrictions can be placed on future distribution of the source. May 05 06:32:00 I assume the sveasoft developers were paid, I never cared enough to investigate. May 05 06:32:00 Bartman007: that's right May 05 06:38:06 build #49 of etrax is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/etrax/builds/49 May 05 06:40:33 cshore: does savedynamic have a Makefile? May 05 06:42:31 yes May 05 06:42:50 or at least it should May 05 06:43:23 and it will likely be used by strongswan as well May 05 06:43:29 the makefile is just to copy the file though May 05 06:43:55 it missed the checkin May 05 06:44:06 ah, ok, just a asec May 05 06:46:21 cshore * r21370 /packages/libs/savedynamic/Makefile: packages: libs/savedynamic: Add missing Makefile. May 05 06:47:20 build #49 of gemini is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/gemini/builds/49 May 05 06:58:58 build #48 of goldfish is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/goldfish/builds/48 May 05 07:16:17 build #49 of mpc52xx is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/mpc52xx/builds/49 May 05 07:49:46 cshore: "savedynamic" is a bit generic, you should rename it to "firewall-mod-save" maybe May 05 08:08:49 xMff: ok May 05 08:09:12 use svn rename to rpeserve the history May 05 08:09:53 xMff: is depending on this going to be a problem for miniupnp (i.e. will it break backfire)? May 05 08:10:05 don't think so May 05 08:10:29 tbh I want to get the new firewall included in 10.03.1 May 05 08:10:36 but it needs more testing May 05 08:10:40 right May 05 08:10:43 see the issue yesterday May 05 08:11:20 once it is in place we can modify miniupnpd to depend on mod-save May 05 08:11:40 and to register a save section in /e/c/firewall in the postinstall May 05 08:11:46 ok...that's what I was thinking May 05 08:12:05 'register a save section?' May 05 08:12:05 you took maintainership of miniupnpd, right? May 05 08:12:21 oh yes, I see a uci-default script May 05 08:12:22 yes May 05 08:12:36 we should also rework the firewall glue May 05 08:12:47 I wrote it a while ago but its far too complex and unreliable May 05 08:12:54 imho May 05 08:14:51 now that I got my crypto accelerator running I'll look into ipsec May 05 08:14:59 hmmm...ok....that means I have to figure out how the new firewall gets called...I looked at it a bit, but it's really not that clear to me May 05 08:15:15 what you mean? May 05 08:15:49 its the same rule layout May 05 08:15:55 I mean I'm not sure what/how I need to call to add a chain/rule (e.g. when miniupnpd wants to add a forwarding) May 05 08:16:16 I mean in the code (not the config....the config is the same) May 05 08:16:19 miniupnpd does the forwarding entries itself, it talks directly to the kernel May 05 08:16:42 you only need to ensure that the MINIUPNP chains get created and that they're referenced May 05 08:17:03 ah, ok, that's what you meant by glue? May 05 08:17:07 yes May 05 08:18:34 gmorning May 05 08:26:37 xMff: while I'm at it, should firewall-mod-save be under libs, or maybe with iptables? May 05 08:26:52 the new firewall should get a menu May 05 08:27:02 and the mod packages grouped under it May 05 08:27:42 Under base? May 05 08:27:49 yes May 05 08:28:07 maybe ask {Nico} he did a lot of reorganizing lately May 05 08:28:30 Should mod-save be in base system then? May 05 08:28:37 probably May 05 08:28:46 but ask the others before May 05 08:29:05 ok, I'll just do the rename where it is for now then May 05 08:29:10 yeah May 05 08:30:31 one thing is that the script itself can't have dashes in it - that makes for bad function name errors (since the function has to begin with the name of the script) May 05 08:30:52 *name of the script May 05 08:30:58 just replace it with underscores May 05 08:31:16 ping nbd May 05 08:31:38 won't that confuse it when it looking for ${name}_${pp}_${hk}_cb ? May 05 08:31:48 no May 05 08:32:18 ah, ok, then I'll do that May 05 08:33:11 just call it mod_save ? May 05 08:33:20 build #48 of ifxmips is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ifxmips/builds/48 May 05 08:33:25 moment May 05 08:34:04 just save.sh May 05 08:35:17 the hook would be called save_post_forwarding_cb() May 05 08:35:35 or save_post_zone_cb() May 05 08:35:43 whatever is appropriate May 05 08:35:54 see grep -nr fw_callback package/firewall/files/lib/ May 05 08:36:08 save_pre_stop_cb() and save_post_core_cb May 05 08:36:18 yep, better May 05 08:36:20 just remove the dynamic May 05 08:37:19 and maybe call it 20_save.sh May 05 08:37:53 sn 20_save_pre_stop_cb? May 05 08:37:57 *so May 05 08:37:59 no the number is sstripped May 05 08:38:13 its just for the order of inclusion and processing May 05 08:38:25 this becomes important for multiple extensions May 05 08:38:33 yes May 05 08:45:20 If I do rename but not delete, will that leave the old version hanging around? May 05 08:45:27 no May 05 08:45:40 how do I undo only the last svn command? May 05 08:45:41 if you svn rename a file you'll see one A and one D May 05 08:46:00 svn revert [-R] path May 05 08:47:16 I didn't realize that...you mean you have to revert more than once if you've done more than one thing to a file? May 05 08:47:33 no May 05 08:47:41 (if you wanted it pristine) May 05 08:47:51 revert just rolls everything back to the repo state May 05 08:48:58 build #46 of adm5120 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/adm5120/builds/46 May 05 08:52:14 cshore * r21371 /packages/libs/ (6 files in 4 dirs): packages: libs/savedynamic: rename package to firewall-mod-save and script to 20_save.sh (from savedynamic.sh). Thanks jow May 05 08:59:44 cshore: the maintainer variable should be moved inside Package/firewall-mod-save/Default May 05 09:00:25 xMff: crypto changes broke about every platform May 05 09:00:30 build #46 of iop32x is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/iop32x/builds/46 May 05 09:00:36 swalker: ? May 05 09:02:28 <[florian]> xMff: I have a fix that I commit in a couple of seconds May 05 09:02:40 <[florian]> xMff: the hifnPP pci id tables are not null terminated, causing this error May 05 09:02:47 ah May 05 09:02:54 xMff: see the buildbot May 05 09:03:18 swalker: I didn't see it succeed at all in the last weeks therfore I didn't pay attention :) May 05 09:03:23 although a few are md5sum/download script busted May 05 09:03:30 xMff: yeah :( May 05 09:03:57 florian * r21372 /trunk/target/linux/rdc/patches-2.6.30/011-tune_lzma_options.patch: [rdc] fix bad lzma tuning options bad hunks May 05 09:03:57 swalker: can you paste relevant excertp if you have it open already? It takes ages to load May 05 09:04:01 florian * r21373 /trunk/target/linux/generic-2.6/files/crypto/ocf/hifn/hifnHIPP.c: [kernel] fix build error on hifnHIPP spotted by buildbot May 05 09:04:10 I don't really think ignoring build errors in general helps the reliablity May 05 09:04:22 yes I know May 05 09:04:34 part of the problem is that loading the logs is so painful May 05 09:04:52 a tail -n 1000 or so would already help imo May 05 09:04:57 <[florian]> definitively May 05 09:04:59 xMff: I had already closed the tab so I'm in the same boat May 05 09:05:44 breakdown of logs by package would be nice; and a hint as to which package broke May 05 09:05:51 thepeople changed some log settings awhile ago that should've helped May 05 09:06:28 [florian]: I think generic-2.4 needs the same fix May 05 09:08:02 <[florian]> xMff: you are right May 05 09:08:03 florian * r21374 /trunk/target/linux/generic-2.4/files/crypto/ocf/hifn/hifnHIPP.c: [kernel] backport r21373 to 2.4 kernel May 05 09:08:15 its the same file actually May 05 09:08:19 files May 05 09:08:32 maybe we should symlink 2.4/files/crypto to 2.6/files/crypto May 05 09:08:51 cshore * r21375 /packages/libs/firewall-mod-save/Makefile: packages: libs/firewall-mod-save: Moved maintainer variable where it belongs. May 05 09:08:59 erm nvm May 05 09:09:12 they were created from a 2.4 specific patch May 05 09:09:24 ignore what I said, I need more coffewe May 05 09:10:29 xMff: is that a kind of caffeinated sheep May 05 09:10:45 lol May 05 09:24:46 build #45 of ixp4xx is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ixp4xx/builds/45 May 05 09:59:28 acoul: ping May 05 10:31:01 build #45 of kirkwood is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/kirkwood/builds/45 May 05 10:34:04 [florian]: ping May 05 10:58:42 build #45 of orion is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/orion/builds/45 May 05 11:04:24 does anybody here has a brcm4710 device around? May 05 11:05:17 build #44 of rb532 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/rb532/builds/44 May 05 11:14:59 build #43 of pxcab is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/pxcab/builds/43 May 05 11:22:46 nico * r21376 /branches/backfire/ (2 files in 2 dirs): [backfire] merge r21261 & r21262 May 05 11:35:44 <[florian]> rtz2: pong May 05 12:01:18 nico * r21377 /branches/backfire/package/kernel/modules/ (netsupport.mk other.mk): [backfire] adjust kmod-crc16 dependencies and dependencies on kmod-crc16 (closes: #7063) May 05 12:08:11 build #43 of rdc is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/rdc/builds/43 May 05 12:56:21 build #42 of s3c24xx is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/s3c24xx/builds/42 May 05 12:57:55 quick q: when a client connects to AP (rb411, madwifi both), the arp table at AP doesn't get the client's mac. Is that normal? May 05 13:46:39 quick q: when a client connects to AP (rb411, madwifi both), the arp table at AP doesn't get the client's mac. Is that normal? May 05 13:56:12 build #42 of ps3 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/42 May 05 14:23:21 xrg: depends on the config. is the ap interface bridged to something else and receiving dhcp from a different device? May 05 14:23:45 no, they are routed, only and static ips. May 05 14:24:07 so, there would not be a reason for any IP packet from client to AP, unless the client wants something. May 05 14:24:26 Anyway, I did put in /etc/hotplug.d/iface/50-ping a ping command. May 05 14:25:20 but, if this is a general phenomenon, we should standardize a "first packet" May 05 14:34:12 xrg: no idea about this May 05 14:34:38 so, think it's normal for a new client not to register itself at the AP's arp table? May 05 14:34:50 (if the client never sends a packet, that is) May 05 14:35:44 build #42 of ar7 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ar7/builds/42 May 05 14:42:07 xrg: if the client never sends a packet, then of course nothing will show up in the arp table May 05 14:42:11 i mean how could it May 05 14:42:39 nbd: yes, that's the case.. May 05 14:43:54 nbd: just any packet (a ping one) activates the connection then (for comm. back from the AP to the client). May 05 15:05:16 ah right, if I recall correctly someone else reported that a while ago May 05 15:05:30 client needs to ping first before the other direction starts working May 05 15:05:59 might have been something else though May 05 15:07:19 jow_laptop: the only tricky part is that a client with static IP will not necessarily know the AP's ip.. May 05 15:10:23 xrg: shouldn't any ping do as long as the system routes it via the wifi link? May 05 15:10:28 e.g. 1.1.1.1 May 05 15:10:46 you can even force the outgoing iface with -I May 05 15:11:12 yes, true.. A broadcast ping on the wifi network could also do. May 05 15:11:52 so, the point is: should we add a few lines at the scripts? May 05 15:13:41 why not ... if it is actually confirmed May 05 15:14:17 nbd: wdyt? May 05 15:35:39 xrg: why do you need the ip addresses to be in the arp cache? May 05 15:36:26 If I don't, packets from AP to client cannot be sent! I get a 'destination unreachable' message. May 05 15:36:43 the ap should do an arp lookup May 05 15:36:52 if that doesn't work, then something is broken May 05 15:37:12 Then, it's broken.. ( I repeat, it's madwifi on rb411, trunk) May 05 15:37:23 any details on your setup? May 05 15:37:27 i mean configuration May 05 15:37:45 ah, it's hostapd-mini and wpa_supplicant, too. May 05 15:38:09 if you want to save space, you could just install wpad-mini or wpad May 05 15:38:14 so, two static interfaces, one AP and one client. associated. May 05 15:38:31 can you show me your wireless and network config? May 05 15:38:38 ok, I'll keep that in mind. May 05 15:39:40 May I paste here, or priv? May 05 15:40:18 use pastebin please May 05 15:40:47 ok. May 05 15:44:22 nbd: http://www.pastebin.org/203344 May 05 15:44:26 build #40 of ubicom32 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/40 May 05 15:44:29 it is trivial May 05 15:46:25 there's an error there May 05 15:46:38 you must not put interface names such as ath0 into /etc/config/network May 05 15:48:52 should they be wlan or sth.? May 05 15:49:05 no May 05 15:49:09 leave out option ifname entirely May 05 15:49:24 you assign interface configs to wlan interfaces by referencing them in /etc/config/wireless May 05 15:50:23 jow * r21378 /trunk/package/firewall/files/lib/fw.sh: [package] firewall: properly unset position for delete command, fixes rule removal in ifdown May 05 15:50:26 but still, this interface gets configured.. Could this only affect the arp cache? May 05 15:51:19 Anyway, I can change that tomorrow. May 05 15:51:21 this can cause the driver to act in weird ways May 05 15:51:29 ok. May 05 15:51:31 madwifi easily gets confused by such stuff May 05 15:51:42 because adding the ath* ifnames to /etc/config/network creates race conditions May 05 15:51:53 because the interface is brought up before it is fully configured May 05 15:52:20 hm.. In that case, let me test (I have to physically go there, tomorrow) and report back .. May 05 15:52:24 thanks a lot! May 05 15:52:25 ok May 05 15:52:27 no problem May 05 15:52:50 i guess i should adjust the hotplug script to ignore hotplug events for ath8 May 05 15:52:53 ath* May 05 15:54:13 some more examples in the wiki could also do. May 05 15:56:03 would be nice if you could add some May 05 15:57:00 nbd * r21379 /trunk/package/base-files/files/etc/hotplug.d/net/10-net: base-files: do not bring up wlan interfaces from hotplug events, so that wrong network/wireless configs are more noticeable May 05 15:57:51 xrg: the is a namespace, doc/recipes May 05 15:57:55 *there May 05 15:58:23 nbd * r21380 /branches/backfire/package/base-files/files/etc/hotplug.d/net/10-net: [backfire] backport network hotplug change from r21379 May 05 15:58:51 xrg: with the above change, your config will fail completely instead of dangling in a half-working state May 05 15:58:59 that should make wrong configs more noticeable May 05 16:03:15 :) thanks. May 05 16:08:41 build #39 of uml is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/uml/builds/39 May 05 16:39:27 hi! May 05 16:39:50 [florian], rtz2, have you received anything...? May 05 16:40:04 <[florian]> alesan: not so far, do you have a tracking number? May 05 16:40:28 no, I just used regular mail. it may take one more week, don't hold your breath :) May 05 16:40:52 I was just wondering because the other day I received a package from sweden, sent with regular mail, in three days May 05 16:43:48 alesan: nope May 05 16:54:06 acoul: ping May 05 16:54:42 hey rtz2 how are you doing? May 05 16:55:26 I have on my spool list to test your ssb code May 05 16:55:52 probably sometime tomorrow May 05 16:56:34 acoul: hold one for some time, there are a bunch of problems with it May 05 16:57:02 oh May 05 16:57:07 acoul: could you create a dump of the register values for the first core? May 05 16:57:19 sure May 05 17:15:38 build #41 of xburst is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/xburst/builds/41 May 05 18:03:33 cshore, thanks for firewall-mod-save! May 05 18:16:42 florian * r21381 /trunk/package/kernel/modules/crypto.mk: [package] kmod-ocf-hifn7751 depends on PCI, spotted by buildbot May 05 19:18:03 hi May 05 19:18:09 ping nbd May 05 19:18:42 is someone of the devs here? May 05 19:18:52 im only a package maintainer :) May 05 19:19:37 . May 05 19:19:38 here May 05 19:38:20 xMff why is noone updating openssl? May 05 19:38:40 i submitted a openssl patch to the ml a long time ago May 05 19:38:56 tripolar: its on my todo May 05 19:39:07 i can also submit a new one for the new ocf patch May 05 19:39:20 I postponed it because we were in the middle of release preparations back then May 05 19:39:27 okay May 05 19:39:37 should i submit the "new" patch? May 05 19:39:43 and after that I simply forgot it, but while doing the ocf update I remembered it May 05 19:40:01 if you want, go ahead, this would speed things up May 05 19:42:53 xMff: by the way, did you see the Add base support for multiple configs over one network interface patch? May 05 19:43:20 rtz2: yes May 05 19:43:40 xMff: any plans for it? May 05 19:44:04 xMff: because I can understand, why this guy is getting grumpy May 05 19:44:12 xMff send it to the ml May 05 19:45:24 brb May 05 19:45:43 rtz2: it looks like a step in the right direction but it is not robust enough May 05 19:46:02 rtz2: if we're going to support that we need a recursive or chained approach May 05 19:46:18 where you can stack more than two levels of interface configs May 05 19:46:50 I also need to investigate how the firewall reacts on it May 05 19:47:06 also I suspect that multiple ppp instances on the same base iface will break May 05 19:55:11 xMff: I would answere him with that May 05 19:55:21 xMff: or do you want to? May 05 19:55:31 i have here an arm board with an imx35 whats the fastet way to get openwrt running on it? May 05 19:55:43 rtz2: feel free to, you may quote me May 05 19:55:52 rtz2: I'll look into it $soon May 05 19:55:58 ok May 05 19:56:10 its on my mid term todo May 05 19:56:15 tripolar: is the cpu already supported in openwrt? May 05 19:56:39 tripolar: and are there linux sources for it around? May 05 19:56:44 yeah May 05 19:56:52 it a freescale cpu May 05 19:56:56 ARM11 May 05 19:57:04 http://www.freescale.com/webapp/sps/site/overview.jsp?code=IMX35_FAMILY May 05 19:57:12 but i dont know if openwrt supports it by now May 05 19:57:42 im using openembedded by now but i think openwrt would be the faster way to devlop for this board May 05 20:00:28 rtz2: actually I think the multi config thing could actually work, even for ppp* as long as you have multiple deidcated config interface sections sharing physical devices May 05 20:01:11 it would also fit into the current uci schema and be supported ootb by both x-wrt and luci May 05 20:01:50 actually the only missing bit is some kind of dependency specification May 05 20:02:06 to define some bringup order for the protocols May 05 20:03:07 one of the most prominent use cases is pptp over something May 05 20:03:14 e.g. pptp over pppoe May 05 20:04:01 with the current approach both ppp instances would be launched simultanously which could lead to race conditions (esp. when both compete for the defaultroute or dns entries) May 05 20:04:50 also certain combinations are undefined, e.g. what happens with static + dhcp, static + static and so on May 05 20:05:25 most protocol backends assume that they're the sole owner of a physical device, so they would end up overwriting each others settings May 05 20:06:42 and handling multiple ips on one interface is a pite with ifconfig May 05 20:07:00 iproute2 would work better but it adds quite some code to the base system May 05 20:07:14 while ifconfig and route come for free May 05 20:07:24 almost that is May 05 20:10:06 so in order to implement this properly, almost all protocol backends need to be changed, the uci firewall needs adaption to work with real aliases (not only rely on -i and -o but use -s and -d to specify subnets on top of a interface carrying multiple protocols) and possibly others I overlooked yet May 05 20:11:19 it boils down to a larger rewrite operation :) May 05 20:11:30 EOF May 05 20:12:39 tripolar: did you reapply the ocf patch? May 05 20:14:29 ? May 05 20:14:43 tripolar: I wonder about the two missing hunks May 05 20:14:48 no this lines where allready in the openssl source May 05 20:14:55 ok May 05 20:15:06 the outbut is a little different but they do the same May 05 20:15:13 output May 05 20:32:15 jow * r21382 /trunk/package/openssl/ (3 files in 2 dirs): [package] openssl: update to v0.9.8n, thanks Peter Wagner May 05 20:34:43 xMff: i will copy the irc log into the mail May 05 20:34:55 xMff: thanks for the explaination May 05 23:24:13 jow * r21383 /trunk/package/ (3 files in 3 dirs): [package] base-files, ppp: remove protocol specific cleanup code from /sbin/ifdown and move it to protocol stop callbacks May 06 00:10:58 nico * r21384 /packages/libs/elfutils/ (5 files in 2 dirs): [packages] elfutils: update to 0.147, use PKG_INSTALL, add MD5SUM, fix URL, cleanup (closes: #7290) May 06 00:16:02 nico * r21385 /packages/utils/lvm2/ (Makefile patches/100-readline-link.patch): [packages] lvm2: update to 2.02.64 (closes: #7291) May 06 00:20:57 ping cshore May 06 00:21:59 pong xMff May 06 00:22:18 when you find time, can you check this pptp patch: http://paste.tksite.gotdns.org/d6c71911d ? May 06 00:22:30 ok May 06 01:14:27 jow * r21386 /trunk/package/ (base-files/files/lib/network/config.sh ppp/files/ppp.sh): [package] base-files, ppp: fix interface shutdown **** ENDING LOGGING AT Thu May 06 02:59:56 2010