**** BEGIN LOGGING AT Wed Aug 12 02:59:57 2009 Aug 12 04:00:18 ping Aug 12 05:02:50 hm.. cross-compiling perl sucks. **** ENDING LOGGING AT Wed Aug 12 05:26:18 2009 **** BEGIN LOGGING AT Wed Aug 12 05:28:14 2009 **** ENDING LOGGING AT Wed Aug 12 05:32:06 2009 **** BEGIN LOGGING AT Wed Aug 12 05:34:16 2009 Aug 12 07:15:11 good morning Aug 12 08:41:50 florian * r17234 /trunk/target/linux/brcm63xx/files/arch/mips/pci/ (ops-bcm63xx.c pci-bcm63xx.c): [brcm63xx] add missing patch which fakes a cardbus controller on top of PCI, thanks SGDA Aug 12 09:32:27 ping Aug 12 09:34:26 pong Aug 12 10:00:36 Hi. I have write some module for LuCI. How I can create separate module for installing via make menuconfig? Aug 12 11:22:44 dkostousov: make them standalone packages, not hard Aug 12 11:24:29 or add them to contrib/package/luci/Makefile Aug 12 11:31:18 xMff: where I can get a docs about creation packages? Aug 12 11:33:45 dkostousov: depends on whether you want ot make them separate (out of tree) or add them to the others Aug 12 11:47:15 xMff: out of tree now Aug 12 11:51:04 dkostousov: hold on I'll prepare an example Aug 12 12:11:34 xMff: are you there? Aug 12 12:49:05 . Aug 12 12:49:14 dkostousov: got my messages? Aug 12 13:19:37 jow * r17235 /trunk/package/opkg/patches/001-fix-double-parsing.patch: [package] opkg: crlf -> lf in 001-fix-double-parsing.patch Aug 12 14:09:29 markus * r17236 /packages/net/openvpn/files/openvpn.init: remove PID-file on TERM. Aug 12 14:11:34 Is there a way to make a kernel module depend on a kernel greater than a certain release (ie 2.6.28)? Aug 12 14:11:41 I'm (slowly) working in gspca support Aug 12 14:15:17 kupesoft: exclude the not supported kernel versions with @!LINUX_2_6_XX Aug 12 14:15:34 2.6.1 through to 2.6.27 ? Aug 12 14:15:45 package/kernel/*/*.mk also contain examples how to test for >2.6.xy Aug 12 14:16:21 xMff: I only see blacklisting Aug 12 14:16:36 @!X Aug 12 14:16:43 I want >=X Aug 12 14:17:39 ifeq ($(strip $(call CompareKernelPatchVer,$(KERNEL_PATCHVER),ge,2.6.24)),1) Aug 12 14:17:56 see package/kernel/modules/crypto.mk Aug 12 14:18:01 Basically blacklist targets in target/linux/generic-2.6/ -- 2.6.21, 2.6.25, 2.6.27 ? Aug 12 14:18:10 xMff: :) Aug 12 14:32:30 nbd: ping Aug 12 14:32:30 Hmm, I still can't find a clean way to require a greater than kernel version strictly in the DEPENDS section Aug 12 14:32:59 you can't Aug 12 14:33:26 you can, however, wrap the DEPENDS:= line in an ifeq Aug 12 14:41:56 xMff: which message you mean? Aug 12 14:42:39 What's the DEPENDS target for disabled, @BROKEN ? Aug 12 14:47:38 Anyone know a why to source files on a Config.in type file, but for all packages in feeds? AIUI source "feeds/*/*/*/another-config.in" fails for a base only compile Aug 12 14:49:50 kupesoft: yes Aug 12 14:50:28 cshore: can you rephrase your question please? I don't understand it after reading it three times :) Aug 12 14:53:37 xMff: I want to include new menuconfig items, to come from packages, and I want them under an existing menu. My solution so far is to use Aug 12 14:53:38 source "feeds/*/*/*/another-config.in" Aug 12 14:53:38 in the file that defines the top-level of the menu, but I think that fails when feeds are not installed and the compile only of the base system Aug 12 14:54:18 yes, likely if you rely on a menu that's defined by some feed package Aug 12 14:55:41 well if the CONFIG_ item is null I can ignore it, that's not a problem, but when I tried to do it before having package with another-config.in the compile failed on the source statement complaining there was no such file Aug 12 14:56:08 is there a way to conditionally include a source statement? Aug 12 14:57:04 wrap it in an "if" ? Aug 12 14:57:37 I couldn't figure out the correct syntax of the statement Aug 12 14:58:16 I tried ifeq/endif and that didn't work Aug 12 14:58:27 or is if in a config file different? Aug 12 14:59:34 so you want to do it within a Makefile? Aug 12 14:59:44 a Config.in type file Aug 12 14:59:59 not possible as far as I know Aug 12 15:00:22 you need to check this a level higher Aug 12 15:00:43 xMff: how do you mean? Aug 12 15:02:18 This gspca patch is going to be a doozy Aug 12 15:02:43 cshore: I *think* that the kconfig langauge does not allow such advanced constructs, so you have to arrange stuff in the makefile that includes the Config.in Aug 12 15:03:48 like for example "cat feeds/*/*/*/another-config.in" > config-tmp.in and then source that Aug 12 15:04:54 ok, thanks Aug 12 15:05:20 ok, btw, what do you think of the preinit patch I posted for comments Aug 12 15:06:49 Should I replace /etc/init.d/firewall from the openwrt with doc/firewall.iptables from the chillispot source tree? Aug 12 15:09:52 dkostousov: I think not, better make a firewall include and put the chillispot specific rules there Aug 12 15:10:33 cshore: I think it's a good idea, we need more hooks like that, just haven't time to actually test the code Aug 12 15:52:02 ifeq ($(strip $(call CompareKernelPatchVer,$(KERNEL_PATCHVER),lt,2.6.31)),1) Aug 12 15:52:02 GSPCA_SUPPORT_TEST:=@BROKEN Aug 12 15:52:02 endif Aug 12 15:52:05 That doesn't seem to work Aug 12 15:52:12 This is giving me a headache Aug 12 15:52:20 s/2.6.31/2.6.28/ Aug 12 15:52:23 ^^ still doesn't work Aug 12 15:52:38 you need to include kernel.mk as well Aug 12 15:52:48 if you haven't already Aug 12 15:53:17 I have Aug 12 15:53:24 I'm editing package/kernel/modules/video.mk Aug 12 15:55:09 I'll just blacklist all possible kernels for now Aug 12 15:57:01 There's just 21, 25, and 27 Aug 12 15:58:14 these conditions do not work for DEPENDS Aug 12 15:59:05 trunk supports only 2.4.xx, 2.6.21, 2.6.25, 2.6.27, 2.6.28 and 2.6.30 Aug 12 16:49:08 wew, 25 gspca kernel module packages auto-parsed, behold, http://openwrt.pastebin.com/m52c65775 Aug 12 17:42:16 architecture question, the gspca driver comes with 27 different modules on 2.6.30 (just under in 2.6.28)-- should I create a kernel package for each? Aug 12 17:44:08 counter question: is a single module useful by itself? then yes Aug 12 17:45:19 I'm not sure how to pack multiple kernel modules in to one Aug 12 17:45:26 There's no option to build them all together Aug 12 17:46:18 And 28 modules loaded in to memory, each on average 30K... Aug 12 17:46:33 Maybe just a submenu on the menuconfig is best Aug 12 17:47:16 brb Aug 12 17:50:18 xMff: Any comment? Aug 12 18:01:17 anyone? Aug 12 18:02:12 kupesoft: make one package / module Aug 12 18:02:27 would really suck to load 27 drivers just to use one of them Aug 12 18:02:36 That's what I'm doing Aug 12 18:02:47 I'm auto-generating what to add to video.mk, so it's not such a pain :P Aug 12 18:02:58 good Aug 12 18:03:16 is there some kind of "base module" or are they all independant? Aug 12 18:03:19 yes Aug 12 18:03:27 the gspca_main Aug 12 18:03:42 ok, so make them all depend on that and you're done Aug 12 18:03:56 But judging by the same of gspca_main.ko compared with others like gspca_x.ko, they're isn't that much common code Aug 12 18:03:57 and maybe create an gspca.mk Aug 12 18:03:59 xMff: I'm one step ahead Aug 12 18:04:12 and not append the stuff to video.mk Aug 12 18:04:48 Sure Aug 12 18:04:52 Make it a submenu of video Aug 12 18:05:12 yep Aug 12 18:05:43 do you need more comments? :P Aug 12 18:06:07 :( Aug 12 18:06:27 Comments? Here's what I got so far... http://priv.kupesoft.com/openwrt/gspca-sofar.patch Aug 12 18:07:17 surprisingly little for a 550+ line diff... Aug 12 18:07:42 you can factor out all the common definitions and replace them with call statements Aug 12 18:08:02 I could, but there's really little Aug 12 18:09:50 actually, that's not a bad idea Aug 12 18:09:58 I do :) Aug 12 18:13:17 http://openwrt.pastebin.com/m76e51947 Aug 12 18:13:49 couldn't find some uppercase/lowercase macros for make, otherwise you could eliminate the second param too I think Aug 12 18:14:28 and KernelPackage/gspca-templateshould be something else but you get the idea Aug 12 18:16:01 I see it Aug 12 18:16:15 Now not all the options are showing up in my submenu, only the first 5 are Aug 12 18:16:16 strange Aug 12 18:17:14 syntax error near the 5th definition? Aug 12 18:17:30 They're all autogenerated though Aug 12 18:17:35 <_trine> does anyone know if theres a problem with the code that controls the switch in the ar71 stuff Aug 12 18:18:57 <_trine> I compiled and flashed a new WRT160NL because the switch appeared to have failed in my TEW632 but the WRT160NL is the same Aug 12 18:19:27 <_trine> so there may have been a change to the code that affects this Aug 12 18:22:36 xMff: It's something to do with menuconfig's cache... the submenus work on a fresh checkout Aug 12 18:44:52 _trine network isn't working, right? Aug 12 18:47:14 <_trine> Dogge_: is that a question or a statement Aug 12 18:47:27 question Aug 12 18:47:30 ? Aug 12 18:48:21 <_trine> on my routers I get an IP assigned on the wan with dhcp but I can never get a connection to the web Aug 12 18:48:54 <_trine> I have reflashed the TEW632 and shall test it in the morning Aug 12 18:49:03 <_trine> with the old code Aug 12 18:50:29 well, here on a nslu2 with trunk i can only ping the ip on the nslu2 itself. ping on ip addresses on my local lan don't work :( Aug 12 18:50:48 <_trine> same here Aug 12 18:51:16 so, someone broke networking across targets :( Aug 12 18:51:26 o_O Aug 12 18:55:00 nslu2 = ixp4xx Aug 12 18:55:17 _trine yours is ar71xx? Aug 12 18:55:24 <_trine> yes Aug 12 18:55:34 not good Aug 12 18:56:05 <_trine> it will get fixed in the goodness of time I'm not bothered Aug 12 18:56:27 same here Aug 12 19:02:48 i'd guess https://dev.openwrt.org/changeset?new=17160@trunk%2Fpackage%2Fbase-files&old=17155@trunk%2Fpackage%2Fbase-files some of these changes broke networking Aug 12 19:04:13 most likely you're guessing wrong Aug 12 19:05:01 then what broke it? Aug 12 19:06:47 xMff ? Aug 12 19:09:15 Dogge_: no idea, I don't have that hw, have never worked with that platform and follow trunk only from time to time, but what I know for sure is that the changes you've shown should _not_ affect the network configuration Aug 12 19:10:23 it's across platforms. mine is atm ixp4xx, _trines is ar71xx... Aug 12 19:11:07 <_trine> sorry I mentioned it now I didn't want to cause a fuss Aug 12 19:11:46 _trine: you didn't cause a fuss, thank you for the notification. Aug 12 19:13:09 <_trine> I am really grateful for having kamikaze Aug 12 19:15:39 would be interesting to know what ifconfig, iptables -nvL, iptables -t nat -nvL, cat /proc/net/arp and dmesg report Aug 12 19:16:04 also whether global ip forwarding is on Aug 12 19:16:36 and what was the last revision known to work Aug 12 19:18:29 <_trine> xMff: I've had an energetic day today so I'm a bit tired but in the morning I'll have a look at what I have Aug 12 19:19:29 from what you've telled it could be anything ranging from bad firewall rules to a borked ethernet driver, therefore more info is required Aug 12 19:19:36 uhm told Aug 12 19:19:46 should take some english lessons Aug 12 19:20:29 <_trine> 'from what you've mentioned ' Aug 12 19:20:40 yeah, something like that :P Aug 12 19:20:42 <_trine> would be better english :) Aug 12 19:21:11 <_trine> but I don't speak anything else so who am I so say Aug 12 19:21:21 <_trine> opps to* Aug 12 19:22:18 Dogge_: what kind of ping reply do you get? "No route to host" or even just timeouts? Aug 12 19:23:19 <_trine> I get network unreachable Aug 12 19:23:29 and what's in /proc/net/arp after your ping attempt? Just 00:00:00:00:00:00 ? Aug 12 19:23:29 xMff i get ping replys from the nslu2 nic. no other ip in my network (cables checked) Aug 12 19:24:19 <_trine> bbl Aug 12 19:24:54 - /proc/net/arp: 192.168.178.1 0x1 0x2 00:1c:4a:64:f8:ca * br-lan Aug 12 19:25:26 178.1 is another host on your network? Aug 12 19:25:47 yeah Aug 12 19:25:53 so arp seems to work Aug 12 19:26:18 have you tried to disable the firewall? Aug 12 19:26:35 this is a image without firewall and iptables Aug 12 19:26:40 okay Aug 12 19:26:58 route -n and stuff is okay? Aug 12 19:27:35 http://openwrt.pastebin.com/d4132dc98 Aug 12 19:28:08 what happens if you disable the bridge? Aug 12 19:29:44 just uci del network.lan.bridge ? Aug 12 19:29:55 "type" not bridge Aug 12 19:29:58 but yes Aug 12 19:31:25 i did: uci del network.lan.type; uci commit network; ifup lan Aug 12 19:31:27 still not working Aug 12 19:31:49 trine_ wrote that dhcp works, so probably not a complete network fsckup but a routing issue for trine_ Aug 12 19:32:46 I only know that kind of behavior from broadcom vaps, they do broadcasts fine but unicast does not work until you put the vap interface into a bridge Aug 12 19:33:43 wtf? :) Aug 12 19:33:47 this one sounds suspicious too: https://dev.openwrt.org/ticket/5687 Aug 12 19:34:10 but the dhcp assignment is unicast, only the discovery is multicast if I'm not mistaken Aug 12 19:34:52 however, the observed behavior was: dhcp is received, ip is configured, pings do not work Aug 12 19:35:09 added a bridge on top: dhcp works, ping works too Aug 12 19:35:37 I have the very same problem on my thinkpad btw, dhcp works, pings do not work, after unloading and loading e1000e it works again Aug 12 19:35:38 hm. so incoming unicast works but outgoing doesn't in that case Aug 12 19:35:58 this happens often after a suspend Aug 12 19:36:03 Dogge_: can you ping the box from the outside while running tcpdump? then you can see if the packets arrive Aug 12 19:36:15 (running tcpdump on the nslu) Aug 12 19:36:41 have to rebuild and reflash an image with tcpdump Aug 12 19:37:00 e the firewall? Aug 12 19:37:09 oops, sorry Aug 12 19:38:04 or maybe this one: http://article.gmane.org/gmane.comp.embedded.openwrt.devel/3838 Aug 12 19:39:34 oh no Aug 12 19:40:20 very likely Aug 12 19:43:47 doesn't help Aug 12 19:50:54 http://openwrt.pastebin.com/d6b231cde Aug 12 19:52:52 Dogge_: what's your kernel version? 2.6.30 ? Aug 12 19:55:43 is "select all packages by default" not working in trunk or is there a special trick? i can select it, but it doesn't select/build additional packages Aug 12 20:06:05 soma: it must be selected during the initial menuconfig run. it merely changes the default setting for any "new" packages. anything that are already set in .config will not be altered. Aug 12 20:08:20 ok, thanks. so i'll try after cleaning Aug 12 20:08:26 xMff: http://luci.subsignal.org/trac/ticket/78/ -- I didn't manage to create a plain testcase. maybe you can point nbd to this Aug 12 20:08:43 soma: rm .config (or mv .config old.config) is enough. Aug 12 20:09:02 Bartman007: how do you prefer the (small) patch to IB, per mail oder pastebin? Aug 12 20:09:34 mail is probably easier. agb openwrt.org Aug 12 20:09:39 soma: what is the function of the patch? Aug 12 20:10:14 it only generates the package lists when something has changed. it takes quite long to rebuild it every time with ~Â1300 packages Aug 12 20:11:28 but before i send it to you i'm going to test it with trunk again. i currently made it for 8.09.1, but since nothing has changed i hope it will also work with trunk. if it does i send it to you Aug 12 20:16:11 i also realized imagebuilder doesn't work when there is no profile. is it easier to add at least one profile for every target or to try to change the IB? i've seen ar71xx already has profiles in trunk while there are none in 8.09.1, so maybe this is already addressed in trunk? Aug 12 20:20:15 xMff 2.6.28.10 Aug 12 20:21:14 I will be addressing it shortly. I also came to the conclusion that creating profiles was probably a better route. Aug 12 20:21:50 ack Aug 12 20:22:53 Dogge_: hm, not exactly bleeding edge Aug 12 20:23:32 i also made a cgi-frontend for the IB, although its currently targeted at mesh-networks, but maybe you want to have a look here too: Aug 12 20:23:34 I mean it's in use since some time, even stranger that ethernet suddenly stops working Aug 12 20:23:36 http://www.wgaugsburg.de/cgi-bin/newkit/wizard.cgi Aug 12 20:28:45 soma: very neat. I know we had a couple people working on a web imagebuilder in the past, not sure what the current state of it is. Aug 12 20:29:49 yes, i heard that too but couldn't find out who, so i tried one myself Aug 12 20:36:31 what language did you write it in? Aug 12 20:36:52 shellscript Aug 12 20:37:35 haha, you'll fit in around here :) Aug 12 20:39:32 if you are interested in the code, i've setup a svn Aug 12 20:39:49 and yes, one day i'm going to learn a real language ;) Aug 12 20:41:20 has anyone fixed igmpproxy spamming the syslog? Aug 12 20:41:30 is that what the -Dflag is? Aug 12 20:41:51 yeah, tried with -D0 ? Aug 12 20:42:12 i'll try that Aug 12 20:42:23 cuz it seems to me it spams to the point where it crashes it Aug 12 20:44:27 and what kind of messages does it spam? Aug 12 20:46:24 xMff: it just fills the log with a bunch of multicast logs Aug 12 20:46:42 whcih on an igmp multicast network come about every 10 seconds Aug 12 20:46:58 Bartman007 freewrt has the openwrt wib running, iirc Aug 12 20:47:01 so basically if you look at the log thtough luci, all you'll see is igmp Aug 12 20:47:20 i'm just trying to figure out where the problem is Aug 12 20:47:54 i might try to write a patch for igmpproxy if its not an openwrt issue, but right now i'm updating to the latest version (it got added to trunk a couple of weeks ago) Aug 12 20:48:02 soma: http://nbd.name/gitweb.cgi?p=openwrt-wib.git;a=summary Aug 12 20:48:13 haven't tried it, just stumbled on it recently Aug 12 20:51:20 russo: according to INSTALL in the igmpproxy tarball, the option is -q or --silent Aug 12 20:51:33 w/o arguments Aug 12 20:51:36 moonflux: thanks Aug 12 20:54:59 xMff: hmm... i think there was an issue with that in the version i had though Aug 12 20:55:05 i'm 90% sure i tried that Aug 12 20:55:30 not a big deal :P the real issue is i lost all my VMs after my raid5 died :P Aug 12 20:55:40 and my openwrt was on there :/ Aug 12 20:56:48 is there somewhere a installation of openwrt-wib to see it in action? Aug 12 21:01:09 soma: I think the one Dogge_ mentioned is at http://wib.freewrt.org/ -- but I don't know if that is based on the openwrt-wib Aug 12 21:24:49 xMff: if you're curious, theres no -q option atleast not in the version in openwrt Aug 12 21:25:04 and it ouputs much less to messages, so thats fine then Aug 12 21:30:36 theres no update-rc.d in openwrt, right? Aug 12 21:31:27 nvm Aug 12 21:31:34 dumb question i remmeber now Aug 12 21:31:51 jeez you forget so much when you stop hacking at openwrt for 2 months :P Aug 12 21:33:40 hehe Aug 12 21:36:00 anyway it seems to me that the new version doesn't output as much debug as the previous version did Aug 12 21:36:11 even when not running in debu Aug 12 21:36:12 g Aug 12 21:36:20 alright i'm off now, thanks xMff Aug 12 21:36:28 np. cu Aug 12 21:52:52 xMff: Figured out what Makefile for tr is? I'm still working on 28 kernel module the behemoth that is gspca Aug 12 21:53:25 s/28 kernel module the/the 28 kernel module/ Aug 12 21:54:01 you mean for upper and lower case? Aug 12 21:55:55 or vice versa Aug 12 21:57:36 nope, the best I came up with is $(shell echo "$(1)" | tr 'a-z' 'A-Z') Aug 12 21:58:44 http://gmsl.sourceforge.net/ :) Aug 12 21:59:45 neat Aug 12 21:59:57 reminds on the httpd written in postscript Aug 12 22:00:12 hehe, yeah Aug 12 22:02:06 moonflux: I've debated adding gmsl multiple times, but have ultimately decided against it. Aug 12 22:02:25 I have started importing a couple functions from it though :) Aug 12 22:03:37 I couldn't believe my eyes the first time I found it.... decided against using it in that project as well Aug 12 22:04:51 the math functions are especially amusing. Aug 12 22:05:48 it converts numbers into strings of "x " so 5 == "x x x x x " Aug 12 22:06:42 ypu start to pull similar stunts if all you have is ash and sed :P Aug 12 22:06:51 okay not that bad Aug 12 22:59:23 xMff: I found the problem with lan clients not getting ip addressed, just for your information Aug 12 23:00:15 spi-gpio new module has bug, I use the deprecated (old) worked fine Aug 12 23:01:23 mankash: great you found a solution! can you open a ticket and describe your findings? Aug 12 23:01:52 I don't know how to open a ticket, give me a url I will post it Aug 12 23:02:01 thanks a lot fo ryour greaqt help Aug 12 23:02:42 but I have wifi prob sometimes, driver doesn't detect wifi card sometimes when I rebbot router Aug 12 23:02:44 https://dev.openwrt.org/newticket fill this out and hit create ticket Aug 12 23:02:51 ok thx Aug 12 23:03:27 try to include relevant information, especially where you got the working version from and what happened when you tried to use the newer one Aug 12 23:03:54 and attach the dmesg output Aug 12 23:05:06 enclosed in Aug 12 23:05:06 {{{ Aug 12 23:05:06 ... Aug 12 23:05:06 }}} Aug 12 23:05:20 ok Aug 13 00:48:54 jow * r17237 /packages/net/openvpn/files/openvpn.init: [packages] openvpn: load config only once in the init script up() and down() sections **** ENDING LOGGING AT Thu Aug 13 01:24:10 2009 **** BEGIN LOGGING AT Thu Aug 13 01:36:15 2009 **** ENDING LOGGING AT Thu Aug 13 01:36:17 2009 **** BEGIN LOGGING AT Thu Aug 13 01:38:14 2009 Aug 13 01:40:37 Anybody having issues with ntpd compiling? It seems to be pulling in which is totally different to what it expects. Aug 13 01:41:17 it's likely that two packages collide when providing md5.h Aug 13 01:42:39 Hm.m. wonder which ones. Aug 13 01:42:56 I don't really need ntpd at the moment. was just looking through the things that didn't compile. Aug 13 01:44:39 another interesting one>> wprobe-util.c:52: warning: format '%lld' expects type 'long long int', but argument 4 has type 'int64_t' Aug 13 01:44:56 (with warnings treated as errors, this stops the compile) Aug 13 01:45:37 frogonwheels: https://dev.openwrt.org/ticket/5595 Aug 13 01:47:50 huh Aug 13 01:47:52 thanks **** ENDING LOGGING AT Thu Aug 13 02:59:57 2009