**** BEGIN LOGGING AT Sat Mar 26 02:59:58 2011 Mar 26 11:28:13 build #77 of mpc52xx is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/mpc52xx/builds/77 Mar 26 12:21:30 build #76 of octeon is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/76 Mar 26 12:23:29 hi, i would like to try to fix https://dev.openwrt.org/ticket/8823 Mar 26 12:24:00 firstly, can I obtain 10.03.1 rc4 by checking out the 10.03 branch? Mar 26 12:24:53 build #81 of iop32x is complete: Failure [failed shell compile_6] Build details are at http://tksite.gotdns.org:8010/builders/iop32x/builds/81 Mar 26 12:25:13 secondly, how (and where) does wireless get initialized? Mar 26 12:29:37 theres an init script in /etc/init.d Mar 26 12:29:49 also the config options are in /etc/config/ Mar 26 12:30:21 sKAApGIF: there is nothing to fix. AP and Ad-Hoc do not work together, if it did in the past it was pure accident Mar 26 12:37:52 ive manually set up an adhoc an ap mode vap's on a wr1043nd with ath9k Mar 26 13:18:12 build #109 of s3c24xx is complete: Failure [failed compile_10] Build details are at http://tksite.gotdns.org:8010/builders/s3c24xx/builds/109 Mar 26 13:48:47 build #95 of ps3 is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/95 Mar 26 14:10:12 nbd * r26302 /trunk/package/mac80211/patches/030-backport_93c86_eeprom.patch: mac80211: remove an obsolete backport patch Mar 26 14:10:15 nbd * r26303 /trunk/package/mac80211/patches/030-disable_tty_set_termios.patch: mac80211: disable an unnecessary backport to reduce code size Mar 26 15:26:00 nbd * r26304 /trunk/rules.mk: strip .gnu.attributes and .reginfo from kernel modules Mar 26 15:33:39 nbd * r26305 /trunk/ (5 files in 3 dirs): Mar 26 15:33:39 Revert "kernel: remove all local symbols from kernel modules at link time to reduce their size" Mar 26 15:33:39 Use -x when stripping kernel modules instead Mar 26 16:19:09 what are the chances of a precompiled binary running on openwrt? Mar 26 16:32:44 0-5% Mar 26 16:33:32 it targets the correct platform Mar 26 16:33:34 im about to try it shortly Mar 26 16:33:39 but im thinking = much fail Mar 26 16:33:39 also Mar 26 16:34:11 problem is more the uclibc abi Mar 26 16:34:26 unless your binary is statically compiled Mar 26 16:34:31 or linked rather Mar 26 16:34:50 yeah altho the binary when trying to link Mar 26 16:34:57 wont it search the various local paths first Mar 26 16:35:06 rather than try and dynamically link the system library? Mar 26 16:35:20 eg.. maybe could hax up that lib seperately just for that binary? Mar 26 16:35:54 altho now haha i cant even get openwrt to compile Mar 26 16:36:07 Library not found: libiptc.so.0 in path: /lib:/usr/lib:/usr/lib/ebtables Mar 26 16:36:27 why on earth is it looking there Mar 26 16:36:40 libtool? Mar 26 16:37:55 someone should fix the dependencies for libc and libgcc Mar 26 16:38:11 whilst you *can* remove them, doing so will cause a kernel panic at boot Mar 26 16:38:33 /home/jesus/opt/openWrt/trunk/staging_dir/host/bin/mklibs.py -D -d /home/jesus/opt/openWrt/trunk/tmp/mklibs-out --sysroot /home/jesus/opt/openWrt/trunk/staging_dir/target-mips_uClibc-0.9.32/root-brcm63xx -L /lib -L /usr/lib -L /usr/lib/ebtables --ldlib /lib/ld-uClibc.so.0 --target mips-openwrt-linux-uclibc `cat /home/jesus/opt/openWrt/trunk/tmp/mklibs-progs` 2>&1 Mar 26 16:38:45 likewise, removing swconfig makes you unable to connect to the router Mar 26 16:38:51 trying to build ebtables Mar 26 16:39:56 anickthatsnottak: disable mklibs, its experimental Mar 26 16:40:31 i think it has something to do with the stripping Mar 26 16:40:56 Olipro: yeah, so does preconfiguring 0.0.0.0 as lan ip. There are endless ways to shoot yourself in the knee :) If you don't know what a particular (preselected!) package is for, don't disable it Mar 26 16:41:29 compiling wrt is so painful Mar 26 16:42:03 anickthatsnottak: again, disable mklibs, its experimental and known to cause troublre Mar 26 16:42:14 anickthatsnottak: stick to the defaults if you don't know you way around Mar 26 16:42:14 yer.. Mar 26 16:42:21 then it won't be painful Mar 26 16:42:22 i ment more compile time Mar 26 16:42:43 xMff: also, if you feel like it, revert cshore's last change to the cups makefile preese Mar 26 16:42:50 it broke the cups web interface Mar 26 16:42:53 and that makes me a sad panda Mar 26 16:43:10 its compiled successfully now Mar 26 16:43:14 it doesn't make cups unusable, but it does make it harder to use Mar 26 16:43:14 also is webif broken? Mar 26 16:43:36 because without the web interface, you must do everything manually with the config files Mar 26 16:43:38 cause i had to manually edit it out of the packages, otherwise i get some stupid info.sh trying to fire up in the browser Mar 26 16:43:52 ls Mar 26 16:43:53 Olipro: why? He jsut moved it to another menu Mar 26 16:44:01 nooo Mar 26 16:44:10 he remove the line from the makefile that builds it Mar 26 16:44:17 ? Mar 26 16:44:18 so now it's completely missing Mar 26 16:44:26 check the Makefile for cups Mar 26 16:44:32 compare cshore's latest revision to the previous Mar 26 16:44:33 https://dev.openwrt.org/changeset/26106 Mar 26 16:44:41 cshore's last change Mar 26 16:44:46 completely innocent Mar 26 16:45:08 humm Mar 26 16:45:23 also /trunk/files/* get included in image yes? Mar 26 16:45:30 anickthatsnottak: yep Mar 26 16:45:40 anickthatsnottak: with webif you mean x-wrt? Mar 26 16:46:00 i assume so, the package is called webif Mar 26 16:46:39 ive only played with openwrt for about 4hours now lol so no idea Mar 26 16:46:50 try luci Mar 26 16:46:59 yer luci was there and working Mar 26 16:47:04 ok Mar 26 16:47:14 then i enabled one of the packages for webif Mar 26 16:47:19 and of course it selected the entire branch Mar 26 16:47:31 then no matter how many packages i disabled i couldnt remove it Mar 26 16:47:41 but with it enabled, breaks the web interface entirely Mar 26 16:47:52 didnt wanna make distclean and loose all my settings Mar 26 16:47:55 heh Mar 26 16:47:59 so hacked it out of the .config Mar 26 16:48:03 yeah Mar 26 16:48:08 that is okay Mar 26 16:48:17 I mean editing .config Mar 26 16:48:29 but you might miss stuff that was selected by webif, like haserl Mar 26 16:48:34 omg now luci is broken Mar 26 16:48:36 wtf Mar 26 16:48:45 in what regard? Mar 26 16:50:29 http://iforce.co.nz/i/ycl0hek4.y03.png Mar 26 16:51:11 which is what webif was doing Mar 26 16:51:19 maybe its the http package(s) i have selected Mar 26 16:51:31 did you select another webserver? Mar 26 16:51:37 build #96 of pxcab is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/pxcab/builds/96 Mar 26 16:51:38 lol i wodner how many writes the flash on this device will survive Mar 26 16:51:39 not uhttpd ? Mar 26 16:51:48 many, don't bother. Mar 26 16:52:09 hold up let me get back in menu config Mar 26 16:52:12 i think i have uhttpd Mar 26 16:52:16 and lighthttp Mar 26 16:52:47 and which one is serving port 80 now? Mar 26 16:53:09 not sure, guess i can fire up tcpdump and look at headers Mar 26 16:53:26 or you could jsut telnet/ssh and look? ;) Mar 26 16:53:36 like netstat -nltp | grep 80 Mar 26 16:53:36 fresh flash Mar 26 16:53:38 both disabled Mar 26 16:53:45 huh? Mar 26 16:53:52 telnet is always enabled if no password is set Mar 26 16:54:10 tru, didnt realise that Mar 26 16:54:36 1297 root 3100 S lighttpd -f /etc/lighttpd/lighttpd.conf Mar 26 16:54:37 thats running Mar 26 16:54:42 i dont see uhttpd Mar 26 16:54:46 so it seems it is not configured for cgi Mar 26 16:55:02 yeah uhttpd didn't start because port 80 is taken by lighttpd Mar 26 16:55:32 yeah hmm, guess i can fix that Mar 26 16:56:07 so either make lighttpd cgi-aware or put it or uhttpd on another port Mar 26 16:56:22 removed lighthttpd and recompiling Mar 26 16:56:43 it must have been selected by a random package Mar 26 16:57:35 build #110 of atheros is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/atheros/builds/110 Mar 26 17:00:23 nbd * r26306 /trunk/package/madwifi/files/lib/wifi/ (madwifi.sh madwifi_countrycodes.txt): madwifi: add support for translating country codes into internal country code numbers (based on a patch by Manuel Munz) Mar 26 17:03:05 root@OpenWrt:/bin# adslctl Mar 26 17:03:05 Usage: adslctl start [--up] [--mod ] [--lpair <(i)nner|(o)uter> Mar 26 17:03:07 the binary works Mar 26 17:03:23 now does the kernel module *cringe* Mar 26 17:04:25 highly doubt it but anus is puckered Mar 26 17:34:11 hmm ok Mar 26 17:34:15 so how do i downgrade the kernel Mar 26 17:34:24 and wtf is up with no insmod -f Mar 26 17:34:55 sure, extremely bad practice - but good to test haha Mar 26 17:35:29 don't even try to use kmods from the repo. build your own Mar 26 17:36:00 this brcm63xxx there are no ADSL drivers Mar 26 17:36:06 no gpl source Mar 26 17:36:14 so thats not a option Mar 26 17:36:21 the binaries work Mar 26 17:36:27 and it SEEMS like these modules might just load Mar 26 17:36:31 with the correct kernel version Mar 26 17:36:33 yeah Mar 26 17:36:41 adsldd: version magic '2.6.21.7 mod_unload MIPS32_R1 32BIT ' should be '2.6.35.11 mod_unload MIPS32_R1 32BIT ' Mar 26 17:36:41 adsldd: version magic '2.6.21.7 mod_unload MIPS32_R1 32BIT ' should be '2.6.35.11 mod_unload MIPS32_R1 32BIT ' Mar 26 17:36:53 the binaries will not work Mar 26 17:36:55 right Mar 26 17:37:00 you should give up on that Mar 26 17:37:00 so i need to try drop back to 2.6.21.7 Mar 26 17:37:04 no Mar 26 17:37:09 you need the original kernel Mar 26 17:37:20 not a vanilla kernel with the same version Mar 26 17:37:20 ideally that too. Mar 26 17:37:30 broadcom modified kernel internals Mar 26 17:37:38 different data structures etc. Mar 26 17:37:55 loading the binary drivers into a vanilla kernel will not work Mar 26 19:30:26 hi all. what's the status of 2.6.38 and LZMA? Mar 26 19:31:32 philipp64|laptop: you mean for squashfs? Mar 26 19:34:17 yup. I can't boot a 2.6.38 image. Mar 26 19:34:38 you need to generate a squashfs image with xz ;p Mar 26 19:35:00 I'm currently looking into updating squashfs to 4.2, which includes xz support Mar 26 19:35:54 why xz? Mar 26 19:36:08 probably better compression? Mar 26 19:36:12 because its the new thing on the block Mar 26 19:36:18 xz = lzma2 Mar 26 19:36:22 I know :P Mar 26 19:36:33 ah... what symbols do I need to change to select this? Mar 26 19:37:01 or at least get a running box until the squashfs patches are committed? Mar 26 19:37:04 philipp64|laptop: it's already selected in the 2.6.38 config, mksquashfs just doesn't create xz compressed images yet Mar 26 19:37:11 It's just that while it does provide a nice boost to compression ratio... it won't make your system run faster :P Mar 26 19:37:11 philipp64|laptop: I don't understand your mail. I didn't revert because : vs true but because the commit broke generation of some required package Mar 26 19:37:27 I have several patches submitted into netdev for 2.6.38 but I'm having to test them under 2.6.37... sigh. Mar 26 19:38:01 xMff: my point was that ":" takes one less fork/exec than "true". Mar 26 19:38:02 philipp64|laptop: as long as you don't use ar71xx, I can give you a patch for creating booting images Mar 26 19:38:08 "true" isn't a builtin. that was my point. Mar 26 19:38:21 philipp64|laptop: I just reverted a change, I didn't even think about the contents Mar 26 19:38:46 KanjiMonster: I'm an x86 guy. Mar 26 19:39:19 xMff: ok, no problem... my quad-core Phenom/FX is plenty faster... I was just looking out for people building on slower systems. :-) Mar 26 19:39:33 philipp64|laptop: okay, then you won't miss the missing lzma options; a grep told me only ar71xx is actually using the lzma options for mksquashfs images ;) Mar 26 19:40:14 btw: it would be nice to have a dhcp client (for CATV and bridged DSL, i.e. PPPoE and RFC-2684 bridged mode) that saves your lease, then reloads it (optionally) on reboot if it hasn't expired. Mar 26 19:40:29 if I reboot, I'd like to have my system ask for the same IP back. Mar 26 19:41:01 KanjiMonster: I'm definitely missing something... it fails to boot. Mar 26 19:41:15 I think I posted a patch with missing 2.6.38 symbols, too. Mar 26 19:41:25 philipp64|laptop: I meant then you can use my patch to produce working images Mar 26 19:42:52 got it. Mar 26 19:42:56 btw... http://fpaste.org/zL6Y/ Mar 26 19:43:16 these are the symbols that are missing for x86, but I'm not sure if I have the correct values. Mar 26 19:43:40 how feasible is the busybox udhcpc patch? Mar 26 19:44:57 xMff: mr wikiman... I was looking for documentation on how to keep the squashfs and overlay filesystems separate so I could install a new image yet keep my modified files intact... is there a best-practice for doing this? Mar 26 19:45:19 I am not aware of any Mar 26 19:45:29 philipp64|laptop: try that one -> http://page.mi.fu-berlin.de/jgorski/openwrt/temp/Update-squashfs-to-4.2.patch Mar 26 19:45:35 xMff: poke Mar 26 19:46:57 philipp64|laptop: /files ? Mar 26 19:47:42 KanjiMonster: I'll fire it up when I get back in a couple of hours... do I need to set any .config or Kconfig symbols? Mar 26 19:50:07 Weedy... no, I mean if I've modified /etc/config on my running system (for instance), and I bring over new binaries... I still want to keep the contents of /overlay... Mar 26 19:50:38 that is what sysupgrade is for usually Mar 26 19:51:56 xMff: how was i supposed to masq when doing loopback conections so i can access the webui of my modem? Mar 26 19:52:11 i added the alias Mar 26 19:55:10 hmm... http://wiki.openwrt.org/doc/howto/generic.sysupgrade is a little thin on details Mar 26 19:57:40 ok, back in a bit... Mar 26 20:15:40 philipp64|laptop: nope, no further configuration needed Mar 26 21:04:09 nbd * r26307 /trunk/package/uci/ (Makefile patches/120-uci_trigger.patch): uci: update to latest version, uses cmake now Mar 26 22:14:54 KanjiMonster: btw. some of the fixes that it drops are still relevant Mar 26 22:15:18 (talking about the squashfs 4.2 patch) Mar 26 22:18:30 nbd: yeah, I already thought so, this is just so that he can start working with 2.6.38 - I intend to forward port the fixes Mar 26 22:19:40 ok Mar 26 22:20:51 nbd: btw, what's the exact reason for the special mksquashfs-lzma for ar71xx? Mar 26 22:23:04 wndr3700 Mar 26 22:23:31 the braindead boot loader reads its kernel from a squashfs partition Mar 26 22:23:49 only compatible to the old format Mar 26 22:24:26 ah, I see Mar 26 22:29:47 nbd * r26308 /trunk/package/uci/ (Makefile patches/ trigger/): remove ucitrigger, it is unused and will be replaced by something different in the future Mar 26 22:32:06 nbd: ah, I just saw, the -lzma one comes from the squashfs package, not squashfs4, so I don't need to fear I'll break it ;) Mar 26 22:32:32 right Mar 26 22:33:01 nbd: btw, I hope its okay if I ignore your lzma options patch and implement it in the same way as the xz options are implemented ;) Mar 26 22:33:16 sure Mar 26 22:34:11 squashfs 4.2 has brought some infrastructure for passing options to the compressors Mar 26 22:34:52 sounds good Mar 26 22:35:18 the lzma thing was just a quick hack anyway, just to be able to quickly compare the efficiency of various lzma parameters Mar 26 22:40:29 there's also support for storing/retreiving non default compressor options in the filesystem (xz uses this; I don't know if this is just convinience or actually necessary) Mar 26 22:45:03 nbd * r26309 /trunk/package/uci/Makefile: uci: install libucimap to the staging dir Mar 26 22:45:46 okay, this seems to be just for convinience for retrieving the actual settings used for creating an image Mar 26 22:57:17 back... what exactly is the difference between mtd and sysupgrade? Mar 26 22:58:05 mtd is a tool for writing data to flash devices, sysupgrade is a set of scripts that use it Mar 26 22:59:53 ah... and how to build the *-squashfs-sysupgrade.bin file? Mar 26 23:00:32 depends on the platform's (or device's) way of handling firmware upgrades Mar 26 23:00:43 and x86 with BIOS. Mar 26 23:00:56 *an x86. Mar 26 23:01:02 actually, a Geos in this case. Mar 26 23:04:15 on x86 the sysupgrade.bin is the same as the combined image Mar 26 23:04:27 ok, wanted to make sure. Mar 26 23:04:27 it uses a mix of mtd + dd to write the image and append the jffs2 data Mar 26 23:04:40 and where does SAVE_CONFIG get set? Mar 26 23:06:39 say, anyone know what the current turnaround is on patches to netdev? I submitted 4 solos-pci driver patches... DaveM has them currently "under review"... Mar 26 23:06:48 nbd: can I pick your (and anyone else here's) brains for devices that are well supported under OpenWRT, have 64MB RAM, and either 32MB flash or USB (or CF etc), supported wireless (preferrably N), and either two mini-PCI slots or FXO/FXS that are already supported (I don't think there are any of those) by asterisk, freeswitch, or Yate Mar 26 23:10:09 cshore: Tolapai. Mar 26 23:11:18 cshore: i don't know any devices matching that description Mar 26 23:11:24 philipp64: know any devices that use it? Mar 26 23:11:33 nbd: hmmm....ok, I was afraid of that Mar 26 23:11:46 nbd: I'm suspsecting an x86 board like Alix is what I'll need Mar 26 23:12:45 yup... it's the ep80579 target. Mar 26 23:13:14 you can also get it with two T1/E1 ports. Mar 26 23:13:15 nbd: do you know any that have FXS/FXO that are not currently supported (i.e. that the only thing missing of what I'm looking for)? Mar 26 23:13:16 I think the vodafone easybox 802's fit that description (lantiq, fxo, 64MB ram, usb) - you even get adsl (although annex b ;) Mar 26 23:13:27 cshore: no, i haven't looked into FXS/FXO at all Mar 26 23:13:44 KanjiMonster: cool thanks Mar 26 23:14:12 philipp64: I mean actual boards Mar 26 23:14:15 cshore: if you mean telephone ports with fxo/fxs Mar 26 23:14:32 KanjiMonster: yep Mar 26 23:14:37 cshore: (it also has an emulated s0 bus) Mar 26 23:14:51 * nbd is currently writing his first ubus based system daemon - netifd Mar 26 23:15:00 cool Mar 26 23:15:03 @nbd Mar 26 23:16:05 cshore: if you try to get one on ebay, there's a slight chance of getting one with different hardware though Mar 26 23:16:13 as soon as i've implemented basic interface command handling, i will start pushing it to my git tree so that people can see where this is going Mar 26 23:16:40 cshore: tolapai is an actual board. Mar 26 23:16:43 KanjiMonster: this would be for production, so not eBay Mar 26 23:17:19 philpp64: oh, ok, I'll have to do a search on it again. Mar 26 23:17:40 i thought tolapai was a SoC design, not a board Mar 26 23:18:22 you can get the sbd's from Intel (the chipset dev kits) as a ready-to-use board. Mar 26 23:18:26 cshore: but be warned, they have ralink wireless, so don't rely to much on it Mar 26 23:18:45 KanjiMonster: ah, ok, that's a problem Mar 26 23:18:55 someone (I forget who... maybe Kontrol) makes a de-populated version of the board for production. Mar 26 23:19:54 cshore: http://ark.intel.com/ProductCollection.aspx?series=34309&wapkw=%28ep80579%29 Mar 26 23:21:59 oops, my bad. it's a PCIe design. Mar 26 23:23:39 nbd: what about the Ubiquity RouterStation Pro or something like that? Mar 26 23:24:25 doesn't have FXS/FXO Mar 26 23:24:42 but has miniPCI, right? Mar 26 23:24:46 yes Mar 26 23:24:58 but not all minipci cards work Mar 26 23:25:14 the controller doesn't have a PCI I/O port range Mar 26 23:25:36 and it has some compatibility issues with some cards Mar 26 23:27:24 ah, ok, damn Mar 26 23:32:18 KanjiMonster: any idea when your squashfs patch for 2.6.38 will be committed? Mar 26 23:38:32 philipp64|laptop: when its finished and submitted ;) Mar 26 23:39:09 philipp64|laptop: I have to forward port the fixes first before I can submit it Mar 26 23:40:10 ahh.... so what else still needs to be finished other than merging with latest? Mar 26 23:42:57 nothing else I think. updating the build fixes and adding the lzma options should be everything needed Mar 27 00:54:07 nbd * r26310 /trunk/package/uci/Makefile: uci: update to fix lua plugin linker issue (#9121) Mar 27 00:55:24 hmmm.... well, sysupgrade didn't quite do what I expected... it seemed to ignore the files and directories in /etc/sysupgrade.conf Mar 27 00:58:13 btw. in case anybody is interested in a better ubus client example than what the ubus tree has to offer, i'll keep adding stuff here: http://nbd.name/gitweb.cgi?p=luci2/netifd.git;a=blob;f=ubus.c;hb=HEAD Mar 27 01:41:55 nbd: ping Mar 27 01:43:44 any chance of http://patchwork.midlink.org/patch/790/ - 793 being reviewed? Mar 27 01:49:46 756 can be marked "applied" **** ENDING LOGGING AT Sun Mar 27 02:59:59 2011