**** BEGIN LOGGING AT Mon Aug 03 02:59:57 2009 Aug 03 03:00:15 .. and you've worked out how to get patches applied to the downloaded source? Aug 03 03:00:35 yeah, I figured that out, pretty simple, I just don't know exactly what to patch :) Aug 03 03:00:56 http://pastie.org/569261 here's the current incarnation, I've tried a lot of different things following what tutorials I've found Aug 03 03:01:22 I think part of my problem is the really weak state of the package's makefile itself Aug 03 03:01:59 hmm.. were all those CXX= CPPFLAGS= CXXFLAGS= necessary? Aug 03 03:02:32 no, not all of them Aug 03 03:02:59 stevecrozz: a lot of that stuff should get passed automatically Aug 03 03:03:32 possibly you wnt to define a Build/Configure section and pass in ./configure options? Aug 03 03:03:47 the package doesn't have a ./configure script Aug 03 03:04:08 oh, ok. hence the options required to the Make ? Aug 03 03:04:50 yeah and the default makefile that comes with the package doesn't use many environment variables, it just explicitly states the compiler as g++ for instance Aug 03 03:04:54 stevecrozz: argh. I'd love to help now, but I gotta work. Aug 03 03:05:17 stevecrozz: hm. ok. That's where patches come in handy Aug 03 03:05:24 I understand that, maybe I can schedule an appointment with you at a more convenient time :) Aug 03 03:05:29 stevecrozz: patch the Makefile so that it _does_ use env variables. Aug 03 03:05:30 I could offer beer Aug 03 03:05:48 stevecrozz: there are a bunch of other people just as clued-up, and mostly more-so. Aug 03 03:06:03 yeah, I have to read a lot of documentation as I've really never written a Makefile Aug 03 03:06:13 it will be good for me, but its really slow going Aug 03 03:06:15 stevecrozz: these Makefiles are kinda 'special' Aug 03 03:06:34 stevecrozz: you've worked out that doing make menuconfig then quitting out, shows you any erros in your Makefile ? Aug 03 03:06:46 yes Aug 03 03:06:49 (as in, if it doesn't pass the buildroot-ng thing) Aug 03 03:06:53 cool Aug 03 03:06:54 my makefile actually works, kinda Aug 03 03:07:02 it downloads the package and attempts to install it Aug 03 03:07:11 but the compiling part fails Aug 03 03:07:22 well post some V=99 dumps of the compiling and somebody will possibly be able to help. Aug 03 03:07:23 gl Aug 03 03:07:35 if I get bored later I'll 'stop by' Aug 03 03:08:35 frogonwheels: thanks Aug 03 03:09:11 my first question is more general. Here's the package's included Makefile... how should it be patched? http://pastie.org/569268 Aug 03 03:11:13 hmm. very simplistic. you could always just not use make and compile it directly. there are examples of that in the packages. Aug 03 03:11:51 look in asterisk-1.4.x search TARGET_CC Aug 03 03:45:33 frogonwheels: actually that's similar to what I'm trying to do, I watched nmap compile (also compiled against uClibc++) and tried to mimick the compile command, but it throws lots of errors Aug 03 05:36:53 well it seems that I've got somewhere, but I get lots of this stuff: warning: ignoring #pragma GCC visibility, is that bad? Aug 03 05:46:10 I have successfully cross-compiled s3fs, but when I run it, it fails with: can't load library 'libstdc++.so.6', which is odd because I compiled it with -nostdinc++ Aug 03 05:54:11 w00000t! Aug 03 05:57:59 finally just got a working s3fs package Aug 03 06:29:13 stevecrozz: hey, well done. Aug 03 06:29:44 frogonwheels: thanks, but all is not quite well because I get this: "fuse: device not found, try 'modprobe fuse' first" Aug 03 06:29:58 and you have fuse installed? Aug 03 06:30:36 well i have libfuse, but i think i need the fuse kernel module as well Aug 03 06:30:40 i don't see a package for that Aug 03 06:31:12 there should be Aug 03 06:31:23 you building it? Aug 03 06:31:47 I have one for my x86 target. Aug 03 06:32:00 kmod-fuse*.ipk Aug 03 06:32:18 any idea what menu its under? Aug 03 06:32:36 kernel module Aug 03 06:32:57 kernel module -> filesystems? Aug 03 06:33:03 type /fuse Aug 03 06:33:14 (within make menuconfig) Aug 03 06:34:07 oh yeah, ive got it Aug 03 06:34:29 its "---" which I think means another thing depends on it and its automatically selected Aug 03 06:34:47 ytep Aug 03 06:34:49 yep Aug 03 06:34:55 is it automatically loaded or do I have to modprobe it? Aug 03 06:35:10 I think i took out the modprobe command to save a few kB Aug 03 06:35:11 stevecrozz: it should automatically load. Aug 03 06:35:22 lsmod doesn't show it Aug 03 06:35:24 stevecrozz: opkg list-installed kmod-fuse* Aug 03 06:36:09 kmod-fuse24 - 2.4.37.1+2.5.3-brcm-2.4-1 - Aug 03 06:36:43 ls /lib/modules/*/ Aug 03 06:37:11 i see fuse.o Aug 03 06:37:24 insmod fuse Aug 03 06:37:37 insmod: unresolved symbol r4k_flush_cache_page Aug 03 06:37:44 oooh. Aug 03 06:37:50 no idea Aug 03 06:37:59 ha, bummer Aug 03 06:38:18 when the other devs are online, they might be able to shed light. Aug 03 06:38:36 well at least ive got a tangible error now Aug 03 06:38:45 hey, it's a start :) Aug 03 06:38:54 probaby need modprobe now ;) Aug 03 06:39:05 though I don't have it. Aug 03 06:39:08 yeah, ill select that again Aug 03 06:39:29 dang it sucks to get so close, i was hoping to get it done tonight Aug 03 06:39:55 google that symbol name perhaps. Aug 03 06:43:24 yeah, ill see what i can come up with, and svn update while I'm at it, and probably post to the mailing list if i still can't get anywhere Aug 03 06:43:58 wait till later in here. you are fighting timezones :) Aug 03 06:45:49 yeah, ok... I'll probably be turning in for the evening myself soon, but I'll stay connected in case anyone has some advice Aug 03 06:46:07 thanks for your help btw Aug 03 06:46:22 np. Aug 03 07:47:15 I want to write some module for LuCI. But the LuCI documentation are very short. Now I have two question: how I can add my module to the main menu or submenu, how I can create a form? Aug 03 07:48:23 juhosg * r17100 /trunk/target/linux/ar71xx/base-files/etc/preinit.arch: [ar71xx] init mac address on the TEW-632BRP boards (closes #5403) Aug 03 07:52:29 dkostousov: check here http://luci.freifunk-halle.net/ Aug 03 07:56:37 I have read http://luci.freifunk-halle.net/Documentation/ModulesHowTo and http://luci.freifunk-halle.net/Documentation/Templates Aug 03 08:00:01 dang, after svn up and recompile I've got mount: mounting sysfs on /sys failed: No such device Aug 03 08:02:36 stevecrozz1: make distclean Aug 03 08:03:08 stevecrozz1: there was a change in lzma algorithm, that could cause the breakage Aug 03 08:03:36 danage: thanks... it would cause it to hang at boot like that? Aug 03 08:04:23 i don't know honestly, but it wouldn't mount the fs for me because of that at some point. but i think it was rootfs Aug 03 08:40:32 There was a similar problem! distclean helping! Aug 03 08:55:00 danage: ping Aug 03 08:56:49 hi glp Aug 03 09:16:54 danage: I didn't find answers for my questions Aug 03 09:20:24 speak Russian? Aug 03 09:21:30 Денис! Aug 03 09:21:57 Да Aug 03 09:22:09 AndyIL: говорю Aug 03 09:22:11 А ты хто? Aug 03 09:22:25 Руль человек! Aug 03 09:22:51 вот ищу мобыть кто из наших интересуется BCM63xx Aug 03 09:23:47 я в этих мыкросхемах не разбираюсь Aug 03 09:24:00 Мне нужно прошивку докрутить под свои нужны. Aug 03 09:24:10 а ты по какому вопросу тут? Aug 03 09:24:20 LuCI Aug 03 09:24:29 (at least utf-8 works) Aug 03 09:24:55 Ну и может кто знает, как OpenWRT на d-link 330 заставить работать. Aug 03 09:25:21 Gottox: heh Aug 03 09:25:26 а какой чип? (давай в личку ато задолбят) Aug 03 09:30:40 hi loswillios Aug 03 09:34:11 hi Aug 03 10:24:17 I have wrote a dummy module for LuCI. How I can add my module to menu at the main page (http://localhost:8080/luci)? Aug 03 11:24:58 kaloz * r17101 /trunk/ (include/package-defaults.mk package/opkg/Makefile): add special handling for the adm5120 target, which uses subtargets for different endianess -- later we should come up with something smarter Aug 03 11:44:30 kaloz * r17102 /trunk/ (include/target.mk target/linux/brcm-2.4/Makefile): brcm-2.4 is the only non-2.6 target, it's pointless to note the kernelversion in other targets' Target-Name Aug 03 11:59:59 hi all! Aug 03 12:00:58 is this the right channel for package makefile creation troubles? Aug 03 12:06:30 u can give it a try Aug 03 12:11:15 okay. my problem is: sources are not in tar's root directory, but in a subdirectory, and I can't find how to tell the configure to use the subdirectory Aug 03 12:12:00 setting pkg build dir to the subdir just resulted in the sources getting extracted to the wrong place Aug 03 13:49:28 lars * r17103 /packages/Xorg/driver/xf86-video-glamo/patches/ (. 001-include-glamo-engine.patch): Aug 03 13:49:28 [packages] xf86-video-glamo: Preliminary patch to make it compile during Aug 03 13:49:28 restructureing of the kernel part. Aug 03 14:44:16 ping jushog Aug 03 14:44:27 err, juhosg Aug 03 14:44:34 * groz got spelling issues this morning Aug 03 14:53:37 florian * r17104 /packages/libs/apr/Makefile: [package] re-enable ipv6 in apr (#5608) Aug 03 15:23:34 groz: btw. i will receive a wrt160nl soon Aug 03 15:23:39 then i'll finish the switch driver stuff Aug 03 15:58:18 groz: pong Aug 03 16:10:12 florian * r17105 /packages/ipv6/radvd/ (Makefile patches/003-linux_old_compat.patch): [package] update radvd to 1.4 Aug 03 16:31:38 [florian]: as you seem to care most about ipv6, any comments on https://dev.openwrt.org/ticket/5356? Aug 03 16:39:02 [florian] Hi! Aug 03 16:39:43 AndyIL: you have new patches? Aug 03 16:39:44 [florian] have questions about MTD_PHYSMAP Aug 03 16:39:45 :) Aug 03 16:40:02 hi! Aug 03 16:46:21 I cannot receive dump from an router Aug 03 17:13:35 * groz back Aug 03 17:13:48 ju you still here ? Aug 03 17:14:46 nbd, I haven't gotten to the driver yet, it's been hectic here, got family showed up to visit for a couple days Aug 03 17:14:49 kind of unexpected Aug 03 17:16:09 all peolpe go to drinking VODKA! Aug 03 17:16:32 AndyIL: all people go to drinking KAMIKAZE! Aug 03 17:21:30 moonflux: I think the cleanest solution is to check for /proc/net/ip6_route in 10-net and then extend the /e/i/network boot() function to re-emit hotplug events for all ipv6 enabled interfaces Aug 03 17:21:37 KAMIKAZE not cool! VODKA cool! Aug 03 17:22:12 xMff: but why should the net script be run out of order anyway? why not just wait until S40network? Aug 03 17:22:41 because interface setup is usually hotplug based Aug 03 17:22:50 the init script is more for humans Aug 03 17:23:14 yeah, but not really on boot (hotplug) because the system just isn't ready Aug 03 17:23:38 other stuff run from /etc/hotplug.d/iface is spilling all over the log as well Aug 03 17:23:46 proper fix is to check whether loading of ipv6 emits another hotplug action, then catch that Aug 03 17:24:32 loading ipv6 doesn't emit a hotplug action Aug 03 17:27:12 then 10-net could do the insmod Aug 03 17:27:19 ah, I just saw: the hotplug/coldplug is triggered by S10boot Aug 03 17:27:27 so why not move that to S40network? Aug 03 17:27:51 hm. or just move the loading of modules before that Aug 03 17:30:33 jepp. moving stuff from S10boot to S40network works as expected Aug 03 17:30:46 will create a patch Aug 03 17:30:57 I could imagine that this breaks other stuff Aug 03 17:31:41 but I'm not sure Aug 03 17:31:59 so then just move the loading of the modules up a few lines for 8.09 and have a try with moving to S40network for trunk? Aug 03 17:32:48 or moving the coldplug net-trigger below the modules is problaly safer Aug 03 17:32:52 yes Aug 03 17:33:02 then the modules can't interfer with the creating of the root device Aug 03 17:35:53 hmm... of course the modules could load network drivers which would lead to two actions for these. better create a list of devices first Aug 03 17:36:35 why two actions? Aug 03 17:37:05 you mean both hot- and coldplug? Aug 03 17:37:42 now, existing interfaces are first manually hotplugged, then the modules are loaded, which could be network drivers, so trigger another event. fine because there's only one event for each device Aug 03 17:38:00 but if I move the hotplug/coldplug after the module loading, I could get two events Aug 03 17:38:08 yes, right Aug 03 17:38:25 you could make the list before and the colplugging after module loading Aug 03 17:38:31 like you proposed Aug 03 17:38:52 jepp, working on it Aug 03 18:35:51 jow * r17106 /packages/net/olsrd/ (Makefile files/olsrd.init): [packages] olsrd: allow pipe signs for service descriptions in olsr init script; bump pkg revision Aug 03 18:51:04 florian * r17107 /trunk/target/linux/ar7/patches-2.6.27/ (160-cpmac_up_and_running.patch 910-cpmac_fixed_phy.patch): [ar7] add back the fixed PHY patches from #3124, until we move this to platform_data Aug 03 19:19:31 [florian] !!!! ping Aug 03 19:19:43 [florian] & Hi! Aug 03 19:21:25 console handover: boot [early0] -> real [ttyS0] Aug 03 19:21:26 physmap platform flash device: 00400000 at 1fc00000 Aug 03 19:21:26 ifconfig eth0 192.168.0.123 Aug 03 19:21:26 nc 192.168.0.120:1234 -l < /dev/mtd0 Aug 03 19:21:26 nc -l -p 1234 >dump Aug 03 19:23:09 root@(none):/# cat /proc/mtd Aug 03 19:23:10 dev: size erasesize name Aug 03 19:23:10 mtd0: 00200000 00010000 "whole-flash" Aug 03 19:23:47 [florian] dump working! with physmap-flash Aug 03 19:30:50 <[florian]> AndyIL: perfect ! Aug 03 19:33:09 <[florian]> try to bcm6332 with spi Aug 03 19:35:27 802.1Q VLAN Support v1.8 Ben Greear Aug 03 19:35:27 All bugs added by David S. Miller Aug 03 19:35:27 Freeing unused kernel memory: 532k freed Aug 03 19:35:27 - preinit - Aug 03 19:35:27 Press CTRL-C for failsafe Aug 03 19:35:27 Please press Enter to activate this console. Aug 03 19:35:37 and dont answer! Aug 03 19:37:02 shit! Aug 03 19:37:20 <[florian]> AndyIL: yes, I told you Aug 03 19:39:31 bcm6332 dont answer after boot! Aug 03 19:41:33 physmap platform flash device: 00400000 at 1fc00000 Aug 03 19:41:34 physmap-flash physmap-flash.0: map_probe failed Aug 03 19:41:34 bcm63xx_wdt started, timer margin: 30 sec Aug 03 19:41:41 spi dont work? Aug 03 19:49:03 groz: i'm here now, although until a few minutes only Aug 03 19:49:26 on phone Aug 03 19:49:44 ok Aug 03 19:51:00 <[florian]> AndyIL: I told you have to make sure bcm63xx_spi works with the flash Aug 03 19:53:36 <[florian]> router have 4 port switch bcm5325 ? possible this problem? Aug 03 19:57:25 needed change setting for eth? Aug 03 19:58:48 <[florian]> AndyIL: flash & switch can be on the SPI bus Aug 03 20:01:39 what to make ? Aug 03 20:01:59 <[florian]> AndyIL: debug bcm63xx_spi and make it "see" the flash Aug 03 20:02:08 <[florian]> AndyIL: see how it is for ar71xx_spi Aug 03 20:02:33 ok! Aug 03 20:02:38 try it! Aug 03 20:17:34 <[florian]> on another device spi work? Aug 03 20:18:22 <[florian]> AndyIL: slic works as far as I have tested them Aug 03 20:19:57 <[florian]> spi driver loan and system hang! Aug 03 20:23:09 <[florian]> lol Aug 03 20:23:22 <[florian]> AndyIL: make sure it accesses the right offsets in the SPI register Aug 03 20:26:30 jow * r17108 /packages/net/sslh/files/sslh.init: Aug 03 20:26:30 [PATCH] corrected test syntax in sslh init script Aug 03 20:26:30 Signed-off-by: Matthias Buecher Aug 03 20:27:35 ju, sorry, that was a real long call Aug 03 20:27:44 the question I had for you, fairly simple I think Aug 03 20:28:05 where would be the correct place to make changes in init scripts, for pulling ethernet mac out of nvram ? Aug 03 20:28:38 i've got nvram going properly here now, the only item left on my 'to do' list, is to actually get the ethernet macs correct right from power up Aug 03 20:29:08 if we need the nvram utility only for the mac addresses, i would like to skip it Aug 03 20:29:13 can you try this patch instead: http://openwrt.pastebin.ca/1516826 ? Aug 03 20:29:29 lemme take a look at the patch Aug 03 20:29:33 <[florian]> ok Aug 03 20:29:44 but I have found, 3 boxes come up with all th esame mac, if i'm not doing something to fix it Aug 03 20:30:02 ok, this looks like it may the the right thing Aug 03 20:31:48 this is working on the tew-632brp already Aug 03 20:32:04 ok, then it'll almost certainly work on this one, if they use the same nvram stuff Aug 03 20:32:43 if this cures the ethernet mac address issue, then, later tonite I'll see about fixing up my access, and start committing what I have Aug 03 20:32:59 which should result in a 'just works' bin file for the 160nl built from trunk Aug 03 20:33:32 the tew-632brp does not use the nvram stuff, but the idea is similar Aug 03 20:34:54 jow * r17109 /packages/net/kismet/Makefile: Aug 03 20:34:54 [PATCH] bump kismet to 2009-06-R1 Aug 03 20:34:54 tested on both x86 & mipsel Aug 03 20:34:54 Signed-off-by: Alexandros C. Couloumbis Aug 03 20:39:15 groz: i have to go, you should ask Kaloz about your access Aug 03 20:39:29 yah, not worried about that till it's all done Aug 03 20:39:33 and, i'm gonna have to go soon to Aug 03 20:39:37 thanks for the help Aug 03 20:40:13 you are welcome Aug 03 20:40:15 cya Aug 03 20:50:59 hehe, ok, he left already Aug 03 20:51:04 but woooooohooooo, it worked Aug 03 20:51:30 ping Kaloz Aug 03 21:32:33 florian * r17110 /trunk/target/linux/ar7/patches-2.6.30/ (5 files): [ar7] add fixed PHY patches from 2.6.27, refresh patches Aug 03 21:49:33 jow * r17111 /packages/net/kismet/ (3 files in 2 dirs): [packages] kismet: drop old atan patch, add new patch to properly link against libm, get rid of libnotimpl dependency, use a cleaner override for $(LD) (followup to r17109) Aug 03 21:54:32 jow * r17112 /packages/net/tor/Makefile: [packages] tor: bump version to 0.2.0.35 Aug 03 21:56:29 [florian] ping! spi hang in this line "bcm_spi_writeb(SPI_INTR_CLEAR_ALL, bs->regs, SPI_INT_MASK);" Aug 03 22:19:06 xMff: https://dev.openwrt.org/attachment/ticket/5356/0001--package-base-files-make-hotplug-load-ipv6-fixes.patch Aug 03 22:19:52 jow * r17113 /trunk/package/dropbear/ (Makefile files/dropbear.init): [package] dropbear: safely support remote restarting of dropbear process; bump pkg revision (#5498) Aug 03 22:21:52 florian * r17114 /trunk/package/ar7-atm/Makefile: [ar7] use updated Annex-A and B firmwares (#5296), bump to D7.04.03-R1 Aug 03 22:24:50 jow * r17115 /trunk/package/firewall/ (Makefile files/uci_firewall.sh): [package] firewall: add icmp_type option to specify the icmp type in rule sections, bump pkg revision (#5554) Aug 03 22:30:13 <[florian]> AndyI: ok, please provide a patch if you fix this Aug 03 22:30:43 florian * r17116 /trunk/package/acx/patches/005-2.6.30_fixes.patch: [package] acx fixes for 2.6.30 Aug 03 22:31:35 <[florian]> i dont understand spi Aug 03 22:32:25 <[florian]> AndyI: what do not you understand? Aug 03 22:33:22 <[florian]> AndyI: you need to define the correct offsets for SPI base register Aug 03 22:33:34 <[florian]> and register offsets in the SPI core Aug 03 22:36:03 moonflux: thanks Aug 03 22:36:27 <[florian]> How it to learn? And where it to register? Aug 03 22:41:19 jow * r17117 /packages/net/samba/ (Makefile patches/103-cp775.patch): Aug 03 22:41:19 [PATCH] Baltic language support for samba 2 Aug 03 22:41:19 Enable Baltic language support in samba2 package Aug 03 22:41:19 Signed-off-by: zintis.petersons@e-mail.lv Aug 04 00:06:49 jow * r17118 /trunk/package/base-files/ (Makefile files/lib/network/config.sh): [package] base-files: ensure that ipv6 is loaded if interfaces have an ip6addr defined, clean trailing whitespace, bump pkg revision (#5356) Aug 04 00:15:31 jow * r17119 /branches/8.09/package/base-files/ (Makefile files/lib/network/config.sh): [8.09] base-files: merge r17118, bump patchlevel Aug 04 00:43:56 nbd * r17120 /trunk/package/uci/trigger/apply_config: fix a typo in apply_config -t **** ENDING LOGGING AT Tue Aug 04 02:59:57 2009