**** BEGIN LOGGING AT Wed Apr 21 02:59:56 2010 Apr 21 05:25:35 build #43 of orion is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/orion/builds/43 Apr 21 07:16:47 swalker, hm? Apr 21 07:18:23 swalker, ah, yep Apr 21 08:39:37 Hi. Can I flash Backfire 10.03 on DLink 320? Is the firmware works on this device? Apr 21 08:40:28 build #38 of uml is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/uml/builds/38 Apr 21 10:17:44 <_lore_> ping nbd Apr 21 10:22:55 What is the reason I loose the ability to choose any other image types, if I choose initrd? Apr 21 10:23:37 Hmm, see, it's actually initramfs, not initrd. Apr 21 10:28:32 yup Apr 21 11:20:00 <[florian]> xl0: indeed Apr 21 11:53:15 Can anybody answer to my question about stability Backfile on DLink DIR-320? Apr 21 12:07:53 jow: found the problem with miniupnpd Apr 21 12:08:14 in http://luci.subsignal.org/trac/changeset/5605/luci/trunk/applications/luci-upnp/luasrc/model/cbi/upnp you made some changes, which were fine, but.. Apr 21 12:08:40 this file should also be removed http://luci.subsignal.org/trac/browser/luci/trunk/applications/luci-upnp/root/etc/uci-defaults/luci-upnp Apr 21 12:08:57 abyrne: yes I recall this, will fix it later today so that it is part of the .1 release Apr 21 13:01:29 Could someone check, are mtd-utils broken only for me? Apr 21 13:01:50 arm_v5te_gcc-4.3.3+cs_eglibc-2.8_eabi Apr 21 13:02:33 <[florian]> broken like they do not work or do not build? Apr 21 13:03:01 Fail at link time. Apr 21 13:03:20 http://dumpz.org/18799/ Apr 21 13:03:55 <[florian]> can you add $(LIBGCC_S) in the list of libraries to link with? Apr 21 13:04:04 <[florian]> give me second to test this actually fix the issue Apr 21 13:05:45 You mean LDFLAGS+=$(LIBGCC_S) Apr 21 13:05:51 <[florian]> yes Apr 21 13:06:10 nico * r21067 /packages/utils/memtester/ (Makefile patches/100-quiet.diff): [packages] memtester: update to 4.1.3 Apr 21 13:10:05 nico * r21068 /packages/utils/bc/Makefile: [packages] bc: update to 1.06.95 Apr 21 13:10:29 <[florian]> xl0: I could not reproduce the issue with uclibc Apr 21 13:10:40 <[florian]> xl0: I suspect this is eglibc specific Apr 21 13:13:26 <{Nico}> xl0: try linking with '-lm $(LIBGCC_S) -lc' Apr 21 13:13:30 Addgin LDFLAGS does not affect the build. Either I am doing something wrong, or LIBGCC_S is not set. Apr 21 13:14:06 <{Nico}> xl0: how are you passing LDFLAGS ? Apr 21 13:15:51 <[florian]> xl0: try using LDFLAGS_mkfs.jffs2 as the variable name instead Apr 21 13:16:03 <[florian]> or override LZOLDFLAGS Apr 21 13:16:06 <[florian]> (not clear) Apr 21 13:16:10 Just += the variable in the Makefile, the build system is passing it to configure, it seems, or am I wrong? Apr 21 13:16:26 <[florian]> xl0: this shall be done in the Build/Compile macro Apr 21 13:16:41 <[florian]> xl0: ermm, in MAKE_FLAGS rather Apr 21 13:17:50 nico * r21069 /packages/net/pepsal/Makefile: [packages] pepsal: update to 1.2.2 Apr 21 13:18:54 [florian]: You mean, LDFLAGS="$(TARGET_LDFLAGS) $(LIBGCC_S)" Apr 21 13:19:10 Or I don't get how the target ldflags are passed. Apr 21 13:19:12 <[florian]> xl0: I mean somehting like this: Apr 21 13:19:48 build #37 of x86 is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/x86/builds/37 Apr 21 13:19:57 <[florian]> @@ -36,6 +36,7 @@ endef Apr 21 13:19:57 <[florian]> MAKE_FLAGS += \ Apr 21 13:19:58 <[florian]> DESTDIR="$(PKG_INSTALL_DIR)" \ Apr 21 13:19:58 <[florian]> BUILDDIR="$(PKG_BUILD_DIR)" \ Apr 21 13:19:58 <[florian]> + LZOLDFLAGS+="$(LIBGCC_S)" \ Apr 21 13:20:02 <[florian]> (not yet tested) Apr 21 13:25:18 mkfs.jffs2 builds fine now, and I see -lgcc_s there, but now it fails at mkfs.ubifs likewise. Apr 21 13:26:15 I suppose the problem was not zlib-related, and you just used the variable to pass the options? Apr 21 13:27:14 <[florian]> change to LDLIBS += $(LIBGCC_S) in the makefile and see if that works for both (it should) Apr 21 13:29:17 <[florian]> xl0: yes I used that variable to pass the option, but it seems to work only for the main directory, not subdirectories Apr 21 13:29:27 <[florian]> thus ubi-utils failing to link the same way Apr 21 13:30:37 Yep, that worked. Apr 21 13:31:07 I'm still a bit confused, can't we pass the additional ld flags at configure time? Apr 21 13:31:30 <[florian]> we could if it used a configure script, but it does not Apr 21 13:31:38 <[florian]> so such thing must be passed a compile time Apr 21 13:31:52 Heh, got it, thanks. ;) Apr 21 13:32:07 Should have looked into the build dir. ;) Apr 21 13:32:16 Or the build logs.. ;) Apr 21 13:32:26 <[florian]> so this is what you have right: http://openwrt.pastebin.com/BUyE8jNf ? Apr 21 13:33:03 Yep Apr 21 13:33:11 <[florian]> ok, will push that Apr 21 13:34:39 florian * r21070 /packages/utils/mtd-utils/Makefile: [package] fix linking failure of mtd-utils on ARM with eglibc, thanks to xl0 for testing. Apr 21 13:40:43 build #40 of xburst is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/xburst/builds/40 Apr 21 13:45:24 whats bet way to patch this file linux/include/net/sock.h "SOCK_DEBUGIN" to disable it as it seems every time it goes overwriten from build proces. Apr 21 13:48:51 maybe it is enabled by the kernel debugfs Apr 21 13:48:59 which is activated by default iirc Apr 21 13:50:53 the socket debug stuff is activated by using setsockopt with the option SO_DEBUG Apr 21 13:50:57 it's on a per-socket basis Apr 21 13:51:23 it's not activated by debugfs Apr 21 13:51:28 nor by a kernel config option Apr 21 14:00:22 well i dont know, i know this this costum apps have that sock debuing enable and its flooding logread , i can disable that in linux/include/net/sock.h or add some rule to logread to ignore these stuff.. this would be used in syslog.conf for syslog http://paste.debian.net/70055/ , best option would be to disable that in costum app but i dont have source nor there is version with that fixed. Apr 21 14:00:51 what kind of an app is that? Apr 21 14:04:02 nbd its card reader/cardserver Apr 21 14:04:33 what kind of loging app openwrt use? Apr 21 14:07:26 busybox syslog Apr 21 14:20:49 nbd and configuration file for it.. Apr 21 14:27:48 nbd I mean any configuraton file for it, as google doesnt help me a lot, everywhere i see syslog.conf which i dont havehere Apr 21 14:28:02 no idea, i've never played with that stuff Apr 21 14:39:19 crow: bb syslogd is quite limited, it has not much config options and certainly does not support filtering of any kind Apr 21 14:39:29 crow: can't you just recompile the software in question? Apr 21 14:39:41 jow_laptop nop i dont have source for it Apr 21 14:39:49 crow: hm Apr 21 14:40:16 i have option to disable sock_debuging in sock.h file or to let syslogd to ignore it Apr 21 14:40:56 now seems only sock.h file option is posible. now the question is how to correct patch that file which is always new exctracted Apr 21 15:01:12 is it right place for patch to be in here target/linux/generic-2.6/patches-2.6.32/costum_sock_h.patch ? Apr 21 15:06:13 it needs to have a number Apr 21 15:19:40 nbd any range of number for costum patch? Apr 21 15:20:01 pick one Apr 21 15:20:06 i'd use 999 or something like that Apr 21 15:23:27 <[florian]> :x Apr 21 15:23:34 <[florian]> grr, wrong window :p Apr 21 16:00:22 nbd ok thnx will pick up some number, now the quesion is how clean is this: http://paste.debian.net/70068/ i never did something like this :) Apr 21 16:23:12 nbd does this should be like this http://paste.debian.net/70072/ (i am no shure from where to make diff (i was doing it from here build_dir/linux-brcm47xx/linux-2.6.32.10$ diff -uNr include/net/sock.h include/net/sock.h_new > 666-sock_h.patch ) Apr 21 17:21:33 acoul * r21071 /trunk/tools/cloog/Makefile: tools: update cloog to 0.15.9 Apr 21 17:29:36 nico * r21072 /packages/admin/sudo/ (Makefile patches/101-install_no_strip.patch): [packages] sudo: update to 1.7.2p6 Apr 21 17:40:01 thepeople: responded to your email Apr 21 17:40:24 ping xMff: Apr 21 17:40:30 ping crow Apr 21 17:40:35 pong cshore Apr 21 17:40:45 xMff: you were looking for me? Apr 21 17:41:02 yes bit I forgot why Apr 21 17:41:06 but Apr 21 17:41:22 heh, I do that too Apr 21 18:09:10 cshore well i found yesterday that bug with ext4, and reported here and jow_laptop backported fix from trunk.. Apr 21 18:09:36 cool Apr 21 18:10:16 cshore also today on #openwrt someone is having big time with block-extroot and sdhc... i am not shure if extroot work with sdhc or not.. Apr 21 18:10:44 hhmmm....I know there was an update for it a little while ago Apr 21 18:13:00 build #58 of brcm63xx is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/brcm63xx/builds/58 Apr 21 18:16:29 cshore its Mr_Smoke who is trying to get it working with sdhc Apr 21 18:31:48 the fundamental issue with sdhc is that it needs user provided params to work Apr 21 18:32:02 and it is not loaded by the regular module mechanism Apr 21 18:45:25 ah, ok, that's something that can't be solved without changing the regular module mechanism Apr 21 18:45:44 right Apr 21 19:07:13 juhosg * r21073 /trunk/target/linux/generic-2.6/ (2 files in 2 dirs): kernel: add KEY_WPS_BUTTON definition Apr 21 19:07:16 juhosg * r21074 /trunk/package/button-hotplug/src/button-hotplug.c: package/button-hotplug: handle KEY_RESTART and KEY_WPS_BUTTON codes as well Apr 21 19:07:21 juhosg * r21075 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/ (21 files): ar71xx: use KEY_RESTART and KEY_WPS_BUTTON code for the appropriate buttons Apr 21 19:07:24 juhosg * r21076 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/mach-nbg460n.c: ar71xx: reduce button threshold value on the NBG460 board Apr 21 19:18:18 hauke * r21077 /trunk/package/mac80211/Makefile: Apr 21 19:18:18 mac80211: add mwl8k driver Apr 21 19:18:18 This closes #7209 Apr 21 20:42:10 nbd * r21078 /trunk/include/prereq-build.mk: fix gnu find prereq check Apr 21 20:43:18 nbd * r21079 /branches/backfire/include/prereq-build.mk: [backfire] merge prereq check fix from r21078 Apr 21 21:45:54 mirko * r21080 /packages/net/cups-bjnp/Makefile: add path to cups, as it does not detect it automatically Apr 21 21:46:31 build #56 of brcm47xx is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/56 Apr 21 22:01:12 build #55 of brcm_2_4 is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/brcm_2_4/builds/55 Apr 22 01:24:11 is there a special reason why the iwlwifi kernel drivers are missing from buildroot menu? if I create a patch would someone commit it? Apr 22 02:43:33 wow.. hi everyone. here's so many person more than i expected. **** ENDING LOGGING AT Thu Apr 22 02:59:57 2010