**** BEGIN LOGGING AT Tue Apr 22 02:59:58 2014 Apr 22 03:15:52 build #579 of cobalt is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/579 Apr 22 05:47:38 build #551 of uml is complete: Failure [failed shell compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/551 Apr 22 06:57:28 build #586 of orion is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/586 Apr 22 08:08:56 blogic r40541 trunk/target/linux/lantiq/patches-3.10/0207-NET-fix-etop-igplen.patch * lantiq: fix maximum ethernet header length Apr 22 08:09:00 blogic r40542 trunk/target/linux/ generic/files/drivers/net/phy/adm6996.h generic/files/drivers/net/phy/adm6996.c * linux/generic: add MIB counters and port status to ADM6996 switch Apr 22 08:09:04 blogic r40543 trunk/target/linux/lantiq/patches-3.10/0019-NET-lantiq-adds-PHY11G-firmware-blobs.patch * Fix vr9_phy22f_a2x.bin firmware linking Apr 22 08:09:08 blogic r40544 trunk/package/boot/uboot-lantiq/Makefile * Depend on lantiq board Apr 22 08:09:13 blogic r40545 trunk/package/boot/uboot-lantiq/Makefile * Create firmware directory if it doesn't exist Apr 22 08:09:17 blogic r40546 trunk/package/boot/uboot-lantiq/ (40 files) * Refresh patches Apr 22 08:09:21 blogic r40547 trunk/package/boot/uboot-lantiq/Makefile * Add Fritz!Box 3370 images Apr 22 08:09:24 blogic r40548 trunk/package/boot/uboot-lantiq/Makefile * Add Zyxel P-2812HNU-Fx images Apr 22 08:09:28 blogic r40549 trunk/target/linux/ ramips/image/Makefile ramips/dts/AR725W.dts ramips/rt288x/config-3.10 * ramips: Add Airlink101 AR725W board support Apr 22 08:09:33 blogic r40550 trunk/target/linux/ (5 files in 4 dirs) * ramips: add Airlink101 AR725W userland support Apr 22 08:09:36 blogic r40551 trunk/tools/firmware-utils/ Makefile src/mkheader_gemtek.c * firmware-utils: add Gemtek header tool Apr 22 08:09:40 blogic r40552 trunk/target/linux/ramips/image/Makefile * ramips: build factory-compatible image for AR725W Apr 22 08:09:44 blogic r40553 trunk/target/linux/ lantiq/xway/profiles/bt.mk lantiq/xway/target.mk lantiq/dts/BTHOMEHUBV2B.dts lantiq/xway/config-default * lantiq: BT Home Hub 2B - overlayfs on top of ubi Apr 22 08:09:49 blogic r40554 trunk/target/linux/ (9 files in 9 dirs) * ramips: Add support for the 7Links PX-4885 wireless router. Apr 22 08:10:01 build #35 of pxa is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/pxa/builds/35 Apr 22 11:06:32 build #35 of omap is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/omap/builds/35 Apr 22 11:34:25 build #37 of x86_64 is complete: Failure [failed compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/x86_64/builds/37 Apr 22 11:36:45 KanjiMonster: ping Apr 22 11:37:00 danitool: pong Apr 22 11:37:34 this is my current script for generating the livebox image http://pastebin.com/7kw26ndV Apr 22 11:38:38 I don't know if it's better to only create rootfs partition in OpenWrT and then let openwrt split rootfs_data partition Apr 22 11:39:14 but when I do it it complains about "cowarly refusing to erase jffs2 filesystem" Apr 22 11:39:45 but it seems it still works saving the configuration Apr 22 11:40:11 sounds like the jffs2 partition isn't properly erased or so Apr 22 11:40:35 it was already erased by Redboot Apr 22 11:41:09 If I port a library to OpenWRT, what should I put in the Makefile file so that the libraries won't be stripped when packaged? Apr 22 11:41:19 if I manually do mtd erase rootfs_data all is made fine Apr 22 11:43:17 my cuestion is, my current script with two partions for rootfs (user_fs) and rootfs_data made within Redboot is OK to integrate it in OpenWrt? Apr 22 11:43:18 danitool: maybe openwrt's view of the jffs2 partition's area differs from redboots so redboot doesn't erase everything, making jffs2 then complain? Apr 22 11:43:32 it seems it's more clean than splitting rootfs Apr 22 11:43:51 Mazilo: what for? Apr 22 11:44:32 cyruff: To retain all the symbols. Apr 22 11:45:01 one time: make package/FOO/{clean,compile} STRIP=/bin/true Apr 22 11:45:22 retain symbols until next clean Apr 22 11:45:28 or change of version etc. Apr 22 11:45:56 otherwise you can disable stripping globally somewhere in menuconfig Apr 22 11:45:56 KanjiMonster: I don't think so, Redboot erases whole flash, then only copy the data for both the kernel and rootfs to their partitions Apr 22 11:46:38 i suppose if you really wanted to you could find some make-variable to "permanently" set STRIP to /bin/true in the makefile itself though i don't see the point in that Apr 22 11:46:55 well whole flash except fis directory, bootloader, and bootloader config Apr 22 11:47:04 cyrusf: If I disable stripping globally, all packages will be built without stripping. All I want, is a single package. Apr 22 11:47:32 cyrusf: I am prefer a 'permanent' solution. Apr 22 11:48:40 danitool: in your script you use -n for creating rootfs_data, for which help says "When set, no image data will be written to the flash. Only the FIS directory will be updated." Apr 22 11:48:51 so the partition is not erased at all Apr 22 11:50:43 well that script is to make two separate partitions, one for rootfs (user_fs) and the other for rootfs_data, this currently works fine Apr 22 11:52:54 KanjiMonster: the another script with only one rootfs partition, to allow openwrt splitting rootfs_data is the one with the problem http://pastebin.com/fie1aXbg Apr 22 11:53:38 so my cuestion is what's the best aproach for generating a firmware Apr 22 11:54:19 cyrusff: did you mean 'make package/FOO/{clean,compile} STRIP=/bin/true' or 'make package/FOO/{clean,compile} STRIP=/bin/false' ? Apr 22 11:54:33 /bin/true Apr 22 11:54:37 otherwise build will fail Apr 22 11:55:02 you could try something like STRIP:=/bin/true in Makefile but not sure if it would work equally well Apr 22 11:55:27 cyrusff: OK. Thanks. Apr 22 11:55:28 also thats not something you should ever (propose to) commit to a feed or openwrt itself Apr 22 11:56:48 cyrusff: This is not in-house used only. Apr 22 12:00:13 how can I run procd in debug-mode without a serial console? Apr 22 12:06:12 danitool: can you post the kernel messages maybe? also note there is a "fis erase" command which should allow you to properly erase the rootfs_data part (or erase the full rootfs partition before writing the partition) Apr 22 12:06:14 KanjiMonster: a bootlog after flashing http://pastebin.com/J9g8AULx Apr 22 12:07:15 nunojpg: pass init_debug=4 in the cmdline Apr 22 12:08:20 fis init -f already erases the partitions included entries in fis directories Apr 22 12:09:02 danitool: the end of filesystem marker should be at 0x0, not 0x10000 - there seems to be something wrong with the partition sizes Apr 22 12:09:56 0x10000 is the end of squashfs partition Apr 22 12:10:09 isn't it the begining of jffs2 partition? Apr 22 12:10:49 the end of filesystem marker is at offset 0x10000 of the jffs2 partition Apr 22 12:10:55 but it should be at the beginning Apr 22 12:12:28 mmm, no idea how to fix it Apr 22 12:12:54 with the another script I have no problems Apr 22 12:14:49 the first one I've linked here Apr 22 12:21:43 cyrusff: I did a search for "STRIP" and found 'package/utils/busybox/Makefile' contains "SKIP_STRIP=y \". Is this the same? Apr 22 12:22:27 Mazilo: no Apr 22 12:22:37 thats a busybox-specific make-flag Apr 22 12:22:44 have you tried STRIP:=/bin/true? Apr 22 12:22:46 cyrusff: IC. Apr 22 12:23:00 cyrusff: Not yet. Apr 22 12:34:41 the deadc0de i appended to the squashfs leads to: jffs2_scan_eraseblock(): End of filesystem marker found at 0x10000 Apr 22 12:35:31 without the deadcode appended that message isn't shown, but the result is the same Apr 22 12:35:41 cowarldly refusing ... bla bla Apr 22 12:38:24 cyrusff: Adding "STRIP:=/bin/true" to the OpenWRT Makefile for the package works as you described, i.e. no stripping on libraries but do stripping on binary programs in /bin, /sbin, /usr/bin, and /usr/sbin directories. Thanks. Apr 22 12:38:45 Mazilo: you're welcome Apr 22 12:52:55 hey jow_laptop. your changes to have a working luci-subpage olsr->neighbors are not fixing the problem on my testdevices with newest trunk :-( Apr 22 12:54:27 the netifd changes from ndb also didnt work proper. so the "wifi" command often kills the wifi Apr 22 13:23:24 blogic: Sorry, where? Apr 22 13:25:24 nunojpg: where what ? Apr 22 13:27:00 blogic: what is cmdline? boot options? where? Apr 22 14:08:46 nbd: tried out USE_SRC_DIR, it... sort of works. but it seems to corrupt the environment for any dependent packages, and it's kidna cumbersome with {clean,compile} not working. the CONFIG_SRC_TREE_OVERRIDE works well, but is perhaps oddly named, as it's still a git tree (and git only) and it requires them to be committed, it doesn't use the local path. Apr 22 14:09:11 still, the CONFIG_SRC_TREE_OVERRIDE is workable enough. Apr 22 14:11:21 I've updated the docs at http://wiki.openwrt.org/doc/devel/packages#working.on.local.application.source with this info too Apr 22 14:45:51 KanjiMonster: the problem was, I was using the roots-squashfs in bin directory, which is padded to 128k Apr 22 14:46:05 padding it to 64k, solved the problem Apr 22 14:48:28 now it seems all is going fine, should I use this partition layout with only one system partition splitted in two, instead using two? Apr 22 14:48:35 that was my first question Apr 22 14:50:30 with only one partition should be enough Apr 22 14:51:25 ok, i'll look on how to integrate this stuff in the build tree Apr 22 15:38:54 hi, how about removing *.orig and *.rej from .gitignore? Apr 22 15:39:55 sometimes, it's convenient to check those stale files with git status. Apr 22 16:59:15 build #35 of adm8668 is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/adm8668/builds/35 Apr 22 17:02:44 can somoene confirm that http://wiki.openwrt.org/doc/howto/build#building.single.packages and http://wiki.openwrt.org/doc/howto/build#errorno.rule.to.make.target are outdated? I've never had to use package/feeds/package/blah/{clean,compile} just staright up package/blah/{¢lean,compile} Apr 22 17:04:12 package/blah/{clean,compile,install} is enough Apr 22 17:06:12 install triggers a rebuild of the rootfs and package index? doesn't make any difference for just building a pkg in bin/platform/packages though right? Apr 22 17:08:02 no Apr 22 17:08:20 it generate the ipk-install/ generation Apr 22 17:08:29 which does the strip call Apr 22 17:09:52 so you just get an unstripped package without install then, (I've only ever used {clean,compile} and haven't noticed any real issues, but then space hasn't been a concern of mine Apr 22 17:10:14 without install no install happens and without that you get no package Apr 22 17:22:27 nbd: are the results from the network.sh api cached somehow per process? My script always fails with timeout: http://pastebin.com/1C6xnu36 Apr 22 17:23:21 but the LAN comes up and has an IP, I can run the script from the console and it works Apr 22 17:24:18 hrm, yeah, I don't see the packages made, I must have just been testing compile worked, and rebuilding the image. Apr 22 17:24:47 still, the lsat thing done when I do a {clean,compile} is "touch blah/root-atherons/stamp/.packagename_installed" Apr 22 17:24:55 and it's clearly not been installed Apr 22 17:29:14 nbd r40555 trunk/package/network/config/netifd/Makefile * netifd: update to latest version, adds support for disabling interfaces Apr 22 17:41:12 blogic: also, my AA branch doesn't build the packages, but trunk does.. http://pastebay.net/1427604 Apr 22 17:44:17 nbd: nevermind, just noticed network_flush_cache Apr 22 18:03:30 adri: anything? Apr 22 18:29:34 build #542 of rb532 is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/542 Apr 22 18:33:05 build #542 of ppc44x is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/542 Apr 22 18:38:49 build #36 of mpc83xx is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/mpc83xx/builds/36 Apr 22 19:04:33 DonkeyHotei: sorry, $JOB took priority Apr 22 19:22:49 build #511 of sibyte is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/511 Apr 22 20:52:05 could someone explain to me UBUS_METHOD and UBUS_METHOD_NOARG? Apr 22 20:52:29 why do we often use UBUS_METHOD_NOARG even if handler expects some argument (it calls blobmsg_parse)? Apr 22 22:02:26 Hello, what is the correct patch numbering if I would love to add a patch for uClibc? Apr 22 22:03:25 (patch that adds definition to libc/sysdeps/linux/common/bits/socket.h ) Apr 22 22:09:48 build #521 of avr32 is complete: Failure [failed shell compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/521 Apr 22 22:19:30 <[florian]> slachta: it is a backport of an upstream patch? Apr 22 22:20:40 yes, it is Apr 22 22:20:46 http://git.uclibc.org/uClibc/commit/libc/sysdeps/linux/common/bits/socket.h?id=1e2e4ac6193ffe0900bd392fa3c596883771eb34 Apr 22 22:20:56 <[florian]> something like 012 then Apr 22 22:21:29 okay :) I am making acpid for x86 and x86_64 working again Apr 22 22:45:49 [florian]: May I ask you to apply this patch? http://patchwork.openwrt.org/patch/5084/ Apr 22 22:48:11 also another fix for gnutls is coming (see buildbot logs, it broke quite a lot of packages) Apr 22 22:48:39 adri: no worries, just lemme know Apr 22 22:51:28 <[florian]> slachta: ok Apr 22 23:25:37 florian r40556 packages/libs/libxslt/Makefile * libxslt: update to version 1.1.28 Apr 23 00:38:21 build #449 of iop32x is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/449 Apr 23 02:05:09 build #587 of orion is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/587 **** ENDING LOGGING AT Wed Apr 23 02:59:58 2014