**** BEGIN LOGGING AT Sun May 27 02:59:59 2012 May 27 04:58:39 build #22 of lantiq is complete: Failure [failed shell_15] Build details are at http://buildbot.openwrt.org:8010/builders/lantiq/builds/22 May 27 06:04:05 builder_: why u wake me upz with bad news May 27 06:05:04 builder_: why u wrong ?! May 27 06:05:04 What you say! May 27 06:05:23 wow May 27 06:05:26 it talks May 27 06:05:27 haha May 27 06:05:41 jow_laptop: whats the story with builder_ ? May 27 10:59:18 blogic: you shouldn't care much about shell_* failures, these just mean buildbot (or its maintainer ;p) screwed up ;) May 27 11:04:42 luckily we have you to point at maintainer screwing up May 27 11:04:44 always a joy May 27 12:33:57 nbd * r31870 /trunk/package/kernel/modules/ (7 files): kernel: remove a few more obsolete CompareKernelPatchVer calls, add dependencies on @TARGET_x86 for a few network drivers that are typically not found in non-PC equipment May 27 13:22:09 jogo * r31871 /trunk/target/linux/brcm63xx/base-files/ (etc/uci-defaults/brcm63xx_fixcrc.sh lib/brcm63xx.sh): May 27 13:22:09 bcm63xx: use the exact board id as board identification string May 27 13:22:09 No need to keep the bcm63xx/ at the start or the chipid info at the end. May 27 13:22:12 jogo * r31872 /trunk/target/linux/brcm63xx/base-files/ (etc/init.d/defconfig lib/brcm63xx.sh): bcm63xx: move the board name workaround to /lib/brcm63xx.sh May 27 13:22:15 jogo * r31873 /trunk/target/linux/brcm63xx/base-files/etc/ (11 files in 11 dirs): bcm63xx: setup network config through uci-defaults May 27 13:22:19 jogo * r31874 /trunk/target/linux/brcm63xx/base-files/etc/uci-defaults/network: bcm63xx: add a default switch configuration for DSL-274XB May 27 13:22:20 jogo * r31875 /trunk/tools/firmware-utils/src/ (4 files): May 27 13:22:21 tools: imagetag: add parameter for padding images May 27 13:22:21 Allow images to be padded to a certain size. This prevents CFE from May 27 13:22:21 flashing them to the second image offset. May 27 13:22:23 jogo * r31876 /trunk/target/linux/brcm63xx/image/Makefile: May 27 13:22:23 bcm63xx: pad Speedport W303 V images to 4 MiB May 27 13:22:23 Prevents CFE from flashing to the second image offset. Fix up the rootfs May 27 13:22:23 length on first boot else CFE will complain. May 27 13:22:26 jogo * r31877 /trunk/target/linux/brcm63xx/patches-3.3/ (32 files): bcm63xx: add flash type detection May 27 13:22:30 jogo * r31878 /trunk/target/linux/brcm63xx/ (21 files in 2 dirs): bcm63xx: add preliminary support for bcm6328 May 27 13:22:34 jogo * r31879 /trunk/target/linux/brcm63xx/ (33 files in 2 dirs): May 27 13:22:34 bcm63xx: add support for the HSSPI controller May 27 13:22:35 Add support for the HSSPI controller found on bcm6328 and SPI attached May 27 13:22:35 flash. May 27 13:22:39 jogo * r31880 /trunk/target/linux/brcm63xx/patches-3.3/ (23 files): bcm63xx: add a fixup for ath9k devices May 27 13:22:42 jogo * r31881 /trunk/target/linux/brcm63xx/patches-3.3/431-bcm63xx_enet-reset-port-link-state-in-bcm_enetsw_ope.patch: May 27 13:22:42 bcm63xx: fix port state after network down May 27 13:22:42 bcm_enetsw_probe did not properly reset port states, resulting in May 27 13:22:42 connected ports staying disabled after bringing up the interface the May 27 13:22:42 second time. May 27 13:22:44 jogo * r31882 /trunk/target/linux/brcm63xx/ (8 files in 3 dirs): bcm63xx: add support for DSL-274XB rev F1 May 27 13:22:47 jogo * r31883 /trunk/target/linux/brcm63xx/config-3.3: May 27 13:22:48 bcm63xx: enable bcm6328 support and required drivers May 27 13:22:48 BCM6328 support is in a usable state, with the following issues May 27 13:22:48 * internal switch port not configurable May 27 13:22:48 * external phys not supported May 27 13:22:48 * broadcom wifis will probably not work correctly May 27 13:22:49 * and all other generic bcm63xx issues (so yes, still no ADSL) May 27 13:22:49 jogo * r31884 /trunk/target/linux/brcm63xx/profiles/102-Atheros.mk: bcm63xx: add ath9k to atheros wifi profile May 27 13:22:53 jogo * r31885 /trunk/target/linux/brcm63xx/image/Makefile: May 27 13:22:53 bcm63xx: build images for DSL-274XB revision F1 May 27 13:22:53 Build images for D-Link DSL-274XB revision F1. Flashing May 27 13:22:53 through original web UI as well as CFE works. May 27 13:22:54 Known issues: May 27 13:22:54 * wifi led inverted (on when wifi is off and vice cersa) May 27 13:22:55 * same issues as generic bcm6328 support May 27 13:23:04 *spams the channel* May 27 13:56:08 spam ? May 27 13:56:28 save some for my breakfast please :) May 27 14:00:54 ;-) May 27 14:43:59 nbd * r31886 /trunk/target/linux/generic/patches-3.3/ (3 files): kernel: add a few patches for avoiding unnecessary skb reallocations - significantly improves ethernet<->wireless performance with mac80211 based drivers May 27 14:46:33 r31886 seems nice May 27 14:49:12 on my ar9344 based board (400 mhz) it went from 130 mbit/s @ 55% cpu load to 170 mbit/s at <50% cpu load May 27 14:50:32 and the patches are quite small ;) May 27 14:51:06 so i should recompile my wr740, wr1043 and dir825 May 27 15:01:05 juhosg * r31887 /trunk/target/linux/cns3xxx/Makefile: cns3xxx: switch to 3.3 and add broken flag May 27 15:01:05 juhosg * r31888 /trunk/target/linux/cns3xxx/ (config-default patches/): cns3xxx: remove 3.1 support May 27 15:01:07 juhosg * r31889 /trunk/target/linux/omap24xx/Makefile: omap24xx: switch to 3.3 and add broken flag May 27 15:01:08 juhosg * r31890 /trunk/target/linux/omap24xx/ (config-3.1 patches-3.1/): omap24xx: remove 3.1 support May 27 15:01:10 juhosg * r31891 /trunk/target/linux/generic/ (config-3.1 patches-3.1/): linux/3.1: R.I.P. May 27 15:01:16 juhosg * r31892 /trunk/target/linux/au1000/ (5 files in 3 dirs): May 27 15:01:16 au1000: add support for 3.3 May 27 15:01:16 Compile tested only. May 27 15:01:18 juhosg * r31893 /trunk/target/linux/au1000/Makefile: au1000: switch to 3.3 and add broken flag May 27 15:01:19 juhosg * r31894 /trunk/target/linux/au1000/ (au1500/config-3.2 au1550/config-3.2 patches-3.2/): au1000: remove 3.2 support May 27 15:01:22 juhosg * r31895 /trunk/target/linux/ep93xx/ (5 files in 2 dirs): May 27 15:01:22 ep93xx: add support for 3.3 May 27 15:01:22 Compile tested only. May 27 15:01:24 juhosg * r31896 /trunk/target/linux/ep93xx/Makefile: ep93xx: switch to 3.3 and add broken flag May 27 15:01:25 juhosg * r31897 /trunk/target/linux/ep93xx/ (config-3.2 patches-3.2/): ep93xx: remove 3.2 support May 27 15:01:27 juhosg * r31898 /trunk/target/linux/cobalt/ (Makefile config-3.3 config-default): May 27 15:01:27 cobalt: switch to 3.3 and add broken flag May 27 15:01:27 Compile tested only. May 27 15:01:29 juhosg * r31899 /trunk/target/linux/ppc44x/ (4 files in 2 dirs): May 27 15:01:29 ppc44x: add support for 3.3 May 27 15:01:29 Compile tested only. May 27 15:01:30 juhosg * r31900 /trunk/target/linux/ppc44x/Makefile: ppx44x: switch to 3.3 and add broken flag May 27 15:01:31 juhosg * r31901 /trunk/target/linux/ppc44x/ (config-default patches/): ppc44x: remove 3.2 support May 27 15:01:33 juhosg * r31902 /trunk/target/linux/xburst/ (27 files in 6 dirs): xburst: add support for 3.3 May 27 15:01:35 juhosg * r31903 /trunk/target/linux/xburst/Makefile: xburst: switch to 3.3 and add broken flag May 27 15:01:37 juhosg * r31904 /trunk/target/linux/xburst/ (11 files in 7 dirs): xburst: remove support for old kernels May 27 15:01:39 juhosg * r31905 /trunk/target/linux/generic/ (config-3.2 patches-3.2/): linux/3.2: R.I.P. May 27 15:16:38 nbd * r31906 /trunk/target/linux/generic/patches-3.3/654-avoid_skb_cow_realloc.patch: kernel: properly pad the allocated headroom in skb_cow to NET_SKB_PAD May 27 15:22:14 nbd * r31907 /trunk/package/kernel/modules/ (7 files): kernel: remove more unnecessary kernel version checks/dependencies May 27 15:47:54 hmmm that r31886 seems interesting indeed :) May 27 16:02:19 blogic * r31908 /trunk/target/linux/lantiq/ (7 files in 7 dirs): May 27 16:02:19 [lantiq] Add basic configuration for new board ZTE H201L. Still missing are button and led gpio values. May 27 16:02:19 Signed-off-by: Luka Perkov May 27 16:02:21 blogic * r31909 /trunk/target/linux/lantiq/files-3.3/arch/mips/lantiq/xway/mach-p2601hnfx.c: May 27 16:02:22 [lantiq] fixes zyxel p2601hnfx May 27 16:02:22 moves the linux partition, removes the config partition and the May 27 16:02:22 buttons settings have small adjustments. May 27 16:02:22 Signed-off-by: Luka Perkov May 27 16:02:24 blogic * r31910 /trunk/target/linux/lantiq/ (3 files in 2 dirs): (log message trimmed) May 27 16:02:24 [lantiq] dgn3500 support with eeprom loading from sysfs May 27 16:02:24 WIFI eeprom: May 27 16:02:24 As discussed, it is impossible for nand and spi flash platforms to have the May 27 16:02:25 eeprom data available from mtd. I suggested to load the eeprom from user-space. May 27 16:02:25 I've looked into regular firmware loading but this is only possible when using May 27 16:02:25 modules. I've created a sysfs entry that allows reading and writing the eeprom May 27 16:13:00 blogic * r31911 /trunk/target/linux/lantiq/ (config-3.2 files-3.2/ patches-3.2/): [lantiq] 3.2 R.I.P. May 27 16:13:03 blogic * r31912 /trunk/target/linux/lantiq/ (39 files in 10 dirs): [lantiq] move files-3.3 -> files May 27 16:40:35 juhosg * r31913 /packages/utils/mtd-utils/ (8 files in 2 dirs): (log message trimmed) May 27 16:40:35 package/mtd-utils: update to 1.4.5 May 27 16:40:35 This updates packages/utils/mtd-utils and includes all patches from May 27 16:40:35 tools/mtd-utils, so the two may eventually be merged. May 27 16:40:35 I also added 150-fix_ubi-utils_static.patch which is a hack needed May 27 16:40:35 to include libcrc32.o in the static binaries. May 27 16:40:36 Having them static makes sense as only ubiformat is required on the target. May 27 16:58:21 [florian]: i'm looking over the state of ar7 now, and it seems the only currently supported kernel patches (3.3) were copied from the 2.6.37 patches, which were in turn submitted by Wipster, which he forked from a very outdated patchset for 2.6.32 that was never updated to what was in 2.6.32 at that time May 27 16:59:17 granted, 2.6.32 was somewhat broken at the time, but i was working on still-uncommitted patches to fix that. i never finished, but i still hope to May 27 17:00:35 problem is reconciling that with what Wipster submitted, and all the patches to it since May 27 17:09:43 juhosg * r31914 /trunk/target/linux/generic/files/drivers/net/phy/rtl8367.c: generic: rtl8367: allow to use VLANs > 31 May 27 17:09:44 juhosg * r31915 /trunk/target/linux/generic/ (5 files in 3 dirs): generic: rtl8366_smi: rename debugfs Kconfig symbol May 27 17:09:47 juhosg * r31916 /trunk/target/linux/generic/files/drivers/net/phy/rtl8366_smi.c: generic: rtl8366_smi: remove duplicated spinlock initialization May 27 17:09:48 juhosg * r31917 /trunk/target/linux/generic/files/drivers/net/phy/rtl8367.c: generic: rtl8367: rename a misnamed variable May 27 17:09:54 juhosg * r31918 /trunk/target/linux/ramips/base-files/etc/hotplug.d/firmware/10-rt2x00-eeprom: ramips: make rt2x00 firmware extraction quiet May 27 17:09:57 juhosg * r31919 /trunk/target/linux/ramips/files/arch/mips/ralink/rt305x/mach-nw718.c: ramips: rt305x: remove superfluous controller_data assignment May 27 17:09:58 juhosg * r31920 /trunk/target/linux/ramips/files/arch/mips/ralink/rt3883/rt3883.c: ramips: rt3883: fix GPIO reservation for the SPI pins May 27 17:10:00 juhosg * r31921 /trunk/target/linux/ramips/patches-3.3/105-ramips-spi-driver.patch: ramips: allow to use the SPI driver on RT3662/RT3883 May 27 17:10:04 juhosg * r31922 /trunk/target/linux/ramips/files/arch/mips/ralink/rt3883/ (devices.c devices.h): ramips: rt3883: add device registration code for the SPI controller May 27 17:21:48 juhosg * r31923 /trunk/target/linux/ar71xx/image/Makefile: ar71xx: fix TL-WR1041 image size May 27 19:13:10 * stappers has DEFAULT_PACKAGES extended in include/target.mk. Doing `touch .config` and `make packages/compile` does NOT show any effect. May 27 19:13:35 rm -r tmp/ May 27 19:13:45 How to trigger "rebuild rootfs May 27 19:13:54 target/linux/install May 27 19:14:17 Will try again when current build has finished May 27 19:35:59 `make target/linux/install` does not build any of the extra packages I want. May 27 19:36:45 right May 27 19:36:51 you asked for "rebuild rootfs" May 27 19:36:57 not for "compile new packages" May 27 19:40:41 Oops May 27 19:43:01 After extending DEFAULT_PACKAGES in include/target, how to compile all packages ( all pkg = previous compiled pkg + new pkg ) ? May 27 19:43:16 rm -r tmp/ didn't help May 27 19:44:45 * stappers means packages that go into the image. ( not extra .opkg packages ) May 27 19:47:40 "make world" May 27 19:50:52 ah and another thing May 27 19:51:04 DEFAULT_PACKAGES simply define the menuconfig defaults May 27 19:51:14 once there is a populated .config, they have zero effect May 27 19:51:30 rm .config; make defconfig/make menuconfig; make would include your defaults May 27 20:00:31 okay, the 'once populated' explains much. May 27 20:06:51 it is also not a hard dependency May 27 20:07:11 it simply sets the initial state of the corrresponding CONFIG_PACKAGE_* symbols to "y" May 27 20:16:57 updated openwrt/upstream, https://home.comcast.net/~sdwalker/uscan/index.html May 27 20:31:17 Which file gets / holds the CONFIG_PACKAGE_pkgname=y May 27 20:35:55 .config May 27 20:41:04 That makes checking modifications in include/target.mk much quicker. Thank you, jow_laptop May 27 20:46:27 Strange ... I have 'CONFIG_DEFAULT_block-mount=y', but I don't see "block-mount" in the output of `make world` May 27 20:48:42 maybe because CONFIG_PACKAGE_block-mount is not set May 27 20:50:58 Yep ( now pasting 5 lines ) May 27 20:51:04 $ grep block .config May 27 20:51:05 CONFIG_DEFAULT_block-mount=y May 27 20:51:05 # CONFIG_PACKAGE_block-mount is not set May 27 20:51:05 # CONFIG_PACKAGE_kmod-block2mtd is not set May 27 20:51:05 # CONFIG_PACKAGE_badblocks is not set May 27 20:51:42 so set it May 27 20:53:04 Euh, that is what i was expecting from my change in include/target.mk May 27 20:53:16 so did you delete .config and started from scratch? May 27 20:53:47 as I said your DEFAULT_PACKAGES list simply holds defaults for menuconfig May 27 20:53:51 it has zero effect May 27 20:53:57 if the .config is already populated May 27 20:54:13 and if you hand edit stuff then changing CONFIG_DEFAULT_* is the wrong way to do so May 27 20:54:41 you would need to both change CONFIG_DEFAULT_* and CONFIG_PACKAGE_* or purge bo5th entries and run make defconfig May 27 20:55:11 however I am not sure what your end goal is, editing include/target.mk to add a set of defaults sounds wrong May 27 20:58:25 nope. I did `make defconfig` and was happy when I saw that .config was touch. Removing .config and `make menuconfig` shows CONFIG_{DEFAULT,PACKAGE}_block-mount=y. Yes \o/ May 27 20:59:02 make defconfig only enforces hard dependencies and fills out missing symbols with defaults May 27 20:59:12 it does not alter existing symbols with changed defaults May 27 20:59:48 Acknowledge on 'make defconfig' May 27 21:00:42 Also Acknowledge on "editing include/target.mk to set default sound wrong". May 27 21:01:09 I mean if you work on a new target or board you can put your extra packages into a profile May 27 21:02:02 if its about automating builds then its usually cleaner to prime .config with a few essential symbols (echo CONFIG_TARGET_foobar=y; echo CONF_PACKAGE_bla=y) > .config and call "make defconfig" on it May 27 21:02:39 nbd * r31924 /trunk/target/linux/generic/patches-3.3/202-reduce_module_size.patch: kernel: fix suppressing -s for linking modules when profiling is enabled May 27 21:02:43 nbd * r31925 /trunk/target/linux/ar71xx/files/ (4 files in 3 dirs): ar71xx: fix MII clock settings for various chips, improves ethernet stability on AR934x May 27 21:04:42 jow_laptop: That is a nice advice. May 28 00:16:41 swalker * r31926 /packages/libs/ldns/Makefile: [packages] ldns: update to 1.6.13 May 28 00:17:02 swalker * r31927 /packages/libs/libssh2/Makefile: [packages] libssh2: update to 1.4.2 May 28 00:17:15 swalker * r31928 /packages/net/unbound/ (Makefile patches/001-conf.patch): [packages] unbound: update to 1.4.17 May 28 00:17:25 swalker * r31929 /packages/net/umurmur/Makefile: [packages] umurmur: update to 0.2.10 May 28 00:17:37 swalker * r31930 /packages/utils/bluelog/Makefile: [packages] bluelog: update to 1.0.4 May 28 00:28:33 how do i do a verbose kernel build to find an error? V=s and V=99 don't work May 28 00:28:50 what kind of verbosity are you interested in? May 28 00:29:08 the kind that won't hide what kills the kernel build May 28 00:29:16 what kills? May 28 00:29:27 it stops with an error May 28 00:29:46 show me what you get with V=99 May 28 00:30:03 i'm trying V=1 now to see whether it's any different May 28 00:30:10 no need May 28 00:30:20 V=1 is definitely wrong May 28 00:30:23 V=99 has the last line start with CC and the source file May 28 00:30:31 and then? May 28 00:30:36 stops May 28 00:30:43 it just goes back to the prompt? May 28 00:30:46 yep May 28 00:31:00 interesting May 28 00:31:04 never seen that one May 28 00:31:11 it seems to even kill the openwrt build system then May 28 00:31:25 otherwise you'd have seen a few more lines May 28 00:31:26 no, the error propagates May 28 00:31:34 so it isn't the last line after all May 28 00:31:46 show me a paste, please May 28 00:31:50 it's the last line that shows progress May 28 00:31:54 i prefer raw data over vague interpretations May 28 00:32:24 trying V=99 again now May 28 00:32:32 i'll paste the result May 28 00:32:34 ok May 28 00:33:48 hmm, not piping through less shows more May 28 00:34:07 you probably forgot to redirect stderr to stdout May 28 00:34:18 the only lines after CC are "Leaving directory" May 28 00:34:26 even with no pipe May 28 00:34:46 wait May 28 00:34:53 ok, i just scrolled WAY up May 28 00:35:08 i went right past the error over and over May 28 00:35:15 did you use -j? May 28 00:35:21 nope May 28 00:35:29 laptop would overheat if i did May 28 00:35:33 k May 28 00:39:21 ok, found the problem: an unclean svn revert May 28 00:42:39 nbd: what do you know of the current state of ar7? May 28 00:42:58 nothing May 28 00:43:04 :/ May 28 00:43:13 it's been many years since i touched ar7 May 28 00:43:16 does it even work atm? May 28 00:43:20 no idea May 28 00:43:30 i guess i'll find out May 28 00:43:57 i highlighted [florian] earlier, when he was only 4 minutes idle, no response May 28 00:44:28 maybe he was busy May 28 00:45:45 one can still buy ar7 devices brand-new in the store, you know May 28 00:46:57 yeah, but those devices all have crappy wifi May 28 00:47:04 ;) May 28 00:47:09 ar7 wifi May 28 00:47:28 better get something lantiq danube based May 28 00:47:29 ;) May 28 00:47:33 mostly the 1350A, which still has no driver May 28 00:48:03 there are no danube devices in the wild in the usa AFAICT May 28 00:48:20 odd May 28 00:50:20 i got my hands on a fonera 2.0n too May 28 00:52:01 build #26 of s3c24xx is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/s3c24xx/builds/26 May 28 00:52:26 jow * r31931 /trunk/package/uhttpd/ (18 files in 3 dirs): (log message trimmed) May 28 00:52:26 [package] uhttpd: May 28 00:52:26 - rewrite large parts of the server, use uloop event driven structure May 28 00:52:26 - support concurrent requests and make the upper limit configurable May 28 00:52:26 - implement initial version of HTTP-to-ubus JSON proxy and session.* namespace May 28 00:52:27 - add compile time support for debug information May 28 00:52:28 - code style changes May 28 00:52:28 jow * r31932 /trunk/package/iwinfo/src/ (iwinfo_lua.c iwinfo_nl80211.c): [package] iwinfo: assume that no tx power information is available if nl80211 returns 0 dBm maximum May 28 00:53:26 btw, why was the whiterussian fakeidentd pkg dropped for kamikaze? May 28 00:54:00 most likely because it didn't build or had a similar problem May 28 00:54:21 no idea. all knowledge about whiterussian stuff was kicked out of my brain by the garbage collector ;) May 28 00:55:03 sometimes i wish my brain had that garbage collector like most other people seem to May 28 00:55:15 jow_laptop: still works fine built manually May 28 00:55:34 DonkeyHotei: it doesn't come preinstalled, you have to put it in yourself ;) May 28 00:56:21 DonkeyHotei: then it was probably forgotton at some point when restructuring the repository May 28 00:56:49 and given that you're the first one I've seen in three years asking for it I assume its not terribly popular either May 28 00:57:08 i was likely the one who asked three years ago May 28 01:04:00 yes, it got moved away when pre-buildroot-ng sources gotr retired May 28 01:04:09 then it never came back May 28 01:05:33 it's a nice li'l proggy May 28 02:55:52 nbd * r31933 /trunk/package/libubox/Makefile: libubox: update to 2012-05-26 May 28 02:56:00 nbd * r31934 /trunk/target/linux/ar71xx/files/drivers/net/ethernet/atheros/ag71xx/ (ag71xx.h ag71xx_main.c): ar71xx: improve rx performance of the ethernet driver by using build_skb to deliver a cache-hot skb to the network stack May 28 02:57:19 gosh so much work ;-) **** ENDING LOGGING AT Mon May 28 02:59:58 2012