**** BEGIN LOGGING AT Tue Oct 30 02:59:58 2012
Oct 30 07:14:24 <OpenWrt-Commits> cyrus r34008 trunk/package/network/ (5 files in 3 dirs) * ipv6: add 6relayd (RD-, DHCPv6- & NDP-Proxy)
Oct 30 07:43:23 <OpenWrt-Commits> cyrus r34009 trunk/package/network/ipv6/6relayd/Makefile * 6relayd: update to work around a raw-socket bug
Oct 30 08:51:34 <russell2> alix2 failing to boot?
Oct 30 08:51:41 <russell2> r34002
Oct 30 08:51:59 <russell2> get grub2, soon as it boots, nothing on console
Oct 30 08:52:40 * russell2 tried make toolchain/clean ; make ... same thing
Oct 30 09:06:53 <russell2> r33504 worked, r33760 (and more recent) doesn't ... trying from a clean tree
Oct 30 09:13:38 <dape8708> guys, latest trunk breaks at this -> http://pastebin.com/p9ZitP1z
Oct 30 09:13:53 <dape8708> something with libubox if i read correctly
Oct 30 13:13:41 <OpenWrt-Commits> jow r34010 trunk/package/network/services/uhttpd/Makefile * uhttpd: update to latest git head, improves handling of invalid requests
Oct 30 13:18:34 <tripolar> hi
Oct 30 13:18:56 <tripolar> i tried to port brcm47xx to 3.6 but i get error like this when i try to compile the kernel
Oct 30 13:19:00 <tripolar> drivers/mtd/bcm47xxpart.c:114:10: error: 'struct mtd_info' has no member named 'read'
Oct 30 13:19:40 <KanjiMonster> tripolar: use mtd_read(foo) instead of mtd->read(foo)
Oct 30 13:20:08 <tripolar> ahh i tried mtd->mtd_read :(
Oct 30 13:20:10 <tripolar> thanks
Oct 30 13:21:22 <KanjiMonster> tripolar: read was renamed to _read to force build errors so users switch to the wrapper methods and don't use the raw function pointers anymore ;)
Oct 30 13:22:20 <OpenWrt-Commits> jow r34011 branches/attitude_adjustment/ package/uhttpd/Makefile package/uhttpd/files/uhttpd.init package/uhttpd/src * AA: uhttpd: backport trunk changes
Oct 30 13:27:54 <tonikasch> do you know how to restore last 64 bits on uboot from dump through serial (arv4518pw)?
Oct 30 13:28:33 <blogic> did you break it ?
Oct 30 13:31:25 <swalker> KanjiMonster: speaking of mtd_read, isn't generic/patches-3.6/401-partial_eraseblock_write.patch incorrectly using it given 994c8409c4591983898182bfb5a0368d3daf1fc1?
Oct 30 13:32:20 <nbd> swalker: that is core code
Oct 30 13:32:23 <tonikasch> blogic, i guess yes, but now i have last 64k backup
Oct 30 13:32:25 <nbd> swalker: it's allowed to use the callbacks directly
Oct 30 13:40:11 <tripolar> KanjiMonster but what do i do with - if (mtd->unlock)
Oct 30 13:40:21 <tripolar>  if (mtd_unlock())
Oct 30 13:40:54 <KanjiMonster> tripolar: i think mtd_unlock() already checks that mtd->_unlock is populated before calling
Oct 30 13:41:20 <tripolar> KanjiMonster do you know where mtd_unlock is defined?
Oct 30 13:42:15 <KanjiMonster> tripolar: http://lxr.linux.no/#linux+v3.6.4/drivers/mtd/mtdcore.c#L947
Oct 30 13:42:24 <tripolar> thanks grep just found it ;)
Oct 30 14:40:17 <tripolar> KanjiMonster: but what would be the best solution for mtd->erase = sflash_mtd_erase; ? mtd->_erase = sflash_mtd_erase ?
Oct 30 14:41:47 <tripolar> its used in 052-mtd-add-serial-flash-driver.patch
Oct 30 14:58:30 <OpenWrt-Commits> juhosg r34012 trunk/target/linux/ (35 files in 3 dirs) * generic: fold yaffs_cvs_2009_04_24 patch to generic/files
Oct 30 14:58:34 <OpenWrt-Commits> juhosg r34013 trunk/target/linux/ (53 files in 3 dirs) * generic: fold yaffs_git_2010_10_20 patch to generic/files
Oct 30 15:59:13 <KanjiMonster> tripolar: yes. filling those pointers is fine, just not calling them directly from non-core code
Oct 30 16:16:54 <raorn> oh. here's the dev channel
Oct 30 16:17:47 <raorn> uboot-envtools fails to compile because of --strip-components
Oct 30 16:17:56 <raorn> looks like path stripping occurs before processing non-option arguments
Oct 30 16:18:15 <raorn> tar xvjf $(DL_DIR)/$(PKG_SOURCE) --strip-components=2 -C $(PKG_BUILD_DIR) $(PKG_DISTNAME)-$(PKG_VERSION)/lib/crc32.c
Oct 30 16:18:29 <raorn> says, tar: u-boot-2012.04.01/lib/crc32.c: Not found in archive
Oct 30 16:18:50 <raorn> and extracting just crc32.c works
Oct 30 16:18:58 <raorn> do i need uboot-envtools for wndr3700 at all?
Oct 30 16:56:21 <OpenWrt-Commits> florian r34014 trunk/include/kernel-defaults.mk * [buildroot] properly reindent Kernel/Prepare/Default
Oct 30 16:56:27 <OpenWrt-Commits> florian r34015 trunk/target/linux/realview/config-3.3 * [realview] set earlyprintk UART to be the default Realview UART
Oct 30 16:58:15 <OpenWrt-Commits> florian r34016 packages/ (5 files in 5 dirs) * [packages] make packages depending on uclibc++ depend on CXX_DEPENDS
Oct 30 17:16:12 <tripolar> i did this in netfilter.mk because the naming changed in 3.6
Oct 30 17:16:12 <tripolar> -$(eval $(call nf_add,IPT_CORE,CONFIG_IP_NF_TARGET_LOG, $(P_V4)ipt_LOG))
Oct 30 17:16:13 <tripolar> +$(eval $(call nf_add,IPT_CORE,CONFIG_NETFILTER_XT_TARGET_LOG, $(P_XT)xt_LOG))
Oct 30 17:16:19 <tripolar> but the option is not enabled
Oct 30 17:16:51 <tripolar> as module when i enable it manually in generic/config-3.6 it works and the module is included in ipt-core - what could be the problem?
Oct 30 17:18:39 <raorn> is it possible to "try" new firmware without reflashing? i almost sure that there was howto on "live images" but can't find it anymore...
Oct 30 17:19:01 <tripolar> @KanjiMonster
Oct 30 18:36:04 <KanjiMonster> tripolar: no idea
Oct 30 18:53:26 <OpenWrt-Commits> juhosg r34017 branches/attitude_adjustment/target/ (34 files in 2 dirs) * AA: generic: fold yaffs_cvs_2009_04_24 patch to generic/files
Oct 30 18:53:30 <OpenWrt-Commits> juhosg r34018 branches/attitude_adjustment/target/ (52 files in 2 dirs) * AA: generic: fold yaffs_git_2010_10_20 patch to generic/files
Oct 30 21:16:39 <OpenWrt-Commits> juhosg r34019 trunk/target/linux/ (24 files in 2 dirs) * generic: rename yaffs patches
Oct 30 21:16:40 <OpenWrt-Commits> juhosg r34020 trunk/target/linux/ generic/patches-3.6/501-yaffs-Fix-directory-unlinking-in-yaffs1-mode.patch generic/patches-3.3/501-yaffs-Fix-directory-unlinking-in-yaffs1-mode.patch generic/patches-3.3/501-yaffs-fix-symlink-bug.patch generic/patches-3.6/501-yaffs-fix-symlink-bug.patch * generic: replace yaffs symlink fix patch
Oct 30 21:16:41 <OpenWrt-Commits> juhosg r34021 trunk/target/linux/ (6 files in 2 dirs) * generic: replace yaffs mutex_fix patch
Oct 30 21:50:24 <OpenWrt-Commits> nbd r34022 trunk/package/libs/libubox/Makefile * libubox: add build dependency on lua
Oct 30 21:57:04 <dape> thx nbd !
Oct 30 21:57:16 <dape> broken since this morning i think
Oct 30 22:20:39 <russell--> reconfirm'd alix2 is broken
Oct 30 22:20:47 <russell--> building from a fresh checkout
Oct 30 22:21:38 <nbd> russell--: details?
Oct 30 22:21:53 <russell--> r34013 is what i just confirmed
Oct 30 22:21:56 <russell--> i get grub
Oct 30 22:22:00 <russell--> or grub2
Oct 30 22:22:17 <russell--> but then, screen blanks, and nothing more on the console
Oct 30 22:22:58 <russell--> going back through my previous builds (that had been untested), i find r33504 was still working, r33760 is not
Oct 30 22:28:34 <russell--> Booting `OpenWrt' two newlines and a blinking cursor, right after a screen blank, is the last thing I see in r34013
Oct 30 22:29:21 <nbd> are you using the alix2 subtaret?
Oct 30 22:29:25 <russell--> yes
Oct 30 22:31:30 <nbd> what's the kernel command line that it uses (when you edit it from within grub2)
Oct 30 22:32:16 <russell--> this is the scripts/diffconfig.sh output (but i saw the same thing from a vanilla build) http://sprunge.us/fZAA
Oct 30 22:32:20 * russell-- looks
Oct 30 22:34:08 <russell--> setparams 'OpenWrt' and then:          linux /boot/vmlinuz block2mtd.block2mtd=/dev/sda2,65536,rootfs root=/dev/mtdblock0 rootfstype=squashfs rootwait console=ttyS0,38400n8 noinitrd
Oct 30 22:34:37 <nbd> ok, and the alix2 uses vga or serial?
Oct 30 22:34:44 <russell--> serial
Oct 30 22:34:48 <nbd> hm
Oct 30 22:34:49 <nbd> ok
Oct 30 22:35:13 <russell--> i see the bios and grub stuff as i am accustomed to
Oct 30 22:35:21 <nbd> have you tried increasing the kernel partition size?
Oct 30 22:35:24 <nbd> via menuconfig
Oct 30 22:35:31 <russell--> no, i haven't
Oct 30 22:35:35 <nbd> ok, try that
Oct 30 22:35:39 <nbd> maybe grub2 needs more space or something
Oct 30 22:37:09 <russell--> mounting the CF on my build machine is telling me that there is 2.5MB free on the boot partition
Oct 30 22:37:16 <nbd> hm, ok
Oct 30 22:39:03 <russell--> it is ext4, which might be new?  but it seemingly can read the /boot/grub/grub.cfg
Oct 30 22:40:12 <nbd> i think the boot partition is just ext2
Oct 30 22:40:56 <russell--> /dev/sdb1 on /media/disk type ext2  ... yes, you are right
Oct 30 22:41:09 <russell--> ubuntu was lying to me
Oct 30 22:43:04 * russell-- can start bisecting this, since i've got some brackets on it, but if there is anything suspicious between 33504 and 33760, maybe i can speed things up?
Oct 30 22:43:16 <nbd> haven't seen anything
Oct 30 22:43:28 <russell--> ok, i'll get started then :-)
Oct 30 22:50:04 <russell--> btw, nbd.name git clone/pulls have been slow recently, i was seeing 12-16 kiB/s earlier today
Oct 30 22:50:15 <nbd> odd, it's very fast for me
Oct 30 22:51:27 <russell--> i noticed it when i was cloning a new tree to be extra sure, maybe hurricane related ;-)
Oct 30 22:53:19 <russell--> yeah, much faster now
Oct 30 23:25:16 <russell--> 17 minute build, with minimal config
Oct 30 23:27:02 <tonikasch> perhaps you could have used make -j 3 if you have a dual-core cpu
Oct 30 23:27:37 <tonikasch> but i haven't measured if it lasts less time, i pretend it does
Oct 30 23:29:40 <russell--> tonikasch: that's with -j8 on an 8 core xeon 3.x ghz
Oct 30 23:30:10 <russell--> from a distclean
Oct 30 23:40:17 <tonikasch> wow
Oct 30 23:40:32 <tonikasch> :)
Oct 30 23:43:03 <russell--> it's also nice -n 19'd
Oct 30 23:45:01 <tonikasch> aah
Oct 30 23:48:22 <russell--> r33638 is okay
Oct 30 23:56:01 <tonikasch> is that 12.09 beta2?
Oct 30 23:56:46 <tonikasch> nope... it's a previous revision
Oct 31 00:04:20 <russell--> r33703 is bad
Oct 31 00:04:38 <russell--> tonikasch: i'm bisecting a problem with alix2
Oct 31 00:07:50 <nbd> were you testing the old revs with grub2?
Oct 31 00:24:04 <OpenWrt-Commits> nbd r34023 trunk/include/toplevel.mk
Oct 31 00:24:04 <OpenWrt-Commits> build: increase file descriptor count limit for the build, some systems (e.g. Mac OS X default to 256, which is too little for some parallel builds)
Oct 31 00:24:08 <OpenWrt-Commits> nbd r34024 trunk/toolchain/eglibc/config/Config.in * eglibc: disable NIS/SUNRPC by default. RPC support is provided by librpc
Oct 31 00:24:12 <OpenWrt-Commits> nbd r34025 trunk/toolchain/eglibc/Makefile * eglibc: enable parallel builds
Oct 31 00:24:18 <OpenWrt-Commits> nbd r34026 trunk/toolchain/ eglibc/include eglibc/patches/2.15/120-use_host_cflags.patch eglibc/common.mk eglibc/include/libintl.h
Oct 31 00:24:18 <OpenWrt-Commits> eglibc: add a fake libintl.h for cross-rpcgen and cross-zic to fix build on mac os x
Oct 31 00:25:41 * russell-- starts to suspect grub2 ... r33670 bad too
Oct 31 00:52:35 <russell--> nbd: r33641 is bad too, so the problem appears between r33638 and r33641
Oct 31 00:53:26 <nbd> did you use grub2 in r33638?
Oct 31 00:54:50 <russell--> i used whatever defconfig gave me
Oct 31 00:55:31 <russell--> where the starting .config basically just selected the target/subtarget and download directory
Oct 31 00:56:41 <nbd> ok, so the issue is probably grub2
Oct 31 00:57:13 <russell--> //sprunge.us/UQOd
Oct 31 00:57:35 <russell--> http://sprunge.us/UQOd
Oct 31 00:57:53 <russell--> grub2 is working on net45xx
Oct 31 01:04:44 <russell--> yeah, confirm r33638 was CONFIG_PACKAGE_grub=y, grub2 is not set
Oct 31 01:05:55 <nbd> no idea why grub2 isn't working on the alix2
Oct 31 01:05:59 <nbd> and unfortunately i don't have such a device
Oct 31 01:06:32 <nbd> couldn't find any useful information on the net either
Oct 31 01:06:44 <russell--> dude makes them in zurich :-)
Oct 31 01:07:25 * russell-- looking for clues in build log files
Oct 31 01:08:13 <nbd> try building padded images and see what happens if you boot the image in qemu
Oct 31 01:08:40 <nbd> in my test, the alix2 kernel didn't boot, but it did print some lines and it didn't hang while still in grub2
Oct 31 01:08:56 <nbd> that way you can check if it's a grub2 compat issue or if your build got messed up somehow
Oct 31 01:09:07 <nbd> if it boots partially in qemu, there's no need to check the build log
Oct 31 01:15:03 <russell--> how to boot in qemu? /me hasn't done that before
Oct 31 01:18:20 <nbd> qemu -nographic openwrt-x86-alix2-combined-ext4.img
Oct 31 01:18:22 <nbd> after gunzipping it
Oct 31 01:21:32 <russell--> lol, how do i kill it?
Oct 31 01:21:48 <russell--> it stopped after "boot" appears
Oct 31 01:23:05 <russell--> that is, r33641 hangs after "boot"
Oct 31 01:23:21 <russell--> (killing qemu from another shell worked)
Oct 31 01:24:24 <russell--> oh, wait, that was r33638, the one that works or real hardware
Oct 31 01:27:13 <russell--> on*
Oct 31 01:35:11 <russell--> aha
Oct 31 01:35:21 <russell--> nbd: r33641 boots in qemu
Oct 31 01:37:32 <russell--> the one that didn't boot on real hardware
Oct 31 01:56:32 <nbd> ok, so the build is correct and grub2 needs to be fixed for some alix2 specific issue
Oct 31 02:31:18 <russell--> nbd: sounds like it
**** ENDING LOGGING AT Wed Oct 31 02:59:58 2012