**** BEGIN LOGGING AT Thu Jun 09 02:59:57 2011 Jun 09 06:46:10 I am compiling a file for mini 2440 through open emmbedded , my arguements for mini 2440 are as follows Jun 09 06:46:47 arm-angstrom-linux-gnueabi-g++ -march=armv4t -mtune=arm920t -mthumb-interwork -mno-thumb -shared.................. Jun 09 06:47:36 are arguements correct .bcoz it gives seg fault on mini 2440 board Jun 09 07:14:36 git.openembedded.org down? Jun 09 07:15:22 cgit as well as git:// timeouts here Jun 09 07:18:42 ka6sox-away: ka6sox-work: moin, ^ Jun 09 07:20:00 really? Jun 09 07:20:02 let me check. Jun 09 07:21:32 maybe load too high again.. because now oe-core-contrib fetched ok Jun 09 07:23:06 kernel bug report...looking Jun 09 07:24:13 [144000.061214] INFO: task apache2:1537 blocked for more than 120 seconds. Jun 09 07:24:13 [144000.061222] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. Jun 09 07:33:20 JaMa|Wrk, rebooting...give us a second. Jun 09 07:36:36 JaMa|Wrk, could you please test? Jun 09 07:39:16 ka6sox-farfarawa: works now, thanks Jun 09 07:39:37 ya, it locked up...will do a post-mortem in the morning Jun 09 07:39:59 gnight :) Jun 09 08:00:09 -) Jun 09 09:53:09 while compiling Xephyr I m getting ..... hostx.c:(.text+0x114): undefined reference to `XSetErrorHandler' ............... Jun 09 09:53:11 ................... Jun 09 09:53:23 can anyone help me why is it so? Jun 09 10:06:29 hard to say. the missing XSetErrorHandler() would suggest you're not linking against libX11. those dots are rather more of a mystery. Jun 09 10:08:45 pb_, http://pastebin.mozilla.org/1246163 ...... Jun 09 10:09:01 Quick question: What recipe provide v4l(2) libs and headers? Jun 09 10:10:07 I think v4l is a support provided by the kernel.... Jun 09 10:10:09 v4l2apps maybe? dunno, I guess find and grep would be able to tell you Jun 09 10:10:32 Yeah, I thought so too Jun 09 10:10:55 I don't know if v4l2apps will provide it, but it's worth a shot Jun 09 10:11:45 At least they need to have dependency on v4l libs that I could check out Jun 09 10:11:56 which apps you want a v4l2 support... Jun 09 10:12:18 I'm developing my own application which is using a webcam Jun 09 10:12:35 I intend to use this application on an embedded system Jun 09 10:12:43 For that, I need v4l2 Jun 09 10:12:48 just give a v4l support in kernel... it vll work.... Jun 09 10:14:00 Hehe Guess what I found in v4l-utils? 'PROVIDES = "libv4l"' Jun 09 10:14:06 heh Jun 09 10:14:09 risca, I had already configured webcam on ARM development board Jun 09 10:14:49 sanket, And all you did was enable v4l in kernel? Jun 09 10:16:13 its a long back.... but surely I had given v4l2 support for that... and use mplayer commands to run my webcam... Jun 09 10:17:30 My application is going to capture frames and process them in a DSP so mplayer is not sufficient Jun 09 10:17:55 I need to have libv4l to be able to manage the webcam myself Jun 09 10:18:28 I do however need v4l2 enabled in the kernel to be able to use the cam at all Jun 09 10:21:26 risca: yes you need v4l2 enabled in the kernel Jun 09 10:24:04 I don't think you strictly _need_ libv4l. There's no magic in there: anything that it does, you could do for yourself. Jun 09 10:24:28 and, iirc, it doesn't actually do a whole lot Jun 09 10:26:21 To speed up the development process I figured it would be easier to just install libv4l and compile my app for the dev. board Jun 09 10:26:35 It is already working on my host Jun 09 10:26:43 ah, fair enough Jun 09 11:43:45 execve("./connMan.bin", ["./connMan.bin"], [/* 21 vars */]) = 0 Jun 09 11:43:45 --- SIGSEGV (Segmentation fault) @ 0 (0) --- Jun 09 11:43:45 +++ killed by SIGSEGV +++ Jun 09 11:43:45 Segmentation fault Jun 09 11:44:04 can any suggest whats happening Jun 09 11:44:10 looks like a segmentation fault Jun 09 11:44:20 i know that Jun 09 13:00:25 i am compling a file called connMan.bin for mini2440 .through openembedded it does not give error on the host machine but it gives on the target board ,it runs well on thinkclients , i am cross complilng it for arm Jun 09 13:54:34 03Richard Purdie  07master * rcd80d7f201 10bitbake.git/lib/bb/persist_data.py: Jun 09 13:54:34 persist_data: Add back code to retry in the case of locked database errors Jun 09 13:54:34 Signed-off-by: Richard Purdie Jun 09 13:54:45 03Richard Purdie  07master * r3c3bd0c2fa 10bitbake.git/ (4 files in 4 dirs): Jun 09 13:54:45 Drop psyco support Jun 09 13:54:45 At best it gave 1-2% improvement now, its 32 bit x86 only and isn't Jun 09 13:54:45 supported after python 2.6. Jun 09 13:54:45 PyPy is probably a better option now. Jun 09 13:54:45 Signed-off-by: Richard Purdie Jun 09 14:47:00 having trouble with boost recipe - seems to have generated a recursive dependency. boost on libboost-filesystem1 and vice-versa Jun 09 14:51:32 pb_: hello, around? Jun 09 14:54:57 ant_work: yo Jun 09 14:55:07 hi Jun 09 14:55:21 I see you're playing with kexec and initramfs :) Jun 09 14:56:08 anyone seen this error before: Jun 09 14:56:13 now, I've seen kexec does not depend on kernel headers Jun 09 14:56:15 git ls-remote http:// master failed with signal 128, output: Jun 09 14:56:15 fatal: Unable to find remote helper for 'http Jun 09 14:56:41 pb_: dis you try kexec-tools-klibc from oe-dev? Jun 09 14:57:20 and about this: can we make klibc not machine-specific? Jun 09 14:57:29 any ideas how this came about? http://pastebin.com/XwEcCj8L <- apparently recursive depends for boost packages Jun 09 14:57:47 (following your previous uclibc commit) Jun 09 14:59:25 pb_: I found teh initial reasosn for +PACKAGE_ARCH = "${MACHINE_ARCH}" here : http://lists.linuxtogo.org/pipermail/openembedded-commits/2007-November/010370.html Jun 09 14:59:40 * ant_work today cannot type... Jun 09 15:01:26 so, in case anyone is tracking my question, the git version that is built out of OE appears to be missing the http capability, but I'm not sure why. Jun 09 15:03:43 ant_work: no, I haven't tried klibc, I think I will probably stick with regular libc for my initramfs Jun 09 15:03:50 it's odd because this was working for the initial compile. Jun 09 15:03:58 I don't really have any space constraints, it just needs to fit on a 2GB SD :-) Jun 09 15:04:30 comfortably Jun 09 15:05:02 mm, that klibc commit about PACKAGE_ARCH sounds a bit bogus to me Jun 09 15:05:24 who is "pfalcon"? Jun 09 15:06:26 "Paul Sokolovsky" : (["pfalcon@openembedded.org" Jun 09 15:06:50 was haret/ipaq maintainer iirc Jun 09 15:07:15 ah yeah Jun 09 15:07:32 unfortunately gone :/ Jun 09 15:09:06 well, if the klibc arch is causing you a problem, I would be inclined to just revert that patch and see what happens Jun 09 15:09:31 "looks at kernel headers" is not generally a good enough reason to make something machine specific Jun 09 15:09:35 no issues, I just aim to a common arm binary Jun 09 15:09:56 and in the years klibc build process has changed Jun 09 15:10:42 what the recipe does is ln -sf ${STAGING_KERNEL_DIR} linux Jun 09 15:11:47 [pid 633] execve("@DATADIR@/udhcpc/default.script", ["@DATADIR@/udhcpc/default.script", "bound"], [/* 15 vars */]) = -1 ENOENT (No such file or directory) Jun 09 15:11:49 heh Jun 09 15:11:53 sux0r Jun 09 15:12:02 no wonder my dhcp is not working very well Jun 09 15:12:17 lovely Jun 09 15:15:20 pb_: I'll try to remove that MACHINE_ARCH. Further step will be test-removing ARM_INSTRUCTION_SET = "arm", ther esince years... Jun 09 15:15:29 I'll let you know, thx Jun 09 15:20:31 Hello. I need some software (such us usbmount in debian) that automount usb flash when it inserted. Is oe support it? Jun 09 15:22:14 or I need solution to easy way to automout usb mass storage Jun 09 15:22:45 what's wrong with udev? Jun 09 15:23:21 anyone know of a specific issue with doing ASSUME_PROVIDED for git? Jun 09 15:23:28 just a little question, i would like to know if there is a way to generate an sdk with oe-core ? Jun 09 15:23:36 i find this doc http://docs.openembedded.org/usermanual/html/commonuse_qte_sdk.html Jun 09 15:24:06 but i would like to know if it's working with oe-core Jun 09 15:24:22 ant_work: righto Jun 09 15:27:12 darknighte: haven't had any problems with it, assuming you mean git-native Jun 09 15:27:52 morning kergoth Jun 09 15:29:24 hey pb_ Jun 09 15:32:06 morning kergoth. I hadn't had any problems with it until this morning. Jun 09 15:33:22 kergoth: suddenly, I'm getting errors from the git-native in "sysroots/i686-linux/usr/i686/bin" Jun 09 15:34:07 kergoth: I did do a rebuild of several host libraries, so it's always possible that something in that is causing it to barf. Jun 09 15:34:39 kergoth: so, for now, I've purged my tmp folder and am starting from scratch. Jun 09 15:34:47 * darknighte sighs and looks at watch Jun 09 15:35:41 * darknighte remembers he doesn't wear a watch anymore, sighs again, and checks cellphone. Jun 09 15:35:44 this may take a bit. Jun 09 15:36:30 darknighte: I had same issues (dismissed watch too;) Jun 09 15:36:46 rebuilding is faster than debugging such cases ;) Jun 09 15:37:19 fingers crossed Jun 09 15:37:21 bbl Jun 09 15:37:30 ant_work: yeah, I got stubborn and spent an hour on it and thanks. Jun 09 15:39:15 kergoth: as a fallback, I may use ASSUME-PROVIDED. Jun 09 15:49:32 hm, these lmbench results are a bit depressing. 620ns for a null syscall on 600MHz arm11. Jun 09 15:53:46 what kernel? Jun 09 15:53:54 2.6.35 I think Jun 09 15:54:05 Linux (none) 2.6.35.7+ #5 PREEMPT Thu Jun 9 16:34:28 BST 2011 armv6l GNU/Linux Jun 09 15:54:48 Any chance you're using a kernel that's git bisect'able back to say .32, for your machine? Jun 09 15:56:06 not easily. I'm not sure if 2.6.32 ran on this hardware or not but I suspect the answer is probably not. Jun 09 15:56:14 do you think the performance got worse since .32? Jun 09 15:57:09 I've got reports from other folks that a hacked up .32 is fine and a less hacked up .37 is not, on armv7l Jun 09 15:57:32 and for that platform there's not a unified tree that goes back that far Jun 09 15:58:02 hm,right Jun 09 15:58:07 and my gut is willing to go with a 'maybe' on it being the same thing Jun 09 15:58:10 do you know what they consider "fine" and "not fine"? Jun 09 15:58:35 300-400% decrease Jun 09 16:00:08 that does sound bad. measured how? Jun 09 16:00:44 I'll have a poke around tomorrow and see how far back this tree goes and what the oldest kernel is that I can boot Jun 09 16:01:38 lmbench Jun 09 16:01:51 right, same thing I am using Jun 09 16:02:11 (don't have direct details, not folks I directly work with) Jun 09 16:06:07 I wonder if turning off PREEMPT would help. Jun 09 16:06:12 might try that tomorrow as well Jun 09 16:07:36 * pb_ go home now Jun 09 16:27:17 pb_: pfalcon is available on irc at #linaro Jun 09 17:02:14 JaMa|Off, any more git issues? Jun 09 17:11:58 03Paul Eggleton  07master * ra71183762e 10bitbake.git/bin/bitbake-layers: Jun 09 17:11:58 bitbake-layers: handle skipped recipes Jun 09 17:11:58 Report bbappends correctly for skipped recipes instead of reporting the Jun 09 17:11:58 bbappends as not having any matching recipe. Jun 09 17:11:58 (From Poky rev: 39d72ecf6bc1da85aefb11e6481719185f0cb953) Jun 09 17:11:58 Signed-off-by: Paul Eggleton Jun 09 17:11:58 Signed-off-by: Richard Purdie Jun 09 17:11:59 03Paul Eggleton  07master * r7d2363f353 10bitbake.git/lib/bb/ (cache.py cooker.py parse/ast.py): (log message trimmed) Jun 09 17:12:00 track skipped packages Jun 09 17:12:00 Add skiplist to cooker that allows access to the list of packages skipped Jun 09 17:12:01 via SkipPackage (this includes COMPATIBLE_MACHINE, INCOMPATIBLE_LICENSE, Jun 09 17:12:01 etc.) This can be used to enhance error reporting. Jun 09 17:12:02 (From Poky rev: 6c12b7b1099c77b87d4431d55e949cf7c5f52ded) Jun 09 17:12:02 Signed-off-by: Paul Eggleton Jun 09 17:18:52 hi, i have built the micro image and unpackeed it onto rootfs of joggler, it boots but throws a load of errors whislt doing so (mostly about read-only filesystem) and most things dont work, i.e. cant bring up network, and even dmesg has some error about missing device. Do i simply need to try booting with 'rw' in the kernel cmdline or is this a bad idea? Jun 09 19:07:01 pb_, is the gnu.org email address good for you Jun 09 19:32:18 Crofton: I have a fix for the libstdc++.so.6.0.14-gdb.py issue Jun 09 19:32:35 Crofton: I posted it to ml please test it in your scenario Jun 09 19:32:45 03Ludovic Desroches  07master * r0052931ada 10openembedded.git/recipes/tiobench/ (tiobench-0.3.3/tiobench-makefile.patch tiobench_0.3.3.bb): Jun 09 19:32:45 tiobench: Add version 0.3.3 (initial recipe) Jun 09 19:32:45 Build tested and run tested with task-base image with the tiobench package on atmel boards by Ludovic Desroches and Jun 09 19:32:45 Build tested using `angstrom-2010.x` and `minimal{,-uclibc}` for `MACHINE = "beagleboard"` by Paul Menzel. Jun 09 19:32:46 Signed-off-by: Ludovic Desroches Jun 09 19:32:46 Signed-off-by: Paul Menzel Jun 09 19:32:50 and add acks if it fixes your problem Jun 09 19:33:32 mangomake: yes try rw Jun 09 19:48:00 khem, thanks Jun 09 19:48:17 I am in MV atm, but will try to fit in testing of this ASAP Jun 09 19:49:01 do you think it will apply direct to the release branch? Jun 09 20:04:31 MV ? Jun 09 20:04:38 I think it should Jun 09 20:22:44 03James Limbouris  07master * r9a799fb346 10openembedded.git/recipes/initscripts/initscripts-1.0/populate-volatile.sh: (log message trimmed) Jun 09 20:22:44 initscripts: Modify 'populate-volatile.sh' to atomically update 'volatile.cache' Jun 09 20:22:44 When power is lost before a new volatile.cache has been flushed to disk, it can Jun 09 20:22:44 be corrupted, leaving the volatiles uncreated and causing software failure Jun 09 20:22:44 during the next reboot. Jun 09 20:22:44 Since only the existence of volative.cache is checked for using a temporary file Jun 09 20:22:45 on the same volume and renaming it avoids this issue. Jun 09 20:47:58 khem: thx, thats does seem to do the trick Jun 09 21:01:27 03Klaus Kurzmann  07master * r45edf62129 10openembedded.git/recipes/linux/ (2 files in 2 dirs): Jun 09 21:01:27 linux-openmoko_2.6.37.bb: allow setting DAI mode even while pcm is active Jun 09 21:01:27 Without this patch the DAI mode can't be set while a pcm stream is Jun 09 21:01:27 running as that might lead to kernel freezing. This patch allows setting Jun 09 21:01:27 it to the same value to make scenario loading work. Jun 09 21:01:27 Signed-off-by: Klaus Kurzmann Jun 09 21:19:36 03Scott Garman  07master * rd000493c09 10bitbake.git/lib/bb/cache.py: Jun 09 21:19:36 cache.py: fix bitbake -s command Jun 09 21:19:36 This uses the correct index of self.pn when setting up cachedata's Jun 09 21:19:36 pkg_pn, fixing the output of bitbake -s. Jun 09 21:19:36 This fixes bug [YOCTO #1149]. Jun 09 21:19:36 Signed-off-by: Scott Garman Jun 09 21:19:36 Signed-off-by: Richard Purdie Jun 09 21:43:28 03Richard Purdie  07master * r899d45b900 10bitbake.git/lib/bb/data_smart.py: (log message trimmed) Jun 09 21:43:28 bitbake/data_smart: Change overrides behaviour to remove expanded variables from the datastore Jun 09 21:43:28 Currently if you do: Jun 09 21:43:28 OVERRIDES = "z" Jun 09 21:43:28 DEPENDS_prepend = "a " Jun 09 21:43:28 DEPENDS = "b" Jun 09 21:43:28 DEPENDS_z = "c" Jun 09 21:43:29 03Richard Purdie  07master * ra981df3cc9 10bitbake.git/lib/bb/ (event.py parse/ast.py ui/knotty.py): Jun 09 21:43:30 bitbake/event/ast: Add RecipePreFinalise event Jun 09 21:43:30 One of the implications is we need to register the event handlers before Jun 09 21:43:31 executing the anonymous python functions. I can't find any issue with making Jun 09 21:43:31 that change in any existing metadata use cases. Jun 09 21:43:32 Signed-off-by: Richard Purdie Jun 09 21:43:32 03Richard Purdie  07master * rc3c2ad6f22 10bitbake.git/ (bin/bitbake lib/bb/__init__.py): Jun 09 21:43:33 Update version to 1.13.1 Jun 09 21:43:33 Signed-off-by: Richard Purdie Jun 09 22:21:30 ka6sox-away: don't know about any now Jun 09 22:56:24 hi Marex **** ENDING LOGGING AT Fri Jun 10 02:59:57 2011