**** BEGIN LOGGING AT Tue Dec 20 02:59:57 2011 Dec 20 03:00:30 CONFIG_BLK_DEV_INITRD=y Dec 20 03:01:09 cmdline is initrd=0x21000000 Dec 20 03:03:01 so initrd support is enabled in kernel Dec 20 03:03:10 yup enabled Dec 20 03:03:27 i read kernel doc Dec 20 03:03:37 it say it must be enabled inorder to pass .cpio.gz Dec 20 03:05:05 initrd should point to cpio.gz image Dec 20 03:05:23 initrd=
does not work Dec 20 03:09:05 ok i will try Dec 20 03:09:15 thanks alot Dec 20 03:13:59 how would the kernel know which address to load if u just provide the filename? Dec 20 04:15:56 any idea why when my kernel boot up there is not " checking if image is initramfs..." Dec 20 04:32:39 any idea why when my kernel boot up there is not " checking if image is initramfs..." Dec 20 06:47:54 any idea why when my kernel boot up there is no " checking if image is initramfs..." Dec 20 07:34:06 How can I build ti hardware codec with gstreamer for beagleboard xm? I have build gstreamer-ti in oe, but there is no Ti codecs plugin found.... Dec 20 07:34:38 any idea why when my kernel boot up there is no " checking if image is initramfs..." Dec 20 08:04:11 good morning Dec 20 08:19:35 any idea why when my kernel boot up there is no " checking if image is initramfs..." Dec 20 10:29:54 When I Compile Bitbake virtual/kernel I get the ERROR: Function 'Fetcher failure for URL: 'file://at91/linux-2.6.30-001-configurable-partition-size.patch.patch'. Pls Help me! Dec 20 10:30:56 hm what help do you want? Dec 20 10:31:21 maybee the command "find" helps you Dec 20 10:31:36 How to fix this error when I compile the kernel Dec 20 10:32:07 maybee you read the error msg again Dec 20 10:38:15 basith if the file linux-2.6.30-001-configurable-partition-size.patch.patch is not in your checkedout dirs Dec 20 10:38:21 it can not be apllied Dec 20 10:41:58 wonglinde: Yes I dont have the file. behind proxy I couldn't download the file linux-2.6.30. So i manually download and put it. Dec 20 10:45:13 I am not talking about the linux-source-tar.bz2 Dec 20 10:45:29 I am talking about the particular patch linux-2.6.30-001-configurable-partition-size.patch.patch Dec 20 10:45:43 you have to put it where it is in original setup Dec 20 10:45:51 no int archives Dec 20 11:03:56 hi all Dec 20 11:04:57 hi pb Dec 20 11:05:49 hi pb_, woglinde, all Dec 20 11:08:20 gm all Dec 20 11:08:45 hi mckoan Dec 20 11:09:32 morning all Dec 20 11:09:53 hi ka6sox Dec 20 11:10:20 almost time for sleep. Dec 20 11:10:48 haha Dec 20 11:10:50 lunch Dec 20 12:02:57 hi Dec 20 12:03:10 i'm having some issues trying to bitbake console-image Dec 20 12:05:04 the output http://www.heypasteit.com/clip/05HX Dec 20 12:16:00 Hello Dec 20 12:18:33 hi otavio Dec 20 12:19:33 any tips? Dec 20 12:20:15 rlrosa thats the problem where -sysroot is not called on g++ and g++ search in wrong dir for headers Dec 20 12:20:26 look at the ml Dec 20 12:20:36 khem posted patches for it Dec 20 12:20:40 oe-core ml Dec 20 12:20:42 When I Compile the kernel I got the error tools/perf directory not found in linux_2.6.30 . When I search about that, It has been from linux_2.6.31. But there is no at91sam9263 patch for linux_2.6.31. what should i do? Dec 20 12:21:27 okay have to go Dec 20 12:21:43 backport the patch Dec 20 12:21:52 or use a more up to date kernel Dec 20 12:21:54 till laterr Dec 20 12:22:22 does anybody know how to point g++ to the correct place? Dec 20 12:41:57 rlrosa: woglinde told you that khem has sent patches for fixing it to ml; please look at it Dec 20 12:42:13 rlrosa: iirc those were fixed in master of oe-core; updating might fix it Dec 20 12:46:08 bluelightning: hi Dec 20 12:46:33 bluelightning: looks like /etc/init.d/blueprobe is missing in rootfs you built for me, is it ok? Dec 20 12:50:55 anarsoul: hi Dec 20 12:51:00 anarsoul: hmm... it really ought to be a dependency; although all we should need is the configuration file that blueprobe outputs Dec 20 12:51:46 AIUI blueprobe is only really needed for devices that might have bluetooth but it is not guaranteed, e.g. some old iPAQs that were effectively offered with and without bluetooth Dec 20 12:51:48 and script to change port parameters (requirement for h1940) Dec 20 12:51:58 port parameters? Dec 20 12:52:01 yep Dec 20 12:52:06 what's the story there? Dec 20 12:53:33 well, it's necessary to change port speed before enabling bt power Dec 20 12:54:05 otherwise hciattach fails (looks like chip hangs) Dec 20 12:56:04 morning Dec 20 12:57:07 bluelightning: so I'm changing port parameters via stty, and then loading rfkill-gpio module Dec 20 12:58:01 (it's not possible to compile rfkill-gpio in kernel, because then it enables bt power immediately, but we're not ready yet) Dec 20 12:58:38 bt chip on h1940 is really weird thing :( Dec 20 12:59:23 hi Dec 20 13:04:47 anarsoul: hmm, I had a bad feeling that hciattach wouldn't be enough for some devices :( Dec 20 13:05:11 anarsoul: I want the init code that's in opie itself to be as generic as possible, avoiding any machine specifics Dec 20 13:05:37 bluelightning: well, no changes in opie is neccesarry (at least for h1940) Dec 20 13:05:50 just a small init script (similar to blueprobe) Dec 20 13:05:51 ok, that's good Dec 20 13:06:47 it just manages rfkill and does hciattach, right? Dec 20 13:07:30 yeah Dec 20 13:07:44 Hi there, is there a way to influence the runtime dependencies of task-core-boot? Dec 20 13:08:29 kenws: VIRTUAL_RUNTIME_foo only Dec 20 13:08:49 JaMa|Wrk: ok, thanks for the hint - I'll have a look Dec 20 13:13:05 anarsoul: btw please let me know if we should add any module recommends for h1940 (e.g. rfkill module) Dec 20 13:13:20 I'm sure I'll discover them once I get around to testing Dec 20 13:13:35 but early warning would also be good :) Dec 20 13:14:55 nope, recommends is not good place for rfkill-gpio; if you plan to resurrect blueprobe script then it's right place Dec 20 13:15:37 (some new var with machine-specific driver name, and modprobe if this var is set) Dec 20 13:15:46 but only after stty Dec 20 13:33:56 Hmm, it looks like the build system somehow cuts the path to the opkg config file and therefore fails at the do_rootfs stage Dec 20 13:34:05 it's line 575 of http://paste.ubuntu.com/776294/ Dec 20 13:34:49 whoops Dec 20 13:36:03 never mind - it's just my screen session over here Dec 20 13:40:37 anarsoul: hmm Dec 20 13:40:49 strange, line 566 of the pastie above still shows that the path to the opkg.conf got cut. Dec 20 13:45:20 calling opkg-cl manually seems to work: http://paste.ubuntu.com/776311 Dec 20 13:48:16 Did anyone encounter such an issue before? Dec 20 13:53:33 (I've double checked the logs and it's not the screen session that fools me as I initially thought) Dec 20 14:02:33 kenws: this is not a path issue; apparently no package provides what another package is requesting Dec 20 14:03:04 kenws: I would suggest bitbake -c cleansstate task-core-boot in this instance Dec 20 14:37:03 i'm having some issues trying to bitbake console-image Dec 20 14:37:06 the output http://www.heypasteit.com/clip/05HX Dec 20 14:37:10 any tips? Dec 20 15:11:35 I read a new thread about fakeroot_1.15.1.bb in ML, any comment? Dec 20 15:12:19 bluelightning: why not? To me it looks like the "opkg-cl install" (line 566) happens before "satisfy_dependencies_for" (line 570 at the pastie) Dec 20 15:15:11 kenws: it's probably just an artifact; you occasionally see that fatal error messages are logged in the middle of log output like that, it's something we ought to fix in bitbake Dec 20 15:15:26 kenws: did you try cleaning task-core-boot? Dec 20 15:16:48 bluelightning: yep, I did run the -ccleanstate task as you suggested and kicked off a rebuilt but it didn't change - same error Dec 20 15:18:59 does "Cannot satisfy the following dependencies for task-core-boot" refer to runtime or build dependencies? Dec 20 15:19:23 I'd have thought it's about runtime dependencies, but I'm not sure Dec 20 15:21:21 I guess I just don't understand what this message really means Dec 20 15:27:12 kenws: this is runtime Dec 20 15:27:25 kenws: it's putting together all the packages to produce the rootfs Dec 20 15:28:50 what's likely happened here is you rebuilt the toolchain after building it and task-core-boot previously, and now its runtime provides are different Dec 20 15:30:29 bluelightning: Ok, thank you for clarifying. Dec 20 15:30:39 although I don't understand why rebuilding task-core-boot would not fix it Dec 20 15:31:26 I'm using an external binary toolchain (based on the CSL recipe I posted last week to the oe-core list) Dec 20 15:31:58 given the error and your explanation, I'd expect the task-core-boot ipk depend on "external-csl-toolchain" Dec 20 15:32:22 but it doesn't (at least `dpkg -I ` doesn't show it) Dec 20 15:33:45 I've made recipy for a package for OpenEmbedded classic that installs some pre-compiled binaries. When I try to build the recipy, I get the following error: "ERROR: QA Issue with scaler: No GNU_HASH in the elf binary" in the do_package step. Is there a way to turn off this QA check so I can build the package anyway? Dec 20 15:36:09 kenws: I wouldn't necessarily trust dpkg to work correctly on ipks, what does opkg-cl say? Dec 20 15:39:54 *checking* Dec 20 15:41:25 dpkg{-deb} -I should work fine on .ipks. I would trust it more than opkg-cl in fact :-} Dec 20 15:41:54 pb_: ok Dec 20 15:42:00 if you search for "external-csl-toolchain (>= arm-2011.03)" in tmp/deploy/*/Packages, do you get any matches? Dec 20 15:42:28 strange, `opkg-cl info ` doesn't print anything Dec 20 15:43:03 I don't think opkg-cl info does quite what you think it does. It's like apt-cache show, it only prints info from the available file. Dec 20 15:43:21 pb_: http://paste.ubuntu.com/776432/ Dec 20 15:43:41 no, that's not quite what I meant Dec 20 15:43:54 tmp/deploy/*/Packages is a set of text files Dec 20 15:44:01 aah Dec 20 15:44:14 if you look at their contents, do you have any lines which mention that external-csl-toolchain? Dec 20 15:44:54 yep Dec 20 15:45:59 the package acl, for example depends on external-csl-toolchain Dec 20 15:46:18 http://paste.ubuntu.com/776436/ Dec 20 15:47:14 so it does Dec 20 15:47:31 well, that doesn't seem quite right. I wonder if bitbake has decided that external-csl-toolchain is providing libgcc_s.so or something. Dec 20 15:49:41 pb_: hm, that could be due to PREFERRED_PROVIDER_libgcc = "external-csl-toolchain" Dec 20 15:50:07 It seemed to work for the CSL toolchain: http://article.gmane.org/gmane.comp.handhelds.openembedded.core/12337 Dec 20 15:51:52 when switching to another binary toolchain (linaro based) I changed these things to "external-linaro-toolchain" of course Dec 20 15:53:13 I think, I need to find out what pulls in the dependency to external-csl-toolchain in my case Dec 20 15:56:37 strange, none of the recipes the build is using contains "external-csl-toolchain" Dec 20 15:56:47 Here is the full build log: http://paste.ubuntu.com/776454/ Dec 20 16:23:10 rlrosa: what version of gcc are you using is it 4.5 or 4.6 Dec 20 16:24:09 khem: that would be gcc (Ubuntu/Linaro 4.5.2-8ubuntu4) 4.5.2, but that's not the one used to cross compile stuff, is it? Dec 20 16:24:19 khem: (that was gcc --version) Dec 20 16:25:07 khem: quick look in the build dir, seems to be 4.5 Dec 20 16:40:57 03Khem Raj  07master * r5337ed8682 10bitbake.git/contrib/vim/syntax/bitbake.vim: Dec 20 16:40:58 vim/syntax: Recognize ?? operator Dec 20 16:40:58 Signed-off-by: Khem Raj Dec 20 16:40:58 Signed-off-by: Richard Purdie Dec 20 16:41:11 03Joshua Lock  07master * r540ba78075 10bitbake.git/lib/bb/ui/crumbs/runningbuild.py: Dec 20 16:41:11 ui/crumbs/runningbuild: handle InvalidTask events Dec 20 16:41:11 The knotty UI just ignores these and so should RunningBuild, if these events Dec 20 16:41:11 aren't handled the UI appears to hang. Dec 20 16:41:11 Fixes [YOCTO #1665] Dec 20 16:41:11 Signed-off-by: Joshua Lock Dec 20 16:41:12 Signed-off-by: Richard Purdie Dec 20 16:41:13 03Joshua Lock  07master * r6c0babf089 10bitbake.git/lib/bb/ui/crumbs/hobprefs.py: (log message trimmed) Dec 20 16:41:24 ui/crumbs/hobprefs: trigger a reparse after changing IMAGE_FSTYPES Dec 20 16:41:24 As reported on the mailing list[1] when changing IMAGE_FSTYPES through the Dec 20 16:41:24 hob preferences a reparse is required before the changes will be picked up Dec 20 16:41:24 by the system. This patch sets the reload_required property of the class to Dec 20 16:41:24 true when the image types have been modified to ensure the reparse is Dec 20 16:41:24 triggered. Dec 20 16:43:08 khem, pong Dec 20 17:03:35 ka6sox-away: any progress on agate and/or garnet Dec 20 17:04:54 rlrosa: the patch I did was for 4.6 Dec 20 17:04:56 not for 4.5 Dec 20 17:04:58 re khm Dec 20 17:05:04 woglinde: hi Dec 20 17:06:25 khem, yes, we are waiting for OSUOSL to hook up their outside ethernet and they will get finished. Dec 20 17:06:28 khem: so i need 4.6 for this to work? where do i choose what version i want to use? Dec 20 17:10:53 anyone cracked open there sheevaplug? Dec 20 17:27:27 khem: I had some issues with latest gcc-4.6 from meta-oe, have you seen them on oe-core ML? Dec 20 17:28:27 mickeyl, hi Dec 20 17:28:29 mickeyl, Dec 07 21:28:34 we should start a buy back program for the older devices Dec 20 17:29:58 ka6sox-away: ok thanks Dec 20 17:30:14 ka6sox-away: any ETA on that > Dec 20 17:30:29 rlrosa: you dont Dec 20 17:30:43 JaMa|Wrk: that error seems unrelated to work I did Dec 20 17:30:54 Khem, unknown...they requested the hookups on Monday, so I wont' even ask till end of day today. Dec 20 17:32:04 khem: and any idea what caused it? Dec 20 17:32:33 khem: it was building fine about a week ago Dec 20 17:37:33 JaMa|Wrk: did gcc change on the build host Dec 20 17:38:28 you mean native one? no Dec 20 17:38:47 yes Dec 20 17:40:58 khem: and it seems to happen mostly for builds from scratch or after cleansstate of toolchain Dec 20 17:41:19 khem: incremental builds suffer only from the last issue (after switching MACHINES) sofar Dec 20 17:44:16 JaMa|Wrk: I did scratch build and incremental builds with oe-core w/o meta-oe patches and it worked well Dec 20 17:44:36 JaMa|Wrk: I dont think this could be something with my patch Dec 20 17:44:56 that's why I expected that there is just something missing in meta-oe Dec 20 17:49:33 khem: Did you get a chance to look at http://thread.gmane.org/gmane.comp.handhelds.openembedded.core/12339 ? Dec 20 17:57:43 kenws: the patches I have glanced over them Dec 20 17:57:55 kenws: someone needs to try them out other than you Dec 20 17:58:07 khem: btw my sstate-cache was empty before starting gcc build Dec 20 17:58:28 khem: ah but not after first machine switch, right Dec 20 17:58:30 khem: ok, I see. The eglibc-locale thingy remains a mystery to me. Dec 20 17:59:33 khem there is no dlvsym support in uclibc right now right? Dec 20 17:59:45 khem: and thanks for looking at it : ) Dec 20 19:31:56 kergoth: OT: you wouldn't know how to add a '.git/config' in a git repository? Dec 20 19:32:45 ? Dec 20 19:33:02 .git/config is part of the repo itself, it exists when you create the repo.. Dec 20 19:35:59 kergoth: sorry. 'a/.git/config' Dec 20 19:36:46 kergoth: I am importing something from mercurial (hg-fast-export) and git chokes on a copy of libxml2 in the source tree. i had hoped you had seen this before Dec 20 19:37:21 you can't import a git repo into another git repo. you'd have to use submodules or something Dec 20 19:38:56 kergoth: do you happen to know where this technical limitation comes from? Dec 20 19:39:24 you can't store an hg repo in an hg repo either, afaik, you have to use subrepos Dec 20 19:39:25 so.. Dec 20 19:39:27 * kergoth shrugs Dec 20 19:43:24 kergoth: thanks. The fun fact is that git-fast-import accepts it, git format-patch can create the patch with a/.git/HEAD... but git am and git checkout-tree refuse it. so looks like an artificial limitation (which is still fine) Dec 20 19:43:44 he zecke Dec 20 20:09:53 woglinde: do u mean dlsym ? Dec 20 20:29:36 dlsym is in Dec 20 20:29:40 dlvsym not Dec 20 22:09:16 Is there no way to specify the SRC_URI and SRCREV for a git repo within a .inc file and then include that file in a bunch of recipes? When attempting to do this I keep on getting do_install errors. It seems like the git repo isn't actually getting downloaded and installed in the WORKDIR. However, cuting the SRC_URI and SRCREV from the .inc file and pasting into the recipe directly works without a problem. Dec 20 22:10:00 that works fine. if it isn't, you have something odd going on, or its user error. bitbake doesn't care what file something is defined in Dec 20 22:10:08 only the order of definition in the case of one var overwriting another Dec 20 22:10:47 if you're using 'include' to pull in the .inc, make sure you use 'require' instead, otherwise it may well not be including your .inc at all, and being quiet about it Dec 20 22:10:51 'include' is optional, 'require' isnt' Dec 20 22:25:47 kergoth, that line of thought is going over poorly Dec 20 22:25:54 people are dropping off :) Dec 20 22:25:59 heh :) Dec 20 22:39:27 kergoth: It seems that it isn't a problem when using SRC_URI = in the include file but using SRC_URI ?= is an issue. Dec 20 22:39:42 fcooper, you're incluidng it in the wrong place. as i said, definition order matters Dec 20 22:39:46 include pulls it in *right there* Dec 20 22:40:04 ?= means 'set only if unset' Dec 20 22:40:10 if it was included after it was already defined, it won't do anything Dec 20 22:41:00 gah Dec 20 22:41:02 | cc1: error: unrecognised debug output level "0-march=i486" Dec 20 22:41:05 anyone ever seen this one? Dec 20 22:41:11 getting it on a qemu-native build Dec 20 22:41:50 I am not defining SRC_URI in the recipe. I just want to conditionally set the variable incase I want the individual recipe to override the value specified in the include file. Dec 20 22:42:24 the recipe can always overwrite what hte .inc does Dec 20 22:42:30 by defining those variables after the include Dec 20 22:42:39 because, for the third time, inclusion occurs right there, an d variable definition order matters Dec 20 22:42:47 a later define will override an earlier Dec 20 22:44:51 I finally get what you mean. Thanks Dec 20 22:45:17 an include is no different from copying and pasting those lines into that spot in the file, really :) Dec 20 22:45:22 np Dec 20 22:53:19 It's definitely a bit confusing, how our file format works. in some cases it's largely declarative, and you don't have to care about under the hood or evaluation order, but other aspects are more imperative and order sensative **** ENDING LOGGING AT Wed Dec 21 02:59:56 2011