**** BEGIN LOGGING AT Thu Oct 27 02:59:58 2011 Oct 27 03:13:04 03Mats Kärrman  07master * rc7179d3903 10openembedded.git/recipes/binutils/binutils_2.18.bb: (log message trimmed) Oct 27 03:13:04 binutils_2.18: update SRC_URI to avoid manual delete Oct 27 03:13:04 For reasons described in [1], the URL and contents of the source archive has Oct 27 03:13:04 changed and the old URL redirects to this. The checksums were updated by a Oct 27 03:13:04 previous commit [2], but only changing the checksums requires a previously Oct 27 03:13:04 downloaded archive to be manually deleted. This patch avoids this by updating Oct 27 03:13:05 the SRC_URI to the new URL. Oct 27 03:13:08 03Ulf Samuelsson  07master * r262f9eb0f0 10openembedded.git/recipes/binutils/ (9 files in 2 dirs): (log message trimmed) Oct 27 03:13:08 binutils-2.20.1: Change source package to binutils-2.20.1a Oct 27 03:13:08 2.20.1 is now 2.20.1a. Oct 27 03:13:08 Patches and diff -r1 backported from openembedded-core Oct 27 03:13:08 meta-openembedded/meta-oe/recipes-devtools/binutils Oct 27 03:13:09 When extracting the 2.20.1a sources, the directory name is still 2.20.1 Oct 27 03:13:09 $ LANG=C diff -rq binutils-2.20.1 binutils-2.20.1a Oct 27 06:15:28 MACHINE_KERNEL_PR in omap3.inc has a comment # Increase this everytime you change something in the kernel Oct 27 06:18:02 I've added some patches to my meta-layer and changed PRINC in my bbappend, but should I also bump MACHINE_KERNEL_PR? Oct 27 06:18:09 to a kernel recipe, that is :) Oct 27 06:19:13 hmm that's harder, but if you're not building any external modules you should be fine without incrementing it Oct 27 06:19:58 great Oct 27 06:32:26 hm. after bumping PRINC it didn't rebuild my kernel when I bitbake my image. Oct 27 07:09:18 03Joshua Lock  07master * r5f8f923b76 10bitbake.git/lib/bb/fetch2/__init__.py: (log message trimmed) Oct 27 07:09:18 fetch2: enable checksum definition as SRC_URI parameter Oct 27 07:09:18 URI parameters should be able to be defined as a parameter of the SRC_URI, Oct 27 07:09:18 this patch enables thus for checksums. Oct 27 07:09:18 An example; Oct 27 07:09:18 SRC_URI = "http://pkgconfig.freedesktop.org/releases/pkg-config-${PV}.tar.gz;md5sum=a3270bab3f4b69b7dc6dbdacbcae9745;sha256sum=3ba691ee2431f32ccb8efa131e59bf23e37f122dc66791309023ca6dcefcd10e" Oct 27 07:09:19 Addresses the remainder of [YOCTO #1399] Oct 27 07:09:28 03Joshua Lock  07master * rdb59297aa1 10bitbake.git/lib/bb/ui/hob.py: Oct 27 07:09:28 hob: fix backtrace when dismissing open dialog Oct 27 07:09:28 Clearly a logic/indentation error - we should only try and load the recipe Oct 27 07:09:28 should the file-chooser return OK. Oct 27 07:09:28 Fixes [YOCTO #1668] Oct 27 07:09:28 Signed-off-by: Joshua Lock Oct 27 07:09:29 03Joshua Lock  07master * r0a6443f9e1 10bitbake.git/doc/manual/usermanual.xml: Oct 27 07:09:29 doc: update manual for checksum changes Oct 27 07:09:30 Document the ability to use SRC_URI parameters for checksums Oct 27 07:09:30 Signed-off-by: Joshua Lock Oct 27 07:09:31 03Joshua Lock  07master * r30ca41e596 10bitbake.git/lib/bb/fetch2/__init__.py: (log message trimmed) Oct 27 07:09:31 fetch2: improve usability of checksums Oct 27 07:09:32 This patch improves the usability of checksums by only requiring one checksum Oct 27 07:09:32 be defined. Oct 27 07:09:33 Further, checksum verification will provide as much information as possible Oct 27 07:09:33 at, rather than a bit at a time. No longer will you need to run fetch, see an Oct 27 07:24:55 good morning Oct 27 07:32:00 My freshly baked OE-image works, has an IP and has dropbear installed, but I can't ssh into the board. Something else to enable/install that I didn't have to do in OE-classic? Oct 27 08:02:15 gm stefan Oct 27 08:10:53 hm. something to make dropbear allow root login most likely Oct 27 08:11:02 ? Oct 27 08:11:08 was always default Oct 27 08:11:14 no passwd Oct 27 08:12:33 woglinde: yeah. that worked by default on classic, but in core I'm not allowed to ssh in Oct 27 08:12:48 hm try to set the passwd Oct 27 08:12:53 via serial Oct 27 08:13:13 hmmaybee you have openssh now running Oct 27 08:15:42 woglinde: actually it was dropbear that had not started :-) Oct 27 08:16:14 haha Oct 27 08:16:19 you see Oct 27 08:17:28 I need more coffee to be able to think for myself today :) Oct 27 08:31:16 mornin Oct 27 08:37:36 ji jield Oct 27 09:36:23 hi Oct 27 09:36:39 mirrors.openembedded.org seems to be down, isn't it? Oct 27 09:54:41 hi obi Oct 27 09:55:37 you've scared him Oct 27 09:55:44 no Oct 27 09:59:40 if some build scripts checks for an executable during build time in order to enable/disable certain features; what would be the proper way to stage such a tool? I think the staging step became obsolete? Oct 27 09:59:55 what's the correct way now? Oct 27 09:59:58 I have to manually start dropbear and networking after booting my board though they are in /etc/init.d/. Isn't that ... odd? Oct 27 10:00:24 are you using systemd? Oct 27 10:00:33 are they in /erc/rc.* ? Oct 27 10:00:35 are they linked to the right runlevel? Oct 27 10:02:06 I am apparently using systemd. It says lots of stuff in dmesg. Oct 27 10:02:28 S40networking in /etc/rcS.d/ Oct 27 10:03:46 any image is longer working on AT91 systems Oct 27 10:03:53 no console-image Oct 27 10:03:57 no x11-image Oct 27 10:03:59 nothing Oct 27 10:04:01 K10dropbear in rc0,1 and 6 Oct 27 10:04:29 I've never looked at systemd before, so I don't know if the info I give is useful or not :) Oct 27 10:04:30 boot freezes Oct 27 10:05:18 sieh aus als muessten wir das rad nicht neu erfinden ;> Oct 27 10:05:20 oops wrong window Oct 27 10:06:33 mckoan: time to send some patches :) Oct 27 10:07:37 tasslehoff: iirc there is .service file for systemd in openssh.. so you can replace dropbear with that or add .service file to it as opensshd does Oct 27 10:08:23 JaMaOffELCE: I've been thinking about using openssh instead, so I guess I'll do that Oct 27 10:08:40 but networking doesn't start either Oct 27 10:09:11 but if you're using empty password dont forget to bbappend config which will allow that Oct 27 10:10:21 tasslehoff: is it networking as whle ot interfaces missing ie auto usb? Oct 27 10:13:02 JaMaOffELCE: I get an eth0 interface, but no ip until I do /etc/networking/start. Oct 27 10:15:29 no such script here, sorry Oct 27 10:16:47 * JaMaOffELCE plans to migrate images to systemd this weekend.. Oct 27 10:17:06 go go Oct 27 10:17:28 _my_ images.. Oct 27 10:17:31 :) Oct 27 10:19:43 sorry, /etc/init.d/networking start, I meant Oct 27 10:43:58 JaMaOffELCE: looks like Atmel eventually did their own overlay Oct 27 13:02:36 hi, all! Oct 27 13:03:08 privet Oct 27 13:03:10 How i am supposed to run various scripts like u-boot's mkimage from python function in .bb? Oct 27 13:03:30 Jin^eLD: how's your powerpc health? Oct 27 13:04:00 slapin: oh well, I fail to find correct settings for the jffs2 image, originally it uses a cramfs as the rootfs, I guess I will have to do the same Oct 27 13:04:31 mckoan: could you please tell where atmel's overlay is? Oct 27 13:05:37 Jin^eLD: you should know your page size and block size for your flash, that's all what jffs needs Oct 27 13:05:45 cramfs is very specific Oct 27 13:06:30 Jin^eLD: I think you might get this info from u-boot or kernel Oct 27 13:07:19 slapin: well, I believe that I got it right, but it still refuses to mount Oct 27 13:07:49 there is another vendor-made /etc partition which is jffs2, and which I did not touch yet, I guess I will boot from NFS and have a closer look at it Oct 27 13:14:09 slapin: http://www.at91.com/linux4sam/bin/view/Linux4SAM/OpenEmbeddedAngstromBuild Oct 27 13:14:24 slapin: ftp://ftp.linux4sam.org/pub/oe Oct 27 13:14:39 slapin: BTW aren't working with main tree Oct 27 13:16:28 Jin^eLD: what errors are there? Oct 27 13:19:12 slapin: I don't have the paste at hand, can't look it up now; but it was complaining about expecting 0x1985 and finding "something else" and then also complaining about possible wrong erase block (I tried various values there, also looked up the flash chip and tried things that should make sense) Oct 27 13:21:24 Jin^eLD: have you checked for insane obvious things like non-erased flash, badly written image, etc? does your image contains oob or not? Oct 27 13:22:13 slapin: well, for flashing I used the vendor script that I found in uboot, it did work for the cramfs... it does seem to erase flash too Oct 27 13:22:30 but too early to tell, I need to investigate a little bit more Oct 27 13:23:50 Jin^eLD: do you use NAND or NOR type flash in device? I mean disk-like or memory-like device? Oct 27 13:24:30 by looks at the driver it seems to be a NAND and its referred to as mtdblock something, so mem like I guess? my hardware experience is very low in that regard Oct 27 13:25:45 I'll try to mount or dump the vendor-created etc partition and will analyse it Oct 27 13:25:49 I think it should provide some answers Oct 27 13:26:10 Jin^eLD: /dev/mtdblock is for both types, you can hook NAND devices with it. What script do you flash it with? does it contain nand+something? or cp.b addr1 addr2? Oct 27 13:26:32 it contains cp.b Oct 27 13:26:34 Jin^eLD: I think dmesg should be enough information Oct 27 13:26:47 Jin^eLD: then it is parallel NOR flash Oct 27 13:26:55 oh.. and that means what? :) Oct 27 13:27:26 Jin^eLD: it is memory-mapped. Oct 27 13:27:44 Jin^eLD: it has cpu=accessible address for every byte of contents Oct 27 13:28:00 Jin^eLD: unlike NAND. Oct 27 13:28:09 which are disk-like Oct 27 13:28:33 this all is for default case of most embedded practice. Oct 27 13:29:21 hmm Oct 27 13:29:30 I wonder what part of the puzzle I am missing then... Oct 27 13:29:31 Jin^eLD: anyway, you just need page size and erase block size Oct 27 13:29:47 for jffs Oct 27 13:30:26 I think that erase block should be 0x2000 Oct 27 13:31:19 not sure about the page size, maybe that is what I am missing? Oct 27 13:43:59 Jin^eLD: probably. Pages vary in sizes, check your chip documentation or u-boot code or dmesg Oct 27 13:45:10 pages vary in size.. is that possible? ok.. I was not aware of that Oct 27 13:52:24 Jin^eLD: they depend on device. Each particular device has the same constant page size Oct 27 13:52:52 Jin^eLD: which should be known to generate proper jffs2 image Oct 27 13:52:59 Jin^eLD: for that flash Oct 27 13:55:02 emmc makes live easier and not Oct 27 13:55:58 slapin: well the flash is a 29LV640BBTC-90G/20918800A1 http://www.semiconductorstore.com/pdf/Macronix/mx29lv640bt-bb-1.1.pdf which should have eitehr an 8kb or 64kb sector size Oct 27 13:56:07 but what is the "page size" is that the sector? Oct 27 13:56:52 64kb looks lke sector size Oct 27 13:56:56 *like Oct 27 13:57:11 check page size Oct 27 13:57:36 usually something in 512-4096 bytes range Oct 27 13:58:50 check where, will that be in the pdf spec? Oct 27 14:00:24 Jin^eLD: ask u-boot for cfi info, it dumps appropriate data Oct 27 14:03:08 aah! http://pastebin.mozilla.org/1365810 Oct 27 14:04:40 so I would add --page-size=8192 to mkfs.jffs2? Oct 27 14:06:31 Jin^eLD: might work for you Oct 27 14:06:43 I'll try it tonight, thanks for the hint! Oct 27 14:20:39 hi jconnolly Oct 27 14:25:06 hello woglinde Oct 27 14:38:05 hi jconnolly woglinde Oct 27 14:54:19 hi gnutoo Oct 27 15:42:16 hi gnutoo Oct 27 16:55:39 I am having problems getting /dev working properly on a verdex6lp. I believe the issue is with the ramfs setup, with the following line in dmesg "ramfs: Bad mount option size". Does anyone here know anything about this? Oct 27 16:57:01 ben22 hm on which hw and which fs you use for the rootfs? Oct 27 16:57:32 woglinde, gumstix verdex6lp (pxa225?), on root fs jffs2 Oct 27 16:57:54 so its nand Oct 27 16:57:58 yes Oct 27 16:58:02 hm okay its ramfs Oct 27 16:58:05 so initrd Oct 27 16:58:07 hm hm Oct 27 16:58:18 can you try without initrd? Oct 27 16:58:32 sorry, how would i do that? Oct 27 16:58:45 what is your kernel cmdline? Oct 27 16:58:52 and whats your bootloader? Oct 27 16:59:04 sec Oct 27 16:59:55 bootloader = u-boot 1.2.0, command line = console=ttyS0,115200n8 root=1f01 rootfstype=jffs2 reboot=cold,hard Oct 27 17:01:10 hm looks like there isnt any ramdisk involved at all hm try with initrd=no Oct 27 17:02:51 ok, the error still shows Oct 27 17:04:40 I assumed you meant append initrd=no to bootargs Oct 27 17:13:44 yes Oct 27 17:13:59 loock if it isin the console output Oct 27 17:14:07 it is Oct 27 17:14:25 okay Oct 27 17:14:25 hm Oct 27 17:15:28 maybee this helps you to find the problem Oct 27 17:15:35 http://www.kernel.org/doc/Documentation/filesystems/ramfs-rootfs-initramfs.txt Oct 27 17:19:35 okay. I read a lot of technical things. it doesn't tell me what is calling the ramfs with the bad size option though. Oct 27 17:28:17 its the kernel self proably, so try to boot and dont let the kernel run a initrd Oct 27 17:32:25 my kernel has no intird! Oct 27 17:32:54 yes but it seems the kernel tries to load one Oct 27 17:33:11 and maybee initrd=no was the false cmdline option Oct 27 17:33:53 so noinitrd instead? Oct 27 17:36:41 thats why I gave you the txt file Oct 27 17:38:09 there's nothing in the docs about command line options Oct 27 17:38:29 hm okay Oct 27 17:38:34 than google yourself Oct 27 17:42:22 okay, i have tried "bootm {kernel_addr}", and "noinitrd". I find little documentation about u-boot-1.2.0, and i have no intention of bricking my device with a u-boot update. Oct 27 17:43:44 i have a question about adding PACKAGES to a recipe Oct 27 17:43:55 when I added something PACKAGES += "FOO" Oct 27 17:44:17 err PACKAGES += "${PN}-FOO" Oct 27 17:44:18 and >> +FILES_${PN} = "${bindir}/strace" Oct 27 17:44:23 grr Oct 27 17:44:25 i can't type Oct 27 17:44:42 FILES_${PN}-FOO = "files/from/install" Oct 27 17:44:54 what happens to the files in the base package? Oct 27 17:45:10 do I need to explicitly add a file listing for FILES_${PN} as well? Oct 27 17:49:45 msm it will not be in the normal package Oct 27 17:50:44 woglinde: i dont quite follow Oct 27 17:51:00 what im trying to do is just pull out one from from ${PN} and put it in ${PN}-foo Oct 27 17:51:22 the default ${PN} will pick up the file though... Oct 27 17:52:22 I guess I could use PACKAGES_prepend? Oct 27 17:52:50 then my new package will be built first - and the default ${PN} won't pick up my new packages items? Oct 27 17:52:59 a few recipes use PACKAGES_prepend Oct 27 18:00:08 no redefine the order than Oct 27 18:00:26 PACKAGES = foo bla moo Oct 27 18:00:53 woglinde: I don't really want to do PACKAGES = foo blah moo… I want to do PACKAGES += foo Oct 27 18:01:24 but my foo packages files will be stolen by a previous one =) I don't want to specify a new match for the defauly packages Oct 27 18:02:08 and? Oct 27 18:02:22 whats your solution now Oct 27 18:06:20 I am having trouble getting my gumstix overo air talking to a mmc card over mmc3 host. The data sheet said that it will send a 80kHz tone out to the reader. My console output shows it sending a 400kHz tone. The boot output is here: http://pastebin.com/SBLENckC Any help on getting the kernel to see mmc3 would be greatly appreciated! Oct 27 18:07:28 the relevant lines where its trying to talk to the card reader begin at line 252 Oct 27 18:11:41 OK. I have successfuly passed 'noinitrd' to the kernel. Now I get the error message that udev is not starting because "udev requires tmpfs support". But I have tmpfs support compiled into the kernel. Oct 27 18:12:06 ben22 as modul? Oct 27 18:12:14 or with y Oct 27 18:12:21 can't be done as a module Oct 27 18:12:27 hm Oct 27 18:12:27 as far as I can tell Oct 27 18:12:50 you are hitting all strange problems Oct 27 18:18:18 calling "mount -t tmpfs -o size=20m tmpfs /mnt/tmp" returns "ramfs: Bad mount option size". Is this unusual? Oct 27 18:20:32 okay, so calling it without the size option, and mounting to dev, results in udev starting... Why is '-t tmpfs' calling a '-t ramfs'? Oct 27 18:21:00 20M Oct 27 18:21:34 tried that. i thought the point of tmpfs was that it needed a size, and ramfs would not need the 'size' option Oct 27 18:21:37 its case sensitive Oct 27 18:21:53 uhm Oct 27 18:22:04 maybee I typo in the kernel Oct 27 18:22:07 cut and paste Oct 27 18:25:15 Well, I have udev working now. Thank you for your help, woglinde. Oct 27 18:30:15 ben22 fine Oct 27 18:38:05 I am trying to build qt4-embedded in bitbake. However, it triggers a bunch of X11-dependencies. Does qt4-embedded package really need X11-libraries? If not, how can I remove x11-dependencies from qt4-embeeded recipe? Oct 27 18:40:12 I am a gentoo user and I hoped that bitbake also would've had something like USE flags, but I couldn't find any like in bitbike manual. Oct 27 18:53:10 nonuse flags Oct 27 18:53:17 they are evil Oct 27 18:53:32 xorg is build do gstreamer and bluez Oct 27 18:53:43 use bitbake -g option Oct 27 18:53:52 to see the deps Oct 27 18:53:56 chains Oct 27 18:54:44 yeah, indeed, I hate when that happens, it always takes me much time to hack out the X11 stuff Oct 27 18:55:21 fix the bluez problem Oct 27 18:55:30 its there since some years Oct 27 18:55:50 but till now it was no such great burden for some one Oct 27 18:56:00 to fix it properly Oct 27 18:56:20 I guess the problem is the "properly" thing Oct 27 18:56:27 it gets pulled in by so many different things Oct 27 18:56:34 no Oct 27 18:56:35 the machine features for instance Oct 27 18:56:49 there is a lot of logic in task-base and machine features Oct 27 18:56:53 bluez sound and gestreamer Oct 27 19:15:50 woglinde: this is workking as expected: Oct 27 19:15:51 +PACKAGES_prepend = "${PN}-graph " Oct 27 19:15:51 +FILES_${PN}-graph = "${bindir}/strace-graph" Oct 27 19:18:54 hm yeah preprend Oct 27 19:18:57 sorry Oct 27 19:19:50 woglinde: keeps things simple =) Oct 27 19:20:07 I'm getting an error when trying to build u-boot that bitbake cannot find the URL: www.sakoman.com/git/u-boot.git;branch=gnome-r13;protocol=git. I've noticed that it's underactive development today. Anyone know if/how to get a copy? Oct 27 19:25:13 brent what oe is this? Oct 27 19:27:53 the gumstix org.openembedded.dev version Oct 27 19:28:30 overo-2011.03 Oct 27 20:23:40 solved my problem. ask me if anyone else has the same problem Oct 27 21:00:20 I'm trying to build "core-image-minimal" for qemuarm but qemu fails during configure. Oct 27 21:00:39 Complains with arm-oe-linux-gnueabi-gcc: unrecognized command line option '-m32' Oct 27 21:00:47 Anyone else seen that? Oct 27 21:46:49 hi! I'm using BeagleBoard-xM rev C and trying ti use Linux kernel 3.1 - it hangs during boot Oct 27 21:51:29 off topic Oct 27 21:51:51 ok, where is correct topic? Oct 28 01:26:42 Greetings. Question, is there a way to tell bitbake to clean all recipie PR's? AKA I do "bitbake dropbear -c clean" but I can see that an old r1 and r2 is in my work dir that it isn't cleaning (only the latest is cleaned). **** ENDING LOGGING AT Fri Oct 28 02:59:57 2011