[00:22] pcacjr, https://bugs.launchpad.net/ubuntu/+source/system-config-printer/+bug/412336 also seems to match [00:22] Ubuntu bug 412336 in system-config-printer (Ubuntu) "Crash while browsing smb printer shares (affects: 4) (heat: 10)" [Undecided,New] [00:22] there's a pcap capture of the network traffic [00:25] and removing all mention of mdns from /etc/nsswitch.conf eliminates the crashes [00:27] it appears that adding wins in front of mdns4 in /etc/nsswitch.conf also removes the crashes [00:31] I'm remaking a fresh image with wins added to verify, it'll be about 30 minutes === fta_ is now known as fta === fta_ is now known as fta === fta_ is now known as fta [03:21] hi, does anyone know if there is a rmadison url for ubuntu-arm? [03:23] There isn't. If you make a local mirror of the packages and sources files, you can use plain madison. [03:24] But the sources should always be the same as any other architecture in Ubuntu, and the binaries ought be similar, excepting when there is a failure or delay to build. [03:24] persia: ok, I'l just use the normal one for now [03:25] acutally, does anyone have any tips for setting up a cross compile toolchain? [03:25] It's what most folks do :) The repo is the same, after all. [03:25] Generally, we don't cross-compile. All the packages are compiled natively. [03:25] I'm currently trying to get buildcross from embedian to work [03:25] Would an emulated build meet your needs, or do you need cross-build for something specific? [03:26] an emulated build is very slow [03:26] Yeah :) [03:26] I was building QT embeded, and it took 43 hours [03:26] and it still wasn't done, so I stopped it [03:31] :-) [03:32] brendan0powers_: you can try the linaro toolchain that hrw|gone is maintaining [03:32] you can cross-compile then [03:34] rsalveti: that looks interesting [03:35] Note that combining cross-compiled and native-compiled stuff may or may not work as expected, depending (basically, widely untested) [03:35] depends a lot on what package you're using and how you're building it [03:35] Yep. [03:36] I generally only cross-compile the kernel, as it's much much faster [03:36] hmm [03:37] the buildcross script I'm fiddling with takes the same version of gcc, libc, etc.. as the distrobution I'm using(jaunty) [03:37] so hopefully that will help some [03:37] I'm really quite suprised at how difficult this is [03:38] got this http://cdimage.ubuntu.com/ubuntu-netbook/ports/daily-preinstalled/current/maverick-preinstalled-netbook-armel+omap.img.gz for testing purpose on non-omap device and X doesn't seem to work. how can I get a login ? [03:38] does it use xf86-video-omapfb ? [03:39] I think it's using omapfb or fbdev, I'm not sure yet [03:39] fbdev would work fine for you [03:39] yeah likely. is there any way to login on tty though ? [03:40] I tried ubuntu/ubuntu login and ubuntu/ doesn't work [03:41] dcordes__: this iso requires you to run it for the first time [03:41] to finish the installation steps, like creating a normal user [03:41] ugh let's see if it will like my known working karmic xorg.conf [03:42] rsalveti: does the linaro toolchain have a PPA? [03:42] rsalveti: I get lost in launchpad [03:42] brendan0powers_: yep, but I believe it works for maverick only, but you can try it in a chroot env [03:42] not a PPA, but an external repo, AFAIK [03:42] rsalveti: if it's so omap specific can I assume rootfs has nice neon accerlation ? [03:43] brendan0powers_: http://people.canonical.com/~hrw/ubuntu-maverick-armel-cross-compilers/ [03:43] dcordes__: neon is not supported by default [03:43] is target for armv7 and thumb2, but not neon [03:43] as not all armv7 hardware supports neon [03:44] I think pixman has neon routines available automatically though, no? [03:44] rsalveti: so currently nothing is compile with neon ? [03:45] would be really cool to have neon ffmpeg ootb [03:45] dcordes__: not that I know [03:45] ok [03:45] yep, but getting a specific ppa for that would be a way to go [03:46] then users could easily install and use it [03:50] rsalveti: looks like many different apt repositories coming :) [03:50] do we have armel ppa's yet? [03:51] dcordes__: yep, but is the way to go, at least for neon, that doesn't affect many packages :-) [03:51] wait a second. what is ppa ? [03:51] cwillu: I believe you can create it by submitting packages by sftp, but didn't test yet [03:52] it ate my xorg.conf very good [03:53] okay, I should play around with that [03:53] thanks [03:53] ok but doesn't work as expected. after clicking the login button screen goes black and then goes back to login screen [03:53] I think it is a problem to not add any extra login possibility [03:55] guess the gui started depends on extensions not provided by xf86-video-fbdev [03:55] dcordes__: I guess it should work, at least the 2d ui [03:55] rsalveti: well is 2d ui default [03:57] I have seen that it has the 2d option with efl installable (on some blog) [03:57] yep [03:57] rsalveti: any pointer where the script is that starts the gui ? [03:57] basically I just need to add a login manually [04:05] dcordes__: hm, don't know, try looking on how gdm request the sections === fta_ is now known as fta [04:11] nite all [04:17] pcacjr: nite === fta_ is now known as fta === fta_ is now known as fta === fta_ is now known as fta === amitk is now known as amitk-afk === tmzt_ is now known as tmzt === amitk-afk is now known as amitk [08:06] OMG cooloney! [08:07] lag: man, morning, hehe [08:08] That pull-request is bonkers [08:09] Are you going to submit an email for every one? [08:10] [1/220] [08:10] [2/220] [08:10] [3/220] [08:10] .. [08:10] oh, no, [08:10] you guys can pull from my branch to review [08:10] i am going to upload a kernel package for testing [08:11] i have no choice, that's the way TI release their code [08:11] :) [08:14] * amitk prepares for 220 emails :) === hrw|gone is now known as hrw [08:31] morning [08:36] michaelh1: 'state of toolchain' news can also interest debian-embedded ML [08:41] hrw: Ta, I'll add that to the list. [08:42] Morning hrw [08:43] michaelh1: they wrote cross rules iirc [08:45] hrw: their own, generating debian packages? [08:48] binutils and gcc cross rules which exists in packaging [08:49] michaelh1: We refactored the cross rules in Debian packages, which had been mostly developed by emdebian / debian-embedded folks [08:49] offtopic: looks like it is time to say goodbye for chromium browser for me. slow use [09:13] lool, i see perl errors in current build logs (dictionaries-common postinst fails) that werent there last week and the failing package wasnt uploaded inbetween (neither was perl) [09:13] (image build logs that is) [09:34] ogra: How's the image going? [09:35] lag, see abive [09:35] *above [09:35] http://paste.ubuntu.com/469721/ [09:36] I don't know much perl [09:36] What does this mean in real terms? [09:36] Will the daily build pass tonight? [09:38] not if thats not fixed [09:38] this breaks package instalation in the image [09:44] ogra: These look like perl warnings, not errors; looks like some debconf is not set properly? [09:46] ogra: Do you have the full livefs build log for armel? [09:46] I only have a truncated email [09:48] ogra: The error is: [09:49] cp: cannot create symbolic link `/tmp/tmp.70yZsJ9bgz/chroot-livecd/usr/share/gnome/help/serverguide/nl/other-apps.xml': No space left on device [09:49] cp: cannot create symbolic link `/tmp/tmp.70yZsJ9bgz/chroot-livecd/usr/share/gnome/help/serverguide/nl/virtualization.xml': No space left on device [09:49] ogra: Would you take care of fixing the livefs logs sync to people.ubuntu.com/~ubuntu-archive? [09:50] ogra: Same issue on the omap4 buildd === amitk is now known as amitk-afk [09:50] So both sycamore and acorn fail in the same way; I think it's lack of space on your loop device [10:17] morning [10:33] lool, thats old [10:34] ogra: It's from today [10:34] not in my mails [10:35] hi lool, do you know if michael or yao made any progress with the mozilla build failures? i stopped looking at it yesterday evening when my head started to hurt ;) [10:35] hmm, right, on the buildds the log is longer :P [10:35] chrisccoulson: I stopped looking as well [10:35] chrisccoulson: yao did some test builds [10:35] chrisccoulson: #linaro? [10:35] michaelh1: ^ [10:35] ogra: As I said, the log is truncated, if you check today's logs you will see the issue I mention... [10:35] yes, i see it [10:36] /dev/loop0 2131676 3192 2128484 1% /tmp/tmp.FOMtHUOcZ5 [10:36] cp: cannot create directory `/tmp/tmp.FOMtHUOcZ5/chroot-livecd/usr/src/linux-hea [10:36] yep [10:36] ogra: Would be nice to fix syncing of logs, and also to fix the length of the emails; probably a SIGPIPE issue [10:36] which is weird since there is a 2.1G image we try to copy 1.6G on [10:36] lool, yes, i will do that, not my highest prio though [10:37] Eh, not my highest priority to look into livefs build logs :-) [10:37] heh, indeed [10:37] seems like /proc is still mounted :/ [10:38] in the chroot when copying === ericm|ubuntu is now known as ericm-afk [11:19] chrisccoulson: sorry for the delay. Yao has been looking into this. The failure does go away in debug/no optimisation mode so we'll track it down further [11:22] michaelh1: chrisccoulson is on #linaro now === amitk-afk is now known as amitk === ericm-afk is now known as ericm [12:06] zumbi: debian-embedded & linaro-toolchain ML [12:58] ogra: ping === zyga is now known as zyga-afk [13:08] Hi! Does anyone know why we can't boot UNE on NFS, while we can with a minimal system? This was observed on both Lucid and Maverick. [13:09] I'm currently investigating (booting with init=/sbin/init --debug), but if some already knows why, it won't spoil my experience ;-) [13:19] mopdenacker: can you elaborate ? what is the exact problem you are facing ? [13:19] hi btw [13:19] :) [13:24] Hi dcordes_ ! I'm trying to boot my board on a root fs on NFS, which contains the netbook edition. I can mount the root filesystem, but something hangs in the boot process: http://pastebin.com/AUKawzvz [13:25] If my rootfs is just a minimal system, I can reach a prompt in the console. [13:26] So, there is one of the UNE services that hangs the boot process. There are many of them, so it's not so easy to find... [13:27] Of course, the same rootfs boots fine from MMC. [13:27] mopdenacker: Do you have an initrd? [13:27] seems not [13:27] mopdenacker: You might have different behavior with and without initrd; I'd also check you fstab [13:27] # [13:27] mount.nfs: remote share not in 'host:dir' format >> points at fstab not being in the right format? [13:34] hi lool ! No, I don't have an initrd. I also have the 'host:dir' message when I boot on the minimal rootfs via NFS. So, it doesn't seem to be the issue here. [13:34] My /etc/fstab just contains /proc. === bjf[bjf] is now known as bjf [13:59] mopdenacker: Do you have networking setup in your UNE fs? if networking doesn't come up, you wont get any ttys [14:00] mopdenacker: Either arrange for n-m to start, or for ifupdown to be installed with some etc/network/interfaces file which has at least a definition for lo [14:04] lool: ouch, I've just realized that it's worse than that. The kernel itself hangs 10 to 15 seconds after starting the init process (it stops responding to pings). [14:05] So, the culprit seems to be the TI kernel I'm using, not user space. [14:05] Eh [14:05] I will go on investigating on that side... [14:05] mopdenacker: Perhaps network is being brought down and breaks NFS which breaks boot? [14:06] lool: that's a possible cause. Thanks for the tip! [14:32] mopdenacker: maybe show the .config you are using for the kernel used. this might also be a question for #ubuntu-arm [14:32] mopdenacker: sorry, that's #ubuntu-kernel [14:32] * dcordes_ also has question [14:33] I have a keypad driver for my device and would like to map alt key to one of the buttons. what is the correct way to identify a code for the button I want to map to, and how do I remap it ? [14:54] Actually, my kernel is not completely dead. The network seems to be dead, and the rsyslogd task is hung (http://pastebin.com/3G5fCpbv) [14:56] The effectively looks hung because of the network problem... [15:00] lag: I'm a bit nervous testing the imx51 kernel on my babbage if it is the same one that blew up two boards at the sprint. [15:00] GrueMaster: I don't know anything about it [15:00] GrueMaster: I would be too [15:01] GrueMaster: You need to speak with cooloney [15:01] lag: thanks, [15:01] It was part of your status update in the meeting, that's why I brought it up. [15:01] GrueMaster: yeah, that's what i'm concerned [15:01] i still don't know why the kernel update will break the board [15:02] * lag is just the collector and presenter of information :) [15:04] lag: you are the man in the middle of US and China [15:05] cooloney: Seemingly [15:07] Yea, Austrailia is between us. :P === amitk is now known as amitk-afk [15:19] ogra: How's the build issue coming along? [15:24] lag, NFC [15:24] still digging [15:25] i'm doing a testbuild that lists all contents of the image now there is no tmpfs mounted or anything, the image is 500M bigger than the filesystem requires but i still get space errors [15:25] NFC? [15:25] Is the first work 'not' and the last word 'clue'? [15:30] hello [15:32] I'm trying to make an image for the BeagleBoard using rootstock, but after I run the rootstock script it don't make me any vmlinuz file [15:32] how do i solve this ? [15:32] lag, yeah, and the middle part is not family friendly [15:32] Quite [15:32] though you make families with it :) [15:32] I guessed as much :) [15:33] Funky-chickens generally aren't [15:33] samuel: you can try to install the kernel you want later, or even giving to rootstock as an argument [15:34] How do I do that ? witch argument ask the rootstock to create me the file? [15:35] samuel: there are some different kernel flavors at the repo, for omap4 for example [15:35] I tried to flow this WIKI http://elinux.org/BeagleBoardUbuntu#Boot_Partition [15:36] samuel: sure, you just need to install the kernel package [15:36] and you'll be fine [15:36] samuel: are you trying it for a beagleboard? [15:36] but i don't have any boot files ... :( [15:36] and what ubuntu version, lucid? [15:36] yeah [15:38] samuel: http://ports.ubuntu.com/ubuntu-ports/pool/main/l/linux-ti-omap/linux-image-2.6.33-502-omap_2.6.33-502.8_armel.deb for example [15:39] ogra: send address for next shipment of awesome [15:39] but this is just a kernel Image, and i need the files to make a uboot file for the fs image [15:40] I have more HW for you [15:40] rsalveti, (thanks by the way) [15:40] or at least fixed HW [15:40] :) [15:40] cool :-) [15:40] hooray [15:41] Also, I need Ivanka's (DX team) email address [15:42] let me ask it differently, Do I need to compile the xload and uimage after I run Rootstock ? [15:42] samuel: if you request to install the correct package, you'll get the uImage and uInitrd, so you can boot ubuntu successfully [15:43] now for xload I believe there's a package for it, but you can try to update it by hand [15:43] cpearson, see PM [15:44] rsalveti, you mean on the host machine I should install another some thing else then rootstock ? or that I need to ask rootstock for it ? [15:45] samuel: you should request rootstock to install it [15:46] when you install the kernel package with rootstock, it'll generate the uIinitrd and uImage [15:46] so you grab them and use the way you like/want [15:46] for xload is different, you can try to update it by flashing at your device [15:46] rcn-ee has some scripts to do the work [15:47] samuel: see http://elinux.org/BeagleBoardUbuntu [15:47] samuel: --seed ubuntu-minimal,linux-image-omap for example [15:47] will install the ubuntu-minimal and the linux-image-omap [15:48] rsalveti, yes I flowed this example. Rootstock finshed to build the linux fs but didn't create any vmlinuz or initrd.img as explined [15:48] on http://elinux.org/BeagleBoardUbuntu#RootStock:_Running [15:49] samuel: yep, try installing giving the linux-image-omap as seed [15:49] you'll get those [15:50] samuel: if you request it by the seed argument, the vmlinuz and initrd.img will be inside the rootfs [15:50] if you request with the kernel-image argument, you'll have these files together with the rootfs at your directory [15:51] * rsalveti needs coffee [15:51] I have requested the kernel image by argument as in the example [15:51] rsalveti, thanks :) [15:51] I will try again :) thank again [15:52] samuel: np, let me know if it worked for you [16:07] rsalveti, I don't know what happed, though it seems to work (to a point). Now I get rootstock: 792: Bad substitution error [16:08] I used this kernel-image http://rcn-ee.net/deb/lucid/v2.6.33.4-l3/linux-image-2.6.33.4-l3_1.0lucid_armel.deb === arun___ is now known as arun_ [16:10] samuel: what version of rootstock are you using? === arun_ is now known as arun [16:10] let me try with mine [16:19] rsalveti, let me check [16:20] rsalveti, I don't have a clue what is the version [16:21] I don't find any indication in the script files [16:22] I download it today with bzr [16:26] samuel: please update your branch, just pushed a typo fix that rcn-ee found [16:26] will grab something to eat and be back in 30m === fta_ is now known as fta [16:58] samuel: just tested here and it worked fine, you'll have initrd.img-2.6.33.4-l3 and vmlinuz-2.6.33.4-l3 after creating the rootfs [16:59] I have commt rootstock back to 94 and still the same error [16:59] :( === hrw is now known as hrw|gone [17:02] samuel: weird [17:02] samuel: what are the arguments you're giving rootstock? [17:02] rsalveti, I'm trying to past it on pastbin [17:12] is there a know problem using rootstock with virtualbox? [17:36] rsalveti, I think the problem is with the VritualBox. I'm checking it now [17:37] samuel_: probably [17:37] it could be a problem. [17:41] can you explain me about qemu ? is it like VM to test my rootfs for embedded device ? === fta_ is now known as fta [17:51] I think Virtual Box is only x86 VM. Qemu will do multiple arches, but has a lit of limitations wrt full compatibility. [17:52] Are you trying to run rootstock/qemu from within a Virtualbox VM? [18:04] hrw|gone: ack [18:06] cpearson: hey, i want hw too! [18:10] he ran away :D [18:10] zumbi: do you fear me too? :D === fta_ is now known as fta [18:15] armin76: eh?! [18:16] armin76: i ignore joins parts and quits :) [18:16] armin76: this is another world [18:47] cpearson: i want hw too! :) [18:49] anyone pretty good with ubiquity and installers and kernel related installer stuff? for instance what on earth is boot=casper and what do I do to my kernel to make this work? ideally so I can grab an installer image from the dailies and just swap out the kernel [18:57] armin76: so sorry chary... [18:58] not enough to go around yet, but you can order Beagles or BeagleXMs now, and others soon [19:04] hm, it seems that the qemu hang issue is related on how apt is using pselect/select and the qemu support for these syscalls [19:05] now I got another problem with the stupid mono package, that wants to run lots of stuff while installing [19:12] Neko, as far as I know, boot=casper is ignored by the kernel and instead checked by an upstart script later [19:13] what I did was basically replace the kernel from the babbage sd card image with my own (efika mx) and then go ahead and try boot from it, but what happened was it just sorta got to mounting a root fs and couldn't find init [19:13] that wouldn't be related to the boot=casper then [19:14] yeah but the kernel is never directed to any other location to find init (it's actually in a squashfs isn't it?) and there's no way the kernel could pick that up [19:14] I don't understand how it gets there or if there's a little ubuntu tweak that does it or so [19:16] there's a hardcoded default location for init [19:16] /bin/init or /usr/bin/init I believe [19:19] there's an initramfs line which is typically used, I think it's in use with the ubuntu arm images as well now [19:21] well I just took the initrd and made it a uinitrd and loaded it and it just sorta stuck. I don't know if it's even using it. [19:21] from the effect it had, I'd say not at all, as if it just ignored it === XorA is now known as XorA|gone [19:22] Neko, there's various "break=" lines you can try which will give you a busybox from inside the initramfs if that's where you're hanging [19:23] ok === bjf is now known as bjf[food] === bjf[food] is now known as bjf [20:03] Does that assemblercode do what i hope/want it to? http://pastebin.com/AEUFfPjX === fta_ is now known as fta [20:20] Andre_H: I suppose that depends on what you want it to do.. [20:21] it should jump to the address in field 1 of .mytable [20:22] do you really mean: mov pc, pc?? [20:23] i learned somewhere that pc points about 8 byte into the future [20:23] also.. PC is usually some number of instructions ahead because of pipeline.. but offhand I'm not sure if the LDR instruction will patch things up to hide that.. [20:23] yup [20:24] when you do a BX type instruction, it gets somehow magically patched up, so that LR has the correct return address.. but, for example, on a data abort, it is not.. [20:25] i would need some simple arm simulator to find out if it patches it up or not... [20:25] I did some prototype for a code coverage tool that worked by doing single instruction writes to the "variables" inserted in the code, but that was years ago and I've forgotten whether I had to account for the PC being ahead of the "current" instruction [20:25] qemu? [20:26] i do my arm work in qemu, so i guess i have to get some more skilled with gdb [20:27] 'stepi' and 'info registers' are all you need to know ;-) [20:27] oh, and disassemble [20:27] (and display/i $pc) [20:29] on unrelated topic, does anyone know if there already exists a PPA somewhere with armel builds of firefox (or chromium) that understand what .webm is? [20:29] new thinkings -> new motivation!!! [20:29] it seems firefox isn't clever enough to know to call totem plugin for .webm [20:29] thx [20:29] np [20:49] pcacjr, figure anything out re: system-config-printer dying? === markos_ is now known as Guest17373 === kmargar is now known as markos_ === andy_ is now known as andyj === Tscheesy_ is now known as Tscheesy [22:06] * ogra_cmpc doesnt get it [22:07] i can exactly copy 1496228k onto the image, after that i get "no space left on device" [22:07] but the image is 2G big [22:08] ogra@babbage2:~$ sudo touch /tmp/tmp.Xr4NggTX5n/blah [22:08] touch: cannot touch `/tmp/tmp.Xr4NggTX5n/blah': No space left on device [22:08] ogra: lack of inodes? [22:10] ogra_cmpc: happens also if you dd something bigger than this? [22:10] or just when populating the fs [22:11] when cp'ing [22:11] it is loop mounted [22:11] ogra@babbage2:~$ df -h |grep tmp [22:11] /dev/loop0 2.1G 1.5G 623M 71% /tmp/tmp.Xr4NggTX5n [22:12] it should have 620M free [22:13] ogra_cmpc: I know, but did you tried giving a dd with a file larger than this inside the loop device? [22:13] i didnt dd inside the loop device [22:13] this is weird, I was having similar issues with genext2fs but the problem was lack of free inodes [22:13] but I believe you're using mkfs.ext3 or similar [22:13] right [22:14] on top of a raw dd'ed file [22:14] which i then loop mount [22:14] the file is 500M bigger thyan the source tree that i copy [22:15] 512M to be exact [22:17] ogra_cmpc: What kernel are you trying this on? Is it possible that there is a regression? [22:20] its a babbage2 with lucid release kernel [22:26] For grins, have you tried your test on an x86 system to see if it performs the same? [22:32] rsalveti, wow, you are right, dumpe2fs shows 0 free inodes [22:33] ogra_cmpc: ouch [22:33] now you just need to fix it :-) [22:33] heh, yeah === bjf is now known as bjf[afk] [23:37] ogra, Hi, I think there is a but you should note [23:39] when trying to create rootfs for arm with rootstock on Ubuntu 9.10 on VirtualBox, and the rootstock was installed via apt0get the operation fails. [23:39] but when creating the rootfs with rootstock [23:40] 0.13 the problem is fixed [23:40] you should also note that I used debootstrap_1.0.20ubuntu1~jaunty1_all.deb package