[02:26] how to add new drivers like ppp? i see that there are no modules for ppp connection using beagleboard. === hrw|gone is now known as hrw [08:09] morning [08:20] * cwillu yawns === ericm|ubuntu is now known as ericm [08:27] ureadahead invoked oom-killer: gfp_mask=0xd0, order=0, oom_adj=0 [08:27] nice. [08:28] what HW is that ? [08:28] beagleboard. [08:28] board, disk type, ram etc [08:28] on SD card or USB key ? [08:28] and do you have swap ? [08:28] Booted from 2GB SD card (that's 90% full), no swap. [08:29] graphical env or only cmdline ? [08:29] oh, and: on host: hub 8-0:1.0: unable to enumerate USB device on port 1 usb 8-1: new full speed USB device using uhci_hcd and address 9 usb 8-1: device descriptor read/64, error -110 [08:29] Commmand line, with some extra stuff (such as deluge). [08:29] Xorg is installed, but no display or login manager. [08:29] hmm, i havent seen a beagle hitting OOM in a cmdline env yet [08:29] is that a rev B ? [08:30] (128M) [08:30] Nope, C4. [08:30] hmm, thats indeed very weird [08:30] where's that ureadahead pack file? [08:33] man ureadahead tells you [08:33] ah, I thought it was /var/lib/ureadahead/pack [08:33] File doesn't exist. [08:33] though ureadahead should gracefully exit on Sd cards anyway [08:33] with code 5 iirc [08:35] argh, tried to load g_audio, and got screenfuls of panic. [08:36] g_audio ? isnt that OTG stuff ? [08:36] yeah. [08:36] OTG is broken in the kernel still [08:37] I hope we'll get musb built-in but all the modules... modular. [08:37] https://bugs.launchpad.net/bugs/566645 [08:37] Launchpad bug 566645 in linux-ti-omap (Ubuntu Maverick) (and 2 other projects) "OTG configuration is broken on omap kernel (affects: 2) (heat: 14)" [High,Confirmed] [08:37] So then I can choose to use g_audio instead of g_ether, or such. [08:37] It also rather oddly hangs lsusb on the host. [08:37] basically we dont have musb atm [08:39] I wonder why it doesn't work as a module. [08:41] i think it didnt work compiled in either [08:41] amitk would know, he ported the patch back then [08:42] no porting involved, it was the state of musb in 2.6.33 [08:42] cooloney, can you make sure that our omap4 kernel has the same config for udebs as the omap3 one ? i think the files under debian.ti-omap/d-i/modules/ rule that, probably you can just copy them over [08:45] ogra: the d-i files in omap4 was copied from omap3 debian.ti-omap/d-i/modules/ [08:46] ogra: they are the same, i never change that. [08:46] they recently changed :) [08:47] i just checked the nic-usb-modules package yesterday and its missing the same modules the omap3 one did [08:48] mathieu added a fix afaik it was accepted to the tree already [08:50] ... and that is why we are moving to a unified debian tree. That will be sync'ed across all branches, all releases. [08:53] ogra: oh yeah, i noticed that. will check it later. maybe just apply mathieu's patch [08:53] that should be fine [08:55] also weird: cnetworkmanager --syscon doesn't list the wireless connection I stuck in /etc/NetworkManager/system-connections/ [08:57] And it also ignores my "Wired Network" connection by that name, and instead creates an "Auto eth0". [09:03] I've never managed to get headless NetworkManager to work well. === XorA|gone is now known as XorA [09:52] Hi everyone [09:53] hey Laibsch [09:53] XorA! [09:53] Laibsch: joining the ubuntu party? [09:53] I'm a bit surprised to find you here [09:53] I think I probably was here earlier than you ;-) [09:53] But probably not as active [09:54] I always you'd never budge from being a staunch Debian user and Ubuntu-loather [09:54] Good to see you here [09:54] * XorA is paid to be here :-D [09:54] are you? [09:54] * XorA is a fedora user [09:54] more power to you [09:55] they should only hire people that at least have a Ubuntu tattoo to show their commitment [09:55] anything less makes no sense ;-) [09:55] hehe [09:55] that might be why at UDS they kept trying to get me intot the "dark" room :-D [09:55] probably [09:55] * XorA is on contract with TI which is working on ubuntu stuff [09:56] I see [09:56] XorA: Are you based in Nice? [09:56] lool: Scotland :-D [09:56] Didn't know TI had a Scotland office [09:56] * XorA has been in Nice [09:57] lool: they dont, I contract [09:57] XorA: Well too bad we failed meeting at UDS then, I dont think I had the chance to chat with you, or if I did I failed I associated name and face [09:57] * XorA is actually based in Dallas office [09:57] anyone used http://elinux.org/BeagleBoardUbuntu#NetInstall_Method ? [09:57] i get a kernel panic on boot :/ [09:57] Oh you're based in Dallas, then I might see who you are [09:57] lool: I was hanging with the TI dudes mostly, you would have found me in Xorg and ALSA talks [09:58] lool: XorA is also one of the OE crowd [09:58] XorA: not everyone is perfect, I hope you don't miss the bus to salvation ;-p [09:58] (just kidding) [09:58] hurry, doors are closing ;-) [09:58] * Laibsch hears that all the time [09:58] http://www.pastie.org/986954 [09:58] any suggestions? [09:58] actually it's "please don't hurry or rush, doors are closing" ;) [09:59] amitk: hehe, Im too old for distro wars [10:01] amitk: better not start - he will try to move you from ubuntu to angstrom [10:02] hrw: I've used OE before for the Zaurus 5500 [10:02] doesn't the Z predate OE? [10:02] * XorA still has a 5500 [10:02] * Laibsch too [10:02] * amitk too [10:02] where is the ubuntu port? [10:02] not more than one meter away ;-) [10:03] XorA: working on it ;-) [10:03] but probably impossible [10:03] Laibsch: no, OE was made as replacement for OpenZaurus build system [10:03] hrw: yes, I know [10:03] amitk, did you have some preinstalled image of ubuntu somewhere? [10:03] * hrw does not have 5500 - donated to 2.6 hacker [10:03] there is a jaunty port for the zaurus somewhere [10:03] Laibsch: if I knew how to rebuild full ubuntu easilly it would be real easy to do an armv4 port [10:03] hrw: therefore the Z was there before OE [10:03] yeah, and it failed miserably last time I tried one. [10:03] Laibsch: I just need a build farm [10:03] Kernel didn't have ucb1x00_ts enabled, or such. [10:04] Laibsch: ah yes... [10:04] neure: preinstall image? Not for Lucid. We will have preinstalled images for Maverick though. [10:04] :/ [10:05] anything that i could try? [10:05] neure, that netinst image uses a kernel from rcn-ee, wait for him to be around [10:06] oh i lied, its hardy that was ported to the zaurus [10:06] http://www.omegamoon.com/blog/index.php?entry=entry081030-074514 === ericm is now known as ericm-afk [10:06] neure: did you try with a different usb-ethernet adapter? [10:07] use an asix or pegasus based one [10:07] amitk, no, i dont think i have one [10:07] i could perhaps get the apple one [10:07] is it known to work? [10:07] also did you file a bug aboout your partitioning error ? [10:08] no, i didnt [10:08] its very likely that you will hit it again even with a netinstall [10:08] maybe [10:08] if i do i can try to file a report [10:08] but its just 'doesnt work' [10:08] well, thats not a proper erro description we accept here :) [10:09] hmm [10:09] i have bootargs=vram=12M omapfb.mode=dvi:800x600-16@60 console=ttyS2,115200n8 root=/dev/mmcblk0p2 rw rootwait [10:10] i wonder what i should have in order to use this rcn-ee thing [10:10] no idea [10:10] rootwait is definately wrong for the official images [10:11] what does it mean? [10:11] but you might need it in rcn"s [10:11] makes kernel wait for rootfs device to appear [10:11] it means to wait for the rootfs device until it shows up [10:11] since ubuntu always uses an initramfs in the official images thats not needed [10:12] we also boot in ro mode by default [10:12] because mountall remounts rw after fsck [10:13] hmm [10:14] mattman wrote me that he managed to boot to a console using our ti-omap kernel + qemu-maemo! :-) [10:14] XorA: lool shared with me a couple of blueprints discussed at UDS: https://blueprints.launchpad.net/sprints/uds-m?searchtext=arm- [10:15] https://blueprints.launchpad.net/ubuntu/+spec/arm-m-cross-compilers looks like something you and I may be interested in [10:15] if i have this 16GB mmc card inserted, i get this: http://www.pastie.org/986970 [10:15] is something wrong? [10:16] lool: have you been to Nice BTW? [10:16] if i dont have it in there, i get http://www.pastie.org/986971 [10:16] XorA: I came to visit, yes [10:16] XorA: Ubuntu on ARM is still very young and evolving, so I have a hard time (and not enough of it)to keep up with the status [10:16] For a couple of days only [10:16] I wouldn't mind going back :-) [10:16] it's an easy ride [10:16] lool: same meeting as persia jumped out of hotel? [10:16] haha yes [10:16] lool: you met me then [10:16] I was his translator in the hospital :-) [10:17] XorA: I missed the morning meeting, but I think I saw you Monday afternoon [10:17] XorA: I am the guy who talked about cross-compilation [10:17] lool: I was all in black wearing New Rocks [10:17] lool: from weds onwards I had my nails painted [10:17] Laibsch, i wouldnt call 1.5 years "young" :) [10:18] I would ;-) [10:18] XorA: I only stayed until Tuesday evening I'm afraid [10:18] I missed the fingernails [10:18] lool, he had them painted at UDS too [10:18] could it be that there is an issue with the sdcard? [10:19] could that be causing issues for server installer? [10:19] * XorA should get his hair hot pink to make it wasier to spot :-D [10:19] I remember how someone asked me 'why XorA paints his fingernails' ;d [10:19] neure, your paste shows that you dont have a second partition [10:20] ogra, that is correct [10:20] libnih - for all those with NIH problem? [10:20] hrw, yeah *g* [10:20] hrw: your answer? [10:20] ogra, that is sdcard with ubuntu installer on it, created by script by rcn-ee.. [10:20] neure, then how do you expect it to boot from mmcblk0p2 ? [10:20] XorA: because he can [10:21] well he doesnt tell what i should have as bootargs :( [10:21] hrw: nice catch :-D [10:21] XorA: I do not care how people look as long as they are fine with it [10:21] i tried to change it to p1 but that didnt help [10:22] hrw: this is why I like working in linux, in the windows world I would always be the freak [10:22] * XorA isnt good with suits [10:22] neure, for netinstall you dont set root= [10:22] it needs to load the initrd as rootfs [10:24] ogra, http://www.pastie.org/986979 -- without root [10:24] how did you load the initrd ? [10:24] i dont know... [10:25] :) [10:26] http://www.pastie.org/986986 [10:26] like this [10:28] bootcmd=mmc init;fatload mmc 0 80300000 uImage;bootm 80300000 [10:28] thats my bootcmd i wonder if thats ok? [10:28] thats definately not right [10:28] great [10:28] wait for rcn-ee [10:28] what should it be?) [10:28] i guess he has a boot.scr file somewhere [10:29] i have no clue about the unofficial images [10:35] lunch [10:36] neure, what's the problem? [10:49] asac, x-loader and u-boot with XM support are in maverick now [10:50] ogra: Cool [10:50] * ogra ponders about the x-loader-omap4 versioning [10:51] lool, i have x-loader-omap4-L24.6-p1git20100520 what would you do with the L ? [10:51] (its TI versioning for the omap4 branches, u-boot has that too) [10:52] i'm inclined to just ignore th elintian warning, but we'll get probs if we ever have a unified x-loader package (not that i expect that to happen) [10:53] ogra: the L? [10:53] -L24.6-p1git20100520 [10:53] is the version [10:53] L24.6-p1 is what TI calls the release [10:53] It shouldn't be a problem for pdkg [10:53] dpkg [10:54] You can have as many dashes as you like [10:54] well, usually x-loader is 1.1.4 or some such [10:54] It would be good to understand how they relate [10:54] Is it a 1.1.4 + patches? [10:54] if we get a unified u-boot or x-loader package we'll struggle with the versions [10:55] ogra: I understand x-loader is a TI only project, so eventually everything will go upstream, right? [10:55] i think its based on 1.1.4 but TI always uses L (for linux) 2.6 (kernel main version) -p* published version [10:55] x-loader will even vanish soon and just become a binary header to u-boot.bimn [10:55] *bin [10:56] but still, the u-boot-omap4 version has the same issue [10:56] i'd actually like to use their versioning, but once jcrigby is done and we have a unified numeric version it will force an epoch [10:56] ogra: If you're taking full tarballs from them, make sure the origin is reflected in the source package name and use their versions [10:57] at least for conflict/replacing the binary versions [10:57] Well we could use different binary package names and provide an upgrade path [10:57] i expect jcrigby's work to result in a u-boot source but u-boot-omap3 and omap4 binary packages [10:57] hmm [10:57] indeed, that would work [10:58] ok, i'll keep the original TI versioning [10:58] lool, thanks [10:59] (effectively it doesnt matter anyway, i dont think anyone will actually use the binaires directly, they are for the image build system) [11:00] ogra: flash-kernel on omap is kinda destructive [11:00] asac, yes, that'll change for maverick [11:00] why dont you use the defaults that are in the uboot source? [11:00] that looks really generic and good enough [11:01] asac, because upstream u-boot has no concept for an initrd [11:01] and we cant use boot.scr if we use NAND [11:01] maverick will change that completely [11:02] re [11:02] ogra: hmm. we cant use boot.scr if we use nand? why is that? [11:02] because boot.scr only works from vfat [11:02] the default that is in uboot seems to do the right thing .. checks if sdcard is there [11:02] right [11:02] otherwise uses nandboot [11:03] but has no concept of initrd at all [11:03] anyway, lucid is done [11:03] well. thats simple to add imo [11:03] maverick will change the whole concept [11:03] ogra: right. problem is that once you install lucid you cant install anything else anymore ... [11:03] since we ship our own u-boot now [11:03] you can update the setup [11:03] i'll take care for that [11:04] ah you didnt ship your own uboot [11:04] kk [11:04] if you have a better concept, feel free to do an SRU [11:04] my image has that [11:04] thats the problem then [11:04] all fine [11:04] oki [11:04] time was to short in lucid to change the whole image build system [11:05] which is why i relied on NAND [11:05] asac, btw, any final word on the naming discussion? [11:05] :) [11:06] i'll start with the tool this afternoon [11:06] (after the mobile meeting) [11:09] ogra: remind me about what naming discussion that is ;) [11:09] asac, casper :) [11:10] see the spec whiteboard [11:10] heh [11:10] ogra: EDONTCARE ;) [11:10] ok [11:10] let me see whats going on on whiteboard [11:10] * asac didnt intent do raise a long discussion [11:10] btw, i didnt bother to rename the specs since th eteam will likely be renamed soon [11:11] and i didnt want to do the renaming twice [11:11] ogra: well. but it was tough to find [11:11] ogra: can you paste the url again ;) [11:11] i wont be able to find it [11:11] yeah, agreed [11:11] https://blueprints.edge.launchpad.net/ubuntu/+spec/preinstalled-sd-card-images-for-omap [11:13] hrm, no NCommander [11:14] he had a task to care for libnih before the meeting [11:15] oh, now upstart failed completely ... no wonder i dont find libnih on the ftbfs list :P [11:16] hmm, funny, i dont see any recent upload [11:17] How does one start to build Ubuntu for ARM from scratch? [11:17] I thought about starting from Jaunty and then recompiling packages [11:17] But Jaunty requires armv5t [11:18] * ogra wonders why you would recompile [11:18] I was wondering how to inflict a lot of pain on myself and base it off armv4 [11:18] Laibsch: you have armv4t? [11:18] aww [11:18] Laibsch: armv4? strongarm? [11:18] XorA seemed to be interested in it [11:18] this is as much about getting binaries as it is about learning the steps [11:19] iirc there was a spec about rebuilding the whole distro for other arm versions [11:19] Laibsch: The way mojo did it and which is proven to work is to rebuild the archive multiple times against itself [11:19] not sure where it went though [11:19] Laibsch: grab as fast arm buildd as possible, install ubuntu and do rebuilds for armv4/eabi? nightmare task [11:19] lool: take Jaunty and rebuild the toolchain packages, then reiterate? [11:20] I see [11:20] Laibsch: Take whatever Ubuntu version, change the toolchain, rebuild the modified archive against itself, rebuild again and again [11:20] Maybe sticking with armv5t is a better idea then (probably plenty of pain as is ;-)) [11:20] I'd recommend taking the latest version of Ubuntu [11:20] (for rebuilds) [11:20] OK [11:20] or get some more recent HW :) [11:21] money seems the easiest solution for your prob [11:21] Where is the armv7 vs. armv5t vs. armv4 switch? [11:21] ogra: you don't seem to understand my prob [11:21] jaunty v5, karmic v6, lucid v7 [11:21] I don't have a particular prob [11:22] the switch is done in the compiler defaults iirc [11:22] ogra: See my comment 19:18:49 [11:22] ogra: thanks [11:22] XorA: ^^^ [11:43] lool: Is there a way to find out if the numbers of compile iterations have been sufficient or not? [11:44] debdiff $deb1 $deb2? [11:44] or something like that? [12:03] back [12:04] cwillu_at_work, i am having hard time installing ubuntu on my beagleboard C2 [12:04] can you be more specific? :) [12:04] are you using rootstock? [12:04] no... 1) installer doesnt recognize my moschip network usb to etherner adapter [12:04] 'doesnt work' [12:04] :P [12:04] 2) server installer fails to format usb stick [12:05] you should use rootstock :D [12:05] i guess so then [12:05] im trying to figure out how to boot http://elinux.org/BeagleBoardUbuntu [12:05] but i dont know what boot environment variables it needs [12:06] which, the kernel, or u-boot? [12:06] any [12:06] c2 will have a pretty old u-boot, but it should still boot fine [12:06] should i upgrade? [12:06] neure, have you done ./setup_sdcard.sh --mmc /dev/sdX --uboot beagle? [12:06] cwillu_at_work, yes [12:07] and you booted with the user button held down? [12:07] yes [12:07] you have a serial line hooked up? [12:07] yes [12:07] i can see kernel booting [12:07] but then it fails to figure out where is root [12:07] what's the kernel panic? [12:07] cwillu_at_work, iirc his u-boot doesnt load boot.scr [12:07] by default [12:08] neure, boot up, and pull up the u-boot prompt over serial [12:08] if i have some root= in boot args, it fails to mount that [12:08] neure, tell me what the bootargs are set to [12:08] bootargs=vram=12M omapfb.mode=dvi:800x600-16@60 console=ttyS2,115200n8 [12:08] neure, and the card mounts fine on a desktop, and you properly unmounted it? [12:08] bootcmd=mmc init;fatload mmc 0 80300000 uImage;bootm 80300000 [12:09] yes [12:09] the card has only one partition [12:09] boot fails it says it cant find init [12:09] neure, you'll need "rootwait root=/..." in bootargs [12:09] root= what? [12:09] root=/dev/mmcblk0p1 [12:09] ogra: hi [12:09] iirc i tried it [12:09] ndec, hey [12:09] let me try again [12:09] or mmcblk0p2 if it's a second partition [12:09] neure, you had rootwait as well? [12:10] yes [12:10] let me try and get you the message [12:10] ogra: if I want to get automatic login on the console by default, what's the easiest way? [12:10] ndec, i think asac has a script for that [12:10] ogra: I don't need root login, so any user is okay [12:10] though thats rootlogin by default [12:10] ogra: that's okay too... [12:11] asac, ?? ^^^ [12:11] ndec, /etc/init/tty1.conf sets up the login prompt on tty1 [12:11] cwillu_at_work, i get [ 24.414520] Kernel panic - not syncing: No init found. Try passing init= option to kernel. [12:12] o_O [12:12] neure, try init=/bin/bash [12:12] oh, wait [12:12] one sec [12:12] neure, your rootstock failed probably [12:12] neure, which qemu packages did you isntall? [12:13] cwillu_at_work: but i want autologin. i have automatic tests that reboot the board and run afterwards, so I need to make sure that I can autologin [12:13] i didnt [12:13] cwillu_at_work, i used http://elinux.org/BeagleBoardUbuntu [12:13] ndec, that's nice. [12:13] cwillu_at_work: what is nice? [12:13] ndec, it's almost like I was giving you exact hint you need, and you missed it ;) [12:15] neure, install qemu-kvm-extras-static, and rerun rootstock and remake the card [12:15] cwillu_at_work, i was trying "NetInstall Method" on that page [12:15] i havent run rootstock at all [12:16] how much rootstock needs diskspace? [12:16] i dont have much on my vmware :/ [12:16] neure, it needs about as much room as the final image size [12:17] ndec: we have a package for that ... [12:17] so 1 GB should be enough? [12:17] re: net install, I presume you got to the part where it said "If boot fails... upgrade x-loader and u-boot ... and clean u-boot environment vars"? [12:17] E: Couldn't find package qemu-kvm-extras-static [12:17] and ignored it, given that you asked if you should update u-boot? [12:17] asac: which package? [12:17] im running ubuntu 9 something in my vmware [12:17] ndec: see msg [12:18] ndec: you want autologin serial console? [12:18] as root? [12:18] or autologin virtual terminal? [12:18] asac: root or not, don't care. [12:18] the package isnt published yet (reason see msg) [12:18] asac, come on, it's a one line change to /etc/init/tty1.conf :p [12:18] ndec: serial or virtual terminal? [12:19] cwillu_at_work: yes, its a one line change, or some magic ;) [12:19] we opted in to magic [12:19] ugh [12:19] bad asac, BAD. [12:19]