[03:22] I'm interested in obtaining a cortex a8 (or a9) tablet... can anyone confirm that Ubuntu will work on one or the other and to what satisfaction? [03:30] Gnea: Other than us probably not having a kernel or bootloader that supports the specific hardware you're buying, the userspace will be just fine. [03:41] as for satisfaction, that depends on a number of factors, including but not limited to availability of xorg drivers [03:46] well, I'm looking specifically at the Eken line of tablets... it looks like their new line has done any extensive market saturation yet, so it may take some time [03:46] s/has/hasn't/g [03:47] infinity: being a kernel-configuring/compiling guru myself, I could probably get a kernel together for it at some point [03:47] bootloader, OTOH, is a bit out of my league :) [03:49] I'm just amazed that someone is making a tablet with ethernet *and* a usb 2.0 port built in [03:55] Gnea: It may be less about configuring and compiling and more about hacking. [03:55] Gnea: Unless the thing ships with linux anyway, then you can just cheat and use the kernel (and bootloader setup) it came with. Ish. [03:56] infinity: they claim it's got android 4.0 [03:58] Well, that's sort of Linux. ;) [03:58] Desktop Linux can run under an Android kernel, generally, with a few potential hiccups. [03:58] But it's good enough to fiddle with, at any rate. [04:00] I've got an x-server running on my droid x, and that can be interesting at times, but not too useful :) [04:31] Gnea: unless you get it for an amazing price the A10 is going to be a nightmare === heathkid|2 is now known as heathkid [08:56] ppisati, urgh ... [08:56] ogra_: that doesn't sound gut... [08:56] ppisati, so i got my install to finish (funnily there is no beagleboard entry in the flash-kernel hardware DB, i had to add it) .... [08:57] that bootloader bit didnt go so well and i now end up in the initrd shell ... [08:57] (where i could fix the issue if ..... ) [08:57] ... sadly i have no kbd or mouse now ... [08:57] but what i found strange is that you had all these problems with netinst... [08:57] could we compile the HID stuff in or some such ? [08:58] well, it were only two problems actually [08:58] dead usb? uhm... [08:58] one is the NIC driver randomly commiting suicide under load [08:58] and the other is flash-kernel actually not knowing about beagleboards [08:59] once i worked around both i had a proper install [08:59] (flash-kernel will be fixed in a minute) [08:59] what *I* find intresting is, how could *you* finish your install with flash-kernel being completely broken [09:00] but you never had any problem with smsc95xx in the other install, right? [09:00] i mean [09:00] the live or preinstalled [09:00] no, but they dont have to work under load [09:00] in preinstalled server the network only gets used to update apt once [09:00] and to find the TZ [09:01] uhm [09:01] right [09:01] in preinstalled/live desktop its similar but there also NM gets into the game [09:01] In no case is a network required, however. [09:01] but my current prob is the missing usb support in initrd [09:02] missing? why? [09:02] no idea, by my kbd definitely doesnt work :) [09:02] Is initramfs-tools not giving you hid stuff? [09:02] seems like [09:02] at least something down the stack is missing [09:02] without kbd thats hard to check :) [09:03] might be USB completely missing or just HID bits [09:03] We specifically add a bunch of hid_* modules. [09:03] It's possible that ARM has differently-named ones. [09:03] Which would be lame. [09:03] right, its more likely the HUB isnt initialized or so [09:03] do you have lights on the nic? [09:04] is the mouse/keyb powered? [09:04] Well, if you can get the system fully-booted at some point, I'd love an lsmod and such. [09:04] * ogra_ is still *massively* surprised that flash-kernel has no trace of beagle support [09:04] To compare against initramfs-tools. [09:04] I'm not shocked about the lack of support. [09:04] ppisati, neither num- nor capslock show any sight of life [09:04] Loic did tell us the new DB didn't have many entries. [09:04] * ppisati yesterday tried but was unable to get an installation at all [09:04] Luckily, omap3 and omap4 support are a sed job away. [09:05] ogra_: so IMO usb is dead [09:05] infinity, heh, be careful ... [09:05] ogra_: kernel version? [09:05] took me a while to find out that it is "OMAP3 Beagle Board" but "OMAP4 Panda board" [09:05] (note the capitalization) [09:05] ogra_: Hah, yeah, I've noticed that before. :) [09:05] ppisati, ugh, whtever is the the most current d-i image ... [09:06] ogra_: Arguably, it's a "bug" with the Pandas, but one they get to live with. [09:06] I wonder if they'll fix it with the Panda5, just to keep confusing you. [09:06] so my sed replacing "OMAP4 PAnda" with "OMAP3 Beagle" didnt work somehow :P [09:06] Or maybe go "OMAP5 panda Board" just for you. [09:06] haha [09:06] or probably Omap5 :) [09:07] Anyhow, after a night of chasing kernel SRU buggery, I think I should find a pillow and get to know it better. [09:07] Have fun with your Beagle madness. [09:07] yeah, get some sleep [09:07] heh, i will [09:08] If anyone wants to send me an xM to replace the piece of &^%!$ on my desk, I'll suddenly feel more excited about helping out with omap3. :P [09:08] talk to steve ? [09:08] we should have some spare XMs in the DC [09:08] He's the one who gave me this one. :P [09:08] Oh, were the buildds xM? [09:08] If so, we have a ton of them. [09:08] right, i think they were [09:08] not sure which rev [09:08] I'll ask elmo's crew what we can do about getting those in the hands of developers. [09:09] i *think* linaro owns them though [09:09] we might need to ask there [09:09] oh [09:09] this boot has a lot more info ! [09:09] Well, rev is less exciting. All xMs are 1GHz/512MB, right? [09:09] udevd segfauilting doesnt help much with having input devices i guess [09:09] ja [09:10] Which sure beats the C4 I have. [09:10] wow [09:10] "mounting /dev on /dev/root failed: no such file or directory" [09:10] err /root/dev indeed [09:11] I assume root isn't mounted? [09:11] that somehow smells like i dont even end up in the initrd [09:11] Or not populated yet? [09:11] shouldnt matter for mount if its empty or not [09:12] its either /dev missing completely or /root/dev [09:12] at the top of the screen i see the end of an oops, something went seriously wrong here [09:13] aha, this time /dev is there but it cant find fstab (which the new flash-kernel needs to read the rootfs UUID from) [09:14] ogra_: but why you don't have any problem with live images then? [09:15] no idea [09:15] i mean [09:15] i havent had a successfull install, i cant really say if i would have probs [09:15] i'm ok if errors are coherent [09:15] but this way... [09:16] no wait [09:16] the installer worked in both [09:16] didn't you say live image was ok? [09:16] apart from the NIC issues which dont matter in live [09:16] no, live image got me into a live session instead of the installer [09:16] and the bioard doesnt have enough ram to run both [09:16] and that's ok, right? [09:17] i mean [09:17] so i didnt bother at all to attempt an install [09:17] ah ok [09:17] the live session itself didnt have any issues [09:17] ok we need a way to bisect this [09:17] both, live and d-i run from SD [09:17] the final install has root on USB [09:17] i'm pretty sure we have some USB issues here [09:18] which the missing kbd in initzrd points to as well [09:18] i thin kwe have more than usb problems here [09:18] see my email with you and Ming [09:18] yes, that too [09:18] ok [09:18] we need to bisect this [09:18] USB might just be a fallout [09:18] right [09:18] how can i recreate the netinst image? [09:18] recreate ? [09:19] i want to recreate one with pure vanilla kernel and omap2plus_defconfig [09:19] so to see if upstream has it too [09:19] thats quite effortful, you need to build d-i [09:19] any way i can "open" the img and replace uImage&c? [09:19] i would suggest trying with the preinstalled server image instead [09:19] no [09:20] uhm [09:20] ogra_: Hrm? A netboot image? You can just write it to an SD, and then replace the bits you don't like. [09:20] It's just a vfat. [09:20] well, the installer *is* your initrd ... thats the prob here, you couold indeed replace the kernel in there but all your udebs wouldnt work and the installer would look for different versions all over the place [09:21] infinity, and the hardcoded versions ? [09:21] (Assuming all you want to replace is the kernel) [09:21] modules etc [09:21] If you're replacing the initrd, yeah, that's harder. [09:21] well, just replacing the kernel will make the initrd contents incompatible [09:21] at least all that contain modules [09:21] Unless you replace it with a monolithic kernel ;) [09:21] indeed :) [09:23] * ppisati is thinking... [09:25] so, problem crops up *only* after the installation, right? [09:25] i never tried netinst in my rev a actually [09:27] yes, on the installed system [09:29] ogra_: so you should encounter it with the preinstalled image too, right? [09:30] ppisati, hmm, so not finding / is my fault [09:30] i messed up the partition when fiddling with flash-kernel [09:30] err, or not [09:30] ignore that [09:30] (i looked at the wrong device) [09:31] ogra_: ok, i'll finish with an O/omap4 kernel release [09:31] k [09:31] ogra_: and then i'll try with my rev c and rev a [09:31] the preinstalled image [09:31] * ogra_ will dig further [09:32] bah, no installer logs were saved [09:32] it looks like a mess [09:32] no, thats normal if the installer cant do its things in the right order [09:33] and if i hack stuff in inbetween [09:34] i know it's a pain, but i'll have to ask you to retry with preinstalled then [09:34] we don't what's going on with this netinst [09:35] *don't know [09:35] i do and want to fix these issues (thats why i picked netinst ) ... but yeah, no prob in testing preinstalled too [09:37] ppisati, we should wait for rcn-ee to show up, he isially works with the bleeding edge kernels for his elinux ubuntu beagle images [09:37] *usually [09:37] yep [09:37] he might know about any issues [09:37] we can check with him if he knows anything is broken upstream [09:38] thats what i mean :) [09:42] * ogra_ sees musb_hdrc.fifo_mode=5 in rcn's boot.scr ... but that seems to be for beagle A-C [09:43] ogra_: where is it? [09:43] http://elinux.org/BeagleBoardUbuntu#boot.scr_-.3E_uEnv.txt [10:14] ogra_: http://cdimage.ubuntu.com/daily-preinstalled/current/ has only ac100, no omap3 [10:19] ppisati, ubuntu-server/daily-preinstalled [10:20] ogra_: ok, found [12:53] ogra_: perhaps i'm crazy but, my preinstalled omap3 here [12:53] ogra_: i mean, the installation process [12:53] ogra_: it never ends [12:54] ogra_: it keeps restarting [12:58] i'm not kidding, it never edns === scottb` is now known as scottb [13:07] and it restarted again [13:07] ... [13:10] weird [13:11] is there an emergency console available? [13:12] not on preinstalled, no [13:12] live has auto-logged-in tty's [13:13] but not on serial [13:16] uhm :( [13:16] then i can't even complete the installation and see what's wrong... [13:16] where does it restart ? [13:17] somewhere in the middle ? [13:17] after installing all the packages [13:17] it goes back to the first screen [13:17] installing packages ? [13:17] on preinstalled ? [13:17] (Choose the language for the installation) [13:17] wait [13:17] let me try again [13:18] so i can tell you exactlty what's the title of the last dialog [13:18] if the installer gets thorugh far enough you should be able to log in on a tty with the user you created [13:18] i'm back to "Selecy a language" now [13:19] it kept all my settings (Full user name, user name, hostname, etcetc) except for the pwd [13:19] if you are sure it went far enough, put the SD into your desktop and remove /var/lib/oem-config/run [13:19] ok [13:19] now i'm at "Software selection" [13:19] then boot it again and apt-get purge ubiquity [13:19] i select "Basic Ubuntu server" + "OpenSSH server" [13:19] right [13:20] Installing system [13:20] check if you can log in with that user... if debconf did set it up already you should eb able to do something with that system [13:20] Configuring keyboard [13:21] Configuring ah! [13:21] ERROR: The OEM installer failed. [13:21] first time [13:21] after 7/8 restarts [13:21] at least i've a console now [13:21] great [13:21] check if the user works [13:21] i'm # [13:22] if so, remove the file and the ubiquity package [13:22] nope [13:22] being in # should pretty much enable you to check for the user ;) [13:22] /etc/passwd doesn't have my user [13:22] crap [13:22] well, for debugging you should be able to just use adduser (still remove the file and the ubiquity package) [13:22] dmesg doesn't show anything [13:22] i mean, no oopses [13:23] or other crap [13:23] /proc, /sys, /dev, etcetc [13:23] are all mounted [13:23] and are ok [13:23] ah! [13:23] /var/log/oem-instaler [13:25] also syslog should have info [13:25] might simply be ubiquity being broken ... there were some bugs that were only recently fixed [13:25] http://paste.ubuntu.com/1078029/ [13:25] most likely the fixes arent on this image [13:33] http://people.canonical.com/~ppisati/Jul062012_omap3_preinstall_brk/ [13:33] ogra_: ^ [13:33] permission denied for syslog [13:34] retry [13:34] well, console-setup seems pretty unhappy [13:34] right [13:34] ubiquity.install_misc.InstallStepError: Plugin console_setup failed with code 1 [13:35] right [13:35] in the oem-config log there more warns/errors [13:35] i guess thats an old bug [13:43] but still relevant [13:43] who is the ubiquity guy? [13:43] ogra_: ^ [13:44] #ubuntu-installer ;) [13:57] ogra_: ok so, i'll restart omap3 testiong on monday when the flash-install fixes hopefully landed [14:02] just uploaded them [14:02] they should be in the next build [14:07] ppisati, have a look at the last comment of bug 709245 [14:08] Launchpad bug 709245 in linux-linaro "ARM SMP scheduler performance bug" [Medium,Fix committed] https://launchpad.net/bugs/709245 [14:08] i wonder if thats related to the smsc95xx issues i see on omap3 [14:09] (it has clearly nothing to do with the scheduler i think) [14:18] ogra_: i agree there's something rotten in omap3 ATM [14:18] well, thats omap4 though [14:19] i thik there is either a general usb prob on the omaps or smsc95xx has an issue [14:19] definitely could be [14:20] as soon as we have a preinstalled working [14:20] i want to try omap2plus_defconfig + vanilla [14:22] sounds good [15:27] helloo all [15:28] I am trying to get perf tool to work on any ubuntu release [15:28] I installed the linux tools but perf tool returns with error "open_counter" returned with error 19 [15:28] did any one get the perf to work on any ubuntu release or daily builds? === prp^2 is now known as prpplague === deffrag_ is now known as deffrag === zyga is now known as zyga-afk [17:39] GrueMaster, ogra_, I have built the 3.2.14 kernel with snd-soc-overo builtin, and it does seem to work. Thanks very much! I'm getting an odd delay on playback.. === rsalveti` is now known as rsalveti [18:20] tim__: Excellent! [18:21] Not sure what to tell you on the delay/ [18:21] its an issue with puredata- === zyga-afk is now known as zyga [21:38] Hi, I have a question about Robert C Nelson's Ubuntu images on Beagle, sorry if this is the wrong channel [21:39] I booted the latest pre-build Oneiric MMC card from here http://rcn-ee.net/deb/rootfs/oneiric/ [21:39] go for it. ;) [21:40] Thanks! Then I built replacement kernel .debs from https://github.com/RobertCNelson/linux-dev.git [21:40] ah, 3.5-rc ? [21:40] display is broken... [21:40] 3.2.0 [21:41] I did a git checkout of a previous version [21:41] use head of https://github.com/RobertCNelson/stable-kernel latest 3.2.x [21:41] OK I will [21:41] linux-dev is more experimental, and stabilized when moved to stable-kernel [21:41] Anyway I copied the .debs onto the Beagle and installed them with dpkg -i [21:42] (worked fine except for one file conflict in linux-firmware) [21:42] but when I reboot, I am still using the old kernel [21:42] don't bother with that firmware packg.. ;) [21:42] OK cool. Do I need to update UBoot somehow? [21:43] use the "./tools/install_image.sh" [21:44] OK I have tools/load_uImage.sh [21:44] it'll install directly to the sd card.. ;) [21:45] Looks like I need to run ./build_kernel.sh instead of ./build_debs.sh then. Trying that now. [21:46] Thanks for your help! [21:46] deb is only useful if you don't have access to the machine. ;) [21:47] Oh I see... if you want to install into a prebuilt image [21:48] ./tools/rebuild.sh running now... on my powerful Atom CPU ;-) [22:04] just a note, that only rebuilds the tree, so if you do a checkout another tag or something, that doesn't udpate the kernel/patachset to the specific checkout.. [22:09] OK... so I should do build_kernel.sh instead of rebuild.sh? [22:10] to give you an idea.. when i'm making small patches/tweaking the config i use the "rebuild.sh"... when starting from scratch with a new checkout use the full "build_kernel.sh"... [22:11] otherwise, the "KERNEL" directory isn't going to be updated with your actual checkout, it'll still contain the old patch/config setup..