[00:57] hi guys [00:57] hows it going [02:29] New here, Is the Panda board the only HW platform that you can purchase to expirement with ubuntu-arm? [02:30] No, we also support BeagleXM and Frescale Quickstart development boards. [02:31] You can also run any other armv7 (Cortex A8/9) compatible system if you have your own kernel. [02:31] Any preferance as to which is the better board? From a Performance, longetivity and/or ease of use point of view? [02:33] really depends on what you want to do. BeagleXM is great for lightweight projects (Kiosk systems, cups server, etc). Panda is a dual processor with HD Video capabilities, Frescale has SATA for better disk I/O. [02:33] All are fairly easy to use. [02:35] That is the best information anyone has provided me yet. Gets me something to think about. Thank you! I must admit my interest is morbid curiousity. I dont know what I want to do with it, other than just something to expirement/test with. [02:36] Heh, I understand. I have several of each for Ubuntu testing. [02:49] pfft disks, just push it out to the san ;-P [02:49] They all have gige right? [02:50] No, 100e, [02:50] lame [02:52] Can't ask too much for a cell phone proc. :P [02:52] Oh I can ask [02:53] OTOH why would a cellphone need SATA [02:53] Last cycle, I had 7 panda's in a CEPH cluster. Painful, but still cool. [02:53] And SATA is far faster than SDHC. [02:54] But why would a cellphone need that [02:54] That's your argument against adding other whizzo features, right? ;-) [02:54] using a SATA based SSD vs eMMC? Speed. Pure unadultrated speed. [02:55] Pft, just load everything into ram [02:55] You only need the SD at boot time and nobody turns their cells off [02:55] No arm system ships with more than 1G. [02:55] 1G should be enough to 100 users! [02:56] grumble grumble [02:57] SOC vendors ship to anyone that wants a proc. It is easier to have features disabled, than non-existent. [02:57] They should ship a disabled gige controller then :-/ [02:57] I can see the Freescale part competing with Marvell for the small NAS box market. [02:58] A $200 NAS appliance ought to do gige [02:58] is the Freescale demo board the I.MX51? [02:58] I think the nic is part of the USB controller. It is on omap3/4. [02:58] mx53. [03:05] I would imagine if you use the mx53 board the easiset thing would be to output via a serial port in ssh/telnet and skimp on the video module. [04:22] Anyone using an MX53? [04:22] I built the entire BSP, but I can't find the GPU driver. Someone said it was gpu.ko but it's not there. Bit mystified :) [04:52] trying to figure out how to ssh into Beaglebone with it installed https://wiki.ubuntu.com/ARM/OMAP [05:01] also hellogl_es2 starts up but the gl widget is black === doko_ is now known as doko [08:31] plop [08:33] doko, how can I know which files are compiled in glibc? Looking at the Makefiles, I have no clue... for example in sysdeps/ieee754, there are *four* e_lgammal_r.c ! [10:01] greetings, with regard to rootstock, it now claims to be replaced by live-build (Which seems to generate iso's). Is there a HOWTO that I haven't found yet for using live-build to build a root tarball [11:49] greetings, with regard to rootstock, it now claims to be replaced by live-build (Which seems to generate iso's). Is there a HOWTO that I haven't found yet for using live-build to build a root tarball [11:49] basically I just need something to copy onto an SD card [12:06] overnight: 39 success, 9 failures (looks like the same debootstrap proxy/skew problem) [12:07] trelane, if you just want a minimal rootfs, have a look at ubuntu-core (it is totally unconfigured though) [12:15] ogra_: noted. I don't need a lot of config, though that looks like it includes some things I don't need [12:16] ?? [12:16] it is the most minimal ubuntu you can build [12:16] pulse, gtk, etc [12:17] by definition just enough OS to run apt [12:17] ok Base Core = Core [12:17] got it [12:17] no [12:17] trying to figure out what Ubuntu-Core is based on the flow chart on the core website [12:17] i'm talking about the ubuntu-core tarballs you can get on cdimage.ubuntu.com [12:18] it contains everything debooorstrap --minbase installs plus apt and its deps [12:18] ok, yeah that's small [12:18] has no users or rootpw configured nor does it have a network setup, hostname etc [12:18] tarball is around 30M... unpacked it should be between 80 and 120M [12:20] ok that's no big deal at all. I can script that and do the construction in qemu :) [12:20] thanks! [12:27] trelane: you might find that useful http://omappedia.org/wiki/OMAP_Ubuntu_Core [12:27] ndec: already found it on my own [12:27] reading it now [12:28] hey, google is cool ;-) [12:28] ndec: I'm actually an idiot, I merely have lighting fast google skills and therefore can fool most people [12:28] you guys might want to put a note on the rootstock page that ubuntu-core is now the preferred solution [12:28] i will update it soon with a section about using mainline kernel and uboot instead of the prebuilt kernel from ubuntu. but basically if you checkout any recent mainline, and build you can use that with ubuntu-core just fine. [12:29] ndec, there is a typo btw [12:30] where? [12:30] sudo C_ALL=C chroot . /bin/bash [12:30] should be LC_ALL [12:30] I'm usually a Funtoo (Gentoo fork) guy, and after a few hours of looking at the OE build system, how they attempted to emulate gentoo, FAILED TOTALLY, then stuck somethign vaguely debian like on top, I gave up [12:30] ogra_: let me check [12:30] I don't know how OE happened but if there wasn't enough alcohol involved to stun a team of oxen... I'd be shocked. [12:32] ogra_: fixed [12:32] great :) [12:32] trelane: this is probably not the right place to talk about OE, even for bashing against OE ... [12:33] noted :) === Ursinha` is now known as Ursinha_ === Ursinha_ is now known as Ursinha === rsalveti` is now known as rsalveti_ === rsalveti_ is now known as rsalveti [15:49] rbasak: did you manage to test the new kernel with the installer? [16:53] ppisati: I just ran netinstall as part of my milestone testing. Seems to have run just fine. I'll check the logs for any unnoticed residue. [16:53] GrueMaster: where can I find a basic intro to installing ubuntu on a pandaboard? [16:54] GrueMaster: cool [16:54] Riddell: I think there is a link off http://wiki.ubuntu.com/ARM [16:54] Also linked from the image download pages on cdimage. [17:10] ogra_: hi. on our private PPA we have a msg like that "A recent upload has resulted in 5 pending builds. " do you know what that means? [17:10] nope [17:10] or davidm if you're around... [17:11] ogra_: argh, not sure if it looks good or not. who should i ask? [17:11] i would assume your packages sit in the queue or some such [17:13] ogra_: another problem is that now that we start pushing kernel in PPA, we are reaching the size limit quickly. [17:14] that shouldnt be a prob, though likely needs davids approval first [17:14] you could ask in #launchpad if they can just bump the size, if they need any additional approval we can manage i guess [17:15] ok. [17:15] * ogra_ takes a look at the queue if there is anything from you in teher [17:17] ogra_: i asked on #launchpad too [17:17] ah, k [17:18] ndec, i see an alsa-utils build that has -ti in the version at https://launchpad.net/builders [17:18] (the PPA builders are the bottom ones) [17:19] seems the builders are busy beyond that, so it could well be that your packages are qeued [17:20] ah, now it also has totem and tibtfm [17:20] and uboot [17:21] ndec, ogra_ : buildd-manager was down, hence the queued builds going nowhere. Got fixed just a while ago. [17:21] ah, k [17:21] ndec: As for increasing PPA size, that doesn't need any management approval or anything, just present a case for it to launchpad folks. [17:22] (Though, they may require a slightly less hand-wavy case for non-Canonical PPAs demanding size bumps, I doubt yours will be an issue) [17:22] * ogra_ thought so ... though it wouldnt have been a prob to get such approval indeed :) === Jack87|Away is now known as Jack87 === trelane_ is now known as Guest46240 [17:35] ppisati: Did you see my earlier message? I may have experienced an irc net-split. [17:38] infinity: ogra_: thx. looks like our builds are going on now. [17:38] :) [17:38] about the size, I will request for more indeed. will ask you if i need support [17:38] dont hesitate :) [17:40] why would i ;-) [17:40] GrueMaster: nope, which one? [17:41] ppisati: I just ran netinstall as part of my milestone testing. Seems to have run just fine. I'll check the logs for any unnoticed residue. === prpplague^2 is now known as prpplague [17:41] Re: omap kernel mac address. [17:41] GrueMaster: what's that? [17:41] ogra_: That was a lot earlier. [17:42] GrueMaster, oh, then your other messages went unnoticed [17:42] ppisati: On the omap kernel, you mentioned something about a change for the mac address? Is it just enabling the "smsc95xx.macaddr=" parameter in the module or is it getting the mac from the soc die id? [17:42] There, reposted. [17:42] GrueMaster: the second one [17:43] GrueMaster: it seems i can't change the mac anymore [17:43] Cool, thanks. I'll double check it here and let you know how well it works. [17:43] GrueMaster: sudo ifconfig eth0 hw ether 00:11:22:33:44:55 [17:44] Nah, just reboot. If it stays the same, my dhcp server will know immediately. [17:44] k [17:45] While you're here, on omap4, is there a camera module built into the kernel now? oem-config is trying to take a picture but there is no camera so it only has a small section of the previous screen. [17:47] it shouldnt offer you to take the pic in the first place if there is no camere attached === rbelem_ is now known as rbelem [17:49] ppisati: no random panics after 48 test installs overnight. I'm using a kernel I built from the git source tree yesterday - not sure what the installer kernel is at the moment. So it's looking good. [17:50] ogra_: Hence why I asked. If ubiquity is detecting a camera module, it is activating that section. [17:50] It doesn't do it on omap or mx5. [17:52] i guess it looks for the actual device in /dev [17:53] GrueMaster, look for /dev/videoX [17:53] yep. /dev/video0 exists. [17:53] that might be it [17:55] Found it in dmesg. "Linux video capture interface: v2.00" [17:55] crap ... i wonder how ubiquity should find out if there is actually a camera attached [17:56] /dev/video0 might as well come from your TV card or some such [17:56] so that bug can happen on such setups too [17:56] That might be...interesting. Have it take a snapshot of a klingon or something. [17:57] heh [17:57] So, I should file a bug against ubiquity for this? [17:57] GrueMaster: yep, there's camera support in it but it should be broken [17:59] GrueMaster, yes [17:59] if there isnt one already [17:59] ok. [17:59] i can easily imagine there are a lot video capture cards out there where people see it too [17:59] * GrueMaster notes this will be the 4th bug filed in 2 days on ubiquity/oem-config. [18:00] awesome [18:00] to sad we dont have an ubiquity maintainer atm [18:00] ??? [18:00] Where's Evan or Colin? [18:00] ev doesnt do it this release [18:01] Ah. [18:01] colin and stgraber step in for him though [18:02] I was looking at that code yesterday, trying to figure out preseeding. What a mess. [18:03] it all becomes clear with enough alcohol in your blood though [18:06] * GrueMaster needs to make an emergency run to the liquor store. [18:06] This camera module, how is it supposed to report no camera detected? [18:07] dont forget to expense it ! [18:17] GrueMaster: ogra_: we did get some problems with 11.10 when using a kernel with v4l2 camera driver on panda... [18:18] some of our panda don't have the actual camera plugged, but the kernel does not detect that... and ubiquity tries to make a picture... [18:18] ndec, yes, we had issues in 12.04 too, thats why it was disabled [18:18] but not completely it seems [18:19] on panda? [18:19] yep [18:19] we have a horrible hack in ubiquity to bypass the camera detection... so if you come up with a proper fix this is better ;-) [18:20] well, as i said above, i dont get how that works on systems with TV capture cards [18:20] they surely also have /dev/video [18:20] yep, i am interested to understand how ubiquity understands there is a camera too... if you figure out... [18:21] well, tobin will file a bug, we will see :) [18:21] please subscribe tiomap-dev on the bug [18:21] k [18:28] Done and done. [18:28] Bug 924419 [18:28] Launchpad bug 924419 in ubiquity "oem-config detecting camera where no camera exists" [Undecided,New] https://launchpad.net/bugs/924419 [18:30] thx ! [18:40] ogra_: Where are the debug symbol packages for arm? ports or somewhere else? [18:41] GrueMaster: Everything's on ddebs.ubuntu.com, but (and this is a big but), not everything is represented. :( [18:41] GrueMaster: pitti keeps culling random old releases and non-primary arches as he runs out of space. [18:42] GrueMaster: Though, we seem to have armel/armhf for precise, at least. [18:43] Ok. ev wants me to run a backtrace on this ubiquity panel crash. [18:44] Since I can't just file it. [19:38] GrueMaster: did you reboot the beable? [19:39] Doing so now. Got caught up in an installer bug on omap4. [19:41] Nope, it didn't have the same mac on reboot. [19:41] It was supposed to get it from the die id, right? [19:44] nope, you get a random one [19:44] but i thought you configured it to get a fixed one [19:45] How do I configure the system to get a fixed mac address? [19:46] via /etc/network/interfaces but i can't make it work anymore, that [19:46] s why i asked, because i thought you did [19:46] There is no way to automate that, especially during a netinstall. [19:47] k [19:47] I need either a die id (like omap4) or a kernel cmdline parameter. [19:49] Die id is preferred as then I don't need to modify the preinstalled images. === dduffey is now known as dduffey_afk === robclark_ is now known as robclark