**** BEGIN LOGGING AT Thu Dec 06 02:59:59 2012 Dec 06 07:49:04 good morning Dec 06 08:24:18 anyone know if the raspberry pi has vfp on it? Dec 06 08:28:49 morning all Dec 06 08:34:15 feasty: morning. Dec 06 13:03:22 hello Dec 06 13:03:33 yo Dec 06 13:25:14 I'm trying to run Ubuntu on my pandaboard, but the 12.10 release only had an installer image, not a preinstalled image. I want to run the rootfs off of my SD card, not a USB stick/disk, is there any way to achieve this? Dec 06 13:48:29 sspiff: have you tried building a linaro image? Dec 06 13:48:59 sspiff: Unofficially, you can use the netboot image, and when d-i/partman asks where to install, tell it to use the "largest contiguous free space". Dec 06 13:49:12 sspiff: Officially, we don't support installing to SD, as the user experience is awful. Dec 06 13:49:44 has anyone tried upgrading the nexus7 from 12.04 -> 13.04 recently? Dec 06 13:49:52 via dist-upgrade Dec 06 13:52:44 sfeole, dholbach did and found a bug, should be fixed by now Dec 06 13:52:57 ogra_: with bluez ? Dec 06 13:53:28 the bluetooth fix wasnt uploaded yet Dec 06 13:53:48 and wemdont have brcm-patchram in raring yet Dec 06 13:54:04 *we don't Dec 06 13:54:24 how to recover a EEE Pad TF101 with battery issues? gets into APX mode but error -71 Dec 06 13:54:34 battery or cable, not sure Dec 06 13:54:38 there were a couple of problems 1) onboard needs an upload to raring (I'm in touch with upstream), 2) bluez did not start/stop correctly, 3) nexus7-firmware wasn't pulled in (fixed by infinity), 4) do-release-upgrade did not work (seems fixed in my last try), 5) unrelated to upgrades, but during an installation I couldn't enter anything into oem-config's text-boxes, neither with onboard, nor attached keyboard Dec 06 13:54:51 sfeole, ogra: ^ that's all I noticed Dec 06 13:55:17 yeah, the last one is really bad Dec 06 13:55:24 did the unity stack land in standard raring already? or should we advise to use the daily-build still? Dec 06 13:55:41 landed yesterday Dec 06 13:55:48 yes baby! Dec 06 13:56:12 in 1.5h there should be images with it Dec 06 13:56:37 (and wit the new serial console by default) Dec 06 13:57:22 no more wlan needed for debugging ;) Dec 06 13:57:42 asiekierka, this probably isn't the place to get support with that, but usually you are looking at an issue with the usb cable Dec 06 13:58:17 lilstevie: there's no place to get a support with that and this looks like the best bet seeing as you're here often Dec 06 13:58:30 #asus-transformer looks kinda more suited Dec 06 13:58:33 oh Dec 06 13:59:38 asked. Dec 06 14:01:01 * lilstevie heads to bed Dec 06 14:02:52 for everyone else watching i meant 12.10 -> 13.04 , I just noticed that Dec 06 14:03:29 heh Dec 06 14:07:25 does this: https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1045741 affect ubuntu on nexus7? Dec 06 14:07:26 Launchpad bug 1045741 in firefox (Ubuntu) "since the switch to unity 3D and the GLES driver by default firefox on the pandaboard reders extremely slow" [Undecided,New] Dec 06 14:09:16 my firefox works fine here Dec 06 14:09:52 * ogra_ is currently typing from his nexus Dec 06 14:10:59 i think firefox is a bit slow, but that could be due to the arm isn't as fast as a desktop...:P Dec 06 14:17:57 I think that ogra_ is spoiled by ac100 speed Dec 06 14:18:33 lol Dec 06 15:04:01 arg still hangs for me when replacing bluez Dec 06 15:04:38 sfeole, I had to kill the process, twice - once for stopping and another time for starting the service Dec 06 15:07:45 dholbach: yea, that did it for me, what I will do is document this for now, I know we are recommended a fresh install to raring but you know there will be some people trying the upgrade path ;P Dec 06 15:08:26 upgrade path == dist-upgrade Dec 06 15:08:30 yeah, I did it too, as the installation from the image did not work for me Dec 06 15:08:35 upgrades should definitely be functional Dec 06 15:09:08 so thats a high prio bug Dec 06 15:09:17 ogra_: agreed, my installation is now continuing .. I'm already filing the bug now Dec 06 15:09:31 great, thx Dec 06 15:12:21 sadly the touchscreen still dies after a while :( Dec 06 15:24:52 hmm the nexus7-firmware0.2 package looks like it was pulled down and made me confirm it, but it failed to install Dec 06 15:25:09 looking into why Dec 06 15:34:41 infinity, ogra_ I am just curious, is armhf going to switch from ports.ubuntu.com ? Dec 06 15:35:20 i dont think so, but ask infinity he might know more ans part of the release team :) Dec 06 15:35:27 *as Dec 06 15:36:32 janimo: Probably not for a long while, if ever. It has nothing to do with official or unofficial support, it has to do with amount of traffic. Dec 06 15:36:51 and mirrors Dec 06 15:36:52 janimo: Mirrors don't want to mirror a whole port just for the sake of 1% of users. Dec 06 15:37:01 ogra_: Yes, mirror traffic was what I meant. Dec 06 15:37:07 yup Dec 06 15:38:17 sounds reasonable Dec 06 15:51:46 sfeole, "screen /dev/ttyACM0 115200" could you run that on your PC with the nexus attached via USB ? hitting enter should get you a login (needs the latest kernel nad the latest default settings) Dec 06 15:52:13 it works for me, just want to make sure it works for everyone Dec 06 15:52:25 ogra_: http://www.elinux.org/PandaBoard-NC Dec 06 15:52:34 ogra_: think there would be any market interested in that? Dec 06 15:53:14 ogra_: i will try after I get upped to Raring Dec 06 15:53:29 yeah,, needs raring indeed Dec 06 15:53:56 ogra_: will let you know Dec 06 15:54:30 prpplague, microsd and no ethernet ? Dec 06 15:55:01 might be intresting for beagle guys that want to go up one level Dec 06 15:55:19 ogra_: ethernet is there, just no connector for it, you have to wire it up yourself Dec 06 15:55:27 yeah Dec 06 15:55:30 ogra_: more for doing hardware dev Dec 06 15:55:36 likewise for the USB hub Dec 06 15:55:41 ogra_: yea Dec 06 15:55:54 ogra_: you can use the standard usb panel connectors Dec 06 15:56:08 well, surely intresting for getting a panda into a slim case Dec 06 15:56:18 ogra_: yea Dec 06 15:56:27 but i wouldnt expect massive sales from it i think Dec 06 15:56:52 seems pretty special cased ... will it be the same priice as a "normal" panda ? Dec 06 15:57:57 cheaper Dec 06 15:59:00 that might be a sales argument then ... i personally would prefer a normal panda if i dont plan to put it into any case or so Dec 06 16:00:11 having the little feet, and all sockets on board makes it easier to use for me ... if i would want to build a server with 10 pandas in a 4u case i would pick the NC Dec 06 16:00:37 or when putting a panda into an ac100 shell :) Dec 06 16:01:41 ogra_: yea that is the thought process Dec 06 16:01:44 brb, need a reboot Dec 06 16:28:52 janimo, ogra_: so is usb-serial turned on in today's image? I'd like to give it a try today if possible Dec 06 16:29:07 plars, see above :) Dec 06 16:29:25 ogra_: do you have the linux-firmware pkg installed on your Raring image? version 1.98? Dec 06 16:29:41 plars, as long aas you have the latest kernel and latest ubuntu-default-settings-nexus7 Dec 06 16:29:50 ogra_: awesome, thanks! Dec 06 16:30:00 sfeole, indeed i do, else i wouldnt have wlan Dec 06 16:30:16 and linux-image-nexus7 Dec 06 16:31:12 ogra_: i'm trying to find out what each pkg does, linux-firmware-nexus7_0.2 and linux-image-nexus7 Dec 06 16:31:48 linux-image-nexus7 is a metapackage that depends on the latest kernel Dec 06 16:32:15 linux-firmware-nexus7 carries the closed binary firmware bits used on the nexus Dec 06 16:32:30 ok Dec 06 16:32:42 both need to always be installed Dec 06 16:33:36 plars, crap, i think we're missing a securetty entry, i get a prompt but cant log in Dec 06 16:44:32 * sfeole cheers as he upgrades to raring from Q, Dec 06 16:44:49 ogra_: i just filed a 2nd bug.. #1087335 Dec 06 16:45:04 bug #1087335 Dec 06 16:45:05 Launchpad bug 1087335 in ubuntu-nexus7 "linux-firmware-nexus7_0.2 fails to install while dist-upgrading from Q -> R" [Critical,Confirmed] https://launchpad.net/bugs/1087335 Dec 06 16:45:18 janimo, ^^^ Dec 06 16:46:35 thats why i was asking about those packages Dec 06 16:46:45 ogra_, are we still supposed to support Q->R upgrades? I thought 12.10 were play only demo images and we should recommend fresh install on R Dec 06 16:46:52 it would save us a lot of trouble Dec 06 16:47:11 janimo, well, we said upgrades should work, even though we dont encourage them Dec 06 16:47:44 ogra_, would we need conflicts markers in the package just because we used linux-firmware in the PPA? Dec 06 16:48:28 I thought the whole point of 12.10 was here, you get to keep both pieces and we do not spend time on anything related besides initial UDS bringup Dec 06 16:49:01 ogra@nexus7:~$ dpkg -l |grep firmware Dec 06 16:49:01 ii linux-firmware 1.98 all Firmware for Linux kernel drivers Dec 06 16:49:01 ii linux-firmware-nexus7 0.2 all Firmware for the Nexus 7 tablet. Dec 06 16:49:23 janimo, conflicts wouldnt help Dec 06 16:49:32 since you want both installed Dec 06 16:50:03 oh, wait, conflicts ony Dec 06 16:50:09 *only Dec 06 16:50:13 yeah, that should work Dec 06 16:50:33 or rarther replaces in that case Dec 06 16:50:54 well whatever marker says some files moved to another package Dec 06 16:51:10 what if we apt-get install linux-firmware Dec 06 16:51:18 this will remove the nexus binaries Dec 06 16:51:25 and then apt-get install linux-firmware-nexus7 Dec 06 16:51:43 so make sure we do not let order be determined by apt Dec 06 16:51:49 yes, that way would work, but isnt how apt does it Dec 06 16:51:58 or even more hackily backport linux-firmware-nexus7 to Q PPA Dec 06 16:52:10 just a replaces should do Dec 06 16:52:21 ogra well would apt do anything else beyond installing the single specified package? Dec 06 16:52:44 ogra_: "screen /dev/ttyACM0 115200" drops me into a login prompt ;) Dec 06 16:52:49 Replaces for a PPA package - so carrying some cruft in a baely opened series feels ugly to me :( Dec 06 16:53:03 sfeole, but can you get in ? i seem to have issues with that here Dec 06 16:53:32 I think people insatlling 12.10 images are really knowledgable anyway and I would rather they sort the upgrade out by themselves or guided by a wiki Dec 06 16:53:41 janimo, well, its a minor change that will save us lots of silly support questions Dec 06 16:57:38 ogra_: yes i'm in, I had to hit "enter" initially to display the password prompt Dec 06 16:57:55 yeah, thats normal Dec 06 16:58:10 ok, i'll blame my chromebook (client side) then Dec 06 16:59:10 http://pastebin.ubuntu.com/1415103/ Dec 06 16:59:10 ogra_: ^^ Dec 06 16:59:45 perfect, thanks for checking ! Dec 06 16:59:49 plars, see above Dec 06 17:02:05 bad chromebook! Dec 06 17:02:28 heh, yeah, i blame hrw Dec 06 17:02:50 he does all the chromebook stuff, so it must be his fault :P Dec 06 17:49:52 ;D Dec 06 18:50:21 fwiw, i blame him for everything. if it weren't for him talking about how awesome the chromebook is, my checking account wouldn't be so light Dec 06 19:12:33 ogra_: so I have the usb serial stuff working fine (Thanks!) but I'm having a bit of trouble getting preseeding to work. I've tried adding my preseed file to the initrd that I'm pushing to it, but I'm guessing that by the time it gets there, it's already remounted userdata as root so I'm not finding it. Any suggestions? Dec 06 20:32:02 \o/ usb serial ftw Dec 06 20:32:08 * vanhoof just flashed todays image Dec 06 20:44:52 vanhoof: yes, it's working pretty nicely I think, but any ideas on how to preseed the remainder of the install? Dec 06 20:45:37 vanhoof: I've tried sticking a preseed in the initrd, as well as in the rootfs, no luck so far, but tbh I don't have much experience with preseeding just the oem-config portion so I'm not confident that I'm doing it right Dec 06 20:46:19 plars: i've never done a oem-config preseed myself :\ Dec 06 20:46:31 * vanhoof takes a look Dec 06 22:33:05 hey, anyone know if the raspberry pi has vfp? Dec 06 22:33:37 yes it does Dec 06 22:48:08 sweet. Dec 06 22:54:32 mjrosenb: It's v6, however. Dec 06 22:54:47 that indeed Dec 06 22:55:34 mjrosenb: Which means that neither Ubuntu nor Debian armhf will run on it (but there's an unofficial Debian v6+hardfloat port called "raspian" just for the Pi) Dec 06 22:57:24 infinity: yup. I need an armv6/vfp machine to test on. Dec 06 22:57:39 infinity: and it certainly sounds like the raspberry pi fits the bill. Dec 06 22:57:40 What an odd requirement. :) Dec 06 22:57:56 well, I already have like 10 armv7 devices Dec 06 22:58:22 As long as you like testing in slooooow motion, you'll have fun with the Pi. Dec 06 22:58:26 It's cute, at least. Dec 06 22:58:43 what is it clocked at? Dec 06 22:58:57 I thought it was like 600 mhz Dec 06 22:59:43 ok, 700 mhz Dec 06 22:59:47 Yeah, 700. Dec 06 23:00:07 so like half the clock rate of the upper end of my armv7's Dec 06 23:00:15 Still, a 700MHz ARM11 isn't anywhere near 7/10 of Cortex-A8. Dec 06 23:00:23 true. Dec 06 23:00:31 And nowhere close to 7/10 of a 1G A9. Dec 06 23:00:44 mjrosenb, why would you want to test on an armv6+vfp Dec 06 23:00:47 The newer model with 512M of RAM might prevent urges to shoot yourself, though. Dec 06 23:01:08 I'm getting my fix for that with a roku 2 :p Dec 06 23:01:19 at least that is optimised for arm11 Dec 06 23:01:41 and the one good thing I have heard about the pi, or really the bcm soc is that it is good at media Dec 06 23:01:48 lilstevie: because firefox is going to attempt to support armv6, and I would like to make sure that we can do that. Dec 06 23:02:00 oh I see Dec 06 23:02:28 but we are requiring armv6 + vfp Dec 06 23:02:47 because javascript uses doubles *EVERYWHERE* Dec 06 23:03:25 Raspbian on a Pi may well be just the right testbed, then. Dec 06 23:03:32 It ain't fast, but it should do the trick. Dec 06 23:03:43 (For the love of god, though, don't compile on the thing) Dec 06 23:03:53 haha Dec 06 23:04:14 compiling on dual and quad core a9s is painful enough Dec 06 23:04:28 I've come to terms with A9 performance. Dec 06 23:04:47 Then again, I came from 68k porting, so I'm a bit weird. Dec 06 23:05:36 infinity: I can't remember the last time I built on an a9 Dec 06 23:05:39 Sadly, software is like goldfish, and it seems to bloat to fit your computer. Dec 06 23:05:42 I cross compile basically everything Dec 06 23:05:47 except on my a15 Dec 06 23:06:11 Cause I swear today's (very fast, in comparison) ARM cores don't really do anything any faster than a 68k 10 years ago performing similar tasks. Dec 06 23:06:26 126 minutes to build firefox! Dec 06 23:07:31 I did get a refresher course in just how many crazy drivers we build in distro kernels the other day. Dec 06 23:07:56 Was doing some benchmarking with an upstream defconfig (cause building the package is just plain no fun as a benchmark)... Turns out that a defconfig on a Panda is about 10 minutes. Dec 06 23:08:06 Whereas our distro kernels take hours. Dec 06 23:08:10 infinity: most distro kernels just enable everything possible as a module Dec 06 23:08:28 Zero_Chaos: Yes, and I know that. I just didn't realise how many modules that IS these days. Dec 06 23:08:34 Zero_Chaos: Apparently, the answer is "a lot". Dec 06 23:08:42 infinity: yeah, make -j64 is nice Dec 06 23:08:56 Not on a Panda, it's not. :P Dec 06 23:09:05 infinity: I would imagine not ;-) Dec 06 23:10:23 Still, a defconfig in 10m isn't bad. It wasn't that long ago I used to build my own barebones monolithic kernels on x86 kit, and that was around 30m. Dec 06 23:10:48 (Might have been around P200ish days?) Dec 06 23:11:03 Okay, maybe I'm getting old, when I imply that the P200 wasn't that long ago. Dec 06 23:12:02 -j5 is about as far as I can take the prime before more jobs make things slower Dec 06 23:12:12 (tegra3) Dec 06 23:12:54 http://paste.ubuntu.com/1413025/ <-- My unscientific quick tests on a PandaES. Dec 06 23:12:55 and also agreed, computers may have gotten faster, but the software has gotten bigger Dec 06 23:13:17 infinity, heh Dec 06 23:13:52 building a kernel deb with pbuilder takes about an hour and a half on my trimslice Dec 06 23:13:52 -j64 would probably make it weep. Dec 06 23:15:13 lilstevie: Do you talk to the Trimslice guys much? Dec 06 23:15:38 lilstevie: I'm kinda curious if they plan to rev the hardware with newer and shinier SoCs, or if it was a one-shot deal. Dec 06 23:15:38 infinity, not really, the only times I have had questions for them I get "why would you want that" Dec 06 23:16:04 Cause I like the idea and the form-factor, but it's obsolete by now. Dec 06 23:16:26 yeah, someone has asked that question, and they said no comment Dec 06 23:16:44 infinity: -j64 brings most systems to their knees pretty quick. -j brings nearly all systems to their knees Dec 06 23:17:12 Zero_Chaos: The build machine we use for internal kernel test builds seems to have a sweet spot around -j256... Ish. Dec 06 23:17:32 infinity, http://emea.kontron.com/products/boards+and+mezzanines/embedded+motherboards/miniitx+motherboards/ktt30mitx.html has my attention, except I am getting to the point where I would rather avoid tegra Dec 06 23:17:41 (It's a 32-core i7-base Xeon with stupid amounts of RAM and a shamelessly fast stripe) Dec 06 23:18:17 infinity, although most of the crippling in tegra systems is slow eMMC Dec 06 23:18:27 infinity: yeah on that I'm guessing you could just rock "-j" and let it fork as much as it likes. even on kernel builds like that I can't keep it >200 jobs for very long so if you have that many procs give up and use -j Dec 06 23:18:42 lilstevie: Curious that that's only clocked at 900MHz. Dec 06 23:18:48 infinity: I have 32 amd cores and build in RAM Dec 06 23:18:53 that board would be an interesting test cause it does have SATA Dec 06 23:19:02 infinity, yeah, I noticed that too Dec 06 23:19:04 infinity: so about 1/2 the intel system you describe Dec 06 23:19:07 infinity, I wonder if it is for cooling Dec 06 23:19:47 infinity, throw a heatsink on it and go nuts Dec 06 23:20:08 on the tf201 they have managed to bring it out to 1.8GHz with all 4 cores active Dec 06 23:20:18 but it does make the metal case nice and toasty Dec 06 23:20:29 lilstevie: Maybe. I'm probably not going to buy any more A9s out of pocket. I suspect I'll pick up one A15 at some point with my own money, and then wait for someone to bless me with aardvarch64 silicon. Dec 06 23:20:55 lilstevie: (Not counting useful items I actually need, like phones, of course. I just mean I'm not going to keep collecting toys) Dec 06 23:21:11 infinity, yeah, I am getting a surface (waiting for it to arrive right now) but after that I am not paying for any more arm systems until aarch64 Dec 06 23:22:05 infinity, the only exception to that may be a quad core S4, cause krait is crazy fast compared to A9 Dec 06 23:24:23 but yeah, I am going to live with my current arm machines for as long as possible Dec 07 00:37:20 Hey anyone here been messing around with ubuntu on the Nexus7? Dec 07 00:37:29 has anyone tried lxde yet? Dec 07 02:35:48 xnox: any ideas on preseeding the oem-config install stuff for n7? Dec 07 02:36:14 xnox: I'm trying to have it finish the setup (time zone, language, user, etc) on the first boot automagically Dec 07 02:37:09 xnox: I tried sticking the preseed in the initrd, and also in the rootfs, but so far haven't had much success with it, nor do I seem to be able to ctrl-alt-f1 to see what's going on under the covers Dec 07 02:37:27 xnox: but tbh, I don't know if maybe there's some difference with the preseeding that I need to account for Dec 07 02:40:51 hi guys, I've heard ubuntu doesn't run as fast on a BBxm when compared to Angstrom, is that still true today? Dec 07 02:41:45 xenome, "heard" is that quantifiable. ;) Dec 07 02:42:41 well I've seen some posts about bogo mips too but I think that has to do with the calculation Dec 07 02:42:55 bogo = bogus... Dec 07 02:42:56 rcn-ee, I believe I'm running your kernel I setup from the wiki Dec 07 02:43:19 is there a good way to benchmark? Dec 07 02:43:25 with ondemand cpufreq, the bogomips calc is useless, as it does not get updated.. Dec 07 02:44:10 well it's different now then at bootup when I cat /proc/cpuinfo Dec 07 02:44:26 xenome, what will get you, post v3.6.x 800Mhz is no longer stable when rebooting, so I had to disable that in my builds.. SO if your Angstrom is running v3.0.x it'll be at 1Ghz, where as i'm at 600Mhz max.. ;) Dec 07 02:45:05 If you don't care if it hardlocks on a softrware reset, you can reenable 800Mhz.. Dec 07 02:45:27 is that a bug in the kernel? Dec 07 02:45:36 or did something change and that capability was just lost Dec 07 02:45:58 looks for Tony Lindgren's email... Dec 07 02:47:34 xenome, http://www.spinics.net/lists/linux-omap/msg78218.html (basically we need the Nokia fixes..) Dec 07 02:47:37 even at 800 if it worked, is there a reason for not hitting 1GHz like on angstrom Dec 07 02:48:30 wait, ubuntu, not stable? unpossible Dec 07 02:48:55 Well, there was hacks to make it work with v3.0.x & v3.2.x but with the 2 or 3 omap power managment rewrites, those patches are useless.. Dec 07 02:49:19 Basicly if you want it, complain to TI. ;) Dec 07 02:49:51 does TI offer a commercial solution? Dec 07 02:50:06 for people who buy this chip, what OS do they recommend to leverage all the features? Dec 07 02:50:38 well, there was the 2.6.32-psp, then 2.6.37-psp and i think a v3.0.x based branch for those customers.. Dec 07 02:53:07 oh so the psp branches are the TI modded kernels? Dec 07 02:53:36 is it possible to setup ubuntu and just drop back to one of their older kernels, or is that a big mess? Dec 07 02:55:31 TI's psp are here: http://arago-project.org/git/projects/ Dec 07 02:57:49 sure, in theory you can go all the way back to 2.6.32 as long as DEVTMPFS is enabled.. Dec 07 02:58:29 but, it'll be a fun adventure, specially with all the gcc bugs. ;) Dec 07 02:59:20 because I'd be using a recent ubuntu dist, or because their code does not work well with gcc? **** ENDING LOGGING AT Fri Dec 07 02:59:59 2012