**** BEGIN LOGGING AT Mon Oct 10 02:59:57 2011 Oct 10 03:06:13 hi! i find myself compiling a kernel for ubuntu for the first time in 5 years :) Oct 10 03:06:31 is make-kpkg still the thing people use when building custom kernel debs? Oct 10 03:16:36 ive noticed android/ubuntu dual boot tablets on youtube but has anyone gotten around having to go through an external pc to change which system you want to boot? Oct 10 03:16:52 like using u-boot or CWM? Oct 10 03:17:43 always having to externally change some zip file everytime I want to `change the system I boot is not a very nice option Oct 10 03:17:44 that would be nice. Oct 10 03:18:04 hoshi411: you have a TF101; its default bootloader will switch to a "rescue" mode if you hold down voldown and power while booting; this rescue mode is simply a different kernel Oct 10 03:18:10 hoshi411: yeah it's like having to stop at a gas station every time you want your car to change gears Oct 10 03:18:27 hoshi411: thus, you can replace the rescue mode of android with a normal ubuntu/debian kernel Oct 10 03:19:06 twb: I see that is an interesting trick... but then how do you designate which kernel you want the rescue mode to boot? Oct 10 03:20:23 is there really no way to communicate with the bootloader, like on windows bios , grub , yaboot etc? Oct 10 03:20:31 hoshi411: the ASUS bootloader is hard-coded to bootstrap specific partitions for normal and rescue mode, so you pick by writing the kernel and ramdisk into the appropriate partition Oct 10 03:21:12 hoshi411: on the TF101 there is no way to communicate with the default bootloader except for that voldn thing, and u-boot currently has no support for the keyboard doc OR the volup/dn buttons, so you cannot provide any input at boot with u-boot. Oct 10 03:21:56 O_o but if you can root an android device then you can install any soft you want on it Oct 10 03:22:12 hoshi411: however if you have a dock, you can put the u-boot scripts or environment onto a SD card or USB (but not micro SD), and u-boot will do whatever the script says. So to change how booting will happen, you can put that SD card into a working box, and reconfigure it. Oct 10 03:22:24 hoshi411: uh, only within android Oct 10 03:22:35 hoshi411: yeahbut someone's gotta write it :) Oct 10 03:22:36 hoshi411: rooting android won't allow you to e.g. replace its kernel Oct 10 03:23:08 And you probably can't easily replace e.g. the shitty android UI with normal X. Oct 10 03:23:12 twb: but rooting android would allow me to run a terminal right Oct 10 03:23:30 hoshi411: I wouldn't know, I do not support or use android. Talk to the android people about that. Oct 10 03:25:47 the ideal way to do a dual boot would be to be able to select the kernel , not at startup but before shutdown Oct 10 03:26:00 if I could somehow in android select which kernel I want to boot from Oct 10 03:26:18 and then boot into it ... then from that distro also to go and select the android kernel again Oct 10 03:26:21 etc.... Oct 10 03:26:28 before shutdown Oct 10 03:26:39 that seems ideal to me Oct 10 03:26:46 That's how Macs used to work; it's a fucking pain because if the main system breaks, you're hosed Oct 10 03:27:03 IMO it's better to have it boot the right thing, and if it screws up you insert a rescue SD card Oct 10 03:27:23 s/main system breaks/current system breaks/ Oct 10 03:27:58 so... any of y'all build a custom kernel .deb lately? :) Oct 10 03:28:27 dash: yes. Oct 10 03:29:01 twb`: what do you use? when I were a lad, make-kpkg was the thing Oct 10 03:29:10 It's built into upstream now Oct 10 03:29:21 can't you just mount the boot partition and alter the boot scripts on the asus device? Oct 10 03:29:24 in particular I want a kernel with NFS server support Oct 10 03:29:32 Just do "make deb-pkg" Oct 10 03:29:35 and i'm not seeing one in oneiric for mx53 Oct 10 03:29:40 twb`: aha. Oct 10 03:30:17 hoshi411: the asus bootloader has boot partitions that are in a hokey android format; you can't mount them Oct 10 03:30:37 hoshi411: you'd have to rebuild them from source components and then reflash them onto the device Oct 10 03:30:39 was thinking about installing ubuntu on my tf101 Oct 10 03:30:46 sounds like i should wait Oct 10 03:30:56 dash: you can have 2.6.36 now Oct 10 03:31:18 dash: but it can't drive the wifi or bt or much else, so it's pretty useless Oct 10 03:31:52 a hokey android format? not ext or ntfs etc? Oct 10 03:31:56 hoshi411: correct. Oct 10 03:31:57 hoshi411: correct. Oct 10 03:32:02 hoshi411: specifically abootimg format Oct 10 03:32:33 abootimg ... hmm Oct 10 03:34:11 twb: wouldn't this allow you to read the files on that partition https://launchpad.net/ubuntu/+source/abootimg Oct 10 03:41:07 hoshi411: I doubt it. Oct 10 03:41:31 Hm, actually it looks like you're right, you could pull apart the bits from that, edit them, reassemble them, then reflash it back on Oct 10 03:42:10 But you'll still need the reflash step, which either means nvflash on an external device or (maybe) flashing it from within a working device, and the default partition table doesn't expose that partition to the OS. Oct 10 03:42:53 twb: and depending on the script that is used to select the kernel on bootup... you might be able to reflash it in a way where you wont have to reflash every time you want to change kernels Oct 10 03:43:33 hoshi411: in magical fairy land, sure. I wouldn't hold your breath waiting for it in the real world. Oct 10 03:44:43 I use yaboot on ppc but have never tinkered with grup on my pc ... but as far as yaboot goes , it is just a script that allows me to select the kernel at bootup... and I can change the script to allow for any kernel I want or as many as I want.... Oct 10 03:44:56 grup > grub Oct 10 03:45:28 arg... I really wanna get inside that boot partition Oct 10 03:45:51 Look, the short version is: bootloading on ARM is fucked up. Oct 10 03:45:58 Do not expect it to be as nice as PPC or x86 Oct 10 03:46:52 twb: yea but I cant give up Oct 10 03:46:59 I need gimp on my transformer Oct 10 03:47:16 otherwise it's almost a paperweight to me Oct 10 03:47:37 and I need the ability to boot back into android whenever I want Oct 10 03:48:00 without haveing to wait to get home just to connect it to my pc and change some zip kernel whatever Oct 10 03:48:30 twb: but thanks for updating me on the state of things on the transformer right now Oct 10 03:48:44 Im gonna try and see if I can get into the boot partition somehow Oct 10 07:24:54 im trying to get ubuntu chroot on my asus transformer Oct 10 07:25:00 do I need windows to do this? Oct 10 07:25:10 or can't I do it from ubuntu Oct 10 07:25:26 my windows partition is acting up and ... it would be nice to be able to do this from the ubutu side Oct 10 07:25:40 I think I have to root my device right? Oct 10 07:25:47 No. Oct 10 07:26:08 really. there is an easier way? Oct 10 07:26:17 hoshi411: if you want to *boot* Ubuntu, you need to follow the notes I showed you in #debian-arm. If you just want a chroot, you need to root android, and that is off-topic for this channel. Oct 10 07:26:43 Rooting android is probably substantially easier, but I don't know. Oct 10 07:26:56 yea I just wanna chroot, cause I really don't wann ahave to connect it to a pc every time I wanna switch back and forth Oct 10 07:27:05 ive been at this all day Oct 10 07:27:09 and I give up Oct 10 07:27:30 chroot will give me an easy switch back and forth Oct 10 07:27:59 would be great if there were gimp on android Oct 10 07:28:28 hoshi411: android is off-topic for this channel. I won't help you with that. Go find the android channel and ask them instead. Oct 10 07:28:40 chroot may be off topic but there is a version of ubuntu that runs on chrooted android that I have seen.... that would be on topic right? Oct 10 07:28:57 hoshi411: I suppose so, *once* you have it working Oct 10 07:29:00 ive seen it on the galaxy tab Oct 10 07:29:34 The vast majority of "get it to chroot" issues will be on the android side Oct 10 07:30:11 aright Oct 10 07:30:14 ill go there Oct 10 07:30:17 twb`: Are you suggesting you've gotten Ubuntu to boot on an ASUS Transformer? Or know of someone who has? Oct 10 07:30:34 soren: lilstevie has a working .36 boot, which I have reproduced Oct 10 07:30:41 *drool* Oct 10 07:30:57 soren: he also allegedly has .38 CrOS, but I can't reproduce that because u-boot isn't working for me Oct 10 07:31:09 soren: it seems that he has done it but .... still no one has figured out how to switch between android and ubuntu without hooking the tablet up to a pc Oct 10 07:31:13 Those things look awesome, but I'm not looking for an *extra* gadget to carry around. Someething that could replace my laptop... maybe. Oct 10 07:31:17 soren: http://cyber.com.au/~twb/doc/tf101.html contains about half my notes so far Oct 10 07:31:49 Yeah, I bought it to replace me Eee PC 1005P, because 1) quadruple the battery life; and 2) ARM is a "fun" challenge. Oct 10 07:31:56 you can dual boot Oct 10 07:32:02 at the cost of recovery Oct 10 07:32:13 but you can still switch without needing to hook it up to the pc Oct 10 07:32:18 lilstevie: dunno if you saw the scrollback, but hoshi411 doesn't seem to be interested in that Oct 10 07:32:49 lolstevie: how do you switch without having to hook it up to a pc? Oct 10 07:33:06 by having one OS kernel in the recovery slot Oct 10 07:33:11 14:17 hoshi411: the ASUS bootloader is hard-coded to bootstrap specific partitions for normal and rescue mode, so you pick by writing the kernel and ramdisk into the appropriate partition Oct 10 07:33:11 lolstevie: are you sure about that Oct 10 07:33:14 14:17 hoshi411: on the TF101 there is no way to communicate with the default bootloader except for that voldn thing, and u-boot currently has no support for the keyboard doc OR the volup/dn buttons, so you cannot provide any input at boot with u-boot. Oct 10 07:33:33 my kit sets it up like that, so yes I am sure Oct 10 07:33:49 and I have over 2000 people using my kit at the moment Oct 10 07:33:51 twb`: Yeah, the crazy long battery life + being able to separate keyboard from the rest makes it excellent for traveling. Oct 10 07:34:10 soren: what would be *really* nice is if the keyboard could talk bt to the tablet while they were apart :-( Oct 10 07:34:11 lolstevie: can you give me a link to "your kit " Oct 10 07:34:24 twb`: Yes. Yes, it would. Oct 10 07:34:47 soren: if you just want a netbook, the Toshiba AC100 is apparently quite similar Oct 10 07:34:53 lilstevie.geek.nz/ports/linux-flash-kit.tar.gz Oct 10 07:34:53 twb: lilstevie says he's got it dual booting back and forth without hooking it up to a pc. are you believing this? Oct 10 07:35:02 hoshi411: yes Oct 10 07:35:14 it is at the cost of recovery mode Oct 10 07:35:17 twb`: How big is the transformer? Specifically the keyboard. Oct 10 07:35:20 >:| but you keep saying it is impossible Oct 10 07:35:36 soren: I dunno; it's about a quarter-inch wider than my 1005P Oct 10 07:35:39 so you only need to use the pc, at the time you want to update or need CWM Oct 10 07:35:49 soren: go down to your local hifi store and look at once Oct 10 07:35:53 s/once/one/ Oct 10 07:35:58 twb`: So slightly larger than a netbook, but still some way to go up to a laptop sized one. Oct 10 07:36:05 soren: it's 10" Oct 10 07:36:17 battery life pwns a netbook Oct 10 07:36:23 twb`: I'm almost 100% sure noone around here stocks stuff like that. Oct 10 07:36:30 soren: obviously, smaller than a 12" or 16" laptop or so Oct 10 07:36:48 soren: where are you again? Uppsala or something? Oct 10 07:37:00 twb`: Aalborg, Denmark. Oct 10 07:37:13 Hum Oct 10 07:37:22 denmark most certainly does Oct 10 07:38:12 I'm sure I can order it, I just doubt they stock it and/or have it on display. Oct 10 07:38:45 There are lots of smallish hifi stores who all carry mostly the same mainstream stuff. Oct 10 07:39:05 lilstevie: im running honeycomb 3.1 on my asus transformer but will your kit force me to downgrade or anything like that? Oct 10 07:39:26 morning Oct 10 07:39:32 hoshi411: it will force you to run what ever rom you want Oct 10 07:40:15 lilstevie: do you have any recommendations of roms based on honeycomb 3.1 or 3.2? Oct 10 07:40:17 lilstevie: I'm planning to email muromec describing my u-boot problems, what I've tried, what the symptoms are, etc, and CC you. Is that cool with you? Oct 10 07:40:32 hoshi411: prime Oct 10 07:40:42 hmm ok Oct 10 07:40:47 twb`: sure, want me to upload a binary for you to test? Oct 10 07:40:57 don't bother writing it just run it in ram Oct 10 07:41:04 to make sure it works Oct 10 07:41:06 lilstevie: also if you can work out how to get me a copy of u-boot.bin that works for you, I can verify that I'm at least uploading it to my device correctly Oct 10 07:41:25 lilstevie: right, understood (re just ram) Oct 10 07:41:47 Hell, you can just email it to me, since it's only 200kB or so Oct 10 07:42:25 kk Oct 10 07:42:31 I was just going to upload it :p Oct 10 07:43:20 Whatever's easiest for you Oct 10 07:43:58 twb`: well I am going to extract it from my device and decrypt it for you Oct 10 07:44:08 Ace Oct 10 07:44:47 I'm going home in a minute, so if it's gonna take you a while just email it me or so, since I won't have time to mess with my tf before the weekend or so anyway Oct 10 07:44:52 no rush Oct 10 07:45:09 shoot me an email and I will shoot it back a bit later Oct 10 07:51:00 lilstevie: later, dude Oct 10 08:17:15 wii infinity Oct 10 08:28:14 infinity [~adconrad@66.79.167.154] Oct 10 08:28:14 ircname : Adam Conrad Oct 10 08:28:15 channels : #ubuntu-desktop #ubuntu-installer #linaro-armhf #ubuntu-release Oct 10 08:28:15 #ubuntu-arm #ubuntu-devel Oct 10 08:28:17 server : wolfe.freenode.net [Manchester, England] Oct 10 08:28:19 : is connecting from *@66.79.167.154 66.79.167.154 Oct 10 08:28:22 idle : 0 days 0 hours 4 mins 37 secs [signon: Sat Oct 8 23:46:12 2011] Oct 10 08:28:24 End of WHOIS Oct 10 08:28:27 Hope that helps. :P Oct 10 08:30:09 lolwit Oct 10 08:30:14 lwut Oct 10 08:30:19 fail Oct 10 08:30:38 lilstevie: That was my response to hrw's wii. :P Oct 10 08:30:56 lol Oct 10 08:33:13 ;D Oct 10 09:09:30 Hi! My BeagleBoard-xM gives me "hub 1-0:1.0: unable to enumerate USB device on port 2". Does that sound familiar to anyone? Oct 10 09:11:33 See my dmesg at http://paste.ubuntu.com/705313/ Oct 10 09:56:34 ogra_, you about? Oct 10 09:56:56 davidm, indeed Oct 10 09:57:08 did you see my pm to you? Oct 10 09:57:30 on freenode ? nope Oct 10 09:57:36 Interesting Oct 10 09:57:47 ah, seeing it now Oct 10 09:58:08 i was testing suspend oin the ac100, seems you opened it while i just fell over Oct 10 09:58:17 yup, will take care Oct 10 10:17:45 ogra_: Reinstalled my ac100 and ran through a bunch of tests here, seems good. Oct 10 10:17:55 yay Oct 10 10:18:08 did you try suspend ? Oct 10 10:18:33 * ogra_ also isnt sure zram was actually a good idea, tell me if you see hangs after some time Oct 10 10:21:26 ogra_: did you see my rants about echo 30000 > /proc/sys/vm/min_free_kbytes Oct 10 10:22:23 ogra_: my experience are that there are some bug in the kernel that makes the kmalloc unable to operate when memory are low even in situations when swap are available Oct 10 10:23:07 ogra_: so I got my openbsd userland cross compiled, then built my kernel fromthat, learned about the fun required to create a proper bootable partition on the SD card, installed uboot, MLO etc to there and pointed it at my bsd kernel only to find out mboot doesn't like my bsd kernel format Oct 10 10:23:12 xranby, what you see is a "normal" message from wlan USB, we should be shipping a sysctl.d file like we do on omap Oct 10 10:23:34 it doesnt do any harm even though it looks like an oops Oct 10 10:23:59 ogra_: how are you working out suspend on the AC100 Oct 10 10:24:17 lilstevie, ask in #ac100, i'm just a consumer of their work ;) Oct 10 10:24:23 haha Oct 10 10:24:24 ok Oct 10 10:25:06 but i think they reverse engineered the nvec commands needed for the device Oct 10 10:25:34 ah Oct 10 10:25:38 nvec isn't used here Oct 10 10:25:48 right, thats why i mention it ;) Oct 10 10:25:52 :p Oct 10 10:26:13 ogra_: if you feel unsure about zram then do not iclude it for the release Oct 10 10:26:34 xranby, a bit late now :P Oct 10 10:26:38 haha ok Oct 10 10:26:41 ogra_: however have been testing out the graphics acceleration with CrOS kernel Oct 10 10:26:58 ogra_: are there anything we can do before release? Oct 10 10:26:59 i could try but i would have to be sure about it adding instability Oct 10 10:27:18 ripping it out if it isnt at fault at all would be really bad Oct 10 10:27:43 * ogra_ would like to see some test results from people seeing hangs ... so we can see if its zram or not Oct 10 10:28:02 having some show stopping bugs though Oct 10 10:28:04 i suspect that OOM doesnt work properly if zram is used Oct 10 10:28:08 ogra_: on my machine i can workaround these hangs by raising the /proc/sys/vm/min_free_kbytes Oct 10 10:28:22 ogra_: i do see userspace hangs without touching that knob Oct 10 10:28:24 xranby, and also by disabling zram at all ? Oct 10 10:28:36 ogra_: i see hand by using swap on usb as well Oct 10 10:28:47 so imho the bug are not zram related Oct 10 10:29:13 mroe a generic kernel have trouble to operate when memory are low issue Oct 10 10:29:14 well, i'd like to know if they go away if you dont use zram, just to prove that Oct 10 10:29:51 (or if they dont go away i should say) Oct 10 10:30:04 ogra_: (14.49.57) juliank: xranby_ac: The bug is not related to swapping, so stop pretending that it is. Lower the value and you'll see those crashes without swap as well Oct 10 10:30:16 hmm, k Oct 10 10:30:27 ogra_: juliank: claim that he starts to see crashes mroe frequently if the threshold gets lowered Oct 10 10:31:16 imho the system should run stable even if the threshold are insanely low Oct 10 10:31:32 else its a kernel bug Oct 10 10:31:56 well, i use 8192 as a default here Oct 10 10:32:18 (for the famous wlan bug) Oct 10 10:32:33 ok. i think the kernel defaults to around 2600 Oct 10 10:33:45 2698, the ubuntu's calculated default for ac100 Oct 10 10:35:13 the reason why using zram makes things worse are because the kernel itself uses more slub's when zram are activated Oct 10 10:35:31 imho Oct 10 10:35:33 infinity, hmm, if we actually do a respin i wonder if we should add an: echo "vm.min_free_kbytes=30000" >/root/etc/sysctl.d/99-ac100 to the installer Oct 10 10:36:54 ogra_: Erk. Why? Oct 10 10:36:59 ogra_: My ac100 seems happy enough. Oct 10 10:37:08 infinity, to fix xranby's hangs after some hours Oct 10 10:37:39 mine too ... but more than one person seems to have issues regarding that Oct 10 10:37:42 infinity: the hang happens when trying to run two memory hogs firefox and thunderbid at the same time Oct 10 10:38:03 hmm, i probably dont see it because i would never do that :) Oct 10 10:38:40 i dont think it will do harm for the working ones and it seems to fix it for the non working ones Oct 10 10:39:43 (and it cant be an SRU since we have no package for such stuff (i guess i should intorduce an ac100-hacks package in P for such a purpose)) Oct 10 10:55:32 Hi ogra. Is there anything I can further do now for helping with launchpad bug 871650? Any more logs needed? Oct 10 10:55:34 Launchpad bug 871650 in linux ""unable to enumerate USB device" with BeagleBoard-xM" [High,Confirmed] https://launchpad.net/bugs/871650 Oct 10 10:57:39 It probably doesn't make sense to try the current daily instead of last Thursday's? Oct 10 11:10:55 ogra_: do OOM killer still go around and slay your chrome? Oct 10 11:15:13 jondo, well, i would recommend to talk to jcrigby about it once he is up Oct 10 11:16:03 xranby, sure, if i use more ram thean i have it doe that Oct 10 11:16:08 *does Oct 10 11:16:24 same goes for firefox but with half the tabvs chromium can handle Oct 10 11:17:22 ogra_: ok not really a bug then Oct 10 11:17:39 but it would be nice if the user got informed that its about to happen Oct 10 11:18:09 Ok. Which timezone is he in? Oct 10 11:20:07 That certainly is https://launchpad.net/~jcrigby? Oct 10 11:22:41 jondo, right, you also should test with the latest kernel, somehow the kernel team bugbot thinks there is a newer one than you tested Oct 10 12:01:06 infinity: hi i got a report from a user who obviously have namaged to install oneiric on his ac100 and then crippled it by pressing the bug shiny install release button from unity http://paste.ubuntu.com/705353/ Oct 10 12:02:02 xranby: That must be from an old image. Oct 10 12:02:15 xranby: We now correctly clean up, so he'd never get a change to press said button. Oct 10 12:02:19 (As in, it won't exist) Oct 10 12:02:25 s/change/chance/ Oct 10 12:03:05 infinity: good.. i told him to go to http://iso.qa.ubuntu.com/ and retest usin ghte image provided there Oct 10 12:03:49 xranby: That should do. Except that he'll need the bootimg that matches, and iso.qa doesn't link both files. Oct 10 12:04:00 oh my... Oct 10 12:04:05 xranby: http://cdimage.ubuntu.com/daily-preinstalled/current/ is a bit less hassle in that regard. Oct 10 12:04:24 xranby: Code limitation, iso.qa can't list more than one file for an image. :P Oct 10 12:04:34 xranby: Not a big deal for testers, we know what we're doing there. Oct 10 12:04:50 xranby: The actual cdimage release bits have the right info. Oct 10 12:05:34 well, iso.qa seems to also point to the omap4 wikipage for installation instructions Oct 10 12:05:56 https://wiki.ubuntu.com/ARM/TEGRA/AC100 might be the best place to point to Oct 10 12:07:17 ogra_: Not sure that it matters deeply. I doubt that many people will be testing ac100 for us. Oct 10 12:07:29 And we really (REALLY) shouldn't be pointing users to iso.qa as a source for images. Oct 10 12:07:37 cdimage.ubuntu.com is the right place. Oct 10 12:07:55 well, i point them to iso.qa beliveing that everything is in order there Oct 10 12:08:00 which it apparently isnt :) Oct 10 12:08:17 the wikipage has the proper instaructions and links to both files you need Oct 10 12:13:10 ogra_: infinity: i was a bit surprised since this user claims he have followed the instuctions from the wiki Oct 10 12:13:41 xranby, i only updated the links from beta to the current daily when i was sure the images build properly Oct 10 12:13:44 i sent him a mail to double check that he are using both the latest .tar.gz and bootimg Oct 10 12:13:58 so beta was linked from there for a few days longer Oct 10 12:15:26 janimo, how do the mx5 images look ? Oct 10 12:15:43 * ogra_ needs to signoff on them fo rthe release team Oct 10 12:18:34 heh... http://marcin.juszkiewicz.com.pl/2006/03/18/openzaurus-354-released/ was nice time - good that I do not have to work on releases anymore Oct 10 12:22:15 ogra_, will ping you back once I test todays. The last ones looked better Oct 10 12:22:30 k, thanks Oct 10 12:59:20 janimo: when i test the daily mx5 image, i find banshee in the sound menu Oct 10 12:59:39 yes, thats a bug :( Oct 10 12:59:45 likely on all images Oct 10 13:00:18 seems there are dependencies that keep banshee in, juts dropping it from the seeds didnt help Oct 10 13:00:27 you should have rhythmbox too though Oct 10 13:02:38 ogra_: there is still no RC image? is that correct? Oct 10 13:02:57 ndec, there wont be an official RC image ... the dailies are RCs Oct 10 13:03:04 ok! Oct 10 13:03:30 ogra_: during oem-setup i only see banshee no rythmbox will check if rythmbox are present when i login into deskto Oct 10 13:03:51 it should be ... but you will likely need to search for it in the dash Oct 10 13:04:11 as long as banshee is installed it claims all the mediaplayer links and launchers Oct 10 13:04:30 talking about banshee... i just dist-upgraded my system, banshee is still there and rythmbox was not installed. Oct 10 13:04:47 do you have ubuntu-desktop installed ? Oct 10 13:05:01 that it doesnt get removed on pre-release images is normal Oct 10 13:05:07 no... Oct 10 13:05:13 now i am wondering why... Oct 10 13:05:14 we exdpect users of alphas and betas to be able to deal with that Oct 10 13:05:25 i have dist-upgraded since alpha2 or something like that. Oct 10 13:05:28 but RB should get installed in any case Oct 10 13:05:32 ;-) Oct 10 13:12:55 xranby, I still find the daily mx5 install very slow Oct 10 13:13:00 you said you saw an imporvement Oct 10 13:13:42 almost 10 minutes to the second bootup, and no output about resizing partitions in the first one Oct 10 13:17:13 hi everybody, where can i find information on the work done re this blueprint: Oct 10 13:17:16 https://blueprints.launchpad.net/ubuntu/+spec/server-o-arm-sysadmin-tools Oct 10 13:17:37 pmatulis, NCommander leads the server stuff in ubuntu-arm Oct 10 13:18:20 pmatulis, though i would start talking to the approver or drafter first Oct 10 13:18:36 looks like cmagina did all the work on it Oct 10 13:18:41 janimo: yes the daily install are still slow Oct 10 13:19:03 janimo: the daily linaro oneiric image are quick in comparision Oct 10 13:19:12 for mx5 Oct 10 13:19:16 xranby, which one was fast then? IIRC you said that a few days ago Oct 10 13:19:21 there finally are linaro oneiric images ? Oct 10 13:19:28 ah, the linaro one. Sure that always was faster Oct 10 13:19:33 heh, yeah Oct 10 13:19:33 for mx5 Oct 10 13:19:48 they just hardcode the autologin and dont have to run jasper Oct 10 13:20:02 that makes the first step lots faster Oct 10 13:20:06 ogra_: alrighty Oct 10 13:20:24 ogra_, I think in this case it is somehing else. kernel/FS/SDcard setting maybe? Although we use the same kernel imagaes Oct 10 13:20:41 well, you said you dont see jasper output Oct 10 13:20:47 as our mx5 is much slower than our omap one so it is not just jasper Oct 10 13:20:49 so there likely also is a jasper bug Oct 10 13:21:02 ogra_, indeed, I did not see any output for over 5 min, board seemed dea Oct 10 13:21:04 d Oct 10 13:21:05 ogra_: janimo: the last linaro oneiric image i testsed was sudo linaro-media-create --rootfs ext4 --mmc /dev/sdc --binary linaro-o-ubuntu-desktop-tar-20111006-1.tar.gz --hwpack hwpack_linaro-lt-mx5_20110929-0_armel_supported.tar.gz --dev mx53loco Oct 10 13:21:21 but it had some issues like metacity crashed Oct 10 13:21:32 apart from that gui was snappy Oct 10 13:28:20 janimo: http://snapshots.linaro.org/oneiric/linaro-o-ubuntu-desktop/ Oct 10 13:29:35 xranby, no images yet for today's directory there Oct 10 13:29:44 correct Oct 10 13:29:47 the last image are from 08 Oct 10 13:29:52 and hw pack http://snapshots.linaro.org/oneiric/imx51-oneiric/20111008/0/images/hwpack/ Oct 10 13:38:17 janimo: my Ubuntu oneiric daily mx5 image did after half an eon complete the oemsetup and i am now viewing the desktop Oct 10 13:38:35 xranby, that's the speed I see too Oct 10 13:38:39 the last part of the oem-setup i was only observing the background image and title bar for about 30min Oct 10 13:39:38 but I get ubiquity restarted at the end Oct 10 13:39:48 I got these with omap images too this cycle Oct 10 13:40:06 as long as it doesnt happen with todays anymore thats fine :) Oct 10 13:44:00 ogra_, it is with today's unfortunatley Oct 10 13:44:23 I am not sure I ever saw an omap image get past instal lw/o this ubiquity restart either Oct 10 13:44:47 so seeing it on the mx5 is not surprising me Oct 10 13:45:00 need to check if there's a crash log, most times there was not even one Oct 10 13:45:14 so you are saying you now see that bug that nobody else but GrueMaster could reproduce for weeks = Oct 10 13:45:16 ? Oct 10 13:45:33 darn, well, at least you can reproduce it too now :/ Oct 10 13:45:57 even though thats to late for release :( Oct 10 13:46:03 after a week of hacking I finally found the source for MLO Oct 10 13:46:29 brandini, you could have asked ;) Oct 10 13:46:38 I know, I was too focused on other things Oct 10 13:46:54 I seriously think I was made to be an embedded hacker, I lvoe this stuff! Oct 10 13:47:02 like spending a week on finding it yourself you mean ? :) Oct 10 13:47:55 I was hacking the pandaboard port for openbsd, cross compiling, etc so I just copied the exising uboot and MLO from linux Oct 10 13:48:08 ah Oct 10 13:48:35 ogra_: it took a special form of art to make a bootable windows partition :) Oct 10 13:48:44 heh Oct 10 13:49:02 needs an even number of sectors to overcome some "bug" in mkfs.vfat and even after knowing that it took me a dozen tries to get it right Oct 10 13:49:48 ogra_, I mean I always( almost always) see this, but there are not crash logs o Oct 10 13:49:59 some of the earlier ones had _Crash files Oct 10 13:50:08 those too were seen by gruemaster but fiex d in the meantime Oct 10 13:55:55 janimo, well, why didnt you speak up when we discussed that, it seemed GrueMaster was the only one seeing it Oct 10 13:56:34 * ogra_ didnt have a single non-explainable restart of ubiquity during this cycle Oct 10 13:59:00 anyway, we should find out why its failing, make sure to file a bug Oct 10 13:59:04 wonder if the eagleboard aims to be a "production" style board Oct 10 13:59:20 as production style as the panda i guess Oct 10 13:59:28 i.e. not for commercial use Oct 10 13:59:35 gotcha Oct 10 14:00:04 * brandini creates a cluster of arm servers on commodity hardware Oct 10 14:09:57 ogra_, hmm beause I thought everyone sees it and it is part of the current status honestly Oct 10 14:10:14 nah, it isnt, sadly Oct 10 14:10:16 the only thing that is suspicions - is /var/og/faillog which is 20K of zeros Oct 10 14:10:40 but no oem-config.log nor crash nor installer log have suspicious content Oct 10 14:11:04 syslog is the intresting one, the rest is usually quite empty anyway Oct 10 14:11:09 also jockey has backtraces but that too must be common for all arms as it looks and fails to find broadcom-wl fglx etc Oct 10 14:11:34 i havent seen any jockey issues, though apport should be disabled on todays images Oct 10 14:11:47 so you might not notice crashes in the bg Oct 10 14:15:13 well, they are python backtraces but it carries on so I think they are harmless Oct 10 14:16:08 well, add them to the bug Oct 10 14:16:14 also faillog is for login attempts, unlikely to be relevant here Oct 10 14:16:24 usually python backtraces are rather serious in ubiquity Oct 10 14:21:31 ogra_, this is in jockey. ubiquity logs are clean Oct 10 14:21:38 ah Oct 10 14:21:55 ogra_, after the installer finished should it prompt for reboot or reboots automatically? Oct 10 14:22:10 it should drop you to lightdm Oct 10 14:22:12 I never get to that step I only see the newly started ubiquity Oct 10 14:22:24 it doesnt reboot at all Oct 10 14:22:31 hmm, never seen being dropped to lightdm. So it logs out then automatically? Oct 10 14:23:13 after it removed itself it restarts the display manager ... since ubiquity-dm is gone you land in lightdm Oct 10 14:27:41 i´m using beagleboard xm with ubuntu 11.04 and kernel 3.0. i have problems with flooding messages in the syslog and kern.log file. the messages look like this: rpm_suspend flags 0x0, rpm_resume flags 0x4, rpm_resume flags 0x5, rpm_resume returns 1. can someone help me with this messages? Oct 10 14:30:37 ogra_, ubiquity-dm is also not removed on this image Oct 10 14:30:58 janimo, yes, else you wouldnt see the issue ;) Oct 10 14:31:05 :) Oct 10 14:52:11 janimo, Oct 10 14:52:12 Oct 10 13:07:30 localhost ubiquity[1108]: log-output -t ubiquity laptop-detect Oct 10 14:52:12 Oct 10 13:07:38 localhost dbus[404]: [system] Activating service name='org.debian.apt' (using servicehelper) Oct 10 14:52:12 Oct 10 13:08:03 localhost dbus[404]: [system] Failed to activate service 'org.debian.apt': timed out Oct 10 14:52:17 that looks suspicious Oct 10 14:52:50 you have a unch of other AptDaemon messages in there Oct 10 14:53:41 and there is: Oct 10 14:53:43 Oct 10 16:20:46 localhost ubiquity: debconf: DbDriver "config": /var/cache/debconf/config.dat is locked by another process: Resource temporarily unavailable Oct 10 14:54:21 that may be from when it already restarted and I tried removing oem-config Oct 10 14:54:28 I am doing a new install now Oct 10 14:54:33 though that might still be locked from the crashed ubiquity run before Oct 10 14:54:39 right Oct 10 14:54:41 it seemed so Oct 10 14:54:54 aptdaemon is the bit thats does the uninstall though Oct 10 14:56:39 there is also one: Oct 10 14:56:40 Oct 10 16:16:49 localhost ubiquity: dpkg-trigger: error: must be called from a maintainer script (or with a --by-package option) Oct 10 14:56:49 though not sure that means anything Oct 10 15:01:16 tbh the installer logs are usually very chatty so it is not easy to find the logs for the really harmful event Oct 10 15:06:20 xranby, did you experience boot problems with USB devices inserted? I need to not have anything plugged in or it won't boot - leds turn off Oct 10 15:06:29 may be a hw issue of my board though Oct 10 15:06:43 Morning all. Nice to know I am not the only one seeing oem-config respawn. Oct 10 15:07:14 janimo: no. i have usbhub + keyboard + mouse inserted on the mx5 when i tested the ubuntu oneiric mx5 image Oct 10 15:07:31 xranby, ok, my hw may have issues then Oct 10 15:07:45 which may be the root of the other errors too Oct 10 15:07:56 I wonder if my setup has issues, since I see such things on omap Oct 10 15:08:46 http://www.theinquirer.net/inquirer/news/2115929/ubuntu-1110-support-arm-chips-fight-red-hat Oct 10 15:09:06 janimo: As long as you don't use your AC100 PS on your panda/beagle/mx5, you shouldn't have too many problems. :P Oct 10 15:09:58 GrueMaster, you never forget do you? Oct 10 15:10:17 luckily the mx5 has (hmm does it??) unique jack Oct 10 15:10:35 Well, you know the old addage, "Elephants never forget", right? Oct 10 15:10:49 The mx5 uses the same jack btw. Oct 10 15:11:26 GrueMaster, but this one comes from a different wall socket not from under the table, so I cannot easily mix them up Oct 10 15:11:48 still a weird situation. Maybe both my hub and my external USB drive are just too much for that supply Oct 10 15:12:00 It isn't the wall part that matters. :P Oct 10 15:12:24 Is your hub and external drive self-powered? Oct 10 15:12:59 Or do they draw from the mx5? Oct 10 15:16:23 both draw from mx5 Oct 10 15:16:32 I may need to check another hub which has its own supply Oct 10 15:16:46 Yea, that would be bad. Oct 10 15:17:33 * GrueMaster wanders over to the coffee pot to fullfill his daily quota. Oct 10 15:29:55 jondo, bug #871650 reminds me of onw GrueMaster reported a while ago and I don't thing we ever resolved. Looking for old bug... Oct 10 15:29:56 Launchpad bug 871650 in linux ""unable to enumerate USB device" with BeagleBoard-xM" [High,Confirmed] https://launchpad.net/bugs/871650 Oct 10 15:35:27 jcrigby: The bug you are looking for is 838200 Oct 10 15:35:40 bug 838200 Oct 10 15:35:41 Launchpad bug 838200 in u-boot-linaro "No network support on Beagle XM" [High,New] https://launchpad.net/bugs/838200 Oct 10 15:35:41 sigh. bug 838200 Oct 10 15:35:50 hehe, i was faster ! Oct 10 15:36:15 GrueMaster, ogra_ you are both faster than me! I was still looking Oct 10 15:36:51 i just reformatted GrueMaster's thought :) i'm not actually that fast Oct 10 15:37:16 or better: LP isnt ... :P Oct 10 15:37:27 But I don't think they are the same, as jondo (and others over the weekend) have been reporting no USB at all. My system has everything enabled, but to get networking to work, I have to unplug the network cable for a fe seconds then plug it in. Oct 10 15:37:49 GrueMaster, yes you are right the old bug has usb working but not network Oct 10 15:38:02 They could be related to the same code. Oct 10 15:38:12 As they are different rev boards. Oct 10 15:38:36 And rev C has some different power scenarios for USB. Oct 10 15:38:37 well, it smells like u-boot or MLO to me ... but i might be totally wrong and its the kernel Oct 10 15:38:49 (I have rev B). Oct 10 15:39:55 I have found that if I use the Natty MLO/U-boot, I get consisten networking. With Oneiric, spotty at best. Oct 10 15:40:04 ogra_, could be however I believe latest linaro images work fine but maybe not on xm rev c Oct 10 15:41:17 jcrigby: linaro boot systems use a master boot partition. Does it ever get updated with the latest u-boot? Oct 10 15:41:49 GrueMaster, you mean the test systems? Oct 10 15:42:00 yes. Oct 10 15:42:50 not sure how often they get updated but you are right it is not with every build because if they did they could get bricked with a bad u-boot build and then a human would have to go replace the sd cards to fix them Oct 10 15:43:25 this bug is so annoying Oct 10 15:43:26 Hence the flaw in the design I mentioned a while ago. Oct 10 15:43:32 do we have any newer xM than C? Oct 10 15:43:46 in theory both kernel and u-boot are handling it correctly Oct 10 15:44:01 I don't think there is one. Not according to beagleboard.org at least. Oct 10 15:44:03 GrueMaster: yours is even more interesting, just eth0 not working Oct 10 15:44:08 what doesn't make any sense to me Oct 10 15:44:35 It is only the link detection circuit that isn't working. Oct 10 15:44:43 The port enumerates fine. Oct 10 15:44:59 possibly nothing to do with this but I found that when switching to the u-boot based MLO on panda the external USB/ethernet phy stopped working until I did the right thing in u-boot Oct 10 15:45:01 that's the interesting thing, it's all the same chip Oct 10 15:45:15 the right thing being setup its clock before powering it on Oct 10 15:45:42 jcrigby: yeah, but I believe this was basically because u-boot wasn't in sync with x-loader Oct 10 15:45:52 yes, exactly Oct 10 15:45:54 rsalveti: It may be the same chip, but I think the code paths are different. Oct 10 15:46:34 Otherwise I would be able to get networking working in u-boot, right? Oct 10 15:46:35 rsalveti, what I am thinking is that this external phy seems to be touchy about its init sequence Oct 10 15:46:50 jcrigby: could be Oct 10 15:47:13 rsalveti, was the the xm that switched the polarity of the reset and the switched back? Oct 10 15:47:14 GrueMaster: yup, but don't know if it's currently working the same way as panda Oct 10 15:47:28 jcrigby: iirc, yes Oct 10 15:47:35 that was the rev C thing Oct 10 15:47:43 rsalveti: It isn't. At least not on my end. Oct 10 15:48:02 jcrigby: I remember we had a bug for rev C support at u-boot a while ago Oct 10 15:48:06 let me check Oct 10 15:49:14 ok, now for some fun Oct 10 15:49:15 mount an nfs share with all my DVD's on it and play them! Oct 10 15:49:44 should I be using gdm or something else? Oct 10 15:49:57 Go brandini, go! Oct 10 15:50:01 Yay! Oct 10 15:50:33 jcrigby: bug 770679 Oct 10 15:50:34 Launchpad bug 770679 in linux "Missing proper support for Beagle XM rev B and C" [Medium,Fix released] https://launchpad.net/bugs/770679 Oct 10 15:50:53 rsalveti: iirc, when we launched Natty, you made a special u-boot & kernel that only updated the board ID table for Rev C. Oct 10 15:51:18 GrueMaster: yup, this was the bug that fixed the support at the kernel side Oct 10 15:51:34 heh, wonder if that's why uboot didn't boot my bsd kernel Oct 10 15:52:59 jcrigby: http://patchwork.ozlabs.org/patch/92334/ need to check the default path when an unkown rev is identified Oct 10 16:00:12 wait, does ubuntu arm support nfs? Oct 10 16:00:39 brandini: Yes. I have it running here on my system I used for builds. Oct 10 16:00:47 ahh ok Oct 10 16:00:52 And I have tested nfs-root as well. Oct 10 16:01:15 NFS-Root is much faster than SD. Oct 10 16:01:23 knowmercy@localhost:~$ sudo mount -t nfs 10.15.32.103:/storepool /media Oct 10 16:01:24 mount: wrong fs type, bad option, bad superblock on 10.15.32.103:/storepool, Oct 10 16:01:29 :( Oct 10 16:01:47 jcrigby: in theory the kernel for both oneiric and linaro are right Oct 10 16:02:09 beagle_config.usb_pwr_level = GPIOF_OUT_INIT_HIGH; for xM only Oct 10 16:02:15 brandini: For grins, try rebooting. Oct 10 16:02:21 GrueMaster: the panda? Oct 10 16:02:23 ops, for A/B only, not C Oct 10 16:02:39 brandini: yes. Oct 10 16:02:42 ok Oct 10 16:02:54 GrueMaster: what display manager works best? Oct 10 16:03:09 I have seen an issue where networking gets confued. Oct 10 16:03:19 Display manager? We ship with lightdm. Oct 10 16:03:28 *confused Oct 10 16:04:47 GrueMaster: do I need mount.nfs? Oct 10 16:05:05 No, mount should just work. Oct 10 16:05:09 seems I do Oct 10 16:05:40 nfs common? Oct 10 16:06:16 aha! Oct 10 16:06:22 that worked Oct 10 16:06:39 rsalveti, ok. One other thing perhaps again no relevant. I noticed when merging ubuntu sauce that the upstreamed beagle changes were a little different than the ubuntu sauce from you and robert Oct 10 16:06:41 now to get a display manager runnin on server Oct 10 16:07:00 GrueMaster: should I be using mplayer to play these movies? Oct 10 16:07:04 or are there better things? Oct 10 16:07:18 jcrigby: hm, u-boot also seems fine Oct 10 16:07:24 default is basically rev C Oct 10 16:07:32 ogra_, ah one reason I probably newver saw the lightm greeter is that when installing I always set autologin on Oct 10 16:07:52 rsalveti, yes just noticed that also Oct 10 16:08:04 brandini: I really don't know. iirc, you want to use something that uses the g-streamer hardware accelerated drivers. Oct 10 16:08:19 janimo: Um, yea. That would do it. Oct 10 16:08:41 now I am trying an install without autologin Oct 10 16:08:46 let's see if anything is different Oct 10 16:10:06 * janimo wonders what the heck is the netsurf web browser and why is it prominently featured in the StarApps section of the installer slideshow Oct 10 16:10:30 brandini: Figured it out. sudo apt-get -y install nfs-common. Then you can mount nfs drives. Oct 10 16:18:09 jcrigby: but I didn't see any specific changes at u-boot for rev c Oct 10 16:18:26 so it seems the kernel or MLO is responsible for setting it up correctly atm Oct 10 16:19:32 Do we have any rev C beagles in Linaro? Oct 10 16:21:26 jcrigby: not sure, need to check the hw spreadsheet Oct 10 16:22:05 I will Oct 10 16:23:54 rsalveti, good news mattwaddel has a rev C Oct 10 16:24:05 jcrigby: oh, even better Oct 10 16:24:15 I am going to swap my B for his C and then I can try some stuff Oct 10 16:30:52 GrueMaster, although once the board is booted up, said peripherals work without external power supply Oct 10 16:31:01 so I still think there's something the kernel could do better Oct 10 16:31:39 possibly., but you could also be very close to hitting the limits of USB. Oct 10 16:31:50 GrueMaster: yeah, I did the nfs common and that did the trick! Oct 10 16:32:09 brandini: Cool. Oct 10 16:32:18 so g-streamer eh? Oct 10 16:32:24 and gdm Oct 10 16:32:53 brandini: What is your end goal? Oct 10 16:47:49 to have X running so I can play movies outputting them on my TV :) Oct 10 16:48:25 You might try mythtv or one of those types of front-ends. Oct 10 16:48:31 :) Oct 10 16:49:12 but my point right now is that I don't have X running since I'm running 11.10 server Oct 10 16:57:36 Ah. Oct 10 16:58:46 I haven't tested anything to that level this cycle. I would suggest experimenting. Of course, it helps to have the accelereated bits, which should be in the ppa already, if not very soon. Oct 10 16:59:07 janimo: Grr. I'm trying to artificially reproduce your oem-config-remove-gtk-not-starting bug, and my Panda is failing to misbehave. Oct 10 17:07:39 What mechanism does it use to launch oem-config-remove-gtk? If it is sending out a dbus call then exiting, X is probably dying before oem-config-gtak-remove has a chance to start. Oct 10 17:08:18 Not sure why dbus is being used to launch apps anyways. Oct 10 17:08:31 Just feels wrong. Oct 10 17:08:46 dbus is used internally by aptdaemon Oct 10 17:08:56 not to call aptdaemon from ubiquity Oct 10 17:09:45 and launching bits through dbus enables privilege control through polkit Oct 10 17:10:00 so it has security advantages Oct 10 17:26:12 infinity, this is on mx5 though Oct 10 17:36:05 janimo: Yeah, I know. Oct 10 17:36:22 janimo: Maybe if I boot nosmp to make my machine even slower. :P Oct 10 17:41:53 why would oem-config-gtk-remove rely on X in the first place? Oct 10 17:42:49 if oem-config needs to uninstall a component of itself it had better do it while ubiquity isn't running, right? don't they share things? Oct 10 19:14:31 hi, any news on 1080p packages for pandaboard running oneric. Did this ever get done for Naty? Oct 10 19:30:29 ? Oct 10 19:32:20 crickets Oct 10 22:36:13 stephen__: nops, was never done for natty Oct 10 22:36:24 we're all waiting ti to publish the newer packages Oct 10 23:35:19 GrueMaster Oct 10 23:35:24 u on? Oct 10 23:35:37 sup? Oct 10 23:36:35 the other day you told me to put 'discard' into /etc/fstab. before the first boot? Oct 10 23:37:41 Yea, forget that. Doesn't work well with SD. Oct 10 23:38:17 do you know another way to get past the start up config? Oct 10 23:39:14 After the first run through, hit , log in as the user you created, and type "sudo oem-config-remove", then reboot. Oct 10 23:39:55 ok, ill give that a try Oct 10 23:44:50 IME -o discard doesn't work at all on SD, and the kernel simply reports "discard not supported, disabling" the first time it tries to GC blocks Oct 10 23:44:58 So AFAICT it's harmless to turn on Oct 10 23:48:27 twb`: I actually saw a performance hit with it enabled, but it also allowed me to get through oem-config. Not sure what is going on yet. Oct 10 23:48:35 Hu Oct 10 23:48:42 What kernel Oct 10 23:51:10 3.0.0-1205.10 Oct 10 23:51:48 What *I* would be doing is adding force-unsafe-io to dpkg.cfg, fwiw Oct 10 23:53:24 while not quite an ubuntu question, I was wondering if anyone had experienced issues using FlexNVM in eeprom mode on the Cortex-M4 (freescale k60 family), I keep seeing resets when trying to initialize the partitions **** ENDING LOGGING AT Tue Oct 11 02:59:57 2011