**** BEGIN LOGGING AT Wed Apr 07 02:59:57 2010 Apr 07 08:33:33 amitk: so the reason why serial isnt enabled by default is that it breaks plymouth Apr 07 08:33:37 so no serial by default Apr 07 08:34:03 e.g. if you have console=* in the line it goes poof and starts without graphical launcher Apr 07 08:34:12 wanted to look at that code if i find a minute Apr 07 08:34:40 asac: I heard, pretty bad. Apr 07 08:35:57 yeah Apr 07 08:36:52 ok i need to finish migration of my mailserver/irc gateway ... let me shut down irc for that and start it up elsewhere Apr 07 09:55:22 asac: heya; https://bugs.launchpad.net/ubuntu/+source/linux-ti-omap/+bug/541030 is milestoned for beta 2 (tomorrow); is an upload still planned for it, or should the milestone be moved? Apr 07 09:55:25 Launchpad bug 541030 in linux-ti-omap (Ubuntu Lucid) (and 1 other project) "omap kernel musb/ehci ports not enabled on beagleboard (affects: 1) (heat: 12)" [High,Triaged] Apr 07 09:57:02 lool: pushed it to final milestone (i doubt there is time for a kernel upload anyway) Apr 07 09:57:06 thx Apr 07 09:58:27 lool: push to final, we can't do any more kernel uploads currently Apr 07 10:00:29 Can't we do one post-beta2 pre-rc? Apr 07 10:02:24 persia: we can try Apr 07 11:06:10 morning Apr 07 12:14:32 JamieBennett: so thats ok upstream code wise ;) Apr 07 12:15:03 yep Apr 07 12:15:05 JamieBennett: hmm. -f doesnt really work Apr 07 12:15:06 if empty Apr 07 12:15:09 just --filename= Apr 07 12:15:40 but thats ok as we use that in .desktop Apr 07 12:15:42 asac: yeah, due to getopt Apr 07 12:16:00 JamieBennett: havent checked looking url Apr 07 12:16:09 is -t irrelevant for that? Apr 07 12:16:38 -t is used to launch the correct service Apr 07 12:16:58 ok so no auto detection by them Apr 07 12:17:04 yeah. so do a release and the packaging Apr 07 12:17:05 ;) Apr 07 12:17:15 will do Apr 07 12:48:25 GrueMaster: hey Apr 07 12:48:42 GrueMaster: so is the focus things for maximus/-efl not an issue for the 3d launcher? Apr 07 12:48:47 quick newbie question, im a bit confused as to how to go about installing ubuntu for ARM on a netbook that comes pre-installed with windows CE... im assuming theres not a similar iso approach like regular ubuntu where i may be able to for instance turn an sd card into a bootable disk and install? Apr 07 12:49:08 styol: what netbook is that? Apr 07 12:51:19 styol: 50% of the effort involved is likely to be figuring out how to boot the device. Once that's sorted, it's easier to make it just work. Apr 07 12:53:00 sorry my IRC app froze for literally the past 3 minutes Apr 07 12:53:27 its some crappy china netbook that with an awesome $90 price tag... Apr 07 12:53:30 Did you get backscroll at least? Apr 07 12:53:46 If you can tell us the model name, we may be able to give more advice. Apr 07 12:53:47 yeah indeed, how to boot the device as in like key combinations? Apr 07 12:54:02 hmm its likely a no name brand, one sec ill grab the link with details Apr 07 12:54:16 http://www.china-wholesale-supplier.com/7-inch-netbookak7802q216wince-50-operating-systemnandflash-2gbfrequency-266mhz-processormini107_p12982.html Apr 07 12:54:33 Or what the bootloader accepts. Some devices can boot USB, some can't. Some can boot from SD, some can't. Some need external special tools to update the NOR to change how they boot. Apr 07 12:54:52 thats a good point that would make sense to be a big difference in approach :) Apr 07 12:56:12 That looks a lot like a Skytone Alpha-400 to me. Apr 07 12:56:37 I wouldn't try to put Ubuntu on it: the RAM is very tight. Apr 07 12:56:46 ive been trying to look up details for the processor AK7802Q216 but only am able to locate other netbooks with the same processor Apr 07 12:56:53 ah ok, any other suggestions? Apr 07 12:57:00 If you do, I suspect you can only run Jaunty (9.04), and not newer releases, because of the processor. Apr 07 12:57:20 the appeal with ubuntu was actually potentially stuff that may be stripped, in particular we're looking to do a kiosk setup with either simply a web page loading or adobe air which may be pushing it Apr 07 12:57:43 You'll want to research more about the device, but I suspect Debian could run on the processor. You'd need to be careful to trim your install to be happy with 128MB RAM though. Apr 07 12:57:44 * an adobe air app (that would essentially only load a web page) Apr 07 12:58:16 yeah will have to make sure to keep that in mind. ultimately we just want a device that display a single web page and does nothing else Apr 07 12:58:31 You don't need a general-purpose OS for that. Apr 07 12:58:44 GO find a kernel that can boot the device, and init into an embedded browser. Apr 07 12:59:27 (Oh, also, I don't think any of the Ubuntu kernels would work there) Apr 07 12:59:46 sorry im a bit new to the more technical side of linux, definitely of the variety that uses iso's with easy installs on common setups hehe.. so this is all new stuff to me Apr 07 13:00:00 ah ok. much appreciated with the suggestions, im soaking everything in that i can Apr 07 13:00:12 do you have any embedded browser recommendations? Apr 07 13:00:36 konqueror? Apr 07 13:00:42 I don't. We don't generally do much embedded stuff (although we run a general purpose OS on hardware that some vendors also sell into embedded markets). Apr 07 13:01:13 It would be some special konqueror: the regular konqueror assumes a general purpose OS (and all of KDE). Apr 07 13:01:20 gotcha. you've been extremely helpful and i really appreciate the guidance, likely have saved me a lot of time already :) Apr 07 13:01:32 styol: midori Apr 07 13:01:47 hrw: Is there a midori-static that can run as init? Apr 07 13:01:49 styol: uses webkit and works on 128MB ram Apr 07 13:01:56 sweeeet! thanks hrw! Apr 07 13:02:08 persia: no idea - in OE we used midori with whole x11 session Apr 07 13:02:10 * XorA has run firefox on 64M device, wasnt that hard Apr 07 13:02:41 heh. The *one* time I actually recommend an embedded-style solution, and everyone points out that general-purpose solutions scale all the way down :) Apr 07 13:02:44 in 2005 I managed to run Minimo (gecko based mobile browser) on 400MHz Xscale with 64MB ram Apr 07 13:03:04 * XorA used firefox on his SL-5500 :-D Apr 07 13:03:11 hrw will remember that device well Apr 07 13:03:23 styol: but forget Adobe Flash support on ARM/Linux Apr 07 13:03:30 oh man all this info, gonna have to save this chat log haha Apr 07 13:03:43 thats fine, i mean we may have wanted to use it for audio but im guessing we can figure something else out Apr 07 13:03:43 styol: Already done: it's at irclogs.ubuntu.com Apr 07 13:03:46 XorA: ba! it get me into ARM Linux which get me to OpenEmbedded which get me to Canonical Apr 07 13:03:55 this place rocks, im leaving this channel on auto-join Apr 07 13:04:04 im extremely interested in this new line of knowledge :) Apr 07 13:04:19 hrw: next release Openzaurus Ostriche? Apr 07 13:04:49 XorA: you mean 11.10? Apr 07 13:04:50 styol: Just be aware that we're targeting newer processors than that in the device you mention, to the degree that there are very few devices currently available in retail that can even run current Ubuntu. Apr 07 13:05:24 To get Zaurus back needs someone to revive the name at Sharp, which needs a better form-factor than Netwalker. Apr 07 13:05:25 styol: grab FriendlyARM devboard with 7" wvga screen Apr 07 13:05:26 persia that is a good point indeed, we would be essentially wanting to provide a setup that never changes aside from the web page loaded so scaling or upgrading wouldnt be in the scope for us Apr 07 13:05:54 persia: I first met hrw working on Openzaurus :-) Apr 07 13:06:01 styol: costs ~170€, is nicely supported in OpenEmbedded Apr 07 13:06:29 styol: but does not run current Ubuntu Apr 07 13:06:32 And for all I tout Ubuntu, OpenEmbedded is probably a better solution at the current time for a low-cost only-shows-this-one-super-magic-webpage device. Apr 07 13:06:47 hrw this device im looking at is $90 usd and $80 usd at wholesale with a screen, keyboard, lan, wifi, speakers, essentially everything we need.. but its 128mb ram and 2gb hd Apr 07 13:07:00 * a netbook basically, really basic one Apr 07 13:07:04 looks like a matel toy Apr 07 13:07:14 On that device, really, run Debian with low-memory apps, and you'll do fine. Apr 07 13:07:38 And doing it that way lets folk use it for anything else once they break into the browser. Apr 07 13:07:45 we really one people that use it to not even be able to goto other web sites, start applications, nothing Apr 07 13:07:50 well, guess logging into wifi is a different story Apr 07 13:08:10 use that case, get prpplague to design an omap4 board to fit inside it :-) Apr 07 13:08:11 wifi is easy if you have working driver Apr 07 13:08:28 styol: Someone *will* break into it if it's a good device anyway. Apr 07 13:08:31 XorA: prpplague will fit BeagleBoard inside Apr 07 13:08:51 persia what do you mean? well, i can put it into perspective if you guys want regarding basically what we're trying to do... Apr 07 13:09:22 essentially will be a device that restaurants can purchase for this free online ordering thing we're doing and it will allow them to basically receive new order notifications, update users of changes to order (on delivery etc), and that sort of stuff Apr 07 13:09:40 styol: That I know of no browser that isn't a security nightmare, so if you permit browser access, someone will find a way to root the device, and use it for something else. This may not impact your plans in any way (it doesn't impact the plans of most folk) Apr 07 13:09:56 its fine honestly if they go through the trouble, they deserve it Apr 07 13:10:02 the standard hostess will likely not do that Apr 07 13:10:06 if they want facebook that bad, have it lol Apr 07 13:10:28 i just dont want it to be use for the general public Apr 07 13:10:33 *easy Apr 07 13:10:49 er... closed audience, not general public in that its like library computer Apr 07 13:11:41 heh the pile driver has started up outside my house Apr 07 13:14:28 I need to find 15cm usb extender cables Apr 07 13:16:57 im in america and converted to inches and search 6 inch usb extender and theres gobbles Apr 07 13:17:19 would you like a link or are you looking for something local (or in your house even) Apr 07 13:17:59 I need to find local - getting them from US would take forever and cost 800% of their value at least Apr 07 13:19:04 wheres local if i may ask? Apr 07 13:21:27 styol: Szczecin, Poland Apr 07 13:22:09 ah i see. need bulk or just one/couple? Apr 07 13:23:38 few Apr 07 13:23:53 but so far I need one - for beagleboard hub Apr 07 13:24:43 ah gotcha Apr 07 13:25:09 btw - how many of already supported by Ubuntu/ARM devices have video-out? (vga/dvi/hdmi/svideo ones) Apr 07 13:45:36 hrw: There's no way to answer that precisely, because there exists no authoritative list of devices that work, but at least all the dev boards I've seen running it have something, and I know that either DisplayLink or SiS/USB works for the rest, if there is USB available. Apr 07 13:46:19 (although there's currently no good integration in terms of UI or autodetection to set xorg.conf for USB solutions) Apr 07 13:46:24 ok Apr 07 14:03:04 lool, why did you move the USB omap bug to high ? it prevents all ways of installation or input Apr 07 14:03:22 * ogra thinks its absolutely criticalk Apr 07 14:03:25 -k Apr 07 14:04:03 Does it really matter? It gets fixed-by-upload about 4 hours after beta2 releases, doesn't it? Apr 07 14:07:13 persia, i dont know, all i currently know is that its not yet fully fixed Apr 07 14:07:24 so it depends if amitk finds a fix Apr 07 14:07:42 Agreed there. I just didn't think "critical" vs. "high" mattered, since we know it needs fixing soonest. Apr 07 14:08:03 persia, right, thats why i wondered why lool changed it :) Apr 07 14:36:15 ogra, persia: critical is when it would affect random other packages, breaking the world Apr 07 14:36:36 It's not the case, it's a high priority bug in a well understood place Apr 07 14:37:07 well, i could argue that it breaks the whole omap world :) Apr 07 14:37:10 Basically, don't use Critical in Ubuntu unless it deserves attention from everybody Apr 07 14:37:17 ok Apr 07 14:37:29 lool: Depends on one's interpretation. It certainly falls under "A bug which has a severe impact on a large portion of Ubuntu users" for every user likely to have that package installed (from https://wiki.ubuntu.com/Bugs/Importance ) Apr 07 14:38:07 heh Apr 07 14:38:20 apart from the fact that it prevents installation thats true :) Apr 07 14:42:40 asac: I don't know about the 3D launcher. I am not testing the netbook image for x86, and we haven't had the 3D launcher in arm images. Apr 07 14:42:53 we do ... Apr 07 14:43:07 its just not started Apr 07 14:45:51 GrueMaster: having this data would help a lot to decide whta to do with it Apr 07 14:47:21 was the openoffice FTBFS discussed in the meeting yesterday ? Apr 07 14:49:16 asac: I don't exactly have a system to test this with. I gave up all of my netbooks this cycle. I would assume it isn't an issue, otherwise it would have been reported. Apr 07 14:49:51 kk Apr 07 14:50:03 focus is sometimes and issue in compiz vs. metacity Apr 07 14:50:06 they behave different Apr 07 14:50:43 well, UNE definately doesnt use compiz anywhere (unless my info is outdated) Apr 07 14:51:16 its surely not depending on the launcher though Apr 07 14:51:22 ogra: doesn't x86 use it? Apr 07 14:51:48 compiz ? Apr 07 14:51:53 i dont think so Apr 07 14:52:21 until recently you couldnt run compiz together with clutter (that changed but i dont know when exactly) Apr 07 14:54:12 definately no compiz in http://cdimage.ubuntu.com/ubuntu-netbook/daily-live/20100406.1/lucid-netbook-i386.manifest Apr 07 14:54:30 That's right. I had compiz & clutter confused. I just got up, need caffeine. Apr 07 14:54:34 so its either maximus or metacity Apr 07 14:55:13 hi all. anyone tried to use git from 10.04 on an ARM target? 10.04 has git 1.7.0, and when I clone any public repo, I have some fatal errors pack signature mismatch and index-pack failed. I am running latest 10.04. On x86 with 10.04 I can clone the same tree without any problem. Apr 07 14:55:16 i would put my bets on maximus since thats the piece that does the fullscreening of the apps *after* the app is up Apr 07 14:55:44 * ogra happily doesnt touch git :) Apr 07 14:55:52 That's what the bug is against. Apr 07 14:56:54 GrueMaster, have a look at /apps/metacity/general/focus_new_windows Apr 07 14:57:21 and in fact, my system is screwed up after git clone failed. for example, I can't apt-get install anymore, it fetches the packages and then fails with Can not write log, openpty() failed (/dev/pts not mounted?) and Sub-process /usr/bin/dpkg received a segmentation fault.... after reboot apt-get install works fine. Apr 07 14:57:23 and fucus_mode Apr 07 14:57:27 *focus Apr 07 14:59:36 sounds like mountall didnt mount your /dev/pts (which ususally does not harm, just spills that warning) Apr 07 14:59:57 focus mode is set to click, and focus_new_windows is set to smart. Apr 07 15:00:17 hmm, i think thats ok Apr 07 15:01:14 ndechesne: if you get some apt-get install errors and random other errors after using a more recent git, it suggests you rendered your hardware or kernel unstable with the new git Apr 07 15:01:30 ogra: don't think so. if i do the same apt-get install before running git clone, it works fine. for example, after git clone fails, running ubuntu-bug gives me a seg fault... after reboot ubuntu-bug works again. any chance you can try to git clone on a real system as well? Apr 07 15:01:38 ndechesne: Typically, dpkg segfaults are attributed quite often to unstable systems Apr 07 15:02:40 lool, ogra: by the way, I downloaded git 1.6.5 source code and built it (native). and this version of git works fine with the same repo. so clearly there is something wrong with the current git version in 10.04 for armel. Apr 07 15:03:39 ndechesne: Yeah but if your system is unstable as a result of using it, it might be more interesting to diagnose the unstability Apr 07 15:03:56 It's probably triggered by accident with a particular version of git Apr 07 15:04:28 ndechesne: If you like to exclude git, you could run it under qemu-system-arm :-) Apr 07 15:04:44 lool: do you have an ARM target handy? can you try to see if it's reproduceable on your side? Apr 07 15:04:56 lool: ;-) I was expecting you would tell that to me! Apr 07 15:05:26 I'm too predictable Apr 07 15:05:35 tsk, these french people, always in agreement Apr 07 15:05:44 :) Apr 07 15:05:56 ndechesne: I have an ARM target ready, but crchhchcrchhh I'm going in a tunnel crhcrchr Apr 07 15:06:12 hahaha Apr 07 15:06:12 * lool drives back home now Apr 07 15:06:15 * lool & Apr 07 15:06:16 ttyl Apr 07 15:11:23 omg... Apr 07 15:11:40 I copied lucid-netbook armel image to sd card Apr 07 15:12:05 I rather expected vfat+ext3 combo then one big vfat partition with 430MB squashfs on it Apr 07 15:12:47 should I dig for usb stick for rootfs? Apr 07 15:12:49 ndechesne, so git clone git://gitorious.org/x-load-omap3/mainline.git with the latest git.-core from lucid and then apt-get install xterm works flawless here on an imx51 system Apr 07 15:12:56 *git-core Apr 07 15:13:36 hrw, if we would use ext3 for our images we would need to have the image build scripts running as root Apr 07 15:14:02 ogra: OE builds ext2/3 images as user Apr 07 15:14:05 in fakeroot Apr 07 15:14:15 the .img files ? Apr 07 15:14:20 yes Apr 07 15:14:22 ogra: argh... i am running on OMAP board, and same repo fails... Apr 07 15:14:34 how do you write to ext3 .img without loop mounting them Apr 07 15:14:43 genext2fs tool Apr 07 15:14:59 hmm, havent heard about that, sounds intresting Apr 07 15:15:06 though not for our live images Apr 07 15:15:11 IMAGE_CMD_ext3 = "genext2fs -b ${ROOTFS_SIZE} -d ${IMAGE_ROOTFS} ${DEPLOY_DIR_IMAGE}/${IMAGE_NAME}.rootfs.ext3 ${EXTRA_IMAGECMD}; tune2fs -j ${DEPLOY_DIR_IMAGE}/${IMAGE_NAME}.rootfs.ext3" Apr 07 15:15:17 there is no point for them to be ext3 Apr 07 15:15:46 since the partition is only a container for the squashfs anyway Apr 07 15:16:55 hi prpplague Apr 07 15:16:56 ndechesne, my /dev/pts isnt mounted either btw (running in a chroot) and i get the same message without subsequent segfault Apr 07 15:17:19 so its clearly system instability of any kind Apr 07 15:17:36 hrw: hey bud Apr 07 15:17:46 hey prpplague Apr 07 15:17:51 any canonical folks here using the beagle board? Apr 07 15:17:54 XorA: hey hey Apr 07 15:18:10 XorA: i heard something on the radio yesterday that made me think of you Apr 07 15:18:19 prpplague, we're still waiting for a working kernel, but yes generally we do Apr 07 15:18:29 prpplague: OH NO! Apr 07 15:18:43 XorA: new flavor of potato crisps, haggis flavor Apr 07 15:18:51 prpplague: I hate haggis Apr 07 15:18:52 lol Apr 07 15:19:11 XorA: hehe, for some reason i thought of you when i heard that Apr 07 15:19:57 ogra: any way i could pick your brain about the things you'd want to do to differently on the beagle to make it easier for desktop dev? Apr 07 15:20:29 well, the 10.04 image will exactly look like the ubuntu-netbook image on other arches Apr 07 15:20:48 beyond that we build a server image and a netinstall image Apr 07 15:20:55 neither of them work properly yet Apr 07 15:21:04 ogra: sorry i mean hardware wise Apr 07 15:21:06 but will do so at release time (i hope :) ) Apr 07 15:21:10 * XorA needs to find room on desk for yet another beagle! Apr 07 15:21:28 ogra: i.e. different connectors or button options Apr 07 15:21:32 prpplague: is zippy2 for sale now? Apr 07 15:21:41 XorA: for about a month Apr 07 15:21:47 XorA: trainer went on sale monday Apr 07 15:21:57 * amitk wonders if someone makes rubber feet for the beagles so that can stack on top of each other... Apr 07 15:22:03 ok, ubuntu boots on my beagleboard Apr 07 15:22:05 for ubuntu I can see having network being essential Apr 07 15:22:20 * prpplague accidentally wore his prescription sunglasses into the office this morning Apr 07 15:22:24 prpplague, hmm, not really, currently we assume a USB disk for the rootfs, USB kbd and mouse and plan to re-use the SD install media as kind-of-bootfloppy to carry /boot Apr 07 15:22:33 now i have to go all the way back to the car to get my other pair Apr 07 15:22:46 (its a quick shot image so we go the path of least resistance with the setup atm) Apr 07 15:23:09 ogra: what about having a usb->rs232 built in rather than the generic rs-232 connector? Apr 07 15:23:29 ogra: or possibily a case of some type like a STB format Apr 07 15:23:45 sure, a case would be cool Apr 07 15:24:07 STB is a bit big though, i would go with nettop format Apr 07 15:24:07 prpplague: if you can do rs232+jtag in one... Apr 07 15:24:15 ogra: if you had a case for it, is there anything special you'd want to add to the case Apr 07 15:24:23 holes :) Apr 07 15:24:40 prpplague: make the top of the case like a lego brick, then we can build beagle walls Apr 07 15:24:44 and indeed the power button would have to go through somehow Apr 07 15:24:48 prpplague: I have a lacie hard disk like that Apr 07 15:24:57 heh.. 1280x720MR does not fit on my 20" lcd ;( Apr 07 15:25:07 stackable beagleboards would be nice Apr 07 15:25:10 probably a power LED would also make sense Apr 07 15:25:12 ogra: might providing a rough dimensions of a case that would be good? Apr 07 15:25:36 prpplague, well, add 1/2 cm around the board Apr 07 15:25:59 and as high as needed plus a bit of space Apr 07 15:26:11 1.5cm or so Apr 07 15:27:16 ogra: i was looking at a case size of 7x5.5x2.5 Apr 07 15:27:32 ogra: but this would include some accessory items inside as well Apr 07 15:27:35 yeah, sounds good Apr 07 15:28:43 * hrw -> out Apr 07 15:28:44 ogra: you'd have two external leds, two external user buttons, power led, built in usb->rs232 for the console uart, proto area , and a secondary sd/mmc slot Apr 07 15:28:46 hrw: later Apr 07 15:29:16 ogra: reset button and power switch Apr 07 15:29:19 how do you get the secondary SD attached ? Apr 07 15:29:32 ogra: via the expansion header Apr 07 15:29:43 ah, thats why you need extra space then Apr 07 15:30:08 but yes, that sounds like a device i'd buy :) Apr 07 15:30:22 ogra: well there are some other features that will go in there, thats just the starting port Apr 07 15:31:00 ogra: what would be the median price range you'd pay for something like that? Apr 07 15:31:19 below 200€ Apr 07 15:31:34 ogra: does that include a beagle? Apr 07 15:31:43 yes, indeed Apr 07 15:32:31 beagle is at 120€ or so atm ... so i would be willing to pay up to 80€ for a case Apr 07 15:32:36 ogra: ahh, target price of the board+case (without beagle) is about 42EUR Apr 07 15:32:44 ah Apr 07 15:32:58 ogra: you'd buy it for that price? Apr 07 15:33:17 yes, but i usually dont look at the price if i want something :) Apr 07 15:33:21 hehe Apr 07 15:33:35 i'm probably the wrong one to actually define a price here Apr 07 15:33:50 ogra: the one item that i still haven't gotten good feedback on is the builtin usb->rs232 for console uart Apr 07 15:33:54 indeed i would also buy it for 42€ Apr 07 15:33:56 ;) Apr 07 15:34:09 ogra: is that something that added value to you? Apr 07 15:34:16 XorA: what about you? Apr 07 15:34:19 well, i'm not a big fan of serial at all Apr 07 15:34:25 prpplague: I prefer real RS-232 as Ive found power cycling the boards seems to reset usb-rs232 chips Apr 07 15:34:36 prpplague: price is awesome though Apr 07 15:34:49 since we started ARM i'm trying to get the ubuntu setup to a point where you never *need* to use serial actually Apr 07 15:35:05 XorA: depends on how you implement the usb-rs232, alot of folks like the gumstix don't implement it properly Apr 07 15:35:18 prpplague: zoom2 is actually quite stable Apr 07 15:35:24 XorA: what about the case size? Apr 07 15:35:33 though given that there will likely still be a long way to go until thats fully achieved having usb->rs232 surely makes sense Apr 07 15:36:31 ogra: you usually use the OTG port on the beagle as a serial gadget or network gadget to work directly with the beagle? Apr 07 15:36:46 XorA: didn't i give you a trainer board? Apr 07 15:36:54 no, i usually use RS232 with my USB/serial adapter Apr 07 15:37:03 ogra: ahh ok Apr 07 15:37:21 ogra: you use the OTG port much ? Apr 07 15:37:22 i usually use the mini USB port for powering the board if i want to work on it on a train etc :) Apr 07 15:37:32 ogra: ahh Apr 07 15:37:39 beyond that i dont use it much, no Apr 07 15:37:44 ahh interesting Apr 07 15:38:03 we might probably play with installation via OTG options later # Apr 07 15:38:06 prpplague: sounds ok with the size, my current case is 3.25x3.5x1.5 Apr 07 15:38:11 ogra: just curious, how do you carry your beagle on the train? got a box or something for it? Apr 07 15:38:18 my bag Apr 07 15:38:28 and two USB cables coming out of it Apr 07 15:38:30 ogra: just leave it in your bag and run the cable out? Apr 07 15:38:35 going to my lappie :) Apr 07 15:38:38 yeah Apr 07 15:38:42 interesting Apr 07 15:39:03 the two cables i assume is one for power and the other to your usb->rs232? Apr 07 15:39:11 right Apr 07 15:39:15 dandy Apr 07 15:39:19 ogra: thanks a bunch for the info Apr 07 15:39:36 welcome :) Apr 07 15:39:40 ogra: you work for canonical? Apr 07 15:39:45 yep Apr 07 15:39:49 dandy Apr 07 15:40:09 XorA: what kind of case are you using? Apr 07 15:40:37 prpplague: clear perspex thing koen sent the board it :-) Apr 07 15:40:43 ogra: oh one more thing, are you planning to purchase a beagleXM ? Apr 07 15:40:48 XorA: ahh Apr 07 15:41:25 prpplague, yes, given that we actually want to provide ubuntu netbook images and the 256M are the lowest edge for running such an image Apr 07 15:41:45 so the C4 doesnt really cut it Apr 07 15:41:52 ogra: thanks Apr 07 15:42:07 * prpplague goes to get his glasses from the car Apr 07 16:02:43 ndechesne: Back now; so what is it you're doing, and how can I reproduce it? Apr 07 16:04:27 lool, Apr 07 16:04:33 W: Bizarre Error - File size is not what the server reported 171880 114587 Apr 07 16:04:58 as i understood, the same fails for ndechesne Apr 07 16:10:19 re Apr 07 16:13:56 prpplague: make a board with: 4xport usb hub, power socket (+ power cable from board to BB), ethernet, reset button Apr 07 16:16:11 hrw: you mean for the RevC boards? Apr 07 16:16:15 yes Apr 07 16:16:40 prpplague: I have B7, C3, C3+beaglebug so prefer to not buy XM just to get some addons Apr 07 16:17:16 hrw: and the case size? the 7x5.5x2.5 , that acceptable? Apr 07 16:18:11 prpplague: I do not need case basically but any which fit BB+expansion and give access to reset/user key, usb ports, serial, sd, hdmi, audio is fine Apr 07 16:20:47 ~curse lucid netbook image for lack of serial use Apr 07 16:21:16 after ~minute screen go black and I do not even know did my keyboard+mouse were recognized Apr 07 16:30:54 hrw: I thought the current issue was the keyboard/mouse are not recognised? Apr 07 16:32:00 XorA: hrw: right, usb is broken at the moment. And we don't have serial enabled by default Apr 07 16:32:32 but you can add serialtty=ttyS2 to your boot.cmd and compile it into boot.scr to get you a console Apr 07 16:32:42 [ 136.328033] Unhandled fault: external abort on non-linefetch (0x1018) at 0x402e6000 Apr 07 16:32:46 known? Apr 07 16:33:00 amitk: I dropped to uboot and used better cmdline then default Apr 07 16:33:08 [ 0.000000] Kernel command line: splash vram=12M omapfb.mode=dvi:1280x800MR-16@60 file=/cdrom/preseed/ubuntu.seed loglevel=9 printk.time=1 console=ttyS2,115200n8 -- boot=casper Apr 07 16:33:17 hrw: sounds like some module was accessed with clocks off Apr 07 16:33:49 anyway without usb it is unusable Apr 07 16:34:05 but atleast 1280x800MR-16@60 fits my LCD Apr 07 16:34:14 https://bugs.edge.launchpad.net/ubuntu/+source/linux-ti-omap/+bug/541030 Apr 07 16:34:16 Launchpad bug 541030 in linux-ti-omap (Ubuntu Lucid) (and 1 other project) "omap kernel musb/ehci ports not enabled on beagleboard (affects: 1) (heat: 12)" [High,Triaged] Apr 07 16:35:19 you can change the configs according to the that bug and you should have usb, and compile your own kernel until we're able to upload a new kernel Apr 07 16:36:18 not today - I need to get openjdk-6 crosscompiled Apr 07 16:50:35 amitk: do the ubuntu boot scripts now auto-build the stuff to go on mmcblk0p1? Apr 07 16:54:12 DanaG: this is the u-boot config: Apr 07 16:54:15 fatload mmc 0:1 0x80000000 /casper/uImage Apr 07 16:54:15 fatload mmc 0:1 0x81600000 /casper/uInitrd Apr 07 16:54:15 setenv bootargs 'splash console=ttyS2,115200n8 console=tty0 serialtty=ttyS2 vram=12M omapfb.mode=dvi:1280x720MR-16@60 file=/cdrom/preseed/ubuntu.seed -- boot=casper' Apr 07 16:54:18 bootm 0x80000000 0x81600000 Apr 07 16:54:44 so, if you're asking if root is hardcoded to mmcblk0, then no Apr 07 16:54:57 this is for the live image, though Apr 07 16:55:02 No, I mean, when you update-initramfs, does it build an initrd? Apr 07 16:55:06 On the installed system. Apr 07 16:55:10 Right now I have to do that manually. Apr 07 16:55:34 hmm, what's "serialtty"? It seems a bit redundant. Apr 07 16:56:13 I also use serial console on my host system, and it sucks that it disables plymouth splash on local tty if a serial console is present. Apr 07 16:56:34 sounds like a ogra question ^^^ Apr 07 16:57:39 DanaG: i think my bootargs might've changed from default as I hack around the deficiencies of the first ubuntu image Apr 07 17:01:02 oh yeah, also... I hope you'll build the g_audio and g_file (and such) drivers. Apr 07 17:03:21 DanaG, serialtty is a casper option while console= is a kernel option Apr 07 17:03:46 ogra_cmpc: no, the question before that Apr 07 17:03:47 it is responsible for creating a getty process during boot in the live image Apr 07 17:04:25 the update-initrsmfs one ? Apr 07 17:05:41 * hrw -> out Apr 07 17:07:22 ogra_cmpc, asac: Would you guys have any issue if we renamed the d-i omap images to omap3? Apr 07 17:07:45 yeah, I was wondering how that copying to the vfat partition would be done automatically. Apr 07 17:08:00 (compare it to x86, where grub does that stuff automatically). Apr 07 17:10:40 ogra_cmpc, asac: Actually I see the kernel is called -omap; I wonder whether it should really be called -omap3 Apr 07 17:11:22 amitk: ^ your thoughts? I see the description properly mentions OMAP3, and given that we're likely to have an omap4 one at some point, I wonder whether it would make sense to name everything -omap3 and avoid the ambiguous -omap name Apr 07 17:19:34 lool, you know i dont have any issues with you renaming stuff but if you plan to do that please do it in all places Apr 07 17:19:58 ogra_cmpc: did I miss any above? Apr 07 17:20:49 lool, well, inside d-i there are several places as well as in the seeds Apr 07 17:21:01 (installer and boot seed specifically) Apr 07 17:21:40 i dont care how the images or the kernel package are called as long as nothing breaks which we have already in place Apr 07 17:22:52 for 10.04 i dont mind at all how anything is called for later stuff we need to touch the names in any case though Apr 07 17:23:13 pamo :D Apr 07 17:39:47 lool: in the future we will want to support omap4, and I don't plan on making two different kernels for omap3 and 4. Apr 07 17:41:29 amitk, you have to Apr 07 17:41:40 ogra_cmpc: why so? Apr 07 17:42:05 the omap4 patchset is incompatible to the omap3 one for now Apr 07 17:42:36 ogra_cmpc: not true, 2.6.34 already has multi-omap support (omap2, 3 and 4 in the same kernel) Apr 07 17:42:37 though i'm sure help on making them work together will be appreciated :) Apr 07 17:43:10 M will probably be 2.6.35+ Apr 07 17:47:20 oh yeah, I think the "M" name (forgot what it was) has one too many syllables. Apr 07 17:47:54 maverick meercat Apr 07 17:48:00 ? Apr 07 17:48:01 5 syllables. Apr 07 17:48:12 Most other names have had only 3 or 4. Apr 07 17:48:13 so did jaunty jackalope Apr 07 17:48:28 that is 5 as well Apr 07 17:48:29 Still sounded a bit smoother than maverICK Apr 07 17:48:37 the sound is rather sharp. =þ Apr 07 17:48:53 the release will be cutting :) Apr 07 18:11:12 Yeah we have to wait till 2.6.34 for mult-omap, even with the rc's i haven't got it to build correctly yet (omap3 + omap4) Apr 07 18:11:37 hi. i'm trying to take arm ubuntu for a spin inside qemu using the netboot iso from lucid alpha Apr 07 18:11:40 *beta Apr 07 18:11:53 rcn-ee: mult-omap? Apr 07 18:12:09 i'm not having any luck with that. i get a black screen. Apr 07 18:12:21 omap2 + omap3 + omap4 boards booting with one kernel (omap2 is armv6 so only debian) Apr 07 18:12:34 * prpplague uses real hardware and not qemu Apr 07 18:12:37 rcn-ee: ahh Apr 07 18:12:46 * desrt doesn't have real hardware :) Apr 07 18:13:39 it makes things very interesting prpplague and good for one image fits all boards distrubtions... but end users will probally disable it when they build their unique images... Apr 07 18:14:16 rcn-ee: yea i wasn't going to make a comment, just figured you had a reason to do that kind of config Apr 07 18:15:59 desrt, if you are to lazy to debug use rootstock to build an ubuntu-minimal image (see topic) Apr 07 18:16:19 ps, congrates, saw your boards where picked up by sparkfun... Apr 07 18:16:19 ya. i'm just installing rootstock now :) Apr 07 18:16:37 btw, FYI for those beagle board users, Trainer boards are now available - http://www.tincantools.com/product.php?productid=16149&cat=0&page=1&featured Apr 07 18:17:36 Hey ogra any chance can i get that 'create initramfs' patch for rootstock merged... I've been playing all week with it, and it speeds up boot times immesilly, my 'beta2' image tomorrow will have a uIntrid bydefault.. Apr 07 18:18:15 rcn-ee, yeah, i havent done the final rootstock upload yet, that patch is still on my list Apr 07 18:19:04 we still need to find out about the quem hang and if there is a way to work around in rootstock, i'm holding back the final release for lucid as long as i can Apr 07 18:19:05 okay awesome! Apr 07 18:19:39 currently omap consumes all my time Apr 07 18:19:42 yeah, i've noticed those when building 'netbook' image and it's sticking on iso-codes again... Apr 07 18:20:37 (and being on the road for the last 10 days didnt really help here) Apr 07 18:20:42 ogra_cmpc: man. this got a lot easier since last time i looked Apr 07 18:20:47 i think i'm going to cheat.. build a minimal image... apt-get the rest on a beagle, then retar it up for release.. Apr 07 18:20:53 ogra_cmpc: are you responsible for this? Apr 07 18:21:02 oh fun.. state side again? Apr 07 18:21:02 desrt, yep Apr 07 18:21:06 props, dude :) Apr 07 18:21:21 desrt, and lots of community people that helped with patches (like rcn-ee ) Apr 07 18:21:50 you almost have me thinking the qemu-user+chroot is the way to go now Apr 07 18:21:57 laughs, thanks.. But we make ogra do all the real work.. ;) Apr 07 18:21:58 heh Apr 07 18:23:02 i'm just the guy with the glue tube and the dirty fingers ;) Apr 07 18:23:42 i love how libdrm-noveau -intel and -radeon are being installed into my arm system Apr 07 18:23:49 lol Apr 07 18:24:25 you never know in 'theory' nouvea might some day work with the 'tegra2' Apr 07 18:24:47 now that would be sweet Apr 07 18:25:14 wtf plymouth support in the minimal install Apr 07 18:25:21 yep Apr 07 18:25:24 that should really be in -desktop Apr 07 18:25:29 nope Apr 07 18:25:55 its used everywhere, even in text boots Apr 07 18:25:56 it looks pretty cool.. Apr 07 18:27:03 cryptfs prompts are routeed through it as i understood Apr 07 18:27:34 no matter which kind of theme is used Apr 07 18:28:08 ah Apr 07 18:28:18 oh by the way, i know oem-config works on freescale hardware, how much ram/swap do you guys set aside? I'm running out of memory on my Bx .. Apr 07 18:28:18 (text is just another form of theme in plymouth) Apr 07 18:28:44 rcn-ee, for the gtk or the text version ? Apr 07 18:28:58 right now i'm just playing with the text version... Apr 07 18:29:11 that should work in 128M Apr 07 18:29:43 worst case enable compcache in your kernel and set the initramfs.conf variable Apr 07 18:29:56 that should extend your ram a bit Apr 07 18:29:59 that's what i thought too... it's dieing with out of memory errors.. (it's also drawing the dmesg console and oem-config console on top of each other onthe screen.. ;) ) Apr 07 18:31:04 though we just found that compcache changed since it entered staging, not sure its still compatible with the ubuntu userspace setup Apr 07 18:32:18 (we use the old version in ubuntu kernels still) Apr 07 18:32:38 thanks.. I'll play with merging that into my 2.6-dev tree and retest... Apr 07 18:40:05 nice. seems to be going. Apr 07 19:30:01 lool: afaiui we are working towards identical kernel for both Apr 07 19:53:46 asac: Ah right, forgot about that multi omap thingie Apr 07 22:33:52 eggonlea: I've been doing some testing with NFS root on ARM, if your interested, I think we can get this to work well with your setup Apr 08 01:23:06 NCommander: Great! Could we use NFS root with default uImage/uInitrd/rootfs now? Apr 08 01:23:27 eggonlea: you have to rebuild the uInitrd for NFS mode Apr 08 01:23:31 eggonlea: its not difficult Apr 08 01:23:41 eggonlea: once you do that, it should just work Apr 08 01:24:01 eggonlea: basically, you'll have to TFTP a boot.scr file that will download a uImage and uInitrd, set the command line, and boot Apr 08 01:24:12 But why? I've looked into the initramfs which should have the ability to mount NFS already. Apr 08 01:24:37 eggonlea: it does, but you need to flip the switch to tell it to actually try and do that ;-) Apr 08 01:24:45 And actually I've tried that before with default initramfs. It did work. Apr 08 01:25:00 eggonlea: hrm, interesting. Where'd you run into problems then? Apr 08 01:25:36 the problem is in 1) /etc/fstab; 2) /etc/network/interface; and 3) /etc/init/network-manager Apr 08 01:25:40 * NCommander got tripped up by finding out you have to set rw on the command line Apr 08 01:25:43 1 is easy Apr 08 01:25:45 2 is easy Apr 08 01:25:48 ah Apr 08 01:25:49 3 needs an off switch Apr 08 01:25:55 yes Apr 08 01:26:05 Here's the fstab I'm using on my board Apr 08 01:26:25 by specifying "nfsroot=xxx" and "boot=nfs" at the same time, I could use the default initramfs. Apr 08 01:26:32 eggonlea: http://paste.ubuntu.com/410807/ Apr 08 01:26:42 eggonlea: ooh, thats good to know, that means the documentation is out of date Apr 08 01:26:42 and I commented out all items in fstab to use NFS. Apr 08 01:26:55 Yes, the doc is really old. Apr 08 01:27:10 You can force network-manager to off by specifying the interface in manual mode in/etc/interfaces/network Apr 08 01:27:26 auto eth0 Apr 08 01:27:26 iface eth0 inet manual Apr 08 01:27:33 Put those lines in, and NM will ignore eth0 Apr 08 01:27:35 Once "boot=nfs" is there, initramfs would deal with "nfsroot=xxx" correctly. Apr 08 01:27:46 eggonlea: seems your one step ahead of me Apr 08 01:27:53 Yes, I deal with interface eth0 as manual Apr 08 01:28:00 hrm, that should turn NM off Apr 08 01:28:04 If it doesn't, that's a bug Apr 08 01:28:13 and can be worked around with apt-get remove network-manager :-) Apr 08 01:28:21 It doesn't work until I remove NM from init. Apr 08 01:28:27 eggonlea: is your eth device 0? Apr 08 01:28:30 y Apr 08 01:28:34 sometimes it likes to move to eth1, eth2, etc. Apr 08 01:28:38 * NCommander got as high as eth37 Apr 08 01:29:03 * NCommander was personally interested in seeing what happened when it got to eth255 ... Apr 08 01:30:15 http://paste.ubuntu.com/410808/ Apr 08 01:30:27 see line 25-53 of my /etc/init/mountall.conf Apr 08 01:30:46 It's a bug. Apr 08 01:30:56 We could follow up this. Apr 08 01:31:15 eggonlea: if interfaces.nfs doesn't exist, it should work properly, my /etc/network/interfaces persists reboots Apr 08 01:31:54 I'm wondering if standard Ubuntu could integrate something like that to deal with NFSROOT and normal boot smartly. :) Apr 08 01:32:27 eggonlea: integrate in which way? Apr 08 01:32:41 * NCommander notes that except for LTSP, we don't do a lot of network booted except for Mythbuntu Apr 08 01:33:13 but the current initramfs DOES include NFS support already. Apr 08 01:33:26 why not make one step forward to complete it? Apr 08 01:34:17 eggonlea: definately something we can look at fixing although its too late for said work to land for 10.04. We can have a spec at UDS and look at improving the existing NFS support there Apr 08 01:34:24 if ethx is configured already, just have NM leave it there (don't re-configure again). Apr 08 01:34:36 That's fine. Apr 08 01:34:44 eggonlea: NM is designed for normal users. We're not normal users :-/ Apr 08 01:35:20 * eggonlea agrees. Apr 08 01:36:20 Could NM be dynamically switched off (when initramfs detects "boot=nfs") by passing some parameter easily? Apr 08 01:36:22 eggonlea: what do you use NFSroot for specifically? Apr 08 01:37:02 eggonlea: hrm, it won'tbe hard to extend its init script to do extactly that. If your doing NFS root, its unlikely to want network-manager (unless you want to configure a WLAN device, but thats really pushing the usecase) Apr 08 01:37:28 I use NFSroot the most time I use Ubuntu. It helps a lot when I transfer files between my development PC, share the powerful utilities in PC, etc. Apr 08 01:38:17 I could even restore a broken rootfs easily by just tar and untar a backup. Apr 08 01:39:41 I only use HD as root when installing a fresh one from livecd and release customized Ubuntu to customers (the released tarball would contains the scripts I pasted just now to support both of HD and NFS). Apr 08 01:40:52 eggonlea: that's true. You can even do squashfs over NFS to get a clean environment every time Apr 08 01:58:49 * DanaG wishes networkmanager wouldn't ignore usb0 ("can't determine driver") Apr 08 02:43:58 DanaG: Please file a bug about that: it's surely fixable. Apr 08 02:44:48 I know why it can't determine driver: /sys/class/net/usb0/device/driver symlink doesn't exist. Apr 08 02:45:22 Then that's a kernel bug (and still fixable) :) **** ENDING LOGGING AT Thu Apr 08 02:59:56 2010