[07:44] Does Ubuntu/Debian have any mechanism for deferring some action after next reboot and not more than once? [08:03] janimo, kernel build failed :( [08:04] (it seems to use Werror ?!?) [08:08] oh, only -Werror=maybe-uninitialized [08:35] janimo, http://paste.ubuntu.com/1247107/ i guess we should drop all these hardcoded -Werrors ;) [09:01] ogra_, I am on it. I built using 4.6 and worked fine :( [09:01] 4.7 is pickier [09:02] I cross-built on precise forgetting there may be toolchain issues if uploading to quantal [09:02] I hate it when I need to bump the ABI with no .deb available and have to do the manual creation of those files and dirs [09:03] ogra_, lots of Werrors in the tegra kernel btw. I'd keep them there thougg but I'll pay more attention when testing the build [09:13] http://marcin.juszkiewicz.com.pl/2012/09/28/lets-take-a-look-at-arm-boards-again/ - opinions? [09:15] hi ! has the ubuntu-arm been installed on commercial device lie : archos 32,43 or other omap based mobile devices [09:16] janimo, well, according to ppisati thats not usual in kernel builds [09:16] hrw: opinion: 500 - Internal Server Error [09:16] likewise [09:16] janimo, i would drop them, in case anyone rebuilds with an exotic toolchain [09:17] ogra_, probably makes sense. No idea. Don't these catch actual bugs as in userspace code? [09:17] Or enforce some coding conventions. Some are really picky I agree. [09:17] usually warnings arent errors :) [09:17] checkng [09:17] bizulk: should be doable. archos even offers an "open" bootloader. [09:17] and kernel builds always spill warnings [09:18] bizulk: there also are positive reports, but no "1-2-3 click here, done" guide. [09:20] * ogra_ hugs ppisati for the fixing of bug 1045855 [09:20] Launchpad bug 1045855 in linux-ti-omap4 "usb keyboard doesn't work during installation of ubuntu-server on panda" [High,Fix committed] https://launchpad.net/bugs/1045855 [09:20] was the udeb actually missing ? [09:22] hrw, "Error establishing a database connection" over here [09:23] it's available on planet.linaro.org tho [09:24] yeah, i read it on my tablet during the coffeebreak i just had ... there it worked fine [09:25] using the link from g+ though [09:25] hrw: you should mention on that the A10 is not only lacking in mainline kernel, but also no public documentation either [09:25] there is [09:26] url? [09:26] (me is just fighting with some 100% chinese PDFs) [09:26] ...from the dark alleys of the internets [09:26] ugh, no idea anymore, its for the script.bin stuff though ... butu googling definitely got me other pdfs too [09:26] suihkulokki: add comment? [09:27] if you want that one i can push it somewhere [09:27] once I get mysql working again [09:27] ogra_: doesn't count as public docs.. more like leaked docs. [09:27] heh, ok [09:27] yeah, might be leaked .... [09:28] their way of doing their own devicetree type of thing is really annoying [09:28] hrw: i know that its not really in the dev board scene already, but what about all that marvell stuff? [09:29] ++ [09:29] once there are armadaxp boards that workstation thing should be possible [09:29] i mean, they pretty much always have the desired sata. [09:29] its just that the widely available stuff from them is still armv6/armv5 [09:29] the armadaxp's i have seen even take normal dram [09:30] xp is v7 [09:30] no gpu or even framebuffer on armadaxp [09:30] ogra_: is there already an affordable kit? [09:30] quad core 1.2GHz [09:30] suihkulokki, PCI slots ;) [09:30] suihkulokki: what would i need that for? [09:30] LetoThe2nd, i only saw pre-prod [09:31] ogra_: ok, so my assumption "not yet" still holds true. [09:31] yeah [09:31] but there should be some buyable ones soon [09:31] ogra_: need x86 emulation for those :) [09:31] i mean, all those cheap qnap mini nas things make nice hackable things after all. [09:32] suihkulokki, do i ? i thought the server boards with AXP on them will have them by default [09:32] ogra_: you are aware how video cards are initialized? [09:33] openrd ultimate has long had pci anyways. [09:33] suihkulokki, something you should be able to fake in a driver, no ? [09:33] LetoThe2nd: openrd is armv5 [09:33] hrw: exactly what i said above. [09:34] i just mean that pci on arm is not very far away. [09:35] don't know if marvell is still so bi***ing about documentation, though [09:35] ogra_: sure, but someone needs to write the code to do so first [09:35] indeed [09:37] huh, *write* code? [09:37] why can't we just *download* the codes? [09:37] someone please send me the exact steps! [09:38] haha [09:41] ok, works again [09:49] I would like to thanks tu arm-ubuntu community for the DSP integration on OMAP, I'm working on this for 4 Weeks. And finally got the video working on 12.04 (thanks for Robert CN scripts that make a got stuff for downloading a clean kernel image with dsp support). I was going onto this conclusion : "Everybody knows what it is, knows somebody that seen it working, but neved seen it himself neither how to make it works" [09:50] bizulk, did you file the bug as i asked you, so ppisati can put it iun the official kernel ? [09:51] no still not. I may do it know [09:51] where shall I post it ? [09:51] on the bugtracker url i gave you last time :) [09:52] https://bugs.launchpad.net/ubuntu/+source/linux [09:52] 8there is a "report a bug" button on the right [09:52] ogra_:ok I do it know [09:53] thanks, give us the bugnumber here afterwards [09:53] Are we talking about the /proc/config.gz missing file, or the tidspbridge stuff ? [09:53] or both [09:54] config.gz isnt a bug [09:54] (we ship the config as text file already) [09:54] that's what I was thinking [09:55] well, i told you last time that its on purpose ... didnt change yet since ;) [09:55] I expect the answer to be : this is not a but as gstreamer does not embbed the gst-dsp plugin on the dist [09:56] well, if the kernel option can be turned on without issues ... even though we dont have the dsp package you shouldnt need t5o recompile your kernel to use it if possible [09:56] hey hey [09:56] heh [09:56] welcome to the club dholbach [09:56] :) [09:57] I was wondering if we had some good instructions somewhere to get 12.10 on the Mele A1000 :) [09:57] no, but there are community people somewhere with images and install instructions [09:57] miele a1000? a dishwasher? [09:58] i'm trying to get an imager for the zatab working before UDS, the a100 shouldnt be to hard to support from that [09:58] LetoThe2nd, no, I'm afraid not - http://dx.com/p/mele-a2000-1080p-android-2-3-network-multi-media-player-w-sata-usb-hdmi-lan-vga-wifi-4gb-131566 :) [09:58] dholbach: ah dammit [09:58] LetoThe2nd, a dishwasher might make a more interesting news story, I agree :) [09:58] http://www.cnx-software.com/2012/04/28/how-to-create-your-own-debian-ubuntu-image-for-mele-a1000-allwinner-a10-based-stb/ [09:58] ogra_: you're right, recompile it on the omap would be a real pain [09:58] looks nice though. but yet another allwinner. [09:59] allwinner isnt that bad ... [09:59] that it is a8 is bad :) [09:59] ogra_: hehehe [09:59] the same CPU with dual core would be awesome [09:59] thanks ogra_ - I'll try that out [10:00] ogra_: my god this site signing in process is a big armor [10:00] given the price tag it might be a nice sidekick device indeed. [10:00] yeah, the zatab i have here has massive issues though ... [10:00] ogra_: like hdmi not working? ;) [10:01] apps often hang in android [10:01] due to the CPU being maxed out [10:01] dholbach: do you know how all that storage/interface things are glued to the cpu? [10:01] badly :) [10:01] LetoThe2nd, I'm afraid I don't [10:02] all usb, i guess. [10:02] * ogra_ tries to get the MMC to work on his zatab kernel since over a week now [10:02] ogra_: Allwinner A15 will be dualcore A7 iirc [10:02] usb works fien so i can run an ubuntu rootfs ... but i cant convince the kernel to keep the MMC powered on once it booted [10:02] hrw, yeah ... will be ... :) [10:03] ogra_: anyway I keep away from anything in mainline kernel (or on a way to it) [10:03] not in mainline you mean ? [10:04] yes ;) [10:04] :) === XorA|gone is now known as XorA [10:18] ogra_: my bug no must be : 1058022 [10:18] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1058022 [10:18] Launchpad bug 1058022 in linux "no tidspbridge support in kernel." [Undecided,New] [10:18] bizulk, thnaks a lot [10:18] ho great stuff this bot. I dream to install such stuff in my office. [10:19] (lunch time now) [10:27] bizulk, oh, i was hoping you would state the exact options that need to be enabled in the bug so ppisati doesnt need to dig around [10:28] bizulk, the page you linked to isnt helpful for finding that option [10:57] bizulk: i'll take a look [11:27] marvin24, http://paste.ubuntu.com/1247332/ ... not sure whats up with your package cache, but squeak-vm is available for me on all armhf platforms === Quintasan_ is now known as Quintasan [11:40] ogra_: marvin24? [11:40] oh, i mixed you up [11:40] lol [11:40] sorry [11:41] ETOOMANYMARCS [11:46] ogra_: I'll post the config file [11:46] ppisati, ^^^ [11:51] ogra_: you mean squeak-vm or the config file? :) [11:51] the config file, unless you want to add smalltalk support to the kernel now :) [11:53] bizulk: add the config diff to the lp bug you just created [12:19] ogra_, I have uploaded a new ac100 package, I hope it gets through soon and actually builds this time [12:19] yay [12:19] ogra_, would you mind looking at the libaio ftbfs? it crashes the buildd machines ... [12:19] lets bribe infinity ;)+ [12:20] doko, i dont see it on http://qa.ubuntuwire.org/ftbfs/ [12:20] janimo: hey, thanks [12:21] ogra_, https://bugs.launchpad.net/ubuntu/+source/libaio/+bug/1058081 [12:21] Launchpad bug 1058081 in libaio "libaio ftbfs on armel/armhf" [High,Confirmed] [12:29] hmm,. the build seems to work just fine ... but these tests even make my sabre cpu start glowing [12:30] there's a bug report about an invalid test case too [12:31] ho can something that builds in 5 seconds take so long for its selftests, tsk [12:33] doko, http://paste.ubuntu.com/1247430/ [12:33] hmm [12:33] no issues [12:38] doko, where is the ftbfs buildlog so i can compare the two ? [12:38] ogra_, there is none. buildd dies [12:39] i can imagine, these tests are really evil [12:39] i have no panda set up for building atm, but will try on a panda as soon as i have one ready [12:42] doko, the debian/rules file has an uption to pass "nocheck", will anone kill me if i just disable the tests on arm for now ? [12:42] *option [12:43] does this boot ubuntu? http://www.hardkernel.com/renewal_2011/products/prdt_info.php?g_code=G133999328931 [12:43] ogra_, or fix https://bugs.launchpad.net/ubuntu/+source/libaio/+bug/607665 [12:43] Launchpad bug 607665 in libaio "Invalid test case on ARM" [Medium,Confirmed] [12:43] doko, well, that doesnt kill the build [12:44] (here) [12:44] ogra_, with a precise kernel? [12:44] 3.2.0 [12:45] we dont have any recent kernel for mx6 [12:45] (or better: more recent) [12:45] thats all quantal [12:46] ok [12:46] the failed tests doesnt stop the build for me [12:47] and in fact that bit scrolls really fast ... it takes like 10 min for the tests before and after [12:47] and i think thats where the buildd is falling over ... [12:49] I added to the tidspbridge bug the config file, and particular option in comments [12:50] great, thanks :) [13:02] I bug is now in triaged state and used the apport-collect tool. Nothing more is needed from me for now ? [13:03] yeah, should be fine [13:32] doko, libaio seems to build for me on ac100/quantal [13:32] no build logs from the crashing machines? [13:32] janimo, apparently not [13:32] I wonder if it's a kernel issue [13:32] no. I could give it back, and you could poll the build log while the build is running ... [13:32] just waiting for my quantal panda to finish [13:33] looks pretty good yet [13:33] although I am suspecting something weird. my build took a few imnutes [13:33] whereas LP shows 8 hours for previous builds [13:33] janimo, the build or the tests ? [13:33] debuild -us uc [13:33] -uc [13:33] my build (even on a panda with a really slow USB key) is done in under a minute [13:34] oh, i just use dpkg-buildpackage -rfakeroot -b [13:34] in the souorce tree [13:34] the tests take between 10 and 30min though [13:34] I wonder why this is 8 hours then? https://launchpad.net/ubuntu/+source/libaio/0.3.109-2ubuntu1/+build/2964075 [13:35] maybe some tests fail early and do not proceed here === Quintasan_ is now known as Quintasan [13:35] might be [13:36] in any case failing tests dont seem to stop the build anyway [13:36] even the precise log has the failures [13:36] indeed [13:36] so i bet we could as well disable them :) [13:37] dh_testdir [13:37] /usr/bin/make [13:37] /bin/sh: 1: git: not found [13:37] /bin/sh: 1: lsdiff: not found [13:37] make[1]: git: Command not found [13:37] heh [13:37] on LP/armhf I think the timeout of 8 hours or something similar kicked in. Or just some obscure bug. The build+tests are short [13:37] ogra_, yes, it just checks whether it's in a git repo [13:37] seems that log doesnt care at all if *anything* fails [13:37] can be ignored, it took just as little time with git installed [13:37] k [13:38] so maybe a lucid kernel or whatver the builders have is the issue [13:38] precise [13:38] but yeah, very likely [13:40] the precise panda kernel wasnt particulary great [13:40] (dieing on idle etc) [13:41] wow, that 14.p test takes a century [13:50] * ppisati goes out to run an errand, brb [13:50] hmm, whatever that test does, it hits the disk realyl hard it seems [13:51] was probably not a good idea ro try that build on my slowest USB key [13:51] *to [14:01] heh, so my panda has a load of 5 and the test eats up all memory [14:02] i can imagine that might fail with anot 100% correct kernel [14:02] *a not [14:13] doko, libaio just finished on my quantal panda (fresh desktop install from wed.) [14:14] so i would say its clearly the buildd kernel [14:14] please leave a comment in the bug report [14:14] will do [14:18] done and buildlog attached for that one as well [14:18] I have installed the xorg omapfb driver but Xorg doesn't seem to use it : sudo apt-get install xserver-xorg-video-omap3 [14:18] you might need an oxrg.conf, not sure [14:19] that package comes from debian as is [14:19] xvinfo --display :0.0: still returns 'no adaptors present [14:20] Where shalll be this file ? I don't remember [14:20] in /etc/X11 [14:21] ubuntu doesnt use that file anymore by default so you might need to create it [14:21] ogra_: I'll check it [14:21] probably the package ships an example [14:22] no it has not :( [14:25] well the wiki presume that the driver is automatically loaded but I could not figure how without a Xorg.conf file [14:25] check the log :) [14:26] I did [14:29] well, all i know about this package is that it has been removed in 12.10 [14:35] why ? does it mean that the Xorg video driver is not anymore part of it ? [14:36] it is a hacked up xfbdev driver nobody maintains afaik [14:36] and isnt compatible to abi 13 anymore [14:39] Does Angstrom not use it (or won't) neither. [14:39] no idea [14:39] we dont do much with oma3 here anymore [14:39] *omap3 [14:40] but once angstrom updates to the current xorg abi they will likely have to port it if they use it [14:41] ogra_, although interestingly it builds on ac100 quantal which has a 3.1 kernel [14:42] panda precise is 3.2 right? [14:42] janimo, well, the precise panda kernel has issues [14:42] we know it dies idling on some boards for example [14:42] are there any non-panda builders left? [14:42] nope [14:42] where are the slow crappy boards when you need them??? [14:42] but i would just propose to backport the quantal kernel and be done [14:43] you could try on the porter box, i think that still runs a natty kernel [14:43] ogra_, if you install the precise kernel in quantal it should fail if that is the cause right? [14:44] I think the porter box has long been dead [14:44] yep [14:44] no, i heard its back [14:44] at least I could not connect to scheat for a long time, so gave up [14:44] scheat [14:44] ok, did not check in the past two months I guess [14:45] i didnt check at all, Daviey claimed its back on #ubuntu-devel [14:45] and i trust Daviey blindly :) [14:54] aie : it seems that Xorg didn't like my conf === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [15:53] janimo, linux-ac100 is in ... time for meta i guess [15:53] yes, but should need at least a test first [15:53] I'll install [15:53] * ogra_ cant ... else i break my screen [15:53] and in as in past NEW? [15:54] well, cjwatson just said he processed it in #ubuntu-release [15:54] ah ok [15:54] so past NEW i would think [15:56] ogra_, uploaded [15:56] thx [15:57] ogra_, are you planning the graphics driver upload for next week? [15:58] or the weekend, yeah [15:58] and if i manae in time also the codecs [15:58] * ogra_ has to rush to the vet now ... back in ~1h [16:06] hi ! boot.scr is a binary file ? [16:07] how do we edit that stuff ? [16:09] okay it's in the host with mkimage [16:09] bizulk, you don't edit the boot.scr directly, you edit the boot.cmd and then make it into a boot.scr with mkimage [16:09] It is actually a text file with a 72 byte crc blob. [16:10] You can strip the first 72 bytes, then edit it as a text file (boot.script). TO recreate the CRC (needed by u-boot), use "mkimage -A arm -O linux -T script -C none -d boot.script boot.scr". [16:11] (I used to have this documented on http://wiki.ubuntu.com/ARM but it has since been either deleted or moved by others. [16:13] there is not other way to put off the splash screen ? well okay [16:15] Sadly, not easily. You could interrupt uboot via serial console, but that is a pain. [16:18] bizulk, the process may sound hard/scary but it is pretty straight forward [16:18] (modifying the boot.scr) [16:20] lilstevie: first of all, new for me :) (but it was just for having debug message on screen/serial as the xorg omapfb bug the boot process) [16:20] The easiest way to strip the file for editing is with "dd if=boot.scr of=boot.script bs=72 skip=1". Then you can edit it and use the above mkimage command to re-bless it. [16:20] GrueMaster: how do you strip the bytes ? dd ? [16:21] bizulk, GrueMaster gave you the command there :) [16:21] Yes. YOu can also do it with a text editor, but it isn't as clean. [16:21] * GrueMaster notices shift key is sticking again and grumbles profusely. [16:24] ho ! sorry I did not see the cmd just below [16:25] I see that uEnv.txt is still supported [16:30] I have strip off quiet and splash, but I've just got a black screen with a penguin :d [16:34] the xorg omapfb driver seems to be loaded but I do think I have to pass more option to activate DVUI output [16:35] If you want earlier kernel messages and boot messages to go to the serial console, add "earlyprintk=ttyO2,115200 console=ttyO2,115200 console=tty1" [16:35] ho yessss [16:35] I did put the console in bootargs without results [16:59] is "xf86-video-omapfb" an alternative to omapfb ? [17:14] janimo: I'm trying to compile your kernel - do you know how to enable ccache in kernel builds? [17:14] marvin24, cross compiling? [17:15] yes [17:15] there must be some magic string added to debuild [17:15] ah, I do not use that. Ijust have /usr/lib/ccache/arm-linux-gnueabihf-gcc -> ../../bin/ccache as symlink [17:16] so it is systemwide not specific to kernel builds [17:16] and I never checked that it actually works :) The builds seem fast though [17:17] yes, the problem is it doesn't [17:17] but the man page says use --prepend-path [17:17] lets try it [17:19] no luck :-( [17:19] no luck as in nothing is put in ~/.ccache? [17:19] * janimo checks [17:21] yes [17:21] neither symlink nor --prepend-path works [23:35] installing armhf [23:36] had to do a tty1->tty7 dance to view X [23:36] maybe i was impatient for ubiquity to come up though.