=== prp^2 is now known as prpplague === Ursinha` is now known as Ursinha === Ursinha is now known as Guest53271 === rsalveti` is now known as rsalveti === zyga-afk is now known as zyga [09:40] ogra_: hi. are you aware of any problem to play mp3 with 12.04? we get segfault with gst/libmad, and if we rebuild libmad with —disable-aso there is no problem. i am surprised that I don't see a bug for that... [09:41] hmm, havent heard of that [09:41] if you don't hear MP3, it's because of that ;-) [09:41] do you have an ARM h/w running ubuntu close by? [09:42] if you could try a gst-launch playbin2 uri=file:// that would be nice. [09:44] nothing with a desktop install atm (doing low level work here) but i will test that [09:45] ok. thx [10:26] does the daily live image (quantal-desktop-armhf+omap4) install onto the same sd card it's booted from? [10:26] I was expecting to be able to install it onto a USB disk but it just went ahead and installed without giving me a choice of destination [10:30] um, have the images ever been different? [10:30] Laney, no, you need an USB disk [10:31] I don't know what happened then [10:31] and you should get the normal ubiquity partitioning dialogs as on x86 ... if not, thats a bug [10:32] "use whole disk" should default to your USB disk [10:32] I chose timezone, keyboard layout, username [10:32] * ogra_ guesses he needs to do a test install to take a look, stgraber also reported some weird stuff yesterday [10:33] and indeed cfdisk confirms the USB disk hasn't been touched [10:33] well, smells a bit like the installer is in oem-config mode ... stephane had an oem-config related error messsage too [10:34] though i have no idea why that would be [10:34] I can believe that, yeah [10:34] if you look at the SD, do you actually have a squashfs on there ? [10:35] (second partition) [10:35] (in the casper/ dir) [10:45] ogra_: I see no casper/ there [10:45] oh [10:46] which image did you download then ? [10:46] http://cdimage.ubuntu.com/daily-live/current/quantal-desktop-armhf+omap4.img [10:46] argh [10:46] hang on ... [10:46] looks ok to me [10:46] spot the difference [10:46] zcat quantal-preinstalled-desktop-armhf+omap4.img.gz | sudo dd bs=4M of=/dev/sdb; sudo sync [10:47] lol [10:47] K [10:47] i should probably add a quantal section to the wiki :P [10:47] "check your tab completion" [10:48] the img file can just be dd'ed [10:48] no need for the zcat [10:48] compressed? [10:48] I think I got that from some wiki page [10:48] it isnt compressed [10:48] right, as i said, i need to update it for quantal [10:48] oh, the desktop one, yeah [10:48] sudo dd if=quantal-desktop-armhf+omap4.img of=/dev/sdb bs=4M [10:49] already doing it [10:49] good [10:49] theoretically that one should work .... [10:49] *theoretically* [10:49] *whistle* ... [10:50] * Laney coughs [10:53] this looks a lot more like ubiquity [10:54] phew [10:55] i hope the bootloader install works now or hw-detect ... it failed in one of these steps [10:55] * ogra_ isnt sure which of the two [11:13] ogra_: ah, bah, it failed [11:13] how ? [11:13] not sure yet [11:13] hopefully apport collects useful info, otherwise i'll look [11:20] ogra_: bug 1024356 [11:20] Launchpad bug 1024356 in ubiquity "quantal armhf+omap4 daily live 2012-07-12 crashed" [Undecided,New] https://launchpad.net/bugs/1024356 [11:21] need any more? [11:21] looks fine, thx [11:21] cool [11:21] debconf.DebconfError: (10, "oem-config/enable doesn't exist") [11:21] aha [11:21] seen before? [11:21] thats the same bug stgraber saw yesterday [11:22] nice, feel free to dup it [11:22] i dont think he filed it actually :) [11:22] but he pinged me about it [11:22] tsk [11:26] * Laney retreats back to precise-server :P [11:34] ppisati, https://launchpadlibrarian.net/110025692/UbiquitySyslog.txt ... lots of "swapper/0: page allocation failure" ... and "smsc95xx 1-1.1:1.0: eth0: kevent 2 may have been dropped" in there [11:36] ogra_: was reading it [11:37] ah, k [11:51] is there a way to "open" the image and replace it with my own kernel? is it the live install? [11:51] ogra_: ^ [11:52] the page allocation failures are a dup of this [11:52] https://bugs.launchpad.net/ubuntu/+source/linux-ti-omap4/+bug/746137 [11:52] Ubuntu bug 746137 in jasper-initramfs "Page allocation failure on Pandaboard and Beagle XM" [Undecided,Fix released] [11:52] and the workaround listed there didn't solve the problem [11:52] do we still use it? [12:20] ppisati, you should be able to replace uImage on the first partition as long as the module symbols match [12:20] ogra_: i'll try [12:20] the initrd is a capser initrd, thats a bit more tricky to rebuild [12:20] *casper [12:20] uhm [12:20] ogra_: btw with live installer [12:21] "Failed to unmoun partitions" [12:21] /cdrom [12:21] that shouldnt cause a fatal error [12:21] (x86 USB installations dont have /cdrom either) [12:22] it's stuck there [12:22] and i can't go further nor back [12:23] todays image ? [12:23] yep [12:23] just downloaded [12:23] there was a ubiquity upload .... [12:23] let me retry [12:23] but that wont be on the image i guess [12:24] (uploaded 11:35 , images were published at 12:00 ... unlikely that it made it on todays build) === zyga is now known as zyga-food [12:34] ogra_: so, /dev/mmcblk0p2 is mounted on /cdrom [12:34] ogra_: and the installer "needs to do changes to the partition table" but can't since a partition is stil mounted (/cdrom indeed) [12:34] err [12:34] are you trying to install to the SD ? [12:35] ogra_: seems it's impossible to installe don the sd card [12:35] thats why the partitioner shows that big fat warning if you select it :) [12:35] it means we can't install anymore on the sd card then [12:35] you can do it but need to create the rootfs partition beforehand and tell the partitioner to not recreate the partition table [12:36] what a mess [12:36] ok [12:36] i'll try on the usd hdd [12:36] (iirc the message the partitioner has says exactly that) [12:36] its an old and known restriction ... [12:36] does it say we can't install on the same card? [12:36] whatever [12:36] let me try with the usb hdd [12:36] you could take over the toolchain maintenance from infinity, i know he has ideas since a while how to fix it, but he never finds the time for it ) [12:37] ;) [12:37] no way, infinity is doing a great work! :) [12:37] btw [12:37] "toolchain maintenance"? don't we get it from linaro? [12:38] i mean the toolchain [12:39] sure, but thats far from prefect ... there are still enough bugs to weed out ... and also bootstrapping a new arch like armhf (or aarm64 next) takes some manpower ... [12:39] i guess so [12:39] which is usually adams :) [12:39] but i though we got it from linaro and we didn't touch it [12:39] we ofte enough do [12:39] *often [12:39] look at who uploads eglibc ;) [12:41] he usually works very close with linaro === zyga-food is now known as zyga [12:41] anyway, i know he has ideas for a fix, but that might require huge amounts of ram i fear (moving the installer to a tmpfs or so) [12:42] the partitioner shouldnt show the device to you though ... i think we should have a bug for this ... [12:42] i.e. if there is a pre-made partition on the device to install to, it should show it ... if the partition isnt there it should be hidden [12:43] that would make far more sense than the current warning nobody reads [12:45] to me the biggest drawback is that user now needs another medium besides the sd card to do the installation [12:45] right [12:45] thats on purposee [12:46] really? [12:46] demoing a system from the slowest media you can use on it isnt really a helpful demo [12:46] true [12:46] and thats largely what these images are, demos [12:47] beyond that, we really want to get rid of any arch specific deltas when not needed [12:48] * ogra_ runs an ubiquity debug session to get more info [13:02] ok [13:02] installation finisehed here [13:02] reboot [13:02] (i can;t type today) [13:04] its friday the 13th ... nobody expects you to be able to type :P [13:04] or to survive... [13:04] heh [13:04] cool [13:05] it took a bit [13:05] but it booted [13:05] nice [13:05] but now i need an sd install only (sprinting next week and i don';t want to bring my hdd around) [13:05] * ppisati -> reinstall preinstalled [13:06] don't we produce omap4 preinsatalled anymore? [13:06] nope [13:06] there's only the server [13:06] i'll use that [13:07] as i said above. you can install to a pre-created partition on the SD [13:08] server is faster [13:08] i'll leave the desktop version @ home [13:08] and [13:09] btw, the splash screen didn't work during the boot [13:09] yep, known [13:09] bug 1018907 [13:09] Launchpad bug 1018907 in plymouth "plymouth in quantal on arm does only boot with black screen" [High,Confirmed] https://launchpad.net/bugs/1018907 [13:22] ogra_: trying with the latest ubiquity it actually worked here... but according to cjwatson and the ubiquity diff, there wasn't any fix in there... [13:22] so it's some kind of race and I got lucky, I guess [13:22] stgraber: omap4 live daily? [13:24] stgraber, my test install is in the last stages, lets see [13:25] ppisati: yep [13:26] ok, crashed here too ... apport running [13:26] (i ran with ubiquity -d, lets see if that shows the issue) [13:27] ogra_: check /var/log/syslog for a stacktrace containing "oem-config/enable doesn't exist" [13:27] ogra_: Colin added some debugging in ubiquity with the latest release to try and catch that bug (or another bug in the same area), so hopefully you'll have something in the log that'll help debugging [13:29] doh... [13:29] it did work here... [13:30] sigh [13:30] i have a totally different error [13:30] thats absolutely not helpful :( [13:30] (massive amounts of ext2 errors in syslog) [13:33] fun [13:35] yeah [13:35] i still dont get why the images default to use an ext2 partition for /boot ... we dont make any use of it === rsalveti` is now known as rsalveti [13:35] rsalveti, hey [13:36] ogra_: hey! [13:37] rsalveti, do you have any ETA for new dkms drivers for omap4 that work with TILT ? [13:40] ogra_: ti pushed a new version at the ppa, I'll be reviewing it later today to see if I can update the driver available at quantal [13:40] but it should work with tilt [13:41] awesome, thanks ... since we have full jockey support all users get offered the installation immediately after first boot [13:41] yeah, makes sense [13:41] so having a fixed package in the archive would help a lot :) [13:41] the newer driver will probably require a few patches on some other packages, but I'll review those as well [13:42] ogra_: ppisati: regarding the smsc95xx messages, the workaround was part of jasper [13:42] and if we want to install the rootfs at a usb driver, we need it around from first boot [13:43] otherwise you'll get tons of messages and it'll make the system slow [13:44] oh, i remember, that was a sysctl workaround [13:56] rsalveti: do you carry it in linaro? the workaround i mean [13:57] ppisati: yup === TheFinder is now known as Neko [17:52] marvin24, ogra_ the free nvflash replacement does only work with unlocked devices of course. Useful since it is easier to install than nvflash if it's packaged but does not do anything which already can't be done [17:53] such as flash locked down hardware :) [17:53] marvin24, if X fails to start with 'TEGRA: Failed to initialize display controller library' where would you look for a fix? [17:53] 2.6.39 kernel [17:53] 2.6.36 works with the same userland/X [17:54] janimo: urg [17:54] update your kernel ... [17:54] to 3.1 ? [17:54] or 3.0 [17:54] transformer not ac100 [17:54] works with 26.36 [17:54] ah, you need the right display drivers for your kernel [17:54] every kernel has a different one [17:54] just thought you know since it is tegra2 stuff [17:54] we just skipped this version ;-) [17:55] heh, nvidia dev site is closed due to attacks at the moment [17:55] but I'll try getting another L4T if that should solve it [17:55] IIRc L4T had one for 2.6.38 [17:55] yes, someone stole 450000 yahoo acounts and tries every site now with the same passports [17:56] I don't know if NV every released L4T for a .39 kernel [17:56] and this is a reason for closing down a site? I am glad google did not follow nvidia devzone's lead :) [17:56] not yet ... [17:56] they are very nervous [17:56] they have kernels for each L4T tagged appropriately but the reverse is not true [18:15] ogra_, did you do some more detective work on the ac100 crash with large initrd? [19:19] janimo: fastboot can only handle 2 MB initrds [19:19] hard coded [19:19] which was more than enough, android uses 150 k ! [19:20] marvin24, is ubuntu already using best compression method? It's lzma right? [19:20] no, I always used xz (which is not so good) [19:20] we could try with lzma though [19:23] I think they are comparable [19:23] we need to get the ubuntu initrd size down for ac100 then I guess [19:24] yeah, nearly equal [19:24] but my testdata is a bit bad for this