**** BEGIN LOGGING AT Wed Jul 11 02:59:58 2012 Jul 11 08:45:45 .c Jul 11 15:33:53 ogra_, marvin24 did the 3.1 kernel issue have something in common with https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/936667 ? Jul 11 15:33:55 Ubuntu bug 936667 in upstart "Upstart early job logging causes boot failure for systems with no initramfs (error is "No available ptys")" [High,Fix released] Jul 11 15:34:17 janimo, well, we always have an initrd Jul 11 15:34:25 any news on that particular nvidia kernel issue? Jul 11 15:34:33 ogra_, is that still holding up quantal ac100 images? Jul 11 15:35:18 and that bug is fixed since mid-precise Jul 11 15:35:55 no, what was holding up was my time to fix ac100-tarball-installer and additionally research the oops i get on boot Jul 11 15:36:05 the installer is fixed now Jul 11 15:36:36 tomorrow if i have new images i'll move on with the kernel issue Jul 11 15:36:37 ogra_, kernel oops on install or first boot? Jul 11 15:36:41 ok Jul 11 15:36:47 sadly ui cant quickly roll images manually atm Jul 11 15:36:57 we only have one builder thats constantly occupied Jul 11 15:37:06 is that a regression? Jul 11 15:37:12 janimo, first boot, some initrd issue i think Jul 11 15:37:29 we dropped all non panda HW from the datacenter Jul 11 15:37:31 I thought the builder situations was only going to get better Jul 11 15:37:45 ah, a good move otherwise Jul 11 15:37:50 and not enough new pandas? Jul 11 15:37:52 and until tere is a new mandabox server we only have one machine for livefs builds Jul 11 15:38:28 i think we have a lot of new pandas, but someone needs to set them up Jul 11 15:51:11 ogra_: build iamge at home? Jul 11 15:51:56 no, thanks, i have better stuff to do than to set up the ubuntu build infrastructure at home Jul 11 16:22:37 * GrueMaster has the infrastructure, but no loinger has the time or desire. Jul 11 16:28:53 the fact our build insfrastructure is so involved as opposed to dead-simple live-build means noone can really take over properly without too much effort Jul 11 16:33:04 janimo, ogra_: I hope the console can be fixed soon Jul 11 16:33:29 otherwise we need to boot with console=tty1 as a workaround Jul 11 16:34:14 janimo: did you tried to compile a 3.5 kernel for tegra already? Jul 11 16:37:26 marvin24, no, since Iheard from you it has no graphics :) Jul 11 16:39:15 it has, but a bit buggy Jul 11 16:40:06 maybe still makes no sense to try it Jul 11 16:40:35 I may give arm-soc a try - it has more tegra stuff than 3.5 Jul 11 16:41:03 I just saw too many kernel tree lately that don't even build, so I am not too attracted to yet another one Jul 11 16:41:04 :) Jul 11 16:42:02 janimo: yes, me too :-) Jul 11 16:47:55 lilstevie, do you know anything about the status of the 3.1 jhinta kernel for tf101? It sort of boots here, but hangs on starting network manager Jul 11 16:48:03 the xda forums are horrendous Jul 11 16:48:59 I wonder if tf101 also has the console problem Jul 11 16:50:51 marvin24, I had to boot with --no-log Jul 11 16:51:15 marvin24, , otherwise not sure how the console problem manifests itself, I did not boot ac100 with 3.1 yet Jul 11 16:51:50 it just prints out the kernel logs, then starting init and a few scripts, pausing on NM Jul 11 16:51:51 yes, it oopes shortly after entering userspace boot Jul 11 16:52:04 but looks random Jul 11 16:52:19 no oops on screen and no ramconsole built in this kernel Jul 11 16:52:40 is plymouth installed? Jul 11 16:53:32 marvin24, yes. Should I make it -x ? Jul 11 16:53:42 meanwhile I'll rebuild the kernel with ramconsole on Jul 11 16:54:28 no, it hangs here only with plymouth installed Jul 11 16:55:16 but uninstall is also only a workaround Jul 11 16:55:26 marvin24, iirc ogra said making lib/plymouthd or some unexecutable is a workaround Jul 11 16:55:38 yes, or add console=tty1 Jul 11 16:55:48 and leave it executable Jul 11 16:56:05 marvin24, do you know if tegra2 kernels can support multiple hw from the same build? Ie. ventana, harmony, ac100, tf101 all built in? Jul 11 16:56:14 marvin24, thanks, I'll try that too Jul 11 16:56:28 will that redirect console output to somewhere invisible? Jul 11 16:56:32 the kernel does, but the bootloader doesn't Jul 11 16:56:42 janimo: yes Jul 11 16:57:05 in mainline, everything will depend on device tree Jul 11 16:57:17 so you can boot the same kernel on all tegras Jul 11 16:57:25 but we ee Jul 11 16:57:35 ... need uboot for this Jul 11 16:58:01 uboot passes device tree to the kernel (same kernel for all) Jul 11 16:58:14 I don't know how other archs are doing it now Jul 11 16:58:39 marvin24, in 3.1 there's no device tree yet afaik Jul 11 16:58:47 yep, only in mainline Jul 11 16:58:59 * janimo wonders if a single zImage could be used on both ac100 and tf101 Jul 11 16:59:17 sure, add tf101 to mainline Jul 11 16:59:25 or add a device-tree Jul 11 17:00:06 as I said, *all* tegras (tegra2 and tegra3) should be bootable from a single kernel Jul 11 17:00:06 so with current 2.6.x/3.0 non-DT, what is uboot specific in the kernel? I see nothing in what I build for in the confis Jul 11 17:00:09 configs Jul 11 17:00:28 marvin24, yes in mainline, I was wondering pre-DT where we are and will likely still be in 12.10 Jul 11 17:00:42 the current bootloader (fastboot) does not support device trees Jul 11 17:00:47 while u-boot does Jul 11 17:01:02 so, mainline + u-boot or fastboot + 3.1 Jul 11 17:01:18 (uboot can also boot 3.1 kernels) Jul 11 17:01:31 but hw support is not finished Jul 11 17:02:04 we have to wait - again :-( Jul 11 17:02:25 but regarding DT support, tegra is pretty complete Jul 11 17:09:48 marvin24, chmod -x /sbin/plymouthd made it boot into X. will try console= too thanks! Jul 11 17:10:29 janimo: good Jul 11 17:10:42 marvin24, do you think it is a nvidia kernel bug? Jul 11 17:11:08 yes, I talked to an engineer from nvidia Jul 11 17:11:22 and he was able to reproduce it on ventana (tegra2 dev board) Jul 11 17:11:33 could also be ubuntu related ... Jul 11 17:11:46 do they plan to address it in 3.1? Jul 11 17:11:50 but the kernel shouldn't oops Jul 11 17:11:55 janimo: yes Jul 11 17:11:58 possibly, ubuntu userspace is a bit different Jul 11 17:12:24 we tried to track it down (and still try) Jul 11 17:13:53 janimo: I will request a new kernel build when it's fixed ;-) Jul 11 17:14:03 marvin24, sure ;) Jul 11 17:14:42 marvin24, back to the previous issue - the idea of single image for multiple kernels is _dependent_ on device tree? Since in menuconfig you can check any number of arches to support Jul 11 17:15:14 with "arches", do you mean omap, tegra, ... Jul 11 17:15:17 or boards? Jul 11 17:16:01 you can select different boards, but our bootloader is buggy and does not accept multi-board kernels Jul 11 17:17:27 janimo: no idea what tf101 bootloader does Jul 11 17:17:55 marvin24, afacit just takes an android bootimg and starts it :) Jul 11 17:18:05 like the ac100 one, not sure about details Jul 11 17:18:41 marvin24, ah so the ac100 bootloader would choke on a kernel image that has any other board than paz00 enabled too? Jul 11 17:18:55 I meant just tegra2 based ones (paz00 and tf101 in particular) Jul 11 17:22:06 in fact, no one knows what it does Jul 11 17:22:47 s*y Jul 11 17:31:45 I had the impression, that the bootloader reads the machine-id from the kernel and feeds it back to the kernel Jul 11 19:12:36 prpplague: howdy pardner Jul 11 19:12:45 greetings Jul 11 19:13:09 prpplague: found time to look at the video i told you? Jul 11 19:13:35 LetoThe2nd: sorry no, yesterday was crazy i had some car troubles from the previous afternoon Jul 11 19:13:49 LetoThe2nd: can you give me the link again? i have it at home Jul 11 19:13:55 prpplague: ah kay... hope you got things roted out? Jul 11 19:14:02 LetoThe2nd: yea Jul 11 19:14:14 sorted out, even :/ Jul 11 19:14:28 LetoThe2nd: i had the car worked on friday, turns out they put one of the nuts back on without the cotter pin Jul 11 19:14:43 aw :( Jul 11 19:15:32 LetoThe2nd: damn near got killed by a large truck on monday afternoon after the front passenger tie rod came off while i was cross an highway intersection Jul 11 19:15:47 prpplague: http://www.youtube.com/watch?v=ufBCoHP-8Yk&list=UUXHs7mzjNipr-xp1Bps2zew&index=2&feature=plcp Jul 11 19:16:02 * prpplague looks Jul 11 19:16:08 LetoThe2nd: you said about 50 seconds in? Jul 11 19:16:11 prpplague: interesting part is from 0:55 onwards or so. Jul 11 19:16:24 prpplague: glad nothing bad happened then Jul 11 19:16:50 LetoThe2nd: this video is work safe correct? Jul 11 19:16:58 prpplague: totally yes. Jul 11 19:17:16 prpplague: its from an educational fair near me. Jul 11 19:18:13 LetoThe2nd: ahh dandy! Jul 11 19:18:27 prpplague: and inside that rc car is a panda :) Jul 11 19:18:48 LetoThe2nd: yea i think those guys were in cambridge a while back at ELC Jul 11 19:18:56 LetoThe2nd: with an early prototype of the design Jul 11 19:20:06 prpplague: will meet those guys sometime in the next two weeks probably, was really surprised there are other panda users in this small town. Jul 11 19:22:08 LetoThe2nd: hehe Jul 11 19:22:15 prpplague: btw got one or two minutes for private? Jul 11 19:22:27 (just noticed we're not even in #pandaboard) Jul 11 19:22:28 * prpplague checks the time Jul 11 19:22:42 LetoThe2nd: yea, got about 5-10 mintues Jul 11 21:04:35 does anyone know where i can find the status of the SGX drivers on a BB XM running 12.04? Jul 11 21:04:50 and audio also Jul 11 22:32:16 smplman: check the linaro website Jul 11 22:32:59 smplman: all the latest linaro releases are based on 12.04 with improved kernels and drivers where available Jul 11 22:34:21 smplman: sorry i think i gave you bad advice.. they only support the pandaboard nowdays Jul 12 00:28:19 just fyi for anyone interested, the call for participation for ELC-E end august 1 - https://events.linuxfoundation.org/events/embedded-linux-conference-europe/cfp Jul 12 01:52:34 janimo: well, personally I do not trust his stuff as far as I can throw him, he has been a massive pain with things like not being even able to identify the root device, and he doesn't really know what he is doing **** ENDING LOGGING AT Thu Jul 12 02:59:58 2012