[00:17] infinity: OK. I have an initial eglibc2.16 aarch64 support patch. What do you want me to do with it? File a bug and attach? [00:17] GrueMaster: Replaced by openjdk-7-jre-headless and python3-software-properties, I'd assume. [00:17] btw it still requires gcc-4.6 - is that right? [00:17] that's a problem for aarch64 cross-build - as there is no such thing [00:17] wookey: Hang on to it until we chat tomorrow? :) [00:17] OK [00:18] wookey: And we'll figure out how to make it work with 4.7, or die trying. [00:18] heh [00:18] infinity: Thanks. Guess I should fix my preseed. === heathkid|2 is now known as heathkid [01:43] http://paritynews.com/software/item/394-linux-37-kernel-to-support-multiple-arm-platforms === dmart_ is now known as dmart === doko_ is now known as doko [11:52] infinity, i'm just uploading tegra and tegra3 drivers with the broken libs, please have a look (i added a changelog note about the libs and documented a manual workaround in the bug) [12:37] Hi [12:38] I have an ARM based hardware, with CPU ARM 926 EJS @216Mhz [12:38] Flash is 32Mbyte NOR [12:38] DRAM is 256Mbyte DDR2 [12:39] Is it possible to use ubuntu arm on this platform? [12:48] ogra_: rsalveti: ndec: I can see that trunk PPA is almost full: 2.7 GiB (89.99%) of 3.0 GiB. Is it possible to increase the size to prevent problems? [12:48] oops... [12:48] iirc you can ask in #launchpad for that, yeah [12:49] stulluk, no, the ARM-926 is ARMv5T from the small bit of googling I just did. support was dropped for that quite a while ago. you might have some luck with debian though [12:53] ogra_: who shall I ask to on #launchpad? [12:58] XavB, hmm, no idea, i would just assk into the room what it takes to raise the limit [13:26] how do I go by making the linux kernel boot in a verbose mode? [13:26] do I have to recompile it with some setting, or can I add some arguments to e.g. the bootloader? [13:27] put earlyprintk on the kernel cmdline and drop quiet [13:28] I see, thanks [13:29] Do I have to have the earlyprintk print to another UART, or can I point both the console and it to the same? [13:30] it will always print to the console you have set [13:30] if oyu have more than one console= option, the first one is for kernel, the second is for userspace [13:31] ah right think I understand [13:31] if you have only one it will be used for both [13:31] I was looking at some u-boot examples here and it specifiesa another uart every time [13:31] a separate one from the console [13:32] if you have none, the default will be used (usually console=tty1) [13:32] right ok [13:40] bah it was already at max verbosity I think [13:40] :[ [14:03] ogra_: heya, around? [14:05] plars, yes, but in meetings for eth rest of the day (atrting in 5 min) [14:05] *starting [14:05] whats up ? [14:06] ogra_: just wanted to catch up on a couple of things I'm seeing on panda, ping me if you have one finish early and have a few minutes? [15:03] plars, shoot [15:04] ogra_: couple of minor but annoying oddities that I'm curious if they're specific to something I have, or if you see them too [15:04] ogra_: 1. on desktop image, when I first boot, sometimes the keyboard doesn't work unless I disconnect/reconnect [15:05] reconnect the keyboard thatis [15:05] hmm, havent had that here ... sounds like a race in udev [15:05] everything being on USB surely causes some high traffic [15:05] not sure if it's specific to this keyboard, I don't have another I can easily use at the moment but if it's something others are seeing, then maybe it's not just this keyboar [15:06] (disc, NIC, input devices) [15:06] yeah [15:06] i havent heard from others about this yet [15:06] file a bug though, i think its bugworthy [15:06] ogra_: suggestions on where it would go though? udev? [15:07] or kernel [15:07] ok [15:07] start with udev though [15:07] sounds good [15:07] 2. audio [15:07] ppisati is busy enough :) [15:07] I get popping noises sometimes on jack audio, but not over hdmi [15:07] hedphones should work OOTB [15:08] ogra_: right, that's the default [15:08] ah, hdmi works (i have no monitor with speakers) [15:08] I do :) [15:08] great [15:08] it doesn't work automatically, you have to switch it by hand [15:08] well, popping means that the amp isnt murted during power up/down [15:08] yeah, it would work if we had pulseaudio profiles [15:09] iirc we have support for them but nobody added any [15:09] ok, so that's pulseaudio related you think? [15:09] no [15:09] the popping is the driver [15:09] no auto switching to hdmi is pulse [15:10] ok [15:10] might probably make a good wishlist bug [15:12] yeah [15:13] ogra_: those were the main things at the moment, I saw that the usb keyboard issue on server installs was fixed, so I'll retry that today too and actually try to do the server on the console rather than over serial [15:13] thanks ogra_ [15:14] plars, i'm not sure ppisati fixed it on panda, i know he looked into begale for the kbd stuff [15:15] ogra_: ok, all the more reason to check it today :) [15:16] :) [15:18] the usb problem was just panda related [15:18] ah [15:18] ATM omap3 is really screwed (at least on my hw) [15:18] i thought you looked into beagle [15:18] i'm looking into beagle now [15:18] ppisati, well, its unsupported anyway [15:18] ogra_: beagle? [15:18] yes [15:19] ah [15:19] we only carry it through because its "for free" [15:19] upstream + one patch and usb is quite ok [15:19] fixing bugs there is best effort [15:19] with our kernel, for some obscrure config screw, the usb hub is ENTIRELY dead [15:19] but not mandatory [15:19] wow [15:20] i;ve been bisecting the hell out of my BEEP for the last 3 days [15:20] still didn't found what's so different from our kernel and upstream [15:47] ogra_, ppisati: just hit a problem with pvr-omap when running dist-upgrade after installing todays -desktop image [15:47] tell me [15:51] waiting on the update to finish [15:56] plars: i'm not pvr-omap maintainer [15:56] and rsalveti is traveling [15:56] (and busy in other projects) [15:57] ppisati: it was the kernel update that was happening when I hit the dkms issue [15:57] oh, dkms [15:57] run-parts: executing /etc/kernel/header_postinst.d/dkms 3.5.0-17-omap /boot/vmlinuz-3.5.0-17-omap [15:58] Error! Bad return status for module build on kernel: 3.5.0-17-omap (armv7l) [15:58] hmpf [15:59] plars, i think there should be a dkms build log somewhere [15:59] ogra_: yes, looking at it now [16:00] looks like it was missing omap_priv.h? [16:00] Also there's an error... [16:00] so kernel header issue .... [16:00] #error "A preemptible Linux kernel is required when using workqueues" [16:00] hmm, priv was gone i thought [16:00] our kernel is preemt, no ? [16:01] *preempt [16:01] yes, it would seem to be [16:01] weird [16:01] dkms is such black magic [16:01] just checked that CONFIG_PREEMPT=y in config [16:02] snap, I forgot to sacrifice the chicken and pour the blood on my laptop before compiling [16:02] :) [16:02] heh [16:04] Sigh, plars, how many times must I show you how to incite the satanic rituals of dkms installs? :P [16:07] iirc, the chicken is used to dust off the monitor (wax on, wax off), while dancing to Gangnam Style. [16:07] nekkid [16:08] With two strips of bacon on each shoulder. [16:08] * plars makes a note to never accept a google hangout invite from GrueMaster or ogra_ [16:08] lol [16:08] lol [16:11] plars: can you paste the logs? [16:11] I know there might be an issue when installing the headers package [16:11] rsalveti: yes, working on it, but desktop on panda is crazy-slow [16:11] which can accidentally pull the omap 3 headers as well [16:11] my browser finally came up [16:12] but my mouse isn't showing up at the moment... [16:12] oh [16:12] that's weird [16:12] plars: the headers/links are kind of messy at this point [16:12] for the dkms packages [16:13] yeah, I don't think it should be this bad [16:13] there's a fair chance it's hung, but I'll give it a minute. ctrl-alt-f1 isn't getting me to a console [16:13] rsalveti, not only for dkms [16:14] kernel headers in arm are a massive mess in general [16:14] dues to no subarch support in dpkg and friends [16:27] exactly [16:27] afaik infinity is working on that [16:32] finally [16:32] bug #1062407 [16:32] Launchpad bug 1062407 in pvr-omap4 "pvr-omap4 1.9.0.5.1.1-0ubuntu4: pvr-omap4 kernel module failed to build" [Undecided,New] https://launchpad.net/bugs/1062407 [16:32] you need faster usb keys [16:32] ogra_, rsalveti ^ [16:33] * ogra_ really doesnt get that preempt stuff [16:34] ogra_: yeah, I do. It's such a gamble buying these things though. I bought 2 of these pny sticks that I'm using now because 2 other people I talked to said they worked great. They are barely usable for arm rootfs, and they are useless for other installs from usb because they can't be booted from [16:34] supertalen ones are pretty great [16:34] but also pretty expensive [16:34] ther have a real SSD attached in the backend [16:34] *supertalent [16:35] "for kernel 3.5.0-17-omap" [16:36] yeah [16:36] plars, ^^^ from your log [16:37] ogra_: yes, but the one I'm on is 3.5.0-212-omap4 after the update [16:37] omap vs omap4 [16:37] it uses the totally wrong headers [16:39] ogra_: so this is related to the kernel header mess you previously mentioned [16:39] yes [16:39] /usr/src/linux-headers-3.5.0-17-omap :-) [16:39] yeah [16:39] these shouldnt be installed [16:39] plars: just remove the omap3 headers package [16:40] rsalveti: how did those get introduced though? The install I started from was today's omap4 image [16:40] guess they might be installed as a dependency from dkms [16:40] the headers-generic or such [16:40] either that or there is some seed mess [16:40] * ogra_ takes a look [16:40] could be that they are seeded for build-essential [16:40] linux-headers-generic-pae | linux-headers-686-pae | linux-headers-amd64 | linux-headers-generic | linux-headers, [16:41] ogra@anubis:~/Devel/seeds/ubuntu.quantal$ grep omap * [16:41] desktop: * (linux-headers-omap4) [armhf armel] [16:41] we should just remove them from recommends [16:41] hmm, no [16:41] yeah [16:42] ogra_: plars: bug 960770 [16:42] Launchpad bug 960770 in dkms "Packages requiring dkms at Pandaboard (omap 4) will also pull linux-headers-generic because current dkms dependencies" [Medium,Confirmed] https://launchpad.net/bugs/960770 [16:43] great, duplicate then [16:44] I believe we had a thread at the kernel ml about that [16:44] * rsalveti searching [16:45] ah [16:51] plars: so while annoying, it should not break anything at your side [16:51] rsalveti: yeah, gotcha [17:35] ogra_, rsalveti: either of you been insane enough to try launching libreoffice on panda recently? seems to be crashing on startup for me [17:45] wookey: I might have to postpone that glibc talk with you until next week, my day's getting busier than expected. [18:08] infinity: wookey is gone already anyway [18:08] rsalveti: I figured, since he said he had plans this evening. :) [18:11] infinity: :-) [18:15] plars, argh, no i havent [18:16] plars: That is usually done by QA. Oh, wait.... [18:16] ogra_: trying to see if I can get someone to reproduce it. I tried all of the lo apps and they all crash on startup for me [18:16] infinity, not sure if you saw that in yur backlog: i'm just uploading tegra and tegra3 drivers with the broken libs, please have a look (i added a changelog note about the libs and documented a manual workaround in the bug) [18:17] ogra_: Fun. :/ [18:17] infinity, its identical to tegra2, just different binaries and versions [18:17] ogra_: Do you have a firm promise from srwarren to get it fixed so you can SRU? [18:17] i have a promise from the actual dev :) but no ETA [18:18] Mmkay. [18:22] plars, if its reproducable i guess we should unseed libO again [18:22] damned [18:23] * ogra_ goes back to friday evening stuff [18:56] infinity: main glibc issue http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33763 seems to be fixed today [18:56] gcc.gnu.org bug 33763 in c "[4.6/4.7/4.8 Regression] Bogus inlining failed in call to `xxx': redefined extern inline functions are not considered for inlining" [Normal,New: ] [18:56] but now it's barfing with check_fds.c:64:1: error: 'AT_FDCWD' undeclared [21:54] Hey guys - I'm trying to get QEMU to work in Ubuntu Precise Pangolin and emulate an ARM system running Precise Pangolin or Lucid Lynx. Has anyone here successfully done this? === Guest84314 is now known as NCommander