=== kentb is now known as kentb-out === maxb_ is now known as maxb [07:44] we have problems getting http://www.kernel.org/pub/linux/kernel/people/mcgrof/firmware/ath6k.tar.gz :) ... wonder if we have that in any ubuntu firmware package in archive? [07:51] found it in linux-firmware (obviously) ... thanks === lool- is now known as lool === chrisccoulson_ is now known as chrisccoulson [09:46] Trying 2a01:450:10:1::10... [10:04] hi, since 2 weeks no updates to mainline kernel. why cant somebody change the git repo? [10:08] Kano, hadn't considered it, we have enough problems due to that breakin as it is [10:08] couldnt be a huge change in your script i think [10:09] also i forgot how to update the kernel config with your package, something with update-config or so? [10:09] probabally not even a change there, probabally only need to change the remote url [10:09] fakeroot debian/rules updateconfigs [10:09] ah [10:10] i did a grep in the rules and did not find it [10:10] also i had to remove the dpkg >... depend, why is that there? [10:11] hmm, maybe fix your target [10:11] debian/scripts/misc/kernelconfig: line 130: /root/kernel/linux-3.0.0/debian/scripts/misc/splitconfig.pl: Permission denied [10:11] in your release package [10:12] thats a 'problem' introduced by the orig+patch format for debian [10:12] those are not really intended for use against the package but in the git repo [10:12] where the permissions are honoured [10:12] chmod 755 there [10:12] same for config-check [10:13] yep, you simply cannot represent that in the package [10:13] but you can add that in the rules [10:13] patch does not carry executable information at all, and can not do so [10:13] thats not impossible [10:14] yeah, but we never use it that way ever, its not our process model, so we'd never notice [10:14] the normal debian way is not chmod, but to change it to like $(PERL) script or $(SHELL) script [10:15] well then use that [10:15] or i could spend my time fixing cves [10:15] its a priority thing, its not a priority as noone working they way we intend would ever hit it [10:15] and they will hit the cves [10:16] but we'd happily accept a patch if someone figured out all the places it is needed [10:16] what was the reason to add dpkg >..? [10:17] not mentioned in changelog [10:17] in which release [10:18] oneric [10:18] UBUNTU: [Config] Make linux-libc-dev coinstallable under multiarch [10:19] ok, i dont build that, can ignore that [10:20] it build still too much, like tools/source/doc with emtpy packages [10:20] yep, we only turn those off for debugging builds, and its simpler to handle the empty packages than add a bunch of handling to the control file generation to turn them off [10:21] i would prefer when they would not build at all when i disable em [10:21] i dont see those in mainline too [10:22] right in mainline i use the uber-complex 'rm' command to remove them before syncing them out [10:23] well if you have got free time maybe consider em not to build [10:23] i have considered it, i have even got patches to do it, and it makes the rules vile and complex [10:23] for something we never use [10:24] and afterall they are our packages to build our kernel not all things to all men [10:24] well they would be more usefull [10:25] i always try to write scripts as generic as possible [10:25] generic for whom? [10:26] so do we but at some point one has to trade genericity for maintainability [10:26] we have to maintain them for years against old kernels and that is not free [10:27] well i just want to script all changes i need against your packages [10:28] so what you are saying is, its too much effort for you to fix the packaging, but you think we should [10:28] well you know the packageing [10:29] indeed, so which is more important for your kernel [10:29] having pretty packaging that allows a derivative to build easily, or having CVEs fixed in your kernle [10:29] as right now, that is what you are stopping me doing [10:29] well i mainly use that for live mode, i need mainline builds... [10:29] for the hd installs [10:31] i write a tiny script for every kernel update, but there is no update [10:31] daily is of course always the same script [10:32] yep the entire kernel.org infrastucture being compromised has been a huge problem [10:32] not least because it has _prioritised_ security fixes above _ALL_ELSE_ [10:32] now i wonder why that would be [10:35] you mean for the older kernels? [10:36] the latest is on github, that should be no problem to build rc6/daily [10:43] yep, indeed and i'll put it on my todo list [10:57] is armel somehow differently handled? not that i would use it but [10:59] http://paste.debian.net/129942/ [11:01] no its not handled differently [11:02] why is there no 1000 hz for it then [11:03] perhaps because it can't do 1000Hz ? [11:03] maybe [11:03] 1000Hz for arm is a bit excessive isn't it? [11:03] 1000HZ == 10% powerformance cost too [11:09] what bootloader is used on arm? [11:12] depends on the board they are all different pretty much [12:16] the wireless-cdra package is weird, it has got no depends but [12:16] crda: error while loading shared libraries: libnl-genl.so.3: cannot open shared object file: No such file or directory [12:16] sounds like a bug to me [12:19] Hello [12:20] I am developing a kernel module under Ubuntu, sometimes this module causes the PC to freeze, is there a way to get dmesg (or whatever debug info) output over the network ? [12:20] AnAnt, netconsole might do it, depending how hard it freezes [12:22] apw: how about serial port (note, the PC does not have a serial port, but I can use a USB/serial converter) [12:22] usb serial can work sometimes, again depending on how hard the kernel is crashing [12:58] apw, need to bounce gomeisa for kernel update [13:00] tgardner, ACK now is as good a time as any [13:02] apw: thanks [13:02] tgardner, let me know when its done i have some builds pending [13:02] apw, its bouncing, should be back up in 90 seconds or so [13:03] apw: how to do it on serial ? [13:03] AnAnt, usb serial? [13:03] apw: yes [13:03] just add console=ttyUSBN I think [13:04] ok, thanks [13:05] apw: so I can only set the serial console during boot? [13:05] i believe so ... never heard of any way to add them [13:07] apw: ok, thanks a lot [13:30] * ogasawara back in 20 [13:33] apw, git://kernel.ubuntu.com/cking/pmdebug.git --> systemtap/s3.stp led_flash() [13:41] herton: lucid was respinned, and the reverted patchset was only the i915, right? i'll invalidate all the derivative arm lucid bug trackers [13:42] ppisati: yep [14:10] $ ps -ef | grep gcc | wc -l [14:10] 282 [14:10] $ ps -ef | grep cc1 | wc -l [14:10] 41 [14:29] so is there a non-kernel.org mirror of Linus' tree right now? [14:29] hallyn, yes [14:29] Need to base some patches on it to send upstream... [14:30] git://github.com/torvalds/linux [14:30] apw: great, thanks [14:49] ubuntu-oneiric$ git push origin :Ubuntu-P-sync [14:49] remote: *** Deleting a tag is not allowed in this repository [14:49] remote: error: hook declined to update refs/tags/Ubuntu-P-sync [14:49] To git+ssh://zinc.canonical.com/srv/kernel.ubuntu.com/git/ubuntu/ubuntu-oneiric.git [14:49] ! [remote rejected] Ubuntu-P-sync (hook declined) [14:49] ogasawara: Error: I am only a bot, please don't think I'm intelligent :) [14:49] error: failed to push some refs to 'git+ssh://zinc.canonical.com/srv/kernel.ubuntu.com/git/ubuntu/ubuntu-oneiric.git' [14:50] apw: ^^ I assume that's due to the new restrictions in our repo? [14:53] ogasawara, yeah, hmmm, thats neither a signed tag, nor is updating them allowed ... hmmm ... push it as a branch for now perhaps [14:53] apw: ack [14:54] we can likely just add an exception for that tag namespace [14:56] :q [14:57] * apw punches unity in the face repeatedly ... _do_not_paste_my_output_in_the_wrong_window_that_might_be_my_password_ [14:58] <_ruben> :q is your password? good to know :) [14:58] heh [14:58] heh luckily not, but it might be safest if it was [14:58] <_ruben> ;) [14:58] "smb, what time is it there?" would make a good password [15:09] hm, laptop isn't even that hot, but just shut itself down... [15:09] nothing in syslog [15:18] hallyn, how quickly [15:19] and its not suspended is it? [15:19] no, it just acted as though i'd hit the power button [15:19] we seem to have a new suspend on idle on AC [15:19] it also wasn't idle, i was in the middle of typing :) [15:19] so ... like you pressed it (full shutdown messages) or like you held it [15:19] like i held it [15:19] no shutdown msg [15:20] nothing in syslog at all for several mins before booting up [15:20] hallyn, running what release, which kernel [15:20] oneiric, uptodate [15:20] not see that, no [15:20] thermal0/temp is showing 84000 - whatever that means [15:20] could be anything, likely as not 84c [15:21] but hte units are not easy to determine at times [15:21] I've had this happen once or twice before, but when it was much hotter, and I do believe that file was around 100 when i rebooted [15:21] 105 is a common 'i am turning of NOW' value [15:21] i think i once deduced it actually did correspond to degrees celsius, don't recall how/why [15:21] indeed they are often listed in /proc/ [15:22] cking, where is the thermal trip stuff hidden in proc ? [15:22] apw - i'm on a call at the mo [15:22] serge@sergelap:/sys/devices/virtual/thermal/thermal_zone0$ cat trip_point_0_temp [15:22] 98000 [15:22] ahh there you go then, but if you are at 84c that in of itself is worrying [15:22] are you fanning? [15:22] I'm down to 70 now [15:23] yeah, fan is always goin gon this thing [15:23] thinkpad ? [15:23] vaio f11 [15:23] what is your machine doing that 70c is a sane temp ? [15:23] I had shut down u1 some time before, so i was just doing a libvirt build, running kvm, and playing a song... [15:23] it'd want to be compileing on something [15:24] so you are kicking the crap out of it, so perhaps 70 is expectable [15:24] well, it's an 8-core 8G laptop, I don't take it easy on it [15:24] no serioualy this thing easily hits 90 [15:24] but i dont' think it had hit 98 [15:24] that sounds like pretty poor thermal control. an 8 core thingy in a laptop, mental [15:24] all right, i guess i'll keep a little display up with the temp values [15:24] yeah good plan try and keep some history [15:25] heh, i had to quickly choose something when i joined canonical. i'd have preferred a good desktop and a netbook. but this thing mostly serves me well [15:25] all right, i'll ping back if it does it again without hitting 98 [15:25] thanks [15:29] hallyn, made the same mistake myself [15:30] I guess in about 1.5 years I'll start looking for a kick-ass desktop :) [16:10] (this is one of those times i'm happy to be using wmii so i can just change my status shellscript :) [16:10] hanging steady at 67 right now [16:28] apw, dpkg-buildpackage is just a script. AFAICT it simply adds -jX to MAKEFLAGS before running debian/rules. [16:28] hmmmm [16:28] I haven't quite been able to simulate it by hand yet [16:50] ppisati, do you want the ti-omap4 kernels uplaoded with the DMA/SMP performance fix ? [16:53] apw, ogasawara: pushed 'UBUNTU: [Config] Fix binary-% build target' to oneiric master-next [16:53] tgardner: ack [16:54] tgardner, nice [16:54] I'll figure out the AMKEFLAGS business and build it into debian/rules somehow [16:54] MAKEFLAGS* [17:12] apw: say do you know of a way to set the trip_point? file isn't writeable... and a simple sbuild of libvirt-bin quickly brought temps to 92 (trip point being 98) [17:13] hallyn, normally they are bios set and bios enforced, they are exposed to allow the OS to help in heat managment i expect [17:13] hallyn, trip points are typically the domain of BIOS/SMI AFAIK [17:13] drat [17:13] i really don't think you should be hitting anywhere near it if the bios was managing fans correctly [17:13] at least 94 seems to be another, 'passive' trip point. I guess I should go read up. must be some discussion online [17:13] it must burn your hands if you go anywhere near the case [17:14] it gets warm, but not that bad. [17:14] maybe one of those fancy powered laptop stands would help :) [17:15] i am suspicious that there are a number of bioss out there which do not handle the fans, but they ship windows drivers to override the fans to make things work [17:15] oh! actually, i guess sony had a bios update about this. [17:15] but the damned thing isn't usable in linux or even in windows 2008 [17:16] now, this time after it hit 94 it started going down. so maybe the 'passive' trip poitn didn't trip last time, and so it kept climbing? [17:17] maybe i should drop by the sony store [17:18] hallyn, maybe you should be using tangerine [17:18] tgardner: i was just compiling libvirt... [17:19] that would feel like abusing resources [17:19] hallyn, well, it beats having your laptop sizzle [17:19] true [17:19] hallyn, i use http://pastebin.ubuntu.com/689370/ to jam the fans on for my thinkpad, don't know what you'd need to change for it to work for you [17:20] bjf: yeah, i might have to bypass the kernel's trip point and have my own little daemon bump up the fan. good idea. [17:20] its gotta have support in the platform driver [17:23] yeah and i don't see a fan control under /sys/module/sony_laptop [17:36] * tgardner --> lunch [17:55] bjf, did you get the magic for bootchart ? as i have just seen one with the lines which was generated this week [17:55] apw, yes been working on it all a.m. :-) [17:59] bjf, can you make sure you also turn on initcall_debug on the test boots and grab the dmesg as well [18:00] apw, ack [18:04] rebooting [18:06] ogasawara: hi! bug 834725 → looks like this is not a bug in linux after all, right? [18:06] Launchpad bug 834725 in linux "Powertop fails to report number of wakeups/events" [Undecided,Confirmed] https://launchpad.net/bugs/834725 [18:06] htorque: was just about to comment there. indeed, I do not believe this to be a kernel issue. so I was going to close out the linux task. [18:07] good, thanks :) [18:37] * jjohansen -> lunch [19:25] smb, do you happen to be around ? [19:25] http://groups.google.com/group/ec2ubuntu/browse_frm/thread/57ff20c6370f7bb9 [19:26] we should have someone look at that. [20:30] * ogasawara bails a little early for an appointment, back on later. === jjohansen is now known as jj-afk [21:26] back on later === kentb is now known as kentb-afk [22:54] well that sucks