=== sconklin is now known as sconklin-away === bjf is now known as bjf-afk [01:48] hmm, anyone know why the kernel-ppa stuff hasn't been updated in a while? the only 34-rc2 kernel there says "karmic" specifically, and the same is true of the latest drm-next kernel there. [01:49] Does it matter that it says "karmic", if I want to use it on Lucid? I've seen it matter before, on ARM, at least.... try a karmic kernel with lucid userspace, it doesn't boot. [02:03] I guess "karmic" means it was compiled with karmic's configuration file? [02:06] beats me. I'm about to try the drm-next "karmic" kernel. [02:06] I've also noticed that the kernel-ppa kernels all have the STAGING dir entirely disabled... as well as MMC_RICOH_MMC [02:07] Something about "symbol 'm' not valid for ... " [12:55] How do I build a (fresh git) kernel when I dont have the debian/ dirs installed and get it installed on the system (preferrably via a deb-file) [12:55] Will make-kpkg handle that? [13:09] preferably (if you don't want to use git) by apt-get source linux-image-2.6.x-y-generic which gives you the kernel source including the debian dirs [13:11] smb: I can't use the ubuntu source, as there are none for my target, so I need to go with a 3rd part kernel branch. [13:11] So I have the kernel source and a working config, but I need to wrap this into ubuntu-world [13:13] You could clone the git tree and copy the debian* dirs into your tree [13:13] smb: Thats all that's needed? What happens to my local kernel config? [13:14] nosse1, You would need to see how you get it integrated with debian.master/configs/... [13:17] nosse1, Maybe it does work with make-kpkg but I never use that [13:17] smb, I'm trying to integrate debian* right now. [13:18] smb, What do you do to compile kernels then? [13:19] nosse1, I only do the Ubuntu kernels. No need for me to do anything special. :-P [13:19] But you use make-kpkg to build them, right? [13:20] nosse1, No. Either debuild, or dpkg-buildpackage or fakeroot debian-rules ... when I am in a hurry [13:21] OK, thanks === sconklin-away is now known as sconklin [13:46] smb: did you see my post here about the kernelconfig script not working because DROOT is not defined? [13:47] mozmck, I cannot recall it. Where did you post it? [13:47] here on IRC a couple days ago. [13:47] To fix it here I just added DROOT=debian to debian/debian.env [13:48] this is in the lucid git tree [13:48] IRC is quite prone to loose things. But on Lucid debian.env should be always set [13:48] Are you using a special tree? [13:49] Actually it should be debian.master which it points to [13:49] no [13:50] ~/src/ubuntu-lucid/kernel$ cat debian/debian.env [13:50] DEBIAN=debian.master [13:50] I'm using kernel.ubuntu.com/ubuntu/ubuntu-lucid.git [13:51] yes, but DROOT is used in the kernelconfig script and is not defined anywhere. [13:51] this is both on karmic and in 10.04 beta1 [13:53] and looking at the kernelconfig script, line 29 is bindir="`pwd`/${DROOT}/scripts/misc" [13:53] which script are you talking about? [13:53] so I figured DROOT had to be 'debian' [13:53] ok found it [13:54] apw, Is that meant to be called directly [13:55] mozmck, smb, nope thats not menat to be run directly [13:55] all the instructions I found have you call "kernelconfig editconfig" to edit the configs. this worked on karmic fine but not in lucid [13:55] its used from debian/rules updateconfigs and the like [13:55] so how are you supposed to edit the configs now? [13:55] where are those instructions? they are wrong [13:55] Right, where does it say to call it directly? [13:55] debian/rules editconfigs would make sense [13:55] yep [13:56] and should work for all releases too [13:56] mozmck, To call "debian/rules editconfig" was always the intended way [13:57] http://blog.avirtualhome.com/2009/11/03/how-to-compile-a-kernel-for-ubuntu-karmic/ [13:57] That's linked to from here: https://help.ubuntu.com/community/Kernel/Compile [13:57] yay ... random community randomness [13:57] https://wiki.ubuntu.com/KernelTeam/KernelForIdiots ;-) [13:58] Keybuk, heh yeah i like that one better :) [13:59] I see. It seems quite hard to find good information on this subject! [13:59] i wonder if we should rename that dummies :) [13:59] mozmck, mostly because there is such a lot of bad information out there [13:59] I've seen that link, but it doesn't mention editconfig, and I am a dummy so I need menuconfig :) [14:00] do feel free to help us make the ForIdiots page better ... Keybuk only did the bits he needed :) [14:01] oh, ok! I forget I can change those pages. [14:01] * apw adds a comment to the random blog telling them they are wrong :) [14:01] thanks for the help. [14:01] it's basically where I put notes so I don't have to keep rediscovering them each time [14:03] is there a way to tell it to just edit the config for one flavour or do you just have to step through them all? [14:03] I just tend to edit the config file itself [14:03] then use updateconfigs to fill in the bits I missed [14:03] mozmck, no you have to step through them all with that interface [14:04] ok [14:04] yeah mostly i am in bulk update mode and want to do them all at once, so i tend to do the update by adding things to the toplevel common files and doing an updateconfigs [14:06] I'm working on a custom flavour with rtai patches for running CNC machines with EMC2, so I just add and change settings for that one flavour. [14:10] 'debian/rules editconfig' -> make: *** No rule to make target `editconfig'. Stop. [14:11] 'debian/rules editconfigs' -> dh_testdir: cannot read debian/control: No such file or directory [14:11] mozmck: fakeroot debian/rules clean first [14:11] ah, that works [14:11] thanks [16:02] ## [16:02] ## Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting [16:02] ## agenda: https://wiki.ubuntu.com/KernelTeam/Meeting [16:02] ## === nosse1 is now known as nosse1_noob === sconklin is now known as sconklin-away [16:57] ogasawara: Congrats!!! awesome,incredible,great news :D [16:57] IdleOne: thanks! === sconklin-away is now known as sconklin [17:27] apw: I'm interested in cleaning up the linux-firmware-nonfree package, partially to learn more about packaging [17:27] rtg is gone for a bit, and I'd like my changes to get into lucid [17:27] would you be able to review them? === nosse1_noob is now known as nosse1 [17:53] cnd sure thing [18:00] apw: https://launchpad.net/~leannogasawara/+archive/ppa/+build/1590208/+files/buildlog_ubuntu-lucid-i386.linux_2.6.34-1.1~oga2_FAILEDTOBUILD.txt.gz [18:12] bug 507416 [18:12] Malone bug 507416 in linux-fsl-imx51 "CONFIG_NEON=y causes platform lockups with certain application/platform combinations" [Medium,Fix committed] https://launchpad.net/bugs/507416 [18:14] JFo, That should be ok as far as I know, they just don't bother update their bug status. :/ [18:14] ok :) [18:33] pgraner: JFo: could you ping Jane S and ask her permission to put a Reported-by in the git changelog for LP 551606, 548661? [18:34] (ideally an e-mail address, too, but I understand if she doesn't want it in the kernel changelog) [18:35] crimsun, will see what I can do about that. :) [18:35] crimsun, you're doing an upstream fix for 551606? [18:37] bjf: yes, I'm about to push the button on git commit -e -a -s [18:38] crimsun, ok, you are doing a quirk for her SSID? [18:38] bjf: no, for all models using that quirk; I'll attach the diff [18:39] crimsun, ack [18:39] bjf: also, I believe I have a fix for her hp mute issue, but there are at least two further debugging steps necessary, one of which requires rolling a new alsa-driver snap (I'll attach a diff to 548661 if the first option is insufficient) [18:41] crimsun, when does takashi "release" new alsa-driver versions? is there a schedule or just when he feels things are "right" [18:42] bjf: he doesn't generally roll those; Jaroslav tends to push the buttons [18:42] crimsun, ah, ok [18:43] apw: while I work with #launchpad to figure out the merge request, feel free to review the last four commits at https://code.launchpad.net/~chasedouglas/lucid/linux-firmware-nonfree [18:51] smb: do you know when the window for 2.6.32.11 closes? I have possibly two more patches to push to stable (both Jane S's bugs), and I'd like to get them in if feasible. [18:52] I think it is tomorrow morning iirc [18:52] but I defer to smb [18:52] crimsun, Not for sure. Normally this amount of activity I would say he might close on Friday [18:52] ah, see I was wrong ;-) [18:52] crimsun, But sometimes he is quicker. [18:52] smb: JFo: ok, thanks [18:53] JFo, ITs hard to say. I have not really found a regular schedule in Greg. [18:53] true, I was thinking of something else entirely I am sure [18:58] apw: ignore my previous bzr branch, I figured out the merge process, so use it instead please [19:02] apw, can you comment on crimsun question ^^ [19:06] (actually I'm awaiting on her ack to submit the first patch :-) [19:06] awaiting her * [19:38] jjohansen: have you had reports of a microtime issue in linux-ec2? [19:38] i couldn't find a relevant bug [19:39] hrmm, not any that I have heard of [19:39] what are you seeing? [19:39] lemme slip it into pastie right quick [19:39] http://pastie.org/895689 [19:41] jjohansen: here's a normal karmic server (bare metal, not in a VM): http://pastie.org/895692 [19:41] which is more of what i'd expect [19:42] if i pop in an old hardy 2.6.24 kernel, all is well [19:43] domU w/2.6.24-23 -> http://pastie.org/895696 [19:44] it's not perfect in 2.6.24, but it's worse in linux-ec2 from lucid [19:44] but it could be a silly kernel config on my part [19:45] hrmm, well that does look odd, I'll have a look and run a few tests [19:46] i can set you up on one of our environments if that'd help [19:48] rackerhacker: sure, that would be worth snooping at [19:49] gimme a second to get one rolling [20:43] jjohansen: want me to pile the details into a launchpad bug? [20:59] rackerhacker: Just for giggles, could you try booting a machine with clocksource=acpi_pm and run the same test? [20:59] rackerhacker: (A physical machine, that is) [21:00] a physical machine running linux-ec2? [21:00] rackerhacker: Running the same as the machine in http://pastie.org/895689 was running. [21:01] soren: that particular example was within a domU [21:01] rackerhacker: Sorry, wrong pastie. [21:01] rackerhacker: http://pastie.org/895692 this one. [21:02] let me see if i can spare that machine for a moment ;-) [21:26] rackerhacker: hrmm, not yet. I haven't been able to replicate on EC2, I have run through 10 different instance size/kernel combinations and every one was good. I haven't tried your config on EC2 yet though [21:32] jjohansen: you might be right, soren walked over and gave me a run down [21:35] crimsun, thumbs up from Jane [21:35] crimsun, [21:35] That's fine - I have no problem with my name and/or email going in there. [21:35] thanks, [21:35] Jane [21:44] jjohansen: call off the search, soren just beat me up with some knowledge [21:44] it turns out i was using clocksource=jiffies [21:44] soren's suggestion of clocksource=acpi_pm fixed it immediately [21:44] ah, yeah that can cause so problems [21:45] soren also says hello [21:45] yeah, say hi to soren for me too [21:46] can do [21:46] jjohansen: i owe you three cases of beer at this point [21:46] my tab is increasing :/ [21:46] hehe, not at all [21:47] after all soren fixed this problem :D [21:47] other than acpi_pm being a little more expensive than jiffies, should i be worried about anything? [21:51] rackerhacker: not really it is consider the best practice to use acpi_pm on more modern kernels [21:52] rackerhacker: Just make sure you keep proper time on the host. [21:55] soren: can do [21:55] jjohansen: thanks ;) [21:56] np [21:56] :-) === sconklin is now known as sconklin-gone