[00:00] I could patch the source manually each time I want to rebuild but I'd like to find a way to automate the process a bit [00:04] There seems to be two guides out there. One that tells you how to build Ubuntu Kernels and one that tells you how to make packages with make-kpkg for a bastardized patched vanilla kernels [00:05] There is nothing about making a clean debian package with patches from a vanilla kernel [00:05] I've been searching this for weeks! [00:08] miconfrink we have a script that can help, let me see if i can find it [00:09] great!! [00:13] miconfrink: start with "git clone git://kernel.ubuntu.com/ubuntu/kteam-tools" [00:13] k one sec [00:13] miconfrink: you want to then look at kteam-tools/mainline-build/README [00:14] miconfrink: essentially, you have your upstream kernel, a reference to one of our kernels and you run the script and it will debian-ize it and build it [00:15] miconfrink: i don't think this will pull our ubuntu-specific patches into your code tree however [00:16] I'm more concerned about how to put Molnar's patches into my tree [00:17] I'm looking at the README now. Thanks bjf [00:26] bjf: So if I understand the README and the scriptfile this will basically give me a Debianized vanilla kernel [00:27] yes [00:27] miconfrink: ^ [00:27] THen I can deconstruct the deb - I guess [00:28] and add Molnar's patch [00:28] miconfrink: have you tried to start with our kernel tree and add the realtime patch to it? [00:29] No I've not tried that but it's a pretty extensive patch [00:29] I know that someone used to keep a Molnar RT kernel in the Ubuntu repo but it got abandoned [00:29] miconfrink: the delta between our kernel and the upstream mainline kernel really isn't that large, it might be worth giving it a try [00:30] okay [00:30] I'll look into it [00:30] my understanding is that debian patches use quilt [00:30] whereas Molnar uses patch [00:31] Is there an easy way to convert between formats? [00:31] miconfrink: we just use git, no quilt [00:31] I thought debian packages use quilt for patching [00:32] miconfrink: they do, but the kernel team package is different, we need to be as compatible with Linus' tree as possible [00:32] I see [00:32] that makes sense [00:33] so how does the kernel team do things differently [00:33] miconfrink: take one of our trees, do a "git am " and see what happens. worse case you've wasted some bandwidth and a little time [00:33] miconfrink: but if it works, you'll be way ahead [00:33] I see [00:34] I've never used git to patch [00:35] miconfrink: you can just use patch if you want [00:36] cool [00:37] How can I tell what vanilla version the git is based on? [00:37] miconfrink: which version do you want? Precise is 3.2 based [00:39] 3.2.16 is the last 3.2 based Molnar patch [00:39] http://www.kernel.org/pub/linux/kernel/projects/rt/3.2/ [00:40] miconfrink: then precise is the one you want, if you do a git log on the master branch you will see we are patched up to 3.2.14 [00:41] miconfrink: if you look at the master-next branch we are patched up to 3.2.16 [00:42] thanks [00:44] can you give me the git path? [00:44] miconfrink: what git patch? [00:45] For either branch [00:45] I'm not quite getting it right [00:45] miconfrink: sorry, i misread what you typed [00:46] miconfrink: "git clone git://kernel.ubuntu.com/ubuntu/ubuntu-precise" [00:46] gotcha thanks a billion! [00:46] miconfrink: in that repo there are both the "master" and "master-next" branch [05:51] good rainy day to everyone :) === smb` is now known as smb [07:28] ppisati, Not right now, but good morning anyway. :) [07:28] smb: a thunderstorm woke me up thus morning, i love when that happens :) [07:30] ppisati, Isn't the relative safety of so many other houses around you to hit something to enjoy? ;) [07:31] smb: when i'm in bed, under a mountain of blanket, i don't really care about other's house safety :D [08:29] hi i've got a question about dkms... [08:30] i'm currently using 10.04 netbook remix with 2.6.32-41-generic [08:30] my netbook has a realtek wireless card [08:30] i have to use this ppa to get drivers which work with my wireless card [08:31] http://www.todoleo.com/2011/01/how-to-get-wifi-working-on-dell-mini.html [08:31] every time the kernel updates, it breaks my wifi [08:31] and i have to apt-get --reinstall the ppa packages [08:32] i'm not sure why i have to do that... shouldn't dkms fix things automagically (or have i misunderstood its purpose)? [08:34] kengyu, ^ that's your ppa [08:34] ahh, yeah, is my ppa. [08:36] ensignkim, rtl8192ce is supposed to to be in mainline already. Did you try to fully uninstall the dkms deb and see if the wifi works? [08:37] kengyu__, he's using 10.04 [08:37] kengyu__, was it in mainline there as well? === kengyu__ is now known as kengyu_ [08:37] yep, i'm currently on lucid still but will upgrade to 12.04 soon [08:40] kengyu_: i'm not sure :) wifi breaks every time there is a kernel upgrade, and i do something to fix it but can't remember what - i think i may do an "apt-get purge" [08:42] ensignkim, then could be my badly packaged dkms. :-( I haven't look at it for a while. the deb could be of low quality since it sees no hope to get in the main archive or so (already one in mainline). [08:44] ppisati: *poke* [08:44] ppisati: Andy tells me that you'd love to have a closer look at https://bugs.launchpad.net/ubuntu/+source/linux-ti-omap4/+bug/986713 [08:44] Launchpad bug 986713 in linux-ti-omap4 "boot on panda shows kernel oops" [Undecided,New] [08:45] kengyu_: so, hopefully wifi should work for me out of the box when i upgrade to 12.04? thanks for the ppa btw, without it my netbook would have been somewhat useless :) [08:48] ensignkim, Realtek engineer worked on upstreaming the driver, so it is supposed to work as well as the one in my ppa. [08:54] infinity: we had these WARNs() since... natty? [08:54] infinity: but ack, i'll take a look [08:54] ppisati, be good to confirm its the known ones at least :) [08:54] ppisati: Nothing wildly urgent, but Andy implied that you have *nothing* to do. ;) [08:54] :) [08:55] infinity: he's kind of right... :) [08:56] at least someone i know someone is testing it... :) [08:57] -ETOOMANYSOMEONE === smb` is now known as smb [11:26] * ppisati -> lunch [11:45] smb, is this something you have a test environment to confirm: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/987715 [11:45] Launchpad bug 987715 in linux "In KVM In a virtual machine, system crashes if vmvga video card is used" [Undecided,Confirmed] [11:46] apw, Without looking into the whole description it sounds like any kvm host/guest could do when changing the graphics card (the emulated) [11:47] * cking reboots his AP [11:48] apw, I might need to reinstall kubuntu... just checking with vmvga and normal ubuntu install [11:51] apw, Unity-2d seems not to crash. Interestingly the default resolution picked is 2360x1770... === yofel_ is now known as yofel [12:11] smb, woh that is pretty big [12:12] apw, Yeah, it made hitting things a bit hard when scaling down into my resolution... :) [12:13] Right now installing kubuntu on the vm [12:19] diwic, https://bugs.launchpad.net/ubuntu/+source/alsa-driver/+bug/987650 <-- might be of interest [12:19] Launchpad bug 987650 in linux "Audio Realtek ALC889, Green/Black/Yellow Line Out, rear jacks: no sound when front jack plugged into headphones" [Undecided,Confirmed] [12:23] apw, sure, is there anything that makes it more or less interesting than all the other audio bugs? [12:42] apw, I cannot confirm a crash, but rendering not being optimal (sometimes large areas not redrawn, splash not being rendered) with the vmvga setting. For this test the host gfx was ati (not sure about the report) [13:40] apw, For the kvm-vmvga bug I think the main problem could be in the host part (which is Natty) and I am not sure whether I still have a host running that around. [14:10] cking, yo ... we have an HP failing to suspend, leading to hang. this is triggering from echo "mem" >... [14:10] cking, whats the next step :) [14:10] apw, hey, is this a regression or not? [14:11] cking, i believe its 'always' been this way [14:11] apw, https://wiki.ubuntu.com/DebuggingKernelSuspend [14:12] https://wiki.ubuntu.com/Kernel/Reference/S3SystemTapDebug [14:13] apw, the latter may provide more info, but takes a while to set up - it was working for precise last month, so I suspect it still should work OK [14:15] cking, sounds good. will get psivaa on the case :) [14:16] I suggest they file a bug, follow the systemtap instructions and attach the logs generated by the debugging tool into the bug [14:16] apw, BTW, what kind of HP is it? [14:16] cking, already done it [14:17] cking, 987840 [14:17] cking, a pavillion g6 [14:18] ** [14:18] ** Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting [14:18] ** [14:22] psivaa, so can you see if you get any logs when following the instructions in: https://wiki.ubuntu.com/Kernel/Reference/S3SystemTapDebug [14:22] cking, just following the debugging information [14:22] psivaa, thanks! [14:23] cking, will let you know once i get the longs [14:23] * logs [14:23] ack [14:26] Hello [14:26] I've opened an upstream bug report at bugzilla.kernel.org [1] but they want that I test a patch. [14:26] 1: https://bugzilla.kernel.org/show_bug.cgi?id=43155 [14:26] It's not a problem except that I don't find the .dsc file linked to the version 3.4-rc4-precise => http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.4-rc4-precise/ [14:26] bugzilla.kernel.org bug 43155 in Sound(ALSA) "hda-intel: Sound card takes time to be initialised" [Normal,New] [14:26] Where can I find this .dsc file? :) [14:35] matttbe: I have not tested by myself, but I guess you have to apply those 3 patches against vanilla 3.4-rc4 sources [14:36] [10:17:03] https://wiki.ubuntu.com/Kernel/MainlineBuilds [14:37] matttbe: ^^ [14:42] dileks: yes I already test the 3.4-rc4 version by installing packages from http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.4-rc4-precise/ [14:42] dileks: now I just want to test a patch. Should I've to download tarball from kernel.org and apply patches from http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.4-rc4-precise/ [14:44] whoever is responsible for the wiki: airlied pulls from danvet's GIT tree (drm-intel-next) [14:44] http://cgit.freedesktop.org/~danvet/drm-intel/log/?h=drm-intel-next [14:44] 3. the tip of the drm-intel-next head of Keith Packard's linux repository daily, [14:54] * ogasawara back in 20 [14:55] dileks: thank you but I just want to easily recompile the kernel with a patch which is not in any git repos mentioned on the wiki [14:56] make deb-pkg [14:57] matttbe: http://paste.ubuntu.com/944190/ [14:57] dileks: thank you :) [15:07] cking: Magic: 11:330:749 maps to hash: b664bb followed by Hash matches: oops_enter() (address: 0) [15:07] cking: is the output of the debug [15:08] psivaa, there may be some logs created, can you attach them to the bug report please [15:08] cking: ack [15:21] jsalisbury: re bug 987052 -- happened with lbm-cw-3.3, added a comment and a pic (FWIW) [15:21] Launchpad bug 987052 in linux "WARNING: at /build/buildd/linux-3.2.0/kernel/softirq.c:159 local_bh_enable_ip+0x7a/0xa0()" [High,Confirmed] https://launchpad.net/bugs/987052 [15:34] psivaa, I've added some more instructions to the bug report. [15:36] cking: ok ill have a look and follow [16:00] apw, cking, if not built into the kernel, how do cpufreq modules like powernow-k8 get loaded? I don't see anything in initramfs-tools that will do it. [16:04] tgardner, it's not clear to me either [16:04] I see some x86cpu id tables [16:06] MODULE_DEVICE_TABLE(x86cpu, ... [16:10] cking, bug #353552 ? [16:10] Launchpad bug 353552 in jockey "nvidia driver crashes on login after fisrt install (dup-of: 349419)" [Undecided,New] https://launchpad.net/bugs/353552 [16:10] Launchpad bug 349419 in python-qt4 "jockey-kde crashed with SIGSEGV in QWidget::metric() on kubuntu 9.04 startup" [Medium,Fix released] https://launchpad.net/bugs/349419 [16:10] 355232 [16:10] bug #355232 [16:10] Launchpad bug 355232 in linux "acpi-cpufreq/powernow-k8 should not be built-in into the kernel image" [Low,In progress] https://launchpad.net/bugs/355232 [16:21] tgardner, they have to be manually specified, i thought that they were all builtin for that reason [16:21] apw, yep, I'm just coming around to that. [16:22] tgardner, iirc there is an annotation to that effevt in the config comparitor [16:23] apw, I wonder if we could add a kernel command line parameter to no load these cpufreq drivers, ala https://bugs.launchpad.net/ubuntu/+source/linux/+bug/355232/comments/15 [16:23] Launchpad bug 355232 in linux "acpi-cpufreq/powernow-k8 should not be built-in into the kernel image" [Low,In progress] [16:24] cking: i have updated the bug report to the effect the suspend on the console mode seems working but every reboot requires pressing of the brightness adjustment key [16:24] tgardner, commit 22ee5756cd21063ba8405598060b0f1ef8d81a6d [16:24] Author: Konrad Rzeszutek Wilk [16:24] Date: Tue Mar 13 19:18:39 2012 -0400 [16:24] provide disable_cpufreq() function to disable the API. [16:25] smb, 448c8b1d07d5342922567e138a4b0108a42c24fb [16:26] Though this looks a bit like one could not use the framework then... [16:26] tgardner, Yep, sorry that was picking up the sha1 of my tree [16:27] smb, yeah, I think what we really want is for the built-in drivers to not register, but leave the infrastructure intact. [16:27] Yeah, agree [16:28] this one is overshooting, ok for Xen since they then use something completely outside the framework [16:28] smb, we can likely manage to do both. [16:43] brendand, bug 926136 - can I get access to this machine so I can do some hands on debugging? [16:43] Launchpad bug 926136 in linux "CPU1 on Dell PowerEdge M610, R715 and IBM X3500 M3 goes offline after exercising frequency governors" [Medium,Triaged] https://launchpad.net/bugs/926136 === herton_ is now known as herton [17:11] * henrix back in 30 [17:12] * cking --> EOD and going to fix his washing machine === jsalisbury changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Ubuntu Kernel Team Meeting - Tues May 1st, 2012 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! [17:24] tgardner: heya - I've come to argue with you over the wontfixing of bug #942846 :) [17:24] Launchpad bug 942846 in grub2 "encrypted install fails to boot as long as vt.handoff=7 is used" [Undecided,Fix released] https://launchpad.net/bugs/942846 [17:24] slangasek, you should likely argue with apw since the vt.handoff patch is his. [17:24] tgardner: I'm not sure if the remaining issue is a kernel bug or not, but I think it's pretty important, UX-wise, that users not be dropped to a black screen instead of an encryption prompt [17:24] well, you wontfixeded it [17:25] apw: I've been redirected to you [17:25] slangasek, well, its pretty low on our list of priorities given some of the other issues we're dealing with. [17:25] slangasek, what the issue ? i thought we didn't pass it anymore from grub [17:26] slangasek, and that we should be saying we wouldn't for encrypted installs [17:26] apw: tgardner says he still encounters a black screen which he works around by pressing Esc [17:26] hmm? [17:26] encrypted installs most certainly still get vt.handoff [17:26] and it works for most of us [17:27] I understand it's not at the top of the todo list, but it's still an important bug (whoever's it actually is) [17:27] then his machine is broken and should be added as broken to the blacklist right? [17:28] so that it doesn't use handoff, as it doesn't work appropriatly in this situation [17:28] apw, where is the blacklist ? grub2 ? [17:28] grub-gfxpayload-lists [17:28] e.g., /usr/share/grub-gfxpayload-lists/blacklist/20_radeon_hd6800 [17:29] yeah ... i think that was the basic plan :) [17:29] ok - I don't think that plan was handed off :) [17:30] slangasek, ok, I'll see if blacklisting works, but I still don't think there is gonna be a kernel fix. [17:30] tgardner: can you get us PCI ID for your affected machine on that bug, then? And I'll open a task for grub-gfxpayload-lists [17:30] slangasek, i think we are a little confused indeed. i think we tried to fix most of them and those that we couldn't we ensured that the blacklist would do the right thing [17:30] slangasek, yep, gimme a bit [17:31] apw: the blacklist is approximately empty by default [17:31] slangasek, and then we needed to find the exclusions and drop them in the blacklist [17:31] slangasek, indeed very much so [17:31] we do additional blacklisting if you install the non-free, non-KMS drivers [17:31] tgardner: do you think it's not a kernel bug, or just not one that will get fixed any time soon? [17:32] slangasek, I've no idea to be honest. [17:32] I just know that I'm not gonna work on it. [17:32] * slangasek nods [17:33] hmm, I should probably take another pass over the plymouth bug list now that RAOF has blacklisted the HD 6800 [17:35] slangasek, i assume its a bug in the initialiseation from 'not the default text mode' in the graphics driver [17:35] that was my approximate thought as well [17:36] slangasek, hmm, apport-collect is giving me shit about not being the bug owner (or some such crap) [17:37] bah [17:37] In Ubuntu 12.04, I installed linux-tools-base, and then tried to install linux-base. /usr/bin/perf is in both of these packages, making this impossible. Is this a known issue? [17:37] tgardner: attach lspci -vvnn by hand? [17:37] slangasek, yeah [17:38] rather specifically linux-tools-common conflicts with linux-base [17:38] RAOF: hey, so bug #975931 is the same issue as bug #971204, but with an incremented PCI ID [17:39] Launchpad bug 975931 in plymouth "purple screen freeze with AMD 6850" [Undecided,New] https://launchpad.net/bugs/975931 [17:39] Launchpad bug 971204 in linux "graphics fails with setgfxpayload=keep, AMD Radeon" [Medium,Confirmed] https://launchpad.net/bugs/971204 [17:39] (1002:6738 vs. 1002:6739) === tgardner is now known as tgardner-lunch [17:53] jsalisbury: any reason to have a kernel meeting before uds ? [17:54] bjf, hmm, probably not, unless there is a specific reason? [17:54] bjf, I can change the next meeting to after UDS? [17:57] +1 [18:06] woot, day-0 uploaded [18:06] Home: https://wiki.ubuntu.com/Kernel/ || Ubuntu Kernel Team Meeting - Tues May 15th, 2012 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! [18:20] * _ruben wonders if that was an attempt to alter the /topic :) [18:25] jsalisbury: you didn't get the topic changed :-) === tgardner-lunch is now known as tgardner [19:15] cnd: https://blueprints.launchpad.net/ubuntu/+spec/other-q-arsenal-report-enhancements [19:16] bjf, ahh, oops [19:17] I think I'll just remove my blueprint then [19:17] and subscribe to this one [19:17] cnd, good plan. add anything to the whiteboard you feel needs to be discussed [19:18] yep [19:18] thanks! [19:48] thought this is a kernel problem but seems to be a nm issue [19:48] https://bugzilla.redhat.com/show_bug.cgi?id=785772 [19:49] bugzilla.redhat.com bug 785772 in NetworkManager "Logs filled by ICMPv6 RA: ndisc_router_discovery() failed to add default route" [Unspecified,On_qa] [19:54] cnd, we were editing the whiteboard at the same time, there doesn't seem to be any locking so my changes were lost [19:56] cnd, i just re-added my text [19:59] bjf, oh, sorry :( [19:59] hey [20:00] do you folks know if Intel Centrino Ultimate-N 6300 AGN is well supported in 12.04? [20:00] I am looking to buy a new laptop [20:01] cnd, not your fault, stupid LP [20:02] jono, tgardner or sforshee would know best but I believe it is [20:02] tgardner, sforshee, any thoughts on this? [20:03] jono, I came in on the tail end of this conversation, but I assume its about wifi ? [20:03] jono, it's supported but I also believe it's one that's had some persistent issues [20:03] tgardner, yeah, I am looking to buy a new laptop, and I have a few options, I want to figure out the best driver support [20:04] my options are: [20:04] ThinkPad 1x1 b/g/n [20:04] Intel WiFi Link 1000 [20:04] Intel Centrino Advanced-N 6205 AGN [20:04] Intel Centrino Ultimate-N 6300 AGN [20:04] Intel CA-N+WiMAX 6250 MOW [20:04] any feedback on which of those is best supported in 12.04? [20:04] sforshee, I see [20:04] jono, Intel is well supported, but there _are_ folks complaining about throughput in certain scenarios [20:05] jono, I've tested all of these models and found them to perform OK. [20:05] thanks tgardner [20:05] I will get the Ultimate-N and see how I get on [20:06] jono, good luck [20:06] thanks tg [20:06] jono: 6205 has worked quite well for me [20:06] thanks tgardner [20:06] vanhoof, oh yeah? [20:06] * tgardner reboots for cpufreq patch testing [20:06] jono, I have a 6205 too and have had good luck with it [20:06] vanhoof, is it working well in 12.04? [20:06] jono: yup, Lenovo x220 [20:06] ok, I am going to do that then :-) [20:06] thanks sforshee, vanhoof [20:07] jono: np [20:09] I am also going with Intel HD Graphics 3000 graphics [20:09] seems to be generally pretty decent [20:09] jono: which machine are you after? [20:09] vanhoof, T520 [20:11] jono: just integrated graphics or nvidia as well? [20:11] vanhoof, just integrated graphics [20:11] jono: ok, you'll be good then [20:11] do you know if Intel HD Graphics 3000 is certified? [20:11] the other is switchable [20:11] right [20:11] oh yeah, its sandybridge [20:11] cool [20:11] the machine itself is also ubuntu certified [20:12] vanhoof, do you know if SATA solid state discs are OK? [20:12] was enabled on 10.10, certified on stock 11.04, 11.10, and 12.04 is in progress [20:12] jono: oh yeah :) [20:12] cool thanks vanhoof [20:12] jono: but you'd likely want to purchase one off of amazon [20:12] * jono buys [20:12] a thinkpad? [20:12] they charge a pretty hefty premium [20:12] for the SSD [20:12] oh really? [20:13] jono: http://www.amazon.com/Intel-Series-2-5-Inch-Solid-State-Retail/dp/B004T0DNQ0/ref=sr_1_2?s=electronics&ie=UTF8&qid=1335298379&sr=1-2 [20:13] looks about the same price...this machine I am getting is in an employee pricing sale [20:14] ah ok === jsalisbury changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Ubuntu Kernel Team Meeting - Tues May 15th, 2012 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! [20:34] * tgardner -> EOD [23:33] glad to see my kernel changes were picked upstream :D