[01:22] hey guys, had a quick question. Bug 332170 is talking about a message printed during boot and seems to be related to a problem some people are having with their systems performing poorly. [01:22] Malone bug 332170 in linux "[Jaunty] Error appearing at boot - cpufreq: No nForce2 chipset" [Undecided,Confirmed] https://launchpad.net/bugs/332170 [01:22] I see the message "cpufreq: No nForce2 chipset" also, but don't experience extra slowness, though I've always used ondemand for cpufreq [01:44] The message is harmless [01:49] thanks mjg59 , i'll let them know [01:50] Should really be KERN_INFO rather than KERN_ERR though [01:51] where do you see that it is ERR? [01:51] The sourcecode [01:51] I'll send a patch upstream [01:51] ah, if it was INFO would it be printed to the screen? [01:51] No [01:52] Ok, so you want me to change the bug to something saying that message should be KERN_INFO instead of KERN_ERR [01:52] ? [01:54] Feel free [01:54] The slowness is because p4-clockmod is binding and getting used with the ondemand governor [01:54] * Rocket2DMn nods [01:54] is it just adapting then and will improve? [01:55] I've told Scott how to fix that [01:56] fix which? the printing or the slowness? [01:57] The slowness [01:57] Ok, so what we really have is two bugs in cpufreq then? One for slowness and one for changing STD_ERR to STD_INFO ? [01:59] Fix slowness: [02:00] sed -i s/1000000/10000001/ arch/x86/kernel/cpu/cpufreq/p4-clockmod.c [02:01] I don't have the kernel source code, you are changing a value of 1 million to 1 million + 1? [02:02] Well, I'll take your word for it :) Shall I include that in the report, or do you know if there is another report open for the slowness problem? [02:04] i thikn i see it [02:04] bug 332017 [02:04] Malone bug 332017 in linux "Significant performance regression in 2.6.28-8.24 due to p4-clockmod" [High,Fix committed] https://launchpad.net/bugs/332017 [02:04] No, one million to ten million and one [02:07] ah, I didn't read close enough, sorry. Is the Scott in the bug I just listed the same Scott you mentioned earlier? [02:07] Scott James Remnant [02:07] Yes [02:08] Ok, great! I'll just have this bug I started with be used to fix the KERN_ERR issue. [02:08] Thank you for your help mjg59 [02:10] Rocket2DMn: Patch just got accepted upstream, so it'll be fixed in 2.6.30. The fix can be backported easily. [02:10] for the slowness? [02:11] For the message [02:11] I'm chatting with the cpufreq maintainer about the slowness issue [02:11] Ah, i just marked as triaged, guess I'll change to Fix Committed. Thanks for the update [02:11] Well, it's not committed to the Ubuntu kernel [02:11] So I don't know that fix ommitted is correct [02:11] that would be Fix Released [02:12] last time i checked anyway [02:12] I've lost track of how launchpad stuff is triaged nowadays [02:12] https://wiki.ubuntu.com/Bugs/Status [02:12] For us, Committed means its committed anywhere, even upstream [02:16] Hey you wouldn't happen to have a link to an upstream bug report or a commit page would you? [02:25] Nope, it hasn't been pushed yet [02:26] heh ok, back to Triaged we go. That's all I have, thank you again for your time mjg59 . [02:28] No problem [02:28] It should turn up in the cpufreq git tree shortly [02:37] hi all, I am getting the error -> kernel: [365385.583926] python[943]: segfault at 8908ec93 ip b7c50c8c sp bf843430 error 4 in libgobject-2.0.so.0.1800.2[b7c46000+3c000] [02:37] is there anyway that I can fix it by add/remove certain package?? [02:41] It's not a kernel problem - the kernel is just reporting that one of your applications crashed [03:28] mjg59, I see. Thanks. [04:21] I isolated a kernel bug a week ago, it's totally ubuntu's fault, not related to upstream, it's a regression and it's easy to fix [04:22] I put a comment on the bug report and emailed the person who committed the bug [04:22] but still no response [04:22] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/297213 [04:22] Malone bug 297213 in linux "Intrepid: Missing hdaps_ec kernel module" [Medium,New] [04:23] is there anyone here willing to look at it? [13:26] quiet channel [13:31] It's weekend. [13:42] you mean you people have a life outside ubuntu? [13:46] morning [13:50] morning [13:51] iulian: I'll take that as a yes [14:05] what's up with the ubuntu kernel naming scheme? [15:04] LLStarks: What do you mean? [15:05] 2.6.28-8 = 2.6.28.6 [15:05] right? [15:10] LLStarks: the -8 is the ABI number [15:10] what does that mean? [15:11] It means that that number will be incremented by the Ubuntu packagers to signify that kernel modules must be rebuilt [15:12] but surely the kernel that gets packaged is based off of a point release, isn't it? [15:12] hello, I posted a bug report with a patch. i wounder if I did everything right: https://bugs.launchpad.net/ubuntu/+source/linux-meta/+bug/330259 [15:12] Malone bug 330259 in linux-meta "Problematic volume control keys and mute key of several notebooks + Patch" [Undecided,In progress] [15:13] what version is the current jaunty kernel based on? [15:16] is there anything else to be done by me ? [15:16] what is the difference between 2.6.28-8 and 2.6.28.6? [15:20] nobody here who can answer my question? [15:21] LLStarks: the key thing to understand is that the .6 and -8 are telling you about totally separate concepts [15:21] * LLStarks is n00b. Explain to me like you would a child. [15:22] The ABI number, which always comes after a hyphen, serves one purpose only: to designate the compatibility (or not, as the case may be) of the built kernel with particular builds of modules [15:23] The point release number designates the number of the upstream point release. [15:23] A point release may *or may not* cause an ABI bump [15:23] Conversely, and ABI bump may be required for reasons other than a point release [15:23] (i.e. an Ubuntu-specific change) [15:24] s/and/an/ [15:24] does -8 imply a point release has occurred? [15:24] No [15:25] the ABI number implies nothing whatsoever about point releases [15:25] (Though it would be somewhat unusual for 7 ABI bumps to have occurred without a single point release happenning) [15:25] (7, because it starts at 1) [15:26] so. when i roll a 20-sided kernel +5 charisma and kms support, what i am getting? [15:26] *kernel for [15:26] To discover which point releases have been integrated into an Ubuntu kernel, you must inspect the package changelog [15:26] **am i [15:27] why is it like that? shouldn't such things like a point release be readily reflected? [15:27] No [15:29] The fact that you need to rebuild your kernel modules is a more significant event than the upstream stable branch happening to release [15:29] i see. [15:29] Information on which point releases are included is always available in the package changelog [15:31] how can i tell which ABI i'm rolling when i use kernelcheck? [15:32] I've not encountered kernelcheck, what is it? [15:33] dot dot dot. [15:33] http://kcheck.sourceforge.net/ [15:33] LLStarks: We now build upstream kernels and provide them in the archives [15:34] i saw. [15:34] no kms though. [15:34] * LLStarks gives angry glare [15:34] not yet, give it time! [15:34] That'll come for Karmic [15:35] can't you just enable i915.modeset? [15:36] You'd have to ask Andy. So far as I'm aware they are being built with the Ubuntu configs... but... I've not actually looked closely at that :) [15:36] LLStarks: I'm not entirely certain I understand your question, but assuming you mean "When I build my own kernel by non-Ubuntu means, which Ubuntu ABI am I getting?", the answer is "Quite possibly none of them, if you're building your own kernel, build your own modules too." [15:38] You can get the abi with "make kernelversion" or "make kernelrelease" [15:39] what folder? [15:42] the root of the kernel source [15:42] that's part of the standard kernel Makefile [15:42] "kernelrelease" only works after "make prepare" [15:42] can I repeat my question from yesterday? [15:43] TimStarling: I saw it... not had time to do anything though [15:43] ok [15:43] I also got a question you got time for me ? [15:44] so should I keep asking people until someone gives in and does it? [15:45] TimStarling: I seem to recall several months ago, something similar happened with another module. Can't remember which one, though :s [15:45] seems kind of silly that a module can just go missing like that [15:46] There could have been a good reason... some (in)compatibility issue that wasn't logged in the changelog [15:46] true [15:47] I have noticed a trivial bug in the lbm prepare-compat-wireless.sh script - a missing $ sign on a shell variable expansion. Is there someone around who could just apply that trivial change? [15:47] that's why I emailled Ben Collins first [15:49] I've had this laptop for almost a year now, and ubuntu has slowly been supporting more and more components in it [15:50] Looking at the commit diff you reference, I'm almost sure I've dealt with that previously. [15:51] the hard drive has had a bit of corruption and I've been wondering if it's about to go, that's what made me think of hdaps [15:55] if you need me to help test it, I'll be around === TimStarling is now known as Tim-away [16:05] ha.. found it! https://lists.ubuntu.com/archives/kernel-team/2008-September/003171.html [16:06] Tim-away: Looks like the same forgetfulness as happened Gutsy > Hardy [16:32] maxb: {WT} ? [16:37] yes [16:37] I guess I could send a patch to the kernel team, but it's a single character :-) [16:37] ok [16:38] they all count :D [20:38] hi amber [20:38] are you going to try to help with the bug jamming [20:38] ? [20:40] er...wrong channel