[00:13] JanC: yes, or any bootloader [00:14] heywood: aside from the echoing the value may not actually bring that core offline? === Tim-away is now known as TimStarling [03:07] the ubuntu mainline kernel builds can be used on any release of ubuntu? even 8.04? [03:08] i'm talking about these - https://wiki.ubuntu.com/KernelMainlineBuilds [08:20] * Panarchy syas Hi [08:20] * Panarchy says Hi [08:22] I have a modified Hardy Kernel (part of Debris Linux). I wish to replace this with the latest Intrepid Kernel. How would I go about doing this? RWW suggested "sudo apt-get install linux-generic", would that be the best way to do this, once I have installed Intrepid repostiories? [08:41] Folks, I just saw https://wiki.ubuntu.com/KernelMainlineBuilds and the instructions are relatively complex due to the manipulation of *.debs; did you consider setting up an APT archive instead? You can do so easily with apt-ftparchive [08:45] hello all [08:45] hi [08:47] dumb question. [08:47] anyone know the difference between passing "noht" to the kernel at boot time (via grub's menu.lst) and issuing something like "echo 0 >/sys/devices/system/node/node0/cpu1/online"? [08:52] heywood:i heard that some kernel dont support the noht param [08:52] so in those cases you can use the sys interface [08:53] any idea where i could look up which do and don't? i'm reading docs for sysfs right now. [08:53] oh, i'm running 2.6.27 [08:54] just check the kernel sources [08:55] IIRC noht is not present in recent kernels anymore, but max_cpus=1 would have the same effect [08:56] or you can write a script that checks /proc/cmdline and accordingly uses sys interface [09:17] @anubhav: basically i'm trying to figure out how to reliably (but temporarily) disable hyperthreading [09:17] i have a task that is sensitive to interrupts, and i want to see if disabling HT will help or hurt performance [09:18] i have full control over the machine it's running on (so no need to make the script too smart or generic) [09:19] so i know in advance what got passed at boot time -- since i get to pass it :) [15:06] does anyone know if the lockd regression fixed in 2.6.29 affects jauntys 2.6.28 kernel? [15:09] Keybuk: does initramfs leave a log somewhere after remounting the real file system? I' [15:09] rtg: no, what would you like to find out? [15:09] I'm trying to figure out 0_kdump which works when run by hand. [15:10] 0_kdump is installed in scripts/init-bottom by kexec-tools [15:10] /usr/share/initramfs-tools/scripts/init-bottom, that is. === TimStarling is now known as Tim-away [15:52] hi there ... i am trying to compile the latest (2.6.28-7) kernel via kernelcheck; the kerenel is built and the .deb file mdae .. but it returns an error in installation: " Failed to process /etc/kernel/postinst.d at /var/lib/dpkg/info/linux-image-2.6.28.7-ultimate.postinst.d at line 1181. [15:53] someone an idea? or a tip how to compile it another way ? [16:00] b10h4z4rd:https://help.ubuntu.com/community/Kernel/Compile [18:27] Hi, i've been trying to upgrade kernel on my debian, tried the normal way and debian way, but there's always some errors end eventually kernel's not upgraded. Today need to upgrade kernel to the new one 2.6.28 or newer. Can someone please go with me through all the steps?? [18:32] amartin83: see here https://help.ubuntu.com/community/Kernel/Compile/ [18:36] amartin83: also see https://help.ubuntu.com/community/CategoryKernel [18:37] wmat, so maby better go through all steps again until errors occurs, but this will take some time couse as i remember it's a long process [19:06] amartin83: it would take far longer for someone to lead you through it. [19:07] amartin83: you should post errors as you encounter them [19:07] amartin83: one note though, this channel is specific to kernel development, so you may not get much help [19:07] ok, i'll do that [19:08] amartin83: you'd likely get more help in #ubuntu [19:08] oh, ok [19:10] just wonder, why there's always so much difficulties with kernel? Seen on many forums that people always get trouble, even if they do exactly like in recipe [19:10] can someone help me troubleshoot a suspend/resume problem with a sata disk? I tried so many things but can never complete a bug report *because* my main disk dies on resume. I suspend to ram and it sleeps in seconds. I power up, takes about a minute and I have nothing but I/O errors to my disk :( [19:11] amartin83: there's really too many variables out there to answer that question conclusively [19:13] wmat, thanks anyway [19:16] when do I do the dmesg > ? ... after trying to resume right? [19:16] before and after according to apw [19:17] if you do it before you have more chance of not needing dmesg to be loaded from the disk [19:17] to run [19:21] humm try: [19:21] $dmesg ; sudo pm-suspend; dmesg [19:22] who knows.. maybe it works [19:22] ok. will try it then... hold up, should I say dmesg > disk ; sudo pm-suspend; dmesg > disk? [19:22] no [19:22] tath would replace the 1st log [19:23] dmesg > disk ; sudo pm-suspend; dmesg >> disk [19:23] or [19:23] dmesg > disk ; sudo pm-suspend; dmesg > disk2 [19:23] ok, will try it then [19:23] apw: is there a flag or echo that can be done to increase the verbose of kernel output [19:23] while suspending/hibernate? [19:24] about to do it. should I say sudo dmesg or no? [19:24] just do it [19:25] boot without quiet set [19:25] no need for that [19:25] dmesg doesn't need sudo [19:25] i would change to VT-1 [19:25] apw, I always do. everything says fine... the only thing that fails is apparmor. talking about couldn't load it or something [19:25] login there, run dmesg, then sudo pm-suspend [19:26] ok, going for it, let me go to tty1 [19:30] vbgunz: ping ?? [19:48] I can never resume from a suspend. I managed to get the dmesg after a fail suspend sent to a MS. here it is. can someone offer a clue as to whats wrong? http://dpaste.com/4775/ [19:48] I mean, well, I can *but* my sata is dead [20:01] vbgunz, yep looks like your disk and controller are out of sync, get that put into a bug on the kernel [20:01] not sure what to do exactly :( [20:02] am going to try pci=nomsi on the kernel parameter with IDE and then AHCI ... see if that works? [20:02] ok but get a new bug in launchpad filed [20:02] and add the information you have to it [20:02] so we can track it [20:02] is there a way, without rebooting and modding the kernel paramter to enable the pci=nomsi now? [20:03] ifconot that i know of [20:05] apw: We've already got a bug report going since last week on this, which I am assigned to [20:05] bug #334644 [20:05] Malone bug 334644 in linux "[MSI MS-7374] suspend/resume failure [non-free: nvidia]" [Undecided,Incomplete] https://launchpad.net/bugs/334644 [20:06] IntuitiveNipple, ahh good, then i can go back to sleep [20:06] thats my bug [20:07] apw: :p [20:07] vbgunz, how many didks do you have? 3? [20:08] the other two seem to have come back up, what is different about them? [20:09] you might want to try switching the two drives round [20:09] (two of the drives) [20:11] apw: The lspci controller info says (non AHCI mode) but ahci reports it is managing the controller. [20:11] apw: Also, vbgunz said "... I saw 2 options for AHCI in bios but they both instead say IDE..." [20:12] ok [20:14] smb_tp_, I fired off an email to mario, looking at the problem it looks like an installer issue [20:14] checking my mobo manual. sata port 5, 6 support AHCI and raid mode only... I am plugged into 1 and see nothing. I'll keep checking :) [20:14] manjo, The BIOS one or the other one? [20:14] the other one [20:15] we had a fix for the raid stuff go in for jaunty [20:15] fixed by cjwatson [20:16] not sure of the bios one that I need to look at it on HW [20:16] I have a MSI mobo, MS-7374 [20:17] smb_tp_, you think mario can give us access to such a hw ? [20:18] manjo, He more likely than any other I would guess. But maybe I am the wrong to ask. ;-) [20:19] :) [20:20] well, let me try then to see if the pci=nomsi would work... brb === zul_ is now known as zul [20:42] rtg: / apw: any thoughts on reenabling PREEMPT for -generic for jaunty? [20:42] what is the driver? [20:42] apw: resolving audio aberrations when PulseAudio is used in glitch-free mode [20:43] compared to turning off glitch-free mode it sounds like high risk adventure [20:44] why would we not turn off the glitch0free as it patently introduced glitches [20:44] apw: i'm fine going either way regarding PREEMPT, but i need to know if i need to hack PulseAudio further [20:46] my feeling is that turning on something like that at this stage is risky [20:47] apw: currently PulseAudio works around alsa-kernel and alsa-lib bugs regarding snd_pcm_avail*() (being debugged in upstream alsa-devel) [20:47] pa has been broken all along, can we not just turn off glitch-free, its intent is only for use where latency is good [20:47] nad if latency is not good it is not appropriate [20:47] (to my reading [20:48] the biggest problem is that the workarounds for snd_pcm_avail*() are only used if glitch-free is enabled [20:48] * apw looks askance at the pile of doo doo that PA seems to be [20:49] so to prevent PA from segfaulting when bogus data is returned, i'd need to alter PA to ignore the glitch-free checks [20:50] its a nightmare eitehr way [20:50] i guess i just need to be diplomatic about suggestion that the checks are changed [20:50] is pa really readdy for primetime at all? [20:50] suggesting* [20:50] i think it's ready if all the stars are aligned, which requires kernel config changes [20:53] well, i'll go ahead and make the changes to PA so at least it won't segfualt [20:53] can't please everyone :) [20:54] apw: thanks for your input [20:57] dtchen_: I don't relish your position! I'd be screaming :)