[00:16] hi, is there a 32 bit kernel with PAE enabled? [00:39] amitk: The ARM porting box (rimu) is online now. [00:39] amitk: FYI. [02:06] Does the cursor get laggy after installing the 2.6.28 tree kernel on 8.10? [02:33] and how do I see the difference between Pentium M and Pentium-4 M= [02:33] ?* [03:30] Question: What should I enable in the kernel to get /proc/profile for kerneltop? === asac_ is now known as asac [16:27] is here sysfs coders? [16:30] or anyone who knows device drivers? [16:33] is here anyone who has made device driver? [17:23] fabbione: hey! [17:23] fabbione: I'm told you're the man to know for sparc stuff :) [17:24] fabbione: I just netbooted a Netfire V210 with the dapper sparc boot.img and the kernel died when I told it which interface to use for DHCP [17:25] fabbione: I'd prefer to be using hardy, but I'm not sure if that will work now that sparc is unofficially supported [17:27] Ng, fabbione uses a self compiled hardy kernel (there is some fix missing in the kernel currently in updates). But the one I gave to compile for -proposed this morning should now be near what he uses... [17:29] smb_tp: aha [17:30] *sigh* Its stuck in the accept queue [17:30] * smb_tp goes nudging some archive maintainer [17:30] smb_tp: so the box in question was running dapper previously and I upgraded it to hardy and the kernel doesn't even start [17:31] Ng, It might be this update from stable that I missed. IIRC fabbione said something like it doesn't boot because of this [17:35] Ng, there were two sparc related changes missed "SPARC64: Loosen checks in exception table handling" and "Fix link errors with gcc-4.3" [17:36] smb_tp: I'm no kernel expert, and maybe there's a quiet option enabled that I'm not seeing in silo.conf, but literally the hardy kernel won't even initialise, it says "Loading linux..." and then sulks indefinitely [17:38] Ng, Right that doesn't sound like the kernel. Unfortunately I only know how to spell SPARC and have never come near one. Is silo somewhat close to lilo code? [17:38] smb_tp: I doubt it, but I don't know. BenC might, he's credited in the manpage ;) [17:39] Ng, Heh, you can try whether he is fit (health wise) enough [17:40] smb_tp: i suspect I've missed fabbione for the night/weekend too, so unless infinity has ideas, I guess I'll have to [17:40] I just know that lilo had some problems when the ramdisk went above a certain limit. [17:41] There you had to give it some "bigmem" option to go on. Otherwise it seemed to crash as well. [17:46] smb_tp: I think the delay on hardy processing may be the finishing out of some -proposed packages that inadvertantly got into the 8.04.2 alt and server images. [17:48] sbeattie, Oh, ok. I was wondering what happened there. fabbione mentioned it this week and first I thought I saw something but then could not find any traces anymore [18:23] guys.. just a few minutes.. i need to catch up on a few things [18:23] Ng: ^^ [18:25] Ng: aren't the V210 the ones in the datacenter? [18:26] fabbione: yep [18:26] Ng: so the installer in hardy is borked.. [18:26] the fix that's missing from hardy will make the machine unbootable after a kernel update [18:26] but smb_tp will fix it, or kittens will die [18:26] last installation I did was: feisty/gutsy -> update to hardy [18:27] fabbione: this was a dapper machine I updated to hardy [18:27] Ng: to get a more detailed output: [18:27] oh [18:27] so you were running hardy [18:27] hold on.. [18:27] i can give you the latest hardy kernel + that one patch to boot hardy [18:28] but it's a UP kernel [18:28] fabbione, I don't want kittens to suffer. The bug fixes are uploaded but stuck in the queue [18:28] or i could give you the patch [18:28] Ng: to boot the kernel in verbose mode on sparc you do: [18:28] boot: Linux -p [18:28] you add the -p [18:29] aha [18:29] i am sure the crash you see is in the MiniRTC driver [18:29] * Ng waits for it to POST to test that [18:30] [ 22.042878] Mini RTC Driver [18:30] [ 22.076708] Unable to handle kernel NULL pointer dereference [18:30] [ 22.143468] tsk->{mm,active_mm}->context = 0000000000000000 [18:30] [ 22.210012] tsk->{mm,active_mm}->pgd = fffff8003042a808 [18:30] [ 22.272400] \|/ ____ \|/ [18:30] [ 22.272407] "@'/ .. \`@" [18:31] [ 22.272413] /_| \__/ |_\ [18:31] [ 22.272419] \__U_/ [18:31] [ 22.448136] swapper(1): Oops [#1] [18:31] something like this [18:31] (much longer) [18:33] fabbione: yep, that's the one [18:34] win [18:34] Ng: http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.24.y.git;a=commit;h=68b498d251d97de9adda518fda42cfe1451063b7 [18:34] and this is the fix [18:34] you can do a chroot on anotehr machine and build + that fix [18:34] or just wait for smb_tp new kerne [18:35] Ng: i hope this helps :) [18:35] Ng: for the installer.. well i haven't isntalled a box in ages [18:35] apt-get dist-upgrade works way too well :) [18:36] the tricky part is getting a fixed kernel onto the machine since I also can't boot the old dapper kernel anymore (I think because in the process of cleaning up after the upgrade, I removed an unused UP kernel and it rebuilt its initrd) [18:36] Ng: just boot a sysrescuecd [18:36] and chroot [18:36] indeed, I meant remotely :) [18:36] thats easy peasy :) [18:36] ah [18:36] well thats a different story [18:37] I'm going to be vaguely near it on monday afternoon, so I'll try and do something then [18:37] fabbione: thanks very much for the info :) [18:38] Ng: you can still netboot from that kernel [18:38] Ng: so get the fixed one built [18:38] hmm, interesting point [18:38] Ng: then check how debian-installer source turn the kernel into a netboot.img [18:38] you won't need to piggyback the initrd as it is already on the disk [18:38] the fix is within vmlinuz [18:39] put that netboot.img to a tftpboot server [18:39] netboot blabla initrd=1/ root=/dev/... [18:39] the only tricky part is to get the initrd values right [18:39] as silo doesn't undestand /dev/sd* [18:39] but IIRC we did setup raid on those boxes [18:39] and you should have something: [18:40] sda1 -> empty [18:40] sda2 -> md0 -> /boot [18:40] sda3 -> md1 -> /restoftheworld [18:40] in that case you would have: [18:41] initrd=1 or 2/initrd-2.6.24-23... [18:41] i don't recall if 1 = sda1 [18:41] or 0 = sda1 [18:41] so in your setup it's either 1 or 2 to get to sda2 [18:41] Ng, BTW, the new kernel package should be on faure. I test build there. Or is that the box you are currently working on? [18:43] smb_tp: it is [18:43] I'm going to pick this up next week, thanks guys :) [18:43] Ng, Murphy [18:44] Ng, I hope until then someone accepted the upload to proposed, so you can get the kernel there [19:03] can any one tell me how to upgrade to latest kernel? [19:04] what is the address of the repository? [19:08] any one who knows how to upgrade the kernel? [19:19] hey guys, i need some help deciding what package to place bug 322610 under [19:19] Malone bug 322610 in linux "[Regression] Screen brightness not 100% at boot or resume from monitor sleep" [Low,Confirmed] https://launchpad.net/bugs/322610 [19:20] I spoke with some guys from the X team, and they think it's probably an acpi kernel issue [19:20] that was my instinct as well - should i keep it filed under linux or place it under something else like acpi-support? [19:21] problem exists for the user in the .27-11 kernel, but not .27-9 [19:21] Rocket2DMn, This might be. So linux is correct. Have you tried with acpi_backlight=vendor or acpi_backlight=video, yet [19:22] what do you mean smb_tp ? [19:24] Rocket2DMn, There have been several issues with acpi and baclight. This is an option which can be placed with the boot options in grub. The default is that backlight is controlled by acpi generic driver and also by vendor specific ones. With that option you can try to select one of them only [19:25] Rocket2DMn, But the package is right [19:25] Rocket2DMn, I mean forthe bug [19:25] smb_tp, in the kern.log's there is some stuff about thinkpad_acpi - it looks like it is surrendering control to the generic driver [19:25] Oh, its a thinkpad [19:26] Then there is another place. Sorry it is a bit complex at the moment [19:26] that happens in both kernels, the only real difference i could find was in lshal where 2 udi's were found in the old kernel and only one in the new kernel [19:26] (see comment 20 for my previous statement) [19:26] smb_tp, i understand it's complicated, that's why i'm here :) [19:26] and it's a lenovo thinkpad, not an ibm [19:27] put an entry "options thinkpad_acpi backlight=1" into /etc/mdprobe.d/options [19:27] Lenovo or IBM doesn't matter in that way :) [19:28] ok smb_tp , ill have my buddy try that, then i'll finish the triage on the bug [19:29] that is just forcing the thinkpad_acpi module to be used, yes? [19:30] Rocket2DMn, Yes, correct. If there is generic support (even if it is not working correctly) it will leave the backlight alone. Please mention the outcome of the option in the bug. [19:30] ok, thanks smb_tp , i expect the triage will be finished later this afternoon [19:32] Rocket2DMn, No hurry. We know of issues there. It is just hard to fix without breaking several others. ;-) [19:32] heh, i know how that goes. thank you for your help [19:32] np :) [20:37] heya [20:37] do ubuntu kernels contain any crashdump facility? i allways wondered why there are analisis utilities like crash but i didnt seem to find anything like redhats netdump or something [20:38] would be great if anyone could enlighten me about that [20:38] realy annoying if there is nothing to trackdown kernel panics [21:21] hi, is something missing in the 2.6.28 kernel so that lvm does not work in the initrd? [21:21] There used to be. What version? [21:21] -6 [21:22] a 2.6.24 kernel boots [21:23] Bug 314879 caused LVM to stop working a few weeks ago, but it's been fixed. [21:23] Malone bug 314879 in lvm2 "root on LVM broken since latest udev 136-2" [High,Fix released] https://launchpad.net/bugs/314879 [21:24] I believe I fixed that the very day it broke. [21:25] but i use udev .114 [21:25] Erm... That sounds.. impossible. [21:26] Current lvm2, mdadm, devmapper, etc. state Breaks: udev (<< 136) or something like that. [21:26] older lvm [21:26] lvm2 2.02.06-4etch1 [21:26] etch? [21:27] Erm... You're on your own, dude. [21:27] so this should work with jaunty? [21:27] It does. [21:27] will try it [21:31] could somebody update module-init-tools in jaunty? it is even older than lenny [21:34] You could? [21:34] i am no u developer [21:35] i test the live images and reuse the kernel [21:35] is there a problem with the current module-init-tools? [21:36] not directly a problem, but when i want to use the u fixes and put it in a repo with lenny it does not update of course [21:37] That maps very poorly to anything resembling a priority for Ubuntu. [21:37] well it would be nice however [21:38] preXX does not looks so good as 3.4 too ;) [21:39] Incidentally, that also maps very poorly to anything resembling a priority :p [21:57] also there must be some keyboard issues with some ps2 (via kvm) or usb (mainly ms wireless) keyboard which are new - 2.6.27 did not have those issues [22:21] soren: OTOH, encrypted / on lvm remains broken in current jaunty (bug 317297, bug 317442) [22:21] Malone bug 317297 in cryptsetup "[jaunty] cryptsetup for root on encrypted lvm not called in initramfs." [High,Confirmed] https://launchpad.net/bugs/317297 [22:21] Malone bug 317442 in cryptsetup "initrd does not contain conf/conf.d/cryptroot file for encrypted root" [Undecided,New] https://launchpad.net/bugs/317442 [22:51] any hints on the crashdump stuff?