[06:05] c === BenC1 is now known as BenC === makx is now known as maks_ [12:33] BenC: ping? [13:42] linux-headers-2.6.24-20-openvz depends on linux-headers-2.6.24-20 however Package linux-headers-2.6.24-20 is not installed. [13:42] any idea guys how to fix that? can I build that headers deb seperately? [13:43] gabbs, you need the basic headers which can be build with the binary-headers target [13:43] fakeroot debian/rules binary-headers [13:43] cheers smb_tp [13:44] gabbs, np :) [14:16] wow, only 2 days to get a patch in Linus's tree. that has to be the fastest ever. '[PATCH] uvcvideo - Add support for MEDION E1210 webcam' [14:33] rtg: that's speedy [14:33] cking: yeah, it should make Kano happy. [14:34] rtg: and all MEDION E1210 webcam users too [14:37] rtg: Are you hammering the Intrepid kernel on any of your boxes? [14:37] cking: nope. I've been totally focused on Hardy LTS issues. [14:38] why? [14:39] fabbione: ? [14:39] rtg: I upgraded my server and had a whole load of ssh lockups - fortunately the current git has a fix for this bug. Good way to waste 2 hours figuring out the problem [14:39] cking: which network driver? [14:39] or was it in the network stack? [14:40] BenC: can you take a look at davis dmesg and tell me what you think? [14:40] rtg I suspect it was this : http://bugzilla.kernel.org/show_bug.cgi?id=10903 [14:42] cking: ah, you have a wireless interface on your server? [14:42] rtg: yep - cause I'm sick. [14:43] cking: perhaps its better then drilling holes for a cable :) [14:43] rtg: it's a cheep RaLink RT2561/RT61 802.11g PCI - does the trick though [14:43] cking: I'm thinking of doing the same for this SuperMicro cabinet I received last week. Its so shrill I can't stand to have it in the house. [14:44] rtg: yep - this was I can shove the box into the loft. It's just a pain when it goes wrong when I fowl up an upgrade [14:45] cking: crawling on hands and knees in the attic? [14:45] rtg: davis dmesg? [14:45] rtg: fortunately my attic is very high [14:45] BenC: davis.canonical.com is ill since the LTS upgrade. [14:56] * cking thinks "Liberate tutemet ex inferis" should be "Liberate tutam me ex inferis".. [15:15] rtg: could you point me to what exactly I am looking for? [15:16] found it [15:16] end of dmesh. [15:16] blah, can't type [15:17] No thoughts on that right this sec, but I'll check into it later [15:17] BenC: its the second time davis has had problems, but this one looks different. [15:18] I wish I'd have kept the first one. Maybe Ng has it squirreled somewhere. [15:24] rtg: the thing I captured is davis:/home/rtg/cmsj-dmesg.txt [15:24] BenC: ^^ [15:24] Ng: thanks [15:26] np, it's the same file that's been there since before, so all I did was check its filename ;) [15:32] Maybe I should break out my G5...but I suspect this is an smp issue and I'm UP [15:33] BenC: it takes awhile, usually I have to thuimp it pretty hard with some kernel builds. [15:36] cking: what package runs when a new ABI kernel is installed in the grub menu? [15:38] rtg: Urrmm... not sure. [15:39] cking: hmm, it has to be a post install hook. Michael Fey is having issue with a netbook. [15:39] s/Fey/Frey/ [15:41] rtg: This is an area where my knowledge is distinctly lacking [15:42] cking: buried somewhere in debian/control-scripts/postinst is the answer [15:42] rtg: ah.. more layers of stuff [15:42] cking: standard debian packaging cruft [15:43] rtg: ..as I said, "more layers of stuff" :) [15:43] I wonder if update-initramfs does it? [15:44] no [15:44] it does try to call lilo to not fuck you up although [15:44] maks_: but what builds the new entry in grup? [15:44] update-grub [15:44] rtg: it's probably /etc/kernel-img.conf stuff? [15:44] is grub smart enough? [15:45] ah, maybe update-grub [15:45] if it is indeed given as hook in /etc/kernel-img.conf [15:45] not maybe rtg :) [15:46] OMG thats an ugly script. [15:46] it appears that /etc/kernel-img.conf have hooks for update-grub [15:47] yep [15:47] rtg: yeah, it was ripped from kernel-package, and I actually diluted it down a bit [15:47] rtg: it could use a rewrite, but I've been of the opinion that "it works, leave it alone" [15:48] you'll get no argument from me. [15:49] BenC: ..even if it is a maintenance nightmare in waiting [15:49] cking: feel free to step up to the plate :) [15:49] BenC: Maybe when all the bugs are fixed :-) [15:50] ..and the sun has turned into a lump of coal [16:04] when I compile the target "custom-binary-openvz", which config do I have to alter ? [16:06] gabbs: debian/binary-custom.d/openvz/config.* [16:06] and my selection of the processor type will decide if its i386 or amd64? [16:07] hmm ... no, that cant be lol - so which one gets taken? and whats figuring that out? [16:11] gabbs: it depends on your platform, but dpkg-architecture tells the build which one to use. You can spoof it using a chroot wherein you can build 32 bit targets using a 64 bit host. [16:13] ah, cheers for that information! ツ [16:14] BenC: drbd module doesn't load in 2.6.26-3 [16:14] BenC: just a plain modprobe will fail. [16:18] fabbione: the only error path I see for that is -EINVAL, which means "connector" cn exists [16:43] /root/openvz/debian/build/custom-source-openvz is not clean, please run 'make mrproper' [16:43] what does that mean? [16:43] I run make mrproper, rerun AUTOBUILD=1 NOEXTRAS=1 fakeroot debian/rules custom-binary-openvz and it fails again [16:45] gabbs: are you using the spurce package, or have to cloned the git repo? mrproper removed the debian directory [16:45] s/spurce/source [16:46] I cloned it from the git repo, yea [16:46] then edited with make menuconfig the config file and tried to build the packages [16:46] fabbione: fixed in git [16:46] gabbs: ok, 'git checkout -f' to recover, but it will also blow away any local changes [16:47] rtg, those local changes thou (changes config) seem to cause this? [16:47] dunno, depends on what changes you've made. [16:47] can you at least build without any local changes? [16:48] well, literally only a make menuconfig, selected the debian/binary-custom.d/openvz/config.i386, exited - agreed to save, started build [16:49] make sure you have a stable build environment first. see if 'debuild -b' will at least start. it ensures that all build deps are satisfied. [16:50] hmm .. whats debuild ? [16:50] install devscripts [16:50] sudo apt-get install devscripts [16:51] well, I am getting an output here [16:51] This package has a Debian revision number but there does not seem to be an appropriate original tar file or .orig directory in the parent directory; [16:51] (expected linux_2.6.24.orig.tar.gz or openvz.orig) [16:51] I assume thats ok if I got it from git ? [16:52] that fine. what command did you run? [16:52] oh damn, I did a -d instead of -b [16:52] * gabbs blushes [16:53] well, I did not confirm the "continue anyway?" [16:53] ok, try 'debuild -b' [16:53] dpkg-buildpackage -rfakeroot -D -us -uc -b failed [16:53] should I switch to a non-priv user first? [16:53] it should be bitching about unmet build dependencies [16:54] you can also build from a standard user login [16:54] oh, I missed that one! kernel-wedge [16:54] * gabbs loads it [16:56] gabbs: once 'debuild -b' has gotten opast the first few lines you can kill it and run ' fakeroot debian/rules clean custom-binary-openvz' [16:57] so I don't have to meet all deps? [16:57] because some seem a bit excessive [16:58] ah, works now [16:58] gabbs: thats what deps are all about. if you don't meet them, then you're on your own [16:58] the above command will start compilation, right? [16:59] because I'd need to modify the config still [16:59] gabbs: make sure you can build from the repo _first_, then make modifications. [17:00] rtg, good point - thou when I execute the command, I get the same error msg [17:00] the weird part is, I did compile a kernel like that roughly 4 hours ago [17:01] soren: is there a secret to installing kvm on Hardy? 'apt-get install kvm' produces "FATAL: Error inserting kvm_intel (/lib/modules/2.6.26-4-generic/kernel/arch/x86/kvm/kvm-intel.ko): Operation not supported" [17:01] soren: never mind. I booted the Intrepid kernel. [17:02] gabbs: what is the error message? [17:02] does any of this give you an idea what might be wrong rtg: http://rafb.net/p/RwgOKt39.html ? [17:02] I notice the "silentoldconfig" - could that cause problems? [17:03] gabbs: 'make mrproper;git checkout -f' [17:04] afterwards issue the command for compilation again? [17:04] fakeroot debian/rules custom-binary-openvz [17:05] without AUTOBUILD=1 NOEXTRAS=1 ? [17:05] yes - without AUTOBUILD=1 NOEXTRAS=1 [17:05] ah, seems to compile now [17:06] should I stop it and try with a changed config or let it run? [17:06] let it finish [17:06] hm, thats gonna take an hour or so, but ok [17:29] btw, does anyone know what a halted boot sequence after "Checking 'hlt' instruction... done." ? [17:39] fabbione: you will need to get drbd utils to build-dep on the new linux-libc-dev (>= 2.6.26-4.10) so that it can pick up the correct linux/connector.h where CN_{IDX,VAL}_DRBD are defined [17:42] fabbione: and make sure to undef the internal ones (idx is conflicting with uvesafb) [17:43] connector idx sync between kernel/userspace is pretty dumb...I would think there would be a way to do connectors by name, but I can't seem to find that mechanism === LifeHacker is now known as aanjhan [20:02] how do I use "make oldconfig" and make it write to debian/build-custom.d/openvz/config.i386 ? === rikai_ is now known as rikai [21:46] gabbs: mkdir build; cp debian/build-custom.d/openvz/config.i386 build/.config; make O=`pwd`/build oldconfig; mv build/.config debian/binary-custom.d/openvz/config.i386 [21:46] rm -rf build === asac_ is now known as asac [23:02] * powitsjj waves [23:04] anyone home? [23:09] good morning [23:10] i need some hardware support [23:10] lol...that doesnt help pizza, we ended up in the same place [23:12] bummer [23:12] yeah [23:12] * powitsjj shoots pizza in the face [23:13] (eats 6 year old chocolate) [23:13] pwned [23:13] i should make an irc channel [23:14] #windows-dhould-die [23:15] chocolate is better when it is aged [23:17] it seems to get a more "firmented" taste to tit [23:17] *it [23:20] hmm [23:21] lol [23:22] i wont say what i said, but you can see i said it [23:22] lol [23:22] im gonna go [23:22] ****************** talk to me on aim thelastwacker@aim.com [23:33] hello. what does preemt mean for a kernel?