UbuntuIRC / 2008 /07 /09 /#ubuntu-kernel.txt
niansa
Initial commit
4aa5fce
[06:05] <TheMuso> c
=== BenC1 is now known as BenC
=== makx is now known as maks_
[12:33] <fabbione> BenC: ping?
[13:42] <gabbs> 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] <gabbs> any idea guys how to fix that? can I build that headers deb seperately?
[13:43] <smb_tp> gabbs, you need the basic headers which can be build with the binary-headers target
[13:43] <smb_tp> fakeroot debian/rules binary-headers
[13:43] <gabbs> cheers smb_tp
[13:44] <smb_tp> gabbs, np :)
[14:16] <rtg> 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] <cking> rtg: that's speedy
[14:33] <rtg> cking: yeah, it should make Kano happy.
[14:34] <cking> rtg: and all MEDION E1210 webcam users too
[14:37] <cking> rtg: Are you hammering the Intrepid kernel on any of your boxes?
[14:37] <rtg> cking: nope. I've been totally focused on Hardy LTS issues.
[14:38] <rtg> why?
[14:39] <BenC> fabbione: ?
[14:39] <cking> 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] <rtg> cking: which network driver?
[14:39] <rtg> or was it in the network stack?
[14:40] <rtg> BenC: can you take a look at davis dmesg and tell me what you think?
[14:40] <cking> rtg I suspect it was this : http://bugzilla.kernel.org/show_bug.cgi?id=10903
[14:42] <rtg> cking: ah, you have a wireless interface on your server?
[14:42] <cking> rtg: yep - cause I'm sick.
[14:43] <rtg> cking: perhaps its better then drilling holes for a cable :)
[14:43] <cking> rtg: it's a cheep RaLink RT2561/RT61 802.11g PCI - does the trick though
[14:43] <rtg> 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] <cking> 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] <rtg> cking: crawling on hands and knees in the attic?
[14:45] <BenC> rtg: davis dmesg?
[14:45] <cking> rtg: fortunately my attic is very high
[14:45] <rtg> 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] <BenC> rtg: could you point me to what exactly I am looking for?
[15:16] <BenC> found it
[15:16] <rtg> end of dmesh.
[15:16] <rtg> blah, can't type
[15:17] <BenC> No thoughts on that right this sec, but I'll check into it later
[15:17] <rtg> BenC: its the second time davis has had problems, but this one looks different.
[15:18] <rtg> I wish I'd have kept the first one. Maybe Ng has it squirreled somewhere.
[15:24] <Ng> rtg: the thing I captured is davis:/home/rtg/cmsj-dmesg.txt
[15:24] <rtg> BenC: ^^
[15:24] <BenC> Ng: thanks
[15:26] <Ng> np, it's the same file that's been there since before, so all I did was check its filename ;)
[15:32] <BenC> Maybe I should break out my G5...but I suspect this is an smp issue and I'm UP
[15:33] <rtg> BenC: it takes awhile, usually I have to thuimp it pretty hard with some kernel builds.
[15:36] <rtg> cking: what package runs when a new ABI kernel is installed in the grub menu?
[15:38] <cking> rtg: Urrmm... not sure.
[15:39] <rtg> cking: hmm, it has to be a post install hook. Michael Fey is having issue with a netbook.
[15:39] <rtg> s/Fey/Frey/
[15:41] <cking> rtg: This is an area where my knowledge is distinctly lacking
[15:42] <rtg> cking: buried somewhere in debian/control-scripts/postinst is the answer
[15:42] <cking> rtg: ah.. more layers of stuff
[15:42] <rtg> cking: standard debian packaging cruft
[15:43] <cking> rtg: ..as I said, "more layers of stuff" :)
[15:43] <rtg> I wonder if update-initramfs does it?
[15:44] <maks_> no
[15:44] <maks_> it does try to call lilo to not fuck you up although
[15:44] <rtg> maks_: but what builds the new entry in grup?
[15:44] <maks_> update-grub
[15:44] <BenC> rtg: it's probably /etc/kernel-img.conf stuff?
[15:44] <rtg> is grub smart enough?
[15:45] <rtg> ah, maybe update-grub
[15:45] <maks_> if it is indeed given as hook in /etc/kernel-img.conf
[15:45] <maks_> not maybe rtg :)
[15:46] <rtg> OMG thats an ugly script.
[15:46] <cking> it appears that /etc/kernel-img.conf have hooks for update-grub
[15:47] <rtg> yep
[15:47] <BenC> rtg: yeah, it was ripped from kernel-package, and I actually diluted it down a bit
[15:47] <BenC> rtg: it could use a rewrite, but I've been of the opinion that "it works, leave it alone"
[15:48] <rtg> you'll get no argument from me.
[15:49] <cking> BenC: ..even if it is a maintenance nightmare in waiting
[15:49] <BenC> cking: feel free to step up to the plate :)
[15:49] <cking> BenC: Maybe when all the bugs are fixed :-)
[15:50] <cking> ..and the sun has turned into a lump of coal
[16:04] <gabbs> when I compile the target "custom-binary-openvz", which config do I have to alter ?
[16:06] <rtg> gabbs: debian/binary-custom.d/openvz/config.*
[16:06] <gabbs> and my selection of the processor type will decide if its i386 or amd64?
[16:07] <gabbs> hmm ... no, that cant be lol - so which one gets taken? and whats figuring that out?
[16:11] <rtg> 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] <gabbs> ah, cheers for that information! ツ
[16:14] <fabbione> BenC: drbd module doesn't load in 2.6.26-3
[16:14] <fabbione> BenC: just a plain modprobe will fail.
[16:18] <BenC> fabbione: the only error path I see for that is -EINVAL, which means "connector" cn exists
[16:43] <gabbs> /root/openvz/debian/build/custom-source-openvz is not clean, please run 'make mrproper'
[16:43] <gabbs> what does that mean?
[16:43] <gabbs> I run make mrproper, rerun AUTOBUILD=1 NOEXTRAS=1 fakeroot debian/rules custom-binary-openvz and it fails again
[16:45] <rtg> gabbs: are you using the spurce package, or have to cloned the git repo? mrproper removed the debian directory
[16:45] <rtg> s/spurce/source
[16:46] <gabbs> I cloned it from the git repo, yea
[16:46] <gabbs> then edited with make menuconfig the config file and tried to build the packages
[16:46] <BenC> fabbione: fixed in git
[16:46] <rtg> gabbs: ok, 'git checkout -f' to recover, but it will also blow away any local changes
[16:47] <gabbs> rtg, those local changes thou (changes config) seem to cause this?
[16:47] <rtg> dunno, depends on what changes you've made.
[16:47] <rtg> can you at least build without any local changes?
[16:48] <gabbs> well, literally only a make menuconfig, selected the debian/binary-custom.d/openvz/config.i386, exited - agreed to save, started build
[16:49] <rtg> 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] <gabbs> hmm .. whats debuild ?
[16:50] <rtg> install devscripts
[16:50] <rtg> sudo apt-get install devscripts
[16:51] <gabbs> well, I am getting an output here
[16:51] <gabbs> 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] <gabbs> (expected linux_2.6.24.orig.tar.gz or openvz.orig)
[16:51] <gabbs> I assume thats ok if I got it from git ?
[16:52] <rtg> that fine. what command did you run?
[16:52] <gabbs> oh damn, I did a -d instead of -b
[16:52] * gabbs blushes
[16:53] <gabbs> well, I did not confirm the "continue anyway?"
[16:53] <rtg> ok, try 'debuild -b'
[16:53] <gabbs> dpkg-buildpackage -rfakeroot -D -us -uc -b failed
[16:53] <gabbs> should I switch to a non-priv user first?
[16:53] <rtg> it should be bitching about unmet build dependencies
[16:54] <rtg> you can also build from a standard user login
[16:54] <gabbs> oh, I missed that one! kernel-wedge
[16:54] * gabbs loads it
[16:56] <rtg> 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] <gabbs> so I don't have to meet all deps?
[16:57] <gabbs> because some seem a bit excessive
[16:58] <gabbs> ah, works now
[16:58] <rtg> gabbs: thats what deps are all about. if you don't meet them, then you're on your own
[16:58] <gabbs> the above command will start compilation, right?
[16:59] <gabbs> because I'd need to modify the config still
[16:59] <rtg> gabbs: make sure you can build from the repo _first_, then make modifications.
[17:00] <gabbs> rtg, good point - thou when I execute the command, I get the same error msg
[17:00] <gabbs> the weird part is, I did compile a kernel like that roughly 4 hours ago
[17:01] <rtg> 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] <rtg> soren: never mind. I booted the Intrepid kernel.
[17:02] <rtg> gabbs: what is the error message?
[17:02] <gabbs> does any of this give you an idea what might be wrong rtg: http://rafb.net/p/RwgOKt39.html ?
[17:02] <gabbs> I notice the "silentoldconfig" - could that cause problems?
[17:03] <rtg> gabbs: 'make mrproper;git checkout -f'
[17:04] <gabbs> afterwards issue the command for compilation again?
[17:04] <rtg> fakeroot debian/rules custom-binary-openvz
[17:05] <gabbs> without AUTOBUILD=1 NOEXTRAS=1 ?
[17:05] <rtg> yes - without AUTOBUILD=1 NOEXTRAS=1
[17:05] <gabbs> ah, seems to compile now
[17:06] <gabbs> should I stop it and try with a changed config or let it run?
[17:06] <rtg> let it finish
[17:06] <gabbs> hm, thats gonna take an hour or so, but ok
[17:29] <gabbs> btw, does anyone know what a halted boot sequence after "Checking 'hlt' instruction... done." ?
[17:39] <BenC> 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] <BenC> fabbione: and make sure to undef the internal ones (idx is conflicting with uvesafb)
[17:43] <BenC> 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] <gabbs> 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] <BenC> 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] <BenC> rm -rf build
=== asac_ is now known as asac
[23:02] * powitsjj waves
[23:04] <powitsjj> anyone home?
[23:09] <Pizzaboy192> good morning
[23:10] <Pizzaboy192> i need some hardware support
[23:10] <powitsjj> lol...that doesnt help pizza, we ended up in the same place
[23:12] <Pizzaboy192> bummer
[23:12] <powitsjj> yeah
[23:12] * powitsjj shoots pizza in the face
[23:13] <Pizzaboy192> (eats 6 year old chocolate)
[23:13] <powitsjj> pwned
[23:13] <Pizzaboy192> i should make an irc channel
[23:14] <Pizzaboy192> #windows-dhould-die
[23:15] <Pizzaboy192> chocolate is better when it is aged
[23:17] <Pizzaboy192> it seems to get a more "firmented" taste to tit
[23:17] <Pizzaboy192> *it
[23:20] <powitsjj> hmm
[23:21] <Pizzaboy192> lol
[23:22] <Pizzaboy192> i wont say what i said, but you can see i said it
[23:22] <Pizzaboy192> lol
[23:22] <Pizzaboy192> im gonna go
[23:22] <Pizzaboy192> ****************** talk to me on aim thelastwacker@aim.com
[23:33] <the-fafa> hello. what does preemt mean for a kernel?