| === Guest34418 is now known as lag | ||
| === lag is now known as Guest66638 | ||
| === yofel_ is now known as yofel | ||
| basix | hello everyone... can someone let me know what the issue is here? I am unable to get a useable kernel module compiled on ANY of the mainline kernels in the ubuntu kernel-ppa repository. Heres the bug and some info from a virtualbox developer who says the issue is that the headers somehow don't match the kernel. https://www.virtualbox.org/ticket/9891 | 20:10 |
|---|---|---|
| basix | i am currently running 3.1.2 but i'll be upgrading to 3.1.6 and checking if the issue still persists | 20:11 |
| basix | this is the kernel ppa i am currently using : http://kernel.ubuntu.com/~kernel-ppa/mainline/ | 20:11 |
| basix | anybody? | 20:20 |
| apw | basix, the compiler the mainline kernels are built with don't match those on the system most likely | 20:21 |
| apw | basix, they are only intended for testing | 20:21 |
| basix | apw, how do i find out the compiler used to build mainline kernels? Also, shouldn't this be an issue with the regular ubuntu kernels as well? I mean a user can always upgrade / downgrade GCC on his system and when a package like virtualbox, vmware tries to build it's kernel modules, it'll cause such issues. | 20:22 |
| basix | apw, i agree they are intended for testing. i wanted to test out the latest virtualbox but unfortunately i cant build any version on my system currently | 20:23 |
| apw | basix, we build the mainline kernels on lucid | 20:23 |
| basix | apw, i am using natty | 20:24 |
| apw | basix, so reboot into an official kernel, those are built in a chroot that matches your system | 20:24 |
| basix | apw, my wireless has a bunch of issues if i boot into an official kernel | 20:24 |
| basix | apw, as well as sleep / power issues and thats the whole reason i am on a mainline kernel | 20:24 |
| basix | apw, are there source packages for the mainline kernel so i can rebuild it natively on my system and that should fix the issue, right? | 20:26 |
| apw | basix, yes i expect if you build it on natty it would work ok | 20:27 |
| basix | apw, so are there any source packages or do i need to build one myself? | 20:28 |
| apw | basix, there are no source packaged, but there is file COMMIT which contains the linus commit to checkout -- and a set of patches which if applied gives you the tree the builder used | 20:28 |
| basix | ok | 20:28 |
| apw | external modules are a royal pain, those guys should jsut submit their stuff upstream | 20:29 |
| basix | apw, alternatively if i setup a lucid chroot and build the modules in there, would that work as well? | 20:29 |
| apw | basix, i would hope that would work also; that might make more sense | 20:30 |
| * apw wonders if we could build vb modules with them ... hmmm | 20:30 | |
| basix | apw, haha yes. also i suspect this is an issue not just affecting virtualbox. i tried installing vmware and it too had the same issue. | 20:31 |
| apw | basix, yep, yet another external module... pain in the ass always | 20:31 |
| basix | any idea what the overall process to build these modules would look like inside a chroot? I haven't build modules in some time let alone inside a chroot. | 20:32 |
| apw | the problem is that people want to use these kernels on a range of series, and we don't really have compute or the space to build for every conceivable combination | 20:32 |
| apw | basix, it should be identicle | 20:32 |
| basix | after i create a chroot for lucid, do i install gcc and other build essentials and build those modules in that chroot? | 20:33 |
| apw | oh and ... you should mention to vb pepople that they are compiler sensitive so their build really ought to check | 20:33 |
| basix | i think i will reopen the bug | 20:33 |
| basix | but i need to be 100% certain it is a compiler | 20:34 |
| basix | apw, so i already have a chroot | 20:34 |
| basix | apw, when i look in /lib/modules inside the chroot there are no headers which kinda makes sense because it's a clean root | 20:34 |
| basix | apw, i will just install the header package inside the chroot as well as virtualbox. it should then go through the process of building the modules as usual and i can then copy those modules out of the chroot. does that sound reasonable? | 20:36 |
| basix | apw, sudo debootstrap --variant=buildd --arch amd64 lucid /srv/chroot/lucid_x86_64 http://archive.ubuntu.com/ubuntu <- does this look ok? | 20:39 |
| apw | basix, heh, it looks believable | 20:41 |
| basix | ?? | 20:41 |
| apw | basix, i have no idea if its purfect of the top of my head, i use some prebuilt ones the majority of the time | 20:43 |
| basix | apw, ok | 20:44 |
| basix | apw, it's a no go. i was able to compile the modules. they got 'modprobed' without any issues within the chroot but when i copied them outside the chroot into /lib/modues/$(uname -r)/misc, it seems the issue persisted. do i need to run depmod or is there some sort of a module cache that i need to clear to ensure the older modules are not loaded? | 21:31 |
| basix | brb | 21:33 |
| basix | apw, ok...so i was able to 'modprobe | 21:50 |
| basix | the modules from within the chroot | 21:50 |
| basix | apw, that worked but outside the chroot, it still causes the panic. i wonder if it has to do with something at the runtime | 21:51 |
| basix | apw, for posterity's sake, i am pinging you. thanks for all your help. i got the modules to build inside the lucid chroot and that fixed the issue. there were some old modules remaining in /lib/modules/3.1.2-030102-generic/updates/dkms which made it seem that the issue was not fixed since modprobe was favoring that path over /lib/modules/$(uname -r)/misc where the newly built modules were. After I replaced the modules in the dkms directory, I was | 23:24 |
| basix | able to modprobe all the virtualbox modules successfully and able to restart virtualbox and run a VM. This is Virtualbox 4.1.8 on kernel 3.1.2. | 23:24 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!