[06:11] <brandon> #ubuntu-sound
[06:13] <brandon> hey, im just here on my quest to get sound to work on my laptop
[06:15] <brandon> how do i use kernel?
[06:17] <ripdisk> i need some help getting your kernel to boot, it's telling me it attempted to kill the idle task
[06:17] <ripdisk> this is for the live cd
[06:18] <ripdisk> i'm trying to install on an old laptop
[07:57] <kraut> moin
[07:58] <ripdisk> i guess nobody has anything to say about the kernel panic i'm getting, then
[09:21] <Kano> hi
[09:24] <Kano> hi rtg 
[09:25] <Kano> when can we expect removed ids from bcm43xx and prism54 and iwl39/49 in lum?
[09:26] <rtg> Kano: I am unaware of this issue. Have you started an LP report?
[09:27] <rtg> Kano: in particular, what is wrong with iwlwifi?
[09:27] <Kano> bcm43xx and ssb share same ids (ssb is used for the b43* drivers and works usually better)
[09:27] <Kano> rtg: nothing is wrong with iwl, but it is disabled in the kernel - marked for inclusion in lum
[09:28] <rtg> Kano: right, but I have not upload either for build. I have some other stuff to work on before I'm ready to upload.
[09:28] <Kano> also the old prism54 module + new p54pci share use same ids, so remove it for both old modules
[09:29] <Kano> rtg: they are disabled in git and when you build it you get no iwl..
[09:30] <rtg> Kano: I'm aware of that, but the git repo is not in a consistent state at the moment.
[09:30] <rtg> I'll get to it this week.
[09:31] <Kano> waiting for it
[09:38] <majikins> hi can anyone help me with a raid1 lvm setup on ubuntu server 7.10?
[09:39] <majikins> when I test by removing one drive, i get dropped into initramfs prompt
[09:39] <majikins> I've googled and have ensured that that the module file includes md and raid1
[09:40] <majikins> ran update-initramfs -u as well
[09:40] <majikins> now I'm stuck
[09:41] <Kano> what do you use as root option
[09:41] <majikins> I don't know what to answer to that? I'm a noob and followed a howto to get lvm and raid setup
[09:42] <Kano> usually you get that problem when the root partition is not found. this can be specified using uuid or /dev/mappper/... in case of lvm
[09:43] <majikins> ok - pls can you help me specify that - where do I go to?
[09:45] <Kano> that was a general hint, i only use u kernels, not u itself
[09:46] <smb> abogani: Hi Alessio, I wanted to quickly follow up on my mail about audit settings. Do you think audit syscall should be configured for -rt?
[10:14] <abogani> smb: Stefan sorry for my poor english and for missed replay.
[10:15] <abogani> smb: At the moment if audit syscall is or isn't configured it is unimportant for -rt.
[10:15] <abogani> smb: Feel free to do the most convenient thing for you. Anyway i prefer to remain stick on -generic's configuration as soon as possible.
[10:15] <abogani> smb: Thank you and sorry again for missed replay.
[10:15] <abogani> s/soon as/
[10:19] <smb> abogani: Alessio, no need to apologize for your english. Mine isn't perfect either. Ok, so (since i just added this to -generic) I will modify the -rt configurations as  well. Thanks.
[10:25] <abogani> smb: Thank you for your kindness. 
[10:56] <smb> abogani: Just noticed you already did the change yourself. Sorry for the bother then.
[10:58] <laga> Kano: hey, do you use aufs or unionfs for your live disks?
[10:58] <Kano> well i did some testing
[10:59] <Kano> standard kanotix does not rely on unionfs or aufs, just when you want it you can enable it
[10:59] <laga> Kano: i did some testing, too, and found unionfs 1.4 (which is in ubuntu) to be unusable with nfs branches. i was just wondering if you made aufs work on ubuntu's 2.6.24 :)
[11:00] <laga> Kano: interesting. what do you use instead? symlinks to a tmpfs?
[11:00] <Kano> but for next gen i need it. an image with network-manager does not work when unionfs is used but aufs makes network-manager working
[11:00] <Kano> debian live based it is just a switch to set it to aufs or unionfs
[11:00] <laga> nice
[11:01] <Kano> i added the security perm patch to be able to compile aufs
[11:01] <laga> adding such a switch is not hard, especially when you compile aufs with unionfs compatibility.
[11:01] <Kano> you just need to add 2 files to the headers manually as they are not there by default
[11:01] <laga> Kano: on ubuntu's 2.6.24?
[11:02] <Kano> the standard headers are not enough
[11:02] <Kano> i guess you could go without the c file if you put it into lum
[11:02] <Kano> as it is only grepped
[11:03] <laga> i wanna put it into lum, but i need to convince the kernel people first i suppose. also need to add some patches to the kernel itself as you said.
[11:03] <Kano> one patch
[11:03] <Kano> if you like i can create 2 kanotix ng images
[11:03] <laga> Kano: do you still have your aufs source tree? i'd be interested in a copy. compilation is broken with the ubuntu kernel because of apparmor patches (IIRC)
[11:03] <Kano> with your latest kernel
[11:04] <Kano> i fixed compiliation of course, when you add the 2 files
[11:04] <laga> Kano: i'm not an official kernel developer, i just need aufs for my work on mythbuntu
[11:04] <Kano> http://kanotix.com/files/kernel/kernel-update-pack-generic-future/source/sec_perm-2.6.24.patch
[11:04] <Kano> thats the only patch needed
[11:05] <Kano> http://kanotix.com/files/thorhammer/updates/aufs/aufs_0+20071217-0+c0.kanotix.1.dsc
[11:05] <Kano> latest snapshot patched for apparmor
[11:06] <laga> Kano: thanks a lot! i'll take a look
[11:09] <laga> debian/patches/06_remove_ia_file.dpatch - heh, i tried the same hack. does that work for you?
[11:15] <Kano> the added patch is 05
[11:16] <Kano> and 06
[11:16] <Kano> maybe 06 is not needed anymore
[11:16] <Kano> did not check
[11:17] <laga> 06 is still needed last time i looked
[11:17] <laga> ie there's no ia_file member anymore
[11:17] <laga> (in the ubuntu kernel)
[11:18] <laga> now that i think about it, maybe my problems were caused by some other missing patches.
[11:19] <Kano> well i adopted the 05 patch from the old aufs from ubuntu
[11:19] <Kano> fixed positions
[11:19] <laga> Kano: i was talking to the aufs guy and he recommended some additional patches. i can forward the mail if you want to
[11:20] <Kano> i wanted to write him too,but my mail was rejected
[11:20] <laga> did you sign up on his mailing list?
[11:20] <Kano> nope
[11:20] <laga> if you want to send email there you have to sign up
[11:21] <laga> your email address is master at kanotix dot com, right?
[11:21] <Kano> yes
[11:21] <Kano> thats not the adress i use to "write" but to get mails
[11:21] <Kano> is basically only a forward
[11:22] <laga> got mail
[11:22] <Kano> laga: btw. i patch my kernels even with dmraid45
[11:23] <Kano> updated dmraid of course too
[11:23] <laga> dmraid is for those "fake" raid adaptors, right?
[11:23] <Kano> yes
[11:23] <Kano> ubuntu does not support raid level 5
[11:23] <laga> raid 5 sounds nice. too bad i'm too cheap for that :)
[11:24] <Kano> did not try it myself yet
[11:24] <laga> ok, i need to run now. thanks a lot for your help, i'll try the patches ASAP.
[11:24] <Kano> but patched the kernel even for 2 ubuntu users + updated dmraid for em *g*
[11:25]  * laga wonders why make-kpkg produces kernel packages > 220MB with linux-sources and default ubuntu config while leaving
[11:27] <Kano> laga: take a look at my scripts
[11:27] <Kano> http://kanotix.com/files/kernel/kernel-update-pack-generic-future-pure/source/
[11:27] <Kano> get-source*
[11:28] <Kano> these compile a patched kernel, you can even add rc8 git patches with ease
[11:30] <zdzichuBG> I believe getting http://lkml.org/lkml/2008/1/17/544 into Hardy kernel is out of question?
[11:47] <Kano> splice-2.6.23.patch, put_filp.patch, and deny_write_access.patch.
[11:47] <Kano> did you use those?
[11:47] <Kano> i think some of em are autenabled
[11:49] <Kano> hmm these are kernel patches, maybe i should add em too
[15:48] <abogani> rtg: Are you around?
[15:49] <abogani> rtg:  I have founded this meta-package: linux-image-debug-rt. Do you want provide a debug version for -rt kernel flavour? BenC always refuse to provide it...
[15:50] <BenC> abogani: and we will still refuse to do so :)
[15:52] <BenC> abogani: that's a huge package to keep around, plus I'm not sure the binary-custom build can handle DEBUG=y kernels+modules
[15:53] <abogani> BenC: :)
[15:54] <abogani> BenC: Could you cut out that meta-package?  For avoid to confuse users... Thanks.
[15:55] <BenC> rtg: you get that?
[15:57] <rtg> BenC, abogani: np.
[15:58] <BenC> rtg: thanks
[16:00] <abogani> rtg: Thanks! 
[16:33] <rtg> abogani: linux-meta uploaded
[16:35] <abogani> rtg: Thank you very much.
[18:14] <Kano> rtg: when do you add iwl to lum?
[18:14] <rtg> Kano: before I do the next upload.
[18:26] <laga> Kano: thanks for your script, i'll try it in a few minutes. have you had the chance to add those additional patches?
[18:26] <Kano> did not try, want to wait for the fixes i proposed
[18:26] <Kano> but nobody likes to do those it seems
[18:26] <laga> :)
[18:27] <laga> ok, i'll add those patches then
[18:29] <Kano> best add latest git patch too
[18:30] <laga> no, i want to use a plain ubuntu kernel
[21:42] <_MMA_> Hi guys. I having a user asking about using more than 3GB of RAM on -RT and -generic. Says server kernels do. Since I don't have 3GB of RAM and haven't had this issue I don't know what to say.
[21:45] <stgraber> he can either switch to amd64 kernels (which shouldn't have the problem) or use the server kernel, highmem is AFAIK not enabled on other kernels
[22:00] <crimsun> yes it is.
[22:00] <crimsun> grep -i highmem /boot/config-2.6.24-4-generic  =>  CONFIG_HIGHMEM4G=y
[22:01] <mjg59> crimsun: If it's mapped above 4GB (because there's PCI stuff in the way, or whatever) then it might need CONFIG_HIGHMEM64G
[22:03] <crimsun> hmph, we don't enable 2/2 or full or whatever
[22:05] <mjg59> That's not relevant for highmem
[22:05] <mjg59> The 3:1 split is fine
[22:06] <crimsun> no, the full option or whatnot
[22:06] <mjg59> Mm?
[22:06] <crimsun> oh, nevermind, misread the help.
[22:07] <mjg59> It wouldn't help - with CONFIG_HIGHMEM4G you're still limited to 4GB of physical address space
[22:07] <mjg59> And PCI has to fit in there
[22:07] <crimsun> (the VMSPLIT_2G_OPT, which isn't even relevant here)
[22:07] <mjg59> So many BIOSes leave a hole between 3 and 4GB and put the PCI stuff in there, and then map the RAM up >4GB
[22:07] <mjg59> Which means you need CONFIG_HIGHMEM64G to get at it
[22:08] <crimsun> fun