[00:50] I don't get this.. I copy the config from /boot/config-2.6.32-9-generic and try to build the kernel with that exact config and I still get a kernel panic when booting :S [00:50] and now I know why... update-grub [00:57] still hates me... I wonder what I'm missing :S [01:01] maybe make clean; reset --hard origin/master; checkout; remake config; make....... [01:35] For a custom kernel, can I use "make xconfig" to change the kernel config or do I have to wade through all the config variants and change things manually? [01:35] mozmck: make xconfig will do just fine. [01:36] make xconfig doesn't seem worth the deps; xconfig is nice :) [01:37] * hyperair uses menuconfig [01:37] I'm using the git repo for karmic. So I can do make xconfig, and then run debian/rules to build packages? What args do I give debian/rules? [01:38] s/xconfig is nice/menuconfig is nice/ [01:38] I want anything other than trying to hand edit the files! [01:41] mozmck: nobody hand edits the files [01:41] mozmck: when I compile my own kernel I grab git repo, copy the config from /boot/, make menuconfig, configure things, save/exit, make -j3; make all modules_install install [01:42] you should probably do make oldconfig prior to make menuconfig [01:42] ok. the instructions I've found so far seem to indicate that some do. [01:43] the only down side to how I do it is there's no .deb to install and it won't be handled via apt [01:43] ah, just what I was going to ask. [01:43] I need to end up with .debs [01:44] * hyperair has a convenient alias for building kernels with make-kpkg [01:44] which ends up with .debs [01:44] so I need to some how run debian/rules I guess. [01:44] if you have kernel-package you can use make-kpkg [01:44] it'll do the debian/rules fancy things [01:44] I've been told make-kpkg is heresy on ubuntu these days :) [01:44] hmph. [01:44] i'm old fashioned [01:44] where's said tutorial? [01:45] http://blog.avirtualhome.com/2009/11/03/how-to-compile-a-kernel-for-ubuntu-karmic/ [01:45] I learned compiling kernels in Gentoo so using make is easiest for me [01:45] lucid has a few things going on that are throwing me for a loop though [01:46] I notice the configs say that were generated with splitconfig.pl [01:46] huh [01:46] I wonder if I can run that or something on a config and create a new flavour... [01:46] mozmck: that tutorial looks quite a bit more complicated than my make-kpkg way... =\ [01:47] ya, that looks ugly [01:47] hyperair: what's your method? [01:47] I normally use make-kpkg (I came from using debian for years), but there were a couple of things that were not quite the same as the official packages that I had to try and fix. [01:48] MTeck-Linux: alias buildkernel='AUTOBUILD=1 CONCURRENCY_LEVEL=${CONCURRENCY_LEVEL:-2} ionice -c 3 schedtool -D -e make-kpkg --initrd --rootcmd=fakeroot --append-to-version=-hyper${KERNELREV:-1} kernel_{headers,image}' [01:48] yeah, it looks like the ubuntu way is harder than any other! [01:51] hyperair: ya, that's massive; not exactly too complicated... you wind up with just a nice pretty .deb with that? [01:52] MTeck-Linux: yep. [01:52] mozmck: i've got custom thinsg added. [01:53] mozmck: what you need is just make-kpkg --rootcmd=fakeroot --append-to-version=-your-own-custom-version-postfix-N kernel_headers kernel_image [01:53] hyperair: single core? [01:53] you can do away with your --append-to-version if you don't really care [01:53] MTeck-Linux: dual core. [01:53] hyperair: that's what I was doing. [01:53] mozmck: then just stick with that =) [01:53] hyperair: why not -3? [01:53] I may do that! [01:54] MTeck-Linux: i use BFS. [01:54] teh brain fuck scheduler [01:54] -j2 works faster on BFS than -j3 on CFS [01:54] and doesn't slow down my system while compiling. =) [01:54] I always use make -j3; I'm guessing that's the same variable [01:54] oh [01:59] that's the CONCURRENCY_LEVEL [02:00] i basically have the whole bunch of ${var:-somevalue} so that i can override it if i want to, or keep it if i want to [03:19] So, what are the staging drivers for? [03:19] dhillon-v10: how's it going? [03:20] A couple of things, but mainly drivers that aren't quite mainline material yet. [03:20] MTeck-Linux, hey what's up :) [03:21] Having them in the mainline repostiory makes it easier to test & gets them more eyeballs, and hopefully more fixes. [03:21] RAOF: what are the chances my system actually uses those drivers? [03:21] Moderate; some wifi drivers are sitting in staging at the moment, and nouveau's in there. [03:22] It's quite possible that a driver from staging could drive some of your hardware. [03:22] dhillon-v10: not too much, trying to compile a lucid kernel which seems to be very hard compared to karmic; also thinking of entirely dropping the locomap project [03:23] MTeck-Linux, come one let's not drop the project it seems to be a nice initiative :) I do understand if you don't have enough time [03:23] RAOF: thanks again :) [03:23] dhillon-v10: It's not just lack of time, there's also a large lack of interest [03:24] RAOF, there is this project going on at Sun called starfish, which has like a template for making device drivers and it makes a *really* nice templates. Is is possible to take that project and change the template to make templates for linux drivers [03:24] MTeck-Linux, really, I think its pretty nice, its just that I haven't had proper time to contribute because of the tests and exams coming up [03:25] dhillon-v10: Probably, but you're asking the wrong person; I'm not proficient in linux driver writing, just packaging. [03:32] I just noticed this - If you are creating a new file, it is helpful to run it through cleanfile and/or Linent before creating a patch [03:32] Shouldn't that me linten? [03:34] nah, nvm - idk where my head is [04:01] can someone help narrow down the issue with this? -- http://paste2.org/p/600785 === mcasadevall is now known as NCommander [04:27] hurray, no matter what I do I get this... mountall main process (477) killed by SEGV [04:27] something like that [04:29] my computer hates the living daylights out of me [04:33] it's complaining that mountall won't finish and drops to a recovery shell; I mount everything myself and there's not issues. I run fsck with no issues. I do that then run the mountall command and it finished booting without any issues..... [04:33] lost [04:43] I wonder if there's something funky that's happening because I used clone --reference [04:44] MTeck-Linux: if that's Lucid, the mountall symptom is known and being debugged [04:44] MTeck-Linux, there's this great video on youtube that I saw the other day about compiling the kernel, wanna have a look at that? [04:45] crimsun: oh..... THANK YOU! I though I was turning into an idiot [04:46] dhillon-v10: no thanks, I've done this a lot; I just couldn't figure out what was goig on [04:46] MTeck-Linux, okay :) [04:46] crimsun: I'm on the mailing list so I suppose I'll actually know about these issues now and when they're fixed :) [04:47] as of this morning I subscribed :) [05:06] This might be the wrong place; but can you explain this to me? http://paste.ubuntu.com/353796/ [05:06] cp -R; the source should be the same size as the dest [05:16] that's weird... do it again and it shows that it's 1002MB; wait a while do it again and it's 1.1GB === Hellow_ is now known as Hellow [07:25] can someone help narrow down the issue with this? -- http://paste2.org/p/600785 === Hellow_ is now known as Hellow [08:32] Hi. There is a regression bug in linux kernel (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/446575). There is also a patch fixing it. What I need to do to get this fix included in karmic-updates? [08:32] Malone bug 446575 in linux "dvb-t on asustek p7131 hybrid looks like poor quality signal" [Undecided,Confirmed] === MTeck-Linux is now known as MTeck-ricer === Whoopie_ is now known as Whoopie [13:15] Hi. There is a regression bug in linux kernel (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/446575). There is also a patch fixing it. What I need to do to get this fix included in karmic-updates? [13:15] Malone bug 446575 in linux "dvb-t on asustek p7131 hybrid looks like poor quality signal" [Undecided,Confirmed] === yofel_ is now known as yofel [16:18] sorry if this is the wrong place to ask, but ... if I write a 1 byte file to a 4k block size file system, then in another process read that 1 byte file, does the kernel have to read 512 bytes from the hard disk (HD block size) or 4k bytes from the hard disk (FS block size) ? === asac_ is now known as asac === SEJeff_ is now known as SEJeff_home === mcasadevall is now known as NCommander [23:42] hello everyone, i think I've found a typo in linux-source-2.6.32.10.14 but I want to confirm it... if anyone have some time for me please write [23:42] or you could just point out the typo :) [23:43] ok, in file ubuntu/omnibook/Makefile there is: [23:43] Line 161:EXTRA_LDFLAGS += $(PWD)/ubuntu/omnibook/sections.lds [23:43] and it should be EXTRA_LDFLAGS += $(PWD) ubuntu/omnibook/sections.lds IMHO [23:45] panaut0lordv: Why do you think it's broken now? The second line doesn't look correct to me. [23:46] well, it wasn't compiling [23:46] cause path was absolute [23:46] at least on my machine [23:47] Want to pastebin the error (and surrounding context)? [23:56] RAOF: here you are, maybe I am simplyfying but for me no errors mean fixed xD http://pastebin.com/m762e87e1 [23:59] I wonder what sets PWD in the normal build.