[12:12] <jbailey> Ah, missed him.
[12:13] <jbailey> I suspect that he's feeding a different root= in grub than he thinks he is.
[02:35] <zul> evening
[02:35] <zul> so...got a question when are you guys gonig to be arriving in montreal? i say we get together for dinner on saturday nightish
[06:08] <iram> Hello all, I'm trying to figure out what the ubuntu specific patches to the kernel are.  I'm need to use kernel 2.6.13 and I'm just not sure what the reprocussions of using a virgin kernel are.  I've looked around a bunch trying to find this information but I can't find it anywhere.  Can anyone give me some pointers on where to find this.
[06:45] <crimsun> iram: install linux-patch-ubuntu-2.6.12
[06:46] <crimsun> iram: if you need to use 2.6.13 for a specific driver, feature, or fix, you may well find it easier to backport it
[07:07] <iram> crisun: what I am trying to get is the ite 8212 driver that is in the 2.6.13 kernel.  I am currently using the driver on a fedora installation but they kernel team for fedora has included it since 2.6.12.  I read the release notes from 2.6.13 and it is now included in the stock kernel (after being tested in fedora).  How would I go about 'backporting' this specific chunk of code.  The other issue is that apparently 
[07:07] <iram> the ide code has been changed a lot in 2.6.13 so it may be kind of hard to just yank that code
[02:31] <typo> coud someone please explain to me how I can create a package just like "linux-image-2.6.12-9-686"?
[02:33] <mjg59> typo: apt-get source linux-image-2.6.12-9-686; cd linux-source-2.6.12-2.6.12; dpkg-buildpackage -rfakeroot
[02:34] <typo> mjg59: great, thanks
[02:38] <typo> mjg59: it doesn't have a .config, is that normal?
[02:38] <mjg59> typo: That's generated when you do the dpkg-buildpackage
[02:38] <mjg59> It builds several different kernels
[02:38] <typo> mjg59: oh, great, and how can I enable ACPI debug?
[02:39] <mjg59> The configs are in debian/config
[02:40] <typo> can I make it build just the -686 kernel?
[02:43] <mjg59> Easiest way is just to delete all the configs except the 686 one
[02:43] <typo> will it try to build the kernels for the other arches?
[02:50] <mjg59> Nope
[03:14] <typo> how can I add an extra version to the packages so that they're parallel instalable?
[03:14] <typo> with the ones in breezy I mean
[03:19] <BenC> it's non-trivial
[03:21] <typo> ok, no problem then
[03:45] <zul> hmmm?
[03:47] <typo> the compile failed with: "drivers/acpi/i2c-acpi-ec.c:322: error: `ACPI_EC_COMPONENT' undeclared (first use in this function)"
[03:47] <typo> anyone know what this is?
[03:47] <BenC> yeah, means you have conflicting patches
[03:47] <typo> all I changed from the original was adding CONFIG_ACPI_DEBUG=y to the config
[03:47] <BenC> ah
[03:47] <typo> BenC: how? I didn't patch anything
[03:48] <BenC> well, could be that the debug code isn't well tested
[03:49] <typo> seems to be the "ACPI_FUNCTION_TRACE" macro
[03:49] <typo> anyone know where this is defined?
[03:51] <zul> grep it
[03:52] <zul> BenC: is there something in the git tree now or am i just crazy?
[03:52] <typo> zul: should I try .h or .c?
[03:53] <typo> found it: include/acpi/acmacros.h
[03:53] <BenC> zul: there is, but after doing a clone, do "git checkout ubuntu-2.6.14"
[03:54] <BenC> I know it's counter to what I wanted to do, but all work will be in a branch :)
[03:55] <typo> yes, the macro is defined to empty when debug is off
[04:00] <typo> so how can I build a kernel with acpi debug?
[04:01] <BenC> just remove that one ACPI_FUNCTION_TRACE line
[04:01] <BenC> or comment it out
[04:02] <typo> BenC: it probably happens in all of those lines
[04:02] <typo> or not?
[04:03] <typo> oh, that file doesn't exist before the ubuntu patches
[04:05] <typo> for a centrino laptop should I use 386 or 686?
[04:05] <typo> for testing I mean
[04:08] <mjg59> 686
[04:09] <typo> ok
[04:09] <typo> and how do I remove a patch from the build?
[04:13] <BenC> you don't remove the patch
[04:15] <BenC> if ACPI_FUNCTION_TRACE is broken, then you're looking at a lot of work, really
[04:15] <typo> BenC: maybe it's just in the i2c patch
[04:15] <typo> removing it is that hard?
[04:16] <BenC> possible
[04:16] <BenC> no, just edit debian/patches/00list-9.23
[04:16] <BenC> first do "fakeroot debian/rules clean"
[04:18] <BenC> git's shared object directories are cool
[04:27] <typo> if I remove that one then the "drivers-acpi-add-devacpi" patch fails
[04:28] <fabbione> hey BenC 
[04:28] <typo> I was trying to debug this: "https://bugzilla.ubuntu.com/show_bug.cgi?id=8140" but I guess it's not to be
[04:28] <BenC> which one did you remove?
[04:29] <BenC> hey fabbione
[04:29] <fabbione> BenC: pulling your tree from kernel.org is quite of a puzzle :)
[04:29] <fabbione> davem 1) Clone Linus's tree
[04:29] <fabbione> davem 2) Clone locally, from that, a tree named "ubuntu-2.6" or whatever
[04:29] <fabbione> davem 3) In the ubuntu-2.6 tree, pull from Ben's with the "ubuntu-2.6" tag as the last arg
[04:29] <fabbione> otherwise we get plenty of errors
[04:29] <BenC> fabbione: that's not how I did it
[04:30] <fabbione> BenC: that's what we had to do :)
[04:30] <typo> BenC: external-drivers-acpi_i2c-acpi-ec.dpatch
[04:30] <fabbione> what's another solution?
[04:31] <BenC> do the command in the topic and then do "git checkout ubuntu-2.6.14"
[04:32] <BenC> I created the tree just like Linus told me
[04:32] <fabbione> ok
[04:32] <BenC> so if it's wrong, blame Linus :)
[04:32] <fabbione> OHH
[04:33] <fabbione> i think it's because i missed the tag at the end
[04:34] <fabbione> BenC: we need to do a git pkg for Ubuntu
[04:34] <fabbione> the one we have inside cogito is old
[04:34] <typo> BenC: any other idea? to do this or to debug it some other way?
[04:34] <fabbione> anyway i need to run off now...
[04:37] <BenC> fabbione: I guess I'll be nominated to maintain git since I'm forcing people to use it :)
[04:38] <fabbione> BenC: nah. i will prepare the pkgs
[04:38] <fabbione> it's okl
[04:38] <fabbione> ok
[04:40] <jbailey> Isn't git in universe? =)
[04:44] <fabbione> jbailey: cogito is
[05:21] <dilinger> isn't git as well?  istr git being broken out of cogito
[05:25] <dilinger> yuck.  cogito's changelog is a bit scary; binaries being removed, readded, renamed...
[05:26] <dilinger> the debian package is also horribly out of date
[05:28] <dilinger> bah
[05:28] <dilinger> i know what i'm doing today
[06:26] <typo> BenC: I now have this: "ABI has changed!  Refusing to continue; please update the ABINAME accordingly."
[06:26] <typo> how do I update ABINAME?
[06:34] <typo> anyone know?
[06:35] <dilinger> the ABINAME is in the package name
[06:36] <typo> and how do I change it?
[06:37] <dilinger> by changing the control file, i would assume
[06:37] <dilinger> linux-image-2.6.12-9-386
[06:37] <dilinger> the ABINAME there is '9'
[06:37] <typo> and that's where I change it?
[06:37] <dilinger> i'm not certain where else it might need to be changed, but debian/control is a good start
[06:38] <typo> I'm guessing debian/abi/ has something to do with it
[06:39] <typo> disabling the abi check would also work
[06:39] <typo> anyone know how to do that?
[06:42] <typo> from debian/rules I'm guessing touching debian/abi/i386.ignore will do that
[07:11] <BenC> typo: debian/rules bumpabi
[07:12] <BenC> and also the ignore will make it continue
[07:12] <BenC> bumabi will change the abi name to -10
[07:54] <fabbione> BenC: that's weird..
[07:54] <fabbione> i did the git pull and git checkout from people
[07:54] <fabbione> but there is no debian/ ?
[07:55] <fabbione> i was on git web that you commited it
[08:07] <BenC> fabbione: git checkout ubuntu-2.6.14
[08:08] <fabbione> ahh
[08:08] <fabbione> so everything needs to be presented with the branch tag
[08:08] <fabbione> thanks :)
[08:09] <BenC> once you checkout with the branchname, you are good
[08:09] <fabbione> yup
[08:10] <BenC> then everything will default to that branch
[08:10] <fabbione> ah perfect
[08:10] <fabbione> so in future i will just git pull to update the tree
[08:10] <BenC> right
[08:10] <fabbione> i will need to figure later how to commit/publish the brach without accessing your repo directly
[08:11] <fabbione> but that's monday task :)
[08:11] <fabbione> BenC: are you using a cronjob to sync with kernel.org?
[08:11] <BenC> commit is local, then you can push it anywhere you want, but back to my repo is best
[08:11] <fabbione> or do you need to do it manually?
[08:11] <BenC> not currently
[08:11] <fabbione> ok
[08:11] <BenC> right now, I am manually syncing to kernel.org so that I can get some sort of checking for commits
[08:11] <fabbione> hmm did you create the repo with the g+s so i can commit?
[08:12] <fabbione> because otherwise that can be an issue
[08:12] <BenC> should be, let me check
[08:12] <BenC> what sort of stuff are you going ot be working on?
[08:13] <fabbione> right now i am only setting up the enviroment
[08:13] <BenC> ok
[08:13] <fabbione> but i will need to commit stuff for cluster and ocfs2
[08:13] <dilinger> bah
[08:13] <fabbione> dilinger: yo
[08:13] <dilinger> cogito upstream has their own debian stuff
[08:13] <fabbione> dilinger: i will make clean gi pkgs next week
[08:13] <fabbione> dilinger: no cogito
[08:13] <fabbione> git even
[08:13] <BenC> keep all your work local until I write up the commit details (log style, etc)
[08:14] <fabbione> BenC: i am not going to do anything without your approval, while we test the bits :)
[08:14] <fabbione> also because at the best, the kernel doesn't even compile
[08:14] <dilinger> fabbione: well,i'm looking at cogito now
[08:14] <BenC> in order to keep everything clean with this commit-all stuff, we need good logging
[08:14] <fabbione> BenC: yup
[08:14] <dilinger> since i want to build bzrweb and update gitweb, might as well
[08:14] <fabbione> dilinger: most kernel hackers don't use cogito for different reasons
[08:14] <BenC> what is cogito?
[08:14] <BenC> as opposed to git
[08:15] <fabbione> BenC: cogito is an simplified UI for git
[08:15] <fabbione> BenC: cogito uses git as backend
[08:15] <fabbione> nothing fancy
[08:15] <BenC> ah
[08:15] <BenC> makes it easier for cvs folks, I assume?
[08:15] <fabbione> the real issue is that cogito Depends: git (= $special_version)
[08:15] <fabbione> when git moves forward much faster
[08:16] <BenC> so far I like git
[08:16] <fabbione> BenC: yeah sort of
[08:16] <BenC> one thing I like is that I have one main linux-2.6 repo locally, and my ubuntu and linux1394 repo's share the majority of hte objects with it, and it saves space
[08:16] <dilinger> fabbione: what reasons are that?
[08:16] <fabbione> dilinger: the missing allignment is the first.. 
[08:17] <fabbione> dilinger: that creates problems for them to keep in sync
[08:18] <fabbione> anyway
[08:18] <fabbione> gotta go
[08:18] <fabbione> cya sunday
[08:18] <fabbione> or later
[08:30] <dilinger> or maybe i should drop gitweb (can't find bzrweb anymore), and break out the code in tailor for reading distribution RCSs
[08:34] <dilinger> http://nautilus.homeip.net/~lele/
[08:34] <dilinger> that's a little disturbing
[08:56] <typo> BenC: it failed because it was expecting to find "debian/build/linux-image-2.6.12-9-386_*i386.deb" and there was a 686 image there instead
[08:56] <typo> I guess deleting the 386 config wasn't a good idea after all
[09:36] <BenC> if it got that far, then it might be done
[09:37] <BenC> is there a .deb image in the directory below your build?
[10:55] <typo> BenC: yes, there's a deb here, I'm going to try this one
[11:01] <typo> BenC: It worked, I got a 2.6.12-9-686 kernel with ACPI debug on
[11:02] <typo> BenC: it hangs after "Executing all Device _STA and _INI methods:........."
[11:05] <typo> BenC: if you need the other stuff I can write it all down
[11:06] <typo> BenC: anything else please ask
[11:06] <typo> BenC: the bug is here: https://bugzilla.ubuntu.com/show_bug.cgi?id=8140