[10:28] <fabbione> BenC: please pull from my repo on people: changes are only redhat-cluster-suite related.
[01:19] <BenC> fabbione: ok
[02:59] <fabbione> BenC: thanks
[03:07] <fabbione> BenC: ping?
[03:16] <zul> heylo
[03:17] <BenC> fabbione: pong
[03:18] <fabbione> BenC: yo
[03:18] <fabbione> BenC: is there any reason why ia64 has no OCFS2?
[03:18] <fabbione> also.. for i386/amd64 it should probably move to -server only
[03:18] <fabbione> it doesn't make much sense to have it in desktop :)
[03:24] <fabbione> BenC: note that OCFS2 pulls in CONFIGFS
[03:24] <fabbione> so they need to move together
[03:24] <fabbione> i did try to play with the config script but it did mess up here
[03:24] <fabbione> so i wasn't sure how to make it properly
[03:24] <fabbione> well.. other than with vi
[03:58] <BenC> fabbione: I'll enable it for ia64 (not sure why it isn't), and get things moved to just -server on i386 and amd64
[03:58] <BenC> OCFS2 needs to move, what about cluster aswell?
[03:58] <fabbione> BenC: you rock!
[03:58] <fabbione> cluster too
[03:58] <fabbione> cluster moves together with GFS
[03:59] <fabbione> as it is now, for me it's more important to get a coherent config where i can install one kernel on each arch and get GFS and OCFS2
[03:59] <fabbione> we can sort them at the sprint in details
[04:01] <BenC> ok, so all desktop kernels (in i386 and amd64 where we have server kernels) should have GFS, CLUSTER and OCFS2 disabled
[04:02] <BenC> is CONFIGFS needed in desktop after that?
[04:02] <BenC> it may end up being needed by a thirdparty module, so maybe just leave it there
[04:03] <fabbione> as it is now, the only client for CONFIGFS is OCFS2
[04:03] <fabbione> i think what you suggest make sense afterall
[04:07] <BenC> ok, done
[04:07] <BenC> btw, benh sent me a patch to support dual-core G5's and iMac G5 with isight
[04:07] <BenC> backported from 2.6.16-git
[04:13] <fabbione> ah cool
[04:13] <fabbione> Mark will like that :)
[04:16] <zul> whats the url for live cds again
[04:17] <fabbione> cdimage.u.c/
[04:18] <zul> thanks i knew it was something like that it just escaped me ofr  a minute
[04:21] <BenC> is it hoary -> warty -> breezy, or warty -> hoary -> breezy?
[04:22] <fabbione> the latter
[04:22] <BenC> ok, thanks
[04:22] <fabbione> warty -> hoary -> breezy
[04:22] <fabbione> whorey -> (w)horey -> 
[04:23] <BenC> starting dapper test builds, and working on security for those three while it's going
[04:23] <BenC> hehe
[04:23] <fabbione> ehhe
[04:23] <fabbione> BenC: i have the old breezy binutils in my ~ on chinstrap
[04:23] <fabbione> i hope i can manage to do to test kernels tomorrow
[04:23] <fabbione> otherwise it will be when i am back
[04:24] <BenC> bcollins@grayson:~$ build-start-all
[04:24] <BenC> green: build started
[04:24] <BenC> emucade: build started
[04:24] <BenC> zachery: build started
[04:24] <BenC> frag: build started
[04:24] <BenC> hippo: build started
[04:24] <fabbione> nice
[04:24] <fabbione> isn't missing one?
[04:24] <fabbione> 6 arches...
[04:24] <BenC> I don't have amd64, and I haven't setup  ronne yet
[04:24] <fabbione> ah ok
[04:25] <BenC> those are ppc, i386, sparc64, ia64, hppa, in that order
[04:25] <fabbione> i think you should push i386 to ronne, ppc to davis and ia64 to halley
[04:25] <BenC> sparc64 always finishes first :)
[04:25] <fabbione> and keep only sparc64 and hppa
[04:26] <fabbione> speaking of zachery..it did crash on me twice building OOo2
[04:26] <BenC> I build them locally, since I also test boot them when I'm done, and doing 80-100 megs download after it's finished doesn't play well with my satellite FAP crap
[04:26] <fabbione> hmm right
[04:26] <fabbione> but on the other side we should publish these images at some point
[04:27] <BenC> all of them builds pretty quick, except i386
[04:27] <zul> BenC: ooh share the script..
[04:27] <BenC> these are going to be 2.6.15-12, so they'll get published within a day :)
[04:27] <fabbione> ehhe
[04:27] <BenC> zul: it's just a couple of shell scripts, really simple
[04:28] <BenC> starts a screened build on each machine
[04:28] <BenC> I can tar them up at some point
[04:28] <zul> yeah but i want to incorporate yours with mine..;)
[04:28] <fabbione> zul: let's do the other way
[04:28] <fabbione> give us your
[04:28] <fabbione> and we will integrate into ours :)
[04:28] <zul> sure
[04:28] <BenC> build-{prep,start,status,stop,watch,attach}{,-all}
[04:28] <BenC> that's all of them
[04:29] <BenC> I can do "build-prep green" to clean the build, and push git, and build-watch just tails, build-attach attaches to the screen session
[04:29] <BenC> watch,attach don't have a -all obviously
[04:31] <BenC> build-prep also checkes for local changes that I might have done on the build system
[04:31] <BenC> keeps me from losing fixes :)
[04:31] <fabbione> ehhe
[04:32] <zul> right now i have 2 scripts..
[04:35] <BenC> none of mine are any longer than 10 lines, but I do have a do_build script on each machine so that it will set the right concurrency and stuff (still less than 10 lines)
[04:39] <zul> first one reads a config file to pull from and the second scripts does the build, rips out the arches that i dont want
[04:39] <fabbione> later guys
[04:39] <BenC> later fabio
[04:40] <fabbione> BenC: meh i forgot to add stuff to debian/changelog
[04:40] <fabbione> do you want to do that for me?
[04:40] <fabbione> it's just the update for the redhat cluster suite
[04:40] <fabbione> to 20060112 CVS
[04:40] <fabbione> nothing fancy
[04:40] <fabbione> gotta run to school
[04:40] <BenC> ok
[04:40] <fabbione> cheers mate
[05:45] <BenC> fabbione: ping
[06:40] <Seveas> --- Topic for #ubuntu-kernel is Ubuntu kernel development discussion ONLY
[06:40] <Seveas> is a question about the kernels used on Ubuntu, which is not directly related to kernel development OK too?
[06:52] <Mithrandir> Seveas: don't ask to ask
[06:53] <Seveas> Mithrandir, hehe, y'never know how strict they are in here
[06:55] <Seveas> ok, here's the deal: somewhere between 2.6.10 and 2.6.12 mmap changed and now returns 'random' addresses (ie cat /proc/self/maps is always the same in 2.6.10 and always different in 2.6.12). Can someone point me to where this change is documented? 
[06:55] <Seveas> It severely ruins my graduation project and the things I've been working on during the last 15 months
[07:02] <zul> you can probably do a git bisect but dont ask me how to do one
[07:04] <dilinger> http://www.kernel.org/pub/linux/kernel/v2.6/testing/ChangeLog-2.6.12-rc1
[07:04] <dilinger> i assume what you're looking for is "mmap rand"
[07:05] <Seveas> dilinger, yes, thank you!
[07:05] <dilinger> np
[07:06] <Seveas> will prelink also make sure ld.so is always in the same location in memory?
[07:07] <dilinger> the comment seems to imply that prelinked binaries will not have randomized mmap base addresses
[07:08] <dilinger> however, the ADDR_NO_RANDOMIZE thing is probably what you'd want if you're looking to disable it
[07:08] <Seveas> then I would have to prelink ld.so itself
[07:08] <Seveas> hmm, then I'll have to look up what personalities are
[07:08] <dilinger> are you seeing ld.so at random addresses?
[07:09] <Seveas> yes
[07:09] <Seveas> and my application kind-of depends on the fact that that does not happen
[07:09] <dilinger> i'm not sure if you can PRELINK ld.so, but i guess it couldn't hurt to try
[07:11] <Seveas> hmm, how do personalities work in debian/ubuntu? apt-file search setarch returns only something completely unrelated
[07:43] <zul> ds: which laptop?
[07:44] <ds> EMachines M5312
[08:14] <ds> zul: do you happen to know if this is a kernel bug, or just a ACPI stupidity that needs to be worked around?
[08:15] <zul> i dont get the same with my laptop you might want to open a bug..
[08:15] <zul> or just acpi stupidity
[08:15] <dilinger> Seveas: *shrug* :)
[08:18] <cjb> ds: The kernel doesn't keep an exhaustive list of ACPI weirdnesses.  I'd say it's unlikely to be a kernel bug, but go ahead and file it.
[08:20] <mjg59> ds: Windows doesn't use that information, so some machines just lie
[08:26] <cjb> Coo, it doesn't?  "Let me guess, it handles lid closing in the video driver."
[08:27] <mjg59> Oh, it picks up the lid events, but yeah - the video driver is responsible for what happens next
[08:34] <Seveas> hmm
[08:34] <Seveas> kinux-kernel-headers is at 2.6.11 on breezy
[08:34] <mjg59> linux-kernel-headers
[08:34] <Seveas> yes :)
[08:35] <Seveas> but it's still weird that it's at 2.6.11
 however, the ADDR_NO_RANDOMIZE thing is probably what you'd want if you're looking to disable it <-- yes, that worked perfectly!
[08:39] <Seveas> dilinger, muchos gracias, you just saved my graduation project :)
[08:39] <zul> dilinger: arent you special ;)
[08:40] <Seveas> more importantly: you saved me having to run my experiments on a crappy old red hat cluster with linux 2.4.1, Now I can continue at the spiffy new sarge HPC cluster :)
[08:40] <Seveas> saved me from*
[08:50] <dilinger> hehe
[08:57] <dilinger> one of these days, i should get myself one of those college degree thingies
[08:57] <zul> you can probably buy one 
[09:00] <Seveas> I have several offers in my mailbox if you're interested ;)
[09:00] <Seveas> they come included with fake rolex and penis enlargements 
[09:34] <dilinger> BenC: my current employer was pretty surpised that i didn't have a BS
[09:34] <dilinger> but that didn't stop them from hiring me
[09:35] <BenC> the only time I was ever asked about it was a guy from Sun that I sent my resume to about 4 years ago
[09:36] <BenC> he actually called me, to tell me that I was wasting my time with "all this open source stuff", that I should switch to applications development instead of embedded/kernel work, and that he pretty much wouldn't hire me, even at a measly $24k/year (when I was making way more than that already), and that I better go get a degree from a major college (not just a degree, but an ivey league degree), else I was doomed to be a loser
[09:36] <dilinger> i'm planning on going back to school soon, but it won't be for CS
[09:36] <dilinger> i'm going to do it for the right reasons this time: girls.
[09:36] <mjg59> dilinger: An excellent plan
[09:37] <BenC> and I'm not kidding about any of that, it's exactly how he put it
[09:37] <mjg59> BenC: Aw GROUP HUG
[09:37] <dilinger> going to school thinking i'd get an education was a huge letdown
[09:37] <BenC> dilinger: that's my only regret about not going to college :)
[09:38] <dilinger> BenC: i would've hung up on him if he started telling me that
[09:38] <BenC> mjg59: see, open source is good...I have a support group :)
[09:38] <mjg59> PhDs are great in that respect
[09:38] <dilinger> "stop wasting my time.  *click*"
[09:38] <mjg59> Actualy, it might just be mine
[09:38] <BenC> you have a PhD?
[09:38] <mjg59> It's my day job
[09:38] <BenC> now PhD's are something to brag about
[09:38] <mjg59> I'll be finished in about a year with a bit of luck
[09:39] <mjg59> But it means I get to be old and mature and have a larger disposable income while surrounded by undergrads
[09:39] <BenC> most CS BS degree holders that I met outside of open source, are just a joke
[09:39] <mjg59> Oh, I'm doing genetics, not CS
[09:39] <BenC> mjg59: undergrads :)
[09:40] <BenC> genetics, that's an interesting departure from ACPI :)
[09:40] <mjg59> I find ACPI an interestig departure from genetics
[09:40] <mjg59> It makes more sense than fruitflies do
[09:40] <BenC> lol
[09:40] <BenC> what is your thesis on?
[09:40] <mjg59> No matter how much crack somebody was on when they designed a specification, it makes more sense than real life does
[09:41] <mjg59> Computational locating of RNA elements involved in gene localisation
[09:41] <dilinger> mjg59: at my prior job, i had to do a lot of interviews.  people w/ CS degrees were a dime a dozen, and in most cases weren't worth hiring.  there were also people w/out degrees, but w/ more real world experience, that seemed a lot more valuable
[09:41] <dilinger> er, s/mjg59/BenC/
[09:42] <BenC> dilinger: one good thing about open source, you don't need to ask for paper to find out if someone is clueful, just search google
[09:43] <dilinger> BenC: i need to do something about my blog coming up as the first entry when you search for my name :)
[09:43] <kozz> I have tried to run mkvmlinuz with the 2.6.15 kernel, but there seems to be some files missing
[09:43] <BenC> heh, my old DPL candidacy letter comes up when you search mine name...I'm embarrased to see it anymore :)
[09:43] <kozz> for 2.6.12 they lie in /usr/lib/linux-image-2.6.12-10-powerpc
[09:43] <kozz> some object files
[09:44] <BenC> kozz: yeah, the build process didn't include them anymore, but I didn't think we had any oldworld users (since we didn't even support oldworld install)
[09:44] <kozz> BenC: it's for Pegasos
[09:44] <BenC> is pegasos CONFIG_PPC_MAPLE?
[09:44] <kozz> no, CHRP
[09:45] <mjg59> Pegasos is CONFIG_PPC_CRACK
[09:45] <BenC> ah, good, MAPLE is going away
[09:45] <mjg59> But yeah, it's basically CHRP
[09:45] <BenC> kozz: sorry, I'm having a hard time keeping non-standard (e.g. non-powermac) support because of all the ARCH=powerpc changes in the kernel
[09:45] <kozz> mjg59: basically?
[09:46] <BenC> but anyone with such a system is more than welcome to send me patches to get the build to do the right thing
[09:46] <mjg59> kozz: Some of the interrupt setup stuff seemed a bit funny
[09:46] <kozz> BenC: I see, I have tried to create them but hasn't managed to do it
[09:46] <mjg59> It never booted with the stock kernel CHRP support
[09:47] <kozz> stock kernel CHRP? it has always worked to boot the zImage.chrp kernel out of the box before
[09:48] <mjg59> kozz: With kernel.org source?
[09:48] <kozz> yes
[09:48] <mjg59> Uhm.
[09:48] <kozz> supported since 2.6.11
[09:48] <mjg59> Right
[09:48] <mjg59> Sorry, I meant "Never used to boot"
[09:48] <mjg59> It needed some specific patches because some of the setup code is different
[09:49] <kozz> yes I know
[09:49] <kozz> but that is not the case anymore ;)
[09:49] <mjg59> Right
[09:49] <mjg59> But it always seemed "mostly CHRP" rather than "CHRP"
[09:49] <kozz> right, I don't know
[09:50] <kozz> hmm, now BenC left
[09:51] <kozz> but how did it work before to create all those files, for the 2.6.12 kernel I mean
[10:06] <fabbione> BenC: pong
[10:29] <BenC> fabbione: nm
[10:30] <fabbione> BenC: ok :)
[10:30] <BenC> I have breezy/hoary/warty ready, just waiting on the last CVE from pitti
[10:30] <fabbione> BenC: ok cool
[10:31] <BenC> green: building
[10:31] <BenC>     building: powerpc-smp(modules)
[10:31] <BenC>     unbuilt : powerpc
[10:31] <BenC>     unbuilt : powerpc64-smp
[10:31] <BenC> emucade: building
[10:31] <BenC>     building: server-bigiron(modules)
[10:31] <BenC>     unbuilt : server k7 686 386
[10:31] <BenC>     unbuilt :
[10:31] <BenC> zachery: idle
[10:31] <BenC> frag: idle
[10:31] <BenC> hippo: idle
[10:31] <fabbione> nice
[10:31] <BenC> first unbuilt should be built, but no matter
[10:31] <fabbione> oh i see
[11:26] <hile> I just filed a trivial bug for 2.6.15-x kernels, #22351 
[11:26] <hile> just wanted to ask if there is any reason _not_ set number of UARTs from 4 to 8 (see bug for details)
[11:44] <makx> you don't want to spam the user with useless tty's
[11:44] <makx> post 2.6.15 it's a boot variable
[11:46] <hile> already with current builds, or later? ok, I'll close the bug then
[11:52] <hile> thanks anyway ;)