[12:35] <jbailey> sistpoty: I don't know about in-kernel stuff, sorry. =)
[12:36] <sistpoty> jbailey: ok, thanks anyway ;)
[12:36] <jbailey> sistpoty: I'd be surprised if kernel modules were expected to use the public syscall interface.
[12:40] <sistpoty> jbailey: no, faumachine isn't only a kernel module, it's s.th. like qemu (actually uses parts of the qemu cpu) and provides a kernel-module as well. 
[12:40] <sistpoty> jbailey: but I know much less about kernel stuff... only that it doesn't compile on edgy *g*
[12:49] <ajmitch> sistpoty: I've got an example for doing it directly via ioctls 
[12:50] <ajmitch> http://paste.ubuntu-nl.org/28739/
[12:51] <sistpoty> ajmitch: nice... thx!
[12:52] <ajmitch> it's a nasty hack of course
[12:54] <sistpoty> ajmitch: it is... but better than nothing :)
[01:16] <sistpoty> gn8 everyone
[01:34] <BenC>  Ping reply from BenC: 0.21 second(s)
[06:06] <infinity> BenC: Woo, you got a real connection, finally?
[06:06] <infinity> BenC: What's that costing you?
[06:07] <BenC> infinity: there difference between this and the satellite is about $450/month
[06:07] <infinity> Ouch.
[06:07] <BenC> trust me, live with satellite for a year, and you'll pony up that extra few bucks too :)
[06:07] <infinity> Time to start selling the neighbour's cows to make ends meet?
[06:08] <BenC> I may attach wireless repeaters to the cows and sell off some b/w to neighbors :)
[06:33] <fabbione> BenC: congratulation!
[06:33] <BenC> hehe, thanks
[06:39] <fabbione> infinity: i see that the kernel is there, but there have been other upload accepted in the archive?
[06:39] <fabbione> infinity: base-files and debootstrap?
[06:40] <fabbione> (and no email to feisty-changes)
[06:41] <infinity> I have a feeling feisty-changes needs moderating to whitelist LP.
[06:42] <fabbione> ah ok
[06:42] <fabbione> why base-files and deboostrap have been approved?
[06:42] <infinity> Because I felt like it. :)
[06:42] <fabbione> aren't they supposed to come after toolchain bootstrapping?
[06:43] <fabbione> yeah right.. :P
[06:43] <infinity> base-files is arch:all, and it'd be nice for feisty systems to claim to be feisty.
[06:43] <infinity> debootstrap builds a bit of C, but the benefit of having it outweighs the "hey, everything should use the new toolchain!!" goal, IMO.
[06:44] <fabbione> hmmm ok
[06:44] <infinity> (Well, okay, base-files isn't TECHNICALLY arch:all, but it doesn't build any C)
[06:44] <infinity> Oh man, hppa is going to be a mess.
[06:44] <fabbione> we are changing binutils to use that hash thingy that's supposed to speed up stuff
[06:45] <fabbione> that's why i was kind of surprised :)
[06:45] <infinity> debhelper/dpkg/libsepol/etc/etc.. I need to remember how I bootstrapped that in early edgy. :)
[06:45] <fabbione> uh why?
[06:45] <fabbione> i don't think it's going to be a big deal
[06:45] <infinity> Define "big deal".
[06:45] <fabbione> with the new bootstrapping sequence you get the toolchain almost out of the box
[06:45] <infinity> I need to manually bootstrap the packaging toolchain, then rebuild it later.
[06:46] <fabbione> that's fun :)
[06:46] <infinity> (Note that the packaging toolchain kinda needs to be there before anything -- including the kernel -- can actually build)
[06:46] <fabbione> well bootstrap it once.. with the big merge of death they will all get rebuilt
[06:47] <fabbione> jb gave me gcc-4.1 packages for hppa..
[06:47] <fabbione> do you want me to upload them somewhere for you?
[06:47] <infinity> Yeah.  That was, more-or-less, the plan.
[06:47] <fabbione> do you want it or not?
[06:47] <fabbione> scp from here takes time.
[06:48] <infinity> May as well.
[06:48] <infinity> Though it's glibc that introduces the ABI change for hppa, so that's really what I want.
[06:48] <fabbione> glibc requires new binutils
[06:49] <fabbione> that we don't have yet
[06:49] <infinity> Yeah.  I know.
[06:49] <fabbione> i am upload gcc/make/module-init-tools
[06:49] <infinity> hppa and I are just playing a bit right now, I'm not doing any real work there.
[06:49] <infinity> I'll do the other 5.
[06:49] <fabbione> don't use the binutils in that tar
[06:49] <fabbione> they are known to be broken
[06:49] <fabbione> yeah i know
[06:49] <infinity> hppa will still not have a chroot for feisty until I'm positive of the best way to get from A to B.
[06:49] <fabbione> yeps make sense
[06:50] <infinity> (Cause a chroot for feisty means hppa will suddenly have several thousand builds queued)
[06:51] <fabbione> yes i know that :)
[06:51] <fabbione> you need to ask Malcc to prepare feisty pockets
[06:51] <fabbione> -updates and -security are not there
[07:04] <fabbione> bah i become addicted to "friends"
[07:05] <infinity> Hahahaha.
[07:14] <fabbione> chinstrap.ubuntu.com:hppa/.
[07:14] <fabbione> my home of course
[07:14] <fabbione> remember binutils there is borked
[07:56] <infinity> BenC: Still alive?
[07:57] <BenC> infinity: yeah
[07:57] <infinity> You have new build failures. :)
[07:57] <infinity> Welcome to feisty!
[07:57] <infinity> (sparc and powerpc so far)
[07:58] <infinity> I'm okay with "not caring until Monday" if you'd prefer... I think we could all use a weekend.
[07:58] <infinity> But if you're feeling like a keener...
[07:58] <fabbione> feh
[07:58] <BenC> if it's failing to build, I blame the chroot :P
[07:58] <BenC> it's been building fine for me for weeks
[07:58] <infinity> Not much in the chroot to blame.
[07:59] <fabbione> infinity: build logs?
[07:59] <BenC> yeah, give me a log to look at
[07:59] <infinity> http://librarian.launchpad.net/4938614/buildlog_ubuntu-feisty-sparc.linux-source-2.6.19_2.6.19-1.1_FAILEDTOBUILD.txt.gz
[07:59] <infinity> http://librarian.launchpad.net/4938613/buildlog_ubuntu-feisty-powerpc.linux-source-2.6.19_2.6.19-1.1_FAILEDTOBUILD.txt.gz
[07:59] <fabbione> sparc is.. special
[08:00] <fabbione> looks like bash dash breakage
[08:00] <infinity> Still using bash in the chroots.
[08:00] <fabbione> infinity: what's the default shell in feisty chroots?
[08:00] <infinity> (Thanks for the reminder, I should switch that now)
[08:00] <fabbione> BenC: ppc is knows.. try to build with linux32 and you will get it too
[08:01] <fabbione> BenC: i think i did put you in the email loop with Benh and Paulus
[08:01] <BenC> stupid linux32 :P
[08:01] <fabbione> but sparc.. is weird
[08:01] <BenC> probably need something better than defconfig
[08:01] <fabbione> BenC: my local workaround: cp ppc64_defconf ppc_defconf
[08:01] <infinity> And that doesn't mess up ppc32 kernels?
[08:02] <BenC> sparc looks like a sparc32 problem
[08:02] <fabbione> no
[08:02] <fabbione> we still use linux32 on sparc, don't we?
[08:02] <infinity> BenC: yeah, we use linux32 on sparc as well.
[08:02] <infinity> A lot of other software in the archive breaks if we don't.
[08:03] <infinity> Well, s/breaks/builds for sparc64/
[08:03] <BenC> something isn't right there
[08:03] <BenC> it looks like it is doing a ARCH=sparc config
[08:04] <BenC> I should build with linux32 on sparc and ppc I guess
[08:04] <infinity> Did we lose workarounds for this in kernel-package at some point?
[08:04] <infinity> No, wait, kernel-package hasn't changed since edgy. :)
[08:04] <infinity> This is all kernel.
[08:05] <BenC> considering it builds fine for me fabbione, I still blame the chroot ;)
[08:05] <BenC> both problems seem related to linux32, so I'll work on that over the weekend
[08:05] <fabbione> BenC: i didn't use linux32 on sparc
[08:05] <fabbione> and i can reproduce the ppc one here
[08:06] <BenC> neither did I
[08:06] <fabbione> i can test it in a few minutes. I just need to finish the ALOM upgrade
[08:07] <infinity> BenC: Thpt.  Blame the chroots or buildds all you want. :)
[08:07] <infinity> BenC: Anything that relies on uname output is broken, period.  It's just easier to fix one thing (the kernel build) than potentially dozens of other broken things.
[08:08] <BenC> infinity: Nah, I know the kernel build needs fixing, I just hate that we still have this bullshit problem
[08:09] <infinity> This reminds me that I really should be using linux32 on i386 buildds too...
[08:10] <infinity> Not that we need it now (we don't), but I'm sure a day will come when elmo will hand me an X86_64 box as an i386 buildd.
[08:14] <infinity> Anyhow, amd64 finished, and i386 and ia64 are still going strong, so I suspect this linux32 issue is the only problem you've got, which isn't so terrible.
[08:30] <infinity> BenC: Can you make a good argument for why linux32 should be used for everything *but* the kernel?  I could special-case it in the script that calls sbuild.
[08:31] <infinity> (I'd prefer not to, but it's trivial)
[08:31] <BenC> infinity: because the kernel should be able to decide for itself, because it's smart enought to? :)
[08:31] <BenC> besides, we _really_ do want to build for sparc64, for example
[08:31] <BenC> unlike the rest of the packages
[08:32] <infinity> That's an argument for not caring if it's linux32 or linux64. :P
[08:32] <infinity> The kernel should build for what you told it to, not care about uname.
[08:40] <fabbione> Missing /usr/src/sparc/ubuntu-2.6/debian/abi/2.6.19-1.1/abiname file.
[08:40] <fabbione> make: *** [clean]  Error 1
[08:40] <fabbione> bah
[08:41] <fabbione> BenC: yeah linux32 makes the mess
[08:42] <fabbione> exec make -f /usr/share/kernel-package/ruleset/minimal.mk debian APPEND_TO_VERSION=-2-sparc64  KPKG_SUBARCH=sparc64  KPKG_STEM=linux 
[08:42] <fabbione> strange is that we tell to look for sparc64
[08:59] <BenC> fabbione: I have a commit left for current git
[08:59] <fabbione> sure..
[09:00] <fabbione> i know about the abi.. that doesn't bother me :)
[11:14] <mhb_> hello all, sorry to bother you, I just wanted to ask you if someone couldn't at least comment on bug 63134, please... it looks quite problematic to me, thank you for anything.