[09:00] <fabbione> hey BenC 
[09:00] <fabbione> BenC: i have a bunch of updates to push to git
[09:01] <fabbione> if you can kindly wait to pull from me, it will spare you another pull in a few hours
[09:05] <BenC> ok, I probably wont pull till I get back up in the morning :)
[09:05] <fabbione> perfect
[09:13] <BenC> hey fabbione, has your powerbook ever booted where capslock and numlock were reverted?
[09:14] <fabbione> hmm i never noticed that
[09:14] <fabbione> i don't use any of them
[09:14] <BenC> last time I rebooted (even in macosx) they were on to start
[09:14] <BenC> it was weird
[09:14] <fabbione> oh..
[09:14] <fabbione> no, never seen
[09:14] <fabbione> i need to check with the latest kernel tho
[09:14] <fabbione> it's a bit i don't update the toy
[09:14] <BenC> well, the lights were on, but they weren't enabled
[09:14] <fabbione> nope.. never seen
[09:15] <BenC> shouldn't be kernel related, since macosx was affected too
[09:15] <BenC> I fear hardware issues :)
[09:15] <fabbione> possibly
[09:15] <fabbione> we can verify the OF versions..
[09:15] <fabbione> i just don't know how
[09:15] <BenC> something in /proc/device-tree
[09:16] <fabbione> ok i will look at it later
[09:16] <fabbione> wife is using it atm
[09:17] <BenC> I'm hoping terra/ibm will do that deal again soon, my wife wants one for herself :)
[09:17] <BenC> but that means I'll get the G5 for building/testing
[09:18] <fabbione> ehehe
[09:30] <fabbione> BenC: ok i am done with the redhat cluster suite
[09:30] <fabbione> you can pull anytime
[09:30] <BenC> ok, thanks
[09:31] <sn9> any more test builds of snd-powermac.ko?
[09:31] <BenC> not yet, but I will be getting to it this coming week
[03:09] <Mithrandir> BenC: prod?
[03:10] <BenC> Mithrandir: ?
[03:10] <Mithrandir> BenC: have you done anything which could have made /dev/null turn up as c 0 259 in the latest kernels?
[03:10] <Mithrandir> (the static /dev/null)
[03:11] <Mithrandir> (when it's on squashfs)
[03:11] <BenC> heh, no, pretty sure I haven't :)
[03:11] <sn9> sounds like a stray udev rule to me
[03:11] <BenC> unless something in squashfs did it
[03:11] <Mithrandir> sn9: static dev.
[03:11] <zul> heylo
[03:12] <Mithrandir> BenC: same image looks fine with an older kernel.
[03:12] <Mithrandir> I see it on both i386 and amd64
[03:12] <sn9> i can check ppc
[03:13] <sn9> a quick ls -l showed 1,3
[03:13] <Mithrandir> BenC: if you pull down http://cdimage.ubuntu.com/livecd-base/current/i386.squashfs and look at it, you'll see that /dev/null is c 0 259
[03:14] <Mithrandir> sn9: how did you check?
[03:14] <BenC> what do you mean "static dev"?
[03:14] <Mithrandir> if you mount that image somewhere and look in /dev
[03:14] <sn9> i did ls -l
[03:14] <BenC> Mithrandir: but what created that file?
[03:15] <Mithrandir> : tfheen@thosu /tmp > sudo mount -o loop amd64.squashfs /mnt -t squashfs
[03:15] <Mithrandir> : tfheen@thosu /tmp > ls -l /mnt/dev/null
[03:15] <Mithrandir> crw-rw-rw- 1 root root 0, 259 2006-02-03 05:53 /mnt/dev/null
[03:15] <Mithrandir> BenC: mksquashfs.
[03:16] <Mithrandir> this worked with -14 but not -15
[03:16] <sn9> oh. i looked at -14
[03:17] <BenC> Mithrandir: no, what created /dev/null in the dir where you used mksquashfs
[03:17] <Mithrandir> BenC: debootstrap/makedev.
[03:17] <BenC> from what I can tell, mksquashfs is like mkisofs, it doesn't create files
[03:17] <sn9> you had said "latest kernels" so i thought -14 would be included
[03:17] <Mithrandir> BenC: it's done on the buildds, which run .12
[03:18] <Mithrandir>   * Updated squashfs to latest CVS.
[03:18] <Mithrandir> I wonder if that can be blamed
[03:18] <BenC> Mithrandir: so maybe debootstrap is at fault?
[03:18] <BenC> Mithrandir: is /dev/null ok in the root where deboostrap creates?
[03:19] <BenC> so if you mount the squashfs under -14, it /dev/null correct?
[03:19] <BenC> if the kernel module is at fault, then the image should be fine, and would work with -14, but not -15
[03:19] <sn9> maybe the version of mksquashfs used isn't in sync with the squashfs in the kernel
[03:19] <Mithrandir> BenC: correct.
[03:19] <Mithrandir> sn9: shouldn't matter.
[03:19] <BenC> Mithrandir: so you verified that this image works correctly with -14?
[03:19] <Mithrandir> BenC: yes.
[03:19] <Mithrandir> BenC: or at least, /dev/null isn't on crack
[03:19] <BenC> Mithrandir: actually it could matter if they made some whacked change in the filesystem
[03:20] <Mithrandir> well, they do actually version the file systems, so if so it'd still be a bug in the kernel module
[03:20] <sn9> just because it shouldn't doesn't mean it doesn't
[03:20] <Mithrandir> it supports 1.0, 2.0 and 2.1, IIRC.
[03:20] <BenC> ok, I'm booted into -14 on my powerbook, but I have -15 installed
[03:20] <BenC> I'll check into this
[03:20] <Mithrandir> coolie, thanks.
[03:21] <BenC> Mithrandir: no, there was compatibility options in the kernel module that I wasn't sure about
[03:21] <BenC> it's possible that I didn't enable/disable one like I should
[03:21] <BenC> is this holding up cd's?
[03:21] <Mithrandir> yes.
[03:21] <BenC> I'll make it a priority for today then
[03:21] <Mithrandir> thanks
[03:21] <BenC> np
[03:22] <fabbione> hey Ben
[03:22] <BenC> how long will you be around?
[03:22] <BenC> hey fabbione
[03:22] <zul> oh hey BenC..
[03:22] <BenC> hey zul
[03:22] <Mithrandir> BenC: I?  An hour more or so.
[03:22] <BenC> I've been spending the past few days out in my barn working on oopses and breezy->dapper regeressions...will be nice to work inside again :)
[03:23] <fabbione> hi zul
[03:23] <BenC> Mithrandir: probably wont be fixed before you leave, but by tomorrow for you :)
[03:24] <fabbione> BenC: you are good to pull from me
[03:24] <Mithrandir> BenC: thanks
[03:24] <fabbione> BenC: tho it might add a symbol, but don't break the ABI
[03:29] <BenC> fabbione: ok, because I'll want to do a quick upload for this squashfs problem with the ABI bump
[03:29] <BenC> I might just skip the pull to be safe, and do it after the upload
[03:30] <fabbione> BenC: if you plan to bump the ABI, it would be good to check OCFS2 and import bug fixes
[03:30] <fabbione> i didn't get that far
[03:30] <fabbione> the pull is safe..
[03:30] <fabbione> but up to you
[03:30] <fabbione> just don't wait long, there is also that security fix in it
[03:31] <BenC> ok
[03:31] <BenC> rebooting to -15, squashfs looked good under -14
[04:49] <BenC> Mithrandir: fixed
[04:49] <BenC> uploading soon
[04:50] <BenC> I would strongly suggest that you update squashfs tools from CVS
[04:57] <Mithrandir> oh, why?  We're in UVF.
[04:59] <BenC> it's not a requirement, just a strong suggestion :)
[04:59] <BenC> the new code isn't being used if you use the v2 tools
[04:59] <BenC> not sure what benefits the new code holds, but it may fix some problems
[05:00] <BenC> maybe you can test with both and see if it helps/hurts? especially on ppc where we have oopses
[05:32] <Mithrandir> I still don't have a workable ppc. :-)
[05:54] <zul> later
[07:29] <mdz> BenC: speaking of unionfs powerpc oopses, I've sent you an email regarding some potential assistance there
[07:55] <mdz> BenC: ack?
[08:08] <BenC> mdz: ok
[08:09] <BenC> mdz: what's the subject of the email, I don't see it
[08:10] <mdz> Subject: [jeremy@au1.ibm.com: Re: Thanks - and guest access card] 
[08:11] <BenC> heh, I might have deleted that as spam considering the subject, can you resend?
[08:12] <mdz> sent again
[08:12] <BenC> thanks
[09:32] <_ace> !help
[09:33] <_ace> hi
[09:33] <_ace> have some questions about make-kpkg
[09:33] <_ace> specifically how to change the name of the kernel
[09:34] <_ace> and later also questions about build-kernelpackage
[09:34] <BenC> what do you mean "name"?
[09:35] <_ace> I made a kernel with make-kpkg and the deb is named kernel-image-2.6.11_2.6.11-0.2_i386.deb
[09:35] <_ace> (I know, it's old, I have reasons though)
[09:35] <BenC> so what name would you like it to be?
[09:36] <sn9> probably linux-image
[09:36] <_ace> Something like kernel-image-2.6.11-w4l-686-smp-megaraid.deb
[09:36] <_ace> sn9
[09:36] <_ace> linux or kernel, there's a whole stoy behind that! want me to tell ya ?
[09:36] <BenC> --append_to_version=-w4l-686-smp-megaraid
[09:37] <BenC> for s/kernel/linux/ do --stem=linux
[09:37] <_ace> ah
[09:37] <_ace> that's 'stem' about thank you.
[09:38] <BenC> not sure where stem came from, but we set it to linux
[09:38] <_ace> will the append-version get rid of the '_2.6.11-0.2_i386' part 
[09:38] <BenC> I think the head is that it can be --stem=bsd at some point
[09:38] <BenC> the 2.6.11-0.2 part is the version
[09:38] <BenC> or revision to be more exact
[09:39] <BenC> all of this is covered in make-kpkg(1)
[09:39] <BenC> and you wont be able to get rid of the _i386 part, that's a dpkg-deb thing, you don't want to mess with it
[09:40] <BenC> in your first example, linux-image-2.6.11 is the package name, 2.6.11-0.2 is the version+revision, and i386 is the architecture
[09:41] <_ace> okay Ill play with that a bit more
[09:41] <_ace> for reference, just chek http://www.suares.an/?page_id=28
[09:42] <BenC> if you change the package name (--append_to_version), it will change the package name and the version+revision
[09:42] <_ace> it also explains the reason I went for the 2.6.11 kernel
[09:42] <_ace> However, in a wiki i read that you can do apt-get source linux-2.6.8.1
[09:42] <_ace> but that gives an error
[09:43] <_ace> you can only do that for 2.6.12
[09:43] <_ace> why is that not possible for 2.6.11 ? 
[09:43] <_ace> also, the only place I found the 2.6.11 kernel with debian patches
[09:44] <_ace> is http://security.ubuntu.com/ubuntu/pool/universe/l/linux-source-2.6.11/
[09:44] <_ace> there are three files (orig, diff and dsc) and I can finally (after some days of toying around
[09:44] <_ace> )
[09:44] <_ace> make a kernel from that.
[09:44] <sn9> it all depends on your sources.list
[09:44] <BenC> because we never shipped a 2.6.11 kernel
[09:45] <_ace> BenC: aha ! thx for the clarification!
[09:45] <BenC> it was a development package, not really supported
[09:45] <_ace> now this kernel that I found
[09:45] <_ace> which *version* is that
[09:45] <_ace> 2.6.11.7 ?
[09:45] <_ace> or some even older one ?
[09:45] <BenC> no idea to be honest, It was before my time here :)
[09:45] <BenC> probably older
[09:45] <_ace> heh
[09:45] <_ace> thought so
[09:46] <_ace> You know the whole problem would be away
[09:46] <_ace> if win4lin had had a patch for 2.6.12 that *does* include smp support
[09:46] <_ace> but that is not the case
[09:46] <_ace> that cost me a lot of time
[09:46] <_ace> win4lin distributes generic kernels
[09:46] <_ace> with 2.6.8.1 
[09:46] <BenC> the 2.6.11 patch doesn't apply to 2.6.12?
[09:47] <_ace> and it doesn't boot on my machine since it misses the megaraid
[09:47] <_ace> drivers..
[09:47] <_ace> eh.. 
[09:47] <_ace> i haven't tried
[09:47] <BenC> and if not, have you tried manually fixing the failed hunks of diff?
[09:47] <BenC> you should try
[09:47] <_ace> but i figured why they publish a patch for 2.6.12 that has no smp support
[09:47] <BenC> just because they say 2.6.11, doesn't mean it wont work with 2.6.12
[09:47] <_ace> if it were that easy...
[09:47] <_ace> guess i should try that...
[09:47] <BenC> probably because the 2.6.11 patch applied cleanly to 2.6.12? :)
[09:48] <_ace> would be very strange
[09:48] <BenC> I'd give it a go, if you get a couple of .rej's you can clean them up by hand usually
[09:48] <_ace> http://www.win4lin.com/component/option,com_remository/Itemid,76/func,select/id,4/
[09:48] <_ace> they *do* have a patch, if the 2.6.11 worked, why make a 2.6.12 patch with no smp support ?
[09:49] <_ace> 2.6.13 doesn't have smp support
[09:49] <_ace> en at 2.6.14.3 suddenly it's back
[09:50] <BenC> just making suggestions, I've no idea what might or might not happen with it
[09:50] <_ace> good idea though, I might try that.
[09:50] <_ace> still, in the end I want a *real* ubuntu kernel
[09:50] <_ace> so now I learned about dpatch
[09:50] <_ace> heh
[09:51] <_ace> and build-kernelpackage (or whatever that command is called) failed miserably with the 2.6.11
[09:51] <_ace> stuff and now I know why :-(
[09:51] <_ace> I'll go off and try some stuff now... brb...
[09:52] <_ace> thanks so far...
[10:18] <bubbalwz> BenC: any idea when the kernel update will be out that'll fix megaraid?
[10:18] <bubbalwz> i thikn fabbione had the changes
[10:18] <BenC> I pulled his changes for dapper earlier
[10:18] <bubbalwz> ok, so nothing for breezy?
[10:19] <bubbalwz> i tested a kernel he sent me ... 2.6.12-10-686.26
[10:19] <bubbalwz> on breezy
[10:20] <bubbalwz> just wonderinf if that would make it into a security update
[10:39] <_ace> whatr's wrong with the megaraid ?
[11:11] <fabbione> bubbalwz: i am uploading to breezy now or tomorrow.
[11:11] <fabbione> bubbalwz: the fix is there
[11:17] <fabbione> bubbalwz: ok 2.6.12-10.27 is up
[11:17] <fabbione> it will take a day or probably 2 to be in the archive
[11:17] <fabbione> with that version you should be fine
[11:18] <fabbione> linux-source-2.6.12 (2.6.12-10.27) breezy-security; urgency=low
[11:18] <fabbione>   Changed by Fabio M. Di Nitto:
[11:18] <fabbione>   * Fix regression from hoary kernel in megaraid scsi driver:
[11:18] <fabbione>     In the breezy kernel a split up of the megaraid and megaraid_m* driver was
[11:18] <fabbione>     done to be able to support all controllers. During the split one pci
[11:18] <fabbione>     subvendor ID was lost due to removal of PCI_ANY catch all.
[11:18] <fabbione>     Thanks to Brian (bubba) Wilson for spotting the issue and testing the fix.
[11:18] <fabbione>     Update drivers-scsi-megaraid_splitup.dpatch with the missing ID.
[11:18] <fabbione>   * [SECURITY] : Fix extra dst release when ip_options_echo fails:
[11:18] <fabbione>     - Add patch CVE-2006-0454.dpatch.
[11:18] <fabbione>  -- Fabio M. Di Nitto <fabbione@ubuntu.com>  Wed, 08 Feb 2006 10:41:12 +0100