[08:58] <tepsipakki> a regression in feisty-kernel: it accepts only 256 characters on the commandline
[08:59] <tepsipakki> where ->edgy accepted 1024
[09:11] <fabbione> tepsipakki: file a bug. 
[09:11] <fabbione> you should know that by now
[09:13] <tepsipakki> ok, I'll reopen the one from a year back ;)
[09:17] <tepsipakki> in fact, pxelinux accepts "only" 512 chars so that needs fixing as well, in another bug
[09:20] <Mithrandir> it hasn't occured to you that if you need to pass in half a k of text you might be doing something wrong?
[09:20] <tepsipakki> no :)
[09:20] <tepsipakki> netcfg is slightly broken
[09:20] <tepsipakki> and we need to use static settings
[09:20] <Mithrandir> use a preseed file
[09:20] <Mithrandir> or fix netcfg
[09:20] <tepsipakki> preseeding doesn't work
[09:21] <tepsipakki> actually, this is preseeding..
[09:21] <tepsipakki> but it can't fetch the file before net is up
[02:31] <lkolbe> Hi, I have a problem with make-kpkg, see http://paste.ubuntu-nl.org/35277/
[02:31] <lkolbe> is this the right place to ask?
[02:55] <BenC> lkolbe: Old version of kernel-package I suspect
[02:57] <lkolbe> Well, I tried with both 9.001ubuntu15 from ubuntu dapper, as well as a rebuilt 10.065 from debian sid
[02:57] <lkolbe> and it always things the version is "kernel-image-.." instead of "kernel-image-2.6.19"
[02:57] <lkolbe> and then it fails ...
[02:58] <lkolbe> thanks for listening so far :)
[03:48] <lkolbe> Oh. I just saw that make-kpkg is not the recommended way to rebuild a kernel on ubuntu?
[04:34] <fabbione> BenC: ping?
[04:34] <BenC> fabbione: pong
[04:34] <fabbione> BenC: did you talk with cjwatson about the kernel upload?
[04:35] <fabbione> I had to push a 7.11 on his behalf to fix missing isofs
[04:35] <fabbione> but my git here still doesn't frigging work
[04:35] <fabbione> it's a one line change for debian/d-i/....
[04:35] <BenC> fabbione: That's ok, because ubuntu-2.6 git is at 2.6.20-git now
[04:35] <fabbione> the rest is only to workaround the ABI checker because i couldn't fetch all the ABI's
[04:35] <BenC> fabbione: If you can send me a diff, I'll include it for git
[04:36] <fabbione> doing
[04:36] <Mithrandir> BenC: how come that it didn't ftbfs?  Isn't isofs required for any udebs?
[04:37] <BenC> Mithrandir: I redid the udeb's for feisty to clean things up
[04:37] <Mithrandir> ok.
[04:37] <BenC> guessing I missed that module
[04:37] <fabbione> Mithrandir: the kernel module is in the deb but not copied in the isofs.. that's not FTBFS
[04:37] <fabbione> hem
[04:37] <fabbione> isofs/udeb
[04:38] <Mithrandir> fabbione: ok, so it was just missing from the correct udeb package list
[04:38] <BenC> right
[04:38] <fabbione> Mithrandir: yeps
[04:39] <fabbione> BenC: http://people.ubuntu.com/~fabbione/cocktastic.diff
[04:39] <fabbione> out of the entire diff you need a one liner :)
[04:40] <fabbione> and the debian/control is still due to LC_ALL=C thingy
[04:41] <BenC> fabbione: Should isofs really be "?"
[04:41] <fabbione> BenC: i was not sure if it was module on all kernels.. so i just made it a failsafe
[04:41] <fabbione> BenC: you are welcome to remove ? if you are sure it's not required
[04:41] <fabbione> i didn't have all the images handy due to port being out of sync and blablabla
[04:41] <BenC> pretty sure it is required
[04:41] <fabbione> so i couldn't verify
[04:42] <fabbione> oh and btw.. we need to update the getabis script to cope with the new lowlatency kernels
[04:42] <Mithrandir> BenC: also, can you make the kernel build faster?  > 4 hours is getting excessive..
[04:42] <Mithrandir> (or our i386 buildd is shit and should be taken out and shot)
[04:42] <fabbione> Mithrandir: the latter :)
[04:43] <fabbione> i so totally blame our buildd
[04:43] <BenC> yeah, it only takes about 40 minutes for me
[04:43] <kylem> let's just disable all the drivers for Mithrandir's hardware
[04:43] <fabbione> Mithrandir: kernel build can fork... ask James to add CPU's.. many many CPU's with tons of sweet ram
[04:43] <Mithrandir> kylem: I think "some" people might complain if X40s suddenly stopped working.
[04:43] <fabbione> kylem: no, you really don't want to do that
[04:44] <kylem> fabbione, no shit.
[04:44] <fabbione> kylem: he has more hw than anybody else
[04:44] <fabbione> kylem: it mean basically # CONFIG_BUILD_KERNEL is not set
[04:44] <Mithrandir> kylem: if you have space for all your hardware in your office, you have too little hardware. :-P
[04:45] <kylem> Mithrandir, heh, i don't
[06:13] <kylem> BenC, how'd we want to go about uploading the security kernels? if you're ok with what's in git, pitti said he would upload the other day.
[06:17] <BenC> kylem: I can upload if you want
[06:18] <kylem> i've done testbuilds of both dapper and edgy -security, and they seem fine
[06:33] <zul> hey
[06:34] <zul> just got back from the ultrasound again baby's heart looks better but still inconclusive
[06:43] <jbailey> BenC, kylem: There?
[06:43] <zul> hey jeff
[06:43] <jbailey> Heya Chck
[06:43] <jbailey> +u
[06:43] <kylem> jbailey, no.
[06:43] <jbailey> kylem: ?
[06:43] <kylem> jbailey, whaddaya want ;-)
[06:44] <jbailey> When we come up with certification bugs on supported hardware.
[06:44] <jbailey> Regressions specifically.
[06:45] <jbailey> What's the best way to get those into the Kernel Team so that it gets prioritised appropriately?
[06:45] <jbailey> We have an escalation process already for the support team into distro for customer things.
[06:45] <jbailey> But curious about the cert side of it now.
[07:28] <BenC> jbailey: that's a good question
[07:28] <jbailey> Keybuk: Heya Scott
[07:28] <jbailey> BenC: It's something to think abotu.
[07:28] <jbailey> I want to make sure we give you the stuff in the most useful way possible.
[07:30] <Keybuk> jbailey: heya
[07:30] <BenC> jbailey: for regressions (like with qla2xxx), then the current way you guys have been doing things seems to be working
[07:31] <BenC> jbailey: I don't know if we need anything more formal than an lp bug, and support pinging the kernel team till they wake up and get on it :)
[07:32] <BenC> jbailey: Maybe you, I and Matt need to formalize how much priority we put on this stuff in comparison to our normal work though
[07:32] <jbailey> Yeah.
[07:32] <jbailey> I'll write something up and send it off.
[07:40] <BenC> Keybuk: ping
[07:41] <BenC> kylem: We have our private dir now...I'll let you know when it's ready
[07:42] <Keybuk> BenC: sorry, on a conf call, can't talk right now
[07:43] <Keybuk> what's the ping about?
[07:44] <BenC> Keybuk: I want to take some of your time to talk about modprobe/udev
[07:45] <BenC> Keybuk: When you get a few minutes can you ping me back?
[07:45] <zul> BenC: still no sign of paravirt in linus' tree yet?
[07:46] <BenC> zul: No, but the merge fest is still on-going
[07:46] <zul> yeah
[07:52] <zul> oh yeah when are we suppose to be doing the security uploads?
[08:06] <BenC> zul: If you have breezy caught up, talk to pitti
[08:10] <zul> ok
[08:31] <Keybuk> BenC: IRC conversation or voip call?
[08:32] <BenC> I don't have canonical voip setup, but my vonage let's me call uk for free
[08:32] <BenC> Keybuk: /msg me a number on that other server? :)
[08:32] <Keybuk> BenC: is this time tomorrow ok?
[08:32] <BenC> yeah, sure
[08:33] <Keybuk> I'm still on the "binary driver" call, and then will be having an evening and decorating the tree :p
[08:33] <BenC> 19:30 GMT tomorrow?
[08:33] <Keybuk> yup
[08:34] <Keybuk> actually, can we make it 1900 ?
[08:34] <Keybuk> TB at 2000 ?
[08:34] <Keybuk> s/ ?$//
[08:37] <BenC> Keybuk: ok
[08:37] <Keybuk> BenC: do you have an ordinary (non-canonical) voip setup?
[08:37] <BenC> it's vonage
[08:38] <Keybuk> because you can just do a SIP call to scott@netsplit.com, if you didn't want to dial a phone number
[08:38] <Keybuk> makes no real difference, except the phone number you'll be dialing is also a SIP gateway :P
[08:39] <Keybuk> (so will get analogued along the way)
[08:39] <dade`> BenC: do you have any news about macbook sleep ? 
[08:39] <BenC> not sure how to setup this phone/voip gw to do SIP
[08:39] <Keybuk> ok, no worries
[08:48] <tepsipakki> kylem: do you have the fix for #72696 in your dapper tree?
[08:49] <kylem> no. will get to that.
[08:49] <kylem> tepsipakki, thanks for pointing it out
[08:50] <kylem> BenC, backporting sky2 wasn't horribly difficult, i assume tg3 will likely be similar and can reuse the same compatibility code. shall i do that?
[08:52] <tepsipakki> kylem: cool :)
[08:53] <zul> kylem: fyi i had some problem with sky2 compiling with the xen crack when i was doing an update for the xen kernel but i got around it
[08:53] <BenC> kylem: Sure
[11:21] <BenC> superm1: I recall looking at lirc before, but I think it was too crappy to include or something
[11:22] <BenC> superm1: I also remember ivtv
[11:22] <superm1> well i know that there was a lot of breakage for it in edgy
[11:22] <superm1> i2dc and gpio wouldnt build
[11:22] <BenC> can you send me URL's to both so I can take a quick look?
[11:23] <superm1> newest ivtv should have no trouble and uses module-assistant. http://packages.ubuntu.com/edgy/x11/ivtv-source
[11:23] <BenC> module assistant means nothing when including it in the proper kernel
[11:23] <superm1> and actually that version won't go against 2.6.19 - its intended for 2.6.17, i can get a package in that will though
[11:23] <superm1> oh i see
[11:24] <BenC> I want upstream source tarballs
[11:24] <superm1> k, let me find the 2.6.19 tarball then.
[11:24] <BenC> 2.6.19+
[11:24] <BenC> we are going to have 2.6.20 in feisty
[11:24] <superm1> http://dl.ivtvdriver.org/ivtv/archive/0.9.x/ivtv-0.9.0.tar.gz
[11:25] <superm1> as of now 0.9.0 is intended for 2.6.19, i'm not sure what their take will be on 2.6.20
[11:26] <superm1> and as for lirc, the CVS snapshots should build against 2.6.19 with less trouble on gpio and i2c, but it would probably be better to get the newer lirc in feisty before adding a snapshot of just the modules
[11:27] <superm1> so probably hold off on the lirc until I can settle that with -motu
[11:30] <BenC> superm1: ivtv seems to build ok
[11:30] <BenC> I'll try to get it included in the next upload
[11:30] <superm1> okay cool
[11:31] <superm1> how soon will the pre2.6.20 kernels start popping up though?
[11:31] <BenC> this week
[11:31] <superm1> and this built fine against the pre 2.6.20 or the 2.6.19 currently there though?
[11:31] <BenC> pre-2.6.20
[11:31] <BenC> my local tree
[11:31] <superm1> okay great then.  dont need to do too much digging to watch out for breakage :)
[11:32] <BenC> I'll depend on you to test if it actually works, I just compile it :)
[11:32] <superm1> hehe okay
[11:33] <superm1> okay i'll see if i can't sort out the lirc mess to make sure we have something that would compile cleanly with all the modules.  i'll ping you again once i sort that with -motu
[11:33] <superm1> thanks for getting this in :)
[11:43] <jbailey> BenC: So it'll be interesting to see if my assertions that the Feisty kernel is fabulous and amazing continue to be true ;)
[11:43] <BenC> jbailey: I hope so...2.6.20 merge fest is breaking all sorts of stuff in the ubuntu/ driver directory
[11:43] <jbailey> Joy.
[11:43] <jbailey> I haven't tried reading any of the git logs so far.
[11:44] <jbailey> I figured I'd catch up at -rc1 =)
[11:45] <jbailey> pre1,  Whatever Linus calls them now. =)
[12:13] <dimach> Hi. Can someone point me to how compile a kernel module? Ubuntu wiki page says it can be done using kernel headers, but I can't find a guide for this.
[12:13] <Bluhd> I've got a question about the timing resolution about the kernel
[12:14] <Bluhd> whenever I run Rosegarden, it complains that the kernel's timing resolution is too low
[12:15] <Bluhd> is this related to the fact that if you use sleep(1); in C++ that it sleeps for one second instead of one millisecond?