[01:56] <Jake333> hi
[12:30] <poningru> hey what do you guys think should go into herd5 'release notes' for kernel?
[12:38] <Nafallo> BenC: ping
[12:38] <Nafallo> hmm
[12:38] <Nafallo> BenC: ping rt61 :-)
[02:17] <fryfrog> is it okay to ask for edgy kernel help in here?
[02:21] <Nafallo> fryfrog: no. this is a development channel. #ubuntu is the place.
[02:27] <fryfrog> blah, such a busy channel :)
[02:27] <fryfrog> is there any place for kernel specific help?
[02:30] <Nafallo> not that I'm aware of.
[02:30] <fryfrog> okay, thanks :)
[02:55] <fabbione> BenC: are we still in contact with that dvb guy that used to show up here?
[02:56] <BenC> fabbione: haven't heard from him in awhile
[02:56] <fabbione> BenC: ok thanks
[03:02] <zul> merky or soemthing like that?
[03:06] <Mithrandir> zul: what is xen-feisty blocked on?  It's listed as blocked in the spec overview.
[03:06] <zul> its not blocked, I havent update it in a while
[03:07] <Mithrandir> can you fix the spec status, then?
[03:07] <zul> yes will do
[03:07] <mjg59> BenC: 86457 sounds like a pretty valid request
[03:09] <kylem> ugh. my eyes aren't working yet this morning.
[03:10] <BenC> kylem: Oooh, I think I found the problem with ASLR
[03:10] <BenC> COMPAT_VDSO is disabled because PARAVIRT is on, but in reality, those should be ok together now
[03:10] <kylem> BenC, wow. ten minutes. that was quick.
[03:11] <kylem> ;-)
[03:12] <mjg59> BenC: Ok. Fancy updating the bug? :)
[03:12] <BenC> just did
[03:13] <mjg59> Cool
[03:14] <mjg59> Did those patches I sent look ok?
[03:15] <BenC> yeah, merging everything now
[03:15] <BenC> if there's no ABI bump, maybe I can convince Mith to let another kernel upload squeeze in for Herd5 :)
[03:16] <mjg59> Heh
[03:17] <mjg59> Should be ok, I think
[03:17] <Mithrandir> BenC: I'd really like to avoid that unless it's "does not boot on any hardware" or a similar problem.
[03:37] <lamont> -11 fails to let me find any USB devices unless I boot with 'irqpoll'
[03:38] <lamont> historically, every now and then USB would go belly up - never tracked it down
[03:40] <lamont> -10 was happy, for the most part
[03:40] <lamont> HP pavillion a1130n, and yes, I apologize
[03:41] <lamont> ati chipset, so feh
[03:49] <BenC> lamont: I think the USB breakage was due to a late change in some USB host setting we had to enable
[03:49] <BenC> it was one of those things where the option was enabled as early as breezy, and some USB wouldn't work at all, or we enable the option, and some people had to use irqpoll
[03:49] <lamont> BenC: is it worth me working on tracking down the change that broke it?
[03:49] <BenC> lamont: I believe it's a config option
[03:50] <BenC> finding out why that broke working setups would be very worth it
[03:50] <lamont> -10 and -11 configs differ only in CONFIG_VERSION_SIGNATURE
[03:50] <BenC> maybe it was something else then
[03:50] <lamont> so, yeah, it's an old option :-)
[03:51] <lamont> would acpi=noirq help maybe?
[03:51] <mjg59> Unlikely
[03:52] <mjg59> But if it does, it gives a pretty clear indication what hte problem is
[03:52] <kylem> it would be nice if people could reproduce such things with newer upstream kernels and file it in kernel.org bugzilla.
[03:52] <mjg59> kylem: Maybe we should package pristine kernels as well
[03:53] <kylem> when a shared irq is raised, every registered handler gets called. because we have 3 usb on the same irq line, i was thinking maybe it got confused.
[03:53] <lamont> can I drop a feisty kernel on an edgy system?
[03:53] <kylem> lamont, shall i just fedex you a kernel.org dvd? :P
[03:53] <mjg59> kylem: It's pretty normal to have multiple USB controllers on one IRQ
[03:53] <lamont> kylem: good bandwidth, high latency.  I'll just fetch it at work
[03:53] <lamont> because fedex same-day service is _EXPENSIVE_
[03:54] <kylem> lamont, if you use git over http, you should be able to pull just a few large packfiles instead.
[03:54] <BenC> lamont: feisty kernel on edgy is ok so long as you also upgrade initramfs-tools, module-init-tools and udev
[03:54] <kylem> (instead of tonnes of small files, etc.)
[03:54] <lamont> BenC: the newer other stuff works with an edgy kernel?
[03:55] <BenC> lamont: That part I don't know :)
[03:55] <lamont> heh
[03:55] <lamont> how stable is feisty these days?
[03:55] <lamont> will herd 5 be safe enough for production?
[03:55] <lamont> production == my primary workstation
[03:56] <BenC> I've been using feisty (kernel AND userspace) on my laptop for about 4 weeks now
[03:56] <lamont> and no blood?
[03:58] <thom> i've been running feisty on my laptop since UDS, as it happens
[03:58] <lamont> thom: nice
[03:59] <thom> it's been occasionally entertaining, but pretty sane on the whole
[03:59] <lamont> lol
[03:59] <Mithrandir> I haven't had a single entertaining episode on my laptop, at least.
[04:00] <kylem> python got out of synch last friday so i had update-notifier crashing every 10 minutes all weekend.
[04:01] <zul> die rpm die!!
[04:18] <zul> BenC: ping...can we get some time to train the guy who is doing the breezy updates because after the 12th I wont be around much
[04:18] <BenC> zul: Sure
[04:19] <zul> when is breezy EOL anyways?
[04:20] <kylem> zul, june.
[04:20] <kylem> is my guess.
[04:20] <kylem> april + migration period, i would imagine
[04:21] <Mithrandir> more like april.  18 months from the release date.
[04:22] <zul> so in theory this could be the last update?
[04:22] <kylem> zul, see, there's always a bright side to things.
[04:23] <zul> kylem: definently
[04:23] <JaccoH> good .. afternoon
[04:23] <kylem> JaccoH, hallo
[04:23] <JaccoH> hallo
[04:23] <JaccoH> you speak dutch? 
[04:24] <kylem> no sir.
[04:24] <JaccoH> but you do know hallo :)
[04:24] <kylem> heheh
[04:25] <JaccoH> or was that a lucky typo?
[04:25] <kylem> i'm just being odd. i had no idea. :)
[04:25] <JaccoH> i see .. hallo is dutch for hello.. close enough
[04:26] <JaccoH> zul i ordered dual e1000 nic to uh .. fix my problem with the bnx2
[04:27] <JaccoH> but its not in yet
[04:27] <zul> ah
[06:50] <kylem> lamont, oi.
[06:51] <lamont> si?
[06:51] <kylem> lamont, how come ia64 doesn't have builds.
[06:51] <kylem> of edgy-security kernels.
[06:51] <lamont> ??
[06:51] <lamont> oh.  dunno.  I should go look
[06:51] <kylem> Fetching itanium...FAILED.
[06:51] <kylem> Fetching mckinley...FAILED.
[06:51] <kylem> the srcpackage is there.
[06:53] <Nafallo> hmm
[06:53] <Nafallo> new security-upload on the way?
[06:53] <kylem> yes.
[06:54] <Nafallo> damnit. I never get the uptimes I had ;-)
[06:54] <Nafallo> I should probably play with the server and XENify it or something... that might help :-)
[07:00] <thom> you still have to reboot for the xen kernel updates, so it doesn't buy much :-)
[07:00] <thom> can we kexec from a running kernel to a new one yet?
[07:01] <lamont> thom: depends on the platform, I think...
[07:02] <lamont> kylem: as mentioned elsewhere, I'll drive fixing that...
[07:02] <kylem> lamont, thanks chief.
[07:02] <kylem> it's blocking me for now, ironically.
[07:03] <poningru> so... anyone want something about kernel to go into herd 5 'release notes'?
[07:03] <poningru> everytime usually BenC puts something in this time nothing :(
[07:03] <kylem> poningru, sure, can you add that "rtl818x/rtl8187 drivers have been blacklisted due to being broken"
[07:04] <kylem> feel free to dress up the language. :)
[07:04] <poningru> heh
[07:04] <poningru> I dont think I can put that in a release note ;)
[07:04] <BenC> poningru: There's not really a whole lot new between herd4/herd5
[07:05] <poningru> cool bugfixes it is
[07:05] <BenC> poningru: Oh, bcm43xx is working again :)
[07:05] <BenC> was broken in herd4
[07:05] <poningru> still requires firmware right?
[07:05] <poningru> or does it do it automagically?
[07:06] <poningru> with bcm-fwcutter
[07:07] <poningru> dude... please dont keep me in suspense I am gonna have a brain aneurysm here, this is one of my pet peeves
[07:08] <poningru> :(
[07:11] <Nafallo> my feisty crashed a lot until I tried rmmod rt61
[07:11] <Nafallo> almost 1d uptime now
[07:18] <poningru> BenC: it still requires firmware right?
[07:19] <BenC> right
[07:26] <bdmurray> Is there a wiki page about debugging kernel crashes via serial connection?
[07:30] <zul> I dont think so
[07:39] <zul> there probably should be
[07:40] <BenC> I think it's generally one of those things where the hardware has to support console on serial
[07:41] <BenC> if it's a crash where the system is still stable enough that the kernel serial console still works, then most likely you don't need serial console
[08:19] <lamont> tifm_7xx1: sd card detected in socket 1
[08:19] <lamont> does that mean I can actually talk to it now?
[08:23] <lamont> hrm.. VESA virtual size - how do I change that in xorg.conf?
[08:30] <BenC> lamont: It works ofr some types of cards, not all
[08:30] <lamont> so if removable media are configured to automount and it doesn't, then no love yet, eh?
[08:32] <BenC> check /lib/modules/`uname -r/kernel/drivers/mmc/
[08:32] <BenC> might have to manually load one of the block drivers and card handlers
[08:32] <lamont> ah, ok
[08:32] <fryfrog> would questions about the linux-restricted-modules debian/rules build script be appropriate here?
[08:38] <pkl_> lamont: check you've got mmc_core and mmc_block loaded
[08:40] <pkl_> lamont: some people are not getting mmc_block automatically loaded.
[09:49] <zul> BenC: er...when is the freeze?
[09:50] <Mithrandir> which of the freezes?
[09:50] <zul> kernel freeze
[09:51] <Mithrandir> april 5th
[09:51] <zul> ah still plenty of time
[09:52] <BenC> we're in semi-freeze right now
[09:52] <BenC> I'm not taking much of anything other than bug fixes
[09:52] <zul> Ill probably have a couple of bug fixes for you on monday
[09:52] <BenC> the 04/05 freeze is "don't touch it unless it wont boot or eats endangered species"
[09:53] <Nafallo> ah good. it can eat children then :-).
[09:55] <Mithrandir> the April 5th freeze is "Ben will pay Tollef cases of beer for each upload". :-)
[09:56] <zul> lol
[10:01] <Nafallo> hehe
[10:01] <zul> right Im heading home..
[10:31] <BenC> Mithrandir: that might get kind of expensive :)
[10:32] <ajmitch> BenC: just try & put all the fixes in 1 or 2 uploads
[10:32] <BenC> there's always last minute fixes
[10:32] <ajmitch> as long as he doesn't specify expensive beer, it should be fine :)
[10:33] <ajmitch> you couldn't claim it as misc work expenses?
[10:36] <crimsun> after a few cases he'll be zonked anyhow, so you can just forego the remainder
[11:28] <psusi> I'm having a little trouble understanding the relationship between struct request, and struct bio... could someone explain?