[06:06]  * abogani waves all!
[06:08] <abogani> apw: I don't take care of it anymore, you have to speak with one of the Ubuntu Studio kernel Team members...
[07:55] <apw> abogani, oh ok, any idea who they are ?
[07:55] <apw> abogani, or more likely what the team is called in LP
[08:00] <smb> mumble seems jetlagged, too
[08:07] <abogani> apw, https://launchpad.net/~ubuntustudio-kernel-team
[08:32] <apw> abogani, thanks
[08:33] <abogani> apw, Thanks to you!
[08:39] <dileks> abogani: isnt renaming from lowlatency to preempt a reason to maintain it :-)?
[08:40] <abogani> dileks: lowlatency turn off tickless and enable thread irqs by default. So no preempt isn't enough.
[08:41] <dileks> thanks for the update, was thinking this is a renamed -rt kernel-flavour
[08:43] <ohsix> disable tickless huh, wasn't there a tool that measured wakeup jitter? :D
[09:11]  * ppisati -> brb
[10:59]  * ppisati -> rush out for lunch and some grocery - later
[11:22] <apw> dileks, right ... there never is a -rt patchset close to the version we are releasing on to even consider
[11:58] <tgardner> apw, did you get this bug report from rick about azure ?
[11:59] <tgardner> apw, nm, I see you commented recently
[12:41] <apw> tgardner, yep, at 95 reboots without reproducing ... so ...
[12:41] <tgardner> apw, ah, this one could be a pain in the ass
[12:41] <tgardner> apw, kind of wish he'd have come talk to us last week.
[12:41] <apw> tgardner, yeah ... they are obviously doing something different ... so ...
[12:42] <apw> tgardner, yeah i did talk with him, and there was no real mention of any urgency
[13:23] <tgardner> cking, what is the expected string from 'uname -m' on an i386? would it be "x86_32" or just "x86" ?
[13:24] <cking> tgardner, not entirely sure, /me looks it up on some old kit
[13:26] <ppisati> tgardner: right, there's even ppc
[13:26] <tgardner> ppisati, right, so I think we need to change your patch to look for x32 before checking on PAE
[13:26] <ppisati> tgardner: yep, better
[13:26] <tgardner> I should have thought of other arches when I wrote it originally
[13:27] <cking> tgardner, uname -m on a 32 bit atom returns 'i686'
[13:27] <tgardner> cking, thanks
[13:28] <tgardner> cking, so, searching on "*86" should be a good reg-ex ?
[13:30] <cking> tgardner, i think so, but make sure it doesn't match x86_64
[13:30] <tgardner> cking, how about 'uname -p' on your atom ?
[13:30] <cking> uname -p --> i686
[13:31] <cking> uname -i --> i386
[13:31] <tgardner> cking, ok. given that we're qualifying the search based on 'pae', I think the general "*86" reg-ex is OK, even for AMD etc
[13:32] <cking> tgardner, perhaps checking uname -i is better
[13:32] <tgardner> cking, yep
[13:32] <tgardner> ppisati, you wanna respin using 'uname -i' and checking for "*86" ?
[13:35] <ppisati> tgardner: will do
[13:35] <ppisati> tgardner: actually i would prefer to avoid using any regexp
[13:38] <ogasawara> tgardner, ppisati: I haven't uploaded meta yet, do we want to respin an upload before I do?
[13:39] <ppisati> ogasawara: yep, doesn't install on arm/ppc
[13:39] <ogasawara> ppisati: ack
[13:40] <tgardner> ogasawara, probably cause this'll break armhf
[13:44] <apw> tgardner, as the builds worked with the tools split, it may be worth you testing backing out the -j change you did for tools (as you can repro it pretty easy)
[13:45] <tgardner> apw, I think I did already
[13:45] <apw> ahh ok
[13:45] <tgardner> apw, its still in quantal, but I dropped it from the other releases
[13:46] <apw> tgardner, you did the backport a bit different, should we standardise on your version
[13:46] <tgardner> apw, I'm almost positive that it was releasted to the perf binaries being removed
[13:46] <tgardner> related*
[13:46] <apw> tgardner, it would cirtainly look similar indeed
[13:46] <tgardner> apw, yeah, we could standardize. its a bit simpler.
[13:46] <apw> tgardner, ack, please do
[13:47] <tgardner> apw, will do
[13:54] <apw> ogasawara, is quantal 'good' ie can i punt some more crack into it?
[13:55] <ogasawara> apw: built fine, just holding off on meta till we get the pae check correct and will re-upload.  feel free to shove any other crack in for the upload.
[13:58] <tgardner> apw, ogasawara: I'm gonna reorganize master-next a bit and cleanup some of our (my) mistakes. 
[13:59] <ogasawara> tgardner: cool, I wanted to rebase some bits out of existence too, you go first.
[13:59] <apw> tgardner, ogasawara, when you are done yell and i'll rebase the bits i have onto the top
[13:59] <apw> (they are re-build testing now)
[14:01] <tgardner> ogasawara, pushed
[14:01] <ogasawara> tgardner: ack
[14:06] <apw> top - 10:06:32 up 14 days, 17:01,  1 user,  load average: 450.34, 296.49, 165.35
[14:06]  * apw whistles quietly
[14:06] <tgardner> apw, tangerine ?
[14:06] <apw> goemeisa
[14:07] <tgardner> I built every release yesterday on tangerine. I imagine it was heating the room.
[14:07] <apw> heh ... i bet ... i am being utterly mean to it, -j512, just to see whether it copes
[14:07] <apw> and its amazingly still up
[14:08] <ppisati> tgardner: i noticed that :)
[14:08] <apw> top - 10:08:08 up 14 days, 17:03,  1 user,  load average: 511.94, 359.70, 200.68
[14:08] <tgardner> apw, pgraner and larry will be at 1SS on Friday at which time I will unf*ck /boot
[14:09] <apw> tgardner, a perfect time to scare the crap out of them by playing tunes on the fans
[14:09] <apw> tgardner, think you can shift the root up a bit, or will you just zap it
[14:09] <tgardner> apw, like they'd be able to hear an individual fan in that place.
[14:10] <tgardner> apw, I'll see if I can shift things around, but failing that I'll just rebuild from scratch.
[14:10] <apw> tgardner, ok then i should make sure the crack builds are pulling off their state properly like they are meant to
[14:10] <apw> tgardner, so i can put it back after
[14:11] <tgardner> apw, yeah, be prepared for total destruction. you _know_ how skilled I am at sys admin :)
[14:11] <apw> tgardner, it all designed with fungible machines in mind, just i've never funged it
[14:12] <apw> i should really kill it all off on gomeisa and just move it to tangerine
[14:12] <apw> to prove it works and is documented
[14:12] <tgardner> apw, well, here is your chance. you could switch to tangerine for awhile.
[14:13] <apw> tgardner, ok friday you say ?
[14:14] <tgardner> apw, that was the last I heard
[14:14]  * apw adds it to his list
[14:15] <dileks> does sth like snapshots.u.c exist?
[14:15]  * dileks wants to downgrade resolvconf
[14:16] <xnox> dileks: it's called launchpadlibrarian
[14:17] <ogasawara> apw: pushed my changes, she's all yours
[14:18] <ppisati> ogasawara: tgardner: patch sent to the ml - in the end i used a =~
[14:18] <tgardner> ppisati, ack
[14:20] <apw> ogasawara, ok pushed overlayfs.v13 update
[14:20] <ogasawara> apw: ack
[14:21] <tgardner> ogasawara, I'll pick up ppisati pae fix
[14:21] <ogasawara> tgardner: ack
[14:22] <ogasawara> apw:  you sure you pushed it?
[14:22] <ogasawara> apw: I'm not seeing it
[14:22] <apw> ogasawara, i failed due to incompetance, try now
[14:22] <ogasawara> apw: much better
[14:23] <tgardner> ogasawara, get your whip out
[14:23] <ogasawara> tgardner: I dunno, he might actually enjoy that
[14:23] <tgardner> kinky englishman
[14:24] <apw> might, how little you know me
[14:24] <ogasawara> hehe
[14:25] <tgardner> ogasawara, ok, I'm gonna whack master-next yet again to cleanup the PAE patch. OK ?
[14:25] <ogasawara> tgardner: ack
[14:26] <arges> ppisati, hello. is there a wiki for pandaboard setup? just got one in the mail, so probably be setting that up sometime
[14:26] <tgardner> ogasawara, pushed. now we should be able to start build and boot testing.
[14:26] <tgardner> ppisati, I assume you've tested your patch on arm ?
[14:26]  * cking upgraded to quantal and sees no breakages
[14:26] <ppisati> tgardner: yep, that
[14:26] <ppisati> s why it took me so long
[14:26] <ppisati> arges: uhm
[14:26] <ogasawara> tgardner: cool, will get things kicked off
[14:26] <ppisati> arges: don't think so...
[14:27] <arges> ppisati, ok just pandaboard.org... read through that?
[14:27] <ppisati> arges: no no, wait
[14:27] <tgardner> ogasawara, yeah, I've got my hoover moaning as well
[14:28] <ppisati> arges: https://wiki.ubuntu.com/ARM/OmapDesktopInstall
[14:28] <ppisati> arges: just dd the img, and you are done
[14:29] <arges> ppisati, perfect thanks
[14:33] <apw> arges, yeah if its harder than dd image -> sd card, insert and power on, you are doing it wrong
[14:35] <dileks> xnox: and how do I get the diverse sources? as "http://librarian.launchpad.net/ is a file repository used by Launchpad. "?
[14:35] <tgardner> cking, do you have a non-PAE atom in your collection ?
[14:36] <arges> apw, yea much easier than the other embedded boards i've played with
[14:36] <cking> tgardner, yep, I can dig one out of the pile
[14:36] <tgardner> cking, it would be good to test the upgrade path from precise -generic  to quantal. 
[14:37] <tgardner> i wanna make sure it fails.
[14:38] <cking> tgardner, just checked, all my old atoms are N270s, which by default run -pae
[14:38] <tgardner> cking, ok
[14:39]  * cking wonders if there is another way to test, how about in virtual box?
[14:39] <tgardner> cking, dunno, I've not used VB
[14:41] <tgardner> ogasawara, I'm still getting build failures in perf, so it looks like it may really not be parallel build safe.
[14:41] <smb> Hm, I think with kvm you can influence the cpu emulation...
[14:42] <cking> looks like VB has PAE enable/disable mode in the config
[14:43] <tgardner> ogasawara, I pushed master-next with a patch to disable $(conc_level) when building perf.
[14:44] <smb> cking, Using virt-manager with kvm seems to allow to set some cpu type and you can change cpu flags... Never tried that, though
[14:49] <xnox> dileks: you need to manually work through http://launchpad.net/ubuntu/+source/package find the binary build you want and download individual debs
[14:51] <dileks> hmm
[14:51] <dileks> https://launchpad.net/ubuntu/precise/amd64/resolvconf
[14:51] <dileks> looks like this gives me an overview of all versions
[14:54] <cking> hrm, can't seem to install precise-i386 with pae disabled in virtualbox
[14:55] <apw> cking, heh you won't be able to, the CD is pae
[14:55] <apw> i think the only official way to get there is to install oneiric i386 and upgrade
[14:56] <cking> doh, yep, /me downloads oneiric -i386 and tries
[14:56] <jsalisbury> apw, just let me know if you want me to help bisect bug 994870
[14:56] <ubot2> Launchpad bug 994870 in linux "kernel panic on boot with hyper-v" [Critical,Confirmed] https://launchpad.net/bugs/994870
[14:56] <apw> jsalisbury, ack though it needs some special patch handing i think and as we have nothing to test on right now ...
[14:56] <jsalisbury> apw, ok.  I'll keep a close eye on it.
[14:59] <ogasawara> tgardner: ack
[15:01] <tgardner> ogasawara, ok, mine built OK this time
[15:02] <ogasawara> tgardner: will re-kick off my builds here
[15:07] <tgardner> ogasawara, I'm bisecting a backlight regression on my Dell 1710 that occurred somewhere between -rc5 and -rc7. Once I know what it was I'll work with upstream to get it fixed.
[15:15] <ogasawara> tgardner: ack
[15:15]  * ogasawara back in 10
[15:19] <apw> ogasawara, man he can eat fast some days
[15:21] <tgardner> apw, its these new management techniques she's been learning
[15:23] <cking> parallel processing perhaps?
[15:27] <apw> cking, shoving half in each end perhaps
[15:27] <cking> apw, one can only speculate...
[15:28] <apw> cking, its likely illegal to do more
[15:29] <cking> :-/
[15:30] <smb> One can speculate about some rolling on the floor going on somewhere... :)
[15:57] <tgardner> ogasawara, quantal build succeed ?
[15:58] <ogasawara> tgardner: looking good so far... powerpc is always the longest to get confirmation
[15:59] <tgardner> ogasawara, thats likely to be OK. I don't think we changed any arch specific stuff, did we ?
[16:00] <ogasawara> tgardner: only big changes were the overlayfs bits but I don't believe that has anything arch specific
[16:00] <ogasawara> tgardner: have we been able to get any test confirmation for non-pae hardware?
[16:01] <tgardner> ogasawara, I think cking is working on a VB test
[16:01] <cking> yep, it's taking a while to install and upgrade in VB :-(
[16:02] <smb> afternoon went where the morning did do just before...
[16:03] <tgardner> smb, I'm gonna add you as s-o-b or acked-by on these xen-acpi patches, then push 'em
[16:04] <smb> tgardner, I should be as s-o-b on those I pushed already
[16:04] <tgardner> smb, right
[16:04] <cking> hrm, "About 4 hours remaining" on the upgrade does not bode well
[16:04] <tgardner> smb, several are cherry-picks from Konrad
[16:05] <smb> tgardner, Ah right, I did not change the s-o-b unless I meddled with one since the version you had
[16:05] <smb> So yeah, feel free to add me to the remainder
[16:06] <tgardner> smb, I'll make sure you're in the chain of responsibility :)
[16:06] <smb> tgardner, I am sure you do :)
[16:07]  * smb really likes the 20W less power usage on idle for his AMD box...
[16:08] <cking> smb, calculate how power is saves and turn it into beer money
[16:08] <cking> s/how/how much/
[16:11]  * smb looks at one additional pint a month...
[16:11] <smb> If my maths are not too much influenced by desire
[16:18] <tgardner> ogasawara, re: backlight regression, -rc5 good, -rc6+ bad
[16:19] <ogasawara> tgardner: anything obvious from the changes between the two?
[16:19] <tgardner> ogasawara, not really. I'm gonna start a real bisect
[16:20] <tgardner> there are a half dozen gpu patches, but I'm not gonna make any assumptions
[16:29] <sforshee> tgardner, have you determined which backlight interface is being used on the machine? That might help limit your bisection
[16:30] <tgardner> sforshee, haven't looked
[16:30] <tgardner> that would require some assumptions
[16:31] <sforshee> how so?
[16:32] <tgardner> sforshee, it assumes the regression is under drivers/gpu right ?
[16:32] <tgardner> it might be a platform issue
[16:32] <sforshee> tgardner, not necessarily
[16:32] <sforshee> you have to check which of the /sys/class/backlight/*/brightness files is changing when you change the backlight
[16:33] <sforshee> it might not be a gpu backlight driver
[16:33] <sforshee> in which case looking under drivers/gpu is pointless
[16:33] <tgardner> sforshee, which is why I'm doing an unrestricted bisect
[16:34] <tgardner> sforshee, acpi_video0 is the backlight class
[16:36] <sforshee> tgardner, nothing in your range from drivers/acpi or drivers/platform/x86 looks related, so I guess the unrestricted bisect probably is your best bet
[16:37] <sforshee> I have a hard time imagining what caused the regression
[16:49]  * ppisati -> gym/workout
[16:55] <tgardner> sforshee, looks like its homing in on nouveau, which is not really surprising
[17:27] <arges> tgardner, hey tim. if i want to test wireless-n performance, is iperf a good tool to use?
[17:27] <tgardner> arges, tahts what I use to sort of smoke test it
[17:27] <tgardner> to smoke*
[17:27] <arges> tgardner, ok just default flags/window size?
[17:28] <tgardner> arges, yep. to make it really whine, though, try a bi-directional test
[17:29] <arges> tgardner, ok i'll give that a shot.. i just can't seem to get more than 50Mb/s on 802.11n using iperf on a couple of machines
[17:30] <arges> both are 2x2s, and I think the router is also a 2x2..but need to make sure
[17:44] <apw> herton, where are we in the precise SRU process
[17:45] <herton> apw, testing week
[17:45] <tgardner> sforshee, I bought a copy of the 802.11-2012 spec for $5 from IEEE. I think its got all of the amendments applied, so its a lot more readable.
[17:45] <herton> apw, new kernels should be prepared next week
[17:46] <apw> herton, i am likely to have a patch which needs to be out as soon as humanly possible
[17:46] <herton> bjf, henrix ^
[17:46] <arges> tgardner, going in the other direction (wired to wireless computer) results in a higher bandwidth reading, not sure if that's normal
[17:46] <apw> ok that sounds ok then i think, this one should be out before we have confirmation
[17:47] <tgardner> arges, I typically use only one wireless station and a wired server
[17:47] <apw> herton, i may ask you to slide out a very thin kernel update in between this and next cycle, so lets circle before you build the next one if thats ok
[17:47] <tgardner> arges, remember that its a half duplex medium
[17:47] <apw> hallyn, hopefully it won't be necessary, and your normal cycle will be quick enough
[17:47] <apw> herton, ^^
[17:47] <arges> tgardner, yea need to probably shut off other devices
[17:47] <apw> herton, so see this as just a head up for now
[17:48] <herton> apw, ok, we still have this precise kernel on testing, bug 991925
[17:48] <ubot2> Launchpad bug 991925 in linux "linux: 3.2.0-24.38 -proposed tracker" [Medium,In progress] https://launchpad.net/bugs/991925
[17:48] <apw> herton, cirtinaly yell if you are respinning for anyting on this update
[17:48] <herton> apw, so we will have to see how to do it if it doesn't get to updates in enough time
[17:51] <apw> herton, indeed.  i think we have time, but we can evaluate before you next cycle
[17:52] <tgardner> sforshee, bisected to b99da31ed8521eb78d5d6930f3128f8ecdb75fae drm/nv10/gpio: fix thinko in mask for gpio lines 2-9
[17:52] <tgardner> am reverting to confirm
[18:01] <greearb> hello!  Back to hacking on casper & USB persistence...anyone around that might know why /cow isn't found once booted?
[18:02] <arges> tgardner, ok yet another wireless question. Running lucid with backported kernel . it needs the broadcom sta wireless restricted driver (bcm4321) . is there a package i need to install, or repository i need to enable?
[18:03] <tgardner> arges, I doubt that is a configuration taht is supportable. you'll need the dkms driver from the backport release.
[18:05] <arges> tgardner,ok thanks
[18:19] <tgardner> ogasawara, commit b99da31ed8521eb78d5d6930f3128f8ecdb75fae causes my regression. email sent upstream.
[18:21] <ogasawara> tgardner-lunch: ack
[19:07] <tgardner> arges, re bug #953494 - in which releases do you want linux-firmware-nonfree uploaded ? Lucid and Precise ?
[19:07] <ubot2> Launchpad bug 953494 in linux-firmware-nonfree "bcm_wimax module hangs when Sierra Wireless 250U Data Card removed" [Undecided,New] https://launchpad.net/bugs/953494
[19:07] <arges> tgardner, yes lucid and precise would be best
[19:08] <tgardner> arges, ack
[19:08] <arges> tgardner, thanks
[19:14] <sforshee> tgardner, I just saw your comment about the 802.11 spec in my backscroll. $5 is cheap, I'm going to pick up a copy too.
[19:14] <tgardner> sforshee, thought you might be interested
[19:24] <HFSPLUS> Katy perry baby, i know your hurting right now after that prick russel brand broke your heart, now all i ask is to give me a chance to prove that i love you. Honey i am nothing like russel brand, and i love you and i will never break your heart katy
[19:26] <mjg59> I'm not sure I can easily enumerate the number of things that have just gone wrong here.
[19:27] <tgardner> kind of makes your head hurt
[19:37] <tgardner> arges, upload linux-firmware-nonfree for Lucid/Precise. Feel free to annoy an AA at your leisure. Without doing so the multiverse approval queue may not get examined for awhile.
[19:38] <arges> tgardner, whats an AA in this context
[19:38] <tgardner> archive admin
[19:38] <arges> ok ok...
[19:39] <arges> thanks
[19:39] <tgardner> arges, the list of folks to annoy are in https://wiki.ubuntu.com/ArchiveAdministration
[20:22]  * tgardner -> EOD