/srv/irclogs.ubuntu.com/2012/03/19/#ubuntu-kernel.txt

=== smb` is now known as smb
* smb Rebooting08:34
ppisatipulseaudio refuses to collaborate this morning08:38
ppisatior maybe is mumble08:38
* smb hates these "experienced and internal problem" messages. So not saying anything08:42
ppisatiuh08:42
ppisatiand i hatre my usb mic not working08:42
smbppisati, Do you have levels in the sound indicator?08:42
ppisatismb: levels?08:43
smbppisati, Some movement on the level-meter08:43
smbif you go to input and your mic08:44
ppisatiif i raise the sensitivity, yes08:44
ppisatiit's alive08:44
smbOk, so its either tuned too low or mumble needs the restart, killpulse, twidle setting dance we all love so much08:45
ppisaticool08:47
ppisatijust killed pulseaudio08:47
ppisatiand now i've lost sound for every app08:47
ppisatiand i can't even close skype08:47
ppisatiit's stuck in a limbo08:47
smbppisati, sounds like another reboot another lucky draw08:48
* ppisati -> back in 10mins09:35
* ppisati scratches his head10:47
ppisatiperf is part of linux-base and linux-tools-common10:50
Davieysmb: Hey, are you around?11:08
Davieysmb: I'd like to raise bug 898127 as crucial, it seems to be present just by showing extreme IO.11:10
ubot2`Launchpad bug 898127 in linux "system hangs and errors at /build/buildd/linux-3.2.0/arch/x86/kernel/apic/ipi.c:113 default_send_IPI_mask_logical+0xdc/0xf0()" [Medium,Confirmed] https://launchpad.net/bugs/89812711:10
DavieyI believe it's that bug anyway...11:10
DavieyI believe i can get you access to hardware where you can see it.11:10
apwDaviey, i think you just missed him popping to lunch11:10
Davieyapw: Oh :(11:10
apwppisati, yeah linux-base is on my todo for today to sort that out, its a debian sync11:13
ppisatiapw: well, actually there's even a skew between kernel version and perf version11:15
ppisatiapw: dunno if it's the same11:15
ppisatiapw: http://paste.ubuntu.com/890466/\11:16
ppisatiand 'perf top' shows nothing11:16
apwppisati, where did you get your 1409 build ?11:17
ppisatiapw: ti-omap4, latest kernel11:17
apwbuilt where11:17
ppisatigot it from dist-upgrade11:17
apwarmel or hf ?11:17
ppisatiarmhf11:17
apwdpkg -S /usr/bin/perf says what11:19
ppisatilinux-tools-common11:19
ppisatiflag@panda:~$ dpkg -S /usr/bin/perf11:19
ppisatilinux-tools-common: /usr/bin/perf11:19
apwppisati, try installing linux-tools instead11:19
ppisatiit's there iirc11:19
apwlinux-ti-omap4-tools-3.2.0-140911:19
ppisatiah11:20
ppisatiuhm11:20
apwis the binary package, but i'd expect the linux-tools to bring it in11:20
ppisatinope11:20
ppisatilinux-tools brought in linux-tools-3.2.0-1911:20
ppisatilet me double check11:20
apwoh hmmm ... yeah11:20
apwmaybe there isn't a linux-tools for ti-omap411:20
apwit would have to have a different name anyhow11:20
ppisatik11:20
* ppisati goes researching the solution11:21
apwactually isn't there a linux-tools-omap4 or something11:21
apwi remmber us having to have a differnet name, and realising we should have had a flavour specific tools name even though it would have been a noop on x8611:21
apwas if you have multiple source packages the linux-tools have to differ in name11:21
ppisatiso, linux-tools tries to install linux-tools-3.2.0-1911:21
apwyep, thats x86 or 'from master' speciific11:22
ppisatiah11:22
ppisatilinux-tools-omap411:22
apwis there a linux-tools-omap4, thats the right one11:22
apwand there is a 'bug' that we don't have linux-tools-generic etc to match it11:22
ppisatianyway, linux-tools-omap4 imported the right one: linux-ti-omap4-tools-3.2.0-140911:22
apwok good... will add to my todo to look at that11:23
ppisatibut perf still doesn't work11:23
ppisatinice11:23
ppisatii've a new toy11:23
ppisatithanks11:23
apwheh whats it do11:23
ppisatinothing11:23
ppisatiPerfTop:       0 irqs/sec  kernel:-nan%  exact:  nan% [1000Hz cycles],  (all, 2 CPUs)11:23
ppisatimute11:23
ppisatisilent11:23
apwhrm, thats a worry, config perhaps ?11:24
ppisatiprobably11:24
ppisatii'm over it11:24
* ppisati -> out for lunch11:35
apwsmb, i've talked to Daviey about his issue, its actually a different bug12:20
smbapw, Hm, a different bug than what? (/me back now)12:52
apwthe one daivy was making you look at12:52
DavieyYes, sorry - i suck.12:54
smbAh ok, so anything I still should look at or are we waiting for him to file information12:54
gemahi guys, couldn't find this in the normal debug symbols apt source: linux-image-3.2.0-17-generic-dbgsym , could someone help there? 12:55
apwgema, may have expired as that is an old kernel12:56
* smb wonders whether the bug he was referring to is offlining cpus as of poewnap12:56
gemaapw: it is the beta1 kernel12:56
apwyes that was the one he was referring to, but the machine is ooming12:56
smbotherwise ddebs.archive.ubuntu.com I think12:56
apwgema, yep but we have -19 in the archive now12:56
apwso if -17 was >14 days ago then it may have gone12:57
smbapw, or was it just ddebs.ubuntu.com12:57
smbIts a pita that any variation exists but some are not good12:57
gemaapw: so you recommend updating ? the problem appeared when rebooting after an alternate install on a production machine, we cannot really afford to reinstall to reproduce.. :/12:58
apwi'd not recommend running a production box on beta1 thats for sure, we're nearly a month on from there12:59
gemaapw: I know, I convinced my other half to use beta 1 for his purposes, because it does what he needs, and we bumped on 3 defects during the install/reboot12:59
gemaapw: so we are trying to raise the kernel one, but without the symbols we only have a picture of the crash13:00
apwgema, so if you can let me see the picture i might recognise it13:00
apwnot sure i would ever install an old snap of precise tho. given we make CD daily13:00
gemaapw: our daily images are not half as tested as beta 113:01
gemaapw: take my word on that one13:01
apwi was under the impression our testing was daily and automated, but hey13:01
apwbut as all the machines i care about are running the crap in the archive right now, and seem good, that'd be my recommendation if you are using P13:01
apwotherwise i'd say O or L was a better plan13:02
gemaapw: ack13:02
gemaapw: there is a lot of manual and varied community hw testing that happens during Beta13:03
gemathat doesn't happen daily13:03
gemanor automated13:03
gemawe are on a road there, but not there yet13:03
gemaapw: sent you that email13:04
apwgema, so this is an EFI machine ?13:06
gemaaye13:06
gemaapw: fwiw, grub was screwed when we tried to reboot, that's another of the problems13:08
smbapw, we can hear you13:08
=== bladernr_afk is now known as bladernr_
apwgema, ok this is looking like an EFI triggered bug, is this machine an EFI bios?13:20
apwif it is we might not expect L to work either on it, but O probabally can, we think13:20
gemaapw: it is an EFI but it emulates the normal bios behaviour13:21
gemaapw: the machine is quite happy after that reboot 13:21
apwhmmm for those efi routines to be being used, that implies its in EFI mode i think ... cking <<13:21
ckingyep definitely in EFI13:21
gemaapw: by L you mean P ? 13:21
apwgema, no i mean L, if tis an EFI mode bios then it probabally won't run L in native EFI mode13:23
apwand i suspect its in that mode from the panic13:23
gemaapw: not sure what the kernel was using at that particular moment, but the normal bios behaviour was there too, as the machine runs DragonFly with no problems, and that wouldn't happen if it was in EFI mode13:24
mjg59Most EFI systems will EFI boot if the media supports it, and fall back to BIOS automatically if it doesn't13:25
mjg59The Ubuntu install media includes EFI support13:25
mjg59So if you tried to boot Dragonfly it'd fall back to BIOS, but if you tried to boot Ubuntu it'd be in EFI13:26
gemamjg59: ack, understood13:26
ckinggema, efI_pstore_write() is failing for you - but w/o seeing the top of the oops it's impossible to say way efi_call5() fails13:28
gemacking: ack, alex is willing to debug it if he finds / can get the debug symbols13:29
ckinggema, yep, with a bug filed please 13:30
gemacking: not a problem, can you get us the symbols for beta1 kernel?13:30
ckinggema, I will build you one13:33
gemacking: thanks a million13:33
tgardnercking, can you change to tangerine? gomeisa is getting relocated.13:39
ckingtgardner, ack13:39
tgardnerogasawara, gomeisa will be down for a bit for relocation13:40
ogasawaratgardner: ack13:41
ogasawaratgardner: where's it relocating to?13:41
tgardnerogasawara, new rack and network switch (I think)13:41
tgardnertangerine will follow in a bit13:41
gemacking: tracking this problem on Bug #959286 13:43
ubot2`Launchpad bug 959286 in linux "EFI related kernel panic on reboot from alternate installer " [Undecided,New] https://launchpad.net/bugs/95928613:43
gemacking: bwalex will be adding his findings there13:45
=== tgardner is now known as tgardner-afk
ckinggema, thanks13:48
=== tgardner-afk is now known as tgardner
apwcking, how long will your build be ... on tang14:30
ckingapw, which one, I'm futzing around doing several14:31
apwcking, how long will you be using tangerine, we want to take ti down for its move, gome is back14:31
ckingI'll log off now14:33
ckingdone14:33
smbapw, 14:34
smbapw, I am using tangerine right now, too14:34
ckingapw, how long will it be offline for?14:35
apwcking, about as long as gomeisa i assume, 10s of minutes14:35
apwsmb, how long :)14:35
smbapw, till now14:36
smb:)14:36
* smb got his kernels off14:36
apweveryone happy ?14:36
apwsmb, cking, got waht you need off her14:36
ckingapw, I was going to pull off a .ddeb, so it's going to take an insane length of time, so just go ahead14:36
smbapw, I am good14:37
apwcking, it is14:37
apwcking, remember you can scp it to zinc14:37
tgardnercking, or even gomeisa14:38
* ogasawara back in 2014:39
ckingok - done14:39
henrixdiwic, Sarvatt: ping14:50
diwichenrix, hi14:50
henrixdiwic: hi! would you be able to help me with bug #956479 ?14:50
ubot2`Launchpad bug 956479 in linux "no analog audio output after HDMI display standby with kernel 3.0.0-17.30" [Medium,Confirmed] https://launchpad.net/bugs/95647914:51
henrixdiwic: it looks like a regression14:51
henrixdiwic: could you please take a look at it, before i ping lkml (and others)?14:51
diwichenrix, so the issue is just that hdmi becomes auto-selected when it previously wasn't, not that anything stops working permanently?14:53
mdeslaurcking: are you aware of bug 952556? could it be related to the power savings work you've done?14:53
ubot2`Launchpad bug 952556 in ubuntu-meta "[Precise] [Hardware-killer] HD restarts every few seconds" [High,Confirmed] https://launchpad.net/bugs/95255614:53
ckingmdeslaur, lemme look14:54
henrixdiwic: yes, that was my understanding. everything seems to work fine14:54
mdeslaurcking: could you hand that off to someone appropriate if you're not the right person...the bug does look important14:54
henrixdiwic: i reverted one of the patches that came from stable, and the problem seems to be solved14:54
diwichenrix, so the issue is probably that from the audio driver, it looks like the monitor was unplugged and replugged14:55
diwichenrix, and in fact, today I'm working on not switching to HDMI for 12.0414:55
henrixdiwic: ok, so you believe the issue is at the audio driver14:56
diwichenrix, so it should hopefully be resolved for 12.04. Anyway, to confirm this theory, you can try commenting out "load-module module-switch-on-port-available" in /etc/pulse/default.pa14:57
diwichenrix, naah, it's probably in the video driver, it shouldn't tell us the monitor is not connected I guess, but at the same time, that's not easy to know14:57
henrixdiwic: hmm... ok, i can ask the reporter to try to comment that and see what we get14:58
diwichenrix, aha, I thought you were the reporter, sorry14:59
henrixdiwic: ah, no. sorry, should have made that clear :)14:59
henrixdiwic: thanks for you hints. i'll try to investigate this a little bit more15:01
diwichenrix, well, so this seems to have made it into 3.315:01
diwicso this commit is not in precise, correct?15:02
henrixdiwic: i haven't check that, but it should be15:02
henrixdiwic: it is reported against oneiric15:02
bjfhenrix, diwic, yes, that commit is in precise15:03
diwicbjf, thanks15:03
henrixyeah, it just. just checked as well15:03
diwichenrix, so I care mostly about 12.04, in which case this would be resolved by the Pulseaudio patch I'm writing today15:05
diwichenrix, for 11.10; well, dunno if it's causing a lot of problems, revert it15:05
henrixdiwic: well, not sure about reverting it. we would need to know what problems it would cause15:06
diwichenrix, as for reporting bugs upstream; it could be worth having the discussion I guess, but it might backfire on us loading the module-switch-on-port-available in the first place15:07
diwichenrix, which is related to the jack detection stuff I've been doing.15:07
henrixdiwic: ah, ok.15:08
henrixdiwic: would that pulse patch be backported into oneiric?15:08
henrixdiwic: or there are no plans for that?15:09
diwichenrix, so question is if this actually a bug fix or rather a change in behaviour15:09
ckingmdeslaur, thanks for the heads-up on that bug, I'm on to it right now15:09
mdeslaurcking: cool, thanks!15:09
henrixdiwic: yeah, i understand what you mean15:09
diwichenrix, I'm not currently planning to backport it, but I don't know...I'm okay with backporting it myself if there is demand for it.15:10
diwichenrix, but that PulseAudio change is also a change in behaviour more than a bug fix15:10
henrixhenrix: ack15:10
diwichenrix, maybe I should ask Wu about what this patch actually resolves?15:11
henrixdiwic: that patch came in from upstream stable. i would need to spend some time looking at the stack in order to actually understand the changes introduced15:12
diwichenrix, I tend to like the patch.15:14
diwichenrix, at least judging from the commit text.15:14
henrixdiwic: well, yeah... but i am unable to measure the consequences just looking at it :)15:15
diwichenrix, it's just a little stupid that we can't tell the difference between a monitor being plugged in, and being returned from standby.15:15
henrixdiwic: so, you already helped me a lot by explaining that this is more a change in the behaviour15:16
henrixdiwic: so, i'll take a better look at the whole thing and try to have some conclusion about all this15:17
diwichenrix, ok15:18
henrixdiwic: and thanks!15:18
dholbachhiya16:06
dholbachcan anyone reply to https://twitter.com/#!/MttCastelli/statuses/181737540187987968 maybe?16:06
tgardnerdholbach, besides the fact that he's running a 3.3 kernel (Precise is 3.2), the Nvidia DKMS driver is owned by OEM.16:09
dholbachsure, I wasn't trying to make this your problem - just wondering if anyone had a reply for the guy16:10
tgardnerdholbach, I don't think any of us tweet16:11
tgardnerdholbach, feel free to paraphrase my response16:11
dholbachI'll see if I can find anyone at OEM who knows what's going on - thanks16:11
tgardnerdholbach, contact tseliot16:12
dholbachyep16:12
dholbachaccording to http://www.nvnews.net/vbulletin/showthread.php?p=2536982 "This issue got fixed and available in next 295 driver."16:13
dholbachI'll mention that to Alberto too16:14
apwdholbach, mainline kernels are not even supported, plus there is a known issue that binary modules do not work with them ...16:44
Sarvattdholbach: it's not compatible with the 3.3 kernels yet, its a really common occurance. the next update we do in precise should be :)16:47
* cking --> in deep thought mode16:47
brendandbjf, bryceh - we have one certification system that can't start x with the new -proposed kernel17:02
bjfbrendand: what's the bug # ?17:02
brendandbjf, bryceh - the tester is telling me that it seems not be related to the kernel itself entirely though17:02
brendandbjf - being raised right now, i'll get you it asap17:02
brendandbjf, bryceh - https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/95949417:06
ubot2`Launchpad bug 959494 in xorg "[HP Probook 6550b] Fails to start graphical environment" [Undecided,New]17:06
brendandbjf, but booting with the -updates kernel also doesn't work17:07
bjfbrendand: well, unfortunately, that makes it _not_ a regression (at least not a regression for this kernel)17:08
brendandis this an argument for including other packages in the kernel SRU cadence?17:09
brendandlike x for example?17:10
bjfbrendand, i don't think so. i think it's a case for better regression testing of other packages.17:11
brendandbjf - that's sort of what i was trying to say, maybe they don't need to follow the same cadence, but the same process17:12
bjfbrendand, yes, i agree that other packages should follow a similar process17:13
brycehbrendand, how has he determined it not to be related to the kernel?17:14
brycehbrendand, if it's an X issue we'll have to see the Xorg.0.log17:14
brendandbryceh, by booting into the -updates kernel. isn't that good enough?17:15
brendandbryceh, it's just an intel chipset too so i'm wondering what other packages could be involved in the regression17:21
brycehbrendand, very little has changed in X on lucid, so I was just asking if he had any better evidence17:21
brendandbryceh, yeah. i always remember you telling me that17:22
brycehbrendand, typically this class of issue is most commonly caused by an error in the xorg.conf, but can also happen if an X driver got uninstalled, or was broken for some reason17:22
brendandbryceh, well - we have a very strictly controlled install process so user error is unlikely to occur17:23
apwbryceh, yo ... did you manage to do any further testing on that edid patch?  i'd want to be commiting things like that right after beta if we're going to have it, as kernel freeze is RSN17:23
brycehin any case, generally an error will get written to Xorg.0.log or gdm.log in this type of failure.17:23
brendandbryceh, so perhaps one of those things happened, but it would most likely have been one of the updates that did it17:23
brendandbryceh, so far we:17:24
brendandinstalled (PxE)17:24
bjfbrendand: we need the log files17:24
brycehapw, heya, yep top of the todo list17:24
bjfbrendand: that bryceh has asked for17:24
brycehapw, last week was a bit crazy with people needing stuff looked into17:24
apwbryceh, ahh cool ... /me wonders if that needs any shining ... i should look17:24
apwbryceh, they are like that people, always wanting things ... good job i am a robot instead17:25
brycehapw, I should turn robot too!17:25
brycehbrendand, also the dpkg.log might be useful in tracking down which packages had gotten updates.  Surprised that isn't already included in your apport hook.17:26
brendandbjf, i know you do - he's just stepped out for lunch though17:26
bjfbrendand: ok, i think further discussion can wait until we see the log17:26
tgardnerogasawara, I have 3.2.12 prepped. shall I push it ?17:36
ogasawaratgardner: yes please17:36
tgardnerogasawara, still doing build tests, but I think it'll be OK. pushed...17:37
tgardneror rather, pushing....17:37
ogasawaratgardner: is it an ABI bumper?17:37
ogasawaratgardner: we could squeeze in an upload before beta-2 freeze this thurs17:38
tgardnerogasawara, dunno yet17:38
tgardnerogasawara, gimme 10 mins17:38
apwtgardner, ogasawara, infinity has confirmed that a UP build of the same commit as your last upload boots ok, the smp only powerpc build does not on his H/W17:41
apwsmb, do your menus appear ok now, or do you still get the partial versions ?17:42
apwsmb, as that bug got closed fixed, adn i can't recall when i last noticed them17:42
ogasawaraapw: hrm, so that would contradict our beliefs that smp should work on non-smp hw17:42
smbapw, No, that seemed to have been fixed around when they closed it17:42
apwogasawara, for one machine at least yes17:43
apwnot sure if that warrent backing it out for beta and some more investigation17:43
apwthere is almost no obvious difference between the configs for the two other than the CPU count17:43
ogasawaraapw: I'd lean towards backing it out until we find a fix.  tgardner thoughts?17:46
tgardnerogasawara, nak from me. lets find out if there is more then 1 machine like it left in the world. I'm tired of supporting antiques.17:47
tgardnerogasawara, no ABI bump on amd6417:53
ogasawaratgardner: ack, I'll prep an upload17:53
ogasawaratgardner: tgardner: for powerpc, I'll ship with non-smp removed for Beta-2 to see if we can flush out any other broken systems.  but I think if we can't find a solution for infinity and others (if any), we may have to reinstate it based on the same argument against dropping generic i386.  We'll give them at least another 5yrs with the LTS, but drop it in Q.17:54
tgardnerogasawara, you're gonna make me cry.17:55
tgardnerogasawara, back in a sec. I'm boot testing 3.2.1217:57
=== tgardner is now known as tgardner-lunch
* henrix will be out for ~20 mins18:10
ogasawaratgardner-lunch: I have an idea baking to make both parties happy for powerpc, mumble when you get back18:15
jjohansenogasawara: when is the last kernel upload?18:20
ogasawarajjohansen: I'll probably squeeze one in today, but then that will be it until after Beta-2 (assuming no kitten killers)18:21
ogasawarajjohansen: we'll probably be able to squeeze in 1 or 2 more uploads in between Beta-2 and Kernel Freeze (Thurs Apr 5)18:22
jjohansenogasawara: okay, /me doesn't have anything that would affect todays upload18:22
jjohansenogasawara: thanks. I will work on getting a rebase request based on the 3.4 version of apparmor together for you post Beta-218:23
ogasawarajjohansen: cool, if I could ask you try to get them to us before the end of March, that would be ideal.  As I hope to upload our final kernel Mon Apr 2.18:24
jjohansenogasawara: yep that should work, just have to wait for linus to suck in the security-next tree18:24
ogasawarajjohansen: ack18:25
=== tgardner-lunch is now known as tgardner
* apw goes for supplies ...18:35
SaveTheRb0tzHi. I've just noticed that I can't build 3.2 precise kernel in lucid chroot without porting new dpkg >= 1.16.0 and dpkg-dev (for new dpkg-architecture -qDEB_HOST_MULTIARCH).. Is there some workarounds or fixes planned?19:12
* apw goes fine beer19:30
tgardnerSaveTheRb0tz, we've not put any effort into supporting that configuration. build in a Preicse schroot, or on a Precise host.19:30
apwSaveTheRb0tz, i am supprised they don't build ... we build mainline builds in lucid chroots with precise packaging19:36
=== yofel_ is now known as yofel
* ogasawara lunch20:10
* tgardner -> EOD20:14
* cking --> EOD too20:25
cking:-( my ThinkPad is dead20:30
vanhoofcking: the x220?20:31
ckingyep20:32
ckingdid "reset BIOS settings to default, F10 save", reboot, completely bricked20:33
ckinguttter pants20:33
vanhoofcking: :\20:33
vanhoofcking: under a year old?20:33
* vanhoof assumes so20:33
vanhoofthey came out about this time last year20:34
vanhoofthey were really easy to work w/ when my lcd went bad20:34
vanhooffixed it in ~3 days20:34
ckingvanhoof, yep, 3 months old20:34
ckingif that20:34
ckingtomorrow's problem20:35
=== bladernr_ is now known as bladernr_afk
SaveTheRb0tzapw: from my little research lucid's dpkg doesn't know anything about multiarch... So it's pretty strange that you somehow got it built on lucid.21:06
brycehapw, the edid write thing tests out well on my hardware.  is there a chance of it being included in precise?21:08
* cking gives up for the day21:59
infinityogasawara: That was the most formal email anyone has ever written to me when our previous conversation involved discussing *censored* and *censored* over beer. :P22:37
mjg59infinity: There's something about you that encourages people to treat you like a gentleman22:39
infinitymjg59: Everyone but you?22:39
pbuckleylol22:41

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!