/srv/irclogs.ubuntu.com/2011/02/11/#ubuntu-kernel.txt

persiaI don't know best practices for kernel rename (kernel packaging confuses me), but I believe you'd want to rename your headers, your -meta, etc. to match, as well as all the relationships.  There may be some packaging helper scripts that need adjustment.00:04
bankixUh.00:07
bankixThat sounds heavy00:08
bankixI thought I could get away with a simple "provides: linux-*"00:08
persiaThat's risky, because your meta will depend on the highest version of linux-*, and versioned provides don't work, and so when then next ABI change happens, you'll be offered an upgrade anyway.00:09
persiaPlus, your headers may not match, which could break building modules.00:09
bankixOK, no renaming.00:10
bankixBut how to avoid updates?00:11
bankixsetting it to "hold" in the package database?00:12
persiawon't prevent offers00:13
bankixThen I'm helpless.00:14
bankixBut thanks for your help00:18
bankixI think the safest but hardest way would be to rename the package. This would prevent all offers and updates, except I offer an update myself.00:19
persiaYes.00:20
bankixOK, something to do next week :-)00:21
bankixAnd I'll set up a quad-core tomorrow to get that compile batch done ...00:26
bankixThanks again and good night.00:28
brycehbjf, apw, JFo: new lpltk 0.4.2 uploaded to natty00:31
brycehugh, my natty desktop is running into some kernel BUG - bug #71681101:13
ubot2Launchpad bug 716811 in linux "[SandyBridge] kernel BUG at /build/buildd/linux-2.6.38/fs/nfsd/nfs4state.c:3132!" [Undecided,New] https://launchpad.net/bugs/71681101:13
brycehJFo, ^^ let me know if there's any more info I could gather that would be of help diagnosing it01:14
jk-wow, that looks like a bogus BUG_ON01:17
* jk- fetches newer git tree01:19
jk-ah, the bogusness has been fixed01:26
brycehJFo, #713864 may be worth adding to the kernel team's list to track; while it doesn't yet have a patch worth pulling it looks like they're narrowing in on one.01:46
brycehJFo, dang it, I meant bug #711591 there.01:47
ubot2Launchpad bug 711591 in nouveau "does not detect resolution correctly(xserver-xorg-video-nouveau causes graphic corruption were you cant do anything)" [Critical,Confirmed] https://launchpad.net/bugs/71159101:47
=== smb` is now known as smb
al-maisanHello, upgraded to the 2.6.38-3.30 kernel this morning and laptop (thinkpad t410) froze twice on resume from suspend (nothing in /var/log/syslog) .. is this a known issue?08:54
al-maisanFWIW the 2.6.38-2.29 kernel does not exhibit this problem08:54
al-maisanthe following upstream kernel change might be the "culprit": drm/i915/lvds: Restore dithering on native modes for gen2/308:55
smbal-maisan, Given that the kernel was uploaded only yesterday there may not be "known" too much. But I don't follow those things not that closely. I could potentially build you a kernel with that one patch removed to test this for sure08:58
apwal-maisan, no it is not known08:58
apwsmb, i don't think that commit is very likely08:59
al-maisansmb: ah, I see .. well, if you make such a kernel available I would certainly give it a try.08:59
apwit changes a dithering order bit in the GPU ... removing some banding08:59
apwal-maisan, why do u pick that one commit as likely ?08:59
apwthe word restore in that commit is not related to resume from suspend08:59
al-maisanapw: I am just guessing .. I had i915 related resume freezes with the maverick kernels .. burnt child syndrome :P09:00
apwhmmm 09:00
apwal-maisan, can you confirm that this is repeatable, that it is every resume with the new kernel ?09:01
al-maisanit occurred twice this morning .. let me reboot into 2.6.38-3.30 and try it one more time09:01
al-maisanbiab09:01
=== al-maisan is now known as almaisan-away
=== almaisan-away is now known as al-maisan
al-maisanapw: ok .. the resume freeze is "reliable" and repeatable .. tried it two more times in a row .. froze both times09:40
al-maisannothing in /var/log/syslog09:40
al-maisanresumes with the 2.6.38-2.29 kernel still work fine09:41
apwal-maisan, hmmm09:44
apwi'd ask you to file a bug, but ... launchpad is in a heap09:44
al-maisanouch ;)09:45
apwal-maisan, assume this is an x86 of some sort09:45
al-maisanx86_6409:45
apwok i have a 64bit install on the -3.30 kernel which is ok ... hmmm09:48
apwal-maisan, what graphics do you have ?09:49
al-maisanI have an integrated i915 and a discrete nvidia chip (which I turned off via the bios)09:49
apw831d52b x86, mm: avoid possible bogus tlb entries by clearing prev mm_cpumask after switching mm09:57
apwal-maisan, you have 64 bit yes?  if i do you a test kernel you available to test it ?10:02
al-maisanapw: yes and yes10:02
=== cking is now known as cking-afk
=== al-maisan is now known as almaisan-away
apwalmaisan-away, sorry for the delay ... i lost the first build: http://people.canonical.com/~apw/amd64-resume-natty/11:15
apwalmaisan-away, could you test that and report back please11:16
=== yofel_ is now known as yofel
=== cking-afk is now known as cking
apw_almaisan-away: did you get a chance to test? lost my scrollback12:28
=== almaisan-away is now known as al-maisan
al-maisanapw: will do!13:13
apwal-maisan, hows it lookin13:30
al-maisanapw: I need to finish a piece of work .. will let you know in 20 minutes .. sorry for the delay.13:31
=== Quintasan_ is now known as Quintasan
apwal-maisan, heh no worries, just impatient like all techies :)13:31
* al-maisan knows the feeling ;)13:32
al-maisanapw: booting into the kernel you prepared now ..13:35
=== al-maisan is now known as almaisan-away
tgardnersconklin-gone, why aren't my linux-fsl-imx51 kernels getting pocket copied from the c-k-t PPA ?13:45
rsajdokWhere is correct link to a git repository contains this patch?13:46
rsajdokhttp://kernel.ubuntu.com/git?p=manjo/ubuntu-maverick.git;a=commitdiff;h=refs/heads/lp71217413:46
=== almaisan-away is now known as al-maisan
apwrsajdok, git://kernel.ubuntu.com/manjo/ubuntu-maverick.git i believe13:47
al-maisanapw: that kernel behaved the same unfortunately .. freeze on resume .. not a single peep in /var/log/syslog13:48
apwal-maisan, well thats perplexing13:48
al-maisanyeah13:49
apwal-maisan, can you test the v2.6.38-rc4 mainline build please13:49
apwif its in mainline then i can start a bisect on it13:50
al-maisanapw: where would I find that kernel?13:50
al-maisanthe mainline one13:50
apwhttp://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.38-rc4-natty/13:50
apwthat is in theory the same as the -3 you tested but without the ubuntu sauce13:50
apwif that fails (and i would expect it to then we can bisect -rc3 -> -rc413:51
jpdsWould anyone happen to know what could be causing bug #709457 ? The dkms.log file isn't really helpful.13:51
ubot2Launchpad bug 709457 in oss4 "package oss4-dkms 4.2-build2002-2 failed to install/upgrade: oss4 kernel module failed to build" [Undecided,Confirmed] https://launchpad.net/bugs/70945713:51
apwjpds, nothing there to work with is there13:53
apwjpds, i think you can run the build by hand with like dkms build (and a few version options)13:53
apwand it'll give you the make command it uses13:53
apwthere may also be a longer log in the output directory somewhere13:53
apwjpds, look in /var/lib/dkms for output directories13:55
jpdsapw: More than http://paste.ubuntu.com/565862/ ?13:55
apwjpds yea13:55
apw/var/lib/dkms/bcmwl/5.100.82.38+bdcom/2.6.37-12-generic/i686/log/make.log13:56
apwas an example i have that in mine for a wireless module, and it has the output of the actual make13:56
apwnote your output even tells you to looki at that file13:56
apwConsult the make.log in the build directory 13:56
apw/var/lib/dkms/oss4/4.2-build2002/build/ for more information. 13:57
apweven telling you where to find it13:57
jpdsI looked at /var/lib/dkms/oss4/4.2-build2002/build/main.log, I'll check make.log.13:57
jpdsapw: Actually, http://paste.ubuntu.com/565870/ is the entire contents of make.log.14:00
apwnot much use is it14:01
jpdsNot really.14:01
apwjpds which kernel are you using ?14:02
apwjpds, can you paste cat /proc/version_signature14:03
apwand can you also paste ls -l /lib/modules/2.6.32-28-generic/source/include/linux/limits.h14:03
jpdsapw: 2.6.32-28-generic14:04
apwjpds, and the other two14:06
rsajdokapw: After that: git log --grep="Manoj"14:08
rsajdokapw: I cannot see this patch: http://kernel.ubuntu.com/git?p=manjo/ubuntu-maverick.git;a=commitdiff;h=refs/heads/lp71217414:08
apwjpds, ok i've updated the bug14:11
apwjpds, i am suspicious that the file will not exist ...14:11
jpdsapw: Excellent, thanks.14:12
apwjpds, perhaps you could also check they have the headers installed for their kernel14:13
jpdsapw: I believe that they're also using VirtualBox with DKMS.14:14
apwcould easily be a bad assumption in this dkms package14:14
apwal-maisan, oh could you file a bug with ubuntu-bug linux when running the -3 kernel ... and let me know the bug number14:15
al-maisanapw: will do.14:15
al-maisanapw: when running the ubuntu variety of the -3 kernel. Right?14:16
apwyeah if you could, then it'll put the right version inforamtion on the bug14:16
al-maisanno problem, will do.14:18
=== sconklin-gone is now known as sconklin
=== al-maisan is now known as almaisan-away
rsajdokapw: any sugestions?14:35
tgardnerogra, what is the appropriate list to send ARM kernel upload announcements? ubuntu-mobile has bounced for me.14:43
ograwe usually use ubuntu-devel, but i guess -kernel would work as well, given that all people interested in kernel uploads of the arm team are subscribed there14:43
apwogra, ok they are going there anyhow14:44
tgardnerogra, works for me14:44
apw(kernel-team@)14:44
ograright, i dont think we need separate notification14:44
apwrsajdok, looking myself14:46
apwrsajdok, worked for me14:47
apweither git clone that URL or add it as a new remote in a maverick tree14:47
apwthen you should have like origin/lp712174 in there which is the same as that patch14:47
apwapw@dm$ git show manjo/lp71217414:48
apwcommit 58835356c5542d623a1110ed40a089e83388f4ab14:48
apwAuthor: Jens Taprogge <jens.taprogge@taprogge.org>14:48
apwDate:   Mon Aug 9 23:48:22 2010 -030014:48
apw    thinkpad-acpi: Add KEY_CAMERA (Fn-F6) for Lenovo keyboards14:48
=== almaisan-away is now known as al-maisan
al-maisanapw: could not report the bug "ubuntu-bug linux" says "This is not a genuine Ubuntu package" :/14:51
al-maisanapw: the mainline kernel exhibits the same behaviour FWIW .. freeze upon resume14:51
apwal-maisan, ok thanks, just file a bug against any old runnable kernel then ... sigh14:52
al-maisanoh, the command is right because I am running "2.6.38-3-generic #30amd64resume201102111045" right now.14:53
apwahhh damn yes14:53
al-maisanI think the original -3 ubuntu kernel has been overwritten14:53
* al-maisan looks for it14:53
apwjust use +filebug, and don't worry about apport info14:53
al-maisanok14:54
apwits me who will whine at your for no apport info, and i'll tell me not to14:54
al-maisanhmmm :)14:54
* smb grumbles about annoying LP features15:05
smbtgardner, could you get my own nomination accepted on https://bugs.launchpad.net/ubuntu/+source/linux-ec2/+bug/71717715:06
ubot2Launchpad bug 717177 in linux-ec2 "Import missing stable update changes into Lucid-ec2" [Medium,In progress]15:06
tgardnersmb, done15:06
smbtgardner, ta15:06
=== diwic is now known as diwic_afk
=== al-maisan is now known as almaisan-away
=== almaisan-away is now known as al-maisan
al-maisanapw: please see bug #71718915:23
ubot2Launchpad bug 717189 in linux "thinkpad t410 freezes reproducibly upon resume from suspend" [Undecided,New] https://launchpad.net/bugs/71718915:23
al-maisanapw: I managed to make "ubuntu-bug linux" by reinstalling the ubuntu -3 kernel i.e. you should have full info there.15:23
=== herton is now known as herton_lunch
=== Artir is now known as afk|Artir
=== afk|Artir is now known as Artir
=== al-maisan is now known as almaisan-away
=== bjf[ack] is now known as bjf
tgardnerapw, jjohansen: why are we considering disabling filename encryption in Natty if long filename support doesn't land in time? IMHO we're no worse off then we've been for Lucid/Maverick.15:49
apwtgardner, i don't think we should be, but people in the peanut gallary were shouting and screaming about how bad it is if you use evolution or something something else15:53
apwtgardner, so i take the view that we should review it in the open with the release team.  my recommendation is status quo15:54
tgardnerapw, frankly, it would be a total regression if we disabled the feature.15:54
apwkirkland i think was of the same mind, others differed15:54
apwtgardner, it would have to be for new installs only indeed, else my shit would break15:54
tgardnerabsolutely15:55
apwso i don't think its trivial, or indeed sensible to regress, but it was reuqested, and that decision is above my head15:55
apwkate can figure it out :)15:55
tgardnerapw, perhaps we could make it an installer option?15:55
tgardnerI'd be OK with that.15:55
jdstrandI think what was discussed was new installs would not have it, but upgrades would not change15:56
apwi think to do so we'd need some engineering to mke it optional, jj would know if that exists already15:56
jdstrandit is hard to know what is going to fail15:56
tgardnerapw, FNEK is already optional, its part of the mount command15:56
apwtgardner, oh ok thats good intel15:57
jdstrandfor example, launchpad scripts die and so I have symlinked out of my encrypted home some launchpad dot directories15:57
apwjsalisbury, i assume you are a proponent of turning it off15:57
tgardnerjdstrand, I feel your pain.15:57
apwjdstrand, even15:57
jdstrandtgardner: it is, but there is no way to adjust it via pam and the tools, aiui15:57
jdstrandand once you have it, there is really no easy way to turn back15:57
apwso basically it would be a foundations effort to make it optional15:57
tgardnerjdstrand, no indeed. its a one way street15:57
apwjdstrand, i guessi a cp -rp could be used to fix it15:58
tgardnerapw, yeah, its not really a kernel bug per se15:58
jdstrandapw: sure, that is what you need to do. use a pivot copy15:58
apwa "make new; cp contents, unmount old, mount new'15:58
apwok ... so if we can fix it, then thats good, but i think someone needs to decide15:58
jdstrandbut that is a pretty scary operation to automate for people15:58
apwwhat the if we can't fix it in time fall back is15:58
tgardnerapw, jjohansen already has it working.15:59
apw1) do nothing, 2) disable by default, 3) disable by default and offer migration back15:59
jsalisburyapw, :-)  bad tab completion 15:59
tgardnerat least on xattr enabled file systems15:59
apwjsalisbury, sorry i hate tab completion :)15:59
jdstrandI was always very, very surprised it was on by default with no pam/tools configurability15:59
apwtgardner, the last version i saw still had limitations in the face of long name hard links15:59
jsalisburyapw, I'm a proponent for turning off bad tab completion lol15:59
apwjdstrand, you'd have to talk to kirkland about the default i recon16:00
jdstrandapw: I already did16:00
apwand what did he say :)16:00
apwpthththth ?16:00
jdstrandapw: you do see the current default, no?16:00
jdstrand:)16:00
jdstrandtbh, I don't think the problems were known at the time the default was set16:00
tgardnerapw, even with the hard link limitation, we'd better off then we are now16:00
jdstrandI mean, the limit was known, but not the affect on the users16:01
apwtgardner, though until its upstream we are risking generating incompatible on disk stuff that we have to live with16:01
apwso we need to be really sure what he is doing will go upstream with the same on disk format, before we enable it16:01
tgardnerapw, true enough16:01
jdstrandideally, I would see filename encryption as opt in and configurable via pam. default off in the installer (without an option in the installer). when filename encryption works right, we can default to on in the installer. upgrades don't change16:02
apwrelease meeting ... gah16:02
jdstrands/via pam/via pam and the tools/16:02
=== herton_lunch is now known as herton
jdstrandapw: not sure you saw my opinion above wrt ecryptfs ^16:27
apwjdstrand, yep see it now ...16:27
jdstrandapw: on a totally unrelated note, I noticed this change in the other-n-security-apparmor bp:16:27
jdstrand- Work items:16:27
jdstrand+ Work items for ubuntu-11.04:16:28
jdstrandapw: may I ask why you did that?16:28
jdstrand(ie, should we be doing something differently?)16:28
apwjdstrand, cause the blueprint didn't have a milestone and nor did the items themselves16:28
apwso they fall in the none bucket ... which is bad ... 16:29
apwyou can fix either by setting them in the main blueprint, or as i did therre16:29
apwbut if you do the former someone can just change that and you loose any idea where they were before16:29
apwand all your planning is affected16:29
jdstrandah16:29
apwso ... i generallyu move them to the release milestone16:29
jdstrandI see it is accepted for natty as the series, but yes, no milestone target16:30
apwand hope that the owner will notice and fix :)16:30
jdstrandapw: ok, thanks for the info16:30
apwthose burny charts cause me loads of pain :)16:30
apwsorry if i freaked you out 16:31
jdstrandapw: our team tries our best to do the right things wrt to those, but since we only commit to essential bps, we don't live and die by the burndowns and we may miss process updates, etc16:31
jdstrandapw: please feel free to call me on it if you are messing you up16:31
jdstrandhehe16:32
jdstrandif we* are messing you up :)16:32
apw:)16:32
apwJFo, hey ... can you drop the 'kernel-tracking-bug' bugs from the tag search ... they are no use to the stable team so lets drop them17:38
JFoon it17:38
* apw realises he has been talking to you on mumble ... without you being therre17:38
bjfJFo, i'm looking at "run-hotlist.sh", it would be nice if there was a comment above each line which indicated who asked for that tag to be tracked and what the tag means17:39
JFoapw, heh sorry :-/17:39
JFobjf, apw, do you guys still want the revert and reapply tagged bugs there?17:40
JFoor no?17:40
JFobjf, ok17:40
bjfsconklin, ^ thoughts ?17:41
bjfJFo, there should be very few of them, and i think it would be nice to see, so i guess keep them17:41
sconklinthinking17:41
JFok17:42
sconklinI think we can leave them, although I don't think we'll do anything automated with them any time soon17:42
JFook17:45
bjfjjohansen, you having problems with the mumble server ?17:51
jjohansenbjf: I have been kicked a few times, and sometime it doesn't log in for a while (say 10+ min)17:52
jjohansenso yeah17:52
bjfjjohansen, same here17:52
tgardnermust be something between the states and Canonical17:53
jpdsAn ocean?17:58
=== sforshee is now known as sforshee-lunch
bjfJFo, there appears to be a problem with the script that is generating the report, the timestamp is 17:11 which is roughly 45 minutes ago18:08
bjfJFo, i see that a new hostlist was generated18:08
bjfs/hostlist/hotlist/18:08
JFobjf, it runs every hour18:10
bjfJFo, ah! that explains it, was thinking it was every 15 minutes18:10
JFoheh :)18:10
bjfjfo, still a problem, in "run-hotlist.sh" you are still cat-ing the SRU_TRACKING_BUGS file into HOTBUGS even though you are not updating SRU_TRACKING_BUGS and the old one still exists18:20
JFoah, forgot that bit18:21
JFofixing now18:21
JFofixed18:22
bjfJFo, can you run the scripts by hand so we get a new report ?18:23
JFosure18:23
bjfthanks18:24
JFomy pleasure :)18:25
apwbjf, it takes like 15 mins to run18:26
bjfapw, do we want it to run faster or are we ok with it taking that long18:29
JFobjf, it is up now18:37
bjfJFo, thanks, that cut 14 off the list18:37
JFonp18:39
JFo<-food18:40
=== sforshee-lunch is now known as sforshee
* tgardner --> lunch18:57
* bjf -> lunch19:28
=== bjf is now known as bjf[afk]
JFosmb, you still around?19:42
=== almaisan-away is now known as al-maisan
=== bjf[afk] is now known as bjf
=== al-maisan is now known as almaisan-away
apwbjf, well personally i hate anything that inefficient ... i fyou can think of tricks to speed it up ... then go for it20:53
bjfapw, if we're happy running it once an hour and if it takes 15 minutes then there are plenty of other things to work on, i've not really looked at it to see if there is anything obvious20:57
JFoI could see about running it every 30 minutes if you think that is better21:07
bjfJFo, if the length of time it was taking to run was keeping us from running more frequently or doing more of what we wanted i'd look at it21:09
bjfJFo, however, if we're happy with it the way it is (and i'm not saying i'm not) then i have no problem just ignoring it21:10
JFoheh, ok21:11
JFoI'm good either way21:11
=== sconklin is now known as sconklin-afk
=== sconklin-afk is now known as sconklin
=== ogra is now known as Guest95972
=== bjf is now known as bjf[afk]

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