/srv/irclogs.ubuntu.com/2011/11/18/#ubuntu-kernel.txt

lifeboyHi all!06:08
lifeboyIs there a channel log somewhere to go read?06:08
apwtseliot, hey ... have you done any testing with 3.2 kernels as yet, for binary drivers10:07
apwlifeboy, yes there are, something like irclogs.ubuntu.com10:08
tseliotapw: no, not yet. Is 3.2 in precise or is it just in the mainline ppa?10:08
apwtseliot, we're working up to uploading it, but i am worried about nvidia particularly for design peeps10:09
apwtseliot, if i put a kernel in a PPA somewhere would that help10:09
* apw notes that kano was just complaining it works on mainline and not our kernel10:09
tseliotapw: yes, definitely, I'll see if I can patch it10:09
apwtseliot, ok so i'll get you a kernel10:10
tseliotapw: thansk10:10
tseliot*thanks10:10
apwtseliot, are raw .debs ok?  that i can get you in much less time10:10
tseliotapw: sure, as long as they are for amd64 ;)10:11
apwanyone else having trouble with LP waiting on fonts.googleapis.com ?10:36
apwtseliot, these are built from the tip of our tree: http://people.canonical.com/~apw/master-next-precise/10:39
tseliotapw: excellent, thanks10:40
apwtseliot, thank you for looki10:40
apwlooking at it10:40
ppisatiwasn't there a script to check if a release was closed properly?11:18
apwppisati, yep, erm ...11:19
apwppisati, stable/verify-src-package perhaps ?11:19
ppisatilet me try11:19
apwi'd guess its meant to be against the source package though11:19
ppisatiuhm no11:19
ppisatiwait11:20
ppisatii use that against the topic branches11:20
ppisatiperhaps it works in this case too11:20
ppisatino wait11:20
apwthe rules should be the same11:20
ppisatiin the topic branches case i use the11:20
ppisatikteam-tools/maintscripts/verify-release-ready11:20
ppisatiand in fact the release wasn't closed properly...11:21
ppisatiok11:21
apwcool11:21
apwodd that its not in the same place ... sigh11:21
ppisatibtw stable/verify-src-package seems a complete different tool11:22
apwjoy11:23
apwherton, just remembered, while you weer away a number of kernels ended up in universe.  not sure if they all got fixed correctly, may be worth checking13:41
ogra_they are universal kernels !13:42
ogra_:)13:42
hertonapw: yep, I noted some of those, and reported the problem on the tracking bugs, mostly all arm kernels ended up in universe13:42
ogra_just ping an archive admin ... thats faster13:43
apwherton, well all the otehre i think as well, i had everything moved for the master branches for everything back to lucid, i ma unsure if hardy got done or not, it was late13:43
apwand hardy is more complex13:44
hertonogra_: I usually do, but it's usually a recurrent issue, I don't know how the copy is done and why it happens13:44
ogra_there is a command the admin runs, sometimes they miss the right option then the kernel goes to universe13:45
hertonapw: ah yes, for the master branches I looked I think on wednesday/yesterday and seems fine now13:45
ogra_usually an oversight 13:45
ogra_and quick to fix 13:45
tseliotapw: can you reupload this file, please? http://people.canonical.com/~apw/master-next-precise/linux-headers-3.2.0-2_3.2.0-2.4~masternext201111180933_all.deb13:45
apwtseliot, reupload it?  s is missing?13:46
tseliotapw: or one which ends with 8093613:47
hertonapw: for hardy I think it didn't work, checking now it's still on universe (http://archive.ubuntu.com/ubuntu/pool/universe/l/linux/)13:48
apwherton, i think hardy got missed as some do belong there and some not and steve wasn't sure13:48
tseliotapw: I'm getting "short read on buffer copy for failed to write to pipe in copy"13:48
apwherton, probabally should re-refer that one to pitti13:48
hertonok13:49
tseliot(also the file is 688kb)13:49
apwtseliot, on it13:49
tseliotthanks13:49
seaniusis there something extra that needs to be done to enable the linux-crashdump support?  i.e. do i need to tweak anything OOTB?  the Kernel/CrashDumpRecipe page doesn't seem to imply i13:51
apwseanius, i thought it was something that needed enabling in /etc/default/somethng13:53
apwother than that i think its meant to be automatic13:53
seaniusapw: yeah, i found /e/d/kexec in which kexec wasn't enabled13:54
seaniuswhich is what i was wondering about actually13:54
* apw isn't a crahsdump expert unfortuanatly, though i am asking abuot13:55
seaniusand *no*, this doesn't have aything to do with that custom kernel stuff i was asking about earlier :)13:55
apwtseliot, ok there now should be a new one in there ... well in about 2:3013:55
seaniuswhat i really want is osmething like redhat's netdump support, but i don't think that exists in debian/ubuntu13:56
seaniusi.e. on crash scp /proc/kcore to a remote server, and reboot13:56
apwthats a pretty nifty idea maybe you could file a bug suggesting it13:57
apwit doesn't sound like so very hard13:57
seaniusheh, they were saying the same thing in #debian-devel13:57
apwyep, we are small compared to redhat so we have less effort to focus13:58
seaniuscould probably be done with the existing stuff and an extra initramfs hook13:58
apwso we don't always see the obvious13:58
apwi'd have thought so indeed13:58
apwtseliot, there now13:59
* cking reboots14:00
seaniushrm, panic seems to still panic, and i get no files in /var/crash14:05
seaniusis there any way to test the crash kernel without triggering a panic?14:09
tseliotapw: ok, thanks14:13
apwseanius, not that i know of14:16
apwseanius, mostly people do that via sysrq14:16
seaniusapw: yes i can do that but it's not loading the crash kernel, hence my wondering.  kexec -p.... seems to work14:18
seaniusbut i get nothing beyond that14:18
apwwhat happens?  just normal boot ?14:18
seaniusthat's just the "prepare" step afaict14:18
apwwhen you ask it to start14:18
apwyou must be able to trigger it like you can a normal boot14:19
apwwhen kexec is installed reboot uses it, i'd have thought -p and whatever it uses to trigger the start might work14:19
ogasawarabug 88859714:25
ubot2Launchpad bug 888597 in linux "linux-image-extra-3.1.0-2-virtual is uninstallable" [High,Confirmed] https://launchpad.net/bugs/88859714:25
ogasawaraapw, tgardner, smb`: so there are test debs for 3.2.0-1.1 on gomeisa in precies-amd64 and precise-i386 in my home dir, but I'm gonna try and fix up 888597 and then respin tests14:26
apwogasawara, ok will wait on that14:28
apwogasawara, i have just been talking to tseliot about checking the nvidia binary drivers14:28
apwogasawara, as that will affect a lot of design people if it doesn't work14:28
apwogasawara, we may want to wait till he reports back before uploading14:29
ogasawaraapw: ack.  also because we've merged server into generic, I also wanted to pick your brain about the best approach for fixing up linux-meta in terms of the control.d files14:29
apwogasawara, ok so is it just called generic now ?14:29
ogasawaraapw: yep14:29
apwogasawara, wil have a poke in a bit and see whats what14:30
ogasawaraapw: awesome thanks14:30
ogra_why do you keep the -generic tag at all ?14:30
ogra_if we only have one kernel that feels somehow redundant14:30
apwwe still ahve two, generci and virtual14:31
ogra_ah14:31
apwand getting rid of it to add it back when someone makes a new one is confusing too14:31
ogra_indeed14:31
tseliotapw: the latest driver in precise seems to work well with linux 3.2 (at least with the one you gave me)14:31
apwtseliot, nvidia binary drivers yes ?14:32
tseliotapw: yep14:32
apwtseliot, would you be interested in testing fgrlx ?14:32
tseliotapw: sure, even though I'm going to update them, either today or next week14:33
apwwill give us some confidence in the viability of our new kernel at least14:33
tseliotapw: I can already tell you that the fglrx module builds, I only have to see if it segfaults ;)14:35
apwtseliot, :) just bound to work14:35
tseliot:)14:38
tgardnerogasawara, I updated the precise chroots on gomeisa14:49
ogasawaratgardner: ack14:49
* herton -> lunch15:05
* ogasawara back in 2015:19
apwogasawara, http://status.ubuntu.com/ubuntu-precise/group/topic-precise-kernel-essential.html <-- should represent the release critical blueprints16:07
ogasawaraapw: awesome, did that just recently appear?16:07
apwi made it about 5 hours ago yep, and its finally in the output :)16:08
tgardnerogasawara, 3.2.0-2-generic - at least suspend now works again on my HP mini16:09
tseliotapw: fglrx works well too :)16:11
apwogasawara, ^^ so i think thats both binary drivers tested :)16:11
ogasawaraapw: cool.  do you want me to pull in that last config change you pushed to master-next before I upload16:12
ogasawaraapw: I think I might wait till we have the linux-meta bits sorted as well16:12
apwogasawara, the config changes i am pushing are not urgent or important16:15
ogasawaraapw: k, I'll leave it for the next one then since I've already prepped the master branch16:15
apwyeah will look at that next, i suspect we want to keep the seaparation at the meta package level, so not a tranisitional package16:16
apwshould be pretty easy in theory, though we don't need it till the builds are complete anyhow16:16
ogasawaraapw: yah, I was hoping to keep it separate in case we do want to fall back to having a server flavor16:16
apwogasawara, how much testing have we had on those builds of your?  you were updating them last i looked and haven't seen if new ones are there16:17
ogasawaraapw: I kicked of the new builds, not sure if they've finished yet16:17
ogasawaraapw: been sidetracked with the release meeting16:18
apwyep no worries16:19
tgardnerpgraner, tangerine.buildd:~rtg/ukb/oneiric/amd64/master-next/linux-image-3.0.0-14-generic_3.0.0-14.23_amd64.deb16:24
apwogasawara, i have just pushed an update to ubuntu-precise-meta which i think is what we want16:39
apwogasawara, obviously it still needs a version bump as and when16:41
apwcking, are you aware of this page: http://status.ubuntu.com/ubuntu-precise/u/colin-king.html16:43
apwcking, that has all your WIs regardless of blueprint, i note you have some duplicates by the look of it16:44
ppisatii'm trying to package a new omap4 kernel for O, but can't get the debian* stuff right i.e. fdr printchanges is always empty16:45
ppisatican anyone give it a look?16:45
ppisatihttp://kernel.ubuntu.com/git?p=ppisati/ubuntu-precise.git;a=shortlog;h=refs/heads/ti-omap416:45
ckingapw, ah, duplication, hrm, well I was unsure if the power management was a general placeholder for all things related to this kind of work16:45
ppisati(btw i bumped abi to 1400 to distinguish it from O/omap4 kernels)16:45
apwppisati, if the previous version differs in the changelog then it won't find it16:45
apwppisati, and you just have to do the changelog bits by hand, common on the first upload for a release16:46
ppisatiapw: i think i did it16:46
apwyou can see how its done in the debian/rules.d16:46
ppisatii can even build the first pkg16:46
apwfdr printenv gives you hints if its not recognising the previous version16:46
ppisatibut now i would like to add some stuff on top of it16:46
apw?16:47
ppisatibut fdr printenv insist in using the OLD abi (1206)16:47
* apw pulls16:47
ppisatiand fdr printchanges is still empty16:47
ogasawaraapw: awesome, thanks16:47
ogasawaraapw: also test builds on gomeisa finished, feel free to pull the debs and test on any kit you have there16:48
apwppisati, do there are no new commits on here after the close ?16:51
ppisatiapw: nope16:51
ppisatilocally, didn't push them16:51
ppisatiwait16:51
ppisatiapw: pushed16:52
apwppisati, so if you want to make a .2 you need to open that with fdr startnewrelease, if you want to replace .1 then you need to rebase the .1 close commit out of existance16:57
ppisatiapw: why shall i obliterate the .1 close commit?16:58
apwppisati, which are you trying to do make a new .1 or move on to .216:58
ppisatimove to .216:58
apwthen you should be ading a startnewrelease16:58
apwfdr startnewrelease, then commit that16:59
ppisatiok16:59
apwthen you can close it to make .216:59
ppisatistart new release add the boilerplate in debian.ti-omap4/changelog16:59
ppisatibut i don't understand why fdk printchanges doesn't show anything16:59
ppisati...17:00
ppisatinever mind17:00
ppisatiprintchanges work only AFTER a new release has been opened...17:00
ppisati"a storm in a changelog..."17:00
ckingpower measuring gets a little repetatititititive after a while17:05
apwppisati, indeed so17:08
apwcking, yep that why you need a program to do it17:34
ckingapw, true, but somebody has to install software, run the test, gather the data for different tests...17:49
ppisatiherton: how does ktl.ubuntu.series_name() knows if a version is in one release or in another one? does it use hardcoded values?17:51
hertonppisati: yes, it checks the version/abi or the package name if necessary, and do the specific checks if needed (verifying abi number for ex.)17:53
hertonppisati: eg. if it's linux-mvl-dove, it just checks abi and returns lucid or maverick17:54
ppisatiherton: ok17:54
ppisatiapw: one last thing, if you are still alive17:56
ppisatiapw: now it complains about "package linux-image-3.0.0-1401-omap4 is not in control info"17:57
ppisatiapw: http://paste.ubuntu.com/742485/17:57
ppisatiapw: and here is the pkg: http://kernel.ubuntu.com/git?p=ppisati/ubuntu-precise.git;a=shortlog;h=refs/heads/ti-omap417:58
EtienneGBeginner's question, guy's: the pae flag in /proc/cpuinfo, that means the CPU supports PAE, right?18:04
EtienneG(I guess so, just want to be sure)18:04
jjohansenEtienneG: yes if pae is listed then the processor should support PAE18:11
EtienneGjjohansen, thanks!18:11
EtienneGreason I asked is the discussion about dropping non-PAE kernel or not in precise.  Not sure if a final decision was made, but I want to be ready in case it is18:12
apwppisati, have you cleaned it properly18:12
EtienneGby making an inventory of the machines that will be non-upgradable18:12
ppisatiapw: i think so18:13
ppisatiapw: let me try with a git clean -fdx18:13
apwppisati, have a look at the control.stub.in and make sure the base version is updated correctly18:13
ppisatiapw: ok, i'll do18:14
apwppisati, and check control afer a clean18:15
mfilipewill Oneiric Kernel be update to 3.2?18:20
apwmfilipe, no it will remain on 3.018:20
mfilipeapw, will be released any package to update the kernel to 3.2?18:22
jjohansenmfilipe: probably not, there will be the backports kernels from the next releases Q, R, S but Q likely will be on of 3.3, 3.4, 3.518:23
jjohansenjust depending on when the upstream kernels come out18:24
mfilipehum... 3.2 will have many improvements in energy management 18:24
mfilipeanyway, I compile it if be necessary 18:25
ppisatimfilipe: why don't you simply install precise kernel?18:25
mfilipejjohansen, apw: thanks for informations18:25
mfilipeppisati, precise kernel?18:25
ppisatiit's 3.2 based18:26
jjohansenmfilipe: oh shoot you where talking Oneiric, and here I was thinking precise, there will be no backports for Oneiric, the are for precise18:27
mfilipehttp://packages.ubuntu.com/precise/kernel-image-3.1.0-2-generic-pae-di18:28
mfilipeit is 3.118:28
jjohansenppisati: err no its not18:29
mfilipewhen torvalds tree tagged 3.2 final release, will you update to 3.2?18:29
ppisatijjohansen: well, not yet... :)18:30
jjohansenright18:31
jjohansenmfilipe: yeah we should move to 3.2, its 3.3 that will be skipped (/me was remembering the wrong revision)18:32
mfilipeok18:32
mfilipethanks a lot!18:32
ogasawaramfilipe: I'm getting ready to upload the first 3.2 based precise kernel in just a bit18:33
jjohansenogasawara: oh really without binary drivers?18:34
ogasawarajjohansen: was told that tseliot tested both nvidia and fglrx this morning on the 3.2 kernel18:35
jjohansenogasawara: ah, I hadn't heard that and was expecting those to be a pita18:35
mfilipeogasawara, but 3.2 stay like release candidate yet 18:36
mfilipehttp://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=summary18:36
ogasawaramfilipe: yep, we rebased to the latest 3.2-rc218:36
ogasawaramfilipe: we'll keep rebasing until 3.2 is final18:36
mfilipeok, nice18:38
apwyeah unexpectedly they still work (the binary drivers)18:42
ricotzthe vbox dkms module will break though19:01
apwricotz, maybe so, they will need to sort it out 19:07
ricotzapw, yeah, 4.1.6 should be compatible which is already in debian19:08
apwricotz, then it should get synced in the mass sync with luck19:08
ricotzapw, it is still in testing so it wont :\19:08
ricotza manual sync would be nice here though19:09
ricotzoh meant "unstable"19:09
GrueMasterppisati: I am seeing several missing CVE issues with the omap4 kernel on maverick (2.6.35).  I'll add them to the kernel release tracking bug.19:53
hertonGrueMaster: what CVEs are missing?20:05
GrueMasterLP #729839 is one.  I have 7 failures, some are CONFIG_ features that are not enabled.20:06
ubot2Launchpad bug 729839 in linux "PR_SET_PTRACER does not work from a thread" [High,Fix released] https://launchpad.net/bugs/72983920:06
GrueMasterI added a log to bug 88856920:07
ubot2Launchpad bug 888569 in kernel-sru-workflow/verification-testing "linux-ti-omap4: 2.6.35-903.27 -proposed tracker" [Undecided,Fix released] https://launchpad.net/bugs/88856920:07
GrueMastertest-kernel log now added as well.20:18
sforsheeogasawara, do you have a 3.2-rc2 precise build somewhere that I can get at easily?20:21
ogasawarasforshee: yep, check on gomeisa in my home dir under the precise-amd64 or precise-i386 dirs20:22
sforsheeogasawara, thanks, now I don't have to wait for a build :)20:22
* herton -> eow20:26
argessforshee, who would be the wifi guru to talk to , just was able to reproduce https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/83625020:33
ubot2Launchpad bug 836250 in network-manager "[Oneiric] [Regression] Intel Corporation Centrino Ultimate-N 6300 poor networking, packet loss and very slow Lenovo X201 and T500 laptops" [High,Invalid]20:33
sforsheearges, the bug is assigned to tgardner and he's the most knowledgeable about wireless anyway, but he's not around right now20:38
argesah, yea I just got hit with it when I tried a new router20:38
sforsheearges, you should probably grab dmesg and syslog while you're seeing the symptoms as a starting point20:39
argesYea, there seems to be quite a bit of info in the bug... so trying to determine if I won't just be adding a 'me too' case to it20:40
sforsheearges, also you might look at http://linuxwireless.org/en/users/Documentation/Reporting_bugs for some things you might want to try20:40
sforsheealthough it seems most of the stuff there is geared towards problems associating, not performance problems20:42
argesCool. I might add some info. One thing was looking at iwlist/iwconfig etc its not obvious which wireless 'mode' i'm connected to 802.11a/g/n etc20:42
argesso I was trying to just figure out how to switch modes from my laptop to isolate it being an 'n' issue vs 'b/g' which is what I've been seeing switching the modes on the router side.20:43
argesthe workaround in the bug seems to agree since it just involves disabling n mode on the laptop. 20:43
argesby disabling it in the module20:44
sforsheearges, if you look at the bit rate in iwconfig you can probably guess at which mode it is20:45
sforsheei don't know a definitive way to tell though20:45
argesyea that's what I was figuring20:45
argesok20:45
sforsheethere's probably a better way to tell, i just don't know what it is20:45
Sarvattno more brcmsmac driver in 3.2.0-1?21:27
mjg59It moved from staging. The config may not have been updated.21:27
argessforshee, so I installed ogasawara's 3.2-rc2 kernel in my system using dpkg -i and its not showing up in grub, do I need to do something special to get that entry made? It looks like it was doing the hooks for the grub update21:40
sforsheearges, update-grub should be run as part of the installation, but I guess you could try running it manually just to be sure21:41
sforsheearges, also run 'dpkg --get-selections | grep ^linux' and make sure the packages really are installed21:43
argessforshee, yea checked if it was installed first nad it is... will re-run update-grub21:48
argestrying again...21:49
sforsheeogasawara, did you see the above comments about brcmsmac?21:58
ogasawaraSarvatt, sforshee: thanks, have added an item to our config blueprint for brcmsmac.  will look into it on Monday.23:49

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