/srv/irclogs.ubuntu.com/2014/02/13/#ubuntu-kernel.txt

=== [home-bjf] is now known as bjf
bjfplars, i'm fine with the btrfs failure01:39
=== gerald is now known as Guest66789
=== fmasi is now known as Guest13772
=== cmagina_ is now known as cmagina
=== xnox_ is now known as xnox
=== lag is now known as Guest67837
=== fmasi is now known as Guest63667
ppisatimoin07:02
=== mwhudson is now known as zz_mwhudson
apwmoin08:56
smbtrue08:59
=== lan3y is now known as Laney
=== Laney is now known as Guest89008
apw(moin_p) ?09:01
=== Guest89008 is now known as Laney
=== zz_mwhudson is now known as mwhudson
smbapw, null pointer dereference ... mornings are pointless ... mwhuah09:06
apwsmb, hehe ... i thought you were talkling lisp at me ..09:34
smbapw, not really. the only fact about lisp I know for sure is that one has to embrace many braces09:38
apwperhaps enbrace to be fuly lisp09:56
=== kloeri_ is now known as kloeri
=== mwhudson is now known as zz_mwhudson
=== rbasak_ is now known as rbasak
=== ikonia_ is now known as ikonia
zequence_apw: I've confirmed that it's threadirqs that causes the freeze on both -lowlatency and -generic11:54
zequence_with the driver rt73usb, at least11:54
zequence_apw: So, best to remove that config for now :)11:56
=== zequence_ is now known as zequence
apwzequence, ok, we can do that for the next upload12:50
apwzequence, got a bug number you are working to, which i can link the config change so we remeber to review it again and turn it back on12:51
=== Guest63667 is now known as fmasi
zequenceapw: bug 127908113:24
ubot2`Launchpad bug 1279081 in linux (Ubuntu) "linux-lowlatency freezes when rt73usb is loaded" [High,Confirmed] https://launchpad.net/bugs/127908113:24
=== gerald is now known as Guest75347
rtgapw, I wonder if this affects our LTS Trusty build: '[PATCH v2] compiler/gcc4: make quirk for asm_volatile_goto unconditional'13:53
matanyahello, regarding https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1276705 the bug was introuduced in one of 95 commits i guess13:55
ubot2`Launchpad bug 1276705 in linux (Ubuntu) "Kernel 3.13 fail to boot with LSI SAS1068E (Dell SAS 6/iR)" [High,Confirmed]13:55
matanyai guess i can help sort it out, any one intersted ?13:55
rtgmatanya, jsalisbury can help you bisect it when he comes online13:56
matanyartg: when would that be roughly?13:56
rtgmatanya, in the next hour or two13:56
matanyathanks rtg 13:56
apwzequence, ok turned off for the next upload, you'll want to test manually tunring it on to see if the stable .3 fixes it 13:58
rtgmatanya, it looks like jsalisbury is already working on that bug.13:58
matanyaok, great, i won't bother :)13:58
rtgmatanya, though it might be to your advantage to start a new bug, describe the last known working version, and provide quick turn around on test kernels.13:59
matanyai'll see if i can be of a help here. thank you14:00
zequenceapw: Alright14:23
rtgjsalisbury, prolly ought to get bug #1276705 on our top 10 until you can get it bisected.14:54
ubot2`Launchpad bug 1276705 in linux (Ubuntu) "Kernel 3.13 fail to boot with LSI SAS1068E (Dell SAS 6/iR)" [High,Confirmed] https://launchpad.net/bugs/127670514:54
jsalisburyrtg, sure thing.  I'll add it.14:54
matanyajsalisbury: do you want my help sorting that one out?14:56
matanyajsalisbury: http://dpaste.com/1614107/14:58
jsalisburymatanya, can you provide some details on your issue?  Do you have a bug ID?  15:00
matanyathe one mentioned above15:00
jsalisburymatanya, ah, ok.  I have a test kernel building now.  I'm going to post it to the bug shortly.  It would be great if you could test it.15:01
matanyaok, thanks15:02
=== erle- is now known as ma
=== ma is now known as matt-
=== matt- is now known as mattt-
=== mattt- is now known as matttt-
=== kamal__ is now known as kamal
_bthello, does the latest trusty kernel have this patch? http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=227ae10f17a5f2fd1307b7e582b603ef7bbb7e9716:35
_btim not sure how to check16:35
_btcurrently, trusty daily is giving this bug our of the box on my hardware: https://bugs.freedesktop.org/show_bug.cgi?id=7492716:36
ubot2`Freedesktop bug 74927 in DRM/Radeon "Screen corruption on Ubuntu 14.04, Kernel 3.13 with Radeon R7 240 (OLAND PRO)" [Normal,New]16:36
rtg_bt, yup16:40
rtghttp://kernel.ubuntu.com/git?p=ubuntu/ubuntu-trusty.git;a=summary16:41
apwrtg, ok i think i have the cloud tools cleaned up as 'cloud-tools', i've pushed that to the repos so you have have a look over it16:43
rtgapw, ack16:43
rtgapw, I assume you've added this to -meta ?16:44
_bthi rtg16:47
_btlooks like it made it into 3.13.0-8.2816:47
_btand i am running 3.13.0-8.2716:47
_btlol16:47
apwrtg, yes both linux and linux-meta16:47
_btbrb16:48
_bttesting16:48
rtgapw, I know I'm being a complete pain in the ass, but I _hate_ that commit subject and body. This is really about splitting out "Could only" tools into a separate package. As such, I would prefer that Hyper-V get little or no mention (other then maybe a small footnote at the end).16:48
rtg"Cloud Only"*16:48
apwrtg, you going to fix it or you want me to?16:49
rtgI'm happy to re-write it :)16:49
apwrtg, feel free :)16:49
rtgapw, ok, pushed. check out what I've written16:54
apwrtg looks fine16:56
rtghallyn, re: bug #1279041 - what is the kernel fix to which you refer ? I can't figure out why ip_local_reserved_ports would be treated different then any other file in that /proc directory.16:59
ubot2`Launchpad bug 1279041 in lxc (Ubuntu) "/proc/sys/net/ipv4/ip_local_reserved_ports not writable because of apparmor" [Medium,Confirmed] https://launchpad.net/bugs/127904116:59
xnoxwho is responsible for creating /dev/md directory?17:03
rtgsmb, ^^17:05
smbI would say udev rules17:05
hallynrtg: ip_local_reserved_ports needs to be namespaced17:05
apwxnox, that'd be udev everything in there is its responsibility17:05
apwxnox, though likely in response to finding something to put in it17:05
rtghallyn, but then why doesn't everything else in that directory need to be name spaced in order to be visible ?17:06
smbGuessing its some /dev/md? appearing17:06
xnoxapw: hm. what if udev couldn't have known about them at the time?17:06
hallynrtg: eh, it's kinda ugly and hard to follow17:07
rtgyeah, I kinf of get that :)17:07
rtgkind*17:07
hallynrtg: some things are namespced;  some are host-wide and should not be namespaced;  some are not yet namespaced17:07
xnoxapw: smb: in the installer - udev fires and is gone, then mdadm kernel modules are loaded, then "mdadm" pretends to find /dev/md/Volume2 which doesn't exist... =)17:07
xnoxlet me assemble it and check what happens.17:07
hallynrtg: I've not had time to do a full survey.  I'll try to do that if I can get the overlayfs thing working (which so far I can't)17:08
rtghallyn, ok, I'll leave you to it17:08
apwxnox, well ... is /dev devtmpfs in the nistaller setup?17:09
xnoxapw: smb: odd, so it did assemble and create /dev/md/imsm0 but /proc/mdstat is out of date it has the array but not the "container" (intel fakeraid entry)17:10
apwwhy is udev not left running by the way17:10
xnoxapw: oh it is running.17:10
hallynrtg: thanks :)17:10
xnoxapw: but there was no mdadm nor modules loaded when the drives first appeared...17:10
smbxnox, /proc should be closest to what the kernel thinks it has17:11
apwright its not really sensible to say it is out of date17:11
apwit is a direct representation of what is in the kernel17:11
xnoxapw: smb: so mdadm did start the container, but mdmon was not autostarted and /proc/mdstat is clearly missing an entry =(17:11
rtgxnox, do you have an entry in /etc/mdadm/mdadm.conf ?17:12
xnoxrtg: ofcourse not, this is in the installer.17:12
smbIn theory it should not be necessary to have one but in practice it can make odd differences17:12
rtgit certainly does on reboot17:13
smbxnox, is that a raid5 setup?17:13
xnoxsmb: raid117:13
smbWondering whether it needs a md device for the container of this setup...17:14
apwmight that not appear in dm instead ?17:14
smbBut I guess you saw it when not in the installer17:14
xnoxsmb: apw: so mdadm --auto-detect (after kernel modules were installed) and then assembling helped17:14
xnoxand how i have two arrays in /proc/mdstat =)17:14
xnoxand it went into resync.17:15
xnoxlet's see if i can reboot and repeat everything....17:15
smbrtg, In theory nowadays everything should just be assembeld by the devices appearing... I n practice I found at some point in the past at least that kernel/userspace started to have differenct options of device names in that case17:15
rtgsmb, yeah, which is why I always write the config to that conf file17:16
smbYeah, its the safe way... for some reason17:16
xnoxretoaded: /etc/mdadm/mdadm.conf if there isn't one, one is generated for the initramfs. and even if initramfs is without one, one would be generated inside initramfs and then attempt to assemble everything.17:16
smbI wanted to investigate but ... *sigh*17:17
rtggrumble. the apparmor userspace/kernel mismatches are annoying.17:21
apwrtg, test booting your kernels on precise userspace ?17:22
rtgyup17:22
rtgcan't start a container17:23
rtgjust gonna grub apparmor=017:23
=== DarkPlayer_ is now known as DarkPlayer
apwsmb, can you remember what options were bad for luks18:37
smbapw, errrrrr no (I think)18:38
smbapw, Bad in what way ?18:38
xnoxapw: xts is the good one, aes is the flaky one.18:41
xnox(IV vector that is)18:41
=== zz_mwhudson is now known as mwhudson
=== mwhudson is now known as zz_mwhudson
=== zz_mwhudson is now known as mwhudson

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