/srv/irclogs.ubuntu.com/2016/02/08/#ubuntu-kernel.txt

apwzkanda, which series are you trying to run it, what is your root disk stored on?08:40
zkandaapw:  model is UX501VW, the root disk is on one for the partition. But have a weird naming scheme. Not the usual sda.08:43
apwzkanda, what does you disk get named as ?08:44
apwand which series of ubuntu are you installing on it08:46
zkandaWait, let me look at it.08:52
zkandaapw: disk get name as something like  this: nvme0n1p2, latest 15.10 ubuntu.08:54
tjaaltonyou'd need newer initramfs-tools09:09
tjaaltonnvme module location changed, so 4.4 and up won't boot since the initrd doesn't include it09:10
zkandatjaalton: How do I do that? Just a link where I could follow is good. Searching didn09:13
zkanda'didn't yield good information.09:13
tjaaltonhttps://bugs.launchpad.net/ubuntu/+source/initramfs-tools/+bug/152487909:16
ubot5`Launchpad bug 1524879 in initramfs-tools (Ubuntu) "initramfs-tools, Xenial is missing NVME kernel driver" [High,Fix released]09:16
tjaaltonapw: looks like it regressed something09:17
apwtjaalton, do we know zkanda is even running xenial ?09:18
apwzkanda, what version of initramfs-tools is installed on your system (dpkg -l initramfs-tools)09:19
tjaaltonapw: using mainline kernels on any release has the same bug09:19
tjaaltonit's 15.1009:19
tjaaltonmainline kernels 4.4 and up that is09:19
apwtjaalton, right .. then then we would expect mainline 4.4 to not work on anything before 16.0409:19
tjaaltonyep09:19
zkandaOk, any advice what should I do from this? Sorry I'm not too familiar on how this things work. :(09:21
tjaaltonyou could just modify the hook manually09:21
apwi suspect that is the "simplest" as you are unlikely to be able to install initramfs-tools from xenial without a bunch of pre-requisites09:22
apwi seem to remember merging up a bunch of stuff to make that work09:23
zkandatjaalton: How do I do that? Don't know how hook works.09:23
zkandaCan I just try Xenial Alpha then? Would that work? I actually tried the Xenial Mate, but it didn't boot either.09:24
zkandaI forgot the actual error when trying it out though.09:24
tjaaltonedit /usr/share/initramfs-tools/hook-functions, add what's on the first comment on https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=80700009:25
ubot5`Debian bug 807000 in initramfs-tools "initramfs-tools: module nvme not included in block modules on kernel > 4.2" [Normal,Fixed]09:25
tjaaltonthen run update-initramfs -u -k all09:26
apwtjaalton, i don't think they want to us -k all09:26
apwuntil they know they ahve done _one_ right09:26
tjaalton4.2 still boots09:27
tjaaltonor -k $whatever09:28
tjaaltondunno what it would be for some random mainline09:28
zkandaOk, I'll try that later, still at work so I can't experiment around. haha09:28
xnoxapw, we really should backport that change.... all the way back to trusty?10:44
xnoxapw, are we doing kernel-lts-xenial for trusty?10:44
apwxnox, yes, we are, infact vrey much now ...10:53
apwxnox, i'll file a bug for that now ... as it owuld be a pre-requisite for that indeed10:53
apwxnox, i've added a task to LP: #1524879 ... note the fix got refixed in a later upload, anyhow will put it on my todo10:59
ubot5`Launchpad bug 1524879 in initramfs-tools (Ubuntu Trusty) "initramfs-tools, Xenial is missing NVME kernel driver" [High,Triaged] https://launchpad.net/bugs/152487910:59
zkandatjaalton: That didn't seem to fix the issue, disk still not found11:26
tjaaltonzkanda: did you run update-initramfs?11:29
zkandaYes I did.11:29
zkandaNote this is kernel 4.511:29
tjaaltondoesn't matter11:29
zkandatjaalton: is there anything I can look to verify that the nvme module is updated?11:34
tjaaltonzkanda: unpack the initrd to see it's there. if it isn't you didn't fix hook-functions enough12:08
jjohansenapw: So the lockinng patchbomb, is at git://kernel.ubuntu.com/jj/ubuntu-xenial.git patch-bomb12:41
jjohansenIt has the mount patch that tim reverted with a fixup: patch for it that fixes an rcu locking problem that the original patch had. I haven't folded that fixup in yet so you guys could easily see the diff12:41
apwjjohansen, ack12:42
jjohansenI haven't managed to get to cleaning up the full queue yet, there are several fixups that need to be folded into other patches a few bug links to add etc.12:42
jjohansenapw: if you want to just pull the mount patch + fixup, its the 2nd and 3rd patches in the queue12:43
jjohansenthe 1st one is for a minor mount auditing issue12:43
jjohansenthe audit patch does slightly change the mount sleep fix patch, I should have probably changed the ordering12:44
jjohansenapw: its passed my testing but I haven't run stgraber's tests nor have I tested on anything but amd64 yet12:45
apwjjohansen, 39 patches, ugg12:45
rtgjjohansen, wow, you want all of these ?12:45
apwyou've got a lot of unannotated titles in there (no apparmor prefixes) is that expected ?12:46
jjohansenapw: like I said I need to clean it up yet12:46
apwjjohansen, oh ok ... rtg ^12:46
jjohansenrtg: sadly yes, its all backported fixups to locking and ref counting, and fallout around it12:47
jjohansenit fixes kees's bug12:47
apwrtg, if we applied the cleaned version to something and uploaded it to unstable, we could get the adt tests for lxc etc run against it12:47
rtgjjohansen, well, we can wait until you are ready. no use in doing this twice.12:47
rtgapw, yeah, could do that12:48
jjohansenrtg: if you want we could cherry-pick out the mount patch + its fixup and just apply that atm,12:48
rtgjjohansen, that is the one causing problems right now, correct ?12:48
jjohansenrtg: yes, its the fix you reverted, patch #3 is a fixup to that patch12:49
apwsounds like we should take that one for master-next, and shove the whole thing on a test kernel12:49
rtgjjohansen, 'fix-up: kern_mount fail path should not be doing put_buffers()' I presume12:50
jjohansenrtg: yes12:50
jjohansenrtg: give me a minute to rebase the patch series, as it stands patch1 causes a conflct with #2 and #3 so you can clean cherry-pick them12:52
rtgjjohansen, yeah, I was just noticing that12:52
rtgjjohansen, just holler when you are ready12:52
jjohansenrtg: ok repushed with -f , the mount patch and its fixup are now the 1st and 2nd patch in the queue, and the old patch #1 gets completely dropped because it turns out to be superfluous with the reorder13:10
rtgjjohansen, got it. I'll upload the whole pile to our unstable PPA for some testing.13:11
rtgand cherry-pick just the 2 for master-next13:11
jjohansenrtg: okay, thanks. I'll work on cleanup the commit messages and folding some of the patches together. No point in keeping fixups to patches in the queue13:11
jjohansenthat is fixups separate13:12
rtgjjohansen, agreed13:12
rtgjjohansen, in fact, would it hurt if I just squashed those 2 ? it might prevent a bisect problem for jsalisbury  sometime in the future.13:13
jjohansenrtg: no, I would squash them, I just kept them separate so you and apw could see what the fix was13:14
rtgcool13:14
jdstrandogasawara, cking: thanks!14:19
ckingjdstrand, my pleasure, I'll wait to see what the upstream developer thinks14:19
jdstrandcking: ack. let me know if you need anything14:20
ckingjdstrand, ok, I may ask you to test the fix at some point14:20
jdstrandsure thing14:20
manjortg, when I cross build on tangerine for arm64 .. /lib/modules/$(uname -r)/build/scripts/<blah> are all built for x64 .. is there a way to force to build everything arm64?16:07
rtgdpkg-buildpackage -d -aarm6416:08
manjoyep that was my cmd line 16:09
rtgmanjo, within the right chroot ?16:09
manjoyeah within xenial chroot 16:09
manjowily/xenial does the same thing 16:10
rtgmanjo, you could try '/usr3/ubuntu/kteam-tools/buildscripts/ukb-make --release=$RELEASE --arch=$ARCH --branch=$BRANCH --clean --repo=`pwd`'16:13
rtgwhich is how I do all of my builds16:13
manjortg, so I don't need to dchroot blah .. just run this on the pwd ie kernel source 16:14
rtgyup16:14
manjowhat is branch ? 16:14
rtgthe branch in your local repo. I assume you are developing on master16:15
manjoah yes master 16:15
manjoso if I drop --branch it would use master ? 16:15
rtgmanjo, can't remember16:15
manjook I will figure out 16:15
manjortg, oops that script won't work if I did a dpkg-source -x extract of a dsc ... 16:20
rtgmanjo, only works from a repo16:20
manjortg, do you have an arm64 cross compiled headers-generic  package I can use? I want to see if your script builds stuff under linux/scripts correctly16:24
manjortg, I think that makefile might not support cross compile flags 16:25
manjoprobably that is the right thing to do ie build those x64 when you are building on x64 ?16:27
rtgmanjo, y'all ain't makin' no sense to me - wtf are you trying to do ?16:28
ogra_dont you need to set ARCH= and CROSS_COMPILE= too ? 16:30
manjo☺ there are a bunch of tools under linux/scripts/*.c which gets built as x64 when I cross compile for arm64... from looking that makefile May be it does not support cross compile16:30
manjoand it could be the right thing to do ie .. build them for the native platform coz some of theose tools are used for the build itself 16:30
ogra_iirc "dpkg-buildpackage -d -aarm64" isnt enough16:30
rtgI've specifically disable tools generation when cross compiling. its just too much of a PITA16:31
ogra_yeah, the deps will bite16:31
rtgmanjo, if you want tools, then you'll have to upload to a PPA for a native build16:32
manjoyeah that is a PITA ☺ coz it takes like 3hrs on native PPA 16:33
rtgI can't help that16:33
ogra_get a dragonboard ;)16:33
manjoI know .. just saying 16:33
rtgjsalisbury, are you pursuing bug #1540731 with upstream now that you've identified the regression ?16:56
ubot5`bug 1540731 in Mir "SocketMessenger::update_session_creds() fails to get client PID, causing "[ FAILED ] PromptSessionClientAPI.client_pid_is_associated_with_session" on kernel 4.4 (but kernel 4.3 works)" [Critical,Confirmed] https://launchpad.net/bugs/154073116:56
jsalisburyrtg, yes.  I have an email to respond to16:58
jsalisburyrtg, and there is a new patch to replace the revert, so I can get it tested16:58
rtgjsalisbury, cool16:59
melodiehello18:39
melodieI am looking for hardware supported info, especially the Intel CPU Skylake : either i5 or i7 (4 brands AFAIK)18:41
melodieI read here: http://www.phoronix.com/scan.php?page=news_item&px=intel-skl-prelim-support18:41
melodieit's of last summer, how is it now?18:41
melodiecan we have accelerated/3D and all working well?18:41
melodieI have to help someone build the specs of his new machine and his favorite distro is Kubuntu. He does current office work and reworks his pictures (gimp mostly). He wants it to be very snappy. (the machine will have 8 GB ram to start with)18:42
melodiewhat do you think?18:42
melodieno one around at this time, for insights about hardware?18:45
Pithegorushye how do i customize ubuntu kernel?19:20
=== Pithegorus is now known as Deadly
=== Deadly is now known as KillerSingam
melodieKillerSingam by recompiling it19:35
melodiehi garrettr 19:36
melodiedo you have knowledge about the intel CPUs and the kernels available at Ubuntu? I'm told that for a i5 or i7 Skylake I would need a Xenial  and I wonder if there is another way to get a 4.3 or 4.4 kernel image in let's say, Wily?19:37
melodieI have to provide a machine to a friend who uses Kubuntu (and a good machine)19:37
melodiehe knows not about hardware and software, but I have to give him the best and stable possible because he also uses it for his work19:38
melodiewhat do you think?19:38
melodie(I asked around here one hour ago, and no answer so far)19:38
KillerSingamhow do i recompile my kernel19:39
melodieKillerSingam you seek for the documentation first19:39
melodiehttps://help.ubuntu.com/community/Kernel/Compile19:40
melodiethanks to Qwant, the search engine. ;)19:40
KillerSingamqwant?19:40
melodiehttps://www.qwant.com/19:41
melodieyes, I've qwanted for you19:41
KillerSingamcan i use that qwantum  for my girlfriend?19:42
melodiethis is not a bouquet of flowers, nor a perfume. It's a search engine19:48
melodieso go compile your kernel now! what are you waiting for,19:48
melodie?19:48
KillerSingamif i did compile my kernel.. how do i install it19:50
argesmelodie: Xenial works just fine on a skylake i7 for me. Basing on an LTS release (like xenial) would provide the most stable path depending on your timeframe.20:06
melodiearges how stable is Kubuntu Xenial is something I should test, I guess20:10
argesmelodie: of course, and also release the actual release isn't until April20:10
melodiemy timeframe is not large, if you are referring to how long to I have to build the specs for this tower. A couple days, then it has to be ordered20:10
melodieI know it's not before April, however I have upgraded my own custom edition (Bento Openbox Remix) and it works fine (it was better with nouveau than with the nvidia driver, but my tower is an old box)20:11
melodiearges the person for whom I'm building the specs needs it for his daily work, which is common current work, and also to edit pictures, he makes photos and edits them with Gimp20:12
melodiehe will have the machine start with 8 GB RAM and we can add more if needed. Just now I'm not quite sure which version Kubuntu and which version processor I should pick up20:13
melodieI was looking here: http://ark.intel.com/fr/products/88192/Intel-Core-i7-6600U-Processor-4M-Cache-up-to-3_40-GHz20:13
melodieand to the latest here:20:13
melodiehttp://ark.intel.com/fr/products/family/88392/6th-Generation-Intel-Core-i7-Processors#@Desktop20:14
melodiearges may I continue?20:16
argesmelodie: I'm not entirely sure which works best. But if you do run into issues please file bugs. : )20:16
melodiearges I don't get to provide specs and use, I only provide specs and the end user will have to get the perfect choice :D20:18
melodiethis is why the choice is difficult (where would be the fun otherwise?) XD20:19
melodiehttp://www.ldlc.com/informatique/pieces-informatique/processeur/c4300/+fb-C000000192+fv579-5524+fv160-13685.html20:19
melodieI have been shown the first one on the list, then when the guy said "Skylake" I thought omg! that's new! how is that with the kernel?20:19
melodie:)20:19
melodieand after I did a search I thought the best would be asking the chans for insights20:19
argesmelodie: I understand. And yes, I was mainly trying to answer 'does the Ubuntu kernel support Skylake CPUs?'20:21
melodiearges now I'm going to grab a Kubuntu Xenial, install it on a box (I have enough computers to give it a try on bare metal) and see how it goes.20:23
melodiearges I thank you for your kindness20:24
argesmelodie: good luck20:24
melodie:)20:24
melodiethanks!20:24
KillerSingammelodie20:35
KillerSingamhi20:35
KillerSingammelodie how to recompile my kernel where do i get the source20:36
melodieKillerSingam you read the page that I pointed you to, then you follow the first link you find in it. surely you know how to read? https://help.ubuntu.com/community/Kernel/Compile20:42
=== garrett__ is now known as Guest26380

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