/srv/irclogs.ubuntu.com/2013/07/09/#ubuntu-kernel.txt

=== henrix_ is now known as henrix
* apw yawns08:01
=== fmasi_afk is now known as fmasi
=== fmasi is now known as fmasi_afk
psivaaapw: Updated bug #1195710 with some more information, not sure it will help. Have time to look into it?09:01
ubot2`Launchpad bug 1195710 in linux (Ubuntu Saucy) "'Kernel bug - invalid opcode: 0000 [#1] SMP' is reported at 'Preparing linux-image-extra-3.10.0-0-generic' stage of multi-lvm installations of amd64 saucy server " [High,Confirmed] https://launchpad.net/bugs/119571009:01
apwpsivaa, oh hmmm, increasing memory helped ...09:03
apwpsivaa, as neither smb nor I could reproduce it, but 1G is my default size09:04
* ppisati -> reboot09:04
psivaaapw: isn't 512 the minimum allowed for server?09:05
=== fmasi_afk is now known as fmasi
apwpsivaa, oh quite probabally, i just mean the default in the box when i make new instances09:09
smbBy default I use 1G but I probably have done 512 as well09:09
smbCannot remember right now09:09
apwis 1G and that sems 'small' and therefore wasn't going to hit it ... doh09:09
=== AlexB_ is now known as AlexB
apw/srv/images192.0.2.0/24(ro,sync,no_subtree_check)09:17
apw/srv/images 192.0.2.0/24(ro,sync,no_subtree_check)09:17
=== yofel_ is now known as yofel
apwError creating pool: Could not start storage pool: cannot open path '/var/lib/libvirt/images/iso': Permission denied09:33
psivaaapw: is this one of the utah errors in trying to reproduce the above bug?09:39
apwpsivaa, oh no, this is me trying to export some images so i can start some new vms with less ram09:45
apwpsivaa, i am chatting with smb about how retarded libvirt is and how much i would like to meet its mother09:45
psivaaapw: ack :)09:46
apw(with and axe)09:47
=== fmasi is now known as fmasi_afk
=== fmasi_afk is now known as fmasi
apwpsivaa, do you have the whole preseed you are using for this install ?10:22
psivaaapw: it's here: http://bazaar.launchpad.net/~ubuntu-server-dev/ubuntu-test-cases/server-tests-raring/view/head:/preseeds/multi-lvm.preseed 10:26
psivaaattached to the bug as a file as well now10:27
apwpsivaa, brilliant ..10:27
=== ayan_ is now known as ayan
apwsmb, ok so using the preseed manually is pretty easy, just change the kernel command line and remove the file=fooo.preseed and replace it with like url=http://192..../foo.preseed10:43
apwpsivaa, ok with that i was able to repro this, i think ... thanks10:49
psivaaapw: ack, thank you10:50
apwpsivaa, will see what i can find10:51
apwpsivaa, ok yeah made it do it twice, seems to happen later without utah, utah must be consuming a chunk of memory i recon11:03
psivaaapw: that's probable. I always used utah. the client logs are saved in there before retrieving to the server11:05
=== fmasi is now known as fmasi_afk
* ppisati -> out for lunch12:09
* henrix follows ppisati12:14
jjohansenrtg: so I have a branch with the aa3 sync against pure upstream 3.11, so it doesn't have the ubuntu configs as part of it. Do you want a pull request using this or do you want to point me to your 3.11 rebase?13:16
rtgjjohansen, a pull request agsint the saucy unstable branch would be nice13:19
jjohansenrtg: ah, got it. I'll have that for you in a few minutes13:20
rtgjjohansen, no rush. we won't transition to 3.11 until about rc3 or 413:20
jjohansenrtg: ack13:21
jjohansenrtg: okay, I've pushed it to my saucy tree on zinc, I'll do some build testing etc, before sending out the request13:25
rtgjjohansen, ack13:25
FernandoMiguelhowdy13:27
FernandoMiguelever since kernel  3.9.0-7, I haven't managed to get a proper CPU scheduler management 13:27
FernandoMiguel3.10 only offers two states13:27
FernandoMigueland seems to use much more battery13:27
rtgsforshee, re: bug #1190225. Is that in drivers/power/smb347-charger.c ?13:41
ubot2`Launchpad bug 1190225 in linux-grouper (Ubuntu) "opening /sys/devices/platform/tegra-i2c.4/i2c-4/4-006a/reg_status (as user) causes immediate reboot" [Undecided,New] https://launchpad.net/bugs/119022513:41
SarvattFernandoMiguel: update again, i think that was fixed a few weeks ago?13:49
FernandoMiguelSarvatt: every kernel I get, I boot into it, and all are the same13:50
Sarvattor boot with intel_pstate=disable13:50
sforsheertg: looking13:52
rtgsforshee, I'm pretty sure that it13:52
rtgroot@ubuntu-phablet:/# dmesg|grep charger13:53
rtg[    2.096531] smb347_charger: [smb347_init] project_id=0, pcba_ver=3, dock_in_gpio=16413:53
rtg[    2.098920] [charger] Disable AICL, retval=93 setting=8313:53
rtg[    2.099397] [charger] set cahrger limmit, limit=900 retval =73 setting=7313:53
rtg[    2.099802] [charger] re-enable AICL, setting=9313:53
rtg[    2.627268] smb347_charger: [cable_type_detect] Reg39 : 0x1013:53
rtg[    2.627539] smb347_charger: [cable_type_detect] Reg3F : 0xc013:53
rtg[    2.627681] smb347_charger: [cable_type_detect] USB_IN13:53
rtg[    2.627949] smb347_charger: [cable_type_detect] Reg3E : 0x0c13:53
Sarvattapw: didn't you make that default to off? i'm not seeing it in the changelog13:53
apwSarvatt, yeah that was disabled in the 3.9 kernel, the rumours were that 3.10 was 'better'13:53
rtgSarvatt, also NO_HZ_FULL=n13:54
apwSarvatt, so it wasn't switched there, on the assumption we would never ever test it there13:54
apwSarvatt, unless it was on, ... are we seeing issues now we are in 3.1013:54
Sarvattoh 3.10 made it start working on ivybridge too and it was bad here once that kicked in13:54
apwSarvatt, so still shit in 3.10 yes13:54
Sarvattyeah, 15C higher idle even in power save mode :(13:54
apwif so i can pull that change up to there, and leave it off in 3.11 which rtg is managing13:54
apwSarvatt, well i call that tested in the negative ... 13:55
FernandoMiguelSarvatt: what will that parameter do ?13:56
sforsheertg: looks like it is smb347-charger13:56
apwSarvatt, what the heck was the bug number13:56
rtgsforshee, yep13:56
FernandoMiguelapw: so it's a known bug/limitation ?13:56
Sarvatti think they expect you to use that intel thermal management daemon to control fans if you use intel_pstate (https://01.org/linux-thermal-daemon)13:56
=== Guest86601 is now known as mfisch
FernandoMiguelcause with 40Âșc here, I can't just leave it in 3.10 Power Save ... it will kill kittens13:57
apwpstate_idle being pants, it is 'known of' in 3.9, and now 3.10 yeah13:57
apwFernandoMiguel, sounds fun indeed13:57
FernandoMiguelapw: it works fine in 3.9.0.713:57
=== mfisch is now known as Guest37969
apwFernandoMiguel, and pstate_idle=off or whatever Sarvatt said it was, did the trick for you13:57
FernandoMigueland all before that13:57
apwFernandoMiguel, well it was13:57
FernandoMiguelthe few after and 3.10 broke 13:57
apwFernandoMiguel, well it was not enabled for some your kit at all in 3.9 13:58
FernandoMiguelthere was a 3.10.1 that worked too. but broke again on .213:58
FernandoMiguelso the advice is to pass a kernel option for 3.10 till this is improved, correct?13:59
FernandoMiguelany bug # I can track?13:59
Sarvatthmm its not on https://bugs.launchpad.net/~adconrad/+reportedbugs so it must have been closed13:59
SarvattFernandoMiguel: still trying to find it..13:59
FernandoMiguelthanks14:00
apwSarvatt, found it ... bug #118864714:00
ubot2`Launchpad bug 1188647 in linux (Ubuntu) "Please change intel_pstate default to disable" [High,Fix released] https://launchpad.net/bugs/118864714:00
FernandoMiguelGRUB_CMDLINE_LINUX="intel_pstate=disable"14:00
FernandoMiguelhow will this affect 3.9.x kernels?14:00
FernandoMiguelthanks guys!14:01
apwFernandoMiguel, if you arn't using there they won't notice it14:02
FernandoMiguelapw: the bug is fix released, but it's not fixed in 3.10.x14:02
apwrtg, i have pulled that pstate_idle change up to master-next, but _not_ applied it to unstable as per discussion above14:02
apwFernandoMiguel, reload the bug14:02
FernandoMiguelah :914:03
sforsheertg: the offending function is smb347_reg_show. One thing that is obvious is that it's reading more data than what its buffers are sized for.14:03
FernandoMiguelapw: need any info on my system to help triage it ?14:03
apwFernandoMiguel, the fix was deliberatly applied and releasd against the v3.9 kernel but not applied to unstable (v3.10 based) at the time14:03
apwFernandoMiguel, nope, i have turned it off, we suspected it was not ready for mainstream in v3.10 and your confirmation that is sucks is sufficient for me to just disable it pending it getting less crap14:04
rtgapw, so 3.11 should work better ? 14:04
apwrtg, i think the bigger question is "is v3.11 any better", for v3.10 we left it not applied in unstable so we would find out, i am proposing the same for v3.1114:04
FernandoMiguelapw: how many users actually even change schedulers, like I do ?14:04
apwrtg, as infinity as an avid user of affected kit we find out pretty smartly if it is utter crap14:05
rtgapw, ack.14:05
apwFernandoMiguel, heh dunno, but this is a default change so anyone with the kit is getting it by default14:05
apwit is new and shiney and meant to give you yet more performance for less power consumption14:05
apwwhich it clearly fails dismally to do14:05
FernandoMiguelon battery, it sure changes a lot 14:06
FernandoMiguelI've gonne from 5h+  to 2h14:06
sforsheertg: that function has so many things wrong with it14:06
FernandoMigueland that's ondemand14:06
rtgsforshee, off by one everywhere14:06
FernandoMiguelnow I need to track down unity folks.... unity service is hogging a core.... :\14:06
apwFernandoMiguel, heh only one, you are lucky14:06
FernandoMiguelbye! if you need any update or testing, I'm subbed to the bug and can be found on #+114:06
rtgsforshee, do you suppose it _ever_ worked ?14:07
sforsheertg: I doubt it14:07
FernandoMiguelapw: $ killall -9 unity-panel-service 14:07
FernandoMiguelthat's how I've been surviving 14:07
FernandoMigueleverytime I hear the fans go crazy14:07
rtgsforshee, well, I'm gonna hack on it to see if I can get it to stop faulting14:07
sforsheertg: I added a comment to the bug, definitely looks like stack corruption14:08
rtgsforshee, agreed14:08
apwFernandoMiguel, hopefully the peeps on #ubuntu-unity have it covered14:10
FernandoMiguelthanks14:10
jsalisbury**14:52
jsalisbury** Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting14:52
jsalisbury**14:52
QuintasanHi, I've recently discovered that my system stopped reading blanks DVDs. Since I didn't want to open up a new bug without any searching I did some digging around and last time anyone had a problem with it was Lucid. I did check the disk itself on another PC and it's blank for sure. I also did check it on Windows, blank there too. Output of `sudo /lib/udev/cdrom_id --debug /dev/sr0` is  http://paste.ubuntu.com/5858658/ and output of `sudo 15:13
Quintasanstrace -vvfo /tmp/cdrom_id_new.trace -s 1024 /lib/udev/cdrom_id --debug /dev/sr0` is http://paste.ubuntu.com/5858660 . The bug I found was bug #56158515:13
ubot2`Launchpad bug 561585 in udev (Ubuntu Lucid) "Blank cd/dvd not recognized [LUCID] " [High,Fix released] https://launchpad.net/bugs/56158515:13
rtgjdstrand, I'll have a look16:16
jdstrandrtg: thanks! :)16:16
rtgjdstrand, hmm, I would have expected all netfilter options to have been enabled. I'll look closer.16:18
jdstrandyeah16:21
rtgits not like the LOG target is new16:21
Sarvattapw: so I'm on 3.10.0-0 that has # CONFIG_NO_HZ_IDLE is not set CONFIG_NO_HZ_FULL=y, i'll verify intel_pstate is still crazy on -2 now16:22
apwrtg, grouper was missing a few last i looked, that was one of the patches i did, out of date now of course16:23
Sarvattah 3.10.0-2 still has that combo, someone responded on https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1188647/comments/616:23
ubot2`Ubuntu bug 1188647 in linux (Ubuntu) "Please change intel_pstate default to disable" [High,Fix committed]16:23
rtgapw, I thought I normalized all netfilter config options. 'UBUNTU: [Config] Enable and modularize all netfilter matches'16:25
rtgjdstrand, CONFIG_NETFILTER_XT_TARGET_NFLOG=y for mako16:26
apwrtg, ahh so you did16:28
jdstrandrtg: [    0.000000] Linux version 3.4.0-3-mako (buildd@tritons) (gcc version 4.7.3 (Ubuntu/Linaro 4.7.3-5ubuntu2) ) #13-Ubuntu SMP PREEMPT Tue Jun 25 17:07:54 UTC 201316:28
jdstrandrtg: is that too old? ^16:28
rtgjdstrand, nope, thats current16:29
rtgjdstrand, maybe I'm looking at the wrong option. 'CONFIG_IP_NF_TARGET_ULOG is not set'16:29
jdstrandthis isn't ulog16:30
Sarvatti'm building a kernel with CONFIG_NO_HZ_IDLE=y CONFIG_NO_HZ_FULL=n CONFIG_NO_HZ_FULL_ALL=n to test that guys theory16:30
jdstrandufw doesn't support ulog. I was going to, then no one did it for ipv6 and said it was bad or something, and so I said forget it16:30
jdstrandrtg: my desktop generic kernel has:16:31
jdstrandCONFIG_NETFILTER_NETLINK_LOG=m16:31
jdstrandCONFIG_NETFILTER_XT_TARGET_LOG=m16:31
jdstrandCONFIG_NETFILTER_XT_TARGET_NFLOG=m16:31
rtgjdstrand, yeah, looks like I missed some16:31
jdstrandCONFIG_NETFILTER_XT_TARGET_LOG is what we want I'm quite sure16:31
jdstrandNFLOG is the successor for ULOG16:31
rtgjdstrand, 'This option adds a `LOG' target, which allows you to create rules...'16:31
jdstrandok, cool16:31
=== Guest37969 is now known as mfisch
=== mfisch is now known as Guest41197
ogasawaraapw, rtg: bug 1196597, I've a one line patch to shove in for Saucy, I'll submit the SRU for Raring16:43
ubot2`Launchpad bug 1196597 in linux (Ubuntu Raring) "nic-modules udeb does not contain the qlcnic driver for Qlogic" [Medium,In progress] https://launchpad.net/bugs/119659716:43
rtgogasawara, ack16:44
rtgogasawara, get it in both master-next and unstable for saucy16:44
ogasawarartg: just want to make sure I won't mess anyone up if I shove this up16:44
ogasawarartg: ack16:44
apwogasawara, not touching anything here, shove away16:46
jsalisbury##16:55
jsalisbury## Kernel team meeting in 5 minutes16:55
jsalisbury##16:55
=== jsalisbury changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Ubuntu Kernel Team Meeting - Tues July 16th, 2013 - 17:00 UTC || If you have a question just ask, and do wait around for an answer!
* ppisati -> gym17:15
=== psivaa is now known as psivaa-afk
* rtg -> lunch17:34
* henrix -> EOD18:25
=== Guest41197 is now known as mfisch
=== mfisch is now known as Guest1021
rtgbjf, is there a Raring bug for 'UBUNTU: [Config] CONFIG_ARM_ERRATA_643719=y' ? Shall I do the same for saucy ?19:24
bjfrtg, i didn't create a bug. it was a FTBS fix :-(19:25
rtghuh19:26
rtgbecause of a stable update ?19:26
bjfrtg, yes19:26
bjfrtg, i forgot to run updateconfigs before submitting19:26
bjfrtg, i'd have to go look at where it came it19:27
rtgbjf, no big deal. I just wanna be sure saucy is consistent19:27
bjfack19:27
bjfrtg, it came in the 3.8.13.4 upstream stable release  bug 1199100  "ARM: 7752/1: errata: LoUIS bit field in CLIDR register is incorrect"19:31
ubot2`Launchpad bug 1199100 in linux (Ubuntu Raring) "Raring update to 3.8.13.4 stable release" [Undecided,New] https://launchpad.net/bugs/119910019:31
bjfjust fyi19:31
rtgbjf, I've got it set in both saucy branches19:31
=== Guest82317 is now known as jefferai
=== jefferai is now known as Guest55882
=== ferai is now known as jefferai
* rtg -> EOD20:10
hallyn_apw: just fyi, that sigchld delivery weirdness?  I can't reproduce it in 3.2, 3.5 or 3.7.  (mind you what we found before was a usage bug, but fixing that in lxc didn't help - so it looks like i ws facing one kernel and one userspace bug)20:46
hallyn_hopefully i don't offend when i say that bisecting ubuntu kernels is something i like to avoid20:46
apwhallyn_, it is something we all like to avoid20:46
hallyn_"Bisecting: a merge base must be tested"20:47
hallyn_so i'm just sort of manually bisecting by Ubuntu-3.7.0-7.15 versions, then i'll try and do upstream git commit (when i can nail down a reasonable ubuntu-ified config)20:48
hallyn_actually i guess i must be just about there.  3.7..3.8 shouldn't require too much config wrangling20:50
apwhallyn_, sounds sane21:06
=== JayF_ is now known as JayF
=== kentb is now known as kentb-out
xnoxso grouper kernel is old, such that avahi doesn't work out of the box, and needs it's parameters tweaked for kernels << 3.9.4 or something like that.22:35
ohsixwhat.22:36
ohsixsince when does avahi have 'parameters', and since when does it do anything but very basic network access22:37
=== RAOF_ is now known as RAOF

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