=== a16g_ is now known as ypwong | ||
=== mfisch` is now known as mfisch | ||
=== mfisch is now known as Guest10530 | ||
ppisati | moin | 05:40 |
---|---|---|
=== fmasi_afk is now known as fmasi | ||
apw | ppisati, moin | 07:57 |
ppisati | apw: moin | 07:57 |
apw | man that was an early one too | 07:57 |
ppisati | apw: neh, i usually wake up way earlier than that | 07:57 |
apw | wierd gym lover :) | 07:58 |
=== neunon_ is now known as neunon | ||
zequence | Anyone available for a fairly simple upload (artwork update)? lp:ubuntustudio-menu | 08:32 |
zequence | Ah, sorry. Wrong channel :) | 08:33 |
=== DavidDuffey is now known as dduffey | ||
apw | zequence, i might be able to help if it is truly trivial :) | 09:25 |
=== ghostcube_ is now known as ghostcube | ||
apw | oh i so did not want to know anything about autopkgtest environments | 10:50 |
lerra | Hey guys, im having a couple of machines running different versions of the kernel 2.6.32-*-server (mostly 40 and 42), I have a vmcore file that I am trying to analyse but I cant find a debug kernel for that but I manage to analyse the binary with strings and manage to get the following content from http://pastebin.com/hCW6W8p9 | 10:56 |
lerra | Anybody seen this ? Its basically a postgres hammering (300-600mb/s) the system (on a mounted xfs partition) | 10:57 |
apw | lerra, not seen that befoer indeed. the oops says that the thread overran the stack, which if accurate is very bad | 10:59 |
apw | lerra, have you looked on ddebs.ubuntu.com for the ddebs ? | 11:01 |
apw | lerra, and those seem anchient kernels | 11:02 |
apw | lerra, ok i can see -40 on ddebs.u.c, but -42 seems to be awol, but the latest is -51 so we are a heap of pain away from soemthing current here | 11:06 |
apw | lerra, i would if you can see if you can get something a bit more recent in there | 11:06 |
lerra | in ddebs ? I did check it and noticed the same as you :) | 11:21 |
zequence | apw: Thanks. I had forgotten there's a procedure for this, i.e. file a bug and subscribe ubuntu-sponsors. Seems we already got it uploaded :) | 11:22 |
lerra | Where did you find the -40 kernel ? and what do you mean with awol ? | 11:23 |
lerra | found http://ddebs.ubuntu.com/pool/main/l/linux/linux-image-2.6.32-40-server-dbgsym_2.6.32-40.87_amd64.ddeb | 11:25 |
apw | zequence, ok good | 11:25 |
apw | lerra, yeah -40 seems to be where you just indicated, the -42 is missing, but that is so old now i suspect there is no hope of recovery | 11:25 |
lerra | I only got a vmcore from a -42 but I will try and see how far I can go with that | 11:25 |
apw | well that stack you indicated clearly shows a very very deep xfs stack | 11:26 |
apw | and deep stacks are illegal, so if it is real, and it looks believeable at least | 11:26 |
apw | then xfs is probabally at fault. there may be fixes for that in later kernls and you are some 10 abi's behine | 11:27 |
apw | behind | 11:27 |
apw | (awol == absent with out leave, ie missing) | 11:27 |
lerra | oki, will see if I can dist upgrade it to 12.04 from 10.04 | 11:30 |
lerra | or maybe a backported kernel | 11:31 |
apw | lerra, i was more thinking the latest kernel on the release | 11:31 |
apw | lerra, at least there if it reporoduces you know you are on the latest and we have debug symbols | 11:32 |
lerra | basically to the -51 version, right ? | 11:33 |
lerra | even 49, I will see how I can proceed | 11:34 |
apw | lerra, i would say the one in -updates if you are going anywhere | 11:34 |
lerra | thanks apw, very helpfull, not to mutch documentation on this out there | 11:34 |
apw | lerra, np latest is: linux | 2.6.32-51.113 | lucid-updates | source | 11:35 |
jafeha | hi, i'm looking for someone with insights into intel lynnfield cpus on ubuntu 12.04 lts (enablement stack / kernel 3.8.0-30-generic). can anyone help me? | 12:21 |
=== thomi_ is now known as thomi | ||
jafeha | my question is: i'm running my ubuntu 12.04 with an lynnfield i5 760, which officially has no hyperthreading support. if i check /proc/cpuinfo i get the cpuflag "ht" which should indicate hyperthreading. ht is not activated on the host as far as i can see: lscpu tells me that i've got only one thread per core. so could i actually turn on the hyperthreading bootflag and acutally get more threads per core? or is th | 12:43 |
=== bladernr` is now known as bladernr_ | ||
cking | jafeha, i was under the impression that the i5 760 has 4 cores and not hyperthreads | 13:03 |
apw | i have the feeling that the ht flag is a lie in some cases, "else windows does not see all teh CPUs" or something | 13:12 |
jafeha | ok thank you cking and apw thats what i already thought. i just wanted to be sure. there were some discussions on the i5 760 enabling ht, but it did not make it into the specs. so how much can i actually trust those flags being shown in cpuinfo? | 13:53 |
=== kentb is now known as kentb-afk | ||
cking | jafeha, the cpuinfo flags come from what cpuid is reporting on the CPU, so one has to trust the CPU not to be lying I guess | 13:54 |
jafeha | hum, what a mess. :D | 13:55 |
jafeha | what would actually happen if i tried turning it on in the bootflags? | 13:55 |
apw | jafeha, i would say that the key is how many siblings are reported, as the kernel will use all the ones it finds it does not use the HT flag to determin which cpus to enable | 13:58 |
jafeha | last thing i found out about that was that the generic kernel won't activate ht without using ht=on as a boot parameter. thats why i'm asking. | 14:01 |
arges | hey guys, is there a tool that automagically pulls in patch deps (based on context) when cherry-picking? doing it manually with git blame / log is fun and all | 14:01 |
apw | arges, not that i know of | 14:02 |
arges | i'll put that on my list of 'projects to work on when you have the time', if i ever have the time | 14:03 |
apw | jefferai, are you sure, where have you seen that documented | 14:04 |
apw | arges, i generally use selective log to see what all else changes the same files | 14:04 |
arges | yea git log -S'blah' /path | 14:04 |
apw | jafeha, are you sure, where have you seen that documented | 14:07 |
apw | jafeha, i can find no documentation for that and it doesn't fit my experience either | 14:07 |
apw | jafeha, i have tended to get the CPU name out of cpuinfo and look that up on intel (or whoevers) website to be sure | 14:08 |
=== Guest10530 is now known as mfisch | ||
xnox | sforshee: hello, stgraber says you are good with backlit brightness. I can echo values between 100 -> 4882 into /sys/class/backlight/intel_backlight/brightness but the keyboard buttons / brigtness settings in the gnome control centre. | 14:22 |
=== mfisch is now known as Guest22239 | ||
xnox | sforshee: any idea what I can tweak to make the buttons work? =) | 14:22 |
sforshee | xnox: are there other devices under /sys/class/backlight? | 14:23 |
xnox | sforshee: yes, acpi_video0 | 14:24 |
sforshee | xnox: userspace is going to try to use that ahead of intel_backlight | 14:25 |
sforshee | does writing values to its brightness file work? | 14:25 |
xnox | sforshee: nope, no effect. | 14:26 |
sforshee | xnox: a workaround is to pass acpi_backlight=vendor to the kernel | 14:27 |
sforshee | but you should probably file a bug | 14:27 |
xnox | sforshee: i had that, didn't work. (had that in grub cmd line) | 14:27 |
stgraber | xnox: that's not the line I gave you, did you try that specific one before? | 14:27 |
xnox | stgraber: yeah, i had acpi_backlight=vendor before trying yours. (found it somewhere online) | 14:28 |
stgraber | k | 14:29 |
sforshee | xnox: do you have the acpi_video0 backlight when you boot with it? | 14:29 |
xnox | Hm: https://bugzilla.kernel.org/show_bug.cgi?id=50551 | 14:29 |
ubot2` | bugzilla.kernel.org bug 50551 in Power-Video "ACPI uses acpi_video0 instead of intel_backlight for Intel GMA 4500M" [Normal,Closed: will_not_fix] | 14:29 |
sforshee | xnox: that's not a bug per se, the rules are that firmware interfaces should be preferred over raw interfaces (which is what a backlight from a graphics driver would be) | 14:30 |
sforshee | the bug is that acpi_video backlight doesn't work | 14:30 |
sforshee | but it's probably a firmware problem | 14:31 |
xnox | sforshee: stgraber: right so with acpi_backlight=vendor I get under /sys/class/backlight ideapad and intel_backlight | 14:33 |
xnox | intel_backlight works, ideapad doesn't. | 14:33 |
sforshee | xnox: bah. Yeah, and ideapad will be preferred over intel. | 14:34 |
sforshee | xnox: ikepanhc might be able to help you with the ideapad backlight | 14:38 |
sforshee | xnox: for the acpi one, file a bug and point me at it so I can look at your acpi tables | 14:38 |
xnox | sforshee: right. Well in the end I added xorg.conf.d snippet bug 1158934 | 14:50 |
xnox | https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1158934/comments/4 | 14:50 |
xnox | I guess idepad should also be blacklisted just like the HP video. | 14:50 |
=== hggdh_ is now known as hggdh | ||
xnox | sforshee: is bug 1227188 ok? | 15:00 |
xnox | bug #1227188 | 15:01 |
xnox | ubot2`: awake? | 15:01 |
ubot2` | Factoid 'awake?' not found | 15:01 |
xnox | https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/1227188 | 15:01 |
sforshee | xnox, got it | 15:01 |
sforshee | xnox: can you also attach acpi tables? acpidump > acpi_tables.txt | 15:03 |
xnox | sforshee: attached: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1158934/+attachment/3827417/+files/acpi_tables.txt.gz | 15:04 |
sforshee | xnox: ta | 15:05 |
phillw | Hi folks, we seem to have narrowed down an issue with freezing to being a version of zram that doesn't work correctly in kernel 3.11.x How should this be reported? | 15:09 |
jsalisbury | phillw, A launchpad bug is best | 15:11 |
phillw | against kernel? | 15:11 |
sforshee | xnox: there's nothing obvious in the acpi tables to explain why the acpi_video0 backlight doesn't work. I wonder if the backlight iplementation there just doesn't work. | 15:11 |
jsalisbury | phillw, yes. You can use the linux package. For example, from a terminal run: ubuntu-bug linux | 15:12 |
phillw | thanks, will do and will add in all the we have found (the last part of the jig saw arrived about an hour ago) | 15:12 |
xnox | sforshee: i believe it just plain doesn't work, and the other one does =) | 15:13 |
sforshee | xnox: well some lenovos have another workaround which works, but not yours | 15:14 |
sforshee | it seems like lenovo doesn't test the acpi backlight implementation in their firmware on machines shipping with win8 | 15:14 |
xnox | =/ i did have windows 8 pre-installed on it but no obvious "Certified for Windows 8" logos or some such. | 15:19 |
xnox | like I don't have "Windows 8 compatible sticker" only windows 8 sticker on the back. | 15:19 |
=== kentb-afk is now known as kentb | ||
=== Guest22239 is now known as mfisch | ||
=== mfisch is now known as Guest20042 | ||
phillw | jsalisbury: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1227202 I've just asked those also affected to mark themselves as such. | 15:27 |
jsalisbury | phillw, great, thanks | 15:27 |
ppisati | infinity: you there? | 15:29 |
kdub | how would one compile a kernel for ubuntu touch? | 16:08 |
=== psivaa is now known as psivaa-afk-bbl | ||
=== Guest20042 is now known as mfisch | ||
=== mfisch is now known as Guest66657 | ||
phillw | jsalisbury: I'd need some instructions as to how get the kernel into my machine and then see if the Error: /dev/zram0: unrecognised disk label and Error: /dev/zram1: unrecognised disk label disappear (that is assuming this fix will also solve that issue). | 17:45 |
jsalisbury | phillw, From a terminal, you can run: sudo dpkg -i PACKAGE_NAME to install the two kernel .deb packages. | 17:46 |
jsalisbury | phillw, It would be the linux-image and linux-image-extra .deb packages | 17:46 |
jsalisbury | phillw, A reboot is then needed | 17:46 |
phillw | so, a wget into what directory ? | 17:47 |
phillw | is my home directory okay? | 17:47 |
jsalisbury | phillw, sure, any directory should work | 17:48 |
phillw | downloading (sorry for being a n00b to this!) | 17:49 |
jsalisbury | phillw, You can go to that link and click the two package names to download them: linux-image-3.11.0-7-generic_3.11.0-7.14~lp1227202v1_amd64.deb and linux-image-extra-3.11.0-7-generic_3.11.0-7.14~lp1227202v1_amd64.deb | 17:49 |
phillw | wget'ing them into my ~/phillw directory now. | 17:50 |
phillw | jsalisbury: will it be | 17:53 |
phillw | sudo dpkg -i linux-image-extra-3.11.0-7-generic_3.11.0-7.14 | 17:53 |
phillw | to install them? | 17:53 |
jsalisbury | phillw, use the full file name, so: sudo dpkg -i linux-image-3.11.0-7-generic_3.11.0-7.14~lp1227202v1_amd64.deb | 17:54 |
jsalisbury | and sudo dpkg -i linux-image-extra-3.11.0-7-generic_3.11.0-7.14~lp1227202v1_amd64.deb | 17:54 |
phillw | okies, thanks! | 17:54 |
jsalisbury | phillw, install linux-image first then linux-image-extra | 17:54 |
phillw | will do, no hardware update bit, I take. | 17:56 |
=== fmasi is now known as fmasi_afk | ||
phillw | see you shortly! | 18:01 |
phillw | jsalisbury: sudo parted -l still has the Error: /dev/zram0: unrecognised disk label and Error: /dev/zram1: unrecognised disk label | 18:06 |
phillw | To test against an ISO it would need injecting into one. And that is way, way beyond anything I even begin to understand! | 18:08 |
jsalisbury | phillw, thanks for testing. Does the test kernel resolve bug 1227202, which is the freeze when using zram? | 18:15 |
jsalisbury | phillw, the other issue may need to be a seperate bug | 18:15 |
phillw | jsalisbury: that only appears when installing from an ISO. that kernel would have to be on the kernel of the ISO? | 18:16 |
=== Guest66657 is now known as mfisch | ||
jsalisbury | phillw, ahh, ok. | 18:16 |
=== mfisch is now known as Guest6677 | ||
jsalisbury | phillw, yeah, creating a custom iso isn't very easy | 18:17 |
phillw | jsalisbury: the only other way would be to add it to the build list of lubuntu desktop amd64 daily and my set the system to do a rebuild of just that one ISO ? | 18:18 |
phillw | but, again, that is well beyond my limited abilities - possibly someone on the release team could help? We (lubuntu testers) would be more than happy to test it for you. | 18:20 |
jsalisbury | phillw, ack, I've been looking into doing that for another bug | 18:20 |
phillw | if you add it to that ISO, just give me a ping and we'll happily descend on it like a pack of vultures :) | 18:21 |
phillw | 2 days before freeze for next beta, so we do still have (a little) time. | 18:22 |
=== psivaa-afk-bbl is now known as psivaa | ||
phillw | having said that, a poorly kernel is not really a candidate for beta testing.... But, not my call :) | 18:23 |
kdub | where does the nexus 4 kernel come from? I heard its stock CM kernel with some patches and config file changes... | 18:49 |
=== fmasi_afk is now known as fmasi | ||
apw | kdub, that is a pretty accurate description | 19:51 |
kdub | apw, so... just the changes in the 'porting guide' is what's needed? | 19:52 |
apw | kdub, not sure what the porting guides says, but the end of the kernel log will show you what we did to it | 19:53 |
=== fmasi is now known as fmasi_afk | ||
=== fmasi_afk is now known as fmasi | ||
=== fmasi is now known as fmasi_afk | ||
zequence-work | apw: orig file for linux-lowlatency saucy? | 20:44 |
infinity | apw: Did you want to do tools things to lowlatency too? | 21:27 |
=== kentb is now known as kentb-out | ||
=== Guest6677 is now known as mfisch | ||
=== mfisch is now known as Guest72565 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!