/srv/irclogs.ubuntu.com/2010/06/04/#ubuntu-kernel.txt

nealmcbHow do we find out what changed in the recent kernel update (to 2.6.32-22-generic #35-Ubuntu)?00:10
nealmcbThe changelog just seems to cover abi changes00:10
nealmcb (I was looking in /usr/share/doc/linux-image-generic/changelog.gz - that may not be the right place....)00:10
SquideshiDoes the kernel ever get updated before the next version of Ubuntu is released? For example, will there be a kernel update before Maverick?00:17
jpdsSquideshi: https://wiki.ubuntu.com/KernelFreeze00:18
ogasawaranealmcb: https://edge.launchpad.net/ubuntu/+source/linux and click the drop down for 2.6.32-22.3500:18
ogasawaranealmcb: although there was a kvm regression, so that patch was backed out and an updated 2.6.32-22.36 kernel is in the process of being uploaded00:19
nealmcbogasawara: thanks!00:20
SquideshiMaybe I should just ask the question I really want to know: I'm having a really difficult time troubleshooting the video on my system, because I think I have at least three different problems. First, Compiz doesn't work at all--had to be uninstalled. Second, the stock kernel that ships with Lucid has some problem with my Intel 845G that causes some display glitches and constant GPU lockups....00:25
Squideshi...Third, my screen often goes blank (backlight not lit) with a reboot being the only recovery, which I think is a separate problem from the last.00:25
SquideshiGood news is that I installed a drm-intel-next kernel and the second problem appears fixed.00:25
SquideshiAlthough, I wish I knew how to find out WHEN it was fixed (i.e. with what patch) so I could tell you guys.00:26
SquideshiAll of this isn't very easy to log as a bug because I don't know what's really affecting what.00:26
SquideshiI figure that the GPU lockups with the stock Lucid kernel has already been fixed in newer kernels.00:27
SquideshiI don't know if there's anything I should do about that.00:27
ogasawaraSquideshi: you could test the latest 2.6.35-rc1 mainline builds to see if it00:27
ogasawara's resolved00:28
SquideshiMy guess is that I should start looking at the blank (no backlight) problem.00:28
keesnealmcb: to answer your specific question, you can find the changelog in /usr/share/doc/linux-image-$(uname -r)/changelog.Debian.gz00:28
ogasawaraSquideshi: https://wiki.ubuntu.com/X/Bugs/Lucidi8xxFreezes might also help00:28
Squideshiogasawara: What's the difference between the drm-intel, drm-intel-next, and mainline kernels?00:28
ogasawaraSquideshi: drm-intel and drm-intel-next I presume are staging trees for patches which eventually get pushed to mainline assuming no ill effects00:30
Squideshiogasawara: Is blacklisting Intel 8xx hardware for KMS going to work, considering that Intel has removed ALL user mode setting in the newer drivers?00:30
RAOFSquideshi: Yes, because you'll get VESA - or, if you install it, an old forked -intel driver from Universe.00:31
SquideshiHmmm... That's a shame because it seems fixed with the newer kernel I installed.00:32
SquideshiThe glitching and GPU lockups that is.00:32
nealmcbkees: ding ding ding - just what I hoped for....  thanks :)00:33
keesnealmcb: np :)00:33
RAOFSquideshi: I'm surprised.  The GTT incoherency bug is still open upstream, and didn't appear to be being worked on.  You'll still be able to enable KMS and use the new intel drivers if you want - we only change the default, not forcibly disable KMS.00:34
SquideshiRAOF: I don't know if "GTT incoherency" is the same thing that I'm experiencing.00:34
RAOFSquideshi: You will be, but it's possible that it's sufficiently infrequent for it to appear as background noise.00:35
SquideshiRAOF: Like I said, it's difficult to separate everything out. I was so happy when the newer drm-intel-next kernel I installed resolved the flicker and GPU lockups that I thought I might be able to jut focus on the frequent screen blanking (no backlight.)00:36
SquideshiRAOF: After those two, I thought only then would it be appropriate to figure out Compiz.00:37
stentenIs there a kernel parameter to set the dmesg font during boot? Trying to catch a trace that keep scrolling off the screen.02:27
=== xfg is now known as zul
=== kamal is now known as kamal-away
=== oubiwann is now known as oubiwann_
=== hrw|gone is now known as hrw
laghttp://www-307.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-6788108:53
* hughhalf steps away for a bit09:02
ckinglag, http://kernel.ubuntu.com/git?p=cking/debug-code/.git;a=blob;f=led-flash-user-space/led.c;h=61131fc7caefcebc50326f48b630726e77e402b3;hb=ebb3c4b8e2c5ff1628f8b42076e6b9e969ed5f5209:02
=== ericm|ubuntu is now known as ericm-afk
laghttp://frommel.net/weblog/images/htc-desire.jpg09:25
=== cooloney is now known as cooloney-afk
rafiyrwhat's kzalloc?09:53
rafiyroh, nm, found the desc09:54
rafiyrstrange though09:55
jk-strange?09:55
rafiyrthere was a comment from dmitry on linux input suggesting kcalloc instead of kzalloc09:56
rafiyrseems to me they are pretty damn similar09:56
rafiyronly thing I see is the potential to take advantage of the knowledge of array structure in the future if calloc is used09:57
rafiyrwell that and I guess one extra check for kcalloc against ULONG_MAX09:58
jk-just makes it explicit that you're allocating for an array of stuff10:02
rafiyrOh, right there's the whole question of readability too :)10:50
rafiyrI was thinking in terms of implementation and possible optimizations.10:50
rafiyrFor example, if you were to run a kernel on an FPGA, you can actually do something useful knowing that a block of memory is actually an array10:51
* apw discovers xchat is not running ... hrm ... i thought you lot were very quiet10:53
amitkjk-: around? care to join #linaro?11:06
l3onHi all, someone of you can tell me if there is something wrong/missing in bug 589598?11:30
ubot2Launchpad bug 589598 in linux (Ubuntu) "B2C2 FlexCopII DVB chip / Technisat SkyStar2 DVB card does not work anymore (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/58959811:30
ograapw, amitk, https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/589624 for your radar11:47
ubot2Launchpad bug 589624 in linux (Ubuntu) "omap flavour does not work on beagle XM board (affects: 1) (heat: 8)" [High,New]11:47
amitkogra: ack11:48
amitkogra: I thought you said it worked before11:48
apwamitk, i suspect that is a bug in the omapfb driver ... as it failed to init, but the cursor flash timer is running11:50
ograamitk, nope, i was thinking i see the typical "no rootfs found" oops here, but after apw had me verify it with an actual rootfs in place it showed its really a bug11:52
amitkI'm not surprised if it is a DSS2 bug11:53
ograyeah, i hope we get that working in time for all these different devices11:53
apwamitk, are you still owning omap, or do i need to lean on someone else11:53
* ogra has to actually try the touchbook and zoom11:53
amitkapw: I don't _own_ it. Mathieu and cooloney-afk do. I'll just help them with bits and pieces. In this case I have the HW.11:54
apwamitk, ok so i'll hastle them then11:56
amitkapw: I am obviously interested though. So what will you hassle them about? ;)11:56
apwamitk, it not working obviously :)11:57
apws/hastle/make sure they are aware it is an issue and needs fixing/11:57
amitkheh11:57
apwogra, does it work on a regular beagle ?11:57
ograapw, havent tried yet :)11:58
ograwill do after the release meeting, but i would expect so11:58
apwogra, is the release meeting on ?  seems to be gone from my cal12:26
=== amitk is now known as amitk-afk
ograapw, according to a mail i got from davidm i have to attend12:57
ograapw, but i dont see it on my cal either12:57
apwogra, seems to be gone and i have this email12:57
apwFrom: Robbie Williamson <robbie.williamson@canonical.com>12:58
apwTo: Andy Whitcroft <andy.whitcroft@canonical.com>12:58
apwSubject: Canceled Event: Maverick Weekly Release Meeting @ Weekly from 16:0012:58
apwso i think its not there any more ...12:58
apwi don't have agenda either12:58
* apw asks on #ubuntu-release12:59
apwogra, seems not till next week13:05
ograsweet !13:06
ograextra free hours with that wonderful weather !13:06
apwyeah indeed so13:06
=== amitk-afk is now known as amitk
apwJFo, about ?14:24
JFoI am indeed apw14:24
apwJFo, roomies for sprint ?14:24
JFosounds good to me :)14:24
apwcool i'll put you in then ... i note you are missing from the list right now14:25
JFoah yes, still hashing out travel with Atlas14:25
JFo;-/14:25
apwcome on, you'd love to row all the way14:26
apwjjohansen, about?14:28
dupondjeJFo: commented my bug again, added another dmesg output. Totally other error .. guess my BIOS/Mobo if00bar :)14:45
JFoheh, it happens sadly enough :-/14:46
dupondjeIts a new bios that supports new cpu's, but seems like its quite bad support :(14:46
dupondjeah well, I close the bug as 'Invalid' ?14:46
ograapw, so i'm seeing the ARM meeting minutes on the kernel ML, thats not talking about our kernel we ship on the ubuntu images, is it ?14:48
=== sconklin-gone is now known as sconklin
apwogra, thats talking about how the linaro kernels might make it into the archive14:50
ograapw, ok, but in separeated out binaries please14:50
apwso that should be 'all' the arm branches but not those in master14:50
ograi dont want to risk image stability for omap3/4 images14:50
apwogra, right this is not fixing the 'one actual binary to rule the world' issue14:51
apwits likely the package would produce 5-6 binary image sets14:51
ograthats fine as long as i have a stable version for omap3 and a stable version for omap414:51
ograi just dont want experimental kernels in my images14:52
JFodupondje, sounds good15:00
JFowho could blame you ogra :)15:01
=== bjf[afk] is now known as bjf
bjfmoin all15:25
apwmoin15:28
JFomoin bjf15:32
=== Hrww is now known as hrw
JFoapw, did you get the chance to create my testing directory on the kernel PPA?15:43
JFoor was it you I was even talking with about it?15:43
* JFo forgets15:43
Kanohi apw 15:46
Kanois there a custom enforce file too? that i dont need to change the other one that will override the default enforce?15:47
apwJFo, can't remember ... can do it now15:47
apwKano, not currently there is just one per series15:47
JFoif you like, was just trying to remember if we had discussed it at all15:47
apwJFo, the scripts running as you i assume yes15:48
Kanoapw: it would be a good idea however, something thats used after the default, so i could put my changes in there15:48
JFoI didn't want to blindside you with it if we hadn't15:48
JFoapw, think they are running as bjf15:48
JFobut I defer to him15:48
Kanoapw: the n value is for unset or?15:48
Kanousually there is no n15:48
apwKano, yes n means #FOO is not set15:48
bjfJFo, which scripts?15:48
JFothe createiso stuff15:49
JFobjf, was seeing if apw minded setting us up a dir in the PPA15:49
JFofor the testing ISOs15:49
apwyep seems fine15:49
apw1) what you want to call it ?15:49
bjfJFo, apw yes they are running as a personal cronjob on emerald15:49
apw2) who needs to be able to write to it15:49
Kanoapw: do you know why mantis driver has still no pci ids enabled15:50
JFotesting is fine with me, what do you think bjf?15:50
Kanoapw: echo "MODULE_DEVICE_TABLE(pci, mantis_pci_table);" >> v4l/mantis_cards.c15:50
Kanosomething like that would add those15:50
JFoas to who writes to it, hmmm I assume bjf will be copying over.15:50
apwKano, no idea, not something i've run into15:50
JFoapw, I think he is generating them on emerald15:50
apwJFo, will play a minute thanks15:50
JFoso they just need to copy over iirc15:50
Kanothats stupid because when you are used to dvb-s2 lip15:50
JFok15:51
Kanothen there was always a pci-id table15:51
Kanojust the normal mantis in the kernel does not15:51
bjfthat's a pretty broad name, as far as who, we can make it anyone, is this something I am going to own or is someone else going to take ownership (doesn't matter to me)15:51
JFoI actually prefer not to lock it down to you now that you mention15:51
JFobus test and all, God forbid15:51
bjfdo we have a non-real user launchpad user that this could run as?15:53
JFokernel-janitor15:53
achiangdoes anyone know if the latest released security kernel has other fixes from -stable?15:53
JFoachiang, apw knows... poor man15:54
apwachiang, it has only the single revert15:54
* JFo fixes apw a large beer15:54
* apw thanks JFo 15:54
JFomy pleasure15:54
achiangapw: hm, i wasn't paying attention to what was released yesterday. it only had some patch that broke kvm and no -stable updates? and now the most recent kernel reverts that single patch. do i understand that correctly?15:55
apw.35 had about 5 security related updates, .36 reverted just one of them15:56
apwthere was no -stable in either of them15:56
achiangapw: ok, thanks15:56
manjoJFo, 1/2 day bug day is morning or afternoon? are we just picking bugs at random from the list ? 16:12
JFoeither morning or afternoon16:12
JFoyou work through the ones on the Top5016:12
JFoto get them moved off if possible16:12
apwbjf got a sec ?16:14
apwmanjo, whenever you want16:14
manjoapw, yep I just started looking at some so was not sure 16:15
manjoJFo, ^16:15
JFocool16:15
bjfapw, what can i do for you? (i cringe at the thought)16:19
* apw mumbles at you16:20
JFoheh16:20
bjfapw, give me a sec. mumble non-responsive16:21
apwbjf, heh what fun16:21
* JFo realizes he's not mumble ready... sitting here with the headset on for no reason apparently16:21
JFosad given I've had a pot of coffee already today16:22
apwno sleep for you16:22
apwJFo, doh16:22
JFodidn't get any last night16:22
JFothat is the reason for the coffee16:22
JFo:(16:22
JFostupid headache16:24
JFoogasawara, are you around yet?16:25
ogasawaraJFo: yep, here16:25
JFocool16:25
JFoI am going through my list of things not to forget...16:25
JFodid I tell you that James Westby asked me if we wanted him to turn on kerneloops again?16:26
JFoor something like that16:26
JFoit is what I have written down from UDS16:26
apwwe norally have it on for development16:26
JForight16:26
ogasawaraJFo: we can, it's your call now :) usually we'd crank in on around alpha116:27
JFook, I'll ping him on it then :)16:27
JFohe made it seem as if it was my call, but I wanted to verify before I made the change16:27
ogasawaraJFo: go for it16:28
JFocool16:28
JFotgardner, you around?16:33
tgardnerJFo, yo16:35
JFotgardner,  cjwatson indicates that the bnx2 module was broken in bug 58930416:36
ubot2Launchpad bug 589304 in linux (Ubuntu) "Broadcom bcm5708 ethernet not initialized during install (affects: 1) (heat: 6)" [High,Triaged] https://launchpad.net/bugs/58930416:36
JFowas just wondering what my next step should be16:36
JFoand I wanted to ask you based on your intimate familiarity :)16:36
cndapw: http://kernel.ubuntu.com/git?p=cndougla/hedley.git;a=shortlog16:37
tgardnerJFo, looking...16:38
JFok, thanks16:38
tgardnerJFo, its a missing firmware problem. seems like a regression, but I dunno why.16:38
JFohe seemed to think it was just broken... and at that point I was lost :)16:39
JFotgardner, looks like fader removed and reinserted the firmware but that didn't help16:40
JFoso it looks like it was there, just not functional for some reason16:40
tgardnerJFo, looks like the maverick driver wants an updated firmware file which we donot yet have.16:41
JFoah, I see16:41
tgardnerJFo, I've updated the LP report and taken ownership16:44
JFothank you sir16:45
manjoapw, http://kernel.ubuntu.com/~kernel-ppa/mainline/daily/current this is correct? I get 40416:48
=== amitk is now known as amitk-afk
apwmanjo, i removed one to rebuild it so current may currently not point anywhere while it builds16:49
manjoapw, ah so it will come back later ? 16:50
manjowhat version are you building right now ? so that I can put that in my report 16:50
apwso it will come back later yes in theory16:53
apwwhat report ?16:53
apwogasawara, if you see a late build failure related to aliases, let me know (with your new kernel)16:58
apwogasawara, i think i know what it is and am working on it16:58
* apw pops out16:58
ogasawaraapw: building cleanly for me on i386 and amd64, just failing on the ports16:58
jjohansenapw: whats up?17:04
=== hrw is now known as hrw|gone
cndgit rebase --onto mvl-dove branchpoint merge17:09
tgardnerJFo, have you seen any bugs complaining about server consoles selecting the wrong display resolution if there is a KVM in the way? I have a case where the KVM supports up to 1680x1024, but the attached monitor only supports 1024x768.17:29
tgardner1680x2048*17:29
tgardneror something....17:29
JFotgardner, not off the top of my head, but I can look17:30
JFoactually, I think i have seen some17:30
tgardnerthanks17:30
JFolet me dig a bit17:30
cndwhen upgrading, 2.6.32-22.21~v1 is seen as later (i.e. newer, should be upgraded to) than 2.6.32-22.2117:33
cndcorrect?17:33
tgardnercnd, yep17:33
cndtgardner: thanks17:33
Kanousually ~ is before17:34
tgardnercnd, lemme retract, 2.6.32-22.21~v1 is ealier17:34
Kano- after17:34
cndwait, so I have it backwards?17:34
tgardneryes17:34
cndok17:34
tgardnerthe tilde is like a minus17:34
cndis the ~ special?17:35
tgardneryes17:35
cndohhhh17:35
cndso I would have been right if I appended 'v1' instead of just '~v1'?17:35
tgardnera tilde lexicographically subtracts from the version17:35
cndok17:35
=== cking_ is now known as cking
=== cking is now known as cking-afk
JFoany reason why we haven't built 2.6.31.13 in the mainline PPA or do we stop doing that at a certain point?18:31
apwJFo, hrm not sure18:32
JFosaw the question in a bug and thought I'd ask... after verifying that there was, in fact, a .13 version :)18:33
apwJFo, but there is a 2.6.31.1318:35
JFohmmm18:36
JFo.me looks in mainline18:36
JFoah, it is the http://kernel.ubuntu.com/~kernel-ppa/info/kernel-version-map.html which doesn't show it18:37
JFoonly has one image and one header while the others have 2 of each. that sound correct apw?18:38
JFoerrr in some cases 3 of each18:38
apwahh you meant its there and its broken18:38
JFocould be18:38
* JFo is not sure what he means today18:39
JFobeen a very bad day18:39
apwJFo, ok its not obvious that its not a bad buids, so rebuilding it18:48
=== cking-afk is now known as cking
JFok, thanks apw18:48
=== bjf is now known as bjf[afk]
* bjf[afk] -> lunch19:31
* kees starts using ccache ...19:31
jjohansen-> lunch19:46
ckingmanjo, did you get that PC?19:53
manjocking, was just receiving it 20:01
manjocking, that is fast shipping 20:01
ckingmanjo, note that it's already pre-installed with karmic - so it may be a good idea to see how that's configured 20:15
kassah_what does the extra minor version mean on proposed kernels? i.e. 2.6.32.14.5 (The last .5)20:15
ckingif you have any questions, send me an email, I'm having to finish up for the day now20:15
cking^^ manjo20:15
manjocking, yep will power it up as soon as I get a  power cable20:15
manjocking, play with it over the weekend and ping you on Monday20:16
ckingmanjo, perhaps we can find time next week to walk through it20:16
manjocking, good idea20:16
ckingcool - have a good weekend!20:16
manjocking, you too20:16
ckingta! bye!20:16
keeswhat's the best way to keep my ubuntu-maverick git tree on zinc up to date ?20:22
sconklinkees: it depends - do you just want to keep the branches from the origin updated, or do you want to rebase your own branches or anything like that?20:31
sconklinkees: I haven't done it on zinc, but I have run cron jobs on my server at home to fetch from upstreams20:32
keessconklin: well, I think I've sorted it with the alternate file20:32
keessconklin: but I'm curious how to push from my local git, in a branch named "maverick-ptrace" to my remote maverick's master.20:33
sconklingit push remotename localbranchname:remotebranchname20:33
sconklinso git push remotename maverick-ptrace:master20:34
keesah-ha, colon.  more man page fail.20:34
sconklinyeah, that one was not obvious at all to me20:34
keesokay.20:35
sconklinbe aware (I think) that this is one of the situations where pushing a blank branch name deletes the remote branch.20:35
keeshttp://kernel.ubuntu.com/git?p=kees/ubuntu-maverick.git;a=summary  <- the tag for 5.14 is missing here, and I don't know why20:35
sconklinoh, ask apw but I think unless you push the tag as the identifier or push --tags, they don't all go. But --tags can push a lot of crud you don't want, since tags are across all branches20:36
keeshm20:37
sconklinI need to reread that chapter in the book, actually20:37
keesheh20:37
apwsconklin, kees, right you have to push them explicitly which may mean ogasawara didn't push it20:41
ogasawaraoops, /me check20:42
keesno, it's my tree that is missing it20:42
apwkees, i have seen tags get lost on fetch20:43
keeshttp://kernel.ubuntu.com/git?p=ubuntu/ubuntu-maverick.git;a=summary vs http://kernel.ubuntu.com/git?p=kees/ubuntu-maverick.git;a=summary20:43
apwyou only get a tag if the tag points to a commit when you get that commit20:43
apwso if ogasawara pushed, you fetch then she pushed the tag you may not get it20:43
apwkees, git fetch --tags origin should get you any missing ones20:44
keesah, so I have a single local repo that has both linux-2.6 and ubuntu-maverick as remotes.  How do I get both sets of tags?20:45
keessounds like I just want to add --tags to my "fetch" uses20:46
apwkees, normally you should get them automaticall20:46
apwit should get any tags pointing to any new commits you get, i've only seen it out of sync once or twice and used --tags to fizx20:46
apwfix20:46
keeshunh, still can't seem to push the tags.20:50
keesif I add --tags to the push, it tries to upload the entire change history.20:51
JFoogasawara, kerneloops re-enabled now per James_w21:04
=== bjf[afk] is now known as bjf
apwJFo, that 31.13 finished ok this time21:26
JFoyep, I saw it. Thank you sir :)21:27
lagNight all!22:05
* lag is sleepy!22:05
=== sconklin is now known as sconklin-gone

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