/srv/irclogs.ubuntu.com/2010/07/29/#ubuntu-kernel.txt

amitkmorning07:46
lagMorning07:47
aboganiMorning07:50
cooloneymorning, eu08:10
apwmoin08:37
jk-hey apw 08:45
apwhi08:45
ericm_apw, cking, is smb on vacation this week?08:59
jk-all of the UK is taking a swap day tomorrow09:00
trapmaxi'm trying to create an lvm-snapshot. my machine just hungs there. here's the kern.log http://pastebin.com/KUzVGZSQ09:01
ericm_jk-, ok09:02
ericm_apw, so I have to bug you again09:02
apwericm_, smb is in today09:02
ericm_after mvl-dove branch is rebased, all the tags could possibly be inconsistent with the current history right?09:03
* smb is there09:03
apwericm_, the tags are correct in that the point to the tree at the time the release was made09:03
ericm_morning smb09:03
apwthe release commits in the latest copy of the tree do not correspond to the tags ... but the tags are right09:03
ericm_apw, yeah - but now the rebased commit isn't09:04
ericm_I mean the equivalent one, you know what I mean09:04
apwright ... that is correct, but only the tag matters09:04
ericm_apw, ok I see09:04
smbThe tags should still be there but point to something not linearly connectect to the releas09:04
apwyou don't use the release commit to find the tree for the release, you use the tag, that is what the release was really made from09:05
ericm_smb, apw, OK I see09:05
ericm_smb, one other thing - why don't we just put maint-startnewrelease to replace the fdr startnewrelease?09:07
smbericm_, Mostly because I just recently wrote it09:07
ericm_smb, mostly I much enjoyed using that script :-)09:08
smbericm_, Another thing that is a bit of a problem it that those scripts use other libraries in the kteam-tools domain09:08
smbericm_, Thanks, I enjoy those too. Take some of the manual labour away09:09
ericm_smb, ok - I copied the lib directly to my /usr/lib/python, though not sure if that's the right way09:10
ericm_smb, maybe for the scripts in kernel package, we could just hardcode those libs into the scripts, but that creates maintenance problems :-/09:11
smbericm_, From my intention you would just add the kteam-tools/maintsacripts to your PATH and all should work09:11
smbAll scripts look for libraries relative to their own path in kteam-tools/lib09:12
ericm_Grr.... I copied those to ~/bin, ugly hacks09:12
smbHmm, seems its only me that thinks adding a path is much simpler. :-P09:13
* apw notes he mentioned that smb adds that dir to their path ...09:14
smbapw, That dir? Which one09:14
apwkteam-tools/maint-scripts09:15
apwthe one you said just now, i mentioned it yesterday when i mentioned the scirpts for doing start new release, which is how he knows about it!09:15
smbAh, ok. Understand now09:15
apwsmb, that the scripts don't cope correctly still if you run them directly ... is still annoying, and something i mean to fix09:15
smbapw Hm, well should be a relative simple fix09:16
apwyep, should be, just not obthered to look yet09:16
smbapw, Just that I never saw that much need as I got it in my path09:16
apwits a proformer fragment of code needed to do it09:17
smbBecause I am lazy and don't want to type the complete path all the time09:17
smbapw, The problem likely is this needs to go into all scripts looking for a lib09:17
apwand everything looking for a binary too09:18
smbHuh?09:18
apwother binaries called from the binaries (or scripts) needs it too, so i don't have to add the path09:18
smbWell, thats what PATH is for09:19
apwbut for a small code change it can work both ways and ~/git/kteam-tools/maint-foo/maint-bar09:19
smbI find it somewhat strange to expect a script or binary will find anything somewhere09:19
apwwill work as expected instread of exploding randomly for some commands09:19
apwplus actually just using path is dangerous as i could have two versions installed09:20
apwand get the libraries or binaries from the other one09:20
apwthe tools come as a kit, and should ensure they use the ones from the kit09:20
apwi am not asking you to do the work, i'll get to it when it next annoys me09:21
smbapw, Somehow you sound like asking for bulletproof shoes while waving around your gun :-P09:21
apwsmb, its a personal way of working that i use commonly with other applications is all09:22
apwand its odd that that one does not follow the same rule09:22
apwi tend to not have random scripts on my path that have low utility, by which i mean i use them seldom09:22
smbAs odd as it is to me that you don't work like me. 09:22
smbBut ok09:22
apwespecially ones which mess arround with the current directory state09:22
apwas there is inhearnt risk of running the wrong thing and shitting all over what you are doing09:23
apwso if i am doing a review i can see adding that dir to the path for that session and using lots of scripts for a bit there09:23
apwbut normally to just get one function out of it, i'll just pick it by absolute path09:23
smbI can try to put in a more robust lookup and you can do the others09:23
apwas i say, don't feel you need to do anything09:24
apwi have lots of experience of doing this mod, as i have it in everything i write for the same reason :)09:24
smbI would do it just once in one script to have a reference for it. Also in case I do more scripts09:25
apwwatch out for the path case, ie there are three cases ./xxx ~/foo/xxx and xxx via path, they all present differenet via $009:26
apwin my scripts i typically have a $here which is instantiated from $0 and basically allow me to find everything else .... include $here/../lib/foo $here/oother-binary09:27
smbI believe there is an abspath which returns the complete path, which I can use instead of getcwd alone09:27
apwyeah there may even be a python lib fn for finding  'here'09:27
apwhere=`dirname $0`09:28
apwcase "$here" in09:28
apw*)  here="`pwd`/$here" ;;09:28
apwesac09:28
apwis what i do in shell09:28
smbhere = os.path.abspath(os.getcwd()) 09:28
smbI think09:28
apwno that is 'here' of the user, not the here of the program09:29
apwok a line is missing from that fragment ... cause it starts with a /09:29
apwdamn irc client09:29
smbright, replace that the path of the program09:29
apw        here=`dirname $0`09:29
apw        case "$here" in09:29
apw        /*) ;;09:29
apw        *)  here="`pwd`/$here" ;;09:29
apw        esac09:29
apwthe problem is that $0 is not always fully qualified09:30
apwit is generally what the user typed, which can be relative09:30
apwso you have to be sneaky09:30
smbHm, so does maint-getabis currently fail for you that way. It seems at least for the libpath I already use abspath09:31
ericm_smb, hrm... maint-getabis doesn't work with version number like: EE: Ubuntu-2.6.32-206.19foo1 does not seem to be a valid name09:37
ericm_it's in maint-getabis09:37
smbericm_, Did you tag it like that?09:37
ericm_smb, oh no - let me tag it again09:37
ericm_smb, now it works09:38
ericm_smb, hrmm... downloading abi failed09:39
ericm_smb, it's an OEM kernel, so the kernel image package isn't on the server09:39
smbWell the script is geared towards official releases. So it looks on official host archives only09:39
ericm_smb, yeah - that makes sense - but is it possible to provide kernel package manually?09:40
ericm_smb, I'm thinking of maintaining the OEM kernel tree as our official ones with same branch management09:40
smbericm_, Not yet. You could add that to a wish list.09:40
smbericm_, Though I probably would make it an additional  'location' to look for packages09:41
ericm_smb, yeah that's a good idea09:41
=== ogra_ is now known as ogra
cooloneyapw: just saw your patch about readahead, will this one be summitted to upstream?09:55
apwcooloney, likely yes, whether they will take it is another question as it uses a page bit, and they are very scarese09:57
smbapw, So maint-getabis seems to work for me without having the path set. So it seems lib lookup is ok. when calling other scripts it seems to require a little more. I changed maint-startnewrelease accordingly. Maybe there is a simpler way by adding it to the right path within python but this seems to work for me.09:58
apwsmb ta, will try it out09:59
cooloneyapw: understood09:59
cooloneyapw: ureadahead needs this patch, i guess.10:00
apwureadahead2 needs it yes10:00
ikepanhcwhen sprint, Tim said when UDS, we make agreements, its means "Each team makes agreements for next release" or "Kernel team makes agreements to other team"?10:05
apweach team agrees with the other teams what it will deliver for the release10:06
apwso both i think10:06
ikepanhcapw: thanks10:06
* ikepanhc hates writing slide :(10:06
jk-ike: if you like, I'm happy to read & review10:23
ikepanhcjk-: sure, but please wait me finish :)10:27
jk-ikepanhc: of course :)10:27
ericm_smb, in maint-getabis:DownloadPackage(), it looks like if I download the .deb package manually, it won't bother to download10:49
=== ericm_ is now known as ericm|ubuntu
ericm|ubuntusmb, yet the problem is where, I downloaded the package in the current ubuntu-lucid/ directory, but doesn't seem to be recognized as already existing10:50
smbericm|ubuntu, Might work, though you would have to do it either on the remote host or say --local10:51
=== amitk is now known as amitk-afk
ericm|ubuntusmb, but maint-startnewrelease doesn't accept --local10:52
smbHm, I should add that as passthrough then10:53
diwicwhat (Maverick) kernel repo/ppa should I enable if I want to help out with testing?10:56
apw~kernel-ppa/pre-proposed11:00
diwicwhoa! That was an interesting resume effect11:02
apwheh what did you see ?11:03
diwicMy (other) laptop makes screen effects that reminds me of the demo scene in the 90's11:03
* apw is intregeded, but has no idea what that might look like11:04
diwicIt purely hang under Lucid, so this is maverick. Every other scanline is black, and  the others are fading between white and blue-black11:05
diwichmm...seem to have settled as being grey + black now11:06
diwica few blue pixels in the top and bottom though11:06
=== ericm|ubuntu is now known as ericm-afk
marco74hello anybody here in this room11:17
marco74I have a problem concerning the ubuntu-kernel of Lucid11:17
marco74can anybody help me?11:17
apwwtf why to people just appear ask and go without waiting?11:34
apwdiwic, sounds mad enugh11:35
apwoh _and_ logoff irc completely ... 11:36
apwcking, hey11:36
=== apw changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Maverick Kernel Version: 2.6.35 || Ubuntu Kernel Team Meeting - August-03 - 17:00 UTC || If you have a question just ask, and do wait around for an answer!
ckingapw, what?11:44
apwcking, was trying to work out if you could hear me on mbl11:45
apwas you clearly wern't there ... 11:45
=== amitk-afk is now known as amitk
diwicapw, you have a movie in bug #61125412:41
ubot2Launchpad bug 611254 in linux (Ubuntu) "Resume hangs and shows funny patterns on screen (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/61125412:41
apwdiwic, what do you use to play it?  totem just ignores it12:48
diwicapw, it works in VLC. It's straight from the phone12:49
apwdiwic, so it does, cute effect12:50
diwicapw, I'll just wait for our suspend/resume guy (still haven't learned all names and who does what) to take care of it :-)12:55
apwcking, http://launchpadlibrarian.net/52687216/MOV00503.MP412:55
apwdiwic, i wonder if thats just the video ram decaying cause its not being refreshed cause we are crashed13:10
Sarvattdiwic: does booting with vesafb=fubar fix it? :)13:10
apwliterally =fubar ?13:11
Sarvattinvalid parameter so it wont load13:11
apwSarvatt, do we have any idea why vesafb is relevant ?13:12
diwicIf that matters, it's Via/OpenChrome13:13
Sarvatti'm just curious because i've seen similar caused by vga16fb. I saw it was openchrome, I'm sorry diwic :)13:14
apwframebuffer handling is mostly broken .. i am working on what is becoming a major patch for it now13:15
diwicSarvatt, I tried it and as we already suspected, it didn't work, i e no difference13:17
diwicapw, video ram decaying was the best theory out of 1 theories so far13:18
* apw likes those odds13:33
* apw bets his theory is the closest to the truth13:33
apwsmb, these new ext4 patches are they the ones we already have or new ones?13:45
apw(the ones in the stable queue for .32)13:45
smbapw, May or may not. I have not done a comparison yet to find out whether the sets are 1:113:45
apwthere seems to be 52 of them13:46
smbthat sounds not too bad13:46
apw142 in total on the queue, 52 ext413:46
smbLet me run over them with a script to be sure its all we had13:46
diwicshould I put something into a git kernel commit to make it automatically close a LP bug?13:50
apwdiwic, yep, a BugLink: link13:51
* smb is too slow13:51
apwBugLink: http://bugs.launchpad.net/bugs/58152513:51
ubot2Ubuntu bug 581525 in linux-meta (Ubuntu) (and 1 other project) "Lucid: system becomes unstable randomly, seems related with apparmor (affects: 5) (heat: 32)" [Undecided,In progress]13:51
manjoapw, how does that automatically close an LP bug ?13:52
diwicapw, thanks13:52
apwinsert changes copied them over to LP: #nnnnn in the debian changelog13:52
apwand things so marked in there are closed by the janitor when the package is published13:52
apwmanjo, ^^13:53
manjodid now know that ... thanks 13:53
apwt'is why they are important to have in the bugs13:53
apwthe - LP: bits of the debian changelog are made from them13:53
manjoapw, I always added them not knowing their side effects 13:54
apw  * SAUCE: Yama: search for PTRACE exceptions via thread group leader13:54
apw    - LP: #60371613:54
apwlike that ... and that is grepped out by the janitor13:54
sorenmanjo: Technically, it's dpkg-genchanges that does it. It spots the stuff in debian/changelog and adds a line to the _source.changes file (X-Launchpad-Bugs-Fixed, IIRC), which in turn is handled by Launchpad.13:56
apwsoren, heh even more complex13:56
sorenWell, the magic has to happen somewhere :)13:57
* manjo is feeling dizzy 13:57
sorenThis way, if the distro team feels like changing preferred format for it in the debian changelog, it doesn't require changes in Launchpad.13:57
apw/home/apw/upload/arm/linux-fsl-imx51_2.6.31-104.12_source.changes:Launchpad-Bugs-Fixed: 310760 410933 410933 416325 416325 427948 430694 430694 430694 430809 430809 433904 435958 438687 445580 44659513:58
apwand there it is lurking in the changes file13:58
sorenOk, so no "X-" prefix. My bad :)13:59
manjoso if a patch fixes multiple bugs we need to have multiple Buglink: ... in the commit comments ? 13:59
apwyeah such inaccuracy you should be fired13:59
apwmanjo, yes13:59
sorenapw: You can't fire people who've already left. :)13:59
apwand they end up as - LP: #xxx, #yyy13:59
apwand so on13:59
apwsoren, i didn't say who should fire you did i ?   :)14:00
sorenapw: point :)14:00
manjocease fire14:00
=== ericm-afk is now known as ericm
ericmsmb, when you rebase mvl-dove branch against master, what is the command you use, 'git rebase Ubuntu-2.6.xx-xx.xx' or 'git rebase --onto Ubuntu-2.6.xx-xx.xx <some start point> mvl-dove'14:02
smbericm, The second form14:02
ericmsmb, as I expected, ok14:03
smbericm, Though I started to run it as a script: maint-rebase-branch14:03
=== yofel_ is now known as yofel
smbericm, This would end on a branch named auto-tmp-rebase which one can check and then reset the target branch to that14:05
ericmsmb, ah - script again, thanks - I love your scripts :-)14:05
smbericm, Thanks :) These are quite ego-centric for the moment, so feel free to let me know anything that can be improved.14:07
=== gnomefreak76 is now known as gnomefreak
=== bjf[afk] is now known as bjf
=== BenC__ is now known as BenC
* manjo going down for reboot15:32
=== sconklin-gone is now known as sconklin
manjoI will be offline for a few mts my electric company is cutting my power to replace the meter 17:05
apwyeah right :)17:12
apwyou will be gone for HOURS :)17:12
ckingdays even17:33
* cking hopes manjo has a lot of batteries17:34
jjohansenogasawara: I'm going to have another AA sync request for tonight / friday morning.  Its not that different from what we have but brings it up to match with the latest updates.18:07
ogasawarajjohansen: cool, I'll wait for it then18:07
jjohansenogasawara: well no real need to wait, I can push it when you want.  Just the longer I wait the more likely it will match whats hopefully going upstream18:08
jjohansenits done now, just waiting on feedback18:09
ogasawarajjohansen: maybe just wait and see if you get any feedback by tomorrow.  I'd like to be as close to what's upstream as possible.18:09
ogasawarajjohansen: I'm gonna hold off the upload as long as I can anyways as I'm hoping 2.6.35 final might land18:09
jjohansenogasawara: also I have a pv-ops request coming too, no nasty xen branch :)18:10
ogasawarajjohansen: sweet18:10
apwjjohansen, that sounds good18:15
apwogasawara, i shot you a summary email on all the patches i've puked out today18:15
* apw is wandering off for some beer shortly, i'll try and check my email tonight in case you have any issues18:16
ogasawaraapw: yep, saw it.  Just getting through the last set.  looking okay so far.  I'm gonna fire up some tests across some of my systems before I push.18:17
apwsounds good.  doesn't solve the fact that plymout18:17
apwplymouth fires up too soon on a lot of drm h/w.  but at least it not longer blows the kernel up (for me at least)18:17
* manjo getting lunch will be back soon18:33
* vanhoof just loves updating wikis :)18:43
ckingvanhoof, watch out - you may become a wiki gardener soon19:00
jjohansencking: beat me too it19:00
vanhooflol19:06
* vanhoof *hides*19:06
=== sconklin is now known as sconklin-lunch
=== sconklin-lunch is now known as sconklin
jjohansen-> lunch20:09
bjfjjohansen, i added you as UDS-N roomie :-)20:19
jjohansenbjf: wow you are on top of things :)20:42
* ogasawara lunch20:43
* manjo going to post office ... brb 30mts 21:34
=== achiang is now known as smallakin1
=== smallakin1 is now known as achiang
bjfogasawara, hope you don't mind if i poke at qa about testing proposed :-)22:19
ogasawarabjf: don't mind at all, was actually hoping others would chime in :)22:19
ogasawarabjf: I was thinking the exact same questions you posted.22:19
jjohansenbjf: have fun tilting at windmills22:20
ogasawaraheh22:20
bjfjjohansen, i'm willing to have a go at this one for a bit, it's been years since I sent some really flaming email22:21
jjohansen:)22:21
ogasawarabjf: after the chat with them at the rally I thought we'd figured out the preseed bit, and modification of the dist.info file to search for results of testing from -proposed, and tweaking their prober or whatever to detect new packages in -proposed22:22
ogasawarabjf: I swear ronald even sent an email about it22:23
bjfogasawara, yes i had the same thoughts as well, I guess we were mistaken22:23
komputesogasawara: hi there! :)22:49
* komputes hands ogasawara a present - https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/61147422:49
ubot2Ubuntu bug 611474 in linux (Ubuntu) "boot failure SAS hard drives controller (affects: 1) (heat: 6)" [Undecided,New]22:49
apwogasawara, yo ... checking in ... all good?  any problems?22:50
ogasawaraapw: all good22:52
ogasawaraapw: just have one more system I want to install and test on22:52
ogasawaracool, the touchscreen on the hp-mini is working again22:56
apwogasawara, awsome :)   ... night23:30
daishadarcan someone give me a general idea of what will probably stop working if i install a mainline kernel?  what are ubuntu specific patches and drivers for?23:58

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