| === jmp_ is now known as _jmp_ | ||
| snadge | jerusalem | 01:05 |
|---|---|---|
| === jmp_ is now known as _jmp_ | ||
| === smb` is now known as smb | ||
| * smb mornings | 07:21 | |
| * apw yawns | 07:52 | |
| smb | apw, sync;sync;sync | 07:52 |
| apw | reboot | 07:53 |
| apw | oops | 07:53 |
| smb | heh | 07:53 |
| ppisati | bloody jetlag... | 08:13 |
| * dileks thinks if there where a hwclock.service for internal human clock | 08:17 | |
| ppisati | bug 985354 | 08:52 |
| ubot2 | Launchpad bug 985354 in ubiquity "Lubuntu 12.04 20120418 install fail" [Undecided,Confirmed] https://launchpad.net/bugs/985354 | 08:52 |
| apw | cking_, did we decide to enable ALPM for Q ? | 08:56 |
| cking_ | apw, it was not discussed, but it probably should be to see what fall out we hit | 08:57 |
| === cking_ is now known as cking | ||
| apw | Bryan Wu (1): | 09:05 |
| apw | MAINTAINERS: add maintainer for LED subsystem | 09:05 |
| apw | cooloney, congrats | 09:05 |
| * cking notes that there is only so much coffee the brain can handle to try to overcome jetlag | 10:50 | |
| * ppisati -> goes looking for food | 11:57 | |
| tgardner | apw, hows it coming on collapsing generic and virtual ? | 12:00 |
| apw | just looking at those patches now | 12:01 |
| tgardner | apw, I promised the preempt guy that we'd add his flavour only if we could collapse virt and generic | 12:02 |
| apw | tgardner, yeah ... we should add an item for that, then i can follow through when this is done | 12:02 |
| tgardner | apw, to the blueprint ? | 12:03 |
| apw | yeah | 12:03 |
| tgardner | apw, I started the lts-backport-quantal branch in precise yesterday. still a bit of messing around to do. | 12:04 |
| apw | tgardner, ok cool. do we have a PPA for that yet? | 12:04 |
| tgardner | apw, not yet, thought I'd get the kernel building first. | 12:04 |
| apw | i more meant have they decided where the combo ppa will land | 12:05 |
| tgardner | apw, I think we can just use an existing PPA, can't we ? | 12:05 |
| apw | we probabally want another team which holds it | 12:05 |
| apw | as its gonna be for us and X | 12:05 |
| tgardner | oh, right, forgot about X | 12:05 |
| apw | we need to poke bryceh about what might work | 12:06 |
| tgardner | apw, I'll hassle byrce when I've got it building | 12:06 |
| Gup | hi, issue with networking/suspend since kernel update. details: http://pastebin.com/VdMVRWQb anyone recognise the symptoms, or have an idea what to try next? | 12:11 |
| apw | Gup, sounds like you have a known working and known not working kernel version, so get a bug filed and let us know the number, we would want to then to a bisect between the two to work out whats bust things | 12:13 |
| apw | Gup, there is a page on doing bisects if you want to attempt it yourself | 12:13 |
| apw | jsalisbury, can you point us to the bisect guide ? | 12:13 |
| Gup | i'll certainly file a bug. i'll have a look at the bisect guide and see what it involves | 12:14 |
| Gup | i tried to find a changelog to get an idea but it looked like it contained lots of changes that might have been the cause | 12:15 |
| apw | Gup, probabally a stable update or similar | 12:15 |
| Gup | i'm not overly familiar with the kernel and its workings, i did compile one a couple of times, but thats been a few years | 12:17 |
| apw | smb, ok you'll be pleased to know all of the patches in that intel-idle branch you pointed me to are now in mainline | 12:24 |
| apw | smb, is this something we can test works ? | 12:24 |
| apw | smb, with the quantal kernel, before i commit removing this thing :) | 12:29 |
| Gup | the link to the bisect guide on here is broken :s https://wiki.ubuntu.com/KernelTeam/KernelTeamBugPolicies | 12:37 |
| apw | jsalisbury, you have been doing a number of these, would you be able to fix the docs :) | 12:42 |
| apw | https://wiki.ubuntu.com/Kernel/KernelBisection | 12:43 |
| apw | Gup, have you seend the above one ?? | 12:43 |
| Gup | ahha thanks apw, i hadn't found that yet, i will have a go | 12:44 |
| tgardner | apw, I see that the hv folks are getting traction with the Debian kernel community | 12:47 |
| * tgardner bounces for quantal kernel backport to precise | 12:58 | |
| ogasawara | hrmph, https://launchpadlibrarian.net/105081769/buildlog_ubuntu-quantal-amd64.linux_3.4.0-2.4_BUILDING.txt.gz | 13:01 |
| ogasawara | bah, https://launchpadlibrarian.net/105086088/buildlog_ubuntu-quantal-i386.linux_3.4.0-2.4_FAILEDTOBUILD.txt.gz | 13:02 |
| tgardner | ogasawara, puny little buildds | 13:02 |
| tgardner | ogasawara, I have a patch in master-next to fix the perf build problem | 13:04 |
| * ogasawara fetches | 13:05 | |
| tgardner | ogasawara, I thought the erro you were pointing at was because the buildd ran out of disk, but I see that its really a packaging error | 13:05 |
| ogasawara | tgardner: I was confused as why it would fail on i386 and not amd64, but seeing your patch explains it | 13:06 |
| tgardner | ogasawara, it seems its a change from precise | 13:06 |
| tgardner | ogasawara, I had at least one build trace that failed within the perf makefile, so I can only assume its a concurrency level issue. | 13:07 |
| smb | apw, About those patches, I can confirm on intel and on my amd that it enables the cpufreq in xen. There were no bad effects. | 13:26 |
| apw | smb, ok cool. so as all of them are in quantal things should be good. i'll clean up and get this merger in | 13:27 |
| smb | To see whether it gains something, one needs to check performance | 13:27 |
| apw | yeah not expecting it to be better, just not worse | 13:27 |
| smb | In theory it should be better at least on intel as it should allow usage of turbo mode. | 13:28 |
| jsalisbury | apw, I'll review and fix up the kernel bisection docs. | 13:32 |
| apw | jsalisbury, thank | 13:32 |
| jsalisbury | tgardner, ogasawara, Do you still want to have the top ten meeting today, since all the top ten bugs are currently precise bugs? | 13:34 |
| ogasawara | jsalisbury: anything new for precise on the list? | 13:34 |
| jsalisbury | ogasawara, nothing yet, but I'm still catching up from pre-uds. | 13:34 |
| ogasawara | jsalisbury: if there's nothing new atm, we can probably skip it. just ping us if anything does pop up that need discussing | 13:35 |
| jsalisbury | ogasawara, sounds good. Thanks. | 13:35 |
| tgardner | jsalisbury, yeah, what she said. | 13:41 |
| jsalisbury | tgardner, ack | 13:42 |
| cooloney | apw: sorry for missing your msg | 13:47 |
| * cooloney is still in jetlag, | 13:48 | |
| apw | cooloney, heh, when i heard you had fun with a bird in SFO i assumed you would be off | 13:48 |
| cooloney | apw: yeah, it's really fun. hah, we heard a very loud crashing noise in the air. | 13:50 |
| apw | cooloney, man that must have been a fright and no mistake | 13:50 |
| ogasawara | cooloney: heh, I wouldn't call hearing a crashing noise while in flight "fun" | 13:50 |
| cooloney | apw: after a while, pilot said our engine was hit by a bird. OMG, I just sat close to the engine | 13:50 |
| tgardner | cooloney, shit, thats just the kind of noise you _never_ want to hear on a plane. | 13:51 |
| * apw bets that the engine does not enjoy eating birds | 13:51 | |
| cooloney | so we flew back to SFO. | 13:52 |
| cooloney | i saw some dirty liquid in the engine when I was getting off the aircraft | 13:53 |
| cooloney | so we had to stayed in SFO for one more night | 13:53 |
| apw | cooloney, sounds awful. especially staying another night | 13:56 |
| cooloney | apw: actually i was upgraded to economic comfort classed in first flight | 13:57 |
| cooloney | apw: after a bird hitting, i had no chance to upgrade | 13:58 |
| apw | double buggeration | 13:58 |
| smb | ... another ass hurting... | 13:59 |
| cooloney | smb: heh, any wiki or lp link for lxc stuff? i intend to try the testcase on my pandaboard. | 14:01 |
| smb | cooloney, test cases, we do have test cases? ;-P | 14:02 |
| smb | cooloney, But joke aside. I am not sure... | 14:04 |
| apw | cooloney, hallyn is probabally the right person to talk to about how we are using it and so what to test | 14:04 |
| smb | He and stgraber | 14:05 |
| smb | For a quick test, I just installed the package, made sure the cgroup fs is mounted (which should be automatic now) and tried running some of the quick examples from the man page | 14:06 |
| hallyn | cooloney: I'm in the middle of an update that i haven't tested yet, but lp:~serge-hallyn/+junk/lxc-test has my testcases | 14:07 |
| hallyn | (high-level ones mostly, besides the reboot ones) | 14:07 |
| smb | cooloney, There is also a chapter in the server guide https://help.ubuntu.com/12.04/serverguide/lxc.html | 14:08 |
| cooloney | smb: yeah, i heard hallyn mentioned that during that lxc session. | 14:09 |
| cooloney | hallyn: thanks for pointing out this. i will try that | 14:09 |
| cooloney | and i believe arm server folks already run them on arm server | 14:10 |
| hallyn | cooloney: yup, that's what ahs3 said. But thanks, the more people running it the better | 14:12 |
| hallyn | i guess this means i better go test my changes :) | 14:12 |
| apw | hallyn, cooloney soon will :) | 14:13 |
| hallyn | heh, good point, on to my next work item :) | 14:13 |
| cooloney | hallyn: sure, | 14:14 |
| * cooloney setting up pandaboard | 14:14 | |
| hallyn | j/k. i couldn't test at end of uds bc of bandwidth. need to try it out | 14:14 |
| === yofel_ is now known as yofel | ||
| hallyn | jsalisbury: did we pick kernel-kvm as the tag for fwded bugs? or kvm-kernel? :) | 14:36 |
| gema | tyhicks: ping | 14:38 |
| gema | tyhicks: can you please fix the work items for the ecryptfs blueprint? | 14:38 |
| gema | tyhicks: it seems not to want to take my changes and it is messing up the work items ... sorry | 14:38 |
| apw | herton, does shank-bot make bugs for linux-lowlatency ? | 14:38 |
| herton | apw, no, just for our kernels at the moment | 14:39 |
| apw | herton, are we doing linux-amardaxp ? | 14:39 |
| bjf | apw, yes | 14:39 |
| herton | apw, we should be on next cycle, bjf added the code there | 14:39 |
| * apw notes lowlatency is lagging severely already ... | 14:40 | |
| apw | herton, bjf, cool stuff | 14:40 |
| bjf | apw, that's a "community" kernel is it not? | 14:40 |
| apw | bjf, yes currently so, wondered if it would make sense to make bugs for them | 14:40 |
| bjf | apw, we don't for ports kernels either | 14:41 |
| apw | bjf, there are no ports kernels which arn't in master are there? | 14:41 |
| bjf | apw, ah, true | 14:41 |
| apw | so we may be being lucky there more than anything | 14:42 |
| tyhicks | gema: Hey - what changes would you like me to make? | 14:42 |
| * apw will chat to abogani about it, and maybe ask you again | 14:42 | |
| bjf | apw, ack | 14:42 |
| herton | bjf, btw, armadaxp lacks a entry for the meta-package on ubuntu.py. I was wondering also if we should make Ike the owner of prepare-package on workflow.py | 14:42 |
| apw | abogani, your lowlatency kernel is lagging, who am i meant to poke about that, you ? | 14:42 |
| apw | herton, for now ike is on the hook for everything armada related | 14:43 |
| bjf | apw, wasn't tgardner considering us picking up lowlatency (uds session) | 14:43 |
| tgardner | bjf, preempt in Quantal | 14:43 |
| apw | bjf, it may become portlike in Q yes | 14:43 |
| bjf | herton, yes, we should hadd the meta package. i thought there was a LP "team" that got assigned, that should be used | 14:44 |
| bjf | tgardner: you are correct as always | 14:44 |
| smb | bjf, herton Just happened to notice that the vcs-git line of precise linux-meta refers to Karmic... Wonder whether fixing this should require a bug report and all the fun... | 14:44 |
| bjf | smb, lol | 14:44 |
| apw | smb, you might file 'one bug' and use it for all the releases, cause i bet its wrong in all of them | 14:44 |
| herton | bjf, I think the team thing never flied, I'm not sure Ike is on ubuntu-armel-kernel and watches it | 14:44 |
| * apw says stike ike on, he'll soon ask you to change it to a team once he starts drowning | 14:45 | |
| smb | apw, Likely at least between karmic and now... :) | 14:45 |
| bjf | herton, lets coordinate with ike. i want it to be a team and not a person | 14:45 |
| herton | bjf, ok | 14:45 |
| vanhoof | bjf: +1 | 14:46 |
| bjf | herton, even if it's a team of one right now | 14:46 |
| * apw suggests we make a team with just vanhoof on it | 14:46 | |
| vanhoof | bjf: we have a few teams now, would you like something trimmed down in size? | 14:46 |
| * vanhoof *hides* | 14:46 | |
| herton | yes, it's better being a team | 14:46 |
| * ogasawara back in 20 | 14:47 | |
| smb | bjf, herton opened bug 999726 for tracking | 14:51 |
| ubot2 | Launchpad bug 999726 in linux-meta "Fix Vcs-Git in linux-meta" [Undecided,New] https://launchpad.net/bugs/999726 | 14:51 |
| bjf | smb, ack and thanks | 14:51 |
| bjf | henrix: ^ | 14:51 |
| henrix | bjf: ack | 14:52 |
| smb | bjf, Was just about to ask whether I should do something about it ... :) But this wfm2 | 14:52 |
| bjf | smb, nah, we'll take it | 14:53 |
| kirkland | tyhicks: I reckon I probably didn't take notes in the modern work-item manner | 14:53 |
| kirkland | tyhicks: apologies | 14:53 |
| kirkland | tyhicks: behind on the times | 14:53 |
| * bjf will brb | 14:53 | |
| jsalisbury | hallyn, yes, kernel-kvm is the tag | 14:53 |
| * tgardner biab | 14:54 | |
| hallyn | thanks, written down :) | 14:54 |
| jsalisbury | ogasawara, whould you like me to change "[TOPIC] Status: Precise Development Kernel (ogasawara)" to Quantal? | 15:02 |
| apw | jsalisbury, makes sense | 15:02 |
| jsalisbury | apw, sounds like a plan then | 15:03 |
| apw | tell me we don't already have a meeting today ... yawn | 15:03 |
| jsalisbury | apw, yup. Top Ten is canceled, but the IRC meeting is still on, for 17:00 UTC | 15:04 |
| apw | good grief do we get no time off | 15:04 |
| jsalisbury | heh | 15:05 |
| === BenC__ is now known as BenC | ||
| * bjf is back | 15:15 | |
| tyhicks | gema: I *think* that I fixed the work items according to the comment you left. Let me know if it still isn't right. | 15:16 |
| ogasawara | tgardner: was reviewing https://wiki.ubuntu.com/Kernel/Release/Rolling, the 14.04 section, should that be linux-image-<flavor> will reference the current supported kernel, not linux-image-<version>-<flavor> ? | 15:18 |
| tgardner | ogasawara, prolly | 15:18 |
| tgardner | ogasawara, feel free to correct that | 15:19 |
| ogasawara | tgardner: ack | 15:19 |
| gema | tyhicks: looks good to me, as long as you are sure you haven't lost any other task, I am happy | 15:43 |
| tyhicks | gema: I'll take a second look at the notification emails to ensure that nothing was lost. Thanks! | 15:43 |
| gema | tyhicks: ack | 15:49 |
| tyhicks | (everything looks good) | 15:49 |
| * cking punches fwts json logging into submission | 15:49 | |
| === BenC__ is now known as BenC | ||
| smb | O_o lucid pointing at karmic | 16:20 |
| henrix | smb: yeah, a couple of branches had that problem too :) | 16:22 |
| smb | Somehow I would have understood them pointing at something earlier but forward is confusion | 16:25 |
| * henrix is confused... isn't karmic earlier than lucid? | 16:27 | |
| smb | err yes | 16:28 |
| * smb needs to get his letters right... | 16:29 | |
| henrix | smb: btw, i tried to add lucid to the bug you filled... but couldn't figure out how to do that :-/ | 16:38 |
| smb | henrix, No you cannot by bad design | 16:40 |
| henrix | smb: ah, cool! | 16:40 |
| smb | henrix, its approved now | 16:40 |
| henrix | smb: thanks | 16:41 |
| smb | hm, and to make this small thing really big, I probably should add the topic branch packages, too.. | 16:42 |
| smb | henrix, Oh, but fsl-imx51 and mvl-dove not relevant anymore... | 16:44 |
| henrix | smb: yeah, i was not sure if all of them were relevant... so you can just NACK those two | 16:44 |
| smb | henrix, yup, oh seems natty/ti-omap4 is pointing at maverick | 16:46 |
| apw | ogasawara, i just did a test build and it failed with the same (i think) perf thing even with the top of master-next (inc tims fix) | 16:50 |
| apw | install: cannot stat `/home/apw/build/ubuntu-quantal/ubuntu-quantal/debian/build/tools/tools/perf/perf': No such file or directory | 16:51 |
| ogasawara | apw: bah, my test build passed. | 16:51 |
| * ogasawara holds off on the upload | 16:51 | |
| ogasawara | apw: that was indeed the same failure. were you using one of out build machines or something local? | 16:52 |
| apw | ogasawara, that was a gomeisa build, doing a 'full_build=true' build to test dropping -virtual | 16:52 |
| smb | henrix, Ok, I hope I have added all the tasks you would need for that bug | 16:52 |
| ogasawara | apw: I'll try to reproduce | 16:53 |
| tgardner | apw, so wtf is going on with perf ? | 16:53 |
| henrix | smb: ok, cool. i missed the natty/ti-omap4, as i just grep'ed for karmic. i should have manually checked everything | 16:53 |
| apw | well the base error says perf wasn't built, or it got removed before we made the package i guess | 16:53 |
| henrix | smb: will check them all again | 16:53 |
| tgardner | apw, thats different then the intermediate failure | 16:53 |
| tgardner | apw, how could it get removed? | 16:54 |
| apw | tgardner, yeah and i think its what leann saw | 16:54 |
| smb | henrix, topic-branch madness. Looks like all ti-omaps like maverick | 16:54 |
| apw | tgardner, reading the logs now, hopefully i can see where it broke | 16:54 |
| henrix | smb: ok, i'll edit the bug description so that it's a little bit more generic | 16:55 |
| smb | henrix, Cool, thanks | 16:55 |
| henrix | smb: i'll replace the "karmic" in the desc by something else | 16:56 |
| apw | tgardner, ogasawara, ok it looks like when we build the indep version the manual build is now wacking it | 16:56 |
| apw | rm -rf /home/apw/build/ubuntu-quantal/ubuntu-quantal/debian/build/tools | 16:57 |
| apw | ... | 16:57 |
| smb | Yeah, can now be Vcs-Git is wrong just about anywhere in linux-meta... :-P | 16:57 |
| apw | make[1]: Leaving directory `/home/apw/build/ubuntu-quantal/ubuntu-quanta | 16:57 |
| apw | l/debian/build/tools/tools/perf' | 16:57 |
| tgardner | apw, hmm, wonder why ? | 16:57 |
| herton | smb, you know, copy & paste everywhere :) | 16:57 |
| apw | rm /home/apw/build/ubuntu-quantal/ubuntu-quantal/debian/build/tools/tools | 16:57 |
| apw | install -s -m755 /home/apw/build/ubuntu-quantal/ubuntu-quantal/debian/build/tools/tools/perf/perf \ | 16:57 |
| apw | so i think we build and prep the mans and the tools binaries in the same directory | 16:58 |
| apw | so depending which order we do them depends on which we have to package | 16:58 |
| tgardner | apw, that seemsl ike a bad thing | 16:58 |
| jsalisbury | ** | 16:58 |
| jsalisbury | ** Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting | 16:58 |
| jsalisbury | ** | 16:58 |
| smb | herton, sure. :) | 16:58 |
| apw | tgardner, indeed it does | 16:58 |
| * apw pokes ... | 16:58 | |
| apw | tgardner, this has almost cirtainly come out of our paralellisation work, and likely in P we were just lucky | 17:00 |
| tgardner | apw, it has been bugging me for awhile | 17:00 |
| tgardner | it just hasn't happened enough to got to the top of my list | 17:01 |
| apw | tgardner, its just wrong, we do two things which can be done in parallel on i386 on the same tmp dir... just working out which is easier to move | 17:01 |
| ogasawara | I'm surprised we were lucky all of Precise to not get hit by this | 17:01 |
| tgardner | apw, seems simple in retrospect | 17:01 |
| tgardner | ogasawara, -j1 on the buildds | 17:01 |
| apw | ogasawara, i think we do more parallelism now that we did | 17:01 |
| tgardner | apw, we could just serialize tools packaging. its not like it takes that long | 17:02 |
| apw | tgardner, its not tools which need serializing, they need serialising wrt 'indep' in general, which isn't reasonable. they should be in differnet places | 17:04 |
| tgardner | apw, which is gonna be a bit of churn. I think Leann is OK to upload given the buildds are at most -j2, don't you think ? | 17:05 |
| apw | tgardner, it looks trivial to move the perarch indep, so i'd say give me an hour to work it and decide after | 17:06 |
| tgardner | apw, ack | 17:06 |
| apw | tgardner, the bid buildds are -j16 arn't they ? | 17:06 |
| tgardner | apw, whats a bid buildd? | 17:06 |
| apw | tgardner, big ... there are two which are monsters relativly speaking | 17:06 |
| tgardner | apw, ah, didn't realize we had anything other then the -j1 xen instances | 17:07 |
| apw | tgardner, oh is this going to a PPA, then yes | 17:07 |
| apw | tgardner, in that case shove it in | 17:07 |
| tgardner | apw, no, she's gonna upload it to the quantal archive | 17:07 |
| ogasawara | apw: I'm not uploading to a PPA | 17:07 |
| apw | so it all depends on the buildd then that it hits | 17:08 |
| tgardner | apw, we could clamp it to -j1 just for the buildds, but thats kind of a hack. | 17:08 |
| ogasawara | apw: I think we want the patch rather than rolling the dice on buildd's | 17:09 |
| ogasawara | apw: or we could have infinity or someone retry the current i386 build on a more susceptible buildd | 17:10 |
| apw | yeah you could just put it back and see what happens | 17:11 |
| apw | i think i have a fix so will go test build it to confir | 17:11 |
| apw | tgardner, ogasawara, we may be simply hitting it more on quantal cause of a make change for instance | 17:15 |
| === jsalisbury changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Ubuntu Kernel Team Meeting - Tues May 22nd, 2012 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! | ||
| tgardner | apw, its certainly been annoying me more of late | 17:15 |
| apw | tgardner, we have be going more and more parallel and releasing the bits which stop it going fast, so we should be hitting it more really | 17:18 |
| tgardner | apw, indeed, that is my empiracal experience | 17:18 |
| apw | tgardner, and possibly depending how you build it you could make it happen, i think the act of building it with dpkg-build package, the make build-foo; fakeroot make foo; ought to near guarentee it happens i think | 17:19 |
| apw | well i suspect we either lose the binaries or the manuals, but the manuals we don't install by name we just take * so we'lll only notice one way round perhaps | 17:19 |
| tgardner | bjf, just pushed some cruft onto ubuntu-precise-meta master. please review prior to uploading. | 17:20 |
| bjf | tgardner, ack | 17:23 |
| bjf | henrix, herton ^ | 17:23 |
| herton | ack | 17:23 |
| henrix | bjf: ack | 17:24 |
| LightScry | Hi, I'm trying to implement "syscall" and having trouble. My code looks like this: http://cl.ly/3Q2n12113a2X3737191R | 17:24 |
| LightScry | It always returns -1. | 17:25 |
| apw | tgardner, that seems to only refer to generic and server, but we have generic-pae on i386 and only that right ? | 17:25 |
| * henrix will be back in ~30mins | 17:25 | |
| tgardner | apw, since its a meta package its abstracted by one level. it should refer to -pae in i386 | 17:26 |
| bjf | cking: https://chrome.google.com/webstore/detail/bcjindcccaagfpapjjmafapmmgkkhgoa | 17:26 |
| apw | tgardner, plus there are only linux-image ... shouldn't there be linux-headers-* as well | 17:26 |
| tgardner | apw, not sure if we wanna propogate that level of complexity, do we ? do we need the headers ? | 17:27 |
| cking | bjf, I'm using: http://jsoneditor.net/ | 17:27 |
| apw | tgardner, anyone needing dkms needs a meta package for their headers | 17:27 |
| apw | tgardner, so either make them linux-hwe-<flavour> and dep on both headers and image, or you need both me thinks | 17:28 |
| tgardner | apw, they can always reference the header meta-package directly. I don't think we should support automatic dkms upgrades from release to release. | 17:28 |
| apw | tgardner, but then they update their kernel automatically and any dkms they have just stops building even if its compatible | 17:29 |
| tgardner | I think dkms is guaranteed to break. look at nVidia. | 17:29 |
| apw | tgardner, well we should be testing those before we inject the new kernels if they are supported if thats going to be the only kernel in a while | 17:30 |
| apw | tgardner, especially for binary drivers for graphics | 17:30 |
| apw | as those are no longer opting in really are they, they are getting updated | 17:30 |
| tgardner | apw, well, I guess I'm OK with depending on the headers from the -hwe- and -current- meta packages. I just didn't want the combinatorial explosion. | 17:32 |
| apw | we might want to take out 'image' to make it clear if we do that | 17:32 |
| tgardner | apw, yeah, I'll update it in a bit. need fuud first | 17:33 |
| === tgardner is now known as tgardner-lunch | ||
| === tgardner-lunch is now known as tgardner | ||
| apw | tgardner-lunch, i can see why my builds never see this normally as i build the indep and dep parts separatly always | 17:54 |
| apw | ogasawara, have you been able to reproduce the perf thing locally, you would have to build 'binary' to see it i think ... as it requires a perarch and an indep run to be in parallel | 17:57 |
| apw | ogasawara, anyhow if you are, i have a patch which seems to resolve it ... shall i shove it | 17:57 |
| ogasawara | apw: go ahead and shove it. I have not been able to reproduce yet | 17:57 |
| tgardner | yeah, its kind of tweaky | 17:58 |
| apw | building binary-gneeric will never show it, nor will indep, but for me a binary always tiggers it at -j256 | 17:58 |
| ogasawara | apw: lemme try that. In the mean time I did kick off a re-build for i386. It landed on aatxe. | 17:59 |
| apw | ogasawara, ok pushed. | 17:59 |
| ogasawara | apw: I'll clean up the tree since I never did officially upload 3.4.0-2.5 | 18:00 |
| apw | ogasawara, i also have an overlayfs.v13 update pending, but don't want to throw that in if you are struggling to build | 18:00 |
| ogasawara | apw: ack | 18:01 |
| tgardner | apw, repushed ubuntu-precise-meta | 18:11 |
| apw | tgardner, so i wonder if the package is going to bring image and headers should it be just linux-hwe-generic | 18:13 |
| apw | as 1) its different and 2) if we ever do need them separate for some reason, like for the CDs we have space to put them | 18:13 |
| tgardner | apw, ah, forgot that. will fix. | 18:13 |
| tgardner | apw, fixed | 18:14 |
| tgardner | now I need to fix the descriptons as well | 18:15 |
| apw | tgardner, i wonder if the linux-generic references should be linux-image-generic | 18:16 |
| tgardner | apw, why? | 18:16 |
| apw | tgardner, we were talking aboue making linux-gneeric be both perhaps, or removing them, so it might be nice to miss the double indirect | 18:16 |
| apw | tgardner, either way right now they are just a double indirect, do we want that | 18:16 |
| tgardner | it seems simpler form a conceptual viewpoint, but its really 6 of one, half dozen of another | 18:17 |
| apw | ie we'd have linux-hwe-generic -> linux-generic -> linux-image-generic -> linux-image-xxxx-generic | 18:17 |
| apw | and linux-hew-generic seems same as linux-generic in the hierachy, so perhaps we should point to l-i-generic | 18:18 |
| apw | matters not much at this point as its not affecting any new names or anything | 18:18 |
| tgardner | apw, ok, repushed again | 18:28 |
| apw | tgardner, so the only thing i find confusing now is that we have -generic and not -generic-pae at the top level for i386 | 18:29 |
| tgardner | apw, you mean linux-hwe-generic ? | 18:29 |
| apw | yeah i'd expect that to be amd64 specific i think, and have a -pae i386 specific | 18:30 |
| tgardner | apw, I think -pae will disappear on our way to 14.04. this is supposed to be an abstract meta-package name | 18:31 |
| tgardner | apw, frankly, I;d like to change Quantal so that -pae disappears | 18:32 |
| apw | tgardner, i almost want to it be -desktop and -server or something, but thats probabally reaching bikeshedding | 18:32 |
| tgardner | apw, quantal should have _just_ a generic flavour (that also happens to be the PAE kernel) | 18:33 |
| tgardner | then we've homogonized across i386 and amd64 | 18:33 |
| apw | yeah makes sense, and we should do that rsn | 18:33 |
| tgardner | apw, I can make that happen :) | 18:33 |
| apw | tgardner, it will conflict heavily with my -virtual work, so perhaps i should do it on top of that | 18:34 |
| tgardner | as soon as ogasawara is done faffing about | 18:34 |
| tgardner | apw, when you do that, remember to go back and fix ubuntu-precise-meta lts-backport-quantal | 18:35 |
| apw | ACK | 18:35 |
| tgardner | and ubuntu-precise lts-backport-quantal | 18:36 |
| apw | tgardner, *cough* lts-hwe-12.10 surely | 18:36 |
| tgardner | apw, its a branch name | 18:36 |
| tgardner | who cares | 18:36 |
| apw | :) | 18:37 |
| tgardner | apw, am wondering, given your builddirpa patch, if perf build really is parallel safe. | 18:50 |
| apw | tgardner, it may well be, i have not tested it reverting your change to -j1 yet | 18:58 |
| apw | i figured you'd probabally re-test once this upload was up and we could revert | 18:58 |
| tgardner | apw, yep. I'm backporting it to older releases, so it'll get some good testing. tangerine is gonna get worked | 18:59 |
| apw | tgardner, i assume the issue is somewhere back to O in theory, not really sure why we've never noticed before | 19:02 |
| apw | tgardner, so you are doing the backport for the tools build bit yes ? | 19:02 |
| tgardner | apw, dunno, but it is pretty racy | 19:02 |
| tgardner | apw, yep | 19:02 |
| apw | then i can crack a beer and let these virtuals build | 19:03 |
| tgardner | thats common all the way back to lucid | 19:03 |
| apw | la la la ... | 19:03 |
| tgardner | apw, I'm also thinking this is bogus: install-tools: install-source $(stampdir)/stamp-build-perarch | 19:03 |
| apw | there are two linkages there, install-source to get source done at all, and stamp-* to build the perarch tools | 19:04 |
| apw | i think its odd but right, the build-arch (ish) which has install-tools on it, should have install-source i think, but i'd be less keen to fix that back to L | 19:05 |
| tgardner | apw, its the stamp-build-perarch dependency that is unorthogonal | 19:05 |
| apw | ahh ... we wen't through that phase of getting rid of the direct build-xxx things forgetting that they were used to allow make build-foo; fakeroot make foo | 19:06 |
| apw | tgardner, so it should be build-tools and build-tools: stamp-* with the @; to make make happy thing we see elsewhere ... probabally | 19:07 |
| apw | but we optimised a number out, that was inserted after that optimisaiton | 19:07 |
| tgardner | apw, yeah, lemme noodle on it for awhile. you go have a beer. | 19:08 |
| apw | ogasawara, i've done two test builds with that perf fix 'full_build=true' stylee on gomeisa and no problems, i was 2/2 bad before | 19:08 |
| ogasawara | apw: I just reproduced here as well w/o the patch, am going to redo with it. assuming it's successful I'm going to upload. | 19:08 |
| ogasawara | apw: I've pushed the proposed upload to master-next | 19:09 |
| apw | ogasawara, nice ... fingers crossed, its pretty blatently wrong | 19:09 |
| ogasawara | apw: the re-build for i386 on the buildd failed with the same error | 19:10 |
| * tgardner -> EOD | 19:57 | |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!