[00:02] Is there an ETA for the lucid mini 9 sound regression caused by kernel 2.6.32-34? (LP: #875300)? [00:03] or do I have to apply workaround for all family and friends? [00:44] more details are here: http://stackoverflow.com/questions/8088901/ [00:46] redzarf: whats up :) [00:46] jjohansen: I was hoping you might be around :) [00:46] still can't get the image deb file to build [00:46] yeah I gathered that :) [00:47] redzarf: what kernel source are you using, and on what version of the OS are you building [00:47] kernel: linux-ec2-2.6.32 [00:47] OS is lucid (latest on EC2) [00:48] alright and where is it failing [00:48] Been trying from the plain Ubuntu AMI's [00:48] I think it's during the fakeroot debian/rules binary [00:49] the stackoverflow page has the error message that comes up at the end [00:49] redzarf: okay, can you paste the error some where? paste.ubunt.com ? [00:49] redzarf: ah, okay looking again [00:49] thx [00:51] redzarf: in the source directory, what debian directories exist? is it debian and debian.ec2 or debian and debian.master [00:52] all 3 [00:53] redzarf: hrmm, try binary-ec2 [00:54] it should be hooked up so that binary, does binary-ec2 /me is wondering if there is a bug there [00:55] still similar errors as before (no image file) [00:56] redzarf: what is under debian/build/ [00:56] i'll try redownloading source and doing it from scratch, then paste the full output [00:56] i'll try redownloading source and doing it from scratch, then paste the full output to paste.ubuntu.com [00:57] redzarf: oh wait did you do a fakeroot debian/rules clean first [00:57] i did [00:57] but i'll do it from scratch to be sure [00:58] would you recommend I get linux-image-2.6.32-318-ec2 or just linux-ec2? [00:58] redzarf: honestly I would recommend the git tree [00:59] what settings would i need to change for ec2? [00:59] git clone git://kernel.ubuntu.com/ubuntu/ubuntu-lucid.git [00:59] git checkout --track ec2 -b ec2 origin/ec2 [01:00] second command from within the directory the first creates [01:00] ah, excellent :-) [01:00] nearly finished using the apt-get version... [01:00] I know the source tarball should be good enough, but I always use the git tree [01:02] http://paste.ubuntu.com/738824/ [01:03] looking [01:03] btw, i think the file that is touch'ed on line 11 doesn't exist - the place it's copied from doesn't have that file [01:05] redzarf: ah yeah, okay do [01:05] make mrproper [01:05] git checkout -f [01:06] and try building again [01:06] ok, doing the clone now [01:07] then I do the first checkout above? (--track ec2) [01:07] then those 2 commands? [01:08] do i make menuconfig just before doing the debian/rules clean? [01:08] oh, sorry no [01:08] within a git tree, those last to will let you get past the mrproper error and build, but not from the source tarball [01:09] so the the clone and checkout commands [01:09] then do [01:09] fakeroot debian/rules editconfigs [01:10] if you do menu config before clean, its changes will get blown away [01:10] actually you will need to do an [01:10] fakeroot debian/rules clean [01:10] before editconfigs [01:11] editconfigs is a wrapper around make menuconfigs [01:11] i'd been doing make menuconfig then the fakeroot debian/rules clean then fakeroot debian/rules binary, so i guess they were out of order [01:11] right [01:12] well you can do it but, I won't guarentee that it will work [01:12] so do i still do make mrproper and git checkout -f ? [01:12] actually I think make menuconfig will but then you hit the mrproper problem [01:12] no [01:13] so this is what i'm about to run: [01:13] git checkout --track ec2 -b ec2 origin/ec2 [01:13] fakeroot debian/rules clean [01:13] fakeroot debian/rules editconfigs [01:13] fakeroot debian/rules binary [01:14] fatal: git checkout: updating paths is incompatible with switching branches. [01:15] oops [01:15] git checkout --track -b ec2 origin/ec2 [01:15] * jjohansen messed up the checkout syntax [01:15] thats ok, the noob who doesn't know any better will forgive you :) [01:17] well even for someone who uses git daily sometimes its syntax is uhm less than friendly [01:20] the binary command is running - taking much longer than it ever has before, so I'll take that as a good sign [02:01] jjohansen: it's finished and the image file is there!! :) [02:01] so I just run dpkg -i linux-*.deb then restart - anything else? [02:02] redzarf: that should be it [02:03] great, thanks for your help! [02:05] redzarf: np [08:39] GrueMaster, did it end up in universe perhaps [08:55] morning .+ [08:58] morning [10:01] morning morning === smb` is now known as smb [10:56] Hi, I'd like to help with the ASPM bug, how do I know that I've installed the patched kernel correctly? [10:57] does it simply replace the existing one or should it be selectable on boot? [10:57] seb03, It would replace the existing one if you take the same version [10:57] Theoretically you should notice a difference in uname -a after boot [10:58] but in doubt it is the one compiled on Sat-12 [10:58] sure do: uname -a Linux snoopy 3.0.0-13-generic #22+mjgaspmfix SMP Mon Nov 14 18:02:48 UTC 2011 i686 i686 i386 GNU/Linux [10:58] So that is the patched one [10:59] so how do I get rid of it again after I've done my power readings? [10:59] install the meta package from the repo again? [11:02] seb03, In doubt you can get the package from there https://launchpad.net/ubuntu/oneiric/+source/linux/3.0.0-13.22 [11:02] There is also some way to force a reinstall with specifying a version, but I personally find it rather complicated. :) [11:04] thanks [11:10] hi [11:11] https://bugs.launchpad.net/ubuntu/+source/linux-firmware/+bug/869502 can you have a look at this bug to give me infos how i can support you on a better way to get it fixed [11:11] Launchpad bug 869502 in linux "Kernel-Panic with 3.0.0.12-generic on asus eee pcs and msi wind (both using rt2800 wifi chipset)" [High,Confirmed] [11:14] see also http://article.gmane.org/gmane.linux.kernel/1211290 and http://rt2x00.serialmonkey.com/phpBB/viewtopic.php?f=5&t=6192 [12:08] hi, quick question: I'm maintaining a local kernel package based off the ubuntu-lucid git repo. I'm wondering how I should make commits to incrememnt the changelog frmo the ubuntu version [12:09] is there some magic string i need to put in the commit to get it show up in debian/changelog? and is there anything else I need to do to increment the ubuntu version? [12:16] seanius, kernel maintenance is described somewhere in https://wiki.ubuntu.com/Kernel [12:20] seanius: everything you commit (except if you use the Ignore: Yes string in the commit log) should show up in the changelog [12:20] seanius: about the version, it's your duty to increment it [12:21] seanius: you mean the last 2 numbers, right? [12:21] seanius: e.g. linux-image-3.0.0-1206-omap4_3.0.0-1206.12_armel.deb [12:21] seanius: the 1206 and 12 numbers in this pkg [12:31] * ppisati -> lunch [12:48] ppisati: none of my commits show up in the changelog, but like I said I don't know if there's some documented syntatic sugar i'm missing in the commit or some script/template etc [12:49] tgardner: i will poke around and see what i can find [12:49] seanius, you probabally need a fakeroot debian/rules insertchanges to get them into the changelog, of course that implies you have followed standard proceedure in opening the changelog entry (fakeroot debian/rules newrelease) [12:50] all of this is documented (poorly) in the kernel-team wiki [12:50] looking at https://wiki.ubuntu.com/KernelTeam/StableHandbook/StablePatchApplication right now [12:51] basically, i have a set of about 50 backported patches which i need to maintain on top of one of the ubuntu-lucid kernels [12:51] and periodically rebase against the ubuntu updates int the repo [12:54] each time i rebase the list of patches get a bit smaller but i don't see the problem ever going away entirely [12:55] at least until 3.2 is released and backported :) [12:55] it looks like apw's suggestion agrees with the above wiki page, so i'll try working with that [12:56] thanks! [13:12] one more question: what provides maint-getabis? [13:19] there is a getabis script in the tree, maint-getabis is more for use by team members, as it talks to our PPAs etc [13:20] (which would be in the kteam-tools git repo) [13:20] right, thx [13:22] and for changes to be inserted via debian/rules insertchanges they need to have an "UBUNTU:" line in them? [13:24] nope [13:24] for them to have your name against them ye [13:24] yes, otherwise no [13:24] or... i guess the new release commit has to go before them [13:25] nope, not that either [13:25] it looks for the previous release commit [13:25] the format of the release commit title is important [13:25] insertchanges turns my new release changelog into a blank entry [13:25] seanius: did you close it properly? [13:26] its all simple shell in debian/rules.d so you should be able to debug it pretty easy [13:26] its definatly orientied to our use models [13:26] understandable [13:26] okay so this is what i'm doing so far [13:26] normally an empty entry is triggered by it failing to find the previous release commit [13:27] git pull --rebase to get some updates [13:27] and fakeroot debian/rules printenv will tell you what it things current and previous are [13:27] git clean -dffx && git reset --hard [13:27] fakeroot debian/rules clean [13:27] fakeroot debian/rules startnewrelease [13:27] git commit -s -F debian/commit-templates/newrelease [13:28] oh, oops, there's a git add debian/changelog in there [13:28] er, debian.natty/changelog [13:28] and then fakeroot debian/rules insertchanges, which sets the changelog to be empty [13:29] ok, as i say thats probabally because its failing to find its previous entry [13:29] fdr printenv will show you the version it is looking for [13:31] i also see an error from expr when running fdr startnewrelease [13:31] apw: will look at printenv after i clean up again [13:32] then likely your version format is probabally not working for the tooling, and the tools may need a tweak [13:32] but i haven't added any version yet [13:32] but this is in the natty-lts-backport tree so maybe it's something existing in the repo [13:33] if your version is triggering the expr then likely the previos version is being broken [13:33] which means we won't find it to do the entry detection [13:33] you can just ignore the tooling there and use like [13:35] if you look at the implementation of insertchanges, you can see how it works [13:35] you can use git log of your own in the same command line [13:36] ah, yeah, you have a bug in rules.d :) [13:37] revision = 13.52~lucid1 [13:37] @nextminor=$(shell expr `echo $(revision) | awk -F. '{print $$2}'` + 1); [13:37] yep, probabally never used on lts backports [13:38] we don't generate those that way [13:38] but really, if all i'm interested is sticking a version on top of the existing backport version, and don't care about good changelog documentation, i could just directly edit the debian.natty/changelog file? [13:39] yes [13:39] okay, simplicity ftw then :) === ericm|ubuntu is now known as ericm-Zzz [15:29] * ogasawara back in 20 === yofel_ is now known as yofel [15:41] smb, whats the latest official relase of your drm33 tree ? [15:42] * smb believes 2.6.32.48.21 [15:43] apw, ^ Should be that. Am I slacking somewhere? [15:43] smb, no making sure the flip back to kernel.org worked [15:43] and that we are up to date [15:43] Ah, ok [15:45] apw, We should be up to date from before. I think I saw builds happening while I still was bringing k.org back to life [15:46] url = git://kernel.ubuntu.com/smb/linux-2.6.32.y-drm33.z.git [15:46] oh we are on the old one, which one do you want to use [15:46] are you keeping our mirror up to date anyhow? [15:48] apw, Sort of automatically as I use it to transfer to local builders to test builds. I had been a bit more lax with the tag, but I guess I will use it as a real mirror now [15:48] Otherwise: git://git.kernel.org/pub/scm/linux/kernel/git/smb/linux-2.6.32.y-drm33.z.git [15:54] smb, ok will leave it pointing at your ubuntu one for nwo [16:01] ## [16:01] ## Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting [16:01] ## agenda: https://wiki.ubuntu.com/KernelTeam/Meeting [16:01] ## [16:55] ## [16:55] ## Ubuntu Kernel Team Meeting - in 5 minutes - #ubuntu-meeting [16:55] ## === ogasawara changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Oneiric Kernel Version: 3.0.0 || Ubuntu Kernel Team Meeting - Tues Nov 22 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! [20:02] ogasawara, ok i've redone the configuration review and dumped a slew of new inconsistant items on your plate :) [20:02] apw: ack [20:02] ogasawara, did we add doing a re-review at rally to the work items ? [20:03] ogasawara, I was sure ther would be a few. it was kind of mind numbing with all the new options [20:03] apw: yep, and I thought you added it to the agenda already [20:03] tgardner, yeah epecially as you see them all 20 times [20:03] ogasawara, will check and confirm [20:04] ogasawara, yep its there good [20:04] tgardner: how's the sprint going? [20:04] * ogasawara rebases precise to 3.2-rc2 [20:05] ogasawara, neck deep in server configuration.... [20:10] ogasawara, I assume there is no patches for the 3.1.0 -> 3.1.1 stuff [20:10] apw: which stuff? [20:10] Replace compat-wireless 3.1.0 with 3.1.1 [20:11] that stuff, i assume your thread is not meant to contains patches for the code itself [20:11] apw: not exactly, it just dumps the 3.1.1 tarball over the 3.1.0 bits [20:13] ogasawara, what are we naming the meta packages nwo [20:14] i was sort of expecting the relase to be 3.1 in the meta package, pointing to 3.1.x whatever it is called in the lbm package [20:14] apw: for oneiric it's linux-backports-modules-cw-3.1.1-oneiric- [20:14] apw: I suppose we could make it just be 3.1 [20:15] we use 2.6.38 presumably for the meta package in previous releases yes? [20:15] ie it points to the base version? then i think the new ones should be 3.0 3.1 etc as they are meant to upgrade [20:16] apw: yep, in lucid for we used 3.0.0, so I followed suit and used 3.1.1 [20:17] well 3.0.0, and 3.1.0 (where the 0 is always 0) might also make sense as it matches our kernel naming [20:17] but the .1 is like patch releases right? if so then we don't want that bit to change in the meta package name [20:17] so either not there, or zapped to 0, so we get upgrades [20:19] apw: yah, I think I like the 2 digit convention [20:19] apw: eg just 3.1 [20:20] ogasawara, yeah [20:20] apw: I'll send a revised patch to the list [20:20] tgardner: who do I need to ping for SRU regressions? [20:22] mdeslaur, start with the stable team [20:22] who is on the stable team? [20:23] mdeslaur, bjf and herton in the first instance [20:24] apw: thanks! [20:24] mdeslaur, fileing a bug against the version with regresion-proposed or regression-updates would get their attention too [20:24] what you found, anything fun ? [20:26] trying to test the latest natty kernel in natty-proposed, but it seems like linux-image-2.6.38-13-generic is not in the archive. Is it still building? [20:26] apw: there's already a bug open (#875300)...we broke sound on mini 9's running lucid...and now, a bunch of friends have started pestering me :P [20:26] apw: never _ever_ recommend machines to friends :P [20:26] EtienneG, shouldn't work that way as things are copied in as a chunk [20:27] EtienneG, which flavour are you having issues with [20:27] apw, -generic, amd64 [20:27] apw, it shows on https://launchpad.net/ubuntu/natty/amd64/linux-image-2.6.38-13-generic/2.6.38-13.52, but it not apt-getable from us.a.u.c or archive.u.c directly [20:28] also, linux-image-generic do depend on linux-image-2.6.38-13-generic, but it is not there ... [20:28] I am perplexed [20:29] might just be me doing something wrong [20:30] EtienneG, they appear to be in universe [20:31] * apw investigates [20:31] ah! [20:31] weird, but he [20:32] EtienneG, an inexperienced archive admin likely forgot the overrides (again) [20:32] tgardner, looks like it yep [20:32] * apw starts poking on #ubuntu-devel [20:32] tgardner, I am shocked! is there such a thing as an "inexperienced archive admin"? :) [20:32] * apw wishes it was not so [20:33] though the kernel copies is meant to be scripted so in thory that could be part of the scirpting [20:33] EtienneG, adding universe temp should let you get testing [20:33] anyway, that was it. my sources.list entry for natty-proposed only had main and restricted [20:33] so there [20:34] thanks for the pointer, as long as I can it, I am happy [20:34] I guess it will be fixed in time [20:36] yep, will poke that aa's to get it resolved [20:36] looking likely will be tommorrow now [21:04] ogasawara, under the new rougime wherein the archive is meant to work all the time, i wonder if we should upload a semi-official kernel to the kernelppa or something so we can get some testing on it before unleashing it on the unsuspecting public [21:04] apw: sounds reasonable [21:05] of course that assumes we have PPAs [21:05] apw: I know there are still some powerpc build failures that I need to fix us (assuming they're not fixed with -rc2) [21:06] ogasawara, ok :) let me know if i can help, else i'll keep grinding on the configuration check tommorrow [21:06] * apw drifts away [21:07] apw: ack, I'm gonna kick off test builds and see what's still breaking. Likely won't get to fixing them up till tomorrow since I drop off here in a bit. [21:07] ogasawara, sounds like a plan to me, now where is that beer [22:01] apw, you might find this thread intresting https://lists.ubuntu.com/archives/ubuntu-users/2011-November/254160.html [22:02] (looks like overlayfs /cow contents dont go into swap) [22:08] ogra_, could be === ericm-Zzz is now known as ericm|ubuntu