=== smb` is now known as smb [07:30] smb, moin, can you remember the bot-stop- tag name [07:30] ppisati, moin [07:30] apw, The thing about robots.thingy [07:36] moin moin === fmasi_afk is now known as fmasi [08:08] tseliot, i see that 3.10-rc5 kernel finally built, how did the dkms hacking go [08:09] apw: I have started working on nvidia. It's going to be a rather invasive patch because of some changes in the kernel [08:09] tseliot, nasty ... [08:10] apw: yes, I have to switch to proc_create_data and proc_create, among other things [08:11] I bet it will be the same for fglrx and broadcom === arun_ is now known as arun === rbasak_ is now known as rbasak === foka_ is now known as foka [12:20] brb [13:10] apw, should be bot-stop-nagging [13:10] bjf, ta [13:32] apw, any objection to uploading saucy rebased on v3.9.5 today ? 3.10-rc5 still doesn't have aufs support, so I'm not sure we should be switching over just yet. [13:33] rtg_, none, tseliot was having some trouble too, so i thin 3.9.5 is a good place to be [13:33] apw, k, I'm just about done compile and smoke testing. [13:33] rtg_, do you want me to do it, or are you already ... ok then [13:33] apw, have you tested overlayfs in 3.10 yet ? [13:33] apw, rtg_: yes please, I'm not done fixing the drivers yet [13:34] rtg_, i am playing in that sandbox right now, not tested that specific combination but that is about to be a sane test [13:34] BenC: You want to do some install/reboot/fiddle testing of the PPC kernels in raring-proposed on some of your hardware? [13:34] apw, aufs hasn't been updated since march, so we might end up dropping it completely if it is no longer maintained. [13:35] rtg_, they will be pleased :) [13:35] * infinity will mourn the loss of aufs. [13:35] Though, if the overlayfs/inotify argument ever amounts to anything, I might stop caring. [13:37] so on the flipped container touch images i get a reboot loop on grouper (all other subarches seem to work and i can boot past the failure point using the grouper desktop kernel) ... [13:37] http://paste.ubuntu.com/5754882/ [13:39] i assume it is the "Warning: unable to open an initial console" that makes init die ... any obvious idea how i could get that console (without having to enable fbcon which will break the android container) [13:40] (teh kernel panic happens right after run-init switched us to upstart) [13:47] ogra_, I'm not seeing an FBCON option for grouper. CONFIG_FB=y and CONFIG_FB_TEGRA=y are set. [13:47] well, it is there since we use it on the desktop kernel ... [13:48] might depend on something else thats off as well [13:48] but fbcon isnt an option anyway [13:48] it would break the android container [13:48] ogra_, does saucy grouper still work on the desktop image ? [13:48] i tried randomly setting CONFIG_VT and stuff, but didnt get much further [13:49] well, the kernel is the same [13:49] or at least largely ... [13:49] just different configs [13:49] ogra_, configs can make a difference [13:50] so for testing i can replace zImage in the bootimg we use for touch ... indeed it fails miserably later in the boot but it gets past the panic point [13:50] err, yes, i know, thats why i'm here :) [13:50] obviously the -nexus7 kernel has something enabled to make it boot ... which the -grouper one is missing ... [13:51] my way to debug that is to enable options one by one and to do a test boot after a fresh build ... but thats very time consuming so i was hoping you guys have a better idea or see something thats obviously missing or so [13:52] ogasawara, lemme take a stab at diffing the configs [13:52] :) [14:11] ogra_, there are quite a few differences. http://kernel.ubuntu.com/~rtg/config.diff [14:11] CONFIG_HW_CONSOLE looks suspicious [14:11] oh, yeah [14:47] ogra_, just pushed a patch to ubuntu-saucy grouper (CONFIG_VT=y). Can you give that a whirl ? [14:47] sure [14:48] i have a call soon, so only afterwards [14:48] I'll flash my N7 with raring in the meantime to make sure it still works [14:48] but i think i tried with CONFIG_VT=y alone already [14:49] CONFIG_HW_CONSOLE sounds more plausible [14:49] ogra_, the whole patch has several more console options enabled [14:49] ah, k [14:57] apw: does 3.9.5 have this commit? https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=80e928f7ebb958f4d79d4099d1c5c0a015a23b93 [14:58] tseliot, not obviously [14:58] apw: ok, just checking [14:59] tseliot, i cannot see it by sha reference or title, so very likely not [14:59] tseliot, something you need [15:00] apw: no, something that breaks the driver [15:05] hello how do I get BFS and BFQ? [15:08] Roberth1990, i suspect we need a little more context to have any hope of helping you. [15:09] they are a part of the ck patchset for the kernel [15:09] which is there again a part og zen and pf patchsets [15:10] if it is part of ck, then its not something we have any trees for, you'd have to apply it yourself [15:11] ** [15:11] ** Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting [15:11] ** [15:59] hallyn, yo ... bug #1098378 ... test kernels in the bug, if you could check them out and report back [15:59] Launchpad bug 1098378 in linux (Ubuntu Raring) "chroot+overlayfs seems to cause umount mis-behavior" [High,Confirmed] https://launchpad.net/bugs/1098378 [16:00] hallyn, and can you confirm this was new in raring, i think from the fix I expect it was === vanhoof_ is now known as vanhoof [16:16] Daviey, That is probably outside the scope of the meeting but if you want dkms packages checked imo its up to you to provide a list of interest. And I cannot recall "all" being agreed anywhere ever. [16:25] smb: not agreed, but that is what i requested [16:25] it seems easy enough to test all, as it does a handful. [16:25] apw: will look, thanks [16:25] And for a test that seems really cheap to run, it seems like a good idea to the ubuntu ecosystem to do [16:27] Daviey, Things are always easy when one has not to do it. :-P But anyway, I would say this is something the qa team will be responsible for anyway. So if they do, I won't complain. [16:30] smb: it sounded like it was already in progress, from jibel. [16:30] Daviey, It did, and plars just write that email. So we will see [16:31] smb: Have you spotted complexity to it that i have blissfully sailed over? [16:33] Daviey: sorry to jump in here, but I've been chatting with jibel about automated DKMS package testing [16:33] super! [16:33] Daviey: I actually have a meeting with him on thurs, he has something already prototyped for us to go through [16:34] Daviey, To detect the build failing probably is simple enough (beside resource and time). Just the question what is done with that [16:34] Daviey: but if you can get me the list of dkms packages that your team is specifically interested in, I'll be sure they try to land on the list [16:34] Daviey: initiall, it's just getting a quick build test to ensure they still build with major version bumps of the kernel [16:35] Daviey: specific testing of the modules/packages beyond that would be something I'd consider phase 2 [16:35] ogasawara: Right. Just validating that the dkms packages still cleanly install sounds super. === ehw is now known as ehw|afk [16:35] Daviey: yep, that's my initial goal [16:35] I can get a list.. but as a ubuntu developer, i'd love all to be covered :) [16:35] Daviey: indeed, and I suggested to him to to all the packages that rdepends spit out for dkms [16:36] ogasawara: I will make sure I have a favoured list. [16:36] Daviey: great, thanks [16:37] Daviey: I was assuming for you guys is was something like openvs and iscsitarget [16:37] Daviey: not sure what else though [16:37] someone in server raised oss4-dkms as another... not sure why i care about that. [16:38] (HINT, i don't think i do) [16:38] openswan module is interesting, but not demanded [16:39] dahdi-dkms is a server universe package [16:39] openvswitch-datapath-dkms is the priority [16:40] One of the real challenges we face is making these packages work on current development version of ubuntu, and precise (all supported kernels) concurrently [16:41] That oss4 may have been somehow misguided, but that sort of gets me back to the point that a list (not on irc) of interest and to what degree would help [16:43] ogasawara: `reverse-depends dkms` isn't a remarkably long list, I wonder if jibel could just test the lot? [16:43] ogasawara: Obviously, we have a priority list of what we care about fixing first, but I want to know if any of them are broken, so we can decide to fix or remove. [16:45] infinity: indeed, I was suggesting to him to just test the lot since it didn't seem incredibly large. I've got a meeting on thurs to sync with him so I'm happy to bring up everything with him and circle back with you et al. [16:46] infinity: or I can just add you to the meeting if you wanted to sit in [16:48] ogasawara: You can add me, and I'll see how social I'm feeling when it rolls around. :) [16:48] infinity: ack [16:48] infinity, and how awake... [16:48] And that. [17:01] ## [17:01] ## Meeting starting now [17:01] ## [17:14] Home: https://wiki.ubuntu.com/Kernel/ || Ubuntu Kernel Team Meeting - Tues June 18th, 2013 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! === jsalisbury changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Ubuntu Kernel Team Meeting - Tues June 18th, 2013 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! [17:22] * ppisati -> gym [17:25] * rtg_ -> lunch [17:37] has anyone here figure out how to upload .crash data from a kernel oops into a new launchpad bug? ubuntu-bug *.crash asks if I want to submit it but gives no feedback afterwards [18:06] ok looks like bug 994921 [18:06] Launchpad bug 994921 in apport (Ubuntu Quantal) "'ubuntu-bug /var/crash/app.crash' (and even more so, 'apport-cli -c /var/crash/app.crash') should still allow manual bug filing in stable releases" [Medium,Triaged] https://launchpad.net/bugs/994921 [18:23] 2291 apw 9 -11 568m 5904 4792 S 200.0 0.1 1349154h pulseaudio [18:33] infinity: Yeah === fmasi is now known as fmasi_afk [18:45] bjf, http://people.canonical.com/~apw/unstable-saucy/ [18:45] apw, ta [18:47] apw: I'm not sure I want to work too close to you now I see you are unstable and saucy ;) [18:48] davmor2, :) [18:49] BenC: Kay, I already marked it passed for regression-testing based on my own testing, but do let me know if it breaks for any of your platforms, so we can revert that. :P [18:50] infinity: it worked when I built it, so should be good. I'll let you know if otherwise. [18:51] BenC: Compiler bugs and bitflips happen. There's also the possibility that I effed up the checkout. Better safe than sorry. :) [18:54] BenC: Alternately: "It works on 750, 970, POWER5, and POWER7, so sucks to be you if your flavours don't work, nyah nyah." ;) === Guest72281 is now known as jpds [20:22] * rtg_ -> EOD === fmasi_afk is now known as fmasi === fmasi is now known as fmasi_afk