[00:08] kees, sconklin: please see bug 810807 on the Maverick backport to lucid [00:08] Launchpad bug 810807 in linux-lts-backport-natty "kernel-test-security multiple errors on backported Maverick kernel" [Undecided,New] https://launchpad.net/bugs/810807 [02:30] ohsix: yeah, it moved to /var/log/syslog iirc [02:31] hggdh: oh, er, yeah. they are inverted tests. it's checking the release not the kernel [02:32] hggdh: e.g. CONFIG_DEBUG_SET_MODULE_RONX is enabled (a good thing) when it wasn't expecting it for lucid. [02:32] hggdh: let me work on that and get the tests updated. that's going to take a bit of work... [02:41] apw: poke, if you're about [02:42] apw: my desktop dropped back into the failure mode. I was able to ssh in, and check up on it. Turns out compiz is taking up 100% of one core, and xorg is taking up 100% of another. [02:43] apw: is this what happens with the vsync bug? [02:44] process (compiz) is stuck in state "Rl" [02:45] If you're thinking of “disconnecting outputs can cause vblank events to be missed”, then, no, it's not that bug. [02:46] Ah, thanks RAOF. He suggested that as what might be going on with my box. [02:47] humm, it looks like compiz is just oustandingly borked [02:47] The symptom of that bug is that compiz is not taking up CPU because it's waiting in the kernel for a vblank event that will never arrive. It can do that quite efficiently :) [02:47] RAOF: gotcha. Not kernelspace at all :) [02:48] RAOF: cheers, thanks for the help, I'm outa here! [02:56] kees: cool, no problem [03:00] kees: better start brainstorming on how to give users the option of opting in to mainline PAX_USERCOPY, which has a 6-9% performance hit on some benchmarks :-/ [03:00] vasiliy is making some good headway i think [03:27] bliss: yeah, been watching that. :) === _LibertyZero is now known as LibertyZero [06:56] * smb yawns [07:19] hey smb [07:19] heya :) [08:13] can someone help me setup kdump on ubuntu [08:14] kdump doesnt seem to start after panic. crash kernel loads though [08:15] i'm running maverick 2.6.35-28 [08:49] #776964 : compiling my own kernel again. Frowny face. ;-P [12:22] good morning to all. If I may ask, why isn't there a Natty backport for server-i386? [12:39] hggdh, if you mean a flavour -server? That would be generic-pae... [13:04] awilkins: just posted a comment to your bug regarding status of the patch [13:04] OOh, ta [13:05] awilkins: it took us a little longer to get the first natty update out the door due to regressions, but your patch should land in the next round of updates. [13:07] It's ok, my local build has finished... discovered a workaround in the process of patching it also, so wasn't a showstopper, used HDAanalyzer to tweak the pin config parameter so I could use sound card today [13:08] But at least now I won't have to build another kernel [13:10] sconklin: do you guys keep a list of bugs to shove in the changelog when you close a release so it gets auto closed by the janitor when you upload? or would you prefer I just add the BugLink to the commit and re-push to the repo. [13:11] re: bug 776964 [13:11] Launchpad bug 776964 in linux "p5n32e-plus HDA Nvidia AD1988B microphone input not working" [Low,Fix committed] https://launchpad.net/bugs/776964 [13:40] ogasawara: we don't actually have any defined process for getting bug numbers into the changelog if they don't get put into the SRU request. This is mostly because we don't have a good way to associate fixes with existing bugs. This is a problem that needs solving. We usually just leave it to the reporters to close them. [13:40] ogasawara: but the worst part of all that is that reporters don't get notified when something is in -proposed that might fix their issue [13:42] sconklin: right. do the scripts which shove a comment into the bug regarding testing, I assume it parses the changelog? [13:44] actually, no. I think It only hits bugs which are tagged as needing verification. And this isn't all that well scripted and requires manual work. In the case of upstream patches we get via a longterm release, there's only one bug open for all those patches, and we don't associate other bugs like the one you linked [13:45] sconklin: ah [13:45] This whole area is probably the next big area we need to focus on [13:46] We had a similar problem determining when upstream commits close CVEs, but the security team handles that now afaik [13:46] seeing as this patch came to us through stable, so I supposed I could mark it as a duplicate of the master bug which covered the pull from stable [13:48] that will get it closed, it won't result in messages to the bug. But - good idea [13:48] and that's a really good solution until we do something better === kentb-out is now known as kentb [14:32] * ogasawara back in 20 [14:42] sconklin, herton: I'm pushing a change to the LTS branches in Lucid that fix some udeb/firmware problems reported by Etienne. [14:45] tgardner: ack [14:47] herton, be sure to commit the new files that are copied by debian.maverick/etc/update-from-maverick-master: [14:47] debian.maverick/d-i/exclude-modules.amd64-virtual [14:47] debian.maverick/d-i/exclude-modules.i386-virtual [14:49] same with natty [14:50] hmm ok, yeah I remember we had to add some extra files to commit already, we will note this ones as well [14:50] perhaps the script could do git add on them automatically [14:51] herton, it does automatically add files that are under git control, but doesn't typically add new files. this is a one time thing (for now) [14:55] yep, it could do the add of required new files, if possible (don't remember the exact details, I think it's possible to do) [14:58] anyone else notice launchpad throwing away their comment if setting status and importance at the same time? [15:00] more pain, less gain [15:01] herton, there really shouldn't be that much churn with the backports from released kernels, so I'm not sure its worth the complexity in the update script. [15:04] cnd: I read you've developed some clickpad patches. In your opinion, what's the proper way to support clickpads (bug 582809) ? [15:04] Launchpad bug 582809 in linux "Synaptics Clickpad touchpad buttons are not working" [High,Triaged] https://launchpad.net/bugs/582809 [15:04] Seems like there's a divergence of opinion on whether kernel patches are needed. === shadeslayer_ is now known as shadeslayer [15:19] ogasawara, do you know how to change the focus in a bug report from the upstream package to the Ubuntu package so one can make release nominations ? [15:20] e.g., bug #509180 [15:20] Launchpad bug 509180 in ecryptfs "ecryptfs sometimes seems to add trailing garbage to encrypted files" [High,Fix released] https://launchpad.net/bugs/509180 [15:20] tgardner: I usually hand edit the url [15:20] ogasawara, hmm, I tried that. what you you change it to ? [15:20] use https://bugs.launchpad.net/ubuntu/+source/linux/+bug/509180 [15:20] Ubuntu bug 509180 in ecryptfs "ecryptfs sometimes seems to add trailing garbage to encrypted files" [High,Fix released] [15:21] ah, I missed the +source [15:21] ogasawara, shit, what a ginormous pain in the ass that is. [15:21] tgardner: totally [15:22] tgardner: not intuitive at all === jpds_ is now known as jpds [15:34] kees: I think you wanted to change Security-signoff to Invalid in bug 806586, instead of "linux-lts-backport-natty (Ubuntu)" ? [15:35] Launchpad bug 806586 in kernel-sru-workflow/security-signoff "linux-lts-backport-natty: 2.6.38-10.46~lucid1 -proposed tracker" [Undecided,In progress] https://launchpad.net/bugs/806586 [15:39] ppisati: have you had a chance to investigate https://bugs.launchpad.net/ubuntu/+source/linux/+bug/806113 [15:39] Ubuntu bug 806113 in linux "Series of segfaults early in the kernel boot process on omap." [Undecided,New] [15:39] ppisati: it was just raising in the release team meeting [15:39] s/raising/raised/ [15:48] ogasawara: nope, let me see [16:10] herton: thanks! I don't know how I managed to click that :( fixing [16:11] johanbr, it's been a while since I've looked into clickpads in detail [16:11] off the top of my head I think there really doesn't need to be any extra kernel patches to get the functionality people want [16:11] it can more easily be handled in xserver-xorg-input-synaptics [16:11] right [16:12] but that's really off the top of my head [16:12] so don't hold me to it if I'm forgetting an important detail :) [16:13] it appears anything to do with synaptics and multitouch is a big pita [16:13] for the synaptics driver, something like https://patchwork.kernel.org/patch/92435/ I guess [16:14] I think I'll try building a synaptics driver with that patch for testing, then bug the X guys [16:14] thank you! [16:16] johanbr, that kernel patch has been thrown around a lot [16:16] I think there's good reason why it's not in the upstream kernel tree by now [16:16] but I couldn't tell you what that reason was [16:16] I can only guess that people thought it was better handled in userspace [16:17] sorry, wrong URL... I meant http://patchwork.freedesktop.org/patch/3160/ [16:17] ogasawara: just tested latest oneiric/master and booted ok [16:18] ogasawara: i guess latest rebased to -rc7 solved it [16:18] s/rebased/rebase/ [16:18] ppisati: cool, care to just post a comment and close out the bug [16:18] ogra_: ^^ [16:19] awesome ! [16:20] done [16:20] * ogra_ loves self-resolving bugs [16:20] :) [16:20] johanbr, yeah, that looks like a good starting point :) [16:44] * ppisati -> gym [16:44] have a nice weekend [16:44] see you === tgardner is now known as tgardner-afk [19:17] sconklin: all tests on bug 806586 are failing the same way; this seems to be a test script issue. If kees agrees to, I can mark the QA task complete and qa-testing-passed [19:17] Launchpad bug 806586 in linux-lts-backport-natty "linux-lts-backport-natty: 2.6.38-10.46~lucid1 -proposed tracker" [Medium,In progress] https://launchpad.net/bugs/806586 === kamalmostafa is now known as kamal [19:18] hggdh: looking. Let's not force a pass unless we understand what's happening [19:18] hggdh: is it only the security tests which are failing? [19:19] sconklin: indeed, this is why I stated 'if kees agrees'. With no postion from him I will tag it failed [19:19] sconklin: yes [19:19] ok, I agree that it's his call (kees) [19:29] hggdh: I'm so close to having it fixed. been working on it all morning [19:30] kees: I am ok on waiting [19:30] hggdh: I've got it down to 2 failures, and should have it finished soon. [19:30] hggdh: cool, thanks. === yofel_ is now known as yofel [19:48] hggdh: okay, it's a really massive change, but I've tested on all my KVM instances and on the natty backport kernel. I haven't retested on ARM or weird stuff (xen, etc) so there might still be some subtle fixes needed, but qrt rev 1347 should work for the backport kernel now [19:49] kees: thanks, I will update the tarball and try again (including xen) [19:52] hggdh: thanks! let me know if anything looks weird. I think I've got the nastier of the corner-cases already (mostly relating to kptr_restrict and its change) [19:52] kees: certainly will ;-) as an aside, all 8 failures were indeed due to checking on ubuntu version? [19:53] kees: so, what do we do about inet_diag (re: kptr_restrict) [19:54] hggdh: yeah [19:54] bliss: magic randomized id numbers? [19:55] specifically, getting david miller to accept that [19:56] I thought that was his idea? [19:56] he seems resistant to the whole thing [20:56] * herton --> eow [21:06] kees: seems to work perfectly on xen also :-) [21:06] kees: thank you, sir [21:06] ogasawara, around? [21:07] bjf: yep [21:07] ogasawara, seems like this team should probably die: Ubuntu Kernel ACPI Team https://launchpad.net/~ubuntu-kernel-acpi [21:07] bjf: wow, didn't even know we had such a team [21:08] ogasawara, exactly [21:08] bjf: you'll probably have to get Tim to delete the team since he looks to be the admin [21:24] sconklin: OK. Almost ready to end on bug 806586. Xen i386 shows http://pastebin.ubuntu.com/645023/ which, I guess, is sort of expected. But I want confirmation on it, if you do not mind [21:24] Launchpad bug 806586 in linux-lts-backport-natty "linux-lts-backport-natty: 2.6.38-10.46~lucid1 -proposed tracker" [Medium,In progress] https://launchpad.net/bugs/806586 [21:25] hggdh: yes that's expected for this release. The fix is already committed for the next release [21:25] so that's a pass [21:26] sconklin: roger. Thank you. Tagging it a pass. [21:48] hggdh: \o/ [21:55] kees: again, thank you :-) === rsalveti` is now known as rsalveti [22:04] hggdh: you bet, thanks for doing the testing :) [23:32] Hey, I'm trying to narrow what regressed in linux to cause https://bugs.launchpad.net/ubuntu/+source/linux/+bug/811214 but I don't think I can bisect across Ubuntu rebases, and I'm not sure how to generate a config sensible for mainline [23:32] Ubuntu bug 811214 in linux "Hangs while suspending with iwlagn on Intel Corporation PRO/Wireless 5350 AGN [Echo Peak]" [Undecided,New] [23:32] any advice on setting things up? [23:34] hmm apparenlty bisect will take the tree as a base [23:36] ok, it is not working simply because I need the packaging stuff to build .debs [23:37] because I have LVM, I need an initrd and that makes things rather painful to iterate over