[03:08] apw: I've finished the 3-phase old kernel bug tracker sync stuff now. I've done a bunch of status and description updates as well. [03:09] apw: there are a few more corner cases I ran into, but I've been documenting it in the uct README file. of most note is how to add and remove SHAs from the bug descriptions. === smb` is now known as smb [07:27] * smb -> BOD [08:39] * apw yawns [08:41] * smb waves [08:43] * abogani waves back to all [08:56] were there some scripts to aid debugging when suspend fails to work? [08:56] my snb desktop hangs while trying that on oneiric [08:57] Not sure how much Colin built into fwts, but he also had some system tap scripts somewhere [08:58] But there also was some bug (I fail to remember atm) that system tap in oneiric was broken.. [08:58] ah [08:58] i'll try fwts [08:59] bug 815944 [08:59] Launchpad bug 815944 in systemtap "systemtap on Oneiric breaks because of kernel commit 449a66fd1fa75d36dca917704827c40c8f416bca" [High,Confirmed] https://launchpad.net/bugs/815944 [09:00] thanks [09:02] smb, i suspect that systemtap needs to change to follow, it is its lot (to break ceaslesly) [09:04] apw, Yeah, I don't think we want to follow a road of changing back the kernel, when the upstream path is to have it removed. [12:30] Do I need to do anything special in order to create a deb source of the kernel that LP can build? [12:30] Maybe a magic wiki article that I missed? [12:33] !kernel [12:33] The core of Ubuntu is the Linux kernel: see https://help.ubuntu.com/community/Kernel - You shouldn't have to compile your own, and if you need to troubleshoot issues, you can try a !Mainline kernel instead, but if you insist, see https://help.ubuntu.com/community/Kernel/Compile (see also !Stages) [12:33] marcoceppi, Something like that? [12:33] !kernel-faq [12:33] A list of common questions about the Ubuntu Kernel can be found in http://wiki.ubuntu.com/Kernel/FAQ [12:33] marcoceppi, how did it fail ? we covered clean yesterday i think [12:36] Yeah, here is the last line from the build line [12:38] enable deprecated sysfs features which may confuse old userspace tools (SYSFS_DEPRECATED_V2) [Y/?] y [12:38] make deprecated sysfs layout dynamically (SYSFS_DEPRECATED_DYN) [Y/n/?] (NEW) aborted! [12:38] Console input/output is redirected. Run 'make oldconfig' to update configuration. [12:38] make[5]: *** [silentoldconfig] Error 1 [12:38] make[4]: *** [silentoldconfig] Error 2 [12:38] make[3]: *** [sub-make] Error 2 [12:38] make[2]: *** [silentoldconfig] Error 2 [12:38] make[1]: *** [sub-make] Error 2 [12:38] make[1]: Leaving directory `/build/buildd/linux-2.6.32-33.72' [12:38] make: *** [/build/buildd/linux-2.6.32-33.72/debian/stamps/stamp-prepare-tree-generic] Error 2 [12:38] dpkg-buildpackage: error: debian/rules build gave error exit status 2 [12:38] The source was built properly, lp build spit this back. I ran make oldconfig in my source tree and it built the .config without an issue. So I wonder if it's not getting my saved config somehow [12:39] you seem to have an out of date config [12:40] fakeroot debian/rules updateconfigs [12:43] So running that seems to produce check-config errors. I assume that I'll need to do something about that. [12:44] ls [12:46] herton, That new lucid kernel upload, I guess that means I should prepare an ec2 tree. Though I cannot remember having seen any bot or person bugging me about that... [12:49] smb: the new lucid kernel was just a respin to revert 2 changes for bug 811745. We should be creating new tracking bugs for you, but this isn't ready yet [12:49] Launchpad bug 811745 in linux "Whole system freeze after safely remove external usb drive" [High,Fix committed] https://launchpad.net/bugs/811745 [12:49] smb: not a respin, but a new update I mean just for that bug [12:51] herton, Ah, ok. So I better wait until you are ready. I was just wondering whether I missed something there [12:51] Unplugging an external usb drive in the cloud is difficult anyways... [12:52] hehe [12:52] smb, does it not have usb emulation ? [12:54] apw, Not the normal usage. There could be ide or scsi emulation but actually only the virtual block devices are used as the are quicker [13:58] Thanks for your help thus far. I'm getting closer. The build servers are telling me that the source isn't clean and that I need to make mrproper - however, when I run that in my local source it removes a lot of the debian control items. [13:58] marcoceppi, you need to "rmdir include/config" normally to fix that [14:00] Ah, I'll give that a try. Thank you [14:24] * ogasawara back in 20 [14:30] ## [14:30] ## Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting [14:30] ## agenda: https://wiki.ubuntu.com/KernelTeam/Meeting [14:30] ## === kamalmostafa is now known as kamal [16:00] ## [16:00] ## Kernel team meeting in one hour [16:00] ## [16:13] bjf: I noticed bug 819632 from the daily bug report [16:13] Launchpad bug 819632 in linux "WARNING: at /build/buildd/linux-2.6.31/net/sched/sch_generic.c:246 dev_watchdog 0x1f6/0x210()" [Undecided,New] https://launchpad.net/bugs/819632 [16:13] bjf: it's against a karmic kernel (which your scripts correctly detected) [16:14] bjf: just curious if you'd want to auto close it rather than have it appear on the report? [16:14] ogasawara, yes, but it didn't get the right tag (and it's an unsupported series) [16:14] ogasawara, i'll look into it [16:25] herton: i just got an unrelated panic (in the wifi stack) on wakeup with -9, so i'm running -11 now and i'll get a screen asap [16:30] ohsix: ok [16:55] Hum, so I removed include/config and I'm still getting mrproper error during the build. Is there anything else I might need to do? [16:55] marcoceppi: make mrproper [16:55] marcoceppi, well the tree should be made clean, so checkin anything you have changed [16:56] then git clean -f -x -d, note this will DELETE any files not tracked by git [16:56] then fakeroot debian/rules clean, then package it [16:56] mrproper removes all the debian/ files - so I won't be able to run debuild [16:56] marcoceppi: mv debian ../debian.orig; make mrproper; mv ../debian.orig debian [16:57] Ah, makes sense. Let me try that [16:58] Shuould I rape mrproper in fakeroot? [16:58] WRAP* [16:58] don't think so [17:01] ## [17:01] ## Meeting starting now [17:01] ## [17:11] bdmurray, i'd be interested in more details of the changes if you able to send them out [17:13] apw: the changes already made or what I think should happen now? [17:14] bdmurray, right, but of what we might find different [17:19] hi [17:20] just reported bug #819925 [17:20] Launchpad bug 819925 in linux "iwlwifi driver exports all packets through netlink" [Undecided,New] https://launchpad.net/bugs/819925 [17:20] ogasawara, ^^ that sounds like something to put on your config-review list [17:21] * ogasawara looks [17:21] * ogasawara adds a work item [17:21] :) [17:21] that should default to N i would assume =) [17:22] apw, hah! you beat the script [17:22] bjf, yeah stop the spam [17:23] well, that's all i got, thanks guys =) [17:23] alfmatos: I'll get it flipped to N and it'll be in the next upload after Alpha3 [17:24] cool [17:29] hell all... [17:29] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/776999 [17:29] Ubuntu bug 776999 in linux "[Lenovo W520] laptop freezes on ACPI-related actions" [Undecided,Confirmed] [17:30] was looking for a little bit of advice on how to move this from confirmed to triage [17:30] I have a W520 and could assist gather some information [17:35] I seem to be getting a lot closer to it actually compiling. The seemingly last hurdle is that 7 configuration values fail check-config. Should I just remove those? Is it maybe they no longer exist in this version of the kernel and are just old? [17:47] marcoceppi, what version of the kernel is the config from and what is the version of the kerenl [17:53] apw I have the latest from ubuntu-lucid.git -33.72 I copied the config from a lucid VM with the latest installed from the repo [17:54] marcoceppi, then no i would not exect the config to change at all [17:55] bjf: pastebin me the minutes, I'll shove em to voices [17:55] bjf, i say don't bother to publish them [17:56] its not a high priority for us apparently [17:56] apw: The build fails at these 7 entries: http://paste.ubuntu.com/657368/ [17:56] marcoceppi, but the config for a lucid kernle would have those set [17:56] apw, ogasawara i've been assured that this will be fixed today, if not i will send email to pgraner that we will no longer be blogging our meeting minutes [17:56] and set correctly else the same failure would have prevented the kernel you have from building in the first place [17:56] bjf: ack [17:58] apw: Makes sense. I'll try to figure out where I went wrong. I may try re-copying the configurations over and running updateconfig again [17:58] ogasawara, http://pastebin.ubuntu.com/657372/ (thanks) [17:58] why not use the config which is in the tree? given it is for the same kernel [17:59] I was just following the instructions from https://wiki.ubuntu.com/KernelTeam/GitKernelBuild I may have gotten confused and copied when I shouldn't have. [18:02] So should I just checkout the configuration files from the last clean commit in the ubuntu-lucid.git? I assume it's in debian/build/ directory? [18:02] marcoceppi, https://wiki.ubuntu.com/Kernel/BuildYourOwnKernel is a good, basic starting point [18:04] bjf: Thanks, I was linked to it yesterday - I'll probably just start from their this time [18:05] marcoceppi: ping me if you run into problems [18:08] I'm trying to avoid starting from scratch. It took me about 5 hours to patch the kernel. Soo, let me see if I can back up until right after I patched [18:16] Okay, I think I've gotten to a point where I can try to perform a clean build. Would it be beneficial to build the kernel locally before trying to get LP to build it? [18:20] marcoceppi, depends how fast your local build is [18:25] It's still failing at those configurations. I wonder if it's because I patched v2.6.32 in the git line, then merged down the changes up to 2.6.32-33.72 [18:25] * tgardner --> lunch [18:25] * smb -Y eod [18:25] If I remove those values, I assume it would make me re-enter them if it really needed it, right? [18:26] marcoceppi, nope, it will then not have things set you need in order to boot most likely [18:26] if you mean removing them from the config check [18:26] I mean removing them from the build's .config [18:26] it will ask you to set thme yes on updateconfigs [18:27] Then what would be causing check-config to throw the error? Could it be expecting a different value? [18:28] what values do the config options have in your config ... git grep FOO debian.master/config [18:28] When I search the debian/build/build-generic/.config file those keys don't exist [18:28] if you got the configs and the enforcer file from the same or nearly the same version of the kernel then they should match [18:29] marcoceppi, thats generated during the build and not interesting as much as the master for them [18:29] Let me check the master [18:31] All of these keys and values exist in config/enforce config/config.common.ubunut and config/config.common.ports [18:33] So, I should run debian/control updateconfigs then try again? [18:33] rules* [18:34] if you do a fakeroot debian/rules prepare-generic (or whatver falvour you care about) [18:34] you can see if the built cdonfig has what you need in it [18:35] It built, but failed with the same 7 check-config errors like before [18:36] you said they were present and correct in the config master files yes ? [18:36] so then the question is what patches have you applied [18:36] I applied the OpenVZ patch [18:37] for 2.6.32 [18:37] so perhaps they prevent those options being set, and you are in a world of hurt [18:37] does the openvz patch set include any configuration (Kconfig) file updates, add any additional depends ? [18:38] It changes a few kconfigs. It doesn't add any external dependencies that I know of [18:38] well if an option is in the master configs and not in the resultant built .config in debian/build [18:38] then there must be a configuration dependancy that is stopping them [18:39] so you need to pick one and validate its pre-requisites [18:39] and find out why its now off/difffernet [18:39] * apw eyes the beer on his desk ... you are mine [18:40] I see, so it's it's time to chase down "why"? [18:40] yep sadly so [18:41] Well the majority is related to SELINUX and APPARMOR which I know does not work with OpenVZ [18:41] then that would explain why they go missing [18:41] if they are incompatible [18:41] So, updating the enforce file would be...the solution? [18:41] then you need to remove the enforcement for them, as they can never be set [18:41] and pay the consequences if userspace needs them [18:42] Well, I'll chase down the others to make sure it's something that is supposed to be missing. Thanks for your help apw enjoy your beer :) [18:43] *glug& [18:51] Awesome, OpenVZ doesn't support DEVTMPFS as well. This is going to be a pain now that I look at it. [18:56] now that is going to be a tricky one [18:56] upstart may cope [18:58] I hope so [18:59] I mean, if people are able to use the linus 2.6.32 kernel on 10.04 I *should* be able to use 2.6.32-33.72 with the patch? right? [18:59] well _maybe_ [19:00] except that that supports devtmpfs [19:01] what I mean was 2.6.32 with the OpenVZ patch (which it was created for) [19:02] if the openvz patch is known to work on ubuntu then yours should, is what i think you are saying [19:07] apw: how did yesterday's bug status sync look to you? [19:10] apw: Yes, only I didn't articulate it very well. [19:13] kees, didn't see anything out of order [19:13] apw: excellent [19:13] wil keep a weather eye [19:14] herton: do you know who is doing the publication of the maverick linux package? it seems to have only hit -updates... [19:14] kees: no idea [19:14] herton: okay, thanks [19:56] kees: sconklin is looking into it, he was told slangasek did the publishing [19:56] kees: it's being worked [19:59] bjf apw: Thanks for your help. I'm able to get past the configuration issues from earlier - now I just need to fix the code errors! [20:00] sconklin: thanks! [20:18] * tgardner -> EOD [20:20] * jjohansen -> lunch [20:27] apw, sconklin: do either of you know the status of ti-omap4 for maverick? [20:27] kees: no [20:28] sconklin: what's needed to make it show up on http://people.canonical.com/~kernel/reports/sru-report.html ? just open bugs? [20:28] a tracking bug that's in process [20:29] sconklin: hm, okay. I guess I don't know what takes stuff from http://people.canonical.com/~apw/cve/pkg/ALL-linux.html in "pending" to showing up on http://people.canonical.com/~kernel/reports/sru-report.html [20:29] I'll ping apw tomorrow [20:29] kees: is this helpful? http://people.canonical.com/~kernel/reports/versions.html [20:29] probably not [20:30] sconklin: I think that just shows me that there isn't something in the PPA waiting for workflow progress [20:30] that's right, and nothing in -proposed [20:33] sconklin: oh, what about linux-lts-backport-maverick? it seems to be missing a tracking bug? [20:35] kees: looking [20:37] sconklin, kees: for linux-lts-backport-maverick it's bug 811215, but it doesn't show on sru-report [20:37] Launchpad bug 811215 in kernel-sru-workflow "linux-lts-backport-maverick: 2.6.35-30.56~lucid1 -proposed tracker" [Undecided,In progress] https://launchpad.net/bugs/811215 [20:37] herton: ah, thanks. [20:37] herton: thanks, I was waiting for git to fetch [20:38] we'll figure out why that's not in the report [20:44] sconklin, i'm looking at the kernel-sru-worflow and the tracking bug shows there, heading to the sru-report [20:44] bjf: thanks! [20:47] yep, the sru-report is the one which misses it === yofel_ is now known as yofel [21:49] cnd, the recent 10.04.3 kernel update "broke" my clickpad in Lucid. I've got an HP ProBook 4320s and stayed on Lucid as Maverick forward don't work properly with the clickpad. [21:49] now that forward problem has come back to haunt me. [21:49] Any suggestions? [21:50] medberry, can you provide more details on what's wrong? [21:52] the left center right areas are treated as one area cnd. [21:52] and it's also really "touchy". Making it much more difficult to click and drag now. [21:52] I suspect the 10.04.3 kernel SRU pulled in some mouse/input changes (similar to what were in Mav and Nat all along.) [21:53] medberry, can you describe the layout of your trackpad? what do you mean by left center right areas? [21:54] sure: the clickpad has a "T" shape overlay on the bottom 20% where the middle of the T is in center bottom. [21:55] representing "left" and "right" mouse buttons. Moreover, the bar of the "T" functioned as a middle click. [21:55] Now, I'm back to the whole area representing a left click (irrespective of the markings.) [21:57] medberry, I'm guessing there was a kernel patch that was in lucid that was not upstream to handle these trackpads [21:57] so "left" is more less the bottom 20-25% and the left 40%, center is the bottom 20-25% and the middle 20% and right is the right most 40% of the bottom 20-25%. I can send a picture if it helps. [21:58] based on a vague memory of what's happened in this area [21:58] sconklin, do you remember any changes in drivers/input/mouse/synaptics.c that went into the latest lucid kernel updates? [21:59] cnd: no but it's really easy for me to check [21:59] medberry, I'll add that upstream still doesn't have a great solution for this, IIRC [21:59] cnd, to clarify, afaict, this is still broken/crippled in M/N/O-alpha as well. [21:59] yeah [21:59] cnd, yes, I think that's the issue. [21:59] the kernel should provide the raw coordinates, which is what occurs in M/N/O/etc. [22:00] did lucid release with a kernel that didn't have the pointer stuff "in kernel". That's my understanding (or misunderstanding) [22:00] but xserver-xorg-input-synaptics needs to handle clickpads properly [22:00] which it doesn't right now [22:00] okay. [22:00] cnd: I don't see any recent changes to that at all [22:01] i.e. last patch was in April 2010 [22:01] I'm going to double check my assumption that it was a kernel change that made the difference. Please disregard until I've completely verified that is the delta. [22:01] medberry, can you check to see what events are emitted by the kernel in the good and bad kernels? [22:01] you can use evtest to see the events [22:01] sure. [22:01] I'll do that and get back to you. Thanks. [22:01] the "good" kernel probably is emitting BTN_LEFT, BTN_RIGHT, and BTN_MIDDLE [22:02] whereas the "bad" kernel is probably only emitting BTN_LEFT [22:02] nod. [22:10] cnd: I may have found what happened with input [22:10] https://bugs.launchpad.net/ubuntu/+source/linux-backports-modules-2.6.32/+bug/814186 [22:10] Ubuntu bug 814186 in linux-backports-modules-2.6.32 "linux-backports-modules-wwan and -input packaged incorrectly" [Medium,Fix committed] [22:11] previously the input driver from lbm wasn't getting loaded, and apparently now it is [22:12] medberry, do you have linux-backports-modules-synaptics installed? [22:12] sconklin, I'm not sure that works out though [22:12] so if lbm is installed, the synaptics module may be the server version [22:13] and so it would fail to load [22:13] but then you'd end up without any input modules loaded [22:13] yeah, it only applies if lbm is installed [22:13] actually, that might work out [22:13] I dunno :) [22:13] well, the kernel also has some input drivers without lbm, so I believe you would just get a different driver [22:13] yeah [22:14] it could be picking up a static standard PS2 driver instead [22:14] and I wouldn't be too surprised if these clickpads worked better in PS2 mode than in full synaptics mode :) [22:17] cnd: a bug with collected info would be helpful. And is this from -proposed? because that change hasn't been released yet [22:17] medberry: are you running -proposed? [22:17] sconklin, meet medberry :) [22:19] medberry: the most helpful thing now is to run 'ubuntu-bug linux' and get all your info into a bug [22:21] sconklin, cnd, nod. [22:22] sconklin, dpkg -l linux-backport\* reports nothing [22:22] note that several touchpad "regressions" have been reported after the update to -33.70; 811420, 811753, 811962, 815357 === medberry is now known as med_brb === med_brb is now known as med === med is now known as medberry [22:40] jwi, bug 811753 is against .32-32.62 where the other three are .32-33.70 [22:40] Launchpad bug 811753 in linux "Synaptics Touchpad keys stoped working" [Undecided,Confirmed] https://launchpad.net/bugs/811753 [22:41] sconklin, my experience is similar to 811753. cnd: evtest reported can't get version: Inappropriate ioctl for device (no matter which mouse dev I chose.) [22:41] gah, the stupid evdev protocol version issue [22:42] medberry, get the oneiric evtest package and install it [22:42] jwi, how do you happen to be tracking those bugs, are you watching input bugs or lucid regressions or ... ? [22:42] it should work just fine [22:42] nod, will do. [22:42] ogasawara, ping (just if you happen to be around right now) [23:00] cnd, oneiric evtest behaved the same: Inappropriate ioctl for device [23:00] I then ran "xinput test ##" with my device on both [23:01] the new kernel reports only button 1 events (and they occur anywhere on the clickpad). The old kernel reports all three buttons but only in the lower 20-25% of the screen. [23:01] Both report movement but the old kernel only reports mouse motion in the top 75-80% of the clickpad. The new kernel reports all over. So it's just one giant button/mouse in its behavior. [23:02] (which as I understand accurately reflects the hardware--software heuristics are required to get the B1, B2, B3 and to distinguish between the "mouse" and "button" areas.) [23:03] my work around for now is to use the 2.6.32-32 kernel [23:06] bjf: incoming regressions, mostly. re 811753: note the user's description, I guess it was reported from the old/working kernel [23:07] jwi, i was looking at the bootdmesg [23:08] jwi, bjf : if you need a guinea pig for these issues let me know. My issue most closely resembles 811753. [23:19] medberry, i have nothing to offer right now [23:21] no worries. Thanks guys. [23:31] sconklin, bjf: where's the code for the workflow mgr bot? [23:31] cnd kteam-tools [23:31] ok [23:33] cnd, it's kind of complicated because it's trying to manage a specific workflow, but it wouldn't be bad to start with, gut it and do your own [23:33] yeah [23:33] bjf, how do you develop? what's your testbed? [23:33] cnd, unfortunately the best thing to use is the qastaging server [23:34] cnd, if you look at the code, there is a --staging command line flag that you can use [23:34] what's the qastaging server? [23:34] is it some semi-permanent instance? [23:34] cnd, you can create any bugs over on that server/service (a testing LP instance) [23:35] cnd, it's got a snapshot of the DB, so any changes you make there are not seen on the production server [23:36] ahh, cool [23:37] cnd, the problem with it is that it isn't real stable [23:37] oh.. [23:37] * cnd just hit an error [23:38] bjf, I assume things are wiped every so often? [23:38] cnd, yes, something like weekly [23:38] ok [23:39] cnd, https://bugs.qastaging.launchpad.net/ubuntu/+source/evolution/+bug/777777 [23:39] Ubuntu bug 777777 in evolution "Mark read / unread only works once in message window" [Low,New] [23:39] cnd that worked for me [23:39] bjf, where's the documentation for the launchpad api? [23:40] https://edge.launchpad.net/+apidoc/ [23:40] cnd the bot uses the LPLTK which is a slick on top of the LP api [23:40] ok [23:46] cnd bzr+ssh://bazaar.launchpad.net/~arsenal-devel/arsenal/python-launchpadlib-toolkit/ [23:50] cnd https://launchpad.net/~arsenal-devel/+archive/ppa