=== chiluk_away is now known as chiluk === chiluk is now known as chiluk_away === chiluk_away is now known as chiluk === chiluk is now known as chiluk_away [08:09] moin [08:47] * apw waves to ppisati === smb` is now known as smb [08:59] * smb waves to apw and ppisati [09:00] moin [09:01] * apw notes that mounting encrypted drives is "not allowed again" and mumble is broke, and until my update completes i cannot reboot for either [09:04] apw is truly in Monday hell [09:10] truly indeed [09:10] though i have tea, so something is ok [09:16] things _seem_ to be working ok [09:18] (famous last words) [09:19] all true === henrix_` is now known as henrix [10:31] brb [12:11] * henrix -> lunch [12:29] apw, bug 1113396 [12:29] Launchpad bug 1113396 in linux-nexus7 (Ubuntu) "syslog spammed with cpu_up/cpu_down messages after upgrade to linux-nexus7 3.1.10-9.24" [Wishlist,New] https://launchpad.net/bugs/1113396 [12:30] not sure if there is an easy way to quieten it [12:33] ogra, i am not sure i would expect to see the kernel turning its own cpus on and off [12:35] well, i dont think the interactive giovernor is a mainline thing [12:38] ogra, and which version was this ok in ? [12:38] none, we only switched to interactive pretty recently [12:38] ogra, as i was runing .24 and it was not saying it, i rebooted to run .25 and it is ... [12:38] ok so this is triggered by a userspace change [12:39] the governor code itself is noisy, i assume android simply de-routesd the messages into their own log or so [12:39] right, it is triggered by actually activating the governor [12:40] (interactive means there is userspace config, we only put that into place recently (it is identical with teh android setup now)) [12:40] see /etc/init.d/ondemand on an up to date nexus7 [12:40] ogra, so all of these messages are at standard low levels, so they won't appear on the console [12:41] ogra, but they will be logged to dmesg, and then to syslog indeed [12:41] yes, but they make dmesg unreadable [12:41] and syslog, yeah [12:43] those messages are key messages during boot so one knows that your CPUs have come up right [12:43] its the dynaminc "core up/down" messages that are noisy [12:44] * ogra doesnt care about boot ... but about readable logs :) [12:44] you will if it doesn't [12:44] heh, indeed [12:44] and this has a measurable difference in something which makes it worthwhile [12:45] definitely [12:45] with all other governors i have a constant system load above 1.0 [12:45] its the only way to get the CPU to behave sanely it seems [12:45] other than load, which could well just be a lie [12:46] battery life definitely is extended (i didnt measure exactly how much yet, but it is noticeable) [12:46] ok i'll have a think about it [12:46] i marked it as wishlist for a reason though ... its not super high prio [12:47] (i guess you can filter with grep/sed/whatever if actually needed) === rsalveti_ is now known as rsalveti [13:40] henrix, when are you spinning the next -proposed kernels? [13:41] brendand: a new SRU cycle just started [13:41] brendand: so, we should have new kernels in -proposed by EOW [13:41] henrix, so the testing week will be 18th to 22nd? [13:42] brendand: yes, i believe so === chiluk_away is now known as chiluk [15:03] * infinity eyeballs the oneiric SRU... [15:03] henrix: I thought we were taking an SRU break? :P [15:03] infinity: i believe that break is now over :) [15:03] Huh. Time flies. [15:04] Though we're still skipping this one for P/Q, I guess? [15:04] infinity: we skiped one cycle only, so... a new one starts this week [15:04] Or just queueing things up in the PPA, but not accepting them. [15:04] So as to not disrupt 12.04.2 [15:05] infinity: no, i believe we're rolling all the kernels this cycle. we have a bunch of stuff already in master-next branches. time to flush, i believe :) [15:05] but i may be wrong :/ [15:05] bjf: any thoughts? ^^^ [15:06] Well, precise is frozen until release, so you can roll what you want, but it's not going anywhere. That's all. [15:07] (And it releases in 10 days) [15:07] So, that may be decent timing anyway. I can do the PPA->proposed copies as soon as the release is out. [15:08] infinity: makes sense to me. but i'll confirm this with bjf once he's around. he may prefer to skip P again [15:10] * ext4: quiet 'unused variables' compile warnings [15:10] - LP: #1071314 [15:10] - CVE-2012-4508 [15:10] Launchpad bug 1071314 in linux (Ubuntu Hardy) "CVE-2012-4508" [Medium,New] https://launchpad.net/bugs/1071314 [15:10] ^-- How on earth does that warrant a CVE? [15:11] infinity: where are you seen that? [15:12] henrix: Your changelog. [15:12] infinity: right, but in which kernel? [15:12] Probably just a copy-and-waste oops. [15:12] henrix: oneiric. [15:12] infinity: ok, let me check... [15:13] infinity: ah! so, the prob was that the fix for that CVE required some adjustments (this was commit dee1f973ca341c266229faa5a1a5bb268bed3531) [15:14] henrix: Stacked commits, fair enough. Still seems like an odd way to attribute the bug, but I don't really care either. :P [15:14] infinity: so, the CVE fix contains 2 commits [15:15] infinity: yeah, i just added that to track the commit. i should probably add a note next time... [15:44] * ogasawara back in 20 [16:05] infinity, unless something happens, I think we are skipping P and Q again this cycle. it means the following cycle is going to have a crapload of commits in those kernels. [16:06] bjf: That sounds fair to me. [16:06] hggdh, brendand, vanhoof ^ [16:09] bjf - oh ok. so we're not looking at new P and Q until the beginning of march? [16:09] brendand, that's what it's looking like [16:10] bjf - what's the motivation exactly? [16:11] brendand: Not disrupting the point release. [16:11] brendand, if you read more of the scrollback you can see the discussion. it's about getting the .2 release out and not allowing P kernels into -proposed until .2 is released [16:11] bjf: To be fair, we COULD do the cycle as normal and let the kernels into -proposed, and just not promote them until post-release. [16:12] bjf: I'm open to that option too, since I'm the one who tends to do all the archive side anyway. [16:12] infinity, not promote them to where until post-release? [16:12] bjf: To updates/security. [16:12] infinity, they shouldn't go to updates/security until the following week anyway [16:12] bjf - ah yeah. i forgot that's delayed until next week [16:13] infinity, i thought you did all the point release "work up" in -proposed [16:13] bjf: We'll be switching image builds to -updates only pretty soon now. [16:13] infinity, when will that switch happen? what is "pretty soon now"? [16:14] Like, this week, maybe today or tomorrow. [16:14] bah! [16:14] henrix, hggdh, vanhoof, brendand I take it all back, we are turning the crank on everything. [16:15] infinity, if i can get into -proposed by next monday, that's good enough for us [16:15] bjf: ack :-) [16:16] bjf: Alright. I may artificially delay for a few days to make sure all our point release ducks are in a row, but "by next Monday" seems entirely reasonable. [16:17] bjf: ack [16:18] * bjf feels much better only skipping a single cycle [16:20] * infinity is reminded that he needs to do a d-i upload to precise for the last SRU round... [16:27] yes please don't forget poor little precise ;p === henrix is now known as henrix_ === henrix_ is now known as henrix [17:01] * ppisati -> gym === henrix is now known as henrix_ === henrix_ is now known as henrix [17:20] ogasawara, i'm starting to look at the release notes for 12.04.2. have you done anything there yet? [17:20] bjf: not yet, but it is on my todo list for today [17:20] ogasawara, so i can just ignore it :-P [17:21] ? [17:21] bjf: yah, let me take a first stab at it. I'll probably ask you to review/edit tomorrow. === kamal1 is now known as kamal [17:21] ogasawara, wfm [18:00] where do i get me a vmlinux for perf nowadays, last time i had to do it it was ugly and it didn't end up in /boot and stuff [18:04] apw, ogasawara should 3.8.0-4 have a linux-tools package? I see the latest one available is linux-tools-3.8.0-2 . bug 1115047 was opened for this [18:04] Launchpad bug 1115047 in linux (Ubuntu) "perf tool missing for Linux 3.8.0 (raring)" [Medium,Confirmed] https://launchpad.net/bugs/1115047 [18:05] jsalisbury, i think i would expect us to be able to make one, it was off because it needed libraries from universe, but i thought tim fixed that [18:06] apw, Actually, I think I see it here: https://launchpad.net/ubuntu/raring/amd64/linux-tools-3.8.0-4 [18:07] jsalisbury, indeed we do seem to be building one [18:07] apw, and it seems accesible from https://launchpad.net/ubuntu/raring/+source/linux [18:07] commit fdfb1fd6dccabf8ff4f4979d01d4f47b47fc5994 [18:07] Author: Tim Gardner [18:07] Date: Thu Jan 10 12:41:32 2013 -0700 [18:07] UBUNTU: [packaging] Add macro to selectively disable building perf [18:07] [18:07] Fixes FTBS until libaudit-dev is promoted to main. [18:07] [18:07] so the report must be wrong now at least [18:07] Signed-off-by: Tim Gardner [18:08] ogasawara, i think the right thing to do is not MIR anything, but to pull tools out of the main kernel package [18:08] ogasawara, i did start looking at it, or more specifically looking at the version in debian where they do the same thing [18:08] and i think it is probabally viable, at least then it can link to anything [18:09] apw: makes sense to me to separate it [18:09] jsalisbury, ok so we have a package, but no perf in it, so the bug is technically correct [18:09] apw, ok, thanks [18:09] jsalisbury, can you assign that bug to me, and i'll try and make a proper determination tommorrow [18:09] jsalisbury, as to whether we can just rip it out [18:10] apw, will do. [18:10] ogasawara, apw, thanks! [18:45] bjf: wasnt attached to this session, turning the crank due to regression? [18:46] vanhoof, all good here. we are turning the crank on everything this week. [18:46] wont change 12.04.2 will it? [18:46] nope [18:46] gotcha [19:07] diwic, i have a n7 wh [19:07] diwic, which is refusing to make noises [19:08] apw, did you see today's update to bug 1068804 ? [19:08] Launchpad bug 1068804 in ubuntu-nexus7 "sound only works after suspend/resume cycle" [Critical,In progress] https://launchpad.net/bugs/1068804 [19:08] diwic, that is the kernel i am running [19:09] and it doesn't work before or after a s/r now [19:09] apw, so you're running a kernel with the patch from comment 31? [19:10] diwic, yes, that is why i am whining at you [19:10] diwic, works fine for me ... [19:11] but andy seems to suddenly have sound issues (though that looks more like userspace) [19:11] 1/2 isn't a good balance [19:11] so ogra and apw have both independently compiled their own kernels with the patch from comment #31, with completely opposite results? [19:11] ogra, can i have your asounds.state ... [19:11] diwic, no i made the kernel for both of us [19:11] diwic, nah [19:11] i was lazy [19:12] http://paste.ubuntu.com/1610080/ [19:12] how about the alsa-store.conf override/disable, are both apw and ogra running that? [19:12] apw, ^^ [19:12] diwic, no idea what that is, so assume not [19:12] diwic, oh, i removed the override files :P [19:13] apw, whats the version of your ubuntu-defaults-nexus7 package [19:13] 0.50 [19:13] that doesnt have the override files [19:14] so we should be on the same page [19:14] ls -l /etc/init/*.override [19:14] just to be sure [19:14] well, if the asound.state has become wrong it will be kept wrong without the overrides, that's a feature [19:15] yeah [19:15] ls: cannot access /etc/init/*.override: No such file or directory [19:15] great [19:15] thats what we want [19:16] apw, let's have a look at your alsa info [19:16] why don't we ship that damn script with the machine? [19:16] apw, /usr/share/alsa-base/alsa-info.sh [19:16] apw, because the driver needs to DTRT [19:16] IIRC [19:17] apw, the script is generated on first boot ... thats like asking why we dont ship 1G of different initrds on the CD for all possible machines [19:17] :) [19:17] are we talking about the alsainfo script or the asound.state script? [19:17] the state file [19:18] diwic, i was talking about the one you showed me _was_ installed :) [19:18] http://paste.ubuntu.com/1610099/ [19:18] diwic, ^^ [19:18] apw, your speaker is muted [19:19] great, its a shame none of the UI elements tell em that [19:19] apw, try muting and unmuting and see if it helps [19:20] diwic, ok my speaker is no longer muted, but its still mut [19:20] mute [19:20] ok [19:21] I've seen this happen a couple of times actually; the speaker alsamixer control spontaneously mutes itself. Probably some powersave feature or something [19:23] diwic, damn this thing has a lot of mutes, any idea which of the 100s to try next? [19:23] int spk [19:24] that is already 00 [19:24] apw, don't try manipulating the 100s of mutiee [19:24] mutes, it's more likely to screw something up. [19:25] oh any i twiddle i am putting back [19:25] not that i have tried many [19:25] apw, instead try: delete asound.state and move away /etc/init/alsa-store.conf [19:25] you can cause pretty nasty results with that [19:25] apw, then power off the nexus7, wait a little while and put it back on. [19:25] * ogra remembers a full day with a squeeking n7 next to him until he got it back to a sane state [19:26] hopefully that brings it into a sane state [19:26] and you couldnt adjust the volume ... and it started when the driver was initialized [19:27] diwic, ok will try that [19:27] ogra, what do you mean "couldnt adjust the volume"? [19:27] ogra, I mean, if it's muted it's muted [19:27] diwic, that was before you entered the project :) i played randomly with the mixer and got into such a state [19:28] not even muting solved it [19:28] ogra, aha, ok [19:28] the driver was just caught in a feedbackish squeek [19:28] i probably routed some random stuff to some other random stuff that should never have been connected :) [19:29] diwic, ok now i have the volume plips, and the 'alert sound' previews back, but nothing for 'test sound' [19:29] just like hrw, i just didnt fry my speakers ;) [19:29] ogra, maybe we should hide more of those controls, all that we don't know what they do [19:29] apw, try pressing 'test sound' a few times [19:29] yeah, but thats extra work [19:30] theoretically people shouldnt touch alsamixer if pulkse works [19:30] or they should know what they do ... [19:30] diwic, nothing obvious, i get the 'tink' to say i've hit Test, but no actual speaking bit [19:31] apw, yeah, for some reason the 'test sound' is sometimes a 'blupp' and sometimes 'front left'. Both count as success. [19:31] diwic, really, thats ... not obvious [19:32] apw, I haven't investigated them. [19:32] totem can't play an mp3 either [19:33] even though it did install *-bad first time [19:33] apw, btw, speaking of funny alsamixer controls. Did you know what the one I disabled did? It was no regular control. It was a debug feature to set/get individual codec registers. Left channel was the codec register address, right channel was the data to get/set. [19:33] heh now that is, er nice [19:33] apw, you can imagine what happens when alsactl tries to save/restore that kind of stuff. [19:34] it is not going to end well for sure [19:35] diwic, neither can rhythmbox even though it knows what it is [19:35] (especially as codec register address 0 was the default, and that likely causes some kind of codec reset. Not that what it does is really documented, just a guess.) [19:35] so whatever those apps use is not connected to anywhere [19:36] apw, does rythmbox show up under applications in sound settings when it's playing something? [19:36] yep, and full volumn [19:36] but no sound [19:36] ? [19:37] that is correct [19:37] apw, while it plays back, try the test speaker feature again in sound settings, is it still working? [19:38] diwic, nope [19:38] apw, can you go into alsamixer and see if "speaker" has muted itself again? [19:38] diwic, alertsound still works [19:39] diwic, as does plip on volume change [19:39] apw, so you have volume change 'plip' but not test speaker 'blupp'? [19:39] correct [19:39] now that's a combination I've never seen before. [19:40] diwic, speaker not muted [19:42] diwic, i can only assume it is s/w somehow [19:42] apw, 'pactl list' output? [19:42] 'pacmd list' [19:43] http://paste.ubuntu.com/1610165/ [19:43] http://paste.ubuntu.com/1610166/ [19:45] ok since running those two, if i start playback again i get like a nice chainsaw [19:46] which is volume controlled [19:47] apw, yeah, it looks like it's constantly reconnecting to pulseaudio or something. Or that's my guess, given that rythmbox is "sink input #82" and I doubt that you've started 81 other clients before [19:47] but ogra has working rhythmbox? [19:47] i havent tried rb [19:48] i played a webm video, tried the test sound and have the plop from the volume up/down here [19:49] client: 27 [19:49] diwic, ^ [19:49] diwic, playing the HBR1 ambient station from RB now [19:49] index: 111 [19:49] works fine [19:52] diwic, and quitting and restarting rb has gone back to silence [19:54] apw, your kernel is working for me [19:54] ogra, maybe we should add more alsamixer controls, put them first and call them "WARNING" and "DO NOT TAMPER" to avoid more people frying their speakers :-) [19:54] haha [19:55] diwic, ahhh got it, i had to install -ugly [19:55] diwic, when you try the first time it tells you to install -bad [19:55] after you do that, it never says another thing [19:56] apw, now what do we think of RB doing that? Is it bad, or just ugly? :-) [19:56] but without, it is just silent [19:56] diwic, well it doing something dumb is ... ugly [19:56] diwic, thanks ... and bjf thanks as well [19:56] apw, anyway, you got successful playback of mp3 files now? [19:56] diwic, i do thanks, once you fixed my plips i was fixed [19:58] yw [19:58] apw, it seems to always come up "mute" when i reboot [19:58] bjf, i thought we had a saver for that. i'll give it a go next [19:58] bjf, is that with or without the alsa-store.conf override? [19:59] diwic, i've not done anything special other than install apw's kernel [19:59] sudo rm /etc/init/alsa*.override [19:59] ls /etc/init.d/*.override [19:59] yes, those exist [19:59] remove them [20:00] they need to go away once the kernel is in [20:00] interesting, so then they do have a purpose [20:00] I was thinking pulseaudio was enough with storing/restoring volume [20:00] it should have been [20:01] it seems to store the volume here [20:01] but still comes up muted [20:01] it is at full volume if i had it there and unmute after a reboot [20:01] maybe it has something to do with the spontaneous mutes of alsamixer controls, haven't really figured out why that happens and what to do about it [20:02] ogra, indeed, that seems to have "fixed" it. i rebooted and it came up in the same state (un-muted) [20:02] so the only bit left is the muting [20:03] apw, bjf so it looks like my patch actually fixes the no-sound-before-s3 problem, can I trust one of you to apply it to the official nexus7 kernel? [20:04] diwic, yep i have applied it already, i'll be uploading it once i am happy the other fixes are good [20:04] ogra, were you happy with dmesg [20:04] apw, thanks [20:05] apw, absolutely [20:05] ok, got to do some laundry now, happy evening everyone [20:05] you too, and thanks ! [20:06] :-) [20:06] bye! [21:36] bjf: I've taken a first pass at https://wiki.ubuntu.com/PrecisePangolin/ReleaseNotes/UbuntuDesktop [21:37] bjf: specifically the "LTS Hardware Enablement Stack" and "Ubuntu Kernel 3.5.0-23.35" sections [21:37] bjf: I still need to scrub the known kernel issues section [21:38] bjf: but let me know what additional edits I should make or info I should add === henrix is now known as henrix_ [22:01] ogasawara, ack [23:12] can anyone please suggest how I should check whether a particular LP bugfix has been applied to the mainline kernel? [23:13] vk1266, which LP bugfix are you referring to? [23:14] LP: #1040557 [23:14] Launchpad bug 1040557 in linux (Ubuntu) "UEFI boot live-usb bricks SAMSUNG 530U3C,np700z5c laptop" [Critical,Fix committed] https://launchpad.net/bugs/1040557 [23:16] vk1266, i can tell you that bugfix (or a variant of it) is in Linus' tree and was part of 3.8-rc6 [23:17] so I guess I am reasonably safe to try the mainline kernel on my Samsung, right? [23:17] vk1266, you are building the kernel from the sources yourself? [23:18] no, I need to install deb's to verify if another bug still persists in the mainline kernel [23:20] vk1266, are they ubuntu debs? [23:20] yes [23:21] vk1266, so for that bug you can see that the fix was released for oneiric, precise and quantal. so if you are installing a recent kernel from one of them you should be fine. [23:22] bjf - perfect, thank you, that's what I needed to know. Now I am off to installing the mainline kernel for my Samsung and checking that other bug! [23:22] vk1266, wait! [23:22] yes... [23:23] vk1266, there is a difference between an ubuntu kernel and a "mainline" kernel [23:23] vk1266, which "mainline" kernel are you installing? where did it come from? [23:24] http://kernel.ubuntu.com/~kernel-ppa/mainline [23:24] vk1266, which mainline kernel are you going to install? [23:24] vk1266, that patch just went into mainline late last week [23:26] vk1266, if you are installing the v3.8-rc6-raring you should be ok [23:27] vk1266, is there a specific bug that you are testing that kernel for? [23:28] I have quantal; I guess I cannot try v3.7-rc2, it is dated 20-Oct-2012 - must stick with 3.5.7.4 [23:28] I need to run the test for LP: #1114856 [23:28] Launchpad bug 1114856 in linux (Ubuntu) "power button not working when Samsung laptop is booted in CSM mode; works fine in EFI mode" [Medium,Confirmed] https://launchpad.net/bugs/1114856 [23:29] vk1266, ok, hold off while i look at that [23:30] Considering that 3.8 kernel is not yet available to my quantal installation, and that my current kernel is 3.5.0-23.35, I am wondering how much value will be in checking out 3.5.7.4 [23:31] vk1266, i would not test a mainline kernel for that case. i'm going to add a comment to that effect. [23:31] jsalisbury, still around? [23:32] thank you bjf! maybe I should try installing raring on my Samsung, but for just checking one bug it may be a little too much [23:32] vk1266, if you can wait, there will be a new Quantal kernel in -proposed by next week that might help [23:33] sure I can -- if the LP bug waits for me [23:34] vk1266, i think there is a good chance it will still be around [23:35] bjf, thank you - I will keep checking -proposed [23:37] bjf, I am going to make a note in the LP bug to that extent [23:45] vk1266, sounds good [23:47] bjf, I have posted a note to LP: #1114856 - thank you, I will now be checking -proposed and the mainline kernel ppa [23:47] Launchpad bug 1114856 in linux (Ubuntu) "power button not working when Samsung laptop is booted in CSM mode; works fine in EFI mode" [Medium,Confirmed] https://launchpad.net/bugs/1114856 [23:47] vk1266, ack, thanks [23:56] bjf, hey [23:56] jsalisbury, hey, if you read the scrollback you get some indication of what i was going to mention [23:57] bjf, ok