[12:01] <nessita> hello everyone. diwic, re: bug # 1201528, on Saturday I installed a fresh precise, and audio broke when I installed the raring kernel. Verbose pulseaudio log attached to the bug
[12:02] <diwic> nessita, yeah, I saw. So, it looks like the kernel is at fault here...
[12:02] <nessita> diwic, apparently...
[12:03] <nessita> diwic, I still have the installation on a separated partition, so if you need me to run more tests, just let me know
[12:06] <diwic> nessita, I guess rtg could set up a git bisect between 3.2 and 3.8 for you 
[12:06] <nessita> diwic, sounds good
[12:07] <nessita> diwic, do we have a newer kernel in saucy? I could try that too
[12:07] <diwic> nessita, saucy is 3.10.
[12:07] <nessita> diwic, would you recommend that path?
[12:08] <diwic> nessita, I think you tried saucy at one point and it was still broken, right?
[12:08] <nessita> diwic, no, at that point you considered it wasn't going to help
[12:08] <nessita> or it likely wasn't ging to help
[12:09] <diwic> nessita, I still think it's more likely it does not help, but feel free to try
[12:10] <nessita> diwic, I have no leftover space in my disk for a separated partition, so I would have to upgrade this raring to saucy. Would you know if the transition is "smooth"?
[12:10] <diwic> nessita, try saucy from a live USB stick.
[12:10] <nessita> or is there a package with saucy's kernel for raring?
[12:10] <nessita> ah, can do that
[12:10] <diwic> nessita, btw, is your computer home-built?
[12:11] <diwic> nessita, or does it have a common name (like "Dell inspiron D505")?
[12:11] <nessita> diwic, I bought the separated parts and I assembled it
[12:11] <nessita> so, I guess the answer is "yes"
[12:11] <diwic> nessita, okay, what graphics card is in there?
[12:12] <nessita> nvidia gforce (grabbing exact model)
[12:12] <nessita> 8400 GS
[12:13] <nessita> diwic, FYI, I did not install the propietary drivers in the precise installation
[12:13] <diwic> nessita, since it has been the graphics showing up on the wakeup_rt tracer reports - do you have another card around that you don't use, you could try switching and see what happens?
[12:13] <nessita> diwic, hum, no, I don't have, and this MB does not have integrated graphics
[12:13] <diwic> nessita, okay
[12:19] <diwic> nessita, so, a long boring git bisect, or possibly switching graphics cards are my best bets at the moment - but saucy could be worth a try too if it is easy for you to test
[12:20] <diwic> nessita, but I'm far from sure it's the graphics card at all so don't go buy a new one just for this :-)
[12:21] <nessita> diwic, right, I wasn't planning on getting a new video card... is hard for me to understand how video can mess up with audio like this. Specially since this computer has "a lot" of cores, so I would expect audio and video to not mess each other
[12:22] <nessita> diwic, will do the sacuy test. I'm also happy to do the git bisect, but I'd need instructions/packages for that
[12:22] <diwic> nessita, yeah, I too think that is surprising. I'm not an expert on scheduling. And I'm not ruling out audio either, it's just that that biggest weirdness are these latencies.
[12:23] <ohsix> oh my god, i can't even believe bfq and deadline came up in discussion as a reasonable thing to do
[12:24] <ohsix> bug #1201528
[12:24] <ubot2`> Launchpad bug 1201528 in alsa-driver (Ubuntu) "[Realtek ALC889] - Audio Playback Unavailable" [Undecided,New] https://launchpad.net/bugs/1201528
[12:27] <ohsix> heh oh man, raymond is on launchpad now ;O
[13:07] <ohsix> diwic: doesn't have to be a scheduling problem, but that is indeed weird stuff
[13:08] <ohsix> i'd be interested to see if some of the power state latencies and stuff were correctly reported and if it's entering them
[13:14] <rtg> apw, I updated overlayfs. looks like it builds OK. maybe you could do your testing on it ?
[13:15] <apw> rtg, is that post what i build this morning ... assume so
[13:15] <rtg> apw, just now. updated to v19
[13:15] <apw> rtg, ok
[13:42] <diwic> nessita, could you attach the output of /proc/interrupts ? It does not have to be any specific kernel or release.
[13:45] <nessita> diwic, sure, will grab them from this session. Do I need to "break" audio first?
[13:46] <diwic> nessita, does not matter.
[13:47] <nessita> ack, attached https://bugs.launchpad.net/ubuntu/+source/alsa-driver/+bug/1201528/comments/41
[13:47] <ubot2`> Ubuntu bug 1201528 in alsa-driver (Ubuntu) "[Realtek ALC889] - Audio Playback Unavailable" [Undecided,New]
[13:48] <nessita> diwic, properly formatted https://pastebin.canonical.com/95432/
[13:48] <ohsix> 45 is a busy one
[13:49] <nessita> FYI, audio playback is still working (haven't opened mumble yet)
[13:50] <diwic> nessita, just curious if nvidia and snd-hda-intel were on the same cpu but it seems like nvidia is spread across all CPUs.
[13:51] <nessita> and nvidia interrupts values seem "reasonable"
[13:51] <nessita> snd_hda_intel interrupts are crazy high
[13:52] <diwic> nessita, and this was raring, right?
[13:52] <nessita> diwic, yes, 13.04 with 3.8.0-26-generic
[13:52] <diwic> ok
[13:53] <nessita> that makes me remember... /me needs to download saucy iso
[14:01] <diwic> nessita, are you running with tsched=0 ?
[14:01] <nessita> diwic, not that I know of, let me check
[14:01] <diwic> nessita, if you are that could explain the large values on snd-hda-intel
[14:02] <nessita> diwic, can you please remind me where to look?
[14:02] <diwic> nessita, /etc/pulse/default.pa
[14:03] <nessita> diwic, full content of the file https://pastebin.canonical.com/95438/
[14:03] <nessita> i don't see tsched there
[14:04] <diwic> nessita, nope, looks like the origina
[14:04] <diwic> l
[14:04] <nessita> yeah, my recall was that I reverted all the changes I tried
[14:10] <rtg> apw, laid down some hacks on lttng in unstable. can you have a look to make sure what I did makes sense ?
[16:12] <brandon__> hello
[16:13] <brandon__> hi
[17:24] <nessita> ohsix, I see that diwic left, but just wanted to mention that booted a saucy live from pendrive, and auido playback broke as soon as I opened mumble, attached to the bug the pulse verbose log (with lots of underuns), and the /proc/interrupts from before