/srv/irclogs.ubuntu.com/2014/07/11/#ubuntu-kernel.txt

=== ayan is now known as Guest70899
=== ikonia is now known as Guest70019
=== kgunn is now known as Guest2511
=== hyperair is now known as Guest84654
=== neunon is now known as Guest28427
=== msx is now known as Guest82103
=== Guest82103 is now known as msx
apwxevious, in normal operation no you should not need to add addition sync commands07:53
=== apb_ is now known as apb1963
=== Guest70019 is now known as ikonia
agmhenrix, I used your -proposed kernel12:23
agmit seems to fix the nfs issue12:23
agmhow do I appropriately report my success?12:23
henrixagm: is there a 'verification-needed-trusty' tag in that bug?12:24
agmyes12:24
agmI mean, do I just say it seemed to work and tag it12:24
agmor is more information required12:25
henrixagm: then, just replace that tag by 'verification-done-trusty' :)12:25
agmOk, I just made my launchpad.net account today12:25
agmthanks12:25
henrixagm: of course you can add more relevant info to the bug, that could eventually help other people testing12:25
agmok, I'll describe my setup12:25
henrixagm: and thanks for testing and reporting back!12:25
agmthanks for fixing it! I just saw the issue last night, googled what I was seeing, saw the bug, and saw your fix12:26
agmfast!12:26
henrixyeah, sometimes that approach works :)12:27
xnoxmy processes tell me that they no longer can initialise inotify12:36
xnoxdid i run out of watches/discriptors, or i have something rougue running on my system, or is kernel buggy? or how do i check that?12:36
=== psivaa is now known as psivaa-lunch
agmDo I add a tag just by appending that text to the bottom of a post?12:53
agmoh nevermind found it12:56
xeviousapw: Can you think of an operation that would require manually running sync?13:49
apwxevious, a full sync, no13:54
=== psivaa-lunch is now known as psivaa
awe_sforshee, ogasawara, can someone on the kernel team take a look at a WiFi adhoc driver bug for us?14:48
awe_https://bugs.launchpad.net/ubuntu/+source/linux-mako/+bug/133775314:48
ubot5Ubuntu bug 1337753 in wpa (Ubuntu) "[mako] can not create an ad-hoc network" [Medium,Incomplete]14:48
sforsheeawe_: ugh, if android doesn't use ibss then there's a good chance that it isn't well-tested in the driver14:58
awe_yup 14:58
sforsheeI'll try to take a peek at the code a little later today, but that driver makes my eyes bleed14:59
awe_sforshee, same problems with bq; although different driver15:02
awe_sforshee, this may be endemic of Android Wi-Fi drivers15:03
awe_that said, Wellark also claims that bq doesn't have "ap" support either15:03
awe_( which is what I think we should be using for hotspot )15:03
awe_sforshee, one last comment... in the syslog, I see reference to a qcom platform drive that NM thinks is the WiFi driver;  it's not clear to me if the actual N4 WiFi driver is a blob or not...15:05
sforsheeawe_: yeah, if android doesn't require ibss support then we probably need to assume that we can't rely on it when we're using those drivers15:09
sforsheeawe_: I don't think the n4 wifi driver is a binary blob, but I'll double check15:11
awe_sforshee, well...the driver claims support, and without, we're missing the underlying framework to support hotspots15:11
sforsheeawe_: but if android is using ap mode then that's the code that we know is tested and _should_ work, so why wouldn't we use that for hotspots instead?15:15
awe_so...we have existing code in Ubuntu to support hotspot, and it's tied to adhoc unfortunately.  I document why this isn't optimal in: https://bugs.launchpad.net/ubuntu/+source/linux-mako/+bug/133775315:19
ubot5Ubuntu bug 1337753 in wpa (Ubuntu) "[mako] can not create an ad-hoc network" [Medium,Incomplete]15:19
awe_sorry...https://bugs.launchpad.net/ubuntu/+source/linux-mako/+bug/1337753/comments/515:19
sforsheeawe_: okay ... so it's easier to go fix every android driver which has sub-par ibss support than to change our hotspot code?15:20
sforsheeand based on that comment, there are other compelling reasons to use infrastructure mode rather than ibss15:23
awe_I haven't yet had this discussion with others15:23
awe_NM apparently does support AP-mode hotspot, but the discussion hasn't been had yet15:24
sforsheeawe_: so as far as I can tell the options are a) continue with adhoc, create a lot of extra work, and have an overall crappy hotspot experience, or b) switch to infrastructure mode for hotspots and have fewer driver problems and a better user experience15:25
sforsheethe answer seems pretty obvious to me15:25
awe_sforshee, also not all of the drivers we're interested in support ap mode15:26
awe_sforshee, it depends on how much code is involved15:26
sforsheeawe_: then these same devices don't support hotspots under android I assume?15:26
awe_in theory, that's correct15:27
sforsheewhat about in practice?15:27
awe_;)15:27
sforsheebecause if android doesn't support hotspots on those devices it seems unreasonable to demand that we should15:27
awe_agreed15:27
awe_sforshee, just asked abeato for the 'iw list' output for krillin15:28
sforsheeawe_: so honestly I don't feel inclined to invest much effort at all in tring to fix adhoc on the n4 if it seems likely we'll just switch to ap mode anyway15:29
sforsheeand to me that seems like the far more reasonable solution, all things considered15:30
awe_sforshee, we can't make a decision till we've looked at all the pieces...  I'd like at least some opinion from you or someone from the kernel team as to why fixing the driver is too much effort15:32
awe_it might be something obvious like building the driver with p2p support might conflict with adhoc mode15:32
sforsheeawe_: okay, I'll take a look this afternoon15:33
awe_I *will* have the conversation re: adhoc mode today15:33
awe_thanks15:33
awe_don't spend a lot of time on it15:33
sforsheeI don't plan to ;-)15:33
awe_cool15:33
awe_thanks much!15:33
kdeuser56how would I get the ubuntu kernel config file for a specific version from commandline?17:11
infinitykdeuser56: If you mean for the installed/running kernel, it's /boot/config-$(uname -r)17:23
kdeuser56infinity: not an installed one17:23
kdeuser56infinity: I was looking for the .config of any kernel ubuntu ever played with17:24
kdeuser56infinity: is there some archive?17:24
infinityI don't think anyone's bothered to extract the configs from every binary package and archive them, no.17:24
infinityBut if you know which binary package you're interested in, you can download and unpack it.17:25
infinityOur git trees show the changes on a more granular level, but there's no .config in there, as it's broken into pieces and constructed at build time.17:25
=== kees_ is now known as kees
BLZbubbahi guys, I'm trying to build 3.15.4 using "fakeroot debian/rules binary-headers binary-generic" and it builds just fine18:42
BLZbubbahowever, the deb only inculdes about 600 modules whereas the binary on kernel.ubuntu.com has over 300018:43
BLZbubbai started looking at the generic inclusion list but it seems to match the one on the standard 3.13 kernel.  any idea where to look to make it include the correct set of modules in the kernel binary deb?18:46
SarvattBLZbubba: look in the linux-image-extra deb, if you want to build it all into one deb you'd want to fakeroot debian/rules do_extras_package=0 binary-headers binary-generic18:59
trippehalso might want to consider 3.15.519:03
BLZbubbabut there is no extras deb here: kernel.ubuntu.com/~kernel-ppa/mainline/v3.15.4-utopic/19:07
BLZbubbai'm just curious how to build the packages exactly the same way as these19:13
BLZbubbaSarvatt: you're right, extras had the missing ko files for useful things like the keyboard19:23
BLZbubbaso how are they doing it differently for the images published to kernel.ubuntu.com?19:25
=== cmagina_ is now known as cmagina
=== slangase` is now known as slangasek

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!