[03:40] <h00k> If I am trying to load a module on boot by adding it to /etc/modules (netconsole, troubleshooting a panic on 12.04), and it fails to load, is that an issue with that module or with the kernel?
[03:41] <h00k> However, if I load it manually (sudo modprobe netconsole), it ignores the /etc/modprobe.d/netconsole options I have, so I'd have to load all options with the 'sudo modprobe netconsole...etc' line.
[03:44] <h00k> I don't know enough to know whether it's with the kernel itself, or with the module :(
[05:04] <infinity> h00k: Try /etc/modprobe.d/netconsole.conf
[05:04] <infinity> h00k: If I recall, modprobe ignores certain filenames.
[05:04] <infinity> h00k: (It might tell you that with -v)
[07:09] <ppisati> back in 20mins
[07:23] <smb> Daviey, If your hammering had any outcome I should know about, let it be known to me... ;)
[07:43]  * ppisati needs a boatload of coffee...
[07:46]  * smb looks for a suitable vessel
[07:46] <apw> yawn
[07:47] <smb> Seems that needs to be a cruise around Europe...
[07:48] <apw> kernel coffee yaught sounds like a plan to me
[08:28] <Daviey> smb: wilco
[09:19] <cooloney> morning guys
[09:19] <cooloney> cking: i've start ceph on my local host now. thanks for your help
[09:19] <cking> cooloney, yay \o/
[09:19] <cooloney> cking: one thing is weird. i need to run 'sudo service ceph -a start' to start ceph
[09:19] <cooloney> w/o '-a' it won't start ceph
[09:20] <cking> cooloney, hrm, thats unexpected
[09:20] <cooloney> i think -a == --allhost
[09:20] <cooloney> and $ ceph health
[09:20] <cking> I've not had to do that, it just worked for me on multi-hosted and single hosted configs
[09:20] <cooloney> HEALTH_WARN 2 near full osd(s)
[09:21] <cooloney> and i got 2 WARNs, heh
[09:22] <cking> well, I suspect the default location for the OSD data is filling up - the default locations are a not sane IMHO
[09:22] <cooloney> i got an 1G jounal file in osd directory
[09:22] <cking> cooloney, that sounds sane to me
[09:23] <cooloney> cking: wow, so clever, my rootfs is about out-of-space
[09:23] <cking> cooloney, :-/
[09:48] <henrix> smb: from your last comment in bug #999755 i see you were not able to reproduce it in oneiric
[09:48] <ubot2> Launchpad bug 999755 in linux "Kernel crash in rb_next doing ohai loops" [Undecided,In progress] https://launchpad.net/bugs/999755
[09:48] <henrix> smb: i spend some time trying to trigger it (both in VM and bare metal), without any luck
[09:48] <smb> henrix, correct, still running without a crash
[09:49] <henrix> smb: ok
[09:49] <smb> henrix, What I'd probably do is to upgrade the kernel and make sure it runs as long with the new one
[09:50] <henrix> smb: ack
[09:51] <henrix> but what if we fail to reproduce it? would you be comfortable to tag it as verified (as the patch has been accepted upstreams already)? or would you rather revert it on oneiric?
[09:52] <smb> henrix, As I tried to say in the bug comment. For Oneiric go on and accept it verified. For natty, stop worrying ...
[09:53] <henrix> smb: yep, that's what i understood from your comment. just wanted to make sure :)
[09:53] <henrix> smb: thanks Stefan
[09:53] <smb> :)
[09:58] <smb> apw, There would be the question for Quantal. There will not be a crash but (probably a very theoretical) race. I am not sure about it ever make it into a stable later than 3.2 for that reason. Should we still bother to get it in there or just move on and pretend we did not see anything (which is actually true)
[10:26] <apw> smb, thats a tricky one without knowing the nature of the race.  i will trust you :)
[10:30] <smb> apw, Actually its even tricky when knowing it. But I believe the worst case is that a task is moved over and placed into the wrong run queue for some time. I guess we can live with that every now and then (which may be as often as any leap seconds). 
[10:33] <apw> yeah as long as smoke isn't coming out when it happens
[11:21]  * ppisati -> lunch
[11:43]  * apw goes splish
[12:21] <rtg> ogra_, apw, can you remember who was doing the ac100 kernel ?
[12:24] <cooloney> rtg: i think jani is working on ac100 kernel.
[12:24] <rtg> cooloney, ack, thanks
[12:26] <ppisati> rtg: he is just packaging it
[12:27] <ppisati> rtg: the guy porting patches&c is marvin24
[12:27] <ppisati> rtg: a guy from community
[12:27] <rtg> ppisati, well, wrt to the mailing list query, redirecting to Jani is sufficient.
[13:36] <josepht> jsalisbury: the test kernel for bug #1029547 works for me, commented in the bug.
[13:36] <ubot2> Launchpad bug 1029547 in linux "8086:0091 WARNING: at /home/apw/COD/linux/drivers/net/wireless/iwlwifi/iwl-trans.h:565 iwlagn_mac_flush+0x1a3/0x1b0 [iwlwifi]()" [Medium,In progress] https://launchpad.net/bugs/1029547
[13:37] <jsalisbury> josepht, thanks for the update.  I'll take a look at that commit and see what the next step should be
[14:00] <ppisati> cool, i received an ssd for my desktop: is there a way to install P on it without requiring to boot an entire box with it attached?
[14:00] <ppisati> e.g. can i attach it to my running desktop, launch an "installer", select the ssd and let it install P on it?
[14:02] <smb> ppisati, Use kvm? 
[14:03] <ppisati> smb: uhm... qemu is an option indeed
[14:04] <ppisati> smb: but axctually the install is an application that runs on top of a live system
[14:04] <ppisati> smb: so i guess it's doable even on my desktop without any system emulation
[14:05] <smb> ppisati, Maybe. Though I thought you were asking about a simple solution. :)
[14:05] <ppisati> doh!
[14:06] <ppisati> ubiquity requires all the qt libs!
[14:06] <ppisati> and even some kde packages...
[14:06] <ppisati> kille me, now...
[14:06] <ppisati> ok, qemu...
[14:07] <apw> ppisati, can't you debootstrap the disk ?
[14:07] <ppisati> apw: i wonder if there're some particular steps (like /etc files&c) that are handled only via the installer
[14:08] <ppisati> but yes, deboostrap is another option
[14:08] <apw> i'd just attach it to a test box and install on it as a usb disk
[14:08] <apw> then switch it in when its done
[14:19] <bjf> apw, bug 1021174. can you add the appropriate verification?
[14:19] <ubot2> Launchpad bug 1021174 in linux "include all binary packages in module checks" [Medium,Fix committed] https://launchpad.net/bugs/1021174
[14:23] <rtg> ppisati, in the Linaro nano image, how do you get the dhcp client to run on boot ?
[14:24] <rtg> hack it into /etc/rc.local ?
[14:24] <ppisati> rtg: never used Linaro images, but i guess rc.local is an option
[14:24] <apw> bjf ack
[14:27] <ppisati> rtg: playing with the guru plug?
[14:28] <smb> Sounds a bit dodgy...
[14:32] <ogra_> rtg, pointing /etc/network/interfaces to dhcp doesnt work ?
[14:32] <rtg> ppisati, no, still the sabrelite stuff
[14:32] <ogra_> oh, nano ... thats likely special since tons of stuff get hacked out of the system to make it small
[14:32] <rtg> ogra_, ah, tahts likely it. doh!
[14:32] <ogra_> so it might or might not work the usual way
[14:33] <ogra_> but /e/n/i would be my first thing to try
[14:33] <rtg> ogra_, it should. the nano image comes with the ISC client installed
[14:33] <ogra_> yeah, then it should
[14:35] <apw> bjf, done
[14:35] <bjf> apw, thanks
[14:59]  * ogasawara back in 20
[15:53]  * herton -> lunch
[16:00]  * smb -> gone
[16:06] <Daviey> dammit.
[16:06]  * cking notes Daviey is not having a fun day
[16:10] <cking> rtg, i meant "Natty" instead of "Oneiric" on that eCryptfs patch, can you apply it to Natty?
[16:10] <rtg> cking, yep
[16:10] <cking> bit stupid of me, especially 'cos I tested it on Natty to make sure it worked OK
[16:17] <infinity> cking: Oh hey, I suck at bug mail.  Feel free to ping me on IRC if you do things like spin test kernels for me. ;)
[16:17] <infinity> cking: (re: 871891)
[16:20] <cking> infinity, yep, if you could try that kernel out, I'd appreciate it, I'm not sure if one of those testers has a sane set-up
[16:21] <ogra_> why do you assume infinity has one ?
[16:21] <cking> ogra_, 'cos it appears he filed the bug
[16:22] <cking> unless I'm totally mad
[16:22] <cking> which is possible
[16:25] <ppisati> "[PATCH 00/24] Introduce Xen support on ARM"
[16:25] <ppisati> "this patch series implements Xen support for ARMv7 with virtualization
[16:25] <ppisati> extensions."
[16:26] <ppisati> still A15 only i guess
[16:26] <infinity> cking: I did indeed file the original bug, yes.  Though, I think it's suffered bug bloat since, as someone else said that nohpet didn't work for them (and it does for me).
[16:26] <infinity> cking: Which, I assume, means two different S10-3s, internally, and two different bugs.
[16:26] <brendand> jsalisbury, how would i load samsung-firmware?
[16:26] <brendand> sorry, samsung-laptop
[16:26] <cking> infinity, does the kernel I provided w/o the nohpet work?
[16:27] <infinity> cking: I'll spin it up in a bit over here and let you know.
[16:27] <cking> ok, no rush
[16:27] <infinity> cking: Well, I'll be scientific and make sure that the current precise still sucks too.  I haven't run without the cmdline option since I filed the bug. :P
[16:28]  * cking notes infinity has an Atari logo on his LP page. perhaps cking should put the commodore chickenhead on his
[16:29] <infinity> cking: See, I used to own a lot more C= hardware than Atari, but I went with the less ugly logo. :P
[16:29] <infinity> cking: (I used to run a lot of both, mind you, when I was deeply involved in 68k porting)
[16:30] <cking> infinity, yep, I had a C64 and Atari 1040ST
[16:30] <brendand> jsalisbury, modprobe -l samsung-laptop shows something underneath kernel/drivers/platform/x86
[16:40] <jsalisbury> brendand, I believe it is samsung_laptop
[16:41]  * jsalisbury checks
[16:42] <brendand> jsalisbury, it's listed in /proc/modules. does that means it's loaded?
[16:42] <dileks> yupp, samsung_laptop
[16:43] <jsalisbury> brendand, you can run lsmod | grep samsung
[16:44] <jsalisbury> brendand, If not loaded, you can run: sudo modprobe samsung_laptop
[16:45] <brendand> jsalisbury, it's definitely loaded
[16:45] <brendand> jsalisbury, always has been
[16:46] <jsalisbury> brendand, did you have this bug in precise or earlier releases?
[16:47] <brendand> jsalisbury, i can check
[16:47] <jsalisbury> brendand, ok
[16:47] <brendand> jsalisbury, let's do this in the bug. put a comment asking me to test precise
[16:49] <jsalisbury> brendand, will do
[16:52] <smoser> ok. its that time again.
[16:52] <smoser> smoser's noob questions
[16:52] <smoser> i'm just curious. if i kexec from one kernel to another, does memory get zero'd ?
[16:53] <smoser> i know that in the kdump use case it clearly does not, but kdump explicitly has the second kernel only "knowing" about a small portion of the memory.
[16:54] <infinity> cking: So... I can't tell you if your fix fixes anything, because current precise (-27-) seems to work fine without nohpet.
[16:55] <cking> infinity, 'current' as in the one I gave you to test?
[16:57] <infinity> cking: No, you gave me a -29-
[16:57] <infinity> cking: I'm saying that I'm testing the -27- from updates, and it's fine.
[16:57] <infinity> cking: (Like I said, I haven't booted without nohpet since oneiric, so this could have magically been fixed ages ago)
[16:58] <cking> infinity, OK, well thanks for testing, this info makes it just a little more mysterious
[17:05] <infinity> cking: Okay, switching from unity-2d to 3d, on resume with the old kernel, my root window didn't get painted, so I couldn't log in.
[17:05] <infinity> cking: But that could have just been a random bug.
[17:05] <infinity> cking: (Rebooting into your test kernel, I didn't have that issue, though)
[17:11] <infinity> cking: Okay, just reproduced the root window painting bug with your kernel too, so that's unrelated, and probably just unity being crap.
[17:11] <cking> infinity, surely not!
[17:11] <infinity> cking: So, conclusion is that (for my hardware), your kernel changed nothing.
[17:11] <cking> infinity, OK - thanks for testing
[17:11] <infinity> The upshot is that my hardware seems to have been fixed somewhere between 3.0 and 3.2.whatever. :P
[17:12] <cking> looks like another random S3 bug to me
[17:12]  * ppisati -> gym
[17:13] <infinity> cking: I'd just close the bug, if it wasn't for the metooing on it.
[17:14]  * cking nods, well I will look at that other users bug and see what's busted and maybe then ask them to file a new bug
[17:18]  * infinity nods.
[17:19] <infinity> Another 10 resume cycles on the old kernel, and I'm satisfied this all works correctly now.
[17:19] <infinity> So, I dunno what fixed it, or when, but thanks? :P
[17:19] <cking> thanks to upstream
[17:19] <infinity> It's harder to buy upstream beer.
[17:20] <infinity> I guess that's a cost savings for me.
[17:20] <cking> heh
[17:21] <dileks> infinity: you know thorsten (mira)?
[17:21] <infinity> dileks: Nein.
[17:22] <dileks> as you were talking about m68k porting - mira did a happy revival for debian
[17:22] <dileks> ...or still doing
[17:24] <infinity> dileks: Oh, yeah, wouter and I were talking about that at Debconf.
[17:24] <infinity> dileks: I've moved on by now, to a different set of slow hardware that actually has a future (ARM).
[17:24] <dileks> hehe
[17:24]  * dileks had a500 a2000 and 2x a3000
[17:25]  * infinity had a 2000 and an unholy collection of 4000s.
[17:26] <infinity> Well, and a 1200 that I used to use to write music, but it never ran Debian.
[17:26] <dileks> 4000 was not stylish, come on :-) 4000t(ower) was
[17:26] <infinity> They were Ts, I was too lazy to add the character. :P
[17:26] <dileks> good choice
[17:36]  * rtg -> lunch
[18:15]  * cking saunters off
[18:40]  * ogasawara lunch
[19:47] <balloons> bjf, jsalisbury how's things going with the community kernel testing on your end?
[19:47] <balloons> from my perspective it seems like you are having some good interactions and feedback from everyone
[19:47] <bjf> balloons: is community kernel testing happening?
[19:47] <balloons> sorry, I mean the 12.10 on 12.04 userspace testing
[19:48] <balloons> I shouldn't confuse :-0
[19:48] <bjf> balloons: i don't really pay any attention to it
[19:49] <balloons> bjf, ahh, perhaps I'm mistaken then.. I thought it was mainly you and joe responding to the bugs raised
[19:49] <bjf> balloons: you confuse me with a bot that runs as my lp user :-)
[19:49] <balloons> bjf, ahaha
[19:50] <balloons> yes yes of course!
[19:51] <balloons> Well jsalisbury interactions might have some "bot" in them as well, but I know he's personally sent some messages
[19:51] <balloons> any feedback you have would be great.. positive or negative. I want to make sure things are going well from your perspective
[20:01]  * rtg -> EOD
[20:21] <pbuckley> how is 12.10 doing?
[20:24] <joshhunt_> I see that 3.2.0-29.46 is listed as proposed as of July 28. I was wondering if there is a release schedule for this?
[20:25] <joshhunt_> i believe it has all of the leap second fixes which i'd like to work into my current release
[20:29] <ogasawara> joshhunt_: https://wiki.ubuntu.com/QuantalQuetzal/ReleaseInterlock
[20:29] <ogasawara> joshhunt_: by that it should move to -updates next week
[20:30] <ogasawara> joshhunt_: but I've been told they might try to have it available earlier so that it can land in the 12.04.1 point release
[20:32] <bjf> joshhunt_: we are shooting for the .1 point release
[20:33] <bjf> joshhunt_: those commits are part of the reason i want it as the point release kernel
[22:16] <joshhunt_> ogasawara, bjf: thanks. i stepped away for a bit.
[23:23] <cheako> How come there are so many CVE bugs marked as new?
[23:27] <cheako> bug 1007089
[23:27] <ubot2> Launchpad bug 1007089 in linux "overlayfs alters /proc/self/exe link(s), making result a dead link." [Medium,Confirmed] https://launchpad.net/bugs/1007089
[23:31] <cheako> How can I re-open a bug that I closed? bug 381896
[23:31] <ubot2> Launchpad bug 381896 in linux "bad router advertisement lead to disabilitation of ipv6 privacy extension" [Medium,Fix released] https://launchpad.net/bugs/381896