[04:24] <JanC> hm, why did Ubuntu kernel move from 300 to 250 HZ ?
[04:25] <JanC> (I'd expect a change in the other direction...)
[16:13] <apw> JanC, which version was it on 300?
[16:13] <apw> thats an odd number for sure
[16:16] <JanC> I thought it was on 300?  or am I so wrong?  (and 300 is one of the default values)
[16:24]  * apw goes check
[16:25] <apw> seems to have been 250 since edgy as far as i can tell
[16:25] <apw> on generic and 100 on server
[17:06] <vadi2_> Where can I find the process for filing a bug related to the kernel? (kacpid using more cpu time than xorg and firefox combined)
[17:21] <amitk> vadi2_: does this help? https://help.ubuntu.com/community/ReportingBugs
[17:22] <vadi2_> I'll give that a try
[17:22] <vadi2_> hum
[17:22] <vadi2_> nothing happened when I did ubuntu-bug
[17:23] <vadi2_> (it asked to be run as root at first, did sudo after, and it just quit)
[17:59] <apw> vadi2_, that doesn't sound right
[17:59] <apw> the bug you mention sound familiar, what version are you running; cat /proc/version_signature
[18:00] <vadi2_> Ubuntu 2.6.27-11.27-generic
[18:00] <maco> is this a kernel problem or dhclient problem? when i switch APs within the same network / using the same ESSID, dhclient always gets no working leases. if i modprobe -r then modprobe iwlagn, dhclient works fine.
[18:02] <maco> er it says no offers received and no leases in the db
[18:03] <IntuitiveNipple> Does sudo iwlist <if> report both APs? I've seen issues where the 'old' weaker one is targeted rather than the one with the stronger signal.
[18:04] <maco> i dont know. ill have to check next time, i guess.
[18:05] <vadi2_> apw: need anything else from me?
[18:05] <IntuitiveNipple> I also had an issue where APs linked via WDS had lost their link and so when switching to the 'remote' AP the PC couldn't get a DHCP lease (since it would come from the 'primary')
[18:05] <maco> if iwlist is still seeing the old one (2 blocks away?) which conclusion would that count toward?
[18:06] <IntuitiveNipple> maco: Sounds like stale records - what does the signal strength report?
[18:07] <maco> well i'm connected to the 2nd one now. i just ran into this for the 3rd time i was switching buildings for classes and trying to reconnect to get back into #ubuntu-meeting and watch
[18:07] <IntuitiveNipple> I can get a signal from 1/2 mile away when I'm out in the fields, but I do have a rooftop high-gain omni-directional mast.
[18:07] <maco> so ive already unloaded and reloaded iwlagn
[18:08] <maco> there are 15 APs in this spot with the same ESSID
[18:08] <maco> they're supposed to let you move from one to the other without a blip
[18:08] <IntuitiveNipple> It sounds like some kind of affinity to the original AP, can't think quite how though.
[18:09] <IntuitiveNipple> Are any of them on the same channel?
[18:10] <IntuitiveNipple> What you really need is to do a site-survey to determine which APs (MACs) and what signal strengths the adapter is seeing at the problem location. Also, what happens if you get on the 'second' AP first, then move back to the 'first'? Is that transition successful?
[18:13] <maco_> welll there's today's panic
[18:15] <maco> IntuitiveNipple: they vary between channels 1, 6, and 11
[18:15] <maco> 7 on chan 1, 3 on chan 6, 4 on chan 11 right now
[18:16] <maco> the one i'm on right now is one of the chan 1 APs
[18:16] <IntuitiveNipple> maco: Are they positioned so as not to overlap on each where they share a channel?
[18:17] <maco> i doubt it, since i can see this many from one spot
[18:17] <maco> its some sort of cisco mesh thing the school does to try (and fail) to cover the entire campus
[18:19] <IntuitiveNipple> It's always a problem using b/g with the limited channels available.
[18:22] <maco> IntuitiveNipple: does that mean "tough cookies"?
[18:23] <IntuitiveNipple> maco: It shouldn't; the client should handle it provided the signal/ratio isn't too low
[18:23] <IntuitiveNipple> signal/noise ratio 
[19:07] <UnixOne> Any explanation? I get the error "kernel maps protect is an unknown key" on bootup. Linux SGC-Abydoss 2.6.28.7-ultimate #1 SMP Tue Mar 3 18:19:03 CET 2009 i686 GNU/Linux
[19:07] <UnixOne> Just compiled the latest stable kernel
[19:11] <IntuitiveNipple> UnixOne: Using an atheros network device?
[19:13] <UnixOne> IntuitiveNipple: I've no wlan. its a wired intel gigabit
[19:14] <IntuitiveNipple> You need to uncomment /etc/sysctl.conf "kernel.maps_protect=1" I think
[19:15] <IntuitiveNipple> See man 8 sysctl
[19:16] <UnixOne> IntuitiveNipple: cool, I will try that. Thank you^^ 
[19:16] <UnixOne> IntuitiveNipple: somehow everything is commented out in that file. is that normal?
[19:17] <IntuitiveNipple> Default settings 
[19:18] <IntuitiveNipple> something you've got installed is trying to check that key
[19:18] <UnixOne> IntuitiveNipple: http://rafb.net/p/Kqsr9F60.html I commented it out. Are the other settings ok?
[19:22] <IntuitiveNipple> Two things. maps_protect was removed by kernel commit 3bbfe05 2008-10-10 which came with v2.6.28-rc1
[19:22] <IntuitiveNipple> Second. The key should not have the "sys." prefix.
[19:23] <IntuitiveNipple> You must have some 'older' utility trying to make use of the key.
[19:23] <UnixOne> IntuitiveNipple: ? I compiled it with kernelcheck. 
[19:30] <UnixOne> will reboot. cu
[19:42] <UnixOne> IntuitiveNipple: All errors fixed I just disabled #kernel.maps_protect = 1 it was enabled in /etc/sysctl.d/10-process-security.conf
[19:44] <UnixOne> I don't understand how hardcore cracks @gentoo manage to speedup and hack their kernel that it reboot into a fully working system withing 7seconds..
[19:44] <UnixOne> my machine took about ~25secs 
[19:45] <UnixOne> which is still very fast. because I've a slim system with kde-core only
[19:47] <UnixOne> ah... parallel execution^^ 
[19:52] <mnemo> can anyone repro this bug? is it a kernel bug? --> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/336771
[19:52] <ubot3> Malone bug 336771 in linux "system locks up when running "strace gdmsetup"" [Undecided,New] 
[22:31] <maxb> "Latest kernel upload:" in topic is woefully out of date. Perhaps it should be removed outright?
[23:57] <maxb> Is there a helper program/script which git tags using the version from the changelog, with a leading Ubuntu- as per ubuntu kernel conventions?
[23:57] <maxb> I know about debcommit -r, but the leader doesn't seem to be customizable