[04:05] <sconklin> http://www.engadget.com/2010/09/18/intel-wants-to-charge-50-to-unlock-stuff-your-cpu-can-already-d/
[07:44] <bullgard4> At startup my Maverick Beta computer writes after 1,8 s : "Magic number: 10:86:987". What does this number refer to?
[11:44] <bullgard4> At startup my Maverick Beta computer writes after 1,8 s : "Magic number: 10:86:987". What does this number refer to?
[11:52] <bullgard4> I understand that it designates USERHASH:FILEHASH:DEVHASH. But why does Ubuntu publish these 3 numbers?
[19:46] <u456503> Hi all
[19:48] <u456503> Is somebody here ?
[19:50] <u456503> I'm trying to compile the kernel using this tutorial: https://wiki.ubuntu.com/KernelTeam/GitKernelBuild
[19:50] <u456503> I'm using Ubuntu 10.10
[19:53] <u456503> but there are some files *.patch to be applied
[19:54] <u456503> can some one help ?
[20:01] <u456503> hi oracle
[20:02] <oracle> hi 
[20:02] <oracle> so many kernel updates recently
[20:03] <u456503> stable candidates, yes
[20:04] <u456503> Do you run an Ubuntu ststem ?
[20:06] <u456503> http://twitter.com/gregkh/status/24873562442 from Greg
[20:07] <oracle> yes i run two ubuntus
[20:08] <oracle> 2.6.27
[20:08] <oracle> weird
[20:11] <u456503> oh, can you explain how to apply the ubuntu patches, on the latest mainile ?
[20:12] <u456503> the links: http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.36-rc4-maverick/
[20:13] <u456503> and: https://wiki.ubuntu.com/KernelTeam/GitKernelBuild
[20:13] <u456503> I am not an expert, if you can help ?
[20:17] <u456503> hi chrisccoulson
[20:17] <chrisccoulson> hi u456503
[20:18] <u456503> can you view the last discution ?
[20:19] <u456503> I'm trying to compile the latest kernel
[20:21] <u456503> but there are some patces, not included in mainline
[20:22] <u456503> there is somebody to upstream them ?
[20:22] <u456503> or I am on the wrong channel ?
[20:45] <u456503> in 5 minutes I will quit
[21:35] <etheretic> 'ello
[21:39] <etheretic> running iotop, get error "CONFIG_TASK_DELAY_ACCT not enabled in kernel, cannot determine SWAPIN and IO %" making it impossible to analyze the iowait pest i'm afflicted with. ubuntu 10.04 kernel 2.6.32-4.
[21:40] <lifeless> ttps://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/532490
[21:40] <etheretic> bug or feature?
[21:40] <etheretic> ah.
[21:40] <etheretic> h
[21:41] <lifeless> its going to be reenabled according to the master bug
[21:41] <lifeless> perhaps you need a newer kernel ?
[21:42] <etheretic> er. have one handy? not very competent w. linux, btw.
[21:42] <lifeless> no, I don't sorry :)
[21:43] <etheretic> thought 32-24 was latest. pardon above typo.
[21:43] <etheretic> 2.6.32-24
[21:47] <etheretic> lifeless: i have rtfm'ed on the iowait issue, but found no applicable solutions. it arises when running firefox/oolite/amule, and causes said programs to crash after having paralysed my comp (acer notebook) for 2-30 minutes.
[21:47] <etheretic> not sure how kernel-related it is. but #ubuntu was mum on the issue.
[21:52] <lifeless> I suggest reading up on disk tuning - DMA etc. Could be that.
[21:54] <etheretic> hdparm?
[21:58] <etheretic> timings look ok.
[22:01] <etheretic> 580/46 cached/buffered. nothing spectacular, but not abysmal either.
[22:20] <sfw> Does anyone know how does kernel module loading in Ubuntu work? Is it any different from Gentoo? This is NOT a troll question or something. For example I try to switch cpu into ondemand mode and it explicitly requires I modprobe "powernow-k8". How does Ubuntu handle this? Im just learning a bit.
[22:23] <mjg59> sfw: Which requires it? Ubuntu or Gentoo?
[22:24] <mjg59> There's no modaliases for CPUs, so it's always necessary to explicitly load cpufreq modules
[22:24] <mjg59> Ubuntu (and most sane distributions) have a startup script that looks at your cpu vendor, family and flags and uses that to figure out which module to load
[22:25] <sfw> mjg59: So ubuntu does not use kerneld and uses kernel autoloader? What startupscript are you refering for please, init in initramfs?
[22:26] <mjg59> sfw: kerneld?
[22:26] <mjg59> It's been a while since I checked exactly how it works in Ubuntu - I /believe/ that it's called /etc/init.d/powernow.something
[22:26] <sfw> I remember pretty good I wasnt required to modprobe powernow with 10.04. It was already somehow loaded.
[22:27] <sfw> Kerneld is from here: http://tldp.org/HOWTO/Kerneld/introduction.html    And I got info about it from here:http://tldp.org/HOWTO/Module-HOWTO/index.html
[22:29] <Nafallo> likely /etc/init.d/ondemand I think
[22:29] <sfw> mjg59, so ubuntu probably uses init.d to automatically load the module. Hm. I just wonder if a program can force kernel module be loaded automatically. In the case of devices its done by udev AFAIK
[22:29] <Nafallo> hmm. no. that doesn't actually load modules
[22:29] <sfw> I mean it detects devices and issues modload.
[22:31] <mjg59> sfw: kerneld doesn't exist in modern kernels
[22:31] <Nafallo> ah. looks like the powernow stuff is compiled into the kernel, and then /etc/init.d/ondemand sets the ondemand for cpus that supports it.
[22:31] <sfw> mjg59: this is what I was fearing)
[22:32] <mjg59> sfw: In the modern world, when a device is created it generates a uevent. That gets picked up by udev, which reads the modalias file and asks modprobe to load any modules with a modalias that matches the device's. CPUs are system devices and don't have modaliases.
[22:32] <sfw> Nafallo: Ok, that pretty clears things up, ty
[22:34] <sfw> mjg59: Thanks a lot for explanation! So software usually should not load or ask to load any kernel modules?
[23:53] <melkor> are there any maverick kernels that have fixed this issue?  http://www.h-online.com/open/news/item/Hole-in-Linux-kernel-provides-root-rights-1081317.html
[23:53] <melkor> I'm using the lts maverick backports and it is not fixed.