[01:03] <`sam`> where do i get the 2.6.21 kernel?
[01:04] <Nafallo> gutsy
[12:17] <Asad2005> I have a multi port serial card that is always detected well (all ports) in kernel 2.6.15 but other kernel only detects few of the ports, I want to know what makes kernel 2.6.15 different that i can add to other kernels as a module or a patch
[12:18] <fabbione> Asad2005: define other kernels?
[12:19] <Asad2005> fabbione, before 2.6.15 and later ones
[12:19] <cjwatson> pkl_: did you get anywhere with tracking down what was going wrong on 256MB systems (the apparent memory pressure thing)?
[12:19] <Asad2005> fabbione, But all are 2.6
[12:19] <cjwatson> I've had a report of a hardlock when starting the installer, and we have regularly reproducible hardlocks on PS3 under some circumstances
[12:19] <fabbione> Asad2005: from vanilla or from ubuntu?
[12:20] <cjwatson> the way that userspace memory use wasn't obviously going up in Scott's tests suggests that it could be a kernel memory leak?
[12:20] <Asad2005> fabbione, ubuntu kernels but i think its not ubuntu its the linux kernel that have this different
[12:21] <fabbione> Asad2005: i recall patching dapper specifically to init more than 4 ports
[12:21] <fabbione> Asad2005: there is either a module option or a value you need to change in the code to init more
[12:21] <fabbione> Asad2005: before .15 the default was 48
[12:21] <fabbione> but it sucks memory.. 
[12:21] <Asad2005> fabbione, i tried at least one other debian based distro with 2.6.15 and worked fine
[12:21] <fabbione> and it might be that edgy or feisty lost the patch
[12:22] <fabbione> .15 is ok... you are telling me that others are not.. so it might be the patch has been lost by mistake
[12:22] <fabbione> Asad2005: [  107.317850]  Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing enabled
[12:22] <fabbione> check this line in dmesg
[12:23] <fabbione> 4 ports is the pre-allocated number
[12:23] <fabbione> you will see in .15 it's 48
[12:24] <Asad2005> fabbione, where can i get this patch and how to apply it to edgy/feisty
[12:24] <fabbione> Asad2005: it's in git tree on kernel.org
[12:26] <Asad2005> fabbione, I am a newbi so could you help more, is it a module what modprobe name to use to add it or if it is a patch how can i apply it
[12:27] <fabbione> Asad2005: sorry i don't have time to coach on kernel hacking. The patch is in git.. there are a lot of tutorials around you can use for that. including wiki.ubuntu.com/Kernel*
[12:39] <pkl_> cjwatson: I was in the middle of doing the preparatory work (get squashfs dumps and writing the analysis tool), but BenC wanted me to concentrate on getting a Feisty SRU out.
[12:40] <pkl_> cjwatson: SO, I've done nothing since last Tuesday on it.  But it is definately something I want to do when I get time.  Probably after the SRU and UDS.
[12:41] <cjwatson> ok
[12:41] <cjwatson> if it turns out to be relatively simple, I'd like to have it in the first feisty point release
[12:41] <cjwatson> (since it prevents installation)
[12:41] <cjwatson> but I agree the first SRU is more important than what may be a wild goose chase ...
[12:43] <pkl_> cjwatson: yes, I'd like to at least track the cause down.  Once we know what is happening, decisions can be made as its seriousness.  What worries me at the moment is not knowing if it is serious or not.
[12:44] <cjwatson> indeed
[12:45] <cjwatson> and not knowing where it is and whether it might affect ordinary systems
[12:45] <cjwatson> I was wondering if it would be worth trying to mount a squashfs on a non-live system and play about with it there, in case it's easier to debug when the root filesystem isn't involved
[12:46] <cjwatson> (or if you've already tried that)
[12:48] <pkl_> I've done some tests on a non-live system, and got some Squashfs dumps.  From an initial analysis everything is working as expected WRT Squashfs and disk I/O.  But, this is a manual eyeball of the raw data, much more meaningful information will be available once I've written the analysis tool.
[12:49] <pkl_> My plan was to write the analysis tool, play around with non-live Squashfs filesystems to check the tool was giving useful information.  Later if, neceessary, generate liveCDs with the instrumented Squashfs module, and get dumps from there.
[01:09] <jordiR> Somebody knows what is this new package named linux-ubuntu-modules?
[01:13] <cjwatson> jordiR: you don't need to ask; just see the package description
[01:14] <cjwatson> and/or look at its contents
[01:15] <jordiR> cjwatson: the description only says that it contains modules supplied by Ubuntu
[01:15] <Nafallo> jordiR: there you go then :-)
[01:15] <cjwatson> the package contents will make it pretty clear which modules
[01:16] <jordiR> and I want to know what contains before install it, that is why I ask
[01:17] <cjwatson> it's just split out from the old linux-image packages
[01:17] <cjwatson> it's nothing you don't already have if you're running Ubuntu
[01:17] <jordiR> can I see package contents without install it?
[01:17] <cjwatson> download the .deb by hand and use dpkg -c
[01:17] <jordiR> cjwatson: ok, thank you
[01:18] <Amaranth> if it's not in linux-restricted-modules and not in mainline it's in linux-ubuntu-modules
[01:18] <Amaranth> i think :)
[01:32] <zul> BenC: around how do you a git pull/merge?
[01:58] <cjwatson> jordiR: oh, also, see ubuntu-devel-announce, where the linux-ubuntu-modules package was described recently
[02:13] <jordiR> cjwatson: thanks, there is the explanation I was searching for.
[02:15] <Amaranth> "the following week is UDS where we'll all be drinking the night away in some spanish saloon, maybe dodging bulls in our spare time."
[02:15] <Amaranth> sounds fun, count me in ;)
[02:20] <Mithrandir> spare time at an UDS?
[02:20] <Amaranth> yeah well...
[02:20] <Amaranth> we can dream? :)
[02:48] <Asad2005> fabbione, I have found the git page it seems to have two files .c and .h how do i compile a module for the extra ports
[12:10] <jdachary> hi: looks like the vserver patch is not supported by ubuntu, or am I mistaken ?