[12:07] <BenC> later
[12:31] <zul> hey
[01:09] <BenC> yo
[01:14] <zul> how is it going
[01:16] <zul> http://www.youtube.com/watch?v=y6vhTSwoSA8
[01:17] <BenC> writing wiki pages isn't fun
[01:17] <zul> heh...im doing some grub stuff
[02:51] <zul> yay..new grub works
[09:03] <kimo> I am using 2.6.17-1.1, everything in official dapper kernel is fixed :) Problem, is waking from sus2ram => kernel oops. Should I report this ? where ?
[09:05] <crimsun_> sure.  [https://launchpad.net/distros/ubuntu/+source/linux-source-2.6.17/+filebug] 
[09:39] <kimo> Thanks, bug reported
[01:02] <kimo> anyone knows the difference bet a process's nice value & its priority value (as shown in top for example)!
[01:04] <infinity> Priority is the kernel assigned priority, which relates to the niceness, but not always in an obvious 1:1 way.
[01:05] <kimo> thanks ... any relation, or totally random!
[01:05] <kimo> depends on load ?
[01:18] <Keybuk> PRI = Nice + -20
[01:18] <Keybuk> iirc
[01:20] <fabbione> last i checked it was slightly different
[01:20] <kimo> if its that straightforward, then its redundant
[01:20] <fabbione> priority indicates the amount of CPU the process is allowed to get and it's calculated according to the amount of processes
[01:20] <fabbione> nice is a value that indicates how quickly a process can be deprioritized 
[01:21] <fabbione> (or prioritized according to the POV)
[01:21] <fabbione> so a process can be high priority because it's really CPU consuming
[01:21] <fabbione> but it can be nice 20 and gets deprioritized very fast if other process with same or higher priority get in the game
[01:21] <fabbione> see for example stuff like setiathome
[01:22] <fabbione> they are very cpu intesive (high priority) but they get trashed on the floor if anything else needs CPU (high nice level)
[01:23] <kimo> this way, the priority, is an internal (scheduler) thing, right. The user shouldnt & doesnt want to see it, and cant change it, right ?
[01:23] <Keybuk> users never need to use top
[01:24] <kimo> admins I mean
[01:24] <fabbione> kimo: assume you have a 1 sec sample of time
[01:24] <fabbione> n one sec a lot of things happen in the CPU
[01:24] <fabbione> tons of processes are running
[01:25] <fabbione> they need to get scheduled and for only a very precise amount of time
[01:25] <fabbione> the priority is used to calculate that time
[01:27] <fabbione> right
[01:28] <kimo> cool ... thanks a lot for the verbose explanation
[01:29] <fabbione> kimo: please deposit 1.000 USD on my bank account
[01:29] <fabbione> ;)
[01:31] <kimo> :)
[01:35] <zul> heylo
[04:00] <hub> hi
[04:00] <hub> how can I get older kernel packages from the Dapper development cycle?
[04:01] <hub> I have a freeze in -23 that happen less often in -22, but used to never have problems with older kernel
[04:01] <hub> I want to track it down to a specific release
[04:09] <jbailey_> BenC: Hey, is this the patch from the guy who's stuff is in the -mm tree?
[04:09] <jbailey_> (re: the lkh mail you sent me today)
[04:09] <jbailey_> If yes, I was hoping to drag you aside and look at it in Paris sometime. =)
[04:09] <BenC> Not sure, but it looked like something you'd be interested in
[04:09] <BenC> ok, we can do that
[04:10] <jbailey_> Or during the toolchain workshop or something.
[04:11] <jbailey_> About 12 hours for me.
[04:11] <jbailey_> The flight was much cheaper.
[04:11] <jbailey_> But I thought it was tomorrow.  I'm glad I looked at my tickets a couple days ago.
[04:16] <BenC> 2 more days for me
[04:21] <zul> dont get into too much trouble
[04:22] <zul> jbailey_: customs agents at trudeau might recieve a call about you on your way back though ;)
[04:22] <jbailey_> Err.  Why?
[04:22] <jbailey_> What did you do?
[04:23] <jbailey_> Oh
[04:23] <jbailey_> misread.
[04:23] <jbailey_> I know where you live.
[04:23] <zul> you know the city, not the street :)
[04:23] <jbailey_> I just need a sufficiently high grade weapon.
[04:23] <zul> true
[04:23] <jbailey_> ;)
[04:30] <BenC> when I went to Montreal
[04:30] <BenC> my friend got flagged in Baltimore for random search
[04:31] <zul> last time i flew out of montreal there was a bomb scare on the airplane that i was on.
[04:31] <BenC> while he was being frisked by a large security person behind a glass wall, I leaned over and told the guard "He's got something hidden up his butt, might want to check it out"
[04:32] <BenC> the guard said "If he's got something up there, it's going to stay there"
[04:32] <BenC> my friend was not happy :)
[04:32] <zul> .hehe
[04:36] <hub> zul: customs agent in YUL are assholes
[04:37] <hub> zul: next time I'll ask what 272 means on the custom form
[04:37] <zul> heh
[04:38] <hub> so no archive of the kernel packages anywhere?
[04:40] <zul> isnt there a morgue?
[04:40] <hub> you tell me
[04:41] <hub> I'm looking for older like 2.6.15-19 to track down a regression
[04:41] <hub> my laptop freeze on wake up, very often
[04:41] <Keybuk> hub: on Launchpad
[04:41] <hub> -22 seems better than -23
[04:41] <Keybuk> go to the kernel source page, and it lists all the versions
[04:41] <Keybuk> pick one and it lets you download the binaries
[04:42] <hub> ok. now I need to locate them in Launchpad
[04:42] <Keybuk> linux-source-2.6.15
[04:43] <BenC> https://launchpad.net/distros/ubuntu/+source/linux-source-2.6.15
[04:43] <hub> BenC: yeah reached that
[04:43] <hub> thanks
[04:44] <Keybuk> now pick a version
[04:45] <hub> yep
[04:45] <Keybuk> oh, weird, the binaries are missing
[04:45] <Keybuk> iz lp bug
[04:45] <hub> Keybuk: no biggie, I'll debuild it
[04:45] <Keybuk> ah
[04:45] <Keybuk> no
[04:45] <Keybuk> you have to go to the build
[04:45] <Keybuk> https://launchpad.net/+builds/+build/189805
[04:45] <Keybuk> and then pick a binary
[04:45] <BenC> hub: Only choice if lp doesn't have them is to get the GIT repo, and use the tags to build
[04:46] <infinity> Yeah, binaries are associated with a build, not with the source.
[04:46] <hub> BenC: I'll build from the debian package
[04:46] <infinity> Which actually does make sense, internally.  Honest.
[04:46] <hub> BenC: I have some time anyway
[04:46] <BenC> hub: then you'll be downloading quite a few 80Meg tarballs if you don't get the right one the first time :)
[04:46] <hub> my problem is that I can't even get enough diagnostic for the bug to be usefull
[04:47] <BenC> plus if you do find the right one, still need to use git-bisect to find the exact commit
[04:47] <hub> BenC: I have no clue were it freeze
[04:47] <hub> BenC: I just want to put in the bug, version foo does not
[04:47] <hub> BenC: and stick to that for my laptop
[04:47] <BenC> hub: which is really useless
[04:47] <hub> freeze on wakeup is a waste of time
[04:48] <hub> BenC: then I'll just keep it for myself
[04:48] <hub> I have NO output anyway
[04:48] <BenC> so you don't want the bug fixed for everyone else, and for security updates?
[04:48] <hub> my problem is that dapper final is less stable that developement
[04:48] <BenC> I'm not asking for output
[04:48] <BenC> you can use git to find the exact commit that causes the problem
[04:49] <BenC> if you are willing to take the time to rebuild a kernel that works for you, it would be nice to take a little more to actually find the problem
[04:49] <hub> I have never used git 
[04:49] <hub> BenC: I'll see what I can do on my copious spare time
[04:50] <BenC> hub: if you're willing to do the testing, I'm willing to help you through git
[04:50] <BenC> URL in topic shows how to obtain the source
[04:51] <hub> BenC: ok
[04:52] <hub> the thing is that this is my only machine
[04:52] <hub> the otherone does not have a keyboard
[04:52] <hub> and I'd rather not use it to build kernel
[04:52] <BenC> doing something like "git-checkout 2.6.15-22.33" will get you the source for that version
[04:53] <hub> .34 freeze sometime, but less that 23.39
[04:53] <hub> 2.6.15-19.29 did not
[04:53] <hub> because that one was actually installed but has been purged since
[04:53] <hub> I'll start from there, 
[04:56] <hub> BenC: ubuntu-dapper.git I should get?
[04:58] <hub> BenC: the bug is that one https://launchpad.net/distros/ubuntu/+source/linux-source-2.6.15/+bug/49078
[04:59] <BenC> yeah, dapper.git is what you want
[05:00] <hub> currently cloning
[05:01] <BenC> it's very similar to bk
[05:02] <mkrufky> except git is open source
[05:03] <hub> mkrufky: yeah. that is something that was bugging me
[05:04] <hub> mkrufky: but the only reason I used it is because I did doing kernel development at that time, for a short living
[05:05] <mkrufky> bk was good while it was used -- i'm not knocking it ;-)
[05:07] <hub> mkrufky: the worse was that $BOSS didn't trust people so we had to send hime patches
[05:08] <mkrufky> huh?
[05:08] <hub> yeah we had our own bk tree
[05:08] <mkrufky> oh ok
[05:08] <hub> and all the change had to go thru $BOSS
[05:08] <hub> who didn't care about upstream
[05:08] <hub> unless it was to pull the changes
[05:09] <mkrufky> who is "we" ?
[05:10] <hub> mkrufky: me and other coworker at that time
[05:10] <hub> mkrufky: working on a ADSL settop box for the #2 ISP in France
[05:11] <hub> but this is getting off-topic
[05:12] <hub> does building the kernel use ccache?
[05:12] <hub> if available?
[05:13] <zul> yes you can use ccache
[05:16] <fabbione> BenC: no just a tad slower ;)
[06:09] <hub> BenC: I did a git clone, and if I do a 'git-checkout 2.6.15-19.29', it says 
[06:09] <hub> git checkout: you need to specify a new branch name
[06:10] <BenC> git-branch 2.6.15-19.29 2.6.15-19.29
[06:10] <BenC> try that
[06:11] <BenC> if that doesn't work, cat .git/refs/tags/2.6.15-19.29
[06:11] <BenC> and do git-branch 2.6.15-19.29 <sha from cat>
[06:12] <hub> git-branch did work
[06:12] <hub> or at least didn't complain
[06:13] <BenC> ok, from there you can git-checkout 2.6.15-19.29
[06:13] <hub> in progress
[06:14] <hub> done
[06:15] <BenC> are you using 386, 686, k7?
[06:15] <BenC> sudo apt-get build-dep linux-source-2.6.15
[06:15] <BenC> then do:
[06:15] <BenC> fakeroot debian/rules binary-debs flavours=686
[06:16] <BenC> change 686 to which ever you are using
[06:16] <hub> Pentium M
[06:16] <hub> so likely to be 686
[06:16] <hub> actually it is -386
[06:16] <hub> 2.6.15-22-386
[06:20] <hub> should have I installed a -686 kernel?
[06:20] <zul> BenC: the one kail is talking about is 0.54
[06:22] <BenC> hub: no, doesn't matter
[06:22] <hub> ok
[06:22] <BenC> -386 should work everywhere, while -686 is a little optimized (and has SMP)
[06:23] <hub> I'll test with -386 to not add variety to the testing conditions
[06:28] <KaiL> fabbione, so you want the bug # here again? ;)
[06:28] <KaiL> bug 49870
[06:29] <KaiL> ok, no bot 
[06:29] <fabbione> KaiL: why? i am not a kernel maintainer so i am not going to fix anyway :)
[06:29] <KaiL> https://launchpad.net/distros/ubuntu/+source/linux-source-2.6.15/+bug/49870
[06:29] <KaiL> I thought, you are? only for edgy?
[06:29] <fabbione> nope
[06:29] <fabbione> i was.. for breezy
[06:29] <KaiL> ah!
[06:29] <fabbione> partially breezy sorry
[06:29] <fabbione> mostly hoary
[06:30] <zul> btw 0.53 apparently fixes this from nvidia
[06:30] <KaiL> there's also a 0.54, maybe also bugfix only?
[06:31] <KaiL>  *	0.54: 21 Mar 2006: Fix spin locks for multi irqs and cleanup.
[06:31] <KaiL> uhm, yes...
[06:31] <zul> BenC: ill put it my tree..
[09:17] <zul> hmmm...whats wrong with this bug? #49789
[09:36] <bluefoxicy> BenC:  side note, do you know anyone particularly good with linux kernel mm/ internals?  Particularly the swap code
[09:36] <BenC> not off hand
[09:37] <bluefoxicy> ah damn.  Oh well, maybe when I get most of this thing coded I can coerce someone into doing that part.
[09:38] <cjb> bluefoxicy: #kernelnewbies on oftc?
[09:39] <bluefoxicy> cjb:  they bounce me source fil eto source file but the functions are non-intuitive.  :)
[09:39] <bluefoxicy> cjb:  I'm trying to break up the swap code and make a point where the kernel thinks a page has been swapped out but it hasn't yet been sent to the actual swap medium yet (i.e. disk)
[09:39] <bluefoxicy> which of course has code that just shoves it into the swap medium
[09:40] <bluefoxicy> I have something interesting to do inbetween though.
[09:40] <bluefoxicy> (I'll break suspend to disk probably but that'll be a one line fix)
[09:41] <cjb> Well, how does, say, the encrypted swap support do it?
[09:41] <bluefoxicy> encrypted swap?  I dunno.
[09:42] <bluefoxicy> I'm not doing anything to the swap file though.
[09:42] <cjb> Okay; thought it might be using the hooks you're after.
[09:43] <cjb> Since you want take page -> ? -> store page on disk, and "encrypt page" is one such action.  :)
[09:43] <bluefoxicy> cjb:  http://bluefox.kicks-ass.org/mediawiki-1.5.5/index.php/Project_Zone/Design
[09:44] <bluefoxicy> cjb:  That's what I'm doing :)
[09:47] <WebMaven> BenC: Hi Ben!
[09:47] <WebMaven> BenC: I just upgraded the laptop, and the fix works! Thanks!
[09:47] <WebMaven> Yay!
[09:47] <BenC> cool
[10:10] <BenC> sweet, -25 is fixing bug after bug
[11:27] <zul> heylo
[11:29] <zul> BenC: ping
[11:30] <zul> can you have a look at http://zulinux.homeinux.net/ubuntu/kernel/vmware-player
[11:41] <KaiL_> connection refused ;)
[11:41] <zul> oops
[11:41] <zul> can you have a look at http://zulinux.homelinux.net/ubuntu/kernel/vmware-player
[11:41] <KaiL_> hmm, would it be possible to write a (userspace?) tool, which sends everything accessing /dev/dsp to esddsp or aoss..?
[11:42] <KaiL_> that should finally end the oss-Hell
[11:47] <KaiL_> ..wrong channel? ;)
[11:50] <zul> BenC: compiles and works for me