[04:37] <Micksa> mjg59: I HAVE STR!
[04:37] <Micksa> I'm so happy
[04:39] <Micksa> has anyone suggested having ubuntu set "echo 10 >/proc/sys/vm/swappiness" on startup?
[08:55] <fabbione> morning
[08:55] <fabbione> Micksa: and what would be the benefit to have that setup?
[09:06] <crimsun> it'd be fairly dubious to make such a change without extensive testing
[09:17] <fabbione> i am not going to anyway
[09:17] <fabbione> but i would like to know the possible benefit of it
[09:20] <Micksa> general idea is it "improves desktop response"
[09:20] <fabbione> lamont: no we still don't have ipv6 stateful support
[09:20] <fabbione> Micksa: the general idea without specific data out of benchmarks are only subjective feelings ala gentoo
[09:20] <Micksa> I *think* it's the max percent of memory free before the kernel starts preemptively swapping stuff out
[09:21] <fabbione> "hey i compiled my kernel with -O31337 and everything crashes as twice as fast!"
[09:21] <Micksa> example: setting it to a lower value means if you copy a gig of files it won't swap all your apps out in the process
[09:22] <Micksa> what kind of testing? stableness or performance?
[09:22] <fabbione> Micksa: what is your default value?
[09:22] <fabbione> not the one you set.. the one that's there after a fresh boot
[09:22] <Micksa> well the default is 60 (it's 1-100) but I've set mine to 10 to see how it goes
[09:23] <fabbione> how much ram do you have on your system?
[09:23] <Micksa> 1G
[09:23] <Micksa> on this one
[09:23] <fabbione> same here
[09:23] <crimsun> [http://www.ussg.iu.edu/hypermail/linux/kernel/0409.0/1937.html for an earlier thread] 
[09:23] <fabbione> on the others?
[09:23] <Micksa> so like, I don't think it makes so much difference for me :)
[09:23] <fabbione> i am curious to know if it is calculate at boot time
[09:23] <Micksa> the only other machine I have running 2.6 at the moment has 512M, but I don't have it in linux anyway
[09:24] <Micksa> I think it's just set at 60
[09:24] <Micksa> I think it's a percentage
[09:24] <Micksa> I don't know for sure
[09:24] <Micksa> all I know is that it's in a range of 0-100, and a higher value gives more preference to cache, less to apps
[09:25] <Micksa> we could argue forever about where it should be set 8)
[09:25] <Micksa> google for it if you don't believe me
[09:25] <fabbione> Micksa: dude.. i didn't say i don't believe
[09:25] <crimsun> the higher the value, the more likely the kernel will swap out an app instead of reclaiming from cache
[09:25] <fabbione> i said i want to see numbers
[09:26] <Micksa> I just thought I'd mention it.  It can at least improve *percieved* performance in certain situations
[09:26] <fabbione> and already the first page of crimsun thread would give you a perfect idea on why you shouldn't touch it
[09:26] <fabbione> "For example, at a swappiness value of 0, Kernel 2.6.5 swapped out 0 bytes,
[09:26] <fabbione> whereas Kernel 2.6.9-rc1-mm3 swapped out 10 GB."
[09:27] <Micksa> heh, ah, well :)
[09:27] <fabbione> since it's not a consistent value across kernels
[09:27] <fabbione> i think the default will be fine for everybody
[09:27] <fabbione> given that the benchmarks says that 60 is where the best interaction is at the moment
[09:28] <Micksa> BUT IT WILL MAKE EVERYTHING BETTER
[09:28] <Micksa> :)
[09:29] <fabbione> probably.. on desktop.. what about servers?
[09:29] <Micksa> it's probably best left at 60ish for servers
[09:29] <fabbione> and how do you expect to detect a server install from a desktop one?
[09:29] <crimsun> another reason to leave it alone: http://bugs.gentoo.org/show_bug.cgi?id=54560
[09:29] <Micksa> I dunno, I thought you guys had that covered :)
[09:29] <crimsun> I think we all recognize the name greg k-h
[09:29] <fabbione> Micksa: there is no way to distinguish installations
[09:30] <Micksa> damn.
[09:31] <Micksa> wouldn't it be nice if 2.6 got stable
[09:32] <Micksa> mmm, google searchable irc logs :)
[09:34] <Micksa> argh dammit, it's not in here
[09:34] <Micksa> oh these are yours! :)
[09:35] <Micksa> dammit. could you grep for a http:// from mjg59 around 7 days ago?...
[09:35] <Micksa> he pointed me at a patch which I've since lost
[09:35] <fabbione> http://people.ubuntu.com/~fabbione/irclogs/ <- kthxbye
[09:37] <Micksa> FINE!
[09:37] <fabbione> Micksa: it's several MB :)
[09:37] <Micksa> *fuming* I know....
[09:38] <fabbione> and a wget -r will parse the .html
[09:38] <fabbione> so basically you will endup downloading the internet :)
[09:39] <Micksa> wget -r -np. whatever.
[09:39] <Micksa> anyway I've found it
[09:39] <Micksa> you coulda just grepped for me and it would have saved us all this ;)
[09:39] <Micksa> but, I guess you had fun
[09:39] <fabbione> Micksa: probably you miss a point
[09:40] <fabbione> it's sunday... it's weekend.. i am not even going to ssh to people to grep :)
[09:40] <Micksa> I sure am a whiny bitch huh
[09:40] <fabbione> that too :)
[09:41] <fabbione> and it doesn't make you gaining any point in here
[09:41] <Micksa> that's not fair
[09:41] <Micksa> this is me!
[09:41] <Micksa> well sorta.
[09:41] <fabbione> whining doesn't make you gain point
[09:41] <Micksa> I just like to play the whiny bitch every now and then for fun
[09:41] <fabbione> indipendently who you are :)
[09:41] <Micksa> I have to hold myself back if I'm not considered a regular.
[09:41] <Micksa> It's no fun.
[09:42] <fabbione> Micksa: you can have fun without being a bitch :)
[09:42] <fabbione> but clearly that depends on the kind of fun you like
[09:43] <fabbione> "slap me! i am your bitch!" <- we do NOT need/want to know :P
[09:43] <Micksa> >:)
[09:43] <Micksa> I tend to be the one doing the slapping
[09:43] <Micksa> which of course helps to increase my popularity
[09:48] <Micksa> gotta go, back laterish
[09:59] <Micksa> boing boing boing
[11:22] <fabbione> Mithrandir, smurfix: ping?
[11:23] <fabbione> Linux trider-g7 2.4.30-xen0 #4 Sun Jun 12 10:56:56 CEST 2005 i686 GNU/Linux
[11:23] <fabbione> :)
[12:05] <Mithrandir> fabbione: yaypong. :-)
[12:08] <fabbione> Mithrandir: did you finish with Uni?
[12:09] <Mithrandir> yeah, handed in my thesis friday.  So I'm freeeeeeeeeeee now
[12:09] <fabbione> ahah cool!
[12:11] <Mithrandir> that's a 2.4 kernel, though?
[12:11] <fabbione> yes
[12:11] <fabbione> i can't run 2.6 on that box yet
[12:11] <Mithrandir> oh, ok.
[12:11] <Mithrandir> anyhow, you got this stuff in arch?
[12:12] <fabbione> uh?
[12:12] <fabbione> i am just using upstream
[12:12] <fabbione> manual patch /build
[12:12] <Mithrandir> ok
[12:12] <fabbione> but there is an issue building the xenU kernel
[12:12] <fabbione> so i need to see if i can boot a subdomain with the domain0 kernel
[12:12] <fabbione> in theory is the same
[12:12] <Mithrandir> what's the problem building the xenU kernel?
[12:13] <fabbione> it's an unreferenced call to something
[12:13] <fabbione> linking error basically
[12:13] <Mithrandir> hm
[12:14] <fabbione> arch/xen/drivers/netif/drv.o(.text+0x37ec): In function `netif_connect':
[12:14] <fabbione> : undefined reference to `direct_remap_area_pages'
[12:14] <fabbione> so i need to understand why is missing
[12:15] <fabbione> probably i just miscompiled it
[12:15] <fabbione> i forgot that 2.4 is a bit behind with kbuild
[12:26] <fabbione> hmm no
[12:26] <fabbione> nothing to do with compilation
[12:26] <fabbione> crap
[12:42] <fabbione> ah hmmm
[12:43] <fabbione> stupi config error i think
[12:59] <fabbione> ah there it is :)
[01:00] <fabbione> fixed
[02:47] <fabbione> ok perfect...
[02:47] <fabbione> xenU boots :)
[02:48] <fabbione> Mithrandir: whatever xen config you are going to do.. remember to disable devfs :)
[02:48] <Mithrandir> heh, ok.
[03:01] <Mithrandir> hmf.  Sourcepuller doesn't work with bkbits, apparently.
[03:18] <Mithrandir> fabbione: you've just pulled the regular stable xen version rather than messing with bk, right?
[05:32] (fabbione/#ubuntu-kernel) freenode should die
[05:33] <Mithrandir> we could start by switching to another network ourselves. :-P
[05:35] <fabbione> actually that's an idea
[06:10] <Micksa> okay, so I'm trying to compile a patched kernel
[06:10] <Micksa> using the linux-source package works, except firmware doesn't get compiled in
[06:11] <Micksa> so like, how should I compile from the *real* source for just the one subarch?
[06:30] <mjg59> Micksa: Cool. What did you add?
[06:31] <Micksa> hi
[06:32] <Micksa> the pcie patch you pointed me at fixed the graphics problem
[06:33] <mjg59> Cool
[06:33] <mjg59> Can you feed that to fabbione?
[06:33] <Micksa> now I just have to put it all together
[06:33] <Micksa> haha
[06:33] <Micksa> no
[06:33] <Micksa> he hates me
[06:34] <Micksa> *chuckles*
[06:34] <Micksa> what does he need, just an email saying "these patches fix these problems with the i6000"?
[06:38] <mjg59> Yeah, but he won't apply the SATA one yet
[06:46] <fabbione> no because upstream said that the patch is broken
[06:46] <fabbione> also is this pcie patch upstream already?
[06:46] <Micksa> *shrug*
[06:50] <mjg59> I don't believe so, no
[06:50] <mjg59> It only appeared last week
[06:51] <fabbione> mjg59: so let's wait before we start pulling too many patches from different places
[06:51] <fabbione> i would like to get to 2.6.12 final first
[06:51] <mjg59> fabbione: I'd prefer to have a kernel that we can actually expect to work on laptops
[06:51] <mjg59> We've no idea how long 2.6.12 final is going to be, but we *need* a kernel that can be used for laptop testing
[06:52] <fabbione> mjg59: .12 should be out pretty soon according to linus
[06:53] <fabbione> mjg59: with this "pull in all patches from everywhere" we fucked up hoary badly
[06:53] <fabbione> specially for 2 SATA controllers
[06:53] <fabbione> thanks god they are not used in the installer
[06:54] <Micksa> does it fry them? 8)
[06:54] <fabbione> and we will need to do soon a point release
[06:54] <fabbione> so for breezy we are going to be a bit more careful
[06:54] <fabbione> i trust your acpi crack
[06:54] <fabbione> but i don't want to end up like hoary
[06:54] <fabbione> Micksa: data loss
[06:54] <Micksa> aw, that's no fun
[06:54] <mjg59> fabbione: How soon is "pretty soon"?
[06:55] <fabbione> mjg59: max a week
[06:55] <mjg59> We need to start this stuff within the next couple of weeks
[06:55] <mjg59> Ok, cool
[06:55] <fabbione> mjg59: .12 is not even in main
[06:55] <fabbione> so you can start whenever you want
[06:55] <mjg59> Uhm
[06:55] <fabbione> push the crack to people
[06:55] <Micksa> it's really cool seeing an announcement from _A_ saying "don't use kernel x if you have device y, it'll fry it. no, really."
[06:55] <fabbione> and have fun :)
[06:55] <mjg59> I can't ask people to build their own kernels from scratch. They're not all going to be competent to
[06:55] <Micksa> haha
[06:56] <Micksa> what about a laptop-test kernel?
[06:56] <mjg59> Yeah, that might be an idea
[06:56] <fabbione> mjg59: if you prefer to handle your own kernel for testing go ahead
[06:56] <Micksa> "this kernel might make your laptop work. also, it might cause it to explode."
[06:56] <Micksa> "we don't know."
[06:56] <fabbione> right now i am not too happy to push more than required
[06:57] <mjg59> Ok, fair enough
[06:57] <fabbione> specially because there is still hell of a lot of work to do to move .12 in main
[06:58] <fabbione> mjg59: anyway.. it's sunday and we are talking about work
[06:58] <Micksa> heh
[06:58] <fabbione> me not happy very much work weekend
[06:58] <Micksa> just hit monday here :)
[06:58] <mjg59> Heh
[06:58] <fabbione> mjg59: let's talk about it tomorrow
[06:58] <mjg59> Yup, that's no problem
[07:45] <zul> heylo
[07:50] <zul> oi...its so freaking hot
[08:54] <zul> hey fabbione
[09:43] <fabbione> hey zul
[10:29] <fabbione> lamont: ping??