[01:12] <zul> mmm...beer..
[01:19] <BenC> Mithrandir: Between the first and last yes
[01:19] <Mithrandir> goodie
[04:18] <poningru> guys writing the herd 3 release notes
[04:18] <poningru> any suggestion I can put under kernel?
[04:23] <poningru> so no one??
[04:24] <crimsun> try the changelog.
[04:24] <poningru> looking at it
[04:24] <poningru> seems just like bug fixes
[04:25] <poningru> nothing big that can be 'featured'
[04:25] <crimsun> "just" is an understatement.
[04:25] <poningru> oh ofcourse
[04:25] <poningru> keep in mind marketing speak
[04:25] <poningru> ;)
[04:27] <BenC> poningru: paravirt + vmware VMI ops, but it may be premature since there's a small chance that it might not make it in release
[04:31] <poningru> meh I'll put it in there
[04:31] <poningru> wait its not already in there?
[04:32] <BenC> it's in there for Herd 3
[04:32] <poningru> oh so it might get taken away? doesnt matter as long as its in there for herd 3
[04:32] <BenC> but mark it in testing phase
[04:32] <poningru> kk
[04:32] <poningru> thanks :)
[04:32] <BenC> we're still investigating the performance inpact on bare metal systems
[04:32] <BenC> impact
[01:21] <kylem> morning folks.
[02:07] <BenC> morning kyle
[02:09] <kylem> s'up?
[02:13] <BenC> not much
[02:26] <BenC> mjg59: pinger
[02:37] <zul> hey
[03:09] <mjg59> BenC: Hi
[03:10] <BenC> mjg59: Hey, rt2x00 drivers don't build with the dscape you sent me
[03:10] <BenC> everything else seems to work fine
[03:10] <BenC> I tried the stock rt2x00 and the one you sent with d80211
[03:10] <BenC> it's like it's expecting a whole different API
[03:13] <mjg59> Hm. I hit workqueue issues, but I thought they looked ok otherwise
[03:13] <mjg59> But that patch is pretty ancient now...
[03:20] <kylem> mjg59, i thought we had rt2x00 building before?
[03:20] <mjg59> I didn't think so for dscape
[03:20] <mjg59> Everything else, yeah
[03:26] <BenC> everything else compiles fine, and I removed the old stuff, and disabled stock bcm43xx
[03:26] <BenC> mjg59: So what do people need to do to get v4 firmware?
[03:26] <BenC> obviously it isn't out in the wild as much because softmac can't use v4
[03:27] <mjg59> Run fwcutter on a v4 image
[03:27] <mjg59> There's some listed in fwcutter's readme
[03:27] <mjg59> We probably want to hack fwcutter to postfix the v4 name and add a default fwpostfix argument to the dscape bcm43xx driver
[03:28] <mjg59> In order to support co-installation
[03:37] <cjwatson> is there any v4 firmware on the web anywhere, like the older linksys firmware?
[03:37] <mjg59> Yes
[03:37] <mjg59> See the fwcutter readme
[03:51] <BenC> mjg59: I'm not so sure I want to bother with multiple drivers for the same hardware
[03:52] <BenC> I either want to switch to the d80211 or stick with stock
[03:52] <BenC> at least for bcm43xx
[03:54] <BenC> Mithrandir: I fixed the kexec-tools build failure on ppc if you wouldn't mind allowing ubuntu2 through for universe
[03:55] <mjg59> BenC: It would be good to get these things tested, but I'm really not in a position to make the call
[03:55] <BenC> mjg59: I've got testers for the new driver
[03:55] <BenC> there were like 4 people in oslo ready to kill me because of how crappy bcm43xx is :)
[03:55] <mjg59> By testing I mean at a distro-user sort of level :)
[03:56] <mjg59> Right
[03:56] <BenC> right, but I assume that if a user already has setup bcm43xx once, they will know how to do it again
[03:56] <mjg59> Once we've got some idea of how good it is in wireless-dev, we probably want to look at pulling Michael Busch's tree
[03:56] <kylem> security is uploaded.
[03:57] <BenC> mjg59: The way I have it setup now, I can enable/disable it on demand from .config's so it is at least in our tree
[03:57] <mjg59> BenC: Yeah
[03:58] <mjg59> Oh my.
[03:59] <mjg59> USB bcm43xx is apparently pretty much RNDIS
[03:59] <mjg59> So. Much. Crack.
[03:59] <Mithrandir> BenC: sure, I'll do
[04:06] <BenC> Mithrandir: thanks
[04:16] <BenC> kylem: Stone cold rejected!
[04:16] <kylem> lol.
[04:17] <kylem> wrong alias.
[04:17] <BenC> at least it isn't just me that does that :)
[04:18] <kylem> it would be nice if we only had to upload to one place to hit all 3.
[04:20] <BenC> this sucks...I setup acpi scripts to disable my second core, and force my first to 1GHz when I'm on battery and I still only get an hour of battery
[04:21] <Lure> BenC: how old is battery (mine gets at that level after 8-9 months)
[04:21] <BenC> it's just a few months old...been that way since I bought it
[04:22] <BenC> I should install winxp and see if it's the same
[04:22] <kylem> pfft. old hat.
[04:22] <kylem> you want vista now.
[04:22] <kylem> ;-)
[04:22] <BenC> hehe
[04:22] <Lure> kylem: vista ultimate signature edition to be precise ;-)
[04:23] <kylem> heh
[04:24] <Lure> BenC: -6 kernel includes new serio we were talking recently? 
[04:24] <mjg59> BenC: How are you disabling the second core?
[04:24] <mjg59> There should be pretty much no power benefit to doing so
[04:24] <Nafallo> thanks etc :-)
[04:24] <zul> kylem: meh...windows 3.1 is better its shiney
[04:24] <BenC> mjg59: echo 0 > /sys/.../cpu1/enable
[04:25] <BenC> Lure: Should
[04:25] <zul> mmmm....shiney..
[04:25] <mjg59> BenC: Hm. I have a feeling that that may do little other than just prevent tasks being scheduled on it
[04:25] <Lure> BenC: ok, then it does not help for HP problems
[04:26] <BenC> Lure: Check changelog to make sure
[04:26] <BenC> mjg59: Would be nice if it physically disabled and powered down the CPU :/
[04:26] <Lure> BenC: will now build custom kernel with patch from kernel bug 7689
[04:26] <mjg59> BenC: Don't think that's possible on any existing hardware
[04:26] <kylem> BenC, easy way to check... is there still interrupts occuring on cpu1?
[04:27] <BenC> guess I'll have to disable it in the bios when I travel to get any decent battery life
[04:27] <BenC> kylem: /proc/interrupts doesn't even show the second one after I disable it
[04:27] <kylem> huh.
[04:27] <BenC> and it shows migration of tasks/interrupts to the first cpu when I do
[04:28] <BenC> so the second cpu is atleast not doing anything
[04:28] <kylem> it's probably just ready for hot-unplug then.
[04:28] <BenC> but I guess it's still drawning power
[04:29] <BenC> stupid xchat
[04:30] <mjg59> BenC: As I said, as far as I know even disabling it in the bios won't have any significant effect on power saving
[04:30] <BenC> mjg59: Only getting an hour of battery totally sucks then
[04:30] <mjg59> Well, yes
[04:30] <BenC> If this is hardware design, then I'm never buying an HP again :/
[04:31] <mjg59> What model is it?
[04:31] <BenC> I would rather it be a linux kernel bug
[04:31] <BenC> dv5000
[04:31] <maks_> x41 doesn't do much more battery life..
[04:31] <BenC> coreduo
[04:31] <mjg59> It's consumer, so it's always likely to be worse than business-class
[04:31] <mjg59> But yes, Linux is poor at PM
[04:31] <BenC> is it just a shit battery?
[04:32] <mjg59> What does /proc/acpi/battery/state say when on battery?
[04:32] <BenC> I willing to fork cash for a higher rated battery if it will help
[04:32] <BenC> present:                 yes
[04:32] <BenC> capacity state:          ok
[04:32] <BenC> charging state:          discharging
[04:32] <BenC> present rate:            0 mA
[04:32] <BenC> remaining capacity:      4356 mAh
[04:32] <BenC> present voltage:         14800 mV
[04:32] <BenC> the thing that gets me is that rate is never shown
[04:32] <BenC> but I tested the lifetime at idle by wall-clock to see how long it lasted
[04:33] <BenC> dead idle, no screen blank, 1 hour
[04:33] <mjg59> And it lasts an hour?
[04:34] <BenC> wonder if I can fix the rate state with a BIOS update
[04:34] <mjg59> That implies it's drawing 60 watts
[04:35] <BenC> I watched cpu state too, and it shows them both at 1GHz (1.6GHz is max)
[04:35] <BenC> I wish it would drop lower, to maybe like 800MHz or so
[04:36] <mjg59> 1GHz is probably the lowest it goes
[04:36] <BenC> it is
[04:36] <mjg59> You're into rapidly diminishing returns after that
[04:36] <BenC> 3 levels, 1, 1.3 and 1.6
[04:36] <mjg59> If the processor is idle, it makes no real difference what clock speed it's at
[04:36] <mjg59> Most of the chip will be switched off anyway
[04:36] <BenC> right
[04:37] <BenC> if I do a make -j3, battery life turns into quick death :)
[04:37] <BenC> I'd be lucky to get 20 minutes out of it
[04:37] <mjg59> So there's no point in dropping the frequency further unless you can drop the voltage further
[04:38] <mjg59> And my suspicion is that dropping the voltage further is impractical for some reason, so they might as well nail it at 1GHz
[04:45] <BenC> guessing I just need a better battery
[04:46] <mjg59> BenC: Given that power draw? No
[04:46] <mjg59> (Assuming it's reporting accurately)
[04:47] <mjg59> 60 watts is insane
[04:48] <BenC> is 6600mAh capacity average?
[04:53] <BenC> stupid hp...I gotta install a windows program that checks my bios and downloads a newer version if I need it
[04:58] <BenC> zul: nice job on the 2.6.19 xen
[05:00] <zul> BenC: thanks
[05:01] <zul> ill upload the xen-meta tonight so it will be easier to install
[05:03] <BenC> zul: Any chance of it getting into feisty with 2.6.20?
[05:04] <zul> i could try but dont count on it
[05:04] <zul> its not very trivial ;)
[05:23] <kylem> BenC, any idea if we have a "sun galaxy x4100/x4200" anywhere in the company?
[05:24] <BenC> I think montreal has those boxes
[05:24] <kylem> k.
[05:24] <BenC> if not, then data center has them
[05:24] <kylem> BenC, btw, i'll be a little late starting tomorrow, i'm heading into mtl to pick up some stuff tomorrow.
[05:24] <BenC> I know we have some somewhere :)
[05:24] <BenC> cool
[05:25] <JanC> hm, my "cheap" MSI-based (Centrino) Core Duo does > 2h on battery...
[05:26] <kylem> my macbook core2 lasts 6 hours...
[05:26] <kylem> ... running macos... :)
[05:26] <BenC> I got spoiled with 3.5hours on my powerbook
[05:26] <JanC> kylem: I guess that's with a better battery...  :)
[05:26] <kylem> my x30 lasted for ~4 hours on its 4 cell...
[05:26] <maks_> my x41 lasts one hour .. :P
[05:27] <kylem> maks_, ouch dude.
[05:27] <kylem> heya phillip!
[05:28] <plougher> hi
[05:28] <JanC> BenC: disabling the second core might actually be counter-productive, if it doesn't power-down...  :)
[05:42] <zul> right lunch time...bbl
[07:00] <zul> *grumble*
[07:00] <kylem> ?
[07:01] <zul> ExceedOnDemand is just so much fun
[07:04] <cjb> zul: Hi, got a sec?  On edgy, booting domU Ubuntu guests works fine, but as soon as I change the kernel= line to point to a Fedora kernel instead, I get an "Error 22".  Any ideas?
[07:04] <cjb> Something like "Upgrade to feisty" or "Downgrade to dapper" or "You have to build Xen from their dist" is all good, if that's the answer.
[07:05] <zul> try feisty but I dont know where the Error 22 is coming from, need more info
[07:05] <BenC> HL is unable to deliver to the address supplied to us for your manual order.  Please either forward a correct physical address via reply to this email.  Please supply this information within 48 hours or we will direct DHL to destroy the shipment.
[07:05] <BenC> I hate DHL
[07:06] <kylem> "destroy the shipment" wtf.
[07:06] <cjb> zul: Thanks.  What kind of more info is good?  The log says "failed to balloon" and not much else.
[07:06] <cjb> (But I can certainly get it online.)
[07:06] <zul> cjb: which kernel which arch
[07:06] <BenC> cjb: Sounds like you are trying to boot a non-domU kernel, which wont work
[07:06] <BenC> kylem: No doubt...what a waste
[07:07] <BenC> DHL sucks in my area...they can never deliver here, unlike UPS or Fedex
[07:07] <kylem> ouch.
[07:07] <zul> BenC: ask them how are they going to destroy the shipment :)
[07:07] <cjb> BenC: Hm, so, I had this VM running inside an FC5 machine, and the kernel is the same that it uses, which I presumed was a domU.
[07:07] <cjb> Fedora's /usr/bin/pygrub stuff confuses me.
[07:07] <BenC> last time I got a DHL shipment, the guy brought it in a cadillac because he didn't want t waste the gas with the delivery truck
[07:08] <cjb> BenC: DHL held up our first OLPC prototype laptop in Alaska so the FDA could clear it.  ;-)
[07:08] <BenC> FDA!?
[07:08] <cjb> that's what we said.  :)
[07:08] <kylem> uh. isn't alaska in the us? :)
[07:08] <cjb> yeah, it's a customs point for DHL.
[07:09] <kylem> weird.
[07:09] <zul> kylem: its a freak state like hawaii
[07:12] <BenC> Alaska is a US state!?!
[07:12] <BenC> pshaw
[07:12] <kylem> it would really fit in better as part of the yukon.
[07:12] <zul> one of those freak ones...like peurto rico
[07:12] <BenC> I feel bad for puerto rico, they can't even vote
[07:13] <mjg59> The Russians must be kicking themselves for selling it
[07:14] <BenC> I know, right. Countries only appreciate over time...lose money selling it at current rates
[07:14] <zul> like the french with the louisana purchase?
[07:14] <kylem> mjg59, i think they kicked themselves hardest during the cold war.
[07:15] <kylem> mjg59, would have needed smaller missiles if they still had it. 8)
[07:15] <zul> kylem: the yukon would be more milatarilistic then..
[08:14] <pina> hi did ubuntu for server need array conf, for proliant ML 370 ??, configuragion array to raid for scsi 
[08:20] <kylem> blink.
[08:41] <zul> hmmm...alcatel just layed off 400 people
[10:01] <cjb> ah -- I see "ERROR: Unrecognized image format" in the xend-debug log.  
[10:01] <cjb> the kernel I'm using is "vmlinuz-2.6.15-1.2054_FC5xenU".  Any ideas?
[10:37] <TheMuso> p/c
[10:53] <kylem> BenC, have you been keeping lrm in git?
[10:53] <BenC> no
[10:53] <kylem> hmm. possibly we could do that and put it in /srv/.../private?
[10:53] <BenC> maybe the debian part
[10:53] <BenC> the .orig.tar.gz doesn't make sense there
[10:54] <BenC> debian/* I mean
[10:54] <Nafallo> bzr <3 :-)
[10:55] <Nafallo> :-P
[12:13] <cjb> hmph:  "at least one of the specified interfaces does not exist!" in the xen-hotplug.log.