[01:42] <Roey> hi
[01:42] <Roey> can I borrow anyone for fs driver help in #ubuntu, please?
[05:58] <fabbione> +  Changes by Matthew Garrett:
[05:58] <fabbione> +
[05:58] <fabbione> +  * Make hardware sleep sequence more compliant to ACPI specifications:
[05:58] <fabbione> +    - Add patch acpi_gpe_spec.dpatch.
[05:58] <fabbione> +
[05:58] <fabbione> mjg59: ^^
[05:58] <fabbione> does it look ok?
[06:36] <zul> gday
[06:40] <fabbione> hey zul
[06:40] <zul> hey fabbione good holiday?
[06:40] <fabbione> zul: more or less.. a bit heavy 
[06:40] <fabbione> had to do quite a lot of work in the house
[06:40] <zul> cool..
[06:40] <zul> hah hah :)
[06:41] <zul> i still get today off
[06:41] <fabbione> it was off today too
[06:41] <fabbione> but i am working on the xen kernels
[06:41] <fabbione> i have almost finished :)
[06:41] <zul> ah how are they coming?
[06:41] <zul> neat
[06:41] <fabbione> they are building :-)
[06:41] <zul> i was doing some kernel janitorial stuff this weekend
[06:41] <zul> and playing with 2.6.12-rc1
[06:42] <fabbione> neat
[06:42] <fabbione> -32 is ready for upload
[06:42] <fabbione> we need the usual round of buildd
[06:42] <fabbione> but i will wait tomorrow to do it
[06:42] <zul> more security updates?
[06:42] <fabbione> yes
[06:42] <zul> seesh
[06:42] <fabbione> and an urgent acpi fix
[06:43] <zul> oh?
[06:43] <fabbione> it's something simple, but fixes a lot of "go to sleep" issues
[06:43] <zul> cool
[06:43] <fabbione> as soon as RC is out and -32 uploaded
[06:43] <fabbione> we need to start focusing on breezy
[06:44] <zul> already have ;)
[06:44] <fabbione> yeah we need to put down a TODO list
[06:44] <zul> i started a todo list on the wiki and i have a list of third party drivers ill put in the wiki as well
[06:45] <fabbione> zul: it's in the debian/ dir too
[06:45] <zul> is is? i didnt see it :)
[06:45] <fabbione> external-drivers or something like that
[06:45] <zul> ah ok
[06:45] <fabbione> yeah i did start writing and updating it
[06:45] <fabbione> there is stuff missing there
[06:45] <fabbione> like the mppe driver
[06:45] <zul> k ill update it as well
[06:45] <fabbione> that i figured only a few days ago
[06:46] <fabbione> we will need to split the tasks for the first 2/3 weeks
[06:46] <zul> the 64bit ppp-mppe?
[06:46] <fabbione> yeah that one.. 
[06:46] <fabbione> but it's not 64bit clean
[06:48] <zul> there were some patches i saw for ibm thinkpad notebooks as well i saw on lkml a while ago
[06:48] <fabbione> yes
[06:48] <fabbione> i saw them too
[06:48] <fabbione> but we will not add it back until it is upstream
[06:48] <fabbione> there were several problems with that patch
[06:48] <zul> yeah i saw
[06:49] <fabbione> so i had rather avoid another war of include/exclude
[06:49] <zul> sure..
[06:49] <fabbione> when it is upstream, we will include
[06:50] <zul> put the todo list on the agenda for the meeting as well
[06:50] <fabbione> yeah ... let's keep it in the aob
[06:50] <zul> sure
[06:57] <zul> hostap is not in the list?
[06:58] <fabbione> probably yes
[06:58] <fabbione> i would like to chase the list down first
[06:59] <fabbione> and see what is the status of the actual drivers
[06:59] <fabbione> i see 3 steps here
[06:59] <fabbione> 1) clean
[06:59] <fabbione> 2) forwardport to the new kernel
[06:59] <fabbione> 3) add new drivers
[07:02] <zul> heh...i think i figured out why gentoo can distribute win4lin http://www.digitalriver.com/dr/v2/ec_MAIN.Entry17c?CID=0&SID=40113&SP=10007&PN=5&PID=481292&DSP=&CUR=124&PGRP=0&CACHE_ID=0
[07:03] <fabbione> brrrrr
[07:03] <zul> loosers :)
[07:35] <jbailey> I might have a bug that involves firmware loading on an amd64 (8224) - Is that generally expected to work?
[07:36] <jbailey> (I don't have an amd64, so no exposure)
[07:37] <zul> no amd64 here either..
[07:39] <zul> fabbione: updated the external-drivers in my baz
[07:40] <fabbione> zul: ok
[07:40] <fabbione> jbailey: checking the bug, but i don't have amd64
[07:41] <fabbione> jbailey: it would be more important imho to get udev fixed for firewire
[07:41] <fabbione> mantiena was talkign about it on #u-d
[07:42] <fabbione> i will look into the firware issue in a few sec.
[07:42] <fabbione> i already had something like that for another driver
[07:42] <jbailey> Yup.  I'm just going over my list of new bugs to see if they are things I actually have the skills to fix. =)
[07:42] <fabbione> jbailey: the bug is still assigned to me
[07:43] <fabbione> #3609
[07:43] <jbailey> Oh?  Hmm.  Bugzilla was sending me emails about it, so I assumed it was assigned to me.
[07:43] <fabbione> the changes to do are pretty simple
[07:43] <fabbione> probably you are in CC
[07:53] <zul> bbiab must clean the guinea pig cage
[08:01] <fabbione> xen images are almost ready...
[08:01] <fabbione> i am getting the abi files now
[08:02] <jbailey> I should learn more about xen sometime.
[08:03] <fabbione> jbailey: it is really neat
[08:03] <fabbione> i have been playing with it the last week
[08:03] <fabbione> basically it allows you to do hw partitioning
[08:03] <fabbione> + other nice stuff
[08:04] <fabbione> jbailey: i am pretty sure you are familiar with the concept of execution domain
[08:04] <jbailey> Yeah.
[08:04] <fabbione> each domain is like a real machine
[08:04] <jbailey> We had large sparc setups at my last job.
[08:04] <fabbione> but it's not
[08:04] <fabbione> jbailey: exactly like the e10k
[08:04] <fabbione> except that is done in software
[08:04] <jbailey> These were sunfire racks, but yeah.
[08:05] <fabbione> you run a master domain that boots the real machine and identify/access the hardware
[08:05] <fabbione> and small subdomains
[08:05] <fabbione> to which you assign hardware
[08:05] <fabbione> that's basically what xen does
[08:05] <fabbione> but it has some extra features
[08:05] <fabbione> like the possiblity to migrate one domain to another host in the same lan
[08:05] <jbailey> Does it do virtualisation for sharing resources like NICs and disks?
[08:05] <jbailey> Oh nice. =)
[08:06] <fabbione> from the master domain you export one or more block devices to each domain
[08:06] <fabbione> it can also be nfs or whatever
[08:06] <fabbione> more domains can share the same device
[08:06] <jbailey> Cool, so it's not necessarily whole disk.
[08:06] <fabbione> with the usual rw/ro limitations
[08:06] <fabbione> not at all
[08:06] <fabbione> you can export a partition or a lvm partition
[08:07] <fabbione> one limitation that i need to check is how it handles the ram allocation
[08:07] <fabbione> right now each domain has its own ram allocated
[08:07] <fabbione> that makes it a bit less dynamic than UML
[08:07] <fabbione> that can see the whole memory
[08:07] <fabbione> but the domains can't interfere with each other
[08:08] <fabbione> with xen, you assign N MB and the domain can see only that
[08:08] <fabbione> so basically you are limited in that way
[08:08] <jbailey> Still nice, though.
[08:09] <jbailey> I can imagine some nice stuff being done with a SAN and a heavily SMP box for ISPs and whatnot.
[08:09] <fabbione> actually SMP is not supported yet (as 2.6.10)
[08:09] <fabbione> or at least it looks disabled as config option and it just does it
[08:10] <fabbione> because it is mentioned everywhere that you can bind a specific domain to a certain CP
[08:10] <fabbione> CPU
[08:10] <fabbione> or keep it in loadsharing
[08:10] <fabbione> btw the scheduler is highly tunable
[08:10] <fabbione> i am off for dinner :-)
[08:10] <fabbione> bbl
[08:11] <jbailey> 'bye Fabio.  Thanks for the explanation. =)
[08:13] <zul> c ya
[08:28] <zul> heh i had mcdicks for lunch
[08:28] <fabbione> ehehe
[08:29] <fabbione> there we go... abi files generated!
[11:18] <zul> cool...i just sent my first patches to kernel-janitors ml