[03:44] <bluefoxicy> startkeylogger
[10:32] <heatxsink> so 2.6.15-16-23 doesn't have that bluez patch?
[01:32] <fabbione> hey BenC 
[01:33] <BenC> hey fabio
[01:33] <fabbione> BenC: sparc buildds are up and running
[01:33] <fabbione> :)
[01:33] <BenC> kick ass
[01:34] <fabbione> BenC: if you can be so kind to pull/push from me i have a few |Viagara bug fixes from davem
[01:34] <fabbione> RTC driver and SMP
[01:34] <BenC> fabbione: btw, this time around, I kept all the ABI files for all 6 architectrures :)
[01:34] <fabbione> now it is stable :)
[01:34] <BenC> ok
[01:34] <fabbione> yeah i saw
[01:34] <fabbione> no abi changes
[01:34] <BenC> great, now Sun can send me one so I can use it :)
[01:34] <fabbione> i wish i could get one too :)
[01:34] <BenC> if they send me two, I'll sneek one into Germany for you :)
[01:35] <fabbione> eheh
[01:35] <fabbione> too bad i won't be in Germany
[01:35] <fabbione> my wife delivery is right at that time
[01:35] <fabbione> so i might or not come..
[01:35] <fabbione> it depends how i "fell"
[01:35] <BenC> ah, yeah, don't want to take a chance of missing that delivery
[01:35] <fabbione> yeah exactly
[01:36] <fabbione> https://launchpad.net/+builds 
[01:38] <fabbione> BenC: i wanted to talk to you about silo too
[01:38] <fabbione> how bad would it be to move it to git? (including history)
[01:38] <fabbione> tailor did convert the svn repo without a glitch
[01:38] <fabbione> or gime davem and me svn commit rights? :)
[01:39] <BenC> would probably be easier to convert to git
[01:39] <BenC> drop it on master.k.o
[01:39] <fabbione> yeah
[01:40] <fabbione> http://people.ubuntu.com/~fabbione/silo.tailor.conf
[01:40] <fabbione> use this tailor config to import trunk into git
[01:40] <fabbione> with history
[01:40] <fabbione> of course change the paths to what fits you best :)
[05:07] <jbailey> BenC: *poke*
[05:07] <BenC> jbailey: yo
[05:07] <jbailey> BenC: FCare to join us on the call? =)
[05:07] <BenC> oh crap
[05:07] <BenC> give me about 2 minutes
[05:07] <jbailey> *lol*
[05:15] <mjg59> Hm.
[07:16] <BenC> shitty...both my USB2 systems work fine with a USB2 storage device
[07:17] <BenC> why can't I get a broken ehci like everyone else
[07:17] <mjg59> BenC: How's your IDE knowledge?
[07:17] <BenC> suitable for questions, not sure about answers though :)
[07:17] <mjg59> I have a PATA controller here
[07:17] <mjg59> If I drive it with ata_piix (the libata driver), it works fine
[07:18] <BenC> via, intel, promise?
[07:18] <mjg59> Intel
[07:18] <BenC> intel, ok
[07:18] <BenC> ahci is broken?
[07:18] <mjg59> If I drive it with piix (the drivers/ide one), it doesn't work
[07:18] <mjg59> No, this isn't a SATA controller
[07:18] <BenC> ah, right, piix
[07:18] <BenC> what sort of error messages from piix?
[07:19] <mjg59> Or rather, it seems fine until I load ide-cd, at which point I get a spew of "failed opcode" errors and then interrupts rise at about 1000 a second
[07:19] <BenC> this seems familiar...I think there's a bug like this in lp
[07:20] <BenC> I can't wait for 2.6.16 when most everything is in libata
[07:20] <BenC> will be a nightmare for upgrades, but it will be worth it
[07:22] <BenC> can't seem to find it right now
[07:24] <BenC> want to try the 2.6.16-git piix.c in dapper's kernel?
[07:26] <BenC> nm, no code changes in 2.6.16-git
[07:36] <jbailey> BenC: Why a nightmare for upgrades?
[07:36] <jbailey> initramfs-tools ought to just load the right driver based on the PCI ID, shouldn't iT?
[07:37] <BenC> the switch to libata based IDE will make things go from /dev/hda to /dev/sda
[07:37] <BenC> if we had been using UUID based root and filesystem mounting all along, it wouldn't be a problem
[07:38] <BenC> however, there's also initramfs support for find-root or something like that
[07:38] <BenC> maybe it can help detect the changes
[07:40] <mjg59> Or udev can magically make them appear as hdwhatever anyway
[07:40] <mjg59> (I'm not convinced, but...)
[07:41] <BenC> hmm...I wonder if udev can rename devices for PATA to a corresponding hd instead of sd :)
[07:43] <jbailey> Ouch, that's nasty.
[07:43] <jbailey> Is this a permanent change?
[07:44] <doko> BenC, mjg59: what package/setting does turn on power management for hd's? shutting down hd's in a server install doesn't seem to make sense
[07:44] <jbailey> Like will all IDE devices then be sda'ish evermore?>
[07:44] <BenC> doko: not sure...
[07:44] <mjg59> doko: To the best of my knowledge, no package does it
[07:44] <BenC> jbailey: pretty much...I think the long term plan is to move everything to libata/PATA
[07:45] <BenC> doko: do you have the -server install, or just normal ubuntu?
[07:45] <jbailey> BenC: Ah, okay.
[07:45] <doko> BenC: -server install
[07:45] <doko> so normal install CD, but -server install, server kernel
[07:46] <BenC> doko: complain to fabbione...he's responsible for much of -server stuff
[07:46] <BenC> doko: dapper?
[07:46] <doko> yes
[07:46] <BenC> yeah, he'd be interested in that
[07:46] <fabbione> BenC: nope.. it's a .15 thingy to shut down disks all the time
[07:46] <doko> fabbione: ^^^
[07:46] <fabbione> doko: ^^
[07:46] <BenC> fabbione: are you sure it's not some default in hdparm.conf?
[07:47] <BenC> or perhaps even a BIOS setting
[07:47] <fabbione> even if i boot with .15 and stay at "Choose language" in d-i the disks are turned off
[07:47] <fabbione> BenC: it didn't happen with .12
[07:47] <fabbione> even my raid5 of 6 disks does go to sleep once in a while
[07:47] <BenC> can it be overridden with hdparm?
[07:47] <doko> I can confirm this, .12 didn't spin down the disks
[07:47] <BenC> -server could have customer settings
[07:47] <BenC> custom
[07:48] <fabbione> BenC: dunno...
[07:48] <fabbione> i am not a fun of hdparm
[07:48] <mjg59> BenC: Probably best to figure out why the kernel is setting it...
[07:48] <fabbione> mjg59++
[07:48] <BenC> could be a good reason though :)
[07:48] <BenC> most people want that...but I guess it could be changed for -server
[07:49] <mjg59> BenC: We can set the timeout from userspace
[07:49] <fabbione> no userland tricks pretty please
[07:49] <BenC> can you get the hdparm output for .12 and .15 and see if there's some difference?
[07:49] <fabbione> BenC: yes i think i can
[07:49] <fabbione> BenC: but not right away
[07:49] <mjg59> But by default, the kernel ought to be leaving them runnning
[07:49] <BenC> then I can track down in the kernel
[07:49] <BenC> np, just email it when you can
[07:49] <fabbione> BenC: i also got hints from elmo about -server kernel
[07:49] <fabbione> BenC: like using a more sane deadline scheduler by default
[07:50] <BenC> looks like we'll be getting a powerpc-pseries kernel back, BTW
[07:50] <BenC> IBM is offering to test it for us
[07:50] <fabbione> ah nice
[07:50] <fabbione> ask benh for a config
[07:50] <fabbione> he has access to one
[07:51] <BenC> I wish we could get someone to get mkvmlinux+oldworld boot working again too
[07:51] <fabbione> why?
[07:51] <fabbione> BenC: you could ask sven luther
[07:52] <fabbione> but that would also mean shipping 20 CDs for powerpc
[07:52] <BenC> uh, heh, might pass
[07:52] <fabbione> one for each possible arch/subarch/underarch combination
[07:52] <fabbione> :P
[07:52] <BenC> don't forget the uberarch's
[08:02] <fabbione> ehe
[08:04] <jbailey> Are there any oldworld systems that people still care about?
[08:04] <jbailey> I thought you had to go back quite a ways to find them.
[08:25] <BenC> infinity claims he has a G3 that needs oldworld boot
[08:26] <BenC> but I don't think he cares enough about it to get it working :)
[09:35] <bluefoxicy> Can somebody give this bug a proper looking at and set appropriate status and severity so it gets assigned faster
[09:35] <bluefoxicy> https://launchpad.net/distros/ubuntu/+source/linux-meta/+bug/32072
[09:36] <bluefoxicy> kernel NULL pointer dereference in 2.6.15-14-amd64-k8
[09:36] <bluefoxicy> don't know if it was found and fixed by -17