[12:44] <mjg59> BenC: Why is EDD turned off?
[12:45] <BenC> remind me what the hell it is
[12:45] <mjg59> It's used to let the OS know what the BIOS thinks about attached drives
[12:45] <mjg59> Which lets us work out which drive corresponds to 0x80, which to 0x81 and so on
[12:46] <BenC> it also totally borks a whole class of system to where it can't boot
[12:47] <mjg59> Then we need to fix that
[12:47] <mjg59> But, uh, wasn't it built as a module?
[12:48] <BenC>   * 33939: Disable CONFIG_EDD. Causing problems, and not needed anyway.
[12:48] <BenC>     Symptons were seemingly locked up after grub jumped to kernel, but
[12:48] <BenC>     actually would boot after 2 minute delay.
[12:48] <BenC> EDD enables things in arch/ or something
[12:48] <BenC> disabled over a year ago
[12:49] <BenC> so was disabled for dapper release
[12:49] <BenC> left disabled in edgy
[12:50] <mjg59> Well, we're going to need it to fix a class of bugs
[12:50] <BenC> rgrep CONFIG_EDD arch/i386/boot/
[12:50] <BenC> it does some junk in the boot wrapper it seems
[12:50] <BenC> maybe we can just disable that part
[12:50] <mjg59> Hm. Yeah, I guess it has to do the call at boot time.
[12:51] <mjg59> However, there's no other way of working out whether sda or sdb is the boot drive
[12:52] <BenC> is that because of libata not working as well as ide drivers for getting that info?
[12:52] <BenC> or was it like hda == 0x80 always?
[12:53] <mjg59> hda was /usually/ 0x80
[12:53] <mjg59> But there's no way of guaranteeing that
[12:54] <BenC> hmm...I think I might have a machine that hung with EDD enabled
[12:55] <mjg59> Now that ide and scsi drives are in the same namespace, there's more risk of collisions
[12:56] <mrec> BenC: do you know about any issues with uvc webcams?
[12:57] <BenC> mrec: I don't know about uvc at all, just that I put the module in there :)
[12:57] <mrec> I wonder because I received such a device and the linux-uvc driver (usb video class) just crashes it
[12:58] <mrec> the company which sent the cam to me told me it will work with 1 gig ram (most of it will be used for the driver)
[12:59] <mrec> with video4linux/dvb regarding support for the 60 additional devices we had several discussions on the ML now
[01:00] <mrec> code will be merged regardless of what some people say since they aren't willing to help either
[01:00] <mrec> (well just 2 people were against it :)
[01:06] <BenC> ok
[01:28] <defendguin> still working on feisty bugs?
[01:34] <jb-home> defendguin: Critical ones only.
[01:34] <defendguin> my touchpad is critically broken after suspend 
[01:35] <defendguin> very very annoying if you forget your mouse
[01:38] <defendguin> bug 82860
[01:39] <jb-home> If there's a bug filed, I'm sure it'll be triaged and such.
[01:39] <jb-home> Given that the release just happened, I wouldn't expect to hear from anyone this weekend though.
[01:41] <defendguin> i've been very proactive about the bugs that i have and can help with
[01:45] <mjg59> defendguin: In most cases, the only ways you can be helpful are to test patches or fix the bug yourself
[01:46] <mjg59> These things are very hard to track down without access to the hardware
[01:46] <defendguin> providing timely access to information?
[01:47] <defendguin> i'm sure i could test patches
[03:29] <steb> how do i boot to install a new kernel if none of my kernels work?
[03:48] <jb-home> steb: #ubuntu is better for these types of questions.
[03:48] <jb-home> steb: But I'd suggest a known-working livecd.
[05:22] <vimalg2> can anyone explain why the power management on ubuntu reduces my laptop's battery life by half. Aren't APM and ACPI industry standards?
[05:22] <poningru> yes and no
[05:22] <poningru> acpi is the modern industry standard
[05:23] <vimalg2> oh ok.
[05:23] <poningru> the only trouble is that most laptop manufacturers implement it 
[05:23] <poningru> poorly
[05:23] <vimalg2> why do some lapt-top installation guides reccomend turning off acpi during installation?
[05:23] <poningru> whether this is malicious or not is not known
[05:23] <poningru> vimalg2: because other wise you cant install it at all
[05:23] <vimalg2> poningru: thanks . 
[05:24] <poningru> but those cases have been mostly removed... but BenC or cjwatson are better candidates to answer that particular question
[05:24] <poningru> vimalg2: for example there are many many scripts for acpi that are setup for specific laptops
[05:25] <vimalg2> poningru: oh. I've never had to do the acpi=off setting ever for debian-based distros. I've had trouble getting it on my buddies notebooks
[05:25] <BenC> vimalg2: battery life is probably more to do with your cpufreq than acpi
[05:26] <BenC> never do acpi=off on a laptop...then you have pretty much a desktop on a battery
[05:26] <BenC> no suspend, not battery reporting, no nothing
[05:26] <vimalg2> BenC: I'm on Nehemiah. The associated kernel module is langhaul I believe for cpu-scaling.
[05:26] <BenC> vimalg2: then check your cpu state in /sys/
[05:26] <vimalg2> Is it in a usable state in Feisty.(I'm still downloading the ISO's)
[05:26] <BenC> check the freq, scaling governor, etc.
[05:27] <vimalg2> Via nehemian CPUID 0698
[05:28] <vimalg2> BenC: i'm still figuring out how to get internet access to my hone(It doesnt have a PPP dial-up profile) Ironically the cell phone runs a 2.4.x moontavista kernel. So i have to play with NAT and the route command to get the traffic into Ubuntu/*nix. 
[05:29] <vimalg2> So i'm stuck with et access in windows
[05:29] <vimalg2> *net
[05:29] <BenC> you're actually getting into #ubuntu type questions and topics here
[05:30] <BenC> most of this stuff can be answered in the wiki, or from folks in that channel
[05:30] <vimalg2> BenC: thank you for your time. I was just curious about the state of power management and cpu-scaling for CMOV instruction enabled Nehemiah procesors
[05:31] <vimalg2> I'll just poke around i guess. thanks :)
[05:31] <BenC> vimalg2: we've enabled everything the kernel supports :)
[05:31] <BenC> np
[02:51] <\sh> guys, anyone has a clue how to replace linux/config.h with something which is in linux-headers?
[02:51] <\sh> (feisty 2.6.20-15-generic image) just trying to compile cisco vpn client
[03:06] <jb-home> \sh: Just don't include it and hardcode the defines.
[03:06] <jb-home> The use of config.h is trying to detect what's in your current kernel.
[03:06] <jb-home> It's generally just a Bad Idea(tm)
[03:07] <jb-home> If you're compiling something for distribution, it's always better to do a runtime check for features.
[03:07] <\sh> jb-home: well, I found a patch already
[05:10] <maks_> congrats for the release :)
[05:11] <kylem> thanks maks.
[05:14] <BenC> well, guess I'm going to give this 2.6.21 build a test run...brb
[05:25] <BenC> well, it booted :)
[05:25] <maks_> dynticks?
[05:26] <BenC> $ grep NO_HZ /boot/config-2.6.21-1-generic 
[05:26] <BenC> CONFIG_NO_HZ=y
[05:26] <kylem> woot.
[05:27] <BenC> not running it right now...I've got to get lrm compiled with it
[05:40] <penguin42> Can anyone suggest anything for checking my kernels idea of timing? movie player is playing way faster than real time and I've got a 'Your time source seems to be instable or some driver is hogging interrupts' error - so I wonder if something is going odd (on Feisty)
[05:43] <penguin42> it's a laptop so I also wonder if something is trying to be too clever and then not coping with varying cpu clocks
[05:44] <BenC> penguin42: is it an AMD cpu?
[05:47] <penguin42> no, core2 duo (in 64 bit mode)
[05:48] <BenC> I've not heard of those issues
[05:48] <BenC> have you tried in 32-bit?
[05:48] <BenC> to be honest, 64-bit mode really doesn't buy you a whole lot, and it keeps you from being able to use things like flashplayer
[05:48] <penguin42> no I haven't yet - most stuff seems to be running fine; but as I say movie player is odd (which I've not seen before)
[05:48] <penguin42> I've got flash going using nspluginwrapper - that is reasonably good
[05:54] <jb-home> penguin42: Ah nice.  I had no luck with nspluginwrapper.
[05:55] <jb-home> penguin42: Is it hard to build?  It would be really nice to have that in the archive.
[05:55] <penguin42> jb-home: I got a prebuilt one - there is a bug filed somewhere asking for it to be packaged
[05:56] <penguin42> jb-home: It's not perfect (e.g. I see errors in the logs about some plugin trying to use the gnome theme engine and of course it's trying to pick up the 64bit one); and I've got a couple of flash issues I've never seen on a native flash
[05:59] <BenC> penguin42: have you installed ia32-libs-gtk?
[05:59] <penguin42> BenC: Yep
[05:59] <BenC> penguin42: also, you may be able to disable HPET in your BIOS and see if that helps
[05:59] <BenC> I've seen problems with it before
[06:00] <penguin42> I wouldn't bet that the BIOS can disable it - I assume there is a kernel param?
[06:35] <BenC> hpet=off
[06:35] <BenC> I know for sure at least 3 of my boxes can disable hpet in BIOS
[06:36] <penguin42> ok, I'll give that a go on my next reboot
[06:36] <penguin42> Thanks
[06:43] <BenC> np
[06:49] <penguin42> hmm I suspect HPET isn't used anyway - I found a test program in Documentation/hpet.txt  and whenever it tries to do anything there is an error in dmesg 'IRQ handler type mismatch for IRQ 8   current handler: rtc' and a backtrace
[07:06] <BenC> penguin42: actually, that looks more like it is used, since rtc module is already grabbing the IRQ
[07:07] <BenC> or not, hard to tell from that
[07:15] <penguin42> there are so many timer related messages in dmesg it's difficult for me to know - there is mention of APIC timer,then Using 14.318180 MHz WALL HPET GTOD HPET timer, ah then there is mention of hpet0 (3 timrs with IRQs 2,8,0) so I guess it is used
[08:16] <BenC> yay, lrm builds against 2.6.21 without a hitch
[08:17] <kylem> good work old bean.
[08:17] <kylem> :)
[08:18] <BenC> time to reboot and really give this a test
[08:19] <BenC> cheerio chap
[08:26] <BenC> Linux cunning 2.6.21-1-generic #1 SMP Sat Apr 21 13:37:35 EDT 2007 i686 GNU/Linux
[12:04] <simira> BenC: do you know if the laptop/hibernate/acpi problem were solved before release?
[12:05] <dade`> BenC: !!!
[12:05] <dade`> where are you ?
[12:09] <dade`> i can try now if you remember me that command.
[12:09] <dade`> for piix