[12:34] <lamont> BenC: if I'm silly enough to want to try 2.6.15 on breezy....
[12:51] <jbailey> lamont-away: 2.6.15 requires a newer udev.
[01:36] <zul> heylo
[02:03] <mjg59> BenC: No idea
[02:03] <mjg59> It's a bit weird looking
[02:19] <BenC> mjg59: what should i do with that bug? :)
[02:20] <BenC> lamont-away: hppa kernels built, but they aren't in the archive yet...any idea why?
[02:20] <mjg59> BenC: Erm... :)
[02:20] <mjg59> Honestly don't know (assuming you're talking about the weird USB one)
[02:20] <BenC> no, the IDE suspend one
[02:21] <mjg59> Oh
[02:21] <mjg59> That's not the bug number you pasted
[02:21] <BenC> let me check it
[02:22] <BenC> 10314, sorry
[02:22] <mjg59> Oh, that
[02:22] <mjg59> Yeah, I can knock up a patch for it
[02:22] <mjg59> It's pretty easy
[02:22] <BenC> ok
[02:32] <lamont-away> BenC: NEW
[02:32] <lamont-away> is my guess
[02:42] <BenC> lamont-away: sparc and ia64 went in pretty quickly
[02:42] <BenC> everything is built, but hppa is the only one not in the archive
[02:48] <lamont-away> I think it's a matter of timing...
[02:48] <BenC> yeah, just odd that it is taking so long
[03:30] <jbailey> BenC: Got bandwidth? =)
[03:30] <jbailey> BenC: If you could test a klibc build on sparc, I'd appreciate it.
[03:36] <BenC> from archive?
[03:38] <BenC> gcc-3.3? :)
[03:39] <BenC> jbailey: should it really be using linux-headers-2.6.12-9?
[03:39] <jbailey> Nah, the one I'm updating.
[03:39] <jbailey> Lemme post it, just a sec.
[03:41] <jbailey> BenC: The gcc-3.3 is a gcc bug.  With my sparc unstable, I probably can't fix it, though.
[03:41] <BenC> I'll try building with gcc-4
[03:41] <BenC> gcc-3.4 doesn't work either?
[03:41] <jbailey> It should be an easy fix, libgcc is leaking a reference to .umul and shouldn't be.
[03:41] <jbailey> Nope, gcc-3.4 and 4.0 both have the bug.
[03:42] <jbailey> My patience for toolchain hacking on an U5 is strained at the best of times. =)
[03:42] <BenC> does it fail when linking?
[03:43] <BenC> I'll do a compile with gcc-3.3, and look at the .umul problem later
[03:43] <jbailey> Yes.
[03:47] <jbailey> Hmm.  I just updated to the 2.6.15-7 headers instead of 2.6.12-9 and it broke locally again.  I might need to fix this tomorrow.
[03:47] <BenC> still want me to do a sparc build?
[03:48] <jbailey> Not yet. =)
[03:48] <jbailey> I will, though.  Unless you're heading out?
[03:48] <BenC> nah, I'll be here for anohter hour or so
[03:55] <jbailey> Hmm
[03:55] <jbailey> I think I might go spend time with my wife, though.
[03:55] <jbailey> I'll catch up with this in the morning.
[03:55] <jbailey> I've already sent 3 klibc patches in today. =)
[06:27] <infinity> Not sure how much of a win NUMA would be a dual CPU amd64.
[06:27] <HrdwrBoB> I'm not sure either
[06:28] <HrdwrBoB> It'd be interesting to do some comparative benchmarks
[06:28] <HrdwrBoB> but realistically - why not
[06:28] <infinity> I know there were stability issues with it in the past, that would be my only argument against enabling it unconditionally.
[06:29] <HrdwrBoB> the only documented stability issues I can find date back to 2.4/2.5
[06:29] <fabbione> morning
[06:29] <HrdwrBoB> though obviously that's no exactly exhaustive testing
[06:29] <fabbione> HrdwrBoB: it's only overhead if you don't have a real NUMA system
[06:29] <HrdwrBoB> yes
[06:29] <fabbione> so pointless in a dual CPU amd64
[06:30] <infinity> fabbione : But, as has been pointed out, all amd64 systems with >1 CPU are NUMA.
[06:30] <infinity> (Because of the embedded memory controller on the CPU)
[06:30] <infinity> A quick Googling taught me that Linux actually pulls some tricks to assign RAM ownership per CPU.
[06:30] <fabbione> is it actually used in such config?
[06:31] <infinity> Learn something new every day.
[06:31] <fabbione> also my BIOS claims that i have a HT cpu
[06:31] <fabbione> but my mobo can't use it
[06:31] <HrdwrBoB> all smp opteron boards have the same features though
[06:32] <HrdwrBoB> there isn't any low-end cheap dodgy motherboards for dual opterons that I've seen
[06:32] <HrdwrBoB> and even if they were, I'm not sure how they wouldn't implement it
[07:10] <fabbione> BenC: did you just reboot the e3k?
[07:32] <fabbione> BenC: meh.. ok i found a way to have a reproducible crash :/
[10:00] <infinity> mjg59 : Is vbetool supposed to build/work on ia64?..  If so, can you fix it?.. If not, can you make acpi-support stop requiring it on ia64?
[10:01] <Mithrandir> it is supposed to work.
[12:16] <siretart> hi
[12:16] <siretart> didn't we agree that we don't ship kernel-patches in ubuntu (not even in universe)?
[12:17] <fabbione> yes...
[12:17] <siretart> I notice that kernel-patch-vserver is being synced over and over, so this seems to be a good candidate to be blacklisted/removed
[12:17] <siretart> but I wanted to ask here first
[12:20] <fabbione> it seems sensible
[12:23] <siretart> what is the procedure for this? just asking elmo to remove/blacklist it?
[12:23] <siretart> or filing a bug?
[12:24] <fabbione> ask elmo i think
[12:24] <siretart> ok
[02:05] <BenC> fabbione: e3k died, thanks, now I have to walk in the rain to the cold barn to power cycle it :)
[02:05] <fabbione> BenC: dude.. it's a reproducible crash
[02:05] <fabbione> it did it twice
[02:06] <fabbione> but the first time it managed to resurrect
[02:06] <fabbione> sorry :(
[02:06] <BenC> hehe, no problem :)
[02:06] <BenC> I've caused it to lock 10 times in a day once, so once for you is ok :)
[02:06] <jbailey> BenC: What about the big breaker to out there you talked about? =)
[02:06] <jbailey> Or is that only for when it's snowing? =)
[02:07] <BenC> rebooting all the machines is more of a pain
[02:07] <BenC> brb
[02:19] <BenC> fabbione: do you know how to reproduce the crash?
[02:19] <fabbione> BenC: yeps :)
[02:19] <fabbione> i was trying to use the machine a bit more
[02:19] <fabbione> created 3 chroots
[02:20] <fabbione> 2 started building gcj-4.0 and gcj-4.1
[02:20] <fabbione> the 3rd one was unpacking the B-D for ghc6
[03:57] <zul> heylo
[06:03] <AcidPils1> hi
[06:03] <AcidPils1> http://paste.ubuntulinux.nl/5558  any ideas about that? 
[06:04] <AcidPils1> infinity: the firmware doesn't work :( alike all other i tryed :(
[09:31] <zul> its quiet today