[01:21] <mjg59> bcollins: Around?
[01:22] <mjg59> bcollins: I suspect there may have been a mismerge when you were trying the new ACPI patches
[01:22] <mjg59> (The ones you reverted, that is)
[01:22] <bcollins> shouldn't have been
[01:22] <bcollins> I did git-revert
[01:22] <mjg59> bcollins: On boot, I get a pile of "Could not allocate new owner id", and then ACPI gets disabled
[01:22] <bcollins> in reverse order
[01:22] <mjg59> bcollins: When you were applying them in the first place
[01:23] <mjg59> Before you reverted them
[01:23] <mjg59> I've just reapplied them to figure out what went wrong
[01:23] <bcollins> ah
[01:23] <bcollins> it's probably got something to do with the 64bit long id thing to 256 max
[01:23] <mjg59> We had a patch applied to deal with the owner_id stuff, which conflicted with the acpi-ca patches
[01:23] <bcollins> that was the only real conflict in applying them
[01:23] <mjg59> Yeah
[01:23] <bcollins> right
[01:23] <mjg59> So I'm suspicious about these errors
[01:29] <mjg59> bcollins: Indeed - just found a line that was missing
[01:29] <mjg59> acpi_gbl_next_owner_id_offset never gets set back to 0, so everything explodes
[01:31] <BenC> ah
[01:31] <mjg59> And yes, that works much better now
[01:32] <mjg59> Though pnpacpi has got broken - looks like there's a patch for that, though
[01:32] <mjg59> BenC: Any chance we can have those back, then? :)
[01:32] <BenC> yeah, want to send me your diff? :)
[01:33] <mjg59> Hmm. Though interrupt routing still seems a touch confused.
[01:35] <mjg59> Yeah, a pile of stuff is ending up on irq 0. Hm.
[01:43] <mjg59> BenC: How can I disable MSI? I'm sure there's a boot parameter, but I can't find it
[01:49] <mjg59> Ah - actually, this could be due to slightly changed acpi semantics - I may just need to rebuild all my modules...
[02:16] <mjg59> BenC: Ok, got a fix for the interrupts
[02:23] <mjg59> BenC: So I've sent you two patches, both to be applied after the acpi-ca patches are un-reverted. One fixes ACPI owner-id stuff, and the other fixes gsi allocation so PCI devices get interrupts
[02:39] <zul> heylo
[02:39] <zul> BenC: dont be goofy...thats evil..:)
[02:43] <BenC> zul: I have macox running and using my rt2500 pci wireless card
[02:44] <zul> sweet..
[02:44] <zul> can you dual boot with it?
[02:44] <BenC> yeah, I have it dual booting with ubuntu
[02:44] <BenC> it's cool, the osx86 website has a howto for ubuntu install/dualboot of osx86 :)
[02:44] <BenC> with a caption that says "the one that started it all"
[02:44] <zul> mmmmmm...need osx
[02:45] <BenC> we need to figure out how osx does it's splash screen, it looks the same as it does on ppc hw
[02:45] <BenC> and it looks perfect on my nvidia/tv-out mame system, whereas our usplash looks a little corrupted
[02:46] <zul> hmm..maybe i should do it at work...tell my boss is for educational purposes
[02:47] <mjg59> BenC: Where did you get an rtl2500 driver?
[02:47] <BenC> it'll need a p4 or athlon system (sse2 minimum, sse3 prefered)
[02:47] <mjg59> Or is it just ported from BSD?
[02:47] <BenC> mjg59: posted on a osx86 forum from someone who got it from the hw vendor
[02:48] <BenC> it's a native driver
[02:48] <mjg59> Sweet
[02:48] <BenC> i386/ppc fat
[02:48] <mjg59> I note that the iMacs still have Broadcom wireless
[02:48] <mjg59> BenC: Does it do suspend/resume?
[02:49] <BenC> hmm...haven't tried that yet
[02:49] <mjg59> Oh, Pentium Ms have sse2 too
[02:50] <BenC> oh crap
[02:50] <BenC> I almost forgot, I've been wanting to test the darwine
[02:50] <BenC> it's x86 only right now, but it's the whole reason I installed osx86
[02:51] <BenC> it doesn't seem to support my built-in UHCI host controller either
[02:51] <BenC> had to install a OHCI card
[02:52] <BenC> UHCI is supported but my 82801 chipset seems to not work well (as reported by others too)
[02:54] <BenC> oh wow, now that I spent all this time installing osx86, there's a ppc version of darwine that runs win32 apps
[02:54] <BenC> using qemu
[02:58] <mjg59> Haha
[03:19] <zul> yay i have qemu working again
[05:40] <bluefoxicy> you know
[05:41] <bluefoxicy> the amd64 kernels in dapper are -VERY- -UNSTABLE-.
[05:41] <bluefoxicy> I can barely stay up 15 minutes.
[06:02] <infinity> They're stable for me.  Guess I'm just lucky...
[06:02] <infinity> (Or you have hardware we don't like... Have you filed bugs?)
[06:06] <bluefoxicy> I can't determine the root cause to file a bug.
[06:06] <bluefoxicy> Last time I installed (about 1 hour ago) it froze when I tried to even boot a -k8 kernel, screen got green lines
[06:06] <bluefoxicy> now, I have to boot a k8 kernel, the generic 2.6.15-15 does the same.
[06:07] <bluefoxicy> this means the problem is not repeatabel.
[06:07] <bluefoxicy> mm ok
[06:07] <bluefoxicy> I can address 85TiB of virutual memory space
[06:08] <bluefoxicy> I should not have problems with gimp dying when I hit 1GiB of allocated shit anymore.
[06:08] <fabbione> bluefoxicy: try to boot with noapic
[06:09] <bluefoxicy> fabbione:  i will the next time this thing freezes.
[06:09] <bluefoxicy> fabbione:  would that be localized to x86-64 only?  I have been running a 32 bit dapper stabilly for quite  while.
[06:11] <fabbione> it might be
[06:12] <bluefoxicy> fabbione:  is there anything that need be said about JFS re not using it ever?
[06:13] <fabbione> bluefoxicy: -EPARSE
[06:13] <bluefoxicy> I tried to install with a jfs root, it seemed slightly tempermental. . . the system froze 2 or 3 times, and after that last one the dpkg database was hosed (I did not use apt or synaptic on that boot), the GDM Human theme was destroyed (was not written to at any point during that boot cycle), and gnome ceased to function properly (panel wouldn't start)
[06:13] <fabbione> dunno i don't use JFS
[06:14] <fabbione> file a bug
[06:14] <bluefoxicy> so i'm on xfs now, which I have experienced things such as reset-button-while-writing etc etc and have a /home on xfs that's like forever old and has been through many failures and never lost data
[06:14] <bluefoxicy> and I'm thinking, "Damn, it looks like JFS might just be flakey."
[06:14] <bluefoxicy> fabbione:  yes but what kind of bug?
[06:14] <bluefoxicy> "JFS appears to be very shakey and easy to damage, it should probably never be used" doesn't seem very descript.
[06:15] <fabbione> a bug in which you write whatever is happening?
[06:15] <fabbione> Ben will ask you for info
[06:15] <fabbione> to see what he needs to know
[06:15] <bluefoxicy> alright
[06:16] <bluefoxicy> out of curiousity, do any of the ubuntu devs use jfs?
[06:16] <fabbione> i don't
[06:16] <fabbione> i am not stupid enough to use these useless filesystem
[06:16] <infinity> I just use ext3, since that's the default.  I'm lazy that way.
[06:17] <bluefoxicy> I installed with ext3 recently (a few months back) and it ate 900 megs more space or so o_o
[06:17] <infinity> I used to use xfs, when I was more interested in tweaking to the Nth degree.
[06:17] <bluefoxicy> I used to use reiserfs, when I didn't mind using piles of trash o.o
[06:17] <infinity> I'd recommend ext3 or xfs to anyone, tell poeple to stay way the hell away from reiser, and I have no real opinion of jfs.
[06:17] <bluefoxicy> trash that works pretty well, but still trash
[06:18] <bluefoxicy> heh, you noticed it's a horribly implemented maze of shit too huh?
.. The code could ber perfectly elegant and lovely, and it wouldn't change the fact that it eats filesystems for breakfast.
[06:18] <infinity> s/ber/be/
[06:19] <bluefoxicy> eh, reiser always seemed stable to me integrity-wise
[06:19] <bluefoxicy> but its underlying design is a bastardization
[06:19] <infinity> You were lucky. :)
[06:19] <bluefoxicy> hans reiser was and is like
[06:19] <bluefoxicy> "OMG, XATTRS?  NOOOOOOOOOOOOO IDIOT."
[06:20] <bluefoxicy> that was the beginning.  Let's not implement what normal people implement in our own way; let's change what we want to say a file system is and does.
[06:20] <bluefoxicy> it went farther with Reiser4; someone described Reiser4 as having a symptom where "all directories are symlinks" or something
[06:21] <bluefoxicy> besides the fact that Reiser4 added a new syscall that goes DIRECTLY into reiser4's code, past the kernel vfs
[06:22] <bluefoxicy> when I started to realize what a psychopath hans was, I decided I should move away from reiserfs
[06:22] <bluefoxicy> Right now I think he's an apprentice to other famous loons, including Theo de Raadt and RMS
[06:22] <bluefoxicy> I'm sure he'll master the art of lunacy in time.
[06:26] <bluefoxicy> bug filed.
[06:27] <bluefoxicy> fabbione:  any ideas on the grub-install bug btw?
[06:27] <bluefoxicy> fabbione:  grub-install hangs if you run it with /boot on xfs
[06:27] <fabbione> known xfs problem
[06:28] <fabbione> sometimes searching and reading docs will be helpful for you
[06:28] <bluefoxicy> fabbione:  yes but what confuses me is
[06:28] <bluefoxicy> I just run grub-install to prepare /boot/grub
[06:28] <bluefoxicy> then kill it
[06:28] <bluefoxicy> and run grub
[06:28] <fabbione> did you read the documentation?
[06:28] <bluefoxicy> "root (hd0,6)"  (or whatever partition is / with /boot on it) and "setup (hd0)" and it's good to go.
[06:28] <fabbione> if so you wouldn't be confused
[06:28] <bluefoxicy> no, not really
[06:29] <bluefoxicy> do you have a link
[06:29] <fabbione> it's in grub-install each time you attempt to run it
[06:29] <bluefoxicy> um, grub-install doesn't give me output
[06:29] <fabbione> that means you didn't even read the output from grub-install
[06:29] <fabbione> it does
[06:30] <bluefoxicy> Due to a bug in xfs_freeze, the following command might produce a segmentation
[06:30] <bluefoxicy> fault when /boot/grub is not in an XFS filesystem. This error is harmless and
[06:30] <bluefoxicy> can be ignored.
[06:30] <bluefoxicy> fabbione:  not helpful.  *digs at command*
[06:30] <fabbione> bbl
[06:31] <bluefoxicy> oh come on
[06:31] <bluefoxicy> you give me a load of crap and then run off
[06:31] <bluefoxicy> so does anyone else know?
[06:31] <fabbione> bluefoxicy: i need to go to the bathroom.. do you mind?
[06:32] <bluefoxicy> oh fine
[06:32] <fabbione> or should i send you to #ubuntu
[06:32] <fabbione> ?
[06:32] <fabbione> you decide
[06:32] <bluefoxicy> just go
[06:33] <bluefoxicy> i'm gonna sleep
[06:33] <bluefoxicy> I still find it odd though that grub-install freezes but grub works perfectly
[06:33] <bluefoxicy> more odd that some idiot wrote an xfs_freeze call into grub-install
[06:34] <bluefoxicy> considering if /boot is on / the result is grub-install -> xfs_freeze -> access / -> xfs is frozen -> deadlock
[06:34] <bluefoxicy> (happens during the ubuntu installation process if you try to install grub in said situation)
[12:24] <Mithrandir> is there any way to extract the contents of a tmpfs in such a way that you can shove it into a file with ext[23]  or something?
[12:25] <fabbione> uh?
[12:25] <Mithrandir> I have a tmpfs unionfs mounted on top of another file system.
[12:25] <fabbione> ok...
[12:25] <Mithrandir> I'd like to pull all the files out of the tmpfs and put them somewhere (in an ext2 filesystem)
[12:26] <fabbione> tar ?
[12:26] <Mithrandir> how can I get the stuff which is just in the tmpfs?
[12:26] <fabbione> ahhh
[12:26] <fabbione> that way..
[12:26] <Mithrandir> note the "unionfs" part. :-)
[12:26] <fabbione> i don't think you can at all
[12:27] <fabbione> or perhaps you can use timestamps?
[12:27] <fabbione> see at what time you mount
[12:27] <fabbione> and copy everything that has been modified since than?
[12:27] <Mithrandir> ew.  I guess that'd be possible, though ugly.
[12:29] <fabbione> i don't think you have any other choise
[12:29] <Mithrandir> hmm, I think I can work around this.
[12:29] <Mithrandir> it'll be slow, but work.
[01:34] <fabbione> BenC: ping?
[02:00] <BenC> fabbione: pong
[02:00] <fabbione> hey BenC 
[02:00] <fabbione> git pull rsync://rsync.kernel.org/pub/scm/linux/kernel/git/davem/ubuntu-niagara-2.6.git
[02:00] <BenC> hey
[02:00] <fabbione> it's almost there :)
[02:00] <BenC> sweet
[02:00] <fabbione> he is doing a test build
[02:00] <BenC> dave rocks
[02:01] <fabbione> i am working on a prebeta installer
[02:01] <fabbione> http://vger.kernel.org/~davem/cgi-bin/blog.cgi/index.html
[02:01] <fabbione> he does :)
[02:01] <mjg59> BenC: Those acpi patches look ok?
[02:01] <fabbione> BenC: he said to wait to pull.. but i am testing locally too...
[02:02] <BenC> mjg59: just woke up, getting to them in a little bit
[02:02] <mjg59> BenC: No problem
[02:04] <BenC> [    9.389903]  Total of 32 processors activated (9895.93 BogoMIPS).
[02:04] <BenC> gotta love that
[02:04] <BenC> quad gige too
[02:08] <fabbione> yeah
[02:09] <fabbione> but there is a lot of tuning that needs to be done
[02:09] <fabbione> look at the calibration delay loops per CPU and bogomips
[02:09] <fabbione> some CPU's come up to 410 Bogo
[02:09] <fabbione> other at 280
[02:14] <BenC> fabbione: how did I get an email for upload of 9.24 that I did 4 months ago?
[02:15] <fabbione> BenC: -ENOCLUE
[02:15] <fabbione> ask Kinni or infinity?
[02:15] <BenC> it was uploaded to launchpad, and aimed at breezy-updates
[02:16] <BenC> and it used my gpg key :)
[03:30] <Mithrandir> BenC: where did you see the squashfs 3.0 stuff?
[03:33] <BenC> Mithrandir: no
[03:33] <BenC> is it in?
[03:34] <BenC> yes, I did
[03:34] <BenC> Mithrandir: it's in CVS
[03:35] <BenC> :pserver:anonymous@cvs.sourceforge.net:/cvsroot/squashfs
[03:35] <BenC> module is squashfs
[03:35] <BenC> they may have a tarball, but I just grabbed CVS
[03:36] <BenC> already tested booting the liveCD with v3 filesystem, so it should be good
[03:36] <Mithrandir> they haven't released anything, so I'm a bit wary of just grabbing CVS.
[03:40] <BenC> well, v2 is broken we know that, and v3 booted fine with no oopses, so I think the best bet is to move to v3
[03:40] <BenC> Mithrandir: you can always repackage squashfs tools with a v2 and v3 mksquashfs
[03:40] <mxpxpod> BenC: so, I've figured out that bcm43xx is causing syslog to go crazy
[03:41] <BenC> mxpxpod: yeah, I've heard, I removed some debug output from ieee80211 in the next kernel upload
[03:41] <BenC> Mithrandir: and then you could just use v3 for ppc
[03:41] <mxpxpod> BenC: no, I mean that it's causing syslogd to spawn new processes
[03:41] <BenC> mksquashfs == v2, mksquashfs3 == v3
[03:41] <BenC> mxpxpod: oh, I don't see how that's possible
[03:41] <mxpxpod> BenC: I don't either
[03:42] <mxpxpod> BenC: but last night I had about 50 syslogd's running
[03:42] <Mithrandir> BenC: true dat.
[03:42] <mjg59> mxpxpod: If you're using our suspend script, it ought to be unloaded and reloaded over suspend
[03:42] <mjg59> mxpxpod: Is that happening?
[03:42] <mxpxpod> mjg59: which suspend script?
[03:42] <mjg59> Oh, is this on PPC?
[03:42] <mxpxpod> mjg59: yup
[03:42] <mjg59> That's probably more complicated
[03:43] <mxpxpod> BenC: and the syslogd's didn't spawn until I ifup'ed my ae
[03:47] <mxpxpod> at first I thought it might be a daemon I was running that was doing it (apache, avahi, postgresql, etc.)... but I shut all those down and it still happened
[04:17] <fabbione> BenC: do you think you have time to pull from Dave and test the resulting kernel on your e3k ?
[04:18] <BenC> yeah
[04:18] <fabbione> just to check if there are no "easy to spot" regression
[04:18] <fabbione> i will do the same here as soon as i am done with a d-i image for hom
[04:18] <fabbione> him
[05:13] <BenC> mjg59: started a build with acpica patches un-reverted plus the two you sent me
[05:14] <mjg59> BenC: Thanks
[05:22] <zul> heylo
[05:23] <zul> BenC: i should have some patches for you this weekend as well
[05:27] <BenC> hopefully I'll be doing an upload later tonight
[05:27] <zul> ok...
[05:37] <fabbione> BenC: did you remember to pull from me?
[05:37] <BenC> not yet, but I will
[05:37] <fabbione> thanks
[07:42] <zul> heh...bug #31698 is funny
[09:01] <BenC> mjg59: kernel seems to boot fine, but I see a few of these
[09:01] <BenC> [4294671.887000]  pnp: PnPACPI: unknown resource type 7
[09:01] <BenC> [4294671.887000]  pnp: PnPACPI: METHOD_NAME__CRS failure for PNP0c01
[09:03] <mjg59> BenC: Yeah. I think it's basically harmless, but I'll get you a patch for that
[09:03] <BenC> ok
[09:03] <crimsun> mjg59: belated pong
[09:03] <mjg59> crimsun: Hi - you know alsa magic, right?
[09:03] <crimsun> mjg59: depends on the card(s)
[09:03] <crimsun> mjg59: which issue(s)?
[09:06] <mjg59> crimsun: i810
[09:07] <mjg59> crimsun: On an HP 6220 laptop - after suspend, the ac97 registers are identical but no sound comes out of the speakers
[09:07] <mjg59> Headphones work, though
[09:07] <mjg59> It has a headphone sense switch, but I made sure that the state of that was being restored
[09:07] <mjg59> It's an ad19something codec
[09:08] <crimsun> ah, that's the exact opposite of the ThinkPad issue. I've picked a PM support patch and will test locally.
[09:08] <crimsun> (adds PM support to intel8x0)
[09:09] <mjg59> Cool. 
[09:09] <mjg59> The Thinkpad one doesn't seem to apply to all Thinkpads - my X40 is fine over suspend/resume
[09:09] <mjg59> Have you got a patch I can test?
[09:09] <mjg59> (I have the hardware)
[09:10] <crimsun> not yet, let me get a url so you can look at the upstream changes
[09:10] <mjg59> Thanks
[09:24] <crimsun> mjg59: (sorry, got called away) http://cvs.sourceforge.net/viewcvs.py/alsa/alsa-kernel/pci/intel8x0.c?r1=1.228&r2=1.229
[09:26] <crimsun> (back to conference)