[02:41] <mjg59> Hmm. There's a Yukon-2 driver in the netdev tree.
[02:43] <BenC> isn't that the broadcom variant that skge wont work for?
[02:45] <mjg59> Yes
[02:46] <BenC> netdev is in a git tree, isn't it?
[02:48] <jbailey> BenC: Will you suck in Chuck's patch for 11237 soonish?
[02:48] <mjg59> Hmm. Allegedly, anyway. I can't find it.
[02:50] <mjg59> Wuh?
[02:50] <mjg59> It's in -mm, but I can't find it in netdev's git tree
[02:50] <mjg59> Madness
[02:52] <BenC> I need checks baz archive url again
[02:52] <BenC> chucks
[02:55] <mjg59> Ah, found it
[02:56] <mjg59> BenC: http://www.kernel.org/git/?p=linux/kernel/git/jgarzik/netdev-2.6.git;a=commitdiff_plain;h=821dc07095d73b757bfc9d5f162e2efb954172f8;hp=caf39e87cc1182f7dae84eefc43ca14d54c78ef9
[02:56] <mjg59> Doesn't quite apply cleanly, but looks like the fixup is trivial
[02:56] <mjg59> If we don't have sk98lin yet, can we have that instead?
[02:57] <BenC> yeah
[02:58] <BenC> odd that it needs a seperate driver
[02:58] <BenC> makes me think of the whole 8139cp/8139too driver
[02:59] <BenC> I mean if syskonnect can make a single driver for all of them, why not the kernel devs :)
[03:00] <BenC> ok, I've got the diff, I'll add it to the repo today
[03:00] <BenC> thanks for the pointer
[03:11] <doko> BenC: could you have a look at http://bugzilla.ubuntu.com/show_bug.cgi?id=15092? I cannot reproduce that, neither in unstable nor breezy
[03:25] <mjg59> BenC: Could you possibly grab http://www.codon.org.uk/~mjg59/tmp/swsusp.diff ?
[03:25] <mjg59> Fixes bogus error checking and actually outputs some sensible information in failure cases
[03:38] <BenC> doko: me either, all those devs are created on my machines
[03:40] <BenC> doko: doing the test shown in the bug report works for me (all devs recreated)
[03:40] <BenC> maybe it's a udev bug?
[03:40] <BenC> the bug report shows 0.068, and I have 0.060 on my machine
[03:41] <BenC> mjg59: got it, will add it today
[03:41] <mjg59> Ta
[03:41] <doko> BenC: ok, thanks
[03:42] <BenC> anyone know why my initrd shows up as a ASCII cpio (reported by 'file')?
[03:42] <BenC> and how do I mount the thing manually
[03:42] <BenC> (ppc)
[03:43] <mjg59> BenC: Uhm. Because it's an initramfs?
[03:44] <mjg59> It's just uncpioed into a tmpfs
[03:44] <mjg59> You don't mount it
[03:44] <BenC> ah
[03:44] <BenC> always used an initrd with a filesystem  put into ramdisk
[03:44] <mjg59> Yeah. Now we have the future! :)
[03:45] <mjg59> initramfs allows all kinds of cool stuff - you can just combine multiple cpio files, and they'll all be extracted into the filesystem
[03:45] <BenC> for some reason, that makes me feel like an old fart
[03:45] <BenC> "I remember when we used to have a filesystem for the initrd"
[03:45] <mjg59> We've only just changed over
[03:45] <BenC> cool
[03:45] <jbailey> BenC: This is why I need your help on the ibook things.
[03:46] <jbailey> BenC: It looks like the kernels are doing the wrong things in thoses cases, it never ever runs /init
[03:47] <mjg59> Hm. Maybe I should try to get an ibook out of Mark.
[03:47] <mjg59> It would make some things much easier.
[03:47] <BenC> I should get an ibook for my laptop too (good excuse to get one)
[03:47] <jbailey> I think it's only older ones. =(
[03:48] <jbailey> mjg59: So you need another 6 laptops at your door.
[03:48] <BenC> jbailey: the problem seems to be related to the xfs filesystem
[03:48] <jbailey> BenC: I don't see how it could be.
[03:48] <jbailey> Nothing knows the filesystem at that point.
[03:48] <BenC> well, if it never gets to init then true
[03:49] <BenC> I can't see any of it being directly related to it being an ibook either though
[03:49] <BenC> I can't see how that could possibly affect it
[03:50] <BenC> you know, when I had this similar problem on my i386 that I upgraded from debian to breezy, I had an ext2 filesystem
[03:51] <jbailey> BenC: My theory was that OF was doing something weird with the initrd.
[03:51] <jbailey> But I can't even prove that.
[03:51] <BenC> I also don't think it is the kernel
[03:51] <BenC> he reported that rebuilding his 2.6.12-3 initrd caused that kernel to fail the same way
[03:52] <BenC> one thing is for sure though, the ramdisk_size is too small
[03:53] <BenC> hmm, I thought mjg59 said it was using tmpfs, is that newer than this bug report?
[03:53] <jbailey> BenC: No, he rebuilt it with initramfs and got it to fail.
[03:53] <jbailey> No, the changeover was late in the -6 cycle from mkinitrd to mkinitramfs
[03:53] <BenC> maybe I misunderstood the bug report
[03:54] <jbailey> My understanding is that every instance, old and new of him using mkinitramfs fails to execute /init on the initramfs.
[03:54] <jbailey> And that every instance of him using mkinitrd succeeds in doing the initrd bootup (/linuxrc)
[03:55] <jbailey> What I'd like is if there were some way to have the kernel printk the equivalent of an 'ls' in a test kernel before it chains
[03:55] <jbailey> Just to prove that the stuff is actually on the tmpfs.
[03:55] <jbailey> tmpfs/rootfs/ramfs, whatever it is.
[04:01] <jbailey> Woohoo, another vmware bug.
[04:01] <jbailey> Desperately need that BusLogic fix. =(
[04:01] <jbailey> And I don't understand how hotplug is picking it up now for VmWare installations.
[04:02] <BenC> vmware 5 works
[04:02] <BenC> vmware 4 doesn't
[04:03] <BenC> you want me to port ls to the kernel? :)
[04:03] <jbailey> BenC: Ah, does vmware 5 use something other than BusLogic?
[04:04] <jbailey> BenC: Well, I was hoping for an iteration over the directory entries, or something.  Dunno, I always imagines there'd be a debug helper fuction in there already for that.
[04:04] <BenC> yeah, it uses a fibre channel controller
[04:04] <jbailey> I can't imagine I'm the first one to want to know what the kernel thinks of a filesystem.
[04:04] <jbailey> BenC: Zul said he had a patch to make BusLogic hotpluggable.
[04:04] <BenC> hmm, my initrd has /init, not /sbin/init or /linuxrc
[04:04] <jbailey> Right.
[04:04] <BenC> yeah, I think he sent it to me
[04:04] <jbailey> initramfs boots with /init now.
[04:05] <BenC> or it's in his repo that I will sync from today
[04:05] <jbailey> Cool.
[04:05] <jbailey> When you have a test i386 kernel handy, can you let me know?
[04:05] <jbailey> I'll put out a call for testers to these people.
[04:05] <BenC> yeah, sure
[04:05] <BenC> will be late this evening before I have a build done
[04:06] <jbailey> Tjat
[04:06] <jbailey> That's fine.
[04:06] <jbailey> I need to afk for about 30m.
[04:08] <BenC> the quiet var in the kern command line for the initramfs needs to be a little more rebust
[04:08] <BenC> no way to to quiet=no or noisy to counteract it temporarily
[04:10] <BenC> does ybin need to be rerun if you edit /etc/yaboot.conf?
[05:11] <BenC> jbailey?
[05:13] <jbailey> I think so, yes.
[05:14] <jbailey> BEcause it needs to copy it to a partition type that OF understands.
[05:14] <jbailey> But that might just be for yaboot upgrades.
[05:14] <jbailey> (Likeyaboot might just know how to read it and where to find it)
[05:17] <BenC> that guy with the ibook fstype problem, is he on IRC?
[05:18] <BenC> jbailey: oh, and in regards to the G5 ppc64 kernel no rebooting, I can confirm and reproduce it
[05:18] <jbailey> Two of them are, yes.  Lemme dig up that
[05:18] <jbailey> BenC: Oh, is that a known bug (The G5)
[05:18] <jbailey> It's really annoying.
[05:18] <jbailey> I assumed I had b0rked my setup somehow.
[05:18] <BenC> my G5 with the colony 4 CD doesn't reboot when I tell it to
[05:18] <BenC> it says "Rebooting" and drops to a shell
[05:18] <jbailey> My g5 current breezy doesn't reboot.
[05:18] <jbailey> Well, I get a Ctrl-D to shell or enter to continue.
[05:18] <BenC> colony 4 live cd
[05:18] <jbailey> Like it's fliped to single user mode.
[05:19] <jbailey> Now that I know it's a real bug, I can troubleshoot it.  Do you have a bugzilla number?
[05:19] <BenC> probably because I'm using live-cd, there's no password
[05:19] <jbailey> The sideffect is that it also leaves the last mounted date on the volume wrong.
[05:19] <BenC> just goes right to shell
[05:19] <jbailey> So I have a 60gb fsck every boot.
[05:19] <BenC> heh, hold a sec
[05:19] <BenC> weird, someone just reported I bug I just experienced with loop device
[05:19] <jbailey> This machine is about due for a pave and replace, but I don't have other drives big enough to store the data in the eantime.
[05:19] <BenC> I think it's a mount bug though
[05:21] <BenC> jbailey: 14898 mentions it
[05:21] <BenC> and 15162 has the actual bug
[05:22] <jbailey> I've confirmed 15162 and added myself to the cc: list.
[05:23] <jbailey> I'll try to look at it later.
[05:23] <jbailey> I really need to squish a few initramfs bugs first.
[05:23] <BenC> I'm going to try to get a kernel for that ibook bug
[05:29] <jbailey> BenC: Thanks.
[05:42] <fabbione> MEHHH
[05:42] <fabbione> guys.. how do i disable a module from being loaded?
[05:42] <Lathiat> by hotplug?
[05:42] <fabbione> the isp2100 driver is borked
[05:42] <Lathiat> or in general?
[05:42] <mjg59>  /etc/hotplug/blacklist
[05:42] <fabbione> and it doesn't let me boot
[05:42] <fabbione> NO
[05:42] <fabbione> i need to do it at boot..
[05:42] <fabbione> first install
[05:42] <mjg59> Oh
[05:42] <mjg59> With trouble
[05:43] <fabbione> the driver is an isp1020
[05:43] <fabbione> or it seems like
[05:43] <Lathiat> fabbione: isp or ipw?w
[05:43] <fabbione> and i only have console here to irc
[05:43] <fabbione> isp
[05:43] <fabbione> BenC: dude?
[05:43] <fabbione> it's a scsi driver...
[05:43] <mjg59> fabbione: There's no kernel way of doing it. You'd need to ask kamion.
[05:46] <fabbione> i guess there is only one way for now..
[05:46] <fabbione> remove the controller...
[05:48] <BenC> yeah, probably the only way
[05:49] <fabbione> yeah
[05:49] <fabbione> too bad because it's a combo scsi + lan from a sparc64
[05:49] <fabbione> nice board
[05:49] <fabbione> and i only needed the lan to be hounest
[05:49] <Lathiat> btw, a boot option to exclude specific modules or somethign could possibly be usefull
[05:50] <BenC> debian-install/exclude-modules=isp1020,xxx,yyy
[05:52] <BenC> what would be nice is the installer to store something in nvram about which module it is about to load, and then on reboot it can say "Last time we tried to load module XXX, it failed, do you want to skip it this time?"
[05:53] <fabbione> hmm i am gonna try that
[05:53] <fabbione> i did swap the card already...
[05:53] <fabbione> worth a test
[05:53] <BenC> oh, that debian-installer thing was made up, it wont work :)
[05:53] <BenC> just a suggestion for implementation
[05:54] <fabbione> DIE!
[07:39] <jbailey> fabbione: Don't kill Ben, you'll have to do kernel work again if you do...
[08:14] <fabbione> jbailey: i need blood :P
[08:16] <jbailey> fabbione: Meat eater.
[08:17] <fabbione> ehhe
[08:25] <jbailey> Ugh, I tought 28 days were done with.
[08:26] <jbailey> How is it supposed to get below zero at this rate for November?
[08:44] <jbailey> do we support qemu on our kernels?
[09:00] <Lathiat> jbailey: qemu doesnt need qemu support unless you want the accelerator
[09:02] <jbailey> Ah, neat.
[09:03] <jbailey> 'k, I'll look at that then.
[09:03] <Lathiat> err
[09:03] <Lathiat> the second qemu was supposed to be 'kernel'
[09:18] <jbailey> It's all good. =)
[11:56] <Nafallo> morning all :-)
[11:58] <Nafallo> we have rt2400 and rt2500. maybe we should have rt2570 (usb rt2500) also? 
[12:00] <Nafallo> hmm, seems dead. I'll come back another day :-).