[03:47] <Chipzz> hi
[03:47] <Chipzz> do you guys realize that the latest initramfs setup has broken home-made kernels built with make-kpkg beyond all belief?
[03:48] <Chipzz> :/
[04:36] <Mithrandir> Chipzz: uh, how so?
[04:38] <Chipzz> networking not starting at all
[04:38] <Chipzz> which is due to /etc/network/ifstate not being removed
[04:39] <Chipzz> and output of e2fsck being spewed all over the console with brown * next to it
[04:39] <Chipzz> (like failure messages)
[04:44] <Mithrandir> that has nothing to do with custom or regular kernels.
[04:45] <zooko> I have a custom kernel built with make-kpkg, and I have different problems than that.
[04:58] <Chipzz> oh *sigh*
[04:58] <Chipzz> this time it fucks up in the initramfs
[04:58] <Chipzz> :/
[04:59] <Chipzz> this is exactly why I have home-made kernels: to avoid the need for any crap like initrds and initramfs'es *at all*
[04:59] <Chipzz> :(
[05:01] <Mithrandir> maybe it fucks up because you did something silly when compiling your kernel, then?
[05:11] <Chipzz> this kernel has worked ever since I installed ubuntu, and before that several months on a debian install
[05:12] <Chipzz> (to be more precise: I recompiled the kernel form my debian system with the same config on ubuntu with make-kpkg)
[05:12] <Mithrandir> yes, and?  If you're not using initrds, the initrd/initramfs can't break your boot.
[05:13] <Chipzz> the reason it breaks is exactly because initramfs being run
[05:13] <Mithrandir> don't use initramfs, then.
[05:13] <Chipzz> ok, little question...
[05:14] <Chipzz> why is /dev mkdir'ed in the initramfs?
[05:14] <Mithrandir> because else it won't exist? :-)
[05:14] <Chipzz> duh :)
[05:14] <Chipzz> but why don't you put it in the initramfs in the first place>?
[05:14] <Mithrandir> that'd work too.
[05:15] <Chipzz> I was going to suggest the following:
[05:15] <Chipzz> on debian I had devfs earlier
[05:15] <Chipzz> but just in case that crapped up
[05:15] <Mithrandir> newer kernels don't support devfs, though.
[05:15] <Chipzz> I had the necessary device nodes in /dev
[05:15] <Chipzz> (for the root filesystem)
[05:16] <Chipzz> so even if for some reason the devfsd maintainer screwed up big time
[05:16] <Chipzz> at the very least the / fs would still mount
[05:17] <Mithrandir> well, we discover stuff at boot-time rather than relying on /dev/sda being the right node each time.
[05:17] <Chipzz> maybe a good idea to actually ship /dev in the initramfs, with the necessary device nodes to mount / (and possibly /usr and /var and /tmp?)
[05:17] <Mithrandir> or rather, you need to load the relevant module 
[05:17] <Mithrandir> what's needed to mount the root filesystem?
[05:17] <Chipzz> one device node ;P
[05:18] <Chipzz> in my case, /dev/hda1 ;)
[05:18] <Mithrandir> udev creates that when the device becomes available
[05:18] <Chipzz> yes
[05:18] <Chipzz> and what if the udev maintainer fucks up and udev doesn't start?
[05:18] <Chipzz> ;)
[05:19] <Chipzz> thing is
[05:19] <Mithrandir> then your system becomes unbootable, but you switch back to your previous initramfs and boot happily using that?
[05:19] <Chipzz> you *know* what device node will be needed to mount / at initramfs creation time
[05:19] <Mithrandir> no, you don't.
[05:20] <Chipzz> wouldn't that be the same device node your / fs is mounted at at that moment?
[05:21] <Mithrandir> what's if your root device is an USB hard drive?
[05:23] <Chipzz> ok, in most sane cases... ;P
[05:27] <Mithrandir> installing to an USB hard drive is not a sane case?  A live cd is not a sane case?
[08:04] <dilinger> i dunno who normally handles this, but i just uploaded ndiswrapper 1.8 to debian
[08:04] <dilinger> there are misc bugs in 1.5 (which is in dapper) that are fixed in 1.8
[08:04] <dilinger> might want to consider resynching w/ debian
[08:05] <BenC> we don't sync that package with debian, plus we are in an upstream feature freeze
[08:05] <BenC> but thanks
[08:05] <BenC> version freeze I mean
[08:07] <dilinger> ok
[08:07] <dilinger> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=338309
[08:07] <dilinger> stuff like that is what you'll get when using ndiswrapper 1.5 w/ a newer kernel
[08:07] <BenC> how new?
[08:08] <BenC> we aren't going past 2.6.15 for dapper
[08:08] <dilinger> that bug report is with 2.6.14
[08:08] <dilinger> 1.5 is quite old
[08:08] <BenC> 1.5 was released not too long after dapper started, which was only a few months ago
[08:09] <dilinger> well, in terms of kernel releases, it's old
[08:09] <zooko> Heh.
[08:09] <dilinger> i assume you had to patch it to make it compile?
[08:10] <dilinger> looks like 1.5 was released on halloween
[08:10] <BenC> don't think so, maybe just the USB URB thing
[08:11] <dilinger> a bunch of people compiled that 1.5 didn't compile w/ 2.6.14
[08:11] <dilinger> complained, rather :)
[08:11] <dilinger> anyways, just a suggestion
[08:12] <BenC> ok
[08:29] <zul> heylo
[08:56] <mxpxpod> BenC: ping
[10:04] <dilinger> ugh
[10:04] <dilinger> this dapper box is having issues
[10:04] <dilinger> disk read issues :/
[10:54] <dilinger> rock!
[10:55] <dilinger> i just upgraded to the latest initramfs/kernel packages in dapper
[10:55] <dilinger> it looks like it fixed all the udev problems
[10:55] <dilinger>  /dev/null and friends now even have the correct permissions
[11:23] <BenC> mxpxpod: pong