**** BEGIN LOGGING AT Thu Oct 27 02:59:58 2011
Oct 27 06:38:03 <PaulFertser> Hey Psi 
Oct 27 09:34:41 <michelem> aloha folks
Oct 27 09:35:10 <michelem> PaulFertser: I had the opportunity to test the Qi patch I sent in. Confirming it works.
Oct 27 09:37:24 <PaulFertser> michelem: cool, thanks.
Oct 27 09:38:56 <michelem> PaulFertser: you are not debian-based, are you?
Oct 27 09:48:41 <PaulFertser> michelem: i'm using only Debian with self-compiled qi, kernel, and FSO2.
Oct 27 09:48:52 <PaulFertser> And with Emacs UI.
Oct 27 09:49:40 <Sazius> wow, Emacs UI :)
Oct 27 09:50:40 <PaulFertser> The only telephony UI i'm using since long.
Oct 27 09:51:35 <michelem> I've been trying to build a read only system based on unionfs, but no unionfs drivers are in the kernel
Oct 27 09:51:39 <michelem> any tip on that?
Oct 27 09:52:23 <Sazius> PaulFertser: how does that work? with the virtual keyboard??
Oct 27 09:53:07 <PaulFertser> Sazius: yep, and i can press the AUX button to answer the currently pending call.
Oct 27 09:53:36 <Sazius> wow :)
Oct 27 09:53:39 <PaulFertser> michelem: recompiling the kernel probably, what can i say...
Oct 27 09:53:58 <michelem> anyone ever used union filesystems here?
Oct 27 09:54:05 <PaulFertser> michelem: i also like the idea of the combined squashfs+jffs2 images that's used by OpenWrt.
Oct 27 09:54:22 <michelem> ubuntu favors aufs but some tell me it's not very stable
Oct 27 09:54:25 <Sazius> PaulFertser: is there some particular telephony-mode or similar for emacs ?
Oct 27 09:54:40 <PaulFertser> michelem: i guess everybody has, booting some livecds :)
Oct 27 09:54:46 <PaulFertser> Sazius: the one i wrote myself, yes.
Oct 27 09:55:14 <PaulFertser> Sazius: http://markmail.org/thread/mugcsgznpaj76yqv
Oct 27 09:56:26 <Sazius> PaulFertser: wow, that's pretty cool :) thanks for the link
Oct 27 09:56:55 <PaulFertser> Sazius: if the code doesn't work with latest FSO, just drop me a note, i'll try to promptly fix it.
Oct 27 09:57:03 <michelem> PaulFertser: aufs fails to patch on the kernel tree advised on OpenMoko. What kernel version are you with?
Oct 27 09:57:31 <PaulFertser> michelem: i'm stuck with 2.6.29 andy-tracking since i stopped following the kernel development effort :(
Oct 27 09:57:47 <michelem> ok.
Oct 27 10:03:07 <Wonka> I'd like a newer kernel than the 2.6.34 that's latest in debian...
Oct 27 10:03:43 <Wonka> also, I'd like a way to use internet via gprs and do phone calls under debian. as a debian package.
Oct 27 11:24:37 <michelem> anyone tried patching the sources with AuFS? I can't figure out which aufs version to combine with the kernel
Oct 27 11:26:23 <michelem> and if I get the full already-patched repository from aufs, it won't cross-compile failing "arm-angstrom-linux-gnueabi-ld: unrecognized option '-Wl,-rpath-link,/usr/local/openmoko/arm/arm-angstrom-linux-gnueabi/lib'"
Oct 27 13:10:39 <michelem> PaulFertser: I need some help with a kernel nuisance
Oct 27 13:11:26 <michelem> I fetched the 2.6.34 kernel sources from http://pkg-fso.alioth.debian.org/debian/pool/main/l/linux-2.6-openmoko/ , and the Aufs2 standalone module (see http://aufs.sourceforge.net/README.aufs2 )
Oct 27 13:11:27 <PaulFertser> michelem: well, i can try...
Oct 27 13:12:05 <michelem> the kernel compiles fine. When compiling the aufs2 module, instead, I get this error: http://dpaste.de/iVGkO/
Oct 27 13:12:44 <michelem> so I added #include <linux/workqueue.h> on top of the failing file wkq.c . Still the compiler error
Oct 27 13:13:41 <michelem> all other modules compile just fine, and I verified that's the file being compiled because it fails differently if I include junk in it
Oct 27 13:14:12 <PaulFertser> michelem: http://git390.marist.edu/cgi-bin/gitweb.cgi?p=linux-2.6.git;a=commitdiff_plain;h=ca1cab37d91cbe8a8333732540d43cabb54cfa85
Oct 27 13:14:19 <PaulFertser> michelem: it was added much later
Oct 27 13:14:46 <michelem> uhm, I have it in my repos
Oct 27 13:15:09 <michelem> what file tells me the version of the kernel in the sources ?
Oct 27 13:15:10 <PaulFertser> Ah, i've misread the commit
Oct 27 13:15:29 <PaulFertser> I always use git, and "git describe" tells me about the current checked out version.
Oct 27 13:17:11 <PaulFertser> michelem: are you sure you're compiling that module against the kernel sources you're supposed to?
Oct 27 13:17:14 <michelem> ok, confirming it's 2.6.34
Oct 27 13:18:31 <michelem> PaulFertser: yes: ./include/linux/aufs_type.h:#define AUFS_VERSION	"2-standalone.tree-34-20100823"
Oct 27 13:19:17 <michelem> its gotta be some ifdef. If I "#ifndef INIT_WORK_ON_STACK .. #error" I get the error
Oct 27 13:19:34 <michelem> If I #ifndef right after #include <linux/workqueue.h>
Oct 27 13:22:37 <PaulFertser> michelem: what about INIT_WORK_ONSTACK ?
Oct 27 13:23:48 <michelem> why without _ ?
Oct 27 13:24:14 <PaulFertser> michelem: later kernels have it without. I suspect you might be trying an inappropriate kernel version or something like that.
Oct 27 13:25:31 <PaulFertser> That change went in 2.6.37 though.
Oct 27 13:28:30 <PaulFertser> michelem: are you sure the way you specify KDIR is correct?
Oct 27 13:30:02 <michelem> yep
Oct 27 13:30:35 <michelem> I'll include garbage in the .h
Oct 27 13:33:36 <michelem> i'm a roaming idiot
Oct 27 13:34:35 <michelem> no I aint
Oct 27 13:36:29 <michelem> ok.
Oct 27 13:36:44 <michelem> the workqueue.h file is taken from /lib/modules/ … not from the kernel source.
Oct 27 13:37:00 <PaulFertser> michelem: i guess you are not specifying the KDIR then.
Oct 27 13:56:55 <michelem> for some reason workqueue.h is taken from /lib/modules instead of the actual kernel source
Oct 27 14:01:08 <michelem> PaulFertser: you were right. the module's Makefile was overriding my KDIR variable.
Oct 27 14:01:18 <michelem> thanks a lot.
Oct 27 14:04:08 <michelem> hell. Apparently aufs is written for x86 only.
Oct 27 15:39:29 <PaulFertser> michelem: take a look at OpenWrt.
Oct 27 15:39:38 <PaulFertser> michelem: they're using unionfs all right.
Oct 27 15:40:22 <PaulFertser> michelem: http://wiki.openwrt.org/doc/techref/filesystems#explanation.2
Oct 27 15:41:14 <michelem> PaulFertser: thanks. I finally managed to get aufs to work
Oct 27 15:41:14 <PaulFertser> michelem: this clearly suggests aufs too works on arm: http://code.google.com/p/aufs-desire/wiki/KernelHowTo
Oct 27 15:41:58 <michelem> my current problem is: I mount $ROOT at kernel time, but then /lib/init/rw cancels and re-does everything from /etc/fstab
Oct 27 18:12:12 <PBeck> hi
**** ENDING LOGGING AT Fri Oct 28 02:59:57 2011