[02:39] <nat2610> hey, I'm curious, how do you compile everything for arm. IO
[02:40] <nat2610> I'm looking at doing some cross compilation and generating an application that I'm working on, on my i386 linux and I'd like to generate an arm binrary
[02:41] <nat2610> I'm stock on the system library, for instance, I link libobjc.so.2 to my app but ld if I use arm-none-linux-gnueabi-gcc, ld complain that it can't find that lib since the one my system have is compiled for i386
[02:42] <nat2610> but I know I have it on my arm system with ubuntu
[02:43] <nat2610> hmmm let me try to say that again
[06:59] <playya> nat2610, ubuntu prefers native builds
[06:59] <playya> you need a arm5 device to compile it there
[18:48] <nat2610> playya, so I can't do any cross compiling at all ? I saw some stuff like http://www.freedesktop.org/wiki/Software/sbox2 but I don't really know who uses that (I was hopping ubuntu would be ;) )
[18:49] <nat2610> playya,  I forgot to add, I have a arm5 device but it's really slow so I'm trying to build everything on my pc i386 and just copy over the binraries generated
[19:18] <armin76> nat2610: ubuntu is about using pre-built binaries :)
[19:22] <[g2]> nat2610: depending on what you are trying to cross-compile, you may want setup a full environment, or just run the arm-rootfs in qemu and cross-compile from within a native environment
[21:35] <kblin> hi folks
[21:36] <kblin> I'm running jaunty on a beagle board using one of ogra's kernels, and it seems like there's something weird going on with /proc
[21:37] <kblin> specifically, /proc/<pid>/maps and /proc/<pid>/smaps are empty
[21:37] <kblin> any idea what could be causing this?
[22:14] <Martyn> kblin : An incomplete implementation
[22:15] <Martyn> kblin : find one of the OMAP3 git trees, and you'll have to build a custom kernel
[22:15] <kblin> ok, I'll have to look into that then
[22:15] <Martyn> There is a good 2.6.29 (and now .30) kernel that compiles well.  Don't use the tree that has the alternate video driver (ubuntu doesn't know how to drive it, and you'll end up with just a text console)
[22:16] <kblin> hey, I can fix the ip6tables modules while at that :)
[22:16] <kblin> Martyn: it's a headless ubuntu-server box, I don't mind :)
[22:16] <Martyn> oh!
[22:16] <Martyn> Well, then...
[22:16] <kblin> I'm trying to turn it into a file server, active directory domain controller
[22:17] <Martyn> you're going to find that IO performance on the beagle is ATROCIOUS
[22:17] <Martyn> Seriously poo-poo
[22:17] <kblin> yeah, I know
[22:17] <Martyn> I am using USB->SATA on a Beagle C3, and frankly I just want to throw it at a wall most days
[22:17] <Martyn> Then again, I'm getting /really/ used to working with the power of the new Cortex-A8 processors, and even Cortex A9 in the software FastModel simulators
[22:18] <kblin> especially as it seems like networking flakes out on my revB if I use a 2.0 USB hdd
[22:18] <Martyn> There's nothing like having 2, even 4 cores to really make things go zoom
[22:18] <Martyn> Yes
[22:18] <Martyn> there's a power issue on the RevB
[22:18] <Martyn> don't even bother with the B
[22:18] <kblin> well, that's what I got here..
[22:19] <kblin> btw, power issue? would that matter on a powered hub?
[22:20] <kblin> anyway, this is more or less a proof of concept to show off at the storage developer's conference
[22:20] <Martyn> kblin : no, it's a matter of the on-board 100mw
[22:20] <Martyn> kblin : For some reason, the USB OTG hub tends to 'brown out' for no good reason
[22:21] <kblin> ok, weird :)
[22:21] <Martyn> which causes the hub to go into a bad state
[22:21] <Martyn> the host controller, rather
[22:21] <Martyn> yeah, feh :)
[22:21] <Martyn> C2/C3 doesn't suffer from the issue
[22:21] <Martyn> and has a PROPER port on it
[22:21] <kblin> that matches what I see in dmesg
[22:21] <kblin> well, crud :)
[22:22] <kblin> I guess if I really want to do this with a decent performance, I should be using a system with on-board ethernet anyway :)
[22:22] <Martyn> this is similar to the problem my company is trying to solve.  Yep.
[22:23] <Martyn> kblin : There will be  (soon)  some boards based on the OMAP3 and i.mx51 that are similar to the beagle
[22:23] <Martyn> with more onboard devices
[22:23] <Martyn> there are also beagle-clones that /do/ have ethernet
[22:23]  * kblin nods
[22:23] <Martyn> one even has a USB->SATA bridge
[22:24] <kblin> as I said, this one is a proof of concept
[22:24] <kblin> I'm sure if someone wants to turn this into a product, they'll bring the embedded know-how to build a system with the correct specs
[22:25] <kblin> I just want to go to a storage conference showing off that opensource can run an active directory domain controller with 2 Watts
[22:26] <broonie> You want to redo the old ARM demo and arrange to power it with the waste heat from a PC. :)
[22:26] <kblin> hehe
[22:30] <kblin> I'm not sure if a stirling engine-type generator won't look too much like a bomb on X-ray :)
[22:32] <suihkulokki> kblin: if you want decent headless network server, you should look at something like the sheevaplug
[22:39] <kblin> suihkulokki: yeah, planning to
[22:41] <kblin> suihkulokki: I'm currently looking for a way to get shipping to be less than 70% of the unit costs