Programs built using buildroot g++ complier for MIPS32 platform results in segmentation fault

[Rob Landley]

On Wednesday 14 December 2005 23:04, Garrett Kajmowicz wrote:

> Unfortunately, I do not personally have the facilities with which to test
> uClibc++ on MIPS - I only have a x86 machine.  Somebody once tested it and
> told me that it worked, so I've listed it in the arch list for testing
> purposes.

All I have is an x86 machine too, but I've been playing with qemu.  It's a 
marvelous thing, and can currently emulate not just i386 but mips, ppc, 
sparc, and x86_64 systems.  (It can also emulate arm and armeb processors, 
but doesn't have system definitions for those so it doesn't know what kind of 
fake hardware to and busses to attach to the fake processor, so you can't 
actually boot a linux kernel on that yet...)

The coolest thing I've seen in a while is booting a knoppix ISO image under 
qemu (it boots in a window) and playing frozen bubble very, very slowly...

Getting non-x86 root filesystem images to test this against is a bit of a 
pain, though.  I'm told buildroot can do it, but haven't found preprepared 
partitioned hard drive images anywhere.  (I've made them for x86, and can 
give the script.  But haven't gotten around to making a working cross-compile 
environment yet, and I don't use buildroot...)

Rob
-- 
Steve Ballmer: Innovation!  Inigo Montoya: You keep using that word.
I do not think it means what you think it means.