On Thu, 27 Dec 2012, Martin Pieuchot wrote:
I'm not sure this is related to (all) G3 processors. I can't reproduce this behavior on my PowerBook G3 with:cpu0 at mainbus0: 750 (Revision 0x8300): 400 MHz: 1MB backside cache See below for a similar register output taken right after executing TCB_THREAD() in pthread_self(). Without much investigation it looks to me that the value in your %r2 register got truncated...
That's weird... Any B&W G3 Rev1 user out there? I know the hard-drive in my B&W is quite old, could it cause this kind of problem? I reinstalled the newer snapshots from scratch every time, but maybe libpthread.so is stored on some bad sectors, or something similar?
