pthread* necessary support, mmap* required changes.Reentrant MiNTlib's functions, virtual memory, etc...Or a different kernel... I voiced this idea before...What about porting GEM and related layer on top of the Linux-m68k kernel? That could be fun because one would probably learn something and it would not be perceived as useless in the CV. Digging deep in the Linux kernelcould be a dream job offer resulting experience.
Well, of course, depending on the job area, implementing VM and other OS basics could result in such offer too ;)
Standa