Re: %gs:0 thread pseudoregister in oskit-mach

[Roland McGrath]

> 0xc0000004 gives the same problems. Note that the assertion was right
> in the first place, the GS segment didn't trigger it, only the DS
> segment.

Humph!  You're right, the %gs:0 word should never have a fault anyway
because those pages and page tables are always in core.  I backed out my
change to the assert.

So you're saying that accessing 0xc0000004 also gets a page fault and 
hits the assert rather than get a GP fault?