"Perry E. Metzger" <perry@piermont.com> writes:

> Klaus Klein <kleink@netbsd.org> writes:
> > 	syssrc/sys/kern: syscalls.master
> > 	syssrc/sys/uvm: uvm_mmap.c
> > 
> > Log Message:
> > XSH5: change function signature to `void *sbrk(intptr_t)'.
> 
> I think you may have fallen in to a bit of a trap, Klaus. Not a big
> one -- this was harmless -- but a bit of one anyway.
> 
> You see, syscalls.master is NOT where sbrk's signature is
> defined. Have a close look. The sbrk() in syscalls.master is NEVER
> used. I described this in one of my pr's on the subject.

I was aware of that prior to going over your PR, actually (though I
could really have left the unimplemented syscall alone, since it's not
even used in the automatic lint stub generation process).  Any
comments/objections about tagging sys_sbrk() OBSOL and removing the
in-kernel stub?  (If emulations should happen to depend on *not*
posting a SIGSYS if called (as it is now), this should be addressed
where required only.)


-kjk