>> So, in addition to eliminating deprecated write_sp, would it be ok to 
>> move the write SP code to the end of the push_dummy_call methods?
> 
> 
> I can't think of any problems that would arise from moving the "write
> SP" code to the end of the various push_dummy_call() methods.  But
> just in case, when you change it, please note how it used to be done
> and why doing it in a different location *shouldn't* be a problem. 
> (It may someday make it easier to debug that obscure target which
> randomly picks a stopped thread's stack to service an interrupt...)

This is what I've checked in.

Andrew