Rafal Boni rafal@pobox.com  said;

> Not that it really affects anything, but __setjmp14.S also has some bugs
> around reorder/noreorder (ie, the explicit zero of v0 in the delay slot
> of the __sigprocmask14 call doesn't really happen there) and there are
> a few bits I might clean up if I had reason to touch it (put the REG_
> EPILOGUE in a saner place, add a nop after the call to abort).

Yes, it's my theory this moment.  It's uncerntain for me compiler works
correct.

- trap/syscall insn has delay slot,
- jump insn has delay slot,
- jump insn in delay slot is illegal or undefined,

Sprinkling fancy MACRO makes code smart.  But the practice tends
easily to confuse resulting code.  It's time to re-write __setjmp14.S
anyway even if it was not the real cause of our trouble.

Toru Nishimura/ALKYL Technology