bug in propagate_one_insn

[Daniel Berlin]

>   In message < 87n15p39o6.fsf@cgsoftware.com >you write:
>   > As i expected, if a libcall's arguments end up getting set in a block
>   > before the call itself, we end up crashing because propagate_one_insn
>   > returns an insn before the start of the basic block we were
>   > propagating liveness for (and it only terminates if insn == bb->head):
> Err, how did the arguments end up getting set in a different block from
> the libcall itself?
>
> jeff
>
>


The scheduler has no idea that libcall arguments shouldn't be moved, so 
interblock/speculative motion will happily move them into other blocks.
-fno-sched-spec makes the problem go away.
--Dan