On 03/30/2013 07:10 AM, Steven Bosscher wrote:
Right, the PA "explodes" table jumps to individual jumps in
pa_reorg().
Yup. One of the many things I'd like to forget. Unlike other jumps on
the PA, we can't nullify the delay slot to save space. The jump must
always be 2 instructions. So we have to have a separate pattern, even
though it looks like the normal jump. Ugh.
I'm more
concerned about some of the back ends that use
{next,prev}_{real,active}_insn to find jump table data after a
JUMP_INSN instead of using tablejump_p, but we'll see and I'll
obviously be on the hook to fix any fall-out.
I won't be surprised if we find this stuff, but I don't think there'll
be anything inherently difficult about fixing it.
jeff
