HAIFA scheduler bug?

[Jeffrey A Law]

  In message < 199803252338.PAA14521@rtl.cygnus.com >you write:
  > The example doesn't fail with the current sources because we get slightly
  > different RTL (due I think to reload changes), and the code in jump.c in
  > delete_computation has a == rtl test which fails in this case.  The bug is
  > still there though.
Actually we made some changes to reduce the change that jump will bogusly
use a REG_DEAD note after reload.  I suspect we didn't catch all of
them, but we may have caught the one in question.

  > Perhaps we should just add a new pass before the second jump pass
  > to recompute the REG_DEAD notes?
It's a possibility.  I had changes once to rerun flow analysis after
reload.  It exposed some problems, but non that I would consider
unsolveable.

If we had accurate REG_DEAD notes, jump and reorg become simpler
since they could depend on them instead of having to compute
such information on their own.

Jeff