This is the mail archive of the
mailing list for the GCC project.
Re: Use-and-clobber insns and REG_DEAD notes (fwd)
- From: Richard Henderson <rth at redhat dot com>
- To: Ulrich Weigand <weigand at i1 dot informatik dot uni-erlangen dot de>
- Cc: gcc at gcc dot gnu dot org
- Date: Tue, 15 Jul 2003 15:47:12 -0700
- Subject: Re: Use-and-clobber insns and REG_DEAD notes (fwd)
- References: <200307142313.BAA08681@faui11.informatik.uni-erlangen.de>
On Tue, Jul 15, 2003 at 01:13:05AM +0200, Ulrich Weigand wrote:
> As part of this patch you have removed a comment:
> ! [snip] Which begs the whole
> ! question of whether we should in fact have death notes for registers
> ! used and clobbered (but not set) in the same insn. [snip]
> This is exactly the question I have here ;-) Does the fact that
> you removed the comment imply that you had answered the question?
I think that the death note SHOULD exist in the clobber case.
> Is there a way for combine to call back into flow to ask whether
> a dead note is required?
Well, there's a way to get combine to call back into flow when it
gets too confused, yes (see refresh_blocks) but that's usually used
for when we've changed global life. I'd expect us to be able to
get this right inside combine.