This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: dead_or_predictable
- From: Richard Henderson <rth at redhat dot com>
- To: Jan Hubicka <jh at suse dot cz>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Wed, 9 Jan 2002 14:11:29 -0800
- Subject: Re: dead_or_predictable
- References: <20020109111346.GA17206@atrey.karlin.mff.cuni.cz>
On Wed, Jan 09, 2002 at 12:13:46PM +0100, Jan Hubicka wrote:
> if_case 2 converison may result in conditional jumping to next instruction.
> In such case we die due to duplicate edge. Sorry for lact of testcase
> but I was unable to simplify it to something trivial :(
I'd be interested in seeing something even non-trivial here.
> ! /* We may result in conditional jumping to the next instruction.
> ! Delete the duplicate edge. */
> !
> ! if (FALLTHRU_EDGE (test_bb)->dest == new_dest)
> {
> ! FALLTHRU_EDGE (test_bb)->count += BRANCH_EDGE (test_bb)->count;
> ! FALLTHRU_EDGE (test_bb)->probability = REG_BR_PROB_BASE;
> ! }
Where does the duplicate actually get removed?
Doesn't the jump-to-next imply that we had an empty block, and
if so, why did we create such a thing?
r~