This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug tree-optimization/61385] [4.10 Regression] ICE on valid code at -O2 and -O3 on x86_64-linux-gnu
- From: "glisse at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Mon, 02 Jun 2014 09:54:27 +0000
- Subject: [Bug tree-optimization/61385] [4.10 Regression] ICE on valid code at -O2 and -O3 on x86_64-linux-gnu
- Auto-submitted: auto-generated
- References: <bug-61385-4 at http dot gcc dot gnu dot org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61385
--- Comment #3 from Marc Glisse <glisse at gcc dot gnu.org> ---
Ah, right, I use last_and_only_stmt (middle_bb) but that doesn't prove there
are no PHI nodes there. Easiest fix, something like:
if (!gimple_seq_empty_p (phi_nodes (middle_bb)))
return 0;
A better fix would be to move the PHI nodes from middle_bb to the same block as
we are moving the assignment. I am just not completely sure how safe that is
(or how to do that, but I'll probably find examples nearby).