This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug middle-end/68870] [6 Regression] ICE on valid code at -O1, -O2 and -O3 on x86_64-linux-gnu
- From: "law at redhat dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Mon, 14 Dec 2015 16:34:18 +0000
- Subject: [Bug middle-end/68870] [6 Regression] ICE on valid code at -O1, -O2 and -O3 on x86_64-linux-gnu
- Auto-submitted: auto-generated
- References: <bug-68870-4 at http dot gcc dot gnu dot org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68870
Jeffrey A. Law <law at redhat dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |law at redhat dot com
--- Comment #4 from Jeffrey A. Law <law at redhat dot com> ---
A post-pass over the CFG doing trivial simplifications on the conditions
wouldn't be terribly hard. It would have to limit itself to not looking at
anything in the SSA graph since it's inconsistent until after update_ssa.
My concern is that doing so just papers over the deeper problems we're having
with the interactions with SSA_NAME recycling and the folders.
Would it make sense to define objects that are on the pending free list as
still being referencable by the IL? So they still have a valid type, def_stmt,
etc. It feels like that's the direction we're already going anyway.