This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug tree-optimization/69039] [6 regression] segfault with ftree-parallelize-loops=2
- From: "vries at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Sat, 26 Dec 2015 22:39:01 +0000
- Subject: [Bug tree-optimization/69039] [6 regression] segfault with ftree-parallelize-loops=2
- Auto-submitted: auto-generated
- References: <bug-69039-4 at http dot gcc dot gnu dot org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69039
--- Comment #9 from vries at gcc dot gnu.org ---
(In reply to vries from comment #7)
> (In reply to vries from comment #6)
> > Created attachment 37132 [details]
> > tentative patch
>
> The tentative patch fixes the reduced test-case, but we run into another ICE
> with the full test-case:
> ...
> ira-color.c: In function â<built-in>â:
> ira-color.c:110833:8: error: invalid PHI argument
> r < conflict_hregno + conflict_nregs;
> ^
>
> &conflicting_regs
> ira-color.c:110833:8: error: location references block not in block tree
> &conflicting_regs
> prephitmp_59 = PHI <_60(22), &conflicting_regs(11)>
> ira-color.c:110833:8: internal compiler error: verify_gimple failed
> 0x11b7aca verify_gimple_in_cfg(function*, bool)
> src/gcc/tree-cfg.c:5119
> 0xf9c3d2 execute_function_todo
> src/gcc/passes.c:1958
> 0xf9b4a5 do_per_function
> src/gcc/passes.c:1645
> 0xf9c5ae execute_todo
> src/gcc/passes.c:2010
> Please submit a full bug report,
> with preprocessed source if appropriate.
> Please include the complete backtrace with any bug report.
> See <http://gcc.gnu.org/bugs.html> for instructions.
> ...
Filed as: PR69062 - invalid PHI argument error with ftree-parallelize-loops=2