https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82016
Segher Boessenkool <segher at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Last reconfirmed| |2017-08-28
Ever confirmed|0 |1
--- Comment #5 from Segher Boessenkool <segher at gcc dot gnu.org> ---
The ccp1 pass makes it return 0 instead of a+b:
===
Simulating block 2
Visiting statement:
_3 = a_1(D) + b_2(D);
which is likely CONSTANT
Match-and-simplified a_1(D) + b_2(D) to 0.0
Lattice value changed to CONSTANT 0.0. Adding SSA edges to worklist.
marking stmt to be not simulated again
===
Confirmed. (float128-type-2.c fails in the same way).
