[Bug target/32191] ICE with complex __float128

[ubizjak at gmail dot com]

------- Comment #3 from ubizjak at gmail dot com  2007-06-03 06:42 -------
(In reply to comment #2)

> #0  build_call_expr (fndecl=0x0, n=4) at ../../trunk/gcc/builtins.c:10379

fndecl is already NULL there.

> #1  0x000000000068d0bd in expand_complex_libcall (bsi=0x7fbfffed20, 

There is a pack of suspicious code there dealing with MULT_EXPR and RDIV_EXPR
specifically in expand_complex_libcall. It is called only in c99 mode.

> #2  0x000000000068ff78 in tree_lower_complex_O0 ()

Executed only at -O0.

Indeed, original testcase fails only for -O0, -std=c99. All other optimizations
level work OK, and it works when multiplication is changed to addition od
substitution.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32191