[Bug target/60604] New: GCC incorrectly compiles s_csinh function on MIPS

["sje at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60604

            Bug ID: 60604
           Summary: GCC incorrectly compiles s_csinh function on MIPS
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: sje at gcc dot gnu.org
            Target: mips*-*-*

Created attachment 32410
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=32410&action=edit
Test case

While looking at a GCC fortran test failure
(gfortran.dg/complex_intrinsic_3.f90) I tracked the problem down to GCC
miscompiling the routine.  I have attached a cutdown test case to this report.

When compiled for mips32r2, big-endian, hard-float, and using -O0 or -O1
the attached program prints:

x = 0.634964 1.298458
rcls = 4, icls = 4, t = 709
In __csinh(1), 0.634964 1.29846
In __csinh(2), 0.634964 1.29846
x = 0.000000 0.000000

When compiled with -O2 or -O3 it prints:

x = 0.634964 1.298458
rcls = 4, icls = 4, t = 709
In __csinh(1), 0.634964 1.29846
In __csinh(2), 1.00959e+116 1.29846
In if statement
x = 0.000000 0.000000

Somehow the optimized code is corrupting the input value of csinh.