[Bug math/17899] New: [powerpc] floorl returns negative zero with FE_DOWNWARD

["msebor at redhat dot com" <sourceware-bugzilla at sourceware dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=17899

            Bug ID: 17899
           Summary: [powerpc] floorl returns negative zero with
                    FE_DOWNWARD
           Product: glibc
           Version: 2.21
            Status: NEW
          Severity: normal
          Priority: P2
         Component: math
          Assignee: unassigned at sourceware dot org
          Reporter: msebor at redhat dot com

Created attachment 8089
  --> https://sourceware.org/bugzilla/attachment.cgi?id=8089&action=edit
floorl test case.

There are many failures in test_{ildoubl,ldouble} on ppc64 (1648 and 3074,
respectively). At least four of them are due to the floorl function returning a
negative zero for a positive argument that's less than 1. See the snippet from
test-ldouble.out below. A simple test case independent of the test suite that
demonstrates the problem along with its output on ppc64 is attached.

Failure: Test: floor_downward (0.1)
Result:
 is:         -0.00000000000000000000e+00  -0x0.00000000000000000000p+0
 should be:   0.00000000000000000000e+00   0x0.00000000000000000000p+0
 difference:  0.00000000000000000000e+00   0x0.00000000000000000000p+0
 ulp       :  0.0000
 max.ulp   :  0.0000

Failure: Test: floor_downward (0.1)
Result:
 is:         -0.00000000000000000000e+00  -0x0.00000000000000000000p+0
 should be:   0.00000000000000000000e+00   0x0.00000000000000000000p+0
 difference:  0.00000000000000000000e+00   0x0.00000000000000000000p+0
 ulp       :  0.0000
 max.ulp   :  0.0000
Failure: Test: floor_downward (0.25)
Result:
 is:         -0.00000000000000000000e+00  -0x0.00000000000000000000p+0
 should be:   0.00000000000000000000e+00   0x0.00000000000000000000p+0
 difference:  0.00000000000000000000e+00   0x0.00000000000000000000p+0
 ulp       :  0.0000
 max.ulp   :  0.0000
Failure: Test: floor_downward (0.625)
Result:
 is:         -0.00000000000000000000e+00  -0x0.00000000000000000000p+0
 should be:   0.00000000000000000000e+00   0x0.00000000000000000000p+0
 difference:  0.00000000000000000000e+00   0x0.00000000000000000000p+0
 ulp       :  0.0000
 max.ulp   :  0.0000

-- 
You are receiving this mail because:
You are on the CC list for the bug.