[LLVMbugs] [Bug 16729] New: wrong code (SIGFPE) on x86_64-linux-gnu (both 32-bit and 64-bit modes)

bugzilla-daemon at llvm.org bugzilla-daemon at llvm.org
Sun Jul 28 22:29:04 PDT 2013


http://llvm.org/bugs/show_bug.cgi?id=16729

            Bug ID: 16729
           Summary: wrong code (SIGFPE) on x86_64-linux-gnu (both 32-bit
                    and 64-bit modes)
           Product: clang
           Version: unspecified
          Hardware: PC
                OS: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
          Assignee: unassignedclangbugs at nondot.org
          Reporter: su at cs.ucdavis.edu
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified

The following code is miscompiled by the current clang trunk (as well as clang
3.3) on x86_64-linux-gnu at -O3 (also at -O2 for clang trunk) in both 32-bit
and 64-bit modes, resulting in a SIGFPE. This is a regression from clang 3.2. 


$ clang-trunk -v
clang version 3.4 (trunk 187322)
Target: x86_64-pc-linux-gnu
Thread model: posix
$ clang-trunk -O2 small.c
$ a.out
Floating point exception (core dumped)
$ clang-trunk -O3 small.c
$ a.out
Floating point exception (core dumped)
$ clang-3.3 -O3 small.c
$ a.out
Floating point exception (core dumped)
$ clang-3.2 -O3 small.c
$ a.out
0
$ 


----------------------------------------------------


int printf (const char *, ...);

unsigned int *a[1], *c, **const d = &a[0];

int b, e, f;

int main ()
{
  int g;
  unsigned int **h = &c;
  for (; f < 8; ++f)
    {
      e = h == d;
      g = e == 0 ? 0 : 1 / e;
      b |= g;
    }
  printf ("%d\n", 0);
  return 0;
}

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20130729/41923ba7/attachment.html>


More information about the llvm-bugs mailing list