strtod: Convert 64 bit double to 64 bit int during computation

The gdtoa implementation uses the type long, defined as Long, in lots
of code.  For historical reason newlib defines Long as int32_t instead.

This works fine, as long as floating point exceptions are not enabled.
The conversion to 32 bit int can lead to a FE_INVALID situation.

Example:

  const char *str = "121645100408832000.0";
  char *ptr;

  feenableexcept (FE_INVALID);
  strtod (str, &ptr);

This leads to the following situation in strtod

  double aadj;
  Long L;

  [...]
  L = (Long)aadj;

For instance, on x86_64 the code here is

  cvttsd2si %xmm0,%eax

At this point, aadj is 2529648000.0 in our example.  The conversion to
32 bit %eax results in a negative int value, thus the conversion is
invalid.  With feenableexcept (FE_INVALID), a SIGFPE is raised.

Fix this by always using 64 bit ints here if double is not a 32 bit type
to avoid this type of FP exceptions.

Signed-off-by: Corinna Vinschen <corinna@vinschen.de>

1 files changed, 9 insertions, 0 deletions

diff --git a/newlib/libc/stdlib/strtod.c b/newlib/libc/stdlib/strtod.c
index 402510cdf..0cfa9e6ae 100644
--- a/newlib/libc/stdlib/strtod.c
+++ b/newlib/libc/stdlib/strtod.c
@@ -1186,7 +1186,16 @@ _strtod_l (struct _reent *ptr, const char *__restrict s00, char **__restrict se,
 #endif
 		if (y == z) {
 			/* Can we stop now? */
+#ifndef _DOUBLE_IS_32BITS
+			/* If FE_INVALID floating point exceptions are
+			   enabled, a conversion to a 32 bit value is
+			   dangerous.  A positive double value can result
+			   in a negative 32 bit int, thus raising SIGFPE.
+			   To avoid this, always convert into 64 bit here. */
+			__int64_t L = (__int64_t)aadj;
+#else
 			L = (Long)aadj;
+#endif
 			aadj -= L;
 			/* The tolerances below are conservative. */
 			if (dsign || dword1(rv) || dword0(rv) & Bndry_mask) {