Correct min_of_type handling of _Float128 [committed]
Commit Message
The math_private.h macro min_of_type has broken _Float128 handling:
instead of passing its type argument to the key __EXPR_FLT128 macro,
it passes x, which is not a macro argument but whatever variable
called x happens to be visible in the calling function. If that
variable has the wrong type, the wrong one of long double and
_Float128 can get chosen. In particular, this applies to some
_Complex long double functions (where x happens to have type _Complex
long double, resulting in min_of_type returning a _Float128 value when
it should return a long double value). For some reason, this only
caused test failures for me on x86_64 with GCC 6 but not GCC 7 (I
suspect it triggers known bugs in conversions from x86 long double to
_Float128 that are present in GCC 6's soft-fp).
Tested for x86_64 (in conjunction with float128 patches). Committed.
2017-06-22 Joseph Myers <joseph@codesourcery.com>
* sysdeps/generic/math_private.h (__EXPR_FLT128): Do not apply
typeof to argument passed to __builtin_types_compatible_p.
(min_of_type): Pass type argument, not x, to __EXPR_FLT128.
@@ -304,7 +304,7 @@ extern void __docos (double __x, double __dx, double __v[]);
#if __HAVE_DISTINCT_FLOAT128
# define __EXPR_FLT128(x, yes, no) \
__builtin_choose_expr (__builtin_types_compatible_p \
- (__typeof (x), long double), no, yes)
+ (x, long double), no, yes)
#else
# define __EXPR_FLT128(x, yes, no) no
#endif
@@ -318,7 +318,7 @@ extern void __docos (double __x, double __dx, double __v[]);
__builtin_choose_expr \
(__builtin_types_compatible_p (type, double), \
DBL_MIN, \
- __EXPR_FLT128 (x, FLT128_MIN, LDBL_MIN)))
+ __EXPR_FLT128 (type, FLT128_MIN, LDBL_MIN)))
/* If X (which is not a NaN) is subnormal, force an underflow
exception. */