26 lines
1.3 KiB
C
26 lines
1.3 KiB
C
$NetBSD: patch-qemu-xen_fpu_softfloat-specialize.h,v 1.1 2013/05/26 20:27:43 bouyer Exp $
|
|
|
|
--- qemu-xen/fpu/softfloat-specialize.h.orig 2013-05-26 21:19:30.000000000 +0200
|
|
+++ qemu-xen/fpu/softfloat-specialize.h 2013-05-26 21:20:25.000000000 +0200
|
|
@@ -89,8 +89,8 @@
|
|
#define floatx80_default_nan_low LIT64( 0xC000000000000000 )
|
|
#endif
|
|
|
|
-const floatx80 floatx80_default_nan = make_floatx80(floatx80_default_nan_high,
|
|
- floatx80_default_nan_low);
|
|
+const floatx80 floatx80_default_nan = { .high = floatx80_default_nan_high,
|
|
+ .low = floatx80_default_nan_low};
|
|
|
|
/*----------------------------------------------------------------------------
|
|
| The pattern for a default generated quadruple-precision NaN. The `high' and
|
|
@@ -104,8 +104,8 @@
|
|
#define float128_default_nan_low LIT64( 0x0000000000000000 )
|
|
#endif
|
|
|
|
-const float128 float128_default_nan = make_float128(float128_default_nan_high,
|
|
- float128_default_nan_low);
|
|
+const float128 float128_default_nan = {.high = float128_default_nan_high,
|
|
+ .low = float128_default_nan_low};
|
|
|
|
/*----------------------------------------------------------------------------
|
|
| Raises the exceptions specified by `flags'. Floating-point traps can be
|