Add issubnormal.

TS 18661-1 adds an issubnormal classification macro to <math.h>.  This
patch implements it for glibc.  There are no new underlying functions
in libm because the implementation uses fpclassify; any optimizations
for this macro should be done through adding __builtin_subnormal in
GCC and using it in the header for suitable GCC versions, not through
adding other optimized inline or out-of-line versions to glibc.

The intended structure of the NEWS entry for <math.h> features from TS
18661-1 is like:

* New <math.h> features are added from TS 18661-1:2014:

  - Nearest integer functions: roundeven, roundevenf, roundevenl.

  - Comparison macros: iseqsig.

  - Classification macros: iscanonical, issubnormal, iszero.

(that is, following the grouping of interfaces in TS 18661-1:2014,
with any group where any interfaces are new in glibc 2.25 being listed
like that).

Tested for x86_64 and x86.

	* math/math.h [__GLIBC_USE (IEC_60559_BFP_EXT)] (issubnormal): New
	macro.
	* math/libm-test.inc (issubnormal_test_data): New array.
	(issubnormal_test): New function.
	* manual/arith.texi (Floating Point Classes): Document
	issubnormal.
	* manual/libm-err-tab.pl: Update comment on interfaces without
	ulps tabulated.

1 files changed, 4 insertions, 0 deletions

diff --git a/NEWS b/NEWS
index f59a27d..23ca5ab 100644
--- a/NEWS
+++ b/NEWS
@@ -49,6 +49,10 @@ Version 2.25
   UINTMAX_WIDTH, PTRDIFF_WIDTH, SIG_ATOMIC_WIDTH, SIZE_WIDTH, WCHAR_WIDTH,
   WINT_WIDTH.
 
+* New <math.h> features are added from TS 18661-1:2014:
+
+  - Classification macros: issubnormal.
+
 * The <sys/quota.h> header now includes the <linux/quota.h> header.  Support
   for the Linux quota interface which predates kernel version 2.4.22 has
   been removed.