From patchwork Thu Aug 4 20:51:53 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Joseph Myers X-Patchwork-Id: 14318 Received: (qmail 126986 invoked by alias); 4 Aug 2016 20:52:10 -0000 Mailing-List: contact libc-alpha-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: libc-alpha-owner@sourceware.org Delivered-To: mailing list libc-alpha@sourceware.org Received: (qmail 126976 invoked by uid 89); 4 Aug 2016 20:52:09 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-1.9 required=5.0 tests=AWL, BAYES_00, RCVD_IN_DNSWL_NONE, SPF_PASS autolearn=ham version=3.3.2 spammy=whatsoever, 65000, readily, 16, 7 X-HELO: relay1.mentorg.com Date: Thu, 4 Aug 2016 20:51:53 +0000 From: Joseph Myers To: Subject: Do not call __nan in scalb functions [committed] Message-ID: User-Agent: Alpine 2.20 (DEB 67 2015-01-07) MIME-Version: 1.0 When libm functions return a NaN: if it is for NaN input, it should be computed from that input (e.g. adding it to itself), so that payloads are propagated and signaling NaNs quieted, while if it is for non-NaN input, it should be produced by a computation such as (x - x) / (x - x), which raises "invalid" at the same time as producing an appropriate NaN, so avoiding any need for a call to feraiseexcept. Various libm functions, however, call __nan ("") (or __nanf or __nanl) to determine the NaN to return, together with using feraiseexcept (FE_INVALID) to raise the exception. sysdeps/generic/math_private.h has an optimization for those functions with constant "" argument so this doesn't actually involve a call to the __nan function, but it is still not the preferred approach for producing NaNs. (The optimized code also always uses the NAN macro, i.e. produces a default NaN for float converted to whatever the target type is, and on some architectures that may not be the same as the preferred default NaN for double or long double.) This patch fixes the scalb functions to use the conventional method of generating NaNs and raising "invalid" with an appropriate computation. (Most instances of this issue are in the complex functions, where it can more readily be fixed once they have been made type-generic and so only a third as many places need fixing. Some of the complex functions use __nan ("") + __nan (""), where the addition serves no purpose whatsoever.) Tested for x86_64 and x86. Committed. 2016-08-04 Joseph Myers * math/e_scalb.c: Do not include . (invalid_fn): Do calculation resulting in NaN instead of raising FE_INVALID and returning a NaN explicitly. * math/e_scalbf.c: Do not include . (invalid_fn): Do calculation resulting in NaN instead of raising FE_INVALID and returning a NaN explicitly. * math/e_scalbl.c: Do not include . (invalid_fn): Do calculation resulting in NaN instead of raising FE_INVALID and returning a NaN explicitly. diff --git a/math/e_scalb.c b/math/e_scalb.c index 7f61ce0..692f105 100644 --- a/math/e_scalb.c +++ b/math/e_scalb.c @@ -16,7 +16,6 @@ License along with the GNU C Library; if not, see . */ -#include #include #include @@ -26,10 +25,7 @@ __attribute__ ((noinline)) invalid_fn (double x, double fn) { if (__rint (fn) != fn) - { - __feraiseexcept (FE_INVALID); - return __nan (""); - } + return (fn - fn) / (fn - fn); else if (fn > 65000.0) return __scalbn (x, 65000); else diff --git a/math/e_scalbf.c b/math/e_scalbf.c index 7377f6e..8df86f4 100644 --- a/math/e_scalbf.c +++ b/math/e_scalbf.c @@ -16,7 +16,6 @@ License along with the GNU C Library; if not, see . */ -#include #include #include @@ -26,10 +25,7 @@ __attribute__ ((noinline)) invalid_fn (float x, float fn) { if (__rintf (fn) != fn) - { - feraiseexcept (FE_INVALID); - return __nan (""); - } + return (fn - fn) / (fn - fn); else if (fn > 65000.0f) return __scalbnf (x, 65000); else diff --git a/math/e_scalbl.c b/math/e_scalbl.c index 53e9ca2..02956ac 100644 --- a/math/e_scalbl.c +++ b/math/e_scalbl.c @@ -16,7 +16,6 @@ License along with the GNU C Library; if not, see . */ -#include #include #include @@ -26,10 +25,7 @@ __attribute__ ((noinline)) invalid_fn (long double x, long double fn) { if (__rintl (fn) != fn) - { - feraiseexcept (FE_INVALID); - return __nan (""); - } + return (fn - fn) / (fn - fn); else if (fn > 65000.0L) return __scalbnl (x, 65000); else