This is the mail archive of the newlib@sourceware.org mailing list for the newlib project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

[PATCH, ARM] Check for 64 FP instructions availability in libm before using them

From: "Thomas Preud'homme" <thomas dot preudhomme at arm dot com>
To: <newlib at sourceware dot org>
Date: Wed, 9 Dec 2015 11:00:55 +0800
Subject: [PATCH, ARM] Check for 64 FP instructions availability in libm before using them
Authentication-results: sourceware.org; auth=none

Currently, double precision math functions in newlib/libm/machine/arm detect whether neon instructions can be used for double precision computation by checking the architecture version (>= 8) and the availability of floating-point instructions. However, these instructions would not be available if targeting fpv5-sp-d16 as the FPU, which by definition does not have 64bit float instructions. This patch adds a check that __ARM_FP advertises 64bit float instructions.

ChangeLog entry is as follows:


*** newlib/ChangeLog ***

2015-11-03  Thomas Preud'homme  <thomas.preudhomme@arm.com>

        * libm/machine/arm/s_ceil.c: Also check that 64bit FP instructions are
        available in the guard.
        * libm/machine/arm/s_floor.c: Likewise.
        * libm/machine/arm/s_nearbyint.c: Likewise.
        * libm/machine/arm/s_rint.c: Likewise.
        * libm/machine/arm/s_round.c: Likewise.
        * libm/machine/arm/s_trunc.c: Likewise.


diff --git a/newlib/libm/machine/arm/s_ceil.c b/newlib/libm/machine/arm/s_ceil.c
index ceb5bbf..ed1e85a 100644
--- a/newlib/libm/machine/arm/s_ceil.c
+++ b/newlib/libm/machine/arm/s_ceil.c
@@ -24,7 +24,7 @@
    NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
    SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */
 
-#if __ARM_ARCH >= 8 && !defined (__SOFTFP__)
+#if __ARM_ARCH >= 8 && (__ARM_FP & 0x8) && !defined (__SOFTFP__)
 #include <math.h>
 
 double
diff --git a/newlib/libm/machine/arm/s_floor.c b/newlib/libm/machine/arm/s_floor.c
index 5001981..d25132f 100644
--- a/newlib/libm/machine/arm/s_floor.c
+++ b/newlib/libm/machine/arm/s_floor.c
@@ -24,7 +24,7 @@
    NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
    SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */
 
-#if __ARM_ARCH >= 8 && !defined (__SOFTFP__)
+#if __ARM_ARCH >= 8 && (__ARM_FP & 0x8) && !defined (__SOFTFP__)
 #include <math.h>
 
 double
diff --git a/newlib/libm/machine/arm/s_nearbyint.c b/newlib/libm/machine/arm/s_nearbyint.c
index c3b50dc..7ead69b 100644
--- a/newlib/libm/machine/arm/s_nearbyint.c
+++ b/newlib/libm/machine/arm/s_nearbyint.c
@@ -24,7 +24,7 @@
    NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
    SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */
 
-#if __ARM_ARCH >= 8 && !defined (__SOFTFP__)
+#if __ARM_ARCH >= 8 && (__ARM_FP & 0x8) && !defined (__SOFTFP__)
 #include <math.h>
 
 double
diff --git a/newlib/libm/machine/arm/s_rint.c b/newlib/libm/machine/arm/s_rint.c
index 9050424..02c1022 100644
--- a/newlib/libm/machine/arm/s_rint.c
+++ b/newlib/libm/machine/arm/s_rint.c
@@ -24,7 +24,7 @@
    NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
    SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */
 
-#if __ARM_ARCH >= 8 && !defined (__SOFTFP__)
+#if __ARM_ARCH >= 8 && (__ARM_FP & 0x8) && !defined (__SOFTFP__)
 #include <math.h>
 
 double
diff --git a/newlib/libm/machine/arm/s_round.c b/newlib/libm/machine/arm/s_round.c
index 97f0296..a8f2466 100644
--- a/newlib/libm/machine/arm/s_round.c
+++ b/newlib/libm/machine/arm/s_round.c
@@ -24,7 +24,7 @@
    NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
    SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */
 
-#if __ARM_ARCH >= 8 && !defined (__SOFTFP__)
+#if __ARM_ARCH >= 8 && (__ARM_FP & 0x8) && !defined (__SOFTFP__)
 #include <math.h>
 
 double
diff --git a/newlib/libm/machine/arm/s_trunc.c b/newlib/libm/machine/arm/s_trunc.c
index 2f38a35..6fb6968 100644
--- a/newlib/libm/machine/arm/s_trunc.c
+++ b/newlib/libm/machine/arm/s_trunc.c
@@ -24,7 +24,7 @@
    NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
    SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */
 
-#if __ARM_ARCH >= 8 && !defined (__SOFTFP__)
+#if __ARM_ARCH >= 8 && (__ARM_FP & 0x8) && !defined (__SOFTFP__)
 #include <math.h>
 
 double


With this patch, newlib can be built with -mfpu=fpv5-sp-d16 while it cannot without.

Is this ok for master branch?

Best regards,

Thomas

Follow-Ups:
- Re: [PATCH, ARM] Check for 64 FP instructions availability in libm before using them
  - From: Corinna Vinschen

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]