Use VFPv3-D16 FPU for ARM builds

Upstream-Bug: https://ghc.haskell.org/trac/ghc/ticket/5914

Jani writes: The D16 part was Debian/Ubuntu specific, IIRC we define hardfloat
in that particular variant (16 double registers) or we had a different naming
for some reason.

Gbp-Pq: Name ARM-VFPv3D16

diff --git a/m4/get_arm_isa.m4 b/m4/get_arm_isa.m4
index 9b458bb09f5dcd451c4d52dc9e217255b2f2e93f..22bba18b161c1cad0d316873e501b4c640162dbf 100644 (file)
--- a/m4/get_arm_isa.m4
+++ b/m4/get_arm_isa.m4
@@ -81,7 +81,7 @@ AC_DEFUN([GET_ARM_ISA],
                 fi],
                 [changequote(, )dnl
                  ARM_ISA=ARMv7
-                 ARM_ISA_EXT="[VFPv3,NEON]"
+                 ARM_ISA_EXT="[VFPv3D16,NEON]"
                  changequote([, ])dnl
                 ])
         ])