acos.s fix: use the formula acos(x) = atan2(sqrt(1-x),sqrt(1+x))

the old formula atan2(1,sqrt((1+x)/(1-x))) was faster but
could give nan result at x=1 when the rounding mode is
FE_DOWNWARD (so 1-1 == -0 and 2/-0 == -inf), the new formula
gives -0 at x=+-1 with downward rounding.

src/math/i386/acos.s

@@ -18,12 +18,10 @@ acos:
 	fld %st(1)
 	fld1
 	fsubp
+	fsqrt
 	fxch %st(2)
 	faddp
-	fdivp
 	fsqrt
-	fld1
-	fxch %st(1)
 	fpatan
 	fld1
 	fld1

src/math/x86_64/acosl.s

@@ -6,12 +6,10 @@ acosl:
 	fld %st(1)
 	fld1
 	fsubp
+	fsqrt
 	fxch %st(2)
 	faddp
-	fdivp
 	fsqrt
-	fld1
-	fxch %st(1)
 	fpatan
 	fld1
 	fld1