fix double-processing of DT_RELR relocations in ldso relocating itself
[musl] / src / math / expm1.c
index f8f32c4..ac1e61e 100644 (file)
@@ -31,7 +31,7 @@
  *          R1(r**2) = 6/r *((exp(r)+1)/(exp(r)-1) - 2/r)
  *                   = 6/r * ( 1 + 2.0*(1/(exp(r)-1) - 1/r))
  *                   = 1 - r^2/60 + r^4/2520 - r^6/100800 + ...
- *      We use a special Reme algorithm on [0,0.347] to generate
+ *      We use a special Remez algorithm on [0,0.347] to generate
  *      a polynomial of degree 5 in r*r to approximate R1. The
  *      maximum error of this polynomial approximation is bounded
  *      by 2**-61. In other words,
 #include "libm.h"
 
 static const double
-huge        = 1.0e+300,
-tiny        = 1.0e-300,
 o_threshold = 7.09782712893383973096e+02, /* 0x40862E42, 0xFEFA39EF */
 ln2_hi      = 6.93147180369123816490e-01, /* 0x3fe62e42, 0xfee00000 */
 ln2_lo      = 1.90821492927058770002e-10, /* 0x3dea39ef, 0x35793c76 */
@@ -122,39 +120,27 @@ Q5 = -2.01099218183624371326e-07; /* BE8AFDB7 6E09C32D */
 
 double expm1(double x)
 {
-       double y,hi,lo,c,t,e,hxs,hfx,r1,twopk;
-       int32_t k,xsb;
-       uint32_t hx;
-
-       GET_HIGH_WORD(hx, x);
-       xsb = hx&0x80000000;  /* sign bit of x */
-       hx &= 0x7fffffff;     /* high word of |x| */
+       double_t y,hi,lo,c,t,e,hxs,hfx,r1,twopk;
+       union {double f; uint64_t i;} u = {x};
+       uint32_t hx = u.i>>32 & 0x7fffffff;
+       int k, sign = u.i>>63;
 
        /* filter out huge and non-finite argument */
        if (hx >= 0x4043687A) {  /* if |x|>=56*ln2 */
-               if (hx >= 0x40862E42) {  /* if |x|>=709.78... */
-                       if (hx >= 0x7ff00000) {
-                               uint32_t low;
-
-                               GET_LOW_WORD(low, x);
-                               if (((hx&0xfffff)|low) != 0) /* NaN */
-                                       return x+x;
-                               return xsb==0 ? x : -1.0; /* exp(+-inf)={inf,-1} */
-                       }
-                       if(x > o_threshold)
-                               return huge*huge; /* overflow */
-               }
-               if (xsb != 0) { /* x < -56*ln2, return -1.0 with inexact */
-                       /* raise inexact */
-                       if(x+tiny<0.0)
-                               return tiny-1.0;  /* return -1 */
+               if (isnan(x))
+                       return x;
+               if (sign)
+                       return -1;
+               if (x > o_threshold) {
+                       x *= 0x1p1023;
+                       return x;
                }
        }
 
        /* argument reduction */
        if (hx > 0x3fd62e42) {  /* if  |x| > 0.5 ln2 */
                if (hx < 0x3FF0A2B2) {  /* and |x| < 1.5 ln2 */
-                       if (xsb == 0) {
+                       if (!sign) {
                                hi = x - ln2_hi;
                                lo = ln2_lo;
                                k =  1;
@@ -164,17 +150,17 @@ double expm1(double x)
                                k = -1;
                        }
                } else {
-                       k  = invln2*x + (xsb==0 ? 0.5 : -0.5);
+                       k  = invln2*x + (sign ? -0.5 : 0.5);
                        t  = k;
                        hi = x - t*ln2_hi;  /* t*ln2_hi is exact here */
                        lo = t*ln2_lo;
                }
-               STRICT_ASSIGN(double, x, hi - lo);
+               x = hi-lo;
                c = (hi-x)-lo;
        } else if (hx < 0x3c900000) {  /* |x| < 2**-54, return x */
-               /* raise inexact flags when x != 0 */
-               t = huge+x;
-               return x - (t-(huge+x));
+               if (hx < 0x00100000)
+                       FORCE_EVAL((float)x);
+               return x;
        } else
                k = 0;
 
@@ -186,9 +172,9 @@ double expm1(double x)
        e  = hxs*((r1-t)/(6.0 - x*t));
        if (k == 0)   /* c is 0 */
                return x - (x*e-hxs);
-       INSERT_WORDS(twopk, 0x3ff00000+(k<<20), 0);  /* 2^k */
        e  = x*(e-c) - c;
        e -= hxs;
+       /* exp(x) ~ 2^k (x_reduced - e + 1) */
        if (k == -1)
                return 0.5*(x-e) - 0.5;
        if (k == 1) {
@@ -196,24 +182,20 @@ double expm1(double x)
                        return -2.0*(e-(x+0.5));
                return 1.0+2.0*(x-e);
        }
-       if (k <= -2 || k > 56) {  /* suffice to return exp(x)-1 */
-               y = 1.0 - (e-x);
+       u.i = (uint64_t)(0x3ff + k)<<52;  /* 2^k */
+       twopk = u.f;
+       if (k < 0 || k > 56) {  /* suffice to return exp(x)-1 */
+               y = x - e + 1.0;
                if (k == 1024)
                        y = y*2.0*0x1p1023;
                else
                        y = y*twopk;
                return y - 1.0;
        }
-       t = 1.0;
-       if (k < 20) {
-               SET_HIGH_WORD(t, 0x3ff00000 - (0x200000>>k));  /* t=1-2^-k */
-               y = t-(e-x);
-               y = y*twopk;
-       } else {
-               SET_HIGH_WORD(t, ((0x3ff-k)<<20));  /* 2^-k */
-               y = x-(e+t);
-               y += 1.0;
-               y = y*twopk;
-       }
+       u.i = (uint64_t)(0x3ff - k)<<52;  /* 2^-k */
+       if (k < 20)
+               y = (x-e+(1-u.f))*twopk;
+       else
+               y = (x-(e+u.f)+1)*twopk;
        return y;
 }