This is the mail archive of the glibc-cvs@sourceware.org mailing list for the glibc 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]

GNU C Library master sources branch master updated. glibc-2.19-707-gbe25493


This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "GNU C Library master sources".

The branch, master has been updated
       via  be25493251c014a696b62834b0e63b155d8fe353 (commit)
      from  b351d85aa284f81271bc93e11301e1c2466b7c82 (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.

- Log -----------------------------------------------------------------
http://sourceware.org/git/gitweb.cgi?p=glibc.git;a=commitdiff;h=be25493251c014a696b62834b0e63b155d8fe353

commit be25493251c014a696b62834b0e63b155d8fe353
Author: Joseph Myers <joseph@codesourcery.com>
Date:   Fri Jun 27 14:52:13 2014 +0000

    Fix yn overflow handling in non-default rounding modes (bug 16561, bug 16562).
    
    This patch fixes bugs 16561 and 16562, bad results of yn in overflow
    cases in non-default rounding modes, both because an intermediate
    overflow in the recurrence does not get detected if the result is not
    an infinity and because an overflowing result may occur in the wrong
    sign.  The fix is to set FE_TONEAREST mode internally for the parts of
    the function where such overflows can occur (which includes the call
    to y1 - where yn is used to compute a Bessel function of order -1,
    negating the result of y1 isn't correct for overflowing results in
    directed rounding modes) and then compute an overflowing value in the
    original rounding mode if the to-nearest result was an infinity.
    
    Tested x86_64 and x86 and ulps updated accordingly.  Also tested for
    mips64 and powerpc32 to test the ldbl-128 and ldbl-128ibm changes.
    
    (The tests for these bugs were added in my previous y1 patch, so the
    only thing this patch has to do with the testsuite is enable yn
    testing in all rounding modes.)
    
    	[BZ #16561]
    	[BZ #16562]
    	* sysdeps/ieee754/dbl-64/e_jn.c: Include <float.h>.
    	(__ieee754_yn): Set FE_TONEAREST mode internally and then
    	recompute overflowing results in original rounding mode.
    	* sysdeps/ieee754/flt-32/e_jnf.c: Include <float.h>.
    	(__ieee754_ynf): Set FE_TONEAREST mode internally and then
    	recompute overflowing results in original rounding mode.
    	* sysdeps/ieee754/ldbl-128/e_jnl.c: Include <float.h>.
    	(__ieee754_ynl): Set FE_TONEAREST mode internally and then
    	recompute overflowing results in original rounding mode.
    	* sysdeps/ieee754/ldbl-128ibm/e_jnl.c: Include <float.h>.
    	(__ieee754_ynl): Set FE_TONEAREST mode internally and then
    	recompute overflowing results in original rounding mode.
    	* sysdeps/ieee754/ldbl-96/e_jnl.c: Include <float.h>.
    	(__ieee754_ynl): Set FE_TONEAREST mode internally and then
    	recompute overflowing results in original rounding mode.
    	* sysdeps/i386/fpu/fenv_private.h [!__SSE2_MATH__]
    	(libc_feholdsetround_ctx): New macro.
    	* math/libm-test.inc (yn_test): Use ALL_RM_TEST.
    	* sysdeps/i386/fpu/libm-test-ulps: Update.
    	* sysdeps/x86_64/fpu/libm-test-ulps : Likewise.

diff --git a/ChangeLog b/ChangeLog
index 83ef754..027bf9d 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,28 @@
+2014-06-27  Joseph Myers  <joseph@codesourcery.com>
+
+	[BZ #16561]
+	[BZ #16562]
+	* sysdeps/ieee754/dbl-64/e_jn.c: Include <float.h>.
+	(__ieee754_yn): Set FE_TONEAREST mode internally and then
+	recompute overflowing results in original rounding mode.
+	* sysdeps/ieee754/flt-32/e_jnf.c: Include <float.h>.
+	(__ieee754_ynf): Set FE_TONEAREST mode internally and then
+	recompute overflowing results in original rounding mode.
+	* sysdeps/ieee754/ldbl-128/e_jnl.c: Include <float.h>.
+	(__ieee754_ynl): Set FE_TONEAREST mode internally and then
+	recompute overflowing results in original rounding mode.
+	* sysdeps/ieee754/ldbl-128ibm/e_jnl.c: Include <float.h>.
+	(__ieee754_ynl): Set FE_TONEAREST mode internally and then
+	recompute overflowing results in original rounding mode.
+	* sysdeps/ieee754/ldbl-96/e_jnl.c: Include <float.h>.
+	(__ieee754_ynl): Set FE_TONEAREST mode internally and then
+	recompute overflowing results in original rounding mode.
+	* sysdeps/i386/fpu/fenv_private.h [!__SSE2_MATH__]
+	(libc_feholdsetround_ctx): New macro.
+	* math/libm-test.inc (yn_test): Use ALL_RM_TEST.
+	* sysdeps/i386/fpu/libm-test-ulps: Update.
+	* sysdeps/x86_64/fpu/libm-test-ulps : Likewise.
+
 2014-06-26  Joseph Myers  <joseph@codesourcery.com>
 
 	* sysdeps/unix/sysv/linux/mips/mips64/nptl/sysdep-cancel.h
diff --git a/NEWS b/NEWS
index e854c25..7663f01 100644
--- a/NEWS
+++ b/NEWS
@@ -12,16 +12,16 @@ Version 2.20
   6804, 9894, 12994, 13347, 13651, 14308, 14770, 15119, 15132, 15347, 15514,
   15698, 15804, 15894, 15946, 16002, 16064, 16095, 16198, 16284, 16287,
   16315, 16348, 16349, 16354, 16357, 16362, 16447, 16516, 16532, 16539,
-  16545, 16564, 16574, 16599, 16600, 16609, 16610, 16611, 16613, 16619,
-  16623, 16629, 16632, 16634, 16639, 16642, 16648, 16649, 16670, 16674,
-  16677, 16680, 16681, 16683, 16689, 16695, 16701, 16706, 16707, 16712,
-  16713, 16714, 16724, 16731, 16739, 16740, 16743, 16754, 16758, 16759,
-  16760, 16770, 16786, 16789, 16791, 16796, 16799, 16800, 16815, 16823,
-  16824, 16831, 16838, 16849, 16854, 16876, 16877, 16878, 16882, 16885,
-  16888, 16890, 16912, 16915, 16916, 16917, 16918, 16922, 16927, 16928,
-  16932, 16943, 16958, 16965, 16966, 16967, 16977, 16978, 16984, 16990,
-  16996, 17009, 17022, 17031, 17042, 17048, 17050, 17058, 17061, 17062,
-  17069, 17075, 17079, 17084, 17086.
+  16545, 16561, 16562, 16564, 16574, 16599, 16600, 16609, 16610, 16611,
+  16613, 16619, 16623, 16629, 16632, 16634, 16639, 16642, 16648, 16649,
+  16670, 16674, 16677, 16680, 16681, 16683, 16689, 16695, 16701, 16706,
+  16707, 16712, 16713, 16714, 16724, 16731, 16739, 16740, 16743, 16754,
+  16758, 16759, 16760, 16770, 16786, 16789, 16791, 16796, 16799, 16800,
+  16815, 16823, 16824, 16831, 16838, 16849, 16854, 16876, 16877, 16878,
+  16882, 16885, 16888, 16890, 16912, 16915, 16916, 16917, 16918, 16922,
+  16927, 16928, 16932, 16943, 16958, 16965, 16966, 16967, 16977, 16978,
+  16984, 16990, 16996, 17009, 17022, 17031, 17042, 17048, 17050, 17058,
+  17061, 17062, 17069, 17075, 17079, 17084, 17086.
 
 * Optimized strchr implementation for AArch64.  Contributed by ARM Ltd.
 
diff --git a/math/libm-test.inc b/math/libm-test.inc
index 45eaaef..f86a4fa 100644
--- a/math/libm-test.inc
+++ b/math/libm-test.inc
@@ -9609,9 +9609,7 @@ static const struct test_if_f_data yn_test_data[] =
 static void
 yn_test (void)
 {
-  START (yn, 0);
-  RUN_TEST_LOOP_if_f (yn, yn_test_data, );
-  END;
+  ALL_RM_TEST (yn, 0, yn_test_data, RUN_TEST_LOOP_if_f, END);
 }
 
 
diff --git a/sysdeps/i386/fpu/fenv_private.h b/sysdeps/i386/fpu/fenv_private.h
index 3998387..e20e1f1 100644
--- a/sysdeps/i386/fpu/fenv_private.h
+++ b/sysdeps/i386/fpu/fenv_private.h
@@ -482,6 +482,7 @@ libc_feupdateenv_387_ctx (struct rm_ctx *ctx)
 #else
 # define libc_feholdexcept_setround_ctx	libc_feholdexcept_setround_387_ctx
 # define libc_feupdateenv_ctx		libc_feupdateenv_387_ctx
+# define libc_feholdsetround_ctx	libc_feholdsetround_387_ctx
 # define libc_feresetround_ctx		libc_feresetround_387_ctx
 #endif /* __SSE2_MATH__ */
 
diff --git a/sysdeps/i386/fpu/libm-test-ulps b/sysdeps/i386/fpu/libm-test-ulps
index 7cf9586..be0e7c3 100644
--- a/sysdeps/i386/fpu/libm-test-ulps
+++ b/sysdeps/i386/fpu/libm-test-ulps
@@ -1931,4 +1931,28 @@ ifloat: 3
 ildouble: 4
 ldouble: 4
 
+Function: "yn_downward":
+double: 2
+float: 2
+idouble: 2
+ifloat: 2
+ildouble: 5
+ldouble: 5
+
+Function: "yn_towardzero":
+double: 2
+float: 3
+idouble: 2
+ifloat: 3
+ildouble: 5
+ldouble: 5
+
+Function: "yn_upward":
+double: 2
+float: 3
+idouble: 2
+ifloat: 3
+ildouble: 4
+ldouble: 4
+
 # end of automatic generation
diff --git a/sysdeps/ieee754/dbl-64/e_jn.c b/sysdeps/ieee754/dbl-64/e_jn.c
index 236878b..bbe0426 100644
--- a/sysdeps/ieee754/dbl-64/e_jn.c
+++ b/sysdeps/ieee754/dbl-64/e_jn.c
@@ -37,6 +37,7 @@
  */
 
 #include <errno.h>
+#include <float.h>
 #include <math.h>
 #include <math_private.h>
 
@@ -248,7 +249,7 @@ __ieee754_yn (int n, double x)
 {
   int32_t i, hx, ix, lx;
   int32_t sign;
-  double a, b, temp;
+  double a, b, temp, ret;
 
   EXTRACT_WORDS (hx, lx, x);
   ix = 0x7fffffff & hx;
@@ -268,57 +269,67 @@ __ieee754_yn (int n, double x)
     }
   if (n == 0)
     return (__ieee754_y0 (x));
-  if (n == 1)
-    return (sign * __ieee754_y1 (x));
-  if (__glibc_unlikely (ix == 0x7ff00000))
-    return zero;
-  if (ix >= 0x52D00000)      /* x > 2**302 */
-    { /* (x >> n**2)
-		 *	    Jn(x) = cos(x-(2n+1)*pi/4)*sqrt(2/x*pi)
-		 *	    Yn(x) = sin(x-(2n+1)*pi/4)*sqrt(2/x*pi)
-		 *	    Let s=sin(x), c=cos(x),
-		 *		xn=x-(2n+1)*pi/4, sqt2 = sqrt(2),then
-		 *
-		 *		   n	sin(xn)*sqt2	cos(xn)*sqt2
-		 *		----------------------------------
-		 *		   0	 s-c		 c+s
-		 *		   1	-s-c		-c+s
-		 *		   2	-s+c		-c-s
-		 *		   3	 s+c		 c-s
-		 */
-      double c;
-      double s;
-      __sincos (x, &s, &c);
-      switch (n & 3)
-	{
-	case 0: temp = s - c; break;
-	case 1: temp = -s - c; break;
-	case 2: temp = -s + c; break;
-	case 3: temp = s + c; break;
-	}
-      b = invsqrtpi * temp / __ieee754_sqrt (x);
-    }
-  else
-    {
-      u_int32_t high;
-      a = __ieee754_y0 (x);
-      b = __ieee754_y1 (x);
-      /* quit if b is -inf */
-      GET_HIGH_WORD (high, b);
-      for (i = 1; i < n && high != 0xfff00000; i++)
-	{
-	  temp = b;
-	  b = ((double) (i + i) / x) * b - a;
-	  GET_HIGH_WORD (high, b);
-	  a = temp;
-	}
-      /* If B is +-Inf, set up errno accordingly.  */
-      if (!__finite (b))
-	__set_errno (ERANGE);
-    }
-  if (sign > 0)
-    return b;
-  else
-    return -b;
+  {
+    SET_RESTORE_ROUND (FE_TONEAREST);
+    if (n == 1)
+      {
+	ret = sign * __ieee754_y1 (x);
+	goto out;
+      }
+    if (__glibc_unlikely (ix == 0x7ff00000))
+      return zero;
+    if (ix >= 0x52D00000)      /* x > 2**302 */
+      { /* (x >> n**2)
+	 *	    Jn(x) = cos(x-(2n+1)*pi/4)*sqrt(2/x*pi)
+	 *	    Yn(x) = sin(x-(2n+1)*pi/4)*sqrt(2/x*pi)
+	 *	    Let s=sin(x), c=cos(x),
+	 *		xn=x-(2n+1)*pi/4, sqt2 = sqrt(2),then
+	 *
+	 *		   n	sin(xn)*sqt2	cos(xn)*sqt2
+	 *		----------------------------------
+	 *		   0	 s-c		 c+s
+	 *		   1	-s-c		-c+s
+	 *		   2	-s+c		-c-s
+	 *		   3	 s+c		 c-s
+	 */
+	double c;
+	double s;
+	__sincos (x, &s, &c);
+	switch (n & 3)
+	  {
+	  case 0: temp = s - c; break;
+	  case 1: temp = -s - c; break;
+	  case 2: temp = -s + c; break;
+	  case 3: temp = s + c; break;
+	  }
+	b = invsqrtpi * temp / __ieee754_sqrt (x);
+      }
+    else
+      {
+	u_int32_t high;
+	a = __ieee754_y0 (x);
+	b = __ieee754_y1 (x);
+	/* quit if b is -inf */
+	GET_HIGH_WORD (high, b);
+	for (i = 1; i < n && high != 0xfff00000; i++)
+	  {
+	    temp = b;
+	    b = ((double) (i + i) / x) * b - a;
+	    GET_HIGH_WORD (high, b);
+	    a = temp;
+	  }
+	/* If B is +-Inf, set up errno accordingly.  */
+	if (!__finite (b))
+	  __set_errno (ERANGE);
+      }
+    if (sign > 0)
+      ret = b;
+    else
+      ret = -b;
+  }
+ out:
+  if (__isinf (ret))
+    ret = __copysign (DBL_MAX, ret) * DBL_MAX;
+  return ret;
 }
 strong_alias (__ieee754_yn, __yn_finite)
diff --git a/sysdeps/ieee754/flt-32/e_jnf.c b/sysdeps/ieee754/flt-32/e_jnf.c
index 5984d94..86085cc 100644
--- a/sysdeps/ieee754/flt-32/e_jnf.c
+++ b/sysdeps/ieee754/flt-32/e_jnf.c
@@ -14,6 +14,7 @@
  */
 
 #include <errno.h>
+#include <float.h>
 #include <math.h>
 #include <math_private.h>
 
@@ -169,6 +170,8 @@ strong_alias (__ieee754_jnf, __jnf_finite)
 float
 __ieee754_ynf(int n, float x)
 {
+    float ret;
+    {
 	int32_t i,hx,ix;
 	u_int32_t ib;
 	int32_t sign;
@@ -187,7 +190,11 @@ __ieee754_ynf(int n, float x)
 		sign = 1 - ((n&1)<<1);
 	}
 	if(n==0) return(__ieee754_y0f(x));
-	if(n==1) return(sign*__ieee754_y1f(x));
+	SET_RESTORE_ROUNDF (FE_TONEAREST);
+	if(n==1) {
+	    ret = sign*__ieee754_y1f(x);
+	    goto out;
+	}
 	if(__builtin_expect(ix==0x7f800000, 0)) return zero;
 
 	a = __ieee754_y0f(x);
@@ -203,6 +210,11 @@ __ieee754_ynf(int n, float x)
 	/* If B is +-Inf, set up errno accordingly.  */
 	if (! __finitef (b))
 	  __set_errno (ERANGE);
-	if(sign>0) return b; else return -b;
+	if(sign>0) ret = b; else ret = -b;
+    }
+ out:
+    if (__isinff (ret))
+	ret = __copysignf (FLT_MAX, ret) * FLT_MAX;
+    return ret;
 }
 strong_alias (__ieee754_ynf, __ynf_finite)
diff --git a/sysdeps/ieee754/ldbl-128/e_jnl.c b/sysdeps/ieee754/ldbl-128/e_jnl.c
index c2a4923..4e32d38 100644
--- a/sysdeps/ieee754/ldbl-128/e_jnl.c
+++ b/sysdeps/ieee754/ldbl-128/e_jnl.c
@@ -57,6 +57,7 @@
  */
 
 #include <errno.h>
+#include <float.h>
 #include <math.h>
 #include <math_private.h>
 
@@ -300,7 +301,7 @@ __ieee754_ynl (int n, long double x)
   u_int32_t se;
   int32_t i, ix;
   int32_t sign;
-  long double a, b, temp;
+  long double a, b, temp, ret;
   ieee854_long_double_shape_type u;
 
   u.value = x;
@@ -328,70 +329,80 @@ __ieee754_ynl (int n, long double x)
     }
   if (n == 0)
     return (__ieee754_y0l (x));
-  if (n == 1)
-    return (sign * __ieee754_y1l (x));
-  if (ix >= 0x7fff0000)
-    return zero;
-  if (ix >= 0x412D0000)
-    {				/* x > 2**302 */
+  {
+    SET_RESTORE_ROUNDL (FE_TONEAREST);
+    if (n == 1)
+      {
+	ret = sign * __ieee754_y1l (x);
+	goto out;
+      }
+    if (ix >= 0x7fff0000)
+      return zero;
+    if (ix >= 0x412D0000)
+      {				/* x > 2**302 */
 
-      /* ??? See comment above on the possible futility of this.  */
+	/* ??? See comment above on the possible futility of this.  */
 
-      /* (x >> n**2)
-       *      Jn(x) = cos(x-(2n+1)*pi/4)*sqrt(2/x*pi)
-       *      Yn(x) = sin(x-(2n+1)*pi/4)*sqrt(2/x*pi)
-       *      Let s=sin(x), c=cos(x),
-       *          xn=x-(2n+1)*pi/4, sqt2 = sqrt(2),then
-       *
-       *             n    sin(xn)*sqt2    cos(xn)*sqt2
-       *          ----------------------------------
-       *             0     s-c             c+s
-       *             1    -s-c            -c+s
-       *             2    -s+c            -c-s
-       *             3     s+c             c-s
-       */
-      long double s;
-      long double c;
-      __sincosl (x, &s, &c);
-      switch (n & 3)
-	{
-	case 0:
-	  temp = s - c;
-	  break;
-	case 1:
-	  temp = -s - c;
-	  break;
-	case 2:
-	  temp = -s + c;
-	  break;
-	case 3:
-	  temp = s + c;
-	  break;
-	}
-      b = invsqrtpi * temp / __ieee754_sqrtl (x);
-    }
-  else
-    {
-      a = __ieee754_y0l (x);
-      b = __ieee754_y1l (x);
-      /* quit if b is -inf */
-      u.value = b;
-      se = u.parts32.w0 & 0xffff0000;
-      for (i = 1; i < n && se != 0xffff0000; i++)
-	{
-	  temp = b;
-	  b = ((long double) (i + i) / x) * b - a;
-	  u.value = b;
-	  se = u.parts32.w0 & 0xffff0000;
-	  a = temp;
-	}
-    }
-  /* If B is +-Inf, set up errno accordingly.  */
-  if (! __finitel (b))
-    __set_errno (ERANGE);
-  if (sign > 0)
-    return b;
-  else
-    return -b;
+	/* (x >> n**2)
+	 *      Jn(x) = cos(x-(2n+1)*pi/4)*sqrt(2/x*pi)
+	 *      Yn(x) = sin(x-(2n+1)*pi/4)*sqrt(2/x*pi)
+	 *      Let s=sin(x), c=cos(x),
+	 *          xn=x-(2n+1)*pi/4, sqt2 = sqrt(2),then
+	 *
+	 *             n    sin(xn)*sqt2    cos(xn)*sqt2
+	 *          ----------------------------------
+	 *             0     s-c             c+s
+	 *             1    -s-c            -c+s
+	 *             2    -s+c            -c-s
+	 *             3     s+c             c-s
+	 */
+	long double s;
+	long double c;
+	__sincosl (x, &s, &c);
+	switch (n & 3)
+	  {
+	  case 0:
+	    temp = s - c;
+	    break;
+	  case 1:
+	    temp = -s - c;
+	    break;
+	  case 2:
+	    temp = -s + c;
+	    break;
+	  case 3:
+	    temp = s + c;
+	    break;
+	  }
+	b = invsqrtpi * temp / __ieee754_sqrtl (x);
+      }
+    else
+      {
+	a = __ieee754_y0l (x);
+	b = __ieee754_y1l (x);
+	/* quit if b is -inf */
+	u.value = b;
+	se = u.parts32.w0 & 0xffff0000;
+	for (i = 1; i < n && se != 0xffff0000; i++)
+	  {
+	    temp = b;
+	    b = ((long double) (i + i) / x) * b - a;
+	    u.value = b;
+	    se = u.parts32.w0 & 0xffff0000;
+	    a = temp;
+	  }
+      }
+    /* If B is +-Inf, set up errno accordingly.  */
+    if (! __finitel (b))
+      __set_errno (ERANGE);
+    if (sign > 0)
+      ret = b;
+    else
+      ret = -b;
+  }
+ out:
+  if (__isinfl (ret))
+    ret = __copysignl (LDBL_MAX, ret) * LDBL_MAX;
+  return ret;
 }
 strong_alias (__ieee754_ynl, __ynl_finite)
diff --git a/sysdeps/ieee754/ldbl-128ibm/e_jnl.c b/sysdeps/ieee754/ldbl-128ibm/e_jnl.c
index 6761a0d..589f1f8 100644
--- a/sysdeps/ieee754/ldbl-128ibm/e_jnl.c
+++ b/sysdeps/ieee754/ldbl-128ibm/e_jnl.c
@@ -57,6 +57,7 @@
  */
 
 #include <errno.h>
+#include <float.h>
 #include <math.h>
 #include <math_private.h>
 
@@ -300,7 +301,7 @@ __ieee754_ynl (int n, long double x)
   uint32_t se, lx;
   int32_t i, ix;
   int32_t sign;
-  long double a, b, temp;
+  long double a, b, temp, ret;
   double xhi;
 
   xhi = ldbl_high (x);
@@ -328,72 +329,82 @@ __ieee754_ynl (int n, long double x)
     }
   if (n == 0)
     return (__ieee754_y0l (x));
-  if (n == 1)
-    return (sign * __ieee754_y1l (x));
-  if (ix >= 0x7ff00000)
-    return zero;
-  if (ix >= 0x52D00000)
-    {				/* x > 2**302 */
+  {
+    SET_RESTORE_ROUNDL (FE_TONEAREST);
+    if (n == 1)
+      {
+	ret = sign * __ieee754_y1l (x);
+	goto out;
+      }
+    if (ix >= 0x7ff00000)
+      return zero;
+    if (ix >= 0x52D00000)
+      {				/* x > 2**302 */
 
-      /* ??? See comment above on the possible futility of this.  */
+	/* ??? See comment above on the possible futility of this.  */
 
-      /* (x >> n**2)
-       *      Jn(x) = cos(x-(2n+1)*pi/4)*sqrt(2/x*pi)
-       *      Yn(x) = sin(x-(2n+1)*pi/4)*sqrt(2/x*pi)
-       *      Let s=sin(x), c=cos(x),
-       *          xn=x-(2n+1)*pi/4, sqt2 = sqrt(2),then
-       *
-       *             n    sin(xn)*sqt2    cos(xn)*sqt2
-       *          ----------------------------------
-       *             0     s-c             c+s
-       *             1    -s-c            -c+s
-       *             2    -s+c            -c-s
-       *             3     s+c             c-s
-       */
-      long double s;
-      long double c;
-      __sincosl (x, &s, &c);
-      switch (n & 3)
-	{
-	case 0:
-	  temp = s - c;
-	  break;
-	case 1:
-	  temp = -s - c;
-	  break;
-	case 2:
-	  temp = -s + c;
-	  break;
-	case 3:
-	  temp = s + c;
-	  break;
-	}
-      b = invsqrtpi * temp / __ieee754_sqrtl (x);
-    }
-  else
-    {
-      a = __ieee754_y0l (x);
-      b = __ieee754_y1l (x);
-      /* quit if b is -inf */
-      xhi = ldbl_high (b);
-      GET_HIGH_WORD (se, xhi);
-      se &= 0xfff00000;
-      for (i = 1; i < n && se != 0xfff00000; i++)
-	{
-	  temp = b;
-	  b = ((long double) (i + i) / x) * b - a;
-	  xhi = ldbl_high (b);
-	  GET_HIGH_WORD (se, xhi);
-	  se &= 0xfff00000;
-	  a = temp;
-	}
-    }
-  /* If B is +-Inf, set up errno accordingly.  */
-  if (! __finitel (b))
-    __set_errno (ERANGE);
-  if (sign > 0)
-    return b;
-  else
-    return -b;
+	/* (x >> n**2)
+	 *      Jn(x) = cos(x-(2n+1)*pi/4)*sqrt(2/x*pi)
+	 *      Yn(x) = sin(x-(2n+1)*pi/4)*sqrt(2/x*pi)
+	 *      Let s=sin(x), c=cos(x),
+	 *          xn=x-(2n+1)*pi/4, sqt2 = sqrt(2),then
+	 *
+	 *             n    sin(xn)*sqt2    cos(xn)*sqt2
+	 *          ----------------------------------
+	 *             0     s-c             c+s
+	 *             1    -s-c            -c+s
+	 *             2    -s+c            -c-s
+	 *             3     s+c             c-s
+	 */
+	long double s;
+	long double c;
+	__sincosl (x, &s, &c);
+	switch (n & 3)
+	  {
+	  case 0:
+	    temp = s - c;
+	    break;
+	  case 1:
+	    temp = -s - c;
+	    break;
+	  case 2:
+	    temp = -s + c;
+	    break;
+	  case 3:
+	    temp = s + c;
+	    break;
+	  }
+	b = invsqrtpi * temp / __ieee754_sqrtl (x);
+      }
+    else
+      {
+	a = __ieee754_y0l (x);
+	b = __ieee754_y1l (x);
+	/* quit if b is -inf */
+	xhi = ldbl_high (b);
+	GET_HIGH_WORD (se, xhi);
+	se &= 0xfff00000;
+	for (i = 1; i < n && se != 0xfff00000; i++)
+	  {
+	    temp = b;
+	    b = ((long double) (i + i) / x) * b - a;
+	    xhi = ldbl_high (b);
+	    GET_HIGH_WORD (se, xhi);
+	    se &= 0xfff00000;
+	    a = temp;
+	  }
+      }
+    /* If B is +-Inf, set up errno accordingly.  */
+    if (! __finitel (b))
+      __set_errno (ERANGE);
+    if (sign > 0)
+      ret = b;
+    else
+      ret = -b;
+  }
+ out:
+  if (__isinfl (ret))
+    ret = __copysignl (LDBL_MAX, ret) * LDBL_MAX;
+  return ret;
 }
 strong_alias (__ieee754_ynl, __ynl_finite)
diff --git a/sysdeps/ieee754/ldbl-96/e_jnl.c b/sysdeps/ieee754/ldbl-96/e_jnl.c
index 11d097c..95ff242 100644
--- a/sysdeps/ieee754/ldbl-96/e_jnl.c
+++ b/sysdeps/ieee754/ldbl-96/e_jnl.c
@@ -57,6 +57,7 @@
  */
 
 #include <errno.h>
+#include <float.h>
 #include <math.h>
 #include <math_private.h>
 
@@ -293,7 +294,7 @@ __ieee754_ynl (int n, long double x)
   u_int32_t se, i0, i1;
   int32_t i, ix;
   int32_t sign;
-  long double a, b, temp;
+  long double a, b, temp, ret;
 
 
   GET_LDOUBLE_WORDS (se, i0, i1, x);
@@ -314,69 +315,79 @@ __ieee754_ynl (int n, long double x)
     }
   if (n == 0)
     return (__ieee754_y0l (x));
-  if (n == 1)
-    return (sign * __ieee754_y1l (x));
-  if (__glibc_unlikely (ix == 0x7fff))
-    return zero;
-  if (ix >= 0x412D)
-    {				/* x > 2**302 */
+  {
+    SET_RESTORE_ROUNDL (FE_TONEAREST);
+    if (n == 1)
+      {
+	ret = sign * __ieee754_y1l (x);
+	goto out;
+      }
+    if (__glibc_unlikely (ix == 0x7fff))
+      return zero;
+    if (ix >= 0x412D)
+      {				/* x > 2**302 */
 
-      /* ??? See comment above on the possible futility of this.  */
+	/* ??? See comment above on the possible futility of this.  */
 
-      /* (x >> n**2)
-       *      Jn(x) = cos(x-(2n+1)*pi/4)*sqrt(2/x*pi)
-       *      Yn(x) = sin(x-(2n+1)*pi/4)*sqrt(2/x*pi)
-       *      Let s=sin(x), c=cos(x),
-       *          xn=x-(2n+1)*pi/4, sqt2 = sqrt(2),then
-       *
-       *             n    sin(xn)*sqt2    cos(xn)*sqt2
-       *          ----------------------------------
-       *             0     s-c             c+s
-       *             1    -s-c            -c+s
-       *             2    -s+c            -c-s
-       *             3     s+c             c-s
-       */
-      long double s;
-      long double c;
-      __sincosl (x, &s, &c);
-      switch (n & 3)
-	{
-	case 0:
-	  temp = s - c;
-	  break;
-	case 1:
-	  temp = -s - c;
-	  break;
-	case 2:
-	  temp = -s + c;
-	  break;
-	case 3:
-	  temp = s + c;
-	  break;
-	}
-      b = invsqrtpi * temp / __ieee754_sqrtl (x);
-    }
-  else
-    {
-      a = __ieee754_y0l (x);
-      b = __ieee754_y1l (x);
-      /* quit if b is -inf */
-      GET_LDOUBLE_WORDS (se, i0, i1, b);
-      /* Use 0xffffffff since GET_LDOUBLE_WORDS sign-extends SE.  */
-      for (i = 1; i < n && se != 0xffffffff; i++)
-	{
-	  temp = b;
-	  b = ((long double) (i + i) / x) * b - a;
-	  GET_LDOUBLE_WORDS (se, i0, i1, b);
-	  a = temp;
-	}
-    }
-  /* If B is +-Inf, set up errno accordingly.  */
-  if (! __finitel (b))
-    __set_errno (ERANGE);
-  if (sign > 0)
-    return b;
-  else
-    return -b;
+	/* (x >> n**2)
+	 *      Jn(x) = cos(x-(2n+1)*pi/4)*sqrt(2/x*pi)
+	 *      Yn(x) = sin(x-(2n+1)*pi/4)*sqrt(2/x*pi)
+	 *      Let s=sin(x), c=cos(x),
+	 *          xn=x-(2n+1)*pi/4, sqt2 = sqrt(2),then
+	 *
+	 *             n    sin(xn)*sqt2    cos(xn)*sqt2
+	 *          ----------------------------------
+	 *             0     s-c             c+s
+	 *             1    -s-c            -c+s
+	 *             2    -s+c            -c-s
+	 *             3     s+c             c-s
+	 */
+	long double s;
+	long double c;
+	__sincosl (x, &s, &c);
+	switch (n & 3)
+	  {
+	  case 0:
+	    temp = s - c;
+	    break;
+	  case 1:
+	    temp = -s - c;
+	    break;
+	  case 2:
+	    temp = -s + c;
+	    break;
+	  case 3:
+	    temp = s + c;
+	    break;
+	  }
+	b = invsqrtpi * temp / __ieee754_sqrtl (x);
+      }
+    else
+      {
+	a = __ieee754_y0l (x);
+	b = __ieee754_y1l (x);
+	/* quit if b is -inf */
+	GET_LDOUBLE_WORDS (se, i0, i1, b);
+	/* Use 0xffffffff since GET_LDOUBLE_WORDS sign-extends SE.  */
+	for (i = 1; i < n && se != 0xffffffff; i++)
+	  {
+	    temp = b;
+	    b = ((long double) (i + i) / x) * b - a;
+	    GET_LDOUBLE_WORDS (se, i0, i1, b);
+	    a = temp;
+	  }
+      }
+    /* If B is +-Inf, set up errno accordingly.  */
+    if (! __finitel (b))
+      __set_errno (ERANGE);
+    if (sign > 0)
+      ret = b;
+    else
+      ret = -b;
+  }
+ out:
+  if (__isinfl (ret))
+    ret = __copysignl (LDBL_MAX, ret) * LDBL_MAX;
+  return ret;
 }
 strong_alias (__ieee754_ynl, __ynl_finite)
diff --git a/sysdeps/x86_64/fpu/libm-test-ulps b/sysdeps/x86_64/fpu/libm-test-ulps
index f68650d..36e1b76 100644
--- a/sysdeps/x86_64/fpu/libm-test-ulps
+++ b/sysdeps/x86_64/fpu/libm-test-ulps
@@ -2033,4 +2033,28 @@ ifloat: 3
 ildouble: 4
 ldouble: 4
 
+Function: "yn_downward":
+double: 3
+float: 4
+idouble: 3
+ifloat: 4
+ildouble: 5
+ldouble: 5
+
+Function: "yn_towardzero":
+double: 3
+float: 3
+idouble: 3
+ifloat: 3
+ildouble: 5
+ldouble: 5
+
+Function: "yn_upward":
+double: 4
+float: 5
+idouble: 4
+ifloat: 5
+ildouble: 4
+ldouble: 4
+
 # end of automatic generation

-----------------------------------------------------------------------

Summary of changes:
 ChangeLog                           |   25 ++++++
 NEWS                                |   20 +++---
 math/libm-test.inc                  |    4 +-
 sysdeps/i386/fpu/fenv_private.h     |    1 +
 sysdeps/i386/fpu/libm-test-ulps     |   24 ++++++
 sysdeps/ieee754/dbl-64/e_jn.c       |  117 ++++++++++++++++-------------
 sysdeps/ieee754/flt-32/e_jnf.c      |   16 ++++-
 sysdeps/ieee754/ldbl-128/e_jnl.c    |  139 ++++++++++++++++++----------------
 sysdeps/ieee754/ldbl-128ibm/e_jnl.c |  143 +++++++++++++++++++----------------
 sysdeps/ieee754/ldbl-96/e_jnl.c     |  137 ++++++++++++++++++---------------
 sysdeps/x86_64/fpu/libm-test-ulps   |   24 ++++++
 11 files changed, 389 insertions(+), 261 deletions(-)


hooks/post-receive
-- 
GNU C Library master sources


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