diff options
Diffstat (limited to 'sysdeps/x86_64/fpu')
-rw-r--r-- | sysdeps/x86_64/fpu/e_log10l.S | 5 | ||||
-rw-r--r-- | sysdeps/x86_64/fpu/e_log2l.S | 5 | ||||
-rw-r--r-- | sysdeps/x86_64/fpu/e_logl.S | 12 | ||||
-rw-r--r-- | sysdeps/x86_64/fpu/e_powl.S | 19 | ||||
-rw-r--r-- | sysdeps/x86_64/fpu/math_private.h | 21 | ||||
-rw-r--r-- | sysdeps/x86_64/fpu/s_copysign.S | 5 | ||||
-rw-r--r-- | sysdeps/x86_64/fpu/s_copysignf.S | 3 | ||||
-rw-r--r-- | sysdeps/x86_64/fpu/s_log1pl.S | 5 |
8 files changed, 54 insertions, 21 deletions
diff --git a/sysdeps/x86_64/fpu/e_log10l.S b/sysdeps/x86_64/fpu/e_log10l.S index b4343bef45..633234b744 100644 --- a/sysdeps/x86_64/fpu/e_log10l.S +++ b/sysdeps/x86_64/fpu/e_log10l.S @@ -42,7 +42,7 @@ ENTRY(__ieee754_log10l) fxam fnstsw fld %st // x : x : log10(2) - andb $1,%ah + testb $1, %ah jnz 3f // in case x is NaN or ħInf 4: fsubl MO(one) // x-1 : x : log10(2) fld %st // x-1 : x-1 : x : log10(2) @@ -59,7 +59,8 @@ ENTRY(__ieee754_log10l) fyl2x // log10(x) ret -3: jp 4b // in case x is ħInf +3: testb $4, %ah + jnz 4b // in case x is ħInf fstp %st(1) fstp %st(1) ret diff --git a/sysdeps/x86_64/fpu/e_log2l.S b/sysdeps/x86_64/fpu/e_log2l.S index 7a89b94d9f..f04d30a05a 100644 --- a/sysdeps/x86_64/fpu/e_log2l.S +++ b/sysdeps/x86_64/fpu/e_log2l.S @@ -39,7 +39,7 @@ ENTRY(__ieee754_log2l) fxam fnstsw fld %st // x : x : 1 - andb $1,%ah + testb $1, %ah jnz 3f // in case x is NaN or ħInf 4: fsub %st(2), %st // x-1 : x : 1 fld %st // x-1 : x-1 : x : 1 @@ -56,7 +56,8 @@ ENTRY(__ieee754_log2l) fyl2x // log(x) ret -3: jp 4b // in case x is ħInf +3: testb $4, %ah + jnz 4b // in case x is ħInf fstp %st(1) fstp %st(1) ret diff --git a/sysdeps/x86_64/fpu/e_logl.S b/sysdeps/x86_64/fpu/e_logl.S index a0bed663c8..2ba91eedfd 100644 --- a/sysdeps/x86_64/fpu/e_logl.S +++ b/sysdeps/x86_64/fpu/e_logl.S @@ -38,8 +38,12 @@ limit: .double 0.29 ENTRY(__ieee754_logl) fldln2 // log(2) fldt 8(%rsp) // x : log(2) + fxam + fnstsw fld %st // x : x : log(2) - fsubl MO(one) // x-1 : x : log(2) + testb $1, %ah + jnz 3f // in case x is NaN or +-Inf +4: fsubl MO(one) // x-1 : x : log(2) fld %st // x-1 : x-1 : x : log(2) fabs // |x-1| : x-1 : x : log(2) fcompl MO(limit) // x-1 : x : log(2) @@ -53,4 +57,10 @@ ENTRY(__ieee754_logl) 2: fstp %st(0) // x : log(2) fyl2x // log(x) ret + +3: testb $4, %ah + jnz 4b // in case x is +-Inf + fstp %st(1) + fstp %st(1) + ret END (__ieee754_logl) diff --git a/sysdeps/x86_64/fpu/e_powl.S b/sysdeps/x86_64/fpu/e_powl.S index 85f4deb3c7..4959bea7ac 100644 --- a/sysdeps/x86_64/fpu/e_powl.S +++ b/sysdeps/x86_64/fpu/e_powl.S @@ -1,5 +1,6 @@ /* ix87 specific implementation of pow function. - Copyright (C) 1996, 1997, 1998, 1999, 2001, 2004 Free Software Foundation, Inc. + Copyright (C) 1996, 1997, 1998, 1999, 2001, 2004, 2007 + Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Ulrich Drepper <drepper@cygnus.com>, 1996. @@ -146,10 +147,11 @@ ENTRY(__ieee754_powl) 2: /* y is a real number. */ fxch // x : y fldl MO(one) // 1.0 : x : y - fld %st(1) // x : 1.0 : x : y - fsub %st(1) // x-1 : 1.0 : x : y - fabs // |x-1| : 1.0 : x : y - fcompl MO(limit) // 1.0 : x : y + fldl MO(limit) // 0.29 : 1.0 : x : y + fld %st(2) // x : 0.29 : 1.0 : x : y + fsub %st(2) // x-1 : 0.29 : 1.0 : x : y + fabs // |x-1| : 0.29 : 1.0 : x : y + fucompp // 1.0 : x : y fnstsw fxch // x : 1.0 : y test $4500,%eax @@ -190,9 +192,10 @@ ENTRY(__ieee754_powl) // y == ħinf .align ALIGNARG(4) 12: fstp %st(0) // pop y - fldt 8(%rsp) // x - fabs - fcompl MO(one) // < 1, == 1, or > 1 + fldl MO(one) // 1 + fldt 8(%rsp) // x : 1 + fabs // abs(x) : 1 + fucompp // < 1, == 1, or > 1 fnstsw andb $0x45, %ah cmpb $0x45, %ah diff --git a/sysdeps/x86_64/fpu/math_private.h b/sysdeps/x86_64/fpu/math_private.h new file mode 100644 index 0000000000..4febcbb5ec --- /dev/null +++ b/sysdeps/x86_64/fpu/math_private.h @@ -0,0 +1,21 @@ +#ifndef _MATH_PRIVATE_H + +#define math_opt_barrier(x) \ +({ __typeof(x) __x; \ + if (sizeof (x) <= sizeof (double)) \ + __asm ("" : "=x" (__x) : "0" (x)); \ + else \ + __asm ("" : "=t" (__x) : "0" (x)); \ + __x; }) +#define math_force_eval(x) \ +do \ + { \ + if (sizeof (x) <= sizeof (double)) \ + __asm __volatile ("" : : "x" (x)); \ + else \ + __asm __volatile ("" : : "f" (x)); \ + } \ +while (0) + +#include <math/math_private.h> +#endif diff --git a/sysdeps/x86_64/fpu/s_copysign.S b/sysdeps/x86_64/fpu/s_copysign.S index f3d9b0cbb4..f1ebcf8bf1 100644 --- a/sysdeps/x86_64/fpu/s_copysign.S +++ b/sysdeps/x86_64/fpu/s_copysign.S @@ -1,5 +1,5 @@ /* copy sign, double version. - Copyright (C) 2002, 2006 Free Software Foundation, Inc. + Copyright (C) 2002 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Andreas Jaeger <aj@suse.de>, 2002. @@ -31,8 +31,6 @@ signmask: .byte 0, 0, 0, 0, 0, 0, 0, 0x80 .byte 0, 0, 0, 0, 0, 0, 0, 0 - ASM_SIZE_DIRECTIVE(signmask) - ASM_TYPE_DIRECTIVE(othermask,@object) othermask: .byte 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x7f .byte 0, 0, 0, 0, 0, 0, 0, 0 @@ -44,7 +42,6 @@ othermask: #define MO(op) op #endif - .text ENTRY(__copysign) andpd MO(othermask),%xmm0 andpd MO(signmask),%xmm1 diff --git a/sysdeps/x86_64/fpu/s_copysignf.S b/sysdeps/x86_64/fpu/s_copysignf.S index 0fbe1d4c96..f5dc5f78ad 100644 --- a/sysdeps/x86_64/fpu/s_copysignf.S +++ b/sysdeps/x86_64/fpu/s_copysignf.S @@ -1,5 +1,5 @@ /* copy sign, double version. - Copyright (C) 2002, 2006 Free Software Foundation, Inc. + Copyright (C) 2002 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Andreas Jaeger <aj@suse.de>, 2002. @@ -38,7 +38,6 @@ mask: #define MO(op) op #endif - .text ENTRY(__copysignf) movss MO(mask),%xmm3 andps %xmm3,%xmm0 diff --git a/sysdeps/x86_64/fpu/s_log1pl.S b/sysdeps/x86_64/fpu/s_log1pl.S index 7fbd0e5aaa..ac2bd22a4f 100644 --- a/sysdeps/x86_64/fpu/s_log1pl.S +++ b/sysdeps/x86_64/fpu/s_log1pl.S @@ -45,7 +45,7 @@ ENTRY(__log1pl) fxam fnstsw fld %st - andb $1,%ah + testb $1, %ah jnz 3f // in case x is NaN or ħInf 4: fabs @@ -62,7 +62,8 @@ ENTRY(__log1pl) 2: fyl2xp1 ret -3: jp 4b // in case x is ħInf +3: testb $4, %ah + jnz 4b // in case x is ħInf fstp %st(1) fstp %st(1) ret |