'long double' support for Darwin
Geoffrey Keating
gkeating@apple.com
Tue Jan 6 23:48:00 GMT 2004
This isn't strictly a bug fix, but it makes it much easier to test
the other fixes I'm making, and I don't think it can hurt anything.
Bootstrapped & tested on powerpc-darwin, and an extra dejagnu run with
-mlong-double-128.
--
- Geoffrey Keating <geoffk@apple.com>
===File ~/patches/rs6000-longdoublelib.patch================
2004-01-06 Geoffrey Keating <geoffk@apple.com>
* config/rs6000/t-darwin (LIB2FUNCS_EXTRA): Compile darwin-ldouble.c.
(TARGET_LIBGCC2_CFLAGS): Use -mlong-double-128.
* config/rs6000/darwin-ldouble.c: New.
Index: config/rs6000/t-darwin
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/rs6000/t-darwin,v
retrieving revision 1.10
diff -u -p -u -p -r1.10 t-darwin
--- config/rs6000/t-darwin 21 Nov 2003 04:53:12 -0000 1.10
+++ config/rs6000/t-darwin 6 Jan 2004 22:57:48 -0000
@@ -1,3 +1,7 @@
-# Library code must include trampoline support.
-LIB2FUNCS_EXTRA = $(srcdir)/config/rs6000/darwin-tramp.asm
+# Add trampoline and long double support to libgcc.
+LIB2FUNCS_EXTRA = $(srcdir)/config/rs6000/darwin-tramp.asm \
+ $(srcdir)/config/rs6000/darwin-ldouble.c
+# For libgcc, we always want 128-bit long double, since a libgcc built with
+# that will work without it.
+TARGET_LIBGCC2_CFLAGS = -mlong-double-128
Index: config/rs6000/darwin-ldouble.c
===================================================================
RCS file: config/rs6000/darwin-ldouble.c
diff -N config/rs6000/darwin-ldouble.c
--- /dev/null 1 Jan 1970 00:00:00 -0000
+++ config/rs6000/darwin-ldouble.c 6 Jan 2004 22:57:48 -0000
@@ -0,0 +1,221 @@
+/* 128-bit long double support routines for Darwin.
+ Copyright (C) 1993, 2003 Free Software Foundation, Inc.
+
+This file is part of GCC.
+
+GCC is free software; you can redistribute it and/or modify it under
+the terms of the GNU General Public License as published by the Free
+Software Foundation; either version 2, or (at your option) any later
+version.
+
+In addition to the permissions in the GNU General Public License, the
+Free Software Foundation gives you unlimited permission to link the
+compiled version of this file into combinations with other programs,
+and to distribute those combinations without any restriction coming
+from the use of this file. (The General Public License restrictions
+do apply in other respects; for example, they cover modification of
+the file, and distribution when not linked into a combine
+executable.)
+
+GCC is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or
+FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+for more details.
+
+You should have received a copy of the GNU General Public License
+along with GCC; see the file COPYING. If not, write to the Free
+Software Foundation, 59 Temple Place - Suite 330, Boston, MA
+02111-1307, USA. */
+
+/* Implementations of floating-point long double basic arithmetic
+ functions called by the IBM C compiler when generating code for
+ PowerPC platforms. In particular, the following functions are
+ implemented: _xlqadd, _xlqsub, _xlqmul, and _xlqdiv. Double-double
+ algorithms are based on the paper "Doubled-Precision IEEE Standard
+ 754 Floating-Point Arithmetic" by W. Kahan, February 26, 1987. An
+ alternative published reference is "Software for Doubled-Precision
+ Floating-Point Computations", by Seppo Linnainmaa, ACM TOMS vol 7
+ no 3, September 1961, pages 272-283. */
+
+#define fabs(x) __builtin_fabs(x)
+
+#define unlikely(x) __builtin_expect ((x), 0)
+
+/* All these routines actually take two long doubles as parameters,
+ but GCC currently generates poor code when a union is used to turn
+ a long double into a pair of doubles. */
+
+extern long double _xlqadd (double, double, double, double);
+extern long double _xlqsub (double, double, double, double);
+extern long double _xlqmul (double, double, double, double);
+extern long double _xlqdiv (double, double, double, double);
+
+typedef union
+{
+ long double ldval;
+ double dval[2];
+} longDblUnion;
+
+static const double FPKINF = 1.0/0.0;
+
+/* Add two 'long double' values and return the result. */
+long double
+_xlqadd (double a, double b, double c, double d)
+{
+ longDblUnion z;
+ double t, tau, u, FPR_zero, FPR_PosInf;
+
+ FPR_zero = 0.0;
+ FPR_PosInf = FPKINF;
+
+ if (unlikely (a != a) || unlikely (c != c)) {
+ z.dval[0] = a + c; /* NaN result. */
+ z.dval[1] = a + c; /* NaN result. */
+ return z.ldval;
+ }
+
+ /* Ordered operands are arranged in order of their magnitudes. */
+
+ /* Switch inputs if |(c,d)| > |(a,b)|. */
+ if (fabs (c) > fabs (a))
+ {
+ t = a;
+ tau = b;
+ a = c;
+ b = d;
+ c = t;
+ d = tau;
+ }
+
+ /* b <- second largest magnitude double. */
+ if (fabs (c) > fabs (b))
+ {
+ t = b;
+ b = c;
+ c = t;
+ }
+
+ /* Thanks to commutivity, sum is invariant w.r.t. the next
+ conditional exchange. */
+ tau = d + c;
+
+ /* Order the smallest magnitude doubles. */
+ if (fabs (d) > fabs (c))
+ {
+ t = c;
+ c = d;
+ d = t;
+ }
+
+ t = (tau + b) + a; /* Sum values in ascending magnitude order. */
+
+ /* Infinite or zero result. */
+ if (unlikely (fabs (t) == FPR_PosInf) || unlikely (t == FPR_zero))
+ {
+ z.dval[0] = t;
+ z.dval[1] = t >= 0.0 ? (fabs (t) >= 0.0 ? t : 0.0) : -0.0;
+ return z.ldval;
+ }
+
+ /* Usual case. */
+ tau = (((a-t) + b) + c) + d;
+ u = t + tau;
+ z.dval[0] = u; /* Final fixup for long double result. */
+ z.dval[1] = (u - t) + tau;
+ return z.ldval;
+}
+
+long double
+_xlqsub (double a, double b, double c, double d)
+{
+ return _xlqadd (a, b, -c, -d);
+}
+
+long double
+_xlqmul (double a, double b, double c, double d)
+{
+ longDblUnion z;
+ double t, tau, u, v, w, FPR_zero, FPR_PosInf;
+
+ FPR_zero = 0.0;
+ FPR_PosInf = FPKINF;
+
+ t = a * c; /* Highest order double term. */
+
+ if (unlikely (t != t) || unlikely (t == FPR_zero))
+ {
+ /* NaN or zero result. */
+ z.dval[0] = t;
+ z.dval[1] = t;
+ return z.ldval;
+ }
+
+ if (unlikely (fabs(t) == FPR_PosInf))
+ {
+ /* Infinite result. */
+ z.dval[0] = t;
+ z.dval[1] = t >= 0 ? 0.0 : -0.0;
+ return z.ldval;
+ }
+
+ /* Finite nonzero result requires summing of terms of two highest
+ orders. */
+
+ /* Use fused multiply-add to get low part of a * c. */
+ asm ("fmsub %0,%1,%2,%3" : "=f"(tau) : "f"(a), "f"(c), "f"(t));
+ v = a*d;
+ w = b*c;
+ tau += v + w; /* Add in other second-order terms. */
+ u = t + tau;
+
+ /* Construct long double result. */
+ z.dval[0] = u;
+ z.dval[1] = (u - t) + tau;
+ return z.ldval;
+}
+
+long double
+_xlqdiv (double a, double b, double c, double d)
+{
+ longDblUnion z;
+ double s, sigma, t, tau, u, v, w, FPR_zero, FPR_PosInf;
+
+ FPR_zero = 0.0;
+ FPR_PosInf = FPKINF;
+
+ t = a / c; /* highest order double term */
+
+ if (unlikely (t != t) || unlikely (t == FPR_zero))
+ {
+ /* NaN or zero result. */
+ z.dval[0] = t;
+ z.dval[1] = t;
+ return z.ldval;
+ }
+
+ if (unlikely (fabs (t) == FPR_PosInf))
+ {
+ /* Infinite result. */
+ z.dval[0] = t;
+ z.dval[1] = t >= 0.0 ? 0.0 : -0.0;
+ return z.ldval;
+ }
+
+ /* Finite nonzero result requires corrections to the highest order term. */
+
+ s = c * t; /* (s,sigma) = c*t exactly. */
+ w = -(-b + d * t); /* Written to get fnmsub for speed, but not
+ numerically necessary. */
+
+ /* Use fused multiply-add to get low part of c * t. */
+ asm ("fmsub %0,%1,%2,%3" : "=f"(sigma) : "f"(c), "f"(t), "f"(s));
+ v = a - s;
+
+ tau = ((v-sigma)+w)/c; /* Correction to t. */
+ u = t + tau;
+
+ /* Construct long double result. */
+ z.dval[0] = u;
+ z.dval[1] = (t - u) + tau;
+ return z.ldval;
+}
============================================================
More information about the Gcc-patches
mailing list