Update. cvs/libc-ud-970904

1997-09-05 03:15  Ulrich Drepper  <drepper@cygnus.com>

	* Makefile (subdirs): Reorganize order so that nss follows add-ons.
	* config.make.in (LDAP): Add.
	* configure.in: Prepare for ldap add-on.
	* shlib-versions: Add version numbers for ldap NSS module.
	* nss/Makefile: Add rules to build ldap NSS module.

	* db/btree/bt_page.c (__bt_free): Update meta data information.

	Update db from db-2.3.6.
	* db2/db.h: Updated.
	* db2/btree/bt_close.c: Updated.
	* db2/btree/bt_compare.c: Updated.
	* db2/btree/bt_conv.c: Updated.
	* db2/btree/bt_cursor.c: Updated.
	* db2/btree/bt_delete.c: Updated.
	* db2/btree/bt_put.c: Updated.
	* db2/btree/bt_rec.c: Updated.
	* db2/btree/bt_recno.c: Updated.
	* db2/btree/bt_search.c: Updated.
	* db2/btree/bt_split.c: Updated.
	* db2/btree/bt_stat.c: Updated.
	* db2/common/db_appinit.c: Updated.
	* db2/common/db_apprec.c: Updated.
	* db2/common/db_byteorder.c: Updated.
	* db2/common/db_err.c: Updated.
	* db2/common/db_region.c: Updated.
	* db2/db/db.c: Updated.
	* db2/db/db_conv.c: Updated.
	* db2/db/db_dup.c: Updated.
	* db2/db/db_pr.c: Updated.
	* db2/db/db_ret.c: Updated.
	* db2/db185/db185.c: Updated.
	* db2/dbm/dbm.c: Updated.
	* db2/hash/hash_dup.c: Updated.
	* db2/include/db_am.h: Updated.
	* db2/include/db_page.h: Updated.
	* db2/include/mp.h: Updated.
	* db2/include/queue.h: Updated.
	* db2/log/log.c: Updated.
	* db2/log/log_findckp.c: Updated.
	* db2/log/log_get.c: Updated.
	* db2/log/log_rec.c: Updated.
	* db2/mp/mp_bh.c: Updated.
	* db2/mp/mp_fopen.c: Updated.
	* db2/mp/mp_pr.c: Updated.
	* db2/mp/mp_sync.c: Updated.
	* db2/mutex/x86.gcc: Updated.
	* db2/os/db_os_dir.c: Updated.
	* db2/os/db_os_fid.c: Updated.
	* db2/progs/db_archive/db_archive.c: Updated.
	* db2/progs/db_checkpoint/db_checkpoint.c: Updated.
	* db2/progs/db_deadlock/db_deadlock.c: Updated.
	* db2/progs/db_dump/db_dump.c: Updated.
	* db2/progs/db_dump185/db_dump185.c: Updated.
	* db2/progs/db_load/db_load.c: Updated.
	* db2/progs/db_printlog/db_printlog.c: Updated.
	* db2/progs/db_recover/db_recover.c: Updated.
	* db2/progs/db_stat/db_stat.c: Updated.
	* db2/txn/txn.c: Updated.
	* db2/txn/txn_rec.c: Updated.

	* db2/Makefile: Add rule to build db_printlog.
	* db2/config.h: Don't include endian.h and don't define
	WORDS_BIGENDIAN.  Only define HAVE_ENDIAN_H.

	* elf/dl-open.c (_dl_open): Make thread-safe.
	* elf/dl-close.c (_dl_close): Likewise.
	* elf/dlclose.c: Pretty print.
	* elf/dl-version.c (make_string): Use __strdup not strdup.
	(find_needed): Don't use l_searchlist of _dl_loaded, use the
	l_next list.
	* elf/dl-deps.c (_dl_map_object_deps): Remove variable head.

	* manual/math.texi: Add note about == and != for FP values.
	* manual/message.texi: Mention problems with relative paths and
	binstextdomain.
	* manual/string.texi: Mark strdupa and strndupa correctly as macros.

	* sunrpc/Makefile (CFLAGS-xkey_prot.c): Define to prevent warnings.
	* sunrpc/rpc_cout.c: Avoid needless races around return value.

	* sysdeps/generic/stpcpy.c: Include <config.h> if wanted.

	Update from tzdata-1997h.
	* time/africa: Updated.
	* time/europe: Updated.
	* time/iso3166.tab: Updated.
	* time/northamerica: Updated.
	* time/southamerica: Updated.
	* time/zone.tab: Updated.

1997-09-04 13:19  Richard Henderson  <rth@cygnus.com>

	* sysdeps/alpha/w_sqrt.S: Removed.
	* sysdeps/alpha/fpu/e_sqrt.c: New.  Obey -mieee and -mieee-with-inexact
	and build a version that is as fast as possible given the constraint.

1997-08-30  Andreas Jaeger  <aj@arthur.rhein-neckar.de>

	* manual/maint.texi: Document some defaults for configuration
	parameters.

1997-09-04 15:57  Ulrich Drepper  <drepper@cygnus.com>

	* termios/cfsetspeed.c (cfsetspeed): Change return value to int.
	* termios/termios.h: Change prototype accordingly.
	Reported by Andreas Jaeger <aj@arthur.rhein-neckar.de>.

1997-07-26 14:42  H.J. Lu  <hjl@gnu.ai.mit.edu>

	* libio/stdfiles.c (_libio_using_thunks): Define if
	_G_USING_THUNKS is defined.

	* libio/_G_config.h: Moved to ...
	* sysdeps/generic/_G_config.h: ...here.

	* sysdeps/unix/sysv/linux/_G_config.h: New,
	Add vtable thunks support.

1997-09-02  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/m68k/Makefile (pic-ccflag): Override it for more
	efficient code (together with a change in gcc).

	* Makeconfig (pic-ccflag): New variable.
	(CFLAGS-.os): Use it instead of hardcoding -fPIC.

1997-09-02  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/m68k/dl-machine.h (elf_machine_load_address): Use
	_dl_start instead of a local label, to avoid getting an unreadable
	label name.

1997-09-02  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* nss/Makefile ($(objpfx)libnss_db.so): Use $(dbobjdir) to find
	libdb.so.

	* Makeconfig (dbobjdir): Use db2.

1997-08-31  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* manual/maint.texi (Porting to Unix): Update description of the
	implementation of Unix system calls.
	(Contributors): Use real umlaut in tege's name.

1997-08-30  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/libm-ieee754/w_gamma.c: Make compilable with
	_IEEE_LIBM.
	* sysdeps/libm-ieee754/w_gammaf.c: Likewise.
	* sysdeps/libm-ieee754/w_gammal.c: Likewise.

10 files changed, 403 insertions, 164 deletions

diff --git a/sysdeps/alpha/fpu/e_sqrt.c b/sysdeps/alpha/fpu/e_sqrt.c
new file mode 100644
index 0000000000..76fa015622
--- /dev/null
+++ b/sysdeps/alpha/fpu/e_sqrt.c
@@ -0,0 +1,256 @@
+/* Copyright (C) 1996, 1997 Free Software Foundation, Inc.
+   Contributed by David Mosberger (davidm@cs.arizona.edu).
+
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Library General Public License as
+   published by the Free Software Foundation; either version 2 of the
+   License, or (at your option) any later version.
+
+   The GNU C Library 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
+   Library General Public License for more details.
+
+   You should have received a copy of the GNU Library General Public
+   License along with the GNU C Library; see the file COPYING.LIB.  If not,
+   write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+   Boston, MA 02111-1307, USA.  */
+
+/*
+ * We have three versions, depending on how exact we need the results.
+ */
+
+#if defined(_IEEE_FP) && defined(_IEEE_FP_INEXACT)
+
+/* Most demanding: go to the original source.  */
+#include <libm-ieee754/e_sqrt.c>
+
+#else
+
+/* Careful with rearranging this without consulting the assembly below.  */
+const static struct sqrt_data_struct {
+	unsigned long dn, up, half, almost_three_half;
+	unsigned long one_and_a_half, two_to_minus_30, one, nan;
+	const int T2[64];
+} sqrt_data = {
+	0x3fefffffffffffff,	/* __dn = nextafter(1,-Inf) */
+	0x3ff0000000000001,	/* __up = nextafter(1,+Inf) */
+	0x3fe0000000000000,	/* half */
+	0x3ff7ffffffc00000,	/* almost_three_half = 1.5-2^-30 */
+	0x3ff8000000000000,	/* one_and_a_half */
+	0x3e10000000000000,	/* two_to_minus_30 */
+	0x3ff0000000000000,	/* one */
+	0xffffffffffffffff,	/* nan */
+
+	{ 0x1500, 0x2ef8, 0x4d67, 0x6b02, 0x87be, 0xa395, 0xbe7a, 0xd866,
+	0xf14a, 0x1091b,0x11fcd,0x13552,0x14999,0x15c98,0x16e34,0x17e5f,
+	0x18d03,0x19a01,0x1a545,0x1ae8a,0x1b5c4,0x1bb01,0x1bfde,0x1c28d,
+	0x1c2de,0x1c0db,0x1ba73,0x1b11c,0x1a4b5,0x1953d,0x18266,0x16be0,
+	0x1683e,0x179d8,0x18a4d,0x19992,0x1a789,0x1b445,0x1bf61,0x1c989,
+	0x1d16d,0x1d77b,0x1dddf,0x1e2ad,0x1e5bf,0x1e6e8,0x1e654,0x1e3cd,
+	0x1df2a,0x1d635,0x1cb16,0x1be2c,0x1ae4e,0x19bde,0x1868e,0x16e2e,
+	0x1527f,0x1334a,0x11051,0xe951, 0xbe01, 0x8e0d, 0x5924, 0x1edd }
+};
+
+#ifdef _IEEE_FP
+/*
+ * This version is much faster than the standard one included above,
+ * but it doesn't maintain the inexact flag.
+ */
+
+#define lobits(x) (((unsigned int *)&x)[0])
+#define hibits(x) (((unsigned int *)&x)[1])
+
+static inline double initial_guess(double x, unsigned int k,
+	const struct sqrt_data_struct * const ptr)
+{
+	double ret = 0.0;
+
+	k = 0x5fe80000 - (k >> 1);
+	k = k - ptr->T2[63&(k>>14)];
+	hibits(ret) = k;
+	return ret;
+}
+
+/* up = nextafter(1,+Inf), dn = nextafter(1,-Inf) */
+
+#define __half			(ptr->half)
+#define __one_and_a_half	(ptr->one_and_a_half)
+#define __two_to_minus_30	(ptr->two_to_minus_30)
+#define __one			(ptr->one)
+#define __up			(ptr->up)
+#define __dn			(ptr->dn)
+#define __Nan			(ptr->nan)
+
+#define Double(x) (*(double *)&x)
+
+/* Multiply with chopping rounding.. */
+#define choppedmul(a,b,c) \
+  __asm__("multc %1,%2,%0":"=&f" (c):"f" (a), "f" (b))
+
+double
+__ieee754_sqrt(double x)
+{
+  const struct sqrt_data_struct * const ptr = &sqrt_data;
+  unsigned long k, bits;
+  double y, z, zp, zn;
+  double dn, up, low, high;
+  double half, one_and_a_half, one, two_to_minus_30;
+
+  *(double *)&bits = x;
+  k = bits;
+
+  /* Negative or NaN or Inf */
+  if ((k >> 52) >= 0x7ff)
+    goto special;
+  y = initial_guess(x, k >> 32, ptr);
+  half = Double(__half);
+  one_and_a_half = Double(__one_and_a_half);
+  y = y*(one_and_a_half - half*x*y*y);
+  dn = Double(__dn);
+  two_to_minus_30 = Double(__two_to_minus_30);
+  y = y*((one_and_a_half - two_to_minus_30) - half*x*y*y);
+  up = Double(__up);
+  z = x*y;
+  one = Double(__one);
+  z = z + half*z*(one-z*y);
+
+  choppedmul(z,dn,zp);
+  choppedmul(z,up,zn);
+
+  choppedmul(z,zp,low);
+  low = low - x;
+  choppedmul(z,zn,high);
+  high = high - x;
+
+  /* I can't get gcc to use fcmov's.. */
+  __asm__("fcmovge %2,%3,%0"
+	  :"=f" (z)
+	  :"0" (z), "f" (low), "f" (zp));
+  __asm__("fcmovlt %2,%3,%0"
+	  :"=f" (z)
+	  :"0" (z), "f" (high), "f" (zn));
+  return z;	/* Argh! gcc jumps to end here */
+
+special:
+  /* throw away sign bit */
+  k <<= 1;
+  /* -0 */
+  if (!k)
+    return x;
+  /* special? */
+  if ((k >> 53) == 0x7ff) {
+    /* NaN? */
+    if (k << 11)
+      return x;
+    /* sqrt(+Inf) = +Inf */
+    if (x > 0)
+      return x;
+  }
+
+  x = Double(__Nan);
+  return x;
+}
+
+#else
+/*
+ * This version is much faster than generic sqrt implementation, but
+ * it doesn't handle exceptional values or the inexact flag.
+ */
+
+asm ("\
+  /* Define offsets into the structure defined in C above.  */
+	$DN = 0*8
+	$UP = 1*8
+	$HALF = 2*8
+	$ALMOST_THREE_HALF = 3*8
+	$NAN = 7*8
+	$T2 = 8*8
+
+  /* Stack variables.  */
+	$K = 0
+	$Y = 8
+
+	.text
+	.align	3
+	.globl	__ieee754_sqrt
+	.ent	__ieee754_sqrt
+__ieee754_sqrt:
+	ldgp	$29, 0($27)
+	subq	$sp, 16, $sp
+	.frame	$sp, 16, $26, 0\n"
+#ifdef PROF
+"	lda	$28, _mcount
+	jsr	$28, ($28), _mcount\n"
+#endif
+"	.prologue 1
+
+	stt	$f16, $K($sp)
+	lda	$4, sqrt_data			# load base address into t3
+	fblt	$f16, $negative
+
+  /* Compute initial guess.  */
+
+	.align 3
+
+	ldah	$2, 0x5fe8			# e0    :
+	ldq	$3, $K($sp)			# .. e1 :
+	ldt	$f12, $HALF($4)			# e0    :
+	ldt	$f18, $ALMOST_THREE_HALF($4)	# .. e1 :
+	srl	$3, 33, $1			# e0    :
+	mult	$f16, $f12, $f11		# .. fm : $f11 = x * 0.5
+	subl	$2, $1, $2			# e0    :
+	addt	$f12, $f12, $f17		# .. fa : $f17 = 1.0
+	srl	$2, 12, $1			# e0    :
+	and	$1, 0xfc, $1			# .. e1 :
+	addq	$1, $4, $1			# e0    :
+	ldl	$1, $T2($1)			# .. e1 :
+	addt	$f12, $f17, $f15		# fa    : $f15 = 1.5
+	subl	$2, $1, $2			# .. e1 :
+	sll	$2, 32, $2			# e0    :
+	ldt	$f14, $DN($4)			# .. e1 :
+	stq	$2, $Y($sp)			# e0    :
+	ldt	$f13, $Y($sp)			# e1    :
+
+	mult	$f11, $f13, $f10	# fm    : $f10 = (x * 0.5) * y
+	mult	$f10, $f13, $f10	# fm    : $f10 = ((x * 0.5) * y) * y
+	subt	$f15, $f10, $f1		# fa    : $f1 = (1.5 - 0.5*x*y*y)
+	mult	$f13, $f1, $f13         # fm    : yp = y*(1.5 - 0.5*x*y*y)
+ 	mult	$f11, $f13, $f11	# fm    : $f11 = x * 0.5 * yp
+	mult	$f11, $f13, $f11	# fm    : $f11 = (x * 0.5 * yp) * yp
+	subt	$f18, $f11, $f1		# fa    : $f1= (1.5-2^-30) - 0.5*x*yp*yp
+	mult	$f13, $f1, $f13		# fm    : ypp = $f13 = yp*$f1
+	subt	$f15, $f12, $f1		# fa    : $f1 = (1.5 - 0.5)
+	ldt	$f15, $UP($4)		# .. e1 :
+	mult	$f16, $f13, $f10	# fm    : z = $f10 = x * ypp
+	mult	$f10, $f13, $f11	# fm    : $f11 = z*ypp
+	mult	$f10, $f12, $f12	# fm    : $f12 = z*0.5
+	subt	$f1, $f11, $f1		# .. fa : $f1 = 1 - z*ypp
+	mult	$f12, $f1, $f12		# fm    : $f12 = z*0.5*(1 - z*ypp)
+	addt	$f10, $f12, $f0		# fa    : zp=res=$f0= z + z*0.5*(1 - z*ypp)
+
+	mult/c	$f0, $f14, $f12		# fm    : zmi = zp * DN
+	mult/c	$f0, $f15, $f11		# fm    : zpl = zp * UP
+	mult/c	$f0, $f12, $f1		# fm    : $f1 = zp * zmi
+	mult/c	$f0, $f11, $f15		# fm    : $f15 = zp * zpl
+
+	subt    $f1, $f16, $f13		# fa    : y1 = zp*zmi - x
+	subt    $f15, $f16, $f15	# fa    : y2 = zp*zpl - x
+
+	fcmovge	$f13, $f12, $f0		# res = (y1 >= 0) ? zmi : res
+	fcmovlt	$f15, $f11, $f0		# res = (y2 <  0) ? zpl : res
+
+	addq	$sp, 16, $sp		# e0    :
+	ret				# .. e1 :
+
+$negative:
+	ldt	$f0, $NAN($4)
+	addq	$sp, 16, $sp
+	ret
+
+	.end	__ieee754_sqrt");
+
+#endif /* _IEEE_FP */
+#endif /* _IEEE_FP && _IEEE_FP_INEXACT */
diff --git a/sysdeps/alpha/w_sqrt.S b/sysdeps/alpha/w_sqrt.S
deleted file mode 100644
index 32b06887d5..0000000000
--- a/sysdeps/alpha/w_sqrt.S
+++ /dev/null
@@ -1,161 +0,0 @@
-/* Copyright (C) 1996 Free Software Foundation, Inc.
-   This file is part of the GNU C Library.
-   Contributed by David Mosberger <davidm@cs.arizona.edu>, 1996.
-   Based on public-domain C source by Linus Torvalds.
-
-   The GNU C Library is free software; you can redistribute it and/or
-   modify it under the terms of the GNU Library General Public License as
-   published by the Free Software Foundation; either version 2 of the
-   License, or (at your option) any later version.
-
-   The GNU C Library 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
-   Library General Public License for more details.
-
-   You should have received a copy of the GNU Library General Public
-   License along with the GNU C Library; see the file COPYING.LIB.  If not,
-   write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
-   Boston, MA 02111-1307, USA.  */
-
-/* This version is much faster than generic sqrt implementation, but
-   it doesn't handle exceptional values or the inexact flag.  Don't use
-   this if _IEEE_FP or _IEEE_FP_INEXACT is in effect. */
-
-#ifndef _IEEE_FP
-
-#define _ERRNO_H
-#include <bits/errno.h>
-#include <sysdep.h>
-
-	.set noreorder
-
-#ifdef __ELF__
-	.section .rodata
-#else
-	.rdata
-#endif
-	.align 5        # align to cache line
-
-	/* Do all memory accesses relative to sqrtdata.  */
-sqrtdata:
-
-#define DN                     0x00
-#define UP                     0x08
-#define HALF                   0x10
-#define ALMOST_THREE_HALF      0x18
-#define T2                     0x20
-
-	.quad 0x3fefffffffffffff        /* DN = next(1.0) */
-	.quad 0x3ff0000000000001        /* UP = prev(1.0) */
-	.quad 0x3fe0000000000000        /* HALF = 0.5 */
-	.quad 0x3ff7ffffffc00000        /* ALMOST_THREE_HALF = 1.5-2^-30 */
-
-/* table T2: */
-.long   0x1500, 0x2ef8,   0x4d67,  0x6b02,  0x87be,  0xa395,  0xbe7a,  0xd866
-.long   0xf14a, 0x1091b, 0x11fcd, 0x13552, 0x14999, 0x15c98, 0x16e34, 0x17e5f
-.long  0x18d03, 0x19a01, 0x1a545, 0x1ae8a, 0x1b5c4, 0x1bb01, 0x1bfde, 0x1c28d
-.long  0x1c2de, 0x1c0db, 0x1ba73, 0x1b11c, 0x1a4b5, 0x1953d, 0x18266, 0x16be0
-.long  0x1683e, 0x179d8, 0x18a4d, 0x19992, 0x1a789, 0x1b445, 0x1bf61, 0x1c989
-.long  0x1d16d, 0x1d77b, 0x1dddf, 0x1e2ad, 0x1e5bf, 0x1e6e8, 0x1e654, 0x1e3cd
-.long  0x1df2a, 0x1d635, 0x1cb16, 0x1be2c, 0x1ae4e, 0x19bde, 0x1868e, 0x16e2e
-.long  0x1527f, 0x1334a, 0x11051,  0xe951,  0xbe01,  0x8e0d,  0x5924,  0x1edd
-
-/*
- * Stack variables:
- */
-#define K      16(sp)
-#define Y      24(sp)
-#define FSIZE  32
-
-	.text
-
-LEAF(__sqrt, FSIZE)
-	lda	sp, -FSIZE(sp)
-	ldgp	gp, .-__sqrt(pv)
-	stq	ra, 0(sp)
-#ifdef PROF
-	lda	AT, _mcount
-	jsr	AT, (AT), _mcount
-#endif
-	.prologue 1
-
-	stt	$f16, K
-	lda	t3, sqrtdata			# load base address into t3
-
-	fblt	$f16, $negative
-
-	/* Compute initial guess.  */
-
-	.align 3
-
-	ldah	t1, 0x5fe8			# e0    :
-	ldq	t2, K				# .. e1 :
-	ldt	$f12, HALF(t3)			# e0    :
-	ldt	$f18, ALMOST_THREE_HALF(t3)	# .. e1 :
-	srl	t2, 33, t0			# e0    :
-	mult	$f16, $f12, $f11		# .. fm : $f11 = x * 0.5
-	subl	t1, t0, t1			# e0    :
-	addt	$f12, $f12, $f17		# .. fa : $f17 = 1.0
-	srl	t1, 12, t0			# e0    :
-	and	t0, 0xfc, t0			# .. e1 :
-	addq	t0, t3, t0			# e0    :
-	ldl	t0, T2(t0)			# .. e1 :
-	addt	$f12, $f17, $f15		# fa    : $f15 = 1.5
-	subl	t1, t0, t1			# .. e1 :
-	sll	t1, 32, t1			# e0    :
-	ldt	$f14, DN(t3)			# .. e1 :
-	stq	t1, Y				# e0    :
-	ldt	$f13, Y				# e1    :
-	addq	sp, FSIZE, sp			# e0    :
-
-	mult	$f11, $f13, $f10	# fm    : $f10 = (x * 0.5) * y
-	mult	$f10, $f13, $f10	# fm    : $f10 = ((x * 0.5) * y) * y
-	subt	$f15, $f10, $f1		# fa    : $f1 = (1.5 - 0.5*x*y*y)
-	mult	$f13, $f1, $f13         # fm    : yp = y*(1.5 - 0.5*x*y*y)
- 	mult	$f11, $f13, $f11	# fm    : $f11 = x * 0.5 * yp
-	mult	$f11, $f13, $f11	# fm    : $f11 = (x * 0.5 * yp) * yp
-	subt	$f18, $f11, $f1		# fa    : $f1= (1.5-2^-30) - 0.5*x*yp*yp
-	mult	$f13, $f1, $f13		# fm    : ypp = $f13 = yp*$f1
-	subt	$f15, $f12, $f1		# fa    : $f1 = (1.5 - 0.5)
-	ldt	$f15, UP(t3)		# .. e1 :
-	mult	$f16, $f13, $f10	# fm    : z = $f10 = x * ypp
-	mult	$f10, $f13, $f11	# fm    : $f11 = z*ypp
-	mult	$f10, $f12, $f12	# fm    : $f12 = z*0.5
-	subt	$f1, $f11, $f1		# .. fa : $f1 = 1 - z*ypp
-	mult	$f12, $f1, $f12		# fm    : $f12 = z*0.5*(1 - z*ypp)
-	addt	$f10, $f12, $f0		# fa    : zp=res=$f0= z + z*0.5*(1 - z*ypp)
-
-	mult/c	$f0, $f14, $f12		# fm    : zmi = zp * DN
-	mult/c	$f0, $f15, $f11		# fm    : zpl = zp * UP
-	mult/c	$f0, $f12, $f1		# fm    : $f1 = zp * zmi
-	mult/c	$f0, $f11, $f15		# fm    : $f15 = zp * zpl
-
-	subt    $f1, $f16, $f13		# fa    : y1 = zp*zmi - x
-	subt    $f15, $f16, $f15	# fa    : y2 = zp*zpl - x
-
-	fcmovge	$f13, $f12, $f0		# res = (y1 >= 0) ? zmi : res
-	fcmovlt	$f15, $f11, $f0		# res = (y2 <  0) ? zpl : res
-
-	ret
-
-$negative:
-	lda	t1, -1
-	stq	t1, K
-	lda	t1, EDOM
-	stl	t1, errno
-#ifdef _LIBC_REENTRANT
-	jsr	ra, __errno_location
-	lda	t1, -1
-	ldq	ra, 0(sp)
-	stl	t1, 0(v0)
-#endif
-	ldt	$f0, K			# res = (double) 0xffffffffffffffff
-	addq	sp, FSIZE, sp
-	ret
-
-	END(__sqrt)
-
-weak_alias(__sqrt, sqrt)
-
-#endif /* !_IEEE_FP */
diff --git a/sysdeps/generic/_G_config.h b/sysdeps/generic/_G_config.h
new file mode 100644
index 0000000000..84a552b41e
--- /dev/null
+++ b/sysdeps/generic/_G_config.h
@@ -0,0 +1,65 @@
+/* This file is needed by libio to define various configuration parameters.
+   These are always the same in the GNU C library.  */
+
+#ifndef _G_config_h
+#define _G_config_h 1
+
+/* Define types for libio in terms of the standard internal type names.  */
+
+#include <bits/types.h>
+#define __need_size_t
+#define __need_wint_t
+#include <stddef.h>
+#ifndef _WINT_T
+/* Integral type unchanged by default argument promotions that can
+   hold any value corresponding to members of the extended character
+   set, as well as at least one value that does not correspond to any
+   member of the extended character set.  */
+#define _WINT_T
+typedef unsigned int wint_t;
+#endif
+#define _G_size_t	size_t
+#define _G_fpos_t	__off_t
+#define _G_ssize_t	__ssize_t
+#define _G_off_t	__off_t
+#define	_G_pid_t	__pid_t
+#define	_G_uid_t	__uid_t
+#define _G_wint_t	wint_t
+
+typedef int _G_int16_t __attribute__ ((__mode__ (__HI__)));
+typedef int _G_int32_t __attribute__ ((__mode__ (__SI__)));
+typedef unsigned int _G_uint16_t __attribute__ ((__mode__ (__HI__)));
+typedef unsigned int _G_uint32_t __attribute__ ((__mode__ (__SI__)));
+
+#define _G_HAVE_BOOL 1
+
+
+/* These library features are always available in the GNU C library.  */
+#define _G_HAVE_ATEXIT 1
+#define _G_HAVE_SYS_CDEFS 1
+#define _G_HAVE_SYS_WAIT 1
+#define _G_NEED_STDARG_H 1
+#define _G_va_list __gnuc_va_list
+
+#define _G_HAVE_PRINTF_FP 1
+#define _G_HAVE_MMAP 1
+#define _G_HAVE_LONG_DOUBLE_IO 1
+
+/* This is defined by <bits/stat.h> if `st_blksize' exists.  */
+#define _G_HAVE_ST_BLKSIZE defined (_STATBUF_ST_BLKSIZE)
+
+#define _G_BUFSIZ 8192
+
+/* These are the vtbl details for ELF.  */
+#define _G_NAMES_HAVE_UNDERSCORE 0
+#define _G_VTABLE_LABEL_PREFIX "_vt."
+#define _G_VTABLE_LABEL_HAS_LENGTH 1
+
+
+#if defined (__cplusplus) || defined (__STDC__)
+#define _G_ARGS(ARGLIST) ARGLIST
+#else
+#define _G_ARGS(ARGLIST) ()
+#endif
+
+#endif	/* _G_config.h */
diff --git a/sysdeps/generic/stpcpy.c b/sysdeps/generic/stpcpy.c
index 9800ace993..0bf84cc700 100644
--- a/sysdeps/generic/stpcpy.c
+++ b/sysdeps/generic/stpcpy.c
@@ -16,6 +16,10 @@
    write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
    Boston, MA 02111-1307, USA.  */
 
+#ifdef HAVE_CONFIG_H
+# include <config.h>
+#endif
+
 #include <string.h>
 
 #ifndef weak_alias
diff --git a/sysdeps/libm-ieee754/w_gamma.c b/sysdeps/libm-ieee754/w_gamma.c
index 5c3f27a8c1..02a3d47669 100644
--- a/sysdeps/libm-ieee754/w_gamma.c
+++ b/sysdeps/libm-ieee754/w_gamma.c
@@ -30,10 +30,12 @@ static char rcsid[] = "$NetBSD: w_gamma.c,v 1.7 1995/11/20 22:06:43 jtc Exp $";
 #endif
 {
         double y;
+#ifndef _IEEE_LIBM
 	if (_LIB_VERSION == _SVID_)
 	  y = __ieee754_lgamma_r(x,&signgam);
 	else
 	  {
+#endif
 	    int local_signgam;
 	    y = __ieee754_gamma_r(x,&local_signgam);
 	    if (local_signgam < 0) y = -y;
diff --git a/sysdeps/libm-ieee754/w_gammaf.c b/sysdeps/libm-ieee754/w_gammaf.c
index 5988e54a60..7d78a52908 100644
--- a/sysdeps/libm-ieee754/w_gammaf.c
+++ b/sysdeps/libm-ieee754/w_gammaf.c
@@ -28,10 +28,12 @@ static char rcsid[] = "$NetBSD: w_gammaf.c,v 1.4 1995/11/20 22:06:48 jtc Exp $";
 #endif
 {
         float y;
+#ifndef _IEEE_LIBM
 	if (_LIB_VERSION == _SVID_)
 	  y = __ieee754_lgammaf_r(x,&signgam);
 	else
 	  {
+#endif
 	    int local_signgam;
 	    y = __ieee754_gammaf_r(x,&local_signgam);
 	    if (local_signgam < 0) y = -y;
diff --git a/sysdeps/libm-ieee754/w_gammal.c b/sysdeps/libm-ieee754/w_gammal.c
index 0107d551b7..1b4ae2cc7b 100644
--- a/sysdeps/libm-ieee754/w_gammal.c
+++ b/sysdeps/libm-ieee754/w_gammal.c
@@ -33,10 +33,12 @@ static char rcsid[] = "$NetBSD: $";
 #endif
 {
         long double y;
+#ifndef _IEEE_LIBM
 	if (_LIB_VERSION == _SVID_)
 	  y = __ieee754_lgammal_r(x,&signgam);
 	else
 	  {
+#endif
 	    int local_signgam;
 	    y = __ieee754_gammal_r(x,&local_signgam);
 	    if (local_signgam < 0) y = -y;
diff --git a/sysdeps/m68k/Makefile b/sysdeps/m68k/Makefile
index 0f811bc62e..a056360ac6 100644
--- a/sysdeps/m68k/Makefile
+++ b/sysdeps/m68k/Makefile
@@ -1,4 +1,4 @@
-# Copyright (C) 1993, 1994, 1996 Free Software Foundation, Inc.
+# Copyright (C) 1993, 1994, 1996, 1997 Free Software Foundation, Inc.
 # This file is part of the GNU C Library.
 
 # The GNU C Library is free software; you can redistribute it and/or
@@ -25,6 +25,8 @@ endif
 
 asm-CPPFLAGS += $(m68k-syntax-flag)
 
+pic-ccflag = -fpic
+
 # Make sure setjmp.c is compiled with a frame pointer
 CFLAGS-setjmp.c := -fno-omit-frame-pointer
 
diff --git a/sysdeps/m68k/dl-machine.h b/sysdeps/m68k/dl-machine.h
index 75434cd750..44eefb20f5 100644
--- a/sysdeps/m68k/dl-machine.h
+++ b/sysdeps/m68k/dl-machine.h
@@ -56,8 +56,8 @@ static inline Elf32_Addr
 elf_machine_load_address (void)
 {
   Elf32_Addr addr;
-  asm ("1: lea 1b(%%pc), %0\n"
-       "   sub.l 1b@GOTPC(%%pc), %0"
+  asm ("lea _dl_start(%%pc), %0\n\t"
+       "sub.l _dl_start@GOTPC(%%pc), %0"
        : "=a" (addr));
   return addr;
 }
diff --git a/sysdeps/unix/sysv/linux/_G_config.h b/sysdeps/unix/sysv/linux/_G_config.h
new file mode 100644
index 0000000000..8dc9221519
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/_G_config.h
@@ -0,0 +1,67 @@
+/* This file is needed by libio to define various configuration parameters.
+   These are always the same in the GNU C library.  */
+
+#ifndef _G_config_h
+#define _G_config_h 1
+
+/* Define types for libio in terms of the standard internal type names.  */
+
+#include <bits/types.h>
+#define __need_size_t
+#define __need_wint_t
+#include <stddef.h>
+#ifndef _WINT_T
+/* Integral type unchanged by default argument promotions that can
+   hold any value corresponding to members of the extended character
+   set, as well as at least one value that does not correspond to any
+   member of the extended character set.  */
+#define _WINT_T
+typedef unsigned int wint_t;
+#endif
+#define _G_size_t	size_t
+#define _G_fpos_t	__off_t
+#define _G_ssize_t	__ssize_t
+#define _G_off_t	__off_t
+#define	_G_pid_t	__pid_t
+#define	_G_uid_t	__uid_t
+#define _G_wint_t	wint_t
+
+typedef int _G_int16_t __attribute__ ((__mode__ (__HI__)));
+typedef int _G_int32_t __attribute__ ((__mode__ (__SI__)));
+typedef unsigned int _G_uint16_t __attribute__ ((__mode__ (__HI__)));
+typedef unsigned int _G_uint32_t __attribute__ ((__mode__ (__SI__)));
+
+#define _G_HAVE_BOOL 1
+
+
+/* These library features are always available in the GNU C library.  */
+#define _G_HAVE_ATEXIT 1
+#define _G_HAVE_SYS_CDEFS 1
+#define _G_HAVE_SYS_WAIT 1
+#define _G_NEED_STDARG_H 1
+#define _G_va_list __gnuc_va_list
+
+#define _G_HAVE_PRINTF_FP 1
+#define _G_HAVE_MMAP 1
+#define _G_HAVE_LONG_DOUBLE_IO 1
+
+/* This is defined by <bits/stat.h> if `st_blksize' exists.  */
+#define _G_HAVE_ST_BLKSIZE defined (_STATBUF_ST_BLKSIZE)
+
+#define _G_BUFSIZ 8192
+
+/* These are the vtbl details for ELF.  */
+#define _G_NAMES_HAVE_UNDERSCORE 0
+#define _G_VTABLE_LABEL_HAS_LENGTH 1
+#define _G_USING_THUNKS
+#define _G_VTABLE_LABEL_PREFIX "__vt_"
+#define _G_VTABLE_LABEL_PREFIX_ID __vt_
+
+
+#if defined (__cplusplus) || defined (__STDC__)
+#define _G_ARGS(ARGLIST) ARGLIST
+#else
+#define _G_ARGS(ARGLIST) ()
+#endif
+
+#endif	/* _G_config.h */