Also create and use ldbl-compat-choose.h.

This patch makes the glibc build generate an additional header
ldbl-compat-choose.h that defines LONG_DOUBLE_COMPAT_CHOOSE_* macros
for each libc and libm symbol, which select one or the other of their
arguments based on whether the symbol was added before a change to
long double != double.

The effect of this is that it is then possible to define a macro
maybe_long_double_symbol that automatically acts as either
long_double_symbol or weak_alias depending on when the symbol being
defined was added.  This can be used when building long double
functions from type-generic templates.  Thus, with this patch ldbl-opt
no longer needs special long double implementations of each new libm
function added using such a template, and the existing such
implementations are removed.

This is a step towards being able more generally to use common macros
to create all the aliases needed for a libm function, so reducing the
amount of special-case code needed in ldbl-opt and ldbl-64-128, and
facilitating subsequently adding *f32 / *f64 / *f128 / *f32x / *f64x
aliases to existing functions (where the set of aliases that a
function should have may depend on the architecture in various ways).

Tested with build-many-glibcs.py.  Except for on
powerpc64le-linux-gnu, installed stripped shared libraries are
unchanged by the patch.  powerpc64le-linux-gnu is the unique
configuration which used ldbl-opt from the start rather than adding a
new long double choice after originally only having had long double =
double.  The effect of the patch there is that various cases that
previously used long_double_symbol unconditionally now use weak_alias
instead, so .os files contain e.g. a symbol cabsl instead of
cabsl@@GLIBC_2.17.  The final dynamic symbols and versions in the
resulting shared libraries are unchanged (ABI tests pass), as is the
disassembly of the shared libraries, but the differences in the .os
files still result in different .gnu_hash contents in libm.so; the
differences are of no significance and logically using weak_alias is
what's most appropriate in those cases.

	* scripts/versions.awk: Generate ldbl-compat-choose.h.
	* sysdeps/ieee754/ldbl-opt/math-type-macros-ldouble.h: Include
	<ldbl-compat-choose.h>.
	(maybe_long_double_symbol): New macro.
	[!declare_mgen_alias] (declare_mgen_alias): Use
	maybe_long_double_symbol.
	* sysdeps/ieee754/ldbl-opt/s_canonicalizel.c: Remove.
	* sysdeps/ieee754/ldbl-opt/s_fmaxmagl.c: Likewise.
	* sysdeps/ieee754/ldbl-opt/s_fminmagl.c: Likewise.
	* sysdeps/ieee754/ldbl-opt/s_nextdownl.c: Likewise.
	* sysdeps/ieee754/ldbl-opt/w_llogbl.c: Likewise.
	* Makerules [$(build-shared) = yes && !avoid-generated]
	(before-compile): Add $(common-objpfx)ldbl-compat-choose.h.
	[$(build-shared) = yes && !avoid-generated]
	($(common-objpfx)ldbl-compat-choose.h): New target.

6 files changed, 8 insertions, 26 deletions

diff --git a/sysdeps/ieee754/ldbl-opt/math-type-macros-ldouble.h b/sysdeps/ieee754/ldbl-opt/math-type-macros-ldouble.h
index da926dabc9..20873ae6b5 100644
--- a/sysdeps/ieee754/ldbl-opt/math-type-macros-ldouble.h
+++ b/sysdeps/ieee754/ldbl-opt/math-type-macros-ldouble.h
@@ -19,12 +19,19 @@
 #ifndef _MATH_TYPE_MACROS_LDOUBLE
 
 #include <math_ldbl_opt.h>
+#include <ldbl-compat-choose.h>
+
+#define maybe_long_double_symbol(lib, from, to)				\
+  LONG_DOUBLE_COMPAT_CHOOSE_ ## lib ## _ ## to (long_double_symbol (lib, \
+								    from, \
+								    to), \
+						weak_alias (from, to))
 
 /* Use properly versioned symbols for long double on platforms where
    it was not always a distinct type.  */
 #if !defined declare_mgen_alias
 # define declare_mgen_alias(from, to) \
-  long_double_symbol (libm, from ## l, to ## l);
+  maybe_long_double_symbol (libm, from ## l, to ## l);
 #endif
 
 #include_next <math-type-macros-ldouble.h>
diff --git a/sysdeps/ieee754/ldbl-opt/s_canonicalizel.c b/sysdeps/ieee754/ldbl-opt/s_canonicalizel.c
deleted file mode 100644
index d5d67dfe0c..0000000000
--- a/sysdeps/ieee754/ldbl-opt/s_canonicalizel.c
+++ /dev/null
@@ -1,5 +0,0 @@
-/* canonicalizel is not subject to complex aliasing rules.  It was
-   added in glibc 2.25.  */
-#define declare_mgen_alias(from, to) weak_alias (M_SUF (from), M_SUF (to))
-#include <math-type-macros-ldouble.h>
-#include <s_canonicalize_template.c>
diff --git a/sysdeps/ieee754/ldbl-opt/s_fmaxmagl.c b/sysdeps/ieee754/ldbl-opt/s_fmaxmagl.c
deleted file mode 100644
index 7e23761b58..0000000000
--- a/sysdeps/ieee754/ldbl-opt/s_fmaxmagl.c
+++ /dev/null
@@ -1,5 +0,0 @@
-/* fmaxmagl is not subject to complex aliasing rules.  It was added in
-   glibc 2.25.  */
-#define declare_mgen_alias(from, to) weak_alias (M_SUF (from), M_SUF (to))
-#include <math-type-macros-ldouble.h>
-#include <s_fmaxmag_template.c>
diff --git a/sysdeps/ieee754/ldbl-opt/s_fminmagl.c b/sysdeps/ieee754/ldbl-opt/s_fminmagl.c
deleted file mode 100644
index 46c10514c9..0000000000
--- a/sysdeps/ieee754/ldbl-opt/s_fminmagl.c
+++ /dev/null
@@ -1,5 +0,0 @@
-/* fminmagl is not subject to complex aliasing rules.  It was added in
-   glibc 2.25.  */
-#define declare_mgen_alias(from, to) weak_alias (M_SUF (from), M_SUF (to))
-#include <math-type-macros-ldouble.h>
-#include <s_fminmag_template.c>
diff --git a/sysdeps/ieee754/ldbl-opt/s_nextdownl.c b/sysdeps/ieee754/ldbl-opt/s_nextdownl.c
deleted file mode 100644
index c92c02edf6..0000000000
--- a/sysdeps/ieee754/ldbl-opt/s_nextdownl.c
+++ /dev/null
@@ -1,5 +0,0 @@
-/* nextdownl is not subject to complex aliasing rules.  It was
-   added in glibc 2.24.  */
-#define declare_mgen_alias(from, to) weak_alias (M_SUF (from), M_SUF (to))
-#include <math-type-macros-ldouble.h>
-#include <s_nextdown_template.c>
diff --git a/sysdeps/ieee754/ldbl-opt/w_llogbl.c b/sysdeps/ieee754/ldbl-opt/w_llogbl.c
deleted file mode 100644
index 8abe54bc85..0000000000
--- a/sysdeps/ieee754/ldbl-opt/w_llogbl.c
+++ /dev/null
@@ -1,5 +0,0 @@
-/* llogbl is not subject to complex aliasing rules.  It was added in
-   glibc 2.25.  */
-#define declare_mgen_alias(from, to) weak_alias (M_SUF (from), M_SUF (to))
-#include <math-type-macros-ldouble.h>
-#include <w_llogb_template.c>