[PATCH v3 0/6] IEEE binary128 long double on powerpc64le

public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed

* [PATCH v3 0/6] IEEE binary128 long double on powerpc64le
@ 2020-04-06 19:12 Paul E. Murphy
  2020-04-06 19:12 ` [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731 Paul E. Murphy
                   ` (5 more replies)
  0 siblings, 6 replies; 22+ messages in thread
From: Paul E. Murphy @ 2020-04-06 19:12 UTC (permalink / raw)
  To: libc-alpha

New for V3:

Cleanup patch to enforce GCC 7.4.  Originally it was GCC 7.3.
The fix for __builtin_signbit(long double) arrived in 7.4.

Cleanup C++ workaround for some GCC 9 compilers, and confirmed
it is a GCC 9 compiler bug fixed between 9.2 and 9.3.

Tested with GCC {7.4, 8.3, 9.3} x {ppc64le, power8, power9}.

See <https://sourceware.org/pipermail/libc-alpha/2020-March/112248.html>
for start of v2 thread.

Gabriel F. T. Gomes (1):
  powerpc64le: Enable support for IEEE long double

Paul E. Murphy (5):
  ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731
  Rename __LONG_DOUBLE_USES_FLOAT128 to
    __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI
  powerpc64le: raise GCC requirement to 7.4 for long double transition
  powerpc64le: bump binutils version requirement to >= 2.26
  powerpc64le: blacklist broken GCC compilers (e.g GCC 7.5.0)

 INSTALL                                       |  10 ++
 NEWS                                          |   8 +-
 argp/argp.h                                   |   2 +-
 bits/long-double.h                            |  15 +-
 elf/tst-addr1.c                               |   2 +-
 include/math.h                                |   6 +-
 include/monetary.h                            |   8 +
 include/printf.h                              |   8 +
 include/stdio.h                               |  16 +-
 include/stdlib.h                              |  11 +-
 include/sys/cdefs.h                           |   2 +-
 include/wchar.h                               |   9 ++
 libio/bits/stdio-ldbl.h                       |   4 +-
 libio/stdio.h                                 |   8 +-
 manual/install.texi                           |   9 ++
 math/complex.h                                |   4 +-
 math/math.h                                   |  12 +-
 math/test-ldouble.h                           |   2 +-
 misc/err.h                                    |   2 +-
 misc/error.h                                  |   2 +-
 misc/sys/cdefs.h                              |   4 +-
 misc/sys/syslog.h                             |   2 +-
 stdio-common/printf.h                         |   2 +-
 stdio-common/tst-vfprintf-user-type.c         |   2 +-
 stdlib/bits/stdlib-ldbl.h                     |   2 +-
 stdlib/monetary.h                             |   2 +-
 stdlib/stdlib.h                               |   2 +-
 sysdeps/ieee754/ldbl-128/bits/long-double.h   |   2 +-
 sysdeps/ieee754/ldbl-128ibm-compat/Makefile   |  12 --
 .../ldbl-128ibm-compat/bits/long-double.h     |   7 +-
 .../ieee754/ldbl-128ibm-compat/math_ldbl.h    |   2 +-
 .../ieee754/ldbl-128ibm/bits/iscanonical.h    |   2 +-
 sysdeps/ieee754/ldbl-128ibm/ieee754.h         |   6 +-
 .../ldbl-128ibm/include/bits/iscanonical.h    |   2 +-
 sysdeps/ieee754/ldbl-96/bits/long-double.h    |   2 +-
 sysdeps/ieee754/ldbl-opt/bits/long-double.h   |   2 +-
 sysdeps/mips/ieee754/bits/long-double.h       |   2 +-
 sysdeps/powerpc/fpu/libm-test-ulps            |   4 +
 sysdeps/powerpc/powerpc64/le/Implies-before   |   1 +
 sysdeps/powerpc/powerpc64/le/Makefile         |   4 +
 sysdeps/powerpc/powerpc64/le/configure        | 151 ++++++++++++++++++
 sysdeps/powerpc/powerpc64/le/configure.ac     |  56 +++++++
 .../powerpc64/le/ldbl-128ibm-compat-abi.h     |   8 +
 .../linux/powerpc/powerpc64/le/libc.abilist   |  93 +++++++++++
 .../linux/powerpc/powerpc64/le/libm.abilist   | 117 ++++++++++++++
 .../linux/sparc/sparc32/bits/long-double.h    |   2 +-
 .../linux/sparc/sparc64/bits/long-double.h    |   2 +-
 wcsmbs/bits/wchar-ldbl.h                      |   4 +-
 wcsmbs/wchar.h                                |   8 +-
 49 files changed, 573 insertions(+), 72 deletions(-)
 create mode 100644 sysdeps/unix/sysv/linux/powerpc/powerpc64/le/ldbl-128ibm-compat-abi.h

-- 
2.21.1


^ permalink raw reply	[flat|nested] 22+ messages in thread

* [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731
  2020-04-06 19:12 [PATCH v3 0/6] IEEE binary128 long double on powerpc64le Paul E. Murphy
@ 2020-04-06 19:12 ` Paul E. Murphy
  2020-04-15 14:20   ` Paul E Murphy
  2020-04-22 15:11   ` Florian Weimer
  2020-04-06 19:12 ` [PATCH v3 2/6] Rename __LONG_DOUBLE_USES_FLOAT128 to __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI Paul E. Murphy
                   ` (4 subsequent siblings)
  5 siblings, 2 replies; 22+ messages in thread
From: Paul E. Murphy @ 2020-04-06 19:12 UTC (permalink / raw)
  To: libc-alpha

GCC 9 has a bug (BZ 90731) whereby __typeof does not correctly copy
exception specifiers[1].  Surprisingly, this can be quieted by declaring
"#pragma system_header", or if the headers are installed in a system
directory.

Work around this by using the pragma for any gcc version between
9.0 and 9.2 to ensure tests continue to compile.

[1] Example error from g++ 9.2.1:

In file included from ../include/sys/cdefs.h:3,
                 from ../include/features.h:465,
                 from ../bits/libc-header-start.h:33,
                 from ../math/math.h:27,
                 from ../include/math.h:7,
                 from test-math-isinff.cc:21:
../libio/bits/stdio-ldbl.h:25:20: error: declaration of ‘int sprintf(char*, const char*, ...)’ has a different exception specifier
   25 | __LDBL_REDIR_DECL (sprintf)
      |                    ^~~~~~~
../misc/sys/cdefs.h:461:26: note: in definition of macro ‘__LDBL_REDIR_DECL’
  461 |   extern __typeof (name) name __asm (__ASMNAME ("__" #name "ieee128"));
      |                          ^~~~
In file included from ../include/stdio.h:5,
                 from test-math-isinff.cc:22:
../libio/stdio.h:334:12: note: from previous declaration ‘int sprintf(char*, const char*, ...) throw ()’
  334 | extern int sprintf (char *__restrict __s,
      |            ^~~~~~~
---
 include/monetary.h |  8 ++++++++
 include/printf.h   |  8 ++++++++
 include/stdio.h    | 10 ++++++++++
 include/stdlib.h   |  9 +++++++++
 include/wchar.h    |  9 +++++++++
 5 files changed, 44 insertions(+)

diff --git a/include/monetary.h b/include/monetary.h
index 240925e87d..b6da7326c7 100644
--- a/include/monetary.h
+++ b/include/monetary.h
@@ -1,3 +1,11 @@
+/* Workaround BZ 90731 with GCC 9 when using ldbl redirects in C++. */
+#include <bits/floatn.h>
+#if defined __cplusplus && __LONG_DOUBLE_USES_FLOAT128 == 1
+# if __GNUC_PREREQ (9, 0) && !__GNUC_PREREQ (9, 3)
+#   pragma GCC system_header
+# endif
+#endif
+
 #include <stdlib/monetary.h>
 #ifndef _ISOMAC
 #include <stdarg.h>
diff --git a/include/printf.h b/include/printf.h
index d051514119..7430f9a833 100644
--- a/include/printf.h
+++ b/include/printf.h
@@ -1,5 +1,13 @@
 #ifndef	_PRINTF_H
 
+/* Workaround BZ 90731 with GCC 9 when using ldbl redirects in C++. */
+#include <bits/floatn.h>
+#if defined __cplusplus && __LONG_DOUBLE_USES_FLOAT128 == 1
+# if __GNUC_PREREQ (9, 0) && !__GNUC_PREREQ (9, 3)
+#   pragma GCC system_header
+# endif
+#endif
+
 #include <stdio-common/printf.h>
 
 # ifndef _ISOMAC
diff --git a/include/stdio.h b/include/stdio.h
index 6718af4108..8ce73c768b 100644
--- a/include/stdio.h
+++ b/include/stdio.h
@@ -2,8 +2,18 @@
 # if !defined _ISOMAC && defined _IO_MTSAFE_IO
 #  include <stdio-lock.h>
 # endif
+
+/* Workaround BZ 90731 with GCC 9 when using ldbl redirects in C++. */
+# include <bits/floatn.h>
+# if defined __cplusplus && __LONG_DOUBLE_USES_FLOAT128 == 1
+#  if __GNUC_PREREQ (9, 0) && !__GNUC_PREREQ (9, 3)
+#    pragma GCC system_header
+#  endif
+# endif
+
 # include <libio/stdio.h>
 # ifndef _ISOMAC
+
 #  define _LIBC_STDIO_H 1
 #  include <libio/libio.h>
 
diff --git a/include/stdlib.h b/include/stdlib.h
index 926f965f69..d1c4d41ba6 100644
--- a/include/stdlib.h
+++ b/include/stdlib.h
@@ -3,6 +3,15 @@
 #ifndef _ISOMAC
 # include <stddef.h>
 #endif
+
+/* Workaround BZ 90731 with GCC 9 when using ldbl redirects in C++. */
+#include <bits/floatn.h>
+#if defined __cplusplus && __LONG_DOUBLE_USES_FLOAT128 == 1
+# if __GNUC_PREREQ (9, 0) && !__GNUC_PREREQ (9, 3)
+#   pragma GCC system_header
+# endif
+#endif
+
 #include <stdlib/stdlib.h>
 
 /* Now define the internal interfaces.  */
diff --git a/include/wchar.h b/include/wchar.h
index 617906eb14..e70d275cad 100644
--- a/include/wchar.h
+++ b/include/wchar.h
@@ -1,4 +1,13 @@
 #ifndef _WCHAR_H
+
+/* Workaround BZ 90731 with GCC 9 when using ldbl redirects in C++. */
+# include <bits/floatn.h>
+# if defined __cplusplus && __LONG_DOUBLE_USES_FLOAT128 == 1
+#  if __GNUC_PREREQ (9, 0) && !__GNUC_PREREQ (9, 3)
+#   pragma GCC system_header
+#  endif
+# endif
+
 # include <wcsmbs/wchar.h>
 # ifndef _ISOMAC
 
-- 
2.21.1


^ permalink raw reply	[flat|nested] 22+ messages in thread

* [PATCH v3 2/6] Rename __LONG_DOUBLE_USES_FLOAT128 to __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI
  2020-04-06 19:12 [PATCH v3 0/6] IEEE binary128 long double on powerpc64le Paul E. Murphy
  2020-04-06 19:12 ` [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731 Paul E. Murphy
@ 2020-04-06 19:12 ` Paul E. Murphy
  2020-04-15 14:21   ` Paul E Murphy
  2020-04-06 19:12 ` [PATCH v3 3/6] powerpc64le: raise GCC requirement to 7.4 for long double transition Paul E. Murphy
                   ` (3 subsequent siblings)
  5 siblings, 1 reply; 22+ messages in thread
From: Paul E. Murphy @ 2020-04-06 19:12 UTC (permalink / raw)
  To: libc-alpha

per feedback and disucssion on libc-alpha [2].  Likewise, add
documentation to the macro in the common bits/long-double.h header.
This is effectively V2 of [2].

Per the feedback from Joseph [1].  Good comments and a more
self-explanatory macro name will be very helpful when this
macro is able to assume a non-zero value.

[1] <https://sourceware.org/legacy-ml/libc-alpha/2020-02/msg00687.html>
[2] <https://sourceware.org/pipermail/libc-alpha/2020-February/111148.html>

---8<---

Improve the commentary to aid future developers who will stumble
upon this novel, yet not always perfect, mechanism to support
alternative formats for long double.

Likewise, rename __LONG_DOUBLE_USES_FLOAT128 to
__LDOUBLE_REDIRECTS_TO_FLOAT128_ABI now that development work
has settled down.  The command used was

git grep -l __LONG_DOUBLE_USES_FLOAT128 ':!./ChangeLog*' | \
  xargs sed -i 's/__LONG_DOUBLE_USES_FLOAT128/__LDOUBLE_REDIRECTS_TO_FLOAT128_ABI/g'
---
 argp/argp.h                                       |  2 +-
 bits/long-double.h                                | 15 ++++++++++++++-
 elf/tst-addr1.c                                   |  2 +-
 include/math.h                                    |  6 +++---
 include/monetary.h                                |  2 +-
 include/printf.h                                  |  2 +-
 include/stdio.h                                   |  8 ++++----
 include/stdlib.h                                  |  4 ++--
 include/sys/cdefs.h                               |  2 +-
 include/wchar.h                                   |  2 +-
 libio/bits/stdio-ldbl.h                           |  4 ++--
 libio/stdio.h                                     |  8 ++++----
 math/complex.h                                    |  4 ++--
 math/math.h                                       | 12 ++++++------
 math/test-ldouble.h                               |  2 +-
 misc/err.h                                        |  2 +-
 misc/error.h                                      |  2 +-
 misc/sys/cdefs.h                                  |  4 ++--
 misc/sys/syslog.h                                 |  2 +-
 stdio-common/printf.h                             |  2 +-
 stdio-common/tst-vfprintf-user-type.c             |  2 +-
 stdlib/bits/stdlib-ldbl.h                         |  2 +-
 stdlib/monetary.h                                 |  2 +-
 stdlib/stdlib.h                                   |  2 +-
 sysdeps/ieee754/ldbl-128/bits/long-double.h       |  2 +-
 .../ieee754/ldbl-128ibm-compat/bits/long-double.h |  7 ++-----
 sysdeps/ieee754/ldbl-128ibm-compat/math_ldbl.h    |  2 +-
 sysdeps/ieee754/ldbl-128ibm/bits/iscanonical.h    |  2 +-
 sysdeps/ieee754/ldbl-128ibm/ieee754.h             |  6 +++---
 .../ldbl-128ibm/include/bits/iscanonical.h        |  2 +-
 sysdeps/ieee754/ldbl-96/bits/long-double.h        |  2 +-
 sysdeps/ieee754/ldbl-opt/bits/long-double.h       |  2 +-
 sysdeps/mips/ieee754/bits/long-double.h           |  2 +-
 .../sysv/linux/sparc/sparc32/bits/long-double.h   |  2 +-
 .../sysv/linux/sparc/sparc64/bits/long-double.h   |  2 +-
 wcsmbs/bits/wchar-ldbl.h                          |  4 ++--
 wcsmbs/wchar.h                                    |  8 ++++----
 37 files changed, 74 insertions(+), 64 deletions(-)

diff --git a/argp/argp.h b/argp/argp.h
index 358d9b43cd..a5b861a4d3 100644
--- a/argp/argp.h
+++ b/argp/argp.h
@@ -555,7 +555,7 @@ __NTH (__option_is_end (const struct argp_option *__opt))
 #endif /* Use extern inlines.  */
 
 #include <bits/floatn.h>
-#if defined __LDBL_COMPAT || __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __LDBL_COMPAT || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # include <bits/argp-ldbl.h>
 #endif
 
diff --git a/bits/long-double.h b/bits/long-double.h
index 6e16447e65..bdde78f49a 100644
--- a/bits/long-double.h
+++ b/bits/long-double.h
@@ -37,4 +37,17 @@
 #ifndef __NO_LONG_DOUBLE_MATH
 # define __NO_LONG_DOUBLE_MATH	1
 #endif
-#define __LONG_DOUBLE_USES_FLOAT128 0
+
+/* The macro __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI is used to determine the
+   choice of the underlying ABI of long double.  It will always assume
+   a constant value for each translation unit.
+
+   If the value is non-zero, any API which is parameterized by the long
+   double type (i.e the scanf/printf family of functions or the explicitly
+   parameterized math.h functions) will be redirected to a compatible
+   implementation using _Float128 ABI via symbols suffixed with ieee128.
+
+   The mechanism this macro uses to acquire may be a function
+   of architecture, or target specific options used to invoke the
+   compiler.  */
+#define __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI 0
diff --git a/elf/tst-addr1.c b/elf/tst-addr1.c
index 27dc0f00f4..93ce669cbd 100644
--- a/elf/tst-addr1.c
+++ b/elf/tst-addr1.c
@@ -15,7 +15,7 @@ do_test (void)
   if (i.dli_sname == NULL)
     return 1;
 
-#if __LONG_DOUBLE_USES_FLOAT128 == 1
+#if __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
   /* On architectures which redirect long double to
      _Float128 (e.g powerpc64le), printf will resolve
      to __printfieee128 due to header redirects.  There
diff --git a/include/math.h b/include/math.h
index 1b3199b3f6..3979c47400 100644
--- a/include/math.h
+++ b/include/math.h
@@ -19,7 +19,7 @@ hidden_proto (__isinff)
 hidden_proto (__isnanf)
 
 #  if !defined __NO_LONG_DOUBLE_MATH \
-      && __LONG_DOUBLE_USES_FLOAT128 == 0
+      && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 0
 hidden_proto (__finitel)
 hidden_proto (__isinfl)
 hidden_proto (__isnanl)
@@ -42,7 +42,7 @@ libm_hidden_proto (__expf)
 libm_hidden_proto (__roundeven)
 
 #  if !defined __NO_LONG_DOUBLE_MATH \
-      && __LONG_DOUBLE_USES_FLOAT128 == 0
+      && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 0
 libm_hidden_proto (__fpclassifyl)
 libm_hidden_proto (__issignalingl)
 libm_hidden_proto (__expl)
@@ -143,7 +143,7 @@ fabsf128 (_Float128 x)
   MATH_REDIRECT_LDBL (FUNC, PREFIX, ARGS)			\
   MATH_REDIRECT_F128 (FUNC, PREFIX, ARGS)
 #   if defined __NO_LONG_DOUBLE_MATH 				\
-       || __LONG_DOUBLE_USES_FLOAT128 == 1
+       || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #    define MATH_REDIRECT_LDBL(FUNC, PREFIX, ARGS)
 #   else
 #    define MATH_REDIRECT_LDBL(FUNC, PREFIX, ARGS)			\
diff --git a/include/monetary.h b/include/monetary.h
index b6da7326c7..06fd59fd73 100644
--- a/include/monetary.h
+++ b/include/monetary.h
@@ -1,6 +1,6 @@
 /* Workaround BZ 90731 with GCC 9 when using ldbl redirects in C++. */
 #include <bits/floatn.h>
-#if defined __cplusplus && __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __cplusplus && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # if __GNUC_PREREQ (9, 0) && !__GNUC_PREREQ (9, 3)
 #   pragma GCC system_header
 # endif
diff --git a/include/printf.h b/include/printf.h
index 7430f9a833..13b7f211fc 100644
--- a/include/printf.h
+++ b/include/printf.h
@@ -2,7 +2,7 @@
 
 /* Workaround BZ 90731 with GCC 9 when using ldbl redirects in C++. */
 #include <bits/floatn.h>
-#if defined __cplusplus && __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __cplusplus && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # if __GNUC_PREREQ (9, 0) && !__GNUC_PREREQ (9, 3)
 #   pragma GCC system_header
 # endif
diff --git a/include/stdio.h b/include/stdio.h
index 8ce73c768b..1534022f7b 100644
--- a/include/stdio.h
+++ b/include/stdio.h
@@ -5,7 +5,7 @@
 
 /* Workaround BZ 90731 with GCC 9 when using ldbl redirects in C++. */
 # include <bits/floatn.h>
-# if defined __cplusplus && __LONG_DOUBLE_USES_FLOAT128 == 1
+# if defined __cplusplus && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #  if __GNUC_PREREQ (9, 0) && !__GNUC_PREREQ (9, 3)
 #    pragma GCC system_header
 #  endif
@@ -22,10 +22,10 @@
 /*  Some libc_hidden_ldbl_proto's do not map to a unique symbol when
     redirecting ldouble to _Float128 variants.  We can therefore safely
     directly alias them to their internal name.  */
-# if __LONG_DOUBLE_USES_FLOAT128 == 1 && IS_IN (libc)
+# if __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1 && IS_IN (libc)
 #  define stdio_hidden_ldbl_proto(p, f) \
   extern __typeof (p ## f) p ## f __asm (__ASMNAME ("___ieee128_" #f));
-# elif __LONG_DOUBLE_USES_FLOAT128 == 1
+# elif __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #  define stdio_hidden_ldbl_proto(p,f) __LDBL_REDIR1_DECL (p ## f, p ## f ## ieee128)
 # else
 #  define stdio_hidden_ldbl_proto(p,f) libc_hidden_proto (p ## f)
@@ -101,7 +101,7 @@ libc_hidden_proto (__isoc99_vfscanf)
 #   define sscanf __isoc99_sscanf
 #  endif
 
-#  if __LONG_DOUBLE_USES_FLOAT128 == 1  && IS_IN (libc)
+#  if __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1  && IS_IN (libc)
 /* These are implemented as redirects to other public API.
    Therefore, the usual redirection fails to avoid PLT.  */
 extern __typeof (__isoc99_sscanf) ___ieee128_isoc99_sscanf __THROW;
diff --git a/include/stdlib.h b/include/stdlib.h
index d1c4d41ba6..54630de44b 100644
--- a/include/stdlib.h
+++ b/include/stdlib.h
@@ -6,7 +6,7 @@
 
 /* Workaround BZ 90731 with GCC 9 when using ldbl redirects in C++. */
 #include <bits/floatn.h>
-#if defined __cplusplus && __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __cplusplus && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # if __GNUC_PREREQ (9, 0) && !__GNUC_PREREQ (9, 3)
 #   pragma GCC system_header
 # endif
@@ -216,7 +216,7 @@ libc_hidden_proto (____strtoull_l_internal)
 #include <bits/floatn.h>
 libc_hidden_proto (strtof)
 libc_hidden_proto (strtod)
-#if __LONG_DOUBLE_USES_FLOAT128 == 0
+#if __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 0
 libc_hidden_proto (strtold)
 #endif
 libc_hidden_proto (strtol)
diff --git a/include/sys/cdefs.h b/include/sys/cdefs.h
index 9f44ac5a0a..6a76160ed4 100644
--- a/include/sys/cdefs.h
+++ b/include/sys/cdefs.h
@@ -16,7 +16,7 @@ rtld_hidden_proto (__chk_fail)
 /* If we are using redirects internally to support long double,
    we need to tweak some macros to ensure the PLT bypass tricks
    continue to work in libc. */
-#if __LONG_DOUBLE_USES_FLOAT128 == 1 && IS_IN (libc) && defined SHARED
+#if __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1 && IS_IN (libc) && defined SHARED
 
 # undef __LDBL_REDIR_DECL
 # define __LDBL_REDIR_DECL(func) \
diff --git a/include/wchar.h b/include/wchar.h
index e70d275cad..41812935fb 100644
--- a/include/wchar.h
+++ b/include/wchar.h
@@ -2,7 +2,7 @@
 
 /* Workaround BZ 90731 with GCC 9 when using ldbl redirects in C++. */
 # include <bits/floatn.h>
-# if defined __cplusplus && __LONG_DOUBLE_USES_FLOAT128 == 1
+# if defined __cplusplus && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #  if __GNUC_PREREQ (9, 0) && !__GNUC_PREREQ (9, 3)
 #   pragma GCC system_header
 #  endif
diff --git a/libio/bits/stdio-ldbl.h b/libio/bits/stdio-ldbl.h
index 3bb9f7dd2c..28c44666e3 100644
--- a/libio/bits/stdio-ldbl.h
+++ b/libio/bits/stdio-ldbl.h
@@ -31,7 +31,7 @@ __LDBL_REDIR_DECL (vsprintf)
 __LDBL_REDIR1_DECL (fscanf, __nldbl___isoc99_fscanf)
 __LDBL_REDIR1_DECL (scanf, __nldbl___isoc99_scanf)
 __LDBL_REDIR1_DECL (sscanf, __nldbl___isoc99_sscanf)
-# elif __LONG_DOUBLE_USES_FLOAT128 == 1
+# elif __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 __LDBL_REDIR1_DECL (fscanf, __isoc99_fscanfieee128)
 __LDBL_REDIR1_DECL (scanf, __isoc99_scanfieee128)
 __LDBL_REDIR1_DECL (sscanf, __isoc99_sscanfieee128)
@@ -55,7 +55,7 @@ __LDBL_REDIR_DECL (vsnprintf)
 __LDBL_REDIR1_DECL (vfscanf, __nldbl___isoc99_vfscanf)
 __LDBL_REDIR1_DECL (vscanf, __nldbl___isoc99_vscanf)
 __LDBL_REDIR1_DECL (vsscanf, __nldbl___isoc99_vsscanf)
-#  elif __LONG_DOUBLE_USES_FLOAT128 == 1
+#  elif __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 __LDBL_REDIR1_DECL (vfscanf, __isoc99_vfscanfieee128)
 __LDBL_REDIR1_DECL (vscanf, __isoc99_vscanfieee128)
 __LDBL_REDIR1_DECL (vsscanf, __isoc99_vsscanfieee128)
diff --git a/libio/stdio.h b/libio/stdio.h
index 18e29752c4..21ef36ae70 100644
--- a/libio/stdio.h
+++ b/libio/stdio.h
@@ -401,11 +401,11 @@ extern int sscanf (const char *__restrict __s,
 
 /* For historical reasons, the C99-compliant versions of the scanf
    functions are at alternative names.  When __LDBL_COMPAT or
-   __LONG_DOUBLE_USES_FLOAT128 are in effect, this is handled in
+   __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI are in effect, this is handled in
    bits/stdio-ldbl.h.  */
 #include <bits/floatn.h>
 #if !__GLIBC_USE (DEPRECATED_SCANF) && !defined __LDBL_COMPAT \
-    && __LONG_DOUBLE_USES_FLOAT128 == 0
+    && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 0
 # ifdef __REDIRECT
 extern int __REDIRECT (fscanf, (FILE *__restrict __stream,
 				const char *__restrict __format, ...),
@@ -451,7 +451,7 @@ extern int vsscanf (const char *__restrict __s,
 /* Same redirection as above for the v*scanf family.  */
 # if !__GLIBC_USE (DEPRECATED_SCANF)
 #  if defined __REDIRECT && !defined __LDBL_COMPAT \
-      && __LONG_DOUBLE_USES_FLOAT128 == 0
+      && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 0
 extern int __REDIRECT (vfscanf,
 		       (FILE *__restrict __s,
 			const char *__restrict __format, __gnuc_va_list __arg),
@@ -872,7 +872,7 @@ extern int __overflow (FILE *, int);
 #endif
 
 #include <bits/floatn.h>
-#if defined __LDBL_COMPAT || __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __LDBL_COMPAT || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # include <bits/stdio-ldbl.h>
 #endif
 
diff --git a/math/complex.h b/math/complex.h
index b6d175e800..7c06add10c 100644
--- a/math/complex.h
+++ b/math/complex.h
@@ -126,7 +126,7 @@ __BEGIN_DECLS
 #  undef __MATHDECL_1
 #  define __MATHDECL_1(type, function, args) \
   extern type __REDIRECT_NTH(__MATH_PRECNAME(function), args, function)
-# elif __LONG_DOUBLE_USES_FLOAT128 == 1
+# elif __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #  undef __MATHDECL_1
 #  undef __MATHDECL
 #  define __REDIR_TO(function) \
@@ -142,7 +142,7 @@ __BEGIN_DECLS
 # define __MATH_PRECNAME(name)	name##l
 # include <bits/cmathcalls.h>
 # if defined __LDBL_COMPAT \
-     || __LONG_DOUBLE_USES_FLOAT128 == 1
+     || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #  undef __REDIR_TO
 #  undef __MATHDECL_1
 #  undef __MATHDECL
diff --git a/math/math.h b/math/math.h
index 5c3dcb9194..baa2e5e11b 100644
--- a/math/math.h
+++ b/math/math.h
@@ -343,7 +343,7 @@ extern long double __REDIRECT_NTH (nexttowardl,
 #   define __MATHDECL_1(type, function,suffix, args) \
   __MATHREDIR(type, function, suffix, args, __CONCAT(function,suffix))
 
-#  elif __LONG_DOUBLE_USES_FLOAT128 == 1
+#  elif __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #   ifdef __REDIRECT_NTH
 #    ifdef __USE_ISOC99
 extern float __REDIRECT_NTH (nexttowardf, (float __x, long double __y),
@@ -390,7 +390,7 @@ extern double __REDIRECT_NTH (nexttoward, (double __x, long double __y),
 #  undef __MATH_DECLARING_FLOATN
 
 #  if defined __LDBL_COMPAT \
-      || __LONG_DOUBLE_USES_FLOAT128 == 1
+      || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #   undef __REDIRTO
 #   undef __REDIRTO_ALT
 #   undef __MATHDECL_1
@@ -562,7 +562,7 @@ extern double __REDIRECT_NTH (nexttoward, (double __x, long double __y),
 #  undef __MATHCALL_NARROW
 #  define __MATHCALL_NARROW(func, redir, nargs) \
   __MATHCALL_NARROW_REDIR (func, redir, nargs)
-# elif __LONG_DOUBLE_USES_FLOAT128 == 1
+# elif __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #  define __MATHCALL_REDIR_NAME(name) __ ## f32 ## name ## ieee128
 #  undef __MATHCALL_NARROW
 #  define __MATHCALL_NARROW(func, redir, nargs) \
@@ -573,7 +573,7 @@ extern double __REDIRECT_NTH (nexttoward, (double __x, long double __y),
 # undef _Marg_
 # undef __MATHCALL_NAME
 # if defined __LDBL_COMPAT \
-     || __LONG_DOUBLE_USES_FLOAT128 == 1
+     || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #  undef __MATHCALL_REDIR_NAME
 #  undef __MATHCALL_NARROW
 #  define __MATHCALL_NARROW(func, redir, nargs) \
@@ -588,7 +588,7 @@ extern double __REDIRECT_NTH (nexttoward, (double __x, long double __y),
 #  undef __MATHCALL_NARROW
 #  define __MATHCALL_NARROW(func, redir, nargs) \
   __MATHCALL_NARROW_REDIR (func, redir, nargs)
-# elif __LONG_DOUBLE_USES_FLOAT128 == 1
+# elif __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #  define __MATHCALL_REDIR_NAME(name) __ ## f64 ## name ## ieee128
 #  undef __MATHCALL_NARROW
 #  define __MATHCALL_NARROW(func, redir, nargs) \
@@ -599,7 +599,7 @@ extern double __REDIRECT_NTH (nexttoward, (double __x, long double __y),
 # undef _Marg_
 # undef __MATHCALL_NAME
 # if defined __LDBL_COMPAT \
-     || __LONG_DOUBLE_USES_FLOAT128 == 1
+     || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 #  undef __MATHCALL_REDIR_NAME
 #  undef __MATHCALL_NARROW
 #  define __MATHCALL_NARROW(func, redir, nargs) \
diff --git a/math/test-ldouble.h b/math/test-ldouble.h
index b9ef6f41f9..3d62c92e14 100644
--- a/math/test-ldouble.h
+++ b/math/test-ldouble.h
@@ -29,7 +29,7 @@
 /* On architectures which redirect long double to _Float128 ABI, we must
    choose the float128 ulps.  Similarly, on such architectures, the ABI
    used may be dependent on how the compiler was invoked.  */
-#elif __LONG_DOUBLE_USES_FLOAT128 == 1
+#elif __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # define TYPE_STR "float128"
 # define ULP_IDX ULP_FLT128
 #else
diff --git a/misc/err.h b/misc/err.h
index d146a55054..e56cd73c0e 100644
--- a/misc/err.h
+++ b/misc/err.h
@@ -53,7 +53,7 @@ extern void verrx (int __status, const char *, __gnuc_va_list)
      __attribute__ ((__noreturn__, __format__ (__printf__, 2, 0)));
 
 #include <bits/floatn.h>
-#if defined __LDBL_COMPAT || __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __LDBL_COMPAT || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # include <bits/err-ldbl.h>
 #endif
 
diff --git a/misc/error.h b/misc/error.h
index bb8a3b26a4..41d54c568f 100644
--- a/misc/error.h
+++ b/misc/error.h
@@ -48,7 +48,7 @@ extern unsigned int error_message_count;
 extern int error_one_per_line;
 
 #include <bits/floatn.h>
-#if defined __LDBL_COMPAT || __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __LDBL_COMPAT || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # include <bits/error-ldbl.h>
 #else
 /* Do not inline error and error_at_line when long double has the same
diff --git a/misc/sys/cdefs.h b/misc/sys/cdefs.h
index dd5f69f833..9fa371ab86 100644
--- a/misc/sys/cdefs.h
+++ b/misc/sys/cdefs.h
@@ -452,7 +452,7 @@
 #include <bits/wordsize.h>
 #include <bits/long-double.h>
 
-#if __LONG_DOUBLE_USES_FLOAT128 == 1
+#if __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # ifdef __REDIRECT
 
 /* Alias name defined automatically.  */
@@ -503,7 +503,7 @@ _Static_assert (0, "IEEE 128-bits long double requires redirection on this platf
   __LDBL_REDIR1_NTH (name, proto, __nldbl_##alias)
 # endif
 #endif
-#if (!defined __LDBL_COMPAT && __LONG_DOUBLE_USES_FLOAT128 == 0) \
+#if (!defined __LDBL_COMPAT && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 0) \
     || !defined __REDIRECT
 # define __LDBL_REDIR1(name, proto, alias) name proto
 # define __LDBL_REDIR(name, proto) name proto
diff --git a/misc/sys/syslog.h b/misc/sys/syslog.h
index c000b8e915..dc3b0e7ef8 100644
--- a/misc/sys/syslog.h
+++ b/misc/sys/syslog.h
@@ -208,7 +208,7 @@ extern void vsyslog (int __pri, const char *__fmt, __gnuc_va_list __ap)
 #endif
 
 #include <bits/floatn.h>
-#if defined __LDBL_COMPAT || __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __LDBL_COMPAT || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # include <bits/syslog-ldbl.h>
 #endif
 
diff --git a/stdio-common/printf.h b/stdio-common/printf.h
index 00bedebfdb..7dccfcd6fb 100644
--- a/stdio-common/printf.h
+++ b/stdio-common/printf.h
@@ -183,7 +183,7 @@ extern int printf_size_info (const struct printf_info *__restrict
      __THROW;
 
 #include <bits/floatn.h>
-#if defined __LDBL_COMPAT || __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __LDBL_COMPAT || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # include <bits/printf-ldbl.h>
 #endif
 
diff --git a/stdio-common/tst-vfprintf-user-type.c b/stdio-common/tst-vfprintf-user-type.c
index 40d714fdb1..0abd61cea3 100644
--- a/stdio-common/tst-vfprintf-user-type.c
+++ b/stdio-common/tst-vfprintf-user-type.c
@@ -147,7 +147,7 @@ do_test (void)
 
   /* Alias declaration for asprintf, to avoid the format string
      attribute and the associated warning.  */
-#if __LONG_DOUBLE_USES_FLOAT128 == 1
+#if __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
   extern int asprintf_alias (char **, const char *, ...) __asm__ ("__asprintfieee128");
 #else
   extern int asprintf_alias (char **, const char *, ...) __asm__ ("asprintf");
diff --git a/stdlib/bits/stdlib-ldbl.h b/stdlib/bits/stdlib-ldbl.h
index 2469474532..cc8831dd16 100644
--- a/stdlib/bits/stdlib-ldbl.h
+++ b/stdlib/bits/stdlib-ldbl.h
@@ -51,7 +51,7 @@ __LDBL_REDIR1_DECL (qfcvt, fcvt)
 __LDBL_REDIR1_DECL (qgcvt, gcvt)
 __LDBL_REDIR1_DECL (qecvt_r, ecvt_r)
 __LDBL_REDIR1_DECL (qfcvt_r, fcvt_r)
-# elif __LONG_DOUBLE_USES_FLOAT128 == 1
+# elif __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 __LDBL_REDIR1_DECL (qecvt, __qecvtieee128)
 __LDBL_REDIR1_DECL (qfcvt, __qfcvtieee128)
 __LDBL_REDIR1_DECL (qgcvt, __qgcvtieee128)
diff --git a/stdlib/monetary.h b/stdlib/monetary.h
index 4a6d61fd2d..c9d3c64e14 100644
--- a/stdlib/monetary.h
+++ b/stdlib/monetary.h
@@ -51,7 +51,7 @@ extern ssize_t strfmon_l (char *__restrict __s, size_t __maxsize,
 #endif
 
 #include <bits/floatn.h>
-#if defined __LDBL_COMPAT || __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __LDBL_COMPAT || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # include <bits/monetary-ldbl.h>
 #endif
 
diff --git a/stdlib/stdlib.h b/stdlib/stdlib.h
index a7c0c71a80..9b7537c545 100644
--- a/stdlib/stdlib.h
+++ b/stdlib/stdlib.h
@@ -1018,7 +1018,7 @@ extern int ttyslot (void) __THROW;
 #endif
 
 #include <bits/floatn.h>
-#if defined __LDBL_COMPAT || __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __LDBL_COMPAT || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # include <bits/stdlib-ldbl.h>
 #endif
 
diff --git a/sysdeps/ieee754/ldbl-128/bits/long-double.h b/sysdeps/ieee754/ldbl-128/bits/long-double.h
index 941e60505f..dd64a06db6 100644
--- a/sysdeps/ieee754/ldbl-128/bits/long-double.h
+++ b/sysdeps/ieee754/ldbl-128/bits/long-double.h
@@ -18,4 +18,4 @@
 
 /* long double is distinct from double, so there is nothing to
    define here.  */
-#define __LONG_DOUBLE_USES_FLOAT128 0
+#define __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI 0
diff --git a/sysdeps/ieee754/ldbl-128ibm-compat/bits/long-double.h b/sysdeps/ieee754/ldbl-128ibm-compat/bits/long-double.h
index 91dddbdc8b..ef834c7748 100644
--- a/sysdeps/ieee754/ldbl-128ibm-compat/bits/long-double.h
+++ b/sysdeps/ieee754/ldbl-128ibm-compat/bits/long-double.h
@@ -22,8 +22,5 @@
 #  define __NO_LONG_DOUBLE_MATH		1
 # endif
 #endif
-/* On platforms that reuse the _Float128 implementation for IEEE long
-   double, access to the correct long double functions is selected based
-   on the long double mode being used during the compilation.  On
-   powerpc64le, this is true when -mabi=ieeelongdouble is in use.  */
-#define __LONG_DOUBLE_USES_FLOAT128 (__LDBL_MANT_DIG__ == 113)
+
+#define __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI (__LDBL_MANT_DIG__ == 113)
diff --git a/sysdeps/ieee754/ldbl-128ibm-compat/math_ldbl.h b/sysdeps/ieee754/ldbl-128ibm-compat/math_ldbl.h
index 352617e178..5490b400dd 100644
--- a/sysdeps/ieee754/ldbl-128ibm-compat/math_ldbl.h
+++ b/sysdeps/ieee754/ldbl-128ibm-compat/math_ldbl.h
@@ -23,7 +23,7 @@
 
 /* Trampoline in the ldbl-128ibm headers if building against the
    old abi.  Otherwise, we have nothing to add. */
-#if __LONG_DOUBLE_USES_FLOAT128 == 0
+#if __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 0
 #include_next <math_ldbl.h>
 #endif
 
diff --git a/sysdeps/ieee754/ldbl-128ibm/bits/iscanonical.h b/sysdeps/ieee754/ldbl-128ibm/bits/iscanonical.h
index b2021b5d64..e60b3017eb 100644
--- a/sysdeps/ieee754/ldbl-128ibm/bits/iscanonical.h
+++ b/sysdeps/ieee754/ldbl-128ibm/bits/iscanonical.h
@@ -20,7 +20,7 @@
 # error "Never use <bits/iscanonical.h> directly; include <math.h> instead."
 #endif
 
-#if defined (__NO_LONG_DOUBLE_MATH) || __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined (__NO_LONG_DOUBLE_MATH) || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # define iscanonical(x) ((void) (__typeof (x)) (x), 1)
 #else
 extern int __iscanonicall (long double __x)
diff --git a/sysdeps/ieee754/ldbl-128ibm/ieee754.h b/sysdeps/ieee754/ldbl-128ibm/ieee754.h
index fe4892d84f..b66e3b9869 100644
--- a/sysdeps/ieee754/ldbl-128ibm/ieee754.h
+++ b/sysdeps/ieee754/ldbl-128ibm/ieee754.h
@@ -112,7 +112,7 @@ union ieee754_double
 #define IEEE754_DOUBLE_BIAS	0x3ff /* Added to exponent.  */
 
 
-#if __LONG_DOUBLE_USES_FLOAT128 == 1
+#if __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 /* long double is IEEE 128 bit */
 union ieee854_long_double
   {
@@ -170,7 +170,7 @@ union ieee854_long_double
 #endif
 
 
-#if __LONG_DOUBLE_USES_FLOAT128 == 0 || __GNUC_PREREQ (7, 0)
+#if __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 0 || __GNUC_PREREQ (7, 0)
 /* IBM extended format for long double.
 
    Each long double is made up of two IEEE doubles.  The value of the
@@ -183,7 +183,7 @@ union ieee854_long_double
    NaN is don't-care.  */
 union ibm_extended_long_double
   {
-# if __LONG_DOUBLE_USES_FLOAT128 == 1 && __GNUC_PREREQ (7, 0)
+# if __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1 && __GNUC_PREREQ (7, 0)
    __ibm128 ld;
 # else
    long double ld;
diff --git a/sysdeps/ieee754/ldbl-128ibm/include/bits/iscanonical.h b/sysdeps/ieee754/ldbl-128ibm/include/bits/iscanonical.h
index 8cbd6f74bf..60d5417089 100644
--- a/sysdeps/ieee754/ldbl-128ibm/include/bits/iscanonical.h
+++ b/sysdeps/ieee754/ldbl-128ibm/include/bits/iscanonical.h
@@ -1,5 +1,5 @@
 #include_next <bits/iscanonical.h>
 
-#if !defined _ISOMAC && (__LONG_DOUBLE_USES_FLOAT128 == 0)
+#if !defined _ISOMAC && (__LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 0)
 libm_hidden_proto (__iscanonicall)
 #endif
diff --git a/sysdeps/ieee754/ldbl-96/bits/long-double.h b/sysdeps/ieee754/ldbl-96/bits/long-double.h
index b77124f842..9a2cd0dfe4 100644
--- a/sysdeps/ieee754/ldbl-96/bits/long-double.h
+++ b/sysdeps/ieee754/ldbl-96/bits/long-double.h
@@ -18,4 +18,4 @@
 
 /* long double is distinct from double, so there is nothing to
    define here.  */
-#define __LONG_DOUBLE_USES_FLOAT128 0
+#define __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI 0
diff --git a/sysdeps/ieee754/ldbl-opt/bits/long-double.h b/sysdeps/ieee754/ldbl-opt/bits/long-double.h
index 242e58410d..48688ad7e1 100644
--- a/sysdeps/ieee754/ldbl-opt/bits/long-double.h
+++ b/sysdeps/ieee754/ldbl-opt/bits/long-double.h
@@ -22,4 +22,4 @@
 #  define __NO_LONG_DOUBLE_MATH		1
 # endif
 #endif
-#define __LONG_DOUBLE_USES_FLOAT128 0
+#define __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI 0
diff --git a/sysdeps/mips/ieee754/bits/long-double.h b/sysdeps/mips/ieee754/bits/long-double.h
index 57007a84d6..1ad74e4497 100644
--- a/sysdeps/mips/ieee754/bits/long-double.h
+++ b/sysdeps/mips/ieee754/bits/long-double.h
@@ -21,4 +21,4 @@
 #if !defined __NO_LONG_DOUBLE_MATH && _MIPS_SIM == _ABIO32
 # define __NO_LONG_DOUBLE_MATH	1
 #endif
-#define __LONG_DOUBLE_USES_FLOAT128 0
+#define __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI 0
diff --git a/sysdeps/unix/sysv/linux/sparc/sparc32/bits/long-double.h b/sysdeps/unix/sysv/linux/sparc/sparc32/bits/long-double.h
index 7046cd40d7..107c5d179b 100644
--- a/sysdeps/unix/sysv/linux/sparc/sparc32/bits/long-double.h
+++ b/sysdeps/unix/sysv/linux/sparc/sparc32/bits/long-double.h
@@ -24,4 +24,4 @@
 #  define __NO_LONG_DOUBLE_MATH        1
 # endif
 #endif
-#define __LONG_DOUBLE_USES_FLOAT128 0
+#define __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI 0
diff --git a/sysdeps/unix/sysv/linux/sparc/sparc64/bits/long-double.h b/sysdeps/unix/sysv/linux/sparc/sparc64/bits/long-double.h
index 7046cd40d7..107c5d179b 100644
--- a/sysdeps/unix/sysv/linux/sparc/sparc64/bits/long-double.h
+++ b/sysdeps/unix/sysv/linux/sparc/sparc64/bits/long-double.h
@@ -24,4 +24,4 @@
 #  define __NO_LONG_DOUBLE_MATH        1
 # endif
 #endif
-#define __LONG_DOUBLE_USES_FLOAT128 0
+#define __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI 0
diff --git a/wcsmbs/bits/wchar-ldbl.h b/wcsmbs/bits/wchar-ldbl.h
index 75d547f680..1b063d1235 100644
--- a/wcsmbs/bits/wchar-ldbl.h
+++ b/wcsmbs/bits/wchar-ldbl.h
@@ -32,7 +32,7 @@ __LDBL_REDIR_DECL (vswprintf);
 __LDBL_REDIR1_DECL (fwscanf, __nldbl___isoc99_fwscanf)
 __LDBL_REDIR1_DECL (wscanf, __nldbl___isoc99_wscanf)
 __LDBL_REDIR1_DECL (swscanf, __nldbl___isoc99_swscanf)
-#  elif __LONG_DOUBLE_USES_FLOAT128 == 1
+#  elif __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 __LDBL_REDIR1_DECL (fwscanf, __isoc99_fwscanfieee128)
 __LDBL_REDIR1_DECL (wscanf, __isoc99_wscanfieee128)
 __LDBL_REDIR1_DECL (swscanf, __isoc99_swscanfieee128)
@@ -57,7 +57,7 @@ __LDBL_REDIR1_DECL (wcstold, __wcstoieee128)
 __LDBL_REDIR1_DECL (vfwscanf, __nldbl___isoc99_vfwscanf)
 __LDBL_REDIR1_DECL (vwscanf, __nldbl___isoc99_vwscanf)
 __LDBL_REDIR1_DECL (vswscanf, __nldbl___isoc99_vswscanf)
-#  elif __LONG_DOUBLE_USES_FLOAT128 == 1
+#  elif __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 __LDBL_REDIR1_DECL (vfwscanf, __isoc99_vfwscanfieee128)
 __LDBL_REDIR1_DECL (vwscanf, __isoc99_vwscanfieee128)
 __LDBL_REDIR1_DECL (vswscanf, __isoc99_vswscanfieee128)
diff --git a/wcsmbs/wchar.h b/wcsmbs/wchar.h
index fb2cec8167..9cf8b05a87 100644
--- a/wcsmbs/wchar.h
+++ b/wcsmbs/wchar.h
@@ -634,10 +634,10 @@ extern int swscanf (const wchar_t *__restrict __s,
 
 /* For historical reasons, the C99-compliant versions of the scanf
    functions are at alternative names.  When __LDBL_COMPAT or
-   __LONG_DOUBLE_USES_FLOAT128 are in effect, this is handled in
+   __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI are in effect, this is handled in
    bits/wchar-ldbl.h.  */
 #if !__GLIBC_USE (DEPRECATED_SCANF) && !defined __LDBL_COMPAT \
-     && __LONG_DOUBLE_USES_FLOAT128 == 0
+     && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 0
 #  ifdef __REDIRECT
 extern int __REDIRECT (fwscanf, (__FILE *__restrict __stream,
 				 const wchar_t *__restrict __format, ...),
@@ -691,7 +691,7 @@ extern int vswscanf (const wchar_t *__restrict __s,
 # if !__GLIBC_USE (DEPRECATED_SCANF) \
      && (!defined __LDBL_COMPAT || !defined __REDIRECT) \
      && (defined __STRICT_ANSI__ || defined __USE_XOPEN2K) \
-     && __LONG_DOUBLE_USES_FLOAT128 == 0
+     && __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 0
 #  ifdef __REDIRECT
 extern int __REDIRECT (vfwscanf, (__FILE *__restrict __s,
 				  const wchar_t *__restrict __format,
@@ -853,7 +853,7 @@ extern size_t wcsftime_l (wchar_t *__restrict __s, size_t __maxsize,
 #endif
 
 #include <bits/floatn.h>
-#if defined __LDBL_COMPAT || __LONG_DOUBLE_USES_FLOAT128 == 1
+#if defined __LDBL_COMPAT || __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI == 1
 # include <bits/wchar-ldbl.h>
 #endif
 
-- 
2.21.1


^ permalink raw reply	[flat|nested] 22+ messages in thread

* [PATCH v3 3/6] powerpc64le: raise GCC requirement to 7.4 for long double transition
  2020-04-06 19:12 [PATCH v3 0/6] IEEE binary128 long double on powerpc64le Paul E. Murphy
  2020-04-06 19:12 ` [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731 Paul E. Murphy
  2020-04-06 19:12 ` [PATCH v3 2/6] Rename __LONG_DOUBLE_USES_FLOAT128 to __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI Paul E. Murphy
@ 2020-04-06 19:12 ` Paul E. Murphy
  2020-04-22 15:19   ` Florian Weimer
  2020-04-06 19:12 ` [PATCH v3 4/6] powerpc64le: bump binutils version requirement to >= 2.26 Paul E. Murphy
                   ` (2 subsequent siblings)
  5 siblings, 1 reply; 22+ messages in thread
From: Paul E. Murphy @ 2020-04-06 19:12 UTC (permalink / raw)
  To: libc-alpha

Add compiler feature tests to ensure we can build ieee128 long double.
These test for -mabi=ieeelongdouble, -mno-gnu-attribute, and -Wno-psabi.

Likewise, verify some compiler bugs have been addressed.  These aren't
helpful for building glibc, but may cause test failures when testing
the new long double.  See notes below from Raji.

On powerpc64le, some older compiler versions give error for the function
signbit() for 128-bit floating point types. This is fixed by PR83862
in gcc 8.0 and backported to gcc6 and gcc7. This patch adds a test
to check compiler version to avoid compiler errors during make check.

Likewise, test for -mno-gnu-attribute support which was

On powerpc64le, a few files are built on IEEE long double mode
(-mabi=ieeelongdouble), whereas most are built on IBM long double mode
(-mabi=ibmlongdouble, the default for -mlong-double-128). Since binutils
2.31, linking object files with different long double modes causes
errors similar to:

  ld: libc_pic.a(s_isinfl.os) uses IBM long double,
      libc_pic.a(ieee128-qefgcvt.os) uses IEEE long double.
  collect2: error: ld returned 1 exit status
  make[2]: *** [../Makerules:649: libc_pic.os] Error 1

The warnings are fair and correct, but in order for glibc to have
support for both long double modes on powerpc64le, they have to be
ignored.  This can be accomplished with the use of -mno-gnu-attribute
option when building the few files that require IEEE long double mode.

However, -mno-gnu-attribute is not available in GCC 6, the minimum
version required to build glibc, so this patch adds a test for this
feature in powerpc64le builds, and fails early if it's not available.

Co-Authored-By: Rajalakshmi Srinivasaraghavan  <raji@linux.vnet.ibm.com>
Co-Authored-By: Gabriel F. T. Gomes <gabrielftg@linux.ibm.com>
---
 INSTALL                                   |  6 +++
 NEWS                                      |  3 +-
 manual/install.texi                       |  5 ++
 sysdeps/powerpc/powerpc64/le/configure    | 59 +++++++++++++++++++++++
 sysdeps/powerpc/powerpc64/le/configure.ac | 35 ++++++++++++++
 5 files changed, 107 insertions(+), 1 deletion(-)

diff --git a/INSTALL b/INSTALL
index 242cb06f91..d7c109358c 100644
--- a/INSTALL
+++ b/INSTALL
@@ -467,6 +467,12 @@ build the GNU C Library:
      better code.  As of release time, GCC 9.2.1 is the newest compiler
      verified to work to build the GNU C Library.
 
+     For PowerPC 64-bits little-endian (powerpc64le), a GCC version with
+     support for '-mno-gnu-attribute' is required.  As of release time,
+     this means GCC 7.4 (excepting 7.5) or newer.  The compiler option
+     is required for building the GNU C Library with support for IEEE
+     long double.
+
      For multi-arch support it is recommended to use a GCC which has
      been built with support for GNU indirect functions.  This ensures
      that correct debugging information is generated for functions
diff --git a/NEWS b/NEWS
index 3cf738adfb..07da1545b4 100644
--- a/NEWS
+++ b/NEWS
@@ -20,7 +20,8 @@ Deprecated and removed features, and other changes affecting compatibility:
 
 Changes to build and runtime requirements:
 
-  [Add changes to build and runtime requirements here]
+* powerpc64le requires GCC 7.4 or newer.  This is required for supporting
+  long double redirects.
 
 Security related changes:
 
diff --git a/manual/install.texi b/manual/install.texi
index 71bf47cac6..66de8ea59b 100644
--- a/manual/install.texi
+++ b/manual/install.texi
@@ -507,6 +507,11 @@ the newest version of the compiler that is known to work for building
 release time, GCC 9.2.1 is the newest compiler verified to work to build
 @theglibc{}.
 
+For PowerPC 64-bits little-endian (powerpc64le), a GCC version with
+support for @option{-mno-gnu-attribute} is required.  As of release time,
+this means GCC 7.4 (excepting 7.5) or newer. The compiler option is required
+for building the GNU C Library with support for IEEE long double.
+
 For multi-arch support it is recommended to use a GCC which has been built with
 support for GNU indirect functions.  This ensures that correct debugging
 information is generated for functions selected by IFUNC resolvers.  This
diff --git a/sysdeps/powerpc/powerpc64/le/configure b/sysdeps/powerpc/powerpc64/le/configure
index bca80cec76..4b4d6cff6f 100644
--- a/sysdeps/powerpc/powerpc64/le/configure
+++ b/sysdeps/powerpc/powerpc64/le/configure
@@ -32,4 +32,63 @@ if test "$libc_cv_target_power8_ok" != "yes"; then :
 fi
 CFLAGS="$OLD_CFLAGS"
 
+OLD_CFLAGS="$CFLAGS"
+CFLAGS="$CFLAGS -mno-gnu-attribute"
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking if the compiler supports -mno-gnu-attribute" >&5
+$as_echo_n "checking if the compiler supports -mno-gnu-attribute... " >&6; }
+if ${libc_cv_no_gnu_attr_ok+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+
+cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+
+
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+  libc_cv_no_gnu_attr_ok=yes
+else
+  libc_cv_no_gnu_attr_ok=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $libc_cv_no_gnu_attr_ok" >&5
+$as_echo "$libc_cv_no_gnu_attr_ok" >&6; }
+if test "$libc_cv_no_gnu_attr_ok" != "yes"; then :
+  critic_missing="$critic_missing A compiler with -mno-gnu-attribute is required on powerpc64le."
+fi
+CFLAGS="$OLD_CFLAGS"
+
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking if $CC compiles signbit with 128-bit floating point type" >&5
+$as_echo_n "checking if $CC compiles signbit with 128-bit floating point type... " >&6; }
+if ${libc_cv_compiler_powerpc64le_ice+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  save_CFLAGS="$CFLAGS"
+CFLAGS="$CFLAGS -mabi=ieeelongdouble -Wno-psabi"
+cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+
+int sbr (long double a) { return __builtin_signbit (a); }
+int sbm (long double *a) { return __builtin_signbit (*a); }
+int sbo (long double *a) { return __builtin_signbit (a[4]); }
+int sbi (long double *a, unsigned long n) { return __builtin_signbit (a[n]); }
+void sbs (int *p, long double a) { *p = __builtin_signbit (a); }
+
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+  libc_cv_compiler_powerpc64le_ice=yes
+else
+  libc_cv_compiler_powerpc64le_ice=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+CFLAGS="$save_CFLAGS"
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $libc_cv_compiler_powerpc64le_ice" >&5
+$as_echo "$libc_cv_compiler_powerpc64le_ice" >&6; }
+if test "$libc_cv_compiler_powerpc64le_ice" != "yes"; then :
+  critic_missing="$critic_missing __builtin_signbit is broken. GCC 7.4 or newer is required to resolve (BZ 83862)."
+fi
+
+
 test -n "$critic_missing" && as_fn_error $? "*** $critic_missing" "$LINENO" 5
diff --git a/sysdeps/powerpc/powerpc64/le/configure.ac b/sysdeps/powerpc/powerpc64/le/configure.ac
index 7f2c9fbf1f..26a54beaa0 100644
--- a/sysdeps/powerpc/powerpc64/le/configure.ac
+++ b/sysdeps/powerpc/powerpc64/le/configure.ac
@@ -17,4 +17,39 @@ AS_IF([test "$libc_cv_target_power8_ok" != "yes"],
       [critic_missing="$critic_missing POWER8 or newer is required on powerpc64le."])
 CFLAGS="$OLD_CFLAGS"
 
+dnl Require support for -mno-gnu-attribute
+OLD_CFLAGS="$CFLAGS"
+CFLAGS="$CFLAGS -mno-gnu-attribute"
+AC_CACHE_CHECK([if the compiler supports -mno-gnu-attribute],
+	       libc_cv_no_gnu_attr_ok, [
+AC_COMPILE_IFELSE([AC_LANG_SOURCE([[
+]])],
+	       [libc_cv_no_gnu_attr_ok=yes],
+	       [libc_cv_no_gnu_attr_ok=no])])
+AS_IF([test "$libc_cv_no_gnu_attr_ok" != "yes"],
+      [critic_missing="$critic_missing A compiler with -mno-gnu-attribute is required on powerpc64le."])
+CFLAGS="$OLD_CFLAGS"
+
+dnl Some old compiler versions give error for the function signbit() for
+dnl 128-bit floating point types. This is fixed by BZ 83862 and backported
+dnl to gcc6 and gcc7. This test is to check if we are using the compiler
+dnl that has this bug. (available in GCC 8.0).
+AC_CACHE_CHECK([if $CC compiles signbit with 128-bit floating point type],
+	       libc_cv_compiler_powerpc64le_ice, [dnl
+save_CFLAGS="$CFLAGS"
+CFLAGS="$CFLAGS -mabi=ieeelongdouble -Wno-psabi"
+AC_COMPILE_IFELSE([AC_LANG_SOURCE([[
+int sbr (long double a) { return __builtin_signbit (a); }
+int sbm (long double *a) { return __builtin_signbit (*a); }
+int sbo (long double *a) { return __builtin_signbit (a[4]); }
+int sbi (long double *a, unsigned long n) { return __builtin_signbit (a[n]); }
+void sbs (int *p, long double a) { *p = __builtin_signbit (a); }
+]])],
+		  [libc_cv_compiler_powerpc64le_ice=yes],
+		  [libc_cv_compiler_powerpc64le_ice=no])
+CFLAGS="$save_CFLAGS"])
+AS_IF([test "$libc_cv_compiler_powerpc64le_ice" != "yes"],
+      [critic_missing="$critic_missing __builtin_signbit is broken. GCC 7.4 or newer is required to resolve (BZ 83862)."])
+
+
 test -n "$critic_missing" && AC_MSG_ERROR([*** $critic_missing])
-- 
2.21.1


^ permalink raw reply	[flat|nested] 22+ messages in thread

* [PATCH v3 4/6] powerpc64le: bump binutils version requirement to >= 2.26
  2020-04-06 19:12 [PATCH v3 0/6] IEEE binary128 long double on powerpc64le Paul E. Murphy
                   ` (2 preceding siblings ...)
  2020-04-06 19:12 ` [PATCH v3 3/6] powerpc64le: raise GCC requirement to 7.4 for long double transition Paul E. Murphy
@ 2020-04-06 19:12 ` Paul E. Murphy
  2020-04-22 15:20   ` Florian Weimer
  2020-04-06 19:12 ` [PATCH v3 5/6] powerpc64le: blacklist broken GCC compilers (e.g GCC 7.5.0) Paul E. Murphy
  2020-04-06 19:12 ` [PATCH v3 6/6] powerpc64le: Enable support for IEEE long double Paul E. Murphy
  5 siblings, 1 reply; 22+ messages in thread
From: Paul E. Murphy @ 2020-04-06 19:12 UTC (permalink / raw)
  To: libc-alpha

This is a small step up from 2.25 which brings in support for
rewriting the .gnu.attributes section of libc/libm.so.

Reviewed-by: Tulio Magno Quites Machado Filho <tuliom@linux.ibm.com>
---
 INSTALL                                   |  4 ++
 manual/install.texi                       |  4 ++
 sysdeps/powerpc/powerpc64/le/configure    | 65 +++++++++++++++++++++++
 sysdeps/powerpc/powerpc64/le/configure.ac |  6 +++
 4 files changed, 79 insertions(+)

diff --git a/INSTALL b/INSTALL
index d7c109358c..551ca9434e 100644
--- a/INSTALL
+++ b/INSTALL
@@ -495,6 +495,10 @@ build the GNU C Library:
      moment.  As of release time, GNU 'binutils' 2.32 is the newest
      verified to work to build the GNU C Library.
 
+     For PowerPC 64-bits little-endian (powerpc64le), binutils is
+     required to support the '--update-section'.  This option requires
+     binutils 2.26 or newer.
+
    * GNU 'texinfo' 4.7 or later
 
      To correctly translate and install the Texinfo documentation you
diff --git a/manual/install.texi b/manual/install.texi
index 66de8ea59b..acbcc8c091 100644
--- a/manual/install.texi
+++ b/manual/install.texi
@@ -533,6 +533,10 @@ No other assembler or linker has the necessary functionality at the
 moment. As of release time, GNU @code{binutils} 2.32 is the newest
 verified to work to build @theglibc{}.
 
+For PowerPC 64-bits little-endian (powerpc64le), binutils is required
+to support the @option{--update-section}.  This option requires binutils
+2.26 or newer.
+
 @item
 GNU @code{texinfo} 4.7 or later
 
diff --git a/sysdeps/powerpc/powerpc64/le/configure b/sysdeps/powerpc/powerpc64/le/configure
index 4b4d6cff6f..b08c3e0587 100644
--- a/sysdeps/powerpc/powerpc64/le/configure
+++ b/sysdeps/powerpc/powerpc64/le/configure
@@ -90,5 +90,70 @@ if test "$libc_cv_compiler_powerpc64le_ice" != "yes"; then :
   critic_missing="$critic_missing __builtin_signbit is broken. GCC 7.4 or newer is required to resolve (BZ 83862)."
 fi
 
+# Binutils (objcopy) 2.26 or newer required to support the --update-section
+# feature for fixing up .gnu.attribute section with IEEE ldbl.
+for ac_prog in $OBJCOPY
+do
+  # Extract the first word of "$ac_prog", so it can be a program name with args.
+set dummy $ac_prog; ac_word=$2
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for $ac_word" >&5
+$as_echo_n "checking for $ac_word... " >&6; }
+if ${ac_cv_prog_OBJCOPY+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  if test -n "$OBJCOPY"; then
+  ac_cv_prog_OBJCOPY="$OBJCOPY" # Let the user override the test.
+else
+as_save_IFS=$IFS; IFS=$PATH_SEPARATOR
+for as_dir in $PATH
+do
+  IFS=$as_save_IFS
+  test -z "$as_dir" && as_dir=.
+    for ac_exec_ext in '' $ac_executable_extensions; do
+  if as_fn_executable_p "$as_dir/$ac_word$ac_exec_ext"; then
+    ac_cv_prog_OBJCOPY="$ac_prog"
+    $as_echo "$as_me:${as_lineno-$LINENO}: found $as_dir/$ac_word$ac_exec_ext" >&5
+    break 2
+  fi
+done
+  done
+IFS=$as_save_IFS
+
+fi
+fi
+OBJCOPY=$ac_cv_prog_OBJCOPY
+if test -n "$OBJCOPY"; then
+  { $as_echo "$as_me:${as_lineno-$LINENO}: result: $OBJCOPY" >&5
+$as_echo "$OBJCOPY" >&6; }
+else
+  { $as_echo "$as_me:${as_lineno-$LINENO}: result: no" >&5
+$as_echo "no" >&6; }
+fi
+
+
+  test -n "$OBJCOPY" && break
+done
+
+if test -z "$OBJCOPY"; then
+  ac_verc_fail=yes
+else
+  # Found it, now check the version.
+  { $as_echo "$as_me:${as_lineno-$LINENO}: checking version of $OBJCOPY" >&5
+$as_echo_n "checking version of $OBJCOPY... " >&6; }
+  ac_prog_version=`$OBJCOPY --version 2>&1 | sed -n 's/^.*GNU objcopy.* \([0-9]*\.[0-9.]*\).*$/\1/p'`
+  case $ac_prog_version in
+    '') ac_prog_version="v. ?.??, bad"; ac_verc_fail=yes;;
+    2.1[0-9][0-9]*|2.2[6-9]*|2.[3-9][0-9]*|[3-9].*|[1-9][0-9]*)
+       ac_prog_version="$ac_prog_version, ok"; ac_verc_fail=no;;
+    *) ac_prog_version="$ac_prog_version, bad"; ac_verc_fail=yes;;
+
+  esac
+  { $as_echo "$as_me:${as_lineno-$LINENO}: result: $ac_prog_version" >&5
+$as_echo "$ac_prog_version" >&6; }
+fi
+if test $ac_verc_fail = yes; then
+  AS=: critic_missing="$critic_missing objcopy >= 2.26 is required on powerpc64le"
+fi
+
 
 test -n "$critic_missing" && as_fn_error $? "*** $critic_missing" "$LINENO" 5
diff --git a/sysdeps/powerpc/powerpc64/le/configure.ac b/sysdeps/powerpc/powerpc64/le/configure.ac
index 26a54beaa0..630086bd88 100644
--- a/sysdeps/powerpc/powerpc64/le/configure.ac
+++ b/sysdeps/powerpc/powerpc64/le/configure.ac
@@ -51,5 +51,11 @@ CFLAGS="$save_CFLAGS"])
 AS_IF([test "$libc_cv_compiler_powerpc64le_ice" != "yes"],
       [critic_missing="$critic_missing __builtin_signbit is broken. GCC 7.4 or newer is required to resolve (BZ 83862)."])
 
+# Binutils (objcopy) 2.26 or newer required to support the --update-section
+# feature for fixing up .gnu.attribute section with IEEE ldbl.
+AC_CHECK_PROG_VER(OBJCOPY, $OBJCOPY, --version,
+		  [GNU objcopy.* \([0-9]*\.[0-9.]*\)],
+		  [2.1[0-9][0-9]*|2.2[6-9]*|2.[3-9][0-9]*|[3-9].*|[1-9][0-9]*],
+		  AS=: critic_missing="$critic_missing objcopy >= 2.26 is required on powerpc64le")
 
 test -n "$critic_missing" && AC_MSG_ERROR([*** $critic_missing])
-- 
2.21.1


^ permalink raw reply	[flat|nested] 22+ messages in thread

* [PATCH v3 5/6] powerpc64le: blacklist broken GCC compilers (e.g GCC 7.5.0)
  2020-04-06 19:12 [PATCH v3 0/6] IEEE binary128 long double on powerpc64le Paul E. Murphy
                   ` (3 preceding siblings ...)
  2020-04-06 19:12 ` [PATCH v3 4/6] powerpc64le: bump binutils version requirement to >= 2.26 Paul E. Murphy
@ 2020-04-06 19:12 ` Paul E. Murphy
  2020-04-22 15:23   ` Florian Weimer
  2020-04-06 19:12 ` [PATCH v3 6/6] powerpc64le: Enable support for IEEE long double Paul E. Murphy
  5 siblings, 1 reply; 22+ messages in thread
From: Paul E. Murphy @ 2020-04-06 19:12 UTC (permalink / raw)
  To: libc-alpha

GCC 7.5.0 (BZ# 94200) will refuse to compile if both -mabi=% and
-mlong-double-128 are passed on the command line.  Surprisingly,
it will work happily if the latter is not.  For the sake of
maintianing status quo, test for and blacklist such compilers.

Tested with a GCC 8.3.1 and GCC 7.5.0 compiler for ppc64le.
---
 sysdeps/ieee754/ldbl-128ibm-compat/Makefile | 12 ---------
 sysdeps/powerpc/powerpc64/le/configure      | 27 +++++++++++++++++++++
 sysdeps/powerpc/powerpc64/le/configure.ac   | 15 ++++++++++++
 3 files changed, 42 insertions(+), 12 deletions(-)

diff --git a/sysdeps/ieee754/ldbl-128ibm-compat/Makefile b/sysdeps/ieee754/ldbl-128ibm-compat/Makefile
index 0ac52a53bc..7b42746c34 100644
--- a/sysdeps/ieee754/ldbl-128ibm-compat/Makefile
+++ b/sysdeps/ieee754/ldbl-128ibm-compat/Makefile
@@ -249,13 +249,6 @@ ldbl-extra-routines += err \
 CFLAGS-ieee128-qefgcvt.c += -mabi=ieeelongdouble -Wno-psabi -mno-gnu-attribute
 CFLAGS-ieee128-qefgcvt_r.c += -mabi=ieeelongdouble -Wno-psabi -mno-gnu-attribute
 
-# Remove -mlong-double-128 because it does not work correctly with
-# -mabi=ibmlongdoule in some GCC versions, e.g. GCC 7.
-$(foreach o,$(all-object-suffixes),$(objpfx)ieee128-qefgcvt$(o)): \
-  sysdep-CFLAGS := $(filter-out -mlong-double-128, $(sysdep-CFLAGS))
-$(foreach o,$(all-object-suffixes),$(objpfx)ieee128-qefgcvt_r$(o)): \
-  sysdep-CFLAGS := $(filter-out -mlong-double-128, $(sysdep-CFLAGS))
-
 tests-internal += tst-ibm128-warn tst-ieee128-warn
 tests-internal += tst-ibm128-error tst-ieee128-error
 tests-internal += tst-ibm128-efgcvt tst-ieee128-efgcvt
@@ -348,11 +341,6 @@ ldbl-ibm128-files := $(objpfx)test-%-ibm128^ \
 		     $(foreach r,$(ldbl-extra-routines),$(objpfx)$(r)-internal^)
 obj-suf-foreach = $(foreach suf,$(all-object-suffixes),$(subst ^,$(suf),$(1)))
 
-# Remove -mlong-double-128 because it does not work correctly with
-# -mabi=ibmlongdoule in some GCC versions, e.g. GCC 7.
-$(call obj-suf-foreach,$(ldbl-ibm128-files)): \
-  sysdep-CFLAGS := $(filter-out -mlong-double-128, $(sysdep-CFLAGS))
-
 # Explicitly add -mabi=ibm-long-double to required files.
 $(call obj-suf-foreach,$(ldbl-ibm128-files)): \
   sysdep-CFLAGS += -mabi=ibmlongdouble
diff --git a/sysdeps/powerpc/powerpc64/le/configure b/sysdeps/powerpc/powerpc64/le/configure
index b08c3e0587..f7f0804b7f 100644
--- a/sysdeps/powerpc/powerpc64/le/configure
+++ b/sysdeps/powerpc/powerpc64/le/configure
@@ -90,6 +90,33 @@ if test "$libc_cv_compiler_powerpc64le_ice" != "yes"; then :
   critic_missing="$critic_missing __builtin_signbit is broken. GCC 7.4 or newer is required to resolve (BZ 83862)."
 fi
 
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking if $CC compiles with -mabi=ieeelongdouble and -mlong-double-128" >&5
+$as_echo_n "checking if $CC compiles with -mabi=ieeelongdouble and -mlong-double-128... " >&6; }
+if ${libc_cv_compiler_powerpc64le_ldbl128_mabi+:} false; then :
+  $as_echo_n "(cached) " >&6
+else
+  save_CFLAGS="$CFLAGS"
+CFLAGS="$CFLAGS -mabi=ieeelongdouble -mlong-double-128"
+cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h.  */
+
+long double x;
+
+_ACEOF
+if ac_fn_c_try_compile "$LINENO"; then :
+  libc_cv_compiler_powerpc64le_ldbl128_mabi=yes
+else
+  libc_cv_compiler_powerpc64le_ldbl128_mabi=no
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+CFLAGS="$save_CFLAGS"
+fi
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $libc_cv_compiler_powerpc64le_ldbl128_mabi" >&5
+$as_echo "$libc_cv_compiler_powerpc64le_ldbl128_mabi" >&6; }
+if test "$libc_cv_compiler_powerpc64le_ldbl128_mabi" == "no"; then :
+  critic_missing="$critic_missing The compiler must support -mabi=ieeelongdouble and -mlongdouble simultaneously."
+fi
+
 # Binutils (objcopy) 2.26 or newer required to support the --update-section
 # feature for fixing up .gnu.attribute section with IEEE ldbl.
 for ac_prog in $OBJCOPY
diff --git a/sysdeps/powerpc/powerpc64/le/configure.ac b/sysdeps/powerpc/powerpc64/le/configure.ac
index 630086bd88..dde956505c 100644
--- a/sysdeps/powerpc/powerpc64/le/configure.ac
+++ b/sysdeps/powerpc/powerpc64/le/configure.ac
@@ -51,6 +51,21 @@ CFLAGS="$save_CFLAGS"])
 AS_IF([test "$libc_cv_compiler_powerpc64le_ice" != "yes"],
       [critic_missing="$critic_missing __builtin_signbit is broken. GCC 7.4 or newer is required to resolve (BZ 83862)."])
 
+dnl Some old compiler versions give out error messages when combining
+dnl -mabi=% and -mlong-double-128.  i.e GCC 7.5.0 (BZ# 94200)
+AC_CACHE_CHECK([if $CC compiles with -mabi=ieeelongdouble and -mlong-double-128],
+	       libc_cv_compiler_powerpc64le_ldbl128_mabi, [dnl
+save_CFLAGS="$CFLAGS"
+CFLAGS="$CFLAGS -mabi=ieeelongdouble -mlong-double-128"
+AC_COMPILE_IFELSE([AC_LANG_SOURCE([[
+long double x;
+]])],
+		  [libc_cv_compiler_powerpc64le_ldbl128_mabi=yes],
+		  [libc_cv_compiler_powerpc64le_ldbl128_mabi=no])
+CFLAGS="$save_CFLAGS"])
+AS_IF([test "$libc_cv_compiler_powerpc64le_ldbl128_mabi" == "no"],
+      [critic_missing="$critic_missing The compiler must support -mabi=ieeelongdouble and -mlongdouble simultaneously."])
+
 # Binutils (objcopy) 2.26 or newer required to support the --update-section
 # feature for fixing up .gnu.attribute section with IEEE ldbl.
 AC_CHECK_PROG_VER(OBJCOPY, $OBJCOPY, --version,
-- 
2.21.1


^ permalink raw reply	[flat|nested] 22+ messages in thread

* [PATCH v3 6/6] powerpc64le: Enable support for IEEE long double
  2020-04-06 19:12 [PATCH v3 0/6] IEEE binary128 long double on powerpc64le Paul E. Murphy
                   ` (4 preceding siblings ...)
  2020-04-06 19:12 ` [PATCH v3 5/6] powerpc64le: blacklist broken GCC compilers (e.g GCC 7.5.0) Paul E. Murphy
@ 2020-04-06 19:12 ` Paul E. Murphy
  2020-04-22 15:03   ` Paul E Murphy
  2020-04-22 15:15   ` Florian Weimer
  5 siblings, 2 replies; 22+ messages in thread
From: Paul E. Murphy @ 2020-04-06 19:12 UTC (permalink / raw)
  To: libc-alpha

From: "Gabriel F. T. Gomes" <gabrielftg@linux.ibm.com>

Changes since v5:

  - Rebase against inline math test removal

Changes since v4:

  - More Makefile massaging
  - Add ULPs for ieee128 only _Float128 redirects
  - Update for glibc 2.32
  - Add short NEWS entry

Changes since v4:

  - Removed the redirections code (moved to a separate commit).

No changes since v3.

Changes since v2:

  - Added definition of LDBL_IBM128_COMPAT_VERSION and
    LDBL_IBM128_VERSION (moved from a previous commit).

Changes since v1:

  - Use __LONG_DOUBLE_USES_FLOAT128 directly.

-- 8< --
On platforms where long double may have two different formats, i.e.: the
same format as double (64-bits) or something else (128-bits), building
with -mlong-double-128 is the default and function calls in the user
program match the name of the function in Glibc.  When building with
-mlong-double-64, Glibc installed headers redirect such calls to the
appropriate function.

Likewise, the internals of glibc are now built against IEEE long double.
However, the only (minimally) notable usage of long double is difftime.
---
 NEWS                                          |   5 +
 sysdeps/powerpc/fpu/libm-test-ulps            |   4 +
 sysdeps/powerpc/powerpc64/le/Implies-before   |   1 +
 sysdeps/powerpc/powerpc64/le/Makefile         |   4 +
 .../powerpc64/le/ldbl-128ibm-compat-abi.h     |   8 ++
 .../linux/powerpc/powerpc64/le/libc.abilist   |  93 ++++++++++++++
 .../linux/powerpc/powerpc64/le/libm.abilist   | 117 ++++++++++++++++++
 7 files changed, 232 insertions(+)
 create mode 100644 sysdeps/unix/sysv/linux/powerpc/powerpc64/le/ldbl-128ibm-compat-abi.h

diff --git a/NEWS b/NEWS
index 07da1545b4..a54c0e0f65 100644
--- a/NEWS
+++ b/NEWS
@@ -14,6 +14,11 @@ Major new features:
 * The GNU C Library now loads audit modules listed in the DT_AUDIT and
   DT_DEPAUDIT dynamic section entries of the main executable.
 
+* powerpc64le supports IEEE128 long double libm/libc redirects when
+  using the -mabi=ieeelongdouble to compile C code on supported GCC
+  toolchains.  It is recommended to use GCC 8 or newer when testing
+  this option.
+
 Deprecated and removed features, and other changes affecting compatibility:
 
   [Add deprecations, removals and changes affecting compatibility here]
diff --git a/sysdeps/powerpc/fpu/libm-test-ulps b/sysdeps/powerpc/fpu/libm-test-ulps
index 73f18cf1df..695c5eee7e 100644
--- a/sysdeps/powerpc/fpu/libm-test-ulps
+++ b/sysdeps/powerpc/fpu/libm-test-ulps
@@ -1264,21 +1264,25 @@ ldouble: 1
 Function: "gamma":
 double: 3
 float: 4
+float128: 5
 ldouble: 3
 
 Function: "gamma_downward":
 double: 4
 float: 4
+float128: 8
 ldouble: 15
 
 Function: "gamma_towardzero":
 double: 4
 float: 3
+float128: 5
 ldouble: 16
 
 Function: "gamma_upward":
 double: 4
 float: 5
+float128: 8
 ldouble: 11
 
 Function: "hypot":
diff --git a/sysdeps/powerpc/powerpc64/le/Implies-before b/sysdeps/powerpc/powerpc64/le/Implies-before
index 7c20db4e97..2139f4dae8 100644
--- a/sysdeps/powerpc/powerpc64/le/Implies-before
+++ b/sysdeps/powerpc/powerpc64/le/Implies-before
@@ -1,4 +1,5 @@
 # On PowerPC we use the IBM extended long double format.
+ieee754/ldbl-128ibm-compat
 ieee754/ldbl-128ibm
 ieee754/ldbl-opt
 ieee754/dbl-64
diff --git a/sysdeps/powerpc/powerpc64/le/Makefile b/sysdeps/powerpc/powerpc64/le/Makefile
index 6afa1228dd..05ab13d85c 100644
--- a/sysdeps/powerpc/powerpc64/le/Makefile
+++ b/sysdeps/powerpc/powerpc64/le/Makefile
@@ -20,6 +20,10 @@ no-gnu-attribute-CFLAGS = -mno-gnu-attribute
 # We built an object file and extract the desired .gnu.attributes section
 # to be inserted into lib{c,m}.so.
 ifeq ($(build-shared),yes)
+
+# Build everything with IEEE 128-bit long double.
+sysdep-CFLAGS += -mabi=ieeelongdouble -Wno-psabi $(no-gnu-attribute-CFLAGS)
+
 before-compile += $(common-objpfx)no_ldbl_gnu_attribute.bin
 
 $(common-objpfx)no_ldbl_gnu_attribute.bin: $(..)sysdeps/powerpc/powerpc64/le/no_ldbl_gnu_attribute.c
diff --git a/sysdeps/unix/sysv/linux/powerpc/powerpc64/le/ldbl-128ibm-compat-abi.h b/sysdeps/unix/sysv/linux/powerpc/powerpc64/le/ldbl-128ibm-compat-abi.h
new file mode 100644
index 0000000000..34c7b6bed6
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/powerpc/powerpc64/le/ldbl-128ibm-compat-abi.h
@@ -0,0 +1,8 @@
+/* ABI version for long double switch to IEEE 128-bit floating point..
+   This is used by the Versions and math_ldbl_opt.h files in
+   sysdeps/ieee754/ldbl-128ibm-compat/.  It gives the ABI version where
+   long double == ibm128 was replaced with long double == _Float128
+   for libm *l functions and libc functions using long double.  */
+
+#define LDBL_IBM128_VERSION		GLIBC_2.32
+#define LDBL_IBM128_COMPAT_VERSION	GLIBC_2_32
diff --git a/sysdeps/unix/sysv/linux/powerpc/powerpc64/le/libc.abilist b/sysdeps/unix/sysv/linux/powerpc/powerpc64/le/libc.abilist
index 2229a1dcc0..c0ca0d6907 100644
--- a/sysdeps/unix/sysv/linux/powerpc/powerpc64/le/libc.abilist
+++ b/sysdeps/unix/sysv/linux/powerpc/powerpc64/le/libc.abilist
@@ -2247,3 +2247,96 @@ GLIBC_2.30 getdents64 F
 GLIBC_2.30 gettid F
 GLIBC_2.30 tgkill F
 GLIBC_2.30 twalk_r F
+GLIBC_2.32 __argp_errorieee128 F
+GLIBC_2.32 __argp_failureieee128 F
+GLIBC_2.32 __asprintf_chkieee128 F
+GLIBC_2.32 __asprintfieee128 F
+GLIBC_2.32 __dprintf_chkieee128 F
+GLIBC_2.32 __dprintfieee128 F
+GLIBC_2.32 __errieee128 F
+GLIBC_2.32 __error_at_lineieee128 F
+GLIBC_2.32 __errorieee128 F
+GLIBC_2.32 __errxieee128 F
+GLIBC_2.32 __fprintf_chkieee128 F
+GLIBC_2.32 __fprintfieee128 F
+GLIBC_2.32 __fscanfieee128 F
+GLIBC_2.32 __fwprintf_chkieee128 F
+GLIBC_2.32 __fwprintfieee128 F
+GLIBC_2.32 __fwscanfieee128 F
+GLIBC_2.32 __isoc99_fscanfieee128 F
+GLIBC_2.32 __isoc99_fwscanfieee128 F
+GLIBC_2.32 __isoc99_scanfieee128 F
+GLIBC_2.32 __isoc99_sscanfieee128 F
+GLIBC_2.32 __isoc99_swscanfieee128 F
+GLIBC_2.32 __isoc99_vfscanfieee128 F
+GLIBC_2.32 __isoc99_vfwscanfieee128 F
+GLIBC_2.32 __isoc99_vscanfieee128 F
+GLIBC_2.32 __isoc99_vsscanfieee128 F
+GLIBC_2.32 __isoc99_vswscanfieee128 F
+GLIBC_2.32 __isoc99_vwscanfieee128 F
+GLIBC_2.32 __isoc99_wscanfieee128 F
+GLIBC_2.32 __obstack_printf_chkieee128 F
+GLIBC_2.32 __obstack_printfieee128 F
+GLIBC_2.32 __obstack_vprintf_chkieee128 F
+GLIBC_2.32 __obstack_vprintfieee128 F
+GLIBC_2.32 __printf_chkieee128 F
+GLIBC_2.32 __printf_sizeieee128 F
+GLIBC_2.32 __printfieee128 F
+GLIBC_2.32 __qecvtieee128 F
+GLIBC_2.32 __qecvtieee128_r F
+GLIBC_2.32 __qfcvtieee128 F
+GLIBC_2.32 __qfcvtieee128_r F
+GLIBC_2.32 __qgcvtieee128 F
+GLIBC_2.32 __scanfieee128 F
+GLIBC_2.32 __snprintf_chkieee128 F
+GLIBC_2.32 __snprintfieee128 F
+GLIBC_2.32 __sprintf_chkieee128 F
+GLIBC_2.32 __sprintfieee128 F
+GLIBC_2.32 __sscanfieee128 F
+GLIBC_2.32 __strfmon_lieee128 F
+GLIBC_2.32 __strfmonieee128 F
+GLIBC_2.32 __strfromieee128 F
+GLIBC_2.32 __strtoieee128 F
+GLIBC_2.32 __strtoieee128_l F
+GLIBC_2.32 __swprintf_chkieee128 F
+GLIBC_2.32 __swprintfieee128 F
+GLIBC_2.32 __swscanfieee128 F
+GLIBC_2.32 __syslog_chkieee128 F
+GLIBC_2.32 __syslogieee128 F
+GLIBC_2.32 __vasprintf_chkieee128 F
+GLIBC_2.32 __vasprintfieee128 F
+GLIBC_2.32 __vdprintf_chkieee128 F
+GLIBC_2.32 __vdprintfieee128 F
+GLIBC_2.32 __verrieee128 F
+GLIBC_2.32 __verrxieee128 F
+GLIBC_2.32 __vfprintf_chkieee128 F
+GLIBC_2.32 __vfprintfieee128 F
+GLIBC_2.32 __vfscanfieee128 F
+GLIBC_2.32 __vfwprintf_chkieee128 F
+GLIBC_2.32 __vfwprintfieee128 F
+GLIBC_2.32 __vfwscanfieee128 F
+GLIBC_2.32 __vprintf_chkieee128 F
+GLIBC_2.32 __vprintfieee128 F
+GLIBC_2.32 __vscanfieee128 F
+GLIBC_2.32 __vsnprintf_chkieee128 F
+GLIBC_2.32 __vsnprintfieee128 F
+GLIBC_2.32 __vsprintf_chkieee128 F
+GLIBC_2.32 __vsprintfieee128 F
+GLIBC_2.32 __vsscanfieee128 F
+GLIBC_2.32 __vswprintf_chkieee128 F
+GLIBC_2.32 __vswprintfieee128 F
+GLIBC_2.32 __vswscanfieee128 F
+GLIBC_2.32 __vsyslog_chkieee128 F
+GLIBC_2.32 __vsyslogieee128 F
+GLIBC_2.32 __vwarnieee128 F
+GLIBC_2.32 __vwarnxieee128 F
+GLIBC_2.32 __vwprintf_chkieee128 F
+GLIBC_2.32 __vwprintfieee128 F
+GLIBC_2.32 __vwscanfieee128 F
+GLIBC_2.32 __warnieee128 F
+GLIBC_2.32 __warnxieee128 F
+GLIBC_2.32 __wcstoieee128 F
+GLIBC_2.32 __wcstoieee128_l F
+GLIBC_2.32 __wprintf_chkieee128 F
+GLIBC_2.32 __wprintfieee128 F
+GLIBC_2.32 __wscanfieee128 F
diff --git a/sysdeps/unix/sysv/linux/powerpc/powerpc64/le/libm.abilist b/sysdeps/unix/sysv/linux/powerpc/powerpc64/le/libm.abilist
index d479a64fca..aa45e2639c 100644
--- a/sysdeps/unix/sysv/linux/powerpc/powerpc64/le/libm.abilist
+++ b/sysdeps/unix/sysv/linux/powerpc/powerpc64/le/libm.abilist
@@ -1097,3 +1097,120 @@ GLIBC_2.31 totalordermagf32x F
 GLIBC_2.31 totalordermagf64 F
 GLIBC_2.31 totalordermagf64x F
 GLIBC_2.31 totalordermagl F
+GLIBC_2.32 __acoshieee128 F
+GLIBC_2.32 __acosieee128 F
+GLIBC_2.32 __asinhieee128 F
+GLIBC_2.32 __asinieee128 F
+GLIBC_2.32 __atan2ieee128 F
+GLIBC_2.32 __atanhieee128 F
+GLIBC_2.32 __atanieee128 F
+GLIBC_2.32 __cabsieee128 F
+GLIBC_2.32 __cacoshieee128 F
+GLIBC_2.32 __cacosieee128 F
+GLIBC_2.32 __canonicalizeieee128 F
+GLIBC_2.32 __cargieee128 F
+GLIBC_2.32 __casinhieee128 F
+GLIBC_2.32 __casinieee128 F
+GLIBC_2.32 __catanhieee128 F
+GLIBC_2.32 __catanieee128 F
+GLIBC_2.32 __cbrtieee128 F
+GLIBC_2.32 __ccoshieee128 F
+GLIBC_2.32 __ccosieee128 F
+GLIBC_2.32 __ceilieee128 F
+GLIBC_2.32 __cexpieee128 F
+GLIBC_2.32 __cimagieee128 F
+GLIBC_2.32 __clog10ieee128 F
+GLIBC_2.32 __clogieee128 F
+GLIBC_2.32 __conjieee128 F
+GLIBC_2.32 __copysignieee128 F
+GLIBC_2.32 __coshieee128 F
+GLIBC_2.32 __cosieee128 F
+GLIBC_2.32 __cpowieee128 F
+GLIBC_2.32 __cprojieee128 F
+GLIBC_2.32 __crealieee128 F
+GLIBC_2.32 __csinhieee128 F
+GLIBC_2.32 __csinieee128 F
+GLIBC_2.32 __csqrtieee128 F
+GLIBC_2.32 __ctanhieee128 F
+GLIBC_2.32 __ctanieee128 F
+GLIBC_2.32 __erfcieee128 F
+GLIBC_2.32 __erfieee128 F
+GLIBC_2.32 __exp10ieee128 F
+GLIBC_2.32 __exp2ieee128 F
+GLIBC_2.32 __expieee128 F
+GLIBC_2.32 __expm1ieee128 F
+GLIBC_2.32 __f32addieee128 F
+GLIBC_2.32 __f32divieee128 F
+GLIBC_2.32 __f32mulieee128 F
+GLIBC_2.32 __f32subieee128 F
+GLIBC_2.32 __f64addieee128 F
+GLIBC_2.32 __f64divieee128 F
+GLIBC_2.32 __f64mulieee128 F
+GLIBC_2.32 __f64subieee128 F
+GLIBC_2.32 __fabsieee128 F
+GLIBC_2.32 __fdimieee128 F
+GLIBC_2.32 __floorieee128 F
+GLIBC_2.32 __fmaieee128 F
+GLIBC_2.32 __fmaxieee128 F
+GLIBC_2.32 __fmaxmagieee128 F
+GLIBC_2.32 __fminieee128 F
+GLIBC_2.32 __fminmagieee128 F
+GLIBC_2.32 __fmodieee128 F
+GLIBC_2.32 __frexpieee128 F
+GLIBC_2.32 __fromfpieee128 F
+GLIBC_2.32 __fromfpxieee128 F
+GLIBC_2.32 __getpayloadieee128 F
+GLIBC_2.32 __hypotieee128 F
+GLIBC_2.32 __ilogbieee128 F
+GLIBC_2.32 __j0ieee128 F
+GLIBC_2.32 __j1ieee128 F
+GLIBC_2.32 __jnieee128 F
+GLIBC_2.32 __ldexpieee128 F
+GLIBC_2.32 __lgammaieee128 F
+GLIBC_2.32 __lgammaieee128_r F
+GLIBC_2.32 __llogbieee128 F
+GLIBC_2.32 __llrintieee128 F
+GLIBC_2.32 __llroundieee128 F
+GLIBC_2.32 __log10ieee128 F
+GLIBC_2.32 __log1pieee128 F
+GLIBC_2.32 __log2ieee128 F
+GLIBC_2.32 __logbieee128 F
+GLIBC_2.32 __logieee128 F
+GLIBC_2.32 __lrintieee128 F
+GLIBC_2.32 __lroundieee128 F
+GLIBC_2.32 __modfieee128 F
+GLIBC_2.32 __nanieee128 F
+GLIBC_2.32 __nearbyintieee128 F
+GLIBC_2.32 __nextafterieee128 F
+GLIBC_2.32 __nextdownieee128 F
+GLIBC_2.32 __nexttoward_to_ieee128 F
+GLIBC_2.32 __nexttowardf_to_ieee128 F
+GLIBC_2.32 __nexttowardieee128 F
+GLIBC_2.32 __nextupieee128 F
+GLIBC_2.32 __powieee128 F
+GLIBC_2.32 __remainderieee128 F
+GLIBC_2.32 __remquoieee128 F
+GLIBC_2.32 __rintieee128 F
+GLIBC_2.32 __roundevenieee128 F
+GLIBC_2.32 __roundieee128 F
+GLIBC_2.32 __scalbieee128 F
+GLIBC_2.32 __scalblnieee128 F
+GLIBC_2.32 __scalbnieee128 F
+GLIBC_2.32 __setpayloadieee128 F
+GLIBC_2.32 __setpayloadsigieee128 F
+GLIBC_2.32 __significandieee128 F
+GLIBC_2.32 __sincosieee128 F
+GLIBC_2.32 __sinhieee128 F
+GLIBC_2.32 __sinieee128 F
+GLIBC_2.32 __sqrtieee128 F
+GLIBC_2.32 __tanhieee128 F
+GLIBC_2.32 __tanieee128 F
+GLIBC_2.32 __tgammaieee128 F
+GLIBC_2.32 __totalorderieee128 F
+GLIBC_2.32 __totalordermagieee128 F
+GLIBC_2.32 __truncieee128 F
+GLIBC_2.32 __ufromfpieee128 F
+GLIBC_2.32 __ufromfpxieee128 F
+GLIBC_2.32 __y0ieee128 F
+GLIBC_2.32 __y1ieee128 F
+GLIBC_2.32 __ynieee128 F
-- 
2.21.1


^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731
  2020-04-06 19:12 ` [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731 Paul E. Murphy
@ 2020-04-15 14:20   ` Paul E Murphy
  2020-04-22 14:58     ` Paul E Murphy
  2020-04-22 15:11   ` Florian Weimer
  1 sibling, 1 reply; 22+ messages in thread
From: Paul E Murphy @ 2020-04-15 14:20 UTC (permalink / raw)
  To: libc-alpha



On 4/6/20 2:12 PM, Paul E. Murphy via Libc-alpha wrote:
> GCC 9 has a bug (BZ 90731) whereby __typeof does not correctly copy
> exception specifiers[1].  Surprisingly, this can be quieted by declaring
> "#pragma system_header", or if the headers are installed in a system
> directory.
> 
> Work around this by using the pragma for any gcc version between
> 9.0 and 9.2 to ensure tests continue to compile.
> 
> [1] Example error from g++ 9.2.1:
> 
> In file included from ../include/sys/cdefs.h:3,
>                   from ../include/features.h:465,
>                   from ../bits/libc-header-start.h:33,
>                   from ../math/math.h:27,
>                   from ../include/math.h:7,
>                   from test-math-isinff.cc:21:
> ../libio/bits/stdio-ldbl.h:25:20: error: declaration of ‘int sprintf(char*, const char*, ...)’ has a different exception specifier
>     25 | __LDBL_REDIR_DECL (sprintf)
>        |                    ^~~~~~~
> ../misc/sys/cdefs.h:461:26: note: in definition of macro ‘__LDBL_REDIR_DECL’
>    461 |   extern __typeof (name) name __asm (__ASMNAME ("__" #name "ieee128"));
>        |                          ^~~~
> In file included from ../include/stdio.h:5,
>                   from test-math-isinff.cc:22:
> ../libio/stdio.h:334:12: note: from previous declaration ‘int sprintf(char*, const char*, ...) throw ()’
>    334 | extern int sprintf (char *__restrict __s,
>        |            ^~~~~~~
> ---
>   include/monetary.h |  8 ++++++++
>   include/printf.h   |  8 ++++++++
>   include/stdio.h    | 10 ++++++++++
>   include/stdlib.h   |  9 +++++++++
>   include/wchar.h    |  9 +++++++++
>   5 files changed, 44 insertions(+)

Ping?

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 2/6] Rename __LONG_DOUBLE_USES_FLOAT128 to __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI
  2020-04-06 19:12 ` [PATCH v3 2/6] Rename __LONG_DOUBLE_USES_FLOAT128 to __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI Paul E. Murphy
@ 2020-04-15 14:21   ` Paul E Murphy
  2020-04-22 14:59     ` Paul E Murphy
  0 siblings, 1 reply; 22+ messages in thread
From: Paul E Murphy @ 2020-04-15 14:21 UTC (permalink / raw)
  To: libc-alpha



On 4/6/20 2:12 PM, Paul E. Murphy via Libc-alpha wrote:
> per feedback and disucssion on libc-alpha [2].  Likewise, add
> documentation to the macro in the common bits/long-double.h header.
> This is effectively V2 of [2].
> 
> Per the feedback from Joseph [1].  Good comments and a more
> self-explanatory macro name will be very helpful when this
> macro is able to assume a non-zero value.
> 
> [1] <https://sourceware.org/legacy-ml/libc-alpha/2020-02/msg00687.html>
> [2] <https://sourceware.org/pipermail/libc-alpha/2020-February/111148.html>
> 
> ---8<---
> 
> Improve the commentary to aid future developers who will stumble
> upon this novel, yet not always perfect, mechanism to support
> alternative formats for long double.
> 
> Likewise, rename __LONG_DOUBLE_USES_FLOAT128 to
> __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI now that development work
> has settled down.  The command used was
> 
> git grep -l __LONG_DOUBLE_USES_FLOAT128 ':!./ChangeLog*' | \
>    xargs sed -i 's/__LONG_DOUBLE_USES_FLOAT128/__LDOUBLE_REDIRECTS_TO_FLOAT128_ABI/g'
> ---
>   argp/argp.h                                       |  2 +-
>   bits/long-double.h                                | 15 ++++++++++++++-
>   elf/tst-addr1.c                                   |  2 +-
>   include/math.h                                    |  6 +++---
>   include/monetary.h                                |  2 +-
>   include/printf.h                                  |  2 +-
>   include/stdio.h                                   |  8 ++++----
>   include/stdlib.h                                  |  4 ++--
>   include/sys/cdefs.h                               |  2 +-
>   include/wchar.h                                   |  2 +-
>   libio/bits/stdio-ldbl.h                           |  4 ++--
>   libio/stdio.h                                     |  8 ++++----
>   math/complex.h                                    |  4 ++--
>   math/math.h                                       | 12 ++++++------
>   math/test-ldouble.h                               |  2 +-
>   misc/err.h                                        |  2 +-
>   misc/error.h                                      |  2 +-
>   misc/sys/cdefs.h                                  |  4 ++--
>   misc/sys/syslog.h                                 |  2 +-
>   stdio-common/printf.h                             |  2 +-
>   stdio-common/tst-vfprintf-user-type.c             |  2 +-
>   stdlib/bits/stdlib-ldbl.h                         |  2 +-
>   stdlib/monetary.h                                 |  2 +-
>   stdlib/stdlib.h                                   |  2 +-
>   sysdeps/ieee754/ldbl-128/bits/long-double.h       |  2 +-
>   .../ieee754/ldbl-128ibm-compat/bits/long-double.h |  7 ++-----
>   sysdeps/ieee754/ldbl-128ibm-compat/math_ldbl.h    |  2 +-
>   sysdeps/ieee754/ldbl-128ibm/bits/iscanonical.h    |  2 +-
>   sysdeps/ieee754/ldbl-128ibm/ieee754.h             |  6 +++---
>   .../ldbl-128ibm/include/bits/iscanonical.h        |  2 +-
>   sysdeps/ieee754/ldbl-96/bits/long-double.h        |  2 +-
>   sysdeps/ieee754/ldbl-opt/bits/long-double.h       |  2 +-
>   sysdeps/mips/ieee754/bits/long-double.h           |  2 +-
>   .../sysv/linux/sparc/sparc32/bits/long-double.h   |  2 +-
>   .../sysv/linux/sparc/sparc64/bits/long-double.h   |  2 +-
>   wcsmbs/bits/wchar-ldbl.h                          |  4 ++--
>   wcsmbs/wchar.h                                    |  8 ++++----
>   37 files changed, 74 insertions(+), 64 deletions(-)

Ping?

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731
  2020-04-15 14:20   ` Paul E Murphy
@ 2020-04-22 14:58     ` Paul E Murphy
  0 siblings, 0 replies; 22+ messages in thread
From: Paul E Murphy @ 2020-04-22 14:58 UTC (permalink / raw)
  To: libc-alpha



On 4/15/20 9:20 AM, Paul E Murphy via Libc-alpha wrote:
> 
> 
> On 4/6/20 2:12 PM, Paul E. Murphy via Libc-alpha wrote:
>> GCC 9 has a bug (BZ 90731) whereby __typeof does not correctly copy
>> exception specifiers[1].  Surprisingly, this can be quieted by declaring
>> "#pragma system_header", or if the headers are installed in a system
>> directory.
>>
>> Work around this by using the pragma for any gcc version between
>> 9.0 and 9.2 to ensure tests continue to compile.
>>
>> [1] Example error from g++ 9.2.1:
>>
>> In file included from ../include/sys/cdefs.h:3,
>>                   from ../include/features.h:465,
>>                   from ../bits/libc-header-start.h:33,
>>                   from ../math/math.h:27,
>>                   from ../include/math.h:7,
>>                   from test-math-isinff.cc:21:
>> ../libio/bits/stdio-ldbl.h:25:20: error: declaration of ‘int 
>> sprintf(char*, const char*, ...)’ has a different exception specifier
>>     25 | __LDBL_REDIR_DECL (sprintf)
>>        |                    ^~~~~~~
>> ../misc/sys/cdefs.h:461:26: note: in definition of macro 
>> ‘__LDBL_REDIR_DECL’
>>    461 |   extern __typeof (name) name __asm (__ASMNAME ("__" #name 
>> "ieee128"));
>>        |                          ^~~~
>> In file included from ../include/stdio.h:5,
>>                   from test-math-isinff.cc:22:
>> ../libio/stdio.h:334:12: note: from previous declaration ‘int 
>> sprintf(char*, const char*, ...) throw ()’
>>    334 | extern int sprintf (char *__restrict __s,
>>        |            ^~~~~~~
>> ---
>>   include/monetary.h |  8 ++++++++
>>   include/printf.h   |  8 ++++++++
>>   include/stdio.h    | 10 ++++++++++
>>   include/stdlib.h   |  9 +++++++++
>>   include/wchar.h    |  9 +++++++++
>>   5 files changed, 44 insertions(+)
> 
> Ping?

Ping. I will commit this friday pending no additional feedback.

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 2/6] Rename __LONG_DOUBLE_USES_FLOAT128 to __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI
  2020-04-15 14:21   ` Paul E Murphy
@ 2020-04-22 14:59     ` Paul E Murphy
  0 siblings, 0 replies; 22+ messages in thread
From: Paul E Murphy @ 2020-04-22 14:59 UTC (permalink / raw)
  To: libc-alpha



On 4/15/20 9:21 AM, Paul E Murphy via Libc-alpha wrote:
> 
> 
> On 4/6/20 2:12 PM, Paul E. Murphy via Libc-alpha wrote:
>> per feedback and disucssion on libc-alpha [2].  Likewise, add
>> documentation to the macro in the common bits/long-double.h header.
>> This is effectively V2 of [2].
>>
>> Per the feedback from Joseph [1].  Good comments and a more
>> self-explanatory macro name will be very helpful when this
>> macro is able to assume a non-zero value.
>>
>> [1] <https://sourceware.org/legacy-ml/libc-alpha/2020-02/msg00687.html>
>> [2] 
>> <https://sourceware.org/pipermail/libc-alpha/2020-February/111148.html>
>>
>> ---8<---
>>
>> Improve the commentary to aid future developers who will stumble
>> upon this novel, yet not always perfect, mechanism to support
>> alternative formats for long double.
>>
>> Likewise, rename __LONG_DOUBLE_USES_FLOAT128 to
>> __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI now that development work
>> has settled down.  The command used was
>>
>> git grep -l __LONG_DOUBLE_USES_FLOAT128 ':!./ChangeLog*' | \
>>    xargs sed -i 
>> 's/__LONG_DOUBLE_USES_FLOAT128/__LDOUBLE_REDIRECTS_TO_FLOAT128_ABI/g'
>> ---
>>   argp/argp.h                                       |  2 +-
>>   bits/long-double.h                                | 15 ++++++++++++++-
>>   elf/tst-addr1.c                                   |  2 +-
>>   include/math.h                                    |  6 +++---
>>   include/monetary.h                                |  2 +-
>>   include/printf.h                                  |  2 +-
>>   include/stdio.h                                   |  8 ++++----
>>   include/stdlib.h                                  |  4 ++--
>>   include/sys/cdefs.h                               |  2 +-
>>   include/wchar.h                                   |  2 +-
>>   libio/bits/stdio-ldbl.h                           |  4 ++--
>>   libio/stdio.h                                     |  8 ++++----
>>   math/complex.h                                    |  4 ++--
>>   math/math.h                                       | 12 ++++++------
>>   math/test-ldouble.h                               |  2 +-
>>   misc/err.h                                        |  2 +-
>>   misc/error.h                                      |  2 +-
>>   misc/sys/cdefs.h                                  |  4 ++--
>>   misc/sys/syslog.h                                 |  2 +-
>>   stdio-common/printf.h                             |  2 +-
>>   stdio-common/tst-vfprintf-user-type.c             |  2 +-
>>   stdlib/bits/stdlib-ldbl.h                         |  2 +-
>>   stdlib/monetary.h                                 |  2 +-
>>   stdlib/stdlib.h                                   |  2 +-
>>   sysdeps/ieee754/ldbl-128/bits/long-double.h       |  2 +-
>>   .../ieee754/ldbl-128ibm-compat/bits/long-double.h |  7 ++-----
>>   sysdeps/ieee754/ldbl-128ibm-compat/math_ldbl.h    |  2 +-
>>   sysdeps/ieee754/ldbl-128ibm/bits/iscanonical.h    |  2 +-
>>   sysdeps/ieee754/ldbl-128ibm/ieee754.h             |  6 +++---
>>   .../ldbl-128ibm/include/bits/iscanonical.h        |  2 +-
>>   sysdeps/ieee754/ldbl-96/bits/long-double.h        |  2 +-
>>   sysdeps/ieee754/ldbl-opt/bits/long-double.h       |  2 +-
>>   sysdeps/mips/ieee754/bits/long-double.h           |  2 +-
>>   .../sysv/linux/sparc/sparc32/bits/long-double.h   |  2 +-
>>   .../sysv/linux/sparc/sparc64/bits/long-double.h   |  2 +-
>>   wcsmbs/bits/wchar-ldbl.h                          |  4 ++--
>>   wcsmbs/wchar.h                                    |  8 ++++----
>>   37 files changed, 74 insertions(+), 64 deletions(-)
> 
> Ping?

Ping. I will commit this friday pending no additional feedback.

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 6/6] powerpc64le: Enable support for IEEE long double
  2020-04-06 19:12 ` [PATCH v3 6/6] powerpc64le: Enable support for IEEE long double Paul E. Murphy
@ 2020-04-22 15:03   ` Paul E Murphy
  2020-04-22 15:15   ` Florian Weimer
  1 sibling, 0 replies; 22+ messages in thread
From: Paul E Murphy @ 2020-04-22 15:03 UTC (permalink / raw)
  To: Paul E. Murphy, libc-alpha



On 4/6/20 2:12 PM, Paul E. Murphy via Libc-alpha wrote:
> From: "Gabriel F. T. Gomes" <gabrielftg@linux.ibm.com>
> 
> Changes since v5:
> 
>    - Rebase against inline math test removal
> 
> Changes since v4:
> 
>    - More Makefile massaging
>    - Add ULPs for ieee128 only _Float128 redirects
>    - Update for glibc 2.32
>    - Add short NEWS entry
> 
> Changes since v4:
> 
>    - Removed the redirections code (moved to a separate commit).
> 
> No changes since v3.
> 
> Changes since v2:
> 
>    - Added definition of LDBL_IBM128_COMPAT_VERSION and
>      LDBL_IBM128_VERSION (moved from a previous commit).
> 
> Changes since v1:
> 
>    - Use __LONG_DOUBLE_USES_FLOAT128 directly.
> 
> -- 8< --
> On platforms where long double may have two different formats, i.e.: the
> same format as double (64-bits) or something else (128-bits), building
> with -mlong-double-128 is the default and function calls in the user
> program match the name of the function in Glibc.  When building with
> -mlong-double-64, Glibc installed headers redirect such calls to the
> appropriate function.
> 
> Likewise, the internals of glibc are now built against IEEE long double.
> However, the only (minimally) notable usage of long double is difftime.
> ---
>   NEWS                                          |   5 +
>   sysdeps/powerpc/fpu/libm-test-ulps            |   4 +
>   sysdeps/powerpc/powerpc64/le/Implies-before   |   1 +
>   sysdeps/powerpc/powerpc64/le/Makefile         |   4 +
>   .../powerpc64/le/ldbl-128ibm-compat-abi.h     |   8 ++
>   .../linux/powerpc/powerpc64/le/libc.abilist   |  93 ++++++++++++++
>   .../linux/powerpc/powerpc64/le/libm.abilist   | 117 ++++++++++++++++++
>   7 files changed, 232 insertions(+)
>   create mode 100644 sysdeps/unix/sysv/linux/powerpc/powerpc64/le/ldbl-128ibm-compat-abi.h
> 
> diff --git a/NEWS b/NEWS
> index 07da1545b4..a54c0e0f65 100644

I will commit the remaining ppc64le specific enablement patches Friday 
if there are no further requests for improvement.

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731
  2020-04-06 19:12 ` [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731 Paul E. Murphy
  2020-04-15 14:20   ` Paul E Murphy
@ 2020-04-22 15:11   ` Florian Weimer
  2020-04-22 21:17     ` Paul E Murphy
  1 sibling, 1 reply; 22+ messages in thread
From: Florian Weimer @ 2020-04-22 15:11 UTC (permalink / raw)
  To: Paul E. Murphy via Libc-alpha

* Paul E. Murphy via Libc-alpha:

> GCC 9 has a bug (BZ 90731) whereby __typeof does not correctly copy
> exception specifiers[1].  Surprisingly, this can be quieted by declaring
> "#pragma system_header", or if the headers are installed in a system
> directory.

I think GCC developers prefer “PR90731” to their bugs.  This occurs
multiple times in the commit message and throughout the patch.

This change will still cause errors with -Wsystem-headers in the build
flags, right?  I think we should mention this somewhere, maybe on the
release wiki?

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 6/6] powerpc64le: Enable support for IEEE long double
  2020-04-06 19:12 ` [PATCH v3 6/6] powerpc64le: Enable support for IEEE long double Paul E. Murphy
  2020-04-22 15:03   ` Paul E Murphy
@ 2020-04-22 15:15   ` Florian Weimer
  2020-04-22 16:20     ` Paul E Murphy
  1 sibling, 1 reply; 22+ messages in thread
From: Florian Weimer @ 2020-04-22 15:15 UTC (permalink / raw)
  To: Paul E. Murphy via Libc-alpha

* Paul E. Murphy via Libc-alpha:

> diff --git a/NEWS b/NEWS
> index 07da1545b4..a54c0e0f65 100644
> --- a/NEWS
> +++ b/NEWS
> @@ -14,6 +14,11 @@ Major new features:
>  * The GNU C Library now loads audit modules listed in the DT_AUDIT and
>    DT_DEPAUDIT dynamic section entries of the main executable.
>  
> +* powerpc64le supports IEEE128 long double libm/libc redirects when
> +  using the -mabi=ieeelongdouble to compile C code on supported GCC
> +  toolchains.  It is recommended to use GCC 8 or newer when testing
> +  this option.

Why do you write “testing this option”?

To me, this implies that the feature is not ready yet.  Is that the
case?

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 3/6] powerpc64le: raise GCC requirement to 7.4 for long double transition
  2020-04-06 19:12 ` [PATCH v3 3/6] powerpc64le: raise GCC requirement to 7.4 for long double transition Paul E. Murphy
@ 2020-04-22 15:19   ` Florian Weimer
  2020-04-22 19:33     ` Paul E Murphy
  0 siblings, 1 reply; 22+ messages in thread
From: Florian Weimer @ 2020-04-22 15:19 UTC (permalink / raw)
  To: Paul E. Murphy via Libc-alpha

* Paul E. Murphy via Libc-alpha:

> diff --git a/manual/install.texi b/manual/install.texi
> index 71bf47cac6..66de8ea59b 100644
> --- a/manual/install.texi
> +++ b/manual/install.texi
> @@ -507,6 +507,11 @@ the newest version of the compiler that is known to work for building
>  release time, GCC 9.2.1 is the newest compiler verified to work to build
>  @theglibc{}.
>  
> +For PowerPC 64-bits little-endian (powerpc64le), a GCC version with
> +support for @option{-mno-gnu-attribute} is required.  As of release time,
> +this means GCC 7.4 (excepting 7.5) or newer. The compiler option is required
> +for building the GNU C Library with support for IEEE long double.

Why “excepting 7.5”?  Somewhere, there should be an explanation, maybe
in a @c comment in the Texinfo sources if it's too confusing to end
users.

Also, missing two spaces at the end of sentences.

And quick question for my understanding: Could we get the effect of
-mno-gnu-attribute using strip if we wanted?  (Not that I think this
change is required, I'm fine with bumping the minimum GCC
requirement.)

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 4/6] powerpc64le: bump binutils version requirement to >= 2.26
  2020-04-06 19:12 ` [PATCH v3 4/6] powerpc64le: bump binutils version requirement to >= 2.26 Paul E. Murphy
@ 2020-04-22 15:20   ` Florian Weimer
  0 siblings, 0 replies; 22+ messages in thread
From: Florian Weimer @ 2020-04-22 15:20 UTC (permalink / raw)
  To: Paul E. Murphy via Libc-alpha

* Paul E. Murphy via Libc-alpha:

> diff --git a/manual/install.texi b/manual/install.texi
> index 66de8ea59b..acbcc8c091 100644
> --- a/manual/install.texi
> +++ b/manual/install.texi
> @@ -533,6 +533,10 @@ No other assembler or linker has the necessary functionality at the
>  moment. As of release time, GNU @code{binutils} 2.32 is the newest
>  verified to work to build @theglibc{}.
>  
> +For PowerPC 64-bits little-endian (powerpc64le), binutils is required
> +to support the @option{--update-section}.  This option requires binutils
> +2.26 or newer.

I think it's @command{objcopy} that needs to support the option,
right?  So perhaps replace the first “binutils” with that?

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 5/6] powerpc64le: blacklist broken GCC compilers (e.g GCC 7.5.0)
  2020-04-06 19:12 ` [PATCH v3 5/6] powerpc64le: blacklist broken GCC compilers (e.g GCC 7.5.0) Paul E. Murphy
@ 2020-04-22 15:23   ` Florian Weimer
  0 siblings, 0 replies; 22+ messages in thread
From: Florian Weimer @ 2020-04-22 15:23 UTC (permalink / raw)
  To: Paul E. Murphy via Libc-alpha

* Paul E. Murphy via Libc-alpha:

> GCC 7.5.0 (BZ# 94200) will refuse to compile if both -mabi=% and
> -mlong-double-128 are passed on the command line.  Surprisingly,
> it will work happily if the latter is not.  For the sake of
> maintianing status quo, test for and blacklist such compilers.

GCC PR reference syntax (as mentioned before), and typo: maintianing

> diff --git a/sysdeps/powerpc/powerpc64/le/configure.ac b/sysdeps/powerpc/powerpc64/le/configure.ac
> index 630086bd88..dde956505c 100644
> --- a/sysdeps/powerpc/powerpc64/le/configure.ac
> +++ b/sysdeps/powerpc/powerpc64/le/configure.ac
> @@ -51,6 +51,21 @@ CFLAGS="$save_CFLAGS"])
>  AS_IF([test "$libc_cv_compiler_powerpc64le_ice" != "yes"],
>        [critic_missing="$critic_missing __builtin_signbit is broken. GCC 7.4 or newer is required to resolve (BZ 83862)."])
>  
> +dnl Some old compiler versions give out error messages when combining
> +dnl -mabi=% and -mlong-double-128.  i.e GCC 7.5.0 (BZ# 94200)

PR reference.

> +AC_CACHE_CHECK([if $CC compiles with -mabi=ieeelongdouble and -mlong-double-128],
> +	       libc_cv_compiler_powerpc64le_ldbl128_mabi, [dnl
> +save_CFLAGS="$CFLAGS"
> +CFLAGS="$CFLAGS -mabi=ieeelongdouble -mlong-double-128"
> +AC_COMPILE_IFELSE([AC_LANG_SOURCE([[
> +long double x;
> +]])],
> +		  [libc_cv_compiler_powerpc64le_ldbl128_mabi=yes],
> +		  [libc_cv_compiler_powerpc64le_ldbl128_mabi=no])
> +CFLAGS="$save_CFLAGS"])
> +AS_IF([test "$libc_cv_compiler_powerpc64le_ldbl128_mabi" == "no"],
> +      [critic_missing="$critic_missing The compiler must support -mabi=ieeelongdouble and -mlongdouble simultaneously."])
> +
>  # Binutils (objcopy) 2.26 or newer required to support the --update-section
>  # feature for fixing up .gnu.attribute section with IEEE ldbl.
>  AC_CHECK_PROG_VER(OBJCOPY, $OBJCOPY, --version,

This change looks reasonable to me.

Maybe add the @c comment to the install.texi file with this patch?
Something like this:

@c See this autoconf check:
@c if $CC compiles with -mabi=ieeelongdouble and -mlong-double-128

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 6/6] powerpc64le: Enable support for IEEE long double
  2020-04-22 15:15   ` Florian Weimer
@ 2020-04-22 16:20     ` Paul E Murphy
  2020-04-22 16:23       ` Florian Weimer
  0 siblings, 1 reply; 22+ messages in thread
From: Paul E Murphy @ 2020-04-22 16:20 UTC (permalink / raw)
  To: Florian Weimer, Paul E. Murphy via Libc-alpha



On 4/22/20 10:15 AM, Florian Weimer wrote:
> * Paul E. Murphy via Libc-alpha:
> 
>> diff --git a/NEWS b/NEWS
>> index 07da1545b4..a54c0e0f65 100644
>> --- a/NEWS
>> +++ b/NEWS
>> @@ -14,6 +14,11 @@ Major new features:
>>   * The GNU C Library now loads audit modules listed in the DT_AUDIT and
>>     DT_DEPAUDIT dynamic section entries of the main executable.
>>   
>> +* powerpc64le supports IEEE128 long double libm/libc redirects when
>> +  using the -mabi=ieeelongdouble to compile C code on supported GCC
>> +  toolchains.  It is recommended to use GCC 8 or newer when testing
>> +  this option.
> 
> Why do you write “testing this option”?
> 
> To me, this implies that the feature is not ready yet.  Is that the
> case?
> 
I still consider the IEEE128 ldouble feature experimental on ppc64le. 
Thus, why we need to explicitly invoke extra compiler options today. 
This is another necessary step towards maturing the feature.  The ABI 
needs to be stablized to continue gcc enablement, and I think the ABI 
piece is ready.

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 6/6] powerpc64le: Enable support for IEEE long double
  2020-04-22 16:20     ` Paul E Murphy
@ 2020-04-22 16:23       ` Florian Weimer
  0 siblings, 0 replies; 22+ messages in thread
From: Florian Weimer @ 2020-04-22 16:23 UTC (permalink / raw)
  To: Paul E Murphy; +Cc: Paul E. Murphy via Libc-alpha

* Paul E. Murphy:

> On 4/22/20 10:15 AM, Florian Weimer wrote:
>> * Paul E. Murphy via Libc-alpha:
>> 
>>> diff --git a/NEWS b/NEWS
>>> index 07da1545b4..a54c0e0f65 100644
>>> --- a/NEWS
>>> +++ b/NEWS
>>> @@ -14,6 +14,11 @@ Major new features:
>>>   * The GNU C Library now loads audit modules listed in the DT_AUDIT and
>>>     DT_DEPAUDIT dynamic section entries of the main executable.
>>>   
>>> +* powerpc64le supports IEEE128 long double libm/libc redirects when
>>> +  using the -mabi=ieeelongdouble to compile C code on supported GCC
>>> +  toolchains.  It is recommended to use GCC 8 or newer when testing
>>> +  this option.
>> 
>> Why do you write “testing this option”?
>> 
>> To me, this implies that the feature is not ready yet.  Is that the
>> case?
>> 
> I still consider the IEEE128 ldouble feature experimental on ppc64le. 
> Thus, why we need to explicitly invoke extra compiler options today. 
> This is another necessary step towards maturing the feature.  The ABI 
> needs to be stablized to continue gcc enablement, and I think the ABI 
> piece is ready.

Ahh, thanks.  I think then the text is fine, and we can revisit the
overall presentation of this feature in the NEWS file later, closer to
the glibc 2.32 release.

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 3/6] powerpc64le: raise GCC requirement to 7.4 for long double transition
  2020-04-22 15:19   ` Florian Weimer
@ 2020-04-22 19:33     ` Paul E Murphy
  0 siblings, 0 replies; 22+ messages in thread
From: Paul E Murphy @ 2020-04-22 19:33 UTC (permalink / raw)
  To: Florian Weimer, Paul E. Murphy via Libc-alpha

On 4/22/20 10:19 AM, Florian Weimer wrote:
> * Paul E. Murphy via Libc-alpha:
> 
>> diff --git a/manual/install.texi b/manual/install.texi
>> index 71bf47cac6..66de8ea59b 100644
>> --- a/manual/install.texi
>> +++ b/manual/install.texi
>> @@ -507,6 +507,11 @@ the newest version of the compiler that is known to work for building
>>   release time, GCC 9.2.1 is the newest compiler verified to work to build
>>   @theglibc{}.
>>   
>> +For PowerPC 64-bits little-endian (powerpc64le), a GCC version with
>> +support for @option{-mno-gnu-attribute} is required.  As of release time,
>> +this means GCC 7.4 (excepting 7.5) or newer. The compiler option is required
>> +for building the GNU C Library with support for IEEE long double.
> 
> Why “excepting 7.5”?  Somewhere, there should be an explanation, maybe
> in a @c comment in the Texinfo sources if it's too confusing to end
> users.

FSF GCC 7.5 does not accept both the -mabi=* and -mlong-double-128 
options being present on the command line.  Ironically, it errors 
instructing you to pass the -mlong-double-128 option.

> Also, missing two spaces at the end of sentences.
> 
> And quick question for my understanding: Could we get the effect of
> -mno-gnu-attribute using strip if we wanted?  (Not that I think this
> change is required, I'm fine with bumping the minimum GCC
> requirement.)

Yes, I think one could post-process each object file to strip the 
.gnu.attributes section.  Though, you do highlight an impedance mismatch 
between the commit message and documentation regarding the -mabi 
requirements.

How about the following instead?

For PowerPC 64-bits little-endian (powerpc64le), a GCC version with 
support for @option{-mno-gnu-attribute}, @option{-mabi=ieeelongdouble}, 
and @option{-mabi=ibmlondouble} is required.  Likewise, the compiler 
must also support passing @options{-mlong-double-128} with the preceding 
options.  As of release, this implies GCC 7.4 and newer (excepting GCC 
7.5.0, see GCC PR94200).  These additional features are required for 
building the GNU C Library with support for IEEE long double.

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731
  2020-04-22 15:11   ` Florian Weimer
@ 2020-04-22 21:17     ` Paul E Murphy
  2020-04-23  5:48       ` Florian Weimer
  0 siblings, 1 reply; 22+ messages in thread
From: Paul E Murphy @ 2020-04-22 21:17 UTC (permalink / raw)
  To: Florian Weimer, Paul E. Murphy via Libc-alpha



On 4/22/20 10:11 AM, Florian Weimer wrote:
> * Paul E. Murphy via Libc-alpha:
> 
>> GCC 9 has a bug (BZ 90731) whereby __typeof does not correctly copy
>> exception specifiers[1].  Surprisingly, this can be quieted by declaring
>> "#pragma system_header", or if the headers are installed in a system
>> directory.
> 
> I think GCC developers prefer “PR90731” to their bugs.  This occurs
> multiple times in the commit message and throughout the patch.

Ack.  Will fix here and in the subsequent patches too.

> 
> This change will still cause errors with -Wsystem-headers in the build
> flags, right?  I think we should mention this somewhere, maybe on the
> release wiki?
> 

Unfortunately, ppc64le glibc will fail to build with GCC 9.2.0 with this 
option irrespective of this feature.  Though, you are correct the 
redirects will generate additional warnings when used with such toolchains.

Likewise, --disable-werror does avoid breaking things when using 
-Wsystem-headers.  The warnings are not fatal.

I suspect the -Wsystem-headers will cause some noise for most users with 
GCC 8/9 and likely others.  I can add a note, but I suspect I will need 
to be granted wiki editing privileges.

^ permalink raw reply	[flat|nested] 22+ messages in thread

* Re: [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731
  2020-04-22 21:17     ` Paul E Murphy
@ 2020-04-23  5:48       ` Florian Weimer
  0 siblings, 0 replies; 22+ messages in thread
From: Florian Weimer @ 2020-04-23  5:48 UTC (permalink / raw)
  To: Paul E Murphy; +Cc: Paul E. Murphy via Libc-alpha, Paul E. Murphy

* Paul E. Murphy:

> I suspect the -Wsystem-headers will cause some noise for most users with 
> GCC 8/9 and likely others.  I can add a note, but I suspect I will need 
> to be granted wiki editing privileges.

I can try to help with that.  What's your Wiki account name?

^ permalink raw reply	[flat|nested] 22+ messages in thread

end of thread, other threads:[~2020-04-23  5:48 UTC | newest]

Thread overview: 22+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-04-06 19:12 [PATCH v3 0/6] IEEE binary128 long double on powerpc64le Paul E. Murphy
2020-04-06 19:12 ` [PATCH v3 1/6] ldbl-128ibm-compat: workaround GCC 9 C++ BZ 90731 Paul E. Murphy
2020-04-15 14:20   ` Paul E Murphy
2020-04-22 14:58     ` Paul E Murphy
2020-04-22 15:11   ` Florian Weimer
2020-04-22 21:17     ` Paul E Murphy
2020-04-23  5:48       ` Florian Weimer
2020-04-06 19:12 ` [PATCH v3 2/6] Rename __LONG_DOUBLE_USES_FLOAT128 to __LDOUBLE_REDIRECTS_TO_FLOAT128_ABI Paul E. Murphy
2020-04-15 14:21   ` Paul E Murphy
2020-04-22 14:59     ` Paul E Murphy
2020-04-06 19:12 ` [PATCH v3 3/6] powerpc64le: raise GCC requirement to 7.4 for long double transition Paul E. Murphy
2020-04-22 15:19   ` Florian Weimer
2020-04-22 19:33     ` Paul E Murphy
2020-04-06 19:12 ` [PATCH v3 4/6] powerpc64le: bump binutils version requirement to >= 2.26 Paul E. Murphy
2020-04-22 15:20   ` Florian Weimer
2020-04-06 19:12 ` [PATCH v3 5/6] powerpc64le: blacklist broken GCC compilers (e.g GCC 7.5.0) Paul E. Murphy
2020-04-22 15:23   ` Florian Weimer
2020-04-06 19:12 ` [PATCH v3 6/6] powerpc64le: Enable support for IEEE long double Paul E. Murphy
2020-04-22 15:03   ` Paul E Murphy
2020-04-22 15:15   ` Florian Weimer
2020-04-22 16:20     ` Paul E Murphy
2020-04-22 16:23       ` Florian Weimer

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).