This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
[PATCH 0/7] Remaining ldbl-128ibm-compat support
- From: "Paul E. Murphy" <murphyp at linux dot vnet dot ibm dot com>
- To: libc-alpha at sourceware dot org
- Date: Fri, 14 Feb 2020 10:10:44 -0600
- Subject: [PATCH 0/7] Remaining ldbl-128ibm-compat support
This patch series contains the outstanding work
to support the IEEE 128 long double redirects from
within ldbl-128ibm-compat.
In conjunction with with the remaining patches on
the mailing list:
<https://sourceware.org/ml/libc-alpha/2020-02/msg00401.html>
and the remaing 5 header patches below
<https://sourceware.org/ml/libc-alpha/2020-02/msg00049.html>
This should complete the backing support for the long
double redirects. A final set of patches will enable the
redirects for ppc64le. There is still more arch specific
Makefile wrangling to build this subdir.
Patches 1,3,5 are straightforward enablement.
Patch 2 enables the nexttoward/nextafterl redirects. It is
nearly trivial, but fixes up some ldbl-128 files to use the
_Float128 name instead of 'long double' to promote reuse.
Patch 4 touches common math Makefile to support testing the
ibm128 type similar to other types. It also modifies the
test-<type>.h to choose correct ULPs when redirects are
being used.
Patch 6 updates the ldbl-128ibm Makefile to support adding
extra compiler flags, primarily -mabi=ibmlongdouble. It
also happened to be sitting between Patch 5 and 7.
Patch 7 allows using the redirects with libc while also
avoiding internal PLT usage. This is kind of ugly, and I
would appreciate feedback on whether this approach is
acceptable, and if so, whether the macros are located in
the most suitable places.
Gabriel F. T. Gomes (1):
ldbl-128ibm-compat: Provide nexttoward functions
Paul E. Murphy (4):
ldbl-128ibm-compat: enforce correct abi flags on internal file
ldbl-128ibm-compat: fixup subdir location of several funcs
ldbl-128ibm-compat: enforce ibm128 on compat tests
RFC: ldbl-128ibm-compat: PLT redirects for using ldbl redirects
internally
Rajalakshmi Srinivasaraghavan (1):
ldbl-128ibm-compat: Add tests for IBM long double functions
Tulio Magno Quites Machado Filho (1):
ldbl-128ibm: Let long double files have specific compiler flags
include/err.h | 12 ++---
include/stdio.h | 51 ++++++++++++------
include/sys/cdefs.h | 23 ++++++++
include/sys/syslog.h | 2 +-
include/wchar.h | 2 +-
math/Makefile | 7 ++-
math/test-float128.h | 2 +-
math/test-float64x.h | 2 +-
math/test-ibm128.h | 19 +++++++
math/test-ldouble.h | 6 +++
sysdeps/ieee754/ldbl-128/s_nexttoward.c | 4 +-
sysdeps/ieee754/ldbl-128/s_nexttowardf.c | 4 +-
sysdeps/ieee754/ldbl-128ibm-compat/Makeconfig | 3 ++
sysdeps/ieee754/ldbl-128ibm-compat/Makefile | 53 +++++++++++++------
sysdeps/ieee754/ldbl-128ibm-compat/Versions | 3 ++
.../ldbl-128ibm-compat/ieee128-asprintf.c | 5 +-
.../ldbl-128ibm-compat/ieee128-dprintf.c | 1 +
.../ieee754/ldbl-128ibm-compat/ieee128-err.c | 7 +++
.../ldbl-128ibm-compat/ieee128-fprintf.c | 1 +
.../ieee128-isoc99_sscanf.c | 1 +
.../ieee128-isoc99_vfscanf.c | 2 +
.../ieee128-isoc99_vsscanf.c | 1 +
.../ldbl-128ibm-compat/ieee128-sprintf.c | 1 +
.../ldbl-128ibm-compat/ieee128-syslog.c | 1 +
.../ldbl-128ibm-compat/ieee128-vfprintf.c | 1 +
.../ldbl-128ibm-compat/s_nextafterf128.c | 20 +++++++
.../ldbl-128ibm-compat/s_nexttowardf128.c | 38 +++++++++++++
sysdeps/ieee754/ldbl-128ibm/Makefile | 22 ++++++--
.../ldbl-128ibm/include/bits/iscanonical.h | 2 +-
sysdeps/powerpc/powerpc64/le/Makefile | 50 ++++++++++++++---
30 files changed, 288 insertions(+), 58 deletions(-)
create mode 100644 math/test-ibm128.h
create mode 100644 sysdeps/ieee754/ldbl-128ibm-compat/Makeconfig
create mode 100644 sysdeps/ieee754/ldbl-128ibm-compat/s_nextafterf128.c
create mode 100644 sysdeps/ieee754/ldbl-128ibm-compat/s_nexttowardf128.c
--
2.21.1