gcc/libatomic/configure.tgt
Xi Ruoyao db80262475
libatomic: x86_64: Always try ifunc
We used to skip ifunc check when CX16 is available.  But now we use
CX16+AVX+Intel/AMD for the "perfect" 16b load implementation, so CX16
alone is not a sufficient reason not to use ifunc (see PR104688).

This causes a subtle and annoying issue: when GCC is built with a
higher -march= setting in CFLAGS_FOR_TARGET, ifunc is disabled and
the worst (locked) implementation of __atomic_load_16 is always used.

There seems no good way to check if the CPU is Intel or AMD from
the built-in macros (maybe we can check every known model like __skylake,
__bdver2, ..., but it will be very error-prune and require an update
whenever we add the support for a new x86 model).  The best thing we can
do seems "always try ifunc" here.

libatomic/ChangeLog:

	* configure.tgt: For x86_64, always set try_ifunc=yes.
2023-06-10 21:43:56 +08:00

203 lines
5.0 KiB
Bash

# -*- shell-script -*-
# Copyright (C) 2012-2023 Free Software Foundation, Inc.
# Contributed by Richard Henderson <rth@redhat.com>.
#
# This file is part of the GNU Atomic Library (libatomic).
#
# Libatomic is free software; you can redistribute it and/or modify it
# under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# Libatomic is distributed in the hope that it will be useful, but WITHOUT ANY
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for
# more details.
#
# Under Section 7 of GPL version 3, you are granted additional
# permissions described in the GCC Runtime Library Exception, version
# 3.1, as published by the Free Software Foundation.
#
# You should have received a copy of the GNU General Public License and
# a copy of the GCC Runtime Library Exception along with this program;
# see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
# <http://www.gnu.org/licenses/>.
# Map the target cpu to an ARCH sub-directory. At the same time,
# work out any special compilation flags as necessary.
# Give operating systems the opportunity to discard XCFLAGS modifications based
# on ${target_cpu}. For example to allow proper use of multilibs.
configure_tgt_pre_target_cpu_XCFLAGS="${XCFLAGS}"
case "${target_cpu}" in
alpha*)
# fenv.c needs this option to generate inexact exceptions.
XCFLAGS="${XCFLAGS} -mfp-trap-mode=sui"
ARCH=alpha
;;
rs6000 | powerpc*) ARCH=powerpc ;;
riscv*) ARCH=riscv ;;
sh*) ARCH=sh ;;
aarch64*)
ARCH=aarch64
case "${target}" in
aarch64*-*-linux*)
if test -n "$enable_aarch64_lse"; then
try_ifunc=yes
fi
;;
esac
XCFLAGS="${XCFLAGS} -mno-outline-atomics"
;;
arm*)
ARCH=arm
case "${target}" in
arm*-*-freebsd* | arm*-*-netbsd*)
;;
*)
# ??? Detect when -march=armv7 is already enabled.
try_ifunc=yes
;;
esac
;;
sparc)
case " ${CC} ${CFLAGS} " in
*" -m64 "*)
;;
*)
if test -z "$with_cpu"; then
XCFLAGS="${XCFLAGS} -mcpu=v9"
fi
esac
ARCH=sparc
;;
sparc64|sparcv9)
case " ${CC} ${CFLAGS} " in
*" -m32 "*)
XCFLAGS="${XCFLAGS} -mcpu=v9"
;;
esac
ARCH=sparc
;;
i[3456]86 | x86_64)
cat > conftestx.c <<EOF
#ifndef __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4
#error need -march=i486
#endif
EOF
if ${CC} ${CFLAGS} -E conftestx.c > /dev/null 2>&1; then
:
else
if test "${target_cpu}" = x86_64; then
XCFLAGS="${XCFLAGS} -march=i486 -mtune=generic"
else
XCFLAGS="${XCFLAGS} -march=i486 -mtune=${target_cpu}"
fi
XCFLAGS="${XCFLAGS} -fomit-frame-pointer"
fi
cat > conftestx.c <<EOF
#ifdef __x86_64__
#error ifunc is always wanted for 16B atomic load
#else
#ifndef __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8
#error need -march=i686
#endif
#endif
EOF
if ${CC} ${CFLAGS} -E conftestx.c > /dev/null 2>&1; then
try_ifunc=no
else
try_ifunc=yes
fi
rm -f conftestx.c
ARCH=x86
;;
*) ARCH="${target_cpu}" ;;
esac
# The cpu configury is always most relevant.
if test -d ${srcdir}/config/$ARCH ; then
config_path="$ARCH"
fi
tmake_file=
# Other system configury
case "${target}" in
aarch64*-*-linux*)
# OS support for atomic primitives.
config_path="${config_path} linux/aarch64 posix"
;;
arm*-*-linux* | arm*-*-uclinux*)
# OS support for atomic primitives.
config_path="${config_path} linux/arm posix"
;;
s390*-*-linux*)
# OS support for atomic primitives.
config_path="${config_path} s390 posix"
;;
powerpc*-*-aix*)
config_path="${config_path} posix"
tmake_file="t-aix"
;;
*-*-linux* | *-*-gnu* | *-*-k*bsd*-gnu \
| *-*-netbsd* | *-*-freebsd* | *-*-openbsd* | *-*-dragonfly* \
| *-*-solaris2* | *-*-sysv4* | *-*-irix6* | *-*-osf* | *-*-hpux11* \
| *-*-darwin* | *-*-aix* | *-*-cygwin*)
# POSIX system. The OS is supported.
config_path="${config_path} posix"
;;
*-*-mingw*)
# OS support for atomic primitives.
case ${target_thread_file} in
win32 | mcf | single)
config_path="${config_path} mingw"
;;
posix)
config_path="${config_path} posix"
;;
esac
;;
*-*-rtems*)
XCFLAGS="${configure_tgt_pre_target_cpu_XCFLAGS}"
config_path="rtems"
;;
*-*-elf*)
# ??? No target OS. We could be targeting bare-metal kernel-mode,
# or user-mode for some custom OS. If the target supports TAS,
# we can build our own spinlocks, given there are no signals.
# If the target supports disabling interrupts, we can work in
# kernel-mode, given the system is not multi-processor.
UNSUPPORTED=1
;;
nvptx*-*-*)
;;
*)
# Who are you?
UNSUPPORTED=1
;;
esac
# glibc will pass hwcap to ifunc resolver functions as an argument.
# The type may be different on different architectures.
case "${target}" in
aarch64*-*-*)
IFUNC_RESOLVER_ARGS="uint64_t hwcap"
;;
*)
IFUNC_RESOLVER_ARGS="void"
;;
esac