* [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32)
@ 2020-06-03 16:25 Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 01/18] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64 Alistair Francis via Libc-alpha
` (18 more replies)
0 siblings, 19 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
This patch set contains the glibc port for 32-bit RISC-V.
This is based on the original work from Zong Li [1] and has been
updated to use a 64-bit time_t.
This requires a 5.4+ kernel and all of the testing has been done using
the 5.4 stable kernel.
Nothing fails when running ./scripts/build-many-glibcs.py (for all
targets) on my x86-64 machine.
This is the current list of tests that fail when running inside QEMU RV32
system emulation on the 5.4 kernel:
FAIL: elf/tst-ldconfig-ld_so_conf-update
FAIL: io/tst-lockf
FAIL: misc/test-errno-linux
FAIL: nss/tst-nss-files-hosts-long
FAIL: resolv/tst-resolv-res_init-thread
FAIL: stdio-common/bug22
FAIL: stdlib/tst-strfrom
FAIL: stdlib/tst-strfrom-locale
---Links---
1: https://sourceware.org/ml/libc-alpha/2018-07/msg00892.html
The latest version of my work can be found here: https://github.com/alistair23/glibc/tree/alistair/rv32.next
This specific version can be found here: https://github.com/alistair23/glibc/tree/alistair/rv32.2
Alistair Francis (9):
RISC-V: Use 64-bit time_t and off_t for RV32 and RV64
RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit
RISC-V: Add support for 32-bit vDSO calls
RISC-V: Add arch-syscall.h for RV32
RISC-V: nptl: update default pthread-offsets.h
riscv32: Add an architecture ipctypes.h
RISC-V: Add ABI lists
RISC-V: Add the RV32 libm-test-ulps
riscv32: Specify the arch_minimum_kernel as 5.4
Zong Li (9):
RISC-V: Support dynamic loader for the 32-bit
RISC-V: Add path of library directories for the 32-bit
RISC-V: The ABI implementation for 32-bit
RISC-V: Hard float support for 32-bit
RISC-V: Fix llrint and llround missing exceptions on RV32
RISC-V: Build Infastructure for 32-bit
RISC-V: Add rv32 path to RTLDLIST in ldd
Documentation for the RISC-V 32-bit port
Add RISC-V 32-bit target to build-many-glibcs.py
NEWS | 6 +
README | 1 +
scripts/build-many-glibcs.py | 15 +
sysdeps/riscv/bits/wordsize.h | 4 +-
sysdeps/riscv/nptl/bits/pthreadtypes-arch.h | 10 +-
sysdeps/riscv/nptl/bits/struct_rwlock.h | 27 +-
sysdeps/riscv/nptl/pthread-offsets.h | 13 +-
sysdeps/riscv/preconfigure | 6 +-
sysdeps/riscv/rv32/Implies-after | 1 +
.../riscv/rv32/fix-fp-int-convert-overflow.h | 38 +
sysdeps/riscv/rv32/rvd/Implies | 3 +
sysdeps/riscv/rv32/rvd/libm-test-ulps | 1401 +++++++++++
sysdeps/riscv/rv32/rvd/libm-test-ulps-name | 1 +
sysdeps/riscv/rv32/rvd/s_lrint.c | 31 +
sysdeps/riscv/rv32/rvd/s_lround.c | 31 +
sysdeps/riscv/rv32/rvf/Implies | 1 +
sysdeps/riscv/rv32/rvf/s_lrintf.c | 31 +
sysdeps/riscv/rv32/rvf/s_lroundf.c | 31 +
sysdeps/riscv/sfp-machine.h | 27 +-
sysdeps/riscv/sys/asm.h | 5 +-
sysdeps/unix/sysv/linux/riscv/Makefile | 4 +-
.../unix/sysv/linux/riscv/bits/environments.h | 85 +
sysdeps/unix/sysv/linux/riscv/bits/ipctypes.h | 35 +
sysdeps/unix/sysv/linux/riscv/bits/time64.h | 36 +
sysdeps/unix/sysv/linux/riscv/bits/timesize.h | 22 +
sysdeps/unix/sysv/linux/riscv/c++-types.data | 67 +
sysdeps/unix/sysv/linux/riscv/configure | 43 +
sysdeps/unix/sysv/linux/riscv/configure.ac | 12 +
sysdeps/unix/sysv/linux/riscv/dl-cache.h | 17 +-
.../unix/sysv/linux/riscv/jmp_buf-macros.h | 53 +
sysdeps/unix/sysv/linux/riscv/kernel_stat.h | 23 +
sysdeps/unix/sysv/linux/riscv/ld.abilist | 5 +
sysdeps/unix/sysv/linux/riscv/ldconfig.h | 2 +-
sysdeps/unix/sysv/linux/riscv/ldd-rewrite.sed | 2 +-
.../sysv/linux/riscv/libBrokenLocale.abilist | 1 +
sysdeps/unix/sysv/linux/riscv/libanl.abilist | 4 +
sysdeps/unix/sysv/linux/riscv/libc.abilist | 2099 +++++++++++++++++
.../unix/sysv/linux/riscv/libcrypt.abilist | 2 +
sysdeps/unix/sysv/linux/riscv/libdl.abilist | 9 +
sysdeps/unix/sysv/linux/riscv/libm.abilist | 940 ++++++++
.../unix/sysv/linux/riscv/libpthread.abilist | 213 ++
.../unix/sysv/linux/riscv/libresolv.abilist | 79 +
sysdeps/unix/sysv/linux/riscv/librt.abilist | 35 +
.../sysv/linux/riscv/libthread_db.abilist | 40 +
sysdeps/unix/sysv/linux/riscv/libutil.abilist | 6 +
sysdeps/unix/sysv/linux/riscv/rv32/Implies | 3 +
.../unix/sysv/linux/riscv/rv32/arch-syscall.h | 283 +++
sysdeps/unix/sysv/linux/riscv/shlib-versions | 10 +-
sysdeps/unix/sysv/linux/riscv/sysdep.h | 80 +-
49 files changed, 5868 insertions(+), 25 deletions(-)
create mode 100644 sysdeps/riscv/rv32/Implies-after
create mode 100644 sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h
create mode 100644 sysdeps/riscv/rv32/rvd/Implies
create mode 100644 sysdeps/riscv/rv32/rvd/libm-test-ulps
create mode 100644 sysdeps/riscv/rv32/rvd/libm-test-ulps-name
create mode 100644 sysdeps/riscv/rv32/rvd/s_lrint.c
create mode 100644 sysdeps/riscv/rv32/rvd/s_lround.c
create mode 100644 sysdeps/riscv/rv32/rvf/Implies
create mode 100644 sysdeps/riscv/rv32/rvf/s_lrintf.c
create mode 100644 sysdeps/riscv/rv32/rvf/s_lroundf.c
create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/environments.h
create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/ipctypes.h
create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/time64.h
create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/timesize.h
create mode 100644 sysdeps/unix/sysv/linux/riscv/c++-types.data
create mode 100644 sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h
create mode 100644 sysdeps/unix/sysv/linux/riscv/kernel_stat.h
create mode 100644 sysdeps/unix/sysv/linux/riscv/ld.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libBrokenLocale.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libanl.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libc.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libcrypt.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libdl.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libm.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libpthread.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libresolv.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/librt.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libthread_db.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libutil.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/rv32/Implies
create mode 100644 sysdeps/unix/sysv/linux/riscv/rv32/arch-syscall.h
--
2.26.2
^ permalink raw reply [flat|nested] 55+ messages in thread
* [PATCH v2 01/18] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-07-07 22:06 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 02/18] RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit Alistair Francis via Libc-alpha
` (17 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
Using the original glibc headers under bits/ let's make small
modifications to use 64-bit time_t and off_t for both RV32 and RV64.
For the typesizes.h, here are justifications for the changes from the
generic version (based on Arnd's very helpful feedback):
- All the !__USE_FILE_OFFSET64 types (__off_t, __ino_t, __rlim_t, ...) are
changed to match the 64-bit replacements.
- __time_t is defined to 64 bit, but no __time64_t is added. This makes sense
as we don't have the time64 support for other 32-bit architectures yet, and
it will be easy to change when that happens.
- __suseconds_t is 64-bit. This matches what we use the kernel ABI for the
few drivers that are relying on 'struct timeval' input arguments in
ioctl, as well as the adjtimex system call. It means that timeval has to
be defined without the padding, unlike timespec, which needs padding.
---
.../unix/sysv/linux/riscv/bits/environments.h | 85 +++++++++++++++++++
sysdeps/unix/sysv/linux/riscv/bits/time64.h | 36 ++++++++
sysdeps/unix/sysv/linux/riscv/bits/timesize.h | 22 +++++
sysdeps/unix/sysv/linux/riscv/kernel_stat.h | 23 +++++
4 files changed, 166 insertions(+)
create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/environments.h
create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/time64.h
create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/timesize.h
create mode 100644 sysdeps/unix/sysv/linux/riscv/kernel_stat.h
diff --git a/sysdeps/unix/sysv/linux/riscv/bits/environments.h b/sysdeps/unix/sysv/linux/riscv/bits/environments.h
new file mode 100644
index 0000000000..c1768af17b
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/bits/environments.h
@@ -0,0 +1,85 @@
+/* Copyright (C) 2019-2020 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#ifndef _UNISTD_H
+# error "Never include this file directly. Use <unistd.h> instead"
+#endif
+
+#include <bits/wordsize.h>
+
+/* This header should define the following symbols under the described
+ situations. A value `1' means that the model is always supported,
+ `-1' means it is never supported. Undefined means it cannot be
+ statically decided.
+
+ _POSIX_V7_ILP32_OFF32 32bit int, long, pointers, and off_t type
+ _POSIX_V7_ILP32_OFFBIG 32bit int, long, and pointers and larger off_t type
+
+ _POSIX_V7_LP64_OFF32 64bit long and pointers and 32bit off_t type
+ _POSIX_V7_LPBIG_OFFBIG 64bit long and pointers and large off_t type
+
+ The macros _POSIX_V6_ILP32_OFF32, _POSIX_V6_ILP32_OFFBIG,
+ _POSIX_V6_LP64_OFF32, _POSIX_V6_LPBIG_OFFBIG, _XBS5_ILP32_OFF32,
+ _XBS5_ILP32_OFFBIG, _XBS5_LP64_OFF32, and _XBS5_LPBIG_OFFBIG were
+ used in previous versions of the Unix standard and are available
+ only for compatibility.
+*/
+
+#if __WORDSIZE == 64
+
+/* We can never provide environments with 32-bit wide pointers. */
+# define _POSIX_V7_ILP32_OFF32 -1
+# define _POSIX_V7_ILP32_OFFBIG -1
+# define _POSIX_V6_ILP32_OFF32 -1
+# define _POSIX_V6_ILP32_OFFBIG -1
+# define _XBS5_ILP32_OFF32 -1
+# define _XBS5_ILP32_OFFBIG -1
+/* We also have no use (for now) for an environment with bigger pointers
+ and offsets. */
+# define _POSIX_V7_LPBIG_OFFBIG -1
+# define _POSIX_V6_LPBIG_OFFBIG -1
+# define _XBS5_LPBIG_OFFBIG -1
+
+/* By default we have 64-bit wide `long int', pointers and `off_t'. */
+# define _POSIX_V7_LP64_OFF64 1
+# define _POSIX_V6_LP64_OFF64 1
+# define _XBS5_LP64_OFF64 1
+
+#else /* __WORDSIZE == 32 */
+
+/* RISC-V requires 64-bit off_t
+ # undef _POSIX_V7_ILP32_OFF32
+ # undef _POSIX_V6_ILP32_OFF32
+ # undef _XBS5_ILP32_OFF32
+ */
+
+# define _POSIX_V7_ILP32_OFFBIG 1
+# define _POSIX_V6_ILP32_OFFBIG 1
+# define _XBS5_ILP32_OFFBIG 1
+
+/* We can never provide environments with 64-bit wide pointers. */
+# define _POSIX_V7_LP64_OFF64 -1
+# define _POSIX_V7_LPBIG_OFFBIG -1
+# define _POSIX_V6_LP64_OFF64 -1
+# define _POSIX_V6_LPBIG_OFFBIG -1
+# define _XBS5_LP64_OFF64 -1
+# define _XBS5_LPBIG_OFFBIG -1
+
+/* CFLAGS. */
+#define __ILP32_OFFBIG_CFLAGS "-D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64"
+
+#endif /* __WORDSIZE == 32 */
diff --git a/sysdeps/unix/sysv/linux/riscv/bits/time64.h b/sysdeps/unix/sysv/linux/riscv/bits/time64.h
new file mode 100644
index 0000000000..7ba0dd1cfd
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/bits/time64.h
@@ -0,0 +1,36 @@
+/* bits/time64.h -- underlying types for __time64_t. Generic version.
+ Copyright (C) 2019-2020 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#ifndef _BITS_TYPES_H
+# error "Never include <bits/time64.h> directly; use <sys/types.h> instead."
+#endif
+
+#ifndef _BITS_TIME64_H
+#define _BITS_TIME64_H 1
+
+/* Define __TIME64_T_TYPE so that it is always a 64-bit type. */
+
+#if __WORDSIZE == 64
+/* If we already have 64-bit time type then use it. */
+# define __TIME64_T_TYPE __TIME_T_TYPE
+#else
+/* Define a 64-bit time type alongsize the 32-bit one. */
+# define __TIME64_T_TYPE __SQUAD_TYPE
+#endif
+
+#endif /* bits/time64.h */
diff --git a/sysdeps/unix/sysv/linux/riscv/bits/timesize.h b/sysdeps/unix/sysv/linux/riscv/bits/timesize.h
new file mode 100644
index 0000000000..78394b7489
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/bits/timesize.h
@@ -0,0 +1,22 @@
+/* Bit size of the time_t type at glibc build time, general case.
+ Copyright (C) 2019-2020 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include <bits/wordsize.h>
+
+/* RV32 and RV64 both use 64-bit time_t */
+#define __TIMESIZE 64
diff --git a/sysdeps/unix/sysv/linux/riscv/kernel_stat.h b/sysdeps/unix/sysv/linux/riscv/kernel_stat.h
new file mode 100644
index 0000000000..25e698b9bf
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/kernel_stat.h
@@ -0,0 +1,23 @@
+/* Copyright (C) 2019-2020 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library. If not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include <bits/wordsize.h>
+
+#define STAT_IS_KERNEL_STAT 1
+
+#define XSTAT_IS_XSTAT64 1
+#define STATFS_IS_STATFS64 0
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 02/18] RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 01/18] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64 Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-07-08 0:09 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 03/18] RISC-V: Add support for 32-bit vDSO calls Alistair Francis via Libc-alpha
` (16 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
---
sysdeps/unix/sysv/linux/riscv/sysdep.h | 61 ++++++++++++++++++++++++++
1 file changed, 61 insertions(+)
diff --git a/sysdeps/unix/sysv/linux/riscv/sysdep.h b/sysdeps/unix/sysv/linux/riscv/sysdep.h
index 83e4adf6a2..aa61e8b04d 100644
--- a/sysdeps/unix/sysv/linux/riscv/sysdep.h
+++ b/sysdeps/unix/sysv/linux/riscv/sysdep.h
@@ -116,6 +116,67 @@
#include <sysdeps/unix/sysdep.h>
+#if __riscv_xlen == 32
+/* Define the __NR_futex as __NR_futex64 as RV32 doesn't have a
+ * __NR_futex syscall.
+ */
+# ifndef __NR_futex
+# define __NR_futex __NR_futex_time64
+# endif
+
+# ifndef __NR_rt_sigtimedwait
+# define __NR_rt_sigtimedwait __NR_rt_sigtimedwait_time64
+# endif
+
+# ifndef __NR_ppoll
+# define __NR_ppoll __NR_ppoll_time64
+# endif
+
+# ifndef __NR_utimensat
+# define __NR_utimensat __NR_utimensat_time64
+# endif
+
+# ifndef __NR_pselect6
+# define __NR_pselect6 __NR_pselect6_time64
+# endif
+
+# ifndef __NR_recvmmsg
+# define __NR_recvmmsg __NR_recvmmsg_time64
+# endif
+
+# ifndef __NR_semtimedop
+# define __NR_semtimedop __NR_semtimedop_time64
+# endif
+
+# ifndef __NR_mq_timedreceive
+# define __NR_mq_timedreceive __NR_mq_timedreceive_time64
+# endif
+
+# ifndef __NR_mq_timedsend
+# define __NR_mq_timedsend __NR_mq_timedsend_time64
+# endif
+
+# ifndef __NR_clock_getres
+# define __NR_clock_getres __NR_clock_getres_time64
+# endif
+
+# ifndef __NR_timerfd_settime
+# define __NR_timerfd_settime __NR_timerfd_settime64
+# endif
+
+# ifndef __NR_timerfd_gettime
+# define __NR_timerfd_gettime __NR_timerfd_gettime64
+# endif
+
+# ifndef __NR_sched_rr_get_interval
+# define __NR_sched_rr_get_interval __NR_sched_rr_get_interval_time64
+# endif
+
+# ifndef __NR_clock_adjtime
+# define __NR_clock_adjtime __NR_clock_adjtime64
+# endif
+#endif /* __riscv_xlen == 32 */
+
#undef SYS_ify
#define SYS_ify(syscall_name) __NR_##syscall_name
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 03/18] RISC-V: Add support for 32-bit vDSO calls
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 01/18] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64 Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 02/18] RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-07-08 1:01 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 04/18] RISC-V: Support dynamic loader for the 32-bit Alistair Francis via Libc-alpha
` (15 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
---
sysdeps/unix/sysv/linux/riscv/sysdep.h | 19 +++++++++++++------
1 file changed, 13 insertions(+), 6 deletions(-)
diff --git a/sysdeps/unix/sysv/linux/riscv/sysdep.h b/sysdeps/unix/sysv/linux/riscv/sysdep.h
index aa61e8b04d..9f6ed2001b 100644
--- a/sysdeps/unix/sysv/linux/riscv/sysdep.h
+++ b/sysdeps/unix/sysv/linux/riscv/sysdep.h
@@ -182,14 +182,21 @@
#ifndef __ASSEMBLER__
-# define VDSO_NAME "LINUX_4.15"
-# define VDSO_HASH 182943605
+# if __riscv_xlen == 32
+# define VDSO_NAME "LINUX_5.4"
+# define VDSO_HASH 61765876
/* List of system calls which are supported as vsyscalls. */
-# define HAVE_CLOCK_GETRES64_VSYSCALL "__vdso_clock_getres"
-# define HAVE_CLOCK_GETTIME64_VSYSCALL "__vdso_clock_gettime"
-# define HAVE_GETTIMEOFDAY_VSYSCALL "__vdso_gettimeofday"
-# define HAVE_GETCPU_VSYSCALL "__vdso_getcpu"
+# else
+# define VDSO_NAME "LINUX_4.15"
+# define VDSO_HASH 182943605
+
+/* List of system calls which are supported as vsyscalls. */
+# define HAVE_CLOCK_GETRES64_VSYSCALL "__vdso_clock_getres"
+# define HAVE_CLOCK_GETTIME64_VSYSCALL "__vdso_clock_gettime"
+# define HAVE_GETTIMEOFDAY_VSYSCALL "__vdso_gettimeofday"
+# define HAVE_GETCPU_VSYSCALL "__vdso_getcpu"
+# endif
# define INTERNAL_SYSCALL(name, nr, args...) \
internal_syscall##nr (SYS_ify (name), args)
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 04/18] RISC-V: Support dynamic loader for the 32-bit
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (2 preceding siblings ...)
2020-06-03 16:25 ` [PATCH v2 03/18] RISC-V: Add support for 32-bit vDSO calls Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-07-08 1:35 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 05/18] RISC-V: Add path of library directories " Alistair Francis via Libc-alpha
` (14 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
From: Zong Li <zongbox@gmail.com>
Add the LD_SO_ABI definition for RISC-V 32-bit.
---
sysdeps/unix/sysv/linux/riscv/ldconfig.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/sysdeps/unix/sysv/linux/riscv/ldconfig.h b/sysdeps/unix/sysv/linux/riscv/ldconfig.h
index 3d7065b293..35badb7cf0 100644
--- a/sysdeps/unix/sysv/linux/riscv/ldconfig.h
+++ b/sysdeps/unix/sysv/linux/riscv/ldconfig.h
@@ -24,7 +24,7 @@
#if __riscv_xlen == 64
# define LD_SO_ABI "riscv64-lp64"
#else
-# error "rv32i-based targets are not supported"
+# define LD_SO_ABI "riscv32-ilp32"
#endif
#define SYSDEP_KNOWN_INTERPRETER_NAMES \
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 05/18] RISC-V: Add path of library directories for the 32-bit
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (3 preceding siblings ...)
2020-06-03 16:25 ` [PATCH v2 04/18] RISC-V: Support dynamic loader for the 32-bit Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-07-08 18:42 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 06/18] RISC-V: Add arch-syscall.h for RV32 Alistair Francis via Libc-alpha
` (13 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
From: Zong Li <zongbox@gmail.com>
For the recommand of 64 bit version, we add the libraries path of 32 bit
in this patch.
The status of RV32 binaries under RV64 kernels is that the ISA
optionally supports having different XLEN for user and supervisor modes,
but AFAIK there's no silicon that implements this feature, and the Linux
kernel doesn't support it yet.
For the recommand of 64 bit version, we add the libraries path of 32 bit
in this patch. This includes a fix to avoid an out of bound array check
when building with GCC 8.2.
---
sysdeps/unix/sysv/linux/riscv/dl-cache.h | 17 +++++++++++++++--
1 file changed, 15 insertions(+), 2 deletions(-)
diff --git a/sysdeps/unix/sysv/linux/riscv/dl-cache.h b/sysdeps/unix/sysv/linux/riscv/dl-cache.h
index c297dfe84f..60fc172edb 100644
--- a/sysdeps/unix/sysv/linux/riscv/dl-cache.h
+++ b/sysdeps/unix/sysv/linux/riscv/dl-cache.h
@@ -34,6 +34,8 @@
RISC-V, libraries can be found in paths ending in:
- /lib64/lp64d
- /lib64/lp64
+ - /lib32/ilp32d
+ - /lib32/ilp32
- /lib (only ld.so)
so this will add all of those paths.
@@ -49,9 +51,16 @@
do \
{ \
size_t len = strlen (dir); \
- char path[len + 9]; \
+ char path[len + 10]; \
memcpy (path, dir, len + 1); \
- if (len >= 12 && ! memcmp(path + len - 12, "/lib64/lp64d", 12)) \
+ if (len >= 13 && ! memcmp(path + len - 13, "/lib32/ilp32d", 13)) \
+ { \
+ len -= 9; \
+ path[len] = '\0'; \
+ } \
+ if (len >= 12 \
+ && (! memcmp(path + len - 12, "/lib32/ilp32", 12) \
+ || ! memcmp(path + len - 12, "/lib64/lp64d", 12))) \
{ \
len -= 8; \
path[len] = '\0'; \
@@ -64,6 +73,10 @@
add_dir (path); \
if (len >= 4 && ! memcmp(path + len - 4, "/lib", 4)) \
{ \
+ memcpy (path + len, "32/ilp32d", 10); \
+ add_dir (path); \
+ memcpy (path + len, "32/ilp32", 9); \
+ add_dir (path); \
memcpy (path + len, "64/lp64d", 9); \
add_dir (path); \
memcpy (path + len, "64/lp64", 8); \
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 06/18] RISC-V: Add arch-syscall.h for RV32
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (4 preceding siblings ...)
2020-06-03 16:25 ` [PATCH v2 05/18] RISC-V: Add path of library directories " Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-07-08 19:33 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 07/18] RISC-V: nptl: update default pthread-offsets.h Alistair Francis via Libc-alpha
` (12 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
---
.../unix/sysv/linux/riscv/rv32/arch-syscall.h | 283 ++++++++++++++++++
1 file changed, 283 insertions(+)
create mode 100644 sysdeps/unix/sysv/linux/riscv/rv32/arch-syscall.h
diff --git a/sysdeps/unix/sysv/linux/riscv/rv32/arch-syscall.h b/sysdeps/unix/sysv/linux/riscv/rv32/arch-syscall.h
new file mode 100644
index 0000000000..9f9a2d5912
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/rv32/arch-syscall.h
@@ -0,0 +1,283 @@
+/* AUTOGENERATED by update-syscall-lists.py. */
+#define __NR_accept 202
+#define __NR_accept4 242
+#define __NR_acct 89
+#define __NR_add_key 217
+#define __NR_bind 200
+#define __NR_bpf 280
+#define __NR_brk 214
+#define __NR_capget 90
+#define __NR_capset 91
+#define __NR_chdir 49
+#define __NR_chroot 51
+#define __NR_clock_adjtime64 405
+#define __NR_clock_getres_time64 406
+#define __NR_clock_gettime64 403
+#define __NR_clock_nanosleep_time64 407
+#define __NR_clock_settime64 404
+#define __NR_clone 220
+#define __NR_close 57
+#define __NR_connect 203
+#define __NR_copy_file_range 285
+#define __NR_delete_module 106
+#define __NR_dup 23
+#define __NR_dup3 24
+#define __NR_epoll_create1 20
+#define __NR_epoll_ctl 21
+#define __NR_epoll_pwait 22
+#define __NR_eventfd2 19
+#define __NR_execve 221
+#define __NR_execveat 281
+#define __NR_exit 93
+#define __NR_exit_group 94
+#define __NR_faccessat 48
+#define __NR_fadvise64_64 223
+#define __NR_fallocate 47
+#define __NR_fanotify_init 262
+#define __NR_fanotify_mark 263
+#define __NR_fchdir 50
+#define __NR_fchmod 52
+#define __NR_fchmodat 53
+#define __NR_fchown 55
+#define __NR_fchownat 54
+#define __NR_fcntl64 25
+#define __NR_fdatasync 83
+#define __NR_fgetxattr 10
+#define __NR_finit_module 273
+#define __NR_flistxattr 13
+#define __NR_flock 32
+#define __NR_fremovexattr 16
+#define __NR_fsconfig 431
+#define __NR_fsetxattr 7
+#define __NR_fsmount 432
+#define __NR_fsopen 430
+#define __NR_fspick 433
+#define __NR_fstatfs64 44
+#define __NR_fsync 82
+#define __NR_ftruncate64 46
+#define __NR_futex_time64 422
+#define __NR_get_mempolicy 236
+#define __NR_get_robust_list 100
+#define __NR_getcpu 168
+#define __NR_getcwd 17
+#define __NR_getdents64 61
+#define __NR_getegid 177
+#define __NR_geteuid 175
+#define __NR_getgid 176
+#define __NR_getgroups 158
+#define __NR_getitimer 102
+#define __NR_getpeername 205
+#define __NR_getpgid 155
+#define __NR_getpid 172
+#define __NR_getppid 173
+#define __NR_getpriority 141
+#define __NR_getrandom 278
+#define __NR_getresgid 150
+#define __NR_getresuid 148
+#define __NR_getrusage 165
+#define __NR_getsid 156
+#define __NR_getsockname 204
+#define __NR_getsockopt 209
+#define __NR_gettid 178
+#define __NR_getuid 174
+#define __NR_getxattr 8
+#define __NR_init_module 105
+#define __NR_inotify_add_watch 27
+#define __NR_inotify_init1 26
+#define __NR_inotify_rm_watch 28
+#define __NR_io_cancel 3
+#define __NR_io_destroy 1
+#define __NR_io_pgetevents_time64 416
+#define __NR_io_setup 0
+#define __NR_io_submit 2
+#define __NR_io_uring_enter 426
+#define __NR_io_uring_register 427
+#define __NR_io_uring_setup 425
+#define __NR_ioctl 29
+#define __NR_ioprio_get 31
+#define __NR_ioprio_set 30
+#define __NR_kcmp 272
+#define __NR_kexec_file_load 294
+#define __NR_kexec_load 104
+#define __NR_keyctl 219
+#define __NR_kill 129
+#define __NR_lgetxattr 9
+#define __NR_linkat 37
+#define __NR_listen 201
+#define __NR_listxattr 11
+#define __NR_llistxattr 12
+#define __NR_llseek 62
+#define __NR_lookup_dcookie 18
+#define __NR_lremovexattr 15
+#define __NR_lsetxattr 6
+#define __NR_madvise 233
+#define __NR_mbind 235
+#define __NR_membarrier 283
+#define __NR_memfd_create 279
+#define __NR_migrate_pages 238
+#define __NR_mincore 232
+#define __NR_mkdirat 34
+#define __NR_mknodat 33
+#define __NR_mlock 228
+#define __NR_mlock2 284
+#define __NR_mlockall 230
+#define __NR_mmap2 222
+#define __NR_mount 40
+#define __NR_move_mount 429
+#define __NR_move_pages 239
+#define __NR_mprotect 226
+#define __NR_mq_getsetattr 185
+#define __NR_mq_notify 184
+#define __NR_mq_open 180
+#define __NR_mq_timedreceive_time64 419
+#define __NR_mq_timedsend_time64 418
+#define __NR_mq_unlink 181
+#define __NR_mremap 216
+#define __NR_msgctl 187
+#define __NR_msgget 186
+#define __NR_msgrcv 188
+#define __NR_msgsnd 189
+#define __NR_msync 227
+#define __NR_munlock 229
+#define __NR_munlockall 231
+#define __NR_munmap 215
+#define __NR_name_to_handle_at 264
+#define __NR_nfsservctl 42
+#define __NR_open_by_handle_at 265
+#define __NR_open_tree 428
+#define __NR_openat 56
+#define __NR_openat2 437
+#define __NR_perf_event_open 241
+#define __NR_personality 92
+#define __NR_pidfd_getfd 438
+#define __NR_pidfd_open 434
+#define __NR_pidfd_send_signal 424
+#define __NR_pipe2 59
+#define __NR_pivot_root 41
+#define __NR_pkey_alloc 289
+#define __NR_pkey_free 290
+#define __NR_pkey_mprotect 288
+#define __NR_ppoll_time64 414
+#define __NR_prctl 167
+#define __NR_pread64 67
+#define __NR_preadv 69
+#define __NR_preadv2 286
+#define __NR_prlimit64 261
+#define __NR_process_vm_readv 270
+#define __NR_process_vm_writev 271
+#define __NR_pselect6_time64 413
+#define __NR_ptrace 117
+#define __NR_pwrite64 68
+#define __NR_pwritev 70
+#define __NR_pwritev2 287
+#define __NR_quotactl 60
+#define __NR_read 63
+#define __NR_readahead 213
+#define __NR_readlinkat 78
+#define __NR_readv 65
+#define __NR_reboot 142
+#define __NR_recvfrom 207
+#define __NR_recvmmsg_time64 417
+#define __NR_recvmsg 212
+#define __NR_remap_file_pages 234
+#define __NR_removexattr 14
+#define __NR_renameat2 276
+#define __NR_request_key 218
+#define __NR_restart_syscall 128
+#define __NR_riscv_flush_icache 259
+#define __NR_rseq 293
+#define __NR_rt_sigaction 134
+#define __NR_rt_sigpending 136
+#define __NR_rt_sigprocmask 135
+#define __NR_rt_sigqueueinfo 138
+#define __NR_rt_sigreturn 139
+#define __NR_rt_sigsuspend 133
+#define __NR_rt_sigtimedwait_time64 421
+#define __NR_rt_tgsigqueueinfo 240
+#define __NR_sched_get_priority_max 125
+#define __NR_sched_get_priority_min 126
+#define __NR_sched_getaffinity 123
+#define __NR_sched_getattr 275
+#define __NR_sched_getparam 121
+#define __NR_sched_getscheduler 120
+#define __NR_sched_rr_get_interval_time64 423
+#define __NR_sched_setaffinity 122
+#define __NR_sched_setattr 274
+#define __NR_sched_setparam 118
+#define __NR_sched_setscheduler 119
+#define __NR_sched_yield 124
+#define __NR_seccomp 277
+#define __NR_semctl 191
+#define __NR_semget 190
+#define __NR_semop 193
+#define __NR_semtimedop_time64 420
+#define __NR_sendfile64 71
+#define __NR_sendmmsg 269
+#define __NR_sendmsg 211
+#define __NR_sendto 206
+#define __NR_set_mempolicy 237
+#define __NR_set_robust_list 99
+#define __NR_set_tid_address 96
+#define __NR_setdomainname 162
+#define __NR_setfsgid 152
+#define __NR_setfsuid 151
+#define __NR_setgid 144
+#define __NR_setgroups 159
+#define __NR_sethostname 161
+#define __NR_setitimer 103
+#define __NR_setns 268
+#define __NR_setpgid 154
+#define __NR_setpriority 140
+#define __NR_setregid 143
+#define __NR_setresgid 149
+#define __NR_setresuid 147
+#define __NR_setreuid 145
+#define __NR_setsid 157
+#define __NR_setsockopt 208
+#define __NR_setuid 146
+#define __NR_setxattr 5
+#define __NR_shmat 196
+#define __NR_shmctl 195
+#define __NR_shmdt 197
+#define __NR_shmget 194
+#define __NR_shutdown 210
+#define __NR_sigaltstack 132
+#define __NR_signalfd4 74
+#define __NR_socket 198
+#define __NR_socketpair 199
+#define __NR_splice 76
+#define __NR_statfs64 43
+#define __NR_statx 291
+#define __NR_swapoff 225
+#define __NR_swapon 224
+#define __NR_symlinkat 36
+#define __NR_sync 81
+#define __NR_sync_file_range 84
+#define __NR_syncfs 267
+#define __NR_sysinfo 179
+#define __NR_syslog 116
+#define __NR_tee 77
+#define __NR_tgkill 131
+#define __NR_timer_create 107
+#define __NR_timer_delete 111
+#define __NR_timer_getoverrun 109
+#define __NR_timer_gettime64 408
+#define __NR_timer_settime64 409
+#define __NR_timerfd_create 85
+#define __NR_timerfd_gettime64 410
+#define __NR_timerfd_settime64 411
+#define __NR_times 153
+#define __NR_tkill 130
+#define __NR_truncate64 45
+#define __NR_umask 166
+#define __NR_umount2 39
+#define __NR_uname 160
+#define __NR_unlinkat 35
+#define __NR_unshare 97
+#define __NR_userfaultfd 282
+#define __NR_utimensat_time64 412
+#define __NR_vhangup 58
+#define __NR_vmsplice 75
+#define __NR_waitid 95
+#define __NR_write 64
+#define __NR_writev 66
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 07/18] RISC-V: nptl: update default pthread-offsets.h
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (5 preceding siblings ...)
2020-06-03 16:25 ` [PATCH v2 06/18] RISC-V: Add arch-syscall.h for RV32 Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-07-09 0:14 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 08/18] riscv32: Add an architecture ipctypes.h Alistair Francis via Libc-alpha
` (11 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
Update the RISC-V pthread-offsets.h values to support RV32.
---
sysdeps/riscv/nptl/pthread-offsets.h | 13 +++++++++++--
1 file changed, 11 insertions(+), 2 deletions(-)
diff --git a/sysdeps/riscv/nptl/pthread-offsets.h b/sysdeps/riscv/nptl/pthread-offsets.h
index 31f0587bec..a85c752a1f 100644
--- a/sysdeps/riscv/nptl/pthread-offsets.h
+++ b/sysdeps/riscv/nptl/pthread-offsets.h
@@ -1,3 +1,12 @@
-#define __PTHREAD_MUTEX_KIND_OFFSET 16
+#if __WORDSIZE == 64
+# define __PTHREAD_MUTEX_KIND_OFFSET 16
+#else
+# define __PTHREAD_MUTEX_KIND_OFFSET 12
+#endif
-#define __PTHREAD_RWLOCK_FLAGS_OFFSET 48
+
+#if __WORDSIZE == 64
+# define __PTHREAD_RWLOCK_FLAGS_OFFSET 48
+#else
+# define __PTHREAD_RWLOCK_FLAGS_OFFSET 24
+#endif
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 08/18] riscv32: Add an architecture ipctypes.h
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (6 preceding siblings ...)
2020-06-03 16:25 ` [PATCH v2 07/18] RISC-V: nptl: update default pthread-offsets.h Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-07-09 2:46 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 09/18] RISC-V: The ABI implementation for 32-bit Alistair Francis via Libc-alpha
` (10 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
Add a RISC-V specific ipctypes.h file so we can define __IPC_CMD_TIME64
as 0.
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
---
sysdeps/unix/sysv/linux/riscv/bits/ipctypes.h | 35 +++++++++++++++++++
1 file changed, 35 insertions(+)
create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/ipctypes.h
diff --git a/sysdeps/unix/sysv/linux/riscv/bits/ipctypes.h b/sysdeps/unix/sysv/linux/riscv/bits/ipctypes.h
new file mode 100644
index 0000000000..c73019bffb
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/bits/ipctypes.h
@@ -0,0 +1,35 @@
+/* bits/ipctypes.h -- Define some types used by SysV IPC/MSG/SHM.
+ Copyright (C) 2012-2020 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <https://www.gnu.org/licenses/>. */
+
+#ifndef _SYS_IPC_H
+# error "Never use <bits/ipctypes.h> directly; include <sys/ipc.h> instead."
+#endif
+
+#ifndef _BITS_IPCTYPES_H
+#define _BITS_IPCTYPES_H 1
+
+/* Used in `struct shmid_ds'. */
+# if __WORDSIZE == 32
+typedef unsigned short int __ipc_pid_t;
+# else
+typedef int __ipc_pid_t;
+# endif
+
+#define __IPC_CMD_TIME64 0x0
+
+#endif /* bits/ipctypes.h */
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 09/18] RISC-V: The ABI implementation for 32-bit
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (7 preceding siblings ...)
2020-06-03 16:25 ` [PATCH v2 08/18] riscv32: Add an architecture ipctypes.h Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-07-09 23:33 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 10/18] RISC-V: Hard float support " Alistair Francis via Libc-alpha
` (9 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
From: Zong Li <zongbox@gmail.com>
This patch adds the ABI implementation about 32 bit version. It contains
the Linux-specific and RISC-V architecture code, I've collected here.
---
sysdeps/riscv/bits/wordsize.h | 4 +-
sysdeps/riscv/nptl/bits/pthreadtypes-arch.h | 10 +++-
sysdeps/riscv/nptl/bits/struct_rwlock.h | 27 +++++++++-
sysdeps/riscv/sfp-machine.h | 27 +++++++++-
sysdeps/riscv/sys/asm.h | 5 +-
.../unix/sysv/linux/riscv/jmp_buf-macros.h | 53 +++++++++++++++++++
6 files changed, 121 insertions(+), 5 deletions(-)
create mode 100644 sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h
diff --git a/sysdeps/riscv/bits/wordsize.h b/sysdeps/riscv/bits/wordsize.h
index faccc71828..ee430d9036 100644
--- a/sysdeps/riscv/bits/wordsize.h
+++ b/sysdeps/riscv/bits/wordsize.h
@@ -25,5 +25,7 @@
#if __riscv_xlen == 64
# define __WORDSIZE_TIME64_COMPAT32 1
#else
-# error "rv32i-based targets are not supported"
+# define __WORDSIZE_TIME64_COMPAT32 0
+# define __WORDSIZE32_SIZE_ULONG 0
+# define __WORDSIZE32_PTRDIFF_LONG 0
#endif
diff --git a/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h b/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h
index c3c72d6c10..363034c38a 100644
--- a/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h
+++ b/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h
@@ -32,7 +32,15 @@
# define __SIZEOF_PTHREAD_BARRIER_T 32
# define __SIZEOF_PTHREAD_BARRIERATTR_T 4
#else
-# error "rv32i-based systems are not supported"
+# define __SIZEOF_PTHREAD_ATTR_T 32
+# define __SIZEOF_PTHREAD_MUTEX_T 32
+# define __SIZEOF_PTHREAD_MUTEXATTR_T 4
+# define __SIZEOF_PTHREAD_COND_T 48
+# define __SIZEOF_PTHREAD_CONDATTR_T 4
+# define __SIZEOF_PTHREAD_RWLOCK_T 48
+# define __SIZEOF_PTHREAD_RWLOCKATTR_T 8
+# define __SIZEOF_PTHREAD_BARRIER_T 20
+# define __SIZEOF_PTHREAD_BARRIERATTR_T 4
#endif
#define __LOCK_ALIGNMENT
diff --git a/sysdeps/riscv/nptl/bits/struct_rwlock.h b/sysdeps/riscv/nptl/bits/struct_rwlock.h
index acfaa75e1b..b478da0132 100644
--- a/sysdeps/riscv/nptl/bits/struct_rwlock.h
+++ b/sysdeps/riscv/nptl/bits/struct_rwlock.h
@@ -32,14 +32,39 @@ struct __pthread_rwlock_arch_t
unsigned int __writers_futex;
unsigned int __pad3;
unsigned int __pad4;
+#if __riscv_xlen == 64
int __cur_writer;
int __shared;
unsigned long int __pad1;
unsigned long int __pad2;
unsigned int __flags;
+#else
+# if __BYTE_ORDER == __BIG_ENDIAN
+ unsigned char __pad1;
+ unsigned char __pad2;
+ unsigned char __shared;
+ unsigned char __flags;
+# else
+ unsigned char __flags;
+ unsigned char __shared;
+ unsigned char __pad1;
+ unsigned char __pad2;
+# endif
+ int __cur_writer;
+#endif
};
-#define __PTHREAD_RWLOCK_INITIALIZER(__flags) \
+#if __riscv_xlen == 64
+# define __PTHREAD_RWLOCK_INITIALIZER(__flags) \
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, __flags
+#else
+# if __BYTE_ORDER == __BIG_ENDIAN
+# define __PTHREAD_RWLOCK_INITIALIZER(__flags) \
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, __flags, 0
+# else
+# define __PTHREAD_RWLOCK_INITIALIZER(__flags) \
+ 0, 0, 0, 0, 0, 0, __flags, 0, 0, 0, 0
+# endif
+#endif
#endif
diff --git a/sysdeps/riscv/sfp-machine.h b/sysdeps/riscv/sfp-machine.h
index 08a84fd701..aef8c61a67 100644
--- a/sysdeps/riscv/sfp-machine.h
+++ b/sysdeps/riscv/sfp-machine.h
@@ -22,7 +22,32 @@
#if __riscv_xlen == 32
-# error "rv32i-based targets are not supported"
+# define _FP_W_TYPE_SIZE 32
+# define _FP_W_TYPE unsigned long
+# define _FP_WS_TYPE signed long
+# define _FP_I_TYPE long
+
+# define _FP_MUL_MEAT_S(R, X, Y) \
+ _FP_MUL_MEAT_1_wide (_FP_WFRACBITS_S, R, X, Y, umul_ppmm)
+# define _FP_MUL_MEAT_D(R, X, Y) \
+ _FP_MUL_MEAT_2_wide (_FP_WFRACBITS_D, R, X, Y, umul_ppmm)
+# define _FP_MUL_MEAT_Q(R, X, Y) \
+ _FP_MUL_MEAT_4_wide (_FP_WFRACBITS_Q, R, X, Y, umul_ppmm)
+
+# define _FP_MUL_MEAT_DW_S(R, X, Y) \
+ _FP_MUL_MEAT_DW_1_wide (_FP_WFRACBITS_S, R, X, Y, umul_ppmm)
+# define _FP_MUL_MEAT_DW_D(R, X, Y) \
+ _FP_MUL_MEAT_DW_2_wide (_FP_WFRACBITS_D, R, X, Y, umul_ppmm)
+# define _FP_MUL_MEAT_DW_Q(R, X, Y) \
+ _FP_MUL_MEAT_DW_4_wide (_FP_WFRACBITS_Q, R, X, Y, umul_ppmm)
+
+# define _FP_DIV_MEAT_S(R, X, Y) _FP_DIV_MEAT_1_udiv_norm (S, R, X, Y)
+# define _FP_DIV_MEAT_D(R, X, Y) _FP_DIV_MEAT_2_udiv (D, R, X, Y)
+# define _FP_DIV_MEAT_Q(R, X, Y) _FP_DIV_MEAT_4_udiv (Q, R, X, Y)
+
+# define _FP_NANFRAC_S _FP_QNANBIT_S
+# define _FP_NANFRAC_D _FP_QNANBIT_D, 0
+# define _FP_NANFRAC_Q _FP_QNANBIT_Q, 0, 0, 0
#else
diff --git a/sysdeps/riscv/sys/asm.h b/sysdeps/riscv/sys/asm.h
index bd2de17e17..e314133ce4 100644
--- a/sysdeps/riscv/sys/asm.h
+++ b/sysdeps/riscv/sys/asm.h
@@ -26,7 +26,10 @@
# define REG_S sd
# define REG_L ld
#elif __riscv_xlen == 32
-# error "rv32i-based targets are not supported"
+# define PTRLOG 2
+# define SZREG 4
+# define REG_S sw
+# define REG_L lw
#else
# error __riscv_xlen must equal 32 or 64
#endif
diff --git a/sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h b/sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h
new file mode 100644
index 0000000000..7e48f24345
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h
@@ -0,0 +1,53 @@
+/* jump buffer constants for RISC-V
+ Copyright (C) 2017-2020 Free Software Foundation, Inc.
+
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library. If not, see
+ <http://www.gnu.org/licenses/>. */
+
+/* Produced by this program:
+
+ #include <stdio.h>
+ #include <unistd.h>
+ #include <setjmp.h>
+ #include <stddef.h>
+
+ int main (int argc, char **argv)
+ {
+ printf ("#define JMP_BUF_SIZE %d\n", sizeof (jmp_buf));
+ printf ("#define JMP_BUF_ALIGN %d\n", __alignof__ (jmp_buf));
+ printf ("#define SIGJMP_BUF_SIZE %d\n", sizeof (sigjmp_buf));
+ printf ("#define SIGJMP_BUF_ALIGN %d\n", __alignof__ (sigjmp_buf));
+ printf ("#define MASK_WAS_SAVED_OFFSET %d\n", offsetof (struct __jmp_buf_tag, __mask_was_saved));
+ printf ("#define SAVED_MASK_OFFSET %d\n", offsetof (struct __jmp_buf_tag, __saved_mask));
+ } */
+
+#if defined __riscv_float_abi_soft
+# define JMP_BUF_SIZE 188
+# define JMP_BUF_ALIGN 4
+# define SIGJMP_BUF_SIZE 188
+# define SIGJMP_BUF_ALIGN 4
+# define MASK_WAS_SAVED_OFFSET 56
+# define SAVED_MASK_OFFSET 60
+#elif defined __riscv_float_abi_double
+# define JMP_BUF_SIZE 288
+# define JMP_BUF_ALIGN 8
+# define SIGJMP_BUF_SIZE 288
+# define SIGJMP_BUF_ALIGN 8
+# define MASK_WAS_SAVED_OFFSET 152
+# define SAVED_MASK_OFFSET 156
+#else
+# error "Unknown RISC-V floating-point ABI"
+#endif
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 10/18] RISC-V: Hard float support for 32-bit
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (8 preceding siblings ...)
2020-06-03 16:25 ` [PATCH v2 09/18] RISC-V: The ABI implementation for 32-bit Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-07-11 0:49 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 11/18] RISC-V: Add ABI lists Alistair Francis via Libc-alpha
` (8 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
From: Zong Li <zongbox@gmail.com>
This patch contains hardware floating-point support for the RV32IF and
RV32IFD
---
sysdeps/riscv/rv32/rvd/s_lrint.c | 31 ++++++++++++++++++++++++++++++
sysdeps/riscv/rv32/rvd/s_lround.c | 31 ++++++++++++++++++++++++++++++
sysdeps/riscv/rv32/rvf/s_lrintf.c | 31 ++++++++++++++++++++++++++++++
sysdeps/riscv/rv32/rvf/s_lroundf.c | 31 ++++++++++++++++++++++++++++++
4 files changed, 124 insertions(+)
create mode 100644 sysdeps/riscv/rv32/rvd/s_lrint.c
create mode 100644 sysdeps/riscv/rv32/rvd/s_lround.c
create mode 100644 sysdeps/riscv/rv32/rvf/s_lrintf.c
create mode 100644 sysdeps/riscv/rv32/rvf/s_lroundf.c
diff --git a/sysdeps/riscv/rv32/rvd/s_lrint.c b/sysdeps/riscv/rv32/rvd/s_lrint.c
new file mode 100644
index 0000000000..df406aacb6
--- /dev/null
+++ b/sysdeps/riscv/rv32/rvd/s_lrint.c
@@ -0,0 +1,31 @@
+/* lrint(). RISC-V version.
+ Copyright (C) 2017-2020 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library. If not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include <math.h>
+#include <libm-alias-double.h>
+#include <stdint.h>
+
+long int
+__lrint (double x)
+{
+ int32_t res;
+ asm ("fcvt.w.d %0, %1" : "=r" (res) : "f" (x));
+ return res;
+}
+
+libm_alias_double (__lrint, lrint)
diff --git a/sysdeps/riscv/rv32/rvd/s_lround.c b/sysdeps/riscv/rv32/rvd/s_lround.c
new file mode 100644
index 0000000000..72aa2b179d
--- /dev/null
+++ b/sysdeps/riscv/rv32/rvd/s_lround.c
@@ -0,0 +1,31 @@
+/* lround(). RISC-V version.
+ Copyright (C) 2017-2020 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library. If not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include <math.h>
+#include <libm-alias-double.h>
+#include <stdint.h>
+
+long int
+__lround (double x)
+{
+ int32_t res;
+ asm ("fcvt.w.d %0, %1, rmm" : "=r" (res) : "f" (x));
+ return res;
+}
+
+libm_alias_double (__lround, lround)
diff --git a/sysdeps/riscv/rv32/rvf/s_lrintf.c b/sysdeps/riscv/rv32/rvf/s_lrintf.c
new file mode 100644
index 0000000000..08d44fa738
--- /dev/null
+++ b/sysdeps/riscv/rv32/rvf/s_lrintf.c
@@ -0,0 +1,31 @@
+/* lrintf(). RISC-V version.
+ Copyright (C) 2017-2018 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library. If not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include <math.h>
+#include <libm-alias-float.h>
+#include <stdint.h>
+
+long int
+__lrintf (float x)
+{
+ int32_t res;
+ asm ("fcvt.w.s %0, %1" : "=r" (res) : "f" (x));
+ return res;
+}
+
+libm_alias_float (__lrint, lrint)
diff --git a/sysdeps/riscv/rv32/rvf/s_lroundf.c b/sysdeps/riscv/rv32/rvf/s_lroundf.c
new file mode 100644
index 0000000000..9ff5d63d1c
--- /dev/null
+++ b/sysdeps/riscv/rv32/rvf/s_lroundf.c
@@ -0,0 +1,31 @@
+/* lroundf(). RISC-V version.
+ Copyright (C) 2017-2020 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library. If not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include <math.h>
+#include <libm-alias-float.h>
+#include <stdint.h>
+
+long int
+__lroundf (float x)
+{
+ int32_t res;
+ asm ("fcvt.w.s %0, %1, rmm" : "=r" (res) : "f" (x));
+ return res;
+}
+
+libm_alias_float (__lround, lround)
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 11/18] RISC-V: Add ABI lists
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (9 preceding siblings ...)
2020-06-03 16:25 ` [PATCH v2 10/18] RISC-V: Hard float support " Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-07-12 20:54 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 12/18] RISC-V: Add the RV32 libm-test-ulps Alistair Francis via Libc-alpha
` (7 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
Use the check-api and update-abi to generate the abilist for rv32.
---
sysdeps/unix/sysv/linux/riscv/c++-types.data | 67 +
sysdeps/unix/sysv/linux/riscv/ld.abilist | 5 +
.../sysv/linux/riscv/libBrokenLocale.abilist | 1 +
sysdeps/unix/sysv/linux/riscv/libanl.abilist | 4 +
sysdeps/unix/sysv/linux/riscv/libc.abilist | 2099 +++++++++++++++++
.../unix/sysv/linux/riscv/libcrypt.abilist | 2 +
sysdeps/unix/sysv/linux/riscv/libdl.abilist | 9 +
sysdeps/unix/sysv/linux/riscv/libm.abilist | 940 ++++++++
.../unix/sysv/linux/riscv/libpthread.abilist | 213 ++
.../unix/sysv/linux/riscv/libresolv.abilist | 79 +
sysdeps/unix/sysv/linux/riscv/librt.abilist | 35 +
.../sysv/linux/riscv/libthread_db.abilist | 40 +
sysdeps/unix/sysv/linux/riscv/libutil.abilist | 6 +
13 files changed, 3500 insertions(+)
create mode 100644 sysdeps/unix/sysv/linux/riscv/c++-types.data
create mode 100644 sysdeps/unix/sysv/linux/riscv/ld.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libBrokenLocale.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libanl.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libc.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libcrypt.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libdl.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libm.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libpthread.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libresolv.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/librt.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libthread_db.abilist
create mode 100644 sysdeps/unix/sysv/linux/riscv/libutil.abilist
diff --git a/sysdeps/unix/sysv/linux/riscv/c++-types.data b/sysdeps/unix/sysv/linux/riscv/c++-types.data
new file mode 100644
index 0000000000..0fb0143ae7
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/c++-types.data
@@ -0,0 +1,67 @@
+blkcnt64_t:x
+blkcnt_t:x
+blksize_t:i
+caddr_t:Pc
+clockid_t:i
+clock_t:l
+daddr_t:i
+dev_t:y
+fd_mask:l
+fsblkcnt64_t:y
+fsblkcnt_t:y
+fsfilcnt64_t:y
+fsfilcnt_t:y
+fsid_t:8__fsid_t
+gid_t:j
+id_t:j
+ino64_t:y
+ino_t:y
+int16_t:s
+int32_t:i
+int64_t:x
+int8_t:a
+intptr_t:i
+key_t:i
+loff_t:x
+mode_t:j
+nlink_t:j
+off64_t:x
+off_t:x
+pid_t:i
+pthread_attr_t:14pthread_attr_t
+pthread_barrier_t:17pthread_barrier_t
+pthread_barrierattr_t:21pthread_barrierattr_t
+pthread_cond_t:14pthread_cond_t
+pthread_condattr_t:18pthread_condattr_t
+pthread_key_t:j
+pthread_mutex_t:15pthread_mutex_t
+pthread_mutexattr_t:19pthread_mutexattr_t
+pthread_once_t:i
+pthread_rwlock_t:16pthread_rwlock_t
+pthread_rwlockattr_t:20pthread_rwlockattr_t
+pthread_spinlock_t:i
+pthread_t:m
+quad_t:x
+register_t:i
+rlim64_t:y
+rlim_t:y
+sigset_t:10__sigset_t
+size_t:j
+socklen_t:j
+ssize_t:i
+suseconds_t:x
+time_t:x
+u_char:h
+uid_t:j
+uint:j
+u_int:j
+u_int16_t:t
+u_int32_t:j
+u_int64_t:y
+u_int8_t:h
+ulong:m
+u_long:m
+u_quad_t:y
+useconds_t:j
+ushort:t
+u_short:t
diff --git a/sysdeps/unix/sysv/linux/riscv/ld.abilist b/sysdeps/unix/sysv/linux/riscv/ld.abilist
new file mode 100644
index 0000000000..048f17c848
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/ld.abilist
@@ -0,0 +1,5 @@
+GLIBC_2.32 __libc_stack_end D 0x4
+GLIBC_2.32 __stack_chk_guard D 0x4
+GLIBC_2.32 __tls_get_addr F
+GLIBC_2.32 _dl_mcount F
+GLIBC_2.32 _r_debug D 0x14
diff --git a/sysdeps/unix/sysv/linux/riscv/libBrokenLocale.abilist b/sysdeps/unix/sysv/linux/riscv/libBrokenLocale.abilist
new file mode 100644
index 0000000000..b0869cec1f
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/libBrokenLocale.abilist
@@ -0,0 +1 @@
+GLIBC_2.32 __ctype_get_mb_cur_max F
diff --git a/sysdeps/unix/sysv/linux/riscv/libanl.abilist b/sysdeps/unix/sysv/linux/riscv/libanl.abilist
new file mode 100644
index 0000000000..ba513bd028
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/libanl.abilist
@@ -0,0 +1,4 @@
+GLIBC_2.32 gai_cancel F
+GLIBC_2.32 gai_error F
+GLIBC_2.32 gai_suspend F
+GLIBC_2.32 getaddrinfo_a F
diff --git a/sysdeps/unix/sysv/linux/riscv/libc.abilist b/sysdeps/unix/sysv/linux/riscv/libc.abilist
new file mode 100644
index 0000000000..9f354d8438
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/libc.abilist
@@ -0,0 +1,2099 @@
+GLIBC_2.32 _Exit F
+GLIBC_2.32 _IO_2_1_stderr_ D 0xa0
+GLIBC_2.32 _IO_2_1_stdin_ D 0xa0
+GLIBC_2.32 _IO_2_1_stdout_ D 0xa0
+GLIBC_2.32 _IO_adjust_column F
+GLIBC_2.32 _IO_adjust_wcolumn F
+GLIBC_2.32 _IO_default_doallocate F
+GLIBC_2.32 _IO_default_finish F
+GLIBC_2.32 _IO_default_pbackfail F
+GLIBC_2.32 _IO_default_uflow F
+GLIBC_2.32 _IO_default_xsgetn F
+GLIBC_2.32 _IO_default_xsputn F
+GLIBC_2.32 _IO_do_write F
+GLIBC_2.32 _IO_doallocbuf F
+GLIBC_2.32 _IO_fclose F
+GLIBC_2.32 _IO_fdopen F
+GLIBC_2.32 _IO_feof F
+GLIBC_2.32 _IO_ferror F
+GLIBC_2.32 _IO_fflush F
+GLIBC_2.32 _IO_fgetpos F
+GLIBC_2.32 _IO_fgetpos64 F
+GLIBC_2.32 _IO_fgets F
+GLIBC_2.32 _IO_file_attach F
+GLIBC_2.32 _IO_file_close F
+GLIBC_2.32 _IO_file_close_it F
+GLIBC_2.32 _IO_file_doallocate F
+GLIBC_2.32 _IO_file_finish F
+GLIBC_2.32 _IO_file_fopen F
+GLIBC_2.32 _IO_file_init F
+GLIBC_2.32 _IO_file_jumps D 0x54
+GLIBC_2.32 _IO_file_open F
+GLIBC_2.32 _IO_file_overflow F
+GLIBC_2.32 _IO_file_read F
+GLIBC_2.32 _IO_file_seek F
+GLIBC_2.32 _IO_file_seekoff F
+GLIBC_2.32 _IO_file_setbuf F
+GLIBC_2.32 _IO_file_stat F
+GLIBC_2.32 _IO_file_sync F
+GLIBC_2.32 _IO_file_underflow F
+GLIBC_2.32 _IO_file_write F
+GLIBC_2.32 _IO_file_xsputn F
+GLIBC_2.32 _IO_flockfile F
+GLIBC_2.32 _IO_flush_all F
+GLIBC_2.32 _IO_flush_all_linebuffered F
+GLIBC_2.32 _IO_fopen F
+GLIBC_2.32 _IO_fprintf F
+GLIBC_2.32 _IO_fputs F
+GLIBC_2.32 _IO_fread F
+GLIBC_2.32 _IO_free_backup_area F
+GLIBC_2.32 _IO_free_wbackup_area F
+GLIBC_2.32 _IO_fsetpos F
+GLIBC_2.32 _IO_fsetpos64 F
+GLIBC_2.32 _IO_ftell F
+GLIBC_2.32 _IO_ftrylockfile F
+GLIBC_2.32 _IO_funlockfile F
+GLIBC_2.32 _IO_fwrite F
+GLIBC_2.32 _IO_getc F
+GLIBC_2.32 _IO_getline F
+GLIBC_2.32 _IO_getline_info F
+GLIBC_2.32 _IO_gets F
+GLIBC_2.32 _IO_init F
+GLIBC_2.32 _IO_init_marker F
+GLIBC_2.32 _IO_init_wmarker F
+GLIBC_2.32 _IO_iter_begin F
+GLIBC_2.32 _IO_iter_end F
+GLIBC_2.32 _IO_iter_file F
+GLIBC_2.32 _IO_iter_next F
+GLIBC_2.32 _IO_least_wmarker F
+GLIBC_2.32 _IO_link_in F
+GLIBC_2.32 _IO_list_all D 0x4
+GLIBC_2.32 _IO_list_lock F
+GLIBC_2.32 _IO_list_resetlock F
+GLIBC_2.32 _IO_list_unlock F
+GLIBC_2.32 _IO_marker_delta F
+GLIBC_2.32 _IO_marker_difference F
+GLIBC_2.32 _IO_padn F
+GLIBC_2.32 _IO_peekc_locked F
+GLIBC_2.32 _IO_popen F
+GLIBC_2.32 _IO_printf F
+GLIBC_2.32 _IO_proc_close F
+GLIBC_2.32 _IO_proc_open F
+GLIBC_2.32 _IO_putc F
+GLIBC_2.32 _IO_puts F
+GLIBC_2.32 _IO_remove_marker F
+GLIBC_2.32 _IO_seekmark F
+GLIBC_2.32 _IO_seekoff F
+GLIBC_2.32 _IO_seekpos F
+GLIBC_2.32 _IO_seekwmark F
+GLIBC_2.32 _IO_setb F
+GLIBC_2.32 _IO_setbuffer F
+GLIBC_2.32 _IO_setvbuf F
+GLIBC_2.32 _IO_sgetn F
+GLIBC_2.32 _IO_sprintf F
+GLIBC_2.32 _IO_sputbackc F
+GLIBC_2.32 _IO_sputbackwc F
+GLIBC_2.32 _IO_sscanf F
+GLIBC_2.32 _IO_str_init_readonly F
+GLIBC_2.32 _IO_str_init_static F
+GLIBC_2.32 _IO_str_overflow F
+GLIBC_2.32 _IO_str_pbackfail F
+GLIBC_2.32 _IO_str_seekoff F
+GLIBC_2.32 _IO_str_underflow F
+GLIBC_2.32 _IO_sungetc F
+GLIBC_2.32 _IO_sungetwc F
+GLIBC_2.32 _IO_switch_to_get_mode F
+GLIBC_2.32 _IO_switch_to_main_wget_area F
+GLIBC_2.32 _IO_switch_to_wbackup_area F
+GLIBC_2.32 _IO_switch_to_wget_mode F
+GLIBC_2.32 _IO_un_link F
+GLIBC_2.32 _IO_ungetc F
+GLIBC_2.32 _IO_unsave_markers F
+GLIBC_2.32 _IO_unsave_wmarkers F
+GLIBC_2.32 _IO_vfprintf F
+GLIBC_2.32 _IO_vsprintf F
+GLIBC_2.32 _IO_wdefault_doallocate F
+GLIBC_2.32 _IO_wdefault_finish F
+GLIBC_2.32 _IO_wdefault_pbackfail F
+GLIBC_2.32 _IO_wdefault_uflow F
+GLIBC_2.32 _IO_wdefault_xsgetn F
+GLIBC_2.32 _IO_wdefault_xsputn F
+GLIBC_2.32 _IO_wdo_write F
+GLIBC_2.32 _IO_wdoallocbuf F
+GLIBC_2.32 _IO_wfile_jumps D 0x54
+GLIBC_2.32 _IO_wfile_overflow F
+GLIBC_2.32 _IO_wfile_seekoff F
+GLIBC_2.32 _IO_wfile_sync F
+GLIBC_2.32 _IO_wfile_underflow F
+GLIBC_2.32 _IO_wfile_xsputn F
+GLIBC_2.32 _IO_wmarker_delta F
+GLIBC_2.32 _IO_wsetb F
+GLIBC_2.32 ___brk_addr D 0x4
+GLIBC_2.32 __adjtimex F
+GLIBC_2.32 __after_morecore_hook D 0x4
+GLIBC_2.32 __argz_count F
+GLIBC_2.32 __argz_next F
+GLIBC_2.32 __argz_stringify F
+GLIBC_2.32 __asprintf F
+GLIBC_2.32 __asprintf_chk F
+GLIBC_2.32 __assert F
+GLIBC_2.32 __assert_fail F
+GLIBC_2.32 __assert_perror_fail F
+GLIBC_2.32 __backtrace F
+GLIBC_2.32 __backtrace_symbols F
+GLIBC_2.32 __backtrace_symbols_fd F
+GLIBC_2.32 __bsd_getpgrp F
+GLIBC_2.32 __bzero F
+GLIBC_2.32 __check_rhosts_file D 0x4
+GLIBC_2.32 __chk_fail F
+GLIBC_2.32 __clone F
+GLIBC_2.32 __close F
+GLIBC_2.32 __cmsg_nxthdr F
+GLIBC_2.32 __confstr_chk F
+GLIBC_2.32 __connect F
+GLIBC_2.32 __ctype_b_loc F
+GLIBC_2.32 __ctype_get_mb_cur_max F
+GLIBC_2.32 __ctype_tolower_loc F
+GLIBC_2.32 __ctype_toupper_loc F
+GLIBC_2.32 __curbrk D 0x4
+GLIBC_2.32 __cxa_at_quick_exit F
+GLIBC_2.32 __cxa_atexit F
+GLIBC_2.32 __cxa_finalize F
+GLIBC_2.32 __cxa_thread_atexit_impl F
+GLIBC_2.32 __cyg_profile_func_enter F
+GLIBC_2.32 __cyg_profile_func_exit F
+GLIBC_2.32 __daylight D 0x4
+GLIBC_2.32 __dcgettext F
+GLIBC_2.32 __default_morecore F
+GLIBC_2.32 __dgettext F
+GLIBC_2.32 __dprintf_chk F
+GLIBC_2.32 __dup2 F
+GLIBC_2.32 __duplocale F
+GLIBC_2.32 __endmntent F
+GLIBC_2.32 __environ D 0x4
+GLIBC_2.32 __errno_location F
+GLIBC_2.32 __explicit_bzero_chk F
+GLIBC_2.32 __fbufsize F
+GLIBC_2.32 __fcntl F
+GLIBC_2.32 __fdelt_chk F
+GLIBC_2.32 __fdelt_warn F
+GLIBC_2.32 __ffs F
+GLIBC_2.32 __fgets_chk F
+GLIBC_2.32 __fgets_unlocked_chk F
+GLIBC_2.32 __fgetws_chk F
+GLIBC_2.32 __fgetws_unlocked_chk F
+GLIBC_2.32 __finite F
+GLIBC_2.32 __finitef F
+GLIBC_2.32 __finitel F
+GLIBC_2.32 __flbf F
+GLIBC_2.32 __fork F
+GLIBC_2.32 __fpending F
+GLIBC_2.32 __fprintf_chk F
+GLIBC_2.32 __fpu_control D 0x4
+GLIBC_2.32 __fpurge F
+GLIBC_2.32 __fread_chk F
+GLIBC_2.32 __fread_unlocked_chk F
+GLIBC_2.32 __freadable F
+GLIBC_2.32 __freading F
+GLIBC_2.32 __free_hook D 0x4
+GLIBC_2.32 __freelocale F
+GLIBC_2.32 __fsetlocking F
+GLIBC_2.32 __fwprintf_chk F
+GLIBC_2.32 __fwritable F
+GLIBC_2.32 __fwriting F
+GLIBC_2.32 __fxstat F
+GLIBC_2.32 __fxstat64 F
+GLIBC_2.32 __fxstatat F
+GLIBC_2.32 __fxstatat64 F
+GLIBC_2.32 __getauxval F
+GLIBC_2.32 __getcwd_chk F
+GLIBC_2.32 __getdelim F
+GLIBC_2.32 __getdomainname_chk F
+GLIBC_2.32 __getgroups_chk F
+GLIBC_2.32 __gethostname_chk F
+GLIBC_2.32 __getlogin_r_chk F
+GLIBC_2.32 __getmntent_r F
+GLIBC_2.32 __getpagesize F
+GLIBC_2.32 __getpgid F
+GLIBC_2.32 __getpid F
+GLIBC_2.32 __gets_chk F
+GLIBC_2.32 __gettimeofday F
+GLIBC_2.32 __getwd_chk F
+GLIBC_2.32 __gmtime_r F
+GLIBC_2.32 __h_errno_location F
+GLIBC_2.32 __isalnum_l F
+GLIBC_2.32 __isalpha_l F
+GLIBC_2.32 __isascii_l F
+GLIBC_2.32 __isblank_l F
+GLIBC_2.32 __iscntrl_l F
+GLIBC_2.32 __isctype F
+GLIBC_2.32 __isdigit_l F
+GLIBC_2.32 __isgraph_l F
+GLIBC_2.32 __isinf F
+GLIBC_2.32 __isinff F
+GLIBC_2.32 __isinfl F
+GLIBC_2.32 __islower_l F
+GLIBC_2.32 __isnan F
+GLIBC_2.32 __isnanf F
+GLIBC_2.32 __isnanl F
+GLIBC_2.32 __isoc99_fscanf F
+GLIBC_2.32 __isoc99_fwscanf F
+GLIBC_2.32 __isoc99_scanf F
+GLIBC_2.32 __isoc99_sscanf F
+GLIBC_2.32 __isoc99_swscanf F
+GLIBC_2.32 __isoc99_vfscanf F
+GLIBC_2.32 __isoc99_vfwscanf F
+GLIBC_2.32 __isoc99_vscanf F
+GLIBC_2.32 __isoc99_vsscanf F
+GLIBC_2.32 __isoc99_vswscanf F
+GLIBC_2.32 __isoc99_vwscanf F
+GLIBC_2.32 __isoc99_wscanf F
+GLIBC_2.32 __isprint_l F
+GLIBC_2.32 __ispunct_l F
+GLIBC_2.32 __isspace_l F
+GLIBC_2.32 __isupper_l F
+GLIBC_2.32 __iswalnum_l F
+GLIBC_2.32 __iswalpha_l F
+GLIBC_2.32 __iswblank_l F
+GLIBC_2.32 __iswcntrl_l F
+GLIBC_2.32 __iswctype F
+GLIBC_2.32 __iswctype_l F
+GLIBC_2.32 __iswdigit_l F
+GLIBC_2.32 __iswgraph_l F
+GLIBC_2.32 __iswlower_l F
+GLIBC_2.32 __iswprint_l F
+GLIBC_2.32 __iswpunct_l F
+GLIBC_2.32 __iswspace_l F
+GLIBC_2.32 __iswupper_l F
+GLIBC_2.32 __iswxdigit_l F
+GLIBC_2.32 __isxdigit_l F
+GLIBC_2.32 __ivaliduser F
+GLIBC_2.32 __key_decryptsession_pk_LOCAL D 0x4
+GLIBC_2.32 __key_encryptsession_pk_LOCAL D 0x4
+GLIBC_2.32 __key_gendes_LOCAL D 0x4
+GLIBC_2.32 __libc_allocate_rtsig F
+GLIBC_2.32 __libc_calloc F
+GLIBC_2.32 __libc_current_sigrtmax F
+GLIBC_2.32 __libc_current_sigrtmin F
+GLIBC_2.32 __libc_free F
+GLIBC_2.32 __libc_freeres F
+GLIBC_2.32 __libc_init_first F
+GLIBC_2.32 __libc_mallinfo F
+GLIBC_2.32 __libc_malloc F
+GLIBC_2.32 __libc_mallopt F
+GLIBC_2.32 __libc_memalign F
+GLIBC_2.32 __libc_pvalloc F
+GLIBC_2.32 __libc_realloc F
+GLIBC_2.32 __libc_sa_len F
+GLIBC_2.32 __libc_start_main F
+GLIBC_2.32 __libc_valloc F
+GLIBC_2.32 __longjmp_chk F
+GLIBC_2.32 __lseek F
+GLIBC_2.32 __lxstat F
+GLIBC_2.32 __lxstat64 F
+GLIBC_2.32 __malloc_hook D 0x4
+GLIBC_2.32 __mbrlen F
+GLIBC_2.32 __mbrtowc F
+GLIBC_2.32 __mbsnrtowcs_chk F
+GLIBC_2.32 __mbsrtowcs_chk F
+GLIBC_2.32 __mbstowcs_chk F
+GLIBC_2.32 __memalign_hook D 0x4
+GLIBC_2.32 __memcpy_chk F
+GLIBC_2.32 __memmove_chk F
+GLIBC_2.32 __mempcpy F
+GLIBC_2.32 __mempcpy_chk F
+GLIBC_2.32 __memset_chk F
+GLIBC_2.32 __monstartup F
+GLIBC_2.32 __morecore D 0x4
+GLIBC_2.32 __nanosleep F
+GLIBC_2.32 __newlocale F
+GLIBC_2.32 __nl_langinfo_l F
+GLIBC_2.32 __nss_configure_lookup F
+GLIBC_2.32 __nss_hostname_digits_dots F
+GLIBC_2.32 __obstack_printf_chk F
+GLIBC_2.32 __obstack_vprintf_chk F
+GLIBC_2.32 __open F
+GLIBC_2.32 __open64 F
+GLIBC_2.32 __open64_2 F
+GLIBC_2.32 __open_2 F
+GLIBC_2.32 __openat64_2 F
+GLIBC_2.32 __openat_2 F
+GLIBC_2.32 __overflow F
+GLIBC_2.32 __pipe F
+GLIBC_2.32 __poll F
+GLIBC_2.32 __poll_chk F
+GLIBC_2.32 __posix_getopt F
+GLIBC_2.32 __ppoll_chk F
+GLIBC_2.32 __pread64 F
+GLIBC_2.32 __pread64_chk F
+GLIBC_2.32 __pread_chk F
+GLIBC_2.32 __printf_chk F
+GLIBC_2.32 __printf_fp F
+GLIBC_2.32 __profile_frequency F
+GLIBC_2.32 __progname D 0x4
+GLIBC_2.32 __progname_full D 0x4
+GLIBC_2.32 __ptsname_r_chk F
+GLIBC_2.32 __pwrite64 F
+GLIBC_2.32 __rawmemchr F
+GLIBC_2.32 __rcmd_errstr D 0x4
+GLIBC_2.32 __read F
+GLIBC_2.32 __read_chk F
+GLIBC_2.32 __readlink_chk F
+GLIBC_2.32 __readlinkat_chk F
+GLIBC_2.32 __realloc_hook D 0x4
+GLIBC_2.32 __realpath_chk F
+GLIBC_2.32 __recv_chk F
+GLIBC_2.32 __recvfrom_chk F
+GLIBC_2.32 __register_atfork F
+GLIBC_2.32 __res_init F
+GLIBC_2.32 __res_nclose F
+GLIBC_2.32 __res_ninit F
+GLIBC_2.32 __res_randomid F
+GLIBC_2.32 __res_state F
+GLIBC_2.32 __riscv_flush_icache F
+GLIBC_2.32 __rpc_thread_createerr F
+GLIBC_2.32 __rpc_thread_svc_fdset F
+GLIBC_2.32 __rpc_thread_svc_max_pollfd F
+GLIBC_2.32 __rpc_thread_svc_pollfd F
+GLIBC_2.32 __sbrk F
+GLIBC_2.32 __sched_cpualloc F
+GLIBC_2.32 __sched_cpucount F
+GLIBC_2.32 __sched_cpufree F
+GLIBC_2.32 __sched_get_priority_max F
+GLIBC_2.32 __sched_get_priority_min F
+GLIBC_2.32 __sched_getparam F
+GLIBC_2.32 __sched_getscheduler F
+GLIBC_2.32 __sched_setscheduler F
+GLIBC_2.32 __sched_yield F
+GLIBC_2.32 __select F
+GLIBC_2.32 __send F
+GLIBC_2.32 __setmntent F
+GLIBC_2.32 __setpgid F
+GLIBC_2.32 __sigaction F
+GLIBC_2.32 __signbit F
+GLIBC_2.32 __signbitf F
+GLIBC_2.32 __signbitl F
+GLIBC_2.32 __sigpause F
+GLIBC_2.32 __sigsetjmp F
+GLIBC_2.32 __sigsuspend F
+GLIBC_2.32 __snprintf_chk F
+GLIBC_2.32 __sprintf_chk F
+GLIBC_2.32 __stack_chk_fail F
+GLIBC_2.32 __statfs F
+GLIBC_2.32 __stpcpy F
+GLIBC_2.32 __stpcpy_chk F
+GLIBC_2.32 __stpncpy F
+GLIBC_2.32 __stpncpy_chk F
+GLIBC_2.32 __strcasecmp F
+GLIBC_2.32 __strcasecmp_l F
+GLIBC_2.32 __strcasestr F
+GLIBC_2.32 __strcat_chk F
+GLIBC_2.32 __strcoll_l F
+GLIBC_2.32 __strcpy_chk F
+GLIBC_2.32 __strdup F
+GLIBC_2.32 __strerror_r F
+GLIBC_2.32 __strfmon_l F
+GLIBC_2.32 __strftime_l F
+GLIBC_2.32 __strncasecmp_l F
+GLIBC_2.32 __strncat_chk F
+GLIBC_2.32 __strncpy_chk F
+GLIBC_2.32 __strndup F
+GLIBC_2.32 __strsep_g F
+GLIBC_2.32 __strtod_internal F
+GLIBC_2.32 __strtod_l F
+GLIBC_2.32 __strtof_internal F
+GLIBC_2.32 __strtof_l F
+GLIBC_2.32 __strtok_r F
+GLIBC_2.32 __strtol_internal F
+GLIBC_2.32 __strtol_l F
+GLIBC_2.32 __strtold_internal F
+GLIBC_2.32 __strtold_l F
+GLIBC_2.32 __strtoll_internal F
+GLIBC_2.32 __strtoll_l F
+GLIBC_2.32 __strtoul_internal F
+GLIBC_2.32 __strtoul_l F
+GLIBC_2.32 __strtoull_internal F
+GLIBC_2.32 __strtoull_l F
+GLIBC_2.32 __strverscmp F
+GLIBC_2.32 __strxfrm_l F
+GLIBC_2.32 __swprintf_chk F
+GLIBC_2.32 __sysconf F
+GLIBC_2.32 __syslog_chk F
+GLIBC_2.32 __sysv_signal F
+GLIBC_2.32 __timezone D 0x4
+GLIBC_2.32 __toascii_l F
+GLIBC_2.32 __tolower_l F
+GLIBC_2.32 __toupper_l F
+GLIBC_2.32 __towctrans F
+GLIBC_2.32 __towctrans_l F
+GLIBC_2.32 __towlower_l F
+GLIBC_2.32 __towupper_l F
+GLIBC_2.32 __ttyname_r_chk F
+GLIBC_2.32 __tzname D 0x8
+GLIBC_2.32 __uflow F
+GLIBC_2.32 __underflow F
+GLIBC_2.32 __uselocale F
+GLIBC_2.32 __vasprintf_chk F
+GLIBC_2.32 __vdprintf_chk F
+GLIBC_2.32 __vfork F
+GLIBC_2.32 __vfprintf_chk F
+GLIBC_2.32 __vfscanf F
+GLIBC_2.32 __vfwprintf_chk F
+GLIBC_2.32 __vprintf_chk F
+GLIBC_2.32 __vsnprintf F
+GLIBC_2.32 __vsnprintf_chk F
+GLIBC_2.32 __vsprintf_chk F
+GLIBC_2.32 __vsscanf F
+GLIBC_2.32 __vswprintf_chk F
+GLIBC_2.32 __vsyslog_chk F
+GLIBC_2.32 __vwprintf_chk F
+GLIBC_2.32 __wait F
+GLIBC_2.32 __waitpid F
+GLIBC_2.32 __wcpcpy_chk F
+GLIBC_2.32 __wcpncpy_chk F
+GLIBC_2.32 __wcrtomb_chk F
+GLIBC_2.32 __wcscasecmp_l F
+GLIBC_2.32 __wcscat_chk F
+GLIBC_2.32 __wcscoll_l F
+GLIBC_2.32 __wcscpy_chk F
+GLIBC_2.32 __wcsftime_l F
+GLIBC_2.32 __wcsncasecmp_l F
+GLIBC_2.32 __wcsncat_chk F
+GLIBC_2.32 __wcsncpy_chk F
+GLIBC_2.32 __wcsnrtombs_chk F
+GLIBC_2.32 __wcsrtombs_chk F
+GLIBC_2.32 __wcstod_internal F
+GLIBC_2.32 __wcstod_l F
+GLIBC_2.32 __wcstof_internal F
+GLIBC_2.32 __wcstof_l F
+GLIBC_2.32 __wcstol_internal F
+GLIBC_2.32 __wcstol_l F
+GLIBC_2.32 __wcstold_internal F
+GLIBC_2.32 __wcstold_l F
+GLIBC_2.32 __wcstoll_internal F
+GLIBC_2.32 __wcstoll_l F
+GLIBC_2.32 __wcstombs_chk F
+GLIBC_2.32 __wcstoul_internal F
+GLIBC_2.32 __wcstoul_l F
+GLIBC_2.32 __wcstoull_internal F
+GLIBC_2.32 __wcstoull_l F
+GLIBC_2.32 __wcsxfrm_l F
+GLIBC_2.32 __wctomb_chk F
+GLIBC_2.32 __wctrans_l F
+GLIBC_2.32 __wctype_l F
+GLIBC_2.32 __wmemcpy_chk F
+GLIBC_2.32 __wmemmove_chk F
+GLIBC_2.32 __wmempcpy_chk F
+GLIBC_2.32 __wmemset_chk F
+GLIBC_2.32 __woverflow F
+GLIBC_2.32 __wprintf_chk F
+GLIBC_2.32 __write F
+GLIBC_2.32 __wuflow F
+GLIBC_2.32 __wunderflow F
+GLIBC_2.32 __xmknod F
+GLIBC_2.32 __xmknodat F
+GLIBC_2.32 __xpg_basename F
+GLIBC_2.32 __xpg_sigpause F
+GLIBC_2.32 __xpg_strerror_r F
+GLIBC_2.32 __xstat F
+GLIBC_2.32 __xstat64 F
+GLIBC_2.32 _authenticate F
+GLIBC_2.32 _dl_mcount_wrapper F
+GLIBC_2.32 _dl_mcount_wrapper_check F
+GLIBC_2.32 _environ D 0x4
+GLIBC_2.32 _exit F
+GLIBC_2.32 _flushlbf F
+GLIBC_2.32 _libc_intl_domainname D 0x5
+GLIBC_2.32 _longjmp F
+GLIBC_2.32 _mcleanup F
+GLIBC_2.32 _mcount F
+GLIBC_2.32 _nl_default_dirname D 0x12
+GLIBC_2.32 _nl_domain_bindings D 0x4
+GLIBC_2.32 _nl_msg_cat_cntr D 0x4
+GLIBC_2.32 _null_auth D 0xc
+GLIBC_2.32 _obstack_allocated_p F
+GLIBC_2.32 _obstack_begin F
+GLIBC_2.32 _obstack_begin_1 F
+GLIBC_2.32 _obstack_free F
+GLIBC_2.32 _obstack_memory_used F
+GLIBC_2.32 _obstack_newchunk F
+GLIBC_2.32 _res D 0x200
+GLIBC_2.32 _res_hconf D 0x30
+GLIBC_2.32 _rpc_dtablesize F
+GLIBC_2.32 _seterr_reply F
+GLIBC_2.32 _setjmp F
+GLIBC_2.32 _sys_errlist D 0x21c
+GLIBC_2.32 _sys_nerr D 0x4
+GLIBC_2.32 _sys_siglist D 0x104
+GLIBC_2.32 _tolower F
+GLIBC_2.32 _toupper F
+GLIBC_2.32 a64l F
+GLIBC_2.32 abort F
+GLIBC_2.32 abs F
+GLIBC_2.32 accept F
+GLIBC_2.32 accept4 F
+GLIBC_2.32 access F
+GLIBC_2.32 acct F
+GLIBC_2.32 addmntent F
+GLIBC_2.32 addseverity F
+GLIBC_2.32 adjtime F
+GLIBC_2.32 adjtimex F
+GLIBC_2.32 alarm F
+GLIBC_2.32 aligned_alloc F
+GLIBC_2.32 alphasort F
+GLIBC_2.32 alphasort64 F
+GLIBC_2.32 argp_err_exit_status D 0x4
+GLIBC_2.32 argp_error F
+GLIBC_2.32 argp_failure F
+GLIBC_2.32 argp_help F
+GLIBC_2.32 argp_parse F
+GLIBC_2.32 argp_program_bug_address D 0x4
+GLIBC_2.32 argp_program_version D 0x4
+GLIBC_2.32 argp_program_version_hook D 0x4
+GLIBC_2.32 argp_state_help F
+GLIBC_2.32 argp_usage F
+GLIBC_2.32 argz_add F
+GLIBC_2.32 argz_add_sep F
+GLIBC_2.32 argz_append F
+GLIBC_2.32 argz_count F
+GLIBC_2.32 argz_create F
+GLIBC_2.32 argz_create_sep F
+GLIBC_2.32 argz_delete F
+GLIBC_2.32 argz_extract F
+GLIBC_2.32 argz_insert F
+GLIBC_2.32 argz_next F
+GLIBC_2.32 argz_replace F
+GLIBC_2.32 argz_stringify F
+GLIBC_2.32 asctime F
+GLIBC_2.32 asctime_r F
+GLIBC_2.32 asprintf F
+GLIBC_2.32 atof F
+GLIBC_2.32 atoi F
+GLIBC_2.32 atol F
+GLIBC_2.32 atoll F
+GLIBC_2.32 authdes_create F
+GLIBC_2.32 authdes_getucred F
+GLIBC_2.32 authdes_pk_create F
+GLIBC_2.32 authnone_create F
+GLIBC_2.32 authunix_create F
+GLIBC_2.32 authunix_create_default F
+GLIBC_2.32 backtrace F
+GLIBC_2.32 backtrace_symbols F
+GLIBC_2.32 backtrace_symbols_fd F
+GLIBC_2.32 basename F
+GLIBC_2.32 bcmp F
+GLIBC_2.32 bcopy F
+GLIBC_2.32 bind F
+GLIBC_2.32 bind_textdomain_codeset F
+GLIBC_2.32 bindresvport F
+GLIBC_2.32 bindtextdomain F
+GLIBC_2.32 brk F
+GLIBC_2.32 bsd_signal F
+GLIBC_2.32 bsearch F
+GLIBC_2.32 btowc F
+GLIBC_2.32 bzero F
+GLIBC_2.32 c16rtomb F
+GLIBC_2.32 c32rtomb F
+GLIBC_2.32 calloc F
+GLIBC_2.32 callrpc F
+GLIBC_2.32 canonicalize_file_name F
+GLIBC_2.32 capget F
+GLIBC_2.32 capset F
+GLIBC_2.32 catclose F
+GLIBC_2.32 catgets F
+GLIBC_2.32 catopen F
+GLIBC_2.32 cbc_crypt F
+GLIBC_2.32 cfgetispeed F
+GLIBC_2.32 cfgetospeed F
+GLIBC_2.32 cfmakeraw F
+GLIBC_2.32 cfsetispeed F
+GLIBC_2.32 cfsetospeed F
+GLIBC_2.32 cfsetspeed F
+GLIBC_2.32 chdir F
+GLIBC_2.32 chflags F
+GLIBC_2.32 chmod F
+GLIBC_2.32 chown F
+GLIBC_2.32 chroot F
+GLIBC_2.32 clearenv F
+GLIBC_2.32 clearerr F
+GLIBC_2.32 clearerr_unlocked F
+GLIBC_2.32 clnt_broadcast F
+GLIBC_2.32 clnt_create F
+GLIBC_2.32 clnt_pcreateerror F
+GLIBC_2.32 clnt_perrno F
+GLIBC_2.32 clnt_perror F
+GLIBC_2.32 clnt_spcreateerror F
+GLIBC_2.32 clnt_sperrno F
+GLIBC_2.32 clnt_sperror F
+GLIBC_2.32 clntraw_create F
+GLIBC_2.32 clnttcp_create F
+GLIBC_2.32 clntudp_bufcreate F
+GLIBC_2.32 clntudp_create F
+GLIBC_2.32 clntunix_create F
+GLIBC_2.32 clock F
+GLIBC_2.32 clock_adjtime F
+GLIBC_2.32 clock_getcpuclockid F
+GLIBC_2.32 clock_getres F
+GLIBC_2.32 clock_gettime F
+GLIBC_2.32 clock_nanosleep F
+GLIBC_2.32 clock_settime F
+GLIBC_2.32 clone F
+GLIBC_2.32 close F
+GLIBC_2.32 closedir F
+GLIBC_2.32 closelog F
+GLIBC_2.32 confstr F
+GLIBC_2.32 connect F
+GLIBC_2.32 copy_file_range F
+GLIBC_2.32 copysign F
+GLIBC_2.32 copysignf F
+GLIBC_2.32 copysignl F
+GLIBC_2.32 creat F
+GLIBC_2.32 creat64 F
+GLIBC_2.32 ctermid F
+GLIBC_2.32 ctime F
+GLIBC_2.32 ctime_r F
+GLIBC_2.32 cuserid F
+GLIBC_2.32 daemon F
+GLIBC_2.32 daylight D 0x4
+GLIBC_2.32 dcgettext F
+GLIBC_2.32 dcngettext F
+GLIBC_2.32 delete_module F
+GLIBC_2.32 des_setparity F
+GLIBC_2.32 dgettext F
+GLIBC_2.32 difftime F
+GLIBC_2.32 dirfd F
+GLIBC_2.32 dirname F
+GLIBC_2.32 div F
+GLIBC_2.32 dl_iterate_phdr F
+GLIBC_2.32 dngettext F
+GLIBC_2.32 dprintf F
+GLIBC_2.32 drand48 F
+GLIBC_2.32 drand48_r F
+GLIBC_2.32 dup F
+GLIBC_2.32 dup2 F
+GLIBC_2.32 dup3 F
+GLIBC_2.32 duplocale F
+GLIBC_2.32 dysize F
+GLIBC_2.32 eaccess F
+GLIBC_2.32 ecb_crypt F
+GLIBC_2.32 ecvt F
+GLIBC_2.32 ecvt_r F
+GLIBC_2.32 endaliasent F
+GLIBC_2.32 endfsent F
+GLIBC_2.32 endgrent F
+GLIBC_2.32 endhostent F
+GLIBC_2.32 endmntent F
+GLIBC_2.32 endnetent F
+GLIBC_2.32 endnetgrent F
+GLIBC_2.32 endprotoent F
+GLIBC_2.32 endpwent F
+GLIBC_2.32 endrpcent F
+GLIBC_2.32 endservent F
+GLIBC_2.32 endsgent F
+GLIBC_2.32 endspent F
+GLIBC_2.32 endttyent F
+GLIBC_2.32 endusershell F
+GLIBC_2.32 endutent F
+GLIBC_2.32 endutxent F
+GLIBC_2.32 environ D 0x4
+GLIBC_2.32 envz_add F
+GLIBC_2.32 envz_entry F
+GLIBC_2.32 envz_get F
+GLIBC_2.32 envz_merge F
+GLIBC_2.32 envz_remove F
+GLIBC_2.32 envz_strip F
+GLIBC_2.32 epoll_create F
+GLIBC_2.32 epoll_create1 F
+GLIBC_2.32 epoll_ctl F
+GLIBC_2.32 epoll_pwait F
+GLIBC_2.32 epoll_wait F
+GLIBC_2.32 erand48 F
+GLIBC_2.32 erand48_r F
+GLIBC_2.32 err F
+GLIBC_2.32 error F
+GLIBC_2.32 error_at_line F
+GLIBC_2.32 error_message_count D 0x4
+GLIBC_2.32 error_one_per_line D 0x4
+GLIBC_2.32 error_print_progname D 0x4
+GLIBC_2.32 errx F
+GLIBC_2.32 ether_aton F
+GLIBC_2.32 ether_aton_r F
+GLIBC_2.32 ether_hostton F
+GLIBC_2.32 ether_line F
+GLIBC_2.32 ether_ntoa F
+GLIBC_2.32 ether_ntoa_r F
+GLIBC_2.32 ether_ntohost F
+GLIBC_2.32 euidaccess F
+GLIBC_2.32 eventfd F
+GLIBC_2.32 eventfd_read F
+GLIBC_2.32 eventfd_write F
+GLIBC_2.32 execl F
+GLIBC_2.32 execle F
+GLIBC_2.32 execlp F
+GLIBC_2.32 execv F
+GLIBC_2.32 execve F
+GLIBC_2.32 execvp F
+GLIBC_2.32 execvpe F
+GLIBC_2.32 exit F
+GLIBC_2.32 explicit_bzero F
+GLIBC_2.32 faccessat F
+GLIBC_2.32 fallocate F
+GLIBC_2.32 fallocate64 F
+GLIBC_2.32 fanotify_init F
+GLIBC_2.32 fanotify_mark F
+GLIBC_2.32 fchdir F
+GLIBC_2.32 fchflags F
+GLIBC_2.32 fchmod F
+GLIBC_2.32 fchmodat F
+GLIBC_2.32 fchown F
+GLIBC_2.32 fchownat F
+GLIBC_2.32 fclose F
+GLIBC_2.32 fcloseall F
+GLIBC_2.32 fcntl F
+GLIBC_2.32 fcntl64 F
+GLIBC_2.32 fcvt F
+GLIBC_2.32 fcvt_r F
+GLIBC_2.32 fdatasync F
+GLIBC_2.32 fdopen F
+GLIBC_2.32 fdopendir F
+GLIBC_2.32 feof F
+GLIBC_2.32 feof_unlocked F
+GLIBC_2.32 ferror F
+GLIBC_2.32 ferror_unlocked F
+GLIBC_2.32 fexecve F
+GLIBC_2.32 fflush F
+GLIBC_2.32 fflush_unlocked F
+GLIBC_2.32 ffs F
+GLIBC_2.32 ffsl F
+GLIBC_2.32 ffsll F
+GLIBC_2.32 fgetc F
+GLIBC_2.32 fgetc_unlocked F
+GLIBC_2.32 fgetgrent F
+GLIBC_2.32 fgetgrent_r F
+GLIBC_2.32 fgetpos F
+GLIBC_2.32 fgetpos64 F
+GLIBC_2.32 fgetpwent F
+GLIBC_2.32 fgetpwent_r F
+GLIBC_2.32 fgets F
+GLIBC_2.32 fgets_unlocked F
+GLIBC_2.32 fgetsgent F
+GLIBC_2.32 fgetsgent_r F
+GLIBC_2.32 fgetspent F
+GLIBC_2.32 fgetspent_r F
+GLIBC_2.32 fgetwc F
+GLIBC_2.32 fgetwc_unlocked F
+GLIBC_2.32 fgetws F
+GLIBC_2.32 fgetws_unlocked F
+GLIBC_2.32 fgetxattr F
+GLIBC_2.32 fileno F
+GLIBC_2.32 fileno_unlocked F
+GLIBC_2.32 finite F
+GLIBC_2.32 finitef F
+GLIBC_2.32 finitel F
+GLIBC_2.32 flistxattr F
+GLIBC_2.32 flock F
+GLIBC_2.32 flockfile F
+GLIBC_2.32 fmemopen F
+GLIBC_2.32 fmtmsg F
+GLIBC_2.32 fnmatch F
+GLIBC_2.32 fopen F
+GLIBC_2.32 fopen64 F
+GLIBC_2.32 fopencookie F
+GLIBC_2.32 fork F
+GLIBC_2.32 fpathconf F
+GLIBC_2.32 fprintf F
+GLIBC_2.32 fputc F
+GLIBC_2.32 fputc_unlocked F
+GLIBC_2.32 fputs F
+GLIBC_2.32 fputs_unlocked F
+GLIBC_2.32 fputwc F
+GLIBC_2.32 fputwc_unlocked F
+GLIBC_2.32 fputws F
+GLIBC_2.32 fputws_unlocked F
+GLIBC_2.32 fread F
+GLIBC_2.32 fread_unlocked F
+GLIBC_2.32 free F
+GLIBC_2.32 freeaddrinfo F
+GLIBC_2.32 freeifaddrs F
+GLIBC_2.32 freelocale F
+GLIBC_2.32 fremovexattr F
+GLIBC_2.32 freopen F
+GLIBC_2.32 freopen64 F
+GLIBC_2.32 frexp F
+GLIBC_2.32 frexpf F
+GLIBC_2.32 frexpl F
+GLIBC_2.32 fscanf F
+GLIBC_2.32 fseek F
+GLIBC_2.32 fseeko F
+GLIBC_2.32 fseeko64 F
+GLIBC_2.32 fsetpos F
+GLIBC_2.32 fsetpos64 F
+GLIBC_2.32 fsetxattr F
+GLIBC_2.32 fstatfs F
+GLIBC_2.32 fstatfs64 F
+GLIBC_2.32 fstatvfs F
+GLIBC_2.32 fstatvfs64 F
+GLIBC_2.32 fsync F
+GLIBC_2.32 ftell F
+GLIBC_2.32 ftello F
+GLIBC_2.32 ftello64 F
+GLIBC_2.32 ftime F
+GLIBC_2.32 ftok F
+GLIBC_2.32 ftruncate F
+GLIBC_2.32 ftruncate64 F
+GLIBC_2.32 ftrylockfile F
+GLIBC_2.32 fts64_children F
+GLIBC_2.32 fts64_close F
+GLIBC_2.32 fts64_open F
+GLIBC_2.32 fts64_read F
+GLIBC_2.32 fts64_set F
+GLIBC_2.32 fts_children F
+GLIBC_2.32 fts_close F
+GLIBC_2.32 fts_open F
+GLIBC_2.32 fts_read F
+GLIBC_2.32 fts_set F
+GLIBC_2.32 ftw F
+GLIBC_2.32 ftw64 F
+GLIBC_2.32 funlockfile F
+GLIBC_2.32 futimens F
+GLIBC_2.32 futimes F
+GLIBC_2.32 futimesat F
+GLIBC_2.32 fwide F
+GLIBC_2.32 fwprintf F
+GLIBC_2.32 fwrite F
+GLIBC_2.32 fwrite_unlocked F
+GLIBC_2.32 fwscanf F
+GLIBC_2.32 gai_strerror F
+GLIBC_2.32 gcvt F
+GLIBC_2.32 get_avphys_pages F
+GLIBC_2.32 get_current_dir_name F
+GLIBC_2.32 get_myaddress F
+GLIBC_2.32 get_nprocs F
+GLIBC_2.32 get_nprocs_conf F
+GLIBC_2.32 get_phys_pages F
+GLIBC_2.32 getaddrinfo F
+GLIBC_2.32 getaliasbyname F
+GLIBC_2.32 getaliasbyname_r F
+GLIBC_2.32 getaliasent F
+GLIBC_2.32 getaliasent_r F
+GLIBC_2.32 getauxval F
+GLIBC_2.32 getc F
+GLIBC_2.32 getc_unlocked F
+GLIBC_2.32 getchar F
+GLIBC_2.32 getchar_unlocked F
+GLIBC_2.32 getcontext F
+GLIBC_2.32 getcpu F
+GLIBC_2.32 getcwd F
+GLIBC_2.32 getdate F
+GLIBC_2.32 getdate_err D 0x4
+GLIBC_2.32 getdate_r F
+GLIBC_2.32 getdelim F
+GLIBC_2.32 getdents64 F
+GLIBC_2.32 getdirentries F
+GLIBC_2.32 getdirentries64 F
+GLIBC_2.32 getdomainname F
+GLIBC_2.32 getdtablesize F
+GLIBC_2.32 getegid F
+GLIBC_2.32 getentropy F
+GLIBC_2.32 getenv F
+GLIBC_2.32 geteuid F
+GLIBC_2.32 getfsent F
+GLIBC_2.32 getfsfile F
+GLIBC_2.32 getfsspec F
+GLIBC_2.32 getgid F
+GLIBC_2.32 getgrent F
+GLIBC_2.32 getgrent_r F
+GLIBC_2.32 getgrgid F
+GLIBC_2.32 getgrgid_r F
+GLIBC_2.32 getgrnam F
+GLIBC_2.32 getgrnam_r F
+GLIBC_2.32 getgrouplist F
+GLIBC_2.32 getgroups F
+GLIBC_2.32 gethostbyaddr F
+GLIBC_2.32 gethostbyaddr_r F
+GLIBC_2.32 gethostbyname F
+GLIBC_2.32 gethostbyname2 F
+GLIBC_2.32 gethostbyname2_r F
+GLIBC_2.32 gethostbyname_r F
+GLIBC_2.32 gethostent F
+GLIBC_2.32 gethostent_r F
+GLIBC_2.32 gethostid F
+GLIBC_2.32 gethostname F
+GLIBC_2.32 getifaddrs F
+GLIBC_2.32 getipv4sourcefilter F
+GLIBC_2.32 getitimer F
+GLIBC_2.32 getline F
+GLIBC_2.32 getloadavg F
+GLIBC_2.32 getlogin F
+GLIBC_2.32 getlogin_r F
+GLIBC_2.32 getmntent F
+GLIBC_2.32 getmntent_r F
+GLIBC_2.32 getnameinfo F
+GLIBC_2.32 getnetbyaddr F
+GLIBC_2.32 getnetbyaddr_r F
+GLIBC_2.32 getnetbyname F
+GLIBC_2.32 getnetbyname_r F
+GLIBC_2.32 getnetent F
+GLIBC_2.32 getnetent_r F
+GLIBC_2.32 getnetgrent F
+GLIBC_2.32 getnetgrent_r F
+GLIBC_2.32 getnetname F
+GLIBC_2.32 getopt F
+GLIBC_2.32 getopt_long F
+GLIBC_2.32 getopt_long_only F
+GLIBC_2.32 getpagesize F
+GLIBC_2.32 getpass F
+GLIBC_2.32 getpeername F
+GLIBC_2.32 getpgid F
+GLIBC_2.32 getpgrp F
+GLIBC_2.32 getpid F
+GLIBC_2.32 getppid F
+GLIBC_2.32 getpriority F
+GLIBC_2.32 getprotobyname F
+GLIBC_2.32 getprotobyname_r F
+GLIBC_2.32 getprotobynumber F
+GLIBC_2.32 getprotobynumber_r F
+GLIBC_2.32 getprotoent F
+GLIBC_2.32 getprotoent_r F
+GLIBC_2.32 getpt F
+GLIBC_2.32 getpublickey F
+GLIBC_2.32 getpw F
+GLIBC_2.32 getpwent F
+GLIBC_2.32 getpwent_r F
+GLIBC_2.32 getpwnam F
+GLIBC_2.32 getpwnam_r F
+GLIBC_2.32 getpwuid F
+GLIBC_2.32 getpwuid_r F
+GLIBC_2.32 getrandom F
+GLIBC_2.32 getresgid F
+GLIBC_2.32 getresuid F
+GLIBC_2.32 getrlimit F
+GLIBC_2.32 getrlimit64 F
+GLIBC_2.32 getrpcbyname F
+GLIBC_2.32 getrpcbyname_r F
+GLIBC_2.32 getrpcbynumber F
+GLIBC_2.32 getrpcbynumber_r F
+GLIBC_2.32 getrpcent F
+GLIBC_2.32 getrpcent_r F
+GLIBC_2.32 getrpcport F
+GLIBC_2.32 getrusage F
+GLIBC_2.32 gets F
+GLIBC_2.32 getsecretkey F
+GLIBC_2.32 getservbyname F
+GLIBC_2.32 getservbyname_r F
+GLIBC_2.32 getservbyport F
+GLIBC_2.32 getservbyport_r F
+GLIBC_2.32 getservent F
+GLIBC_2.32 getservent_r F
+GLIBC_2.32 getsgent F
+GLIBC_2.32 getsgent_r F
+GLIBC_2.32 getsgnam F
+GLIBC_2.32 getsgnam_r F
+GLIBC_2.32 getsid F
+GLIBC_2.32 getsockname F
+GLIBC_2.32 getsockopt F
+GLIBC_2.32 getsourcefilter F
+GLIBC_2.32 getspent F
+GLIBC_2.32 getspent_r F
+GLIBC_2.32 getspnam F
+GLIBC_2.32 getspnam_r F
+GLIBC_2.32 getsubopt F
+GLIBC_2.32 gettext F
+GLIBC_2.32 gettid F
+GLIBC_2.32 gettimeofday F
+GLIBC_2.32 getttyent F
+GLIBC_2.32 getttynam F
+GLIBC_2.32 getuid F
+GLIBC_2.32 getusershell F
+GLIBC_2.32 getutent F
+GLIBC_2.32 getutent_r F
+GLIBC_2.32 getutid F
+GLIBC_2.32 getutid_r F
+GLIBC_2.32 getutline F
+GLIBC_2.32 getutline_r F
+GLIBC_2.32 getutmp F
+GLIBC_2.32 getutmpx F
+GLIBC_2.32 getutxent F
+GLIBC_2.32 getutxid F
+GLIBC_2.32 getutxline F
+GLIBC_2.32 getw F
+GLIBC_2.32 getwc F
+GLIBC_2.32 getwc_unlocked F
+GLIBC_2.32 getwchar F
+GLIBC_2.32 getwchar_unlocked F
+GLIBC_2.32 getwd F
+GLIBC_2.32 getxattr F
+GLIBC_2.32 glob F
+GLIBC_2.32 glob64 F
+GLIBC_2.32 glob_pattern_p F
+GLIBC_2.32 globfree F
+GLIBC_2.32 globfree64 F
+GLIBC_2.32 gmtime F
+GLIBC_2.32 gmtime_r F
+GLIBC_2.32 gnu_dev_major F
+GLIBC_2.32 gnu_dev_makedev F
+GLIBC_2.32 gnu_dev_minor F
+GLIBC_2.32 gnu_get_libc_release F
+GLIBC_2.32 gnu_get_libc_version F
+GLIBC_2.32 grantpt F
+GLIBC_2.32 group_member F
+GLIBC_2.32 gsignal F
+GLIBC_2.32 gtty F
+GLIBC_2.32 h_errlist D 0x14
+GLIBC_2.32 h_nerr D 0x4
+GLIBC_2.32 hasmntopt F
+GLIBC_2.32 hcreate F
+GLIBC_2.32 hcreate_r F
+GLIBC_2.32 hdestroy F
+GLIBC_2.32 hdestroy_r F
+GLIBC_2.32 herror F
+GLIBC_2.32 host2netname F
+GLIBC_2.32 hsearch F
+GLIBC_2.32 hsearch_r F
+GLIBC_2.32 hstrerror F
+GLIBC_2.32 htonl F
+GLIBC_2.32 htons F
+GLIBC_2.32 iconv F
+GLIBC_2.32 iconv_close F
+GLIBC_2.32 iconv_open F
+GLIBC_2.32 if_freenameindex F
+GLIBC_2.32 if_indextoname F
+GLIBC_2.32 if_nameindex F
+GLIBC_2.32 if_nametoindex F
+GLIBC_2.32 imaxabs F
+GLIBC_2.32 imaxdiv F
+GLIBC_2.32 in6addr_any D 0x10
+GLIBC_2.32 in6addr_loopback D 0x10
+GLIBC_2.32 index F
+GLIBC_2.32 inet6_opt_append F
+GLIBC_2.32 inet6_opt_find F
+GLIBC_2.32 inet6_opt_finish F
+GLIBC_2.32 inet6_opt_get_val F
+GLIBC_2.32 inet6_opt_init F
+GLIBC_2.32 inet6_opt_next F
+GLIBC_2.32 inet6_opt_set_val F
+GLIBC_2.32 inet6_option_alloc F
+GLIBC_2.32 inet6_option_append F
+GLIBC_2.32 inet6_option_find F
+GLIBC_2.32 inet6_option_init F
+GLIBC_2.32 inet6_option_next F
+GLIBC_2.32 inet6_option_space F
+GLIBC_2.32 inet6_rth_add F
+GLIBC_2.32 inet6_rth_getaddr F
+GLIBC_2.32 inet6_rth_init F
+GLIBC_2.32 inet6_rth_reverse F
+GLIBC_2.32 inet6_rth_segments F
+GLIBC_2.32 inet6_rth_space F
+GLIBC_2.32 inet_addr F
+GLIBC_2.32 inet_aton F
+GLIBC_2.32 inet_lnaof F
+GLIBC_2.32 inet_makeaddr F
+GLIBC_2.32 inet_netof F
+GLIBC_2.32 inet_network F
+GLIBC_2.32 inet_nsap_addr F
+GLIBC_2.32 inet_nsap_ntoa F
+GLIBC_2.32 inet_ntoa F
+GLIBC_2.32 inet_ntop F
+GLIBC_2.32 inet_pton F
+GLIBC_2.32 init_module F
+GLIBC_2.32 initgroups F
+GLIBC_2.32 initstate F
+GLIBC_2.32 initstate_r F
+GLIBC_2.32 innetgr F
+GLIBC_2.32 inotify_add_watch F
+GLIBC_2.32 inotify_init F
+GLIBC_2.32 inotify_init1 F
+GLIBC_2.32 inotify_rm_watch F
+GLIBC_2.32 insque F
+GLIBC_2.32 ioctl F
+GLIBC_2.32 iruserok F
+GLIBC_2.32 iruserok_af F
+GLIBC_2.32 isalnum F
+GLIBC_2.32 isalnum_l F
+GLIBC_2.32 isalpha F
+GLIBC_2.32 isalpha_l F
+GLIBC_2.32 isascii F
+GLIBC_2.32 isatty F
+GLIBC_2.32 isblank F
+GLIBC_2.32 isblank_l F
+GLIBC_2.32 iscntrl F
+GLIBC_2.32 iscntrl_l F
+GLIBC_2.32 isctype F
+GLIBC_2.32 isdigit F
+GLIBC_2.32 isdigit_l F
+GLIBC_2.32 isfdtype F
+GLIBC_2.32 isgraph F
+GLIBC_2.32 isgraph_l F
+GLIBC_2.32 isinf F
+GLIBC_2.32 isinff F
+GLIBC_2.32 isinfl F
+GLIBC_2.32 islower F
+GLIBC_2.32 islower_l F
+GLIBC_2.32 isnan F
+GLIBC_2.32 isnanf F
+GLIBC_2.32 isnanl F
+GLIBC_2.32 isprint F
+GLIBC_2.32 isprint_l F
+GLIBC_2.32 ispunct F
+GLIBC_2.32 ispunct_l F
+GLIBC_2.32 isspace F
+GLIBC_2.32 isspace_l F
+GLIBC_2.32 isupper F
+GLIBC_2.32 isupper_l F
+GLIBC_2.32 iswalnum F
+GLIBC_2.32 iswalnum_l F
+GLIBC_2.32 iswalpha F
+GLIBC_2.32 iswalpha_l F
+GLIBC_2.32 iswblank F
+GLIBC_2.32 iswblank_l F
+GLIBC_2.32 iswcntrl F
+GLIBC_2.32 iswcntrl_l F
+GLIBC_2.32 iswctype F
+GLIBC_2.32 iswctype_l F
+GLIBC_2.32 iswdigit F
+GLIBC_2.32 iswdigit_l F
+GLIBC_2.32 iswgraph F
+GLIBC_2.32 iswgraph_l F
+GLIBC_2.32 iswlower F
+GLIBC_2.32 iswlower_l F
+GLIBC_2.32 iswprint F
+GLIBC_2.32 iswprint_l F
+GLIBC_2.32 iswpunct F
+GLIBC_2.32 iswpunct_l F
+GLIBC_2.32 iswspace F
+GLIBC_2.32 iswspace_l F
+GLIBC_2.32 iswupper F
+GLIBC_2.32 iswupper_l F
+GLIBC_2.32 iswxdigit F
+GLIBC_2.32 iswxdigit_l F
+GLIBC_2.32 isxdigit F
+GLIBC_2.32 isxdigit_l F
+GLIBC_2.32 jrand48 F
+GLIBC_2.32 jrand48_r F
+GLIBC_2.32 key_decryptsession F
+GLIBC_2.32 key_decryptsession_pk F
+GLIBC_2.32 key_encryptsession F
+GLIBC_2.32 key_encryptsession_pk F
+GLIBC_2.32 key_gendes F
+GLIBC_2.32 key_get_conv F
+GLIBC_2.32 key_secretkey_is_set F
+GLIBC_2.32 key_setnet F
+GLIBC_2.32 key_setsecret F
+GLIBC_2.32 kill F
+GLIBC_2.32 killpg F
+GLIBC_2.32 klogctl F
+GLIBC_2.32 l64a F
+GLIBC_2.32 labs F
+GLIBC_2.32 lchmod F
+GLIBC_2.32 lchown F
+GLIBC_2.32 lckpwdf F
+GLIBC_2.32 lcong48 F
+GLIBC_2.32 lcong48_r F
+GLIBC_2.32 ldexp F
+GLIBC_2.32 ldexpf F
+GLIBC_2.32 ldexpl F
+GLIBC_2.32 ldiv F
+GLIBC_2.32 lfind F
+GLIBC_2.32 lgetxattr F
+GLIBC_2.32 link F
+GLIBC_2.32 linkat F
+GLIBC_2.32 listen F
+GLIBC_2.32 listxattr F
+GLIBC_2.32 llabs F
+GLIBC_2.32 lldiv F
+GLIBC_2.32 llistxattr F
+GLIBC_2.32 localeconv F
+GLIBC_2.32 localtime F
+GLIBC_2.32 localtime_r F
+GLIBC_2.32 lockf F
+GLIBC_2.32 lockf64 F
+GLIBC_2.32 longjmp F
+GLIBC_2.32 lrand48 F
+GLIBC_2.32 lrand48_r F
+GLIBC_2.32 lremovexattr F
+GLIBC_2.32 lsearch F
+GLIBC_2.32 lseek F
+GLIBC_2.32 lseek64 F
+GLIBC_2.32 lsetxattr F
+GLIBC_2.32 lutimes F
+GLIBC_2.32 madvise F
+GLIBC_2.32 makecontext F
+GLIBC_2.32 mallinfo F
+GLIBC_2.32 malloc F
+GLIBC_2.32 malloc_info F
+GLIBC_2.32 malloc_stats F
+GLIBC_2.32 malloc_trim F
+GLIBC_2.32 malloc_usable_size F
+GLIBC_2.32 mallopt F
+GLIBC_2.32 mallwatch D 0x4
+GLIBC_2.32 mblen F
+GLIBC_2.32 mbrlen F
+GLIBC_2.32 mbrtoc16 F
+GLIBC_2.32 mbrtoc32 F
+GLIBC_2.32 mbrtowc F
+GLIBC_2.32 mbsinit F
+GLIBC_2.32 mbsnrtowcs F
+GLIBC_2.32 mbsrtowcs F
+GLIBC_2.32 mbstowcs F
+GLIBC_2.32 mbtowc F
+GLIBC_2.32 mcheck F
+GLIBC_2.32 mcheck_check_all F
+GLIBC_2.32 mcheck_pedantic F
+GLIBC_2.32 memalign F
+GLIBC_2.32 memccpy F
+GLIBC_2.32 memchr F
+GLIBC_2.32 memcmp F
+GLIBC_2.32 memcpy F
+GLIBC_2.32 memfd_create F
+GLIBC_2.32 memfrob F
+GLIBC_2.32 memmem F
+GLIBC_2.32 memmove F
+GLIBC_2.32 mempcpy F
+GLIBC_2.32 memrchr F
+GLIBC_2.32 memset F
+GLIBC_2.32 mincore F
+GLIBC_2.32 mkdir F
+GLIBC_2.32 mkdirat F
+GLIBC_2.32 mkdtemp F
+GLIBC_2.32 mkfifo F
+GLIBC_2.32 mkfifoat F
+GLIBC_2.32 mkostemp F
+GLIBC_2.32 mkostemp64 F
+GLIBC_2.32 mkostemps F
+GLIBC_2.32 mkostemps64 F
+GLIBC_2.32 mkstemp F
+GLIBC_2.32 mkstemp64 F
+GLIBC_2.32 mkstemps F
+GLIBC_2.32 mkstemps64 F
+GLIBC_2.32 mktemp F
+GLIBC_2.32 mktime F
+GLIBC_2.32 mlock F
+GLIBC_2.32 mlock2 F
+GLIBC_2.32 mlockall F
+GLIBC_2.32 mmap F
+GLIBC_2.32 mmap64 F
+GLIBC_2.32 modf F
+GLIBC_2.32 modff F
+GLIBC_2.32 modfl F
+GLIBC_2.32 moncontrol F
+GLIBC_2.32 monstartup F
+GLIBC_2.32 mount F
+GLIBC_2.32 mprobe F
+GLIBC_2.32 mprotect F
+GLIBC_2.32 mrand48 F
+GLIBC_2.32 mrand48_r F
+GLIBC_2.32 mremap F
+GLIBC_2.32 msgctl F
+GLIBC_2.32 msgget F
+GLIBC_2.32 msgrcv F
+GLIBC_2.32 msgsnd F
+GLIBC_2.32 msync F
+GLIBC_2.32 mtrace F
+GLIBC_2.32 munlock F
+GLIBC_2.32 munlockall F
+GLIBC_2.32 munmap F
+GLIBC_2.32 muntrace F
+GLIBC_2.32 name_to_handle_at F
+GLIBC_2.32 nanosleep F
+GLIBC_2.32 netname2host F
+GLIBC_2.32 netname2user F
+GLIBC_2.32 newlocale F
+GLIBC_2.32 nftw F
+GLIBC_2.32 nftw64 F
+GLIBC_2.32 ngettext F
+GLIBC_2.32 nice F
+GLIBC_2.32 nl_langinfo F
+GLIBC_2.32 nl_langinfo_l F
+GLIBC_2.32 nrand48 F
+GLIBC_2.32 nrand48_r F
+GLIBC_2.32 ntohl F
+GLIBC_2.32 ntohs F
+GLIBC_2.32 ntp_adjtime F
+GLIBC_2.32 ntp_gettime F
+GLIBC_2.32 ntp_gettimex F
+GLIBC_2.32 obstack_alloc_failed_handler D 0x4
+GLIBC_2.32 obstack_exit_failure D 0x4
+GLIBC_2.32 obstack_free F
+GLIBC_2.32 obstack_printf F
+GLIBC_2.32 obstack_vprintf F
+GLIBC_2.32 on_exit F
+GLIBC_2.32 open F
+GLIBC_2.32 open64 F
+GLIBC_2.32 open_by_handle_at F
+GLIBC_2.32 open_memstream F
+GLIBC_2.32 open_wmemstream F
+GLIBC_2.32 openat F
+GLIBC_2.32 openat64 F
+GLIBC_2.32 opendir F
+GLIBC_2.32 openlog F
+GLIBC_2.32 optarg D 0x4
+GLIBC_2.32 opterr D 0x4
+GLIBC_2.32 optind D 0x4
+GLIBC_2.32 optopt D 0x4
+GLIBC_2.32 parse_printf_format F
+GLIBC_2.32 passwd2des F
+GLIBC_2.32 pathconf F
+GLIBC_2.32 pause F
+GLIBC_2.32 pclose F
+GLIBC_2.32 perror F
+GLIBC_2.32 personality F
+GLIBC_2.32 pipe F
+GLIBC_2.32 pipe2 F
+GLIBC_2.32 pivot_root F
+GLIBC_2.32 pkey_alloc F
+GLIBC_2.32 pkey_free F
+GLIBC_2.32 pkey_get F
+GLIBC_2.32 pkey_mprotect F
+GLIBC_2.32 pkey_set F
+GLIBC_2.32 pmap_getmaps F
+GLIBC_2.32 pmap_getport F
+GLIBC_2.32 pmap_rmtcall F
+GLIBC_2.32 pmap_set F
+GLIBC_2.32 pmap_unset F
+GLIBC_2.32 poll F
+GLIBC_2.32 popen F
+GLIBC_2.32 posix_fadvise F
+GLIBC_2.32 posix_fadvise64 F
+GLIBC_2.32 posix_fallocate F
+GLIBC_2.32 posix_fallocate64 F
+GLIBC_2.32 posix_madvise F
+GLIBC_2.32 posix_memalign F
+GLIBC_2.32 posix_openpt F
+GLIBC_2.32 posix_spawn F
+GLIBC_2.32 posix_spawn_file_actions_addchdir_np F
+GLIBC_2.32 posix_spawn_file_actions_addclose F
+GLIBC_2.32 posix_spawn_file_actions_adddup2 F
+GLIBC_2.32 posix_spawn_file_actions_addfchdir_np F
+GLIBC_2.32 posix_spawn_file_actions_addopen F
+GLIBC_2.32 posix_spawn_file_actions_destroy F
+GLIBC_2.32 posix_spawn_file_actions_init F
+GLIBC_2.32 posix_spawnattr_destroy F
+GLIBC_2.32 posix_spawnattr_getflags F
+GLIBC_2.32 posix_spawnattr_getpgroup F
+GLIBC_2.32 posix_spawnattr_getschedparam F
+GLIBC_2.32 posix_spawnattr_getschedpolicy F
+GLIBC_2.32 posix_spawnattr_getsigdefault F
+GLIBC_2.32 posix_spawnattr_getsigmask F
+GLIBC_2.32 posix_spawnattr_init F
+GLIBC_2.32 posix_spawnattr_setflags F
+GLIBC_2.32 posix_spawnattr_setpgroup F
+GLIBC_2.32 posix_spawnattr_setschedparam F
+GLIBC_2.32 posix_spawnattr_setschedpolicy F
+GLIBC_2.32 posix_spawnattr_setsigdefault F
+GLIBC_2.32 posix_spawnattr_setsigmask F
+GLIBC_2.32 posix_spawnp F
+GLIBC_2.32 ppoll F
+GLIBC_2.32 prctl F
+GLIBC_2.32 pread F
+GLIBC_2.32 pread64 F
+GLIBC_2.32 preadv F
+GLIBC_2.32 preadv2 F
+GLIBC_2.32 preadv64 F
+GLIBC_2.32 preadv64v2 F
+GLIBC_2.32 printf F
+GLIBC_2.32 printf_size F
+GLIBC_2.32 printf_size_info F
+GLIBC_2.32 prlimit F
+GLIBC_2.32 prlimit64 F
+GLIBC_2.32 process_vm_readv F
+GLIBC_2.32 process_vm_writev F
+GLIBC_2.32 profil F
+GLIBC_2.32 program_invocation_name D 0x4
+GLIBC_2.32 program_invocation_short_name D 0x4
+GLIBC_2.32 pselect F
+GLIBC_2.32 psiginfo F
+GLIBC_2.32 psignal F
+GLIBC_2.32 pthread_attr_destroy F
+GLIBC_2.32 pthread_attr_getdetachstate F
+GLIBC_2.32 pthread_attr_getinheritsched F
+GLIBC_2.32 pthread_attr_getschedparam F
+GLIBC_2.32 pthread_attr_getschedpolicy F
+GLIBC_2.32 pthread_attr_getscope F
+GLIBC_2.32 pthread_attr_init F
+GLIBC_2.32 pthread_attr_getsigmask_np F
+GLIBC_2.32 pthread_attr_setaffinity_np F
+GLIBC_2.32 pthread_attr_setsigmask_np F
+GLIBC_2.32 pthread_attr_setdetachstate F
+GLIBC_2.32 pthread_attr_setinheritsched F
+GLIBC_2.32 pthread_attr_setschedparam F
+GLIBC_2.32 pthread_attr_setschedpolicy F
+GLIBC_2.32 pthread_attr_setscope F
+GLIBC_2.32 pthread_cond_broadcast F
+GLIBC_2.32 pthread_cond_destroy F
+GLIBC_2.32 pthread_cond_init F
+GLIBC_2.32 pthread_cond_signal F
+GLIBC_2.32 pthread_cond_timedwait F
+GLIBC_2.32 pthread_cond_wait F
+GLIBC_2.32 pthread_condattr_destroy F
+GLIBC_2.32 pthread_condattr_init F
+GLIBC_2.32 pthread_equal F
+GLIBC_2.32 pthread_exit F
+GLIBC_2.32 pthread_getaffinity_np F
+GLIBC_2.32 pthread_getattr_np F
+GLIBC_2.32 pthread_getschedparam F
+GLIBC_2.32 pthread_mutex_destroy F
+GLIBC_2.32 pthread_mutex_init F
+GLIBC_2.32 pthread_mutex_lock F
+GLIBC_2.32 pthread_mutex_unlock F
+GLIBC_2.32 pthread_self F
+GLIBC_2.32 pthread_setcancelstate F
+GLIBC_2.32 pthread_setcanceltype F
+GLIBC_2.32 pthread_setschedparam F
+GLIBC_2.32 pthread_sigmask F
+GLIBC_2.32 ptrace F
+GLIBC_2.32 ptsname F
+GLIBC_2.32 ptsname_r F
+GLIBC_2.32 putc F
+GLIBC_2.32 putc_unlocked F
+GLIBC_2.32 putchar F
+GLIBC_2.32 putchar_unlocked F
+GLIBC_2.32 putenv F
+GLIBC_2.32 putgrent F
+GLIBC_2.32 putpwent F
+GLIBC_2.32 puts F
+GLIBC_2.32 putsgent F
+GLIBC_2.32 putspent F
+GLIBC_2.32 pututline F
+GLIBC_2.32 pututxline F
+GLIBC_2.32 putw F
+GLIBC_2.32 putwc F
+GLIBC_2.32 putwc_unlocked F
+GLIBC_2.32 putwchar F
+GLIBC_2.32 putwchar_unlocked F
+GLIBC_2.32 pvalloc F
+GLIBC_2.32 pwrite F
+GLIBC_2.32 pwrite64 F
+GLIBC_2.32 pwritev F
+GLIBC_2.32 pwritev2 F
+GLIBC_2.32 pwritev64 F
+GLIBC_2.32 pwritev64v2 F
+GLIBC_2.32 qecvt F
+GLIBC_2.32 qecvt_r F
+GLIBC_2.32 qfcvt F
+GLIBC_2.32 qfcvt_r F
+GLIBC_2.32 qgcvt F
+GLIBC_2.32 qsort F
+GLIBC_2.32 qsort_r F
+GLIBC_2.32 quick_exit F
+GLIBC_2.32 quotactl F
+GLIBC_2.32 raise F
+GLIBC_2.32 rand F
+GLIBC_2.32 rand_r F
+GLIBC_2.32 random F
+GLIBC_2.32 random_r F
+GLIBC_2.32 rawmemchr F
+GLIBC_2.32 rcmd F
+GLIBC_2.32 rcmd_af F
+GLIBC_2.32 re_comp F
+GLIBC_2.32 re_compile_fastmap F
+GLIBC_2.32 re_compile_pattern F
+GLIBC_2.32 re_exec F
+GLIBC_2.32 re_match F
+GLIBC_2.32 re_match_2 F
+GLIBC_2.32 re_search F
+GLIBC_2.32 re_search_2 F
+GLIBC_2.32 re_set_registers F
+GLIBC_2.32 re_set_syntax F
+GLIBC_2.32 re_syntax_options D 0x4
+GLIBC_2.32 read F
+GLIBC_2.32 readahead F
+GLIBC_2.32 readdir F
+GLIBC_2.32 readdir64 F
+GLIBC_2.32 readdir64_r F
+GLIBC_2.32 readdir_r F
+GLIBC_2.32 readlink F
+GLIBC_2.32 readlinkat F
+GLIBC_2.32 readv F
+GLIBC_2.32 realloc F
+GLIBC_2.32 reallocarray F
+GLIBC_2.32 realpath F
+GLIBC_2.32 reboot F
+GLIBC_2.32 recv F
+GLIBC_2.32 recvfrom F
+GLIBC_2.32 recvmmsg F
+GLIBC_2.32 recvmsg F
+GLIBC_2.32 regcomp F
+GLIBC_2.32 regerror F
+GLIBC_2.32 regexec F
+GLIBC_2.32 regfree F
+GLIBC_2.32 register_printf_function F
+GLIBC_2.32 register_printf_modifier F
+GLIBC_2.32 register_printf_specifier F
+GLIBC_2.32 register_printf_type F
+GLIBC_2.32 registerrpc F
+GLIBC_2.32 remap_file_pages F
+GLIBC_2.32 remove F
+GLIBC_2.32 removexattr F
+GLIBC_2.32 remque F
+GLIBC_2.32 rename F
+GLIBC_2.32 renameat F
+GLIBC_2.32 renameat2 F
+GLIBC_2.32 revoke F
+GLIBC_2.32 rewind F
+GLIBC_2.32 rewinddir F
+GLIBC_2.32 rexec F
+GLIBC_2.32 rexec_af F
+GLIBC_2.32 rexecoptions D 0x4
+GLIBC_2.32 rindex F
+GLIBC_2.32 rmdir F
+GLIBC_2.32 rpc_createerr D 0x10
+GLIBC_2.32 rpmatch F
+GLIBC_2.32 rresvport F
+GLIBC_2.32 rresvport_af F
+GLIBC_2.32 rtime F
+GLIBC_2.32 ruserok F
+GLIBC_2.32 ruserok_af F
+GLIBC_2.32 ruserpass F
+GLIBC_2.32 sbrk F
+GLIBC_2.32 scalbn F
+GLIBC_2.32 scalbnf F
+GLIBC_2.32 scalbnl F
+GLIBC_2.32 scandir F
+GLIBC_2.32 scandir64 F
+GLIBC_2.32 scandirat F
+GLIBC_2.32 scandirat64 F
+GLIBC_2.32 scanf F
+GLIBC_2.32 sched_get_priority_max F
+GLIBC_2.32 sched_get_priority_min F
+GLIBC_2.32 sched_getaffinity F
+GLIBC_2.32 sched_getcpu F
+GLIBC_2.32 sched_getparam F
+GLIBC_2.32 sched_getscheduler F
+GLIBC_2.32 sched_rr_get_interval F
+GLIBC_2.32 sched_setaffinity F
+GLIBC_2.32 sched_setparam F
+GLIBC_2.32 sched_setscheduler F
+GLIBC_2.32 sched_yield F
+GLIBC_2.32 secure_getenv F
+GLIBC_2.32 seed48 F
+GLIBC_2.32 seed48_r F
+GLIBC_2.32 seekdir F
+GLIBC_2.32 select F
+GLIBC_2.32 semctl F
+GLIBC_2.32 semget F
+GLIBC_2.32 semop F
+GLIBC_2.32 semtimedop F
+GLIBC_2.32 send F
+GLIBC_2.32 sendfile F
+GLIBC_2.32 sendfile64 F
+GLIBC_2.32 sendmmsg F
+GLIBC_2.32 sendmsg F
+GLIBC_2.32 sendto F
+GLIBC_2.32 setaliasent F
+GLIBC_2.32 setbuf F
+GLIBC_2.32 setbuffer F
+GLIBC_2.32 setcontext F
+GLIBC_2.32 setdomainname F
+GLIBC_2.32 setegid F
+GLIBC_2.32 setenv F
+GLIBC_2.32 seteuid F
+GLIBC_2.32 setfsent F
+GLIBC_2.32 setfsgid F
+GLIBC_2.32 setfsuid F
+GLIBC_2.32 setgid F
+GLIBC_2.32 setgrent F
+GLIBC_2.32 setgroups F
+GLIBC_2.32 sethostent F
+GLIBC_2.32 sethostid F
+GLIBC_2.32 sethostname F
+GLIBC_2.32 setipv4sourcefilter F
+GLIBC_2.32 setitimer F
+GLIBC_2.32 setjmp F
+GLIBC_2.32 setlinebuf F
+GLIBC_2.32 setlocale F
+GLIBC_2.32 setlogin F
+GLIBC_2.32 setlogmask F
+GLIBC_2.32 setmntent F
+GLIBC_2.32 setnetent F
+GLIBC_2.32 setnetgrent F
+GLIBC_2.32 setns F
+GLIBC_2.32 setpgid F
+GLIBC_2.32 setpgrp F
+GLIBC_2.32 setpriority F
+GLIBC_2.32 setprotoent F
+GLIBC_2.32 setpwent F
+GLIBC_2.32 setregid F
+GLIBC_2.32 setresgid F
+GLIBC_2.32 setresuid F
+GLIBC_2.32 setreuid F
+GLIBC_2.32 setrlimit F
+GLIBC_2.32 setrlimit64 F
+GLIBC_2.32 setrpcent F
+GLIBC_2.32 setservent F
+GLIBC_2.32 setsgent F
+GLIBC_2.32 setsid F
+GLIBC_2.32 setsockopt F
+GLIBC_2.32 setsourcefilter F
+GLIBC_2.32 setspent F
+GLIBC_2.32 setstate F
+GLIBC_2.32 setstate_r F
+GLIBC_2.32 settimeofday F
+GLIBC_2.32 setttyent F
+GLIBC_2.32 setuid F
+GLIBC_2.32 setusershell F
+GLIBC_2.32 setutent F
+GLIBC_2.32 setutxent F
+GLIBC_2.32 setvbuf F
+GLIBC_2.32 setxattr F
+GLIBC_2.32 sgetsgent F
+GLIBC_2.32 sgetsgent_r F
+GLIBC_2.32 sgetspent F
+GLIBC_2.32 sgetspent_r F
+GLIBC_2.32 shmat F
+GLIBC_2.32 shmctl F
+GLIBC_2.32 shmdt F
+GLIBC_2.32 shmget F
+GLIBC_2.32 shutdown F
+GLIBC_2.32 sigaction F
+GLIBC_2.32 sigaddset F
+GLIBC_2.32 sigaltstack F
+GLIBC_2.32 sigandset F
+GLIBC_2.32 sigblock F
+GLIBC_2.32 sigdelset F
+GLIBC_2.32 sigemptyset F
+GLIBC_2.32 sigfillset F
+GLIBC_2.32 siggetmask F
+GLIBC_2.32 sighold F
+GLIBC_2.32 sigignore F
+GLIBC_2.32 siginterrupt F
+GLIBC_2.32 sigisemptyset F
+GLIBC_2.32 sigismember F
+GLIBC_2.32 siglongjmp F
+GLIBC_2.32 signal F
+GLIBC_2.32 signalfd F
+GLIBC_2.32 sigorset F
+GLIBC_2.32 sigpause F
+GLIBC_2.32 sigpending F
+GLIBC_2.32 sigprocmask F
+GLIBC_2.32 sigqueue F
+GLIBC_2.32 sigrelse F
+GLIBC_2.32 sigreturn F
+GLIBC_2.32 sigset F
+GLIBC_2.32 sigsetmask F
+GLIBC_2.32 sigstack F
+GLIBC_2.32 sigsuspend F
+GLIBC_2.32 sigtimedwait F
+GLIBC_2.32 sigwait F
+GLIBC_2.32 sigwaitinfo F
+GLIBC_2.32 sleep F
+GLIBC_2.32 snprintf F
+GLIBC_2.32 sockatmark F
+GLIBC_2.32 socket F
+GLIBC_2.32 socketpair F
+GLIBC_2.32 splice F
+GLIBC_2.32 sprintf F
+GLIBC_2.32 sprofil F
+GLIBC_2.32 srand F
+GLIBC_2.32 srand48 F
+GLIBC_2.32 srand48_r F
+GLIBC_2.32 srandom F
+GLIBC_2.32 srandom_r F
+GLIBC_2.32 sscanf F
+GLIBC_2.32 ssignal F
+GLIBC_2.32 statfs F
+GLIBC_2.32 statfs64 F
+GLIBC_2.32 statvfs F
+GLIBC_2.32 statvfs64 F
+GLIBC_2.32 statx F
+GLIBC_2.32 stderr D 0x4
+GLIBC_2.32 stdin D 0x4
+GLIBC_2.32 stdout D 0x4
+GLIBC_2.32 stpcpy F
+GLIBC_2.32 stpncpy F
+GLIBC_2.32 strcasecmp F
+GLIBC_2.32 strcasecmp_l F
+GLIBC_2.32 strcasestr F
+GLIBC_2.32 strcat F
+GLIBC_2.32 strchr F
+GLIBC_2.32 strchrnul F
+GLIBC_2.32 strcmp F
+GLIBC_2.32 strcoll F
+GLIBC_2.32 strcoll_l F
+GLIBC_2.32 strcpy F
+GLIBC_2.32 strcspn F
+GLIBC_2.32 strdup F
+GLIBC_2.32 strerror F
+GLIBC_2.32 strerror_l F
+GLIBC_2.32 strerror_r F
+GLIBC_2.32 strfmon F
+GLIBC_2.32 strfmon_l F
+GLIBC_2.32 strfromd F
+GLIBC_2.32 strfromf F
+GLIBC_2.32 strfromf128 F
+GLIBC_2.32 strfromf32 F
+GLIBC_2.32 strfromf32x F
+GLIBC_2.32 strfromf64 F
+GLIBC_2.32 strfromf64x F
+GLIBC_2.32 strfroml F
+GLIBC_2.32 strfry F
+GLIBC_2.32 strftime F
+GLIBC_2.32 strftime_l F
+GLIBC_2.32 strlen F
+GLIBC_2.32 strncasecmp F
+GLIBC_2.32 strncasecmp_l F
+GLIBC_2.32 strncat F
+GLIBC_2.32 strncmp F
+GLIBC_2.32 strncpy F
+GLIBC_2.32 strndup F
+GLIBC_2.32 strnlen F
+GLIBC_2.32 strpbrk F
+GLIBC_2.32 strptime F
+GLIBC_2.32 strptime_l F
+GLIBC_2.32 strrchr F
+GLIBC_2.32 strsep F
+GLIBC_2.32 strsignal F
+GLIBC_2.32 strspn F
+GLIBC_2.32 strstr F
+GLIBC_2.32 strtod F
+GLIBC_2.32 strtod_l F
+GLIBC_2.32 strtof F
+GLIBC_2.32 strtof128 F
+GLIBC_2.32 strtof128_l F
+GLIBC_2.32 strtof32 F
+GLIBC_2.32 strtof32_l F
+GLIBC_2.32 strtof32x F
+GLIBC_2.32 strtof32x_l F
+GLIBC_2.32 strtof64 F
+GLIBC_2.32 strtof64_l F
+GLIBC_2.32 strtof64x F
+GLIBC_2.32 strtof64x_l F
+GLIBC_2.32 strtof_l F
+GLIBC_2.32 strtoimax F
+GLIBC_2.32 strtok F
+GLIBC_2.32 strtok_r F
+GLIBC_2.32 strtol F
+GLIBC_2.32 strtol_l F
+GLIBC_2.32 strtold F
+GLIBC_2.32 strtold_l F
+GLIBC_2.32 strtoll F
+GLIBC_2.32 strtoll_l F
+GLIBC_2.32 strtoq F
+GLIBC_2.32 strtoul F
+GLIBC_2.32 strtoul_l F
+GLIBC_2.32 strtoull F
+GLIBC_2.32 strtoull_l F
+GLIBC_2.32 strtoumax F
+GLIBC_2.32 strtouq F
+GLIBC_2.32 strverscmp F
+GLIBC_2.32 strxfrm F
+GLIBC_2.32 strxfrm_l F
+GLIBC_2.32 stty F
+GLIBC_2.32 svc_exit F
+GLIBC_2.32 svc_fdset D 0x80
+GLIBC_2.32 svc_getreq F
+GLIBC_2.32 svc_getreq_common F
+GLIBC_2.32 svc_getreq_poll F
+GLIBC_2.32 svc_getreqset F
+GLIBC_2.32 svc_max_pollfd D 0x4
+GLIBC_2.32 svc_pollfd D 0x4
+GLIBC_2.32 svc_register F
+GLIBC_2.32 svc_run F
+GLIBC_2.32 svc_sendreply F
+GLIBC_2.32 svc_unregister F
+GLIBC_2.32 svcauthdes_stats D 0xc
+GLIBC_2.32 svcerr_auth F
+GLIBC_2.32 svcerr_decode F
+GLIBC_2.32 svcerr_noproc F
+GLIBC_2.32 svcerr_noprog F
+GLIBC_2.32 svcerr_progvers F
+GLIBC_2.32 svcerr_systemerr F
+GLIBC_2.32 svcerr_weakauth F
+GLIBC_2.32 svcfd_create F
+GLIBC_2.32 svcraw_create F
+GLIBC_2.32 svctcp_create F
+GLIBC_2.32 svcudp_bufcreate F
+GLIBC_2.32 svcudp_create F
+GLIBC_2.32 svcudp_enablecache F
+GLIBC_2.32 svcunix_create F
+GLIBC_2.32 svcunixfd_create F
+GLIBC_2.32 swab F
+GLIBC_2.32 swapcontext F
+GLIBC_2.32 swapoff F
+GLIBC_2.32 swapon F
+GLIBC_2.32 swprintf F
+GLIBC_2.32 swscanf F
+GLIBC_2.32 symlink F
+GLIBC_2.32 symlinkat F
+GLIBC_2.32 sync F
+GLIBC_2.32 sync_file_range F
+GLIBC_2.32 syncfs F
+GLIBC_2.32 sys_errlist D 0x21c
+GLIBC_2.32 sys_nerr D 0x4
+GLIBC_2.32 sys_sigabbrev D 0x104
+GLIBC_2.32 sys_siglist D 0x104
+GLIBC_2.32 syscall F
+GLIBC_2.32 sysconf F
+GLIBC_2.32 sysinfo F
+GLIBC_2.32 syslog F
+GLIBC_2.32 system F
+GLIBC_2.32 sysv_signal F
+GLIBC_2.32 tcdrain F
+GLIBC_2.32 tcflow F
+GLIBC_2.32 tcflush F
+GLIBC_2.32 tcgetattr F
+GLIBC_2.32 tcgetpgrp F
+GLIBC_2.32 tcgetsid F
+GLIBC_2.32 tcsendbreak F
+GLIBC_2.32 tcsetattr F
+GLIBC_2.32 tcsetpgrp F
+GLIBC_2.32 tdelete F
+GLIBC_2.32 tdestroy F
+GLIBC_2.32 tee F
+GLIBC_2.32 telldir F
+GLIBC_2.32 tempnam F
+GLIBC_2.32 textdomain F
+GLIBC_2.32 tfind F
+GLIBC_2.32 tgkill F
+GLIBC_2.32 thrd_current F
+GLIBC_2.32 thrd_equal F
+GLIBC_2.32 thrd_sleep F
+GLIBC_2.32 thrd_yield F
+GLIBC_2.32 time F
+GLIBC_2.32 timegm F
+GLIBC_2.32 timelocal F
+GLIBC_2.32 timerfd_create F
+GLIBC_2.32 timerfd_gettime F
+GLIBC_2.32 timerfd_settime F
+GLIBC_2.32 times F
+GLIBC_2.32 timespec_get F
+GLIBC_2.32 timezone D 0x4
+GLIBC_2.32 tmpfile F
+GLIBC_2.32 tmpfile64 F
+GLIBC_2.32 tmpnam F
+GLIBC_2.32 tmpnam_r F
+GLIBC_2.32 toascii F
+GLIBC_2.32 tolower F
+GLIBC_2.32 tolower_l F
+GLIBC_2.32 toupper F
+GLIBC_2.32 toupper_l F
+GLIBC_2.32 towctrans F
+GLIBC_2.32 towctrans_l F
+GLIBC_2.32 towlower F
+GLIBC_2.32 towlower_l F
+GLIBC_2.32 towupper F
+GLIBC_2.32 towupper_l F
+GLIBC_2.32 tr_break F
+GLIBC_2.32 truncate F
+GLIBC_2.32 truncate64 F
+GLIBC_2.32 tsearch F
+GLIBC_2.32 ttyname F
+GLIBC_2.32 ttyname_r F
+GLIBC_2.32 ttyslot F
+GLIBC_2.32 twalk F
+GLIBC_2.32 twalk_r F
+GLIBC_2.32 tzname D 0x8
+GLIBC_2.32 tzset F
+GLIBC_2.32 ualarm F
+GLIBC_2.32 ulckpwdf F
+GLIBC_2.32 ulimit F
+GLIBC_2.32 umask F
+GLIBC_2.32 umount F
+GLIBC_2.32 umount2 F
+GLIBC_2.32 uname F
+GLIBC_2.32 ungetc F
+GLIBC_2.32 ungetwc F
+GLIBC_2.32 unlink F
+GLIBC_2.32 unlinkat F
+GLIBC_2.32 unlockpt F
+GLIBC_2.32 unsetenv F
+GLIBC_2.32 unshare F
+GLIBC_2.32 updwtmp F
+GLIBC_2.32 updwtmpx F
+GLIBC_2.32 uselocale F
+GLIBC_2.32 user2netname F
+GLIBC_2.32 usleep F
+GLIBC_2.32 utime F
+GLIBC_2.32 utimensat F
+GLIBC_2.32 utimes F
+GLIBC_2.32 utmpname F
+GLIBC_2.32 utmpxname F
+GLIBC_2.32 valloc F
+GLIBC_2.32 vasprintf F
+GLIBC_2.32 vdprintf F
+GLIBC_2.32 verr F
+GLIBC_2.32 verrx F
+GLIBC_2.32 versionsort F
+GLIBC_2.32 versionsort64 F
+GLIBC_2.32 vfork F
+GLIBC_2.32 vfprintf F
+GLIBC_2.32 vfscanf F
+GLIBC_2.32 vfwprintf F
+GLIBC_2.32 vfwscanf F
+GLIBC_2.32 vhangup F
+GLIBC_2.32 vlimit F
+GLIBC_2.32 vmsplice F
+GLIBC_2.32 vprintf F
+GLIBC_2.32 vscanf F
+GLIBC_2.32 vsnprintf F
+GLIBC_2.32 vsprintf F
+GLIBC_2.32 vsscanf F
+GLIBC_2.32 vswprintf F
+GLIBC_2.32 vswscanf F
+GLIBC_2.32 vsyslog F
+GLIBC_2.32 vtimes F
+GLIBC_2.32 vwarn F
+GLIBC_2.32 vwarnx F
+GLIBC_2.32 vwprintf F
+GLIBC_2.32 vwscanf F
+GLIBC_2.32 wait F
+GLIBC_2.32 wait3 F
+GLIBC_2.32 wait4 F
+GLIBC_2.32 waitid F
+GLIBC_2.32 waitpid F
+GLIBC_2.32 warn F
+GLIBC_2.32 warnx F
+GLIBC_2.32 wcpcpy F
+GLIBC_2.32 wcpncpy F
+GLIBC_2.32 wcrtomb F
+GLIBC_2.32 wcscasecmp F
+GLIBC_2.32 wcscasecmp_l F
+GLIBC_2.32 wcscat F
+GLIBC_2.32 wcschr F
+GLIBC_2.32 wcschrnul F
+GLIBC_2.32 wcscmp F
+GLIBC_2.32 wcscoll F
+GLIBC_2.32 wcscoll_l F
+GLIBC_2.32 wcscpy F
+GLIBC_2.32 wcscspn F
+GLIBC_2.32 wcsdup F
+GLIBC_2.32 wcsftime F
+GLIBC_2.32 wcsftime_l F
+GLIBC_2.32 wcslen F
+GLIBC_2.32 wcsncasecmp F
+GLIBC_2.32 wcsncasecmp_l F
+GLIBC_2.32 wcsncat F
+GLIBC_2.32 wcsncmp F
+GLIBC_2.32 wcsncpy F
+GLIBC_2.32 wcsnlen F
+GLIBC_2.32 wcsnrtombs F
+GLIBC_2.32 wcspbrk F
+GLIBC_2.32 wcsrchr F
+GLIBC_2.32 wcsrtombs F
+GLIBC_2.32 wcsspn F
+GLIBC_2.32 wcsstr F
+GLIBC_2.32 wcstod F
+GLIBC_2.32 wcstod_l F
+GLIBC_2.32 wcstof F
+GLIBC_2.32 wcstof128 F
+GLIBC_2.32 wcstof128_l F
+GLIBC_2.32 wcstof32 F
+GLIBC_2.32 wcstof32_l F
+GLIBC_2.32 wcstof32x F
+GLIBC_2.32 wcstof32x_l F
+GLIBC_2.32 wcstof64 F
+GLIBC_2.32 wcstof64_l F
+GLIBC_2.32 wcstof64x F
+GLIBC_2.32 wcstof64x_l F
+GLIBC_2.32 wcstof_l F
+GLIBC_2.32 wcstoimax F
+GLIBC_2.32 wcstok F
+GLIBC_2.32 wcstol F
+GLIBC_2.32 wcstol_l F
+GLIBC_2.32 wcstold F
+GLIBC_2.32 wcstold_l F
+GLIBC_2.32 wcstoll F
+GLIBC_2.32 wcstoll_l F
+GLIBC_2.32 wcstombs F
+GLIBC_2.32 wcstoq F
+GLIBC_2.32 wcstoul F
+GLIBC_2.32 wcstoul_l F
+GLIBC_2.32 wcstoull F
+GLIBC_2.32 wcstoull_l F
+GLIBC_2.32 wcstoumax F
+GLIBC_2.32 wcstouq F
+GLIBC_2.32 wcswcs F
+GLIBC_2.32 wcswidth F
+GLIBC_2.32 wcsxfrm F
+GLIBC_2.32 wcsxfrm_l F
+GLIBC_2.32 wctob F
+GLIBC_2.32 wctomb F
+GLIBC_2.32 wctrans F
+GLIBC_2.32 wctrans_l F
+GLIBC_2.32 wctype F
+GLIBC_2.32 wctype_l F
+GLIBC_2.32 wcwidth F
+GLIBC_2.32 wmemchr F
+GLIBC_2.32 wmemcmp F
+GLIBC_2.32 wmemcpy F
+GLIBC_2.32 wmemmove F
+GLIBC_2.32 wmempcpy F
+GLIBC_2.32 wmemset F
+GLIBC_2.32 wordexp F
+GLIBC_2.32 wordfree F
+GLIBC_2.32 wprintf F
+GLIBC_2.32 write F
+GLIBC_2.32 writev F
+GLIBC_2.32 wscanf F
+GLIBC_2.32 xdecrypt F
+GLIBC_2.32 xdr_accepted_reply F
+GLIBC_2.32 xdr_array F
+GLIBC_2.32 xdr_authdes_cred F
+GLIBC_2.32 xdr_authdes_verf F
+GLIBC_2.32 xdr_authunix_parms F
+GLIBC_2.32 xdr_bool F
+GLIBC_2.32 xdr_bytes F
+GLIBC_2.32 xdr_callhdr F
+GLIBC_2.32 xdr_callmsg F
+GLIBC_2.32 xdr_char F
+GLIBC_2.32 xdr_cryptkeyarg F
+GLIBC_2.32 xdr_cryptkeyarg2 F
+GLIBC_2.32 xdr_cryptkeyres F
+GLIBC_2.32 xdr_des_block F
+GLIBC_2.32 xdr_double F
+GLIBC_2.32 xdr_enum F
+GLIBC_2.32 xdr_float F
+GLIBC_2.32 xdr_free F
+GLIBC_2.32 xdr_getcredres F
+GLIBC_2.32 xdr_hyper F
+GLIBC_2.32 xdr_int F
+GLIBC_2.32 xdr_int16_t F
+GLIBC_2.32 xdr_int32_t F
+GLIBC_2.32 xdr_int64_t F
+GLIBC_2.32 xdr_int8_t F
+GLIBC_2.32 xdr_key_netstarg F
+GLIBC_2.32 xdr_key_netstres F
+GLIBC_2.32 xdr_keybuf F
+GLIBC_2.32 xdr_keystatus F
+GLIBC_2.32 xdr_long F
+GLIBC_2.32 xdr_longlong_t F
+GLIBC_2.32 xdr_netnamestr F
+GLIBC_2.32 xdr_netobj F
+GLIBC_2.32 xdr_opaque F
+GLIBC_2.32 xdr_opaque_auth F
+GLIBC_2.32 xdr_pmap F
+GLIBC_2.32 xdr_pmaplist F
+GLIBC_2.32 xdr_pointer F
+GLIBC_2.32 xdr_quad_t F
+GLIBC_2.32 xdr_reference F
+GLIBC_2.32 xdr_rejected_reply F
+GLIBC_2.32 xdr_replymsg F
+GLIBC_2.32 xdr_rmtcall_args F
+GLIBC_2.32 xdr_rmtcallres F
+GLIBC_2.32 xdr_short F
+GLIBC_2.32 xdr_sizeof F
+GLIBC_2.32 xdr_string F
+GLIBC_2.32 xdr_u_char F
+GLIBC_2.32 xdr_u_hyper F
+GLIBC_2.32 xdr_u_int F
+GLIBC_2.32 xdr_u_long F
+GLIBC_2.32 xdr_u_longlong_t F
+GLIBC_2.32 xdr_u_quad_t F
+GLIBC_2.32 xdr_u_short F
+GLIBC_2.32 xdr_uint16_t F
+GLIBC_2.32 xdr_uint32_t F
+GLIBC_2.32 xdr_uint64_t F
+GLIBC_2.32 xdr_uint8_t F
+GLIBC_2.32 xdr_union F
+GLIBC_2.32 xdr_unixcred F
+GLIBC_2.32 xdr_vector F
+GLIBC_2.32 xdr_void F
+GLIBC_2.32 xdr_wrapstring F
+GLIBC_2.32 xdrmem_create F
+GLIBC_2.32 xdrrec_create F
+GLIBC_2.32 xdrrec_endofrecord F
+GLIBC_2.32 xdrrec_eof F
+GLIBC_2.32 xdrrec_skiprecord F
+GLIBC_2.32 xdrstdio_create F
+GLIBC_2.32 xencrypt F
+GLIBC_2.32 xprt_register F
+GLIBC_2.32 xprt_unregister F
diff --git a/sysdeps/unix/sysv/linux/riscv/libcrypt.abilist b/sysdeps/unix/sysv/linux/riscv/libcrypt.abilist
new file mode 100644
index 0000000000..6bd253453e
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/libcrypt.abilist
@@ -0,0 +1,2 @@
+GLIBC_2.32 crypt F
+GLIBC_2.32 crypt_r F
diff --git a/sysdeps/unix/sysv/linux/riscv/libdl.abilist b/sysdeps/unix/sysv/linux/riscv/libdl.abilist
new file mode 100644
index 0000000000..bf20b0c404
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/libdl.abilist
@@ -0,0 +1,9 @@
+GLIBC_2.32 dladdr F
+GLIBC_2.32 dladdr1 F
+GLIBC_2.32 dlclose F
+GLIBC_2.32 dlerror F
+GLIBC_2.32 dlinfo F
+GLIBC_2.32 dlmopen F
+GLIBC_2.32 dlopen F
+GLIBC_2.32 dlsym F
+GLIBC_2.32 dlvsym F
diff --git a/sysdeps/unix/sysv/linux/riscv/libm.abilist b/sysdeps/unix/sysv/linux/riscv/libm.abilist
new file mode 100644
index 0000000000..a79e180258
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/libm.abilist
@@ -0,0 +1,940 @@
+GLIBC_2.32 __clog10 F
+GLIBC_2.32 __clog10f F
+GLIBC_2.32 __clog10l F
+GLIBC_2.32 __finite F
+GLIBC_2.32 __finitef F
+GLIBC_2.32 __finitel F
+GLIBC_2.32 __fpclassify F
+GLIBC_2.32 __fpclassifyf F
+GLIBC_2.32 __fpclassifyl F
+GLIBC_2.32 __iseqsig F
+GLIBC_2.32 __iseqsigf F
+GLIBC_2.32 __iseqsigl F
+GLIBC_2.32 __issignaling F
+GLIBC_2.32 __issignalingf F
+GLIBC_2.32 __issignalingl F
+GLIBC_2.32 __signbit F
+GLIBC_2.32 __signbitf F
+GLIBC_2.32 __signbitl F
+GLIBC_2.32 __signgam D 0x4
+GLIBC_2.32 acos F
+GLIBC_2.32 acosf F
+GLIBC_2.32 acosf128 F
+GLIBC_2.32 acosf32 F
+GLIBC_2.32 acosf32x F
+GLIBC_2.32 acosf64 F
+GLIBC_2.32 acosf64x F
+GLIBC_2.32 acosh F
+GLIBC_2.32 acoshf F
+GLIBC_2.32 acoshf128 F
+GLIBC_2.32 acoshf32 F
+GLIBC_2.32 acoshf32x F
+GLIBC_2.32 acoshf64 F
+GLIBC_2.32 acoshf64x F
+GLIBC_2.32 acoshl F
+GLIBC_2.32 acosl F
+GLIBC_2.32 asin F
+GLIBC_2.32 asinf F
+GLIBC_2.32 asinf128 F
+GLIBC_2.32 asinf32 F
+GLIBC_2.32 asinf32x F
+GLIBC_2.32 asinf64 F
+GLIBC_2.32 asinf64x F
+GLIBC_2.32 asinh F
+GLIBC_2.32 asinhf F
+GLIBC_2.32 asinhf128 F
+GLIBC_2.32 asinhf32 F
+GLIBC_2.32 asinhf32x F
+GLIBC_2.32 asinhf64 F
+GLIBC_2.32 asinhf64x F
+GLIBC_2.32 asinhl F
+GLIBC_2.32 asinl F
+GLIBC_2.32 atan F
+GLIBC_2.32 atan2 F
+GLIBC_2.32 atan2f F
+GLIBC_2.32 atan2f128 F
+GLIBC_2.32 atan2f32 F
+GLIBC_2.32 atan2f32x F
+GLIBC_2.32 atan2f64 F
+GLIBC_2.32 atan2f64x F
+GLIBC_2.32 atan2l F
+GLIBC_2.32 atanf F
+GLIBC_2.32 atanf128 F
+GLIBC_2.32 atanf32 F
+GLIBC_2.32 atanf32x F
+GLIBC_2.32 atanf64 F
+GLIBC_2.32 atanf64x F
+GLIBC_2.32 atanh F
+GLIBC_2.32 atanhf F
+GLIBC_2.32 atanhf128 F
+GLIBC_2.32 atanhf32 F
+GLIBC_2.32 atanhf32x F
+GLIBC_2.32 atanhf64 F
+GLIBC_2.32 atanhf64x F
+GLIBC_2.32 atanhl F
+GLIBC_2.32 atanl F
+GLIBC_2.32 cabs F
+GLIBC_2.32 cabsf F
+GLIBC_2.32 cabsf128 F
+GLIBC_2.32 cabsf32 F
+GLIBC_2.32 cabsf32x F
+GLIBC_2.32 cabsf64 F
+GLIBC_2.32 cabsf64x F
+GLIBC_2.32 cabsl F
+GLIBC_2.32 cacos F
+GLIBC_2.32 cacosf F
+GLIBC_2.32 cacosf128 F
+GLIBC_2.32 cacosf32 F
+GLIBC_2.32 cacosf32x F
+GLIBC_2.32 cacosf64 F
+GLIBC_2.32 cacosf64x F
+GLIBC_2.32 cacosh F
+GLIBC_2.32 cacoshf F
+GLIBC_2.32 cacoshf128 F
+GLIBC_2.32 cacoshf32 F
+GLIBC_2.32 cacoshf32x F
+GLIBC_2.32 cacoshf64 F
+GLIBC_2.32 cacoshf64x F
+GLIBC_2.32 cacoshl F
+GLIBC_2.32 cacosl F
+GLIBC_2.32 canonicalize F
+GLIBC_2.32 canonicalizef F
+GLIBC_2.32 canonicalizef128 F
+GLIBC_2.32 canonicalizef32 F
+GLIBC_2.32 canonicalizef32x F
+GLIBC_2.32 canonicalizef64 F
+GLIBC_2.32 canonicalizef64x F
+GLIBC_2.32 canonicalizel F
+GLIBC_2.32 carg F
+GLIBC_2.32 cargf F
+GLIBC_2.32 cargf128 F
+GLIBC_2.32 cargf32 F
+GLIBC_2.32 cargf32x F
+GLIBC_2.32 cargf64 F
+GLIBC_2.32 cargf64x F
+GLIBC_2.32 cargl F
+GLIBC_2.32 casin F
+GLIBC_2.32 casinf F
+GLIBC_2.32 casinf128 F
+GLIBC_2.32 casinf32 F
+GLIBC_2.32 casinf32x F
+GLIBC_2.32 casinf64 F
+GLIBC_2.32 casinf64x F
+GLIBC_2.32 casinh F
+GLIBC_2.32 casinhf F
+GLIBC_2.32 casinhf128 F
+GLIBC_2.32 casinhf32 F
+GLIBC_2.32 casinhf32x F
+GLIBC_2.32 casinhf64 F
+GLIBC_2.32 casinhf64x F
+GLIBC_2.32 casinhl F
+GLIBC_2.32 casinl F
+GLIBC_2.32 catan F
+GLIBC_2.32 catanf F
+GLIBC_2.32 catanf128 F
+GLIBC_2.32 catanf32 F
+GLIBC_2.32 catanf32x F
+GLIBC_2.32 catanf64 F
+GLIBC_2.32 catanf64x F
+GLIBC_2.32 catanh F
+GLIBC_2.32 catanhf F
+GLIBC_2.32 catanhf128 F
+GLIBC_2.32 catanhf32 F
+GLIBC_2.32 catanhf32x F
+GLIBC_2.32 catanhf64 F
+GLIBC_2.32 catanhf64x F
+GLIBC_2.32 catanhl F
+GLIBC_2.32 catanl F
+GLIBC_2.32 cbrt F
+GLIBC_2.32 cbrtf F
+GLIBC_2.32 cbrtf128 F
+GLIBC_2.32 cbrtf32 F
+GLIBC_2.32 cbrtf32x F
+GLIBC_2.32 cbrtf64 F
+GLIBC_2.32 cbrtf64x F
+GLIBC_2.32 cbrtl F
+GLIBC_2.32 ccos F
+GLIBC_2.32 ccosf F
+GLIBC_2.32 ccosf128 F
+GLIBC_2.32 ccosf32 F
+GLIBC_2.32 ccosf32x F
+GLIBC_2.32 ccosf64 F
+GLIBC_2.32 ccosf64x F
+GLIBC_2.32 ccosh F
+GLIBC_2.32 ccoshf F
+GLIBC_2.32 ccoshf128 F
+GLIBC_2.32 ccoshf32 F
+GLIBC_2.32 ccoshf32x F
+GLIBC_2.32 ccoshf64 F
+GLIBC_2.32 ccoshf64x F
+GLIBC_2.32 ccoshl F
+GLIBC_2.32 ccosl F
+GLIBC_2.32 ceil F
+GLIBC_2.32 ceilf F
+GLIBC_2.32 ceilf128 F
+GLIBC_2.32 ceilf32 F
+GLIBC_2.32 ceilf32x F
+GLIBC_2.32 ceilf64 F
+GLIBC_2.32 ceilf64x F
+GLIBC_2.32 ceill F
+GLIBC_2.32 cexp F
+GLIBC_2.32 cexpf F
+GLIBC_2.32 cexpf128 F
+GLIBC_2.32 cexpf32 F
+GLIBC_2.32 cexpf32x F
+GLIBC_2.32 cexpf64 F
+GLIBC_2.32 cexpf64x F
+GLIBC_2.32 cexpl F
+GLIBC_2.32 cimag F
+GLIBC_2.32 cimagf F
+GLIBC_2.32 cimagf128 F
+GLIBC_2.32 cimagf32 F
+GLIBC_2.32 cimagf32x F
+GLIBC_2.32 cimagf64 F
+GLIBC_2.32 cimagf64x F
+GLIBC_2.32 cimagl F
+GLIBC_2.32 clog F
+GLIBC_2.32 clog10 F
+GLIBC_2.32 clog10f F
+GLIBC_2.32 clog10f128 F
+GLIBC_2.32 clog10f32 F
+GLIBC_2.32 clog10f32x F
+GLIBC_2.32 clog10f64 F
+GLIBC_2.32 clog10f64x F
+GLIBC_2.32 clog10l F
+GLIBC_2.32 clogf F
+GLIBC_2.32 clogf128 F
+GLIBC_2.32 clogf32 F
+GLIBC_2.32 clogf32x F
+GLIBC_2.32 clogf64 F
+GLIBC_2.32 clogf64x F
+GLIBC_2.32 clogl F
+GLIBC_2.32 conj F
+GLIBC_2.32 conjf F
+GLIBC_2.32 conjf128 F
+GLIBC_2.32 conjf32 F
+GLIBC_2.32 conjf32x F
+GLIBC_2.32 conjf64 F
+GLIBC_2.32 conjf64x F
+GLIBC_2.32 conjl F
+GLIBC_2.32 copysign F
+GLIBC_2.32 copysignf F
+GLIBC_2.32 copysignf128 F
+GLIBC_2.32 copysignf32 F
+GLIBC_2.32 copysignf32x F
+GLIBC_2.32 copysignf64 F
+GLIBC_2.32 copysignf64x F
+GLIBC_2.32 copysignl F
+GLIBC_2.32 cos F
+GLIBC_2.32 cosf F
+GLIBC_2.32 cosf128 F
+GLIBC_2.32 cosf32 F
+GLIBC_2.32 cosf32x F
+GLIBC_2.32 cosf64 F
+GLIBC_2.32 cosf64x F
+GLIBC_2.32 cosh F
+GLIBC_2.32 coshf F
+GLIBC_2.32 coshf128 F
+GLIBC_2.32 coshf32 F
+GLIBC_2.32 coshf32x F
+GLIBC_2.32 coshf64 F
+GLIBC_2.32 coshf64x F
+GLIBC_2.32 coshl F
+GLIBC_2.32 cosl F
+GLIBC_2.32 cpow F
+GLIBC_2.32 cpowf F
+GLIBC_2.32 cpowf128 F
+GLIBC_2.32 cpowf32 F
+GLIBC_2.32 cpowf32x F
+GLIBC_2.32 cpowf64 F
+GLIBC_2.32 cpowf64x F
+GLIBC_2.32 cpowl F
+GLIBC_2.32 cproj F
+GLIBC_2.32 cprojf F
+GLIBC_2.32 cprojf128 F
+GLIBC_2.32 cprojf32 F
+GLIBC_2.32 cprojf32x F
+GLIBC_2.32 cprojf64 F
+GLIBC_2.32 cprojf64x F
+GLIBC_2.32 cprojl F
+GLIBC_2.32 creal F
+GLIBC_2.32 crealf F
+GLIBC_2.32 crealf128 F
+GLIBC_2.32 crealf32 F
+GLIBC_2.32 crealf32x F
+GLIBC_2.32 crealf64 F
+GLIBC_2.32 crealf64x F
+GLIBC_2.32 creall F
+GLIBC_2.32 csin F
+GLIBC_2.32 csinf F
+GLIBC_2.32 csinf128 F
+GLIBC_2.32 csinf32 F
+GLIBC_2.32 csinf32x F
+GLIBC_2.32 csinf64 F
+GLIBC_2.32 csinf64x F
+GLIBC_2.32 csinh F
+GLIBC_2.32 csinhf F
+GLIBC_2.32 csinhf128 F
+GLIBC_2.32 csinhf32 F
+GLIBC_2.32 csinhf32x F
+GLIBC_2.32 csinhf64 F
+GLIBC_2.32 csinhf64x F
+GLIBC_2.32 csinhl F
+GLIBC_2.32 csinl F
+GLIBC_2.32 csqrt F
+GLIBC_2.32 csqrtf F
+GLIBC_2.32 csqrtf128 F
+GLIBC_2.32 csqrtf32 F
+GLIBC_2.32 csqrtf32x F
+GLIBC_2.32 csqrtf64 F
+GLIBC_2.32 csqrtf64x F
+GLIBC_2.32 csqrtl F
+GLIBC_2.32 ctan F
+GLIBC_2.32 ctanf F
+GLIBC_2.32 ctanf128 F
+GLIBC_2.32 ctanf32 F
+GLIBC_2.32 ctanf32x F
+GLIBC_2.32 ctanf64 F
+GLIBC_2.32 ctanf64x F
+GLIBC_2.32 ctanh F
+GLIBC_2.32 ctanhf F
+GLIBC_2.32 ctanhf128 F
+GLIBC_2.32 ctanhf32 F
+GLIBC_2.32 ctanhf32x F
+GLIBC_2.32 ctanhf64 F
+GLIBC_2.32 ctanhf64x F
+GLIBC_2.32 ctanhl F
+GLIBC_2.32 ctanl F
+GLIBC_2.32 daddl F
+GLIBC_2.32 ddivl F
+GLIBC_2.32 dmull F
+GLIBC_2.32 drem F
+GLIBC_2.32 dremf F
+GLIBC_2.32 dreml F
+GLIBC_2.32 dsubl F
+GLIBC_2.32 erf F
+GLIBC_2.32 erfc F
+GLIBC_2.32 erfcf F
+GLIBC_2.32 erfcf128 F
+GLIBC_2.32 erfcf32 F
+GLIBC_2.32 erfcf32x F
+GLIBC_2.32 erfcf64 F
+GLIBC_2.32 erfcf64x F
+GLIBC_2.32 erfcl F
+GLIBC_2.32 erff F
+GLIBC_2.32 erff128 F
+GLIBC_2.32 erff32 F
+GLIBC_2.32 erff32x F
+GLIBC_2.32 erff64 F
+GLIBC_2.32 erff64x F
+GLIBC_2.32 erfl F
+GLIBC_2.32 exp F
+GLIBC_2.32 exp10 F
+GLIBC_2.32 exp10f F
+GLIBC_2.32 exp10f128 F
+GLIBC_2.32 exp10f32 F
+GLIBC_2.32 exp10f32x F
+GLIBC_2.32 exp10f64 F
+GLIBC_2.32 exp10f64x F
+GLIBC_2.32 exp10l F
+GLIBC_2.32 exp2 F
+GLIBC_2.32 exp2f F
+GLIBC_2.32 exp2f128 F
+GLIBC_2.32 exp2f32 F
+GLIBC_2.32 exp2f32x F
+GLIBC_2.32 exp2f64 F
+GLIBC_2.32 exp2f64x F
+GLIBC_2.32 exp2l F
+GLIBC_2.32 expf F
+GLIBC_2.32 expf128 F
+GLIBC_2.32 expf32 F
+GLIBC_2.32 expf32x F
+GLIBC_2.32 expf64 F
+GLIBC_2.32 expf64x F
+GLIBC_2.32 expl F
+GLIBC_2.32 expm1 F
+GLIBC_2.32 expm1f F
+GLIBC_2.32 expm1f128 F
+GLIBC_2.32 expm1f32 F
+GLIBC_2.32 expm1f32x F
+GLIBC_2.32 expm1f64 F
+GLIBC_2.32 expm1f64x F
+GLIBC_2.32 expm1l F
+GLIBC_2.32 f32addf128 F
+GLIBC_2.32 f32addf32x F
+GLIBC_2.32 f32addf64 F
+GLIBC_2.32 f32addf64x F
+GLIBC_2.32 f32divf128 F
+GLIBC_2.32 f32divf32x F
+GLIBC_2.32 f32divf64 F
+GLIBC_2.32 f32divf64x F
+GLIBC_2.32 f32mulf128 F
+GLIBC_2.32 f32mulf32x F
+GLIBC_2.32 f32mulf64 F
+GLIBC_2.32 f32mulf64x F
+GLIBC_2.32 f32subf128 F
+GLIBC_2.32 f32subf32x F
+GLIBC_2.32 f32subf64 F
+GLIBC_2.32 f32subf64x F
+GLIBC_2.32 f32xaddf128 F
+GLIBC_2.32 f32xaddf64 F
+GLIBC_2.32 f32xaddf64x F
+GLIBC_2.32 f32xdivf128 F
+GLIBC_2.32 f32xdivf64 F
+GLIBC_2.32 f32xdivf64x F
+GLIBC_2.32 f32xmulf128 F
+GLIBC_2.32 f32xmulf64 F
+GLIBC_2.32 f32xmulf64x F
+GLIBC_2.32 f32xsubf128 F
+GLIBC_2.32 f32xsubf64 F
+GLIBC_2.32 f32xsubf64x F
+GLIBC_2.32 f64addf128 F
+GLIBC_2.32 f64addf64x F
+GLIBC_2.32 f64divf128 F
+GLIBC_2.32 f64divf64x F
+GLIBC_2.32 f64mulf128 F
+GLIBC_2.32 f64mulf64x F
+GLIBC_2.32 f64subf128 F
+GLIBC_2.32 f64subf64x F
+GLIBC_2.32 f64xaddf128 F
+GLIBC_2.32 f64xdivf128 F
+GLIBC_2.32 f64xmulf128 F
+GLIBC_2.32 f64xsubf128 F
+GLIBC_2.32 fabs F
+GLIBC_2.32 fabsf F
+GLIBC_2.32 fabsf128 F
+GLIBC_2.32 fabsf32 F
+GLIBC_2.32 fabsf32x F
+GLIBC_2.32 fabsf64 F
+GLIBC_2.32 fabsf64x F
+GLIBC_2.32 fabsl F
+GLIBC_2.32 fadd F
+GLIBC_2.32 faddl F
+GLIBC_2.32 fdim F
+GLIBC_2.32 fdimf F
+GLIBC_2.32 fdimf128 F
+GLIBC_2.32 fdimf32 F
+GLIBC_2.32 fdimf32x F
+GLIBC_2.32 fdimf64 F
+GLIBC_2.32 fdimf64x F
+GLIBC_2.32 fdiml F
+GLIBC_2.32 fdiv F
+GLIBC_2.32 fdivl F
+GLIBC_2.32 feclearexcept F
+GLIBC_2.32 fedisableexcept F
+GLIBC_2.32 feenableexcept F
+GLIBC_2.32 fegetenv F
+GLIBC_2.32 fegetexcept F
+GLIBC_2.32 fegetexceptflag F
+GLIBC_2.32 fegetmode F
+GLIBC_2.32 fegetround F
+GLIBC_2.32 feholdexcept F
+GLIBC_2.32 feraiseexcept F
+GLIBC_2.32 fesetenv F
+GLIBC_2.32 fesetexcept F
+GLIBC_2.32 fesetexceptflag F
+GLIBC_2.32 fesetmode F
+GLIBC_2.32 fesetround F
+GLIBC_2.32 fetestexcept F
+GLIBC_2.32 fetestexceptflag F
+GLIBC_2.32 feupdateenv F
+GLIBC_2.32 finite F
+GLIBC_2.32 finitef F
+GLIBC_2.32 finitel F
+GLIBC_2.32 floor F
+GLIBC_2.32 floorf F
+GLIBC_2.32 floorf128 F
+GLIBC_2.32 floorf32 F
+GLIBC_2.32 floorf32x F
+GLIBC_2.32 floorf64 F
+GLIBC_2.32 floorf64x F
+GLIBC_2.32 floorl F
+GLIBC_2.32 fma F
+GLIBC_2.32 fmaf F
+GLIBC_2.32 fmaf128 F
+GLIBC_2.32 fmaf32 F
+GLIBC_2.32 fmaf32x F
+GLIBC_2.32 fmaf64 F
+GLIBC_2.32 fmaf64x F
+GLIBC_2.32 fmal F
+GLIBC_2.32 fmax F
+GLIBC_2.32 fmaxf F
+GLIBC_2.32 fmaxf128 F
+GLIBC_2.32 fmaxf32 F
+GLIBC_2.32 fmaxf32x F
+GLIBC_2.32 fmaxf64 F
+GLIBC_2.32 fmaxf64x F
+GLIBC_2.32 fmaxl F
+GLIBC_2.32 fmaxmag F
+GLIBC_2.32 fmaxmagf F
+GLIBC_2.32 fmaxmagf128 F
+GLIBC_2.32 fmaxmagf32 F
+GLIBC_2.32 fmaxmagf32x F
+GLIBC_2.32 fmaxmagf64 F
+GLIBC_2.32 fmaxmagf64x F
+GLIBC_2.32 fmaxmagl F
+GLIBC_2.32 fmin F
+GLIBC_2.32 fminf F
+GLIBC_2.32 fminf128 F
+GLIBC_2.32 fminf32 F
+GLIBC_2.32 fminf32x F
+GLIBC_2.32 fminf64 F
+GLIBC_2.32 fminf64x F
+GLIBC_2.32 fminl F
+GLIBC_2.32 fminmag F
+GLIBC_2.32 fminmagf F
+GLIBC_2.32 fminmagf128 F
+GLIBC_2.32 fminmagf32 F
+GLIBC_2.32 fminmagf32x F
+GLIBC_2.32 fminmagf64 F
+GLIBC_2.32 fminmagf64x F
+GLIBC_2.32 fminmagl F
+GLIBC_2.32 fmod F
+GLIBC_2.32 fmodf F
+GLIBC_2.32 fmodf128 F
+GLIBC_2.32 fmodf32 F
+GLIBC_2.32 fmodf32x F
+GLIBC_2.32 fmodf64 F
+GLIBC_2.32 fmodf64x F
+GLIBC_2.32 fmodl F
+GLIBC_2.32 fmul F
+GLIBC_2.32 fmull F
+GLIBC_2.32 frexp F
+GLIBC_2.32 frexpf F
+GLIBC_2.32 frexpf128 F
+GLIBC_2.32 frexpf32 F
+GLIBC_2.32 frexpf32x F
+GLIBC_2.32 frexpf64 F
+GLIBC_2.32 frexpf64x F
+GLIBC_2.32 frexpl F
+GLIBC_2.32 fromfp F
+GLIBC_2.32 fromfpf F
+GLIBC_2.32 fromfpf128 F
+GLIBC_2.32 fromfpf32 F
+GLIBC_2.32 fromfpf32x F
+GLIBC_2.32 fromfpf64 F
+GLIBC_2.32 fromfpf64x F
+GLIBC_2.32 fromfpl F
+GLIBC_2.32 fromfpx F
+GLIBC_2.32 fromfpxf F
+GLIBC_2.32 fromfpxf128 F
+GLIBC_2.32 fromfpxf32 F
+GLIBC_2.32 fromfpxf32x F
+GLIBC_2.32 fromfpxf64 F
+GLIBC_2.32 fromfpxf64x F
+GLIBC_2.32 fromfpxl F
+GLIBC_2.32 fsub F
+GLIBC_2.32 fsubl F
+GLIBC_2.32 gamma F
+GLIBC_2.32 gammaf F
+GLIBC_2.32 gammal F
+GLIBC_2.32 getpayload F
+GLIBC_2.32 getpayloadf F
+GLIBC_2.32 getpayloadf128 F
+GLIBC_2.32 getpayloadf32 F
+GLIBC_2.32 getpayloadf32x F
+GLIBC_2.32 getpayloadf64 F
+GLIBC_2.32 getpayloadf64x F
+GLIBC_2.32 getpayloadl F
+GLIBC_2.32 hypot F
+GLIBC_2.32 hypotf F
+GLIBC_2.32 hypotf128 F
+GLIBC_2.32 hypotf32 F
+GLIBC_2.32 hypotf32x F
+GLIBC_2.32 hypotf64 F
+GLIBC_2.32 hypotf64x F
+GLIBC_2.32 hypotl F
+GLIBC_2.32 ilogb F
+GLIBC_2.32 ilogbf F
+GLIBC_2.32 ilogbf128 F
+GLIBC_2.32 ilogbf32 F
+GLIBC_2.32 ilogbf32x F
+GLIBC_2.32 ilogbf64 F
+GLIBC_2.32 ilogbf64x F
+GLIBC_2.32 ilogbl F
+GLIBC_2.32 j0 F
+GLIBC_2.32 j0f F
+GLIBC_2.32 j0f128 F
+GLIBC_2.32 j0f32 F
+GLIBC_2.32 j0f32x F
+GLIBC_2.32 j0f64 F
+GLIBC_2.32 j0f64x F
+GLIBC_2.32 j0l F
+GLIBC_2.32 j1 F
+GLIBC_2.32 j1f F
+GLIBC_2.32 j1f128 F
+GLIBC_2.32 j1f32 F
+GLIBC_2.32 j1f32x F
+GLIBC_2.32 j1f64 F
+GLIBC_2.32 j1f64x F
+GLIBC_2.32 j1l F
+GLIBC_2.32 jn F
+GLIBC_2.32 jnf F
+GLIBC_2.32 jnf128 F
+GLIBC_2.32 jnf32 F
+GLIBC_2.32 jnf32x F
+GLIBC_2.32 jnf64 F
+GLIBC_2.32 jnf64x F
+GLIBC_2.32 jnl F
+GLIBC_2.32 ldexp F
+GLIBC_2.32 ldexpf F
+GLIBC_2.32 ldexpf128 F
+GLIBC_2.32 ldexpf32 F
+GLIBC_2.32 ldexpf32x F
+GLIBC_2.32 ldexpf64 F
+GLIBC_2.32 ldexpf64x F
+GLIBC_2.32 ldexpl F
+GLIBC_2.32 lgamma F
+GLIBC_2.32 lgamma_r F
+GLIBC_2.32 lgammaf F
+GLIBC_2.32 lgammaf128 F
+GLIBC_2.32 lgammaf128_r F
+GLIBC_2.32 lgammaf32 F
+GLIBC_2.32 lgammaf32_r F
+GLIBC_2.32 lgammaf32x F
+GLIBC_2.32 lgammaf32x_r F
+GLIBC_2.32 lgammaf64 F
+GLIBC_2.32 lgammaf64_r F
+GLIBC_2.32 lgammaf64x F
+GLIBC_2.32 lgammaf64x_r F
+GLIBC_2.32 lgammaf_r F
+GLIBC_2.32 lgammal F
+GLIBC_2.32 lgammal_r F
+GLIBC_2.32 llogb F
+GLIBC_2.32 llogbf F
+GLIBC_2.32 llogbf128 F
+GLIBC_2.32 llogbf32 F
+GLIBC_2.32 llogbf32x F
+GLIBC_2.32 llogbf64 F
+GLIBC_2.32 llogbf64x F
+GLIBC_2.32 llogbl F
+GLIBC_2.32 llrint F
+GLIBC_2.32 llrintf F
+GLIBC_2.32 llrintf128 F
+GLIBC_2.32 llrintf32 F
+GLIBC_2.32 llrintf32x F
+GLIBC_2.32 llrintf64 F
+GLIBC_2.32 llrintf64x F
+GLIBC_2.32 llrintl F
+GLIBC_2.32 llround F
+GLIBC_2.32 llroundf F
+GLIBC_2.32 llroundf128 F
+GLIBC_2.32 llroundf32 F
+GLIBC_2.32 llroundf32x F
+GLIBC_2.32 llroundf64 F
+GLIBC_2.32 llroundf64x F
+GLIBC_2.32 llroundl F
+GLIBC_2.32 log F
+GLIBC_2.32 log10 F
+GLIBC_2.32 log10f F
+GLIBC_2.32 log10f128 F
+GLIBC_2.32 log10f32 F
+GLIBC_2.32 log10f32x F
+GLIBC_2.32 log10f64 F
+GLIBC_2.32 log10f64x F
+GLIBC_2.32 log10l F
+GLIBC_2.32 log1p F
+GLIBC_2.32 log1pf F
+GLIBC_2.32 log1pf128 F
+GLIBC_2.32 log1pf32 F
+GLIBC_2.32 log1pf32x F
+GLIBC_2.32 log1pf64 F
+GLIBC_2.32 log1pf64x F
+GLIBC_2.32 log1pl F
+GLIBC_2.32 log2 F
+GLIBC_2.32 log2f F
+GLIBC_2.32 log2f128 F
+GLIBC_2.32 log2f32 F
+GLIBC_2.32 log2f32x F
+GLIBC_2.32 log2f64 F
+GLIBC_2.32 log2f64x F
+GLIBC_2.32 log2l F
+GLIBC_2.32 logb F
+GLIBC_2.32 logbf F
+GLIBC_2.32 logbf128 F
+GLIBC_2.32 logbf32 F
+GLIBC_2.32 logbf32x F
+GLIBC_2.32 logbf64 F
+GLIBC_2.32 logbf64x F
+GLIBC_2.32 logbl F
+GLIBC_2.32 logf F
+GLIBC_2.32 logf128 F
+GLIBC_2.32 logf32 F
+GLIBC_2.32 logf32x F
+GLIBC_2.32 logf64 F
+GLIBC_2.32 logf64x F
+GLIBC_2.32 logl F
+GLIBC_2.32 lrint F
+GLIBC_2.32 lrintf F
+GLIBC_2.32 lrintf128 F
+GLIBC_2.32 lrintf32 F
+GLIBC_2.32 lrintf32x F
+GLIBC_2.32 lrintf64 F
+GLIBC_2.32 lrintf64x F
+GLIBC_2.32 lrintl F
+GLIBC_2.32 lround F
+GLIBC_2.32 lroundf F
+GLIBC_2.32 lroundf128 F
+GLIBC_2.32 lroundf32 F
+GLIBC_2.32 lroundf32x F
+GLIBC_2.32 lroundf64 F
+GLIBC_2.32 lroundf64x F
+GLIBC_2.32 lroundl F
+GLIBC_2.32 modf F
+GLIBC_2.32 modff F
+GLIBC_2.32 modff128 F
+GLIBC_2.32 modff32 F
+GLIBC_2.32 modff32x F
+GLIBC_2.32 modff64 F
+GLIBC_2.32 modff64x F
+GLIBC_2.32 modfl F
+GLIBC_2.32 nan F
+GLIBC_2.32 nanf F
+GLIBC_2.32 nanf128 F
+GLIBC_2.32 nanf32 F
+GLIBC_2.32 nanf32x F
+GLIBC_2.32 nanf64 F
+GLIBC_2.32 nanf64x F
+GLIBC_2.32 nanl F
+GLIBC_2.32 nearbyint F
+GLIBC_2.32 nearbyintf F
+GLIBC_2.32 nearbyintf128 F
+GLIBC_2.32 nearbyintf32 F
+GLIBC_2.32 nearbyintf32x F
+GLIBC_2.32 nearbyintf64 F
+GLIBC_2.32 nearbyintf64x F
+GLIBC_2.32 nearbyintl F
+GLIBC_2.32 nextafter F
+GLIBC_2.32 nextafterf F
+GLIBC_2.32 nextafterf128 F
+GLIBC_2.32 nextafterf32 F
+GLIBC_2.32 nextafterf32x F
+GLIBC_2.32 nextafterf64 F
+GLIBC_2.32 nextafterf64x F
+GLIBC_2.32 nextafterl F
+GLIBC_2.32 nextdown F
+GLIBC_2.32 nextdownf F
+GLIBC_2.32 nextdownf128 F
+GLIBC_2.32 nextdownf32 F
+GLIBC_2.32 nextdownf32x F
+GLIBC_2.32 nextdownf64 F
+GLIBC_2.32 nextdownf64x F
+GLIBC_2.32 nextdownl F
+GLIBC_2.32 nexttoward F
+GLIBC_2.32 nexttowardf F
+GLIBC_2.32 nexttowardl F
+GLIBC_2.32 nextup F
+GLIBC_2.32 nextupf F
+GLIBC_2.32 nextupf128 F
+GLIBC_2.32 nextupf32 F
+GLIBC_2.32 nextupf32x F
+GLIBC_2.32 nextupf64 F
+GLIBC_2.32 nextupf64x F
+GLIBC_2.32 nextupl F
+GLIBC_2.32 pow F
+GLIBC_2.32 powf F
+GLIBC_2.32 powf128 F
+GLIBC_2.32 powf32 F
+GLIBC_2.32 powf32x F
+GLIBC_2.32 powf64 F
+GLIBC_2.32 powf64x F
+GLIBC_2.32 powl F
+GLIBC_2.32 remainder F
+GLIBC_2.32 remainderf F
+GLIBC_2.32 remainderf128 F
+GLIBC_2.32 remainderf32 F
+GLIBC_2.32 remainderf32x F
+GLIBC_2.32 remainderf64 F
+GLIBC_2.32 remainderf64x F
+GLIBC_2.32 remainderl F
+GLIBC_2.32 remquo F
+GLIBC_2.32 remquof F
+GLIBC_2.32 remquof128 F
+GLIBC_2.32 remquof32 F
+GLIBC_2.32 remquof32x F
+GLIBC_2.32 remquof64 F
+GLIBC_2.32 remquof64x F
+GLIBC_2.32 remquol F
+GLIBC_2.32 rint F
+GLIBC_2.32 rintf F
+GLIBC_2.32 rintf128 F
+GLIBC_2.32 rintf32 F
+GLIBC_2.32 rintf32x F
+GLIBC_2.32 rintf64 F
+GLIBC_2.32 rintf64x F
+GLIBC_2.32 rintl F
+GLIBC_2.32 round F
+GLIBC_2.32 roundeven F
+GLIBC_2.32 roundevenf F
+GLIBC_2.32 roundevenf128 F
+GLIBC_2.32 roundevenf32 F
+GLIBC_2.32 roundevenf32x F
+GLIBC_2.32 roundevenf64 F
+GLIBC_2.32 roundevenf64x F
+GLIBC_2.32 roundevenl F
+GLIBC_2.32 roundf F
+GLIBC_2.32 roundf128 F
+GLIBC_2.32 roundf32 F
+GLIBC_2.32 roundf32x F
+GLIBC_2.32 roundf64 F
+GLIBC_2.32 roundf64x F
+GLIBC_2.32 roundl F
+GLIBC_2.32 scalb F
+GLIBC_2.32 scalbf F
+GLIBC_2.32 scalbl F
+GLIBC_2.32 scalbln F
+GLIBC_2.32 scalblnf F
+GLIBC_2.32 scalblnf128 F
+GLIBC_2.32 scalblnf32 F
+GLIBC_2.32 scalblnf32x F
+GLIBC_2.32 scalblnf64 F
+GLIBC_2.32 scalblnf64x F
+GLIBC_2.32 scalblnl F
+GLIBC_2.32 scalbn F
+GLIBC_2.32 scalbnf F
+GLIBC_2.32 scalbnf128 F
+GLIBC_2.32 scalbnf32 F
+GLIBC_2.32 scalbnf32x F
+GLIBC_2.32 scalbnf64 F
+GLIBC_2.32 scalbnf64x F
+GLIBC_2.32 scalbnl F
+GLIBC_2.32 setpayload F
+GLIBC_2.32 setpayloadf F
+GLIBC_2.32 setpayloadf128 F
+GLIBC_2.32 setpayloadf32 F
+GLIBC_2.32 setpayloadf32x F
+GLIBC_2.32 setpayloadf64 F
+GLIBC_2.32 setpayloadf64x F
+GLIBC_2.32 setpayloadl F
+GLIBC_2.32 setpayloadsig F
+GLIBC_2.32 setpayloadsigf F
+GLIBC_2.32 setpayloadsigf128 F
+GLIBC_2.32 setpayloadsigf32 F
+GLIBC_2.32 setpayloadsigf32x F
+GLIBC_2.32 setpayloadsigf64 F
+GLIBC_2.32 setpayloadsigf64x F
+GLIBC_2.32 setpayloadsigl F
+GLIBC_2.32 signgam D 0x4
+GLIBC_2.32 significand F
+GLIBC_2.32 significandf F
+GLIBC_2.32 significandl F
+GLIBC_2.32 sin F
+GLIBC_2.32 sincos F
+GLIBC_2.32 sincosf F
+GLIBC_2.32 sincosf128 F
+GLIBC_2.32 sincosf32 F
+GLIBC_2.32 sincosf32x F
+GLIBC_2.32 sincosf64 F
+GLIBC_2.32 sincosf64x F
+GLIBC_2.32 sincosl F
+GLIBC_2.32 sinf F
+GLIBC_2.32 sinf128 F
+GLIBC_2.32 sinf32 F
+GLIBC_2.32 sinf32x F
+GLIBC_2.32 sinf64 F
+GLIBC_2.32 sinf64x F
+GLIBC_2.32 sinh F
+GLIBC_2.32 sinhf F
+GLIBC_2.32 sinhf128 F
+GLIBC_2.32 sinhf32 F
+GLIBC_2.32 sinhf32x F
+GLIBC_2.32 sinhf64 F
+GLIBC_2.32 sinhf64x F
+GLIBC_2.32 sinhl F
+GLIBC_2.32 sinl F
+GLIBC_2.32 sqrt F
+GLIBC_2.32 sqrtf F
+GLIBC_2.32 sqrtf128 F
+GLIBC_2.32 sqrtf32 F
+GLIBC_2.32 sqrtf32x F
+GLIBC_2.32 sqrtf64 F
+GLIBC_2.32 sqrtf64x F
+GLIBC_2.32 sqrtl F
+GLIBC_2.32 tan F
+GLIBC_2.32 tanf F
+GLIBC_2.32 tanf128 F
+GLIBC_2.32 tanf32 F
+GLIBC_2.32 tanf32x F
+GLIBC_2.32 tanf64 F
+GLIBC_2.32 tanf64x F
+GLIBC_2.32 tanh F
+GLIBC_2.32 tanhf F
+GLIBC_2.32 tanhf128 F
+GLIBC_2.32 tanhf32 F
+GLIBC_2.32 tanhf32x F
+GLIBC_2.32 tanhf64 F
+GLIBC_2.32 tanhf64x F
+GLIBC_2.32 tanhl F
+GLIBC_2.32 tanl F
+GLIBC_2.32 tgamma F
+GLIBC_2.32 tgammaf F
+GLIBC_2.32 tgammaf128 F
+GLIBC_2.32 tgammaf32 F
+GLIBC_2.32 tgammaf32x F
+GLIBC_2.32 tgammaf64 F
+GLIBC_2.32 tgammaf64x F
+GLIBC_2.32 tgammal F
+GLIBC_2.32 totalorder F
+GLIBC_2.32 totalorderf F
+GLIBC_2.32 totalorderf128 F
+GLIBC_2.32 totalorderf32 F
+GLIBC_2.32 totalorderf32x F
+GLIBC_2.32 totalorderf64 F
+GLIBC_2.32 totalorderf64x F
+GLIBC_2.32 totalorderl F
+GLIBC_2.32 totalordermag F
+GLIBC_2.32 totalordermagf F
+GLIBC_2.32 totalordermagf128 F
+GLIBC_2.32 totalordermagf32 F
+GLIBC_2.32 totalordermagf32x F
+GLIBC_2.32 totalordermagf64 F
+GLIBC_2.32 totalordermagf64x F
+GLIBC_2.32 totalordermagl F
+GLIBC_2.32 trunc F
+GLIBC_2.32 truncf F
+GLIBC_2.32 truncf128 F
+GLIBC_2.32 truncf32 F
+GLIBC_2.32 truncf32x F
+GLIBC_2.32 truncf64 F
+GLIBC_2.32 truncf64x F
+GLIBC_2.32 truncl F
+GLIBC_2.32 ufromfp F
+GLIBC_2.32 ufromfpf F
+GLIBC_2.32 ufromfpf128 F
+GLIBC_2.32 ufromfpf32 F
+GLIBC_2.32 ufromfpf32x F
+GLIBC_2.32 ufromfpf64 F
+GLIBC_2.32 ufromfpf64x F
+GLIBC_2.32 ufromfpl F
+GLIBC_2.32 ufromfpx F
+GLIBC_2.32 ufromfpxf F
+GLIBC_2.32 ufromfpxf128 F
+GLIBC_2.32 ufromfpxf32 F
+GLIBC_2.32 ufromfpxf32x F
+GLIBC_2.32 ufromfpxf64 F
+GLIBC_2.32 ufromfpxf64x F
+GLIBC_2.32 ufromfpxl F
+GLIBC_2.32 y0 F
+GLIBC_2.32 y0f F
+GLIBC_2.32 y0f128 F
+GLIBC_2.32 y0f32 F
+GLIBC_2.32 y0f32x F
+GLIBC_2.32 y0f64 F
+GLIBC_2.32 y0f64x F
+GLIBC_2.32 y0l F
+GLIBC_2.32 y1 F
+GLIBC_2.32 y1f F
+GLIBC_2.32 y1f128 F
+GLIBC_2.32 y1f32 F
+GLIBC_2.32 y1f32x F
+GLIBC_2.32 y1f64 F
+GLIBC_2.32 y1f64x F
+GLIBC_2.32 y1l F
+GLIBC_2.32 yn F
+GLIBC_2.32 ynf F
+GLIBC_2.32 ynf128 F
+GLIBC_2.32 ynf32 F
+GLIBC_2.32 ynf32x F
+GLIBC_2.32 ynf64 F
+GLIBC_2.32 ynf64x F
+GLIBC_2.32 ynl F
diff --git a/sysdeps/unix/sysv/linux/riscv/libpthread.abilist b/sysdeps/unix/sysv/linux/riscv/libpthread.abilist
new file mode 100644
index 0000000000..1adcbecc2e
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/libpthread.abilist
@@ -0,0 +1,213 @@
+GLIBC_2.32 _IO_flockfile F
+GLIBC_2.32 _IO_ftrylockfile F
+GLIBC_2.32 _IO_funlockfile F
+GLIBC_2.32 __close F
+GLIBC_2.32 __connect F
+GLIBC_2.32 __errno_location F
+GLIBC_2.32 __h_errno_location F
+GLIBC_2.32 __libc_allocate_rtsig F
+GLIBC_2.32 __libc_current_sigrtmax F
+GLIBC_2.32 __libc_current_sigrtmin F
+GLIBC_2.32 __lseek F
+GLIBC_2.32 __open F
+GLIBC_2.32 __open64 F
+GLIBC_2.32 __pread64 F
+GLIBC_2.32 __pthread_cleanup_routine F
+GLIBC_2.32 __pthread_getspecific F
+GLIBC_2.32 __pthread_key_create F
+GLIBC_2.32 __pthread_mutex_destroy F
+GLIBC_2.32 __pthread_mutex_init F
+GLIBC_2.32 __pthread_mutex_lock F
+GLIBC_2.32 __pthread_mutex_trylock F
+GLIBC_2.32 __pthread_mutex_unlock F
+GLIBC_2.32 __pthread_mutexattr_destroy F
+GLIBC_2.32 __pthread_mutexattr_init F
+GLIBC_2.32 __pthread_mutexattr_settype F
+GLIBC_2.32 __pthread_once F
+GLIBC_2.32 __pthread_register_cancel F
+GLIBC_2.32 __pthread_register_cancel_defer F
+GLIBC_2.32 __pthread_rwlock_destroy F
+GLIBC_2.32 __pthread_rwlock_init F
+GLIBC_2.32 __pthread_rwlock_rdlock F
+GLIBC_2.32 __pthread_rwlock_tryrdlock F
+GLIBC_2.32 __pthread_rwlock_trywrlock F
+GLIBC_2.32 __pthread_rwlock_unlock F
+GLIBC_2.32 __pthread_rwlock_wrlock F
+GLIBC_2.32 __pthread_setspecific F
+GLIBC_2.32 __pthread_unregister_cancel F
+GLIBC_2.32 __pthread_unregister_cancel_restore F
+GLIBC_2.32 __pthread_unwind_next F
+GLIBC_2.32 __pwrite64 F
+GLIBC_2.32 __read F
+GLIBC_2.32 __res_state F
+GLIBC_2.32 __send F
+GLIBC_2.32 __sigaction F
+GLIBC_2.32 __write F
+GLIBC_2.32 _pthread_cleanup_pop F
+GLIBC_2.32 _pthread_cleanup_pop_restore F
+GLIBC_2.32 _pthread_cleanup_push F
+GLIBC_2.32 _pthread_cleanup_push_defer F
+GLIBC_2.32 accept F
+GLIBC_2.32 call_once F
+GLIBC_2.32 close F
+GLIBC_2.32 cnd_broadcast F
+GLIBC_2.32 cnd_destroy F
+GLIBC_2.32 cnd_init F
+GLIBC_2.32 cnd_signal F
+GLIBC_2.32 cnd_timedwait F
+GLIBC_2.32 cnd_wait F
+GLIBC_2.32 connect F
+GLIBC_2.32 flockfile F
+GLIBC_2.32 fsync F
+GLIBC_2.32 ftrylockfile F
+GLIBC_2.32 funlockfile F
+GLIBC_2.32 lseek F
+GLIBC_2.32 lseek64 F
+GLIBC_2.32 msync F
+GLIBC_2.32 mtx_destroy F
+GLIBC_2.32 mtx_init F
+GLIBC_2.32 mtx_lock F
+GLIBC_2.32 mtx_timedlock F
+GLIBC_2.32 mtx_trylock F
+GLIBC_2.32 mtx_unlock F
+GLIBC_2.32 open F
+GLIBC_2.32 open64 F
+GLIBC_2.32 pause F
+GLIBC_2.32 pread F
+GLIBC_2.32 pread64 F
+GLIBC_2.32 pthread_attr_getaffinity_np F
+GLIBC_2.32 pthread_attr_getguardsize F
+GLIBC_2.32 pthread_attr_getstack F
+GLIBC_2.32 pthread_attr_getstackaddr F
+GLIBC_2.32 pthread_attr_getstacksize F
+GLIBC_2.32 pthread_attr_setguardsize F
+GLIBC_2.32 pthread_attr_setstack F
+GLIBC_2.32 pthread_attr_setstackaddr F
+GLIBC_2.32 pthread_attr_setstacksize F
+GLIBC_2.32 pthread_barrier_destroy F
+GLIBC_2.32 pthread_barrier_init F
+GLIBC_2.32 pthread_barrier_wait F
+GLIBC_2.32 pthread_barrierattr_destroy F
+GLIBC_2.32 pthread_barrierattr_getpshared F
+GLIBC_2.32 pthread_barrierattr_init F
+GLIBC_2.32 pthread_barrierattr_setpshared F
+GLIBC_2.32 pthread_cancel F
+GLIBC_2.32 pthread_clockjoin_np F
+GLIBC_2.32 pthread_cond_broadcast F
+GLIBC_2.32 pthread_cond_clockwait F
+GLIBC_2.32 pthread_cond_signal F
+GLIBC_2.32 pthread_cond_timedwait F
+GLIBC_2.32 pthread_cond_wait F
+GLIBC_2.32 pthread_condattr_getclock F
+GLIBC_2.32 pthread_condattr_getpshared F
+GLIBC_2.32 pthread_condattr_setclock F
+GLIBC_2.32 pthread_condattr_setpshared F
+GLIBC_2.32 pthread_create F
+GLIBC_2.32 pthread_detach F
+GLIBC_2.32 pthread_exit F
+GLIBC_2.32 pthread_getattr_default_np F
+GLIBC_2.32 pthread_getconcurrency F
+GLIBC_2.32 pthread_getcpuclockid F
+GLIBC_2.32 pthread_getname_np F
+GLIBC_2.32 pthread_getspecific F
+GLIBC_2.32 pthread_join F
+GLIBC_2.32 pthread_key_create F
+GLIBC_2.32 pthread_key_delete F
+GLIBC_2.32 pthread_kill F
+GLIBC_2.32 pthread_kill_other_threads_np F
+GLIBC_2.32 pthread_mutex_clocklock F
+GLIBC_2.32 pthread_mutex_consistent F
+GLIBC_2.32 pthread_mutex_consistent_np F
+GLIBC_2.32 pthread_mutex_destroy F
+GLIBC_2.32 pthread_mutex_getprioceiling F
+GLIBC_2.32 pthread_mutex_init F
+GLIBC_2.32 pthread_mutex_lock F
+GLIBC_2.32 pthread_mutex_setprioceiling F
+GLIBC_2.32 pthread_mutex_timedlock F
+GLIBC_2.32 pthread_mutex_trylock F
+GLIBC_2.32 pthread_mutex_unlock F
+GLIBC_2.32 pthread_mutexattr_destroy F
+GLIBC_2.32 pthread_mutexattr_getkind_np F
+GLIBC_2.32 pthread_mutexattr_getprioceiling F
+GLIBC_2.32 pthread_mutexattr_getprotocol F
+GLIBC_2.32 pthread_mutexattr_getpshared F
+GLIBC_2.32 pthread_mutexattr_getrobust F
+GLIBC_2.32 pthread_mutexattr_getrobust_np F
+GLIBC_2.32 pthread_mutexattr_gettype F
+GLIBC_2.32 pthread_mutexattr_init F
+GLIBC_2.32 pthread_mutexattr_setkind_np F
+GLIBC_2.32 pthread_mutexattr_setprioceiling F
+GLIBC_2.32 pthread_mutexattr_setprotocol F
+GLIBC_2.32 pthread_mutexattr_setpshared F
+GLIBC_2.32 pthread_mutexattr_setrobust F
+GLIBC_2.32 pthread_mutexattr_setrobust_np F
+GLIBC_2.32 pthread_mutexattr_settype F
+GLIBC_2.32 pthread_once F
+GLIBC_2.32 pthread_rwlock_clockrdlock F
+GLIBC_2.32 pthread_rwlock_clockwrlock F
+GLIBC_2.32 pthread_rwlock_destroy F
+GLIBC_2.32 pthread_rwlock_init F
+GLIBC_2.32 pthread_rwlock_rdlock F
+GLIBC_2.32 pthread_rwlock_timedrdlock F
+GLIBC_2.32 pthread_rwlock_timedwrlock F
+GLIBC_2.32 pthread_rwlock_tryrdlock F
+GLIBC_2.32 pthread_rwlock_trywrlock F
+GLIBC_2.32 pthread_rwlock_unlock F
+GLIBC_2.32 pthread_rwlock_wrlock F
+GLIBC_2.32 pthread_rwlockattr_destroy F
+GLIBC_2.32 pthread_rwlockattr_getkind_np F
+GLIBC_2.32 pthread_rwlockattr_getpshared F
+GLIBC_2.32 pthread_rwlockattr_init F
+GLIBC_2.32 pthread_rwlockattr_setkind_np F
+GLIBC_2.32 pthread_rwlockattr_setpshared F
+GLIBC_2.32 pthread_setaffinity_np F
+GLIBC_2.32 pthread_setattr_default_np F
+GLIBC_2.32 pthread_setcancelstate F
+GLIBC_2.32 pthread_setcanceltype F
+GLIBC_2.32 pthread_setconcurrency F
+GLIBC_2.32 pthread_setname_np F
+GLIBC_2.32 pthread_setschedprio F
+GLIBC_2.32 pthread_setspecific F
+GLIBC_2.32 pthread_sigqueue F
+GLIBC_2.32 pthread_spin_destroy F
+GLIBC_2.32 pthread_spin_init F
+GLIBC_2.32 pthread_spin_lock F
+GLIBC_2.32 pthread_spin_trylock F
+GLIBC_2.32 pthread_spin_unlock F
+GLIBC_2.32 pthread_testcancel F
+GLIBC_2.32 pthread_timedjoin_np F
+GLIBC_2.32 pthread_tryjoin_np F
+GLIBC_2.32 pthread_yield F
+GLIBC_2.32 pwrite F
+GLIBC_2.32 pwrite64 F
+GLIBC_2.32 raise F
+GLIBC_2.32 read F
+GLIBC_2.32 recv F
+GLIBC_2.32 recvfrom F
+GLIBC_2.32 recvmsg F
+GLIBC_2.32 sem_clockwait F
+GLIBC_2.32 sem_close F
+GLIBC_2.32 sem_destroy F
+GLIBC_2.32 sem_getvalue F
+GLIBC_2.32 sem_init F
+GLIBC_2.32 sem_open F
+GLIBC_2.32 sem_post F
+GLIBC_2.32 sem_timedwait F
+GLIBC_2.32 sem_trywait F
+GLIBC_2.32 sem_unlink F
+GLIBC_2.32 sem_wait F
+GLIBC_2.32 send F
+GLIBC_2.32 sendmsg F
+GLIBC_2.32 sendto F
+GLIBC_2.32 sigaction F
+GLIBC_2.32 sigwait F
+GLIBC_2.32 tcdrain F
+GLIBC_2.32 thrd_create F
+GLIBC_2.32 thrd_detach F
+GLIBC_2.32 thrd_exit F
+GLIBC_2.32 thrd_join F
+GLIBC_2.32 tss_create F
+GLIBC_2.32 tss_delete F
+GLIBC_2.32 tss_get F
+GLIBC_2.32 tss_set F
+GLIBC_2.32 write F
diff --git a/sysdeps/unix/sysv/linux/riscv/libresolv.abilist b/sysdeps/unix/sysv/linux/riscv/libresolv.abilist
new file mode 100644
index 0000000000..c5edf99ea9
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/libresolv.abilist
@@ -0,0 +1,79 @@
+GLIBC_2.32 __b64_ntop F
+GLIBC_2.32 __b64_pton F
+GLIBC_2.32 __dn_comp F
+GLIBC_2.32 __dn_count_labels F
+GLIBC_2.32 __dn_expand F
+GLIBC_2.32 __dn_skipname F
+GLIBC_2.32 __fp_nquery F
+GLIBC_2.32 __fp_query F
+GLIBC_2.32 __fp_resstat F
+GLIBC_2.32 __hostalias F
+GLIBC_2.32 __loc_aton F
+GLIBC_2.32 __loc_ntoa F
+GLIBC_2.32 __p_cdname F
+GLIBC_2.32 __p_cdnname F
+GLIBC_2.32 __p_class F
+GLIBC_2.32 __p_class_syms D 0x54
+GLIBC_2.32 __p_fqname F
+GLIBC_2.32 __p_fqnname F
+GLIBC_2.32 __p_option F
+GLIBC_2.32 __p_query F
+GLIBC_2.32 __p_rcode F
+GLIBC_2.32 __p_time F
+GLIBC_2.32 __p_type F
+GLIBC_2.32 __p_type_syms D 0x228
+GLIBC_2.32 __putlong F
+GLIBC_2.32 __putshort F
+GLIBC_2.32 __res_close F
+GLIBC_2.32 __res_dnok F
+GLIBC_2.32 __res_hnok F
+GLIBC_2.32 __res_hostalias F
+GLIBC_2.32 __res_isourserver F
+GLIBC_2.32 __res_mailok F
+GLIBC_2.32 __res_mkquery F
+GLIBC_2.32 __res_nameinquery F
+GLIBC_2.32 __res_nmkquery F
+GLIBC_2.32 __res_nquery F
+GLIBC_2.32 __res_nquerydomain F
+GLIBC_2.32 __res_nsearch F
+GLIBC_2.32 __res_nsend F
+GLIBC_2.32 __res_ownok F
+GLIBC_2.32 __res_queriesmatch F
+GLIBC_2.32 __res_query F
+GLIBC_2.32 __res_querydomain F
+GLIBC_2.32 __res_search F
+GLIBC_2.32 __res_send F
+GLIBC_2.32 __sym_ntop F
+GLIBC_2.32 __sym_ntos F
+GLIBC_2.32 __sym_ston F
+GLIBC_2.32 _getlong F
+GLIBC_2.32 _getshort F
+GLIBC_2.32 inet_net_ntop F
+GLIBC_2.32 inet_net_pton F
+GLIBC_2.32 inet_neta F
+GLIBC_2.32 ns_datetosecs F
+GLIBC_2.32 ns_format_ttl F
+GLIBC_2.32 ns_get16 F
+GLIBC_2.32 ns_get32 F
+GLIBC_2.32 ns_initparse F
+GLIBC_2.32 ns_makecanon F
+GLIBC_2.32 ns_msg_getflag F
+GLIBC_2.32 ns_name_compress F
+GLIBC_2.32 ns_name_ntol F
+GLIBC_2.32 ns_name_ntop F
+GLIBC_2.32 ns_name_pack F
+GLIBC_2.32 ns_name_pton F
+GLIBC_2.32 ns_name_rollback F
+GLIBC_2.32 ns_name_skip F
+GLIBC_2.32 ns_name_uncompress F
+GLIBC_2.32 ns_name_unpack F
+GLIBC_2.32 ns_parse_ttl F
+GLIBC_2.32 ns_parserr F
+GLIBC_2.32 ns_put16 F
+GLIBC_2.32 ns_put32 F
+GLIBC_2.32 ns_samedomain F
+GLIBC_2.32 ns_samename F
+GLIBC_2.32 ns_skiprr F
+GLIBC_2.32 ns_sprintrr F
+GLIBC_2.32 ns_sprintrrf F
+GLIBC_2.32 ns_subdomain F
diff --git a/sysdeps/unix/sysv/linux/riscv/librt.abilist b/sysdeps/unix/sysv/linux/riscv/librt.abilist
new file mode 100644
index 0000000000..fda2b20c01
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/librt.abilist
@@ -0,0 +1,35 @@
+GLIBC_2.32 __mq_open_2 F
+GLIBC_2.32 aio_cancel F
+GLIBC_2.32 aio_cancel64 F
+GLIBC_2.32 aio_error F
+GLIBC_2.32 aio_error64 F
+GLIBC_2.32 aio_fsync F
+GLIBC_2.32 aio_fsync64 F
+GLIBC_2.32 aio_init F
+GLIBC_2.32 aio_read F
+GLIBC_2.32 aio_read64 F
+GLIBC_2.32 aio_return F
+GLIBC_2.32 aio_return64 F
+GLIBC_2.32 aio_suspend F
+GLIBC_2.32 aio_suspend64 F
+GLIBC_2.32 aio_write F
+GLIBC_2.32 aio_write64 F
+GLIBC_2.32 lio_listio F
+GLIBC_2.32 lio_listio64 F
+GLIBC_2.32 mq_close F
+GLIBC_2.32 mq_getattr F
+GLIBC_2.32 mq_notify F
+GLIBC_2.32 mq_open F
+GLIBC_2.32 mq_receive F
+GLIBC_2.32 mq_send F
+GLIBC_2.32 mq_setattr F
+GLIBC_2.32 mq_timedreceive F
+GLIBC_2.32 mq_timedsend F
+GLIBC_2.32 mq_unlink F
+GLIBC_2.32 shm_open F
+GLIBC_2.32 shm_unlink F
+GLIBC_2.32 timer_create F
+GLIBC_2.32 timer_delete F
+GLIBC_2.32 timer_getoverrun F
+GLIBC_2.32 timer_gettime F
+GLIBC_2.32 timer_settime F
diff --git a/sysdeps/unix/sysv/linux/riscv/libthread_db.abilist b/sysdeps/unix/sysv/linux/riscv/libthread_db.abilist
new file mode 100644
index 0000000000..dcbc4a8fbe
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/libthread_db.abilist
@@ -0,0 +1,40 @@
+GLIBC_2.32 td_init F
+GLIBC_2.32 td_log F
+GLIBC_2.32 td_symbol_list F
+GLIBC_2.32 td_ta_clear_event F
+GLIBC_2.32 td_ta_delete F
+GLIBC_2.32 td_ta_enable_stats F
+GLIBC_2.32 td_ta_event_addr F
+GLIBC_2.32 td_ta_event_getmsg F
+GLIBC_2.32 td_ta_get_nthreads F
+GLIBC_2.32 td_ta_get_ph F
+GLIBC_2.32 td_ta_get_stats F
+GLIBC_2.32 td_ta_map_id2thr F
+GLIBC_2.32 td_ta_map_lwp2thr F
+GLIBC_2.32 td_ta_new F
+GLIBC_2.32 td_ta_reset_stats F
+GLIBC_2.32 td_ta_set_event F
+GLIBC_2.32 td_ta_setconcurrency F
+GLIBC_2.32 td_ta_thr_iter F
+GLIBC_2.32 td_ta_tsd_iter F
+GLIBC_2.32 td_thr_clear_event F
+GLIBC_2.32 td_thr_dbresume F
+GLIBC_2.32 td_thr_dbsuspend F
+GLIBC_2.32 td_thr_event_enable F
+GLIBC_2.32 td_thr_event_getmsg F
+GLIBC_2.32 td_thr_get_info F
+GLIBC_2.32 td_thr_getfpregs F
+GLIBC_2.32 td_thr_getgregs F
+GLIBC_2.32 td_thr_getxregs F
+GLIBC_2.32 td_thr_getxregsize F
+GLIBC_2.32 td_thr_set_event F
+GLIBC_2.32 td_thr_setfpregs F
+GLIBC_2.32 td_thr_setgregs F
+GLIBC_2.32 td_thr_setprio F
+GLIBC_2.32 td_thr_setsigpending F
+GLIBC_2.32 td_thr_setxregs F
+GLIBC_2.32 td_thr_sigsetmask F
+GLIBC_2.32 td_thr_tls_get_addr F
+GLIBC_2.32 td_thr_tlsbase F
+GLIBC_2.32 td_thr_tsd F
+GLIBC_2.32 td_thr_validate F
diff --git a/sysdeps/unix/sysv/linux/riscv/libutil.abilist b/sysdeps/unix/sysv/linux/riscv/libutil.abilist
new file mode 100644
index 0000000000..61f73bc34e
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/libutil.abilist
@@ -0,0 +1,6 @@
+GLIBC_2.32 forkpty F
+GLIBC_2.32 login F
+GLIBC_2.32 login_tty F
+GLIBC_2.32 logout F
+GLIBC_2.32 logwtmp F
+GLIBC_2.32 openpty F
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 12/18] RISC-V: Add the RV32 libm-test-ulps
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (10 preceding siblings ...)
2020-06-03 16:25 ` [PATCH v2 11/18] RISC-V: Add ABI lists Alistair Francis via Libc-alpha
@ 2020-06-03 16:25 ` Alistair Francis via Libc-alpha
2020-06-03 17:34 ` Joseph Myers
2020-06-03 16:26 ` [PATCH v2 13/18] RISC-V: Fix llrint and llround missing exceptions on RV32 Alistair Francis via Libc-alpha
` (6 subsequent siblings)
18 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:25 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
Add a libm-test-ulps for RV32, generated by running `make regen-ulps`.
This is different to the RV64 one.
---
sysdeps/riscv/rv32/rvd/libm-test-ulps | 1401 ++++++++++++++++++++
sysdeps/riscv/rv32/rvd/libm-test-ulps-name | 1 +
2 files changed, 1402 insertions(+)
create mode 100644 sysdeps/riscv/rv32/rvd/libm-test-ulps
create mode 100644 sysdeps/riscv/rv32/rvd/libm-test-ulps-name
diff --git a/sysdeps/riscv/rv32/rvd/libm-test-ulps b/sysdeps/riscv/rv32/rvd/libm-test-ulps
new file mode 100644
index 0000000000..35d77ff1ac
--- /dev/null
+++ b/sysdeps/riscv/rv32/rvd/libm-test-ulps
@@ -0,0 +1,1401 @@
+# Begin of automatic generation
+
+# Maximal error of functions:
+Function: "acos":
+float: 1
+ldouble: 1
+
+Function: "acos_downward":
+double: 1
+float: 1
+ldouble: 1
+
+Function: "acos_towardzero":
+double: 1
+float: 1
+ldouble: 1
+
+Function: "acos_upward":
+double: 1
+float: 1
+ldouble: 1
+
+Function: "acosh":
+double: 2
+float: 2
+ldouble: 2
+
+Function: "acosh_downward":
+double: 2
+float: 2
+ldouble: 3
+
+Function: "acosh_towardzero":
+double: 2
+float: 2
+ldouble: 2
+
+Function: "acosh_upward":
+double: 2
+float: 2
+ldouble: 4
+
+Function: "asin":
+float: 1
+ldouble: 1
+
+Function: "asin_downward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "asin_towardzero":
+double: 1
+float: 1
+ldouble: 1
+
+Function: "asin_upward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "asinh":
+double: 1
+float: 2
+ldouble: 3
+
+Function: "asinh_downward":
+double: 3
+float: 3
+ldouble: 4
+
+Function: "asinh_towardzero":
+double: 2
+float: 2
+ldouble: 3
+
+Function: "asinh_upward":
+double: 3
+float: 3
+ldouble: 4
+
+Function: "atan":
+float: 1
+ldouble: 1
+
+Function: "atan2":
+float: 1
+ldouble: 1
+
+Function: "atan2_downward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: "atan2_towardzero":
+double: 1
+float: 2
+ldouble: 3
+
+Function: "atan2_upward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "atan_downward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: "atan_towardzero":
+double: 1
+float: 1
+ldouble: 1
+
+Function: "atan_upward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: "atanh":
+double: 2
+float: 2
+ldouble: 3
+
+Function: "atanh_downward":
+double: 3
+float: 3
+ldouble: 4
+
+Function: "atanh_towardzero":
+double: 2
+float: 2
+ldouble: 2
+
+Function: "atanh_upward":
+double: 3
+float: 3
+ldouble: 4
+
+Function: "cabs":
+double: 1
+ldouble: 1
+
+Function: "cabs_downward":
+double: 1
+ldouble: 1
+
+Function: "cabs_towardzero":
+double: 1
+ldouble: 1
+
+Function: "cabs_upward":
+double: 1
+ldouble: 1
+
+Function: Real part of "cacos":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "cacos":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Real part of "cacos_downward":
+double: 3
+float: 2
+ldouble: 3
+
+Function: Imaginary part of "cacos_downward":
+double: 5
+float: 3
+ldouble: 6
+
+Function: Real part of "cacos_towardzero":
+double: 3
+float: 2
+ldouble: 3
+
+Function: Imaginary part of "cacos_towardzero":
+double: 4
+float: 2
+ldouble: 5
+
+Function: Real part of "cacos_upward":
+double: 2
+float: 2
+ldouble: 4
+
+Function: Imaginary part of "cacos_upward":
+double: 5
+float: 5
+ldouble: 7
+
+Function: Real part of "cacosh":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "cacosh":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Real part of "cacosh_downward":
+double: 4
+float: 2
+ldouble: 5
+
+Function: Imaginary part of "cacosh_downward":
+double: 3
+float: 3
+ldouble: 4
+
+Function: Real part of "cacosh_towardzero":
+double: 4
+float: 2
+ldouble: 5
+
+Function: Imaginary part of "cacosh_towardzero":
+double: 3
+float: 2
+ldouble: 3
+
+Function: Real part of "cacosh_upward":
+double: 4
+float: 3
+ldouble: 6
+
+Function: Imaginary part of "cacosh_upward":
+double: 3
+float: 3
+ldouble: 4
+
+Function: "carg":
+float: 1
+ldouble: 2
+
+Function: "carg_downward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: "carg_towardzero":
+double: 1
+float: 2
+ldouble: 3
+
+Function: "carg_upward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: Real part of "casin":
+double: 1
+float: 1
+ldouble: 2
+
+Function: Imaginary part of "casin":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Real part of "casin_downward":
+double: 3
+float: 2
+ldouble: 3
+
+Function: Imaginary part of "casin_downward":
+double: 5
+float: 3
+ldouble: 6
+
+Function: Real part of "casin_towardzero":
+double: 3
+float: 1
+ldouble: 3
+
+Function: Imaginary part of "casin_towardzero":
+double: 4
+float: 2
+ldouble: 5
+
+Function: Real part of "casin_upward":
+double: 3
+float: 2
+ldouble: 3
+
+Function: Imaginary part of "casin_upward":
+double: 5
+float: 5
+ldouble: 7
+
+Function: Real part of "casinh":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "casinh":
+double: 1
+float: 1
+ldouble: 2
+
+Function: Real part of "casinh_downward":
+double: 5
+float: 3
+ldouble: 6
+
+Function: Imaginary part of "casinh_downward":
+double: 3
+float: 2
+ldouble: 3
+
+Function: Real part of "casinh_towardzero":
+double: 4
+float: 2
+ldouble: 5
+
+Function: Imaginary part of "casinh_towardzero":
+double: 3
+float: 1
+ldouble: 3
+
+Function: Real part of "casinh_upward":
+double: 5
+float: 5
+ldouble: 7
+
+Function: Imaginary part of "casinh_upward":
+double: 3
+float: 2
+ldouble: 3
+
+Function: Real part of "catan":
+double: 1
+float: 1
+ldouble: 1
+
+Function: Imaginary part of "catan":
+double: 1
+float: 1
+ldouble: 1
+
+Function: Real part of "catan_downward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "catan_downward":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Real part of "catan_towardzero":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "catan_towardzero":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Real part of "catan_upward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: Imaginary part of "catan_upward":
+double: 2
+float: 2
+ldouble: 3
+
+Function: Real part of "catanh":
+double: 1
+float: 1
+ldouble: 1
+
+Function: Imaginary part of "catanh":
+double: 1
+float: 1
+ldouble: 1
+
+Function: Real part of "catanh_downward":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "catanh_downward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Real part of "catanh_towardzero":
+double: 3
+float: 3
+ldouble: 3
+
+Function: Imaginary part of "catanh_towardzero":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Real part of "catanh_upward":
+double: 4
+float: 4
+ldouble: 4
+
+Function: Imaginary part of "catanh_upward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: "cbrt":
+double: 3
+float: 1
+ldouble: 1
+
+Function: "cbrt_downward":
+double: 4
+float: 1
+ldouble: 1
+
+Function: "cbrt_towardzero":
+double: 3
+float: 1
+ldouble: 1
+
+Function: "cbrt_upward":
+double: 5
+float: 1
+ldouble: 1
+
+Function: Real part of "ccos":
+double: 1
+float: 1
+ldouble: 1
+
+Function: Imaginary part of "ccos":
+double: 1
+float: 1
+ldouble: 1
+
+Function: Real part of "ccos_downward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: Imaginary part of "ccos_downward":
+double: 3
+float: 3
+ldouble: 2
+
+Function: Real part of "ccos_towardzero":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "ccos_towardzero":
+double: 3
+float: 3
+ldouble: 2
+
+Function: Real part of "ccos_upward":
+double: 1
+float: 2
+ldouble: 3
+
+Function: Imaginary part of "ccos_upward":
+double: 2
+float: 2
+ldouble: 3
+
+Function: Real part of "ccosh":
+double: 1
+float: 1
+ldouble: 1
+
+Function: Imaginary part of "ccosh":
+double: 1
+float: 1
+ldouble: 1
+
+Function: Real part of "ccosh_downward":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "ccosh_downward":
+double: 3
+float: 3
+ldouble: 2
+
+Function: Real part of "ccosh_towardzero":
+double: 2
+float: 3
+ldouble: 2
+
+Function: Imaginary part of "ccosh_towardzero":
+double: 3
+float: 3
+ldouble: 2
+
+Function: Real part of "ccosh_upward":
+double: 3
+float: 2
+ldouble: 3
+
+Function: Imaginary part of "ccosh_upward":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Real part of "cexp":
+double: 2
+float: 1
+ldouble: 1
+
+Function: Imaginary part of "cexp":
+double: 1
+float: 2
+ldouble: 1
+
+Function: Real part of "cexp_downward":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "cexp_downward":
+double: 3
+float: 3
+ldouble: 2
+
+Function: Real part of "cexp_towardzero":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "cexp_towardzero":
+double: 3
+float: 3
+ldouble: 2
+
+Function: Real part of "cexp_upward":
+double: 2
+float: 2
+ldouble: 3
+
+Function: Imaginary part of "cexp_upward":
+double: 3
+float: 2
+ldouble: 3
+
+Function: Real part of "clog":
+double: 3
+float: 3
+ldouble: 2
+
+Function: Imaginary part of "clog":
+float: 1
+ldouble: 1
+
+Function: Real part of "clog10":
+double: 3
+float: 4
+ldouble: 2
+
+Function: Imaginary part of "clog10":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Real part of "clog10_downward":
+double: 5
+float: 5
+ldouble: 3
+
+Function: Imaginary part of "clog10_downward":
+double: 2
+float: 4
+ldouble: 3
+
+Function: Real part of "clog10_towardzero":
+double: 5
+float: 5
+ldouble: 4
+
+Function: Imaginary part of "clog10_towardzero":
+double: 2
+float: 4
+ldouble: 3
+
+Function: Real part of "clog10_upward":
+double: 6
+float: 5
+ldouble: 4
+
+Function: Imaginary part of "clog10_upward":
+double: 2
+float: 4
+ldouble: 4
+
+Function: Real part of "clog_downward":
+double: 4
+float: 3
+ldouble: 3
+
+Function: Imaginary part of "clog_downward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Real part of "clog_towardzero":
+double: 4
+float: 4
+ldouble: 3
+
+Function: Imaginary part of "clog_towardzero":
+double: 1
+float: 3
+ldouble: 2
+
+Function: Real part of "clog_upward":
+double: 4
+float: 3
+ldouble: 4
+
+Function: Imaginary part of "clog_upward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: "cos":
+double: 1
+float: 1
+ldouble: 1
+
+Function: "cos_downward":
+double: 1
+float: 1
+ldouble: 3
+
+Function: "cos_towardzero":
+double: 1
+float: 1
+ldouble: 1
+
+Function: "cos_upward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "cosh":
+double: 1
+float: 2
+ldouble: 1
+
+Function: "cosh_downward":
+double: 2
+float: 1
+ldouble: 2
+
+Function: "cosh_towardzero":
+double: 2
+float: 1
+ldouble: 2
+
+Function: "cosh_upward":
+double: 2
+float: 2
+ldouble: 3
+
+Function: Real part of "cpow":
+double: 2
+float: 5
+ldouble: 4
+
+Function: Imaginary part of "cpow":
+float: 2
+ldouble: 1
+
+Function: Real part of "cpow_downward":
+double: 5
+float: 8
+ldouble: 6
+
+Function: Imaginary part of "cpow_downward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Real part of "cpow_towardzero":
+double: 5
+float: 8
+ldouble: 6
+
+Function: Imaginary part of "cpow_towardzero":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Real part of "cpow_upward":
+double: 4
+float: 1
+ldouble: 3
+
+Function: Imaginary part of "cpow_upward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Real part of "csin":
+double: 1
+float: 1
+ldouble: 1
+
+Function: Imaginary part of "csin":
+ldouble: 1
+
+Function: Real part of "csin_downward":
+double: 3
+float: 3
+ldouble: 2
+
+Function: Imaginary part of "csin_downward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: Real part of "csin_towardzero":
+double: 3
+float: 3
+ldouble: 2
+
+Function: Imaginary part of "csin_towardzero":
+double: 1
+float: 1
+ldouble: 2
+
+Function: Real part of "csin_upward":
+double: 4
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "csin_upward":
+double: 2
+float: 2
+ldouble: 3
+
+Function: Real part of "csinh":
+float: 1
+ldouble: 1
+
+Function: Imaginary part of "csinh":
+double: 1
+float: 1
+ldouble: 1
+
+Function: Real part of "csinh_downward":
+double: 2
+float: 1
+ldouble: 2
+
+Function: Imaginary part of "csinh_downward":
+double: 3
+float: 3
+ldouble: 2
+
+Function: Real part of "csinh_towardzero":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "csinh_towardzero":
+double: 3
+float: 3
+ldouble: 2
+
+Function: Real part of "csinh_upward":
+double: 1
+float: 2
+ldouble: 3
+
+Function: Imaginary part of "csinh_upward":
+double: 1
+float: 2
+ldouble: 2
+
+Function: Real part of "csqrt":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Imaginary part of "csqrt":
+double: 2
+float: 2
+ldouble: 2
+
+Function: Real part of "csqrt_downward":
+double: 5
+float: 4
+ldouble: 4
+
+Function: Imaginary part of "csqrt_downward":
+double: 4
+float: 3
+ldouble: 3
+
+Function: Real part of "csqrt_towardzero":
+double: 4
+float: 3
+ldouble: 3
+
+Function: Imaginary part of "csqrt_towardzero":
+double: 4
+float: 3
+ldouble: 3
+
+Function: Real part of "csqrt_upward":
+double: 5
+float: 4
+ldouble: 4
+
+Function: Imaginary part of "csqrt_upward":
+double: 3
+float: 3
+ldouble: 3
+
+Function: Real part of "ctan":
+double: 1
+float: 1
+ldouble: 3
+
+Function: Imaginary part of "ctan":
+double: 2
+float: 2
+ldouble: 3
+
+Function: Real part of "ctan_downward":
+double: 6
+float: 5
+ldouble: 4
+
+Function: Imaginary part of "ctan_downward":
+double: 2
+float: 2
+ldouble: 5
+
+Function: Real part of "ctan_towardzero":
+double: 5
+float: 2
+ldouble: 4
+
+Function: Imaginary part of "ctan_towardzero":
+double: 2
+float: 2
+ldouble: 5
+
+Function: Real part of "ctan_upward":
+double: 2
+float: 4
+ldouble: 5
+
+Function: Imaginary part of "ctan_upward":
+double: 1
+float: 2
+ldouble: 5
+
+Function: Real part of "ctanh":
+double: 2
+float: 2
+ldouble: 3
+
+Function: Imaginary part of "ctanh":
+double: 2
+float: 1
+ldouble: 3
+
+Function: Real part of "ctanh_downward":
+double: 2
+float: 2
+ldouble: 5
+
+Function: Imaginary part of "ctanh_downward":
+double: 6
+float: 5
+ldouble: 4
+
+Function: Real part of "ctanh_towardzero":
+double: 2
+float: 2
+ldouble: 5
+
+Function: Imaginary part of "ctanh_towardzero":
+double: 5
+float: 2
+ldouble: 3
+
+Function: Real part of "ctanh_upward":
+double: 1
+float: 2
+ldouble: 5
+
+Function: Imaginary part of "ctanh_upward":
+double: 3
+float: 6
+ldouble: 5
+
+Function: "erf":
+double: 1
+float: 1
+ldouble: 1
+
+Function: "erf_downward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "erf_towardzero":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "erf_upward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "erfc":
+double: 2
+float: 2
+ldouble: 2
+
+Function: "erfc_downward":
+double: 4
+float: 4
+ldouble: 5
+
+Function: "erfc_towardzero":
+double: 3
+float: 3
+ldouble: 5
+
+Function: "erfc_upward":
+double: 4
+float: 4
+ldouble: 6
+
+Function: "exp":
+float: 1
+ldouble: 1
+
+Function: "exp10":
+double: 2
+float: 1
+ldouble: 2
+
+Function: "exp10_downward":
+double: 3
+ldouble: 3
+
+Function: "exp10_towardzero":
+double: 3
+ldouble: 3
+
+Function: "exp10_upward":
+double: 2
+float: 1
+ldouble: 4
+
+Function: "exp2":
+double: 1
+ldouble: 1
+
+Function: "exp2_downward":
+double: 1
+ldouble: 1
+
+Function: "exp2_towardzero":
+double: 1
+ldouble: 1
+
+Function: "exp2_upward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "exp_downward":
+double: 1
+float: 1
+
+Function: "exp_towardzero":
+double: 1
+float: 1
+
+Function: "exp_upward":
+double: 1
+float: 1
+
+Function: "expm1":
+double: 1
+float: 1
+ldouble: 1
+
+Function: "expm1_downward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "expm1_towardzero":
+double: 1
+float: 2
+ldouble: 4
+
+Function: "expm1_upward":
+double: 2
+float: 1
+ldouble: 3
+
+Function: "gamma":
+double: 3
+float: 3
+ldouble: 5
+
+Function: "gamma_downward":
+double: 5
+float: 4
+ldouble: 8
+
+Function: "gamma_towardzero":
+double: 4
+float: 3
+ldouble: 5
+
+Function: "gamma_upward":
+double: 4
+float: 5
+ldouble: 8
+
+Function: "hypot":
+double: 1
+ldouble: 1
+
+Function: "hypot_downward":
+double: 1
+ldouble: 1
+
+Function: "hypot_towardzero":
+double: 1
+ldouble: 1
+
+Function: "hypot_upward":
+double: 1
+ldouble: 2
+
+Function: "j0":
+double: 2
+float: 8
+ldouble: 2
+
+Function: "j0_downward":
+double: 2
+float: 4
+ldouble: 4
+
+Function: "j0_towardzero":
+double: 5
+float: 6
+ldouble: 2
+
+Function: "j0_upward":
+double: 4
+float: 5
+ldouble: 5
+
+Function: "j1":
+double: 2
+float: 8
+ldouble: 4
+
+Function: "j1_downward":
+double: 3
+float: 5
+ldouble: 4
+
+Function: "j1_towardzero":
+double: 3
+float: 2
+ldouble: 4
+
+Function: "j1_upward":
+double: 3
+float: 4
+ldouble: 5
+
+Function: "jn":
+double: 4
+float: 4
+ldouble: 7
+
+Function: "jn_downward":
+double: 4
+float: 5
+ldouble: 8
+
+Function: "jn_towardzero":
+double: 4
+float: 5
+ldouble: 8
+
+Function: "jn_upward":
+double: 5
+float: 4
+ldouble: 7
+
+Function: "lgamma":
+double: 3
+float: 3
+ldouble: 5
+
+Function: "lgamma_downward":
+double: 4
+float: 4
+ldouble: 8
+
+Function: "lgamma_towardzero":
+double: 4
+float: 4
+ldouble: 5
+
+Function: "lgamma_upward":
+double: 4
+float: 5
+ldouble: 8
+
+Function: "log":
+ldouble: 1
+
+Function: "log10":
+double: 2
+float: 2
+ldouble: 1
+
+Function: "log10_downward":
+double: 2
+float: 3
+ldouble: 1
+
+Function: "log10_towardzero":
+double: 2
+float: 1
+ldouble: 1
+
+Function: "log10_upward":
+double: 2
+float: 2
+ldouble: 1
+
+Function: "log1p":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "log1p_downward":
+double: 1
+float: 2
+ldouble: 3
+
+Function: "log1p_towardzero":
+double: 2
+float: 2
+ldouble: 3
+
+Function: "log1p_upward":
+double: 2
+float: 2
+ldouble: 4
+
+Function: "log2":
+float: 1
+ldouble: 2
+
+Function: "log2_downward":
+ldouble: 3
+
+Function: "log2_towardzero":
+double: 1
+ldouble: 2
+
+Function: "log2_upward":
+double: 1
+ldouble: 3
+
+Function: "log_downward":
+ldouble: 1
+
+Function: "log_towardzero":
+ldouble: 2
+
+Function: "log_upward":
+ldouble: 1
+
+Function: "pow":
+double: 1
+ldouble: 2
+
+Function: "pow_downward":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "pow_towardzero":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "pow_upward":
+double: 2
+float: 1
+ldouble: 2
+
+Function: "sin":
+double: 1
+float: 1
+ldouble: 1
+
+Function: "sin_downward":
+double: 1
+float: 1
+ldouble: 3
+
+Function: "sin_towardzero":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "sin_upward":
+double: 1
+float: 1
+ldouble: 3
+
+Function: "sincos":
+double: 1
+ldouble: 1
+
+Function: "sincos_downward":
+double: 1
+float: 1
+ldouble: 3
+
+Function: "sincos_towardzero":
+double: 1
+float: 1
+ldouble: 2
+
+Function: "sincos_upward":
+double: 1
+float: 1
+ldouble: 3
+
+Function: "sinh":
+double: 2
+float: 2
+ldouble: 2
+
+Function: "sinh_downward":
+double: 3
+float: 3
+ldouble: 3
+
+Function: "sinh_towardzero":
+double: 2
+float: 2
+ldouble: 3
+
+Function: "sinh_upward":
+double: 3
+float: 3
+ldouble: 4
+
+Function: "tan":
+float: 1
+ldouble: 1
+
+Function: "tan_downward":
+double: 1
+float: 2
+ldouble: 1
+
+Function: "tan_towardzero":
+double: 1
+float: 1
+ldouble: 1
+
+Function: "tan_upward":
+double: 1
+float: 2
+ldouble: 1
+
+Function: "tanh":
+double: 2
+float: 2
+ldouble: 2
+
+Function: "tanh_downward":
+double: 3
+float: 3
+ldouble: 4
+
+Function: "tanh_towardzero":
+double: 2
+float: 2
+ldouble: 3
+
+Function: "tanh_upward":
+double: 3
+float: 3
+ldouble: 3
+
+Function: "tgamma":
+double: 5
+float: 8
+ldouble: 4
+
+Function: "tgamma_downward":
+double: 5
+float: 7
+ldouble: 5
+
+Function: "tgamma_towardzero":
+double: 5
+float: 7
+ldouble: 5
+
+Function: "tgamma_upward":
+double: 4
+float: 8
+ldouble: 4
+
+Function: "y0":
+double: 2
+float: 6
+ldouble: 3
+
+Function: "y0_downward":
+double: 3
+float: 4
+ldouble: 4
+
+Function: "y0_towardzero":
+double: 3
+float: 3
+ldouble: 5
+
+Function: "y0_upward":
+double: 2
+float: 5
+ldouble: 5
+
+Function: "y1":
+double: 3
+float: 2
+ldouble: 2
+
+Function: "y1_downward":
+double: 3
+float: 2
+ldouble: 4
+
+Function: "y1_towardzero":
+double: 3
+float: 2
+ldouble: 5
+
+Function: "y1_upward":
+double: 5
+float: 2
+ldouble: 5
+
+Function: "yn":
+double: 3
+float: 3
+ldouble: 5
+
+Function: "yn_downward":
+double: 3
+float: 4
+ldouble: 5
+
+Function: "yn_towardzero":
+double: 3
+float: 3
+ldouble: 5
+
+Function: "yn_upward":
+double: 4
+float: 5
+ldouble: 5
+
+# end of automatic generation
diff --git a/sysdeps/riscv/rv32/rvd/libm-test-ulps-name b/sysdeps/riscv/rv32/rvd/libm-test-ulps-name
new file mode 100644
index 0000000000..827fcdca19
--- /dev/null
+++ b/sysdeps/riscv/rv32/rvd/libm-test-ulps-name
@@ -0,0 +1 @@
+RISC-V
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 13/18] RISC-V: Fix llrint and llround missing exceptions on RV32
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (11 preceding siblings ...)
2020-06-03 16:25 ` [PATCH v2 12/18] RISC-V: Add the RV32 libm-test-ulps Alistair Francis via Libc-alpha
@ 2020-06-03 16:26 ` Alistair Francis via Libc-alpha
2020-06-03 16:26 ` [PATCH v2 14/18] RISC-V: Build Infastructure for 32-bit Alistair Francis via Libc-alpha
` (5 subsequent siblings)
18 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:26 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
From: Zong Li <zongbox@gmail.com>
Similar to the fix for MIPS, ARM and S/390, RV32 is missing
correct exception on overflow from llrint and llround functions because
cast from floating-point types to long long do not result in correct
exceptions on overflow.
---
.../riscv/rv32/fix-fp-int-convert-overflow.h | 38 +++++++++++++++++++
1 file changed, 38 insertions(+)
create mode 100644 sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h
diff --git a/sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h b/sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h
new file mode 100644
index 0000000000..9137ee0fd8
--- /dev/null
+++ b/sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h
@@ -0,0 +1,38 @@
+/* Fix for conversion of floating point to integer overflow. RISC-V version.
+ Copyright (C) 2015-2020 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#ifndef FIX_FP_INT_CONVERT_OVERFLOW_H
+#define FIX_FP_INT_CONVERT_OVERFLOW_H 1
+
+/* The generic libgcc2.c conversions from floating point
+ to long long may not raise the correct exceptions on overflow (and
+ may raise spurious "inexact" exceptions even in non-overflow cases,
+ see <https://gcc.gnu.org/bugzilla/show_bug.cgi?id=59412>). */
+#define FIX_FLT_LONG_CONVERT_OVERFLOW 0
+#define FIX_FLT_LLONG_CONVERT_OVERFLOW 1
+
+#define FIX_DBL_LONG_CONVERT_OVERFLOW 0
+#define FIX_DBL_LLONG_CONVERT_OVERFLOW 1
+
+#define FIX_LDBL_LONG_CONVERT_OVERFLOW 0
+#define FIX_LDBL_LLONG_CONVERT_OVERFLOW 0
+
+#define FIX_FLT128_LONG_CONVERT_OVERFLOW 0
+#define FIX_FLT128_LLONG_CONVERT_OVERFLOW 0
+
+#endif /* fix-fp-int-convert-overflow.h */
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 14/18] RISC-V: Build Infastructure for 32-bit
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (12 preceding siblings ...)
2020-06-03 16:26 ` [PATCH v2 13/18] RISC-V: Fix llrint and llround missing exceptions on RV32 Alistair Francis via Libc-alpha
@ 2020-06-03 16:26 ` Alistair Francis via Libc-alpha
2020-06-03 16:26 ` [PATCH v2 15/18] riscv32: Specify the arch_minimum_kernel as 5.4 Alistair Francis via Libc-alpha
` (4 subsequent siblings)
18 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:26 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
From: Zong Li <zongbox@gmail.com>
This patch lays out the top-level orginazition of the RISC-V 32-bit port. It
contains all the Implies files as well as various other fragments of
build infastructure for the RISC-V 32-bit port.
---
sysdeps/riscv/preconfigure | 6 +--
sysdeps/riscv/rv32/Implies-after | 1 +
sysdeps/riscv/rv32/rvd/Implies | 3 ++
sysdeps/riscv/rv32/rvf/Implies | 1 +
sysdeps/unix/sysv/linux/riscv/Makefile | 4 +-
sysdeps/unix/sysv/linux/riscv/configure | 39 ++++++++++++++++++++
sysdeps/unix/sysv/linux/riscv/configure.ac | 8 ++++
sysdeps/unix/sysv/linux/riscv/rv32/Implies | 3 ++
sysdeps/unix/sysv/linux/riscv/shlib-versions | 10 ++++-
9 files changed, 67 insertions(+), 8 deletions(-)
create mode 100644 sysdeps/riscv/rv32/Implies-after
create mode 100644 sysdeps/riscv/rv32/rvd/Implies
create mode 100644 sysdeps/riscv/rv32/rvf/Implies
create mode 100644 sysdeps/unix/sysv/linux/riscv/rv32/Implies
diff --git a/sysdeps/riscv/preconfigure b/sysdeps/riscv/preconfigure
index d9adb31b64..1ab5d20f0e 100644
--- a/sysdeps/riscv/preconfigure
+++ b/sysdeps/riscv/preconfigure
@@ -6,11 +6,7 @@ riscv*)
atomic=`$CC $CFLAGS $CPPFLAGS -E -dM -xc /dev/null | grep '#define __riscv_atomic' | cut -d' ' -f2`
case "$xlen" in
- 32)
- echo "glibc does not yet support 32-bit systems" >&2
- exit 1
- ;;
- 64)
+ 64 | 32)
;;
*)
echo "Unable to determine XLEN" >&2
diff --git a/sysdeps/riscv/rv32/Implies-after b/sysdeps/riscv/rv32/Implies-after
new file mode 100644
index 0000000000..39a34c5f57
--- /dev/null
+++ b/sysdeps/riscv/rv32/Implies-after
@@ -0,0 +1 @@
+wordsize-32
diff --git a/sysdeps/riscv/rv32/rvd/Implies b/sysdeps/riscv/rv32/rvd/Implies
new file mode 100644
index 0000000000..1151214e8f
--- /dev/null
+++ b/sysdeps/riscv/rv32/rvd/Implies
@@ -0,0 +1,3 @@
+riscv/rv32/rvf
+riscv/rvd
+riscv/rvf
diff --git a/sysdeps/riscv/rv32/rvf/Implies b/sysdeps/riscv/rv32/rvf/Implies
new file mode 100644
index 0000000000..66c401443b
--- /dev/null
+++ b/sysdeps/riscv/rv32/rvf/Implies
@@ -0,0 +1 @@
+riscv/rvf
diff --git a/sysdeps/unix/sysv/linux/riscv/Makefile b/sysdeps/unix/sysv/linux/riscv/Makefile
index 301b082398..6c11f7fd9a 100644
--- a/sysdeps/unix/sysv/linux/riscv/Makefile
+++ b/sysdeps/unix/sysv/linux/riscv/Makefile
@@ -7,11 +7,13 @@ ifeq ($(subdir),stdlib)
gen-as-const-headers += ucontext_i.sym
endif
-abi-variants := lp64 lp64d
+abi-variants := ilp32 ilp32d lp64 lp64d
ifeq (,$(filter $(default-abi),$(abi-variants)))
$(error Unknown ABI $(default-abi), must be one of $(abi-variants))
endif
+abi-ilp32-condition := !defined __LP64__ && defined __riscv_float_abi_soft
+abi-ilp32d-condition := !defined __LP64__ && defined __riscv_float_abi_double
abi-lp64-condition := defined __LP64__ && defined __riscv_float_abi_soft
abi-lp64d-condition := defined __LP64__ && defined __riscv_float_abi_double
diff --git a/sysdeps/unix/sysv/linux/riscv/configure b/sysdeps/unix/sysv/linux/riscv/configure
index 3018ca8f1b..2b3c77f18c 100755
--- a/sysdeps/unix/sysv/linux/riscv/configure
+++ b/sysdeps/unix/sysv/linux/riscv/configure
@@ -147,6 +147,17 @@ if (eval "$ac_cpp conftest.$ac_ext") 2>&5 |
fi
rm -f conftest*
+cat confdefs.h - <<_ACEOF >conftest.$ac_ext
+/* end confdefs.h. */
+__SIZEOF_INT__ __SIZEOF_LONG__ __SIZEOF_POINTER__
+
+_ACEOF
+if (eval "$ac_cpp conftest.$ac_ext") 2>&5 |
+ $EGREP "4 4 4" >/dev/null 2>&1; then :
+ libc_cv_riscv_int_abi=ilp32
+fi
+rm -f conftest*
+
if test $libc_cv_riscv_int_abi = no; then
as_fn_error $? "Unable to determine integer ABI" "$LINENO" 5
fi
@@ -214,6 +225,34 @@ case "$prefix" in
;;
esac
;;
+ilp32-riscv/rv32/*)
+ test -n "$libc_cv_slibdir" ||
+case "$prefix" in
+/usr | /usr/)
+ libc_cv_slibdir='/lib32/ilp32'
+ libc_cv_rtlddir='/lib'
+ if test "$libdir" = '${exec_prefix}/lib'; then
+ libdir='${exec_prefix}/lib32/ilp32';
+ # Locale data can be shared between 32-bit and 64-bit libraries.
+ libc_cv_complocaledir='${exec_prefix}/lib/locale'
+ fi
+ ;;
+esac
+ ;;
+ilp32d-riscv/rv32/*)
+ test -n "$libc_cv_slibdir" ||
+case "$prefix" in
+/usr | /usr/)
+ libc_cv_slibdir='/lib32/ilp32d'
+ libc_cv_rtlddir='/lib'
+ if test "$libdir" = '${exec_prefix}/lib'; then
+ libdir='${exec_prefix}/lib32/ilp32d';
+ # Locale data can be shared between 32-bit and 64-bit libraries.
+ libc_cv_complocaledir='${exec_prefix}/lib/locale'
+ fi
+ ;;
+esac
+ ;;
esac
ldd_rewrite_script=sysdeps/unix/sysv/linux/riscv/ldd-rewrite.sed
diff --git a/sysdeps/unix/sysv/linux/riscv/configure.ac b/sysdeps/unix/sysv/linux/riscv/configure.ac
index d4819931ca..710d46afcd 100644
--- a/sysdeps/unix/sysv/linux/riscv/configure.ac
+++ b/sysdeps/unix/sysv/linux/riscv/configure.ac
@@ -7,6 +7,8 @@ arch_minimum_kernel=4.15.0
libc_cv_riscv_int_abi=no
AC_EGREP_CPP(4 8 8, [__SIZEOF_INT__ __SIZEOF_LONG__ __SIZEOF_POINTER__
], libc_cv_riscv_int_abi=lp64)
+AC_EGREP_CPP(4 4 4, [__SIZEOF_INT__ __SIZEOF_LONG__ __SIZEOF_POINTER__
+ ], libc_cv_riscv_int_abi=ilp32)
if test $libc_cv_riscv_int_abi = no; then
AC_MSG_ERROR([Unable to determine integer ABI])
fi
@@ -33,6 +35,12 @@ lp64-riscv/rv64/*)
lp64d-riscv/rv64/*)
LIBC_SLIBDIR_RTLDDIR([lib64/lp64d], [lib])
;;
+ilp32-riscv/rv32/*)
+ LIBC_SLIBDIR_RTLDDIR([lib32/ilp32], [lib])
+ ;;
+ilp32d-riscv/rv32/*)
+ LIBC_SLIBDIR_RTLDDIR([lib32/ilp32d], [lib])
+ ;;
esac
ldd_rewrite_script=sysdeps/unix/sysv/linux/riscv/ldd-rewrite.sed
diff --git a/sysdeps/unix/sysv/linux/riscv/rv32/Implies b/sysdeps/unix/sysv/linux/riscv/rv32/Implies
new file mode 100644
index 0000000000..8b7deb33cd
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/riscv/rv32/Implies
@@ -0,0 +1,3 @@
+unix/sysv/linux/riscv
+unix/sysv/linux/generic/wordsize-32
+unix/sysv/linux/generic
diff --git a/sysdeps/unix/sysv/linux/riscv/shlib-versions b/sysdeps/unix/sysv/linux/riscv/shlib-versions
index 98c9b29cc4..919c8ec1fd 100644
--- a/sysdeps/unix/sysv/linux/riscv/shlib-versions
+++ b/sysdeps/unix/sysv/linux/riscv/shlib-versions
@@ -1,9 +1,15 @@
-DEFAULT GLIBC_2.27
-
%if RISCV_ABI_XLEN == 64 && RISCV_ABI_FLEN == 64
+DEFAULT GLIBC_2.27
ld=ld-linux-riscv64-lp64d.so.1
%elif RISCV_ABI_XLEN == 64 && RISCV_ABI_FLEN == 0
+DEFAULT GLIBC_2.27
ld=ld-linux-riscv64-lp64.so.1
+%elif RISCV_ABI_XLEN == 32 && RISCV_ABI_FLEN == 64
+DEFAULT GLIBC_2.32
+ld=ld-linux-riscv32-ilp32d.so.1
+%elif RISCV_ABI_XLEN == 32 && RISCV_ABI_FLEN == 0
+DEFAULT GLIBC_2.32
+ld=ld-linux-riscv32-ilp32.so.1
%else
%error cannot determine ABI
%endif
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 15/18] riscv32: Specify the arch_minimum_kernel as 5.4
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (13 preceding siblings ...)
2020-06-03 16:26 ` [PATCH v2 14/18] RISC-V: Build Infastructure for 32-bit Alistair Francis via Libc-alpha
@ 2020-06-03 16:26 ` Alistair Francis via Libc-alpha
2020-06-03 16:26 ` [PATCH v2 16/18] RISC-V: Add rv32 path to RTLDLIST in ldd Alistair Francis via Libc-alpha
` (3 subsequent siblings)
18 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:26 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
---
sysdeps/unix/sysv/linux/riscv/configure | 4 ++++
sysdeps/unix/sysv/linux/riscv/configure.ac | 4 ++++
2 files changed, 8 insertions(+)
diff --git a/sysdeps/unix/sysv/linux/riscv/configure b/sysdeps/unix/sysv/linux/riscv/configure
index 2b3c77f18c..33139618e7 100755
--- a/sysdeps/unix/sysv/linux/riscv/configure
+++ b/sysdeps/unix/sysv/linux/riscv/configure
@@ -162,6 +162,10 @@ if test $libc_cv_riscv_int_abi = no; then
as_fn_error $? "Unable to determine integer ABI" "$LINENO" 5
fi
+if test $libc_cv_riscv_int_abi = ilp32; then
+ arch_minimum_kernel=5.4.0
+fi
+
libc_cv_riscv_float_abi=no
cat confdefs.h - <<_ACEOF >conftest.$ac_ext
/* end confdefs.h. */
diff --git a/sysdeps/unix/sysv/linux/riscv/configure.ac b/sysdeps/unix/sysv/linux/riscv/configure.ac
index 710d46afcd..9c736415f7 100644
--- a/sysdeps/unix/sysv/linux/riscv/configure.ac
+++ b/sysdeps/unix/sysv/linux/riscv/configure.ac
@@ -13,6 +13,10 @@ if test $libc_cv_riscv_int_abi = no; then
AC_MSG_ERROR([Unable to determine integer ABI])
fi
+if test $libc_cv_riscv_int_abi = ilp32; then
+ arch_minimum_kernel=5.4.0
+fi
+
libc_cv_riscv_float_abi=no
AC_EGREP_CPP(yes, [#ifdef __riscv_float_abi_double
yes
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 16/18] RISC-V: Add rv32 path to RTLDLIST in ldd
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (14 preceding siblings ...)
2020-06-03 16:26 ` [PATCH v2 15/18] riscv32: Specify the arch_minimum_kernel as 5.4 Alistair Francis via Libc-alpha
@ 2020-06-03 16:26 ` Alistair Francis via Libc-alpha
2020-06-03 16:26 ` [PATCH v2 17/18] Documentation for the RISC-V 32-bit port Alistair Francis via Libc-alpha
` (2 subsequent siblings)
18 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:26 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
From: Zong Li <zong.li@sifive.com>
---
sysdeps/unix/sysv/linux/riscv/ldd-rewrite.sed | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/sysdeps/unix/sysv/linux/riscv/ldd-rewrite.sed b/sysdeps/unix/sysv/linux/riscv/ldd-rewrite.sed
index 377a9c6ef4..bc0c4f51a1 100644
--- a/sysdeps/unix/sysv/linux/riscv/ldd-rewrite.sed
+++ b/sysdeps/unix/sysv/linux/riscv/ldd-rewrite.sed
@@ -1 +1 @@
-s_^\(RTLDLIST=\)\(.*lib/\)\(ld-linux\)-\(riscv64\)-\(lp64\)\(d*\)\(\.so\.[0-9.]*\)_\1"\2\3-\4-\5\7 \2\3-\4-\5d\7"_
+s_^\(RTLDLIST=\)\(.*lib\)\(/[^/]*\)\(64\|32\)\(-.*\)\(lp\)\(64\|32\)\(d*\)\(\.so\.[0-9.]*\)[ ]*$_\1"\2\364-\664\9 \2\364-\664\8\9 \2\332-i\632\9 \2\332-i\632\8\9"_
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 17/18] Documentation for the RISC-V 32-bit port
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (15 preceding siblings ...)
2020-06-03 16:26 ` [PATCH v2 16/18] RISC-V: Add rv32 path to RTLDLIST in ldd Alistair Francis via Libc-alpha
@ 2020-06-03 16:26 ` Alistair Francis via Libc-alpha
2020-06-03 16:26 ` [PATCH v2 18/18] Add RISC-V 32-bit target to build-many-glibcs.py Alistair Francis via Libc-alpha
2020-06-15 22:37 ` [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
18 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:26 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
From: Zong Li <zongbox@gmail.com>
There is already RISC-V 64-bit port information in the documentation.
Let's add some documentation entries for the RISC-V 32-bit as well.
---
NEWS | 6 ++++++
README | 1 +
2 files changed, 7 insertions(+)
diff --git a/NEWS b/NEWS
index a660fc59a8..d408c25592 100644
--- a/NEWS
+++ b/NEWS
@@ -31,6 +31,12 @@ Major new features:
pthread_attr_getsigmask_np have been added. They allow applications
to specify the signal mask of a thread created with pthread_create.
+* Support RISC-V port for 32-bit. The ISA and ABI pairs supported as follows:
+
+ - rv32imac ilp32
+ - rv32imafdc ilp32
+ - rv32imafdc ilp32d
+
Deprecated and removed features, and other changes affecting compatibility:
* The deprecated <sys/sysctl.h> header and the sysctl function have been
diff --git a/README b/README
index 31c5da0405..cee86ab178 100644
--- a/README
+++ b/README
@@ -38,6 +38,7 @@ The GNU C Library supports these configurations for using Linux kernels:
powerpc64*-*-linux-gnu Big-endian and little-endian.
s390-*-linux-gnu
s390x-*-linux-gnu
+ riscv32-*-linux-gnu
riscv64-*-linux-gnu
sh[34]-*-linux-gnu
sparc*-*-linux-gnu
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* [PATCH v2 18/18] Add RISC-V 32-bit target to build-many-glibcs.py
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (16 preceding siblings ...)
2020-06-03 16:26 ` [PATCH v2 17/18] Documentation for the RISC-V 32-bit port Alistair Francis via Libc-alpha
@ 2020-06-03 16:26 ` Alistair Francis via Libc-alpha
2020-06-15 22:37 ` [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
18 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-03 16:26 UTC (permalink / raw)
To: libc-alpha; +Cc: alistair.francis
From: Zong Li <zongbox@gmail.com>
Support building three variant of 32 bit RISC-V glibc as follows:
- riscv32-linux-gnu-rv32imac-ilp32
- riscv32-linux-gnu-rv32imafdc-ilp32
- riscv32-linux-gnu-rv32imafdc-ilp32d
---
scripts/build-many-glibcs.py | 15 +++++++++++++++
1 file changed, 15 insertions(+)
diff --git a/scripts/build-many-glibcs.py b/scripts/build-many-glibcs.py
index 6c6615b200..929ad8d0a5 100755
--- a/scripts/build-many-glibcs.py
+++ b/scripts/build-many-glibcs.py
@@ -338,6 +338,21 @@ class Context(object):
self.add_config(arch='powerpc64le',
os_name='linux-gnu',
gcc_cfg=['--disable-multilib', '--enable-secureplt'])
+ self.add_config(arch='riscv32',
+ os_name='linux-gnu',
+ variant='rv32imac-ilp32',
+ gcc_cfg=['--with-arch=rv32imac', '--with-abi=ilp32',
+ '--disable-multilib'])
+ self.add_config(arch='riscv32',
+ os_name='linux-gnu',
+ variant='rv32imafdc-ilp32',
+ gcc_cfg=['--with-arch=rv32imafdc', '--with-abi=ilp32',
+ '--disable-multilib'])
+ self.add_config(arch='riscv32',
+ os_name='linux-gnu',
+ variant='rv32imafdc-ilp32d',
+ gcc_cfg=['--with-arch=rv32imafdc', '--with-abi=ilp32d',
+ '--disable-multilib'])
self.add_config(arch='riscv64',
os_name='linux-gnu',
variant='rv64imac-lp64',
--
2.26.2
^ permalink raw reply related [flat|nested] 55+ messages in thread
* Re: [PATCH v2 12/18] RISC-V: Add the RV32 libm-test-ulps
2020-06-03 16:25 ` [PATCH v2 12/18] RISC-V: Add the RV32 libm-test-ulps Alistair Francis via Libc-alpha
@ 2020-06-03 17:34 ` Joseph Myers
2020-06-05 4:01 ` Alistair Francis via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Joseph Myers @ 2020-06-03 17:34 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> Add a libm-test-ulps for RV32, generated by running `make regen-ulps`.
> This is different to the RV64 one.
*Why* is it different from the RV64 one? (If some functions produce
different results because of use of sysdeps/ieee754/dbl-64/wordsize-64/
variants on RV64, that would be a good reason, for example, but I don't
see ieee754/dbl-64/wordsize-64 mentioned in Implies for RV64.)
--
Joseph S. Myers
joseph@codesourcery.com
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 12/18] RISC-V: Add the RV32 libm-test-ulps
2020-06-03 17:34 ` Joseph Myers
@ 2020-06-05 4:01 ` Alistair Francis via Libc-alpha
0 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-05 4:01 UTC (permalink / raw)
To: Joseph Myers; +Cc: GNU C Library, Alistair Francis
On Wed, Jun 3, 2020 at 10:34 AM Joseph Myers <joseph@codesourcery.com> wrote:
>
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
> > Add a libm-test-ulps for RV32, generated by running `make regen-ulps`.
> > This is different to the RV64 one.
>
> *Why* is it different from the RV64 one? (If some functions produce
> different results because of use of sysdeps/ieee754/dbl-64/wordsize-64/
> variants on RV64, that would be a good reason, for example, but I don't
> see ieee754/dbl-64/wordsize-64 mentioned in Implies for RV64.)
So I just checked and I think they can actually be the same. The
reason they end up different is that the RV32 one is generated on QEMU
while the RV64 one I'm assuming was generated on hardware.
I can replace the RV32 one with the RV64 one.
Alistair
>
> --
> Joseph S. Myers
> joseph@codesourcery.com
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32)
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
` (17 preceding siblings ...)
2020-06-03 16:26 ` [PATCH v2 18/18] Add RISC-V 32-bit target to build-many-glibcs.py Alistair Francis via Libc-alpha
@ 2020-06-15 22:37 ` Alistair Francis via Libc-alpha
18 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-06-15 22:37 UTC (permalink / raw)
To: Alistair Francis; +Cc: GNU C Library
On Wed, Jun 3, 2020 at 9:34 AM Alistair Francis
<alistair.francis@wdc.com> wrote:
>
> This patch set contains the glibc port for 32-bit RISC-V.
>
> This is based on the original work from Zong Li [1] and has been
> updated to use a 64-bit time_t.
>
> This requires a 5.4+ kernel and all of the testing has been done using
> the 5.4 stable kernel.
>
> Nothing fails when running ./scripts/build-many-glibcs.py (for all
> targets) on my x86-64 machine.
>
> This is the current list of tests that fail when running inside QEMU RV32
> system emulation on the 5.4 kernel:
>
> FAIL: elf/tst-ldconfig-ld_so_conf-update
> FAIL: io/tst-lockf
> FAIL: misc/test-errno-linux
> FAIL: nss/tst-nss-files-hosts-long
> FAIL: resolv/tst-resolv-res_init-thread
> FAIL: stdio-common/bug22
> FAIL: stdlib/tst-strfrom
> FAIL: stdlib/tst-strfrom-locale
>
> ---Links---
> 1: https://sourceware.org/ml/libc-alpha/2018-07/msg00892.html
>
> The latest version of my work can be found here: https://github.com/alistair23/glibc/tree/alistair/rv32.next
>
> This specific version can be found here: https://github.com/alistair23/glibc/tree/alistair/rv32.2
>
> Alistair Francis (9):
> RISC-V: Use 64-bit time_t and off_t for RV32 and RV64
> RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit
> RISC-V: Add support for 32-bit vDSO calls
> RISC-V: Add arch-syscall.h for RV32
> RISC-V: nptl: update default pthread-offsets.h
> riscv32: Add an architecture ipctypes.h
> RISC-V: Add ABI lists
> RISC-V: Add the RV32 libm-test-ulps
> riscv32: Specify the arch_minimum_kernel as 5.4
>
> Zong Li (9):
> RISC-V: Support dynamic loader for the 32-bit
> RISC-V: Add path of library directories for the 32-bit
> RISC-V: The ABI implementation for 32-bit
> RISC-V: Hard float support for 32-bit
> RISC-V: Fix llrint and llround missing exceptions on RV32
> RISC-V: Build Infastructure for 32-bit
> RISC-V: Add rv32 path to RTLDLIST in ldd
> Documentation for the RISC-V 32-bit port
> Add RISC-V 32-bit target to build-many-glibcs.py
Ping!
Patch 12 (RISC-V: Add the RV32 libm-test-ulps) had a comment, any
other comments?
I would like to start merging some of the early patches of this series
to avoid continually reposting them without comments or changes.
Alistair
>
> NEWS | 6 +
> README | 1 +
> scripts/build-many-glibcs.py | 15 +
> sysdeps/riscv/bits/wordsize.h | 4 +-
> sysdeps/riscv/nptl/bits/pthreadtypes-arch.h | 10 +-
> sysdeps/riscv/nptl/bits/struct_rwlock.h | 27 +-
> sysdeps/riscv/nptl/pthread-offsets.h | 13 +-
> sysdeps/riscv/preconfigure | 6 +-
> sysdeps/riscv/rv32/Implies-after | 1 +
> .../riscv/rv32/fix-fp-int-convert-overflow.h | 38 +
> sysdeps/riscv/rv32/rvd/Implies | 3 +
> sysdeps/riscv/rv32/rvd/libm-test-ulps | 1401 +++++++++++
> sysdeps/riscv/rv32/rvd/libm-test-ulps-name | 1 +
> sysdeps/riscv/rv32/rvd/s_lrint.c | 31 +
> sysdeps/riscv/rv32/rvd/s_lround.c | 31 +
> sysdeps/riscv/rv32/rvf/Implies | 1 +
> sysdeps/riscv/rv32/rvf/s_lrintf.c | 31 +
> sysdeps/riscv/rv32/rvf/s_lroundf.c | 31 +
> sysdeps/riscv/sfp-machine.h | 27 +-
> sysdeps/riscv/sys/asm.h | 5 +-
> sysdeps/unix/sysv/linux/riscv/Makefile | 4 +-
> .../unix/sysv/linux/riscv/bits/environments.h | 85 +
> sysdeps/unix/sysv/linux/riscv/bits/ipctypes.h | 35 +
> sysdeps/unix/sysv/linux/riscv/bits/time64.h | 36 +
> sysdeps/unix/sysv/linux/riscv/bits/timesize.h | 22 +
> sysdeps/unix/sysv/linux/riscv/c++-types.data | 67 +
> sysdeps/unix/sysv/linux/riscv/configure | 43 +
> sysdeps/unix/sysv/linux/riscv/configure.ac | 12 +
> sysdeps/unix/sysv/linux/riscv/dl-cache.h | 17 +-
> .../unix/sysv/linux/riscv/jmp_buf-macros.h | 53 +
> sysdeps/unix/sysv/linux/riscv/kernel_stat.h | 23 +
> sysdeps/unix/sysv/linux/riscv/ld.abilist | 5 +
> sysdeps/unix/sysv/linux/riscv/ldconfig.h | 2 +-
> sysdeps/unix/sysv/linux/riscv/ldd-rewrite.sed | 2 +-
> .../sysv/linux/riscv/libBrokenLocale.abilist | 1 +
> sysdeps/unix/sysv/linux/riscv/libanl.abilist | 4 +
> sysdeps/unix/sysv/linux/riscv/libc.abilist | 2099 +++++++++++++++++
> .../unix/sysv/linux/riscv/libcrypt.abilist | 2 +
> sysdeps/unix/sysv/linux/riscv/libdl.abilist | 9 +
> sysdeps/unix/sysv/linux/riscv/libm.abilist | 940 ++++++++
> .../unix/sysv/linux/riscv/libpthread.abilist | 213 ++
> .../unix/sysv/linux/riscv/libresolv.abilist | 79 +
> sysdeps/unix/sysv/linux/riscv/librt.abilist | 35 +
> .../sysv/linux/riscv/libthread_db.abilist | 40 +
> sysdeps/unix/sysv/linux/riscv/libutil.abilist | 6 +
> sysdeps/unix/sysv/linux/riscv/rv32/Implies | 3 +
> .../unix/sysv/linux/riscv/rv32/arch-syscall.h | 283 +++
> sysdeps/unix/sysv/linux/riscv/shlib-versions | 10 +-
> sysdeps/unix/sysv/linux/riscv/sysdep.h | 80 +-
> 49 files changed, 5868 insertions(+), 25 deletions(-)
> create mode 100644 sysdeps/riscv/rv32/Implies-after
> create mode 100644 sysdeps/riscv/rv32/fix-fp-int-convert-overflow.h
> create mode 100644 sysdeps/riscv/rv32/rvd/Implies
> create mode 100644 sysdeps/riscv/rv32/rvd/libm-test-ulps
> create mode 100644 sysdeps/riscv/rv32/rvd/libm-test-ulps-name
> create mode 100644 sysdeps/riscv/rv32/rvd/s_lrint.c
> create mode 100644 sysdeps/riscv/rv32/rvd/s_lround.c
> create mode 100644 sysdeps/riscv/rv32/rvf/Implies
> create mode 100644 sysdeps/riscv/rv32/rvf/s_lrintf.c
> create mode 100644 sysdeps/riscv/rv32/rvf/s_lroundf.c
> create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/environments.h
> create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/ipctypes.h
> create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/time64.h
> create mode 100644 sysdeps/unix/sysv/linux/riscv/bits/timesize.h
> create mode 100644 sysdeps/unix/sysv/linux/riscv/c++-types.data
> create mode 100644 sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h
> create mode 100644 sysdeps/unix/sysv/linux/riscv/kernel_stat.h
> create mode 100644 sysdeps/unix/sysv/linux/riscv/ld.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/libBrokenLocale.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/libanl.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/libc.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/libcrypt.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/libdl.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/libm.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/libpthread.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/libresolv.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/librt.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/libthread_db.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/libutil.abilist
> create mode 100644 sysdeps/unix/sysv/linux/riscv/rv32/Implies
> create mode 100644 sysdeps/unix/sysv/linux/riscv/rv32/arch-syscall.h
>
> --
> 2.26.2
>
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 01/18] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64
2020-06-03 16:25 ` [PATCH v2 01/18] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64 Alistair Francis via Libc-alpha
@ 2020-07-07 22:06 ` Maciej W. Rozycki via Libc-alpha
2020-07-10 15:27 ` Alistair Francis via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-07 22:06 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> Using the original glibc headers under bits/ let's make small
> modifications to use 64-bit time_t and off_t for both RV32 and RV64.
>
> For the typesizes.h, here are justifications for the changes from the
> generic version (based on Arnd's very helpful feedback):
> - All the !__USE_FILE_OFFSET64 types (__off_t, __ino_t, __rlim_t, ...) are
^^ ^^
Please remove extraneous spaces above.
> changed to match the 64-bit replacements.
>
> - __time_t is defined to 64 bit, but no __time64_t is added. This makes sense
> as we don't have the time64 support for other 32-bit architectures yet, and
> it will be easy to change when that happens.
>
> - __suseconds_t is 64-bit. This matches what we use the kernel ABI for the
> few drivers that are relying on 'struct timeval' input arguments in
> ioctl, as well as the adjtimex system call. It means that timeval has to
> be defined without the padding, unlike timespec, which needs padding.
^^^
Likewise.
The commit description will be indented by 4 when presented by `git log',
etc., so please reformat so as to stay within 74-75 columns to avoid line
wrapping. Also use two spaces after the end of a sentence throughout.
Please apply these guidelines across the whole series.
> diff --git a/sysdeps/unix/sysv/linux/riscv/bits/environments.h b/sysdeps/unix/sysv/linux/riscv/bits/environments.h
> new file mode 100644
> index 0000000000..c1768af17b
> --- /dev/null
> +++ b/sysdeps/unix/sysv/linux/riscv/bits/environments.h
> @@ -0,0 +1,85 @@
> +/* Copyright (C) 2019-2020 Free Software Foundation, Inc.
> + This file is part of the GNU C Library.
> +
> + The GNU C Library is free software; you can redistribute it and/or
> + modify it under the terms of the GNU Lesser General Public
> + License as published by the Free Software Foundation; either
> + version 2.1 of the License, or (at your option) any later version.
> +
> + The GNU C Library is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + Lesser General Public License for more details.
> +
> + You should have received a copy of the GNU Lesser General Public
> + License along with the GNU C Library; if not, see
> + <http://www.gnu.org/licenses/>. */
We've switched to HTTPS, so: <https://www.gnu.org/licenses/>.
> +/* This header should define the following symbols under the described
> + situations. A value `1' means that the model is always supported,
> + `-1' means it is never supported. Undefined means it cannot be
> + statically decided.
> +
> + _POSIX_V7_ILP32_OFF32 32bit int, long, pointers, and off_t type
> + _POSIX_V7_ILP32_OFFBIG 32bit int, long, and pointers and larger off_t type
> +
> + _POSIX_V7_LP64_OFF32 64bit long and pointers and 32bit off_t type
> + _POSIX_V7_LPBIG_OFFBIG 64bit long and pointers and large off_t type
Stray tab after _POSIX_V7_LP64_OFF32?
> +#else /* __WORDSIZE == 32 */
> +
> +/* RISC-V requires 64-bit off_t
> + # undef _POSIX_V7_ILP32_OFF32
> + # undef _POSIX_V6_ILP32_OFF32
> + # undef _XBS5_ILP32_OFF32
> + */
If 64-bit `off_t' is required, then shouldn't these be #define'd to -1
instead? Otherwise under what circumstances support for these models will
have to be determined at run time?
+# define _POSIX_V7_ILP32_OFFBIG 1
+# define _POSIX_V6_ILP32_OFFBIG 1
+# define _XBS5_ILP32_OFFBIG 1
Please use tabs here like with the rest of the macros.
> +/* CFLAGS. */
> +#define __ILP32_OFFBIG_CFLAGS "-D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64"
Likewise if 64-bit `off_t' is required, then why do we have to define
these macros in CFLAGS and what happens if they aren't defined?
Please use a tab here as well (assuming this macro does have to exist).
> diff --git a/sysdeps/unix/sysv/linux/riscv/bits/time64.h b/sysdeps/unix/sysv/linux/riscv/bits/time64.h
> new file mode 100644
> index 0000000000..7ba0dd1cfd
> --- /dev/null
> +++ b/sysdeps/unix/sysv/linux/riscv/bits/time64.h
> @@ -0,0 +1,36 @@
> +/* bits/time64.h -- underlying types for __time64_t. Generic version.
"RISC-V version." presumably?
> + Copyright (C) 2019-2020 Free Software Foundation, Inc.
> + This file is part of the GNU C Library.
> +
> + The GNU C Library is free software; you can redistribute it and/or
> + modify it under the terms of the GNU Lesser General Public
> + License as published by the Free Software Foundation; either
> + version 2.1 of the License, or (at your option) any later version.
> +
> + The GNU C Library is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + Lesser General Public License for more details.
> +
> + You should have received a copy of the GNU Lesser General Public
> + License along with the GNU C Library; if not, see
> + <http://www.gnu.org/licenses/>. */
<https://www.gnu.org/licenses/>
> diff --git a/sysdeps/unix/sysv/linux/riscv/bits/timesize.h b/sysdeps/unix/sysv/linux/riscv/bits/timesize.h
> new file mode 100644
> index 0000000000..78394b7489
> --- /dev/null
> +++ b/sysdeps/unix/sysv/linux/riscv/bits/timesize.h
> @@ -0,0 +1,22 @@
> +/* Bit size of the time_t type at glibc build time, general case.
"RISC-V case."?
> + Copyright (C) 2019-2020 Free Software Foundation, Inc.
> + This file is part of the GNU C Library.
> +
> + The GNU C Library is free software; you can redistribute it and/or
> + modify it under the terms of the GNU Lesser General Public
> + License as published by the Free Software Foundation; either
> + version 2.1 of the License, or (at your option) any later version.
> +
> + The GNU C Library is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + Lesser General Public License for more details.
> +
> + You should have received a copy of the GNU Lesser General Public
> + License along with the GNU C Library; if not, see
> + <http://www.gnu.org/licenses/>. */
<https://www.gnu.org/licenses/>
> diff --git a/sysdeps/unix/sysv/linux/riscv/kernel_stat.h b/sysdeps/unix/sysv/linux/riscv/kernel_stat.h
> new file mode 100644
> index 0000000000..25e698b9bf
> --- /dev/null
> +++ b/sysdeps/unix/sysv/linux/riscv/kernel_stat.h
> @@ -0,0 +1,23 @@
> +/* Copyright (C) 2019-2020 Free Software Foundation, Inc.
> + This file is part of the GNU C Library.
> +
> + The GNU C Library is free software; you can redistribute it and/or
> + modify it under the terms of the GNU Lesser General Public
> + License as published by the Free Software Foundation; either
> + version 2.1 of the License, or (at your option) any later version.
> +
> + The GNU C Library is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + Lesser General Public License for more details.
> +
> + You should have received a copy of the GNU Lesser General Public
> + License along with the GNU C Library. If not, see
> + <http://www.gnu.org/licenses/>. */
<https://www.gnu.org/licenses/>
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 02/18] RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit
2020-06-03 16:25 ` [PATCH v2 02/18] RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit Alistair Francis via Libc-alpha
@ 2020-07-08 0:09 ` Maciej W. Rozycki via Libc-alpha
2020-07-08 17:08 ` Adhemerval Zanella via Libc-alpha
2020-07-09 17:10 ` Alistair Francis via Libc-alpha
0 siblings, 2 replies; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-08 0:09 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
Alistair,
I think the change heading is too cryptic and does not express the intent
of the change well enough. How about:
RISC-V: Use 64-bit-time syscall numbers with the 32-bit port
and then maybe explain in a little more details in the change description.
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> diff --git a/sysdeps/unix/sysv/linux/riscv/sysdep.h b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> index 83e4adf6a2..aa61e8b04d 100644
> --- a/sysdeps/unix/sysv/linux/riscv/sysdep.h
> +++ b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> @@ -116,6 +116,67 @@
>
> #include <sysdeps/unix/sysdep.h>
This file is weird, as it includes <sysdeps/unix/sysdep.h> twice, first
time indirectly via <sysdeps/unix/sysv/linux/generic/sysdep.h> at the top,
and then second time here. So I think this second inclusion can be
removed (along with the preceding inclusion of <errno.h>, as it does not
appear to change anything), and the following conditional moved to the
top, just after the inclusion of <tls.h>. Oddly <sysdeps/unix/sysdep.h>
has not been protected against multiple inclusion, but its contents do not
trigger compilation warnings if processed more than once.
The two #ifdef/#ifndef __ASSEMBLER__ conditionals will then become
adjacent and can be merged into a single #ifdef/#else one.
This clean-up would probably better be made as a separate preceding
change.
> +#if __riscv_xlen == 32
I think using __WORDSIZE here would be more consistent with the rest of
our code (we do use `__riscv_xlen' in a couple of places, but I think they
ought to be cleaned up).
> +/* Define the __NR_futex as __NR_futex64 as RV32 doesn't have a
> + * __NR_futex syscall.
> + */
> +# ifndef __NR_futex
> +# define __NR_futex __NR_futex_time64
> +# endif
The comment does not match the code.
I think it makes no sense to comment on individual entries as they all
repeat the same pattern and the same purpose, so an introductory comment
covering them all at the beginning of the conditional would be better
instead. I suppose you can then reuse it for the change description too.
> +# ifndef __NR_clock_adjtime
> +# define __NR_clock_adjtime __NR_clock_adjtime64
> +# endif
> +#endif /* __riscv_xlen == 32 */
Since you have multiple inner conditionals separated by an empty line
each I think it will make sense to have an empty line as well between the
final one and the closing of the outer conditional. Likewise at the
beginning.
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 03/18] RISC-V: Add support for 32-bit vDSO calls
2020-06-03 16:25 ` [PATCH v2 03/18] RISC-V: Add support for 32-bit vDSO calls Alistair Francis via Libc-alpha
@ 2020-07-08 1:01 ` Maciej W. Rozycki via Libc-alpha
2020-07-08 18:17 ` Alistair Francis via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-08 1:01 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> diff --git a/sysdeps/unix/sysv/linux/riscv/sysdep.h b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> index aa61e8b04d..9f6ed2001b 100644
> --- a/sysdeps/unix/sysv/linux/riscv/sysdep.h
> +++ b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> @@ -182,14 +182,21 @@
>
> #ifndef __ASSEMBLER__
>
> -# define VDSO_NAME "LINUX_4.15"
> -# define VDSO_HASH 182943605
> +# if __riscv_xlen == 32
Same note on __riscv_xlen vs __WORDSIZE as with 02/18.
Also given it is a two-way conditional we can choose which leg is first
and I'd make that the RV64 variant, as the first to be there and more
developed, as evident here.
> /* List of system calls which are supported as vsyscalls. */
> -# define HAVE_CLOCK_GETRES64_VSYSCALL "__vdso_clock_getres"
> -# define HAVE_CLOCK_GETTIME64_VSYSCALL "__vdso_clock_gettime"
> -# define HAVE_GETTIMEOFDAY_VSYSCALL "__vdso_gettimeofday"
> -# define HAVE_GETCPU_VSYSCALL "__vdso_getcpu"
> +# else
> +# define VDSO_NAME "LINUX_4.15"
> +# define VDSO_HASH 182943605
> +
> +/* List of system calls which are supported as vsyscalls. */
> +# define HAVE_CLOCK_GETRES64_VSYSCALL "__vdso_clock_getres"
> +# define HAVE_CLOCK_GETTIME64_VSYSCALL "__vdso_clock_gettime"
> +# define HAVE_GETTIMEOFDAY_VSYSCALL "__vdso_gettimeofday"
> +# define HAVE_GETCPU_VSYSCALL "__vdso_getcpu"
> +# endif
Also "__vdso_getcpu" does get exported unlike the time calls with RV32
AFAICT:
# Symbols present in the vdso
vdso-syms = rt_sigreturn
ifdef CONFIG_64BIT
vdso-syms += vgettimeofday
endif
vdso-syms += getcpu
vdso-syms += flush_icache
so it should be included (and placed outside the conditional, unlike with
an earlier version of this change). Then the change description (and
preferably a comment as well) ought to explain why the remaining ones have
been left out for RV32.
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 04/18] RISC-V: Support dynamic loader for the 32-bit
2020-06-03 16:25 ` [PATCH v2 04/18] RISC-V: Support dynamic loader for the 32-bit Alistair Francis via Libc-alpha
@ 2020-07-08 1:35 ` Maciej W. Rozycki via Libc-alpha
0 siblings, 0 replies; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-08 1:35 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> From: Zong Li <zongbox@gmail.com>
>
> Add the LD_SO_ABI definition for RISC-V 32-bit.
LGTM.
Reviewed-by: Maciej W. Rozycki <macro@wdc.com>
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 02/18] RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit
2020-07-08 0:09 ` Maciej W. Rozycki via Libc-alpha
@ 2020-07-08 17:08 ` Adhemerval Zanella via Libc-alpha
2020-07-09 17:14 ` Alistair Francis via Libc-alpha
2020-07-16 0:23 ` Maciej W. Rozycki via Libc-alpha
2020-07-09 17:10 ` Alistair Francis via Libc-alpha
1 sibling, 2 replies; 55+ messages in thread
From: Adhemerval Zanella via Libc-alpha @ 2020-07-08 17:08 UTC (permalink / raw)
To: libc-alpha
On 07/07/2020 21:09, Maciej W. Rozycki via Libc-alpha wrote:
> Alistair,
>
> I think the change heading is too cryptic and does not express the intent
> of the change well enough. How about:
>
> RISC-V: Use 64-bit-time syscall numbers with the 32-bit port
>
> and then maybe explain in a little more details in the change description.
>
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
>> diff --git a/sysdeps/unix/sysv/linux/riscv/sysdep.h b/sysdeps/unix/sysv/linux/riscv/sysdep.h
>> index 83e4adf6a2..aa61e8b04d 100644
>> --- a/sysdeps/unix/sysv/linux/riscv/sysdep.h
>> +++ b/sysdeps/unix/sysv/linux/riscv/sysdep.h
>> @@ -116,6 +116,67 @@
>>
>> #include <sysdeps/unix/sysdep.h>
>
> This file is weird, as it includes <sysdeps/unix/sysdep.h> twice, first
> time indirectly via <sysdeps/unix/sysv/linux/generic/sysdep.h> at the top,
> and then second time here. So I think this second inclusion can be
> removed (along with the preceding inclusion of <errno.h>, as it does not
> appear to change anything), and the following conditional moved to the
> top, just after the inclusion of <tls.h>. Oddly <sysdeps/unix/sysdep.h>
> has not been protected against multiple inclusion, but its contents do not
> trigger compilation warnings if processed more than once.
This multiple inclusion is done internally in some places so the file
that wants to override a definition would first include the generic
definitions and then #undef/#define the new one (kernel-features.h
still does it).
I think this is kind of confusing, but this is how sysdep.h is currently
organized. We have been changing bit per bit, but there is a lot of
place where inclusing is done without guards.
>
> The two #ifdef/#ifndef __ASSEMBLER__ conditionals will then become
> adjacent and can be merged into a single #ifdef/#else one.
>
> This clean-up would probably better be made as a separate preceding
> change.
>
>> +#if __riscv_xlen == 32
>
> I think using __WORDSIZE here would be more consistent with the rest of
> our code (we do use `__riscv_xlen' in a couple of places, but I think they
> ought to be cleaned up).
>
>> +/* Define the __NR_futex as __NR_futex64 as RV32 doesn't have a
>> + * __NR_futex syscall.
>> + */
>> +# ifndef __NR_futex
>> +# define __NR_futex __NR_futex_time64
>> +# endif
>
> The comment does not match the code.
>
> I think it makes no sense to comment on individual entries as they all
> repeat the same pattern and the same purpose, so an introductory comment
> covering them all at the beginning of the conditional would be better
> instead. I suppose you can then reuse it for the change description too.
>
>> +# ifndef __NR_clock_adjtime
>> +# define __NR_clock_adjtime __NR_clock_adjtime64
>> +# endif
>> +#endif /* __riscv_xlen == 32 */
>
> Since you have multiple inner conditionals separated by an empty line
> each I think it will make sense to have an empty line as well between the
> final one and the closing of the outer conditional. Likewise at the
> beginning.
I think the correct way to how ARC port is doing which is:
1. Add any required syscall suppression for 32-bit off_t/time_t on
fixup-asm-unistd.h so arch-syscall.h will have only the required
definitions. This might be the case for riscv32 since its kernel
ABI only supports 64-bit off_t/time_t.
2. Regenerate arch-syscall.h if it were the case.
3. On sysdep.h redefine only the syscall where the generic implementation
still does not have actual 64-bit time_t support:
/* "workarounds" for generic code needing to handle 64-bit time_t. */
/* Fix sysdeps/unix/sysv/linux/clock_getcpuclockid.c. */
#define __NR_clock_getres __NR_clock_getres_time64
/* Fix sysdeps/nptl/lowlevellock-futex.h. */
#define __NR_futex __NR_futex_time64
[...]
(with proper guards it should require the #ifndef/#define)
4. Add a comment it is a workaround to handle 64-bit time_t
and on each #define comment for which implementation it intends to.
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 03/18] RISC-V: Add support for 32-bit vDSO calls
2020-07-08 1:01 ` Maciej W. Rozycki via Libc-alpha
@ 2020-07-08 18:17 ` Alistair Francis via Libc-alpha
0 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-07-08 18:17 UTC (permalink / raw)
To: Maciej W. Rozycki; +Cc: GNU C Library, Alistair Francis
On Tue, Jul 7, 2020 at 6:02 PM Maciej W. Rozycki via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
> > diff --git a/sysdeps/unix/sysv/linux/riscv/sysdep.h b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> > index aa61e8b04d..9f6ed2001b 100644
> > --- a/sysdeps/unix/sysv/linux/riscv/sysdep.h
> > +++ b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> > @@ -182,14 +182,21 @@
> >
> > #ifndef __ASSEMBLER__
> >
> > -# define VDSO_NAME "LINUX_4.15"
> > -# define VDSO_HASH 182943605
> > +# if __riscv_xlen == 32
>
> Same note on __riscv_xlen vs __WORDSIZE as with 02/18.
Fixed
>
> Also given it is a two-way conditional we can choose which leg is first
> and I'd make that the RV64 variant, as the first to be there and more
> developed, as evident here.
Fixed.
>
> > /* List of system calls which are supported as vsyscalls. */
> > -# define HAVE_CLOCK_GETRES64_VSYSCALL "__vdso_clock_getres"
> > -# define HAVE_CLOCK_GETTIME64_VSYSCALL "__vdso_clock_gettime"
> > -# define HAVE_GETTIMEOFDAY_VSYSCALL "__vdso_gettimeofday"
> > -# define HAVE_GETCPU_VSYSCALL "__vdso_getcpu"
> > +# else
> > +# define VDSO_NAME "LINUX_4.15"
> > +# define VDSO_HASH 182943605
> > +
> > +/* List of system calls which are supported as vsyscalls. */
> > +# define HAVE_CLOCK_GETRES64_VSYSCALL "__vdso_clock_getres"
> > +# define HAVE_CLOCK_GETTIME64_VSYSCALL "__vdso_clock_gettime"
> > +# define HAVE_GETTIMEOFDAY_VSYSCALL "__vdso_gettimeofday"
> > +# define HAVE_GETCPU_VSYSCALL "__vdso_getcpu"
> > +# endif
>
> Also "__vdso_getcpu" does get exported unlike the time calls with RV32
> AFAICT:
>
> # Symbols present in the vdso
> vdso-syms = rt_sigreturn
> ifdef CONFIG_64BIT
> vdso-syms += vgettimeofday
> endif
> vdso-syms += getcpu
> vdso-syms += flush_icache
>
> so it should be included (and placed outside the conditional, unlike with
> an earlier version of this change). Then the change description (and
> preferably a comment as well) ought to explain why the remaining ones have
> been left out for RV32.
Ok, I have fixed this as well.
Alistair
>
> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 05/18] RISC-V: Add path of library directories for the 32-bit
2020-06-03 16:25 ` [PATCH v2 05/18] RISC-V: Add path of library directories " Alistair Francis via Libc-alpha
@ 2020-07-08 18:42 ` Maciej W. Rozycki via Libc-alpha
2020-07-09 17:03 ` Alistair Francis via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-08 18:42 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> From: Zong Li <zongbox@gmail.com>
>
> For the recommand of 64 bit version, we add the libraries path of 32 bit
> in this patch.
I have troubles comprehending this sentence; also it's repeated literally
below. Please rewrite an remove the duplication.
> The status of RV32 binaries under RV64 kernels is that the ISA
> optionally supports having different XLEN for user and supervisor modes,
> but AFAIK there's no silicon that implements this feature, and the Linux
> kernel doesn't support it yet.
I'm not sure if this note is relevant here.
> For the recommand of 64 bit version, we add the libraries path of 32 bit
> in this patch. This includes a fix to avoid an out of bound array check
> when building with GCC 8.2.
Where exactly is that fix? Shouldn't it be applied as a separate
preparatory change?
> diff --git a/sysdeps/unix/sysv/linux/riscv/dl-cache.h b/sysdeps/unix/sysv/linux/riscv/dl-cache.h
> index c297dfe84f..60fc172edb 100644
> --- a/sysdeps/unix/sysv/linux/riscv/dl-cache.h
> +++ b/sysdeps/unix/sysv/linux/riscv/dl-cache.h
[...]
> @@ -49,9 +51,16 @@
> do \
> { \
> size_t len = strlen (dir); \
> - char path[len + 9]; \
> + char path[len + 10]; \
> memcpy (path, dir, len + 1); \
> - if (len >= 12 && ! memcmp(path + len - 12, "/lib64/lp64d", 12)) \
> + if (len >= 13 && ! memcmp(path + len - 13, "/lib32/ilp32d", 13)) \
> + { \
> + len -= 9; \
> + path[len] = '\0'; \
> + } \
> + if (len >= 12 \
> + && (! memcmp(path + len - 12, "/lib32/ilp32", 12) \
> + || ! memcmp(path + len - 12, "/lib64/lp64d", 12))) \
Please correct indentation above and use tabs throughout.
I think this code should use `else' clauses. While the truncation of the
string will cause subsequent `memcmp' calls to fail, they'll still be
executed, `len' permitting, making this ugly.
Though frankly with the growing number of entries I would rewrite this
sequence of conditionals entirely, as a loop over a (static) array of the
strings processed, cutting the insane number of error-prone hardcoded
string lengths while at it too. It is only ever used in `ldconfig', and
then invoked there at most twice, so it is not that it is critical enough
for performance to justify illegibility, and making additional `strlen'
calls shouldn't be a big deal.
> @@ -64,6 +73,10 @@
> add_dir (path); \
> if (len >= 4 && ! memcmp(path + len - 4, "/lib", 4)) \
> { \
> + memcpy (path + len, "32/ilp32d", 10); \
> + add_dir (path); \
> + memcpy (path + len, "32/ilp32", 9); \
> + add_dir (path); \
Then the same array of strings could be used here (skipping over the
"/lib" prefix throughout).
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 06/18] RISC-V: Add arch-syscall.h for RV32
2020-06-03 16:25 ` [PATCH v2 06/18] RISC-V: Add arch-syscall.h for RV32 Alistair Francis via Libc-alpha
@ 2020-07-08 19:33 ` Maciej W. Rozycki via Libc-alpha
0 siblings, 0 replies; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-08 19:33 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> ---
> .../unix/sysv/linux/riscv/rv32/arch-syscall.h | 283 ++++++++++++++++++
> 1 file changed, 283 insertions(+)
> create mode 100644 sysdeps/unix/sysv/linux/riscv/rv32/arch-syscall.h
LGTM.
Reviewed-by: Maciej W. Rozycki <macro@wdc.com>
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 07/18] RISC-V: nptl: update default pthread-offsets.h
2020-06-03 16:25 ` [PATCH v2 07/18] RISC-V: nptl: update default pthread-offsets.h Alistair Francis via Libc-alpha
@ 2020-07-09 0:14 ` Maciej W. Rozycki via Libc-alpha
2020-07-09 11:47 ` Adhemerval Zanella via Libc-alpha
2020-08-10 17:34 ` Alistair Francis via Libc-alpha
0 siblings, 2 replies; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-09 0:14 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> diff --git a/sysdeps/riscv/nptl/pthread-offsets.h b/sysdeps/riscv/nptl/pthread-offsets.h
> index 31f0587bec..a85c752a1f 100644
> --- a/sysdeps/riscv/nptl/pthread-offsets.h
> +++ b/sysdeps/riscv/nptl/pthread-offsets.h
> @@ -1,3 +1,12 @@
> -#define __PTHREAD_MUTEX_KIND_OFFSET 16
> +#if __WORDSIZE == 64
> +# define __PTHREAD_MUTEX_KIND_OFFSET 16
> +#else
> +# define __PTHREAD_MUTEX_KIND_OFFSET 12
> +#endif
OK, we have:
typedef union
{
struct __pthread_mutex_s __data;
char __size[__SIZEOF_PTHREAD_MUTEX_T];
long int __align;
} pthread_mutex_t;
and then:
struct __pthread_mutex_s
{
int __lock __LOCK_ALIGNMENT;
unsigned int __count;
int __owner;
#if __WORDSIZE == 64
unsigned int __nusers;
#endif
int __kind;
#if __WORDSIZE != 64
unsigned int __nusers;
#endif
which means the `__nusers' and `__kind' members are swapped between 64-bit
and 32-bit hosts. Which I find kind of weird (what for?), but the offset
of `__kind' changes accordingly and the values are correct.
> -#define __PTHREAD_RWLOCK_FLAGS_OFFSET 48
> +
> +#if __WORDSIZE == 64
> +# define __PTHREAD_RWLOCK_FLAGS_OFFSET 48
> +#else
> +# define __PTHREAD_RWLOCK_FLAGS_OFFSET 24
> +#endif
Likewise (as from 09/18):
typedef union
{
struct __pthread_rwlock_arch_t __data;
char __size[__SIZEOF_PTHREAD_RWLOCK_T];
long int __align;
} pthread_rwlock_t;
and:
struct __pthread_rwlock_arch_t
{
unsigned int __readers;
unsigned int __writers;
unsigned int __wrphase_futex;
unsigned int __writers_futex;
unsigned int __pad3;
unsigned int __pad4;
#if __riscv_xlen == 64
int __cur_writer;
int __shared;
unsigned long int __pad1;
unsigned long int __pad2;
unsigned int __flags;
#else
# if __BYTE_ORDER == __BIG_ENDIAN
unsigned char __pad1;
unsigned char __pad2;
unsigned char __shared;
unsigned char __flags;
# else
unsigned char __flags;
unsigned char __shared;
unsigned char __pad1;
unsigned char __pad2;
# endif
int __cur_writer;
#endif
};
so here we have 48 for RV64, 24 for RV32/LE, and 27 for RV/BE, meaning
that your change is wrong. Please fix that.
Also I think this change makes no sense at this point or indeed on its
own as there's been no RV32 support added to <bits/struct_rwlock.h> as at
this commit yet, meaning that the offsets would become inconsistent.
Please fold it into 09/18.
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 08/18] riscv32: Add an architecture ipctypes.h
2020-06-03 16:25 ` [PATCH v2 08/18] riscv32: Add an architecture ipctypes.h Alistair Francis via Libc-alpha
@ 2020-07-09 2:46 ` Maciej W. Rozycki via Libc-alpha
2020-07-09 11:36 ` Adhemerval Zanella via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-09 2:46 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> Add a RISC-V specific ipctypes.h file so we can define __IPC_CMD_TIME64
> as 0.
I gather this has been superseded with upcoming commit 3ca7a464ef33
("sysv: linux: Add 64-bit time_t variant for semctl"), right?
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 08/18] riscv32: Add an architecture ipctypes.h
2020-07-09 2:46 ` Maciej W. Rozycki via Libc-alpha
@ 2020-07-09 11:36 ` Adhemerval Zanella via Libc-alpha
0 siblings, 0 replies; 55+ messages in thread
From: Adhemerval Zanella via Libc-alpha @ 2020-07-09 11:36 UTC (permalink / raw)
To: Maciej W. Rozycki, Alistair Francis; +Cc: libc-alpha
On 08/07/2020 23:46, Maciej W. Rozycki wrote:
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
>> Add a RISC-V specific ipctypes.h file so we can define __IPC_CMD_TIME64
>> as 0.
>
> I gather this has been superseded with upcoming commit 3ca7a464ef33
> ("sysv: linux: Add 64-bit time_t variant for semctl"), right?
>
> Maciej
>
Yes, this is not required anymore. I wrapping some tests before pushed it
upstream.
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 07/18] RISC-V: nptl: update default pthread-offsets.h
2020-07-09 0:14 ` Maciej W. Rozycki via Libc-alpha
@ 2020-07-09 11:47 ` Adhemerval Zanella via Libc-alpha
2020-07-15 19:23 ` Maciej W. Rozycki via Libc-alpha
2020-08-10 17:34 ` Alistair Francis via Libc-alpha
1 sibling, 1 reply; 55+ messages in thread
From: Adhemerval Zanella via Libc-alpha @ 2020-07-09 11:47 UTC (permalink / raw)
To: libc-alpha
On 08/07/2020 21:14, Maciej W. Rozycki via Libc-alpha wrote:
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
>> diff --git a/sysdeps/riscv/nptl/pthread-offsets.h b/sysdeps/riscv/nptl/pthread-offsets.h
>> index 31f0587bec..a85c752a1f 100644
>> --- a/sysdeps/riscv/nptl/pthread-offsets.h
>> +++ b/sysdeps/riscv/nptl/pthread-offsets.h
>> @@ -1,3 +1,12 @@
>> -#define __PTHREAD_MUTEX_KIND_OFFSET 16
>> +#if __WORDSIZE == 64
>> +# define __PTHREAD_MUTEX_KIND_OFFSET 16
>> +#else
>> +# define __PTHREAD_MUTEX_KIND_OFFSET 12
>> +#endif
>
> OK, we have:
>
> typedef union
> {
> struct __pthread_mutex_s __data;
> char __size[__SIZEOF_PTHREAD_MUTEX_T];
> long int __align;
> } pthread_mutex_t;
>
> and then:
>
> struct __pthread_mutex_s
> {
> int __lock __LOCK_ALIGNMENT;
> unsigned int __count;
> int __owner;
> #if __WORDSIZE == 64
> unsigned int __nusers;
> #endif
> int __kind;
> #if __WORDSIZE != 64
> unsigned int __nusers;
> #endif
>
> which means the `__nusers' and `__kind' members are swapped between 64-bit
> and 32-bit hosts. Which I find kind of weird (what for?), but the offset
> of `__kind' changes accordingly and the values are correct.
The default __pthread_mutex_s was modeled from how other architectures
has done so we use the common ground as default (and avoid having to
create even more arch specific struct_mutex.h). I don't exactly why
each port has used this specific layout though.
>
>> -#define __PTHREAD_RWLOCK_FLAGS_OFFSET 48
>> +
>> +#if __WORDSIZE == 64
>> +# define __PTHREAD_RWLOCK_FLAGS_OFFSET 48
>> +#else
>> +# define __PTHREAD_RWLOCK_FLAGS_OFFSET 24
>> +#endif
>
> Likewise (as from 09/18):
>
> typedef union
> {
> struct __pthread_rwlock_arch_t __data;
> char __size[__SIZEOF_PTHREAD_RWLOCK_T];
> long int __align;
> } pthread_rwlock_t;
>
> and:
>
> struct __pthread_rwlock_arch_t
> {
> unsigned int __readers;
> unsigned int __writers;
> unsigned int __wrphase_futex;
> unsigned int __writers_futex;
> unsigned int __pad3;
> unsigned int __pad4;
> #if __riscv_xlen == 64
> int __cur_writer;
> int __shared;
> unsigned long int __pad1;
> unsigned long int __pad2;
> unsigned int __flags;
> #else
> # if __BYTE_ORDER == __BIG_ENDIAN
> unsigned char __pad1;
> unsigned char __pad2;
> unsigned char __shared;
> unsigned char __flags;
> # else
> unsigned char __flags;
> unsigned char __shared;
> unsigned char __pad1;
> unsigned char __pad2;
> # endif
> int __cur_writer;
> #endif
> };
>
> so here we have 48 for RV64, 24 for RV32/LE, and 27 for RV/BE, meaning
> that your change is wrong. Please fix that.
This raises a flags because the wrong __PTHREAD_RWLOCK_FLAGS_OFFSET should
generate a build failure at pthread_rwlock_init.c.
>
> Also I think this change makes no sense at this point or indeed on its
> own as there's been no RV32 support added to <bits/struct_rwlock.h> as at
> this commit yet, meaning that the offsets would become inconsistent.
> Please fold it into 09/18.
>
> Maciej
>
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 05/18] RISC-V: Add path of library directories for the 32-bit
2020-07-08 18:42 ` Maciej W. Rozycki via Libc-alpha
@ 2020-07-09 17:03 ` Alistair Francis via Libc-alpha
0 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-07-09 17:03 UTC (permalink / raw)
To: Maciej W. Rozycki; +Cc: GNU C Library, Alistair Francis
On Wed, Jul 8, 2020 at 11:43 AM Maciej W. Rozycki via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
> > From: Zong Li <zongbox@gmail.com>
> >
> > For the recommand of 64 bit version, we add the libraries path of 32 bit
> > in this patch.
>
> I have troubles comprehending this sentence; also it's repeated literally
> below. Please rewrite an remove the duplication.
>
> > The status of RV32 binaries under RV64 kernels is that the ISA
> > optionally supports having different XLEN for user and supervisor modes,
> > but AFAIK there's no silicon that implements this feature, and the Linux
> > kernel doesn't support it yet.
>
> I'm not sure if this note is relevant here.
>
> > For the recommand of 64 bit version, we add the libraries path of 32 bit
> > in this patch. This includes a fix to avoid an out of bound array check
> > when building with GCC 8.2.
>
> Where exactly is that fix? Shouldn't it be applied as a separate
> preparatory change?
I have re-written the commit message to be clearer.
>
> > diff --git a/sysdeps/unix/sysv/linux/riscv/dl-cache.h b/sysdeps/unix/sysv/linux/riscv/dl-cache.h
> > index c297dfe84f..60fc172edb 100644
> > --- a/sysdeps/unix/sysv/linux/riscv/dl-cache.h
> > +++ b/sysdeps/unix/sysv/linux/riscv/dl-cache.h
> [...]
> > @@ -49,9 +51,16 @@
> > do \
> > { \
> > size_t len = strlen (dir); \
> > - char path[len + 9]; \
> > + char path[len + 10]; \
> > memcpy (path, dir, len + 1); \
> > - if (len >= 12 && ! memcmp(path + len - 12, "/lib64/lp64d", 12)) \
> > + if (len >= 13 && ! memcmp(path + len - 13, "/lib32/ilp32d", 13)) \
> > + { \
> > + len -= 9; \
> > + path[len] = '\0'; \
> > + } \
> > + if (len >= 12 \
> > + && (! memcmp(path + len - 12, "/lib32/ilp32", 12) \
> > + || ! memcmp(path + len - 12, "/lib64/lp64d", 12))) \
>
> Please correct indentation above and use tabs throughout.
Fixed.
>
> I think this code should use `else' clauses. While the truncation of the
> string will cause subsequent `memcmp' calls to fail, they'll still be
> executed, `len' permitting, making this ugly.
>
> Though frankly with the growing number of entries I would rewrite this
> sequence of conditionals entirely, as a loop over a (static) array of the
> strings processed, cutting the insane number of error-prone hardcoded
> string lengths while at it too. It is only ever used in `ldconfig', and
> then invoked there at most twice, so it is not that it is critical enough
> for performance to justify illegibility, and making additional `strlen'
> calls shouldn't be a big deal.
I have converted this to a loop
Alistair
>
> > @@ -64,6 +73,10 @@
> > add_dir (path); \
> > if (len >= 4 && ! memcmp(path + len - 4, "/lib", 4)) \
> > { \
> > + memcpy (path + len, "32/ilp32d", 10); \
> > + add_dir (path); \
> > + memcpy (path + len, "32/ilp32", 9); \
> > + add_dir (path); \
>
> Then the same array of strings could be used here (skipping over the
> "/lib" prefix throughout).
>
> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 02/18] RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit
2020-07-08 0:09 ` Maciej W. Rozycki via Libc-alpha
2020-07-08 17:08 ` Adhemerval Zanella via Libc-alpha
@ 2020-07-09 17:10 ` Alistair Francis via Libc-alpha
1 sibling, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-07-09 17:10 UTC (permalink / raw)
To: Maciej W. Rozycki; +Cc: GNU C Library, Alistair Francis
On Tue, Jul 7, 2020 at 5:09 PM Maciej W. Rozycki via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> Alistair,
>
> I think the change heading is too cryptic and does not express the intent
> of the change well enough. How about:
>
> RISC-V: Use 64-bit-time syscall numbers with the 32-bit port
Fixed.
>
> and then maybe explain in a little more details in the change description.
Done.
>
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
> > diff --git a/sysdeps/unix/sysv/linux/riscv/sysdep.h b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> > index 83e4adf6a2..aa61e8b04d 100644
> > --- a/sysdeps/unix/sysv/linux/riscv/sysdep.h
> > +++ b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> > @@ -116,6 +116,67 @@
> >
> > #include <sysdeps/unix/sysdep.h>
>
> This file is weird, as it includes <sysdeps/unix/sysdep.h> twice, first
> time indirectly via <sysdeps/unix/sysv/linux/generic/sysdep.h> at the top,
> and then second time here. So I think this second inclusion can be
> removed (along with the preceding inclusion of <errno.h>, as it does not
> appear to change anything), and the following conditional moved to the
> top, just after the inclusion of <tls.h>. Oddly <sysdeps/unix/sysdep.h>
> has not been protected against multiple inclusion, but its contents do not
> trigger compilation warnings if processed more than once.
>
> The two #ifdef/#ifndef __ASSEMBLER__ conditionals will then become
> adjacent and can be merged into a single #ifdef/#else one.
>
> This clean-up would probably better be made as a separate preceding
> change.
>
> > +#if __riscv_xlen == 32
>
> I think using __WORDSIZE here would be more consistent with the rest of
> our code (we do use `__riscv_xlen' in a couple of places, but I think they
> ought to be cleaned up).
I have split out these changes into a seperate patch.
>
> > +/* Define the __NR_futex as __NR_futex64 as RV32 doesn't have a
> > + * __NR_futex syscall.
> > + */
> > +# ifndef __NR_futex
> > +# define __NR_futex __NR_futex_time64
> > +# endif
>
> The comment does not match the code.
>
> I think it makes no sense to comment on individual entries as they all
> repeat the same pattern and the same purpose, so an introductory comment
> covering them all at the beginning of the conditional would be better
> instead. I suppose you can then reuse it for the change description too.
>
> > +# ifndef __NR_clock_adjtime
> > +# define __NR_clock_adjtime __NR_clock_adjtime64
> > +# endif
> > +#endif /* __riscv_xlen == 32 */
>
> Since you have multiple inner conditionals separated by an empty line
> each I think it will make sense to have an empty line as well between the
> final one and the closing of the outer conditional. Likewise at the
> beginning.
I have changed this based on what Adhemerval said (the same as the ARC port).
Alistair
>
> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 02/18] RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit
2020-07-08 17:08 ` Adhemerval Zanella via Libc-alpha
@ 2020-07-09 17:14 ` Alistair Francis via Libc-alpha
2020-07-16 0:23 ` Maciej W. Rozycki via Libc-alpha
1 sibling, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-07-09 17:14 UTC (permalink / raw)
To: Adhemerval Zanella; +Cc: GNU C Library
On Wed, Jul 8, 2020 at 10:08 AM Adhemerval Zanella via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
>
>
> On 07/07/2020 21:09, Maciej W. Rozycki via Libc-alpha wrote:
> > Alistair,
> >
> > I think the change heading is too cryptic and does not express the intent
> > of the change well enough. How about:
> >
> > RISC-V: Use 64-bit-time syscall numbers with the 32-bit port
> >
> > and then maybe explain in a little more details in the change description.
> >
> > On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> >
> >> diff --git a/sysdeps/unix/sysv/linux/riscv/sysdep.h b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> >> index 83e4adf6a2..aa61e8b04d 100644
> >> --- a/sysdeps/unix/sysv/linux/riscv/sysdep.h
> >> +++ b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> >> @@ -116,6 +116,67 @@
> >>
> >> #include <sysdeps/unix/sysdep.h>
> >
> > This file is weird, as it includes <sysdeps/unix/sysdep.h> twice, first
> > time indirectly via <sysdeps/unix/sysv/linux/generic/sysdep.h> at the top,
> > and then second time here. So I think this second inclusion can be
> > removed (along with the preceding inclusion of <errno.h>, as it does not
> > appear to change anything), and the following conditional moved to the
> > top, just after the inclusion of <tls.h>. Oddly <sysdeps/unix/sysdep.h>
> > has not been protected against multiple inclusion, but its contents do not
> > trigger compilation warnings if processed more than once.
>
> This multiple inclusion is done internally in some places so the file
> that wants to override a definition would first include the generic
> definitions and then #undef/#define the new one (kernel-features.h
> still does it).
I'm not sure if this is required. I split out the clean up into a
seperate patch and I don't see any regressions.
>
> I think this is kind of confusing, but this is how sysdep.h is currently
> organized. We have been changing bit per bit, but there is a lot of
> place where inclusing is done without guards.
>
> >
> > The two #ifdef/#ifndef __ASSEMBLER__ conditionals will then become
> > adjacent and can be merged into a single #ifdef/#else one.
> >
> > This clean-up would probably better be made as a separate preceding
> > change.
> >
> >> +#if __riscv_xlen == 32
> >
> > I think using __WORDSIZE here would be more consistent with the rest of
> > our code (we do use `__riscv_xlen' in a couple of places, but I think they
> > ought to be cleaned up).
> >
> >> +/* Define the __NR_futex as __NR_futex64 as RV32 doesn't have a
> >> + * __NR_futex syscall.
> >> + */
> >> +# ifndef __NR_futex
> >> +# define __NR_futex __NR_futex_time64
> >> +# endif
> >
> > The comment does not match the code.
> >
> > I think it makes no sense to comment on individual entries as they all
> > repeat the same pattern and the same purpose, so an introductory comment
> > covering them all at the beginning of the conditional would be better
> > instead. I suppose you can then reuse it for the change description too.
> >
> >> +# ifndef __NR_clock_adjtime
> >> +# define __NR_clock_adjtime __NR_clock_adjtime64
> >> +# endif
> >> +#endif /* __riscv_xlen == 32 */
> >
> > Since you have multiple inner conditionals separated by an empty line
> > each I think it will make sense to have an empty line as well between the
> > final one and the closing of the outer conditional. Likewise at the
> > beginning.
>
> I think the correct way to how ARC port is doing which is:
>
> 1. Add any required syscall suppression for 32-bit off_t/time_t on
> fixup-asm-unistd.h so arch-syscall.h will have only the required
> definitions. This might be the case for riscv32 since its kernel
> ABI only supports 64-bit off_t/time_t.
>
> 2. Regenerate arch-syscall.h if it were the case.
>
> 3. On sysdep.h redefine only the syscall where the generic implementation
> still does not have actual 64-bit time_t support:
>
> /* "workarounds" for generic code needing to handle 64-bit time_t. */
>
> /* Fix sysdeps/unix/sysv/linux/clock_getcpuclockid.c. */
> #define __NR_clock_getres __NR_clock_getres_time64
> /* Fix sysdeps/nptl/lowlevellock-futex.h. */
> #define __NR_futex __NR_futex_time64
> [...]
>
> (with proper guards it should require the #ifndef/#define)
>
> 4. Add a comment it is a workaround to handle 64-bit time_t
> and on each #define comment for which implementation it intends to.
Thanks! I have updated this to be similar to the ARC port. I also
paraphrased the points above in the commit message (I hope that is
ok).
Alistair
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 09/18] RISC-V: The ABI implementation for 32-bit
2020-06-03 16:25 ` [PATCH v2 09/18] RISC-V: The ABI implementation for 32-bit Alistair Francis via Libc-alpha
@ 2020-07-09 23:33 ` Maciej W. Rozycki via Libc-alpha
2020-07-10 16:45 ` Alistair Francis via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-09 23:33 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> This patch adds the ABI implementation about 32 bit version. It contains
> the Linux-specific and RISC-V architecture code, I've collected here.
I think this might be a little better worded, and perhaps discuss briefly
at least some choices made.
> diff --git a/sysdeps/riscv/bits/wordsize.h b/sysdeps/riscv/bits/wordsize.h
> index faccc71828..ee430d9036 100644
> --- a/sysdeps/riscv/bits/wordsize.h
> +++ b/sysdeps/riscv/bits/wordsize.h
> @@ -25,5 +25,7 @@
> #if __riscv_xlen == 64
> # define __WORDSIZE_TIME64_COMPAT32 1
> #else
> -# error "rv32i-based targets are not supported"
> +# define __WORDSIZE_TIME64_COMPAT32 0
Hmm, this will be problematic on RV64 hardware running mixed RV64/RV32
userland. This is because files like /var/log/lastlog or /var/log/wtmp
might then be processed by both RV64 and RV32 executables, meaning that
the interpretation will be wrong for executables using the other format.
We made the wrong choice with RV64, anticipating that `__time_t' will be
32-bit on RV32 and will have to live with that, by following whatever the
solution is for ports that hold 32-bit time records in the affected
structures. For now I think that means we have to keep the RV32 time
records 32-bit, i.e. set __WORDSIZE_TIME64_COMPAT32 to 1, consistently
with RV64.
I note there is an extensive discussion on the way to move forward here:
<https://sourceware.org/glibc/wiki/Y2038ProofnessDesign#utmp_types_and_APIs>
We might as well try to implement it right away, so as to avoid being
limited to 32-bit time records here.
At least these records are not supposed to refer to the future, so
nothing will be broken right now, however if someone for whatever reason
wants to keep a single login record for their system past 2038, then
they'll have an issue (a conversion tool would be straightforward though).
NB some existing ports do have __WORDSIZE_TIME64_COMPAT32 set and cleared
for their 64-bit and 32-bit ABIs respectively, as per the note in our
top-level bits/wordsize.h, however this reflects the state as before we
introduced the possibility for `__time_t' to be a 64-bit type with
`__WORDSIZE == 32' ABIs. Given the turn of events I think the note ought
to be updated accordingly; I gather it was missed with commit 07fe93cd9850
("generic/typesizes.h: Add support for 32-bit arches with 64-bit types").
> +# define __WORDSIZE32_SIZE_ULONG 0
> +# define __WORDSIZE32_PTRDIFF_LONG 0
Ack; this matches GCC's <stddef.h>.
> diff --git a/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h b/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h
> index c3c72d6c10..363034c38a 100644
> --- a/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h
> +++ b/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h
> @@ -32,7 +32,15 @@
> # define __SIZEOF_PTHREAD_BARRIER_T 32
> # define __SIZEOF_PTHREAD_BARRIERATTR_T 4
> #else
> -# error "rv32i-based systems are not supported"
> +# define __SIZEOF_PTHREAD_ATTR_T 32
> +# define __SIZEOF_PTHREAD_MUTEX_T 32
> +# define __SIZEOF_PTHREAD_MUTEXATTR_T 4
> +# define __SIZEOF_PTHREAD_COND_T 48
> +# define __SIZEOF_PTHREAD_CONDATTR_T 4
> +# define __SIZEOF_PTHREAD_RWLOCK_T 48
> +# define __SIZEOF_PTHREAD_RWLOCKATTR_T 8
> +# define __SIZEOF_PTHREAD_BARRIER_T 20
> +# define __SIZEOF_PTHREAD_BARRIERATTR_T 4
The values are correct, however some of these macros have the same
expansion regardless of the ABI. For clarity please place them outside
the conditional, as other ports do.
> diff --git a/sysdeps/riscv/nptl/bits/struct_rwlock.h b/sysdeps/riscv/nptl/bits/struct_rwlock.h
> index acfaa75e1b..b478da0132 100644
> --- a/sysdeps/riscv/nptl/bits/struct_rwlock.h
> +++ b/sysdeps/riscv/nptl/bits/struct_rwlock.h
> @@ -32,14 +32,39 @@ struct __pthread_rwlock_arch_t
> unsigned int __writers_futex;
> unsigned int __pad3;
> unsigned int __pad4;
> +#if __riscv_xlen == 64
Same concern about the use of `__riscv_xlen' vs `__WORDSIZE' as before.
> int __cur_writer;
> int __shared;
> unsigned long int __pad1;
> unsigned long int __pad2;
> unsigned int __flags;
> +#else
> +# if __BYTE_ORDER == __BIG_ENDIAN
> + unsigned char __pad1;
> + unsigned char __pad2;
> + unsigned char __shared;
> + unsigned char __flags;
> +# else
> + unsigned char __flags;
> + unsigned char __shared;
> + unsigned char __pad1;
> + unsigned char __pad2;
> +# endif
> + int __cur_writer;
> +#endif
> };
I note with this change the RV32 structure will use the generic layout as
per sysdeps/nptl/bits/struct_rwlock.h, however regrettably RV64 does not.
Would it make sense to instead have the layout the same between RV64 and
RV32, perhaps by redefining `__pad1' and `__pad2' in terms of `unsigned
long long' (which would have alignment implications though) or otherwise?
Is there any benefit from having `__flags' and `__shared' (and the
reserve) grouped within a single 32-bit word? I gather there is, given
the lengths gone to to match the bit lanes across the word regardless of
the endianness. But what is it?
> -#define __PTHREAD_RWLOCK_INITIALIZER(__flags) \
> +#if __riscv_xlen == 64
Again, `__riscv_xlen' vs `__WORDSIZE'.
> +# define __PTHREAD_RWLOCK_INITIALIZER(__flags) \
> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, __flags
> +#else
> +# if __BYTE_ORDER == __BIG_ENDIAN
I would suggest using #elif so as to reduce the number of conditionals
and eliminate nesting.
> diff --git a/sysdeps/riscv/sfp-machine.h b/sysdeps/riscv/sfp-machine.h
> index 08a84fd701..aef8c61a67 100644
> --- a/sysdeps/riscv/sfp-machine.h
> +++ b/sysdeps/riscv/sfp-machine.h
> @@ -22,7 +22,32 @@
>
> #if __riscv_xlen == 32
NB I think it's OK to keep this `__riscv_xlen' reference as soft-fp is
largely an independent subsystem and shared between projects (at least
libgcc and the Linux kernel).
> -# error "rv32i-based targets are not supported"
> +# define _FP_W_TYPE_SIZE 32
> +# define _FP_W_TYPE unsigned long
> +# define _FP_WS_TYPE signed long
> +# define _FP_I_TYPE long
Please align the RHS of these to the same column.
> +# define _FP_DIV_MEAT_S(R, X, Y) _FP_DIV_MEAT_1_udiv_norm (S, R, X, Y)
> +# define _FP_DIV_MEAT_D(R, X, Y) _FP_DIV_MEAT_2_udiv (D, R, X, Y)
> +# define _FP_DIV_MEAT_Q(R, X, Y) _FP_DIV_MEAT_4_udiv (Q, R, X, Y)
> +
> +# define _FP_NANFRAC_S _FP_QNANBIT_S
> +# define _FP_NANFRAC_D _FP_QNANBIT_D, 0
> +# define _FP_NANFRAC_Q _FP_QNANBIT_Q, 0, 0, 0
Likewise. There seems to be an established practice for this header
across architectures to have no space between macro arguments or before
the opening parenthesis. This might help with the alignment.
This looks otherwise OK to me (and virtually the same as libgcc's copy,
except for the extra widening operations defined accordingly for FMA).
> diff --git a/sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h b/sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h
> new file mode 100644
> index 0000000000..7e48f24345
> --- /dev/null
> +++ b/sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h
Shouldn't it go into sysdeps/unix/sysv/linux/riscv/rv32/jmp_buf-macros.h?
> @@ -0,0 +1,53 @@
> +/* jump buffer constants for RISC-V
Please make it a proper sentence.
> + Copyright (C) 2017-2020 Free Software Foundation, Inc.
This is a new file so just 2020 should do.
> +/* Produced by this program:
> +
> + #include <stdio.h>
> + #include <unistd.h>
> + #include <setjmp.h>
> + #include <stddef.h>
> +
> + int main (int argc, char **argv)
> + {
> + printf ("#define JMP_BUF_SIZE %d\n", sizeof (jmp_buf));
> + printf ("#define JMP_BUF_ALIGN %d\n", __alignof__ (jmp_buf));
> + printf ("#define SIGJMP_BUF_SIZE %d\n", sizeof (sigjmp_buf));
> + printf ("#define SIGJMP_BUF_ALIGN %d\n", __alignof__ (sigjmp_buf));
> + printf ("#define MASK_WAS_SAVED_OFFSET %d\n", offsetof (struct __jmp_buf_tag, __mask_was_saved));
> + printf ("#define SAVED_MASK_OFFSET %d\n", offsetof (struct __jmp_buf_tag, __saved_mask));
> + } */
Please reformat according to the GNU coding style.
This file is otherwise OK.
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 01/18] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64
2020-07-07 22:06 ` Maciej W. Rozycki via Libc-alpha
@ 2020-07-10 15:27 ` Alistair Francis via Libc-alpha
0 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-07-10 15:27 UTC (permalink / raw)
To: Maciej W. Rozycki; +Cc: GNU C Library, Alistair Francis
On Tue, Jul 7, 2020 at 3:07 PM Maciej W. Rozycki via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
> > Using the original glibc headers under bits/ let's make small
> > modifications to use 64-bit time_t and off_t for both RV32 and RV64.
> >
> > For the typesizes.h, here are justifications for the changes from the
> > generic version (based on Arnd's very helpful feedback):
> > - All the !__USE_FILE_OFFSET64 types (__off_t, __ino_t, __rlim_t, ...) are
> ^^ ^^
> Please remove extraneous spaces above.
Fixed.
>
> > changed to match the 64-bit replacements.
> >
> > - __time_t is defined to 64 bit, but no __time64_t is added. This makes sense
> > as we don't have the time64 support for other 32-bit architectures yet, and
> > it will be easy to change when that happens.
> >
> > - __suseconds_t is 64-bit. This matches what we use the kernel ABI for the
> > few drivers that are relying on 'struct timeval' input arguments in
> > ioctl, as well as the adjtimex system call. It means that timeval has to
> > be defined without the padding, unlike timespec, which needs padding.
> ^^^
> Likewise.
Fixed.
>
> The commit description will be indented by 4 when presented by `git log',
> etc., so please reformat so as to stay within 74-75 columns to avoid line
> wrapping. Also use two spaces after the end of a sentence throughout.
> Please apply these guidelines across the whole series.
>
> > diff --git a/sysdeps/unix/sysv/linux/riscv/bits/environments.h b/sysdeps/unix/sysv/linux/riscv/bits/environments.h
> > new file mode 100644
> > index 0000000000..c1768af17b
> > --- /dev/null
> > +++ b/sysdeps/unix/sysv/linux/riscv/bits/environments.h
> > @@ -0,0 +1,85 @@
> > +/* Copyright (C) 2019-2020 Free Software Foundation, Inc.
> > + This file is part of the GNU C Library.
> > +
> > + The GNU C Library is free software; you can redistribute it and/or
> > + modify it under the terms of the GNU Lesser General Public
> > + License as published by the Free Software Foundation; either
> > + version 2.1 of the License, or (at your option) any later version.
> > +
> > + The GNU C Library is distributed in the hope that it will be useful,
> > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> > + Lesser General Public License for more details.
> > +
> > + You should have received a copy of the GNU Lesser General Public
> > + License along with the GNU C Library; if not, see
> > + <http://www.gnu.org/licenses/>. */
>
> We've switched to HTTPS, so: <https://www.gnu.org/licenses/>.
Fixed in all patches.
>
> > +/* This header should define the following symbols under the described
> > + situations. A value `1' means that the model is always supported,
> > + `-1' means it is never supported. Undefined means it cannot be
> > + statically decided.
> > +
> > + _POSIX_V7_ILP32_OFF32 32bit int, long, pointers, and off_t type
> > + _POSIX_V7_ILP32_OFFBIG 32bit int, long, and pointers and larger off_t type
> > +
> > + _POSIX_V7_LP64_OFF32 64bit long and pointers and 32bit off_t type
> > + _POSIX_V7_LPBIG_OFFBIG 64bit long and pointers and large off_t type
>
> Stray tab after _POSIX_V7_LP64_OFF32?
Fixed
>
> > +#else /* __WORDSIZE == 32 */
> > +
> > +/* RISC-V requires 64-bit off_t
> > + # undef _POSIX_V7_ILP32_OFF32
> > + # undef _POSIX_V6_ILP32_OFF32
> > + # undef _XBS5_ILP32_OFF32
> > + */
>
> If 64-bit `off_t' is required, then shouldn't these be #define'd to -1
> instead? Otherwise under what circumstances support for these models will
> have to be determined at run time?
I have changed this to -1
>
> +# define _POSIX_V7_ILP32_OFFBIG 1
> +# define _POSIX_V6_ILP32_OFFBIG 1
> +# define _XBS5_ILP32_OFFBIG 1
>
> Please use tabs here like with the rest of the macros.
Fixed.
>
> > +/* CFLAGS. */
> > +#define __ILP32_OFFBIG_CFLAGS "-D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64"
>
> Likewise if 64-bit `off_t' is required, then why do we have to define
> these macros in CFLAGS and what happens if they aren't defined?
I have removed this.
>
> Please use a tab here as well (assuming this macro does have to exist).
>
> > diff --git a/sysdeps/unix/sysv/linux/riscv/bits/time64.h b/sysdeps/unix/sysv/linux/riscv/bits/time64.h
> > new file mode 100644
> > index 0000000000..7ba0dd1cfd
> > --- /dev/null
> > +++ b/sysdeps/unix/sysv/linux/riscv/bits/time64.h
> > @@ -0,0 +1,36 @@
> > +/* bits/time64.h -- underlying types for __time64_t. Generic version.
>
> "RISC-V version." presumably?
Fixed as well as all the comments below.
Thanks for reviewing!
Alistair
>
> > + Copyright (C) 2019-2020 Free Software Foundation, Inc.
> > + This file is part of the GNU C Library.
> > +
> > + The GNU C Library is free software; you can redistribute it and/or
> > + modify it under the terms of the GNU Lesser General Public
> > + License as published by the Free Software Foundation; either
> > + version 2.1 of the License, or (at your option) any later version.
> > +
> > + The GNU C Library is distributed in the hope that it will be useful,
> > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> > + Lesser General Public License for more details.
> > +
> > + You should have received a copy of the GNU Lesser General Public
> > + License along with the GNU C Library; if not, see
> > + <http://www.gnu.org/licenses/>. */
>
> <https://www.gnu.org/licenses/>
>
> > diff --git a/sysdeps/unix/sysv/linux/riscv/bits/timesize.h b/sysdeps/unix/sysv/linux/riscv/bits/timesize.h
> > new file mode 100644
> > index 0000000000..78394b7489
> > --- /dev/null
> > +++ b/sysdeps/unix/sysv/linux/riscv/bits/timesize.h
> > @@ -0,0 +1,22 @@
> > +/* Bit size of the time_t type at glibc build time, general case.
>
> "RISC-V case."?
>
> > + Copyright (C) 2019-2020 Free Software Foundation, Inc.
> > + This file is part of the GNU C Library.
> > +
> > + The GNU C Library is free software; you can redistribute it and/or
> > + modify it under the terms of the GNU Lesser General Public
> > + License as published by the Free Software Foundation; either
> > + version 2.1 of the License, or (at your option) any later version.
> > +
> > + The GNU C Library is distributed in the hope that it will be useful,
> > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> > + Lesser General Public License for more details.
> > +
> > + You should have received a copy of the GNU Lesser General Public
> > + License along with the GNU C Library; if not, see
> > + <http://www.gnu.org/licenses/>. */
>
> <https://www.gnu.org/licenses/>
>
> > diff --git a/sysdeps/unix/sysv/linux/riscv/kernel_stat.h b/sysdeps/unix/sysv/linux/riscv/kernel_stat.h
> > new file mode 100644
> > index 0000000000..25e698b9bf
> > --- /dev/null
> > +++ b/sysdeps/unix/sysv/linux/riscv/kernel_stat.h
> > @@ -0,0 +1,23 @@
> > +/* Copyright (C) 2019-2020 Free Software Foundation, Inc.
> > + This file is part of the GNU C Library.
> > +
> > + The GNU C Library is free software; you can redistribute it and/or
> > + modify it under the terms of the GNU Lesser General Public
> > + License as published by the Free Software Foundation; either
> > + version 2.1 of the License, or (at your option) any later version.
> > +
> > + The GNU C Library is distributed in the hope that it will be useful,
> > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> > + Lesser General Public License for more details.
> > +
> > + You should have received a copy of the GNU Lesser General Public
> > + License along with the GNU C Library. If not, see
> > + <http://www.gnu.org/licenses/>. */
>
> <https://www.gnu.org/licenses/>
>
> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 09/18] RISC-V: The ABI implementation for 32-bit
2020-07-09 23:33 ` Maciej W. Rozycki via Libc-alpha
@ 2020-07-10 16:45 ` Alistair Francis via Libc-alpha
2020-07-11 1:24 ` Maciej W. Rozycki via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-07-10 16:45 UTC (permalink / raw)
To: Maciej W. Rozycki; +Cc: GNU C Library, Alistair Francis
On Thu, Jul 9, 2020 at 4:33 PM Maciej W. Rozycki via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
> > This patch adds the ABI implementation about 32 bit version. It contains
> > the Linux-specific and RISC-V architecture code, I've collected here.
>
> I think this might be a little better worded, and perhaps discuss briefly
> at least some choices made.
>
> > diff --git a/sysdeps/riscv/bits/wordsize.h b/sysdeps/riscv/bits/wordsize.h
> > index faccc71828..ee430d9036 100644
> > --- a/sysdeps/riscv/bits/wordsize.h
> > +++ b/sysdeps/riscv/bits/wordsize.h
> > @@ -25,5 +25,7 @@
> > #if __riscv_xlen == 64
> > # define __WORDSIZE_TIME64_COMPAT32 1
> > #else
> > -# error "rv32i-based targets are not supported"
> > +# define __WORDSIZE_TIME64_COMPAT32 0
>
> Hmm, this will be problematic on RV64 hardware running mixed RV64/RV32
> userland. This is because files like /var/log/lastlog or /var/log/wtmp
> might then be processed by both RV64 and RV32 executables, meaning that
> the interpretation will be wrong for executables using the other format.
>
> We made the wrong choice with RV64, anticipating that `__time_t' will be
> 32-bit on RV32 and will have to live with that, by following whatever the
> solution is for ports that hold 32-bit time records in the affected
> structures. For now I think that means we have to keep the RV32 time
> records 32-bit, i.e. set __WORDSIZE_TIME64_COMPAT32 to 1, consistently
> with RV64.
Aw :(
I have changed this so __WORDSIZE_TIME64_COMPAT32 is 1 for both RV64 and RV32.
>
> I note there is an extensive discussion on the way to move forward here:
> <https://sourceware.org/glibc/wiki/Y2038ProofnessDesign#utmp_types_and_APIs>
> We might as well try to implement it right away, so as to avoid being
> limited to 32-bit time records here.
Is there an advantage of doing it now or can we put this off for the
next release?
>
> At least these records are not supposed to refer to the future, so
> nothing will be broken right now, however if someone for whatever reason
> wants to keep a single login record for their system past 2038, then
> they'll have an issue (a conversion tool would be straightforward though).
>
> NB some existing ports do have __WORDSIZE_TIME64_COMPAT32 set and cleared
> for their 64-bit and 32-bit ABIs respectively, as per the note in our
> top-level bits/wordsize.h, however this reflects the state as before we
> introduced the possibility for `__time_t' to be a 64-bit type with
> `__WORDSIZE == 32' ABIs. Given the turn of events I think the note ought
> to be updated accordingly; I gather it was missed with commit 07fe93cd9850
> ("generic/typesizes.h: Add support for 32-bit arches with 64-bit types").
>
> > +# define __WORDSIZE32_SIZE_ULONG 0
> > +# define __WORDSIZE32_PTRDIFF_LONG 0
>
> Ack; this matches GCC's <stddef.h>.
>
> > diff --git a/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h b/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h
> > index c3c72d6c10..363034c38a 100644
> > --- a/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h
> > +++ b/sysdeps/riscv/nptl/bits/pthreadtypes-arch.h
> > @@ -32,7 +32,15 @@
> > # define __SIZEOF_PTHREAD_BARRIER_T 32
> > # define __SIZEOF_PTHREAD_BARRIERATTR_T 4
> > #else
> > -# error "rv32i-based systems are not supported"
> > +# define __SIZEOF_PTHREAD_ATTR_T 32
> > +# define __SIZEOF_PTHREAD_MUTEX_T 32
> > +# define __SIZEOF_PTHREAD_MUTEXATTR_T 4
> > +# define __SIZEOF_PTHREAD_COND_T 48
> > +# define __SIZEOF_PTHREAD_CONDATTR_T 4
> > +# define __SIZEOF_PTHREAD_RWLOCK_T 48
> > +# define __SIZEOF_PTHREAD_RWLOCKATTR_T 8
> > +# define __SIZEOF_PTHREAD_BARRIER_T 20
> > +# define __SIZEOF_PTHREAD_BARRIERATTR_T 4
>
> The values are correct, however some of these macros have the same
> expansion regardless of the ABI. For clarity please place them outside
> the conditional, as other ports do.
Fixed.
>
> > diff --git a/sysdeps/riscv/nptl/bits/struct_rwlock.h b/sysdeps/riscv/nptl/bits/struct_rwlock.h
> > index acfaa75e1b..b478da0132 100644
> > --- a/sysdeps/riscv/nptl/bits/struct_rwlock.h
> > +++ b/sysdeps/riscv/nptl/bits/struct_rwlock.h
> > @@ -32,14 +32,39 @@ struct __pthread_rwlock_arch_t
> > unsigned int __writers_futex;
> > unsigned int __pad3;
> > unsigned int __pad4;
> > +#if __riscv_xlen == 64
>
> Same concern about the use of `__riscv_xlen' vs `__WORDSIZE' as before.
Fixed.
>
> > int __cur_writer;
> > int __shared;
> > unsigned long int __pad1;
> > unsigned long int __pad2;
> > unsigned int __flags;
> > +#else
> > +# if __BYTE_ORDER == __BIG_ENDIAN
> > + unsigned char __pad1;
> > + unsigned char __pad2;
> > + unsigned char __shared;
> > + unsigned char __flags;
> > +# else
> > + unsigned char __flags;
> > + unsigned char __shared;
> > + unsigned char __pad1;
> > + unsigned char __pad2;
> > +# endif
> > + int __cur_writer;
> > +#endif
> > };
>
> I note with this change the RV32 structure will use the generic layout as
> per sysdeps/nptl/bits/struct_rwlock.h, however regrettably RV64 does not.
> Would it make sense to instead have the layout the same between RV64 and
> RV32, perhaps by redefining `__pad1' and `__pad2' in terms of `unsigned
> long long' (which would have alignment implications though) or otherwise?
I'm not sure which one is better. On one hand it seems better to be
more generic and therefore RV32 should use the generic interface. On
the other hand the more similar they are the better. I'm still leaning
towards we should be generic where possible.
>
> Is there any benefit from having `__flags' and `__shared' (and the
> reserve) grouped within a single 32-bit word? I gather there is, given
> the lengths gone to to match the bit lanes across the word regardless of
> the endianness. But what is it?
I have no idea.
>
> > -#define __PTHREAD_RWLOCK_INITIALIZER(__flags) \
> > +#if __riscv_xlen == 64
>
> Again, `__riscv_xlen' vs `__WORDSIZE'.
Fixed.
>
> > +# define __PTHREAD_RWLOCK_INITIALIZER(__flags) \
> > 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, __flags
> > +#else
> > +# if __BYTE_ORDER == __BIG_ENDIAN
>
> I would suggest using #elif so as to reduce the number of conditionals
> and eliminate nesting.
Fixed.
>
> > diff --git a/sysdeps/riscv/sfp-machine.h b/sysdeps/riscv/sfp-machine.h
> > index 08a84fd701..aef8c61a67 100644
> > --- a/sysdeps/riscv/sfp-machine.h
> > +++ b/sysdeps/riscv/sfp-machine.h
> > @@ -22,7 +22,32 @@
> >
> > #if __riscv_xlen == 32
>
> NB I think it's OK to keep this `__riscv_xlen' reference as soft-fp is
> largely an independent subsystem and shared between projects (at least
> libgcc and the Linux kernel).
No changed here then
>
> > -# error "rv32i-based targets are not supported"
> > +# define _FP_W_TYPE_SIZE 32
> > +# define _FP_W_TYPE unsigned long
> > +# define _FP_WS_TYPE signed long
> > +# define _FP_I_TYPE long
>
> Please align the RHS of these to the same column.
Fixed.
>
> > +# define _FP_DIV_MEAT_S(R, X, Y) _FP_DIV_MEAT_1_udiv_norm (S, R, X, Y)
> > +# define _FP_DIV_MEAT_D(R, X, Y) _FP_DIV_MEAT_2_udiv (D, R, X, Y)
> > +# define _FP_DIV_MEAT_Q(R, X, Y) _FP_DIV_MEAT_4_udiv (Q, R, X, Y)
> > +
> > +# define _FP_NANFRAC_S _FP_QNANBIT_S
> > +# define _FP_NANFRAC_D _FP_QNANBIT_D, 0
> > +# define _FP_NANFRAC_Q _FP_QNANBIT_Q, 0, 0, 0
>
> Likewise. There seems to be an established practice for this header
> across architectures to have no space between macro arguments or before
> the opening parenthesis. This might help with the alignment.
I still think it makes sense to follow the glibc style though, even if
other archs don't.
Let me know if it should be a different way and I'll update it.
>
> This looks otherwise OK to me (and virtually the same as libgcc's copy,
> except for the extra widening operations defined accordingly for FMA).
Great!
>
> > diff --git a/sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h b/sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h
> > new file mode 100644
> > index 0000000000..7e48f24345
> > --- /dev/null
> > +++ b/sysdeps/unix/sysv/linux/riscv/jmp_buf-macros.h
>
> Shouldn't it go into sysdeps/unix/sysv/linux/riscv/rv32/jmp_buf-macros.h?
>
> > @@ -0,0 +1,53 @@
> > +/* jump buffer constants for RISC-V
>
> Please make it a proper sentence.
Fixed
>
> > + Copyright (C) 2017-2020 Free Software Foundation, Inc.
>
> This is a new file so just 2020 should do.
Fixed
>
> > +/* Produced by this program:
> > +
> > + #include <stdio.h>
> > + #include <unistd.h>
> > + #include <setjmp.h>
> > + #include <stddef.h>
> > +
> > + int main (int argc, char **argv)
> > + {
> > + printf ("#define JMP_BUF_SIZE %d\n", sizeof (jmp_buf));
> > + printf ("#define JMP_BUF_ALIGN %d\n", __alignof__ (jmp_buf));
> > + printf ("#define SIGJMP_BUF_SIZE %d\n", sizeof (sigjmp_buf));
> > + printf ("#define SIGJMP_BUF_ALIGN %d\n", __alignof__ (sigjmp_buf));
> > + printf ("#define MASK_WAS_SAVED_OFFSET %d\n", offsetof (struct __jmp_buf_tag, __mask_was_saved));
> > + printf ("#define SAVED_MASK_OFFSET %d\n", offsetof (struct __jmp_buf_tag, __saved_mask));
> > + } */
>
> Please reformat according to the GNU coding style.
Fixed
Alistair
>
> This file is otherwise OK.
>
> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 10/18] RISC-V: Hard float support for 32-bit
2020-06-03 16:25 ` [PATCH v2 10/18] RISC-V: Hard float support " Alistair Francis via Libc-alpha
@ 2020-07-11 0:49 ` Maciej W. Rozycki via Libc-alpha
2020-07-11 15:49 ` Alistair Francis via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-11 0:49 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> This patch contains hardware floating-point support for the RV32IF and
> RV32IFD
Full stop please.
> diff --git a/sysdeps/riscv/rv32/rvd/s_lrint.c b/sysdeps/riscv/rv32/rvd/s_lrint.c
> new file mode 100644
> index 0000000000..df406aacb6
> --- /dev/null
> +++ b/sysdeps/riscv/rv32/rvd/s_lrint.c
> @@ -0,0 +1,31 @@
> +/* lrint(). RISC-V version.
I think this has to mention this is the 32-bit version somehow, like
"32-bit RISC-V" or suchlike ("RV32" might be too cryptic/slang). Feel
free to find a better wording (I'm not particularly happy to start a
sentence with a number).
> + Copyright (C) 2017-2020 Free Software Foundation, Inc.
Again, 2020 only, and likewise throughout. Also I missed one case in
01/18 and may have elsewhere, please double-check the remaining patches.
Otherwise OK as far as code already proposed for this change is
concerned.
What about the other math functions though? We have a lot of optimised
versions in `sysdeps/riscv/rv64/rv{f,d}/', which seem suitable for RV32 as
it stands, such as `s_llrintf.c' or `s_nearbyint.c'. Instead we build
generic `sysdeps/ieee754/{flt-32,dbl-64}/' variants.
Shouldn't we also move the XLEN-agnostic optimised functions to
`sysdeps/riscv/rv{f,d}/' with this change? I think we only need to keep
those that use the `long int' type at the interface in the XLEN-specific
directory.
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 09/18] RISC-V: The ABI implementation for 32-bit
2020-07-10 16:45 ` Alistair Francis via Libc-alpha
@ 2020-07-11 1:24 ` Maciej W. Rozycki via Libc-alpha
2020-08-10 21:29 ` Alistair Francis via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-11 1:24 UTC (permalink / raw)
To: Alistair Francis; +Cc: GNU C Library, Alistair Francis
On Fri, 10 Jul 2020, Alistair Francis wrote:
> > I note there is an extensive discussion on the way to move forward here:
> > <https://sourceware.org/glibc/wiki/Y2038ProofnessDesign#utmp_types_and_APIs>
> > We might as well try to implement it right away, so as to avoid being
> > limited to 32-bit time records here.
>
> Is there an advantage of doing it now or can we put this off for the
> next release?
The change is major enough it'll have to wait for the next development
cycle anyway. It shouldn't matter that much for RV32 glibc deployments
though, given the amount of suitable hardware available.
> > NB some existing ports do have __WORDSIZE_TIME64_COMPAT32 set and cleared
> > for their 64-bit and 32-bit ABIs respectively, as per the note in our
> > top-level bits/wordsize.h, however this reflects the state as before we
> > introduced the possibility for `__time_t' to be a 64-bit type with
> > `__WORDSIZE == 32' ABIs. Given the turn of events I think the note ought
> > to be updated accordingly; I gather it was missed with commit 07fe93cd9850
> > ("generic/typesizes.h: Add support for 32-bit arches with 64-bit types").
Will you be able to look into it?
The context here is before Y2038 changes __WORDSIZE_TIME64_COMPAT32 would
only be clear for 64-bit ABIs with those 64-bit systems that do not have
any 32-bit ABI (compatibility mode) to support, such as the DEC Alpha.
And it would always be clear for 32-bit ABIs, so as to use the proper
`__time_t' type without changing the width of actual data held there in
the structure.
I'm not sure what the story is behind the S/390 port though; perhaps it
does not support ABI coexistence in a single run-time environment.
> > > int __cur_writer;
> > > int __shared;
> > > unsigned long int __pad1;
> > > unsigned long int __pad2;
> > > unsigned int __flags;
> > > +#else
> > > +# if __BYTE_ORDER == __BIG_ENDIAN
> > > + unsigned char __pad1;
> > > + unsigned char __pad2;
> > > + unsigned char __shared;
> > > + unsigned char __flags;
> > > +# else
> > > + unsigned char __flags;
> > > + unsigned char __shared;
> > > + unsigned char __pad1;
> > > + unsigned char __pad2;
> > > +# endif
> > > + int __cur_writer;
> > > +#endif
> > > };
> >
> > I note with this change the RV32 structure will use the generic layout as
> > per sysdeps/nptl/bits/struct_rwlock.h, however regrettably RV64 does not.
> > Would it make sense to instead have the layout the same between RV64 and
> > RV32, perhaps by redefining `__pad1' and `__pad2' in terms of `unsigned
> > long long' (which would have alignment implications though) or otherwise?
>
> I'm not sure which one is better. On one hand it seems better to be
> more generic and therefore RV32 should use the generic interface. On
> the other hand the more similar they are the better. I'm still leaning
> towards we should be generic where possible.
It would be good to get a second opinion here.
> > Is there any benefit from having `__flags' and `__shared' (and the
> > reserve) grouped within a single 32-bit word? I gather there is, given
> > the lengths gone to to match the bit lanes across the word regardless of
> > the endianness. But what is it?
>
> I have no idea.
Especially given this.
> > > +# define _FP_DIV_MEAT_S(R, X, Y) _FP_DIV_MEAT_1_udiv_norm (S, R, X, Y)
> > > +# define _FP_DIV_MEAT_D(R, X, Y) _FP_DIV_MEAT_2_udiv (D, R, X, Y)
> > > +# define _FP_DIV_MEAT_Q(R, X, Y) _FP_DIV_MEAT_4_udiv (Q, R, X, Y)
> > > +
> > > +# define _FP_NANFRAC_S _FP_QNANBIT_S
> > > +# define _FP_NANFRAC_D _FP_QNANBIT_D, 0
> > > +# define _FP_NANFRAC_Q _FP_QNANBIT_Q, 0, 0, 0
> >
> > Likewise. There seems to be an established practice for this header
> > across architectures to have no space between macro arguments or before
> > the opening parenthesis. This might help with the alignment.
>
> I still think it makes sense to follow the glibc style though, even if
> other archs don't.
>
> Let me know if it should be a different way and I'll update it.
There is the issue of the discrepancy compared to the libgcc version, and
while `diff -l' and `patch -l' solve that for manual processing, more
sophisticated tools may not cope and require manual intervention.
Again, I would suggest getting a second opinion.
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 10/18] RISC-V: Hard float support for 32-bit
2020-07-11 0:49 ` Maciej W. Rozycki via Libc-alpha
@ 2020-07-11 15:49 ` Alistair Francis via Libc-alpha
2020-07-11 22:13 ` Maciej W. Rozycki via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-07-11 15:49 UTC (permalink / raw)
To: Maciej W. Rozycki; +Cc: GNU C Library, Alistair Francis
On Fri, Jul 10, 2020 at 5:49 PM Maciej W. Rozycki via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
> > This patch contains hardware floating-point support for the RV32IF and
> > RV32IFD
>
> Full stop please.
Fixed.
>
> > diff --git a/sysdeps/riscv/rv32/rvd/s_lrint.c b/sysdeps/riscv/rv32/rvd/s_lrint.c
> > new file mode 100644
> > index 0000000000..df406aacb6
> > --- /dev/null
> > +++ b/sysdeps/riscv/rv32/rvd/s_lrint.c
> > @@ -0,0 +1,31 @@
> > +/* lrint(). RISC-V version.
>
> I think this has to mention this is the 32-bit version somehow, like
> "32-bit RISC-V" or suchlike ("RV32" might be too cryptic/slang). Feel
> free to find a better wording (I'm not particularly happy to start a
> sentence with a number).
Fixed.
>
> > + Copyright (C) 2017-2020 Free Software Foundation, Inc.
>
> Again, 2020 only, and likewise throughout. Also I missed one case in
> 01/18 and may have elsewhere, please double-check the remaining patches.
Fixed in the whole series.
>
> Otherwise OK as far as code already proposed for this change is
> concerned.
Thanks
>
> What about the other math functions though? We have a lot of optimised
> versions in `sysdeps/riscv/rv64/rv{f,d}/', which seem suitable for RV32 as
> it stands, such as `s_llrintf.c' or `s_nearbyint.c'. Instead we build
> generic `sysdeps/ieee754/{flt-32,dbl-64}/' variants.
I haven't looked into the others, but I think you are right and this
can be improved.
>
> Shouldn't we also move the XLEN-agnostic optimised functions to
> `sysdeps/riscv/rv{f,d}/' with this change? I think we only need to keep
> those that use the `long int' type at the interface in the XLEN-specific
> directory.
They are all xlen dependent though. All of the 64-bit ones use
fcvt.l.d (or something similar) which doesn't exist on RV32.
Although the functions end up being the same, the actual assembly
instruction called is different. We can look at consolidating them
into a single file and do a xlen/__WORDSIZE macro check.
At this stage in glibc development I don't really want to change the
floating point helpers. Can we leave this as is and then for the next
release I can consolidate all of these into single files that do
xlen/__WORDSIZE #ifdefs? That way we will just have a single file (for
each operation) for RISC-V that will call a different assembly
instruction based on xlen or __WORDSIZE.
Alistair
>
> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 10/18] RISC-V: Hard float support for 32-bit
2020-07-11 15:49 ` Alistair Francis via Libc-alpha
@ 2020-07-11 22:13 ` Maciej W. Rozycki via Libc-alpha
2020-07-12 15:34 ` Alistair Francis via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-11 22:13 UTC (permalink / raw)
To: Alistair Francis; +Cc: GNU C Library, Alistair Francis
On Sat, 11 Jul 2020, Alistair Francis wrote:
> > Shouldn't we also move the XLEN-agnostic optimised functions to
> > `sysdeps/riscv/rv{f,d}/' with this change? I think we only need to keep
> > those that use the `long int' type at the interface in the XLEN-specific
> > directory.
>
> They are all xlen dependent though. All of the 64-bit ones use
> fcvt.l.d (or something similar) which doesn't exist on RV32.
Umm, right, I missed this detail, sorry for the confusion.
> Although the functions end up being the same, the actual assembly
> instruction called is different. We can look at consolidating them
> into a single file and do a xlen/__WORDSIZE macro check.
I don't think there is a simple alternative available for RV32 that would
be worth keeping together with the RV64 variant. Did I miss anything?
What instruction(s) do you have in mind?
> At this stage in glibc development I don't really want to change the
> floating point helpers. Can we leave this as is and then for the next
> release I can consolidate all of these into single files that do
> xlen/__WORDSIZE #ifdefs? That way we will just have a single file (for
> each operation) for RISC-V that will call a different assembly
> instruction based on xlen or __WORDSIZE.
Sure, in these circumstances this change is fine with me then.
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 10/18] RISC-V: Hard float support for 32-bit
2020-07-11 22:13 ` Maciej W. Rozycki via Libc-alpha
@ 2020-07-12 15:34 ` Alistair Francis via Libc-alpha
2020-07-12 22:10 ` Maciej W. Rozycki via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-07-12 15:34 UTC (permalink / raw)
To: Maciej W. Rozycki; +Cc: GNU C Library, Alistair Francis
On Sat, Jul 11, 2020 at 3:14 PM Maciej W. Rozycki <macro@wdc.com> wrote:
>
> On Sat, 11 Jul 2020, Alistair Francis wrote:
>
> > > Shouldn't we also move the XLEN-agnostic optimised functions to
> > > `sysdeps/riscv/rv{f,d}/' with this change? I think we only need to keep
> > > those that use the `long int' type at the interface in the XLEN-specific
> > > directory.
> >
> > They are all xlen dependent though. All of the 64-bit ones use
> > fcvt.l.d (or something similar) which doesn't exist on RV32.
>
> Umm, right, I missed this detail, sorry for the confusion.
>
> > Although the functions end up being the same, the actual assembly
> > instruction called is different. We can look at consolidating them
> > into a single file and do a xlen/__WORDSIZE macro check.
>
> I don't think there is a simple alternative available for RV32 that would
> be worth keeping together with the RV64 variant. Did I miss anything?
> What instruction(s) do you have in mind?
Something like this for the generic RISC-V lround:
long int
__lroundf (float x)
{
#if __WORDSIZE == 64
int64_t res;
asm ("fcvt.l.s %0, %1, rmm" : "=r" (res) : "f" (x));
#else
int32_t res;
asm ("fcvt.w.s %0, %1, rmm" : "=r" (res) : "f" (x));
#endif
return res;
}
I'm not sure if it's clearer, but for some of the more complex
functions (roundeven for example) it might be easier.
It also means if there is a bug fixed in one it'll end up fixed for both.
>
> > At this stage in glibc development I don't really want to change the
> > floating point helpers. Can we leave this as is and then for the next
> > release I can consolidate all of these into single files that do
> > xlen/__WORDSIZE #ifdefs? That way we will just have a single file (for
> > each operation) for RISC-V that will call a different assembly
> > instruction based on xlen or __WORDSIZE.
>
> Sure, in these circumstances this change is fine with me then.
Great! Thanks
Alistair
>
> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 11/18] RISC-V: Add ABI lists
2020-06-03 16:25 ` [PATCH v2 11/18] RISC-V: Add ABI lists Alistair Francis via Libc-alpha
@ 2020-07-12 20:54 ` Maciej W. Rozycki via Libc-alpha
2020-07-13 16:14 ` Alistair Francis via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-12 20:54 UTC (permalink / raw)
To: Alistair Francis; +Cc: libc-alpha
Alistair,
Maybe: "RISC-V: Add 32-bit ABI lists"?
On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
> Use the check-api and update-abi to generate the abilist for rv32.
What does `check-api' refer to? I can't see such a Makefile target or a
script. Also something's missing or extraneous here WRT first "the".
> diff --git a/sysdeps/unix/sysv/linux/riscv/c++-types.data b/sysdeps/unix/sysv/linux/riscv/c++-types.data
> new file mode 100644
> index 0000000000..0fb0143ae7
> --- /dev/null
> +++ b/sysdeps/unix/sysv/linux/riscv/c++-types.data
This would better be `sysdeps/unix/sysv/linux/riscv/rv32/c++-types.data',
I think. We have no "default" architecture. Likewise the other files.
Let me know if you disagree.
Otherwise these are all generated files, and I have verified the contents
pass our checks, so it all looks good to me.
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 10/18] RISC-V: Hard float support for 32-bit
2020-07-12 15:34 ` Alistair Francis via Libc-alpha
@ 2020-07-12 22:10 ` Maciej W. Rozycki via Libc-alpha
2020-08-27 18:36 ` Alistair Francis via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-12 22:10 UTC (permalink / raw)
To: Alistair Francis; +Cc: GNU C Library, Alistair Francis
On Sun, 12 Jul 2020, Alistair Francis wrote:
> > I don't think there is a simple alternative available for RV32 that would
> > be worth keeping together with the RV64 variant. Did I miss anything?
> > What instruction(s) do you have in mind?
>
> Something like this for the generic RISC-V lround:
>
> long int
> __lroundf (float x)
> {
> #if __WORDSIZE == 64
> int64_t res;
> asm ("fcvt.l.s %0, %1, rmm" : "=r" (res) : "f" (x));
> #else
> int32_t res;
> asm ("fcvt.w.s %0, %1, rmm" : "=r" (res) : "f" (x));
> #endif
> return res;
> }
>
> I'm not sure if it's clearer, but for some of the more complex
> functions (roundeven for example) it might be easier.
>
> It also means if there is a bug fixed in one it'll end up fixed for both.
Ah, you mean the `float' to `long int' conversion functions, necessarily
ABI-specific due to the changing width of the latter data type. Well, I
meant the operations involving FCVT.L.D/FCVT.D.L. I can see no RISC-V
solution for them that would surpass the generic implementation.
As far as your example above is concerned if we decided to merge the
files at all, I would reduce it to:
#if __WORDSIZE == 64
# define OP "fcvt.l.s"
#elif __WORDSIZE == 32
# define OP "fcvt.w.s"
#else
# error Unsupported
#endif
long int
__lroundf (float x)
{
long int res;
asm (OP "\t%0, %1, rmm" : "=r" (res) : "f" (x));
return res;
}
or suchlike (I'm not sure if there's any gain here from `res' having an
explicit-width data type). Likewise with the rest.
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 11/18] RISC-V: Add ABI lists
2020-07-12 20:54 ` Maciej W. Rozycki via Libc-alpha
@ 2020-07-13 16:14 ` Alistair Francis via Libc-alpha
0 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-07-13 16:14 UTC (permalink / raw)
To: Maciej W. Rozycki; +Cc: GNU C Library, Alistair Francis
On Sun, Jul 12, 2020 at 1:57 PM Maciej W. Rozycki via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> Alistair,
>
> Maybe: "RISC-V: Add 32-bit ABI lists"?
Fixed.
>
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
> > Use the check-api and update-abi to generate the abilist for rv32.
>
> What does `check-api' refer to? I can't see such a Makefile target or a
> script. Also something's missing or extraneous here WRT first "the".
I have corrected this.
>
> > diff --git a/sysdeps/unix/sysv/linux/riscv/c++-types.data b/sysdeps/unix/sysv/linux/riscv/c++-types.data
> > new file mode 100644
> > index 0000000000..0fb0143ae7
> > --- /dev/null
> > +++ b/sysdeps/unix/sysv/linux/riscv/c++-types.data
>
> This would better be `sysdeps/unix/sysv/linux/riscv/rv32/c++-types.data',
> I think. We have no "default" architecture. Likewise the other files.
> Let me know if you disagree.
Agreed! I have moved it into the rv32 directory.
>
> Otherwise these are all generated files, and I have verified the contents
> pass our checks, so it all looks good to me.
Thanks!
Alistair
>
> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 07/18] RISC-V: nptl: update default pthread-offsets.h
2020-07-09 11:47 ` Adhemerval Zanella via Libc-alpha
@ 2020-07-15 19:23 ` Maciej W. Rozycki via Libc-alpha
0 siblings, 0 replies; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-15 19:23 UTC (permalink / raw)
To: Adhemerval Zanella; +Cc: libc-alpha, Alistair Francis
[Added cc's back.]
On Thu, 9 Jul 2020, Adhemerval Zanella via Libc-alpha wrote:
> > so here we have 48 for RV64, 24 for RV32/LE, and 27 for RV/BE, meaning
> > that your change is wrong. Please fix that.
>
> This raises a flags because the wrong __PTHREAD_RWLOCK_FLAGS_OFFSET should
> generate a build failure at pthread_rwlock_init.c.
Possibly no one has ever built this code BE; I haven't.
Now that I looked into it again I think GCC currently has no option for
the big endianness with the RISC-V target, so maybe we can leave it as it
is under your observation that a build failure is expected to happen here
if someone tries it once/if we have such support in the compiler.
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 02/18] RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit
2020-07-08 17:08 ` Adhemerval Zanella via Libc-alpha
2020-07-09 17:14 ` Alistair Francis via Libc-alpha
@ 2020-07-16 0:23 ` Maciej W. Rozycki via Libc-alpha
1 sibling, 0 replies; 55+ messages in thread
From: Maciej W. Rozycki via Libc-alpha @ 2020-07-16 0:23 UTC (permalink / raw)
To: Adhemerval Zanella; +Cc: libc-alpha, Alistair Francis
[Added cc's back.]
On Wed, 8 Jul 2020, Adhemerval Zanella via Libc-alpha wrote:
> >> diff --git a/sysdeps/unix/sysv/linux/riscv/sysdep.h b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> >> index 83e4adf6a2..aa61e8b04d 100644
> >> --- a/sysdeps/unix/sysv/linux/riscv/sysdep.h
> >> +++ b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> >> @@ -116,6 +116,67 @@
> >>
> >> #include <sysdeps/unix/sysdep.h>
> >
> > This file is weird, as it includes <sysdeps/unix/sysdep.h> twice, first
> > time indirectly via <sysdeps/unix/sysv/linux/generic/sysdep.h> at the top,
> > and then second time here. So I think this second inclusion can be
> > removed (along with the preceding inclusion of <errno.h>, as it does not
> > appear to change anything), and the following conditional moved to the
> > top, just after the inclusion of <tls.h>. Oddly <sysdeps/unix/sysdep.h>
> > has not been protected against multiple inclusion, but its contents do not
> > trigger compilation warnings if processed more than once.
>
> This multiple inclusion is done internally in some places so the file
> that wants to override a definition would first include the generic
> definitions and then #undef/#define the new one (kernel-features.h
> still does it).
>
> I think this is kind of confusing, but this is how sysdep.h is currently
> organized. We have been changing bit per bit, but there is a lot of
> place where inclusing is done without guards.
Fair enough, however the RISC-V port does not appear to make use of it as
the removal of the second inclusion does not change anything (I actually
double-checked).
Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 07/18] RISC-V: nptl: update default pthread-offsets.h
2020-07-09 0:14 ` Maciej W. Rozycki via Libc-alpha
2020-07-09 11:47 ` Adhemerval Zanella via Libc-alpha
@ 2020-08-10 17:34 ` Alistair Francis via Libc-alpha
1 sibling, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-08-10 17:34 UTC (permalink / raw)
To: Maciej W. Rozycki; +Cc: GNU C Library, Alistair Francis
On Wed, Jul 8, 2020 at 5:14 PM Maciej W. Rozycki via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> On Wed, 3 Jun 2020, Alistair Francis via Libc-alpha wrote:
>
> > diff --git a/sysdeps/riscv/nptl/pthread-offsets.h b/sysdeps/riscv/nptl/pthread-offsets.h
> > index 31f0587bec..a85c752a1f 100644
> > --- a/sysdeps/riscv/nptl/pthread-offsets.h
> > +++ b/sysdeps/riscv/nptl/pthread-offsets.h
> > @@ -1,3 +1,12 @@
> > -#define __PTHREAD_MUTEX_KIND_OFFSET 16
> > +#if __WORDSIZE == 64
> > +# define __PTHREAD_MUTEX_KIND_OFFSET 16
> > +#else
> > +# define __PTHREAD_MUTEX_KIND_OFFSET 12
> > +#endif
>
> OK, we have:
>
> typedef union
> {
> struct __pthread_mutex_s __data;
> char __size[__SIZEOF_PTHREAD_MUTEX_T];
> long int __align;
> } pthread_mutex_t;
>
> and then:
>
> struct __pthread_mutex_s
> {
> int __lock __LOCK_ALIGNMENT;
> unsigned int __count;
> int __owner;
> #if __WORDSIZE == 64
> unsigned int __nusers;
> #endif
> int __kind;
> #if __WORDSIZE != 64
> unsigned int __nusers;
> #endif
>
> which means the `__nusers' and `__kind' members are swapped between 64-bit
> and 32-bit hosts. Which I find kind of weird (what for?), but the offset
> of `__kind' changes accordingly and the values are correct.
Great
>
> > -#define __PTHREAD_RWLOCK_FLAGS_OFFSET 48
> > +
> > +#if __WORDSIZE == 64
> > +# define __PTHREAD_RWLOCK_FLAGS_OFFSET 48
> > +#else
> > +# define __PTHREAD_RWLOCK_FLAGS_OFFSET 24
> > +#endif
>
> Likewise (as from 09/18):
>
> typedef union
> {
> struct __pthread_rwlock_arch_t __data;
> char __size[__SIZEOF_PTHREAD_RWLOCK_T];
> long int __align;
> } pthread_rwlock_t;
>
> and:
>
> struct __pthread_rwlock_arch_t
> {
> unsigned int __readers;
> unsigned int __writers;
> unsigned int __wrphase_futex;
> unsigned int __writers_futex;
> unsigned int __pad3;
> unsigned int __pad4;
> #if __riscv_xlen == 64
> int __cur_writer;
> int __shared;
> unsigned long int __pad1;
> unsigned long int __pad2;
> unsigned int __flags;
> #else
> # if __BYTE_ORDER == __BIG_ENDIAN
> unsigned char __pad1;
> unsigned char __pad2;
> unsigned char __shared;
> unsigned char __flags;
> # else
> unsigned char __flags;
> unsigned char __shared;
> unsigned char __pad1;
> unsigned char __pad2;
> # endif
> int __cur_writer;
> #endif
> };
>
> so here we have 48 for RV64, 24 for RV32/LE, and 27 for RV/BE, meaning
> that your change is wrong. Please fix that.
I have fixed the BE offset, although I don't think we actually support BE.
>
> Also I think this change makes no sense at this point or indeed on its
> own as there's been no RV32 support added to <bits/struct_rwlock.h> as at
> this commit yet, meaning that the offsets would become inconsistent.
> Please fold it into 09/18.
Done.
Alistair
>
> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 09/18] RISC-V: The ABI implementation for 32-bit
2020-07-11 1:24 ` Maciej W. Rozycki via Libc-alpha
@ 2020-08-10 21:29 ` Alistair Francis via Libc-alpha
2020-08-27 19:43 ` Adhemerval Zanella via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-08-10 21:29 UTC (permalink / raw)
To: Maciej W. Rozycki; +Cc: GNU C Library, Alistair Francis
On Fri, Jul 10, 2020 at 6:24 PM Maciej W. Rozycki <macro@wdc.com> wrote:
>
> On Fri, 10 Jul 2020, Alistair Francis wrote:
>
> > > I note there is an extensive discussion on the way to move forward here:
> > > <https://sourceware.org/glibc/wiki/Y2038ProofnessDesign#utmp_types_and_APIs>
> > > We might as well try to implement it right away, so as to avoid being
> > > limited to 32-bit time records here.
> >
> > Is there an advantage of doing it now or can we put this off for the
> > next release?
>
> The change is major enough it'll have to wait for the next development
> cycle anyway. It shouldn't matter that much for RV32 glibc deployments
> though, given the amount of suitable hardware available.
>
> > > NB some existing ports do have __WORDSIZE_TIME64_COMPAT32 set and cleared
> > > for their 64-bit and 32-bit ABIs respectively, as per the note in our
> > > top-level bits/wordsize.h, however this reflects the state as before we
> > > introduced the possibility for `__time_t' to be a 64-bit type with
> > > `__WORDSIZE == 32' ABIs. Given the turn of events I think the note ought
> > > to be updated accordingly; I gather it was missed with commit 07fe93cd9850
> > > ("generic/typesizes.h: Add support for 32-bit arches with 64-bit types").
>
> Will you be able to look into it?
>
> The context here is before Y2038 changes __WORDSIZE_TIME64_COMPAT32 would
> only be clear for 64-bit ABIs with those 64-bit systems that do not have
> any 32-bit ABI (compatibility mode) to support, such as the DEC Alpha.
> And it would always be clear for 32-bit ABIs, so as to use the proper
> `__time_t' type without changing the width of actual data held there in
> the structure.
>
> I'm not sure what the story is behind the S/390 port though; perhaps it
> does not support ABI coexistence in a single run-time environment.
I was just about to start this but Adhemerval has very kindly sent a
patch series already. I'm going to leave this as is (defining it as 1
for RV32 and RV64).
>
> > > > int __cur_writer;
> > > > int __shared;
> > > > unsigned long int __pad1;
> > > > unsigned long int __pad2;
> > > > unsigned int __flags;
> > > > +#else
> > > > +# if __BYTE_ORDER == __BIG_ENDIAN
> > > > + unsigned char __pad1;
> > > > + unsigned char __pad2;
> > > > + unsigned char __shared;
> > > > + unsigned char __flags;
> > > > +# else
> > > > + unsigned char __flags;
> > > > + unsigned char __shared;
> > > > + unsigned char __pad1;
> > > > + unsigned char __pad2;
> > > > +# endif
> > > > + int __cur_writer;
> > > > +#endif
> > > > };
> > >
> > > I note with this change the RV32 structure will use the generic layout as
> > > per sysdeps/nptl/bits/struct_rwlock.h, however regrettably RV64 does not.
> > > Would it make sense to instead have the layout the same between RV64 and
> > > RV32, perhaps by redefining `__pad1' and `__pad2' in terms of `unsigned
> > > long long' (which would have alignment implications though) or otherwise?
> >
> > I'm not sure which one is better. On one hand it seems better to be
> > more generic and therefore RV32 should use the generic interface. On
> > the other hand the more similar they are the better. I'm still leaning
> > towards we should be generic where possible.
>
> It would be good to get a second opinion here.
The RV64 __pthread_rwlock_arch_t is exactly the same as the AArch64
version. My guess is that it was just copied to be the same.
The generic version also has this comment
/* Generic struct for both POSIX read-write lock. New ports are expected
to use the default layout, however archictetures can redefine it to add
arch-specific extensions (such as lock-elision). The struct have a size
of 32 bytes on both LP32 and LP64 architectures. */
As we aren't adding any arch-specific extensions I think we should
keep the generic one. It's unfortunate we didn't do that for RV64, but
too late now.
The only downside I see in using the generic version is what happens
if the value of _shared or flags overflows a char. There are currently
only 4 flags though, so I think it's ok.
>
> > > Is there any benefit from having `__flags' and `__shared' (and the
> > > reserve) grouped within a single 32-bit word? I gather there is, given
> > > the lengths gone to to match the bit lanes across the word regardless of
> > > the endianness. But what is it?
> >
> > I have no idea.
>
> Especially given this.
>
> > > > +# define _FP_DIV_MEAT_S(R, X, Y) _FP_DIV_MEAT_1_udiv_norm (S, R, X, Y)
> > > > +# define _FP_DIV_MEAT_D(R, X, Y) _FP_DIV_MEAT_2_udiv (D, R, X, Y)
> > > > +# define _FP_DIV_MEAT_Q(R, X, Y) _FP_DIV_MEAT_4_udiv (Q, R, X, Y)
> > > > +
> > > > +# define _FP_NANFRAC_S _FP_QNANBIT_S
> > > > +# define _FP_NANFRAC_D _FP_QNANBIT_D, 0
> > > > +# define _FP_NANFRAC_Q _FP_QNANBIT_Q, 0, 0, 0
> > >
> > > Likewise. There seems to be an established practice for this header
> > > across architectures to have no space between macro arguments or before
> > > the opening parenthesis. This might help with the alignment.
> >
> > I still think it makes sense to follow the glibc style though, even if
> > other archs don't.
> >
> > Let me know if it should be a different way and I'll update it.
>
> There is the issue of the discrepancy compared to the libgcc version, and
> while `diff -l' and `patch -l' solve that for manual processing, more
> sophisticated tools may not cope and require manual intervention.
>
> Again, I would suggest getting a second opinion.
The current option matches the rest of the file, so I think it makes
sense to leave as is.
If this is a problem we can just change the spacing for the entire
file in a future patch.
Alistair
>
> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 10/18] RISC-V: Hard float support for 32-bit
2020-07-12 22:10 ` Maciej W. Rozycki via Libc-alpha
@ 2020-08-27 18:36 ` Alistair Francis via Libc-alpha
0 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-08-27 18:36 UTC (permalink / raw)
To: Maciej W. Rozycki; +Cc: GNU C Library, Alistair Francis
On Sun, Jul 12, 2020 at 3:10 PM Maciej W. Rozycki <macro@wdc.com> wrote:
>
> On Sun, 12 Jul 2020, Alistair Francis wrote:
>
> > > I don't think there is a simple alternative available for RV32 that would
> > > be worth keeping together with the RV64 variant. Did I miss anything?
> > > What instruction(s) do you have in mind?
> >
> > Something like this for the generic RISC-V lround:
> >
> > long int
> > __lroundf (float x)
> > {
> > #if __WORDSIZE == 64
> > int64_t res;
> > asm ("fcvt.l.s %0, %1, rmm" : "=r" (res) : "f" (x));
> > #else
> > int32_t res;
> > asm ("fcvt.w.s %0, %1, rmm" : "=r" (res) : "f" (x));
> > #endif
> > return res;
> > }
> >
> > I'm not sure if it's clearer, but for some of the more complex
> > functions (roundeven for example) it might be easier.
> >
> > It also means if there is a bug fixed in one it'll end up fixed for both.
>
> Ah, you mean the `float' to `long int' conversion functions, necessarily
> ABI-specific due to the changing width of the latter data type. Well, I
> meant the operations involving FCVT.L.D/FCVT.D.L. I can see no RISC-V
> solution for them that would surpass the generic implementation.
>
> As far as your example above is concerned if we decided to merge the
> files at all, I would reduce it to:
>
> #if __WORDSIZE == 64
> # define OP "fcvt.l.s"
> #elif __WORDSIZE == 32
> # define OP "fcvt.w.s"
> #else
> # error Unsupported
> #endif
>
> long int
> __lroundf (float x)
> {
> long int res;
> asm (OP "\t%0, %1, rmm" : "=r" (res) : "f" (x));
> return res;
> }
>
> or suchlike (I'm not sure if there's any gain here from `res' having an
> explicit-width data type). Likewise with the rest.
Sounds good. Now that RV32 support is in I have some patches for this,
I'll send them out soon.
Alistair
>
> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 09/18] RISC-V: The ABI implementation for 32-bit
2020-08-10 21:29 ` Alistair Francis via Libc-alpha
@ 2020-08-27 19:43 ` Adhemerval Zanella via Libc-alpha
2020-09-25 23:03 ` Alistair Francis via Libc-alpha
0 siblings, 1 reply; 55+ messages in thread
From: Adhemerval Zanella via Libc-alpha @ 2020-08-27 19:43 UTC (permalink / raw)
To: libc-alpha
On 10/08/2020 18:29, Alistair Francis via Libc-alpha wrote:
> On Fri, Jul 10, 2020 at 6:24 PM Maciej W. Rozycki <macro@wdc.com> wrote:
>>
>> On Fri, 10 Jul 2020, Alistair Francis wrote:
>>
>>>> I note there is an extensive discussion on the way to move forward here:
>>>> <https://sourceware.org/glibc/wiki/Y2038ProofnessDesign#utmp_types_and_APIs>
>>>> We might as well try to implement it right away, so as to avoid being
>>>> limited to 32-bit time records here.
>>>
>>> Is there an advantage of doing it now or can we put this off for the
>>> next release?
>>
>> The change is major enough it'll have to wait for the next development
>> cycle anyway. It shouldn't matter that much for RV32 glibc deployments
>> though, given the amount of suitable hardware available.
>>
>>>> NB some existing ports do have __WORDSIZE_TIME64_COMPAT32 set and cleared
>>>> for their 64-bit and 32-bit ABIs respectively, as per the note in our
>>>> top-level bits/wordsize.h, however this reflects the state as before we
>>>> introduced the possibility for `__time_t' to be a 64-bit type with
>>>> `__WORDSIZE == 32' ABIs. Given the turn of events I think the note ought
>>>> to be updated accordingly; I gather it was missed with commit 07fe93cd9850
>>>> ("generic/typesizes.h: Add support for 32-bit arches with 64-bit types").
>>
>> Will you be able to look into it?
>>
>> The context here is before Y2038 changes __WORDSIZE_TIME64_COMPAT32 would
>> only be clear for 64-bit ABIs with those 64-bit systems that do not have
>> any 32-bit ABI (compatibility mode) to support, such as the DEC Alpha.
>> And it would always be clear for 32-bit ABIs, so as to use the proper
>> `__time_t' type without changing the width of actual data held there in
>> the structure.
>>
>> I'm not sure what the story is behind the S/390 port though; perhaps it
>> does not support ABI coexistence in a single run-time environment.
>
> I was just about to start this but Adhemerval has very kindly sent a
> patch series already. I'm going to leave this as is (defining it as 1
> for RV32 and RV64).
Don't forget to help on review ;).
>
>
>>
>>>>> int __cur_writer;
>>>>> int __shared;
>>>>> unsigned long int __pad1;
>>>>> unsigned long int __pad2;
>>>>> unsigned int __flags;
>>>>> +#else
>>>>> +# if __BYTE_ORDER == __BIG_ENDIAN
>>>>> + unsigned char __pad1;
>>>>> + unsigned char __pad2;
>>>>> + unsigned char __shared;
>>>>> + unsigned char __flags;
>>>>> +# else
>>>>> + unsigned char __flags;
>>>>> + unsigned char __shared;
>>>>> + unsigned char __pad1;
>>>>> + unsigned char __pad2;
>>>>> +# endif
>>>>> + int __cur_writer;
>>>>> +#endif
>>>>> };
>>>>
>>>> I note with this change the RV32 structure will use the generic layout as
>>>> per sysdeps/nptl/bits/struct_rwlock.h, however regrettably RV64 does not.
>>>> Would it make sense to instead have the layout the same between RV64 and
>>>> RV32, perhaps by redefining `__pad1' and `__pad2' in terms of `unsigned
>>>> long long' (which would have alignment implications though) or otherwise?
>>>
>>> I'm not sure which one is better. On one hand it seems better to be
>>> more generic and therefore RV32 should use the generic interface. On
>>> the other hand the more similar they are the better. I'm still leaning
>>> towards we should be generic where possible.
>>
>> It would be good to get a second opinion here.
>
> The RV64 __pthread_rwlock_arch_t is exactly the same as the AArch64
> version. My guess is that it was just copied to be the same.
>
> The generic version also has this comment
>
> /* Generic struct for both POSIX read-write lock. New ports are expected
> to use the default layout, however archictetures can redefine it to add
> arch-specific extensions (such as lock-elision). The struct have a size
> of 32 bytes on both LP32 and LP64 architectures. */
>
> As we aren't adding any arch-specific extensions I think we should
> keep the generic one. It's unfortunate we didn't do that for RV64, but
> too late now.
>
> The only downside I see in using the generic version is what happens
> if the value of _shared or flags overflows a char. There are currently
> only 4 flags though, so I think it's ok.
I come up with the generic interface by using the already one in place for
mostly 64-bit architecture and the internal layout is mainly to avoid duplicate
the internal type on multiple architectures (so we can use the generic
implementation).
The only usage I can think of where same structure for RV64 and RV32 may
pose a role is to use a shared rwlock between a 32-bit and a 64-bit process
(pthread_rwlockattr_setpshared with PTHREAD_PROCESS_SHARED). But afaik this
is not supported on other architectures.
>
>>
>>>> Is there any benefit from having `__flags' and `__shared' (and the
>>>> reserve) grouped within a single 32-bit word? I gather there is, given
>>>> the lengths gone to to match the bit lanes across the word regardless of
>>>> the endianness. But what is it?
>>>
>>> I have no idea.
>>
>> Especially given this.
The internal member that are most concurrently accessed by the common operations
(pthread_rwlock_{clock,timed}{wr,rd}lock) are the __readers, __writers,
__wrphase_futex, __writers_futex, and __cur_writer. The __shared and __flag
are either accessed with non-atomic operation or at initialization, so I don't
think there is much different in packing or not in a 32-bit word. My
understanding is architectures usually does that to avoid possible internal
padding, since the usual types for both member are uint8_t.
>>
>>>>> +# define _FP_DIV_MEAT_S(R, X, Y) _FP_DIV_MEAT_1_udiv_norm (S, R, X, Y)
>>>>> +# define _FP_DIV_MEAT_D(R, X, Y) _FP_DIV_MEAT_2_udiv (D, R, X, Y)
>>>>> +# define _FP_DIV_MEAT_Q(R, X, Y) _FP_DIV_MEAT_4_udiv (Q, R, X, Y)
>>>>> +
>>>>> +# define _FP_NANFRAC_S _FP_QNANBIT_S
>>>>> +# define _FP_NANFRAC_D _FP_QNANBIT_D, 0
>>>>> +# define _FP_NANFRAC_Q _FP_QNANBIT_Q, 0, 0, 0
>>>>
>>>> Likewise. There seems to be an established practice for this header
>>>> across architectures to have no space between macro arguments or before
>>>> the opening parenthesis. This might help with the alignment.
>>>
>>> I still think it makes sense to follow the glibc style though, even if
>>> other archs don't.
>>>
>>> Let me know if it should be a different way and I'll update it.
>>
>> There is the issue of the discrepancy compared to the libgcc version, and
>> while `diff -l' and `patch -l' solve that for manual processing, more
>> sophisticated tools may not cope and require manual intervention.
>>
>> Again, I would suggest getting a second opinion.
>
> The current option matches the rest of the file, so I think it makes
> sense to leave as is.
>
> If this is a problem we can just change the spacing for the entire
> file in a future patch.
>
> Alistair
>
>>
>> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
* Re: [PATCH v2 09/18] RISC-V: The ABI implementation for 32-bit
2020-08-27 19:43 ` Adhemerval Zanella via Libc-alpha
@ 2020-09-25 23:03 ` Alistair Francis via Libc-alpha
0 siblings, 0 replies; 55+ messages in thread
From: Alistair Francis via Libc-alpha @ 2020-09-25 23:03 UTC (permalink / raw)
To: Adhemerval Zanella; +Cc: GNU C Library
On Thu, Aug 27, 2020 at 12:43 PM Adhemerval Zanella via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
>
>
> On 10/08/2020 18:29, Alistair Francis via Libc-alpha wrote:
> > On Fri, Jul 10, 2020 at 6:24 PM Maciej W. Rozycki <macro@wdc.com> wrote:
> >>
> >> On Fri, 10 Jul 2020, Alistair Francis wrote:
> >>
> >>>> I note there is an extensive discussion on the way to move forward here:
> >>>> <https://sourceware.org/glibc/wiki/Y2038ProofnessDesign#utmp_types_and_APIs>
> >>>> We might as well try to implement it right away, so as to avoid being
> >>>> limited to 32-bit time records here.
> >>>
> >>> Is there an advantage of doing it now or can we put this off for the
> >>> next release?
> >>
> >> The change is major enough it'll have to wait for the next development
> >> cycle anyway. It shouldn't matter that much for RV32 glibc deployments
> >> though, given the amount of suitable hardware available.
> >>
> >>>> NB some existing ports do have __WORDSIZE_TIME64_COMPAT32 set and cleared
> >>>> for their 64-bit and 32-bit ABIs respectively, as per the note in our
> >>>> top-level bits/wordsize.h, however this reflects the state as before we
> >>>> introduced the possibility for `__time_t' to be a 64-bit type with
> >>>> `__WORDSIZE == 32' ABIs. Given the turn of events I think the note ought
> >>>> to be updated accordingly; I gather it was missed with commit 07fe93cd9850
> >>>> ("generic/typesizes.h: Add support for 32-bit arches with 64-bit types").
> >>
> >> Will you be able to look into it?
> >>
> >> The context here is before Y2038 changes __WORDSIZE_TIME64_COMPAT32 would
> >> only be clear for 64-bit ABIs with those 64-bit systems that do not have
> >> any 32-bit ABI (compatibility mode) to support, such as the DEC Alpha.
> >> And it would always be clear for 32-bit ABIs, so as to use the proper
> >> `__time_t' type without changing the width of actual data held there in
> >> the structure.
> >>
> >> I'm not sure what the story is behind the S/390 port though; perhaps it
> >> does not support ABI coexistence in a single run-time environment.
> >
> > I was just about to start this but Adhemerval has very kindly sent a
> > patch series already. I'm going to leave this as is (defining it as 1
> > for RV32 and RV64).
>
> Don't forget to help on review ;).
I'm trying :)
>
> >
> >
> >>
> >>>>> int __cur_writer;
> >>>>> int __shared;
> >>>>> unsigned long int __pad1;
> >>>>> unsigned long int __pad2;
> >>>>> unsigned int __flags;
> >>>>> +#else
> >>>>> +# if __BYTE_ORDER == __BIG_ENDIAN
> >>>>> + unsigned char __pad1;
> >>>>> + unsigned char __pad2;
> >>>>> + unsigned char __shared;
> >>>>> + unsigned char __flags;
> >>>>> +# else
> >>>>> + unsigned char __flags;
> >>>>> + unsigned char __shared;
> >>>>> + unsigned char __pad1;
> >>>>> + unsigned char __pad2;
> >>>>> +# endif
> >>>>> + int __cur_writer;
> >>>>> +#endif
> >>>>> };
> >>>>
> >>>> I note with this change the RV32 structure will use the generic layout as
> >>>> per sysdeps/nptl/bits/struct_rwlock.h, however regrettably RV64 does not.
> >>>> Would it make sense to instead have the layout the same between RV64 and
> >>>> RV32, perhaps by redefining `__pad1' and `__pad2' in terms of `unsigned
> >>>> long long' (which would have alignment implications though) or otherwise?
> >>>
> >>> I'm not sure which one is better. On one hand it seems better to be
> >>> more generic and therefore RV32 should use the generic interface. On
> >>> the other hand the more similar they are the better. I'm still leaning
> >>> towards we should be generic where possible.
> >>
> >> It would be good to get a second opinion here.
> >
> > The RV64 __pthread_rwlock_arch_t is exactly the same as the AArch64
> > version. My guess is that it was just copied to be the same.
> >
> > The generic version also has this comment
> >
> > /* Generic struct for both POSIX read-write lock. New ports are expected
> > to use the default layout, however archictetures can redefine it to add
> > arch-specific extensions (such as lock-elision). The struct have a size
> > of 32 bytes on both LP32 and LP64 architectures. */
> >
> > As we aren't adding any arch-specific extensions I think we should
> > keep the generic one. It's unfortunate we didn't do that for RV64, but
> > too late now.
> >
> > The only downside I see in using the generic version is what happens
> > if the value of _shared or flags overflows a char. There are currently
> > only 4 flags though, so I think it's ok.
>
> I come up with the generic interface by using the already one in place for
> mostly 64-bit architecture and the internal layout is mainly to avoid duplicate
> the internal type on multiple architectures (so we can use the generic
> implementation).
>
> The only usage I can think of where same structure for RV64 and RV32 may
> pose a role is to use a shared rwlock between a 32-bit and a 64-bit process
> (pthread_rwlockattr_setpshared with PTHREAD_PROCESS_SHARED). But afaik this
> is not supported on other architectures.
Thanks for describing this. I'm going to stick with the current
generic implementation for RV32 then.
>
> >
> >>
> >>>> Is there any benefit from having `__flags' and `__shared' (and the
> >>>> reserve) grouped within a single 32-bit word? I gather there is, given
> >>>> the lengths gone to to match the bit lanes across the word regardless of
> >>>> the endianness. But what is it?
> >>>
> >>> I have no idea.
> >>
> >> Especially given this.
>
> The internal member that are most concurrently accessed by the common operations
> (pthread_rwlock_{clock,timed}{wr,rd}lock) are the __readers, __writers,
> __wrphase_futex, __writers_futex, and __cur_writer. The __shared and __flag
> are either accessed with non-atomic operation or at initialization, so I don't
> think there is much different in packing or not in a 32-bit word. My
> understanding is architectures usually does that to avoid possible internal
> padding, since the usual types for both member are uint8_t.
Thanks for the info here as well.
Alistair
>
> >>
> >>>>> +# define _FP_DIV_MEAT_S(R, X, Y) _FP_DIV_MEAT_1_udiv_norm (S, R, X, Y)
> >>>>> +# define _FP_DIV_MEAT_D(R, X, Y) _FP_DIV_MEAT_2_udiv (D, R, X, Y)
> >>>>> +# define _FP_DIV_MEAT_Q(R, X, Y) _FP_DIV_MEAT_4_udiv (Q, R, X, Y)
> >>>>> +
> >>>>> +# define _FP_NANFRAC_S _FP_QNANBIT_S
> >>>>> +# define _FP_NANFRAC_D _FP_QNANBIT_D, 0
> >>>>> +# define _FP_NANFRAC_Q _FP_QNANBIT_Q, 0, 0, 0
> >>>>
> >>>> Likewise. There seems to be an established practice for this header
> >>>> across architectures to have no space between macro arguments or before
> >>>> the opening parenthesis. This might help with the alignment.
> >>>
> >>> I still think it makes sense to follow the glibc style though, even if
> >>> other archs don't.
> >>>
> >>> Let me know if it should be a different way and I'll update it.
> >>
> >> There is the issue of the discrepancy compared to the libgcc version, and
> >> while `diff -l' and `patch -l' solve that for manual processing, more
> >> sophisticated tools may not cope and require manual intervention.
> >>
> >> Again, I would suggest getting a second opinion.
> >
> > The current option matches the rest of the file, so I think it makes
> > sense to leave as is.
> >
> > If this is a problem we can just change the spacing for the entire
> > file in a future patch.
> >
> > Alistair
> >
> >>
> >> Maciej
^ permalink raw reply [flat|nested] 55+ messages in thread
end of thread, other threads:[~2020-09-25 23:14 UTC | newest]
Thread overview: 55+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-06-03 16:25 [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 01/18] RISC-V: Use 64-bit time_t and off_t for RV32 and RV64 Alistair Francis via Libc-alpha
2020-07-07 22:06 ` Maciej W. Rozycki via Libc-alpha
2020-07-10 15:27 ` Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 02/18] RISC-V: Define __NR_* as __NR_*_time64/64 for 32-bit Alistair Francis via Libc-alpha
2020-07-08 0:09 ` Maciej W. Rozycki via Libc-alpha
2020-07-08 17:08 ` Adhemerval Zanella via Libc-alpha
2020-07-09 17:14 ` Alistair Francis via Libc-alpha
2020-07-16 0:23 ` Maciej W. Rozycki via Libc-alpha
2020-07-09 17:10 ` Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 03/18] RISC-V: Add support for 32-bit vDSO calls Alistair Francis via Libc-alpha
2020-07-08 1:01 ` Maciej W. Rozycki via Libc-alpha
2020-07-08 18:17 ` Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 04/18] RISC-V: Support dynamic loader for the 32-bit Alistair Francis via Libc-alpha
2020-07-08 1:35 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 05/18] RISC-V: Add path of library directories " Alistair Francis via Libc-alpha
2020-07-08 18:42 ` Maciej W. Rozycki via Libc-alpha
2020-07-09 17:03 ` Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 06/18] RISC-V: Add arch-syscall.h for RV32 Alistair Francis via Libc-alpha
2020-07-08 19:33 ` Maciej W. Rozycki via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 07/18] RISC-V: nptl: update default pthread-offsets.h Alistair Francis via Libc-alpha
2020-07-09 0:14 ` Maciej W. Rozycki via Libc-alpha
2020-07-09 11:47 ` Adhemerval Zanella via Libc-alpha
2020-07-15 19:23 ` Maciej W. Rozycki via Libc-alpha
2020-08-10 17:34 ` Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 08/18] riscv32: Add an architecture ipctypes.h Alistair Francis via Libc-alpha
2020-07-09 2:46 ` Maciej W. Rozycki via Libc-alpha
2020-07-09 11:36 ` Adhemerval Zanella via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 09/18] RISC-V: The ABI implementation for 32-bit Alistair Francis via Libc-alpha
2020-07-09 23:33 ` Maciej W. Rozycki via Libc-alpha
2020-07-10 16:45 ` Alistair Francis via Libc-alpha
2020-07-11 1:24 ` Maciej W. Rozycki via Libc-alpha
2020-08-10 21:29 ` Alistair Francis via Libc-alpha
2020-08-27 19:43 ` Adhemerval Zanella via Libc-alpha
2020-09-25 23:03 ` Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 10/18] RISC-V: Hard float support " Alistair Francis via Libc-alpha
2020-07-11 0:49 ` Maciej W. Rozycki via Libc-alpha
2020-07-11 15:49 ` Alistair Francis via Libc-alpha
2020-07-11 22:13 ` Maciej W. Rozycki via Libc-alpha
2020-07-12 15:34 ` Alistair Francis via Libc-alpha
2020-07-12 22:10 ` Maciej W. Rozycki via Libc-alpha
2020-08-27 18:36 ` Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 11/18] RISC-V: Add ABI lists Alistair Francis via Libc-alpha
2020-07-12 20:54 ` Maciej W. Rozycki via Libc-alpha
2020-07-13 16:14 ` Alistair Francis via Libc-alpha
2020-06-03 16:25 ` [PATCH v2 12/18] RISC-V: Add the RV32 libm-test-ulps Alistair Francis via Libc-alpha
2020-06-03 17:34 ` Joseph Myers
2020-06-05 4:01 ` Alistair Francis via Libc-alpha
2020-06-03 16:26 ` [PATCH v2 13/18] RISC-V: Fix llrint and llround missing exceptions on RV32 Alistair Francis via Libc-alpha
2020-06-03 16:26 ` [PATCH v2 14/18] RISC-V: Build Infastructure for 32-bit Alistair Francis via Libc-alpha
2020-06-03 16:26 ` [PATCH v2 15/18] riscv32: Specify the arch_minimum_kernel as 5.4 Alistair Francis via Libc-alpha
2020-06-03 16:26 ` [PATCH v2 16/18] RISC-V: Add rv32 path to RTLDLIST in ldd Alistair Francis via Libc-alpha
2020-06-03 16:26 ` [PATCH v2 17/18] Documentation for the RISC-V 32-bit port Alistair Francis via Libc-alpha
2020-06-03 16:26 ` [PATCH v2 18/18] Add RISC-V 32-bit target to build-many-glibcs.py Alistair Francis via Libc-alpha
2020-06-15 22:37 ` [PATCH v2 00/18] glibc port for 32-bit RISC-V (RV32) Alistair Francis via Libc-alpha
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).