* [PATCH 0/2] OpenRISC glibc hard float support
@ 2024-03-29 7:40 Stafford Horne
2024-03-29 7:40 ` [PATCH 1/2] or1k: Add " Stafford Horne
2024-03-29 7:40 ` [PATCH 2/2] build-many-glibcs.py: Add openrisc hard float glibc variant Stafford Horne
0 siblings, 2 replies; 8+ messages in thread
From: Stafford Horne @ 2024-03-29 7:40 UTC (permalink / raw)
To: GLIBC patches; +Cc: Linux OpenRISC, Stafford Horne
This series adds hardware floating point support to the OpenRIS glibc port.
I will keep this cover letter short and point to patch 1/2. It contains details
about the change and testing results.
Stafford Horne (2):
or1k: Add hard float support
build-many-glibcs.py: Add openrisc hard float glibc variant
scripts/build-many-glibcs.py | 5 +-
sysdeps/or1k/fpu/fclrexcpt.c | 44 +
sysdeps/or1k/fpu/fegetenv.c | 32 +
sysdeps/or1k/fpu/fegetmode.c | 29 +
sysdeps/or1k/fpu/fegetround.c | 29 +
sysdeps/or1k/fpu/feholdexcpt.c | 33 +
sysdeps/or1k/fpu/fenv_private.h | 199 +++
sysdeps/or1k/fpu/fesetenv.c | 32 +
sysdeps/or1k/fpu/fesetexcept.c | 35 +
sysdeps/or1k/fpu/fesetmode.c | 39 +
sysdeps/or1k/fpu/fesetround.c | 39 +
sysdeps/or1k/fpu/feupdateenv.c | 33 +
sysdeps/or1k/fpu/fgetexcptflg.c | 29 +
.../or1k/fpu/fix-fp-int-convert-overflow.h | 38 +
sysdeps/or1k/fpu/fraiseexcpt.c | 67 +
sysdeps/or1k/fpu/fsetexcptflg.c | 43 +
sysdeps/or1k/fpu/ftestexcept.c | 27 +
sysdeps/or1k/fpu/get-rounding-mode.h | 38 +
sysdeps/or1k/fpu/libm-test-ulps | 1115 +++++++++++++++++
sysdeps/or1k/fpu/libm-test-ulps-name | 1 +
sysdeps/or1k/fpu_control.h | 61 +
sysdeps/or1k/libm-test-ulps-name | 1 -
sysdeps/or1k/math-tests-snan-payload.h | 26 +
sysdeps/or1k/math-tests-trap.h | 27 +
sysdeps/or1k/{ => nofpu}/libm-test-ulps | 0
sysdeps/or1k/nofpu/libm-test-ulps-name | 1 +
sysdeps/or1k/sfp-machine.h | 18 +-
sysdeps/unix/sysv/linux/or1k/getcontext.S | 6 +
sysdeps/unix/sysv/linux/or1k/setcontext.S | 6 +
sysdeps/unix/sysv/linux/or1k/swapcontext.S | 12 +
sysdeps/unix/sysv/linux/or1k/sys/ucontext.h | 1 +
sysdeps/unix/sysv/linux/or1k/ucontext_i.sym | 3 +
32 files changed, 2065 insertions(+), 4 deletions(-)
create mode 100644 sysdeps/or1k/fpu/fclrexcpt.c
create mode 100644 sysdeps/or1k/fpu/fegetenv.c
create mode 100644 sysdeps/or1k/fpu/fegetmode.c
create mode 100644 sysdeps/or1k/fpu/fegetround.c
create mode 100644 sysdeps/or1k/fpu/feholdexcpt.c
create mode 100644 sysdeps/or1k/fpu/fenv_private.h
create mode 100644 sysdeps/or1k/fpu/fesetenv.c
create mode 100644 sysdeps/or1k/fpu/fesetexcept.c
create mode 100644 sysdeps/or1k/fpu/fesetmode.c
create mode 100644 sysdeps/or1k/fpu/fesetround.c
create mode 100644 sysdeps/or1k/fpu/feupdateenv.c
create mode 100644 sysdeps/or1k/fpu/fgetexcptflg.c
create mode 100644 sysdeps/or1k/fpu/fix-fp-int-convert-overflow.h
create mode 100644 sysdeps/or1k/fpu/fraiseexcpt.c
create mode 100644 sysdeps/or1k/fpu/fsetexcptflg.c
create mode 100644 sysdeps/or1k/fpu/ftestexcept.c
create mode 100644 sysdeps/or1k/fpu/get-rounding-mode.h
create mode 100644 sysdeps/or1k/fpu/libm-test-ulps
create mode 100644 sysdeps/or1k/fpu/libm-test-ulps-name
create mode 100644 sysdeps/or1k/fpu_control.h
delete mode 100644 sysdeps/or1k/libm-test-ulps-name
create mode 100644 sysdeps/or1k/math-tests-snan-payload.h
create mode 100644 sysdeps/or1k/math-tests-trap.h
rename sysdeps/or1k/{ => nofpu}/libm-test-ulps (100%)
create mode 100644 sysdeps/or1k/nofpu/libm-test-ulps-name
--
2.44.0
^ permalink raw reply [flat|nested] 8+ messages in thread* [PATCH 1/2] or1k: Add hard float support 2024-03-29 7:40 [PATCH 0/2] OpenRISC glibc hard float support Stafford Horne @ 2024-03-29 7:40 ` Stafford Horne 2024-04-16 14:04 ` Adhemerval Zanella Netto 2024-03-29 7:40 ` [PATCH 2/2] build-many-glibcs.py: Add openrisc hard float glibc variant Stafford Horne 1 sibling, 1 reply; 8+ messages in thread From: Stafford Horne @ 2024-03-29 7:40 UTC (permalink / raw) To: GLIBC patches; +Cc: Linux OpenRISC, Stafford Horne This patch adds hardware floating point support to OpenRISC. Hardware floating point toolchain builds are enabled by passing the machine specific argument -mhard-float to gcc via CFLAGS. With this enabled GCC generates floating point instructions for single-precision operations and exports __or1k_hard_float__. There are 2 main parts to this patch. - Implement fenv functions to update the FPCSR flags keeping it in sync with sfp (software floating point). - Update machine context functions to store and restore the FPCSR state. *On mcontext_t ABI* This patch adds __fpcsr to mcontext_t. This is an ABI change, but also an ABI fix. The linux kernel has always defined padding in mcontext_t that was was missing from the glibc ABI. This unused space is has now been re-purposed for storing the FPCSR. This patch brings OpenRISC glibc in line with the Linux kernel and other libc implementation (musl). *Hard float ABI* The calling conventions and types do not change with OpenRISC hard-float so glibc hard-float builds continue to use dynamic linker /lib/ld-linux-or1k.so.1. *Testing* I have tested this patch both with hard-float and soft-float builds and the test results look fine to me. Results are as follows: Hard Float # failures FAIL: elf/tst-sprof-basic (Not checked, seen in other ports too) FAIL: gmon/tst-gmon-pie (GCC PIE bug) FAIL: gmon/tst-gmon-pie-gprof (GCC PIE bug) FAIL: nptl/tst-cond24 (timeouts) FAIL: nptl/tst-mutex10 (timeouts) # summary 5 FAIL 4279 PASS 83 UNSUPPORTED 16 XFAIL 2 XPASS # versions Toolchain: or1k-smhfpu-linux-gnu Compiler: gcc version 14.0.1 20240324 (experimental) [master r14-9649-gbb04a11418f] (GCC) Binutils: GNU assembler version 2.42.0 (or1k-smhfpu-linux-gnu) using BFD version (GNU Binutils) 2.42.0.20240324 Linux: Linux buildroot 6.8.0-rc5-00003-g9158c6253560-dirty #78 SMP Fri Mar 22 16:56:54 GMT 2024 openrisc GNU/Linux Soft Float # failures FAIL: elf/tst-sprof-basic FAIL: gmon/tst-gmon-pie FAIL: gmon/tst-gmon-pie-gprof FAIL: iconvdata/iconv-test (timeout, passed when run manually) FAIL: nptl/tst-mutex10 # summary 5 FAIL 4281 PASS 81 UNSUPPORTED 16 XFAIL 2 XPASS # versions Toolchain: or1k-smh-linux-gnu Compiler: gcc version 14.0.1 20240324 (experimental) [master r14-9649-gbb04a11418f] (GCC) Binutils: GNU assembler version 2.42.0 (or1k-smh-linux-gnu) using BFD version (GNU Binutils) 2.42.0.20240324 Linux: Linux buildroot 6.8.0-rc5-00003-g9158c6253560-dirty #78 SMP Fri Mar 22 16:56:54 GMT 2024 openrisc GNU/Linux Documentation: https://raw.githubusercontent.com/openrisc/doc/master/openrisc-arch-1.4-rev0.pdf --- sysdeps/or1k/fpu/fclrexcpt.c | 44 + sysdeps/or1k/fpu/fegetenv.c | 32 + sysdeps/or1k/fpu/fegetmode.c | 29 + sysdeps/or1k/fpu/fegetround.c | 29 + sysdeps/or1k/fpu/feholdexcpt.c | 33 + sysdeps/or1k/fpu/fenv_private.h | 199 +++ sysdeps/or1k/fpu/fesetenv.c | 32 + sysdeps/or1k/fpu/fesetexcept.c | 35 + sysdeps/or1k/fpu/fesetmode.c | 39 + sysdeps/or1k/fpu/fesetround.c | 39 + sysdeps/or1k/fpu/feupdateenv.c | 33 + sysdeps/or1k/fpu/fgetexcptflg.c | 29 + .../or1k/fpu/fix-fp-int-convert-overflow.h | 38 + sysdeps/or1k/fpu/fraiseexcpt.c | 67 + sysdeps/or1k/fpu/fsetexcptflg.c | 43 + sysdeps/or1k/fpu/ftestexcept.c | 27 + sysdeps/or1k/fpu/get-rounding-mode.h | 38 + sysdeps/or1k/fpu/libm-test-ulps | 1115 +++++++++++++++++ sysdeps/or1k/fpu/libm-test-ulps-name | 1 + sysdeps/or1k/fpu_control.h | 61 + sysdeps/or1k/libm-test-ulps-name | 1 - sysdeps/or1k/math-tests-snan-payload.h | 26 + sysdeps/or1k/math-tests-trap.h | 27 + sysdeps/or1k/{ => nofpu}/libm-test-ulps | 0 sysdeps/or1k/nofpu/libm-test-ulps-name | 1 + sysdeps/or1k/sfp-machine.h | 18 +- sysdeps/unix/sysv/linux/or1k/getcontext.S | 6 + sysdeps/unix/sysv/linux/or1k/setcontext.S | 6 + sysdeps/unix/sysv/linux/or1k/swapcontext.S | 12 + sysdeps/unix/sysv/linux/or1k/sys/ucontext.h | 1 + sysdeps/unix/sysv/linux/or1k/ucontext_i.sym | 3 + 31 files changed, 2062 insertions(+), 2 deletions(-) create mode 100644 sysdeps/or1k/fpu/fclrexcpt.c create mode 100644 sysdeps/or1k/fpu/fegetenv.c create mode 100644 sysdeps/or1k/fpu/fegetmode.c create mode 100644 sysdeps/or1k/fpu/fegetround.c create mode 100644 sysdeps/or1k/fpu/feholdexcpt.c create mode 100644 sysdeps/or1k/fpu/fenv_private.h create mode 100644 sysdeps/or1k/fpu/fesetenv.c create mode 100644 sysdeps/or1k/fpu/fesetexcept.c create mode 100644 sysdeps/or1k/fpu/fesetmode.c create mode 100644 sysdeps/or1k/fpu/fesetround.c create mode 100644 sysdeps/or1k/fpu/feupdateenv.c create mode 100644 sysdeps/or1k/fpu/fgetexcptflg.c create mode 100644 sysdeps/or1k/fpu/fix-fp-int-convert-overflow.h create mode 100644 sysdeps/or1k/fpu/fraiseexcpt.c create mode 100644 sysdeps/or1k/fpu/fsetexcptflg.c create mode 100644 sysdeps/or1k/fpu/ftestexcept.c create mode 100644 sysdeps/or1k/fpu/get-rounding-mode.h create mode 100644 sysdeps/or1k/fpu/libm-test-ulps create mode 100644 sysdeps/or1k/fpu/libm-test-ulps-name create mode 100644 sysdeps/or1k/fpu_control.h delete mode 100644 sysdeps/or1k/libm-test-ulps-name create mode 100644 sysdeps/or1k/math-tests-snan-payload.h create mode 100644 sysdeps/or1k/math-tests-trap.h rename sysdeps/or1k/{ => nofpu}/libm-test-ulps (100%) create mode 100644 sysdeps/or1k/nofpu/libm-test-ulps-name diff --git a/sysdeps/or1k/fpu/fclrexcpt.c b/sysdeps/or1k/fpu/fclrexcpt.c new file mode 100644 index 0000000000..44224f9c24 --- /dev/null +++ b/sysdeps/or1k/fpu/fclrexcpt.c @@ -0,0 +1,44 @@ +/* Clear given exceptions in current floating-point environment. + OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fpu_control.h> + +int +feclearexcept (int excepts) +{ + fpu_control_t cw; + fpu_control_t cw_new; + + /* Mask out unsupported bits/exceptions. */ + excepts &= FE_ALL_EXCEPT; + + /* Read the complete control word. */ + _FPU_GETCW (cw); + + cw_new = cw & ~excepts; + + /* Put the new data in effect. */ + if (cw != cw_new) + _FPU_SETCW (cw_new); + + /* Success. */ + return 0; +} +libm_hidden_def (feclearexcept) diff --git a/sysdeps/or1k/fpu/fegetenv.c b/sysdeps/or1k/fpu/fegetenv.c new file mode 100644 index 0000000000..70c75aa0bf --- /dev/null +++ b/sysdeps/or1k/fpu/fegetenv.c @@ -0,0 +1,32 @@ +/* Store current floating-point environment. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fpu_control.h> + +int +__fegetenv (fenv_t *envp) +{ + _FPU_GETCW (*envp); + + /* Success. */ + return 0; +} +libm_hidden_def (__fegetenv) +weak_alias (__fegetenv, fegetenv) +libm_hidden_weak (fegetenv) diff --git a/sysdeps/or1k/fpu/fegetmode.c b/sysdeps/or1k/fpu/fegetmode.c new file mode 100644 index 0000000000..7fffd2e0b5 --- /dev/null +++ b/sysdeps/or1k/fpu/fegetmode.c @@ -0,0 +1,29 @@ +/* Store current floating-point control modes. OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fpu_control.h> + +int +fegetmode (femode_t *modep) +{ + _FPU_GETCW (*modep); + + /* Success. */ + return 0; +} diff --git a/sysdeps/or1k/fpu/fegetround.c b/sysdeps/or1k/fpu/fegetround.c new file mode 100644 index 0000000000..7e993b980a --- /dev/null +++ b/sysdeps/or1k/fpu/fegetround.c @@ -0,0 +1,29 @@ +/* Return current rounding direction. OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <get-rounding-mode.h> + +int +__fegetround (void) +{ + return get_rounding_mode (); +} +libm_hidden_def (__fegetround) +weak_alias (__fegetround, fegetround) +libm_hidden_weak (fegetround) diff --git a/sysdeps/or1k/fpu/feholdexcpt.c b/sysdeps/or1k/fpu/feholdexcpt.c new file mode 100644 index 0000000000..0036e41ba2 --- /dev/null +++ b/sysdeps/or1k/fpu/feholdexcpt.c @@ -0,0 +1,33 @@ +/* Store current floating-point environment and clear exceptions. + OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fenv_private.h> + +int +__feholdexcept (fenv_t *envp) +{ + libc_feholdexcept_or1k (envp); + + /* Success. */ + return 0; +} +libm_hidden_def (__feholdexcept) +weak_alias (__feholdexcept, feholdexcept) +libm_hidden_weak (feholdexcept) diff --git a/sysdeps/or1k/fpu/fenv_private.h b/sysdeps/or1k/fpu/fenv_private.h new file mode 100644 index 0000000000..4f401e7a5a --- /dev/null +++ b/sysdeps/or1k/fpu/fenv_private.h @@ -0,0 +1,199 @@ +/* Private floating point rounding and exceptions handling. OpenRISC version. + Copyright (C) 2024 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 OR1K_FENV_PRIVATE_H +#define OR1K_FENV_PRIVATE_H 1 + +#include <fenv.h> +#include <fpu_control.h> + +static __always_inline void +libc_feholdexcept_or1k (fenv_t *envp) +{ + fpu_control_t cw; + fpu_control_t cw_new; + + /* Get and store the environment. */ + _FPU_GETCW (cw); + *envp = cw; + + /* Clear the exception status flags. */ + cw_new = cw & ~FE_ALL_EXCEPT; + + if (cw != cw_new) + _FPU_SETCW (cw_new); +} + +#define libc_feholdexcept libc_feholdexcept_or1k +#define libc_feholdexceptf libc_feholdexcept_or1k +#define libc_feholdexceptl libc_feholdexcept_or1k + +static __always_inline void +libc_fesetround_or1k (int round) +{ + fpu_control_t cw; + fpu_control_t cw_new; + + _FPU_GETCW (cw); + cw_new = cw & ~_FPU_FPCSR_RM_MASK; + cw_new |= round; + if (cw != cw_new) + _FPU_SETCW (cw_new); +} + +#define libc_fesetround libc_fesetround_or1k +#define libc_fesetroundf libc_fesetround_or1k +#define libc_fesetroundl libc_fesetround_or1k + +static __always_inline void +libc_feholdexcept_setround_or1k (fenv_t *envp, int round) +{ + fpu_control_t cw; + fpu_control_t cw_new; + + /* Get and store the environment. */ + _FPU_GETCW (cw); + *envp = cw; + + /* Clear the status flags and rounding mode. */ + cw_new = cw & ~(FE_ALL_EXCEPT | _FPU_FPCSR_RM_MASK); + + /* Set rounding mode. */ + cw_new |= round; + + if (cw != cw_new) + _FPU_SETCW (cw_new); +} + +#define libc_feholdexcept_setround libc_feholdexcept_setround_or1k +#define libc_feholdexcept_setroundf libc_feholdexcept_setround_or1k +#define libc_feholdexcept_setroundl libc_feholdexcept_setround_or1k + +static __always_inline int +libc_fetestexcept_or1k (int ex) +{ + fpu_control_t cw; + + /* Get current control word. */ + _FPU_GETCW (cw); + + /* Check if any of the queried exception flags are set. */ + return cw & ex & FE_ALL_EXCEPT; +} + +#define libc_fetestexcept libc_fetestexcept_or1k +#define libc_fetestexceptf libc_fetestexcept_or1k +#define libc_fetestexceptl libc_fetestexcept_or1k + +static __always_inline void +libc_fesetenv_or1k (const fenv_t *envp) +{ + if (envp == FE_DFL_ENV) + _FPU_SETCW (_FPU_DEFAULT); + else + _FPU_SETCW (*envp); +} + +#define libc_fesetenv libc_fesetenv_or1k +#define libc_fesetenvf libc_fesetenv_or1k +#define libc_fesetenvl libc_fesetenv_or1k +#define libc_feresetround_noex libc_fesetenv_or1k +#define libc_feresetround_noexf libc_fesetenv_or1k +#define libc_feresetround_noexl libc_fesetenv_or1k + +static __always_inline int +libc_feupdateenv_test_or1k (const fenv_t *envp, int ex) +{ + fpu_control_t cw; + fpu_control_t cw_new; + int excepts; + + /* Get current control word. */ + _FPU_GETCW (cw); + + /* Merge current exception flags with the passed fenv. */ + excepts = cw & FE_ALL_EXCEPT; + cw_new = (envp == FE_DFL_ENV ? _FPU_DEFAULT : *envp) | excepts; + + if (__glibc_unlikely (cw != cw_new)) + _FPU_SETCW (cw_new); + + /* Raise the exceptions if enabled in the new FP state. */ + if (__glibc_unlikely (excepts)) + __feraiseexcept (excepts); + + return excepts & ex; +} + +#define libc_feupdateenv_test libc_feupdateenv_test_or1k +#define libc_feupdateenv_testf libc_feupdateenv_test_or1k +#define libc_feupdateenv_testl libc_feupdateenv_test_or1k + +static __always_inline void +libc_feupdateenv_or1k (const fenv_t *envp) +{ + libc_feupdateenv_test_or1k (envp, 0); +} + +#define libc_feupdateenv libc_feupdateenv_or1k +#define libc_feupdateenvf libc_feupdateenv_or1k +#define libc_feupdateenvl libc_feupdateenv_or1k + +static __always_inline void +libc_feholdsetround_or1k (fenv_t *envp, int round) +{ + fpu_control_t cw; + + _FPU_GETCW (cw); + *envp = cw; + + /* Check whether rounding modes are different. */ + round = (cw ^ round) & _FPU_FPCSR_RM_MASK; + + /* Set new rounding mode if different. */ + if (__glibc_unlikely (round != 0)) + _FPU_SETCW (cw ^ round); +} + +#define libc_feholdsetround libc_feholdsetround_or1k +#define libc_feholdsetroundf libc_feholdsetround_or1k +#define libc_feholdsetroundl libc_feholdsetround_or1k + +static __always_inline void +libc_feresetround_or1k (fenv_t *envp) +{ + fpu_control_t cw; + int round; + + _FPU_GETCW (cw); + + /* Check whether rounding modes are different. */ + round = (*envp ^ cw) & _FPU_FPCSR_RM_MASK; + + /* Restore the rounding mode if it was changed. */ + if (__glibc_unlikely (round != 0)) + _FPU_SETCW (cw ^ round); +} + +#define libc_feresetround libc_feresetround_or1k +#define libc_feresetroundf libc_feresetround_or1k +#define libc_feresetroundl libc_feresetround_or1k + +#include_next <fenv_private.h> + +#endif diff --git a/sysdeps/or1k/fpu/fesetenv.c b/sysdeps/or1k/fpu/fesetenv.c new file mode 100644 index 0000000000..742ca719e0 --- /dev/null +++ b/sysdeps/or1k/fpu/fesetenv.c @@ -0,0 +1,32 @@ +/* Install given floating-point environment. OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fenv_private.h> + +int +__fesetenv (const fenv_t *envp) +{ + libc_fesetenv_or1k (envp); + + /* Success. */ + return 0; +} +libm_hidden_def (__fesetenv) +weak_alias (__fesetenv, fesetenv) +libm_hidden_weak (fesetenv) diff --git a/sysdeps/or1k/fpu/fesetexcept.c b/sysdeps/or1k/fpu/fesetexcept.c new file mode 100644 index 0000000000..43734eac18 --- /dev/null +++ b/sysdeps/or1k/fpu/fesetexcept.c @@ -0,0 +1,35 @@ +/* Set given exception flags. OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fpu_control.h> + +int +fesetexcept (int excepts) +{ + fpu_control_t cw; + fpu_control_t cw_new; + + _FPU_GETCW (cw); + cw_new = cw | (excepts & FE_ALL_EXCEPT); + if (cw != cw_new) + _FPU_SETCW (cw_new); + + /* Success. */ + return 0; +} diff --git a/sysdeps/or1k/fpu/fesetmode.c b/sysdeps/or1k/fpu/fesetmode.c new file mode 100644 index 0000000000..d4556927ce --- /dev/null +++ b/sysdeps/or1k/fpu/fesetmode.c @@ -0,0 +1,39 @@ +/* Install given floating-point control modes. OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fpu_control.h> + +int +fesetmode (const femode_t *modep) +{ + fpu_control_t cw; + fpu_control_t cw_new; + + _FPU_GETCW (cw); + cw_new = cw & ~_FPU_FPCSR_RM_MASK; + if (modep == FE_DFL_MODE) + cw_new |= (_FPU_DEFAULT & _FPU_FPCSR_RM_MASK); + else + cw_new |= (*modep & _FPU_FPCSR_RM_MASK); + if (cw != cw_new) + _FPU_SETCW (cw_new); + + /* Success. */ + return 0; +} diff --git a/sysdeps/or1k/fpu/fesetround.c b/sysdeps/or1k/fpu/fesetround.c new file mode 100644 index 0000000000..c2ada98f1b --- /dev/null +++ b/sysdeps/or1k/fpu/fesetround.c @@ -0,0 +1,39 @@ +/* Set current rounding direction. OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fenv_private.h> + +int +__fesetround (int round) +{ + switch (round) + { + case FE_TONEAREST: + case FE_TOWARDZERO: + case FE_DOWNWARD: + case FE_UPWARD: + libc_fesetround_or1k (round); + return 0; + default: + return round; /* A nonzero value. */ + } +} +libm_hidden_def (__fesetround) +weak_alias (__fesetround, fesetround) +libm_hidden_weak (fesetround) diff --git a/sysdeps/or1k/fpu/feupdateenv.c b/sysdeps/or1k/fpu/feupdateenv.c new file mode 100644 index 0000000000..3355bf6596 --- /dev/null +++ b/sysdeps/or1k/fpu/feupdateenv.c @@ -0,0 +1,33 @@ +/* Install given floating-point environment and raise exceptions. + OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fenv_private.h> + +int +__feupdateenv (const fenv_t *envp) +{ + libc_feupdateenv_or1k (envp); + + /* Success. */ + return 0; +} +libm_hidden_def (__feupdateenv) +weak_alias (__feupdateenv, feupdateenv) +libm_hidden_weak (feupdateenv) diff --git a/sysdeps/or1k/fpu/fgetexcptflg.c b/sysdeps/or1k/fpu/fgetexcptflg.c new file mode 100644 index 0000000000..a954f6a2f1 --- /dev/null +++ b/sysdeps/or1k/fpu/fgetexcptflg.c @@ -0,0 +1,29 @@ +/* Store current state of exceptions. OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fenv_private.h> + +int +fegetexceptflag (fexcept_t *flagp, int excepts) +{ + *flagp = libc_fetestexcept_or1k (excepts); + + /* Success. */ + return 0; +} diff --git a/sysdeps/or1k/fpu/fix-fp-int-convert-overflow.h b/sysdeps/or1k/fpu/fix-fp-int-convert-overflow.h new file mode 100644 index 0000000000..78104289c0 --- /dev/null +++ b/sysdeps/or1k/fpu/fix-fp-int-convert-overflow.h @@ -0,0 +1,38 @@ +/* Fix for conversion of floating point to integer overflow. OpenRISC version. + Copyright (C) 2024 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 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 */ diff --git a/sysdeps/or1k/fpu/fraiseexcpt.c b/sysdeps/or1k/fpu/fraiseexcpt.c new file mode 100644 index 0000000000..bbacfd50bc --- /dev/null +++ b/sysdeps/or1k/fpu/fraiseexcpt.c @@ -0,0 +1,67 @@ +/* Raise given exceptions. OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fpu_control.h> +#include <float.h> +#include <math.h> + +int +__feraiseexcept (int excepts) +{ + if (excepts == 0) + return 0; + + /* Raise exceptions represented by EXPECTS. */ + + if (excepts & FE_INEXACT) + { + float d = 1.0, x = 3.0; + __asm__ volatile ("lf.div.s %0, %0, %1" : "+r" (d) : "r" (x)); + } + + if (excepts & FE_UNDERFLOW) + { + float d = FLT_MIN; + __asm__ volatile ("lf.mul.s %0, %0, %0" : "+r" (d)); + } + + if (excepts & FE_OVERFLOW) + { + float d = FLT_MAX; + __asm__ volatile ("lf.mul.s %0, %0, %0" : "+r" (d) : "r" (d)); + } + + if (excepts & FE_DIVBYZERO) + { + float d = 1.0, x = 0.0; + __asm__ volatile ("lf.div.s %0, %0, %1" : "+r" (d) : "r" (x)); + } + + if (excepts & FE_INVALID) + { + float d = HUGE_VAL, x = 0.0; + __asm__ volatile ("lf.mul.s %0, %1, %0" : "+r" (d) : "r" (x)); + } + + /* Success. */ + return 0; +} +libm_hidden_def (__feraiseexcept) +weak_alias (__feraiseexcept, feraiseexcept) +libm_hidden_weak (feraiseexcept) diff --git a/sysdeps/or1k/fpu/fsetexcptflg.c b/sysdeps/or1k/fpu/fsetexcptflg.c new file mode 100644 index 0000000000..c327e4c5d1 --- /dev/null +++ b/sysdeps/or1k/fpu/fsetexcptflg.c @@ -0,0 +1,43 @@ +/* Set floating-point environment exception handling. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fpu_control.h> + +int +fesetexceptflag (const fexcept_t *flagp, int excepts) +{ + fpu_control_t cw; + fpu_control_t cw_new; + + /* Get the current exceptions. */ + _FPU_GETCW (cw); + + /* Make sure the flags we want restored are legal. */ + excepts &= FE_ALL_EXCEPT; + + /* Now set selected bits from flagp. Note that we ignore all non-flag + bits from *flagp, so they don't matter. */ + cw_new = (cw & ~excepts) | (*flagp & excepts); + + if (cw != cw_new) + _FPU_SETCW (cw_new); + + /* Success. */ + return 0; +} diff --git a/sysdeps/or1k/fpu/ftestexcept.c b/sysdeps/or1k/fpu/ftestexcept.c new file mode 100644 index 0000000000..59f06afa22 --- /dev/null +++ b/sysdeps/or1k/fpu/ftestexcept.c @@ -0,0 +1,27 @@ +/* Test exception in current environment. OpenRISC version. + Copyright (C) 2024 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/>. */ + +#include <fenv.h> +#include <fenv_private.h> + +int +fetestexcept (int excepts) +{ + return libc_fetestexcept_or1k (excepts); +} +libm_hidden_def (fetestexcept) diff --git a/sysdeps/or1k/fpu/get-rounding-mode.h b/sysdeps/or1k/fpu/get-rounding-mode.h new file mode 100644 index 0000000000..a66d553be8 --- /dev/null +++ b/sysdeps/or1k/fpu/get-rounding-mode.h @@ -0,0 +1,38 @@ +/* Determine floating-point rounding mode within libc. OpenRISC version. + + Copyright (C) 2024 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 _OR1K_GET_ROUNDING_MODE_H +#define _OR1K_GET_ROUNDING_MODE_H 1 + +#include <fenv.h> +#include <fpu_control.h> + +/* Return the floating-point rounding mode. */ + +static inline int +get_rounding_mode (void) +{ + fpu_control_t cw; + + _FPU_GETCW (cw); + return cw & _FPU_FPCSR_RM_MASK; +} + +#endif /* get-rounding-mode.h */ diff --git a/sysdeps/or1k/fpu/libm-test-ulps b/sysdeps/or1k/fpu/libm-test-ulps new file mode 100644 index 0000000000..948ec01cdc --- /dev/null +++ b/sysdeps/or1k/fpu/libm-test-ulps @@ -0,0 +1,1115 @@ +# Begin of automatic generation + +# Maximal error of functions: +Function: "acos": +double: 1 +float: 1 + +Function: "acos_downward": +double: 1 +float: 1 + +Function: "acos_towardzero": +double: 1 +float: 1 + +Function: "acos_upward": +double: 1 +float: 1 + +Function: "acosh": +double: 2 +float: 2 + +Function: "acosh_downward": +double: 2 +float: 2 + +Function: "acosh_towardzero": +double: 2 +float: 2 + +Function: "acosh_upward": +double: 2 +float: 2 + +Function: "asin": +double: 1 +float: 1 + +Function: "asin_downward": +double: 1 +float: 1 + +Function: "asin_towardzero": +double: 1 +float: 1 + +Function: "asin_upward": +double: 2 +float: 1 + +Function: "asinh": +double: 2 +float: 2 + +Function: "asinh_downward": +double: 3 +float: 3 + +Function: "asinh_towardzero": +double: 2 +float: 2 + +Function: "asinh_upward": +double: 3 +float: 3 + +Function: "atan": +double: 1 +float: 1 + +Function: "atan2": +float: 2 + +Function: "atan2_downward": +double: 5 +float: 2 + +Function: "atan2_towardzero": +double: 5 +float: 2 + +Function: "atan2_upward": +double: 8 +float: 2 + +Function: "atan_downward": +double: 1 +float: 2 + +Function: "atan_towardzero": +double: 1 +float: 1 + +Function: "atan_upward": +double: 1 +float: 2 + +Function: "atanh": +double: 2 +float: 2 + +Function: "atanh_downward": +double: 3 +float: 3 + +Function: "atanh_towardzero": +double: 2 +float: 2 + +Function: "atanh_upward": +double: 3 +float: 3 + +Function: "cabs": +double: 1 + +Function: "cabs_downward": +double: 1 + +Function: "cabs_towardzero": +double: 1 + +Function: "cabs_upward": +double: 1 + +Function: Real part of "cacos": +double: 1 +float: 2 + +Function: Imaginary part of "cacos": +double: 2 +float: 2 + +Function: Real part of "cacos_downward": +double: 3 +float: 2 + +Function: Imaginary part of "cacos_downward": +double: 5 +float: 3 + +Function: Real part of "cacos_towardzero": +double: 3 +float: 2 + +Function: Imaginary part of "cacos_towardzero": +double: 4 +float: 2 + +Function: Real part of "cacos_upward": +double: 2 +float: 2 + +Function: Imaginary part of "cacos_upward": +double: 5 +float: 7 + +Function: Real part of "cacosh": +double: 2 +float: 2 + +Function: Imaginary part of "cacosh": +double: 1 +float: 2 + +Function: Real part of "cacosh_downward": +double: 4 +float: 2 + +Function: Imaginary part of "cacosh_downward": +double: 3 +float: 3 + +Function: Real part of "cacosh_towardzero": +double: 4 +float: 2 + +Function: Imaginary part of "cacosh_towardzero": +double: 3 +float: 2 + +Function: Real part of "cacosh_upward": +double: 4 +float: 3 + +Function: Imaginary part of "cacosh_upward": +double: 3 +float: 2 + +Function: "carg": +float: 1 + +Function: "carg_downward": +double: 5 +float: 2 + +Function: "carg_towardzero": +double: 5 +float: 2 + +Function: "carg_upward": +double: 8 +float: 2 + +Function: Real part of "casin": +double: 1 +float: 1 + +Function: Imaginary part of "casin": +double: 2 +float: 2 + +Function: Real part of "casin_downward": +double: 3 +float: 2 + +Function: Imaginary part of "casin_downward": +double: 5 +float: 3 + +Function: Real part of "casin_towardzero": +double: 3 +float: 1 + +Function: Imaginary part of "casin_towardzero": +double: 4 +float: 2 + +Function: Real part of "casin_upward": +double: 3 +float: 2 + +Function: Imaginary part of "casin_upward": +double: 5 +float: 7 + +Function: Real part of "casinh": +double: 2 +float: 2 + +Function: Imaginary part of "casinh": +double: 1 +float: 1 + +Function: Real part of "casinh_downward": +double: 5 +float: 3 + +Function: Imaginary part of "casinh_downward": +double: 3 +float: 2 + +Function: Real part of "casinh_towardzero": +double: 4 +float: 2 + +Function: Imaginary part of "casinh_towardzero": +double: 3 +float: 1 + +Function: Real part of "casinh_upward": +double: 5 +float: 7 + +Function: Imaginary part of "casinh_upward": +double: 3 +float: 2 + +Function: Real part of "catan": +double: 1 +float: 1 + +Function: Imaginary part of "catan": +double: 1 +float: 1 + +Function: Real part of "catan_downward": +double: 1 +float: 2 + +Function: Imaginary part of "catan_downward": +double: 2 +float: 2 + +Function: Real part of "catan_towardzero": +double: 1 +float: 2 + +Function: Imaginary part of "catan_towardzero": +double: 2 +float: 2 + +Function: Real part of "catan_upward": +double: 2 +float: 1 + +Function: Imaginary part of "catan_upward": +double: 2 +float: 2 + +Function: Real part of "catanh": +double: 1 +float: 1 + +Function: Imaginary part of "catanh": +double: 1 +float: 1 + +Function: Real part of "catanh_downward": +double: 2 +float: 2 + +Function: Imaginary part of "catanh_downward": +double: 1 +float: 2 + +Function: Real part of "catanh_towardzero": +double: 2 +float: 2 + +Function: Imaginary part of "catanh_towardzero": +double: 1 +float: 2 + +Function: Real part of "catanh_upward": +double: 4 +float: 4 + +Function: Imaginary part of "catanh_upward": +double: 2 +float: 1 + +Function: "cbrt": +double: 4 +float: 1 + +Function: "cbrt_downward": +double: 4 +float: 1 + +Function: "cbrt_towardzero": +double: 3 +float: 1 + +Function: "cbrt_upward": +double: 5 +float: 1 + +Function: Real part of "ccos": +double: 1 +float: 1 + +Function: Imaginary part of "ccos": +double: 1 +float: 1 + +Function: Real part of "ccos_downward": +double: 3 +float: 1 + +Function: Imaginary part of "ccos_downward": +double: 3 +float: 3 + +Function: Real part of "ccos_towardzero": +double: 3 +float: 2 + +Function: Imaginary part of "ccos_towardzero": +double: 3 +float: 3 + +Function: Real part of "ccos_upward": +double: 1 +float: 2 + +Function: Imaginary part of "ccos_upward": +double: 2 +float: 2 + +Function: Real part of "ccosh": +double: 2 +float: 1 + +Function: Imaginary part of "ccosh": +double: 1 +float: 1 + +Function: Real part of "ccosh_downward": +double: 3 +float: 2 + +Function: Imaginary part of "ccosh_downward": +double: 3 +float: 3 + +Function: Real part of "ccosh_towardzero": +double: 3 +float: 3 + +Function: Imaginary part of "ccosh_towardzero": +double: 3 +float: 3 + +Function: Real part of "ccosh_upward": +double: 1 +float: 2 + +Function: Imaginary part of "ccosh_upward": +double: 2 +float: 2 + +Function: Real part of "cexp": +double: 2 +float: 1 + +Function: Imaginary part of "cexp": +double: 1 +float: 2 + +Function: Real part of "cexp_downward": +double: 4 +float: 2 + +Function: Imaginary part of "cexp_downward": +double: 3 +float: 3 + +Function: Real part of "cexp_towardzero": +double: 4 +float: 2 + +Function: Imaginary part of "cexp_towardzero": +double: 3 +float: 3 + +Function: Real part of "cexp_upward": +double: 2 +float: 2 + +Function: Imaginary part of "cexp_upward": +double: 3 +float: 2 + +Function: Real part of "clog": +double: 3 +float: 3 + +Function: Imaginary part of "clog": +double: 1 +float: 1 + +Function: Real part of "clog10": +double: 3 +float: 4 + +Function: Imaginary part of "clog10": +double: 2 +float: 2 + +Function: Real part of "clog10_downward": +double: 7 +float: 5 + +Function: Imaginary part of "clog10_downward": +double: 8 +float: 4 + +Function: Real part of "clog10_towardzero": +double: 5 +float: 5 + +Function: Imaginary part of "clog10_towardzero": +double: 8 +float: 4 + +Function: Real part of "clog10_upward": +double: 6 +float: 5 + +Function: Imaginary part of "clog10_upward": +double: 5 +float: 4 + +Function: Real part of "clog_downward": +double: 4 +float: 3 + +Function: Imaginary part of "clog_downward": +double: 5 +float: 2 + +Function: Real part of "clog_towardzero": +double: 8 +float: 4 + +Function: Imaginary part of "clog_towardzero": +double: 5 +float: 3 + +Function: Real part of "clog_upward": +double: 8 +float: 3 + +Function: Imaginary part of "clog_upward": +double: 8 +float: 2 + +Function: "cos": +double: 1 +float: 1 + +Function: "cos_downward": +double: 1 +float: 1 + +Function: "cos_towardzero": +double: 4 +float: 1 + +Function: "cos_upward": +double: 4 +float: 1 + +Function: "cosh": +double: 2 +float: 2 + +Function: "cosh_downward": +double: 3 +float: 1 + +Function: "cosh_towardzero": +double: 3 +float: 1 + +Function: "cosh_upward": +double: 2 +float: 2 + +Function: Real part of "cpow": +double: 2 +float: 5 + +Function: Imaginary part of "cpow": +float: 2 + +Function: Real part of "cpow_downward": +double: 5 +float: 8 + +Function: Imaginary part of "cpow_downward": +double: 2 +float: 2 + +Function: Real part of "cpow_towardzero": +double: 5 +float: 8 + +Function: Imaginary part of "cpow_towardzero": +double: 2 +float: 2 + +Function: Real part of "cpow_upward": +double: 4 +float: 1 + +Function: Imaginary part of "cpow_upward": +double: 2 +float: 2 + +Function: Real part of "csin": +double: 1 +float: 1 + +Function: Real part of "csin_downward": +double: 3 +float: 3 + +Function: Imaginary part of "csin_downward": +double: 3 +float: 1 + +Function: Real part of "csin_towardzero": +double: 3 +float: 3 + +Function: Imaginary part of "csin_towardzero": +double: 3 +float: 1 + +Function: Real part of "csin_upward": +double: 2 +float: 2 + +Function: Imaginary part of "csin_upward": +double: 1 +float: 2 + +Function: Real part of "csinh": +float: 1 + +Function: Imaginary part of "csinh": +double: 1 +float: 1 + +Function: Real part of "csinh_downward": +double: 3 +float: 1 + +Function: Imaginary part of "csinh_downward": +double: 3 +float: 3 + +Function: Real part of "csinh_towardzero": +double: 3 +float: 2 + +Function: Imaginary part of "csinh_towardzero": +double: 3 +float: 3 + +Function: Real part of "csinh_upward": +double: 1 +float: 2 + +Function: Imaginary part of "csinh_upward": +double: 2 +float: 2 + +Function: Real part of "csqrt": +double: 2 +float: 2 + +Function: Imaginary part of "csqrt": +double: 2 +float: 2 + +Function: Real part of "csqrt_downward": +double: 5 +float: 4 + +Function: Imaginary part of "csqrt_downward": +double: 4 +float: 3 + +Function: Real part of "csqrt_towardzero": +double: 4 +float: 3 + +Function: Imaginary part of "csqrt_towardzero": +double: 4 +float: 3 + +Function: Real part of "csqrt_upward": +double: 5 +float: 4 + +Function: Imaginary part of "csqrt_upward": +double: 3 +float: 3 + +Function: Real part of "ctan": +double: 1 +float: 1 + +Function: Imaginary part of "ctan": +double: 2 +float: 2 + +Function: Real part of "ctan_downward": +double: 6 +float: 5 + +Function: Imaginary part of "ctan_downward": +double: 3 +float: 2 + +Function: Real part of "ctan_towardzero": +double: 5 +float: 3 + +Function: Imaginary part of "ctan_towardzero": +double: 3 +float: 2 + +Function: Real part of "ctan_upward": +double: 2 +float: 4 + +Function: Imaginary part of "ctan_upward": +double: 2 +float: 1 + +Function: Real part of "ctanh": +double: 2 +float: 2 + +Function: Imaginary part of "ctanh": +double: 2 +float: 2 + +Function: Real part of "ctanh_downward": +double: 3 +float: 2 + +Function: Imaginary part of "ctanh_downward": +double: 6 +float: 5 + +Function: Real part of "ctanh_towardzero": +double: 3 +float: 2 + +Function: Imaginary part of "ctanh_towardzero": +double: 5 +float: 3 + +Function: Real part of "ctanh_upward": +double: 2 +float: 2 + +Function: Imaginary part of "ctanh_upward": +double: 2 +float: 3 + +Function: "erf": +double: 1 +float: 1 + +Function: "erf_downward": +double: 1 +float: 1 + +Function: "erf_towardzero": +double: 1 +float: 1 + +Function: "erf_upward": +double: 1 +float: 1 + +Function: "erfc": +double: 5 +float: 3 + +Function: "erfc_downward": +double: 5 +float: 6 + +Function: "erfc_towardzero": +double: 3 +float: 4 + +Function: "erfc_upward": +double: 5 +float: 6 + +Function: "exp": +double: 1 +float: 1 + +Function: "exp10": +double: 2 +float: 1 + +Function: "exp10_downward": +double: 3 +float: 1 + +Function: "exp10_towardzero": +double: 3 +float: 1 + +Function: "exp10_upward": +double: 2 +float: 1 + +Function: "exp2": +double: 1 + +Function: "exp2_downward": +double: 1 + +Function: "exp2_towardzero": +double: 1 + +Function: "exp2_upward": +double: 1 +float: 1 + +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 + +Function: "expm1_downward": +double: 1 +float: 1 + +Function: "expm1_towardzero": +double: 1 +float: 2 + +Function: "expm1_upward": +double: 1 +float: 1 + +Function: "gamma": +double: 4 +float: 7 + +Function: "gamma_downward": +double: 7 +float: 7 + +Function: "gamma_towardzero": +double: 7 +float: 7 + +Function: "gamma_upward": +double: 5 +float: 6 + +Function: "hypot": +double: 1 +float: 1 + +Function: "hypot_downward": +double: 1 + +Function: "hypot_towardzero": +double: 1 + +Function: "hypot_upward": +double: 1 + +Function: "j0": +double: 2 +float: 9 + +Function: "j0_downward": +double: 5 +float: 9 + +Function: "j0_towardzero": +double: 6 +float: 9 + +Function: "j0_upward": +double: 9 +float: 9 + +Function: "j1": +double: 4 +float: 9 + +Function: "j1_downward": +double: 5 +float: 8 + +Function: "j1_towardzero": +double: 4 +float: 8 + +Function: "j1_upward": +double: 9 +float: 9 + +Function: "jn": +double: 4 +float: 4 + +Function: "jn_downward": +double: 7 +float: 9 + +Function: "jn_towardzero": +double: 7 +float: 9 + +Function: "jn_upward": +double: 7 +float: 9 + +Function: "lgamma": +double: 4 +float: 7 + +Function: "lgamma_downward": +double: 7 +float: 7 + +Function: "lgamma_towardzero": +double: 7 +float: 7 + +Function: "lgamma_upward": +double: 5 +float: 6 + +Function: "log10": +double: 2 +float: 2 + +Function: "log10_downward": +double: 2 +float: 3 + +Function: "log10_towardzero": +double: 2 +float: 1 + +Function: "log10_upward": +double: 2 +float: 2 + +Function: "log1p": +double: 1 +float: 1 + +Function: "log1p_downward": +double: 2 +float: 2 + +Function: "log1p_towardzero": +double: 2 +float: 2 + +Function: "log1p_upward": +double: 2 +float: 2 + +Function: "log2": +float: 1 + +Function: "log2_downward": +double: 1 + +Function: "log2_towardzero": +double: 1 + +Function: "log2_upward": +double: 1 + +Function: "pow": +double: 1 + +Function: "pow_downward": +double: 1 +float: 1 + +Function: "pow_towardzero": +double: 1 +float: 1 + +Function: "pow_upward": +double: 1 +float: 1 + +Function: "sin": +double: 1 +float: 1 + +Function: "sin_downward": +double: 4 +float: 1 + +Function: "sin_towardzero": +double: 3 +float: 1 + +Function: "sin_upward": +double: 7 +float: 1 + +Function: "sincos": +double: 1 + +Function: "sincos_downward": +double: 1 +float: 1 + +Function: "sincos_towardzero": +double: 4 +float: 1 + +Function: "sincos_upward": +double: 1 +float: 1 + +Function: "sinh": +double: 2 +float: 2 + +Function: "sinh_downward": +double: 3 +float: 3 + +Function: "sinh_towardzero": +double: 3 +float: 2 + +Function: "sinh_upward": +double: 3 +float: 3 + +Function: "tan": +float: 1 + +Function: "tan_downward": +double: 1 +float: 2 + +Function: "tan_towardzero": +double: 1 +float: 1 + +Function: "tan_upward": +double: 1 +float: 1 + +Function: "tanh": +double: 2 +float: 2 + +Function: "tanh_downward": +double: 3 +float: 3 + +Function: "tanh_towardzero": +double: 2 +float: 2 + +Function: "tanh_upward": +double: 3 +float: 3 + +Function: "tgamma": +double: 9 +float: 8 + +Function: "tgamma_downward": +double: 9 +float: 9 + +Function: "tgamma_towardzero": +double: 9 +float: 8 + +Function: "tgamma_upward": +double: 9 +float: 8 + +Function: "y0": +double: 3 +float: 9 + +Function: "y0_downward": +double: 3 +float: 9 + +Function: "y0_towardzero": +double: 4 +float: 9 + +Function: "y0_upward": +double: 3 +float: 9 + +Function: "y1": +double: 3 +float: 9 + +Function: "y1_downward": +double: 6 +float: 9 + +Function: "y1_towardzero": +double: 3 +float: 9 + +Function: "y1_upward": +double: 7 +float: 9 + +Function: "yn": +double: 3 +float: 3 + +Function: "yn_downward": +double: 6 +float: 8 + +Function: "yn_towardzero": +double: 8 +float: 8 + +Function: "yn_upward": +double: 8 +float: 8 + +# end of automatic generation diff --git a/sysdeps/or1k/fpu/libm-test-ulps-name b/sysdeps/or1k/fpu/libm-test-ulps-name new file mode 100644 index 0000000000..7f72f7a873 --- /dev/null +++ b/sysdeps/or1k/fpu/libm-test-ulps-name @@ -0,0 +1 @@ +OpenRISC hard-float diff --git a/sysdeps/or1k/fpu_control.h b/sysdeps/or1k/fpu_control.h new file mode 100644 index 0000000000..4ac03db7a9 --- /dev/null +++ b/sysdeps/or1k/fpu_control.h @@ -0,0 +1,61 @@ +/* FPU control word bits. OpenRISC version. + Copyright (C) 2024 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 _FPU_CONTROL_H +#define _FPU_CONTROL_H + +#ifndef __or1k_hard_float__ + +# define _FPU_RESERVED 0xffffffff +# define _FPU_DEFAULT 0x00000000 +# define _FPU_GETCW(cw) (cw) = 0 +# define _FPU_SETCW(cw) (void) (cw) + +#else /* __or1k_hard_float__ */ + +/* Layout of FPCSR: + + +-----------+----------------------------+-----+----+ + | 32 - 12 | 11 10 9 8 7 6 5 4 3 | 2-1 | 0 | + +-----------+----------------------------+-----+----+ + | Reserved | DZ IN IV IX Z QN SN UN OV | RM | EE | + +-----------+----------------------------+-----+----+ + + */ + +# define _FPU_RESERVED 0xfffff000 +/* Default: rounding to nearest with exceptions disabled. */ +# define _FPU_DEFAULT 0 +/* IEEE: Same as above with exceptions enabled. */ +# define _FPU_IEEE (_FPU_DEFAULT | 1) + +# define _FPU_FPCSR_RM_MASK (0x3 << 1) + +/* Macros for accessing the hardware control word. */ +# define _FPU_GETCW(cw) __asm__ volatile ("l.mfspr %0,r0,20" : "=r" (cw)) +# define _FPU_SETCW(cw) __asm__ volatile ("l.mtspr r0,%0,20" : : "r" (cw)) + +#endif /* __or1k_hard_float__ */ + +/* Type of the control word. */ +typedef unsigned int fpu_control_t; + +/* Default control word set at startup. */ +extern fpu_control_t __fpu_control; + +#endif /* fpu_control.h */ diff --git a/sysdeps/or1k/libm-test-ulps-name b/sysdeps/or1k/libm-test-ulps-name deleted file mode 100644 index 0af6591fd9..0000000000 --- a/sysdeps/or1k/libm-test-ulps-name +++ /dev/null @@ -1 +0,0 @@ -OpenRISC diff --git a/sysdeps/or1k/math-tests-snan-payload.h b/sysdeps/or1k/math-tests-snan-payload.h new file mode 100644 index 0000000000..62467a371c --- /dev/null +++ b/sysdeps/or1k/math-tests-snan-payload.h @@ -0,0 +1,26 @@ +/* Configuration for math tests: sNaN payloads. OpenRISC version. + Copyright (C) 2024 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 OR1K_MATH_TESTS_SNAN_PAYLOAD_H +#define OR1K_MATH_TESTS_SNAN_PAYLOAD_H 1 + +/* OpenRISC floating-point instructions do not preserve NaN + payloads. */ +#define SNAN_TESTS_PRESERVE_PAYLOAD 0 + +#endif /* math-tests-snan-payload.h */ diff --git a/sysdeps/or1k/math-tests-trap.h b/sysdeps/or1k/math-tests-trap.h new file mode 100644 index 0000000000..a95b42d66d --- /dev/null +++ b/sysdeps/or1k/math-tests-trap.h @@ -0,0 +1,27 @@ +/* Configuration for math tests: support for enabling exception traps. + OpenRISC version. + Copyright (C) 2024 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 OR1K_MATH_TESTS_TRAP_H +#define OR1K_MATH_TESTS_TRAP_H 1 + +#include <fenv.h> + +#define EXCEPTION_ENABLE_SUPPORTED(EXCEPT) ((EXCEPT) == 0) + +#endif /* math-tests-trap.h */ diff --git a/sysdeps/or1k/libm-test-ulps b/sysdeps/or1k/nofpu/libm-test-ulps similarity index 100% rename from sysdeps/or1k/libm-test-ulps rename to sysdeps/or1k/nofpu/libm-test-ulps diff --git a/sysdeps/or1k/nofpu/libm-test-ulps-name b/sysdeps/or1k/nofpu/libm-test-ulps-name new file mode 100644 index 0000000000..76c66a0e53 --- /dev/null +++ b/sysdeps/or1k/nofpu/libm-test-ulps-name @@ -0,0 +1 @@ +OpenRISC soft-float diff --git a/sysdeps/or1k/sfp-machine.h b/sysdeps/or1k/sfp-machine.h index d17fd37730..e58e683969 100644 --- a/sysdeps/or1k/sfp-machine.h +++ b/sysdeps/or1k/sfp-machine.h @@ -36,7 +36,6 @@ #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_loop(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) @@ -90,4 +89,21 @@ #define FP_ROUNDMODE (_fpcsr & FP_RND_MASK) +#ifdef __or1k_hard_float__ +#define FP_INIT_ROUNDMODE \ +do { \ + __asm__ volatile ("l.mfspr %0,r0,20" : "=r" (_fpcsr)); \ +} while (0) + +#define FP_HANDLE_EXCEPTIONS \ +do { \ + if (__builtin_expect (_fex, 0)) \ + { \ + _fpcsr &= ~FP_EX_ALL; \ + _fpcsr |= _fex; \ + __asm__ volatile ("l.mtspr r0,%0,20" : : "r" (_fpcsr)); \ + } \ +} while (0) +#endif /* __or1k_hard_float__ */ + #define _FP_TININESS_AFTER_ROUNDING 0 diff --git a/sysdeps/unix/sysv/linux/or1k/getcontext.S b/sysdeps/unix/sysv/linux/or1k/getcontext.S index a25b377bda..397b9e4c60 100644 --- a/sysdeps/unix/sysv/linux/or1k/getcontext.S +++ b/sysdeps/unix/sysv/linux/or1k/getcontext.S @@ -46,6 +46,12 @@ ENTRY(__getcontext) l.sw (UCONTEXT_MCONTEXT + 28*4)(r3), r28 l.sw (UCONTEXT_MCONTEXT + 30*4)(r3), r30 +#ifdef __or1k_hard_float__ + /* Store the floating point state. */ + l.mfspr r4, r0, 20 + l.sw (MCONTEXT_FPCSR)(r3), r4 +#endif /* __or1k_hard_float__ */ + /* Get signal mask. */ /* rt_sigprocmask (SIG_BLOCK, NULL, &ucp->uc_sigmask, _NSIG8) */ l.ori r6, r0, _NSIG8 diff --git a/sysdeps/unix/sysv/linux/or1k/setcontext.S b/sysdeps/unix/sysv/linux/or1k/setcontext.S index d28a0ac0aa..4bf258b6f1 100644 --- a/sysdeps/unix/sysv/linux/or1k/setcontext.S +++ b/sysdeps/unix/sysv/linux/or1k/setcontext.S @@ -40,6 +40,12 @@ ENTRY(__setcontext) l.bf 1f l.nop +#ifdef __or1k_hard_float__ + /* Restore the floating point state. */ + l.lwz r28, (MCONTEXT_FPCSR)(r30) + l.mtspr r0, r28, 20 +#endif /* __or1k_hard_float__ */ + /* Restore argument registers, for the makecontext case. */ l.lwz r3, (UCONTEXT_MCONTEXT + 3*4)(r30) l.lwz r4, (UCONTEXT_MCONTEXT + 4*4)(r30) diff --git a/sysdeps/unix/sysv/linux/or1k/swapcontext.S b/sysdeps/unix/sysv/linux/or1k/swapcontext.S index d09651a5b2..dfc788812c 100644 --- a/sysdeps/unix/sysv/linux/or1k/swapcontext.S +++ b/sysdeps/unix/sysv/linux/or1k/swapcontext.S @@ -45,6 +45,12 @@ ENTRY(__swapcontext) l.sw (UCONTEXT_MCONTEXT + 28*4)(r3), r28 l.sw (UCONTEXT_MCONTEXT + 30*4)(r3), r30 +#ifdef __or1k_hard_float__ + /* Store the floating point state. */ + l.mfspr r6, r0, 20 + l.sw (MCONTEXT_FPCSR)(r3), r6 +#endif /* __or1k_hard_float__ */ + /* Store ucp to non-argument syscall preserved register. */ l.ori r30, r4, 0 @@ -82,6 +88,12 @@ ENTRY(__swapcontext) l.bf 1f l.nop +#ifdef __or1k_hard_float__ + /* Restore the floating point state. */ + l.lwz r28, (MCONTEXT_FPCSR)(r30) + l.mtspr r0, r28, 20 +#endif /* __or1k_hard_float__ */ + /* Restore argument registers, for the makecontext case. */ l.lwz r3, (UCONTEXT_MCONTEXT + 3*4)(r30) l.lwz r4, (UCONTEXT_MCONTEXT + 4*4)(r30) diff --git a/sysdeps/unix/sysv/linux/or1k/sys/ucontext.h b/sysdeps/unix/sysv/linux/or1k/sys/ucontext.h index b17e919154..1b428592ee 100644 --- a/sysdeps/unix/sysv/linux/or1k/sys/ucontext.h +++ b/sysdeps/unix/sysv/linux/or1k/sys/ucontext.h @@ -38,6 +38,7 @@ typedef struct unsigned long int __gprs[__NGREG]; unsigned long int __pc; unsigned long int __sr; + unsigned long int __fpcsr; } mcontext_t; /* Userlevel context. */ diff --git a/sysdeps/unix/sysv/linux/or1k/ucontext_i.sym b/sysdeps/unix/sysv/linux/or1k/ucontext_i.sym index a8d4db080f..45cd72527d 100644 --- a/sysdeps/unix/sysv/linux/or1k/ucontext_i.sym +++ b/sysdeps/unix/sysv/linux/or1k/ucontext_i.sym @@ -13,6 +13,7 @@ _NSIG8 (_NSIG / 8) -- Offsets of the fields in the ucontext_t structure. #define ucontext(member) offsetof (ucontext_t, member) #define stack(member) ucontext (uc_stack.member) +#define mcontext(member) ucontext (uc_mcontext.member) UCONTEXT_LINK ucontext (uc_link) UCONTEXT_STACK ucontext (uc_stack) @@ -23,4 +24,6 @@ STACK_SP stack (ss_sp) STACK_SIZE stack (ss_size) STACK_FLAGS stack (ss_flags) +MCONTEXT_FPCSR mcontext (__fpcsr) + UCONTEXT_SIZE sizeof (ucontext_t) -- 2.44.0 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH 1/2] or1k: Add hard float support 2024-03-29 7:40 ` [PATCH 1/2] or1k: Add " Stafford Horne @ 2024-04-16 14:04 ` Adhemerval Zanella Netto 2024-04-16 14:53 ` Richard Henderson 0 siblings, 1 reply; 8+ messages in thread From: Adhemerval Zanella Netto @ 2024-04-16 14:04 UTC (permalink / raw) To: Stafford Horne, GLIBC patches; +Cc: Linux OpenRISC On 29/03/24 04:40, Stafford Horne wrote: > This patch adds hardware floating point support to OpenRISC. Hardware > floating point toolchain builds are enabled by passing the machine > specific argument -mhard-float to gcc via CFLAGS. With this enabled GCC > generates floating point instructions for single-precision operations > and exports __or1k_hard_float__. > > There are 2 main parts to this patch. > > - Implement fenv functions to update the FPCSR flags keeping it in sync > with sfp (software floating point). > - Update machine context functions to store and restore the FPCSR > state. > > *On mcontext_t ABI* > > This patch adds __fpcsr to mcontext_t. This is an ABI change, but also > an ABI fix. The linux kernel has always defined padding in mcontext_t > that was was missing from the glibc ABI. This unused space is has now > been re-purposed for storing the FPCSR. This patch brings OpenRISC > glibc in line with the Linux kernel and other libc implementation > (musl). > > *Hard float ABI* > > The calling conventions and types do not change with OpenRISC hard-float > so glibc hard-float builds continue to use dynamic linker > /lib/ld-linux-or1k.so.1. > > *Testing* > > I have tested this patch both with hard-float and soft-float builds and > the test results look fine to me. Results are as follows: > > Hard Float > > # failures > FAIL: elf/tst-sprof-basic (Not checked, seen in other ports too) > FAIL: gmon/tst-gmon-pie (GCC PIE bug) > FAIL: gmon/tst-gmon-pie-gprof (GCC PIE bug) > FAIL: nptl/tst-cond24 (timeouts) > FAIL: nptl/tst-mutex10 (timeouts) > > # summary > 5 FAIL > 4279 PASS > 83 UNSUPPORTED > 16 XFAIL > 2 XPASS > > # versions > Toolchain: or1k-smhfpu-linux-gnu > Compiler: gcc version 14.0.1 20240324 (experimental) [master r14-9649-gbb04a11418f] (GCC) > Binutils: GNU assembler version 2.42.0 (or1k-smhfpu-linux-gnu) using BFD version (GNU Binutils) 2.42.0.20240324 > Linux: Linux buildroot 6.8.0-rc5-00003-g9158c6253560-dirty #78 SMP Fri Mar 22 16:56:54 GMT 2024 openrisc GNU/Linux > > Soft Float > > # failures > FAIL: elf/tst-sprof-basic > FAIL: gmon/tst-gmon-pie > FAIL: gmon/tst-gmon-pie-gprof > FAIL: iconvdata/iconv-test (timeout, passed when run manually) > FAIL: nptl/tst-mutex10 > > # summary > 5 FAIL > 4281 PASS > 81 UNSUPPORTED > 16 XFAIL > 2 XPASS > > # versions > Toolchain: or1k-smh-linux-gnu > Compiler: gcc version 14.0.1 20240324 (experimental) [master r14-9649-gbb04a11418f] (GCC) > Binutils: GNU assembler version 2.42.0 (or1k-smh-linux-gnu) using BFD version (GNU Binutils) 2.42.0.20240324 > Linux: Linux buildroot 6.8.0-rc5-00003-g9158c6253560-dirty #78 SMP Fri Mar 22 16:56:54 GMT 2024 openrisc GNU/Linux > > Documentation: https://raw.githubusercontent.com/openrisc/doc/master/openrisc-arch-1.4-rev0.pdf If I understond correctly the OpenRISC hard-float ABIs uses the same parameter passing ABI from soft-fp, similar of what ARM does with -mfloat-abi=softfp. So the the shared objects built with/without -mhard-float should be interoperable, assuming hardware has hard-fp support. However it does solve the potential problem a binary built with softfp loading a hardfp built shared object, or even trying to link with a static object with different ABI (and assuming a hardware without hard-fp support).. It means that loader and ldconfig won't recognize shared objects with different floating-point objects if they are installed on the same system. I don't think we have a strong policy regarding this, and historical there were ABI variants that followed this (like powerpc soft and hard), but most ABIs that support soft/hard floating-point usually advertise it through ElfXX_Ehdr::e_flags. For instance, with arm: $ cat << EOF > lib.c float foo_float (float x, float y) { return x + y; } double foo_double (double x, double y) { return x + y; } EOF $ arm-glibc-linux-gnueabi-gcc -shared -o lib.so lib.c $ readelf -h lib.so | grep Flags Flags: 0x5000200, Version5 EABI, soft-float ABI $ arm-glibc-linux-gnueabihf-gcc -shared -o lib.so lib.c $ readelf -h lib.so | grep Flags Flags: 0x5000400, Version5 EABI, hard-float ABI This is similar for RISC-V rv64imac-lp64 vs rv64imafdc-lp64d, where the former also defined a soft-float ABI (Flags 0x1) while the latter a a hard-float (Flags: 0x5). With a different ABI, it prevents possible issues with mcontext_t change and related functions (for instance, when a hard-float binary runs on a soft-fp built libc.so and start to see that the fpcsr is not save/restore). And it seems that OpenRISCV also support a double-precision (-mdouble-float), but I think it does not really require a distinction between single and double (as RISC-V does and it would require a different ABI for support it [1]). [1] https://sourceware.org/bugzilla/show_bug.cgi?id=30963 > --- > sysdeps/or1k/fpu/fclrexcpt.c | 44 + > sysdeps/or1k/fpu/fegetenv.c | 32 + > sysdeps/or1k/fpu/fegetmode.c | 29 + > sysdeps/or1k/fpu/fegetround.c | 29 + > sysdeps/or1k/fpu/feholdexcpt.c | 33 + > sysdeps/or1k/fpu/fenv_private.h | 199 +++ > sysdeps/or1k/fpu/fesetenv.c | 32 + > sysdeps/or1k/fpu/fesetexcept.c | 35 + > sysdeps/or1k/fpu/fesetmode.c | 39 + > sysdeps/or1k/fpu/fesetround.c | 39 + > sysdeps/or1k/fpu/feupdateenv.c | 33 + > sysdeps/or1k/fpu/fgetexcptflg.c | 29 + > .../or1k/fpu/fix-fp-int-convert-overflow.h | 38 + > sysdeps/or1k/fpu/fraiseexcpt.c | 67 + > sysdeps/or1k/fpu/fsetexcptflg.c | 43 + > sysdeps/or1k/fpu/ftestexcept.c | 27 + > sysdeps/or1k/fpu/get-rounding-mode.h | 38 + > sysdeps/or1k/fpu/libm-test-ulps | 1115 +++++++++++++++++ > sysdeps/or1k/fpu/libm-test-ulps-name | 1 + > sysdeps/or1k/fpu_control.h | 61 + > sysdeps/or1k/libm-test-ulps-name | 1 - > sysdeps/or1k/math-tests-snan-payload.h | 26 + > sysdeps/or1k/math-tests-trap.h | 27 + > sysdeps/or1k/{ => nofpu}/libm-test-ulps | 0 > sysdeps/or1k/nofpu/libm-test-ulps-name | 1 + > sysdeps/or1k/sfp-machine.h | 18 +- > sysdeps/unix/sysv/linux/or1k/getcontext.S | 6 + > sysdeps/unix/sysv/linux/or1k/setcontext.S | 6 + > sysdeps/unix/sysv/linux/or1k/swapcontext.S | 12 + > sysdeps/unix/sysv/linux/or1k/sys/ucontext.h | 1 + > sysdeps/unix/sysv/linux/or1k/ucontext_i.sym | 3 + > 31 files changed, 2062 insertions(+), 2 deletions(-) > create mode 100644 sysdeps/or1k/fpu/fclrexcpt.c > create mode 100644 sysdeps/or1k/fpu/fegetenv.c > create mode 100644 sysdeps/or1k/fpu/fegetmode.c > create mode 100644 sysdeps/or1k/fpu/fegetround.c > create mode 100644 sysdeps/or1k/fpu/feholdexcpt.c > create mode 100644 sysdeps/or1k/fpu/fenv_private.h > create mode 100644 sysdeps/or1k/fpu/fesetenv.c > create mode 100644 sysdeps/or1k/fpu/fesetexcept.c > create mode 100644 sysdeps/or1k/fpu/fesetmode.c > create mode 100644 sysdeps/or1k/fpu/fesetround.c > create mode 100644 sysdeps/or1k/fpu/feupdateenv.c > create mode 100644 sysdeps/or1k/fpu/fgetexcptflg.c > create mode 100644 sysdeps/or1k/fpu/fix-fp-int-convert-overflow.h > create mode 100644 sysdeps/or1k/fpu/fraiseexcpt.c > create mode 100644 sysdeps/or1k/fpu/fsetexcptflg.c > create mode 100644 sysdeps/or1k/fpu/ftestexcept.c > create mode 100644 sysdeps/or1k/fpu/get-rounding-mode.h > create mode 100644 sysdeps/or1k/fpu/libm-test-ulps > create mode 100644 sysdeps/or1k/fpu/libm-test-ulps-name > create mode 100644 sysdeps/or1k/fpu_control.h > delete mode 100644 sysdeps/or1k/libm-test-ulps-name > create mode 100644 sysdeps/or1k/math-tests-snan-payload.h > create mode 100644 sysdeps/or1k/math-tests-trap.h > rename sysdeps/or1k/{ => nofpu}/libm-test-ulps (100%) > create mode 100644 sysdeps/or1k/nofpu/libm-test-ulps-name > > diff --git a/sysdeps/or1k/fpu/fclrexcpt.c b/sysdeps/or1k/fpu/fclrexcpt.c > new file mode 100644 > index 0000000000..44224f9c24 > --- /dev/null > +++ b/sysdeps/or1k/fpu/fclrexcpt.c > @@ -0,0 +1,44 @@ > +/* Clear given exceptions in current floating-point environment. > + OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fpu_control.h> > + > +int > +feclearexcept (int excepts) > +{ > + fpu_control_t cw; > + fpu_control_t cw_new; > + > + /* Mask out unsupported bits/exceptions. */ > + excepts &= FE_ALL_EXCEPT; > + > + /* Read the complete control word. */ > + _FPU_GETCW (cw); > + > + cw_new = cw & ~excepts; > + > + /* Put the new data in effect. */ > + if (cw != cw_new) > + _FPU_SETCW (cw_new); > + > + /* Success. */ > + return 0; > +} > +libm_hidden_def (feclearexcept) > diff --git a/sysdeps/or1k/fpu/fegetenv.c b/sysdeps/or1k/fpu/fegetenv.c > new file mode 100644 > index 0000000000..70c75aa0bf > --- /dev/null > +++ b/sysdeps/or1k/fpu/fegetenv.c > @@ -0,0 +1,32 @@ > +/* Store current floating-point environment. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fpu_control.h> > + > +int > +__fegetenv (fenv_t *envp) > +{ > + _FPU_GETCW (*envp); > + > + /* Success. */ > + return 0; > +} > +libm_hidden_def (__fegetenv) > +weak_alias (__fegetenv, fegetenv) > +libm_hidden_weak (fegetenv) > diff --git a/sysdeps/or1k/fpu/fegetmode.c b/sysdeps/or1k/fpu/fegetmode.c > new file mode 100644 > index 0000000000..7fffd2e0b5 > --- /dev/null > +++ b/sysdeps/or1k/fpu/fegetmode.c > @@ -0,0 +1,29 @@ > +/* Store current floating-point control modes. OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fpu_control.h> > + > +int > +fegetmode (femode_t *modep) > +{ > + _FPU_GETCW (*modep); > + > + /* Success. */ > + return 0; > +} > diff --git a/sysdeps/or1k/fpu/fegetround.c b/sysdeps/or1k/fpu/fegetround.c > new file mode 100644 > index 0000000000..7e993b980a > --- /dev/null > +++ b/sysdeps/or1k/fpu/fegetround.c > @@ -0,0 +1,29 @@ > +/* Return current rounding direction. OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <get-rounding-mode.h> > + > +int > +__fegetround (void) > +{ > + return get_rounding_mode (); > +} > +libm_hidden_def (__fegetround) > +weak_alias (__fegetround, fegetround) > +libm_hidden_weak (fegetround) > diff --git a/sysdeps/or1k/fpu/feholdexcpt.c b/sysdeps/or1k/fpu/feholdexcpt.c > new file mode 100644 > index 0000000000..0036e41ba2 > --- /dev/null > +++ b/sysdeps/or1k/fpu/feholdexcpt.c > @@ -0,0 +1,33 @@ > +/* Store current floating-point environment and clear exceptions. > + OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fenv_private.h> > + > +int > +__feholdexcept (fenv_t *envp) > +{ > + libc_feholdexcept_or1k (envp); > + > + /* Success. */ > + return 0; > +} > +libm_hidden_def (__feholdexcept) > +weak_alias (__feholdexcept, feholdexcept) > +libm_hidden_weak (feholdexcept) > diff --git a/sysdeps/or1k/fpu/fenv_private.h b/sysdeps/or1k/fpu/fenv_private.h > new file mode 100644 > index 0000000000..4f401e7a5a > --- /dev/null > +++ b/sysdeps/or1k/fpu/fenv_private.h > @@ -0,0 +1,199 @@ > +/* Private floating point rounding and exceptions handling. OpenRISC version. > + Copyright (C) 2024 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 OR1K_FENV_PRIVATE_H > +#define OR1K_FENV_PRIVATE_H 1 > + > +#include <fenv.h> > +#include <fpu_control.h> > + > +static __always_inline void > +libc_feholdexcept_or1k (fenv_t *envp) > +{ > + fpu_control_t cw; > + fpu_control_t cw_new; > + > + /* Get and store the environment. */ > + _FPU_GETCW (cw); > + *envp = cw; > + > + /* Clear the exception status flags. */ > + cw_new = cw & ~FE_ALL_EXCEPT; > + > + if (cw != cw_new) > + _FPU_SETCW (cw_new); > +} > + > +#define libc_feholdexcept libc_feholdexcept_or1k > +#define libc_feholdexceptf libc_feholdexcept_or1k > +#define libc_feholdexceptl libc_feholdexcept_or1k > + > +static __always_inline void > +libc_fesetround_or1k (int round) > +{ > + fpu_control_t cw; > + fpu_control_t cw_new; > + > + _FPU_GETCW (cw); > + cw_new = cw & ~_FPU_FPCSR_RM_MASK; > + cw_new |= round; > + if (cw != cw_new) > + _FPU_SETCW (cw_new); > +} > + > +#define libc_fesetround libc_fesetround_or1k > +#define libc_fesetroundf libc_fesetround_or1k > +#define libc_fesetroundl libc_fesetround_or1k > + > +static __always_inline void > +libc_feholdexcept_setround_or1k (fenv_t *envp, int round) > +{ > + fpu_control_t cw; > + fpu_control_t cw_new; > + > + /* Get and store the environment. */ > + _FPU_GETCW (cw); > + *envp = cw; > + > + /* Clear the status flags and rounding mode. */ > + cw_new = cw & ~(FE_ALL_EXCEPT | _FPU_FPCSR_RM_MASK); > + > + /* Set rounding mode. */ > + cw_new |= round; > + > + if (cw != cw_new) > + _FPU_SETCW (cw_new); > +} > + > +#define libc_feholdexcept_setround libc_feholdexcept_setround_or1k > +#define libc_feholdexcept_setroundf libc_feholdexcept_setround_or1k > +#define libc_feholdexcept_setroundl libc_feholdexcept_setround_or1k > + > +static __always_inline int > +libc_fetestexcept_or1k (int ex) > +{ > + fpu_control_t cw; > + > + /* Get current control word. */ > + _FPU_GETCW (cw); > + > + /* Check if any of the queried exception flags are set. */ > + return cw & ex & FE_ALL_EXCEPT; > +} > + > +#define libc_fetestexcept libc_fetestexcept_or1k > +#define libc_fetestexceptf libc_fetestexcept_or1k > +#define libc_fetestexceptl libc_fetestexcept_or1k > + > +static __always_inline void > +libc_fesetenv_or1k (const fenv_t *envp) > +{ > + if (envp == FE_DFL_ENV) > + _FPU_SETCW (_FPU_DEFAULT); > + else > + _FPU_SETCW (*envp); > +} > + > +#define libc_fesetenv libc_fesetenv_or1k > +#define libc_fesetenvf libc_fesetenv_or1k > +#define libc_fesetenvl libc_fesetenv_or1k > +#define libc_feresetround_noex libc_fesetenv_or1k > +#define libc_feresetround_noexf libc_fesetenv_or1k > +#define libc_feresetround_noexl libc_fesetenv_or1k > + > +static __always_inline int > +libc_feupdateenv_test_or1k (const fenv_t *envp, int ex) > +{ > + fpu_control_t cw; > + fpu_control_t cw_new; > + int excepts; > + > + /* Get current control word. */ > + _FPU_GETCW (cw); > + > + /* Merge current exception flags with the passed fenv. */ > + excepts = cw & FE_ALL_EXCEPT; > + cw_new = (envp == FE_DFL_ENV ? _FPU_DEFAULT : *envp) | excepts; > + > + if (__glibc_unlikely (cw != cw_new)) > + _FPU_SETCW (cw_new); > + > + /* Raise the exceptions if enabled in the new FP state. */ > + if (__glibc_unlikely (excepts)) > + __feraiseexcept (excepts); > + > + return excepts & ex; > +} > + > +#define libc_feupdateenv_test libc_feupdateenv_test_or1k > +#define libc_feupdateenv_testf libc_feupdateenv_test_or1k > +#define libc_feupdateenv_testl libc_feupdateenv_test_or1k > + > +static __always_inline void > +libc_feupdateenv_or1k (const fenv_t *envp) > +{ > + libc_feupdateenv_test_or1k (envp, 0); > +} > + > +#define libc_feupdateenv libc_feupdateenv_or1k > +#define libc_feupdateenvf libc_feupdateenv_or1k > +#define libc_feupdateenvl libc_feupdateenv_or1k > + > +static __always_inline void > +libc_feholdsetround_or1k (fenv_t *envp, int round) > +{ > + fpu_control_t cw; > + > + _FPU_GETCW (cw); > + *envp = cw; > + > + /* Check whether rounding modes are different. */ > + round = (cw ^ round) & _FPU_FPCSR_RM_MASK; > + > + /* Set new rounding mode if different. */ > + if (__glibc_unlikely (round != 0)) > + _FPU_SETCW (cw ^ round); > +} > + > +#define libc_feholdsetround libc_feholdsetround_or1k > +#define libc_feholdsetroundf libc_feholdsetround_or1k > +#define libc_feholdsetroundl libc_feholdsetround_or1k > + > +static __always_inline void > +libc_feresetround_or1k (fenv_t *envp) > +{ > + fpu_control_t cw; > + int round; > + > + _FPU_GETCW (cw); > + > + /* Check whether rounding modes are different. */ > + round = (*envp ^ cw) & _FPU_FPCSR_RM_MASK; > + > + /* Restore the rounding mode if it was changed. */ > + if (__glibc_unlikely (round != 0)) > + _FPU_SETCW (cw ^ round); > +} > + > +#define libc_feresetround libc_feresetround_or1k > +#define libc_feresetroundf libc_feresetround_or1k > +#define libc_feresetroundl libc_feresetround_or1k > + > +#include_next <fenv_private.h> > + > +#endif > diff --git a/sysdeps/or1k/fpu/fesetenv.c b/sysdeps/or1k/fpu/fesetenv.c > new file mode 100644 > index 0000000000..742ca719e0 > --- /dev/null > +++ b/sysdeps/or1k/fpu/fesetenv.c > @@ -0,0 +1,32 @@ > +/* Install given floating-point environment. OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fenv_private.h> > + > +int > +__fesetenv (const fenv_t *envp) > +{ > + libc_fesetenv_or1k (envp); > + > + /* Success. */ > + return 0; > +} > +libm_hidden_def (__fesetenv) > +weak_alias (__fesetenv, fesetenv) > +libm_hidden_weak (fesetenv) > diff --git a/sysdeps/or1k/fpu/fesetexcept.c b/sysdeps/or1k/fpu/fesetexcept.c > new file mode 100644 > index 0000000000..43734eac18 > --- /dev/null > +++ b/sysdeps/or1k/fpu/fesetexcept.c > @@ -0,0 +1,35 @@ > +/* Set given exception flags. OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fpu_control.h> > + > +int > +fesetexcept (int excepts) > +{ > + fpu_control_t cw; > + fpu_control_t cw_new; > + > + _FPU_GETCW (cw); > + cw_new = cw | (excepts & FE_ALL_EXCEPT); > + if (cw != cw_new) > + _FPU_SETCW (cw_new); > + > + /* Success. */ > + return 0; > +} > diff --git a/sysdeps/or1k/fpu/fesetmode.c b/sysdeps/or1k/fpu/fesetmode.c > new file mode 100644 > index 0000000000..d4556927ce > --- /dev/null > +++ b/sysdeps/or1k/fpu/fesetmode.c > @@ -0,0 +1,39 @@ > +/* Install given floating-point control modes. OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fpu_control.h> > + > +int > +fesetmode (const femode_t *modep) > +{ > + fpu_control_t cw; > + fpu_control_t cw_new; > + > + _FPU_GETCW (cw); > + cw_new = cw & ~_FPU_FPCSR_RM_MASK; > + if (modep == FE_DFL_MODE) > + cw_new |= (_FPU_DEFAULT & _FPU_FPCSR_RM_MASK); > + else > + cw_new |= (*modep & _FPU_FPCSR_RM_MASK); > + if (cw != cw_new) > + _FPU_SETCW (cw_new); > + > + /* Success. */ > + return 0; > +} > diff --git a/sysdeps/or1k/fpu/fesetround.c b/sysdeps/or1k/fpu/fesetround.c > new file mode 100644 > index 0000000000..c2ada98f1b > --- /dev/null > +++ b/sysdeps/or1k/fpu/fesetround.c > @@ -0,0 +1,39 @@ > +/* Set current rounding direction. OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fenv_private.h> > + > +int > +__fesetround (int round) > +{ > + switch (round) > + { > + case FE_TONEAREST: > + case FE_TOWARDZERO: > + case FE_DOWNWARD: > + case FE_UPWARD: > + libc_fesetround_or1k (round); > + return 0; > + default: > + return round; /* A nonzero value. */ > + } > +} > +libm_hidden_def (__fesetround) > +weak_alias (__fesetround, fesetround) > +libm_hidden_weak (fesetround) > diff --git a/sysdeps/or1k/fpu/feupdateenv.c b/sysdeps/or1k/fpu/feupdateenv.c > new file mode 100644 > index 0000000000..3355bf6596 > --- /dev/null > +++ b/sysdeps/or1k/fpu/feupdateenv.c > @@ -0,0 +1,33 @@ > +/* Install given floating-point environment and raise exceptions. > + OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fenv_private.h> > + > +int > +__feupdateenv (const fenv_t *envp) > +{ > + libc_feupdateenv_or1k (envp); > + > + /* Success. */ > + return 0; > +} > +libm_hidden_def (__feupdateenv) > +weak_alias (__feupdateenv, feupdateenv) > +libm_hidden_weak (feupdateenv) > diff --git a/sysdeps/or1k/fpu/fgetexcptflg.c b/sysdeps/or1k/fpu/fgetexcptflg.c > new file mode 100644 > index 0000000000..a954f6a2f1 > --- /dev/null > +++ b/sysdeps/or1k/fpu/fgetexcptflg.c > @@ -0,0 +1,29 @@ > +/* Store current state of exceptions. OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fenv_private.h> > + > +int > +fegetexceptflag (fexcept_t *flagp, int excepts) > +{ > + *flagp = libc_fetestexcept_or1k (excepts); > + > + /* Success. */ > + return 0; > +} > diff --git a/sysdeps/or1k/fpu/fix-fp-int-convert-overflow.h b/sysdeps/or1k/fpu/fix-fp-int-convert-overflow.h > new file mode 100644 > index 0000000000..78104289c0 > --- /dev/null > +++ b/sysdeps/or1k/fpu/fix-fp-int-convert-overflow.h > @@ -0,0 +1,38 @@ > +/* Fix for conversion of floating point to integer overflow. OpenRISC version. > + Copyright (C) 2024 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 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 */ > diff --git a/sysdeps/or1k/fpu/fraiseexcpt.c b/sysdeps/or1k/fpu/fraiseexcpt.c > new file mode 100644 > index 0000000000..bbacfd50bc > --- /dev/null > +++ b/sysdeps/or1k/fpu/fraiseexcpt.c > @@ -0,0 +1,67 @@ > +/* Raise given exceptions. OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fpu_control.h> > +#include <float.h> > +#include <math.h> > + > +int > +__feraiseexcept (int excepts) > +{ > + if (excepts == 0) > + return 0; > + > + /* Raise exceptions represented by EXPECTS. */ > + > + if (excepts & FE_INEXACT) > + { > + float d = 1.0, x = 3.0; > + __asm__ volatile ("lf.div.s %0, %0, %1" : "+r" (d) : "r" (x)); > + } > + > + if (excepts & FE_UNDERFLOW) > + { > + float d = FLT_MIN; > + __asm__ volatile ("lf.mul.s %0, %0, %0" : "+r" (d)); > + } > + > + if (excepts & FE_OVERFLOW) > + { > + float d = FLT_MAX; > + __asm__ volatile ("lf.mul.s %0, %0, %0" : "+r" (d) : "r" (d)); > + } > + > + if (excepts & FE_DIVBYZERO) > + { > + float d = 1.0, x = 0.0; > + __asm__ volatile ("lf.div.s %0, %0, %1" : "+r" (d) : "r" (x)); > + } > + > + if (excepts & FE_INVALID) > + { > + float d = HUGE_VAL, x = 0.0; > + __asm__ volatile ("lf.mul.s %0, %1, %0" : "+r" (d) : "r" (x)); > + } > + > + /* Success. */ > + return 0; > +} > +libm_hidden_def (__feraiseexcept) > +weak_alias (__feraiseexcept, feraiseexcept) > +libm_hidden_weak (feraiseexcept) > diff --git a/sysdeps/or1k/fpu/fsetexcptflg.c b/sysdeps/or1k/fpu/fsetexcptflg.c > new file mode 100644 > index 0000000000..c327e4c5d1 > --- /dev/null > +++ b/sysdeps/or1k/fpu/fsetexcptflg.c > @@ -0,0 +1,43 @@ > +/* Set floating-point environment exception handling. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fpu_control.h> > + > +int > +fesetexceptflag (const fexcept_t *flagp, int excepts) > +{ > + fpu_control_t cw; > + fpu_control_t cw_new; > + > + /* Get the current exceptions. */ > + _FPU_GETCW (cw); > + > + /* Make sure the flags we want restored are legal. */ > + excepts &= FE_ALL_EXCEPT; > + > + /* Now set selected bits from flagp. Note that we ignore all non-flag > + bits from *flagp, so they don't matter. */ > + cw_new = (cw & ~excepts) | (*flagp & excepts); > + > + if (cw != cw_new) > + _FPU_SETCW (cw_new); > + > + /* Success. */ > + return 0; > +} > diff --git a/sysdeps/or1k/fpu/ftestexcept.c b/sysdeps/or1k/fpu/ftestexcept.c > new file mode 100644 > index 0000000000..59f06afa22 > --- /dev/null > +++ b/sysdeps/or1k/fpu/ftestexcept.c > @@ -0,0 +1,27 @@ > +/* Test exception in current environment. OpenRISC version. > + Copyright (C) 2024 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/>. */ > + > +#include <fenv.h> > +#include <fenv_private.h> > + > +int > +fetestexcept (int excepts) > +{ > + return libc_fetestexcept_or1k (excepts); > +} > +libm_hidden_def (fetestexcept) > diff --git a/sysdeps/or1k/fpu/get-rounding-mode.h b/sysdeps/or1k/fpu/get-rounding-mode.h > new file mode 100644 > index 0000000000..a66d553be8 > --- /dev/null > +++ b/sysdeps/or1k/fpu/get-rounding-mode.h > @@ -0,0 +1,38 @@ > +/* Determine floating-point rounding mode within libc. OpenRISC version. > + > + Copyright (C) 2024 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 _OR1K_GET_ROUNDING_MODE_H > +#define _OR1K_GET_ROUNDING_MODE_H 1 > + > +#include <fenv.h> > +#include <fpu_control.h> > + > +/* Return the floating-point rounding mode. */ > + > +static inline int > +get_rounding_mode (void) > +{ > + fpu_control_t cw; > + > + _FPU_GETCW (cw); > + return cw & _FPU_FPCSR_RM_MASK; > +} > + > +#endif /* get-rounding-mode.h */ > diff --git a/sysdeps/or1k/fpu/libm-test-ulps b/sysdeps/or1k/fpu/libm-test-ulps > new file mode 100644 > index 0000000000..948ec01cdc > --- /dev/null > +++ b/sysdeps/or1k/fpu/libm-test-ulps > @@ -0,0 +1,1115 @@ > +# Begin of automatic generation > + > +# Maximal error of functions: > +Function: "acos": > +double: 1 > +float: 1 > + > +Function: "acos_downward": > +double: 1 > +float: 1 > + > +Function: "acos_towardzero": > +double: 1 > +float: 1 > + > +Function: "acos_upward": > +double: 1 > +float: 1 > + > +Function: "acosh": > +double: 2 > +float: 2 > + > +Function: "acosh_downward": > +double: 2 > +float: 2 > + > +Function: "acosh_towardzero": > +double: 2 > +float: 2 > + > +Function: "acosh_upward": > +double: 2 > +float: 2 > + > +Function: "asin": > +double: 1 > +float: 1 > + > +Function: "asin_downward": > +double: 1 > +float: 1 > + > +Function: "asin_towardzero": > +double: 1 > +float: 1 > + > +Function: "asin_upward": > +double: 2 > +float: 1 > + > +Function: "asinh": > +double: 2 > +float: 2 > + > +Function: "asinh_downward": > +double: 3 > +float: 3 > + > +Function: "asinh_towardzero": > +double: 2 > +float: 2 > + > +Function: "asinh_upward": > +double: 3 > +float: 3 > + > +Function: "atan": > +double: 1 > +float: 1 > + > +Function: "atan2": > +float: 2 > + > +Function: "atan2_downward": > +double: 5 > +float: 2 > + > +Function: "atan2_towardzero": > +double: 5 > +float: 2 > + > +Function: "atan2_upward": > +double: 8 > +float: 2 > + > +Function: "atan_downward": > +double: 1 > +float: 2 > + > +Function: "atan_towardzero": > +double: 1 > +float: 1 > + > +Function: "atan_upward": > +double: 1 > +float: 2 > + > +Function: "atanh": > +double: 2 > +float: 2 > + > +Function: "atanh_downward": > +double: 3 > +float: 3 > + > +Function: "atanh_towardzero": > +double: 2 > +float: 2 > + > +Function: "atanh_upward": > +double: 3 > +float: 3 > + > +Function: "cabs": > +double: 1 > + > +Function: "cabs_downward": > +double: 1 > + > +Function: "cabs_towardzero": > +double: 1 > + > +Function: "cabs_upward": > +double: 1 > + > +Function: Real part of "cacos": > +double: 1 > +float: 2 > + > +Function: Imaginary part of "cacos": > +double: 2 > +float: 2 > + > +Function: Real part of "cacos_downward": > +double: 3 > +float: 2 > + > +Function: Imaginary part of "cacos_downward": > +double: 5 > +float: 3 > + > +Function: Real part of "cacos_towardzero": > +double: 3 > +float: 2 > + > +Function: Imaginary part of "cacos_towardzero": > +double: 4 > +float: 2 > + > +Function: Real part of "cacos_upward": > +double: 2 > +float: 2 > + > +Function: Imaginary part of "cacos_upward": > +double: 5 > +float: 7 > + > +Function: Real part of "cacosh": > +double: 2 > +float: 2 > + > +Function: Imaginary part of "cacosh": > +double: 1 > +float: 2 > + > +Function: Real part of "cacosh_downward": > +double: 4 > +float: 2 > + > +Function: Imaginary part of "cacosh_downward": > +double: 3 > +float: 3 > + > +Function: Real part of "cacosh_towardzero": > +double: 4 > +float: 2 > + > +Function: Imaginary part of "cacosh_towardzero": > +double: 3 > +float: 2 > + > +Function: Real part of "cacosh_upward": > +double: 4 > +float: 3 > + > +Function: Imaginary part of "cacosh_upward": > +double: 3 > +float: 2 > + > +Function: "carg": > +float: 1 > + > +Function: "carg_downward": > +double: 5 > +float: 2 > + > +Function: "carg_towardzero": > +double: 5 > +float: 2 > + > +Function: "carg_upward": > +double: 8 > +float: 2 > + > +Function: Real part of "casin": > +double: 1 > +float: 1 > + > +Function: Imaginary part of "casin": > +double: 2 > +float: 2 > + > +Function: Real part of "casin_downward": > +double: 3 > +float: 2 > + > +Function: Imaginary part of "casin_downward": > +double: 5 > +float: 3 > + > +Function: Real part of "casin_towardzero": > +double: 3 > +float: 1 > + > +Function: Imaginary part of "casin_towardzero": > +double: 4 > +float: 2 > + > +Function: Real part of "casin_upward": > +double: 3 > +float: 2 > + > +Function: Imaginary part of "casin_upward": > +double: 5 > +float: 7 > + > +Function: Real part of "casinh": > +double: 2 > +float: 2 > + > +Function: Imaginary part of "casinh": > +double: 1 > +float: 1 > + > +Function: Real part of "casinh_downward": > +double: 5 > +float: 3 > + > +Function: Imaginary part of "casinh_downward": > +double: 3 > +float: 2 > + > +Function: Real part of "casinh_towardzero": > +double: 4 > +float: 2 > + > +Function: Imaginary part of "casinh_towardzero": > +double: 3 > +float: 1 > + > +Function: Real part of "casinh_upward": > +double: 5 > +float: 7 > + > +Function: Imaginary part of "casinh_upward": > +double: 3 > +float: 2 > + > +Function: Real part of "catan": > +double: 1 > +float: 1 > + > +Function: Imaginary part of "catan": > +double: 1 > +float: 1 > + > +Function: Real part of "catan_downward": > +double: 1 > +float: 2 > + > +Function: Imaginary part of "catan_downward": > +double: 2 > +float: 2 > + > +Function: Real part of "catan_towardzero": > +double: 1 > +float: 2 > + > +Function: Imaginary part of "catan_towardzero": > +double: 2 > +float: 2 > + > +Function: Real part of "catan_upward": > +double: 2 > +float: 1 > + > +Function: Imaginary part of "catan_upward": > +double: 2 > +float: 2 > + > +Function: Real part of "catanh": > +double: 1 > +float: 1 > + > +Function: Imaginary part of "catanh": > +double: 1 > +float: 1 > + > +Function: Real part of "catanh_downward": > +double: 2 > +float: 2 > + > +Function: Imaginary part of "catanh_downward": > +double: 1 > +float: 2 > + > +Function: Real part of "catanh_towardzero": > +double: 2 > +float: 2 > + > +Function: Imaginary part of "catanh_towardzero": > +double: 1 > +float: 2 > + > +Function: Real part of "catanh_upward": > +double: 4 > +float: 4 > + > +Function: Imaginary part of "catanh_upward": > +double: 2 > +float: 1 > + > +Function: "cbrt": > +double: 4 > +float: 1 > + > +Function: "cbrt_downward": > +double: 4 > +float: 1 > + > +Function: "cbrt_towardzero": > +double: 3 > +float: 1 > + > +Function: "cbrt_upward": > +double: 5 > +float: 1 > + > +Function: Real part of "ccos": > +double: 1 > +float: 1 > + > +Function: Imaginary part of "ccos": > +double: 1 > +float: 1 > + > +Function: Real part of "ccos_downward": > +double: 3 > +float: 1 > + > +Function: Imaginary part of "ccos_downward": > +double: 3 > +float: 3 > + > +Function: Real part of "ccos_towardzero": > +double: 3 > +float: 2 > + > +Function: Imaginary part of "ccos_towardzero": > +double: 3 > +float: 3 > + > +Function: Real part of "ccos_upward": > +double: 1 > +float: 2 > + > +Function: Imaginary part of "ccos_upward": > +double: 2 > +float: 2 > + > +Function: Real part of "ccosh": > +double: 2 > +float: 1 > + > +Function: Imaginary part of "ccosh": > +double: 1 > +float: 1 > + > +Function: Real part of "ccosh_downward": > +double: 3 > +float: 2 > + > +Function: Imaginary part of "ccosh_downward": > +double: 3 > +float: 3 > + > +Function: Real part of "ccosh_towardzero": > +double: 3 > +float: 3 > + > +Function: Imaginary part of "ccosh_towardzero": > +double: 3 > +float: 3 > + > +Function: Real part of "ccosh_upward": > +double: 1 > +float: 2 > + > +Function: Imaginary part of "ccosh_upward": > +double: 2 > +float: 2 > + > +Function: Real part of "cexp": > +double: 2 > +float: 1 > + > +Function: Imaginary part of "cexp": > +double: 1 > +float: 2 > + > +Function: Real part of "cexp_downward": > +double: 4 > +float: 2 > + > +Function: Imaginary part of "cexp_downward": > +double: 3 > +float: 3 > + > +Function: Real part of "cexp_towardzero": > +double: 4 > +float: 2 > + > +Function: Imaginary part of "cexp_towardzero": > +double: 3 > +float: 3 > + > +Function: Real part of "cexp_upward": > +double: 2 > +float: 2 > + > +Function: Imaginary part of "cexp_upward": > +double: 3 > +float: 2 > + > +Function: Real part of "clog": > +double: 3 > +float: 3 > + > +Function: Imaginary part of "clog": > +double: 1 > +float: 1 > + > +Function: Real part of "clog10": > +double: 3 > +float: 4 > + > +Function: Imaginary part of "clog10": > +double: 2 > +float: 2 > + > +Function: Real part of "clog10_downward": > +double: 7 > +float: 5 > + > +Function: Imaginary part of "clog10_downward": > +double: 8 > +float: 4 > + > +Function: Real part of "clog10_towardzero": > +double: 5 > +float: 5 > + > +Function: Imaginary part of "clog10_towardzero": > +double: 8 > +float: 4 > + > +Function: Real part of "clog10_upward": > +double: 6 > +float: 5 > + > +Function: Imaginary part of "clog10_upward": > +double: 5 > +float: 4 > + > +Function: Real part of "clog_downward": > +double: 4 > +float: 3 > + > +Function: Imaginary part of "clog_downward": > +double: 5 > +float: 2 > + > +Function: Real part of "clog_towardzero": > +double: 8 > +float: 4 > + > +Function: Imaginary part of "clog_towardzero": > +double: 5 > +float: 3 > + > +Function: Real part of "clog_upward": > +double: 8 > +float: 3 > + > +Function: Imaginary part of "clog_upward": > +double: 8 > +float: 2 > + > +Function: "cos": > +double: 1 > +float: 1 > + > +Function: "cos_downward": > +double: 1 > +float: 1 > + > +Function: "cos_towardzero": > +double: 4 > +float: 1 > + > +Function: "cos_upward": > +double: 4 > +float: 1 > + > +Function: "cosh": > +double: 2 > +float: 2 > + > +Function: "cosh_downward": > +double: 3 > +float: 1 > + > +Function: "cosh_towardzero": > +double: 3 > +float: 1 > + > +Function: "cosh_upward": > +double: 2 > +float: 2 > + > +Function: Real part of "cpow": > +double: 2 > +float: 5 > + > +Function: Imaginary part of "cpow": > +float: 2 > + > +Function: Real part of "cpow_downward": > +double: 5 > +float: 8 > + > +Function: Imaginary part of "cpow_downward": > +double: 2 > +float: 2 > + > +Function: Real part of "cpow_towardzero": > +double: 5 > +float: 8 > + > +Function: Imaginary part of "cpow_towardzero": > +double: 2 > +float: 2 > + > +Function: Real part of "cpow_upward": > +double: 4 > +float: 1 > + > +Function: Imaginary part of "cpow_upward": > +double: 2 > +float: 2 > + > +Function: Real part of "csin": > +double: 1 > +float: 1 > + > +Function: Real part of "csin_downward": > +double: 3 > +float: 3 > + > +Function: Imaginary part of "csin_downward": > +double: 3 > +float: 1 > + > +Function: Real part of "csin_towardzero": > +double: 3 > +float: 3 > + > +Function: Imaginary part of "csin_towardzero": > +double: 3 > +float: 1 > + > +Function: Real part of "csin_upward": > +double: 2 > +float: 2 > + > +Function: Imaginary part of "csin_upward": > +double: 1 > +float: 2 > + > +Function: Real part of "csinh": > +float: 1 > + > +Function: Imaginary part of "csinh": > +double: 1 > +float: 1 > + > +Function: Real part of "csinh_downward": > +double: 3 > +float: 1 > + > +Function: Imaginary part of "csinh_downward": > +double: 3 > +float: 3 > + > +Function: Real part of "csinh_towardzero": > +double: 3 > +float: 2 > + > +Function: Imaginary part of "csinh_towardzero": > +double: 3 > +float: 3 > + > +Function: Real part of "csinh_upward": > +double: 1 > +float: 2 > + > +Function: Imaginary part of "csinh_upward": > +double: 2 > +float: 2 > + > +Function: Real part of "csqrt": > +double: 2 > +float: 2 > + > +Function: Imaginary part of "csqrt": > +double: 2 > +float: 2 > + > +Function: Real part of "csqrt_downward": > +double: 5 > +float: 4 > + > +Function: Imaginary part of "csqrt_downward": > +double: 4 > +float: 3 > + > +Function: Real part of "csqrt_towardzero": > +double: 4 > +float: 3 > + > +Function: Imaginary part of "csqrt_towardzero": > +double: 4 > +float: 3 > + > +Function: Real part of "csqrt_upward": > +double: 5 > +float: 4 > + > +Function: Imaginary part of "csqrt_upward": > +double: 3 > +float: 3 > + > +Function: Real part of "ctan": > +double: 1 > +float: 1 > + > +Function: Imaginary part of "ctan": > +double: 2 > +float: 2 > + > +Function: Real part of "ctan_downward": > +double: 6 > +float: 5 > + > +Function: Imaginary part of "ctan_downward": > +double: 3 > +float: 2 > + > +Function: Real part of "ctan_towardzero": > +double: 5 > +float: 3 > + > +Function: Imaginary part of "ctan_towardzero": > +double: 3 > +float: 2 > + > +Function: Real part of "ctan_upward": > +double: 2 > +float: 4 > + > +Function: Imaginary part of "ctan_upward": > +double: 2 > +float: 1 > + > +Function: Real part of "ctanh": > +double: 2 > +float: 2 > + > +Function: Imaginary part of "ctanh": > +double: 2 > +float: 2 > + > +Function: Real part of "ctanh_downward": > +double: 3 > +float: 2 > + > +Function: Imaginary part of "ctanh_downward": > +double: 6 > +float: 5 > + > +Function: Real part of "ctanh_towardzero": > +double: 3 > +float: 2 > + > +Function: Imaginary part of "ctanh_towardzero": > +double: 5 > +float: 3 > + > +Function: Real part of "ctanh_upward": > +double: 2 > +float: 2 > + > +Function: Imaginary part of "ctanh_upward": > +double: 2 > +float: 3 > + > +Function: "erf": > +double: 1 > +float: 1 > + > +Function: "erf_downward": > +double: 1 > +float: 1 > + > +Function: "erf_towardzero": > +double: 1 > +float: 1 > + > +Function: "erf_upward": > +double: 1 > +float: 1 > + > +Function: "erfc": > +double: 5 > +float: 3 > + > +Function: "erfc_downward": > +double: 5 > +float: 6 > + > +Function: "erfc_towardzero": > +double: 3 > +float: 4 > + > +Function: "erfc_upward": > +double: 5 > +float: 6 > + > +Function: "exp": > +double: 1 > +float: 1 > + > +Function: "exp10": > +double: 2 > +float: 1 > + > +Function: "exp10_downward": > +double: 3 > +float: 1 > + > +Function: "exp10_towardzero": > +double: 3 > +float: 1 > + > +Function: "exp10_upward": > +double: 2 > +float: 1 > + > +Function: "exp2": > +double: 1 > + > +Function: "exp2_downward": > +double: 1 > + > +Function: "exp2_towardzero": > +double: 1 > + > +Function: "exp2_upward": > +double: 1 > +float: 1 > + > +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 > + > +Function: "expm1_downward": > +double: 1 > +float: 1 > + > +Function: "expm1_towardzero": > +double: 1 > +float: 2 > + > +Function: "expm1_upward": > +double: 1 > +float: 1 > + > +Function: "gamma": > +double: 4 > +float: 7 > + > +Function: "gamma_downward": > +double: 7 > +float: 7 > + > +Function: "gamma_towardzero": > +double: 7 > +float: 7 > + > +Function: "gamma_upward": > +double: 5 > +float: 6 > + > +Function: "hypot": > +double: 1 > +float: 1 > + > +Function: "hypot_downward": > +double: 1 > + > +Function: "hypot_towardzero": > +double: 1 > + > +Function: "hypot_upward": > +double: 1 > + > +Function: "j0": > +double: 2 > +float: 9 > + > +Function: "j0_downward": > +double: 5 > +float: 9 > + > +Function: "j0_towardzero": > +double: 6 > +float: 9 > + > +Function: "j0_upward": > +double: 9 > +float: 9 > + > +Function: "j1": > +double: 4 > +float: 9 > + > +Function: "j1_downward": > +double: 5 > +float: 8 > + > +Function: "j1_towardzero": > +double: 4 > +float: 8 > + > +Function: "j1_upward": > +double: 9 > +float: 9 > + > +Function: "jn": > +double: 4 > +float: 4 > + > +Function: "jn_downward": > +double: 7 > +float: 9 > + > +Function: "jn_towardzero": > +double: 7 > +float: 9 > + > +Function: "jn_upward": > +double: 7 > +float: 9 > + > +Function: "lgamma": > +double: 4 > +float: 7 > + > +Function: "lgamma_downward": > +double: 7 > +float: 7 > + > +Function: "lgamma_towardzero": > +double: 7 > +float: 7 > + > +Function: "lgamma_upward": > +double: 5 > +float: 6 > + > +Function: "log10": > +double: 2 > +float: 2 > + > +Function: "log10_downward": > +double: 2 > +float: 3 > + > +Function: "log10_towardzero": > +double: 2 > +float: 1 > + > +Function: "log10_upward": > +double: 2 > +float: 2 > + > +Function: "log1p": > +double: 1 > +float: 1 > + > +Function: "log1p_downward": > +double: 2 > +float: 2 > + > +Function: "log1p_towardzero": > +double: 2 > +float: 2 > + > +Function: "log1p_upward": > +double: 2 > +float: 2 > + > +Function: "log2": > +float: 1 > + > +Function: "log2_downward": > +double: 1 > + > +Function: "log2_towardzero": > +double: 1 > + > +Function: "log2_upward": > +double: 1 > + > +Function: "pow": > +double: 1 > + > +Function: "pow_downward": > +double: 1 > +float: 1 > + > +Function: "pow_towardzero": > +double: 1 > +float: 1 > + > +Function: "pow_upward": > +double: 1 > +float: 1 > + > +Function: "sin": > +double: 1 > +float: 1 > + > +Function: "sin_downward": > +double: 4 > +float: 1 > + > +Function: "sin_towardzero": > +double: 3 > +float: 1 > + > +Function: "sin_upward": > +double: 7 > +float: 1 > + > +Function: "sincos": > +double: 1 > + > +Function: "sincos_downward": > +double: 1 > +float: 1 > + > +Function: "sincos_towardzero": > +double: 4 > +float: 1 > + > +Function: "sincos_upward": > +double: 1 > +float: 1 > + > +Function: "sinh": > +double: 2 > +float: 2 > + > +Function: "sinh_downward": > +double: 3 > +float: 3 > + > +Function: "sinh_towardzero": > +double: 3 > +float: 2 > + > +Function: "sinh_upward": > +double: 3 > +float: 3 > + > +Function: "tan": > +float: 1 > + > +Function: "tan_downward": > +double: 1 > +float: 2 > + > +Function: "tan_towardzero": > +double: 1 > +float: 1 > + > +Function: "tan_upward": > +double: 1 > +float: 1 > + > +Function: "tanh": > +double: 2 > +float: 2 > + > +Function: "tanh_downward": > +double: 3 > +float: 3 > + > +Function: "tanh_towardzero": > +double: 2 > +float: 2 > + > +Function: "tanh_upward": > +double: 3 > +float: 3 > + > +Function: "tgamma": > +double: 9 > +float: 8 > + > +Function: "tgamma_downward": > +double: 9 > +float: 9 > + > +Function: "tgamma_towardzero": > +double: 9 > +float: 8 > + > +Function: "tgamma_upward": > +double: 9 > +float: 8 > + > +Function: "y0": > +double: 3 > +float: 9 > + > +Function: "y0_downward": > +double: 3 > +float: 9 > + > +Function: "y0_towardzero": > +double: 4 > +float: 9 > + > +Function: "y0_upward": > +double: 3 > +float: 9 > + > +Function: "y1": > +double: 3 > +float: 9 > + > +Function: "y1_downward": > +double: 6 > +float: 9 > + > +Function: "y1_towardzero": > +double: 3 > +float: 9 > + > +Function: "y1_upward": > +double: 7 > +float: 9 > + > +Function: "yn": > +double: 3 > +float: 3 > + > +Function: "yn_downward": > +double: 6 > +float: 8 > + > +Function: "yn_towardzero": > +double: 8 > +float: 8 > + > +Function: "yn_upward": > +double: 8 > +float: 8 > + > +# end of automatic generation > diff --git a/sysdeps/or1k/fpu/libm-test-ulps-name b/sysdeps/or1k/fpu/libm-test-ulps-name > new file mode 100644 > index 0000000000..7f72f7a873 > --- /dev/null > +++ b/sysdeps/or1k/fpu/libm-test-ulps-name > @@ -0,0 +1 @@ > +OpenRISC hard-float > diff --git a/sysdeps/or1k/fpu_control.h b/sysdeps/or1k/fpu_control.h > new file mode 100644 > index 0000000000..4ac03db7a9 > --- /dev/null > +++ b/sysdeps/or1k/fpu_control.h > @@ -0,0 +1,61 @@ > +/* FPU control word bits. OpenRISC version. > + Copyright (C) 2024 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 _FPU_CONTROL_H > +#define _FPU_CONTROL_H > + > +#ifndef __or1k_hard_float__ > + > +# define _FPU_RESERVED 0xffffffff > +# define _FPU_DEFAULT 0x00000000 > +# define _FPU_GETCW(cw) (cw) = 0 > +# define _FPU_SETCW(cw) (void) (cw) > + > +#else /* __or1k_hard_float__ */ > + > +/* Layout of FPCSR: > + > + +-----------+----------------------------+-----+----+ > + | 32 - 12 | 11 10 9 8 7 6 5 4 3 | 2-1 | 0 | > + +-----------+----------------------------+-----+----+ > + | Reserved | DZ IN IV IX Z QN SN UN OV | RM | EE | > + +-----------+----------------------------+-----+----+ > + > + */ > + > +# define _FPU_RESERVED 0xfffff000 > +/* Default: rounding to nearest with exceptions disabled. */ > +# define _FPU_DEFAULT 0 > +/* IEEE: Same as above with exceptions enabled. */ > +# define _FPU_IEEE (_FPU_DEFAULT | 1) > + > +# define _FPU_FPCSR_RM_MASK (0x3 << 1) > + > +/* Macros for accessing the hardware control word. */ > +# define _FPU_GETCW(cw) __asm__ volatile ("l.mfspr %0,r0,20" : "=r" (cw)) > +# define _FPU_SETCW(cw) __asm__ volatile ("l.mtspr r0,%0,20" : : "r" (cw)) > + > +#endif /* __or1k_hard_float__ */ > + > +/* Type of the control word. */ > +typedef unsigned int fpu_control_t; > + > +/* Default control word set at startup. */ > +extern fpu_control_t __fpu_control; > + > +#endif /* fpu_control.h */ > diff --git a/sysdeps/or1k/libm-test-ulps-name b/sysdeps/or1k/libm-test-ulps-name > deleted file mode 100644 > index 0af6591fd9..0000000000 > --- a/sysdeps/or1k/libm-test-ulps-name > +++ /dev/null > @@ -1 +0,0 @@ > -OpenRISC > diff --git a/sysdeps/or1k/math-tests-snan-payload.h b/sysdeps/or1k/math-tests-snan-payload.h > new file mode 100644 > index 0000000000..62467a371c > --- /dev/null > +++ b/sysdeps/or1k/math-tests-snan-payload.h > @@ -0,0 +1,26 @@ > +/* Configuration for math tests: sNaN payloads. OpenRISC version. > + Copyright (C) 2024 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 OR1K_MATH_TESTS_SNAN_PAYLOAD_H > +#define OR1K_MATH_TESTS_SNAN_PAYLOAD_H 1 > + > +/* OpenRISC floating-point instructions do not preserve NaN > + payloads. */ > +#define SNAN_TESTS_PRESERVE_PAYLOAD 0 > + > +#endif /* math-tests-snan-payload.h */ > diff --git a/sysdeps/or1k/math-tests-trap.h b/sysdeps/or1k/math-tests-trap.h > new file mode 100644 > index 0000000000..a95b42d66d > --- /dev/null > +++ b/sysdeps/or1k/math-tests-trap.h > @@ -0,0 +1,27 @@ > +/* Configuration for math tests: support for enabling exception traps. > + OpenRISC version. > + Copyright (C) 2024 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 OR1K_MATH_TESTS_TRAP_H > +#define OR1K_MATH_TESTS_TRAP_H 1 > + > +#include <fenv.h> > + > +#define EXCEPTION_ENABLE_SUPPORTED(EXCEPT) ((EXCEPT) == 0) > + > +#endif /* math-tests-trap.h */ > diff --git a/sysdeps/or1k/libm-test-ulps b/sysdeps/or1k/nofpu/libm-test-ulps > similarity index 100% > rename from sysdeps/or1k/libm-test-ulps > rename to sysdeps/or1k/nofpu/libm-test-ulps > diff --git a/sysdeps/or1k/nofpu/libm-test-ulps-name b/sysdeps/or1k/nofpu/libm-test-ulps-name > new file mode 100644 > index 0000000000..76c66a0e53 > --- /dev/null > +++ b/sysdeps/or1k/nofpu/libm-test-ulps-name > @@ -0,0 +1 @@ > +OpenRISC soft-float > diff --git a/sysdeps/or1k/sfp-machine.h b/sysdeps/or1k/sfp-machine.h > index d17fd37730..e58e683969 100644 > --- a/sysdeps/or1k/sfp-machine.h > +++ b/sysdeps/or1k/sfp-machine.h > @@ -36,7 +36,6 @@ > #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_loop(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) > @@ -90,4 +89,21 @@ > > #define FP_ROUNDMODE (_fpcsr & FP_RND_MASK) > > +#ifdef __or1k_hard_float__ > +#define FP_INIT_ROUNDMODE \ > +do { \ > + __asm__ volatile ("l.mfspr %0,r0,20" : "=r" (_fpcsr)); \ > +} while (0) > + > +#define FP_HANDLE_EXCEPTIONS \ > +do { \ > + if (__builtin_expect (_fex, 0)) \ > + { \ > + _fpcsr &= ~FP_EX_ALL; \ > + _fpcsr |= _fex; \ > + __asm__ volatile ("l.mtspr r0,%0,20" : : "r" (_fpcsr)); \ > + } \ > +} while (0) > +#endif /* __or1k_hard_float__ */ > + > #define _FP_TININESS_AFTER_ROUNDING 0 > diff --git a/sysdeps/unix/sysv/linux/or1k/getcontext.S b/sysdeps/unix/sysv/linux/or1k/getcontext.S > index a25b377bda..397b9e4c60 100644 > --- a/sysdeps/unix/sysv/linux/or1k/getcontext.S > +++ b/sysdeps/unix/sysv/linux/or1k/getcontext.S > @@ -46,6 +46,12 @@ ENTRY(__getcontext) > l.sw (UCONTEXT_MCONTEXT + 28*4)(r3), r28 > l.sw (UCONTEXT_MCONTEXT + 30*4)(r3), r30 > > +#ifdef __or1k_hard_float__ > + /* Store the floating point state. */ > + l.mfspr r4, r0, 20 > + l.sw (MCONTEXT_FPCSR)(r3), r4 > +#endif /* __or1k_hard_float__ */ > + > /* Get signal mask. */ > /* rt_sigprocmask (SIG_BLOCK, NULL, &ucp->uc_sigmask, _NSIG8) */ > l.ori r6, r0, _NSIG8 > diff --git a/sysdeps/unix/sysv/linux/or1k/setcontext.S b/sysdeps/unix/sysv/linux/or1k/setcontext.S > index d28a0ac0aa..4bf258b6f1 100644 > --- a/sysdeps/unix/sysv/linux/or1k/setcontext.S > +++ b/sysdeps/unix/sysv/linux/or1k/setcontext.S > @@ -40,6 +40,12 @@ ENTRY(__setcontext) > l.bf 1f > l.nop > > +#ifdef __or1k_hard_float__ > + /* Restore the floating point state. */ > + l.lwz r28, (MCONTEXT_FPCSR)(r30) > + l.mtspr r0, r28, 20 > +#endif /* __or1k_hard_float__ */ > + > /* Restore argument registers, for the makecontext case. */ > l.lwz r3, (UCONTEXT_MCONTEXT + 3*4)(r30) > l.lwz r4, (UCONTEXT_MCONTEXT + 4*4)(r30) > diff --git a/sysdeps/unix/sysv/linux/or1k/swapcontext.S b/sysdeps/unix/sysv/linux/or1k/swapcontext.S > index d09651a5b2..dfc788812c 100644 > --- a/sysdeps/unix/sysv/linux/or1k/swapcontext.S > +++ b/sysdeps/unix/sysv/linux/or1k/swapcontext.S > @@ -45,6 +45,12 @@ ENTRY(__swapcontext) > l.sw (UCONTEXT_MCONTEXT + 28*4)(r3), r28 > l.sw (UCONTEXT_MCONTEXT + 30*4)(r3), r30 > > +#ifdef __or1k_hard_float__ > + /* Store the floating point state. */ > + l.mfspr r6, r0, 20 > + l.sw (MCONTEXT_FPCSR)(r3), r6 > +#endif /* __or1k_hard_float__ */ > + > /* Store ucp to non-argument syscall preserved register. */ > l.ori r30, r4, 0 > > @@ -82,6 +88,12 @@ ENTRY(__swapcontext) > l.bf 1f > l.nop > > +#ifdef __or1k_hard_float__ > + /* Restore the floating point state. */ > + l.lwz r28, (MCONTEXT_FPCSR)(r30) > + l.mtspr r0, r28, 20 > +#endif /* __or1k_hard_float__ */ > + > /* Restore argument registers, for the makecontext case. */ > l.lwz r3, (UCONTEXT_MCONTEXT + 3*4)(r30) > l.lwz r4, (UCONTEXT_MCONTEXT + 4*4)(r30) > diff --git a/sysdeps/unix/sysv/linux/or1k/sys/ucontext.h b/sysdeps/unix/sysv/linux/or1k/sys/ucontext.h > index b17e919154..1b428592ee 100644 > --- a/sysdeps/unix/sysv/linux/or1k/sys/ucontext.h > +++ b/sysdeps/unix/sysv/linux/or1k/sys/ucontext.h > @@ -38,6 +38,7 @@ typedef struct > unsigned long int __gprs[__NGREG]; > unsigned long int __pc; > unsigned long int __sr; > + unsigned long int __fpcsr; > } mcontext_t; > > /* Userlevel context. */ > diff --git a/sysdeps/unix/sysv/linux/or1k/ucontext_i.sym b/sysdeps/unix/sysv/linux/or1k/ucontext_i.sym > index a8d4db080f..45cd72527d 100644 > --- a/sysdeps/unix/sysv/linux/or1k/ucontext_i.sym > +++ b/sysdeps/unix/sysv/linux/or1k/ucontext_i.sym > @@ -13,6 +13,7 @@ _NSIG8 (_NSIG / 8) > -- Offsets of the fields in the ucontext_t structure. > #define ucontext(member) offsetof (ucontext_t, member) > #define stack(member) ucontext (uc_stack.member) > +#define mcontext(member) ucontext (uc_mcontext.member) > > UCONTEXT_LINK ucontext (uc_link) > UCONTEXT_STACK ucontext (uc_stack) > @@ -23,4 +24,6 @@ STACK_SP stack (ss_sp) > STACK_SIZE stack (ss_size) > STACK_FLAGS stack (ss_flags) > > +MCONTEXT_FPCSR mcontext (__fpcsr) > + > UCONTEXT_SIZE sizeof (ucontext_t) ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH 1/2] or1k: Add hard float support 2024-04-16 14:04 ` Adhemerval Zanella Netto @ 2024-04-16 14:53 ` Richard Henderson 2024-04-16 16:20 ` Adhemerval Zanella Netto 0 siblings, 1 reply; 8+ messages in thread From: Richard Henderson @ 2024-04-16 14:53 UTC (permalink / raw) To: Adhemerval Zanella Netto, Stafford Horne, GLIBC patches; +Cc: Linux OpenRISC On 4/16/24 07:04, Adhemerval Zanella Netto wrote: > If I understond correctly the OpenRISC hard-float ABIs uses the same > parameter passing ABI from soft-fp, similar of what ARM does with > -mfloat-abi=softfp. So the the shared objects built with/without > -mhard-float should be interoperable, assuming hardware has hard-fp > support. > > However it does solve the potential problem a binary built with softfp > loading a hardfp built shared object, or even trying to link with a > static object with different ABI (and assuming a hardware without > hard-fp support).. It means that loader and ldconfig won't recognize > shared objects with different floating-point objects if they are installed > on the same system. > > I don't think we have a strong policy regarding this, and historical > there were ABI variants that followed this (like powerpc soft and hard), > but most ABIs that support soft/hard floating-point usually advertise it > through ElfXX_Ehdr::e_flags. > > For instance, with arm: > > $ cat << EOF > lib.c > float foo_float (float x, float y) > { > return x + y; > } > > double foo_double (double x, double y) > { > return x + y; > } > EOF > $ arm-glibc-linux-gnueabi-gcc -shared -o lib.so lib.c > $ readelf -h lib.so | grep Flags > Flags: 0x5000200, Version5 EABI, soft-float ABI > $ arm-glibc-linux-gnueabihf-gcc -shared -o lib.so lib.c > $ readelf -h lib.so | grep Flags > Flags: 0x5000400, Version5 EABI, hard-float ABI The typical reason for wanting an e_flags bit is because those architectures want the parameter passing to change when floating-point registers are available. But for OpenRISC there is no separate floating-point register set. All floating-point operations use the same general-purpose registers as for all integer operations. So there is no point in a new parameter passing ABI. r~ ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH 1/2] or1k: Add hard float support 2024-04-16 14:53 ` Richard Henderson @ 2024-04-16 16:20 ` Adhemerval Zanella Netto 2024-04-16 19:56 ` Richard Henderson 0 siblings, 1 reply; 8+ messages in thread From: Adhemerval Zanella Netto @ 2024-04-16 16:20 UTC (permalink / raw) To: Richard Henderson, Stafford Horne, GLIBC patches; +Cc: Linux OpenRISC On 16/04/24 11:53, Richard Henderson wrote: > On 4/16/24 07:04, Adhemerval Zanella Netto wrote: >> If I understond correctly the OpenRISC hard-float ABIs uses the same >> parameter passing ABI from soft-fp, similar of what ARM does with >> -mfloat-abi=softfp. So the the shared objects built with/without >> -mhard-float should be interoperable, assuming hardware has hard-fp >> support. >> >> However it does solve the potential problem a binary built with softfp >> loading a hardfp built shared object, or even trying to link with a >> static object with different ABI (and assuming a hardware without >> hard-fp support).. It means that loader and ldconfig won't recognize >> shared objects with different floating-point objects if they are installed >> on the same system. >> >> I don't think we have a strong policy regarding this, and historical >> there were ABI variants that followed this (like powerpc soft and hard), >> but most ABIs that support soft/hard floating-point usually advertise it >> through ElfXX_Ehdr::e_flags. >> >> For instance, with arm: >> >> $ cat << EOF > lib.c >> float foo_float (float x, float y) >> { >> return x + y; >> } >> >> double foo_double (double x, double y) >> { >> return x + y; >> } >> EOF >> $ arm-glibc-linux-gnueabi-gcc -shared -o lib.so lib.c >> $ readelf -h lib.so | grep Flags >> Flags: 0x5000200, Version5 EABI, soft-float ABI >> $ arm-glibc-linux-gnueabihf-gcc -shared -o lib.so lib.c >> $ readelf -h lib.so | grep Flags >> Flags: 0x5000400, Version5 EABI, hard-float ABI > > The typical reason for wanting an e_flags bit is because those architectures want the parameter passing to change when floating-point registers are available. > > But for OpenRISC there is no separate floating-point register set. All floating-point operations use the same general-purpose registers as for all integer operations. So there is no point in a new parameter passing ABI. Right, it was not clear from patch without dig into the ABI documen itself. So the only thing I am not sure is the mcontext_t change. Other ABIs added a symbol version to proper handle it, so maybe OpenRISC would need something similar. ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH 1/2] or1k: Add hard float support 2024-04-16 16:20 ` Adhemerval Zanella Netto @ 2024-04-16 19:56 ` Richard Henderson 2024-04-16 20:37 ` Stafford Horne 0 siblings, 1 reply; 8+ messages in thread From: Richard Henderson @ 2024-04-16 19:56 UTC (permalink / raw) To: Adhemerval Zanella Netto, Stafford Horne, GLIBC patches; +Cc: Linux OpenRISC On 4/16/24 09:20, Adhemerval Zanella Netto wrote: > Right, it was not clear from patch without dig into the ABI documen > itself. So the only thing I am not sure is the mcontext_t change. > Other ABIs added a symbol version to proper handle it, so maybe > OpenRISC would need something similar. Yes, I think some compat symbols are needed for getcontext et al, since the size of ucontext_t changes. I think softfloat glibc should populate the same field (with zero, since neither exceptions nor rounding are supported). r~ ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH 1/2] or1k: Add hard float support 2024-04-16 19:56 ` Richard Henderson @ 2024-04-16 20:37 ` Stafford Horne 0 siblings, 0 replies; 8+ messages in thread From: Stafford Horne @ 2024-04-16 20:37 UTC (permalink / raw) To: Richard Henderson; +Cc: Adhemerval Zanella Netto, GLIBC patches, Linux OpenRISC On Tue, Apr 16, 2024 at 12:56:11PM -0700, Richard Henderson wrote: > On 4/16/24 09:20, Adhemerval Zanella Netto wrote: > > Right, it was not clear from patch without dig into the ABI documen > > itself. So the only thing I am not sure is the mcontext_t change. > > Other ABIs added a symbol version to proper handle it, so maybe > > OpenRISC would need something similar. > > Yes, I think some compat symbols are needed for getcontext et al, since the > size of ucontext_t changes. Just to re-iterate, the size of ucontext_t provided by the kernel never changed. But the change of ucontext_t in glibc changing causes issues. With the glibc ucontext_t size change there will be a problem with running old compiled applications on new versions of glibc if this hard-float series is added as is. I will see if we can add some versioning to ensure binaries will not break when running on new glibc versions after this series. > I think softfloat glibc should populate the same field (with zero, since > neither exceptions nor rounding are supported). Let me double check this is the case. -Stafford ^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH 2/2] build-many-glibcs.py: Add openrisc hard float glibc variant 2024-03-29 7:40 [PATCH 0/2] OpenRISC glibc hard float support Stafford Horne 2024-03-29 7:40 ` [PATCH 1/2] or1k: Add " Stafford Horne @ 2024-03-29 7:40 ` Stafford Horne 1 sibling, 0 replies; 8+ messages in thread From: Stafford Horne @ 2024-03-29 7:40 UTC (permalink / raw) To: GLIBC patches; +Cc: Linux OpenRISC, Stafford Horne This adds the OpenRISC hard float glibc variant to the build many script. We update the compiler for glibc to support hard-float multilibs to allow us to use a single generic compiler for all glibc variants, this requires updating the compiler name. Tested and all builds are passing. --- scripts/build-many-glibcs.py | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/scripts/build-many-glibcs.py b/scripts/build-many-glibcs.py index ecc743e672..b3d4617f62 100755 --- a/scripts/build-many-glibcs.py +++ b/scripts/build-many-glibcs.py @@ -361,8 +361,9 @@ class Context(object): os_name='linux-gnu') self.add_config(arch='or1k', os_name='linux-gnu', - variant='soft', - gcc_cfg=['--with-multilib-list=mcmov']) + gcc_cfg=['--with-multilib-list=mcmov,mhard-float'], + glibcs=[{'variant': 'soft'}, + {'variant': 'hard', 'ccopts': '-mhard-float'}]) self.add_config(arch='powerpc', os_name='linux-gnu', gcc_cfg=['--disable-multilib', '--enable-secureplt'], -- 2.44.0 ^ permalink raw reply related [flat|nested] 8+ messages in thread
end of thread, other threads:[~2024-04-16 20:37 UTC | newest] Thread overview: 8+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2024-03-29 7:40 [PATCH 0/2] OpenRISC glibc hard float support Stafford Horne 2024-03-29 7:40 ` [PATCH 1/2] or1k: Add " Stafford Horne 2024-04-16 14:04 ` Adhemerval Zanella Netto 2024-04-16 14:53 ` Richard Henderson 2024-04-16 16:20 ` Adhemerval Zanella Netto 2024-04-16 19:56 ` Richard Henderson 2024-04-16 20:37 ` Stafford Horne 2024-03-29 7:40 ` [PATCH 2/2] build-many-glibcs.py: Add openrisc hard float glibc variant Stafford Horne
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.