| Message ID | 0ae3af1b-cb2b-367f-8b86-f39971498291@arm.com |
|---|---|
| State | New |
| Headers | show |
| Series | [RFC,1/2] Aarch64: Add simd exp/expf ABI symbols | expand |
Szabolcs Nagy <Szabolcs.Nagy@arm.com> writes: > diff --git a/sysdeps/aarch64/fpu/Versions b/sysdeps/aarch64/fpu/Versions > new file mode 100644 > index 0000000000..da36f3c495 > --- /dev/null > +++ b/sysdeps/aarch64/fpu/Versions > @@ -0,0 +1,5 @@ > +libmvec { > + GLIBC_2.30 { > + _ZGVnN2v_exp; _ZGVnN4v_expf; The prefix _ZGV is reserved for guard variables in the C++ ABI [1] [2]. How do you plan to avoid a conflict there? [1] https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/cp/mangle.c;#l4242 [2] https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/testsuite/g%2B%2B.dg/abi/aarch64_guard1.C
On 08/07/2019 19:00, Tulio Magno Quites Machado Filho wrote: > Szabolcs Nagy <Szabolcs.Nagy@arm.com> writes: > >> diff --git a/sysdeps/aarch64/fpu/Versions b/sysdeps/aarch64/fpu/Versions >> new file mode 100644 >> index 0000000000..da36f3c495 >> --- /dev/null >> +++ b/sysdeps/aarch64/fpu/Versions >> @@ -0,0 +1,5 @@ >> +libmvec { >> + GLIBC_2.30 { >> + _ZGVnN2v_exp; _ZGVnN4v_expf; > > The prefix _ZGV is reserved for guard variables in the C++ ABI [1] [2]. > How do you plan to avoid a conflict there? > > [1] https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/cp/mangle.c;#l4242 > [2] https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/testsuite/g%2B%2B.dg/abi/aarch64_guard1.C i noticed that too when this was designed, but it was decided that it's better to use the intel vector function mangling on aarch64 too than to introduce target specific mangling. i don't know how serious such a conflict is, nor why intel chose a clearly conflicting prefix. i assumed c++ folks were aware of the issue and accepted it when the intel vector abi was introduced.
* Szabolcs Nagy: > On 08/07/2019 19:00, Tulio Magno Quites Machado Filho wrote: >> Szabolcs Nagy <Szabolcs.Nagy@arm.com> writes: >> >>> diff --git a/sysdeps/aarch64/fpu/Versions b/sysdeps/aarch64/fpu/Versions >>> new file mode 100644 >>> index 0000000000..da36f3c495 >>> --- /dev/null >>> +++ b/sysdeps/aarch64/fpu/Versions >>> @@ -0,0 +1,5 @@ >>> +libmvec { >>> + GLIBC_2.30 { >>> + _ZGVnN2v_exp; _ZGVnN4v_expf; >> >> The prefix _ZGV is reserved for guard variables in the C++ ABI [1] [2]. >> How do you plan to avoid a conflict there? >> >> [1] https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/cp/mangle.c;#l4242 >> [2] https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/testsuite/g%2B%2B.dg/abi/aarch64_guard1.C > > i noticed that too when this was designed, but it > was decided that it's better to use the intel vector > function mangling on aarch64 too than to introduce > target specific mangling. > > i don't know how serious such a conflict is, nor why > intel chose a clearly conflicting prefix. i assumed > c++ folks were aware of the issue and accepted it > when the intel vector abi was introduced. I think the mangled names of objects always start with a capital letter, so there is no conflict in practice because the libmvec names use a lower-case letter after the _ZGV prefix. It may make sense to ask the C++ developers for clarification. Thanks, Florian
Szabolcs, I applied your patches (1 and 2) to a glibc tree and tried doing a full build (built ToT binutils, then partial ToT GCC, then ToT glibc with your patches, then full GCC) and I got an error when building the final GCC: /extra/sellcey/gcc-vect/install/aarch64-linux-gnu/bin/ld: cannot find /lib/libmvec_nonshared.a inside /extra/sellcey/gcc-vect/obj-gcc/gcc/../../install collect2: error: ld returned 1 exit status Makefile:1704: recipe for target 'libgfortran.la' failed I don't see a libmvec_nonshared.a object in my build or install directory. There are libmvec.so and libmvec.a libraries. Have you run into this problem? You may not have hit it if you don't config GCC with the Fortran compiler. I think it is necessary to have some object (even if it is not used) that is put in the libmvec-static-only-routines list of objects in sysdeps/aarch64/fpu/Makefile in order to build libmvec_nonshared.a. If you don't have that library the linker script fails when it cannot find it. Steve Ellcey sellcey@marvell.com
On 12/07/2019 22:51, Steve Ellcey wrote: > Szabolcs, > > I applied your patches (1 and 2) to a glibc tree and tried doing a full > build (built ToT binutils, then partial ToT GCC, then ToT glibc with > your patches, then full GCC) and I got an error when building the final > GCC: > > /extra/sellcey/gcc-vect/install/aarch64-linux-gnu/bin/ld: cannot find /lib/libmvec_nonshared.a inside /extra/sellcey/gcc-vect/obj-gcc/gcc/../../install > collect2: error: ld returned 1 exit status > Makefile:1704: recipe for target 'libgfortran.la' failed > > I don't see a libmvec_nonshared.a object in my build or install > directory. There are libmvec.so and libmvec.a libraries. Have you run > into this problem? You may not have hit it if you don't config GCC > with the Fortran compiler. > > I think it is necessary to have some object (even if it is not used) > that is put in the libmvec-static-only-routines list of objects in > sysdeps/aarch64/fpu/Makefile in order to build libmvec_nonshared.a. If > you don't have that library the linker script fails when it cannot find > it. you are right, i think libmvec-static-only-routines = non-existing-routine in that Makefile would fix it: creates an empty libmvec_nonshared.a. i made this change and put the patches on the nsz/mathvec branch.
* Steve Ellcey: > I think it is necessary to have some object (even if it is not used) > that is put in the libmvec-static-only-routines list of objects in > sysdeps/aarch64/fpu/Makefile in order to build libmvec_nonshared.a. If > you don't have that library the linker script fails when it cannot find > it. Maybe you could make the generation of the linker script optional in math/Makefile? Thanks, Florian
On 15/07/2019 11:55, Florian Weimer wrote: > * Steve Ellcey: > >> I think it is necessary to have some object (even if it is not used) >> that is put in the libmvec-static-only-routines list of objects in >> sysdeps/aarch64/fpu/Makefile in order to build libmvec_nonshared.a. If >> you don't have that library the linker script fails when it cannot find >> it. > > Maybe you could make the generation of the linker script optional in > math/Makefile? the linker script is needed anyway to add libmvec.so, just the libmvec_nonshared.a part can be dropped on targets that don't need it. but it's easier to have an empty libmvec_nonshared.a (it would be nicer if all symbols were in libm.so without any linker script and nonshared hack, but i guess it's too late for that)
On Mon, 2019-07-15 at 10:33 +0000, Szabolcs Nagy wrote: > > > > > I think it is necessary to have some object (even if it is not used) > > that is put in the libmvec-static-only-routines list of objects in > > sysdeps/aarch64/fpu/Makefile in order to build libmvec_nonshared.a. If > > you don't have that library the linker script fails when it cannot find > > it. > > you are right, i think > > libmvec-static-only-routines = non-existing-routine > > in that Makefile would fix it: creates an empty libmvec_nonshared.a. > > i made this change and put the patches on the nsz/mathvec branch. Thanks, the other thing I noticed when I tried building with the patches was that no functions got declared as SIMD in math-vector.h or math-vector-fortran.h. Not even exp or expf. Steve Ellcey sellcey@marvell.com
On Jul 15 2019, Szabolcs Nagy <Szabolcs.Nagy@arm.com> wrote: > you are right, i think > > libmvec-static-only-routines = non-existing-routine > > in that Makefile would fix it: creates an empty libmvec_nonshared.a. You should use make-dummy-lib. Andreas.
On 15/07/2019 18:14, Steve Ellcey wrote: > On Mon, 2019-07-15 at 10:33 +0000, Szabolcs Nagy wrote: >> >>> >>> I think it is necessary to have some object (even if it is not used) >>> that is put in the libmvec-static-only-routines list of objects in >>> sysdeps/aarch64/fpu/Makefile in order to build libmvec_nonshared.a. If >>> you don't have that library the linker script fails when it cannot find >>> it. >> >> you are right, i think >> >> libmvec-static-only-routines = non-existing-routine >> >> in that Makefile would fix it: creates an empty libmvec_nonshared.a. >> >> i made this change and put the patches on the nsz/mathvec branch. > > Thanks, the other thing I noticed when I tried building with the > patches was that no functions got declared as SIMD in math-vector.h > or math-vector-fortran.h. Not even exp or expf. yes, because currently it's not possible to declare them correctly. (adding omp declare simd should work today, but when gcc gets sve support that will be incorrect, so you can add such declarations for testing purposes, but it's not valid to distribute glibc headers that way) (and wilco's finite math removal patches will be needed once we add declarations to the headers.)
On 15/07/2019 18:25, Andreas Schwab wrote: > On Jul 15 2019, Szabolcs Nagy <Szabolcs.Nagy@arm.com> wrote: > >> you are right, i think >> >> libmvec-static-only-routines = non-existing-routine >> >> in that Makefile would fix it: creates an empty libmvec_nonshared.a. > > You should use make-dummy-lib. thanks.
diff --git a/sysdeps/aarch64/configure b/sysdeps/aarch64/configure index 5bd355a691..df15cdb02a 100644 --- a/sysdeps/aarch64/configure +++ b/sysdeps/aarch64/configure @@ -172,3 +172,34 @@ else config_vars="$config_vars default-abi = lp64" fi + +{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for pcs attribute support" >&5 +$as_echo_n "checking for pcs attribute support... " >&6; } +if ${libc_cv_gcc_pcs_attribute+:} false; then : + $as_echo_n "(cached) " >&6 +else + cat > conftest.c <<EOF +__attribute__((aarch64_vector_pcs)) extern void foo (void); +EOF +libc_cv_gcc_pcs_attribute=no +if ${CC-cc} -c -Wall -Werror conftest.c -o conftest.o 1>&5 \ + 2>&5 ; then + libc_cv_gcc_pcs_attribute=yes +fi +rm -f conftest* +fi +{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $libc_cv_gcc_pcs_attribute" >&5 +$as_echo "$libc_cv_gcc_pcs_attribute" >&6; } + +# Enable libmvec by default. +if test x"$build_mathvec" = xnotset; then + build_mathvec=yes +fi + +# Only test libmvec if the compiler supports aarch64_vector_pcs. +if test x"$build_mathvec" = xyes; then + if test $libc_cv_gcc_pcs_attribute = yes; then + config_vars="$config_vars +test-mathvec = yes" + fi +fi diff --git a/sysdeps/aarch64/configure.ac b/sysdeps/aarch64/configure.ac index 7851dd4dac..eab411cad4 100644 --- a/sysdeps/aarch64/configure.ac +++ b/sysdeps/aarch64/configure.ac @@ -20,3 +20,27 @@ if test $libc_cv_aarch64_be = yes; then else LIBC_CONFIG_VAR([default-abi], [lp64]) fi + +AC_CACHE_CHECK([for pcs attribute support], + libc_cv_gcc_pcs_attribute, [dnl +cat > conftest.c <<EOF +__attribute__((aarch64_vector_pcs)) extern void foo (void); +EOF +libc_cv_gcc_pcs_attribute=no +if ${CC-cc} -c -Wall -Werror conftest.c -o conftest.o 1>&AS_MESSAGE_LOG_FD \ + 2>&AS_MESSAGE_LOG_FD ; then + libc_cv_gcc_pcs_attribute=yes +fi +rm -f conftest*]) + +# Enable libmvec by default. +if test x"$build_mathvec" = xnotset; then + build_mathvec=yes +fi + +# Only test libmvec if the compiler supports aarch64_vector_pcs. +if test x"$build_mathvec" = xyes; then + if test $libc_cv_gcc_pcs_attribute = yes; then + LIBC_CONFIG_VAR([test-mathvec], [yes]) + fi +fi diff --git a/sysdeps/aarch64/fpu/Makefile b/sysdeps/aarch64/fpu/Makefile index 4a182bd6d6..220b664323 100644 --- a/sysdeps/aarch64/fpu/Makefile +++ b/sysdeps/aarch64/fpu/Makefile @@ -12,3 +12,20 @@ CFLAGS-s_fmaxf.c += -ffinite-math-only CFLAGS-s_fmin.c += -ffinite-math-only CFLAGS-s_fminf.c += -ffinite-math-only endif + +ifeq ($(subdir),mathvec) +libmvec-support += \ + libmvec_double_vlen2_exp \ + libmvec_float_vlen4_expf \ + +endif + +ifeq ($(subdir),math) +ifeq ($(build-mathvec),yes) +double-vlen2-funcs = exp +float-vlen4-funcs = exp +ifeq ($(test-mathvec),yes) +libmvec-tests += double-vlen2 float-vlen4 +endif +endif +endif diff --git a/sysdeps/aarch64/fpu/Versions b/sysdeps/aarch64/fpu/Versions new file mode 100644 index 0000000000..da36f3c495 --- /dev/null +++ b/sysdeps/aarch64/fpu/Versions @@ -0,0 +1,5 @@ +libmvec { + GLIBC_2.30 { + _ZGVnN2v_exp; _ZGVnN4v_expf; + } +} diff --git a/sysdeps/aarch64/fpu/libmvec_double_vlen2.h b/sysdeps/aarch64/fpu/libmvec_double_vlen2.h new file mode 100644 index 0000000000..383980d6ef --- /dev/null +++ b/sysdeps/aarch64/fpu/libmvec_double_vlen2.h @@ -0,0 +1,59 @@ +/* Double-precision 2 element vector function template. + Copyright (C) 2019 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + <http://www.gnu.org/licenses/>. */ + +#include <sysdep.h> + +ENTRY (VECTOR_FUNCTION) + stp x29, x30, [sp, -288]! + cfi_adjust_cfa_offset (288) + cfi_rel_offset (x29, 0) + cfi_rel_offset (x30, 8) + mov x29, sp + stp q8, q9, [sp, 16] + stp q10, q11, [sp, 48] + stp q12, q13, [sp, 80] + stp q14, q15, [sp, 112] + stp q16, q17, [sp, 144] + stp q18, q19, [sp, 176] + stp q20, q21, [sp, 208] + stp q22, q23, [sp, 240] + + // Use per lane load/store to avoid endianness issues. + str q0, [sp, 272] + ldr d0, [sp, 272] + bl SCALAR_FUNCTION + str d0, [sp, 272] + ldr d0, [sp, 280] + bl SCALAR_FUNCTION + str d0, [sp, 280] + ldr q0, [sp, 272] + + ldp q8, q9, [sp, 16] + ldp q10, q11, [sp, 48] + ldp q12, q13, [sp, 80] + ldp q14, q15, [sp, 112] + ldp q16, q17, [sp, 144] + ldp q18, q19, [sp, 176] + ldp q20, q21, [sp, 208] + ldp q22, q23, [sp, 240] + ldp x29, x30, [sp], 288 + cfi_adjust_cfa_offset (288) + cfi_restore (x29) + cfi_restore (x30) + ret +END (VECTOR_FUNCTION) diff --git a/sysdeps/aarch64/fpu/libmvec_double_vlen2_exp.S b/sysdeps/aarch64/fpu/libmvec_double_vlen2_exp.S new file mode 100644 index 0000000000..644405cc4f --- /dev/null +++ b/sysdeps/aarch64/fpu/libmvec_double_vlen2_exp.S @@ -0,0 +1,21 @@ +/* Double-precision 2 element vector e^x function. + Copyright (C) 2019 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + <http://www.gnu.org/licenses/>. */ + +#define SCALAR_FUNCTION exp +#define VECTOR_FUNCTION _ZGVnN2v_exp +#include "libmvec_double_vlen2.h" diff --git a/sysdeps/aarch64/fpu/libmvec_float_vlen4.h b/sysdeps/aarch64/fpu/libmvec_float_vlen4.h new file mode 100644 index 0000000000..2450309c13 --- /dev/null +++ b/sysdeps/aarch64/fpu/libmvec_float_vlen4.h @@ -0,0 +1,65 @@ +/* Single-precision 4 element vector function template. + Copyright (C) 2019 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + <http://www.gnu.org/licenses/>. */ + +#include <sysdep.h> + +ENTRY (VECTOR_FUNCTION) + stp x29, x30, [sp, -288]! + cfi_adjust_cfa_offset (288) + cfi_rel_offset (x29, 0) + cfi_rel_offset (x30, 8) + mov x29, sp + stp q8, q9, [sp, 16] + stp q10, q11, [sp, 48] + stp q12, q13, [sp, 80] + stp q14, q15, [sp, 112] + stp q16, q17, [sp, 144] + stp q18, q19, [sp, 176] + stp q20, q21, [sp, 208] + stp q22, q23, [sp, 240] + + // Use per lane load/store to avoid endianness issues. + str q0, [sp, 272] + ldr s0, [sp, 272] + bl SCALAR_FUNCTION + str s0, [sp, 272] + ldr s0, [sp, 276] + bl SCALAR_FUNCTION + str s0, [sp, 276] + ldr s0, [sp, 280] + bl SCALAR_FUNCTION + str s0, [sp, 280] + ldr s0, [sp, 284] + bl SCALAR_FUNCTION + str s0, [sp, 284] + ldr q0, [sp, 272] + + ldp q8, q9, [sp, 16] + ldp q10, q11, [sp, 48] + ldp q12, q13, [sp, 80] + ldp q14, q15, [sp, 112] + ldp q16, q17, [sp, 144] + ldp q18, q19, [sp, 176] + ldp q20, q21, [sp, 208] + ldp q22, q23, [sp, 240] + ldp x29, x30, [sp], 288 + cfi_adjust_cfa_offset (288) + cfi_restore (x29) + cfi_restore (x30) + ret +END (VECTOR_FUNCTION) diff --git a/sysdeps/aarch64/fpu/libmvec_float_vlen4_expf.S b/sysdeps/aarch64/fpu/libmvec_float_vlen4_expf.S new file mode 100644 index 0000000000..ab76ea0c77 --- /dev/null +++ b/sysdeps/aarch64/fpu/libmvec_float_vlen4_expf.S @@ -0,0 +1,21 @@ +/* Single-precision 4 element vector e^x function. + Copyright (C) 2019 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + <http://www.gnu.org/licenses/>. */ + +#define SCALAR_FUNCTION expf +#define VECTOR_FUNCTION _ZGVnN4v_expf +#include "libmvec_float_vlen4.h" diff --git a/sysdeps/aarch64/fpu/test-double-vlen2-wrappers.c b/sysdeps/aarch64/fpu/test-double-vlen2-wrappers.c new file mode 100644 index 0000000000..6c6c44d6b5 --- /dev/null +++ b/sysdeps/aarch64/fpu/test-double-vlen2-wrappers.c @@ -0,0 +1,28 @@ +/* Wrapper part of tests for aarch64 double vector math functions. + Copyright (C) 2019 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + <http://www.gnu.org/licenses/>. */ + +#include <arm_neon.h> +#include "test-double-vlen2.h" + +#define VEC_TYPE float64x2_t + +/* Hack: VECTOR_WRAPPER declares the vector function without the pcs attribute, + placing it here happens to work, should be fixed in test-math-vector.h. */ +__attribute__ ((aarch64_vector_pcs)) + +VECTOR_WRAPPER (WRAPPER_NAME (exp), _ZGVnN2v_exp) diff --git a/sysdeps/aarch64/fpu/test-float-vlen4-wrappers.c b/sysdeps/aarch64/fpu/test-float-vlen4-wrappers.c new file mode 100644 index 0000000000..5117633f1f --- /dev/null +++ b/sysdeps/aarch64/fpu/test-float-vlen4-wrappers.c @@ -0,0 +1,28 @@ +/* Wrapper part of tests for float aarch64 vector math functions. + Copyright (C) 2019 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + <http://www.gnu.org/licenses/>. */ + +#include <arm_neon.h> +#include "test-float-vlen4.h" + +#define VEC_TYPE float32x4_t + +/* Hack: VECTOR_WRAPPER declares the vector function without the pcs attribute, + placing it here happens to work, should be fixed in test-math-vector.h. */ +__attribute__ ((aarch64_vector_pcs)) + +VECTOR_WRAPPER (WRAPPER_NAME (expf), _ZGVnN4v_expf) diff --git a/sysdeps/aarch64/libm-test-ulps b/sysdeps/aarch64/libm-test-ulps index 585e5bbce7..1ed4af9e55 100644 --- a/sysdeps/aarch64/libm-test-ulps +++ b/sysdeps/aarch64/libm-test-ulps @@ -1601,6 +1601,12 @@ float: 1 idouble: 1 ifloat: 1 +Function: "exp_vlen2": +double: 1 + +Function: "exp_vlen4": +float: 1 + Function: "expm1": double: 1 float: 1 diff --git a/sysdeps/unix/sysv/linux/aarch64/libmvec.abilist b/sysdeps/unix/sysv/linux/aarch64/libmvec.abilist new file mode 100644 index 0000000000..9e178253f7 --- /dev/null +++ b/sysdeps/unix/sysv/linux/aarch64/libmvec.abilist @@ -0,0 +1,2 @@ +GLIBC_2.30 _ZGVnN2v_exp F +GLIBC_2.30 _ZGVnN4v_expf F