The CORE-MATH implementation is correctly rounded (for any rounding mode)
and shows better performance compared to the generic exp2m1f.
The code was adapted to glibc style and to use the definition of
math_config.h (to handle errno, overflow, and underflow). The
only change is to handle FLT_MAX_EXP for FE_DOWNWARD or FE_TOWARDZERO.
The benchmark inputs are based on exp2f ones.
Benchtest on x64_64 (Ryzen 9 5900X, gcc 14.2.1), aarch64 (Neoverse-N1,
gcc 13.3.1), and powerpc (POWER10, gcc 13.2.1):
Latency master patched improvement
x86_64 40.6042 48.7104 -19.96%
x86_64v2 40.7506 35.9032 11.90%
x86_64v3 35.2301 31.7956 9.75%
i686 102.094 94.6657 7.28%
aarch64 18.2704 15.1387 17.14%
power10 11.9444 8.2402 31.01%
reciprocal-throughput master patched improvement
x86_64 20.8683 16.1428 22.64%
x86_64v2 19.5076 10.4474 46.44%
x86_64v3 19.2106 10.4014 45.86%
i686 56.4054 59.3004 -5.13%
aarch64 12.0781 7.3953 38.77%
power10 6.5306 5.9388 9.06%
The generic implementation calls __ieee754_exp2f and x86_64 provides
an optimized ifunc version (built with -mfma -mavx2, not correctly
rounded). This explains the performance difference for x86_64.
Same for i686, where the ABI provides an optimized __ieee754_exp2f
version built with '-msse2 -mfpmath=sse'. When built wth same
flags, the new algorithm shows a better performance:
master patched improvement
latency 102.094 91.2823 10.59%
reciprocal-throughput 56.4054 52.7984 6.39%
Signed-off-by: Alexei Sibidanov <sibid@uvic.ca>
Signed-off-by: Paul Zimmermann <Paul.Zimmermann@inria.fr>
Signed-off-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Reviewed-by: DJ Delorie <dj@redhat.com>
The CORE-MATH implementation is correctly rounded (for any rounding mode)
and shows better performance compared to the generic exp10m1f.
The code was adapted to glibc style and to use the definition of
math_config.h (to handle errno, overflow, and underflow). I mostly
fixed some small issues in corner cases (sNaN handling, -INFINITY,
a specific overflow check).
Benchtest on x64_64 (Ryzen 9 5900X, gcc 14.2.1), aarch64 (Neoverse-N1,
gcc 13.3.1), and powerpc (POWER10, gcc 13.2.1):
Latency master patched improvement
x86_64 45.4690 49.5845 -9.05%
x86_64v2 46.1604 36.2665 21.43%
x86_64v3 37.8442 31.0359 17.99%
i686 121.367 93.0079 23.37%
aarch64 21.1126 15.0165 28.87%
power10 12.7426 8.4929 33.35%
reciprocal-throughput master patched improvement
x86_64 19.6005 17.4005 11.22%
x86_64v2 19.6008 11.1977 42.87%
x86_64v3 17.5427 10.2898 41.34%
i686 59.4215 60.9675 -2.60%
aarch64 13.9814 7.9173 43.37%
power10 6.7814 6.4258 5.24%
The generic implementation calls __ieee754_exp10f which has an
optimized version, although it is not correctly rounded, which is
the main culprit of the the latency difference for x86_64 and
throughp for i686.
Signed-off-by: Alexei Sibidanov <sibid@uvic.ca>
Signed-off-by: Paul Zimmermann <Paul.Zimmermann@inria.fr>
Signed-off-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Reviewed-by: DJ Delorie <dj@redhat.com>
The CORE-MATH implementation is correctly rounded (for any rounding mode).
This can be checked by exhaustive tests in a few minutes since there are
less than 2^32 values to check against for example GNU MPFR.
This patch also adds some bench values for tgammaf.
Tested on x86_64 and x86 (cfarm26).
With the initial GNU libc code it gave on an Intel(R) Core(TM) i7-8700:
"tgammaf": {
"": {
"duration": 3.50188e+09,
"iterations": 2e+07,
"max": 602.891,
"min": 65.1415,
"mean": 175.094
}
}
With the new code:
"tgammaf": {
"": {
"duration": 3.30825e+09,
"iterations": 5e+07,
"max": 211.592,
"min": 32.0325,
"mean": 66.1649
}
}
With the initial GNU libc code it gave on cfarm26 (i686):
"tgammaf": {
"": {
"duration": 3.70505e+09,
"iterations": 6e+06,
"max": 2420.23,
"min": 243.154,
"mean": 617.509
}
}
With the new code:
"tgammaf": {
"": {
"duration": 3.24497e+09,
"iterations": 1.8e+07,
"max": 1238.15,
"min": 101.155,
"mean": 180.276
}
}
Signed-off-by: Alexei Sibidanov <sibid@uvic.ca>
Signed-off-by: Paul Zimmermann <Paul.Zimmermann@inria.fr>
Changes in v2:
- include <math.h> (fix the linknamespace failures)
- restored original benchtests/strcoll-inputs/filelist#en_US.UTF-8 file
- restored original wrapper code (math/w_tgammaf_compat.c),
except for the dealing with the sign
- removed the tgammaf/float entries in all libm-test-ulps files
- address other comments from Joseph Myers
(https://sourceware.org/pipermail/libc-alpha/2024-July/158736.html)
Changes in v3:
- pass NULL argument for signgam from w_tgammaf_compat.c
- use of math_narrow_eval
- added more comments
Changes in v4:
- initialize local_signgam to 0 in math/w_tgamma_template.c
- replace sysdeps/ieee754/dbl-64/gamma_productf.c by dummy file
Changes in v5:
- do not mention local_signgam any more in math/w_tgammaf_compat.c
- initialize local_signgam to 1 instead of 0 in w_tgamma_template.c
and added comment
Changes in v6:
- pass NULL as 2nd argument of __ieee754_gammaf_r in
w_tgammaf_compat.c, and check for NULL in e_gammaf_r.c
Changes in v7:
- added Signed-off-by line for Alexei Sibidanov (author of the code)
Changes in v8:
- added Signed-off-by line for Paul Zimmermann (submitted of the patch)
Changes in v9:
- address comments from review by Adhemerval Zanella
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
As discussed at the patch review meeting
Signed-off-by: Andreas K. Hüttel <dilfridge@gentoo.org>
Reviewed-by: Simon Chopin <simon.chopin@canonical.com>
C23 adds various <math.h> function families originally defined in TS
18661-4. Add the logp1 functions (aliases for log1p functions - the
name is intended to be more consistent with the new log2p1 and
log10p1, where clearly it would have been very confusing to name those
functions log21p and log101p). As aliases rather than new functions,
the content of this patch is somewhat different from those actually
adding new functions.
Tests are shared with log1p, so this patch *does* mechanically update
all affected libm-test-ulps files to expect the same errors for both
functions.
The vector versions of log1p on aarch64 and x86_64 are *not* updated
to have logp1 aliases (and thus there are no corresponding header,
tests, abilist or ulps changes for vector functions either). It would
be reasonable for such vector aliases and corresponding changes to
other files to be made separately. For now, the log1p tests instead
avoid testing logp1 in the vector case (a Makefile change is needed to
avoid problems with grep, used in generating the .c files for vector
function tests, matching more than one ALL_RM_TEST line in a file
testing multiple functions with the same inputs, when it assumes that
the .inc file only has a single such line).
Tested for x86_64 and x86, and with build-many-glibcs.py.
The default <utmp-size.h> is for ports with a 64-bit time_t.
Ports with a 32-bit time_t or with __WORDSIZE_TIME64_COMPAT32=1
need to override it.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
The value of l_scope is only valid post relocation, so this original
check was triggering undefined behavior. Instead just directly check to
see if the object has been relocated, at which point using l_scope is
safe.
Reported-by: Andreas Schwab <schwab@suse.de>
Closes: BZ #31317
Fixes: e0590f41fe ("RISC-V: Enable static-pie.")
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
The memcpy optimization (commit 587a1290a1) has a series
of mistakes:
- The implementation is wrong: the chunk size calculation is wrong
leading to invalid memory access.
- It adds ifunc supports as default, so --disable-multi-arch does
not work as expected for riscv.
- It mixes Linux files (memcpy ifunc selection which requires the
vDSO/syscall mechanism) with generic support (the memcpy
optimization itself).
- There is no __libc_ifunc_impl_list, which makes testing only
check the selected implementation instead of all supported
by the system.
This patch also simplifies the required bits to enable ifunc: there
is no need to memcopy.h; nor to add Linux-specific files.
The __memcpy_noalignment tail handling now uses a branchless strategy
similar to aarch64 (overlap 32-bits copies for sizes 4..7 and byte
copies for size 1..3).
Checked on riscv64 and riscv32 by explicitly enabling the function
on __libc_ifunc_impl_list on qemu-system.
Changes from v1:
* Implement the memcpy in assembly to correctly handle RISCV
strict-alignment.
Reviewed-by: Evan Green <evan@rivosinc.com>
Acked-by: Palmer Dabbelt <palmer@rivosinc.com>
For CPU implementations that can perform unaligned accesses with little
or no performance penalty, create a memcpy implementation that does not
bother aligning buffers. It will use a block of integer registers, a
single integer register, and fall back to bytewise copy for the
remainder.
Signed-off-by: Evan Green <evan@rivosinc.com>
Reviewed-by: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
RISC-V is apparently the first architecture to pass more than one
argument to ifunc resolvers. The helper macros in libc-symbols.h,
__ifunc_resolver(), __ifunc(), and __ifunc_hidden(), are incompatible
with this. These macros have an "arg" (non-final) parameter that
represents the parameter signature of the ifunc resolver. The result is
an inability to pass the required comma through in a single preprocessor
argument.
Rearrange the __ifunc_resolver() macro to be variadic, and pass the
types as those variable parameters. Move the guts of __ifunc() and
__ifunc_hidden() into new macros, __ifunc_args(), and
__ifunc_args_hidden(), that pass the variable arguments down through to
__ifunc_resolver(). Then redefine __ifunc() and __ifunc_hidden(), which
are used in a bunch of places, to simply shuffle the arguments down into
__ifunc_args[_hidden]. Finally, define a riscv-ifunc.h header, which
provides convenience macros to those looking to write ifunc selectors
that use both arguments.
Signed-off-by: Evan Green <evan@rivosinc.com>
Reviewed-by: Florian Weimer <fweimer@redhat.com>
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
The new __riscv_hwprobe() function is designed to be used by ifunc
selector functions. This presents a challenge for applications and
libraries, as ifunc selectors are invoked before all relocations have
been performed, so an external call to __riscv_hwprobe() from an ifunc
selector won't work. To address this, pass a pointer to the
__riscv_hwprobe() function into ifunc selectors as the second
argument (alongside dl_hwcap, which was already being passed).
Include a typedef as well for convenience, so that ifunc users don't
have to go through contortions to call this routine. Users will need to
remember to check the second argument for NULL, to account for older
glibcs that don't pass the function.
Signed-off-by: Evan Green <evan@rivosinc.com>
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
WG14 decided to use the name C23 as the informal name of the next
revision of the C standard (notwithstanding the publication date in
2024). Update references to C2X in glibc to use the C23 name.
This is intended to update everything *except* where it involves
renaming files (the changes involving renaming tests are intended to
be done separately). In the case of the _ISOC2X_SOURCE feature test
macro - the only user-visible interface involved - support for that
macro is kept for backwards compatibility, while adding
_ISOC23_SOURCE.
Tested for x86_64.
This patch referents the commit 374cef3 to add static-pie support. And
because the dummy link map is used when relocating ourselves, so need
not to set __global_pointer$ at this time.
It will also check whether toolchain supports to build static-pie.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Code is mostly inspired from the LoongArch one, which has a similar ABI,
with minor changes to support riscv32 and register differences.
This fixes elf/tst-sprof-basic. This also fixes elf/tst-audit1,
elf/tst-audit2 and elf/tst-audit8 with recent binutils snapshots when
--enable-bind-now is used.
Resolves: BZ #31151
Acked-by: Palmer Dabbelt <palmer@rivosinc.com>
libc_feupdateenv_riscv should check for FE_DFL_ENV, similar to
libc_fesetenv_riscv.
Also extend the test-fenv.c to test fenvupdate.
Checked on riscv under qemu-system.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
XTheadBb has similar instructions like Zbb, which allow optimized
string processing:
* th.ff0: find-first zero is a CLZ instruction.
* th.tstnbz: Similar like orc.b, but with a bit-inverted result.
The instructions are documented here:
https://github.com/T-head-Semi/thead-extension-spec/tree/master/xtheadbb
These instructions can be found in the T-Head C906 and the C910.
Tested with the string tests.
Signed-off-by: Christoph Müllner <christoph.muellner@vrull.eu>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Generated on a VisionFive 2 board running Linux version 6.4.2 and
GCC 13.1.0.
Needed due to commit cf7ffdd8a5 ("added pair of inputs for hypotf in
binary32").
Bump autoconf requirement to 2.71 to allow regenerating configure on
more recent distributions. autoconf 2.71 has been in Fedora since F36
and is the current version in Debian stable (bookworm). It appears to
be current in Gentoo as well.
All sysdeps configure and preconfigure scripts have also been
regenerated; all changes are trivial transformations that do not affect
functionality.
Signed-off-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
In some cases, we do not want to go through the resolver for function
calls. For example, functions with vector arguments will use vector
registers to pass arguments. In the resolver, we do not save/restore the
vector argument registers for lazy binding efficiency. To avoid ruining
the vector arguments, functions with vector arguments will not go
through the resolver.
To achieve the goal, we will annotate the function symbols with
STO_RISCV_VARIANT_CC flag and add DT_RISCV_VARIANT_CC tag in the dynamic
section. In the first pass on PLT relocations, we do not set up to call
_dl_runtime_resolve. Instead, we resolve the functions directly.
Signed-off-by: Hsiangkai Wang <kai.wang@sifive.com>
Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
Reviewed-by: Palmer Dabbelt <palmer@rivosinc.com>
Acked-by: Palmer Dabbelt <palmer@rivosinc.com>
Link: https://inbox.sourceware.org/libc-alpha/20230314162512.35802-1-kito.cheng@sifive.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
It uses the bitmanip extension to optimize index_fist and index_last
with clz/ctz (using generic implementation that routes to compiler
builtin) and orc.b to check null bytes.
Checked the string test on riscv64 user mode.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
This makes it more likely that the compiler can compute the strlen
argument in _startup_fatal at compile time, which is required to
avoid a dependency on strlen this early during process startup.
Reviewed-by: Szabolcs Nagy <szabolcs.nagy@arm.com>
According to the specification of ISO/IEC TS 18661-1:2014,
The strfromd, strfromf, and strfroml functions are equivalent to
snprintf(s, n, format, fp) (7.21.6.5), except the format string contains only
the character %, an optional precision that does not contain an asterisk *, and
one of the conversion specifiers a, A, e, E, f, F, g, or G, which applies to
the type (double, float, or long double) indicated by the function suffix
(rather than by a length modifier). Use of these functions with any other 20
format string results in undefined behavior.
strfromf will convert the arguement with type float to double first.
According to the latest version of IEEE754 which is published in 2019,
Conversion of a quiet NaN from a narrower format to a wider format in the same
radix, and then back to the same narrower format, should not change the quiet
NaN payload in any way except to make it canonical.
When either an input or result is a NaN, this standard does not interpret the
sign of a NaN. However, operations on bit strings—copy, negate, abs,
copySign—specify the sign bit of a NaN result, sometimes based upon the sign
bit of a NaN operand. The logical predicates totalOrder and isSignMinus are
also affected by the sign bit of a NaN operand. For all other operations, this
standard does not specify the sign bit of a NaN result, even when there is only
one input NaN, or when the NaN is produced from an invalid operation.
converting NAN or -NAN with type float to double doesn't need to keep
the signbit. As a result, this test case isn't mandatory.
The problem is that according to RISC-V ISA manual in chapter 11.3 of
riscv-isa-20191213,
Except when otherwise stated, if the result of a floating-point operation is
NaN, it is the canonical NaN. The canonical NaN has a positive sign and all
significand bits clear except the MSB, a.k.a. the quiet bit. For
single-precision floating-point, this corresponds to the pattern 0x7fc00000.
which means that conversion -NAN from float to double won't keep the signbit.
Since glibc ought to be consistent here between types and architectures, this
patch adds copysign to fix this problem if the string is NAN. This patch
adds two different functions under sysdeps directory to work around the
issue.
This patch has been tested on x86_64 and riscv64.
Resolves: BZ #29501
v2: Change from macros to different inline functions.
v3: Add unlikely check to isnan.
v4: Fix wrong commit message header.
v5: Fix style: add space before parentheses.
v6: Add copyright.
Signed-off-by: Letu Ren <fantasquex@gmail.com>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
In the future, this will result in a compilation failure if the
macros are unexpectedly undefined (due to header inclusion ordering
or header inclusion missing altogether).
Assembler sources are more difficult to convert. In many cases,
they are hand-optimized for the mangling and no-mangling variants,
which is why they are not converted.
sysdeps/s390/s390-32/__longjmp.c and sysdeps/s390/s390-64/__longjmp.c
are special: These are C sources, but most of the implementation is
in assembler, so the PTR_DEMANGLE macro has to be undefined in some
cases, to match the assembler style.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
This allows us to define a generic no-op version of PTR_MANGLE and
PTR_DEMANGLE. In the future, we can use PTR_MANGLE and PTR_DEMANGLE
unconditionally in C sources, avoiding an unintended loss of hardening
due to missing include files or unlucky header inclusion ordering.
In i386 and x86_64, we can avoid a <tls.h> dependency in the C
code by using the computed constant from <tcb-offsets.h>. <sysdep.h>
no longer includes these definitions, so there is no cyclic dependency
anymore when computing the <tcb-offsets.h> constants.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Rename atomic_exchange_rel/acq to use atomic_exchange_release/acquire
since these map to the standard C11 atomic builtins.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
We don't need RV32 specific floating point functions, instead make them
generic for RISC-V.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Both RV32 and RV64 should have the same libm-test-ulps, so consolidate
them into a single file.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Minor clean-up, we need to change this part in following patch, clean this up
to prevent we duplicated the change twice.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
The RTLD_BOOTSTRAP branch is used to relocate ld.so itself. It only
needs to handle RELATIVE, GLOB_DAT, and the symbolic relocation type
(R_RISCV_{32,64}). NONE and IRELATIVE can be removed.
The code relies on ld.so having DT_RELACOUNT so that the RTLD_BOOTSTRAP
branch does not need handle RELATIVE. Drop this minor size
optimization for clarity.
Acked-by: Palmer Dabbelt <palmer@rivosinc.com>
Since ad43cac44a the generic code already shuffles the argv/envp/auxv
on the stack to remove the ld.so own arguments and thus _dl_skip_args
is always 0. So there is no need to adjust the argc or argv.
Checked with qemu-user that arguments are correctly passed on both
constructors and main program.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Avoid fiddling with autoconf internals and use AC_DEFINE_UNQUOTED to
define macros in the configuration headers rather than handcoding an
equivalent shell sequence with the use of the `as_echo' undocumented
variable.
Switch to using AC_MSG_ERROR rather than `echo' and `exit' directly for
error handling. Owing to the lack of any kind of error annotation it
makes it difficult to spot the message in the flood in a parallel build
and neither it is logged in `config.log'.
Reviewed-by: Palmer Dabbelt <palmer@rivosinc.com>
Acked-by: Palmer Dabbelt <palmer@rivosinc.com>
PI_STATIC_AND_HIDDEN indicates whether accesses to internal linkage
variables and hidden visibility variables in a shared object (ld.so)
need dynamic relocations (usually R_*_RELATIVE). PI (position
independent) in the macro name is a misnomer: a code sequence using GOT
is typically position-independent as well, but using dynamic relocations
does not meet the requirement.
Not defining PI_STATIC_AND_HIDDEN is legacy and we expect that all new
ports will define PI_STATIC_AND_HIDDEN. Current ports defining
PI_STATIC_AND_HIDDEN are more than the opposite. Change the configure
default.
No functional change.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
The 32-bit and 64-bit variants of RISC-V share the same name - "RISC-V"
- when generating the libm error table for the info pages. This
collision, and the way how the table is generated, mean that the values
in the final table for "RISC-V" may be either for the 32- or 64-bit
variant, with no indication as to which.
As an additional side-effect, this makes the build non-reproducible, as
the error table generated is dependent upon the host filesystem
implementation.
To solve this issue, the libm-test-ulps-name files for both variants
have been modified to include their word size, so as to remove the
collision and provide more accurate information in the table.
An alternative proposed was to merge the two variants' ULP values into a
single file, but this would mean that information about error values is
lost, as the two variants are not identical. Some differences are
considerable, notably the values for the exp() function are large.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
-z combreloc has been the default regadless of the architecture since
binutils commit f4d733664aabd7bd78c82895e030ec9779a92809 (2002). The
configure check added in commit fdde83499a (2001) has long been
unneeded.
We can therefore treat HAVE_Z_COMBRELOC as always 1 and delete dead code
paths in dl-machine.h files (many were copied from commit a711b01d34
and ee0cb67ec2).
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
I used these shell commands:
../glibc/scripts/update-copyrights $PWD/../gnulib/build-aux/update-copyright
(cd ../glibc && git commit -am"[this commit message]")
and then ignored the output, which consisted lines saying "FOO: warning:
copyright statement not found" for each of 7061 files FOO.
I then removed trailing white space from math/tgmath.h,
support/tst-support-open-dev-null-range.c, and
sysdeps/x86_64/multiarch/strlen-vec.S, to work around the following
obscure pre-commit check failure diagnostics from Savannah. I don't
know why I run into these diagnostics whereas others evidently do not.
remote: *** 912-#endif
remote: *** 913:
remote: *** 914-
remote: *** error: lines with trailing whitespace found
...
remote: *** error: sysdeps/unix/sysv/linux/statx_cp.c: trailing lines
And use machine-sp.h instead. The Linux implementation is based on
already provided CURRENT_STACK_FRAME (used on nptl code) and
STACK_GROWS_UPWARD is replaced with _STACK_GROWS_UP.
