By using a combination of mask-and-add instead of the shift-based
index calculation the routines can share the same table as other
variants with no performance degradation.
The tables change name because of other changes in downstream AOR.
Reviewed-by: Wilco Dijkstra <Wilco.Dijkstra@arm.com>
The CORE-MATH exp2m1f implementation showed slight worse latency
when using x86_64 baseline ABI. This patch adds a ifunc variant
with similar performance for x86_64-v3.
Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>
Reviewed-by: DJ Delorie <dj@redhat.com>
The CORE-MATH exp10m1f implementation showed slight worse latency
when using x86_64 baseline ABI. This patch adds a ifunc variant
with similar performance for x86_64-v3.
Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>
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 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>
Also remove the use of builtins in favor of standard names, compiler
already inline them (if supported) with current compiler options.
It also fixes and issue where __builtin_roundeven is not support on
gcc older than version 10.
Checked on x86_64-linux-gnu and i686-linux_gnu.
Signed-off-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Reviewed-by: DJ Delorie <dj@redhat.com>
This will be required by the rseq extensible ABI implementation on all
Linux architectures exposing the '__rseq_size' and '__rseq_offset'
symbols to set the initial value of the 'cpu_id' field which can be used
by applications to test if rseq is available and registered. As long as
the symbols are exposed it is valid for an application to perform this
test even if rseq is not yet implemented in libc for this architecture.
Both code paths are compile tested with build-many-glibcs.py but I don't
have access to any hardware to run the tests.
Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
Reviewed-by: Arjun Shankar <arjun@redhat.com>
Save lr in a non-volatile register before scv in clone/clone3.
For clone, the non-volatile register was unused and already
saved/restored. Remove the dead code from clone.
Signed-off-by: Sachin Monga <smonga@linux.ibm.com>
Reviewed-by: Peter Bergner <bergner@linux.ibm.com>
There are no tests specifically focused on the functions time,
gettimeofday and clock_gettime, although there are some incidental
uses in tests of other functions. Add tests specifically for these
three functions.
Tested for x86_64 and x86.
There are various existing tests that call pthread_attr_init and then
verify properties of the resulting initial values retrieved with
pthread_attr_get* functions. However, those are missing coverage of
the initial values retrieved with pthread_attr_getschedparam and
pthread_attr_getstacksize. Add testing for initial values from those
functions as well.
(tst-attr2 covers pthread_attr_getdetachstate,
pthread_attr_getguardsize, pthread_attr_getinheritsched,
pthread_attr_getschedpolicy, pthread_attr_getscope. tst-attr3 covers
some of those together with pthread_attr_getaffinity_np.
tst-pthread-attr-sigmask covers pthread_attr_getsigmask_np.
pthread_attr_getstack has unspecified results if called before the
relevant attributes have been set, while pthread_attr_getstackaddr is
deprecated.)
Tested for x86_64.
The gilbc manual has some documentation in llio.texi of requirements
for moving between I/O on FILE * streams and file descriptors on the
same open file description.
The documentation of what must be done on a FILE * stream to move from
it to either a file descriptor or another FILE * for the same open
file description seems to match POSIX. However, there is an
additional requirement in POSIX on the *second* of the two handles
being moved between, which is not mentioned in the glibc manual: "If
any previous active handle has been used by a function that explicitly
changed the file offset, except as required above for the first
handle, the application shall perform an lseek() or fseek() (as
appropriate to the type of handle) to an appropriate location.".
Document this requirement on seeking in the glibc manual, limited to
the case that seems relevant to glibc (the new channel is a previously
active stream, on which the seeking previously occurred). Note that
I'm not sure what the "except as required above for the first handle"
is meant to be about, so I haven't documented anything for it. As far
as I can tell, nothing specified for moving from the first handle
actually list calling a seek function as one of the steps to be done.
(Current POSIX doesn't seem to have any relevant rationale for this
section. The rationale in the 1996 edition says "In requiring the
seek to an appropriate location for the new handle, the application is
required to know what it is doing if it is passing streams with seeks
involved. If the required seek is not done, the results are undefined
(and in fact the program probably will not work on many common
implementations)." - which also doesn't help in understanding the
purpose of "except as required above for the first handle".)
Tested with "make info" and "make pdf".
In both routines, reduce register pressure such that GCC 14 emits no
spills for erf and fewer spills for erfc. Also use more efficient
comparison for the special-case in erf.
Benchtests show erf improves by 6.4%, erfc by 1.0%.
In commit c628c22963 (elf: Remove
ldconfig kernel version check), the layout of auxcache entries
changed because the osversion field was removed from
struct aux_cache_file_entry. However, AUX_CACHEMAGIC was not
changed, so existing files are still used, potentially leading
to unintended ldconfig behavior. This commit changes AUX_CACHEMAGIC,
so that the file is regenerated.
Reported-by: DJ Delorie <dj@redhat.com>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
This fixes a buffer overflow in wide character string output, reproducing
when output fails, such as if the output fd is closed or is redirected
to a full device.
Wide character output data attempts to maintain the invariant that
`_IO_buf_base <= _IO_write_base <= _IO_write_end <= _IO_buf_end` (that is,
that the write region is a sub-region of `_IO_buf`). Prior to this commit,
this invariant is violated by the `_IO_wfile_overflow` function as so:
1. `_IO_wsetg` is called, assigning `_IO_write_base` to `_IO_buf_base`
2. `_IO_doallocbuf` is called, which jumps to `_IO_wfile_doallocate` via
the _IO_wfile_jumps vtable. This function then assigns the wide data
`_IO_buf_base` and `_IO_buf_end` to a malloc'd buffer.
Thus the invariant is violated. The fix is simply to reverse the order:
malloc the `_IO_buf` first and then assign `_IO_write_base` to it.
We also take this opportunity to defensively guard the initialization of
the number of unwritten characters via pointer arithmetic. We now check
that the buffer end is not before the buffer beginning; this matches a
similar defensive check in the narrow analogue `fileops.c`.
Add a test which fails without the fix.
Signed-off-by: Peter Ammon <corydoras@ridiculousfish.com>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
The scanf family of functions like sscanf and fscanf currently
ignore nan() and nan(n-char-sequence). This happens because
__vfscanf_internal only checks for 'nan'.
This commit adds support for all valid nan types i.e. nan, nan()
and nan(n-char-sequence), where n-char-sequence can be
[a-zA-Z0-9_]+, thus fixing the bug 30647. Any other representation
of NaN should result in conversion error.
New tests are also added to verify the correct parsing of NaN types for
float, double and long double formats.
Signed-off-by: Avinal Kumar <avinal.xlvii@gmail.com>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
The refactoring did not take the change of variable into account.
Fixes commit 43db5e2c06
("elf: Signal RT_CONSISTENT after relocation processing in dlopen
(bug 31986)").
Previously, a la_activity audit event was generated before
relocation processing completed. This does did not match what
happened during initial startup in elf/rtld.c (towards the end
of dl_main). It also caused various problems if an auditor
tried to open the same shared object again using dlmopen:
If it was the directly loaded object, it had a search scope
associated with it, so the early exit in dl_open_worker_begin
was taken even though the object was unrelocated. This caused
the r_state == RT_CONSISTENT assert to fail. Avoidance of the
assert also depends on reversing the order of r_state update
and auditor event (already implemented in a previous commit).
At the later point, args->map can be NULL due to failure,
so use the assigned namespace ID instead if that is available.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Auditors can call into the dynamic loader again if
LA_ACT_CONSISTENT, and those recursive calls could observe
r_state != RT_CONSISTENT.
We should consider failing dlopen/dlmopen/dlclose if
r_state != RT_CONSISTENT. The dynamic linker is probably not
in a state in which it can handle reentrant calls. This
needs further investigation.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
This is conceptually similar to the reported bug, but does not
depend on auditing. The fix is simple: just complete execution
of the constructors. This exposed the fact that the link map
for statically linked executables does not have l_init_called
set, even though constructors have run.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
This avoids -Werror build issues in strace, which bundles UAPI
headers, but does not include them as system headers.
Fixes commit c444cc1d83
("Linux: Add missing scheduler constants to <sched.h>").
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
tst-popen-fork failed to build for Hurd due to not being linked with
libpthread. This commit fixes that.
Tested with build-many-glibcs.py for i686-gnu.
Reviewed-by: Florian Weimer <fweimer@redhat.com>
Add basic tests of pthread_mutexattr_gettype and
pthread_mutexattr_settype with each valid mutex kind, plus test for
EINVAL with an invalid mutex kind.
Tested for x86_64.
popen modifies its file handler book-keeping under a lock that wasn't
being taken during fork. This meant that a concurrent popen and fork
could end up copying the lock in a "locked" state into the fork child,
where subsequently calling popen would lead to a deadlock due to the
already (spuriously) held lock.
This commit fixes the deadlock by appropriately taking the lock before
fork, and releasing/resetting it in the parent/child after the fork.
A new test for concurrent popen and fork is also added. It consistently
hangs (and therefore fails via timeout) without the fix applied.
Reviewed-by: Florian Weimer <fweimer@redhat.com>
Glibc has two gnu-extension functions that are implemented as
macros but not documented as such: fread_unlocked and
fwrite_unlocked. Document them as such.
Additionally, putc_unlocked and getc_unlocked are documented in
POSIX as possibly being macros. Update the manual to add a warning
about those also, depite glibc not implementing them as macros.
The pthread_timedjoin_np and pthread_clockjoin_np functions do not
check that a valid time has been specified. The documentation for
these functions in the glibc manual isn't sufficiently detailed to say
if they should, but consistency with POSIX functions such as
pthread_mutex_timedlock and pthread_cond_timedwait strongly indicates
that an EINVAL error is appropriate (even if there might be some
ambiguity about exactly where such a check should go in relation to
other checks for whether the thread exists, whether it's immediately
joinable, etc.). Copy the logic for such a check used in
pthread_rwlock_common.c.
pthread_join_common had some logic calling valid_nanoseconds before
commit 9e92278ffa, "nptl: Remove
clockwait_tid"; I haven't checked exactly what cases that detected.
Tested for x86_64 and x86.
This makes b4 use inbox.sourceware.org instead of the default host
lore.kernel.org, so that every b4 user doesn't have to configure this
themselves for the glibc repo.
Reviewed-by: Florian Weimer <fweimer@redhat.com>
The commit 'sparc: Use Linux kABI for syscall return'
(86c5d2cf0c) did not take into account
a subtle sparc syscall kABI constraint. For syscalls that might block
indefinitely, on an interrupt (like SIGCONT) the kernel will set the
instruction pointer to just before the syscall:
arch/sparc/kernel/signal_64.c
476 static void do_signal(struct pt_regs *regs, unsigned long orig_i0)
477 {
[...]
525 if (restart_syscall) {
526 switch (regs->u_regs[UREG_I0]) {
527 case ERESTARTNOHAND:
528 case ERESTARTSYS:
529 case ERESTARTNOINTR:
530 /* replay the system call when we are done */
531 regs->u_regs[UREG_I0] = orig_i0;
532 regs->tpc -= 4;
533 regs->tnpc -= 4;
534 pt_regs_clear_syscall(regs);
535 fallthrough;
536 case ERESTART_RESTARTBLOCK:
537 regs->u_regs[UREG_G1] = __NR_restart_syscall;
538 regs->tpc -= 4;
539 regs->tnpc -= 4;
540 pt_regs_clear_syscall(regs);
541 }
However, on a SIGCONT it seems that 'g1' register is being clobbered after the
syscall returns. Before 86c5d2cf0c, the 'g1' was always placed jus
before the 'ta' instruction which then reloads the syscall number and restarts
the syscall.
On master, where 'g1' might be placed before 'ta':
$ cat test.c
#include <unistd.h>
int main ()
{
pause ();
}
$ gcc test.c -o test
$ strace -f ./t
[...]
ppoll(NULL, 0, NULL, NULL, 0
On another terminal
$ kill -STOP 2262828
$ strace -f ./t
[...]
--- SIGSTOP {si_signo=SIGSTOP, si_code=SI_USER, si_pid=2521813, si_uid=8289} ---
--- stopped by SIGSTOP ---
And then
$ kill -CONT 2262828
Results in:
--- SIGCONT {si_signo=SIGCONT, si_code=SI_USER, si_pid=2521813, si_uid=8289} ---
restart_syscall(<... resuming interrupted ppoll ...>) = -1 EINTR (Interrupted system call)
Where the expected behaviour would be:
$ strace -f ./t
[...]
ppoll(NULL, 0, NULL, NULL, 0) = ? ERESTARTNOHAND (To be restarted if no handler)
--- SIGSTOP {si_signo=SIGSTOP, si_code=SI_USER, si_pid=2521813, si_uid=8289} ---
--- stopped by SIGSTOP ---
--- SIGCONT {si_signo=SIGCONT, si_code=SI_USER, si_pid=2521813, si_uid=8289} ---
ppoll(NULL, 0, NULL, NULL, 0
Just moving the 'g1' setting near the syscall asm is not suffice,
the compiler might optimize it away (as I saw on cancellation.c by
trying this fix). Instead, I have change the inline asm to put the
'g1' setup in ithe asm block. This would require to change the asm
constraint for INTERNAL_SYSCALL_NCS, since the syscall number is not
constant.
Checked on sparc64-linux-gnu.
Reported-by: René Rebe <rene@exactcode.de>
Tested-by: Sam James <sam@gentoo.org>
Reviewed-by: Sam James <sam@gentoo.org>
