tests/lib/libpthread: Fix tests in subdirectory.

Need to use TESTS_SUBDIRS for this, not SUBDIR. Avoid the extra
nesting level while here; can make the library be a subdirectory of
the directory where the tests live.

PR lib/59685: libcrypto should not depend on libpthread

libpthread: append to SUBDIR instead of overwrite it

Noticed and from <riastradh>, thanks!

PR lib/59685: libcrypto should not depend on libpthread

libpthread: Test pthread stubs in threaded vs non-threaded programs.

PR lib/59685: libcrypto should not depend on libpthread

branches: 1.19.2;
tests/lib/libpthread/t_compat_cancel: Make linker warnings non-fatal.

We will get warnings like:

/home/riastradh/netbsd/current/src/tests/lib/libpthread/t_compat_cancel.c:233: warning: warning: reference to compatibility sigsuspend(); include <signal.h> for correct reference

This is intended -- t_compat_cancel deliberately uses the compat
symbols, not the modern symbols, in order to test the compat symbols.

Fixes clang build because bsd.sys.mk enables -Wl,--fatal-warnings in
LDFLAGS by default.

PR lib/59240: POSIX.1-2024: cancellation point audit
PR lib/59247: pthread_cancelstub.c is inadequately tested

Add tests for compat functions as cancellation points.

While here, test kevent too.

PR lib/59240: POSIX.1-2024: cancellation point audit
PR lib/59247: pthread_cancelstub.c is inadequately tested

pthread_cancel(3): Add some automatic tests.

PR lib/59240: POSIX.1-2024: cancellation point audit
PR lib/59134: POSIX-1.2024: pthread_setcancelstate must be
async-signal-safe

branches: 1.16.2;
pthread: Add tests for pthread user stack allocation.

PR lib/57721

XXX pullup-10
XXX pullup-9
XXX pullup-8

branches: 1.15.6;
fix build of h_thread_local_dtor.cpp

branches: 1.14.2;
Add a complete C11 threads(3) implementation

C11 Thread support library is a portable threading C API between OSs,
similar to std::threads in the C++ world.

The library is implemented as a thin shim over POSIX interfaces.

NetBSD implements the API as a part of the POSIX threading library
(libpthread(3)).

C11 threads(3) are in the process of making them an integral part
of the POSIX standard. The interface has been implemented in major
OSs and used with stopgap libraries for older versions of them.

C11 threading library is already used (with a stopgap implementation)
in the NetBSD distribution in MESA.

Original implementation by myself from 2016.

ATF tests are new and cover almost all interfaces.

Proposed on tech-userlevel@.

branches: 1.13.6;
Implement __cxa_thread_atexit and __cxa_thread_atexit_impl. This
functions are used for destructors of thread_local objects.

If a pending destructor exists, prevent unloading of shared objects.
Introduce __dl_cxa_refcount interface for this purpose. When the last
reference is gone and the object has been dlclose'd before, the
unloading is finalized.

Ideally, __cxa_thread_atexit_impl wouldn't exist, but libstdc++ insists
on providing __cxa_thread_atexit as direct wrapper without further
patching.

branches: 1.12.6;
Add new test t_timedmutex

This test is a clone on t_mutex with additional two tests for timed-mutex
specific block.

All simple-mutex (not with the timed property according to the C11 wording)
specific tests are covered by pthread_mutex_timedlock(3) with parameter
ts_lengthy of sufficiently large tv_sec value (right now UINT16_MAX). If,
a test will hang, it won't wait UINT16_MAX seconds, but will be terminated
within the default timeout for ATF tests (right now 300 [sec] in my
NetBSD/amd64 setup).

This test was inspired by a classic selflock test failure of
pthread_mutex_timedlock(3) of the following form:

#include <assert.h>
#include <errno.h>
#include <pthread.h>
#include <stdio.h>
#include <time.h>

int main(int argc, char **argv)
{
pthread_mutex_t mtx;
struct timespec ts;

ts.tv_sec = 0;
ts.tv_nsec = 1000;
printf("ts{.tv_sec = %d, .tv_nsec=%ld}\n", ts.tv_sec, ts.tv_nsec);
fflush(stdout);

printf("mtx_init\n");
assert(pthread_mutex_init(&mtx, NULL) == 0);

printf("mtx_lock\n");
assert(pthread_mutex_lock(&mtx) == 0);

printf("mtx_timedlock\n");
assert(pthread_mutex_timedlock(&mtx, &ts) == ETIMEDOUT);

printf("mtx_unlock\n");
assert(pthread_mutex_unlock(&mtx) == 0);

printf("mtx_destroy\n");
assert(pthread_mutex_destroy(&mtx) == 0);

return 0;
}

Current NetBSD implementation wrongly hangs on this test.

The issue was detected during development of the C11 portable threads.

My local tests in chroot presents that the are further issues:

t_timedmutex (21/25): 10 test cases
mutex1: [0.001142s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:75: *param != 20
mutex2: [0.261499s] Passed.
mutex3: [0.261496s] Passed.
mutex4: [0.001204s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:265: pthread_mutex_timedlock(&mutex, &ts_lengthy): Connection timed out
mutex5: [0.001235s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:337: pthread_mutex_timedlock(&mutex5, &ts_lengthy): Connection timed out
mutex6: [21.218497s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:512: start != 1
mutexattr1: [0.001328s] Passed.
mutexattr2: [0.001175s] Passed.
timedmutex1: [301.119397s] Failed: Test case timed out after 300 seconds
timedmutex2: [301.123081s] Failed: Test case timed out after 300 seconds
[623.990659s]

I'm also receiveing the same failure in the mutex6 test in t_mutex, so
there might be a false positives due to local chroot(8) issues.

Commit approved by <christos>.

branches: 1.11.10;
loosen the test only for qemu.

Add pthread_cond_timedwait(3) test from PR/47703

new dlopen tests for libpthread from manu@

branches: 1.8.2;
setcontext() used to be incompatible with -lpthread since it affected
the TLS pointer, therefore wrecking the pthread environement.

Some ports had _UC_TLSBASE flag or equivalent (_UC_UNIQUE on alpha)
that controlled whether setcontext() would change the TLS pointer.
This change let libpthread override setcontext() with its own version
that unsets _UC_TLSBASE, enabling safe usage of setcontext() with
-lpthread.

We also have the following required changes here:
- rename alpha's _UC_UNIQUE into _UC_TLSBASE
- add _UC_TLSBASE definition in header file for all ports
(powerpc, sh3, sparc and sparc64 lack the implementation for now)
- introduce a libc stub that can be overriden for setcontext()
- modify MD libcs swapcontext() implementations so that they use the
setcontext() libc stub instead of doing a plain system call.

While we are there:
- document various MD _UC_* flags in header file
- add libc and libpthread tests for swapcontext() behavior
(hopefully helpful to spot MD problems introduced with this change)

Future work:
- Deciding whether kernel support or _UC_TLSBASE should be added for
powerpc, sh3, sparc and sparc64 is left to portmasters
sparc64

Approved by core@

branches: 1.7.4;
As per PR lib/44818, remove 'lib/libpthread/t_status'. It takes two minutes
to rewrite this properly if someone misses this.

A dummy conformance-test for pthread_detach(3). I will extend this later.

A dummy conformance-test of pthread_equal(3).

Link t_fpu against libm if the compiler doesn't want to inline sin/cos.

Test alignment of constructor / destructor calls as well as the stack
of new threads. Currently implement for i386 and AMD64.

Add a simple test for pthread_join(3).

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

branches: 1.1.4;
Add tests for compat functions as cancellation points.

While here, test kevent too.

PR lib/59240: POSIX.1-2024: cancellation point audit
PR lib/59247: pthread_cancelstub.c is inadequately tested

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

beware of your shadow

branches: 1.2.2;
Add new test t_timedmutex

This test is a clone on t_mutex with additional two tests for timed-mutex
specific block.

All simple-mutex (not with the timed property according to the C11 wording)
specific tests are covered by pthread_mutex_timedlock(3) with parameter
ts_lengthy of sufficiently large tv_sec value (right now UINT16_MAX). If,
a test will hang, it won't wait UINT16_MAX seconds, but will be terminated
within the default timeout for ATF tests (right now 300 [sec] in my
NetBSD/amd64 setup).

This test was inspired by a classic selflock test failure of
pthread_mutex_timedlock(3) of the following form:

#include <assert.h>
#include <errno.h>
#include <pthread.h>
#include <stdio.h>
#include <time.h>

int main(int argc, char **argv)
{
pthread_mutex_t mtx;
struct timespec ts;

ts.tv_sec = 0;
ts.tv_nsec = 1000;
printf("ts{.tv_sec = %d, .tv_nsec=%ld}\n", ts.tv_sec, ts.tv_nsec);
fflush(stdout);

printf("mtx_init\n");
assert(pthread_mutex_init(&mtx, NULL) == 0);

printf("mtx_lock\n");
assert(pthread_mutex_lock(&mtx) == 0);

printf("mtx_timedlock\n");
assert(pthread_mutex_timedlock(&mtx, &ts) == ETIMEDOUT);

printf("mtx_unlock\n");
assert(pthread_mutex_unlock(&mtx) == 0);

printf("mtx_destroy\n");
assert(pthread_mutex_destroy(&mtx) == 0);

return 0;
}

Current NetBSD implementation wrongly hangs on this test.

The issue was detected during development of the C11 portable threads.

My local tests in chroot presents that the are further issues:

t_timedmutex (21/25): 10 test cases
mutex1: [0.001142s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:75: *param != 20
mutex2: [0.261499s] Passed.
mutex3: [0.261496s] Passed.
mutex4: [0.001204s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:265: pthread_mutex_timedlock(&mutex, &ts_lengthy): Connection timed out
mutex5: [0.001235s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:337: pthread_mutex_timedlock(&mutex5, &ts_lengthy): Connection timed out
mutex6: [21.218497s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:512: start != 1
mutexattr1: [0.001328s] Passed.
mutexattr2: [0.001175s] Passed.
timedmutex1: [301.119397s] Failed: Test case timed out after 300 seconds
timedmutex2: [301.123081s] Failed: Test case timed out after 300 seconds
[623.990659s]

I'm also receiveing the same failure in the mutex6 test in t_mutex, so
there might be a false positives due to local chroot(8) issues.

Commit approved by <christos>.

branches: 1.1.28;
Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

add joins

branches: 1.2.46;
add Makefile.inc everywhere so that we can set WARNS=4 by default. Amazing
how many bugs this found :-)

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

branches: 1.1.4;
Implement __cxa_thread_atexit and __cxa_thread_atexit_impl. This
functions are used for destructors of thread_local objects.

If a pending destructor exists, prevent unloading of shared objects.
Introduce __dl_cxa_refcount interface for this purpose. When the last
reference is gone and the object has been dlclose'd before, the
unloading is finalized.

Ideally, __cxa_thread_atexit_impl wouldn't exist, but libstdc++ insists
on providing __cxa_thread_atexit as direct wrapper without further
patching.

Adjusts tests after import of atf-0.12:

- The use.fs property is gone.
- Mark the tests/fs/t_create:attrs test as broken when using the default
unprivileged-user:_atf setting. This probably deserves a fix somehow
but I'm not sure at this point.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

add Makefile.inc everywhere so that we can set WARNS=4 by default. Amazing
how many bugs this found :-)

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

branches: 1.1.2;
Add a complete C11 threads(3) implementation

C11 Thread support library is a portable threading C API between OSs,
similar to std::threads in the C++ world.

The library is implemented as a thin shim over POSIX interfaces.

NetBSD implements the API as a part of the POSIX threading library
(libpthread(3)).

C11 threads(3) are in the process of making them an integral part
of the POSIX standard. The interface has been implemented in major
OSs and used with stopgap libraries for older versions of them.

C11 threading library is already used (with a stopgap implementation)
in the NetBSD distribution in MESA.

Original implementation by myself from 2016.

ATF tests are new and cover almost all interfaces.

Proposed on tech-userlevel@.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

branches: 1.4.4;
Add tests for compat functions as cancellation points.

While here, test kevent too.

PR lib/59240: POSIX.1-2024: cancellation point audit
PR lib/59247: pthread_cancelstub.c is inadequately tested

libpthread(3): Add missing cancellation points.

- accept4
- tcdrain

XXX Should our nonstandard alias paccept(2) also be a cancellation
point?

PR lib/59240: POSIX.1-2024: cancellation point audit

pthread_cancel(3): Rework.

Make pthread_setcancelstate(3) async-signal-safe. (As a side effect,
this also makes pthread_setcanceltype(3) async-signal-safe, although
that is not required.)

PR lib/59134: POSIX-1.2024: pthread_setcancelstate must be
async-signal-safe

pthread_cancel(3): Add some automatic tests.

PR lib/59240: POSIX.1-2024: cancellation point audit
PR lib/59134: POSIX-1.2024: pthread_setcancelstate must be
async-signal-safe

branches: 1.1.2;
Add a complete C11 threads(3) implementation

C11 Thread support library is a portable threading C API between OSs,
similar to std::threads in the C++ world.

The library is implemented as a thin shim over POSIX interfaces.

NetBSD implements the API as a part of the POSIX threading library
(libpthread(3)).

C11 threads(3) are in the process of making them an integral part
of the POSIX standard. The interface has been implemented in major
OSs and used with stopgap libraries for older versions of them.

C11 threading library is already used (with a stopgap implementation)
in the NetBSD distribution in MESA.

Original implementation by myself from 2016.

ATF tests are new and cover almost all interfaces.

Proposed on tech-userlevel@.

branches: 1.3.4;
t_compat_cancel: Use -1 rather than SIGALRM for expected signal.

On kernels without (e.g.) COMPAT_13, we get SIGSYS rather than
SIGALRM if the cancellation point doesn't cancel the syscall.

Won't help if we try to interrupt in-progress waits, but currently
none of the tests do that; we'll burn that bridge when we come to it,

PR lib/59240: POSIX.1-2024: cancellation point audit
PR lib/59247: pthread_cancelstub.c is inadequately tested

t_compat_cancel: Fix compat numbers to match reality.

- msync is as it existed in 1.2, and is implemented by compat12;
__msync13 was new in 1.3.

- sigsuspend is as it existed in 1.3, and is implemented by compat13;
__sigsuspend14 was new in 1.4.

(But, while select is as it existed in 5.0, and is implemented by
compat50, __select50 was new in _6.0_, not in 5.0.)

No functional change intended other than to rename the test cases for
clarity.

PR lib/59240: POSIX.1-2024: cancellation point audit
PR lib/59247: pthread_cancelstub.c is inadequately tested

Add tests for compat functions as cancellation points.

While here, test kevent too.

PR lib/59240: POSIX.1-2024: cancellation point audit
PR lib/59247: pthread_cancelstub.c is inadequately tested

Adjust cond_timedwait_race to take account of spurious wakeups (which are
completely legit).

GSoC 2016 Charles Cui: Implement thread priority protection based on work
by Andy Doran. Also document the get/set pshared thread calls as not
implemented, and add a skeleton implementation that is disabled.
XXX: document _sched_protect(2).

The cond_timedwait_race test case is no longer expected to fail; it
has been consistently passing since CVS date 2014.01.31.19.22.00.
See also PR lib/44756.

branches: 1.5.4;
fix unused variable warnings

Try to trigger the cond_timedwait_race race several times.

Sometime this tests passes (after all, it's exercising a race condition) and
when it does it's reported as a failure. By giving the test a few chances
to expose the problem, we prevent this noisy signal. When the race is really
addressed, this will start failing consistently as expected.

branches: 1.3.4; 1.3.10;
Add a test case for pthread_cond_timedwait(3) failures reported by
Sad Clouds in PR lib/44756. This was discussed also in:

http://mail-index.netbsd.org/tech-userlevel/2011/03/17/msg004689.html

add Makefile.inc everywhere so that we can set WARNS=4 by default. Amazing
how many bugs this found :-)

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

This test makes (made) a false assumption about the way that
process scheduling works. That a process (or in this case,
a thread) is no longer blocked at time T does not mean that it
will resume execution at time T. The OS is free to devote
resources to other processes/threads instead - all we should
normally be able to expect is that if it is not unblocked before
time T, that it will not start running before then.

In general though, the pthread_cond_*wait() functions don't guarantee
even that - but for this test, the possibility of something else
randomly signalling the condvar isn't believable, so don't worry about
that possibility (but do fail without calling strerror(0) on the off
chance it does happen).

Once we cease testing that the process resumed running before some
particular time, we can stop dealing with qemu timekeeping issues,
it might (seem to) take qemu twice as long as was requested before
the thread resumes, but that's OK - the same thing could happen on
a loaded system for other reasons.

Beyond that, the test also has (had) a race condition. When using
CLOCK_REALTIME though that clock needed to have advanced to T before
the ETIMEDOUT should happen, there is no guarantee that it will stay >T
(CLOCK_REALTIME is allowed to be reset backwards). So, only test
that the current time (after ETIMEDOUT) >= T when we're using
CLOCK_MONOTONIC - for CLOCK_REALTIME the time might have stepped
back between when the ETIMEDOUT happened and when the thread
obtains the current clock reading. For that case, all we can test
is that the ETIMEDOUT actually happens.

With much of what was there now gone, the code can be simplified,
we no longer need to do timespec arithmetic, just one comparison
(simpler to test that Tend >= Tstart+period than Tend-Tstart > period
as we need Tstart+period for the abstime value for the timeout anyway).

Note that this still tests for the issue reported in PR lib/47703
which is where the test came from in the first place.

ps: we seem to be missing pthread_cond_clockwait() which is the same
as pthread_cond_timedwait() except that the clock to use is passed
as a parameter, rather than as an attribute of the condition variable.

fix various typos in comments and log messages.

Re-enable the QEMU specific timing limits, but only an increased upper
limit for now - let's see if that works on the test-bed.

PR lib/54440: adapt the FreeBSD change to this test and calculate time
differences more exact, allowing between 0 and 1 s delay between the
expected wakeup and the actual event happening.
Also convert the QEMU special case code to the same scheme, but for now
disable it (with XXX mark) and see if the proper timing limits fix that
case too.
If not, we will re-enable the QEMU special case.

return return atf_no_error() instead of 0 for consistency.

suggested by moritzbuhl in https://github.com/NetBSD/src/pull/11/

branches: 1.5.12;
PR/51884: Ngie Cooper: Include and message fixes

branches: 1.4.2; 1.4.8; 1.4.14; 1.4.18;
loosen the test only for qemu.

Make cond_wait_mono and cond_wait_real tests accept a wait time in the
range of 90% to 250% of nominal, to allow the test to pass under qemu
which has a known issue where timing can be off by a factor of two.

fix printf formats

Add pthread_cond_timedwait(3) test from PR/47703

PR/51885: Ngie Cooper: Add delays to avoid races.

branches: 1.1.26; 1.1.30;
A dummy conformance-test for pthread_detach(3). I will extend this later.

A dummy conformance-test of pthread_equal(3).

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

PR/51886: Ngie Cooper: Use _exit instead of exit

branches: 1.1.28; 1.1.32;
Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

PR/51887: Ngie Cooper: Include and error message fixes

branches: 1.2.12; 1.2.16;
exitting -> exiting.

branches: 1.1.6; 1.1.12;
Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Fix the t_join test

For pthread_attr_get_np() attr should be initialized prior to the call by
using pthread_attr_init(3). pthread_getattr_np() does this
automatically so switch to it.

branches: 1.9.6;
Export the guard size of the main thread via vm.guard_size. Add a
complementary writable sysctl for the initial guard size of threads
created via pthread_create. Let the existing attribut accessors do the
right thing. Raise the default guard size for threads to 64KB.

branches: 1.8.24;
Also exercise pthread_attr_setstacksize.

Remove bogus check that (a) depends uninitialised memory (b) overflows a
buffer.

branches: 1.6.2;
Requires stdint.h.

Test alignment of constructor / destructor calls as well as the stack
of new threads. Currently implement for i386 and AMD64.

Revert to 1.1 and fix the casts by using uintptr_t. Ok jruoho.

One more cast.

Fix build failure on AMD64, noted by njoly@.

Add a simple test for pthread_join(3).

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Do not destroy mutices that failed to init - fixes a run with
PTHREAD_DIAGASSERT set to "a". Pointed out by joerg.

branches: 1.1.2;
Add a complete C11 threads(3) implementation

C11 Thread support library is a portable threading C API between OSs,
similar to std::threads in the C++ world.

The library is implemented as a thin shim over POSIX interfaces.

NetBSD implements the API as a part of the POSIX threading library
(libpthread(3)).

C11 threads(3) are in the process of making them an integral part
of the POSIX standard. The interface has been implemented in major
OSs and used with stopgap libraries for older versions of them.

C11 threading library is already used (with a stopgap implementation)
in the NetBSD distribution in MESA.

Original implementation by myself from 2016.

ATF tests are new and cover almost all interfaces.

Proposed on tech-userlevel@.

Bump timeout for slow machines.

Since there has been no objection (or even comment) in response
to my message on tech-userlevel ...

Subject: tests/lib/libpthread/t_mutex:mutex6
Date: Thu, 23 Nov 2017 17:34:54 +0700
Message-ID: <28385.1511433294@andromeda.noi.kre.to>

which can be found at:
http://mail-index.netbsd.org/tech-userlevel/2017/11/23/msg011010.html

which analysed the mutex6 test case of this test, and concluded
that it was useless, nonsense, and broken (the whole test is just a
race - not even really using or testing mutexes), let it be henceforth
forever gone.

branches: 1.18.4;
Make the mutex6 test (which fails every know and then, but too rarely
to actually debug it) print some debug info when failing.

Avoid using an uninitialized mutex.

reenable mutex2 and mutex3 on powerpc now that PR 44387 is fixed.

PR/51888: Ngie Cooper: more error checking, missing includes

branches: 1.14.2;
more tests from kamil

fix typos

Merge and fix the timed mutex tests to use absolute time.
NB: the new tests are broken?

Add new test t_timedmutex

This test is a clone on t_mutex with additional two tests for timed-mutex
specific block.

All simple-mutex (not with the timed property according to the C11 wording)
specific tests are covered by pthread_mutex_timedlock(3) with parameter
ts_lengthy of sufficiently large tv_sec value (right now UINT16_MAX). If,
a test will hang, it won't wait UINT16_MAX seconds, but will be terminated
within the default timeout for ATF tests (right now 300 [sec] in my
NetBSD/amd64 setup).

This test was inspired by a classic selflock test failure of
pthread_mutex_timedlock(3) of the following form:

#include <assert.h>
#include <errno.h>
#include <pthread.h>
#include <stdio.h>
#include <time.h>

int main(int argc, char **argv)
{
pthread_mutex_t mtx;
struct timespec ts;

ts.tv_sec = 0;
ts.tv_nsec = 1000;
printf("ts{.tv_sec = %d, .tv_nsec=%ld}\n", ts.tv_sec, ts.tv_nsec);
fflush(stdout);

printf("mtx_init\n");
assert(pthread_mutex_init(&mtx, NULL) == 0);

printf("mtx_lock\n");
assert(pthread_mutex_lock(&mtx) == 0);

printf("mtx_timedlock\n");
assert(pthread_mutex_timedlock(&mtx, &ts) == ETIMEDOUT);

printf("mtx_unlock\n");
assert(pthread_mutex_unlock(&mtx) == 0);

printf("mtx_destroy\n");
assert(pthread_mutex_destroy(&mtx) == 0);

return 0;
}

Current NetBSD implementation wrongly hangs on this test.

The issue was detected during development of the C11 portable threads.

My local tests in chroot presents that the are further issues:

t_timedmutex (21/25): 10 test cases
mutex1: [0.001142s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:75: *param != 20
mutex2: [0.261499s] Passed.
mutex3: [0.261496s] Passed.
mutex4: [0.001204s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:265: pthread_mutex_timedlock(&mutex, &ts_lengthy): Connection timed out
mutex5: [0.001235s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:337: pthread_mutex_timedlock(&mutex5, &ts_lengthy): Connection timed out
mutex6: [21.218497s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:512: start != 1
mutexattr1: [0.001328s] Passed.
mutexattr2: [0.001175s] Passed.
timedmutex1: [301.119397s] Failed: Test case timed out after 300 seconds
timedmutex2: [301.123081s] Failed: Test case timed out after 300 seconds
[623.990659s]

I'm also receiveing the same failure in the mutex6 test in t_mutex, so
there might be a false positives due to local chroot(8) issues.

Commit approved by <christos>.

we require root for sched_fifo, and more verbose messages.

branches: 1.9.2;
fix uninitialized var.

GSoC 2016 Charles Cui: Implement thread priority protection based on work
by Andy Doran. Also document the get/set pshared thread calls as not
implemented, and add a skeleton implementation that is disabled.
XXX: document _sched_protect(2).

PR misc/49356 remove unnecessary references to atf-c/config.h

The function included via this header is not used and is removed in
later versions of atf, so let us avoid it.

Use compiler builtins instead of atf_arch and atf_machine.

The atf_arch and atf_machine configuration variables were removed from
atf-0.19 without me realizing that some tests were querying them directly.

Instead of reintroducing those variables, just rely on compiler builtins
as many other tests already do.

Should fix PR bin/48582.

branches: 1.5.4; 1.5.10;
Do only set a (lower than default) timeout value if we are on powerpc and
expect to run into a real deadrun ofter.
The (too short) timeout otherwise makes slower archs fail.

mutex2/mutex3 are expected to fail on powerpc because of
PR port-powerpc/44387.

XXX the ugly sleep at the end is because ATF will mark an un-triggered
race condition (ie, the test passes unexpectedly) as a test failure otherwise.

The mutex3 test is supposed to use a static initializer so that it differs
to the mutex2 test. This detail was lost during the atfification of these
tests.

Spotted by pooka@ and riz@.

branches: 1.2.2;
Add missing pthread_mutex_init call in mutex3 testcase.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

tests/lib/libpthread/t_once: Test fork and pthread_once race.

Test is disabled by default because it triggers only with very low
probability in each trial, but let's make it easy to run if you want.

PR lib/59125: pthread_once(3) races with concurrent fork

branches: 1.2.16;
PR/49003: Ngie Cooper: add <sys/time.h>

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

add Makefile.inc everywhere so that we can set WARNS=4 by default. Amazing
how many bugs this found :-)

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Check that PTHREAD_RWLOCK_INITIALIZER works.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

PR/51889: Ngie Cooper: add sys/time.h

branches: 1.8.2; 1.8.4;
PR misc/49356 remove unnecessary references to atf-c/config.h

The function included via this header is not used and is removed in
later versions of atf, so let us avoid it.

It is perfectly valid for sem_wait to be interrupted, so loop on EINTR.

Remove libpthread's semaphore implementation and always use the kernel
one. The implementation doesn't provide an async-safe sem_post and can't
without a lot of work on the pthread primitives.

Remove bogus time out requirement in test case, it should have been
a "known failure" if anything.

branches: 1.5.6;
add Makefile.inc everywhere so that we can set WARNS=4 by default. Amazing
how many bugs this found :-)

before_start_one_thread is not racy only in amd64; my confusion. Mark it
as an xfail everywhere and, while doing so, make the test longer so that
we trigger the failure all the time -- of course, being a race this may
still happen but the chances should be pretty low.

Split the before_start test in two: one that does not use threads and one
that does. The former works all the time but the latter gets consistently
stuck on amd64. Mark the latter as an expected timeout (should be a "race
condition" test, but atf does not have such a thing yet[1]).

This clears the test failures, at least, under anita running NetBSD/i386.

From pooka@: this could well be because calling sem_post(3) from a signal
handler can't possibly do the right thing with the pthread implementation.
However, according to signal(7), sem_post(3) is signal-async safe...

While here, program alarms using a timeout shorter than 1 second to speed
up the execution of the tests.

1: Good thing is I finally understand what a "race condition" test looks
like, I believe.

Reduce sleep time for practically the same effect. Until atf can
do parallel test execution, tests which spend 99.9999% of their
execution sleeping are not nice.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

fix unused variable warnings

branches: 1.2.6; 1.2.12;
add Makefile.inc everywhere so that we can set WARNS=4 by default. Amazing
how many bugs this found :-)

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

branches: 1.6.2; 1.6.4; 1.6.6;
pthread: Don't adjust user-allocated stack addresses by guardsize.

PR lib/57721

XXX pullup-10
XXX pullup-9
XXX pullup-8

libpthread/t_stack: Fix format string for size_t.

Tested this on i386 since that had been crashing before, but i386
doesn't see %zu for unsigned int as a problem.

PR lib/57721

XXX pullup-10
XXX pullup-9
XXX pullup-8

libpthread/t_stack: Appease gcc12 maybe-uninitialized warning.

The jmp_buf is not, in fact, uninitialized at the point of use, but
it doesn't hurt to narrow the scope a bit to between when the jmp_buf
is initialized by setjmp, and when the signal handler might be called
after sigaction.

Noted by prlw1.

PR lib/57721

XXX pullup-10
XXX pullup-9
XXX pullup-8

libpthread/t_stack: Omit needless cast in previous.

Arose from an earlier draft of the change.

PR lib/57721

XXX pullup-10
XXX pullup-9
XXX pullup-8

libpthread/t_stack: Make this more robust to the guard size bug.

Make sure to allocate enough space for the thread's stack for a guard
even though there shouldn't be one, so that when we run the thread,
it doesn't start with the stack pointer pointing into someone else's
allocation (like malloc) causing stack frames to trash another data
structure -- or causing the user of that data structure to trash the
stack frames.

PR lib/57721

XXX pullup-10
XXX pullup-9
XXX pullup-8

pthread: Add tests for pthread user stack allocation.

PR lib/57721

XXX pullup-10
XXX pullup-9
XXX pullup-8

As per PR lib/44818, remove 'lib/libpthread/t_status'. It takes two minutes
to rewrite this properly if someone misses this.

add Makefile.inc everywhere so that we can set WARNS=4 by default. Amazing
how many bugs this found :-)

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory. Will investigate later.

Simplify - it's enough to set nctx.uc_link to &octx to restore old
context automatically when swapfunc() returns.

Add missing $ in the RCSID.

threadfunc - we do return from swapcontext() (since revision 1.2),
so drop the NOTREACHED comment.

swapfunc() is not passed to pthread_create, there's no need for it to
have any arguments.

Drop unnecessary cast to makecontext() argument.

Cleanup trailing whitespace.

PR/51890: Ngie Cooper: include and error message fixes.

branches: 1.2.2; 1.2.4;
Go back to the initial context (as tests/lib/libc/sys/t_swapcontext.c does)
after checking pthread_self() didn't change. Otherwise the process exits
outside of atf context.
Should fix "Test case exited normally but failed to create the results file: Results file is empty" reports from atf-run.

branches: 1.1.4; 1.1.12;
setcontext() used to be incompatible with -lpthread since it affected
the TLS pointer, therefore wrecking the pthread environement.

Some ports had _UC_TLSBASE flag or equivalent (_UC_UNIQUE on alpha)
that controlled whether setcontext() would change the TLS pointer.
This change let libpthread override setcontext() with its own version
that unsets _UC_TLSBASE, enabling safe usage of setcontext() with
-lpthread.

We also have the following required changes here:
- rename alpha's _UC_UNIQUE into _UC_TLSBASE
- add _UC_TLSBASE definition in header file for all ports
(powerpc, sh3, sparc and sparc64 lack the implementation for now)
- introduce a libc stub that can be overriden for setcontext()
- modify MD libcs swapcontext() implementations so that they use the
setcontext() libc stub instead of doing a plain system call.

While we are there:
- document various MD _UC_* flags in header file
- add libc and libpthread tests for swapcontext() behavior
(hopefully helpful to spot MD problems introduced with this change)

Future work:
- Deciding whether kernel support or _UC_TLSBASE should be added for
powerpc, sh3, sparc and sparc64 is left to portmasters
sparc64

Approved by core@

Remove extra semicolon.

branches: 1.1.2;
Add a complete C11 threads(3) implementation

C11 Thread support library is a portable threading C API between OSs,
similar to std::threads in the C++ world.

The library is implemented as a thin shim over POSIX interfaces.

NetBSD implements the API as a part of the POSIX threading library
(libpthread(3)).

C11 threads(3) are in the process of making them an integral part
of the POSIX standard. The interface has been implemented in major
OSs and used with stopgap libraries for older versions of them.

C11 threading library is already used (with a stopgap implementation)
in the NetBSD distribution in MESA.

Original implementation by myself from 2016.

ATF tests are new and cover almost all interfaces.

Proposed on tech-userlevel@.

branches: 1.1.4;
Implement __cxa_thread_atexit and __cxa_thread_atexit_impl. This
functions are used for destructors of thread_local objects.

If a pending destructor exists, prevent unloading of shared objects.
Introduce __dl_cxa_refcount interface for this purpose. When the last
reference is gone and the object has been dlclose'd before, the
unloading is finalized.

Ideally, __cxa_thread_atexit_impl wouldn't exist, but libstdc++ insists
on providing __cxa_thread_atexit as direct wrapper without further
patching.

branches: 1.2.2;
Merge and fix the timed mutex tests to use absolute time.
NB: the new tests are broken?

Add new test t_timedmutex

This test is a clone on t_mutex with additional two tests for timed-mutex
specific block.

All simple-mutex (not with the timed property according to the C11 wording)
specific tests are covered by pthread_mutex_timedlock(3) with parameter
ts_lengthy of sufficiently large tv_sec value (right now UINT16_MAX). If,
a test will hang, it won't wait UINT16_MAX seconds, but will be terminated
within the default timeout for ATF tests (right now 300 [sec] in my
NetBSD/amd64 setup).

This test was inspired by a classic selflock test failure of
pthread_mutex_timedlock(3) of the following form:

#include <assert.h>
#include <errno.h>
#include <pthread.h>
#include <stdio.h>
#include <time.h>

int main(int argc, char **argv)
{
pthread_mutex_t mtx;
struct timespec ts;

ts.tv_sec = 0;
ts.tv_nsec = 1000;
printf("ts{.tv_sec = %d, .tv_nsec=%ld}\n", ts.tv_sec, ts.tv_nsec);
fflush(stdout);

printf("mtx_init\n");
assert(pthread_mutex_init(&mtx, NULL) == 0);

printf("mtx_lock\n");
assert(pthread_mutex_lock(&mtx) == 0);

printf("mtx_timedlock\n");
assert(pthread_mutex_timedlock(&mtx, &ts) == ETIMEDOUT);

printf("mtx_unlock\n");
assert(pthread_mutex_unlock(&mtx) == 0);

printf("mtx_destroy\n");
assert(pthread_mutex_destroy(&mtx) == 0);

return 0;
}

Current NetBSD implementation wrongly hangs on this test.

The issue was detected during development of the C11 portable threads.

My local tests in chroot presents that the are further issues:

t_timedmutex (21/25): 10 test cases
mutex1: [0.001142s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:75: *param != 20
mutex2: [0.261499s] Passed.
mutex3: [0.261496s] Passed.
mutex4: [0.001204s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:265: pthread_mutex_timedlock(&mutex, &ts_lengthy): Connection timed out
mutex5: [0.001235s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:337: pthread_mutex_timedlock(&mutex5, &ts_lengthy): Connection timed out
mutex6: [21.218497s] Failed: /usr/src/tests/lib/libpthread/t_timedmutex.c:512: start != 1
mutexattr1: [0.001328s] Passed.
mutexattr2: [0.001175s] Passed.
timedmutex1: [301.119397s] Failed: Test case timed out after 300 seconds
timedmutex2: [301.123081s] Failed: Test case timed out after 300 seconds
[623.990659s]

I'm also receiveing the same failure in the mutex6 test in t_mutex, so
there might be a false positives due to local chroot(8) issues.

Commit approved by <christos>.

branches: 1.1.2;
Add a complete C11 threads(3) implementation

C11 Thread support library is a portable threading C API between OSs,
similar to std::threads in the C++ world.

The library is implemented as a thin shim over POSIX interfaces.

NetBSD implements the API as a part of the POSIX threading library
(libpthread(3)).

C11 threads(3) are in the process of making them an integral part
of the POSIX standard. The interface has been implemented in major
OSs and used with stopgap libraries for older versions of them.

C11 threading library is already used (with a stopgap implementation)
in the NetBSD distribution in MESA.

Original implementation by myself from 2016.

ATF tests are new and cover almost all interfaces.

Proposed on tech-userlevel@.

branches: 1.1.4; 1.1.10;
new dlopen tests for libpthread from manu@

libpthread: Link with -Wl,-z,nodelete.

Can't safely unload libpthread because of the interaction with libc
thread stubs.

PR lib/59784: dlopening and dlclosing libpthread is broken

tests/lib/libpthread: Test unloading libpthread after lazy binding.

If you dlopen libpthread and dlclose it again, the thread stubs like
pthread_mutex_lock need to continue working -- a library might have
calls to it in order to support thread-safety for threaded
applications, but that library needs to continue working even in
non-threaded applications after lazy binding of the libpthread symbol
instead of the libc stub.

PR lib/59784: dlopening and dlclosing libpthread is broken

branches: 1.1.4; 1.1.10; 1.1.48;
new dlopen tests for libpthread from manu@

tests/lib/libpthread: Don't abuse xfail.

Use a signal handler to check for SIGABRT, rather than
atf_tc_expect_signal.

xfail is for when there is a bug that we haven't fixed yet and the
test manifests a symptom of that bug -- a list of xfails is a list of
open bugs to be fixed. In this case, we are verifying that
pthread_create _correctly_ raises SIGABRT (or fails with nonzero
return code -- both are acceptable outcomes, really), and there is no
bug here at the moment.

Prompted by (but unrelated to):

PR lib/59784: dlopening and dlclosing libpthread is broken

branches: 1.1.4; 1.1.10; 1.1.48;
new dlopen tests for libpthread from manu@

branches: 1.1.4; 1.1.10;
new dlopen tests for libpthread from manu@

branches: 1.1.4; 1.1.10;
new dlopen tests for libpthread from manu@

branches: 1.1.4; 1.1.10;
new dlopen tests for libpthread from manu@

branches: 1.2.2;
tests/lib/libpthread: Fix tests in subdirectory.

Need to use TESTS_SUBDIRS for this, not SUBDIR. Avoid the extra
nesting level while here; can make the library be a subdirectory of
the directory where the tests live.

PR lib/59685: libcrypto should not depend on libpthread

libpthread: Test pthread stubs in threaded vs non-threaded programs.

PR lib/59685: libcrypto should not depend on libpthread

branches: 1.1.2;
libpthread: Test pthread stubs in threaded vs non-threaded programs.

PR lib/59685: libcrypto should not depend on libpthread

branches: 1.1.2;
tests/lib/libpthread: Fix tests in subdirectory.

Need to use TESTS_SUBDIRS for this, not SUBDIR. Avoid the extra
nesting level while here; can make the library be a subdirectory of
the directory where the tests live.

PR lib/59685: libcrypto should not depend on libpthread

branches: 1.1.2;
tests/lib/libpthread: Fix tests in subdirectory.

Need to use TESTS_SUBDIRS for this, not SUBDIR. Avoid the extra
nesting level while here; can make the library be a subdirectory of
the directory where the tests live.

PR lib/59685: libcrypto should not depend on libpthread

branches: 1.1.2;
libpthread: Test pthread stubs in threaded vs non-threaded programs.

PR lib/59685: libcrypto should not depend on libpthread

branches: 1.1.2;
libpthread: Test pthread stubs in threaded vs non-threaded programs.

PR lib/59685: libcrypto should not depend on libpthread

branches: 1.1.2;
libpthread: Test pthread stubs in threaded vs non-threaded programs.

PR lib/59685: libcrypto should not depend on libpthread