3fefd3aa2d
vcomp/tests: Add tests for 64-bit atomic instructions.
...
Signed-off-by: Sebastian Lackner <sebastian@fds-team.de>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
2016-02-10 21:50:09 +09:00
a77e14714e
vcomp/tests: Reenable architecture dependent tests.
...
Signed-off-by: Sebastian Lackner <sebastian@fds-team.de>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
2016-02-10 21:49:40 +09:00
0410cf9cbe
vcomp: Implement _vcomp_get_thread_num.
...
Signed-off-by: Sebastian Lackner <sebastian@fds-team.de>
2015-10-07 21:28:49 +09:00
146df7dab1
vcomp: Implement OpenMP regular locking functions.
...
Native OpenMP also immediately terminates the process (without
exception) when calling omp_set_lock while holding the lock from the
same thread.
2015-08-18 14:30:30 +09:00
f212a137a6
vcomp: Implement OpenMP nested locking functions.
...
The test marked with todo_wine suggests that native OpenMP might use a
separate counter for the nesting level. However, to implement the owner
checks required for regular locks we need to access the critical section
private members anyway.
2015-08-18 14:29:53 +09:00
037a079fdf
vcomp: Export omp_init_nest_lock and omp_destroy_nest_lock.
2015-08-18 14:29:47 +09:00
ce18a73304
vcomp: Implement stub for _vcomp_flush and add tests.
2015-08-18 14:29:31 +09:00
15495d19ef
vcomp/tests: Make float test i386 only.
...
amd64 and arm don't pass floats on the stack
2015-08-17 16:58:33 +09:00
e40970bce7
vcomp/tests: Add ARM and ARM64 arch strings.
2015-08-17 16:58:29 +09:00
f0b32b8d57
vcomp/tests: Comment out architecture dependent tests (shift count overflow).
...
Shifting more than variable width adds zeros on ARM to fill up.
2015-08-17 16:58:18 +09:00
7734de65c7
vcomp/tests: Make my_for_static_init() static.
2015-08-07 23:50:19 +09:00
7fd7bcf00e
vcomp/tests: Add tests for _vcomp_enter_critsect and _vcomp_leave_critsect.
2015-08-07 16:11:53 +09:00
89b383740c
vcomp/tests: Add tests for _vcomp_single_begin and _vcomp_single_end.
...
Yes, its intentional that the call to p_vcomp_single_end() is after the if-block.
Somehow MS made a difference here between the master and the single directive...
2015-08-07 16:10:04 +09:00
0ff9a68af3
vcomp/tests: Add tests for _vcomp_master_begin and _vcomp_master_end.
2015-08-07 16:09:45 +09:00
d3c8055b48
vcomp/tests: Add tests for dynamic work scheduling.
2015-08-06 23:36:02 +09:00
4bcdbe0982
vcomp/tests: Print the correct variable in ok() message.
2015-07-30 11:35:54 +02:00
dfe32d6c52
vcomp/tests: Add tests for atomic double functions.
2015-07-28 16:56:04 +02:00
c14db31b9b
vcomp/tests: Add tests for atomic float functions.
2015-07-28 16:55:38 +02:00
c68a63e474
vcomp/tests: Add tests for 32-bit atomic integer functions.
2015-07-28 16:54:57 +02:00
a250c0b29f
vcomp/tests: Add tests for _vcomp_for_static_init.
2015-07-20 06:34:46 +02:00
02054a7522
vcomp/tests: Add tests for _vcomp_for_static_simple_init.
2015-07-20 06:34:31 +02:00
bc10d127b4
vcomp: Implement omp_in_parallel and add tests.
2015-07-20 06:34:03 +02:00
c06167b172
vcomp: Fix handling of _vcomp_fork with ifval == FALSE.
...
Forks with ifval == FALSE do not count as "parallel", so nested forks are
still allowed. Please note that calling _vcomp_fork(FALSE, ...) or directly
calling the callback is still different in some aspects, the synchronization
functions run in a different context for example.
2015-07-20 06:33:47 +02:00
889eba36e9
vcomp: Allow calls to _vcomp_sections_init outside of parallel environment.
2015-07-20 06:33:33 +02:00
57cbcec0c2
vcomp/tests: Add tests for vcomp sections.
2015-07-17 16:12:40 +09:00
9c9d60f020
vcomp/tests: Add additional tests for _vcomp_fork.
2015-07-17 16:11:17 +09:00
c0a664456e
vcomp/tests: Add initial tests.
2015-07-17 16:10:18 +09:00
Sebastian Lackner