| Commit message (Collapse) | Author | Files | Lines |
|---|
|
|
|
Replace STATES_DELAYING with STATES_WAITING_FOR_TIME.
There is no need for separate timeout thread states. The
Thread_Control::Timer::header and Watchdog_Control::cpu members can be
used to figure out the kind of timeout.
|
|
|
|
|
|
Update #2825.
|
|
The fatal is internal indicator is redundant since the fatal source and
error code uniquely identify a fatal error. Keep the fatal user
extension is internal parameter for backward compatibility and set it to
false always.
Update #2825.
|
|
Update #2797.
|
|
Fix thread dispatch profiling of rtems_scheduler_add_processor().
Update #2797.
|
|
Initialize thread queue context early preferably outside the critical
section.
Remove implicit _Thread_queue_Context_initialize() from
_Thread_Wait_acquire().
|
|
Move thread state for _Thread_queue_Enqueue() to the thread queue
context. This reduces the parameter count of _Thread_queue_Enqueue()
from five to four (ARM for example has only four function parameter
registers). Since the thread state is used after several function calls
inside _Thread_queue_Enqueue() this parameter was saved on the stack
previously.
|
|
Turn pthread_spinlock_t into a self-contained object. On uni-processor
configurations, interrupts are disabled in the lock/trylock operations
and the previous interrupt status is restored in the corresponding
unlock operations. On SMP configurations, a ticket lock is a acquired
and released in addition.
The self-contained pthread_spinlock_t object is defined by Newlib in
<sys/_pthreadtypes.h>.
typedef struct {
struct _Ticket_lock_Control _lock;
__uint32_t _interrupt_state;
} pthread_spinlock_t;
This implementation is simple and efficient. However, this test case of
the Linux Test Project would fail due to call of printf() and sleep()
during spin lock ownership:
https://github.com/linux-test-project/ltp/blob/master/testcases/open_posix_testsuite/conformance/interfaces/pthread_spin_lock/1-2.c
There is only limited support for profiling on SMP configurations.
Delete CORE spinlock implementation.
Update #2674.
|
|
Delete unused _Thread_queue_Enqueue() and rename
_Thread_queue_Enqueue_critical() to _Thread_queue_Enqueue().
|
|
Replace the expected thread dispatch disable level with a thread queue
enqueue callout. This enables the use of _Thread_Dispatch_direct() in
the thread queue enqueue procedure. This avoids impossible exection
paths, e.g. Per_CPU_Control::dispatch_necessary is always true.
|
|
Update #2808.
|
|
This function is useful for operations which synchronously block, e.g.
self restart, self deletion, yield, sleep. It helps to detect if these
operations are called in the wrong context. Since the thread dispatch
necessary indicator is not used, this is more robust in some SMP
situations.
Update #2751.
|
|
Update #2797.
|
|
Avoid dead code in non-SMP configurations. Return scheduler identifier
independent of the current processor count of the scheduler via
rtems_scheduler_ident(), since this value may change during run-time.
Check the processor count in _Scheduler_Set() under scheduler lock
protection.
Update #2797.
|
|
Use the home scheduler of the executing thread for the created thread.
This is in line with pthread_create(). Using the current processor may
pick up an unexpected scheduler in case of a temporary migration, e.g.
due to locking protocols.
|
|
This makes it easier to conditionally enable/disable the thread resource
count usage.
|
|
Binary semaphores (not simple binary semaphores) maintain the resource
count since 8797c76addf22a2f0ffc3717ff977695e35b9b0b. Do this also for
initially locked binary semaphores.
|
|
Replace Thread_Scheduler_control::control and
Thread_Scheduler_control::own_control with new
Thread_Scheduler_control::home.
Update #2556.
|
|
Update #2423.
|
|
Use the Thread_Control::resource_count for the no protocol mutexes.
Merge the no protocol and priority inherit CORE mutex seize/surrender
operations.
|
|
Update #2556.
Update #2784.
|
|
Add priority nodes which contribute to the overall thread priority.
The actual priority of a thread is now an aggregation of priority nodes.
The thread priority aggregation for the home scheduler instance of a
thread consists of at least one priority node, which is normally the
real priority of the thread. The locking protocols (e.g. priority
ceiling and priority inheritance), rate-monotonic period objects and the
POSIX sporadic server add, change and remove priority nodes.
A thread changes its priority now immediately, e.g. priority changes are
not deferred until the thread releases its last resource.
Replace the _Thread_Change_priority() function with
* _Thread_Priority_perform_actions(),
* _Thread_Priority_add(),
* _Thread_Priority_remove(),
* _Thread_Priority_change(), and
* _Thread_Priority_update().
Update #2412.
Update #2556.
|
|
Avoid direct access to thread internal data fields.
|
|
Introduce Thread_queue_Lock_context to contain the context necessary for
thread queue lock and thread wait lock acquire/release operations to
reduce the Thread_Control size.
|
|
|
|
Split up the potential thread priority change in the scheduler
release/cancel job operation. Protect the rate monotonic period state
with a dedicated SMP lock. This avoids a race condition during
_Rate_monotonic_Timeout() while _Rate_monotonic_Cancel() is called on
another processor.
|
|
Do not use a deadline value of zero to indicate a job cancellation. Use
a dedicated scheduler operation for this.
|
|
|
|
The _Thread_Lock_acquire() function had a potentially infinite run-time
due to the lack of fairness at atomic operations level.
Update #2412.
Update #2556.
Update #2765.
|
|
|
|
Clock disciplines may be WATCHDOG_RELATIVE, WATCHDOG_ABSOLUTE,
or WATCHDOG_NO_TIMEOUT. A discipline of WATCHDOG_RELATIVE with
a timeout of WATCHDOG_NO_TIMEOUT is equivalent to a discipline
of WATCHDOG_NO_TIMEOUT.
updates #2732
|
|
and maximal alignment.
There is need for unambiguous named and defined cache function
which should be called when code is updated, loaded
or is self-modifying.
There should be function to obtain maximal cache line length
as well. This function can and should be used for allocations
which can be used for data and or code and ensures that
there are no partial cache lines overlaps on start and
end of allocated region.
|
|
Task priorities are only valid within a scheduler instance. The
rtems_task_set_scheduler() directive moves a task from one scheduler
instance to another using the current priority of the thread. However,
the current task priority of the source scheduler instance is undefined
in the target scheduler instance. Add a third parameter to specify the
priority.
Close #2749.
|
|
The use of atomic fences is brittle and may break due to changes in
different areas which is hard to manage.
|
|
A 32-bit Priority_Control limits the uptime to 49 days with a 1ms clock
tick in case the EDF scheduler is used. Increase it to 64-bit to enable
proper operation of the EDF scheduler,
Close 2173.
|
|
Pass the deadline in watchdog ticks to the scheduler.
Update #2173.
|
|
Introduce map/unmap priority scheduler operations to map thread priority
values from/to the user domain to/from the scheduler domain. Use the
map priority operation to validate the thread priority. The EDF
schedulers use this new operation to distinguish between normal
priorities and priorities obtain through a job release.
Update #2173.
Update #2556.
|
|
Use same structure as POSIX API for thread priority conversion to/from
SuperCore.
|
|
This gets rid of the pull in of <stdarg.h> via <rtems.h> via
<rtems/rtems/ratemon.h> via <rtems/print.h>.
|
|
The RTEMS print user need to know nothing about a particular printer
implementation. In particular get rid of the <stdio.h> include which
would be visible via <rtems.h>.
|
|
Destroy the thread queue in case of a priority ceiling violation,
otherwise the SMP profiling data gets corrupted.
|
|
Remove superfluous includes. Use one attribute compare for each
semaphore variant. Text size drops by 10% on PowerPC due to this.
|
|
The Classic binary semaphores without a locking protocol
(RTEMS_BINARY_SEMAPHORE) could be released by everyone, e.g. in contrast
to the POSIX mutexes (all variants) or the Classic binary semphores with
priority inheritance or ceiling, there was no owner check in the release
path.
This behaviour was a bit unexpected and not documented. Add an owner
check to the release path. Update sptests/sp42 accordingly.
This change has nothing to do with the simple binary semaphores
(RTEMS_SIMPLE_BINARY_SEMAPHORE) which have no owner at all.
Update #2725
|
|
|
|
|
|
Move MrsP semaphore operations to a less prominent location. Fix field
name.
|
|
Provide dedicated seize and surrender methods for inherit priority
mutexes. This eliminates CORE_mutex_Attributes.
|
