7.2. Background

7.2.1. Task Definition

Many definitions of a task have been proposed in computer literature. Unfortunately, none of these definitions encompasses all facets of the concept in a manner which is operating system independent. Several of the more common definitions are provided to enable each user to select a definition which best matches their own experience and understanding of the task concept:

  • a “dispatchable” unit.

  • an entity to which the processor is allocated.

  • an atomic unit of a real-time, multiprocessor system.

  • single threads of execution which concurrently compete for resources.

  • a sequence of closely related computations which can execute concurrently with other computational sequences.

From RTEMS’ perspective, a task is the smallest thread of execution which can compete on its own for system resources. A task is manifested by the existence of a task control block (TCB).

7.2.2. Task Control Block

The Task Control Block (TCB) is an RTEMS defined data structure which contains all the information that is pertinent to the execution of a task. During system initialization, RTEMS reserves a TCB for each task configured. A TCB is allocated upon creation of the task and is returned to the TCB free list upon deletion of the task.

The TCB’s elements are modified as a result of system calls made by the application in response to external and internal stimuli. TCBs are the only RTEMS internal data structure that can be accessed by an application via user extension routines. The TCB contains a task’s name, ID, current priority, current and starting states, execution mode, TCB user extension pointer, scheduling control structures, as well as data required by a blocked task.

A task’s context is stored in the TCB when a task switch occurs. When the task regains control of the processor, its context is restored from the TCB. When a task is restarted, the initial state of the task is restored from the starting context area in the task’s TCB.

7.2.3. Task Memory

The system uses two separate memory areas to manage a task. One memory area is the Task Control Block. The other memory area is allocated from the stack space or provided by the user and contains

  • the task stack,

  • the thread-local storage (TLS), and

  • an optional architecture-specific floating-point context.

The size of the thread-local storage is determined at link time. A user-provided task stack must take the size of the thread-local storage into account.

On architectures with a dedicated floating-point context, the application configuration assumes that every task is a floating-point task, but whether or not a task is actually floating-point is determined at runtime during task creation (see Floating Point Considerations). In highly memory constrained systems this potential overestimate of the task stack space can be mitigated through the CONFIGURE_MINIMUM_TASK_STACK_SIZE configuration option and aligned task stack sizes for the tasks. A user-provided task stack must take the potential floating-point context into account.

7.2.4. Task Name

By default, the task name is defined by the task object name given to rtems_task_create(). The task name can be obtained with the pthread_getname_np() function. Optionally, a new task name may be set with the pthread_setname_np() function. The maximum size of a task name is defined by the application configuration option CONFIGURE_MAXIMUM_THREAD_NAME_SIZE.

7.2.5. Task States

A task may exist in one of the following five states:

  • executing - Currently scheduled to the CPU

  • ready - May be scheduled to the CPU

  • blocked - Unable to be scheduled to the CPU

  • dormant - Created task that is not started

  • non-existent - Uncreated or deleted task

An active task may occupy the executing, ready, blocked or dormant state, otherwise the task is considered non-existent. One or more tasks may be active in the system simultaneously. Multiple tasks communicate, synchronize, and compete for system resources with each other via system calls. The multiple tasks appear to execute in parallel, but actually each is dispatched to the CPU for periods of time determined by the RTEMS scheduling algorithm. The scheduling of a task is based on its current state and priority.

7.2.6. Task Priority

A task’s priority determines its importance in relation to the other tasks executing on the processor set owned by a scheduler. Normally, RTEMS supports 256 levels of priority ranging from 0 to 255. The priority level 0 represents a special priority reserved for the operating system. The data type rtems_task_priority is used to store task priorities. The maximum priority level depends on the configured scheduler, see CONFIGURE_MAXIMUM_PRIORITY, Clustered Scheduler Configuration, and Scheduling Concepts.

Tasks of numerically smaller priority values are more important tasks than tasks of numerically larger priority values. For example, a task at priority level 5 is of higher privilege than a task at priority level 10. There is no limit to the number of tasks assigned to the same priority.

Each task has a priority associated with it at all times. The initial value of this priority is assigned at task creation time. The priority of a task may be changed at any subsequent time.

Priorities are used by the scheduler to determine which ready task will be allowed to execute. In general, the higher the logical priority of a task, the more likely it is to receive processor execution time.

7.2.7. Task Mode

A task’s execution mode is a combination of the following four components:

  • preemption

  • ASR processing

  • timeslicing

  • interrupt level

It is used to modify RTEMS’ scheduling process and to alter the execution environment of the task. The data type rtems_task_mode is used to manage the task execution mode.

The preemption component allows a task to determine when control of the processor is relinquished. If preemption is disabled (RTEMS_NO_PREEMPT), the task will retain control of the processor as long as it is in the executing state - even if a higher priority task is made ready. If preemption is enabled (RTEMS_PREEMPT) and a higher priority task is made ready, then the processor will be taken away from the current task immediately and given to the higher priority task.

The timeslicing component is used by the RTEMS scheduler to determine how the processor is allocated to tasks of equal priority. If timeslicing is enabled (RTEMS_TIMESLICE), then RTEMS will limit the amount of time the task can execute before the processor is allocated to another ready task of equal priority. The length of the timeslice is application dependent and specified in the Configuration Table. If timeslicing is disabled (RTEMS_NO_TIMESLICE), then the task will be allowed to execute until a task of higher priority is made ready. If RTEMS_NO_PREEMPT is selected, then the timeslicing component is ignored by the scheduler.

The asynchronous signal processing component is used to determine when received signals are to be processed by the task. If signal processing is enabled (RTEMS_ASR), then signals sent to the task will be processed the next time the task executes. If signal processing is disabled (RTEMS_NO_ASR), then all signals received by the task will remain posted until signal processing is enabled. This component affects only tasks which have established a routine to process asynchronous signals.

The interrupt level component is used to determine which interrupts will be enabled when the task is executing. RTEMS_INTERRUPT_LEVEL(n) specifies that the task will execute at interrupt level n.

RTEMS_PREEMPT

enable preemption (default)

RTEMS_NO_PREEMPT

disable preemption

RTEMS_NO_TIMESLICE

disable timeslicing (default)

RTEMS_TIMESLICE

enable timeslicing

RTEMS_ASR

enable ASR processing (default)

RTEMS_NO_ASR

disable ASR processing

RTEMS_INTERRUPT_LEVEL(0)

enable all interrupts (default)

RTEMS_INTERRUPT_LEVEL(n)

execute at interrupt level n

The set of default modes may be selected by specifying the RTEMS_DEFAULT_MODES constant.

7.2.8. Task Life States

Independent of the task state with respect to the scheduler, the task life is determined by several orthogonal states:

  • protected or unprotected

  • deferred life changes or no deferred life changes

  • restarting or not restarting

  • terminating or not terminating

  • detached or not detached

While the task life is protected, asynchronous task restart and termination requests are blocked. A task may still restart or terminate itself. All tasks are created with an unprotected task life. The task life protection is used by the system to prevent system resources being affected by asynchronous task restart and termination requests. The task life protection can be enabled (PTHREAD_CANCEL_DISABLE) or disabled (PTHREAD_CANCEL_ENABLE) for the calling task through the pthread_setcancelstate() directive.

While deferred life changes are enabled, asynchronous task restart and termination requests are delayed until the task performs a life change itself or calls pthread_testcancel(). Cancellation points are not implemented in RTEMS. Deferred task life changes can be enabled (PTHREAD_CANCEL_DEFERRED) or disabled (PTHREAD_CANCEL_ASYNCHRONOUS) for the calling task through the pthread_setcanceltype() directive. Classic API tasks are created with deferred life changes disabled. POSIX threads are created with deferred life changes enabled.

A task is made restarting by issuing a task restart request through the rtems_task_restart() directive.

A task is made terminating by issuing a task termination request through the rtems_task_exit(), rtems_task_delete(), pthread_exit(), and pthread_cancel() directives.

When a detached task terminates, the termination procedure completes without the need for another task to join with the terminated task. Classic API tasks are created as not detached. The detached state of created POSIX threads is determined by the thread attributes. They are created as not detached by default. The calling task is made detached through the pthread_detach() directive. The rtems_task_exit() directive and self deletion though rtems_task_delete() directive make the calling task detached. In contrast, the pthread_exit() directive does not change the detached state of the calling task.

7.2.9. Accessing Task Arguments

All RTEMS tasks are invoked with a single argument which is specified when they are started or restarted. The argument is commonly used to communicate startup information to the task. The simplest manner in which to define a task which accesses it argument is:

rtems_task user_task(
    rtems_task_argument argument
);

Application tasks requiring more information may view this single argument as an index into an array of parameter blocks.

7.2.10. Floating Point Considerations

Please consult the RTEMS CPU Architecture Supplement if this section is relevant on your architecture. On some architectures the floating-point context is contained in the normal task context and this section does not apply.

Creating a task with the RTEMS_FLOATING_POINT attribute flag results in additional memory being allocated for the task to store the state of the numeric coprocessor during task switches. This additional memory is not allocated for RTEMS_NO_FLOATING_POINT tasks. Saving and restoring the context of a RTEMS_FLOATING_POINT task takes longer than that of a RTEMS_NO_FLOATING_POINT task because of the relatively large amount of time required for the numeric coprocessor to save or restore its computational state.

Since RTEMS was designed specifically for embedded military applications which are floating point intensive, the executive is optimized to avoid unnecessarily saving and restoring the state of the numeric coprocessor. In uniprocessor configurations, the state of the numeric coprocessor is only saved when a RTEMS_FLOATING_POINT task is dispatched and that task was not the last task to utilize the coprocessor. In a uniprocessor system with only one RTEMS_FLOATING_POINT task, the state of the numeric coprocessor will never be saved or restored.

Although the overhead imposed by RTEMS_FLOATING_POINT tasks is minimal, some applications may wish to completely avoid the overhead associated with RTEMS_FLOATING_POINT tasks and still utilize a numeric coprocessor. By preventing a task from being preempted while performing a sequence of floating point operations, a RTEMS_NO_FLOATING_POINT task can utilize the numeric coprocessor without incurring the overhead of a RTEMS_FLOATING_POINT context switch. This approach also avoids the allocation of a floating point context area. However, if this approach is taken by the application designer, no tasks should be created as RTEMS_FLOATING_POINT tasks. Otherwise, the floating point context will not be correctly maintained because RTEMS assumes that the state of the numeric coprocessor will not be altered by RTEMS_NO_FLOATING_POINT tasks. Some architectures with a dedicated floating-point context raise a processor exception if a task with RTEMS_NO_FLOATING_POINT issues a floating-point instruction, so this approach may not work at all.

If the supported processor type does not have hardware floating capabilities or a standard numeric coprocessor, RTEMS will not provide built-in support for hardware floating point on that processor. In this case, all tasks are considered RTEMS_NO_FLOATING_POINT whether created as RTEMS_FLOATING_POINT or RTEMS_NO_FLOATING_POINT tasks. A floating point emulation software library must be utilized for floating point operations.

On some processors, it is possible to disable the floating point unit dynamically. If this capability is supported by the target processor, then RTEMS will utilize this capability to enable the floating point unit only for tasks which are created with the RTEMS_FLOATING_POINT attribute. The consequence of a RTEMS_NO_FLOATING_POINT task attempting to access the floating point unit is CPU dependent but will generally result in an exception condition.

7.2.11. Building a Task Attribute Set

In general, an attribute set is built by a bitwise OR of the desired components. The set of valid task attribute components is listed below:

RTEMS_NO_FLOATING_POINT

does not use coprocessor (default)

RTEMS_FLOATING_POINT

uses numeric coprocessor

RTEMS_LOCAL

local task (default)

RTEMS_GLOBAL

global task

Attribute values are specifically designed to be mutually exclusive, therefore bitwise OR and addition operations are equivalent as long as each attribute appears exactly once in the component list. A component listed as a default is not required to appear in the component list, although it is a good programming practice to specify default components. If all defaults are desired, then RTEMS_DEFAULT_ATTRIBUTES should be used.

This example demonstrates the attribute_set parameter needed to create a local task which utilizes the numeric coprocessor. The attribute_set parameter could be RTEMS_FLOATING_POINT or RTEMS_LOCAL | RTEMS_FLOATING_POINT. The attribute_set parameter can be set to RTEMS_FLOATING_POINT because RTEMS_LOCAL is the default for all created tasks. If the task were global and used the numeric coprocessor, then the attribute_set parameter would be RTEMS_GLOBAL | RTEMS_FLOATING_POINT.

7.2.12. Building a Mode and Mask

In general, a mode and its corresponding mask is built by a bitwise OR of the desired components. The set of valid mode constants and each mode’s corresponding mask constant is listed below:

RTEMS_PREEMPT

is masked by RTEMS_PREEMPT_MASK and enables preemption

RTEMS_NO_PREEMPT

is masked by RTEMS_PREEMPT_MASK and disables preemption

RTEMS_NO_TIMESLICE

is masked by RTEMS_TIMESLICE_MASK and disables timeslicing

RTEMS_TIMESLICE

is masked by RTEMS_TIMESLICE_MASK and enables timeslicing

RTEMS_ASR

is masked by RTEMS_ASR_MASK and enables ASR processing

RTEMS_NO_ASR

is masked by RTEMS_ASR_MASK and disables ASR processing

RTEMS_INTERRUPT_LEVEL(0)

is masked by RTEMS_INTERRUPT_MASK and enables all interrupts

RTEMS_INTERRUPT_LEVEL(n)

is masked by RTEMS_INTERRUPT_MASK and sets interrupts level n

Mode values are specifically designed to be mutually exclusive, therefore bitwise OR and addition operations are equivalent as long as each mode appears exactly once in the component list. A mode component listed as a default is not required to appear in the mode component list, although it is a good programming practice to specify default components. If all defaults are desired, the mode RTEMS_DEFAULT_MODES and the mask RTEMS_ALL_MODE_MASKS should be used.

The following example demonstrates the mode and mask parameters used with the rtems_task_mode directive to place a task at interrupt level 3 and make it non-preemptible. The mode should be set to RTEMS_INTERRUPT_LEVEL(3) | RTEMS_NO_PREEMPT to indicate the desired preemption mode and interrupt level, while the mask parameter should be set to RTEMS_INTERRUPT_MASK | RTEMS_NO_PREEMPT_MASK to indicate that the calling task’s interrupt level and preemption mode are being altered.