7.4 Highest Locker Pattern
The Highest Locker Pattern defines a priority ceiling with each resource. The basic idea is that the task owning the resource runs at the highest-priority ceiling of all the resources that it currently owns, provided that it is blocking one or more higher-priority tasks. This limits priority inversion to at most one level.
The Highest Locker Pattern is another solution to the unbounded blocking/unbounded priority inversion problem. It is perhaps a minor elaboration from the Priority Inheritance Pattern, but it is different enough to have some different properties with respects to schedulability. The Highest Locker Pattern limits priority inversion to a single level as long as a task does not suspend itself while owning a resource. In this case, you may get chained blocking similar to the Priority Inheritance Pattern. Unlike the Priority Inheritance Pattern, however, you cannot get chained blocking if a task is preempted while owning a resource.
The unbounded priority inversion problem is discussed in the chapter introduction in some detail. The problem addressed by this pattern is to limit the maximum amount of priority inversion to a single levelthat is, there is at most a single lower-priority task blocking a higher-priority task from executing.
7.4.3 Pattern Structure
The Highest Locker Pattern is shown in Figure 7-10. The structural elements of the pattern are the same as for the Priority Inheritance Pattern, with the addition of an attribute priorityCeiling for the SharedResource.
Figure 7-10: Highest Locker Pattern
The pattern works by defining each lockable resource with a priority ceiling. The priority ceiling is just greater than the priority of the highest-priority client of the resourcethis is known at design time in a static priority scheme. When the resource is locked, the priority of the locking task is augmented to the priority ceiling of the resource.
7.4.4 Collaboration Roles
The Abstract Thread class is an abstract (noninstantiable) superclass for Concrete Thread. Abstract Thread associates with the Scheduler. Since Concrete Thread is a subclass, it has the same interface to the Scheduler as the Abstract Thread. This enforces interface compliance. The Abstract Thread is an «active» object, meaning that when it is created, it creates an OS thread in which to run. It contains (that is, it has composition relations with) more primitive application objects that execute in the thread of the composite «active» object.
The Concrete Thread is an «active» object most typically constructed to contain passive "semantic" objects (via the composition relation) that do the real work of the system. The Concrete Thread object provides a straightforward means of attaching these semantic objects into the concurrency architecture. Concrete Thread is an instantiable subclass of Abstract Thread.
The Mutex is a mutual exclusion semaphore object that permits only a single caller through at a time. The operations of the Shared Resource invoke it whenever a relevant service is called, locking it prior to starting the service and unlocking it once the service is complete. Threads that attempt to invoke a service when the services are already locked become blocked until the Mutex is in its unlocked state. This is done by the Mutex semaphore signaling the Scheduler that a call attempt was made by the currently active thread, the Mutex ID (necessary to unlock it later when the mutex is released), and the entry pointthe place at which to continue execution of the Thread.
This object orchestrates the execution of multiple threads based on their priority according to a simple rule: Always run the ready thread with the highest priority. When the «active» Thread object is created, it (or its creator) calls the createThread operation to create a thread for the «active» object. Whenever this thread is executed by the Scheduler, it calls the StartAddr address (except when the thread has been blocked or preemptedin which case it calls the EntryPoint address).
In this pattern, the Scheduler has some special duties when the Mutex signals an attempt to access a locked resource. Specifically, it must block the requesting task (done by stopping that task and placing a reference to it in the Blocked Queue (not shownfor details of the Blocked Queue, see the Static Priority Pattern in Chapter 5), and it must elevate the priority of the task owning the resource to the Shared Resource's priorityCeiling.
A resource is an object shared by one or more Threads. For the system to operate properly in all cases, all Shared Resources must either be reentrant (meaning that corruption from simultaneous access cannot occur), or they must be protected. In the case of a protected resource, when a Thread attempts to use the resource, the associated Mutex semaphore is checked, and if locked, the calling task is placed into the Blocked Queue. The task is terminated with its reentry point noted in the TCB.
The SharedResource has a constant attribute (note the «frozen» constraint in Figure 7-10), called priorityCeiling. This is set during design to just greater than the priority of the highest-priority task that can ever access it. In some RTOSs, this means that the priority will be one more (when a larger number indicates a higher priority), and in some it will be one less (when a lower number indicates a higher priority). This ensures that when the resource is locked, no other task using that resource can preempt it.
Task Control Block
The TCB contains the scheduling information for its corresponding Thread object. This includes the priority of the thread, the default start address, and the current entry address if it was preempted or blocked prior to completion. The Scheduler maintains a TCB object for each existing Thread. Note that TCB typically also has a reference off to a call and parameter stack for its Thread, but that level of detail is not shown in Figure 7-10. The TCB tracks both the current priority of the thread (which may have been elevated due to resource access and blocking) and its nominal priority.
The Highest Locker Pattern has even better priority inversion-bounding properties than the Priority Inheritance Pattern. It allows higher-priority tasks to run, but only if they have a priority higher than the priority ceiling of the resource. The priority ceiling can be determined at design time for each resource by examining the clients of a given resource and identifying to which active object they belong and selecting the highest from among those. The priority ceiling is this value augmented by one. Computation of worst-case blocking is the length of the longest critical section (that is, resource locking time) of any task of lesser priority as long as a task never suspends itself while owning a resource.
The pattern has the disadvantage that while it bounds priority inversion to a single level, that level happens more frequently than with some other approaches. For example, if the lowest-priority task locks a resource with the highest-priority ceiling, and during that time an intermediate priority task becomes ready to run, then it is blocked even though in this case one would prefer that the normal priority rules apply. One way to handle that is to elevate the priority of the task owning the resource only when another task attempts to lock it; until then, the locking tasks runs at its nominal priority.
In this pattern, care must be taken to ensure that a task never suspends itself while owning a resource. It is fine if it is preempted, but voluntary preemption while owning a resource can lead to chain blocking, a problem previously identified with the Priority Inheritance Pattern in the previous section. If the system allows tasks to suspend themselves while owning a resource, then the computation of worst-case blocking is computed in the same way as with the Priority Inheritance Patternthe longest case of chain blocked must be traversed.
This pattern avoids deadlock as long as no task suspends itself while owning a resource because no other task is permitted to wait on the resource (condition 4). This is because the locking task runs at a priority higher than any of the other clients of the resource. As previously noted, there is also a consequence of computational overhead associated with the Highest Locker Pattern.
7.4.6 Implementation Strategies
Fewer RTOSs support the Highest Locker Pattern more than the basic Priority Inheritance Pattern. Implementation of this pattern in your own RTOS is fairly straightforward, with the addition of priority ceiling attributes in the Shared Resource. When the mutex is locked, it must notify the Scheduler to elevate the priority of the locking task to that resource's priority ceiling.
7.4.7 Related Patterns
The Highest Locker Pattern exists to help solve a particular problem peculiar to priority-based preemption multitasking, so all of the concurrency patterns having to do with that style of multitasking can be mixed with this pattern.
7.4.8 Sample Model
In the example shown in Figure 7-11, there are four tasks with their priorities shown using constraints, two of which, Waveform Draw and Message Display, share a common resource, Display. The tasks, represented as active objects in order of their priority, are Message Display (priority Low), Switch Monitor (priority Medium Low), Waveform Draw (priority Medium High), and Safety Monitor (priority Very High), leaving priority High unused at the outset. Message Display and Waveform Draw share Display, so the priority ceiling of Display is just above Waveform Draw (that is, High).
The scenario runs as follows: First, the lowest-priority task, Message Display, runs, calling the operation Display.displayMsg(). Because the Display has a mutex semaphore, this locks the resource, and the Scheduler (not shown in Figure 7-11) escalates the priority of the locking task, Message Display, to the priority ceiling of the resourcethat is, the value High.
Figure 7-11: Highest Locker Pattern
While this operation executes, first the Switch Monitor and then the Waveform Draw tasks both become ready to run but cannot because the Message Display task is running at a higher priority than either of them. The Safety Monitor task becomes ready to run. Because it runs at a priority Very High, it can, and does, preempt the Message Display task.
After the Safety Monitor task returns control to the Scheduler, the Scheduler continues the execution of the Message Display task. Once it releases the resource, the mutex signals the Scheduler, and the latter deescalates the priority of the Message Display task to its nominal priority level of Low. At this point, there are two tasks of a higher priority waiting to run, so the higher-priority waiting task (Waveform Draw) runs, and when it completes, the remaining higher-priority task (Switch Monitor) runs. When this last task completes, the Message Display task can finally resume its work and complete.