INtime® 3.1 Software - tenAsys

Chapter 5: Designing RT applications
Interrupt and event processing
Multi-tasking
Interrupts and internal events occur at the bedside units: bedside personnel enter
commands asynchronously and the system computes toxicity levels at regular
intervals.
Toxicity levels, measured as the blood enters the bedside unit, are not subject to abrupt
change. The machine slowly removes toxins while the patient's body, more slowly,
puts toxins back in. The result is a steadily declining toxicity level. The bedside units
must monitor toxicity levels regularly, but not too frequently. For instance, the bedside
units could compute the toxicity levels once every 10 seconds, using a clock for timing.
The measurement thread would measure and compute the toxicity, put the information
in a mailbox for the MCU, and suspend itself for 10 seconds.
Command interrupts from the bedside unit occur when a medical operator types a
command and presses Enter. Interrupts from command entries occur at random times.
The interrupt handler signals the interrupt thread. The interrupt thread performs any
required processing and waits for the next interrupt.
Processing commands from the bedside units: Each time a medical operator types a
command and presses Enter, the bedside unit receives an interrupt signal from the
terminal. The bedside unit stops executing the current instruction and begins to
execute an interrupt handler.
1. The interrupt handler accumulates the characters in a buffer and puts them in
memory. The interrupt handler signals the interrupt thread for bedside commands.
2. The interrupt thread gets the contents of the memory where the handler put the
command. It parses the command and does the required processing.
3. The thread puts the command information, along with the number of the bedside
unit, into a message.
4. The thread sends the message to the predetermined mailbox for the MCU.
5. The interrupt thread waits for the next interrupt. The system returns to its normal
priority-based, preemptive scheduling.
Threads in the application run using preemptive, priority-based scheduling. This
allows the more important threads, such as those that control the dialysis rate, to
preempt lower-priority threads, such as those that update displays. New capabilities
can be added to the system by simply adding new threads.
59