If multiple tasks are trying to write to the same queue, do I need to do anything differently to protect these queue writes from context switches? For example, say I have an application with multiple tasks, each gathering information from different sources. One additional task is responsible for reporting this information through a communications pipe. Can I create one queue that each of the producer tasks writes to and the consumer task reads from, or do I need to have multiple queues, one for each producer task? If I also want to have a status variable that is accessed by each of these tasks, can I merely create a global variable and surround any reads and writes to this variable with: portENTER_CRITICAL(); // write to or read from global variable portEXIT_CRITICAL(); Thanks
You can have multiple readers and multiple writers to a single queue with no problem. Tasks will block on queue reads and writes in their priority order. The kernel takes care of the multiple access issues for you. Regarding the status variable. You can use the critical sections as you note in your post. This is only necessary if there are multiple writers to the variable, or if the data type of the variable is larger than the natural type of the machine (meaning accesses to the variable are not atomic). A task that just reads a variable of the natural type or smaller does not need the critical section Regards.
Thanks. That makes life easy!
Copyright (C) Amazon Web Services, Inc. or its affiliates. All rights reserved.