ephobb wrote on Sunday, January 24, 2016:
Hello,
i want to know the difference between xqueuesend and xqueuesendtobackfromisr.
I know we use api xqueuesend from normal flow and we use api xqueuesendtobackfromisr, when we are sending data to queue from ISR routine.
But i want to know, how they are differ?
richard_damon wrote on Sunday, January 24, 2016:
First, xqueuesend and xquesendtoback are aliases for the same function. So the difference you are looking at is the difference between a non-fromisr function and a fromisr function. The major difference between them is that the non-ISR version give you the option to block the task for awhile for the operation to complete if needed, while the ISR versions do not (you DON’T want to delay an ISR execution) but instead give you a flag to let you decide if you need to run the scheduler.
A second internal difference is the way critical sections are implemented. The non-ISR versions use slightly simpler code knowing they aren’t in an interrupt context.
A further question to this:
So it means that a non-FromISR function will in general also work inside an isr and will not cause a crash or haung-up? Right? Given the waiting time is zero or not used because there is enough space in the queue.
NO! It is a violation of the API to use a “non-FromISR” inside an isr, as they are built on the assumption that they are being used at a task level, and (in general) the critical sections they are using assumes no possible interrupt nesting might be happening.
On SOME ports, it is possible to use a FromISR function at the task level, generally requirements for this are that the port supports nested interrupts (or some of the FromISR routines will effectively not create needed critical sections, as being in an interrupt is itself sufficient protection).