The added overhead for context-switching and SVC handling pays off? With all threads sharing the same address space, user threads can overrun kernel stacks if the code says so.Could anyone with more expertise elaborate on that?
Thank you very much.
Generally speaking, defining a privileged stack for the RTOS, and a separate (unprivileged) stack for the apps makes sense.
However it is not a requirement from the processor point of view (though some RTOS implementations may enforce it).