In the HCS12 port there is a variable named uxCriticalNesting. I do not quite understand why this is necessary, and i did not find it in the AVR port. I would like to hear the reason for why is it necessary.
A task can, if it wishes (and care is taken) yield while it is within a critical section. Therefore each task needs to maintain its own critical nesting state. If a task yields from within a critical section interrupts must again be disabled when the task starts to run again.
The simplest way of implementing a critical section is to push the register that contains the interrupt mask flags onto the stack – then disable interrupts. The interrupt enable flags are then set back to their original state by simply popping the register from the stack again.
However, not all compilers allow you to modify the stack. If the compiler uses stack pointer relative addressing (and does not use a separate frame pointer) then changing the stack value will crash the program. In this case another method is required. A simple method is for each task to maintain a count of the critical nesting depth. When a critical section is entered, interrupts are disable and uxCriticalNesting is incremented. When a critical section is exited uxCriticalNesting is decremented, and if then it is zero the nesting has unwound and interrupts enabled. The critical nesting counter is saved as part of the task context.
The AVR port uses the stack method. The HCS12 port uses the uxCriticalNesting method.