Inconsistent enabling/disabling of interrupts

[mavaneerden]

Sometimes, interrupts are disabled and enabled inconsistently. For example, here interrupts are disabled when checking the next thread:

InK/Kernel/scheduler/scheduler.c

Lines 188 to 193 in 2a95405

	__disable_interrupt();
	// check the ready queue for the last time
	if(!__next_thread()){
	__mcu_sleep();
	__enable_interrupt();
	}

(also in this example, if there IS a next thread, interrupts are not re-enabled)

But then, interrupts are NOT disabled when checking the next thread here:

InK/Kernel/scheduler/scheduler.c

Lines 173 to 177 in 2a95405

	case SCHED_SELECT:
	// the scheduler selects the highest priority task right
	// after it has finished the execution of a single task
	_thread = __next_thread();
	_sched_state = SCHED_BUSY;

I don't really understand why interrupts are not disabled in the second case. In both cases, an ISR could interrupt the priority calculation at the wrong time by signalling a thread, leading to inconsistent readyLevel and readyTable values.