问题
In operating system the scheduler is invoked after the system call api or after a hardware interrupt processing.
I am unable to get and found how and who calls the scheduler???
回答1:
The scheduler is invoked:
- With explicit blocking: mutex, semaphore, waitqueue, etc.
- If
TIF_NEED_RESCHED
flag is set, on the nearest possible occasion:- If the kernel is preemptible:
- in syscall or exception context, at the next outmost
preempt_enable()
- in IRQ context, return from interrupt-handler to preemptible context
- in syscall or exception context, at the next outmost
- If the kernel is not preemptible:
cond_resched()
call- explicit
schedule()
call - return from syscall or exception to user-space
- return from interrupt-handler to user-space
- If the kernel is preemptible:
回答2:
The scheduler will be invoked if the current thread/process is going to sleep/wait for some event/resource to be released.
In one of the cases of worker threads which executes the bottom half in the form of workqueues, it will run in a while loop and check if the workqueue list is empty. If found empty it will mark itself as TASK_INTERRUPTABLE
, calls schedule()
and then goes to sleep.
If the workqueque list is not empty the worker thread marks itself RUNNING
and executes the deferred bottom halfs.
So in general schedule()
is called by a task which wants to sleep and thus hands over the cpu to other processes/tasks.
回答3:
I think in case of round robin algorithm followed for scheduling some process some timer is being programmed based on time slice value which is going to invoke scheduler after that time slice gets over using interrupt functionality. and that preempt the current executing thread save its context and process ready queue figure out which thread to run next restore its context from thread stack and set PC value based on new thread's previous preempt value now new thread starts running. https://en.wikipedia.org/wiki/Programmable_interval_timer
来源:https://stackoverflow.com/questions/17522658/how-scheduler-is-called-in-linux