Operating Systems scheduling project

The Project needs to be done in C language in LINUX. Check the attachment if you are interested. Full work is required. It is about thread scheduling, it may come long but it is to make the Project clear instead of much harder. Clarifications: Whenever you want to update the min_vruntime (rq.min_vruntime below) of the runqueue, update it as follows so that it is monotonically increasing: Let p be the running process and [login to view URL] is the runqueue rbtree (or linked list) if len([login to view URL]) > 0: // if there is a ready process firstnode = [login to view URL][0] // firstnode is the one that has smallest vrtuntime in the tree vruntime = min([login to view URL], [login to view URL]) else: vruntime = [login to view URL] rq.min_vruntime = max(rq.min_vruntime, vruntime) To keep the set of ready processes, you can use any data structure you like: red-black tree, AVL tree, linked list, etc. When we say rbtree below or in any place in the assignment, that means any one of the structure you like. Keep time values in nanoseconds in your program. That means use ns unit in your code. In the workload file, the time unit is ms. Hence we you read those values in your program, convert them into ns in your program. It is better if your program is working with ns as the time unit inside. This will simplify a lot of thing for you. Timeslice value in your program will be an integer and the unit will be ns (nanosecond). Consider the runqueue (structure) to consist of the following fields (at least): 1) red black tree (rbtree) containing processes that are in ready state (ordered wrt to vruntime), 2) currently running process (it is not part of the rbtree, but it is in the runqueue data structure), 3) min_vruntime (a monotonically increasing value; never decreases), 4) number of processes in the runqueue (includes the running process), 5) load (weight) of the runqueue, which is the total weight of the processes in the runqueue (includes the running process and the processes in the rbtree). You don't need to write out the cpu utilization. Don't need to compute it. Timeslice value will be computed for the running process at each timer tick as well. Hence it is only computed for the running process. Whan the running process has run in the cpu for timeslice time, it is suspended. This means that the timeslice values of processes are computed dynamically. When process goes into cpu, at that time and at each timer tick, the timeslice of the process is computed using the load of the process and the load of the runqueue. When process is out of cpu, its timeslice needs not be computing till it is executed again. Timeslice for a process becomes relevant when a process starts executing in the cpu. At each timer tick, the vruntime of the running process is updated. The min_vruntime of the runqueue is also updated. It is also checked if the current running process expired its allowed timeslice or finished its cpu burst. You need to update vruntim of the running process and also the min_vruntime of the runqueue also when the cpu burst of the running process finishes, or when another job needs to be scheduled, before the running job is removed from cpu.