1 / 14

Linux Scheduler

Linux Scheduler. Objectives. Response time Throughput for batch No starvation Accommodate high AND low priority. General Process Classifications. Two schemes for classifying: Traditional classes IO bound CPU bound Alternative classes Interactive Batch Real-time

hayden
Télécharger la présentation

Linux Scheduler

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Linux Scheduler

  2. Objectives • Response time • Throughput for batch • No starvation • Accommodate high AND low priority

  3. General Process Classifications • Two schemes for classifying: • Traditional classes • IO bound • CPU bound • Alternative classes • Interactive • Batch • Real-time • Video, robotics, sensor-based How to determine?

  4. Linux Scheduling Classes • FIFO FIFO real-time process • RR Round Robin real-time process • Normal non-real-time process

  5. Real-timeDetails

  6. Real-time Rules • Priority 0(highest) - 99(lowest) • Program adjustable: • sched_setparam() • sched_setscheduler() • FIFO - non-interruptible except • Higher priority • Blocked • Yields • RR • Adds a time-slice to each thread • Suspends at end of slice

  7. Non-real-time Scheduling(Conventional processes)

  8. Rules • Every process has a “static” priority • 100(highest) – 139 (lowest) • Variable via nice() and setpriority() • Base quantum (not a constant!) • If sp<120 • 20*(140-static priority) • If sp>=120 • 5*(140-static priority) • Dynamic priority (also 100-139) • Max(100, min(sp-bonus+5,139)) • 0<=bonus<=10

  9. Rules (continued) • Bonus • 0-5 is a penalty • 6-10 is a prize • Related to average sleep time • OF THIS PROCESS (not of all Processes) • Based on past history • Some sample values:

  10. Priority Arrays • Two arrays: struct prio_array { int nr_active; unsigned long bitmap[BITMAP_SIZE]; struct list_head queue[MAX_PRIO]; } • With: 140 priority levels (max) wordsize=32 bits • BITMAP_SIZE is 5 (last 20 bits ignored)

  11. The Bitmaps • Active queues • Time-slice-expired queues • 1 bit/queue (140 queues) • Based on the Intel “bsfl” instruction • Select least significant 1-bit

  12. Priority arrays – revisited • Two arrays (active/expired): *listheads prio_array[2][140]; typedef struct listheads { // 2-way list of PDs @ this priority PD *next; PD *prev;} struct PD { PD *next; PD*prev; int PID; int prio; // etc }

  13. Scheduling • Pick highest priority non-empty queue/list • Tasks in queue/list scheduled

  14. References • Lindsey, R., "What's New in the 2.6 Scheduler", Linux Journal, March 2004 • Love, R., Linux Kernel Development, Indianapolis, IN, Sams Publishing, 2004 • Understanding the Linux Kernel, Bovet & Cesati, O’Reilly Press, 2006

More Related