An implementation of the earliest deadline first algorithm in Linux

Authors:
Dario Faggioli;Michael Trimarchi;Fabio Checconi;Marko Bertogna;Antonio Mancina
Affiliations:
ReTiS Lab, Scuola Superiore Sant'Anna, Pisa, Italy;ReTiS Lab, Scuola Superiore Sant'Anna, Pisa, Italy;ReTiS Lab, Scuola Superiore Sant'Anna, Pisa, Italy;Scuola Superiore Sant Anna, Pisa - Italy;Scuola Superiore Sant Anna, Pisa - Italy
Venue:
Proceedings of the 2009 ACM symposium on Applied Computing
Year:
2009

Citing 4
Cited 4

Stack-based scheduling for realtime processes

Real-Time Systems
Scheduling Algorithms for Multiprogramming in a Hard-Real-Time Environment

Journal of the ACM (JACM)
Priority Inheritance Protocols: An Approach to Real-Time Synchronization

IEEE Transactions on Computers
The Limited-Preemption Uniprocessor Scheduling of Sporadic Task Systems

ECRTS '05 Proceedings of the 17th Euromicro Conference on Real-Time Systems

Sporadic Server revisited

Proceedings of the 2010 ACM Symposium on Applied Computing
CPU scheduling and memory management for interactive real-time applications

Real-Time Systems
FPSL, FPCL and FPZL schedulability analysis

Real-Time Systems
Towards a user-mode approach to partitioned scheduling in the seL4 microkernel

ACM SIGBED Review

Quantified Score

Hi-index	0.00

Visualization

Abstract

Recently, many projects have been started to introduce some real-time mechanisms into general purpose operating systems (GPOS) in order to make them capable of providing the users with some temporal guarantees. Many of these projects focused especially on Linux for its capillary and widespread adoption throughout many different research and industrial environments. By tracking the kernel release cycle, we propose an efficient Earliest Deadline First implementation in the form of a patch-set against the 2.6.27 version, that is the latest released one, as of now. Our implementation provides the user with the possibility to choose SCHED_EDF as one of the possible scheduling policies for a task, with an enhanced version of the standard algorithm. In fact, we propose a new approach to shared resources' access which, differently from many other previous existing works, does not require the user to specify any parameters about the critical sections every task will enter during its execution.