Notes

These policies are more complex to understand and implement than it seems! E.g.:

“I observed in the kernel code (to my disgust), the Linux PIP implementation is a nightmare: extremely heavy weight, involving maintenance of a full wait-for graph, and requiring updates for a range of events, including priority changes and interruptions of wait operations.“
“An additional reason is that the original specification of PIP [24], despite being informally “proved” correct, is actually flawed."
Quoted from ”Zhang, X., Urban, C. & Wu, C. Priority Inheritance Protocol Proved Correct. J Autom Reasoning 64, 73-95 (2020).

Equations in Buttazzo’s book

In the equations for computing the blocking times, Buttazzo’s book has a ”-1“ factor affecting the critical region duration.
This is due to the fact that he considers a clock resolution of 1 t.u.
This means that if a t = tk a resource is busy then it was taken at least at t = tk − 1, so the task holding it already executed at least one time unit.
In these slides I’m considering that the clock resolution is much smaller than the duration of tasks and critical sections, so that factor is neglected.
Other than that the equations are similar.

PreviousStack Resource Policy NextJoint scheduling of periodic and aperiodic tasks

Last updated 2 years ago

Notes

These policies are more complex to understand and implement than it seems! E.g.:

“I observed in the kernel code (to my disgust), the Linux PIP implementation is a nightmare: extremely heavy weight, involving maintenance of a full wait-for graph, and requiring updates for a range of events, including priority changes and interruptions of wait operations.“
“An additional reason is that the original specification of PIP [24], despite being informally “proved” correct, is actually flawed."
Quoted from ”Zhang, X., Urban, C. & Wu, C. Priority Inheritance Protocol Proved Correct. J Autom Reasoning 64, 73-95 (2020).

Equations in Buttazzo’s book

In the equations for computing the blocking times, Buttazzo’s book has a ”-1“ factor affecting the critical region duration.
This is due to the fact that he considers a clock resolution of 1 t.u.
This means that if a t = tk a resource is busy then it was taken at least at t = tk − 1, so the task holding it already executed at least one time unit.
In these slides I’m considering that the clock resolution is much smaller than the duration of tasks and critical sections, so that factor is neglected.
Other than that the equations are similar.

PreviousStack Resource Policy NextJoint scheduling of periodic and aperiodic tasks

Last updated 2 years ago