Library prosa.analysis.definitions.priority_inversion
Priority Inversion
In this section, we define the notion of priority inversion for arbitrary processors.
Consider any type of tasks ...
... and any type of jobs associated with these tasks.
Context {Job : JobType}.
Context `{JobTask Job Task}.
Context `{JobArrival Job}.
Context `{JobCost Job}.
Context `{JobTask Job Task}.
Context `{JobArrival Job}.
Context `{JobCost Job}.
Next, consider any kind of processor state model, ...
... any arrival sequence, ...
... and any schedule.
Assume a given JLFP policy.
Consider an arbitrary job.
We say that the job incurs priority inversion if it has higher
priority than the scheduled job. Note that this definition
implicitly assumes that the scheduler is
work-conserving. Therefore, it cannot be applied to models with
jitter or self-suspensions.
Definition priority_inversion (t : instant) :=
~~ scheduled_at sched j t ∧
∃ jlp,
scheduled_at sched jlp t && ~~ hep_job jlp j.
~~ scheduled_at sched j t ∧
∃ jlp,
scheduled_at sched jlp t && ~~ hep_job jlp j.
Similarly, we define a decidable counterpart of
priority_inversion.
Definition priority_inversion_dec (t : instant) :=
~~ scheduled_at sched j t &&
has (fun jlp ⇒ scheduled_at sched jlp t && ~~ hep_job jlp j)
(arrivals_before arr_seq t.+1).
~~ scheduled_at sched j t &&
has (fun jlp ⇒ scheduled_at sched jlp t && ~~ hep_job jlp j)
(arrivals_before arr_seq t.+1).
In the following short section we show that the propositional
and computational versions of priority_inversion are
equivalent.
We assume that all jobs come from the arrival sequence and do
not execute before their arrival nor after completion.
Hypothesis H_jobs_come_from_arrival_sequence : jobs_come_from_arrival_sequence sched arr_seq.
Hypothesis H_jobs_must_arrive_to_execute : jobs_must_arrive_to_execute sched.
Hypothesis H_arrival_times_are_consistent : consistent_arrival_times arr_seq.
Hypothesis H_jobs_must_arrive_to_execute : jobs_must_arrive_to_execute sched.
Hypothesis H_arrival_times_are_consistent : consistent_arrival_times arr_seq.
Then the propositional and computational versions of
priority_inversion are equivalent.
Lemma priority_inversion_P :
∀ t, reflect (priority_inversion t) (priority_inversion_dec t).
Proof.
intros t; apply /introP.
- move⇒ /andP [SCHED /hasP [jlp ARR PRIO]].
by split ⇒ //; (∃ jlp).
- rewrite negb_and /priority_inversion; move⇒ /orP NPRIO.
destruct NPRIO as [NSCHED | NHAS].
+ rewrite Bool.negb_involutive in NSCHED.
by rewrite NSCHED; intros [F _].
+ move ⇒ [NSCHED [jlp /andP[SCHED_jlp NHEP]]].
move: NHAS ⇒ /hasPn NJOBS.
specialize (NJOBS jlp).
have ARR: jlp \in arrivals_before arr_seq t.+1.
{ move: (H_jobs_must_arrive_to_execute jlp t SCHED_jlp) ⇒ ARR.
apply arrived_between_implies_in_arrivals ⇒ //=.
by apply (H_jobs_come_from_arrival_sequence jlp t).
}
move: (NJOBS ARR) ⇒ /negP NN.
by apply NN; apply /andP; split ⇒ //.
Qed.
End PriorityInversionReflection.
∀ t, reflect (priority_inversion t) (priority_inversion_dec t).
Proof.
intros t; apply /introP.
- move⇒ /andP [SCHED /hasP [jlp ARR PRIO]].
by split ⇒ //; (∃ jlp).
- rewrite negb_and /priority_inversion; move⇒ /orP NPRIO.
destruct NPRIO as [NSCHED | NHAS].
+ rewrite Bool.negb_involutive in NSCHED.
by rewrite NSCHED; intros [F _].
+ move ⇒ [NSCHED [jlp /andP[SCHED_jlp NHEP]]].
move: NHAS ⇒ /hasPn NJOBS.
specialize (NJOBS jlp).
have ARR: jlp \in arrivals_before arr_seq t.+1.
{ move: (H_jobs_must_arrive_to_execute jlp t SCHED_jlp) ⇒ ARR.
apply arrived_between_implies_in_arrivals ⇒ //=.
by apply (H_jobs_come_from_arrival_sequence jlp t).
}
move: (NJOBS ARR) ⇒ /negP NN.
by apply NN; apply /andP; split ⇒ //.
Qed.
End PriorityInversionReflection.
Cumulative priority inversion incurred by a job within some time
interval
[t1, t2)
is the total number of time instances
within [t1,t2)
at which job j incurred priority
inversion.
Definition cumulative_priority_inversion (t1 t2 : instant) :=
\sum_(t1 ≤ t < t2) priority_inversion_dec t.
\sum_(t1 ≤ t < t2) priority_inversion_dec t.
We say that the priority inversion experienced by a job j is
bounded by a constant B if the cumulative priority inversion
within any busy interval prefix is bounded by B.
Definition priority_inversion_of_job_is_bounded_by_constant (B : duration) :=
∀ (t1 t2 : instant),
busy_interval_prefix arr_seq sched j t1 t2 →
cumulative_priority_inversion t1 t2 ≤ B.
∀ (t1 t2 : instant),
busy_interval_prefix arr_seq sched j t1 t2 →
cumulative_priority_inversion t1 t2 ≤ B.
More generally, if the priority inversion experienced by job j
depends on its relative arrival time w.r.t. the beginning of its
busy interval at a time t1, we say that the priority inversion
of job j is bounded by a function B : duration → duration if
the cumulative priority inversion within any busy interval
prefix is bounded by B (job_arrival j - t1)
Definition priority_inversion_of_job_is_bounded_by (B : duration → duration) :=
∀ (t1 t2 : instant),
busy_interval_prefix arr_seq sched j t1 t2 →
cumulative_priority_inversion t1 t2 ≤ B (job_arrival j - t1).
End PriorityInversion.
∀ (t1 t2 : instant),
busy_interval_prefix arr_seq sched j t1 t2 →
cumulative_priority_inversion t1 t2 ≤ B (job_arrival j - t1).
End PriorityInversion.
In this section, we define a notion of the bounded priority inversion for tasks.
Consider any type of tasks ...
... and any type of jobs associated with these tasks.
Context {Job : JobType}.
Context `{JobTask Job Task}.
Context `{JobArrival Job}.
Context `{JobCost Job}.
Context `{JobTask Job Task}.
Context `{JobArrival Job}.
Context `{JobCost Job}.
Next, consider any kind of processor state model, ...
... any arrival sequence, ...
... and any schedule.
Assume a given JLFP policy.
Consider an arbitrary task tsk.
We say that priority inversion of task tsk is bounded by a constant
B if all jobs released by the task have cumulative priority inversion
bounded by B.
Definition priority_inversion_is_bounded_by_constant (B : duration) :=
∀ (j : Job),
arrives_in arr_seq j →
job_of_task tsk j →
job_cost j > 0 →
priority_inversion_of_job_is_bounded_by_constant arr_seq sched j B.
∀ (j : Job),
arrives_in arr_seq j →
job_of_task tsk j →
job_cost j > 0 →
priority_inversion_of_job_is_bounded_by_constant arr_seq sched j B.
We say that task tsk has bounded priority inversion if all its
jobs have bounded cumulative priority inversion that depends on
its relative arrival time w.r.t. the beginning of the busy
interval.
Definition priority_inversion_is_bounded_by (B : duration → duration) :=
∀ (j : Job),
arrives_in arr_seq j →
job_of_task tsk j →
job_cost j > 0 →
priority_inversion_of_job_is_bounded_by arr_seq sched j B.
End TaskPriorityInversionBound.
∀ (j : Job),
arrives_in arr_seq j →
job_of_task tsk j →
job_cost j > 0 →
priority_inversion_of_job_is_bounded_by arr_seq sched j B.
End TaskPriorityInversionBound.