Skip to content
GitLab
Projects
Groups
Snippets
Help
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
R
rtproofs
Project overview
Project overview
Details
Activity
Releases
Repository
Repository
Files
Commits
Branches
Tags
Contributors
Graph
Compare
Issues
0
Issues
0
List
Boards
Labels
Service Desk
Milestones
Merge Requests
0
Merge Requests
0
CI / CD
CI / CD
Pipelines
Jobs
Schedules
Operations
Operations
Incidents
Environments
Analytics
Analytics
CI / CD
Repository
Value Stream
Wiki
Wiki
Members
Members
Collapse sidebar
Close sidebar
Activity
Graph
Create a new issue
Jobs
Commits
Issue Boards
Open sidebar
Sophie Quinton
rtproofs
Commits
3fd984a4
Commit
3fd984a4
authored
Apr 04, 2019
by
Sergey Bozhko
Committed by
Sergey Bozhko
Apr 05, 2019
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
Finish FP instantiation of abstract RTA
parent
ccd2b0ca
Changes
6
Expand all
Hide whitespace changes
Inline
Sidebyside
Showing
6 changed files
with
2180 additions
and
0 deletions
+2180
0
model/schedule/uni/limited/fixed_priority/nonpr_reg/fixed/response_time_bound.v
...ited/fixed_priority/nonpr_reg/fixed/response_time_bound.v
+330
0
model/schedule/uni/limited/fixed_priority/nonpr_reg/floating/response_time_bound.v
...d/fixed_priority/nonpr_reg/floating/response_time_bound.v
+232
0
model/schedule/uni/limited/fixed_priority/nonpr_reg/nonpreemptive/response_time_bound.v
...ed_priority/nonpr_reg/nonpreemptive/response_time_bound.v
+159
0
model/schedule/uni/limited/fixed_priority/nonpr_reg/preemptive/response_time_bound.v
...fixed_priority/nonpr_reg/preemptive/response_time_bound.v
+153
0
model/schedule/uni/limited/fixed_priority/nonpr_reg/response_time_bound.v
...ni/limited/fixed_priority/nonpr_reg/response_time_bound.v
+288
0
model/schedule/uni/limited/fixed_priority/response_time_bound.v
...schedule/uni/limited/fixed_priority/response_time_bound.v
+1018
0
No files found.
model/schedule/uni/limited/fixed_priority/nonpr_reg/fixed/response_time_bound.v
0 → 100644
View file @
3fd984a4
This diff is collapsed.
Click to expand it.
model/schedule/uni/limited/fixed_priority/nonpr_reg/floating/response_time_bound.v
0 → 100644
View file @
3fd984a4
This diff is collapsed.
Click to expand it.
model/schedule/uni/limited/fixed_priority/nonpr_reg/nonpreemptive/response_time_bound.v
0 → 100644
View file @
3fd984a4
Require
Import
rt
.
util
.
all
.
Require
Import
rt
.
model
.
arrival
.
basic
.
job
rt
.
model
.
arrival
.
basic
.
task_arrival
rt
.
model
.
priority
.
Require
Import
rt
.
model
.
schedule
.
uni
.
service
rt
.
model
.
schedule
.
uni
.
workload
rt
.
model
.
schedule
.
uni
.
schedule
rt
.
model
.
schedule
.
uni
.
response_time
.
Require
Import
rt
.
model
.
schedule
.
uni
.
nonpreemptive
.
schedule
.
Require
Import
rt
.
model
.
schedule
.
uni
.
limited
.
schedule
rt
.
model
.
schedule
.
uni
.
limited
.
fixed_priority
.
nonpr_reg
.
response_time_bound
.
Require
Import
rt
.
model
.
arrival
.
curves
.
bounds
.
Require
Import
rt
.
analysis
.
uni
.
arrival_curves
.
workload_bound
.
Require
Import
rt
.
model
.
schedule
.
uni
.
limited
.
platform
.
nonpreemptive
.
From
mathcomp
Require
Import
ssreflect
ssrbool
eqtype
ssrnat
seq
path
fintype
bigop
.
(** * RTA for fully nonpreemptive FP model *)
(** In this module we prove the RTA theorem for the fully nonpreemptive FP model. *)
Module
RTAforFullyNonPreemptiveFPModelwithArrivalCurves
.
Import
Epsilon
Job
ArrivalCurves
TaskArrival
Priority
UniprocessorSchedule
NonpreemptiveSchedule
Workload
Service
FullyNonPreemptivePlatform
ResponseTime
MaxArrivalsWorkloadBound
LimitedPreemptionPlatform
RTAforFPwithBoundedNonpreemptiveSegmentsWithArrivalCurves
.
Section
Analysis
.
Context
{
Task
:
eqType
}.
Variable
task_cost
:
Task
>
time
.
Context
{
Job
:
eqType
}.
Variable
job_arrival
:
Job
>
time
.
Variable
job_cost
:
Job
>
time
.
Variable
job_task
:
Job
>
Task
.
(* Consider any arrival sequence with consistent, nonduplicate arrivals. *)
Variable
arr_seq
:
arrival_sequence
Job
.
Hypothesis
H_arrival_times_are_consistent
:
arrival_times_are_consistent
job_arrival
arr_seq
.
Hypothesis
H_arr_seq_is_a_set
:
arrival_sequence_is_a_set
arr_seq
.
(* Consider an arbitrary task set ts. *)
Variable
ts
:
list
Task
.
(* Assume that all jobs come from the task set... *)
Hypothesis
H_all_jobs_from_taskset
:
forall
j
,
arrives_in
arr_seq
j
>
job_task
j
\
in
ts
.
(* ...and the cost of a job cannot be larger than the task cost. *)
Hypothesis
H_job_cost_le_task_cost
:
cost_of_jobs_from_arrival_sequence_le_task_cost
task_cost
job_cost
job_task
arr_seq
.
(* Let max_arrivals be a family of proper arrival curves, i.e., for any task tsk in ts
[max_arrival tsk] is (1) an arrival bound of tsk, and (2) it is a monotonic function
that equals 0 for the empty interval delta = 0. *)
Variable
max_arrivals
:
Task
>
time
>
nat
.
Hypothesis
H_family_of_proper_arrival_curves
:
family_of_proper_arrival_curves
job_task
arr_seq
max_arrivals
ts
.
(* Let tsk be any task in ts. *)
Variable
tsk
:
Task
.
Hypothesis
H_tsk_in_ts
:
tsk
\
in
ts
.
(* Next, consider any uniprocessor nonpreemptive schedule of this arrival sequence...*)
Variable
sched
:
schedule
Job
.
Hypothesis
H_jobs_come_from_arrival_sequence
:
jobs_come_from_arrival_sequence
sched
arr_seq
.
Hypothesis
H_nonpreemptive_sched
:
is_nonpreemptive_schedule
job_cost
sched
.
(* ... where jobs do not execute before their arrival nor after completion. *)
Hypothesis
H_jobs_must_arrive_to_execute
:
jobs_must_arrive_to_execute
job_arrival
sched
.
Hypothesis
H_completed_jobs_dont_execute
:
completed_jobs_dont_execute
job_cost
sched
.
(* Assume we have sequential jobs, i.e, jobs from the same
task execute in the order of their arrival. *)
Hypothesis
H_sequential_jobs
:
sequential_jobs
job_arrival
job_cost
sched
job_task
.
(* Consider an FP policy that indicates a higherorequal priority relation,
and assume that the relation is reflexive and transitive. *)
Variable
higher_eq_priority
:
FP_policy
Task
.
Hypothesis
H_priority_is_reflexive
:
FP_is_reflexive
higher_eq_priority
.
Hypothesis
H_priority_is_transitive
:
FP_is_transitive
higher_eq_priority
.
(* Next, we assume that the schedule is a workconserving schedule which
respects the FP policy under a fully nonpreemptive model. *)
Hypothesis
H_work_conserving
:
work_conserving
job_arrival
job_cost
arr_seq
sched
.
Hypothesis
H_respects_policy
:
respects_FP_policy_at_preemption_point
job_arrival
job_cost
job_task
arr_seq
sched
(
can_be_preempted_for_fully_nonpreemptive_model
job_cost
)
higher_eq_priority
.
(* Let's define some local names for clarity. *)
Let
response_time_bounded_by
:
=
is_response_time_bound_of_task
job_arrival
job_cost
job_task
arr_seq
sched
.
Let
task_rbf
:
=
task_request_bound_function
task_cost
max_arrivals
tsk
.
Let
total_hep_rbf
:
=
total_hep_request_bound_function_FP
task_cost
higher_eq_priority
max_arrivals
ts
tsk
.
Let
total_ohep_rbf
:
=
total_ohep_request_bound_function_FP
task_cost
higher_eq_priority
max_arrivals
ts
tsk
.
(* Next, we define a bound for the priority inversion caused by tasks of lower priority. *)
Definition
blocking
:
=
\
max_
(
tsk_other
<
ts

~~
higher_eq_priority
tsk_other
tsk
)
(
task_cost
tsk_other

ε
).
(* Let L be any positive fixed point of the busy interval recurrence, determined by
the sum of blocking and higherorequalpriority workload. *)
Variable
L
:
time
.
Hypothesis
H_L_positive
:
L
>
0
.
Hypothesis
H_fixed_point
:
L
=
blocking
+
total_hep_rbf
L
.
(* To reduce the time complexity of the analysis, recall the notion of search space. *)
Let
is_in_search_space
A
:
=
(
A
<
L
)
&&
(
task_rbf
A
!=
task_rbf
(
A
+
ε
)).
(* Next, consider any value R, and assume that for any given arrival A from search space
there is a solution of the responsetime bound recurrence which is bounded by R. *)
Variable
R
:
nat
.
Hypothesis
H_R_is_maximum
:
forall
A
,
is_in_search_space
A
>
exists
F
,
A
+
F
=
blocking
+
(
task_rbf
(
A
+
ε
)

(
task_cost
tsk

ε
))
+
total_ohep_rbf
(
A
+
F
)
/\
F
+
(
task_cost
tsk

ε
)
<=
R
.
(* Now, we can reuse the results for the abstract model with fixed preemption points to
establish a responsetime bound for the more concrete model of fully nonpreemptive scheduling. *)
Theorem
uniprocessor_response_time_bound_fully_nonpreemptive_fp
:
response_time_bounded_by
tsk
R
.
Proof
.
move
:
(
posnP
(
task_cost
tsk
))
=>
[
ZERO

POS
].
{
intros
j
ARR
TSK
.
have
ZEROj
:
job_cost
j
=
0
.
{
move
:
(
H_job_cost_le_task_cost
j
ARR
)
=>
NEQ
.
rewrite
/
job_cost_le_task_cost
TSK
ZERO
in
NEQ
.
by
apply
/
eqP
;
rewrite

leqn0
.
}
rewrite
/
is_response_time_bound_of_job
/
completed_by
eqn_leq
;
apply
/
andP
;
split
.

by
apply
H_completed_jobs_dont_execute
.

by
rewrite
ZEROj
.
}
eapply
uniprocessor_response_time_bound_fp_with_bounded_nonpreemptive_segments
with
(
job_max_nps
:
=
fun
j
=>
job_cost
j
)
(
task_max_nps
:
=
fun
tsk
=>
task_cost
tsk
)
(
job_lock_in_service
:
=
fun
j
=>
ε
)
(
task_lock_in_service
:
=
fun
tsk
=>
ε
)
(
L0
:
=
L
)
;
eauto
2
.

by
eapply
fully_nonpreemptive_model_is_correct
;
eauto
2
.

eapply
fully_nonpreemptive_model_is_model_with_bounded_nonpreemptive_regions
;
eauto
2
.

repeat
split
;
try
done
.

intros
j
t
t'
ARR
LE
SERV
NCOMPL
.
rewrite
/
service
in
SERV
;
apply
incremental_service_during
in
SERV
.
move
:
SERV
=>
[
t_first
[/
andP
[
_
H1
]
[
H2
H3
]]].
apply
H_nonpreemptive_sched
with
t_first
;
try
done
.
by
apply
leq_trans
with
t
;
first
apply
ltnW
.

repeat
split
;
try
done
.
Qed
.
End
Analysis
.
End
RTAforFullyNonPreemptiveFPModelwithArrivalCurves
.
\ No newline at end of file
model/schedule/uni/limited/fixed_priority/nonpr_reg/preemptive/response_time_bound.v
0 → 100644
View file @
3fd984a4
Require
Import
rt
.
util
.
all
.
Require
Import
rt
.
model
.
arrival
.
basic
.
job
rt
.
model
.
arrival
.
basic
.
task_arrival
rt
.
model
.
priority
.
Require
Import
rt
.
model
.
schedule
.
uni
.
service
rt
.
model
.
schedule
.
uni
.
workload
rt
.
model
.
schedule
.
uni
.
schedule
rt
.
model
.
schedule
.
uni
.
response_time
.
Require
Import
rt
.
model
.
schedule
.
uni
.
limited
.
platform
.
preemptive
rt
.
model
.
schedule
.
uni
.
limited
.
schedule
rt
.
model
.
schedule
.
uni
.
limited
.
fixed_priority
.
nonpr_reg
.
response_time_bound
.
Require
Import
rt
.
model
.
arrival
.
curves
.
bounds
.
Require
Import
rt
.
analysis
.
uni
.
arrival_curves
.
workload_bound
.
From
mathcomp
Require
Import
ssreflect
ssrbool
eqtype
ssrnat
seq
path
fintype
bigop
.
(** * RTA for fully preemptive FP model *)
(* In this module we prove the RTA theorem for fully preemptive FP model *)
Module
RTAforFullyPreemptiveFPModelwithArrivalCurves
.
Import
Epsilon
Job
ArrivalCurves
TaskArrival
Priority
UniprocessorSchedule
Workload
Service
ResponseTime
MaxArrivalsWorkloadBound
FullyPreemptivePlatform
LimitedPreemptionPlatform
RTAforFPwithBoundedNonpreemptiveSegmentsWithArrivalCurves
.
(* In this section we prove that the maximum among the solutions of the responsetime bound
recurrence for some set of parameters is a response time bound for tsk. *)
Section
Analysis
.
Context
{
Task
:
eqType
}.
Variable
task_cost
:
Task
>
time
.
Context
{
Job
:
eqType
}.
Variable
job_arrival
:
Job
>
time
.
Variable
job_cost
:
Job
>
time
.
Variable
job_deadline
:
Job
>
time
.
Variable
job_task
:
Job
>
Task
.
(* Consider any arrival sequence with consistent, nonduplicate arrivals. *)
Variable
arr_seq
:
arrival_sequence
Job
.
Hypothesis
H_arrival_times_are_consistent
:
arrival_times_are_consistent
job_arrival
arr_seq
.
Hypothesis
H_arr_seq_is_a_set
:
arrival_sequence_is_a_set
arr_seq
.
(* Consider an arbitrary task set ts. *)
Variable
ts
:
list
Task
.
(* Assume that all jobs come from the task set... *)
Hypothesis
H_all_jobs_from_taskset
:
forall
j
,
arrives_in
arr_seq
j
>
job_task
j
\
in
ts
.
(* ...and the cost of a job cannot be larger than the task cost. *)
Hypothesis
H_job_cost_le_task_cost
:
cost_of_jobs_from_arrival_sequence_le_task_cost
task_cost
job_cost
job_task
arr_seq
.
(* Let max_arrivals be a family of proper arrival curves, i.e., for any task tsk in ts
[max_arrival tsk] is (1) an arrival bound of tsk, and (2) it is a monotonic function
that equals 0 for the empty interval delta = 0. *)
Variable
max_arrivals
:
Task
>
time
>
nat
.
Hypothesis
H_family_of_proper_arrival_curves
:
family_of_proper_arrival_curves
job_task
arr_seq
max_arrivals
ts
.
(* Let tsk be any task in ts. *)
Variable
tsk
:
Task
.
Hypothesis
H_tsk_in_ts
:
tsk
\
in
ts
.
(* Next, consider any uniprocessor schedule of the arrival sequence...*)
Variable
sched
:
schedule
Job
.
Hypothesis
H_jobs_come_from_arrival_sequence
:
jobs_come_from_arrival_sequence
sched
arr_seq
.
(* ... where jobs do not execute before their arrival nor after completion. *)
Hypothesis
H_jobs_must_arrive_to_execute
:
jobs_must_arrive_to_execute
job_arrival
sched
.
Hypothesis
H_completed_jobs_dont_execute
:
completed_jobs_dont_execute
job_cost
sched
.
(* Assume we have sequential jobs. *)
Hypothesis
H_sequential_jobs
:
sequential_jobs
job_arrival
job_cost
sched
job_task
.
(* Consider an FP policy that indicates a higherorequal priority relation,
and assume that the relation is reflexive and transitive. *)
Variable
higher_eq_priority
:
FP_policy
Task
.
Hypothesis
H_priority_is_reflexive
:
FP_is_reflexive
higher_eq_priority
.
Hypothesis
H_priority_is_transitive
:
FP_is_transitive
higher_eq_priority
.
(* Next, we assume that the schedule is a workconserving schedule which
respects the FP policy under a fully preemptive model. *)
Hypothesis
H_work_conserving
:
work_conserving
job_arrival
job_cost
arr_seq
sched
.
Hypothesis
H_respects_policy
:
respects_FP_policy_at_preemption_point
job_arrival
job_cost
job_task
arr_seq
sched
(
can_be_preempted_for_fully_preemptive_model
)
higher_eq_priority
.
(* Let's define some local names for clarity. *)
Let
response_time_bounded_by
:
=
is_response_time_bound_of_task
job_arrival
job_cost
job_task
arr_seq
sched
.
Let
task_rbf
:
=
task_request_bound_function
task_cost
max_arrivals
tsk
.
Let
total_hep_rbf
:
=
total_hep_request_bound_function_FP
task_cost
higher_eq_priority
max_arrivals
ts
tsk
.
Let
total_ohep_rbf
:
=
total_ohep_request_bound_function_FP
task_cost
higher_eq_priority
max_arrivals
ts
tsk
.
(* Let L be any positive fixed point of the busy interval recurrence, determined by
the sum of blocking and higherorequalpriority workload. *)
Variable
L
:
time
.
Hypothesis
H_L_positive
:
L
>
0
.
Hypothesis
H_fixed_point
:
L
=
total_hep_rbf
L
.
(* To reduce the time complexity of the analysis, recall the notion of search space. *)
Let
is_in_search_space
A
:
=
(
A
<
L
)
&&
(
task_rbf
A
!=
task_rbf
(
A
+
ε
)).
(* Next, consider any value R, and assume that for any given arrival A from search space
there is a solution of the responsetime bound recurrence which is bounded by R. *)
Variable
R
:
nat
.
Hypothesis
H_R_is_maximum
:
forall
A
,
is_in_search_space
A
>
exists
F
,
A
+
F
=
task_rbf
(
A
+
ε
)
+
total_ohep_rbf
(
A
+
F
)
/\
F
<=
R
.
(* Now, we can reuse the results for the abstract model with fixed preemption
points to establish a responsetime bound for the more concrete model
of fullypreemptive scheduling. *)
Theorem
uniprocessor_response_time_bound_fully_preemptive_fp
:
response_time_bounded_by
tsk
R
.
Proof
.
have
BLOCK
:
RTAforFPwithBoundedNonpreemptiveSegmentsWithArrivalCurves
.
blocking_bound
(
fun
_
=>
ε
)
higher_eq_priority
ts
tsk
=
0
.
{
by
rewrite
/
RTAforFPwithBoundedNonpreemptiveSegmentsWithArrivalCurves
.
blocking_bound
subnn
big1_eq
.
}
eapply
uniprocessor_response_time_bound_fp_with_bounded_nonpreemptive_segments
with
(
task_max_nps
:
=
fun
_
=>
ε
)
(
can_be_preempted
:
=
fun
j
prog
=>
true
)
(
task_lock_in_service
:
=
fun
tsk
=>
task_cost
tsk
)
(
job_lock_in_service
:
=
fun
j
=>
job_cost
j
)
(
job_max_nps
:
=
fun
j
=>
ε
)
;
eauto
2
;
try
done
.

by
eapply
fully_preemptive_model_is_model_with_bounded_nonpreemptive_regions
.

repeat
split
;
try
done
.
intros
?
?
?
ARR
;
move
=>
LE
COMPL
/
negP
NCOMPL
.
exfalso
;
apply
:
NCOMPL
.
apply
completion_monotonic
with
t
;
try
done
.
rewrite
/
completed_by
eqn_leq
;
apply
/
andP
;
split
;
try
done
.

repeat
split
;
try
done
.
rewrite
/
task_lock_in_service_le_task_cost
.
by
done
.
unfold
task_lock_in_service_bounds_job_lock_in_service
.
by
intros
?
ARR
TSK
;
rewrite

TSK
;
apply
H_job_cost_le_task_cost
.

by
rewrite
BLOCK
add0n
.

move
=>
A
/
andP
[
LT
NEQ
].
specialize
(
H_R_is_maximum
A
)
;
feed
H_R_is_maximum
.
{
by
apply
/
andP
;
split
.
}
move
:
H_R_is_maximum
=>
[
F
[
FIX
BOUND
]].
exists
F
;
split
.
+
by
rewrite
BLOCK
add0n
subnn
subn0
.
+
by
rewrite
subnn
addn0
.
Qed
.
End
Analysis
.
End
RTAforFullyPreemptiveFPModelwithArrivalCurves
.
\ No newline at end of file
model/schedule/uni/limited/fixed_priority/nonpr_reg/response_time_bound.v
0 → 100644
View file @
3fd984a4
This diff is collapsed.
Click to expand it.
model/schedule/uni/limited/fixed_priority/response_time_bound.v
0 → 100644
View file @
3fd984a4
This diff is collapsed.
Click to expand it.
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment