Move tg_weight() upward and make cpu_shares_read_u64() use it too. This
makes the weight retrieval shared between cgroup v1 and v2 paths and will be
used to implement cgroup support for sched_ext.

No functional changes.
Signed-off-by: Tejun Heo <tj@kernel.org>

A new BPF extensible sched_class will use css_tg() in the init and exit
paths to visit all task_groups by walking cgroups.

v4: __setscheduler_prio() is already exposed. Dropped from this patch.

v3: Dropped SCHED_CHANGE_BLOCK() as upstream is adding more generic cleanup
    mechanism.

v2: Expose SCHED_CHANGE_BLOCK() too and update the description.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: David Vernet <dvernet@meta.com>
Acked-by: Josh Don <joshdon@google.com>
Acked-by: Hao Luo <haoluo@google.com>
Acked-by: Barret Rhoden <brho@google.com>

During scx_ops_enable(), SCX needs to invoke the sleepable ops.init_task()
on every task. To do this, it does get_task_struct() on each iterated task,
drop the lock and then call ops.init_task().

However, a TASK_DEAD task may already have lost all its usage count and be
waiting for RCU grace period to be freed. If get_task_struct() is called on
such task, use-after-free can happen. To avoid such situations,
scx_ops_enable() skips initialization of TASK_DEAD tasks, which seems safe
as they are never going to be scheduled again.

Unfortunately, a racing sched_setscheduler(2) can grab the task before the
task is unhashed and then continue to e.g. move the task from RT to SCX
after TASK_DEAD is set and ops_enable skipped the task. As the task hasn't
gone through scx_ops_init_task(), scx_ops_enable_task() called from
switching_to_scx() triggers the following warning:

  sched_ext: Invalid task state transition 0 -> 3 for stress-ng-race-[2872]
  WARNING: CPU: 6 PID: 2367 at kernel/sched/ext.c:3327 scx_ops_enable_task+0x18f/0x1f0
  ...
  RIP: 0010:scx_ops_enable_task+0x18f/0x1f0
  ...
   switching_to_scx+0x13/0xa0
   __sched_setscheduler+0x84e/0xa50
   do_sched_setscheduler+0x104/0x1c0
   __x64_sys_sched_setscheduler+0x18/0x30
   do_syscall_64+0x7b/0x140
   entry_SYSCALL_64_after_hwframe+0x76/0x7e

As in the ops_disable path, it just doesn't seem like a good idea to leave
any task in an inconsistent state, even when the task is dead. The root
cause is ops_enable not being able to tell reliably whether a task is truly
dead (no one else is looking at it and it's about to be freed) and was
testing TASK_DEAD instead. Fix it by testing the task's usage count
directly.

- ops_init no longer ignores TASK_DEAD tasks. As now all users iterate all
  tasks, @include_dead is removed from scx_task_iter_next_locked() along
  with dead task filtering.

- tryget_task_struct() is added. Tasks are skipped iff tryget_task_struct()
  fails.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: David Vernet <void@manifault.com>
Cc: Peter Zijlstra <peterz@infradead.org>

scx_ops_disable_workfn() only switches !TASK_DEAD tasks out of SCX while
calling scx_ops_exit_task() on all tasks including dead ones. This can leave
a dead task on SCX but with SCX_TASK_NONE state, which is inconsistent.

If another task was in the process of changing the TASK_DEAD task's
scheduling class and grabs the rq lock after scx_ops_disable_workfn() is
done with the task, the task ends up calling scx_ops_disable_task() on the
dead task which is in an inconsistent state triggering a warning:

  WARNING: CPU: 6 PID: 3316 at kernel/sched/ext.c:3411 scx_ops_disable_task+0x12c/0x160
  ...
  RIP: 0010:scx_ops_disable_task+0x12c/0x160
  ...
  Call Trace:
   <TASK>
   check_class_changed+0x2c/0x70
   __sched_setscheduler+0x8a0/0xa50
   do_sched_setscheduler+0x104/0x1c0
   __x64_sys_sched_setscheduler+0x18/0x30
   do_syscall_64+0x7b/0x140
   entry_SYSCALL_64_after_hwframe+0x76/0x7e
  RIP: 0033:0x7f140d70ea5b

There is no reason to leave dead tasks on SCX when unloading the BPF
scheduler. Fix by making scx_ops_disable_workfn() eject all tasks including
the dead ones from SCX.
Signed-off-by: Tejun Heo <tj@kernel.org>

With sched_ext converted to use put_prev_task() for class switch detection,
there's no user of switch_class() left. Drop it.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>

Now that put_prev_task_scx() is called with @next on task switches, there's
no reason to use sched_class.switch_class(). Rename switch_class_scx() to
switch_class() and call it from put_prev_task_scx().
Signed-off-by: Tejun Heo <tj@kernel.org>

Relocate functions to ease the removal of switch_class_scx(). No functional
changes.
Signed-off-by: Tejun Heo <tj@kernel.org>

Because the BPF scheduler's dispatch path is invoked from balance(),
sched_ext needs to invoke balance_one() on all sibling rq's before picking
the next task for core-sched.

Before the recent pick_next_task() updates, sched_ext couldn't share pick
task between regular and core-sched paths because pick_next_task() depended
on put_prev_task() being called on the current task. Tasks currently running
on sibling rq's can't be put when one rq is trying to pick the next task, so
pick_task_scx() had to have a separate mechanism to pick between a sibling
rq's current task and the first task in its local DSQ.

However, with the preceding updates, pick_next_task_scx() no longer depends
on the current task being put and can compare the current task and the next
in line statelessly, and the pick task logic should be shareable between
regular and core-sched paths.

Unify regular and core-sched pick task paths:

- There's no reason to distinguish local and sibling picks anymore. @local
  is removed from balance_one().

- pick_next_task_scx() is turned into pick_task_scx() by dropping the
  put_prev_set_next_task() call.

- The old pick_task_scx() is dropped.
Signed-off-by: Tejun Heo <tj@kernel.org>

SCX_TASK_BAL_KEEP is used by balance_one() to tell pick_next_task_scx() to
keep running the current task. It's not really a task property. Replace it
with SCX_RQ_BAL_KEEP which resides in rq->scx.flags and is a better fit for
the usage. Also, the existing clearing rule is unnecessarily strict and
makes it difficult to use with core-sched. Just clear it on entry to
balance_one().
Signed-off-by: Tejun Heo <tj@kernel.org>

fd03c5b8 ("sched: Rework pick_next_task()") changed the definition of
pick_next_task() from:

  pick_next_task() := pick_task() + set_next_task(.first = true)

to:

  pick_next_task(prev) := pick_task() + put_prev_task() + set_next_task(.first = true)

making invoking put_prev_task() pick_next_task()'s responsibility. This
reordering allows pick_task() to be shared between regular and core-sched
paths and put_prev_task() to know the next task.

sched_ext depended on put_prev_task_scx() enqueueing the current task before
pick_next_task_scx() is called. While pulling sched/core changes,
70cc76aa0d80 ("Merge branch 'tip/sched/core' into for-6.12") added an
explicit put_prev_task_scx() call for SCX tasks in pick_next_task_scx()
before picking the first task as a workaround.

Clean it up and adopt the conventions that other sched classes are
following.

The operation of keeping running the current task was spread and required
the task to be put on the local DSQ before picking:

  - balance_one() used SCX_TASK_BAL_KEEP to indicate that the task is still
    runnable, hasn't exhausted its slice, and thus should keep running.

  - put_prev_task_scx() enqueued the task to local DSQ if SCX_TASK_BAL_KEEP
    is set. It also called do_enqueue_task() with SCX_ENQ_LAST if it is the
    only runnable task. do_enqueue_task() in turn decided whether to use the
    local DSQ depending on SCX_OPS_ENQ_LAST.

Consolidate the logic in balance_one() as it always knows whether it is
going to keep the current task. balance_one() now considers all conditions
where the current task should be kept and uses SCX_TASK_BAL_KEEP to tell
pick_next_task_scx() to keep the current task instead of picking one from
the local DSQ. Accordingly, SCX_ENQ_LAST handling is removed from
put_prev_task_scx() and do_enqueue_task() and pick_next_task_scx() is
updated to pick the current task if SCX_TASK_BAL_KEEP is set.

The workaround put_prev_task[_scx]() calls are replaced with
put_prev_set_next_task().

This causes two behavior changes observable from the BPF scheduler:

- When a task keep running, it no longer goes through enqueue/dequeue cycle
  and thus ops.stopping/running() transitions. The new behavior is better
  and all the existing schedulers should be able to handle the new behavior.

- The BPF scheduler cannot keep executing the current task by enqueueing
  SCX_ENQ_LAST task to the local DSQ. If SCX_OPS_ENQ_LAST is specified, the
  BPF scheduler is responsible for resuming execution after each
  SCX_ENQ_LAST. SCX_OPS_ENQ_LAST is mostly useful for cases where scheduling
  decisions are not made on the local CPU - e.g. central or userspace-driven
  schedulin - and the new behavior is more logical and shouldn't pose any
  problems. SCX_OPS_ENQ_LAST demonstration from scx_qmap is dropped as it
  doesn't fit that well anymore and the last task handling is moved to the
  end of qmap_dispatch().
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: David Vernet <void@manifault.com>
Cc: Andrea Righi <righi.andrea@gmail.com>
Cc: Changwoo Min <multics69@gmail.com>
Cc: Daniel Hodges <hodges.daniel.scott@gmail.com>
Cc: Dan Schatzberg <schatzberg.dan@gmail.com>

- Resolve trivial context conflicts from dl_server clearing being moved
  around.

- Add @next to put_prev_task_scx() and @prev to pick_next_task_scx() to
  match sched/core.

- Merge sched_class->switch_class() addition from sched_ext with
  tip/sched/core changes in __pick_next_task().

- Make pick_next_task_scx() call put_prev_task_scx() to emulate the previous
  behavior where sched_class->put_prev_task() was called before
  sched_class->pick_next_task().

While this makes sched_ext build and function, the behavior is not in line
with other sched classes. The follow-up patches will address the
discrepancies and remove sched_class->switch_class().
Signed-off-by: Tejun Heo <tj@kernel.org>

In order to tell the previous sched_class what the next task is, add
put_prev_task(.next).

Notable SCX will use this to:

 1) determine the next task will leave the SCX sched class and push
    the current task to another CPU if possible.
 2) statistics on how often and which other classes preempt it
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20240813224016.367421076@infradead.org

When a task is selected through a dl_server, it will have p->dl_server
set, such that it can account runtime to the dl_server, see
update_curr_task().

Currently p->dl_server is set in pick*task() whenever it goes through
the dl_server, clearing it is a bit of a mess though. The trivial
solution is clearing it on the final put (now that we have this
location).

However, this gives a problem when:

	p = pick_task(rq);
	if (p)
		put_prev_set_next_task(rq, prev, next);

picks the same task but through a different path, notably when it goes
from picking through the dl_server to a direct pick or vice-versa. In
that case we cannot readily determine wether we should clear or
preserve p->dl_server.

An additional complication is pick_*task() setting p->dl_server for a
remote pick, it might still need to update runtime before it schedules
the core_pick.

Close all these holes and remove all the random clearing of
p->dl_server by:

 - having pick_*task() manage rq->dl_server

 - having the final put_prev_task() clear p->dl_server

 - having the first set_next_task() set p->dl_server = rq->dl_server

 - complicate the core_sched code to save/restore rq->dl_server where
   appropriate.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20240813224016.259853414@infradead.org

Ensure the last put_prev_task() and the first set_next_task() always
go together.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20240813224016.158454756@infradead.org

The current rule is that:

  pick_next_task() := pick_task() + set_next_task(.first = true)

And many classes implement it directly as such. Change things around
to make pick_next_task() optional while also changing the definition to:

  pick_next_task(prev) := pick_task() + put_prev_task() + set_next_task(.first = true)

The reason is that sched_ext would like to have a 'final' call that
knows the next task. By placing put_prev_task() right next to
set_next_task() (as it already is for sched_core) this becomes
trivial.

As a bonus, this is a nice cleanup on its own.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20240813224016.051225657@infradead.org

With the goal of pushing put_prev_task() after pick_task() / into
pick_next_task().
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20240813224015.943143811@infradead.org

Abide by the simple rule:

  pick_next_task() := pick_task() + set_next_task(.first = true)

This allows us to trivially get rid of server_pick_next() and things
collapse nicely.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20240813224015.837303391@infradead.org

The rule is that:

  pick_next_task() := pick_task() + set_next_task(.first = true)

Turns out, there's still a few things in pick_next_task() that are
missing from that combination.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20240813224015.724111109@infradead.org

Turns out the core_sched bits forgot to use the
set_next_task(.first=true) variant. Notably:

  pick_next_task() := pick_task() + set_next_task(.first = true)
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20240813224015.614146342@infradead.org

__sched_setscheduler() goes through an enqueue/dequeue cycle like so:

  flags := DEQUEUE_SAVE | DEQUEUE_MOVE | DEQUEUE_NOCLOCK;
  prev_class->dequeue_task(rq, p, flags);
  new_class->enqueue_task(rq, p, flags);

when prev_class := fair_sched_class, this is followed by:

  dequeue_task(rq, p, DEQUEUE_NOCLOCK | DEQUEUE_SLEEP);

the idea being that since the task has switched classes, we need to drop
the sched_delayed logic and have that task be deactivated per its previous
dequeue_task(..., DEQUEUE_SLEEP).

Unfortunately, this leaves the task on_rq. This is missing the tail end of
dequeue_entities() that issues __block_task(), which __sched_setscheduler()
won't have done due to not using DEQUEUE_DELAYED - not that it should, as
it is pretty much a fair_sched_class specific thing.

Make switched_from_fair() properly deactivate sched_delayed tasks upon
class changes via __block_task(), as if a
  dequeue_task(..., DEQUEUE_DELAYED)
had been issued.

Fixes: 2e0199df ("sched/fair: Prepare exit/cleanup paths for delayed_dequeue")
Reported-by: "Paul E. McKenney" <paulmck@kernel.org>
Reported-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20240829135353.1524260-1-vschneid@redhat.com

In dl_server_start(), when schedstats is enabled, the following
happens:

  dl_server_start()
    dl_se->dl_server = 1;
    enqueue_dl_entity()
      update_stats_enqueue_dl()
        __schedstats_from_dl_se()
          dl_task_of()
            BUG_ON(dl_server(dl_se));

Since only tasks have schedstats and internal entries do not, avoid
trying to update stats in this case.

Fixes: 63ba8422 ("sched/deadline: Introduce deadline servers")
Signed-off-by: Huang Shijie <shijie@os.amperecomputing.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Juri Lelli <juri.lelli@redhat.com>
Link: https://lkml.kernel.org/r/20240829031111.12142-1-shijie@os.amperecomputing.com

Since 3cf78c5d ("sched_ext: Unpin and repin rq lock from
balance_scx()"), sched_ext's balance path terminates rq_pin in the outermost
function. This is simpler and in line with what other balance functions are
doing but it loses control over rq->clock_update_flags which makes
assert_clock_udpated() trigger if other CPUs pins the rq lock.

The only place this matters is touch_core_sched() which uses the timestamp
to order tasks from sibling rq's. Switch to sched_clock_cpu(). Later, it may
be better to use per-core dispatch sequence number.

v2: Use sched_clock_cpu() instead of ktime_get_ns() per David.
Signed-off-by: Tejun Heo <tj@kernel.org>
Fixes: 3cf78c5d ("sched_ext: Unpin and repin rq lock from balance_scx()")
Acked-by: David Vernet <void@manifault.com>
Cc: Peter Zijlstra <peterz@infradead.org>

When deciding whether a task can be migrated to a CPU,
dispatch_to_local_dsq() was open-coding p->cpus_allowed and scx_rq_online()
tests instead of using task_can_run_on_remote_rq(). This had two problems.

- It was missing is_migration_disabled() check and thus could try to migrate
  a task which shouldn't leading to assertion and scheduling failures.

- It was testing p->cpus_ptr directly instead of using task_allowed_on_cpu()
  and thus failed to consider ISA compatibility.

Update dispatch_to_local_dsq() to use task_can_run_on_remote_rq():

- Move scx_ops_error() triggering into task_can_run_on_remote_rq().

- When migration isn't allowed, fall back to the global DSQ instead of the
  source DSQ by returning DTL_INVALID. This is both simpler and an overall
  better behavior.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Acked-by: David Vernet <void@manifault.com>

The cfi stub for ops.tick was missing which will fail scheduler loading
after pending BPF changes. Add it.
Signed-off-by: Tejun Heo <tj@kernel.org>

ARRAY_ELEM_PTR() is an access macro used to help the BPF verifier not
confused by offseted memory acceeses by yiedling a valid pointer or NULL in
a way that's clear to the verifier. As such, the canonical usage involves
checking NULL return from the macro. Note that in many cases, the NULL
condition can never happen - they're there just to hint the verifier.

In a bpf_loop in scx_central.bpf.c::central_dispatch(), the NULL check was
incorrect in that there was another dereference of the pointer in addition
to the NULL checked access. This worked as the pointer can never be NULL and
the verifier could tell it would never be NULL in this case.

However, this still looks wrong and trips smatch:

  ./tools/sched_ext/scx_central.bpf.c:205 ____central_dispatch()
  error: we previously assumed 'gimme' could be null (see line 201)

  ./tools/sched_ext/scx_central.bpf.c
      195
      196                         if (!scx_bpf_dispatch_nr_slots())
      197                                 break;
      198
      199                         /* central's gimme is never set */
      200                         gimme = ARRAY_ELEM_PTR(cpu_gimme_task, cpu, nr_cpu_ids);
      201                         if (gimme && !*gimme)
				      ^^^^^
  If gimme is NULL

      202                                 continue;
      203
      204                         if (dispatch_to_cpu(cpu))
  --> 205                                 *gimme = false;

Fix the NULL check so that there are no derefs if NULL. This doesn't change
actual behavior.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Link: http://lkml.kernel.org/r/<955e1c3c-ace2-4a1d-b246-15b8196038a3@stanley.mountain>

Since dequeue_task() allowed to fail, there is a compile error:

kernel/sched/ext.c:3630:19: error: initialization of ‘bool (*)(struct rq*, struct task_struct *, int)’ {aka ‘_Bool (*)(struct rq *, struct task_struct *, int)’} from incompatible pointer type ‘void (*)(struct rq*, struct task_struct *, int)’
  3630 |  .dequeue_task  = dequeue_task_scx,
       |                   ^~~~~~~~~~~~~~~~

Allow dequeue_task_scx to fail too.

Fixes: 863ccdbb ("sched: Allow sched_class::dequeue_task() to fail")
Signed-off-by: Yipeng Zou <zouyipeng@huawei.com>
Signed-off-by: Tejun Heo <tj@kernel.org>

To receive 863ccdbb ("sched: Allow sched_class::dequeue_task() to fail")
which makes sched_class.dequeue_task() return bool instead of void. This
leads to compile breakage and will be fixed by a follow-up patch.
Signed-off-by: Tejun Heo <tj@kernel.org>

In the absence of an explicit cgroup slice configureation, make mixed
slice length work with cgroups by propagating the min_slice up the
hierarchy.

This ensures the cgroup entity gets timely service to service its
entities that have this timing constraint set on them.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105030.948188417@infradead.org

Allow applications to directly set a suggested request/slice length using
sched_attr::sched_runtime.

The implementation clamps the value to: 0.1[ms] <= slice <= 100[ms]
which is 1/10 the size of HZ=1000 and 10 times the size of HZ=100.

Applications should strive to use their periodic runtime at a high
confidence interval (95%+) as the target slice. Using a smaller slice
will introduce undue preemptions, while using a larger value will
increase latency.

For all the following examples assume a scheduling quantum of 8, and for
consistency all examples have W=4:

  {A,B,C,D}(w=1,r=8):

  ABCD...
  +---+---+---+---

  t=0, V=1.5				t=1, V=3.5
  A  |------<				A          |------<
  B   |------<				B   |------<
  C    |------<				C    |------<
  D     |------<			D     |------<
  ---+*------+-------+---		---+--*----+-------+---

  t=2, V=5.5				t=3, V=7.5
  A          |------<			A          |------<
  B           |------<			B           |------<
  C    |------<				C            |------<
  D     |------<			D     |------<
  ---+----*--+-------+---		---+------*+-------+---

Note: 4 identical tasks in FIFO order

~~~

  {A,B}(w=1,r=16) C(w=2,r=16)

  AACCBBCC...
  +---+---+---+---

  t=0, V=1.25				t=2, V=5.25
  A  |--------------<                   A                  |--------------<
  B   |--------------<                  B   |--------------<
  C    |------<                         C    |------<
  ---+*------+-------+---               ---+----*--+-------+---

  t=4, V=8.25				t=6, V=12.25
  A                  |--------------<   A                  |--------------<
  B   |--------------<                  B                   |--------------<
  C            |------<                 C            |------<
  ---+-------*-------+---               ---+-------+---*---+---

Note: 1 heavy task -- because q=8, double r such that the deadline of the w=2
      task doesn't go below q.

Note: observe the full schedule becomes: W*max(r_i/w_i) = 4*2q = 8q in length.

Note: the period of the heavy task is half the full period at:
      W*(r_i/w_i) = 4*(2q/2) = 4q

~~~

  {A,C,D}(w=1,r=16) B(w=1,r=8):

  BAACCBDD...
  +---+---+---+---

  t=0, V=1.5				t=1, V=3.5
  A  |--------------<			A  |---------------<
  B   |------<				B           |------<
  C    |--------------<			C    |--------------<
  D     |--------------<		D     |--------------<
  ---+*------+-------+---		---+--*----+-------+---

  t=3, V=7.5				t=5, V=11.5
  A                  |---------------<  A                  |---------------<
  B           |------<                  B           |------<
  C    |--------------<                 C                    |--------------<
  D     |--------------<                D     |--------------<
  ---+------*+-------+---               ---+-------+--*----+---

  t=6, V=13.5
  A                  |---------------<
  B                   |------<
  C                    |--------------<
  D     |--------------<
  ---+-------+----*--+---

Note: 1 short task -- again double r so that the deadline of the short task
      won't be below q. Made B short because its not the leftmost task, but is
      eligible with the 0,1,2,3 spread.

Note: like with the heavy task, the period of the short task observes:
      W*(r_i/w_i) = 4*(1q/1) = 4q

~~~

  A(w=1,r=16) B(w=1,r=8) C(w=2,r=16)

  BCCAABCC...
  +---+---+---+---

  t=0, V=1.25				t=1, V=3.25
  A  |--------------<                   A  |--------------<
  B   |------<                          B           |------<
  C    |------<                         C    |------<
  ---+*------+-------+---               ---+--*----+-------+---

  t=3, V=7.25				t=5, V=11.25
  A  |--------------<                   A                  |--------------<
  B           |------<                  B           |------<
  C            |------<                 C            |------<
  ---+------*+-------+---               ---+-------+--*----+---

  t=6, V=13.25
  A                  |--------------<
  B                   |------<
  C            |------<
  ---+-------+----*--+---

Note: 1 heavy and 1 short task -- combine them all.

Note: both the short and heavy task end up with a period of 4q

~~~

  A(w=1,r=16) B(w=2,r=16) C(w=1,r=8)

  BBCAABBC...
  +---+---+---+---

  t=0, V=1				t=2, V=5
  A  |--------------<                   A  |--------------<
  B   |------<                          B           |------<
  C    |------<                         C    |------<
  ---+*------+-------+---               ---+----*--+-------+---

  t=3, V=7				t=5, V=11
  A  |--------------<                   A                  |--------------<
  B           |------<                  B           |------<
  C            |------<                 C            |------<
  ---+------*+-------+---               ---+-------+--*----+---

  t=7, V=15
  A                  |--------------<
  B                   |------<
  C            |------<
  ---+-------+------*+---

Note: as before but permuted

~~~

From all this it can be deduced that, for the steady state:

 - the total period (P) of a schedule is:	W*max(r_i/w_i)
 - the average period of a task is:		W*(r_i/w_i)
 - each task obtains the fair share:		w_i/W of each full period P
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105030.842834421@infradead.org

Part of the reason to have shorter slices is to improve
responsiveness. Allow shorter slices to preempt longer slices on
wakeup.

    Task                  |   Runtime ms  | Switches | Avg delay ms    | Max delay ms    | Sum delay ms     |

  100ms massive_intr 500us cyclictest NO_PREEMPT_SHORT

  1 massive_intr:(5)      | 846018.956 ms |   779188 | avg:   0.273 ms | max:  58.337 ms | sum:212545.245 ms |
  2 massive_intr:(5)      | 853450.693 ms |   792269 | avg:   0.275 ms | max:  71.193 ms | sum:218263.588 ms |
  3 massive_intr:(5)      | 843888.920 ms |   771456 | avg:   0.277 ms | max:  92.405 ms | sum:213353.221 ms |
  1 chromium-browse:(8)   |  53015.889 ms |   131766 | avg:   0.463 ms | max:  36.341 ms | sum:60959.230  ms |
  2 chromium-browse:(8)   |  53864.088 ms |   136962 | avg:   0.480 ms | max:  27.091 ms | sum:65687.681  ms |
  3 chromium-browse:(9)   |  53637.904 ms |   132637 | avg:   0.481 ms | max:  24.756 ms | sum:63781.673  ms |
  1 cyclictest:(5)        |  12615.604 ms |   639689 | avg:   0.471 ms | max:  32.272 ms | sum:301351.094 ms |
  2 cyclictest:(5)        |  12511.583 ms |   642578 | avg:   0.448 ms | max:  44.243 ms | sum:287632.830 ms |
  3 cyclictest:(5)        |  12545.867 ms |   635953 | avg:   0.475 ms | max:  25.530 ms | sum:302374.658 ms |

  100ms massive_intr 500us cyclictest PREEMPT_SHORT

  1 massive_intr:(5)      | 839843.919 ms |   837384 | avg:   0.264 ms | max:  74.366 ms | sum:221476.885 ms |
  2 massive_intr:(5)      | 852449.913 ms |   845086 | avg:   0.252 ms | max:  68.162 ms | sum:212595.968 ms |
  3 massive_intr:(5)      | 839180.725 ms |   836883 | avg:   0.266 ms | max:  69.742 ms | sum:222812.038 ms |
  1 chromium-browse:(11)  |  54591.481 ms |   138388 | avg:   0.458 ms | max:  35.427 ms | sum:63401.508  ms |
  2 chromium-browse:(8)   |  52034.541 ms |   132276 | avg:   0.436 ms | max:  31.826 ms | sum:57732.958  ms |
  3 chromium-browse:(8)   |  55231.771 ms |   141892 | avg:   0.469 ms | max:  27.607 ms | sum:66538.697  ms |
  1 cyclictest:(5)        |  13156.391 ms |   667412 | avg:   0.373 ms | max:  38.247 ms | sum:249174.502 ms |
  2 cyclictest:(5)        |  12688.939 ms |   665144 | avg:   0.374 ms | max:  33.548 ms | sum:248509.392 ms |
  3 cyclictest:(5)        |  13475.623 ms |   669110 | avg:   0.370 ms | max:  37.819 ms | sum:247673.390 ms |

As per the numbers the, this makes cyclictest (short slice) it's
max-delay more consistent and consistency drops the sum-delay. The
trade-off is that the massive_intr (long slice) gets more context
switches and a slight increase in sum-delay.

Chunxin contributed did_preempt_short() where a task that lost slice
protection from PREEMPT_SHORT gets rescheduled once it becomes
in-eligible.

[mike: numbers]
Co-Developed-by: Chunxin Zang <zangchunxin@lixiang.com>
Signed-off-by: Chunxin Zang <zangchunxin@lixiang.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Tested-by: Mike Galbraith <umgwanakikbuti@gmail.com>
Link: https://lkml.kernel.org/r/20240727105030.735459544@infradead.org

During OSPM24 Youssef noted that migrations are re-setting the virtual
deadline. Notably everything that does a dequeue-enqueue, like setting
nice, changing preferred numa-node, and a myriad of other random crap,
will cause this to happen.

This shouldn't be. Preserve the relative virtual deadline across such
dequeue/enqueue cycles.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105030.625119246@infradead.org

Note that tasks that are kept on the runqueue to burn off negative
lag, are not in fact runnable anymore, they'll get dequeued the moment
they get picked.

As such, don't count this time towards runnable.

Thanks to Valentin for spotting I had this backwards initially.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105030.514088302@infradead.org

'Extend' DELAY_DEQUEUE by noting that since we wanted to dequeued them
at the 0-lag point, truncate lag (eg. don't let them earn positive
lag).
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105030.403750550@infradead.org

Extend / fix 86bfbb7c ("sched/fair: Add lag based placement") by
noting that lag is fundamentally a temporal measure. It should not be
carried around indefinitely.

OTOH it should also not be instantly discarded, doing so will allow a
task to game the system by purposefully (micro) sleeping at the end of
its time quantum.

Since lag is intimately tied to the virtual time base, a wall-time
based decay is also insufficient, notably competition is required for
any of this to make sense.

Instead, delay the dequeue and keep the 'tasks' on the runqueue,
competing until they are eligible.

Strictly speaking, we only care about keeping them until the 0-lag
point, but that is a difficult proposition, instead carry them around
until they get picked again, and dequeue them at that point.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105030.226163742@infradead.org

Since special task states must not suffer spurious wakeups, and the
proposed delayed dequeue can cause exactly these (under some boundary
conditions), propagate this knowledge into dequeue_task() such that it
can do the right thing.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105030.110439521@infradead.org

The special task states are those that do not suffer spurious wakeups,
TASK_FROZEN is very much one of those, mark it as such.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105029.998329901@infradead.org

Doing a wakeup on a delayed dequeue task is about as simple as it
sounds -- remove the delayed mark and enjoy the fact it was actually
still on the runqueue.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105029.888107381@infradead.org

Delayed dequeue's natural end is when it gets picked again. Ensure
pick_next_task() knows what to do with delayed tasks.

Note, this relies on the earlier patch that made pick_next_task()
state invariant -- it will restart the pick on dequeue, because
obviously the just dequeued task is no longer eligible.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105029.747330118@infradead.org

When dequeue_task() is delayed it becomes possible to exit a task (or
cgroup) that is still enqueued. Ensure things are dequeued before
freeing.

Thanks to Valentin for asking the obvious questions and making
switched_from_fair() less weird.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105029.631948434@infradead.org

Just a little sanity test..
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Tested-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lkml.kernel.org/r/20240727105029.486423066@infradead.org