History

sched_ext: Factor out scx_rehome_task() and scx_punt_task()

Factor out scx_rehome_task() and scx_punt_task() from the sub-disable re-home loop and scx_fail_parent(). The upcoming cgroup migration re-homing also needs scx_rehome_task(). No functional changes.

Tejun Heo committed 4 months ago

Commit 0dc90ce1 parent 52478771 file · +56 −28Browse files

Showing the changes to kernel (1 of 1 file). Show all

Unified · Split

kernel/sched/ext/sub.c+56−28
@@ -758,6 +758,57 @@ void drain_descendants(struct scx_sched *sch) · 757 unmodified lines
758758 wait_event(scx_unlink_waitq, list_empty(&sch->children));
759759 }
760760
761+/**
762+ * scx_rehome_task - Move a task to a sched it has been initialized for
763+ * @to: sched taking over @p, @p's init on it already complete
764+ * @p: task to re-home
765+ *
766+ * Exit @p from its current sched and switch it over to @to, overriding the
767+ * state to %SCX_TASK_READY to account for the already completed init. A task
768+ * on a non-ext class, possible under an %SCX_OPS_SWITCH_PARTIAL root, stays
769+ * %READY and is enabled by switching_to_scx() if it switches over.
770+ */
771+static void scx_rehome_task(struct scx_sched *to, struct task_struct *p)
772+{
773+ lockdep_assert_held(&p->pi_lock);
774+ lockdep_assert_rq_held(task_rq(p));
775+
776+ scoped_guard (sched_change, p, DEQUEUE_SAVE | DEQUEUE_MOVE) {
777+ scx_disable_and_exit_task(scx_task_sched(p), p);
778+ scx_set_task_state(p, SCX_TASK_INIT_BEGIN);
779+ scx_set_task_state(p, SCX_TASK_INIT);
780+ scx_set_task_sched(p, to);
781+ scx_set_task_state(p, SCX_TASK_READY);
782+ if (p->sched_class == &ext_sched_class)
783+ scx_enable_task(to, p);
784+ }
785+}
786+
787+/**
788+ * scx_punt_task - Hand a task to a failed sched without initialization
789+ * @to: failed and bypassed sched taking custody of @p
790+ * @p: task to punt
791+ *
792+ * Take @p off its current sched and put it on @to at %SCX_TASK_NONE. @to is
793+ * dying and its teardown will re-home @p properly.
794+ *
795+ * Used when @to must take over @p but failed to initialize it. Bypass keeps
796+ * scheduling decisions away from @to but @p can still trigger its task ops,
797+ * which may confuse the BPF side. @to is dying anyway. The exit paths skip
798+ * %NONE tasks (see __scx_disable_and_exit_task() and switched_from_scx()).
799+ */
800+static void scx_punt_task(struct scx_sched *to, struct task_struct *p)
801+{
802+ lockdep_assert_held(&p->pi_lock);
803+ lockdep_assert_rq_held(task_rq(p));
804+ WARN_ON_ONCE(!READ_ONCE(to->bypass_depth));
805+
806+ scoped_guard (sched_change, p, DEQUEUE_SAVE | DEQUEUE_MOVE) {
807+ scx_disable_and_exit_task(scx_task_sched(p), p);
808+ scx_set_task_sched(p, to);
809+ }
810+}
811+
761812 static void scx_fail_parent(struct scx_sched *sch,
762813 struct task_struct *failed, s32 fail_code)
763814 {
@@ -769,9 +820,9 @@ static void scx_fail_parent(struct scx_sched *sch, · 5 unmodified lines
769820 fail_code, failed->comm, failed->pid);
770821
771822 /*
772− * Once $parent is bypassed, it's safe to put SCX_TASK_NONE tasks into
773− * it. This may cause downstream failures on the BPF side but $parent is
774− * dying anyway.
823+ * Once $parent is bypassed, tasks can be punted into it. This may
824+ * cause downstream failures on the BPF side but $parent is dying
825+ * anyway.
775826 */
776827 scx_bypass(parent, true);
777828
@@ -780,10 +831,7 @@ static void scx_fail_parent(struct scx_sched *sch, · 2 unmodified lines
780831 if (scx_task_on_sched(parent, p))
781832 continue;
782833
783− scoped_guard (sched_change, p, DEQUEUE_SAVE | DEQUEUE_MOVE) {
784− scx_disable_and_exit_task(sch, p);
785− scx_set_task_sched(p, parent);
786− }
834+ scx_punt_task(parent, p);
787835 }
788836 scx_task_iter_stop(&sti);
789837 }
@@ -881,27 +929,7 @@ void scx_sub_disable(struct scx_sched *sch) · 91 unmodified lines
881929 continue;
882930 }
883931
884− scoped_guard (sched_change, p, DEQUEUE_SAVE | DEQUEUE_MOVE) {
885− /*
886− * $p is initialized for $parent and still attached to
887− * @sch. Disable and exit for @sch, switch over to
888− * $parent and override the state to READY to account
889− * for $p having already been initialized.
890− */
891− scx_disable_and_exit_task(sch, p);
892− scx_set_task_state(p, SCX_TASK_INIT_BEGIN);
893− scx_set_task_state(p, SCX_TASK_INIT);
894− scx_set_task_sched(p, parent);
895− scx_set_task_state(p, SCX_TASK_READY);
896−
897− /*
898− * A task on a non-ext class, possible under an
899− * %SCX_OPS_SWITCH_PARTIAL root, stays READY and is
900− * enabled by switching_to_scx() if it switches over.
901− */
902− if (p->sched_class == &ext_sched_class)
903− scx_enable_task(parent, p);
904− }
932+ scx_rehome_task(parent, p);
905933
906934 task_rq_unlock(rq, p, &rf);
907935 put_task_struct(p);

Keyboard shortcuts

General

Show keyboard shortcuts
?
Focus the current search
/

Issues and pull requests

Open filters
f

Issues, pull requests and Inbox

Select next item
j↓
Select previous item
k↑
Open selected item
Enter
Dismiss filters or clear selection
Esc

Repository code

Go to file
t