From 029b6ce733712a41421955194b113f283dcb1026 Mon Sep 17 00:00:00 2001 From: Changwoo Min Date: Sun, 2 Feb 2025 12:37:48 +0900 Subject: [PATCH 1/3] sched_ext: Fix incorrect time delta calculation in time_delta() When (s64)(after - before) > 0, the code returns the result of (s64)(after - before) > 0 while the intended result should be (s64)(after - before). That happens because the middle operand of the ternary operator was omitted incorrectly, returning the result of (s64)(after - before) > 0. Thus, add the middle operand -- (s64)(after - before) -- to return the correct time calculation. Fixes: d07be814fc71 ("sched_ext: Add time helpers for BPF schedulers") Signed-off-by: Changwoo Min Acked-by: Andrea Righi Signed-off-by: Tejun Heo --- tools/sched_ext/include/scx/common.bpf.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/sched_ext/include/scx/common.bpf.h b/tools/sched_ext/include/scx/common.bpf.h index f254a39b86a5..d72b60a0c582 100644 --- a/tools/sched_ext/include/scx/common.bpf.h +++ b/tools/sched_ext/include/scx/common.bpf.h @@ -432,7 +432,7 @@ void bpf_rcu_read_unlock(void) __ksym; */ static inline s64 time_delta(u64 after, u64 before) { - return (s64)(after - before) > 0 ? : 0; + return (s64)(after - before) > 0 ? (s64)(after - before) : 0; } /** From 2fa0fbeb69edd367b7c44f484e8dc5a5a1a311ef Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 7 Feb 2025 10:58:23 -1000 Subject: [PATCH 2/3] sched_ext: Implement auto local dispatching of migration disabled tasks Migration disabled tasks are special and pinned to their previous CPUs. They tripped up some unsuspecting BPF schedulers as their ->nr_cpus_allowed may not agree with the bits set in ->cpus_ptr. Make it easier for BPF schedulers by automatically dispatching them to the pinned local DSQs by default. If a BPF scheduler wants to handle migration disabled tasks explicitly, it can set SCX_OPS_ENQ_MIGRATION_DISABLED. Signed-off-by: Tejun Heo Acked-by: Andrea Righi --- kernel/sched/ext.c | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index a6d6d6dadde5..efdbf4d85a21 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -122,6 +122,19 @@ enum scx_ops_flags { */ SCX_OPS_SWITCH_PARTIAL = 1LLU << 3, + /* + * A migration disabled task can only execute on its current CPU. By + * default, such tasks are automatically put on the CPU's local DSQ with + * the default slice on enqueue. If this ops flag is set, they also go + * through ops.enqueue(). + * + * A migration disabled task never invokes ops.select_cpu() as it can + * only select the current CPU. Also, p->cpus_ptr will only contain its + * current CPU while p->nr_cpus_allowed keeps tracking p->user_cpus_ptr + * and thus may disagree with cpumask_weight(p->cpus_ptr). + */ + SCX_OPS_ENQ_MIGRATION_DISABLED = 1LLU << 4, + /* * CPU cgroup support flags */ @@ -130,6 +143,7 @@ enum scx_ops_flags { SCX_OPS_ALL_FLAGS = SCX_OPS_KEEP_BUILTIN_IDLE | SCX_OPS_ENQ_LAST | SCX_OPS_ENQ_EXITING | + SCX_OPS_ENQ_MIGRATION_DISABLED | SCX_OPS_SWITCH_PARTIAL | SCX_OPS_HAS_CGROUP_WEIGHT, }; @@ -882,6 +896,7 @@ static bool scx_warned_zero_slice; static DEFINE_STATIC_KEY_FALSE(scx_ops_enq_last); static DEFINE_STATIC_KEY_FALSE(scx_ops_enq_exiting); +static DEFINE_STATIC_KEY_FALSE(scx_ops_enq_migration_disabled); static DEFINE_STATIC_KEY_FALSE(scx_ops_cpu_preempt); static DEFINE_STATIC_KEY_FALSE(scx_builtin_idle_enabled); @@ -2014,6 +2029,11 @@ static void do_enqueue_task(struct rq *rq, struct task_struct *p, u64 enq_flags, unlikely(p->flags & PF_EXITING)) goto local; + /* see %SCX_OPS_ENQ_MIGRATION_DISABLED */ + if (!static_branch_unlikely(&scx_ops_enq_migration_disabled) && + is_migration_disabled(p)) + goto local; + if (!SCX_HAS_OP(enqueue)) goto global; @@ -5052,6 +5072,7 @@ static void scx_ops_disable_workfn(struct kthread_work *work) static_branch_disable(&scx_has_op[i]); static_branch_disable(&scx_ops_enq_last); static_branch_disable(&scx_ops_enq_exiting); + static_branch_disable(&scx_ops_enq_migration_disabled); static_branch_disable(&scx_ops_cpu_preempt); static_branch_disable(&scx_builtin_idle_enabled); synchronize_rcu(); @@ -5661,6 +5682,8 @@ static int scx_ops_enable(struct sched_ext_ops *ops, struct bpf_link *link) if (ops->flags & SCX_OPS_ENQ_EXITING) static_branch_enable(&scx_ops_enq_exiting); + if (ops->flags & SCX_OPS_ENQ_MIGRATION_DISABLED) + static_branch_enable(&scx_ops_enq_migration_disabled); if (scx_ops.cpu_acquire || scx_ops.cpu_release) static_branch_enable(&scx_ops_cpu_preempt); From 32966821574cd2917bd60f2554f435fe527f4702 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 7 Feb 2025 10:59:06 -1000 Subject: [PATCH 3/3] sched_ext: Fix migration disabled handling in targeted dispatches A dispatch operation that can target a specific local DSQ - scx_bpf_dsq_move_to_local() or scx_bpf_dsq_move() - checks whether the task can be migrated to the target CPU using task_can_run_on_remote_rq(). If the task can't be migrated to the targeted CPU, it is bounced through a global DSQ. task_can_run_on_remote_rq() assumes that the task is on a CPU that's different from the targeted CPU but the callers doesn't uphold the assumption and may call the function when the task is already on the target CPU. When such task has migration disabled, task_can_run_on_remote_rq() ends up returning %false incorrectly unnecessarily bouncing the task to a global DSQ. Fix it by updating the callers to only call task_can_run_on_remote_rq() when the task is on a different CPU than the target CPU. As this is a bit subtle, for clarity and documentation: - Make task_can_run_on_remote_rq() trigger SCHED_WARN_ON() if the task is on the same CPU as the target CPU. - is_migration_disabled() test in task_can_run_on_remote_rq() cannot trigger if the task is on a different CPU than the target CPU as the preceding task_allowed_on_cpu() test should fail beforehand. Convert the test into SCHED_WARN_ON(). Signed-off-by: Tejun Heo Fixes: 4c30f5ce4f7a ("sched_ext: Implement scx_bpf_dispatch[_vtime]_from_dsq()") Fixes: 0366017e0973 ("sched_ext: Use task_can_run_on_remote_rq() test in dispatch_to_local_dsq()") Cc: stable@vger.kernel.org # v6.12+ --- kernel/sched/ext.c | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index efdbf4d85a21..e01144340d67 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -2333,12 +2333,16 @@ static void move_remote_task_to_local_dsq(struct task_struct *p, u64 enq_flags, * * - The BPF scheduler is bypassed while the rq is offline and we can always say * no to the BPF scheduler initiated migrations while offline. + * + * The caller must ensure that @p and @rq are on different CPUs. */ static bool task_can_run_on_remote_rq(struct task_struct *p, struct rq *rq, bool trigger_error) { int cpu = cpu_of(rq); + SCHED_WARN_ON(task_cpu(p) == cpu); + /* * We don't require the BPF scheduler to avoid dispatching to offline * CPUs mostly for convenience but also because CPUs can go offline @@ -2352,8 +2356,11 @@ static bool task_can_run_on_remote_rq(struct task_struct *p, struct rq *rq, return false; } - if (unlikely(is_migration_disabled(p))) - return false; + /* + * If @p has migration disabled, @p->cpus_ptr only contains its current + * CPU and the above task_allowed_on_cpu() test should have failed. + */ + SCHED_WARN_ON(is_migration_disabled(p)); if (!scx_rq_online(rq)) return false; @@ -2457,7 +2464,8 @@ static struct rq *move_task_between_dsqs(struct task_struct *p, u64 enq_flags, if (dst_dsq->id == SCX_DSQ_LOCAL) { dst_rq = container_of(dst_dsq, struct rq, scx.local_dsq); - if (!task_can_run_on_remote_rq(p, dst_rq, true)) { + if (src_rq != dst_rq && + unlikely(!task_can_run_on_remote_rq(p, dst_rq, true))) { dst_dsq = find_global_dsq(p); dst_rq = src_rq; } @@ -2611,7 +2619,8 @@ static void dispatch_to_local_dsq(struct rq *rq, struct scx_dispatch_q *dst_dsq, } #ifdef CONFIG_SMP - if (unlikely(!task_can_run_on_remote_rq(p, dst_rq, true))) { + if (src_rq != dst_rq && + unlikely(!task_can_run_on_remote_rq(p, dst_rq, true))) { dispatch_enqueue(find_global_dsq(p), p, enq_flags | SCX_ENQ_CLEAR_OPSS); return;