From: Tejun Heo <tj@kernel.org>
To: void@manifault.com, arighi@nvidia.com, multics69@gmail.com
Cc: linux-kernel@vger.kernel.org, sched-ext@lists.linux.dev,
memxor@gmail.com, bpf@vger.kernel.org, Tejun Heo <tj@kernel.org>
Subject: [PATCH 33/46] sched_ext: Factor out scx_dispatch_sched()
Date: Fri, 19 Sep 2025 14:58:56 -1000 [thread overview]
Message-ID: <20250920005931.2753828-34-tj@kernel.org> (raw)
In-Reply-To: <20250920005931.2753828-1-tj@kernel.org>
In preparation of multiple scheduler support, factor out
scx_dispatch_sched() from balance_one(). The function boundary makes
remembering $prev_on_scx and $prev_on_rq less useful. Open code $prev_on_scx
in balance_one() and $prev_on_rq in both balance_one() and
scx_dispatch_sched().
No functional changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
---
kernel/sched/ext.c | 115 ++++++++++++++++++++++++---------------------
1 file changed, 62 insertions(+), 53 deletions(-)
diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
index 44f9cc7f0915..efe01ba84e2d 100644
--- a/kernel/sched/ext.c
+++ b/kernel/sched/ext.c
@@ -2178,14 +2178,68 @@ static void flush_dispatch_buf(struct scx_sched *sch, struct rq *rq)
dspc->cursor = 0;
}
-static int balance_one(struct rq *rq, struct task_struct *prev)
+static bool scx_dispatch_sched(struct scx_sched *sch, struct rq *rq,
+ struct task_struct *prev)
{
- struct scx_sched *sch = scx_root;
struct scx_dsp_ctx *dspc = this_cpu_ptr(scx_dsp_ctx);
bool prev_on_scx = prev->sched_class == &ext_sched_class;
- bool prev_on_rq = prev->scx.flags & SCX_TASK_QUEUED;
int nr_loops = SCX_DSP_MAX_LOOPS;
+ if (consume_global_dsq(sch, rq))
+ return true;
+
+ if (unlikely(!SCX_HAS_OP(sch, dispatch)) ||
+ scx_bypassing(sch, cpu_of(rq)) || !scx_rq_online(rq))
+ return false;
+
+ dspc->rq = rq;
+
+ /*
+ * The dispatch loop. Because flush_dispatch_buf() may drop the rq lock,
+ * the local DSQ might still end up empty after a successful
+ * ops.dispatch(). If the local DSQ is empty even after ops.dispatch()
+ * produced some tasks, retry. The BPF scheduler may depend on this
+ * looping behavior to simplify its implementation.
+ */
+ do {
+ dspc->nr_tasks = 0;
+
+ SCX_CALL_OP(sch, SCX_KF_DISPATCH, dispatch, rq,
+ cpu_of(rq), prev_on_scx ? prev : NULL);
+
+ flush_dispatch_buf(sch, rq);
+
+ if ((prev->scx.flags & SCX_TASK_QUEUED) && prev->scx.slice) {
+ rq->scx.flags |= SCX_RQ_BAL_KEEP;
+ return true;
+ }
+ if (rq->scx.local_dsq.nr)
+ return true;
+ if (consume_global_dsq(sch, rq))
+ return true;
+
+ /*
+ * ops.dispatch() can trap us in this loop by repeatedly
+ * dispatching ineligible tasks. Break out once in a while to
+ * allow the watchdog to run. As IRQ can't be enabled in
+ * balance(), we want to complete this scheduling cycle and then
+ * start a new one. IOW, we want to call resched_curr() on the
+ * next, most likely idle, task, not the current one. Use
+ * __scx_bpf_kick_cpu() for deferred kicking.
+ */
+ if (unlikely(!--nr_loops)) {
+ scx_kick_cpu(sch, cpu_of(rq), 0);
+ break;
+ }
+ } while (dspc->nr_tasks);
+
+ return false;
+}
+
+static int balance_one(struct rq *rq, struct task_struct *prev)
+{
+ struct scx_sched *sch = scx_root;
+
lockdep_assert_rq_held(rq);
rq->scx.flags |= SCX_RQ_IN_BALANCE;
rq->scx.flags &= ~(SCX_RQ_BAL_PENDING | SCX_RQ_BAL_KEEP);
@@ -2204,7 +2258,7 @@ static int balance_one(struct rq *rq, struct task_struct *prev)
rq->scx.cpu_released = false;
}
- if (prev_on_scx) {
+ if (prev->sched_class == &ext_sched_class) {
update_curr_scx(rq);
/*
@@ -2217,7 +2271,7 @@ static int balance_one(struct rq *rq, struct task_struct *prev)
* See scx_disable_workfn() for the explanation on the bypassing
* test.
*/
- if (prev_on_rq && prev->scx.slice &&
+ if ((prev->scx.flags & SCX_TASK_QUEUED) && prev->scx.slice &&
!scx_bypassing(sch, cpu_of(rq))) {
rq->scx.flags |= SCX_RQ_BAL_KEEP;
goto has_tasks;
@@ -2228,60 +2282,15 @@ static int balance_one(struct rq *rq, struct task_struct *prev)
if (rq->scx.local_dsq.nr)
goto has_tasks;
- if (consume_global_dsq(sch, rq))
+ /* dispatch @sch */
+ if (scx_dispatch_sched(sch, rq, prev))
goto has_tasks;
- if (unlikely(!SCX_HAS_OP(sch, dispatch)) ||
- scx_bypassing(sch, cpu_of(rq)) || !scx_rq_online(rq))
- goto no_tasks;
-
- dspc->rq = rq;
-
- /*
- * The dispatch loop. Because flush_dispatch_buf() may drop the rq lock,
- * the local DSQ might still end up empty after a successful
- * ops.dispatch(). If the local DSQ is empty even after ops.dispatch()
- * produced some tasks, retry. The BPF scheduler may depend on this
- * looping behavior to simplify its implementation.
- */
- do {
- dspc->nr_tasks = 0;
-
- SCX_CALL_OP(sch, SCX_KF_DISPATCH, dispatch, rq,
- cpu_of(rq), prev_on_scx ? prev : NULL);
-
- flush_dispatch_buf(sch, rq);
-
- if (prev_on_rq && prev->scx.slice) {
- rq->scx.flags |= SCX_RQ_BAL_KEEP;
- goto has_tasks;
- }
- if (rq->scx.local_dsq.nr)
- goto has_tasks;
- if (consume_global_dsq(sch, rq))
- goto has_tasks;
-
- /*
- * ops.dispatch() can trap us in this loop by repeatedly
- * dispatching ineligible tasks. Break out once in a while to
- * allow the watchdog to run. As IRQ can't be enabled in
- * balance(), we want to complete this scheduling cycle and then
- * start a new one. IOW, we want to call resched_curr() on the
- * next, most likely idle, task, not the current one. Use
- * scx_kick_cpu() for deferred kicking.
- */
- if (unlikely(!--nr_loops)) {
- scx_kick_cpu(sch, cpu_of(rq), 0);
- break;
- }
- } while (dspc->nr_tasks);
-
-no_tasks:
/*
* Didn't find another task to run. Keep running @prev unless
* %SCX_OPS_ENQ_LAST is in effect.
*/
- if (prev_on_rq &&
+ if ((prev->scx.flags & SCX_TASK_QUEUED) &&
(!(sch->ops.flags & SCX_OPS_ENQ_LAST) ||
scx_bypassing(sch, cpu_of(rq)))) {
rq->scx.flags |= SCX_RQ_BAL_KEEP;
--
2.51.0
next prev parent reply other threads:[~2025-09-20 1:00 UTC|newest]
Thread overview: 47+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-09-20 0:58 [PATCHSET RFC] sched_ext: Implement cgroup sub-scheduler support Tejun Heo
2025-09-20 0:58 ` [PATCH 01/46] sched_ext: Use rhashtable_lookup() instead of rhashtable_lookup_fast() Tejun Heo
2025-09-20 0:58 ` [PATCH 02/46] sched_ext: Improve SCX_KF_DISPATCH comment Tejun Heo
2025-09-20 0:58 ` [PATCH 03/46] sched_ext: Fix stray scx_root usage in task_can_run_on_remote_rq() Tejun Heo
2025-09-20 0:58 ` [PATCH 04/46] sched_ext: Use bitfields for boolean warning flags Tejun Heo
2025-09-20 0:58 ` [PATCH 05/46] sched_ext: Add SCX_EFLAG_INITIALIZED to indicate successful ops.init() Tejun Heo
2025-09-20 0:58 ` [PATCH 06/46] sched_ext: Make qmap dump operation non-destructive Tejun Heo
2025-09-20 0:58 ` [PATCH 07/46] tools/sched_ext: scx_qmap: Make debug output quieter by default Tejun Heo
2025-09-20 0:58 ` [PATCH 08/46] sched_ext: Separate out scx_kick_cpu() and add @sch to it Tejun Heo
2025-09-20 0:58 ` [PATCH 09/46] sched_ext: Add the @sch parameter to __bstr_format() Tejun Heo
2025-09-20 0:58 ` [PATCH 10/46] sched_ext: Add the @sch parameter to ext_idle helpers Tejun Heo
2025-09-20 0:58 ` [PATCH 11/46] sched_ext: Drop kf_cpu_valid() Tejun Heo
2025-09-20 0:58 ` [PATCH 12/46] sched_ext: Add the @sch parameter to scx_dsq_insert_preamble/commit() Tejun Heo
2025-09-20 0:58 ` [PATCH 13/46] sched_ext: Drop scx_kf_exit() and scx_kf_error() Tejun Heo
2025-09-20 0:58 ` [PATCH 14/46] sched_ext: Misc updates around scx_sched instance pointer Tejun Heo
2025-09-20 0:58 ` [PATCH 15/46] sched_ext: Keep dying tasks on a separate list Tejun Heo
2025-09-20 0:58 ` [PATCH 16/46] sched_ext: Implement cgroup subtree iteration for scx_task_iter Tejun Heo
2025-09-20 0:58 ` [PATCH 17/46] sched_ext: Add @kargs to scx_fork() Tejun Heo
2025-09-20 0:58 ` [PATCH 18/46] sched/core: Swap the order between sched_post_fork() and cgroup_post_fork() Tejun Heo
2025-09-20 0:58 ` [PATCH 19/46] cgroup: Expose some cgroup helpers Tejun Heo
2025-09-20 0:58 ` [PATCH 20/46] sched_ext: Update p->scx.disallow warning in scx_init_task() Tejun Heo
2025-09-20 0:58 ` [PATCH 21/46] sched_ext: Minor reorganization of enable/disable path Tejun Heo
2025-09-20 0:58 ` [PATCH 22/46] sched_ext: Factor out scx_claim_exit() from scx_disable() Tejun Heo
2025-09-20 0:58 ` [PATCH 23/46] sched_ext: Introduce cgroup sub-sched support Tejun Heo
2025-09-20 0:58 ` [PATCH 24/46] HACK_NOT_FOR_UPSTREAM: BPF: Implement prog grouping hack Tejun Heo
2025-09-20 0:58 ` [PATCH 25/46] sched_ext: Introduce scx_task_sched[_rcu]() Tejun Heo
2025-09-20 0:58 ` [PATCH 26/46] sched_ext: Introduce scx_prog_sched() Tejun Heo
2025-09-20 0:58 ` [PATCH 27/46] sched_ext: Ignore insertions of not-owned tasks into DSQs Tejun Heo
2025-09-20 0:58 ` [PATCH 28/46] sched_ext: scx_dsq_move() should validate the task belongs to the right scheduler Tejun Heo
2025-09-20 0:58 ` [PATCH 29/46] sched_ext: Refactor task init/exit helpers Tejun Heo
2025-09-20 0:58 ` [PATCH 30/46] sched_ext: Make scx_prio_less() handle multiple schedulers Tejun Heo
2025-09-20 0:58 ` [PATCH 31/46] sched_ext: Move bypass_depth into scx_sched Tejun Heo
2025-09-20 0:58 ` [PATCH 32/46] sched_ext: Make bypass mode sub-sched aware Tejun Heo
2025-09-20 0:58 ` Tejun Heo [this message]
2025-09-20 0:58 ` [PATCH 34/46] sched_ext: When calling ops.dispatch() @prev must be on the same scx_sched Tejun Heo
2025-09-20 0:58 ` [PATCH 35/46] sched_ext: Dispatch from all scx_sched instances Tejun Heo
2025-09-20 0:58 ` [PATCH 36/46] sched_ext: Move scx_dsp_ctx and scx_dsp_max_batch into scx_sched Tejun Heo
2025-09-20 0:59 ` [PATCH 37/46] sched_ext: Make watchdog sub-sched aware Tejun Heo
2025-09-20 0:59 ` [PATCH 38/46] sched_ext: Convert scx_dump_state() spinlock to raw spinlock Tejun Heo
2025-09-20 0:59 ` [PATCH 39/46] sched_ext: Support dumping multiple schedulers and add scheduler identification Tejun Heo
2025-09-20 0:59 ` [PATCH 40/46] sched_ext: Implement cgroup sub-sched enabling and disabling Tejun Heo
2025-09-20 0:59 ` [PATCH 41/46] HACK_NOT_FOR_UPSTREAM: sched_ext: Work around @aux__prog prototype mismatch Tejun Heo
2025-09-20 0:59 ` [PATCH 42/46] sched_ext: Wrap global DSQs in per-node structure Tejun Heo
2025-09-20 0:59 ` [PATCH 43/46] sched_ext: Add bypass DSQ for sub-schedulers Tejun Heo
2025-09-20 0:59 ` [PATCH 44/46] sched_ext: Factor out scx_link_sched() and scx_unlink_sched() Tejun Heo
2025-09-20 0:59 ` [PATCH 45/46] sched_ext: Add rhashtable lookup for sub-schedulers Tejun Heo
2025-09-20 0:59 ` [PATCH 46/46] sched_ext: Add basic building blocks for nested sub-scheduler dispatching Tejun Heo
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250920005931.2753828-34-tj@kernel.org \
--to=tj@kernel.org \
--cc=arighi@nvidia.com \
--cc=bpf@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=memxor@gmail.com \
--cc=multics69@gmail.com \
--cc=sched-ext@lists.linux.dev \
--cc=void@manifault.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox