sched/cache: Keep nr_pref_llc_running in the runnable domain, to fix LLC mis-scheduling buggrafted

alb_break_llc() decides whether to break LLC preference during active
load balance. It does so by testing that every runnable fair task on the
source rq prefers its LLC:

	env->src_rq->nr_pref_llc_running == env->src_rq->cfs.h_nr_runnable

But the two counters cover different sets. nr_pref_llc_running is updated
in account_llc_enqueue()/account_llc_dequeue(), next to cfs_rq->nr_queued,
so it follows queued tasks. h_nr_runnable is updated in set_delayed()/
clear_delayed() and drops delay-dequeued tasks.

So under DELAY_DEQUEUE, a preferring task that goes to sleep stays counted
in nr_pref_llc_running while h_nr_runnable falls. The equality then breaks,
alb_break_llc() returns false, and active balance is free to pull a task
off its preferred LLC. Active balance only moves runnable tasks, and this
is the only LLC check it consults: once the stopper runs, LBF_ACTIVE_LB
skips the per-task test in can_migrate_task(). The runnable set is the one
we want.

Fix it on the counter side. A task should be counted in
nr_pref_llc_running exactly while it is both queued on its preferred LLC
(pref_llc_queued) and runnable (!sched_delayed). Define that membership
once in task_pref_llc_runnable(), and adjust the counter only through
pref_llc_running_inc()/pref_llc_running_dec() from the four sites that
change either input: account_llc_enqueue(), account_llc_dequeue(),
set_delayed() and clear_delayed(). Gating every update on the same
predicate keeps the delay, wake and dequeue paths from double-counting
or underflowing; see the comments at those sites for the ordering.

nr_llc_running and sd->llc_counts are not touched and stay on queued
semantics.

Fixes: 714059f79ff0 ("sched/cache: Handle moving single tasks to/from their preferred LLC")
Closes: https://lore.kernel.org/lkml/20260827135000.735138-1-zhanxusheng@xiaomi.com/
Reported-by: Zhan Xusheng <zhanxusheng@xiaomi.com>
Suggested-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Kayra Cizmeci <kayracizmeci@gmail.com>
Cc: <stable@kernel.org> # v7.2.x
Link: https://patch.msgid.link/06af61afedac32e6477f57feb4d658f6c411c3af.1790035273.git.tim.c.chen@linux.intel.com

1 files changed, 117 insertions, 0 deletions

diff --git a/io_uring/tw.h b/io_uring/tw.h
new file mode 100644
index 000000000..3ade5ad57
--- /dev/null
+++ b/io_uring/tw.h
@@ -0,0 +1,117 @@
+// SPDX-License-Identifier: GPL-2.0
+#ifndef IOU_TW_H
+#define IOU_TW_H
+
+#include <linux/sched.h>
+#include <linux/percpu-refcount.h>
+#include <linux/io_uring_types.h>
+
+#include "mpscq.h"
+
+#define IO_LOCAL_TW_DEFAULT_MAX		20
+
+/*
+ * Terminate the request if either of these conditions are true:
+ *
+ * 1) It's being executed by the original task, but that task is marked
+ *    with PF_EXITING as it's exiting.
+ * 2) PF_KTHREAD is set, in which case the invoker of the task_work is
+ *    our fallback task_work.
+ * 3) The ring has been closed and is going away.
+ */
+static inline bool io_should_terminate_tw(struct io_ring_ctx *ctx)
+{
+	return (current->flags & (PF_EXITING | PF_KTHREAD)) || percpu_ref_is_dying(&ctx->refs);
+}
+
+void io_req_task_work_add_remote(struct io_kiocb *req, unsigned flags);
+void tctx_task_work(struct callback_head *cb);
+void io_tctx_fallback_work(struct work_struct *work);
+int io_run_local_work(struct io_ring_ctx *ctx, int min_events, int max_events);
+int io_run_task_work_sig(struct io_ring_ctx *ctx);
+
+__cold void io_cancel_local_task_work(struct io_ring_ctx *ctx);
+int io_run_local_work_locked(struct io_ring_ctx *ctx, int min_events);
+
+void io_req_local_work_add(struct io_kiocb *req, unsigned flags);
+void io_req_normal_work_add(struct io_kiocb *req);
+void tctx_task_work_run(struct io_uring_task *tctx, unsigned int max_entries, unsigned int *count);
+
+static inline void __io_req_task_work_add(struct io_kiocb *req, unsigned flags)
+{
+	if (req->ctx->flags & IORING_SETUP_DEFER_TASKRUN)
+		io_req_local_work_add(req, flags);
+	else
+		io_req_normal_work_add(req);
+}
+
+static inline void io_req_task_work_add(struct io_kiocb *req)
+{
+	__io_req_task_work_add(req, 0);
+}
+
+static inline int io_run_task_work(void)
+{
+	bool ret = false;
+
+	/*
+	 * Always check-and-clear the task_work notification signal. With how
+	 * signaling works for task_work, we can find it set with nothing to
+	 * run. We need to clear it for that case, like get_signal() does.
+	 */
+	if (test_thread_flag(TIF_NOTIFY_SIGNAL))
+		clear_notify_signal();
+	/*
+	 * PF_IO_WORKER never returns to userspace, so check here if we have
+	 * notify work that needs processing.
+	 */
+	if (current->flags & PF_IO_WORKER) {
+		if (test_thread_flag(TIF_NOTIFY_RESUME)) {
+			__set_current_state(TASK_RUNNING);
+			resume_user_mode_work(NULL);
+		}
+		if (current->io_uring) {
+			unsigned int count = 0;
+
+			__set_current_state(TASK_RUNNING);
+			tctx_task_work_run(current->io_uring, UINT_MAX, &count);
+			if (count)
+				ret = true;
+		}
+	}
+	if (task_work_pending(current)) {
+		__set_current_state(TASK_RUNNING);
+		task_work_run();
+		ret = true;
+	}
+
+	return ret;
+}
+
+static inline bool io_local_work_pending(struct io_ring_ctx *ctx)
+{
+	return !mpscq_empty(&ctx->work_list);
+}
+
+static inline bool io_task_work_pending(struct io_ring_ctx *ctx)
+{
+	return task_work_pending(current) || io_local_work_pending(ctx);
+}
+
+static inline void io_tw_lock(struct io_ring_ctx *ctx, io_tw_token_t tw)
+{
+	lockdep_assert_held(&ctx->uring_lock);
+}
+
+static inline bool io_allowed_defer_tw_run(struct io_ring_ctx *ctx)
+{
+	return likely(ctx->submitter_task == current);
+}
+
+static inline bool io_allowed_run_tw(struct io_ring_ctx *ctx)
+{
+	return likely(!(ctx->flags & IORING_SETUP_DEFER_TASKRUN) ||
+		      ctx->submitter_task == current);
+}
+
+#endif