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, 119 insertions, 0 deletions

diff --git a/tools/perf/util/top.c b/tools/perf/util/top.c
new file mode 100644
index 000000000..851a26be6
--- /dev/null
+++ b/tools/perf/util/top.c
@@ -0,0 +1,119 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2011, Red Hat Inc, Arnaldo Carvalho de Melo <acme@redhat.com>
+ *
+ * Refactored from builtin-top.c, see that files for further copyright notes.
+ */
+
+#include "event.h"
+#include "evlist.h"
+#include "evsel.h"
+#include "parse-events.h"
+#include "symbol.h"
+#include "top.h"
+#include "util.h"
+#include <inttypes.h>
+
+#define SNPRINTF(buf, size, fmt, args...) \
+({ \
+	size_t r = snprintf(buf, size, fmt, ## args); \
+	r > size ?  size : r; \
+})
+
+size_t perf_top__header_snprintf(struct perf_top *top, char *bf, size_t size)
+{
+	float samples_per_sec;
+	float ksamples_per_sec;
+	float esamples_percent;
+	struct record_opts *opts = &top->record_opts;
+	struct target *target = &opts->target;
+	size_t ret = 0;
+	int nr_cpus;
+
+	if (top->samples) {
+		samples_per_sec = top->samples / top->delay_secs;
+		ksamples_per_sec = top->kernel_samples / top->delay_secs;
+		esamples_percent = (100.0 * top->exact_samples) / top->samples;
+	} else {
+		samples_per_sec = ksamples_per_sec = esamples_percent = 0.0;
+	}
+
+	if (!perf_guest) {
+		float ksamples_percent = 0.0;
+
+		if (samples_per_sec)
+			ksamples_percent = (100.0 * ksamples_per_sec) /
+							samples_per_sec;
+		ret = SNPRINTF(bf, size,
+			       "   PerfTop:%8.0f irqs/sec  kernel:%4.1f%%"
+			       "  exact: %4.1f%% lost: %" PRIu64 "/%" PRIu64 " drop: %" PRIu64 "/%" PRIu64 " [",
+			       samples_per_sec, ksamples_percent, esamples_percent,
+			       top->lost, top->lost_total, top->drop, top->drop_total);
+	} else {
+		float us_samples_per_sec = top->us_samples / top->delay_secs;
+		float guest_kernel_samples_per_sec = top->guest_kernel_samples / top->delay_secs;
+		float guest_us_samples_per_sec = top->guest_us_samples / top->delay_secs;
+
+		ret = SNPRINTF(bf, size,
+			       "   PerfTop:%8.0f irqs/sec  kernel:%4.1f%% us:%4.1f%%"
+			       " guest kernel:%4.1f%% guest us:%4.1f%%"
+			       " exact: %4.1f%% [", samples_per_sec,
+			       100.0 - (100.0 * ((samples_per_sec - ksamples_per_sec) /
+						 samples_per_sec)),
+			       100.0 - (100.0 * ((samples_per_sec - us_samples_per_sec) /
+						 samples_per_sec)),
+			       100.0 - (100.0 * ((samples_per_sec -
+						  guest_kernel_samples_per_sec) /
+						 samples_per_sec)),
+			       100.0 - (100.0 * ((samples_per_sec -
+						  guest_us_samples_per_sec) /
+						 samples_per_sec)),
+			       esamples_percent);
+	}
+
+	if (evlist__nr_entries(top->evlist) == 1) {
+		struct evsel *first = evlist__first(top->evlist);
+		ret += SNPRINTF(bf + ret, size - ret, "%" PRIu64 "%s ",
+				(uint64_t)first->core.attr.sample_period,
+				opts->freq ? "Hz" : "");
+	}
+
+	ret += SNPRINTF(bf + ret, size - ret, "%s", evsel__name(top->sym_evsel));
+
+	ret += SNPRINTF(bf + ret, size - ret, "], ");
+
+	if (target->pid)
+		ret += SNPRINTF(bf + ret, size - ret, " (target_pid: %s",
+				target->pid);
+	else if (target->tid)
+		ret += SNPRINTF(bf + ret, size - ret, " (target_tid: %s",
+				target->tid);
+	else if (top->uid_str != NULL)
+		ret += SNPRINTF(bf + ret, size - ret, " (uid: %s",
+				top->uid_str);
+	else
+		ret += SNPRINTF(bf + ret, size - ret, " (all");
+
+	nr_cpus = perf_cpu_map__nr(evlist__core(top->evlist)->user_requested_cpus);
+	if (target->cpu_list)
+		ret += SNPRINTF(bf + ret, size - ret, ", CPU%s: %s)",
+				nr_cpus > 1 ? "s" : "",
+				target->cpu_list);
+	else {
+		if (target->tid)
+			ret += SNPRINTF(bf + ret, size - ret, ")");
+		else
+			ret += SNPRINTF(bf + ret, size - ret, ", %d CPU%s)",
+					nr_cpus, nr_cpus > 1 ? "s" : "");
+	}
+
+	perf_top__reset_sample_counters(top);
+	return ret;
+}
+
+void perf_top__reset_sample_counters(struct perf_top *top)
+{
+	top->samples = top->us_samples = top->kernel_samples =
+	top->exact_samples = top->guest_kernel_samples =
+	top->guest_us_samples = top->lost = top->drop = 0;
+}