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, 150 insertions, 0 deletions

diff --git a/tools/perf/util/expr.l b/tools/perf/util/expr.l
new file mode 100644
index 000000000..f16ccff27
--- /dev/null
+++ b/tools/perf/util/expr.l
@@ -0,0 +1,150 @@
+%option prefix="expr_"
+%option reentrant
+%option bison-bridge
+
+%{
+#include <linux/compiler.h>
+#include "expr.h"
+#include "expr-bison.h"
+#include <math.h>
+
+char *expr_get_text(yyscan_t yyscanner);
+YYSTYPE *expr_get_lval(yyscan_t yyscanner);
+
+static double __value(YYSTYPE *yylval, char *str, int token)
+{
+	double num;
+
+	errno = 0;
+	num = strtod(str, NULL);
+	if (errno)
+		return EXPR_ERROR;
+
+	yylval->num = num;
+	return token;
+}
+
+static int value(yyscan_t scanner)
+{
+	YYSTYPE *yylval = expr_get_lval(scanner);
+	char *text = expr_get_text(scanner);
+
+	return __value(yylval, text, NUMBER);
+}
+
+/*
+ * Allow @ instead of / to be able to specify pmu/event/ without
+ * conflicts with normal division.
+ */
+static char *normalize(char *str, int runtime)
+{
+	char *ret = str;
+	char *dst = str;
+
+	while (*str) {
+		if (*str == '\\') {
+			*dst++ = *++str;
+			if (!*str)
+				break;
+		}
+		else if (*str == '?') {
+			char *paramval;
+			int i = 0;
+			int size = asprintf(&paramval, "%d", runtime);
+
+			if (size < 0)
+				*dst++ = '0';
+			else {
+				while (i < size)
+					*dst++ = paramval[i++];
+				free(paramval);
+			}
+		}
+		else
+			*dst++ = *str;
+		str++;
+	}
+
+	*dst = 0x0;
+	return ret;
+}
+
+static int str(yyscan_t scanner, int token, int runtime)
+{
+	YYSTYPE *yylval = expr_get_lval(scanner);
+	char *text = expr_get_text(scanner);
+
+	yylval->str = normalize(strdup(text), runtime);
+	if (!yylval->str)
+		return EXPR_ERROR;
+
+	yylval->str = normalize(yylval->str, runtime);
+	return token;
+}
+
+static int literal(yyscan_t scanner, const struct expr_scanner_ctx *sctx)
+{
+	YYSTYPE *yylval = expr_get_lval(scanner);
+
+	yylval->num = expr__get_literal(expr_get_text(scanner), sctx);
+	if (isnan(yylval->num)) {
+		if (!sctx->is_test)
+			return EXPR_ERROR;
+		yylval->num = 1;
+	}
+	return LITERAL;
+}
+
+static int nan_value(yyscan_t scanner)
+{
+	YYSTYPE *yylval = expr_get_lval(scanner);
+
+	yylval->num = NAN;
+	return NUMBER;
+}
+%}
+
+number		([0-9]+\.?[0-9]*|[0-9]*\.?[0-9]+)(e-?[0-9]+)?
+
+sch		[-,=]
+spec		\\{sch}
+sym		[0-9a-zA-Z_\.:@?]+
+symbol		({spec}|{sym})+
+literal		#[0-9a-zA-Z_\.\-]+
+
+%%
+	struct expr_scanner_ctx *sctx = expr_get_extra(yyscanner);
+
+d_ratio		{ return D_RATIO; }
+max		{ return MAX; }
+min		{ return MIN; }
+if		{ return IF; }
+else		{ return ELSE; }
+source_count	{ return SOURCE_COUNT; }
+aggr_nr		{ return AGGR_NR; }
+has_event	{ return HAS_EVENT; }
+strcmp_cpuid_str	{ return STRCMP_CPUID_STR; }
+NaN		{ return nan_value(yyscanner); }
+{literal}	{ return literal(yyscanner, sctx); }
+{number}	{ return value(yyscanner); }
+{symbol}	{ return str(yyscanner, ID, sctx->runtime); }
+"|"		{ return '|'; }
+"^"		{ return '^'; }
+"&"		{ return '&'; }
+"<"		{ return '<'; }
+">"		{ return '>'; }
+"-"		{ return '-'; }
+"+"		{ return '+'; }
+"*"		{ return '*'; }
+"/"		{ return '/'; }
+"%"		{ return '%'; }
+"("		{ return '('; }
+")"		{ return ')'; }
+","		{ return ','; }
+.		{ }
+%%
+
+int expr_wrap(void *scanner __maybe_unused)
+{
+	return 1;
+}