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, 168 insertions, 0 deletions

diff --git a/drivers/edac/edac_module.c b/drivers/edac/edac_module.c
new file mode 100644
index 000000000..1c9f62382
--- /dev/null
+++ b/drivers/edac/edac_module.c
@@ -0,0 +1,168 @@
+/*
+ * edac_module.c
+ *
+ * (C) 2007 www.softwarebitmaker.com
+ *
+ * This file is licensed under the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ *
+ * Author: Doug Thompson <dougthompson@xmission.com>
+ *
+ */
+#include <linux/edac.h>
+
+#include "edac_mc.h"
+#include "edac_module.h"
+
+#define EDAC_VERSION "Ver: 3.0.0"
+
+#ifdef CONFIG_EDAC_DEBUG
+
+static int edac_set_debug_level(const char *buf,
+				const struct kernel_param *kp)
+{
+	unsigned long val;
+	int ret;
+
+	ret = kstrtoul(buf, 0, &val);
+	if (ret)
+		return ret;
+
+	if (val > 4)
+		return -EINVAL;
+
+	return param_set_int(buf, kp);
+}
+
+/* Values of 0 to 4 will generate output */
+int edac_debug_level = 2;
+EXPORT_SYMBOL_GPL(edac_debug_level);
+
+module_param_call(edac_debug_level, edac_set_debug_level, param_get_int,
+		  &edac_debug_level, 0644);
+MODULE_PARM_DESC(edac_debug_level, "EDAC debug level: [0-4], default: 2");
+#endif
+
+/*
+ * edac_op_state_to_string()
+ */
+char *edac_op_state_to_string(int opstate)
+{
+	if (opstate == OP_RUNNING_POLL)
+		return "POLLED";
+	else if (opstate == OP_RUNNING_INTERRUPT)
+		return "INTERRUPT";
+	else if (opstate == OP_RUNNING_POLL_INTR)
+		return "POLL-INTR";
+	else if (opstate == OP_ALLOC)
+		return "ALLOC";
+	else if (opstate == OP_OFFLINE)
+		return "OFFLINE";
+
+	return "UNKNOWN";
+}
+
+/*
+ * sysfs object: /sys/devices/system/edac
+ *	need to export to other files
+ */
+static const struct bus_type edac_subsys = {
+	.name = "edac",
+	.dev_name = "edac",
+};
+
+static int edac_subsys_init(void)
+{
+	int err;
+
+	/* create the /sys/devices/system/edac directory */
+	err = subsys_system_register(&edac_subsys, NULL);
+	if (err)
+		printk(KERN_ERR "Error registering toplevel EDAC sysfs dir\n");
+
+	return err;
+}
+
+static void edac_subsys_exit(void)
+{
+	bus_unregister(&edac_subsys);
+}
+
+/* return pointer to the 'edac' node in sysfs */
+const struct bus_type *edac_get_sysfs_subsys(void)
+{
+	return &edac_subsys;
+}
+EXPORT_SYMBOL_GPL(edac_get_sysfs_subsys);
+/*
+ * edac_init
+ *      module initialization entry point
+ */
+static int __init edac_init(void)
+{
+	int err = 0;
+
+	edac_printk(KERN_INFO, EDAC_MC, EDAC_VERSION "\n");
+
+	err = edac_subsys_init();
+	if (err)
+		return err;
+
+	/*
+	 * Harvest and clear any boot/initialization PCI parity errors
+	 *
+	 * FIXME: This only clears errors logged by devices present at time of
+	 *      module initialization.  We should also do an initial clear
+	 *      of each newly hotplugged device.
+	 */
+	edac_pci_clear_parity_errors();
+
+	err = edac_mc_sysfs_init();
+	if (err)
+		goto err_sysfs;
+
+	edac_debugfs_init();
+
+	err = edac_workqueue_setup();
+	if (err) {
+		edac_printk(KERN_ERR, EDAC_MC, "Failure initializing workqueue\n");
+		goto err_wq;
+	}
+
+	return 0;
+
+err_wq:
+	edac_debugfs_exit();
+	edac_mc_sysfs_exit();
+
+err_sysfs:
+	edac_subsys_exit();
+
+	return err;
+}
+
+/*
+ * edac_exit()
+ *      module exit/termination function
+ */
+static void __exit edac_exit(void)
+{
+	edac_dbg(0, "\n");
+
+	/* tear down the various subsystems */
+	edac_workqueue_teardown();
+	edac_mc_sysfs_exit();
+	edac_debugfs_exit();
+	edac_subsys_exit();
+}
+
+/*
+ * Inform the kernel of our entry and exit points
+ */
+subsys_initcall(edac_init);
+module_exit(edac_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Doug Thompson www.softwarebitmaker.com, et al");
+MODULE_DESCRIPTION("Core library routines for EDAC reporting");