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, 98 insertions, 0 deletions

diff --git a/drivers/crypto/caam/debugfs.c b/drivers/crypto/caam/debugfs.c
new file mode 100644
index 000000000..718352b7a
--- /dev/null
+++ b/drivers/crypto/caam/debugfs.c
@@ -0,0 +1,98 @@
+// SPDX-License-Identifier: (GPL-2.0+ OR BSD-3-Clause)
+/* Copyright 2019, 2023 NXP */
+
+#include <linux/debugfs.h>
+#include "compat.h"
+#include "debugfs.h"
+#include "regs.h"
+#include "intern.h"
+
+static int caam_debugfs_u64_get(void *data, u64 *val)
+{
+	*val = caam64_to_cpu(*(u64 *)data);
+	return 0;
+}
+
+static int caam_debugfs_u32_get(void *data, u64 *val)
+{
+	*val = caam32_to_cpu(*(u32 *)data);
+	return 0;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(caam_fops_u32_ro, caam_debugfs_u32_get, NULL, "%llu\n");
+DEFINE_DEBUGFS_ATTRIBUTE(caam_fops_u64_ro, caam_debugfs_u64_get, NULL, "%llu\n");
+
+#ifdef CONFIG_CRYPTO_DEV_FSL_CAAM_CRYPTO_API_QI
+/*
+ * This is a counter for the number of times the congestion group (where all
+ * the request and response queueus are) reached congestion. Incremented
+ * each time the congestion callback is called with congested == true.
+ */
+static u64 times_congested;
+
+void caam_debugfs_qi_congested(void)
+{
+	times_congested++;
+}
+
+void caam_debugfs_qi_init(struct caam_drv_private *ctrlpriv)
+{
+	debugfs_create_file("qi_congested", 0444, ctrlpriv->ctl,
+			    &times_congested, &caam_fops_u64_ro);
+}
+#endif
+
+void caam_debugfs_init(struct caam_drv_private *ctrlpriv,
+		       struct caam_perfmon __force *perfmon,
+		       struct dentry *root)
+{
+	/*
+	 * FIXME: needs better naming distinction, as some amalgamation of
+	 * "caam" and nprop->full_name. The OF name isn't distinctive,
+	 * but does separate instances
+	 */
+
+	ctrlpriv->ctl = debugfs_create_dir("ctl", root);
+
+	debugfs_create_file("rq_dequeued", 0444, ctrlpriv->ctl,
+			    &perfmon->req_dequeued, &caam_fops_u64_ro);
+	debugfs_create_file("ob_rq_encrypted", 0444, ctrlpriv->ctl,
+			    &perfmon->ob_enc_req, &caam_fops_u64_ro);
+	debugfs_create_file("ib_rq_decrypted", 0444, ctrlpriv->ctl,
+			    &perfmon->ib_dec_req, &caam_fops_u64_ro);
+	debugfs_create_file("ob_bytes_encrypted", 0444, ctrlpriv->ctl,
+			    &perfmon->ob_enc_bytes, &caam_fops_u64_ro);
+	debugfs_create_file("ob_bytes_protected", 0444, ctrlpriv->ctl,
+			    &perfmon->ob_prot_bytes, &caam_fops_u64_ro);
+	debugfs_create_file("ib_bytes_decrypted", 0444, ctrlpriv->ctl,
+			    &perfmon->ib_dec_bytes, &caam_fops_u64_ro);
+	debugfs_create_file("ib_bytes_validated", 0444, ctrlpriv->ctl,
+			    &perfmon->ib_valid_bytes, &caam_fops_u64_ro);
+
+	/* Controller level - global status values */
+	debugfs_create_file("fault_addr", 0444, ctrlpriv->ctl,
+			    &perfmon->faultaddr, &caam_fops_u32_ro);
+	debugfs_create_file("fault_detail", 0444, ctrlpriv->ctl,
+			    &perfmon->faultdetail, &caam_fops_u32_ro);
+	debugfs_create_file("fault_status", 0444, ctrlpriv->ctl,
+			    &perfmon->status, &caam_fops_u32_ro);
+
+	if (ctrlpriv->optee_en)
+		return;
+
+	/* Internal covering keys (useful in non-secure mode only) */
+	ctrlpriv->ctl_kek_wrap.data = (__force void *)&ctrlpriv->ctrl->kek[0];
+	ctrlpriv->ctl_kek_wrap.size = KEK_KEY_SIZE * sizeof(u32);
+	debugfs_create_blob("kek", 0444, ctrlpriv->ctl,
+			    &ctrlpriv->ctl_kek_wrap);
+
+	ctrlpriv->ctl_tkek_wrap.data = (__force void *)&ctrlpriv->ctrl->tkek[0];
+	ctrlpriv->ctl_tkek_wrap.size = KEK_KEY_SIZE * sizeof(u32);
+	debugfs_create_blob("tkek", 0444, ctrlpriv->ctl,
+			    &ctrlpriv->ctl_tkek_wrap);
+
+	ctrlpriv->ctl_tdsk_wrap.data = (__force void *)&ctrlpriv->ctrl->tdsk[0];
+	ctrlpriv->ctl_tdsk_wrap.size = KEK_KEY_SIZE * sizeof(u32);
+	debugfs_create_blob("tdsk", 0444, ctrlpriv->ctl,
+			    &ctrlpriv->ctl_tdsk_wrap);
+}