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, 91 insertions, 0 deletions

diff --git a/crypto/cipher.c b/crypto/cipher.c
new file mode 100644
index 000000000..c9dab656a
--- /dev/null
+++ b/crypto/cipher.c
@@ -0,0 +1,91 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Cryptographic API.
+ *
+ * Single-block cipher operations.
+ *
+ * Copyright (c) 2002 James Morris <jmorris@intercode.com.au>
+ * Copyright (c) 2005 Herbert Xu <herbert@gondor.apana.org.au>
+ */
+
+#include <crypto/algapi.h>
+#include <crypto/internal/cipher.h>
+#include <linux/kernel.h>
+#include <linux/crypto.h>
+#include <linux/errno.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+#include "internal.h"
+
+static int setkey_unaligned(struct crypto_cipher *tfm, const u8 *key,
+			    unsigned int keylen)
+{
+	struct cipher_alg *cia = crypto_cipher_alg(tfm);
+	unsigned long alignmask = crypto_cipher_alignmask(tfm);
+	int ret;
+	u8 *buffer, *alignbuffer;
+	unsigned long absize;
+
+	absize = keylen + alignmask;
+	buffer = kmalloc(absize, GFP_ATOMIC);
+	if (!buffer)
+		return -ENOMEM;
+
+	alignbuffer = (u8 *)ALIGN((unsigned long)buffer, alignmask + 1);
+	memcpy(alignbuffer, key, keylen);
+	ret = cia->cia_setkey(crypto_cipher_tfm(tfm), alignbuffer, keylen);
+	kfree_sensitive(buffer);
+	return ret;
+
+}
+
+int crypto_cipher_setkey(struct crypto_cipher *tfm,
+			 const u8 *key, unsigned int keylen)
+{
+	struct cipher_alg *cia = crypto_cipher_alg(tfm);
+	unsigned long alignmask = crypto_cipher_alignmask(tfm);
+
+	if (keylen < cia->cia_min_keysize || keylen > cia->cia_max_keysize)
+		return -EINVAL;
+
+	if ((unsigned long)key & alignmask)
+		return setkey_unaligned(tfm, key, keylen);
+
+	return cia->cia_setkey(crypto_cipher_tfm(tfm), key, keylen);
+}
+EXPORT_SYMBOL_NS_GPL(crypto_cipher_setkey, "CRYPTO_INTERNAL");
+
+static inline void cipher_crypt_one(struct crypto_cipher *tfm,
+				    u8 *dst, const u8 *src, bool enc)
+{
+	unsigned long alignmask = crypto_cipher_alignmask(tfm);
+	struct cipher_alg *cia = crypto_cipher_alg(tfm);
+	void (*fn)(struct crypto_tfm *, u8 *, const u8 *) =
+		enc ? cia->cia_encrypt : cia->cia_decrypt;
+
+	if (unlikely(((unsigned long)dst | (unsigned long)src) & alignmask)) {
+		unsigned int bs = crypto_cipher_blocksize(tfm);
+		u8 buffer[MAX_CIPHER_BLOCKSIZE + MAX_CIPHER_ALIGNMASK];
+		u8 *tmp = (u8 *)ALIGN((unsigned long)buffer, alignmask + 1);
+
+		memcpy(tmp, src, bs);
+		fn(crypto_cipher_tfm(tfm), tmp, tmp);
+		memcpy(dst, tmp, bs);
+	} else {
+		fn(crypto_cipher_tfm(tfm), dst, src);
+	}
+}
+
+void crypto_cipher_encrypt_one(struct crypto_cipher *tfm,
+			       u8 *dst, const u8 *src)
+{
+	cipher_crypt_one(tfm, dst, src, true);
+}
+EXPORT_SYMBOL_NS_GPL(crypto_cipher_encrypt_one, "CRYPTO_INTERNAL");
+
+void crypto_cipher_decrypt_one(struct crypto_cipher *tfm,
+			       u8 *dst, const u8 *src)
+{
+	cipher_crypt_one(tfm, dst, src, false);
+}
+EXPORT_SYMBOL_NS_GPL(crypto_cipher_decrypt_one, "CRYPTO_INTERNAL");