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, 143 insertions, 0 deletions

diff --git a/net/sunrpc/auth_null.c b/net/sunrpc/auth_null.c
new file mode 100644
index 000000000..41a633a40
--- /dev/null
+++ b/net/sunrpc/auth_null.c
@@ -0,0 +1,143 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * linux/net/sunrpc/auth_null.c
+ *
+ * AUTH_NULL authentication. Really :-)
+ *
+ * Copyright (C) 1996, Olaf Kirch <okir@monad.swb.de>
+ */
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/sunrpc/clnt.h>
+
+#if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
+# define RPCDBG_FACILITY	RPCDBG_AUTH
+#endif
+
+static struct rpc_auth null_auth;
+static struct rpc_cred null_cred;
+
+static struct rpc_auth *
+nul_create(const struct rpc_auth_create_args *args, struct rpc_clnt *clnt)
+{
+	refcount_inc(&null_auth.au_count);
+	return &null_auth;
+}
+
+static void
+nul_destroy(struct rpc_auth *auth)
+{
+}
+
+/*
+ * Lookup NULL creds for current process
+ */
+static struct rpc_cred *
+nul_lookup_cred(struct rpc_auth *auth, struct auth_cred *acred, int flags)
+{
+	return get_rpccred(&null_cred);
+}
+
+/*
+ * Destroy cred handle.
+ */
+static void
+nul_destroy_cred(struct rpc_cred *cred)
+{
+}
+
+/*
+ * Match cred handle against current process
+ */
+static int
+nul_match(struct auth_cred *acred, struct rpc_cred *cred, int taskflags)
+{
+	return 1;
+}
+
+/*
+ * Marshal credential.
+ */
+static int
+nul_marshal(struct rpc_task *task, struct xdr_stream *xdr)
+{
+	__be32 *p;
+
+	p = xdr_reserve_space(xdr, 4 * sizeof(*p));
+	if (!p)
+		return -EMSGSIZE;
+	/* Credential */
+	*p++ = rpc_auth_null;
+	*p++ = xdr_zero;
+	/* Verifier */
+	*p++ = rpc_auth_null;
+	*p   = xdr_zero;
+	return 0;
+}
+
+/*
+ * Refresh credential. This is a no-op for AUTH_NULL
+ */
+static int
+nul_refresh(struct rpc_task *task)
+{
+	set_bit(RPCAUTH_CRED_UPTODATE, &task->tk_rqstp->rq_cred->cr_flags);
+	return 0;
+}
+
+static int
+nul_validate(struct rpc_task *task, struct xdr_stream *xdr)
+{
+	__be32 *p;
+
+	p = xdr_inline_decode(xdr, 2 * sizeof(*p));
+	if (!p)
+		return -EIO;
+	if (*p++ != rpc_auth_null)
+		return -EIO;
+	if (*p != xdr_zero)
+		return -EIO;
+	return 0;
+}
+
+const struct rpc_authops authnull_ops = {
+	.owner		= THIS_MODULE,
+	.au_flavor	= RPC_AUTH_NULL,
+	.au_name	= "NULL",
+	.create		= nul_create,
+	.destroy	= nul_destroy,
+	.lookup_cred	= nul_lookup_cred,
+};
+
+static
+struct rpc_auth null_auth = {
+	.au_cslack	= NUL_CALLSLACK,
+	.au_rslack	= NUL_REPLYSLACK,
+	.au_verfsize	= NUL_REPLYSLACK,
+	.au_ralign	= NUL_REPLYSLACK,
+	.au_ops		= &authnull_ops,
+	.au_flavor	= RPC_AUTH_NULL,
+	.au_count	= REFCOUNT_INIT(1),
+};
+
+static
+const struct rpc_credops null_credops = {
+	.cr_name	= "AUTH_NULL",
+	.crdestroy	= nul_destroy_cred,
+	.crmatch	= nul_match,
+	.crmarshal	= nul_marshal,
+	.crwrap_req	= rpcauth_wrap_req_encode,
+	.crrefresh	= nul_refresh,
+	.crvalidate	= nul_validate,
+	.crunwrap_resp	= rpcauth_unwrap_resp_decode,
+};
+
+static
+struct rpc_cred null_cred = {
+	.cr_lru		= LIST_HEAD_INIT(null_cred.cr_lru),
+	.cr_auth	= &null_auth,
+	.cr_ops		= &null_credops,
+	.cr_count	= REFCOUNT_INIT(2),
+	.cr_flags	= 1UL << RPCAUTH_CRED_UPTODATE,
+};