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, 112 insertions, 0 deletions

diff --git a/fs/cachefiles/security.c b/fs/cachefiles/security.c
new file mode 100644
index 000000000..fc6611886
--- /dev/null
+++ b/fs/cachefiles/security.c
@@ -0,0 +1,112 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* CacheFiles security management
+ *
+ * Copyright (C) 2007, 2021 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ */
+
+#include <linux/fs.h>
+#include <linux/cred.h>
+#include "internal.h"
+
+/*
+ * determine the security context within which we access the cache from within
+ * the kernel
+ */
+int cachefiles_get_security_ID(struct cachefiles_cache *cache)
+{
+	struct cred *new;
+	int ret;
+
+	_enter("{%u}", cache->have_secid ? cache->secid : 0);
+
+	new = prepare_kernel_cred(current);
+	if (!new) {
+		ret = -ENOMEM;
+		goto error;
+	}
+
+	if (cache->have_secid) {
+		ret = set_security_override(new, cache->secid);
+		if (ret < 0) {
+			put_cred(new);
+			pr_err("Security denies permission to nominate security context: error %d\n",
+			       ret);
+			goto error;
+		}
+	}
+
+	cache->cache_cred = new;
+	ret = 0;
+error:
+	_leave(" = %d", ret);
+	return ret;
+}
+
+/*
+ * see if mkdir and create can be performed in the root directory
+ */
+static int cachefiles_check_cache_dir(struct cachefiles_cache *cache,
+				      struct dentry *root)
+{
+	int ret;
+
+	ret = security_inode_mkdir(d_backing_inode(root), root, 0);
+	if (ret < 0) {
+		pr_err("Security denies permission to make dirs: error %d",
+		       ret);
+		return ret;
+	}
+
+	ret = security_inode_create(d_backing_inode(root), root, 0);
+	if (ret < 0)
+		pr_err("Security denies permission to create files: error %d",
+		       ret);
+
+	return ret;
+}
+
+/*
+ * check the security details of the on-disk cache
+ * - must be called with security override in force
+ * - must return with a security override in force - even in the case of an
+ *   error
+ */
+int cachefiles_determine_cache_security(struct cachefiles_cache *cache,
+					struct dentry *root,
+					const struct cred **_saved_cred)
+{
+	struct cred *new;
+	int ret;
+
+	_enter("");
+
+	/* duplicate the cache creds for COW (the override is currently in
+	 * force, so we can use prepare_creds() to do this) */
+	new = prepare_creds();
+	if (!new)
+		return -ENOMEM;
+
+	cachefiles_end_secure(cache, *_saved_cred);
+
+	/* use the cache root dir's security context as the basis with
+	 * which create files */
+	ret = set_create_files_as(new, d_backing_inode(root));
+	if (ret < 0) {
+		abort_creds(new);
+		cachefiles_begin_secure(cache, _saved_cred);
+		_leave(" = %d [cfa]", ret);
+		return ret;
+	}
+
+	put_cred(cache->cache_cred);
+	cache->cache_cred = new;
+
+	cachefiles_begin_secure(cache, _saved_cred);
+	ret = cachefiles_check_cache_dir(cache, root);
+
+	if (ret == -EOPNOTSUPP)
+		ret = 0;
+	_leave(" = %d", ret);
+	return ret;
+}