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, 177 insertions, 0 deletions

diff --git a/net/wireless/tests/fragmentation.c b/net/wireless/tests/fragmentation.c
new file mode 100644
index 000000000..411fae18c
--- /dev/null
+++ b/net/wireless/tests/fragmentation.c
@@ -0,0 +1,177 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * KUnit tests for element fragmentation
+ *
+ * Copyright (C) 2023-2024 Intel Corporation
+ */
+#include <linux/ieee80211.h>
+#include <net/cfg80211.h>
+#include <kunit/test.h>
+
+static void defragment_0(struct kunit *test)
+{
+	ssize_t ret;
+	static const u8 input[] = {
+		[0] = WLAN_EID_EXTENSION,
+		[1] = 254,
+		[2] = WLAN_EID_EXT_EHT_MULTI_LINK,
+		[27] = 27,
+		[123] = 123,
+		[254 + 2] = WLAN_EID_FRAGMENT,
+		[254 + 3] = 7,
+		[254 + 3 + 7] = 0, /* for size */
+	};
+	u8 *data = kunit_kzalloc(test, sizeof(input), GFP_KERNEL);
+
+	KUNIT_ASSERT_NOT_NULL(test, data);
+
+	ret = cfg80211_defragment_element((void *)input,
+					  input, sizeof(input),
+					  NULL, 0,
+					  WLAN_EID_FRAGMENT);
+	KUNIT_EXPECT_EQ(test, ret, 253);
+	ret = cfg80211_defragment_element((void *)input,
+					  input, sizeof(input),
+					  data, ret,
+					  WLAN_EID_FRAGMENT);
+	KUNIT_EXPECT_EQ(test, ret, 253);
+	KUNIT_EXPECT_MEMEQ(test, data, input + 3, 253);
+}
+
+static void defragment_1(struct kunit *test)
+{
+	ssize_t ret;
+	static const u8 input[] = {
+		[0] = WLAN_EID_EXTENSION,
+		[1] = 255,
+		[2] = WLAN_EID_EXT_EHT_MULTI_LINK,
+		[27] = 27,
+		[123] = 123,
+		[255 + 2] = WLAN_EID_FRAGMENT,
+		[255 + 3] = 7,
+		[255 + 3 + 1] = 0xaa,
+		[255 + 3 + 8] = WLAN_EID_FRAGMENT, /* not used */
+		[255 + 3 + 9] = 1,
+		[255 + 3 + 10] = 0, /* for size */
+	};
+	u8 *data = kunit_kzalloc(test, sizeof(input), GFP_KERNEL);
+	const struct element *elem;
+	int count = 0;
+
+	KUNIT_ASSERT_NOT_NULL(test, data);
+
+	for_each_element(elem, input, sizeof(input))
+		count++;
+
+	/* check the elements are right */
+	KUNIT_ASSERT_EQ(test, count, 3);
+
+	ret = cfg80211_defragment_element((void *)input,
+					  input, sizeof(input),
+					  NULL, 0,
+					  WLAN_EID_FRAGMENT);
+	KUNIT_EXPECT_EQ(test, ret, 254 + 7);
+	ret = cfg80211_defragment_element((void *)input,
+					  input, sizeof(input),
+					  data, ret,
+					  WLAN_EID_FRAGMENT);
+	/* this means the last fragment was not used */
+	KUNIT_EXPECT_EQ(test, ret, 254 + 7);
+	KUNIT_EXPECT_MEMEQ(test, data, input + 3, 254);
+	KUNIT_EXPECT_MEMEQ(test, data + 254, input + 255 + 4, 7);
+}
+
+static void defragment_2(struct kunit *test)
+{
+	ssize_t ret;
+	static const u8 input[] = {
+		[0] = WLAN_EID_EXTENSION,
+		[1] = 255,
+		[2] = WLAN_EID_EXT_EHT_MULTI_LINK,
+		[27] = 27,
+		[123] = 123,
+
+		[257 + 0] = WLAN_EID_FRAGMENT,
+		[257 + 1] = 255,
+		[257 + 20] = 0xaa,
+
+		[2 * 257 + 0] = WLAN_EID_FRAGMENT,
+		[2 * 257 + 1] = 1,
+		[2 * 257 + 2] = 0xcc,
+		[2 * 257 + 3] = WLAN_EID_FRAGMENT, /* not used */
+		[2 * 257 + 4] = 1,
+		[2 * 257 + 5] = 0, /* for size */
+	};
+	u8 *data = kunit_kzalloc(test, sizeof(input), GFP_KERNEL);
+	const struct element *elem;
+	int count = 0;
+
+	KUNIT_ASSERT_NOT_NULL(test, data);
+
+	for_each_element(elem, input, sizeof(input))
+		count++;
+
+	/* check the elements are right */
+	KUNIT_ASSERT_EQ(test, count, 4);
+
+	ret = cfg80211_defragment_element((void *)input,
+					  input, sizeof(input),
+					  NULL, 0,
+					  WLAN_EID_FRAGMENT);
+	/* this means the last fragment was not used */
+	KUNIT_EXPECT_EQ(test, ret, 254 + 255 + 1);
+	ret = cfg80211_defragment_element((void *)input,
+					  input, sizeof(input),
+					  data, ret,
+					  WLAN_EID_FRAGMENT);
+	KUNIT_EXPECT_EQ(test, ret, 254 + 255 + 1);
+	KUNIT_EXPECT_MEMEQ(test, data, input + 3, 254);
+	KUNIT_EXPECT_MEMEQ(test, data + 254, input + 257 + 2, 255);
+	KUNIT_EXPECT_MEMEQ(test, data + 254 + 255, input + 2 * 257 + 2, 1);
+}
+
+static void defragment_at_end(struct kunit *test)
+{
+	ssize_t ret;
+	static const u8 input[] = {
+		[0] = WLAN_EID_EXTENSION,
+		[1] = 255,
+		[2] = WLAN_EID_EXT_EHT_MULTI_LINK,
+		[27] = 27,
+		[123] = 123,
+		[255 + 2] = WLAN_EID_FRAGMENT,
+		[255 + 3] = 7,
+		[255 + 3 + 7] = 0, /* for size */
+	};
+	u8 *data = kunit_kzalloc(test, sizeof(input), GFP_KERNEL);
+
+	KUNIT_ASSERT_NOT_NULL(test, data);
+
+	ret = cfg80211_defragment_element((void *)input,
+					  input, sizeof(input),
+					  NULL, 0,
+					  WLAN_EID_FRAGMENT);
+	KUNIT_EXPECT_EQ(test, ret, 254 + 7);
+	ret = cfg80211_defragment_element((void *)input,
+					  input, sizeof(input),
+					  data, ret,
+					  WLAN_EID_FRAGMENT);
+	KUNIT_EXPECT_EQ(test, ret, 254 + 7);
+	KUNIT_EXPECT_MEMEQ(test, data, input + 3, 254);
+	KUNIT_EXPECT_MEMEQ(test, data + 254, input + 255 + 4, 7);
+}
+
+static struct kunit_case element_fragmentation_test_cases[] = {
+	KUNIT_CASE(defragment_0),
+	KUNIT_CASE(defragment_1),
+	KUNIT_CASE(defragment_2),
+	KUNIT_CASE(defragment_at_end),
+	{}
+};
+
+static struct kunit_suite element_fragmentation = {
+	.name = "cfg80211-element-defragmentation",
+	.test_cases = element_fragmentation_test_cases,
+};
+
+kunit_test_suite(element_fragmentation);