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, 128 insertions, 0 deletions

diff --git a/drivers/misc/altera-stapl/altera-comp.c b/drivers/misc/altera-stapl/altera-comp.c
new file mode 100644
index 000000000..4a63f51cc
--- /dev/null
+++ b/drivers/misc/altera-stapl/altera-comp.c
@@ -0,0 +1,128 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * altera-comp.c
+ *
+ * altera FPGA driver
+ *
+ * Copyright (C) Altera Corporation 1998-2001
+ * Copyright (C) 2010 NetUP Inc.
+ * Copyright (C) 2010 Igor M. Liplianin <liplianin@netup.ru>
+ */
+
+#include <linux/kernel.h>
+#include "altera-exprt.h"
+
+#define	SHORT_BITS		16
+#define	CHAR_BITS		8
+#define	DATA_BLOB_LENGTH	3
+#define	MATCH_DATA_LENGTH	8192
+#define ALTERA_REQUEST_SIZE	1024
+#define ALTERA_BUFFER_SIZE	(MATCH_DATA_LENGTH + ALTERA_REQUEST_SIZE)
+
+static u32 altera_bits_req(u32 n)
+{
+	u32 result = SHORT_BITS;
+
+	if (n == 0)
+		result = 1;
+	else {
+		/* Look for the highest non-zero bit position */
+		while ((n & (1 << (SHORT_BITS - 1))) == 0) {
+			n <<= 1;
+			--result;
+		}
+	}
+
+	return result;
+}
+
+static u32 altera_read_packed(u8 *buffer, u32 bits, u32 *bits_avail,
+							u32 *in_index)
+{
+	u32 result = 0;
+	u32 shift = 0;
+	u32 databyte = 0;
+
+	while (bits > 0) {
+		databyte = buffer[*in_index];
+		result |= (((databyte >> (CHAR_BITS - *bits_avail))
+			& (0xff >> (CHAR_BITS - *bits_avail))) << shift);
+
+		if (bits <= *bits_avail) {
+			result &= (0xffff >> (SHORT_BITS - (bits + shift)));
+			*bits_avail -= bits;
+			bits = 0;
+		} else {
+			++(*in_index);
+			shift += *bits_avail;
+			bits -= *bits_avail;
+			*bits_avail = CHAR_BITS;
+		}
+	}
+
+	return result;
+}
+
+u32 altera_shrink(u8 *in, u32 in_length, u8 *out, u32 out_length, s32 version)
+{
+	u32 i, j, data_length = 0L;
+	u32 offset, length;
+	u32 match_data_length = MATCH_DATA_LENGTH;
+	u32 bits_avail = CHAR_BITS;
+	u32 in_index = 0L;
+
+	if (version > 0)
+		--match_data_length;
+
+	for (i = 0; i < out_length; ++i)
+		out[i] = 0;
+
+	/* Read number of bytes in data. */
+	for (i = 0; i < sizeof(in_length); ++i) {
+		data_length = data_length | (
+			altera_read_packed(in,
+					CHAR_BITS,
+					&bits_avail,
+					&in_index) << (i * CHAR_BITS));
+	}
+
+	if (data_length > out_length) {
+		data_length = 0L;
+		return data_length;
+	}
+
+	i = 0;
+	while (i < data_length) {
+		/* A 0 bit indicates literal data. */
+		if (altera_read_packed(in, 1, &bits_avail,
+						&in_index) == 0) {
+			for (j = 0; j < DATA_BLOB_LENGTH; ++j) {
+				if (i < data_length) {
+					out[i] = (u8)altera_read_packed(in,
+							CHAR_BITS,
+							&bits_avail,
+							&in_index);
+					i++;
+				}
+			}
+		} else {
+			/* A 1 bit indicates offset/length to follow. */
+			offset = altera_read_packed(in, altera_bits_req((s16)
+					(i > match_data_length ?
+						match_data_length : i)),
+					&bits_avail,
+					&in_index);
+			length = altera_read_packed(in, CHAR_BITS,
+					&bits_avail,
+					&in_index);
+			for (j = 0; j < length; ++j) {
+				if (i < data_length) {
+					out[i] = out[i - offset];
+					i++;
+				}
+			}
+		}
+	}
+
+	return data_length;
+}