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, 77 insertions, 0 deletions

diff --git a/sound/firewire/packets-buffer.c b/sound/firewire/packets-buffer.c
new file mode 100644
index 000000000..2093a5fbd
--- /dev/null
+++ b/sound/firewire/packets-buffer.c
@@ -0,0 +1,77 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * helpers for managing a buffer for many packets
+ *
+ * Copyright (c) Clemens Ladisch <clemens@ladisch.de>
+ */
+
+#include <linux/firewire.h>
+#include <linux/export.h>
+#include <linux/slab.h>
+#include "packets-buffer.h"
+
+/**
+ * iso_packets_buffer_init - allocates the memory for packets
+ * @b: the buffer structure to initialize
+ * @unit: the device at the other end of the stream
+ * @count: the number of packets
+ * @packet_size: the (maximum) size of a packet, in bytes
+ * @direction: %DMA_TO_DEVICE or %DMA_FROM_DEVICE
+ */
+int iso_packets_buffer_init(struct iso_packets_buffer *b, struct fw_unit *unit,
+			    unsigned int count, unsigned int packet_size,
+			    enum dma_data_direction direction)
+{
+	unsigned int packets_per_page, pages;
+	unsigned int i, page_index, offset_in_page;
+	void *p;
+	int err;
+
+	b->packets = kmalloc_objs(*b->packets, count);
+	if (!b->packets) {
+		err = -ENOMEM;
+		goto error;
+	}
+
+	packet_size = L1_CACHE_ALIGN(packet_size);
+	packets_per_page = PAGE_SIZE / packet_size;
+	if (WARN_ON(!packets_per_page)) {
+		err = -EINVAL;
+		goto err_packets;
+	}
+	pages = DIV_ROUND_UP(count, packets_per_page);
+
+	err = fw_iso_buffer_init(&b->iso_buffer, fw_parent_device(unit)->card,
+				 pages, direction);
+	if (err < 0)
+		goto err_packets;
+
+	for (i = 0; i < count; ++i) {
+		page_index = i / packets_per_page;
+		p = page_address(b->iso_buffer.pages[page_index]);
+		offset_in_page = (i % packets_per_page) * packet_size;
+		b->packets[i].buffer = p + offset_in_page;
+		b->packets[i].offset = page_index * PAGE_SIZE + offset_in_page;
+	}
+
+	return 0;
+
+err_packets:
+	kfree(b->packets);
+error:
+	return err;
+}
+EXPORT_SYMBOL(iso_packets_buffer_init);
+
+/**
+ * iso_packets_buffer_destroy - frees packet buffer resources
+ * @b: the buffer structure to free
+ * @unit: the device at the other end of the stream
+ */
+void iso_packets_buffer_destroy(struct iso_packets_buffer *b,
+				struct fw_unit *unit)
+{
+	fw_iso_buffer_destroy(&b->iso_buffer, fw_parent_device(unit)->card);
+	kfree(b->packets);
+}
+EXPORT_SYMBOL(iso_packets_buffer_destroy);