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, 175 insertions, 0 deletions

diff --git a/sound/synth/util_mem.c b/sound/synth/util_mem.c
new file mode 100644
index 000000000..21893e561
--- /dev/null
+++ b/sound/synth/util_mem.c
@@ -0,0 +1,175 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ *  Copyright (C) 2000 Takashi Iwai <tiwai@suse.de>
+ *
+ *  Generic memory management routines for soundcard memory allocation
+ */
+
+#include <linux/mutex.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <sound/core.h>
+#include <sound/util_mem.h>
+
+MODULE_AUTHOR("Takashi Iwai");
+MODULE_DESCRIPTION("Generic memory management routines for soundcard memory allocation");
+MODULE_LICENSE("GPL");
+
+#define get_memblk(p)	list_entry(p, struct snd_util_memblk, list)
+
+/*
+ * create a new memory manager
+ */
+struct snd_util_memhdr *
+snd_util_memhdr_new(int memsize)
+{
+	struct snd_util_memhdr *hdr;
+
+	hdr = kzalloc_obj(*hdr);
+	if (hdr == NULL)
+		return NULL;
+	hdr->size = memsize;
+	mutex_init(&hdr->block_mutex);
+	INIT_LIST_HEAD(&hdr->block);
+
+	return hdr;
+}
+
+/*
+ * free a memory manager
+ */
+void snd_util_memhdr_free(struct snd_util_memhdr *hdr)
+{
+	struct list_head *p;
+
+	if (!hdr)
+		return;
+	/* release all blocks */
+	while ((p = hdr->block.next) != &hdr->block) {
+		list_del(p);
+		kfree(get_memblk(p));
+	}
+	kfree(hdr);
+}
+
+/*
+ * allocate a memory block (without mutex)
+ */
+struct snd_util_memblk *
+__snd_util_mem_alloc(struct snd_util_memhdr *hdr, int size)
+{
+	struct snd_util_memblk *blk;
+	unsigned int units, prev_offset;
+	struct list_head *p;
+
+	if (snd_BUG_ON(!hdr || size <= 0))
+		return NULL;
+
+	/* word alignment */
+	units = size;
+	if (units & 1)
+		units++;
+	if (units > hdr->size)
+		return NULL;
+
+	/* look for empty block */
+	prev_offset = 0;
+	list_for_each(p, &hdr->block) {
+		blk = get_memblk(p);
+		if (blk->offset - prev_offset >= units)
+			goto __found;
+		prev_offset = blk->offset + blk->size;
+	}
+	if (hdr->size - prev_offset < units)
+		return NULL;
+
+__found:
+	return __snd_util_memblk_new(hdr, units, p->prev);
+}
+
+
+/*
+ * create a new memory block with the given size
+ * the block is linked next to prev
+ */
+struct snd_util_memblk *
+__snd_util_memblk_new(struct snd_util_memhdr *hdr, unsigned int units,
+		      struct list_head *prev)
+{
+	struct snd_util_memblk *blk;
+
+	blk = kmalloc(sizeof(struct snd_util_memblk) + hdr->block_extra_size,
+		      GFP_KERNEL);
+	if (blk == NULL)
+		return NULL;
+
+	if (prev == &hdr->block)
+		blk->offset = 0;
+	else {
+		struct snd_util_memblk *p = get_memblk(prev);
+		blk->offset = p->offset + p->size;
+	}
+	blk->size = units;
+	list_add(&blk->list, prev);
+	hdr->nblocks++;
+	hdr->used += units;
+	return blk;
+}
+
+
+/*
+ * allocate a memory block (with mutex)
+ */
+struct snd_util_memblk *
+snd_util_mem_alloc(struct snd_util_memhdr *hdr, int size)
+{
+	guard(mutex)(&hdr->block_mutex);
+	return __snd_util_mem_alloc(hdr, size);
+}
+
+
+/*
+ * remove the block from linked-list and free resource
+ * (without mutex)
+ */
+void
+__snd_util_mem_free(struct snd_util_memhdr *hdr, struct snd_util_memblk *blk)
+{
+	list_del(&blk->list);
+	hdr->nblocks--;
+	hdr->used -= blk->size;
+	kfree(blk);
+}
+
+/*
+ * free a memory block (with mutex)
+ */
+int snd_util_mem_free(struct snd_util_memhdr *hdr, struct snd_util_memblk *blk)
+{
+	if (snd_BUG_ON(!hdr || !blk))
+		return -EINVAL;
+
+	guard(mutex)(&hdr->block_mutex);
+	__snd_util_mem_free(hdr, blk);
+	return 0;
+}
+
+/*
+ * return available memory size
+ */
+int snd_util_mem_avail(struct snd_util_memhdr *hdr)
+{
+	guard(mutex)(&hdr->block_mutex);
+	return hdr->size - hdr->used;
+}
+
+
+EXPORT_SYMBOL(snd_util_memhdr_new);
+EXPORT_SYMBOL(snd_util_memhdr_free);
+EXPORT_SYMBOL(snd_util_mem_alloc);
+EXPORT_SYMBOL(snd_util_mem_free);
+EXPORT_SYMBOL(snd_util_mem_avail);
+EXPORT_SYMBOL(__snd_util_mem_alloc);
+EXPORT_SYMBOL(__snd_util_mem_free);
+EXPORT_SYMBOL(__snd_util_memblk_new);