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, 213 insertions, 0 deletions

diff --git a/mm/cma_debug.c b/mm/cma_debug.c
new file mode 100644
index 000000000..523ba4a0f
--- /dev/null
+++ b/mm/cma_debug.c
@@ -0,0 +1,213 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * CMA DebugFS Interface
+ *
+ * Copyright (c) 2015 Sasha Levin <sasha.levin@oracle.com>
+ */
+
+
+#include <linux/debugfs.h>
+#include <linux/cma.h>
+#include <linux/list.h>
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/mm_types.h>
+
+#include "cma.h"
+
+struct cma_mem {
+	struct hlist_node node;
+	struct page *p;
+	unsigned long n;
+};
+
+static int cma_debugfs_get(void *data, u64 *val)
+{
+	unsigned long *p = data;
+
+	*val = *p;
+
+	return 0;
+}
+DEFINE_DEBUGFS_ATTRIBUTE(cma_debugfs_fops, cma_debugfs_get, NULL, "%llu\n");
+
+static int cma_used_get(void *data, u64 *val)
+{
+	struct cma *cma = data;
+
+	spin_lock_irq(&cma->lock);
+	*val = cma->count - cma->available_count;
+	spin_unlock_irq(&cma->lock);
+
+	return 0;
+}
+DEFINE_DEBUGFS_ATTRIBUTE(cma_used_fops, cma_used_get, NULL, "%llu\n");
+
+static int cma_maxchunk_get(void *data, u64 *val)
+{
+	struct cma *cma = data;
+	struct cma_memrange *cmr;
+	unsigned long maxchunk = 0;
+	unsigned long start, end;
+	unsigned long bitmap_maxno;
+	int r;
+
+	spin_lock_irq(&cma->lock);
+	for (r = 0; r < cma->nranges; r++) {
+		cmr = &cma->ranges[r];
+		bitmap_maxno = cma_bitmap_maxno(cma, cmr);
+		for_each_clear_bitrange(start, end, cmr->bitmap, bitmap_maxno)
+			maxchunk = max(end - start, maxchunk);
+	}
+	spin_unlock_irq(&cma->lock);
+	*val = (u64)maxchunk << cma->order_per_bit;
+
+	return 0;
+}
+DEFINE_DEBUGFS_ATTRIBUTE(cma_maxchunk_fops, cma_maxchunk_get, NULL, "%llu\n");
+
+static void cma_add_to_cma_mem_list(struct cma *cma, struct cma_mem *mem)
+{
+	spin_lock(&cma->mem_head_lock);
+	hlist_add_head(&mem->node, &cma->mem_head);
+	spin_unlock(&cma->mem_head_lock);
+}
+
+static struct cma_mem *cma_get_entry_from_list(struct cma *cma)
+{
+	struct cma_mem *mem = NULL;
+
+	spin_lock(&cma->mem_head_lock);
+	if (!hlist_empty(&cma->mem_head)) {
+		mem = hlist_entry(cma->mem_head.first, struct cma_mem, node);
+		hlist_del_init(&mem->node);
+	}
+	spin_unlock(&cma->mem_head_lock);
+
+	return mem;
+}
+
+static int cma_free_mem(struct cma *cma, int count)
+{
+	struct cma_mem *mem = NULL;
+
+	while (count) {
+		mem = cma_get_entry_from_list(cma);
+		if (mem == NULL)
+			return 0;
+
+		if (mem->n <= count) {
+			cma_release(cma, mem->p, mem->n);
+			count -= mem->n;
+			kfree(mem);
+		} else if (cma->order_per_bit == 0) {
+			cma_release(cma, mem->p, count);
+			mem->p += count;
+			mem->n -= count;
+			count = 0;
+			cma_add_to_cma_mem_list(cma, mem);
+		} else {
+			pr_debug("cma: cannot release partial block when order_per_bit != 0\n");
+			cma_add_to_cma_mem_list(cma, mem);
+			break;
+		}
+	}
+
+	return 0;
+
+}
+
+static int cma_free_write(void *data, u64 val)
+{
+	int pages = val;
+	struct cma *cma = data;
+
+	return cma_free_mem(cma, pages);
+}
+DEFINE_DEBUGFS_ATTRIBUTE(cma_free_fops, NULL, cma_free_write, "%llu\n");
+
+static int cma_alloc_mem(struct cma *cma, int count)
+{
+	struct cma_mem *mem;
+	struct page *p;
+
+	mem = kzalloc_obj(*mem);
+	if (!mem)
+		return -ENOMEM;
+
+	p = cma_alloc(cma, count, 0, false);
+	if (!p) {
+		kfree(mem);
+		return -ENOMEM;
+	}
+
+	mem->p = p;
+	mem->n = count;
+
+	cma_add_to_cma_mem_list(cma, mem);
+
+	return 0;
+}
+
+static int cma_alloc_write(void *data, u64 val)
+{
+	int pages = val;
+	struct cma *cma = data;
+
+	return cma_alloc_mem(cma, pages);
+}
+DEFINE_DEBUGFS_ATTRIBUTE(cma_alloc_fops, NULL, cma_alloc_write, "%llu\n");
+
+static void cma_debugfs_add_one(struct cma *cma, struct dentry *root_dentry)
+{
+	struct dentry *tmp, *dir, *rangedir;
+	int r;
+	char rdirname[12];
+	struct cma_memrange *cmr;
+
+	tmp = debugfs_create_dir(cma->name, root_dentry);
+
+	debugfs_create_file("alloc", 0200, tmp, cma, &cma_alloc_fops);
+	debugfs_create_file("free", 0200, tmp, cma, &cma_free_fops);
+	debugfs_create_file("count", 0444, tmp, &cma->count, &cma_debugfs_fops);
+	debugfs_create_file("order_per_bit", 0444, tmp,
+			    &cma->order_per_bit, &cma_debugfs_fops);
+	debugfs_create_file("used", 0444, tmp, cma, &cma_used_fops);
+	debugfs_create_file("maxchunk", 0444, tmp, cma, &cma_maxchunk_fops);
+
+	rangedir = debugfs_create_dir("ranges", tmp);
+	for (r = 0; r < cma->nranges; r++) {
+		cmr = &cma->ranges[r];
+		snprintf(rdirname, sizeof(rdirname), "%d", r);
+		dir = debugfs_create_dir(rdirname, rangedir);
+		debugfs_create_file("base_pfn", 0444, dir,
+			    &cmr->base_pfn, &cma_debugfs_fops);
+		cmr->dfs_bitmap.array = (u32 *)cmr->bitmap;
+		cmr->dfs_bitmap.n_elements =
+			DIV_ROUND_UP(cma_bitmap_maxno(cma, cmr),
+					BITS_PER_BYTE * sizeof(u32));
+		debugfs_create_u32_array("bitmap", 0444, dir,
+				&cmr->dfs_bitmap);
+	}
+
+	/*
+	 * Backward compatible symlinks to range 0 for base_pfn and bitmap.
+	 */
+	debugfs_create_symlink("base_pfn", tmp, "ranges/0/base_pfn");
+	debugfs_create_symlink("bitmap", tmp, "ranges/0/bitmap");
+}
+
+static int __init cma_debugfs_init(void)
+{
+	struct dentry *cma_debugfs_root;
+	int i;
+
+	cma_debugfs_root = debugfs_create_dir("cma", NULL);
+
+	for (i = 0; i < cma_area_count; i++)
+		if (test_bit(CMA_ACTIVATED, &cma_areas[i].flags))
+			cma_debugfs_add_one(&cma_areas[i], cma_debugfs_root);
+
+	return 0;
+}
+late_initcall(cma_debugfs_init);