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, 120 insertions, 0 deletions

diff --git a/include/sound/memalloc.h b/include/sound/memalloc.h
new file mode 100644
index 000000000..9dd475cf4
--- /dev/null
+++ b/include/sound/memalloc.h
@@ -0,0 +1,120 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ *  Copyright (c) by Jaroslav Kysela <perex@perex.cz>
+ *                   Takashi Iwai <tiwai@suse.de>
+ * 
+ *  Generic memory allocators
+ */
+
+#ifndef __SOUND_MEMALLOC_H
+#define __SOUND_MEMALLOC_H
+
+#include <linux/dma-direction.h>
+#include <asm/page.h>
+
+struct device;
+struct vm_area_struct;
+struct sg_table;
+
+/*
+ * buffer device info
+ */
+struct snd_dma_device {
+	int type;			/* SNDRV_DMA_TYPE_XXX */
+	enum dma_data_direction dir;	/* DMA direction */
+	bool need_sync;			/* explicit sync needed? */
+	struct device *dev;		/* generic device */
+};
+
+/*
+ * buffer types
+ */
+#define SNDRV_DMA_TYPE_UNKNOWN		0	/* not defined */
+#define SNDRV_DMA_TYPE_CONTINUOUS	1	/* continuous no-DMA memory */
+#define SNDRV_DMA_TYPE_DEV		2	/* generic device continuous */
+#define SNDRV_DMA_TYPE_DEV_WC		5	/* continuous write-combined */
+#ifdef CONFIG_GENERIC_ALLOCATOR
+#define SNDRV_DMA_TYPE_DEV_IRAM		4	/* generic device iram-buffer */
+#else
+#define SNDRV_DMA_TYPE_DEV_IRAM	SNDRV_DMA_TYPE_DEV
+#endif
+#define SNDRV_DMA_TYPE_VMALLOC		7	/* vmalloc'ed buffer */
+#define SNDRV_DMA_TYPE_NONCONTIG	8	/* non-coherent SG buffer */
+#define SNDRV_DMA_TYPE_NONCOHERENT	9	/* non-coherent buffer */
+#ifdef CONFIG_SND_DMA_SGBUF
+#define SNDRV_DMA_TYPE_DEV_SG		3	/* S/G pages */
+#define SNDRV_DMA_TYPE_DEV_WC_SG	6	/* SG write-combined */
+#else
+#define SNDRV_DMA_TYPE_DEV_SG	SNDRV_DMA_TYPE_DEV /* no SG-buf support */
+#define SNDRV_DMA_TYPE_DEV_WC_SG	SNDRV_DMA_TYPE_DEV_WC
+#endif
+
+/*
+ * info for buffer allocation
+ */
+struct snd_dma_buffer {
+	struct snd_dma_device dev;	/* device type */
+	unsigned char *area;	/* virtual pointer */
+	dma_addr_t addr;	/* physical address */
+	size_t bytes;		/* buffer size in bytes */
+	void *private_data;	/* private for allocator; don't touch */
+};
+
+/*
+ * return the pages matching with the given byte size
+ */
+static inline unsigned int snd_sgbuf_aligned_pages(size_t size)
+{
+	return (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
+}
+
+/* allocate/release a buffer */
+int snd_dma_alloc_dir_pages(int type, struct device *dev,
+			    enum dma_data_direction dir, size_t size,
+			    struct snd_dma_buffer *dmab);
+
+static inline int snd_dma_alloc_pages(int type, struct device *dev,
+				      size_t size, struct snd_dma_buffer *dmab)
+{
+	return snd_dma_alloc_dir_pages(type, dev, DMA_BIDIRECTIONAL, size, dmab);
+}
+
+int snd_dma_alloc_pages_fallback(int type, struct device *dev, size_t size,
+                                 struct snd_dma_buffer *dmab);
+void snd_dma_free_pages(struct snd_dma_buffer *dmab);
+int snd_dma_buffer_mmap(struct snd_dma_buffer *dmab,
+			struct vm_area_struct *area);
+
+enum snd_dma_sync_mode { SNDRV_DMA_SYNC_CPU, SNDRV_DMA_SYNC_DEVICE };
+#ifdef CONFIG_HAS_DMA
+void snd_dma_buffer_sync(struct snd_dma_buffer *dmab,
+			 enum snd_dma_sync_mode mode);
+#else
+static inline void snd_dma_buffer_sync(struct snd_dma_buffer *dmab,
+				       enum snd_dma_sync_mode mode) {}
+#endif
+
+dma_addr_t snd_sgbuf_get_addr(struct snd_dma_buffer *dmab, size_t offset);
+struct page *snd_sgbuf_get_page(struct snd_dma_buffer *dmab, size_t offset);
+unsigned int snd_sgbuf_get_chunk_size(struct snd_dma_buffer *dmab,
+				      unsigned int ofs, unsigned int size);
+
+/* device-managed memory allocator */
+struct snd_dma_buffer *snd_devm_alloc_dir_pages(struct device *dev, int type,
+						enum dma_data_direction dir,
+						size_t size);
+
+static inline struct snd_dma_buffer *
+snd_devm_alloc_pages(struct device *dev, int type, size_t size)
+{
+	return snd_devm_alloc_dir_pages(dev, type, DMA_BIDIRECTIONAL, size);
+}
+
+static inline struct sg_table *
+snd_dma_noncontig_sg_table(struct snd_dma_buffer *dmab)
+{
+	return dmab->private_data;
+}
+
+#endif /* __SOUND_MEMALLOC_H */
+