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, 182 insertions, 0 deletions

diff --git a/include/sound/sof.h b/include/sound/sof.h
new file mode 100644
index 000000000..38d6c8cb5
--- /dev/null
+++ b/include/sound/sof.h
@@ -0,0 +1,182 @@
+/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-3-Clause) */
+/*
+ * This file is provided under a dual BSD/GPLv2 license.  When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * Copyright(c) 2018 Intel Corporation
+ *
+ * Author: Liam Girdwood <liam.r.girdwood@linux.intel.com>
+ */
+
+#ifndef __INCLUDE_SOUND_SOF_H
+#define __INCLUDE_SOUND_SOF_H
+
+#include <linux/pci.h>
+#include <sound/soc.h>
+#include <sound/soc-acpi.h>
+
+struct snd_sof_dsp_ops;
+struct snd_sof_dev;
+
+/**
+ * enum sof_fw_state - DSP firmware state definitions
+ * @SOF_FW_BOOT_NOT_STARTED:	firmware boot is not yet started
+ * @SOF_DSPLESS_MODE:		DSP is not used
+ * @SOF_FW_BOOT_PREPARE:	preparing for boot (firmware loading for exaqmple)
+ * @SOF_FW_BOOT_IN_PROGRESS:	firmware boot is in progress
+ * @SOF_FW_BOOT_FAILED:		firmware boot failed
+ * @SOF_FW_BOOT_READY_FAILED:	firmware booted but fw_ready op failed
+ * @SOF_FW_BOOT_READY_OK:	firmware booted and fw_ready op passed
+ * @SOF_FW_BOOT_COMPLETE:	firmware is booted up and functional
+ * @SOF_FW_CRASHED:		firmware crashed after successful boot
+ */
+enum sof_fw_state {
+	SOF_FW_BOOT_NOT_STARTED = 0,
+	SOF_DSPLESS_MODE,
+	SOF_FW_BOOT_PREPARE,
+	SOF_FW_BOOT_IN_PROGRESS,
+	SOF_FW_BOOT_FAILED,
+	SOF_FW_BOOT_READY_FAILED,
+	SOF_FW_BOOT_READY_OK,
+	SOF_FW_BOOT_COMPLETE,
+	SOF_FW_CRASHED,
+};
+
+/* DSP power states */
+enum sof_dsp_power_states {
+	SOF_DSP_PM_D0,
+	SOF_DSP_PM_D1,
+	SOF_DSP_PM_D2,
+	SOF_DSP_PM_D3,
+};
+
+/* Definitions for multiple IPCs */
+enum sof_ipc_type {
+	SOF_IPC_TYPE_3,
+	SOF_IPC_TYPE_4,
+	SOF_IPC_TYPE_COUNT
+};
+
+struct sof_loadable_file_profile {
+	enum sof_ipc_type ipc_type;
+
+	const char *fw_path;
+	const char *fw_path_postfix;
+	const char *fw_name;
+	const char *fw_lib_path;
+	const char *fw_lib_path_postfix;
+	const char *tplg_path;
+	const char *tplg_name;
+};
+
+/*
+ * SOF Platform data.
+ */
+struct snd_sof_pdata {
+	const char *name;
+	const char *platform;
+
+	/*
+	 * PCI SSID. As PCI does not define 0 as invalid, the subsystem_id_set
+	 * flag indicates that a value has been written to these members.
+	 */
+	unsigned short subsystem_vendor;
+	unsigned short subsystem_device;
+	bool subsystem_id_set;
+
+	struct device *dev;
+
+	/*
+	 * notification callback used if the hardware initialization
+	 * can take time or is handled in a workqueue. This callback
+	 * can be used by the caller to e.g. enable runtime_pm
+	 * or limit functionality until all low-level inits are
+	 * complete.
+	 */
+	void (*sof_probe_complete)(struct device *dev);
+
+	/* descriptor */
+	const struct sof_dev_desc *desc;
+
+	/* platform's preferred IPC type and path overrides */
+	struct sof_loadable_file_profile ipc_file_profile_base;
+
+	/* firmware and topology filenames */
+	const char *fw_filename_prefix;
+	const char *fw_filename;
+	const char *tplg_filename_prefix;
+	const char *tplg_filename;
+	bool disable_function_topology;
+
+	/* loadable external libraries available under this directory */
+	const char *fw_lib_prefix;
+
+	/* machine */
+	struct platform_device *pdev_mach;
+	const struct snd_soc_acpi_mach *machine;
+	const struct snd_sof_of_mach *of_machine;
+
+	void *hw_pdata;
+
+	enum sof_ipc_type ipc_type;
+};
+
+/*
+ * Descriptor used for setting up SOF platform data. This is used when
+ * ACPI/PCI data is missing or mapped differently.
+ */
+struct sof_dev_desc {
+	/* list of machines using this configuration */
+	struct snd_soc_acpi_mach *machines;
+	struct snd_sof_of_mach *of_machines;
+
+	/* alternate list of machines using this configuration */
+	struct snd_soc_acpi_mach *alt_machines;
+
+	bool use_acpi_target_states;
+
+	/* Platform resource indexes in BAR / ACPI resources. */
+	/* Must set to -1 if not used - add new items to end */
+	int resindex_lpe_base;
+	int resindex_pcicfg_base;
+	int resindex_imr_base;
+	int irqindex_host_ipc;
+
+	/* IPC timeouts in ms */
+	int ipc_timeout;
+	int boot_timeout;
+
+	/* chip information for dsp */
+	const void *chip_info;
+
+	/* defaults for no codec mode */
+	const char *nocodec_tplg_filename;
+
+	/* information on supported IPCs */
+	unsigned int ipc_supported_mask;
+	enum sof_ipc_type ipc_default;
+
+	/* The platform supports DSPless mode */
+	bool dspless_mode_supported;
+
+	/* On demand DSP booting is possible on the platform */
+	bool on_demand_dsp_boot;
+
+	/* defaults paths for firmware, library and topology files */
+	const char *default_fw_path[SOF_IPC_TYPE_COUNT];
+	const char *default_lib_path[SOF_IPC_TYPE_COUNT];
+	const char *default_tplg_path[SOF_IPC_TYPE_COUNT];
+
+	/* default firmware name */
+	const char *default_fw_filename[SOF_IPC_TYPE_COUNT];
+
+	const struct snd_sof_dsp_ops *ops;
+	int (*ops_init)(struct snd_sof_dev *sdev);
+	void (*ops_free)(struct snd_sof_dev *sdev);
+};
+
+int sof_dai_get_mclk(struct snd_soc_pcm_runtime *rtd);
+int sof_dai_get_bclk(struct snd_soc_pcm_runtime *rtd);
+int sof_dai_get_tdm_slots(struct snd_soc_pcm_runtime *rtd);
+
+#endif