[linux-block.git] / drivers / base / cpu.c

// SPDX-License-Identifier: GPL-2.0
/*
 * CPU subsystem support
 */

#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/sched.h>
#include <linux/cpu.h>
#include <linux/topology.h>
#include <linux/device.h>
#include <linux/node.h>
#include <linux/gfp.h>
#include <linux/slab.h>
#include <linux/percpu.h>
#include <linux/acpi.h>
#include <linux/of.h>
#include <linux/cpufeature.h>
#include <linux/tick.h>
#include <linux/pm_qos.h>
#include <linux/sched/isolation.h>

#include "base.h"

static DEFINE_PER_CPU(struct device *, cpu_sys_devices);

static int cpu_subsys_match(struct device *dev, struct device_driver *drv)
{
	/* ACPI style match is the only one that may succeed. */
	if (acpi_driver_match_device(dev, drv))
		return 1;

	return 0;
}

#ifdef CONFIG_HOTPLUG_CPU
static void change_cpu_under_node(struct cpu *cpu,
			unsigned int from_nid, unsigned int to_nid)
{
	int cpuid = cpu->dev.id;
	unregister_cpu_under_node(cpuid, from_nid);
	register_cpu_under_node(cpuid, to_nid);
	cpu->node_id = to_nid;
}

static int cpu_subsys_online(struct device *dev)
{
	struct cpu *cpu = container_of(dev, struct cpu, dev);
	int cpuid = dev->id;
	int from_nid, to_nid;
	int ret;

	from_nid = cpu_to_node(cpuid);
	if (from_nid == NUMA_NO_NODE)
		return -ENODEV;

	ret = cpu_device_up(dev);
	/*
	 * When hot adding memory to memoryless node and enabling a cpu
	 * on the node, node number of the cpu may internally change.
	 */
	to_nid = cpu_to_node(cpuid);
	if (from_nid != to_nid)
		change_cpu_under_node(cpu, from_nid, to_nid);

	return ret;
}

static int cpu_subsys_offline(struct device *dev)
{
	return cpu_device_down(dev);
}

void unregister_cpu(struct cpu *cpu)
{
	int logical_cpu = cpu->dev.id;

	unregister_cpu_under_node(logical_cpu, cpu_to_node(logical_cpu));

	device_unregister(&cpu->dev);
	per_cpu(cpu_sys_devices, logical_cpu) = NULL;
	return;
}

#ifdef CONFIG_ARCH_CPU_PROBE_RELEASE
static ssize_t cpu_probe_store(struct device *dev,
			       struct device_attribute *attr,
			       const char *buf,
			       size_t count)
{
	ssize_t cnt;
	int ret;

	ret = lock_device_hotplug_sysfs();
	if (ret)
		return ret;

	cnt = arch_cpu_probe(buf, count);

	unlock_device_hotplug();
	return cnt;
}

static ssize_t cpu_release_store(struct device *dev,
				 struct device_attribute *attr,
				 const char *buf,
				 size_t count)
{
	ssize_t cnt;
	int ret;

	ret = lock_device_hotplug_sysfs();
	if (ret)
		return ret;

	cnt = arch_cpu_release(buf, count);

	unlock_device_hotplug();
	return cnt;
}

static DEVICE_ATTR(probe, S_IWUSR, NULL, cpu_probe_store);
static DEVICE_ATTR(release, S_IWUSR, NULL, cpu_release_store);
#endif /* CONFIG_ARCH_CPU_PROBE_RELEASE */
#endif /* CONFIG_HOTPLUG_CPU */

#ifdef CONFIG_KEXEC
#include <linux/kexec.h>

static ssize_t crash_notes_show(struct device *dev,
				struct device_attribute *attr,
				char *buf)
{
	struct cpu *cpu = container_of(dev, struct cpu, dev);
	unsigned long long addr;
	int cpunum;

	cpunum = cpu->dev.id;

	/*
	 * Might be reading other cpu's data based on which cpu read thread
	 * has been scheduled. But cpu data (memory) is allocated once during
	 * boot up and this data does not change there after. Hence this
	 * operation should be safe. No locking required.
	 */
	addr = per_cpu_ptr_to_phys(per_cpu_ptr(crash_notes, cpunum));

	return sysfs_emit(buf, "%llx\n", addr);
}
static DEVICE_ATTR_ADMIN_RO(crash_notes);

static ssize_t crash_notes_size_show(struct device *dev,
				     struct device_attribute *attr,
				     char *buf)
{
	return sysfs_emit(buf, "%zu\n", sizeof(note_buf_t));
}
static DEVICE_ATTR_ADMIN_RO(crash_notes_size);

static struct attribute *crash_note_cpu_attrs[] = {
	&dev_attr_crash_notes.attr,
	&dev_attr_crash_notes_size.attr,
	NULL
};

static const struct attribute_group crash_note_cpu_attr_group = {
	.attrs = crash_note_cpu_attrs,
};
#endif

static const struct attribute_group *common_cpu_attr_groups[] = {
#ifdef CONFIG_KEXEC
	&crash_note_cpu_attr_group,
#endif
	NULL
};

static const struct attribute_group *hotplugable_cpu_attr_groups[] = {
#ifdef CONFIG_KEXEC
	&crash_note_cpu_attr_group,
#endif
	NULL
};

/*
 * Print cpu online, possible, present, and system maps
 */

struct cpu_attr {
	struct device_attribute attr;
	const struct cpumask *const map;
};

static ssize_t show_cpus_attr(struct device *dev,
			      struct device_attribute *attr,
			      char *buf)
{
	struct cpu_attr *ca = container_of(attr, struct cpu_attr, attr);

	return cpumap_print_to_pagebuf(true, buf, ca->map);
}

#define _CPU_ATTR(name, map) \
	{ __ATTR(name, 0444, show_cpus_attr, NULL), map }

/* Keep in sync with cpu_subsys_attrs */
static struct cpu_attr cpu_attrs[] = {
	_CPU_ATTR(online, &__cpu_online_mask),
	_CPU_ATTR(possible, &__cpu_possible_mask),
	_CPU_ATTR(present, &__cpu_present_mask),
};

/*
 * Print values for NR_CPUS and offlined cpus
 */
static ssize_t print_cpus_kernel_max(struct device *dev,
				     struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "%d\n", NR_CPUS - 1);
}
static DEVICE_ATTR(kernel_max, 0444, print_cpus_kernel_max, NULL);

/* arch-optional setting to enable display of offline cpus >= nr_cpu_ids */
unsigned int total_cpus;

static ssize_t print_cpus_offline(struct device *dev,
				  struct device_attribute *attr, char *buf)
{
	int len = 0;
	cpumask_var_t offline;

	/* display offline cpus < nr_cpu_ids */
	if (!alloc_cpumask_var(&offline, GFP_KERNEL))
		return -ENOMEM;
	cpumask_andnot(offline, cpu_possible_mask, cpu_online_mask);
	len += sysfs_emit_at(buf, len, "%*pbl", cpumask_pr_args(offline));
	free_cpumask_var(offline);

	/* display offline cpus >= nr_cpu_ids */
	if (total_cpus && nr_cpu_ids < total_cpus) {
		len += sysfs_emit_at(buf, len, ",");

		if (nr_cpu_ids == total_cpus-1)
			len += sysfs_emit_at(buf, len, "%u", nr_cpu_ids);
		else
			len += sysfs_emit_at(buf, len, "%u-%d",
					     nr_cpu_ids, total_cpus - 1);
	}

	len += sysfs_emit_at(buf, len, "\n");

	return len;
}
static DEVICE_ATTR(offline, 0444, print_cpus_offline, NULL);

static ssize_t print_cpus_isolated(struct device *dev,
				  struct device_attribute *attr, char *buf)
{
	int len;
	cpumask_var_t isolated;

	if (!alloc_cpumask_var(&isolated, GFP_KERNEL))
		return -ENOMEM;

	cpumask_andnot(isolated, cpu_possible_mask,
		       housekeeping_cpumask(HK_TYPE_DOMAIN));
	len = sysfs_emit(buf, "%*pbl\n", cpumask_pr_args(isolated));

	free_cpumask_var(isolated);

	return len;
}
static DEVICE_ATTR(isolated, 0444, print_cpus_isolated, NULL);

#ifdef CONFIG_NO_HZ_FULL
static ssize_t print_cpus_nohz_full(struct device *dev,
				    struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "%*pbl\n", cpumask_pr_args(tick_nohz_full_mask));
}
static DEVICE_ATTR(nohz_full, 0444, print_cpus_nohz_full, NULL);
#endif

static void cpu_device_release(struct device *dev)
{
	/*
	 * This is an empty function to prevent the driver core from spitting a
	 * warning at us.  Yes, I know this is directly opposite of what the
	 * documentation for the driver core and kobjects say, and the author
	 * of this code has already been publically ridiculed for doing
	 * something as foolish as this.  However, at this point in time, it is
	 * the only way to handle the issue of statically allocated cpu
	 * devices.  The different architectures will have their cpu device
	 * code reworked to properly handle this in the near future, so this
	 * function will then be changed to correctly free up the memory held
	 * by the cpu device.
	 *
	 * Never copy this way of doing things, or you too will be made fun of
	 * on the linux-kernel list, you have been warned.
	 */
}

#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
static ssize_t print_cpu_modalias(struct device *dev,
				  struct device_attribute *attr,
				  char *buf)
{
	int len = 0;
	u32 i;

	len += sysfs_emit_at(buf, len,
			     "cpu:type:" CPU_FEATURE_TYPEFMT ":feature:",
			     CPU_FEATURE_TYPEVAL);

	for (i = 0; i < MAX_CPU_FEATURES; i++)
		if (cpu_have_feature(i)) {
			if (len + sizeof(",XXXX\n") >= PAGE_SIZE) {
				WARN(1, "CPU features overflow page\n");
				break;
			}
			len += sysfs_emit_at(buf, len, ",%04X", i);
		}
	len += sysfs_emit_at(buf, len, "\n");
	return len;
}

static int cpu_uevent(const struct device *dev, struct kobj_uevent_env *env)
{
	char *buf = kzalloc(PAGE_SIZE, GFP_KERNEL);
	if (buf) {
		print_cpu_modalias(NULL, NULL, buf);
		add_uevent_var(env, "MODALIAS=%s", buf);
		kfree(buf);
	}
	return 0;
}
#endif

struct bus_type cpu_subsys = {
	.name = "cpu",
	.dev_name = "cpu",
	.match = cpu_subsys_match,
#ifdef CONFIG_HOTPLUG_CPU
	.online = cpu_subsys_online,
	.offline = cpu_subsys_offline,
#endif
#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
	.uevent = cpu_uevent,
#endif
};
EXPORT_SYMBOL_GPL(cpu_subsys);

/*
 * register_cpu - Setup a sysfs device for a CPU.
 * @cpu - cpu->hotpluggable field set to 1 will generate a control file in
 *	  sysfs for this CPU.
 * @num - CPU number to use when creating the device.
 *
 * Initialize and register the CPU device.
 */
int register_cpu(struct cpu *cpu, int num)
{
	int error;

	cpu->node_id = cpu_to_node(num);
	memset(&cpu->dev, 0x00, sizeof(struct device));
	cpu->dev.id = num;
	cpu->dev.bus = &cpu_subsys;
	cpu->dev.release = cpu_device_release;
	cpu->dev.offline_disabled = !cpu->hotpluggable;
	cpu->dev.offline = !cpu_online(num);
	cpu->dev.of_node = of_get_cpu_node(num, NULL);
	cpu->dev.groups = common_cpu_attr_groups;
	if (cpu->hotpluggable)
		cpu->dev.groups = hotplugable_cpu_attr_groups;
	error = device_register(&cpu->dev);
	if (error) {
		put_device(&cpu->dev);
		return error;
	}

	per_cpu(cpu_sys_devices, num) = &cpu->dev;
	register_cpu_under_node(num, cpu_to_node(num));
	dev_pm_qos_expose_latency_limit(&cpu->dev,
					PM_QOS_RESUME_LATENCY_NO_CONSTRAINT);

	return 0;
}

struct device *get_cpu_device(unsigned int cpu)
{
	if (cpu < nr_cpu_ids && cpu_possible(cpu))
		return per_cpu(cpu_sys_devices, cpu);
	else
		return NULL;
}
EXPORT_SYMBOL_GPL(get_cpu_device);

static void device_create_release(struct device *dev)
{
	kfree(dev);
}

__printf(4, 0)
static struct device *
__cpu_device_create(struct device *parent, void *drvdata,
		    const struct attribute_group **groups,
		    const char *fmt, va_list args)
{
	struct device *dev = NULL;
	int retval = -ENOMEM;

	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
	if (!dev)
		goto error;

	device_initialize(dev);
	dev->parent = parent;
	dev->groups = groups;
	dev->release = device_create_release;
	device_set_pm_not_required(dev);
	dev_set_drvdata(dev, drvdata);

	retval = kobject_set_name_vargs(&dev->kobj, fmt, args);
	if (retval)
		goto error;

	retval = device_add(dev);
	if (retval)
		goto error;

	return dev;

error:
	put_device(dev);
	return ERR_PTR(retval);
}

struct device *cpu_device_create(struct device *parent, void *drvdata,
				 const struct attribute_group **groups,
				 const char *fmt, ...)
{
	va_list vargs;
	struct device *dev;

	va_start(vargs, fmt);
	dev = __cpu_device_create(parent, drvdata, groups, fmt, vargs);
	va_end(vargs);
	return dev;
}
EXPORT_SYMBOL_GPL(cpu_device_create);

#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
static DEVICE_ATTR(modalias, 0444, print_cpu_modalias, NULL);
#endif

static struct attribute *cpu_root_attrs[] = {
#ifdef CONFIG_ARCH_CPU_PROBE_RELEASE
	&dev_attr_probe.attr,
	&dev_attr_release.attr,
#endif
	&cpu_attrs[0].attr.attr,
	&cpu_attrs[1].attr.attr,
	&cpu_attrs[2].attr.attr,
	&dev_attr_kernel_max.attr,
	&dev_attr_offline.attr,
	&dev_attr_isolated.attr,
#ifdef CONFIG_NO_HZ_FULL
	&dev_attr_nohz_full.attr,
#endif
#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
	&dev_attr_modalias.attr,
#endif
	NULL
};

static const struct attribute_group cpu_root_attr_group = {
	.attrs = cpu_root_attrs,
};

static const struct attribute_group *cpu_root_attr_groups[] = {
	&cpu_root_attr_group,
	NULL,
};

bool cpu_is_hotpluggable(unsigned int cpu)
{
	struct device *dev = get_cpu_device(cpu);
	return dev && container_of(dev, struct cpu, dev)->hotpluggable
		&& tick_nohz_cpu_hotpluggable(cpu);
}
EXPORT_SYMBOL_GPL(cpu_is_hotpluggable);

#ifdef CONFIG_GENERIC_CPU_DEVICES
static DEFINE_PER_CPU(struct cpu, cpu_devices);
#endif

static void __init cpu_dev_register_generic(void)
{
#ifdef CONFIG_GENERIC_CPU_DEVICES
	int i;

	for_each_possible_cpu(i) {
		if (register_cpu(&per_cpu(cpu_devices, i), i))
			panic("Failed to register CPU device");
	}
#endif
}

#ifdef CONFIG_GENERIC_CPU_VULNERABILITIES

ssize_t __weak cpu_show_meltdown(struct device *dev,
				 struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "Not affected\n");
}

ssize_t __weak cpu_show_spectre_v1(struct device *dev,
				   struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "Not affected\n");
}

ssize_t __weak cpu_show_spectre_v2(struct device *dev,
				   struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "Not affected\n");
}

ssize_t __weak cpu_show_spec_store_bypass(struct device *dev,
					  struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "Not affected\n");
}

ssize_t __weak cpu_show_l1tf(struct device *dev,
			     struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "Not affected\n");
}

ssize_t __weak cpu_show_mds(struct device *dev,
			    struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "Not affected\n");
}

ssize_t __weak cpu_show_tsx_async_abort(struct device *dev,
					struct device_attribute *attr,
					char *buf)
{
	return sysfs_emit(buf, "Not affected\n");
}

ssize_t __weak cpu_show_itlb_multihit(struct device *dev,
				      struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "Not affected\n");
}

ssize_t __weak cpu_show_srbds(struct device *dev,
			      struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "Not affected\n");
}

ssize_t __weak cpu_show_mmio_stale_data(struct device *dev,
					struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "Not affected\n");
}

ssize_t __weak cpu_show_retbleed(struct device *dev,
				 struct device_attribute *attr, char *buf)
{
	return sysfs_emit(buf, "Not affected\n");
}

static DEVICE_ATTR(meltdown, 0444, cpu_show_meltdown, NULL);
static DEVICE_ATTR(spectre_v1, 0444, cpu_show_spectre_v1, NULL);
static DEVICE_ATTR(spectre_v2, 0444, cpu_show_spectre_v2, NULL);
static DEVICE_ATTR(spec_store_bypass, 0444, cpu_show_spec_store_bypass, NULL);
static DEVICE_ATTR(l1tf, 0444, cpu_show_l1tf, NULL);
static DEVICE_ATTR(mds, 0444, cpu_show_mds, NULL);
static DEVICE_ATTR(tsx_async_abort, 0444, cpu_show_tsx_async_abort, NULL);
static DEVICE_ATTR(itlb_multihit, 0444, cpu_show_itlb_multihit, NULL);
static DEVICE_ATTR(srbds, 0444, cpu_show_srbds, NULL);
static DEVICE_ATTR(mmio_stale_data, 0444, cpu_show_mmio_stale_data, NULL);
static DEVICE_ATTR(retbleed, 0444, cpu_show_retbleed, NULL);

static struct attribute *cpu_root_vulnerabilities_attrs[] = {
	&dev_attr_meltdown.attr,
	&dev_attr_spectre_v1.attr,
	&dev_attr_spectre_v2.attr,
	&dev_attr_spec_store_bypass.attr,
	&dev_attr_l1tf.attr,
	&dev_attr_mds.attr,
	&dev_attr_tsx_async_abort.attr,
	&dev_attr_itlb_multihit.attr,
	&dev_attr_srbds.attr,
	&dev_attr_mmio_stale_data.attr,
	&dev_attr_retbleed.attr,
	NULL
};

static const struct attribute_group cpu_root_vulnerabilities_group = {
	.name  = "vulnerabilities",
	.attrs = cpu_root_vulnerabilities_attrs,
};

static void __init cpu_register_vulnerabilities(void)
{
	struct device *dev = bus_get_dev_root(&cpu_subsys);

	if (dev) {
		if (sysfs_create_group(&dev->kobj, &cpu_root_vulnerabilities_group))
			pr_err("Unable to register CPU vulnerabilities\n");
		put_device(dev);
	}
}

#else
static inline void cpu_register_vulnerabilities(void) { }
#endif

void __init cpu_dev_init(void)
{
	if (subsys_system_register(&cpu_subsys, cpu_root_attr_groups))
		panic("Failed to register CPU subsystem");

	cpu_dev_register_generic();
	cpu_register_vulnerabilities();
}
Commit	Line	Data
989d42e8	1	// SPDX-License-Identifier: GPL-2.0
1da177e4	2	/*
8a25a2fd	3	* CPU subsystem support
1da177e4 LT	4	*/
1da177e4 LT	5
024f7846	6	#include <linux/kernel.h>
1da177e4 LT	7	#include <linux/module.h>
1da177e4 LT	8	#include <linux/init.h>
f6a57033	9	#include <linux/sched.h>
1da177e4 LT	10	#include <linux/cpu.h>
	11	#include <linux/topology.h>
	12	#include <linux/device.h>
76b67ed9	13	#include <linux/node.h>
5a0e3ad6	14	#include <linux/gfp.h>
fad12ac8	15	#include <linux/slab.h>
9f13a1fd	16	#include <linux/percpu.h>
ac212b69	17	#include <linux/acpi.h>
f86e4718	18	#include <linux/of.h>
67bad2fd	19	#include <linux/cpufeature.h>
6570a9a1	20	#include <linux/tick.h>
37efa4b4	21	#include <linux/pm_qos.h>
edb93821	22	#include <linux/sched/isolation.h>
1da177e4	23
a1bdc7aa	24	#include "base.h"
1da177e4	25
8a25a2fd	26	static DEFINE_PER_CPU(struct device *, cpu_sys_devices);
ad74557a	27
ac212b69 RW	28	static int cpu_subsys_match(struct device dev, struct device_driver drv)
	29	{
	30	/* ACPI style match is the only one that may succeed. */
	31	if (acpi_driver_match_device(dev, drv))
	32	return 1;
	33
	34	return 0;
	35	}
	36
1da177e4	37	#ifdef CONFIG_HOTPLUG_CPU
34640468 YI	38	static void change_cpu_under_node(struct cpu *cpu,
	39	unsigned int from_nid, unsigned int to_nid)
	40	{
	41	int cpuid = cpu->dev.id;
	42	unregister_cpu_under_node(cpuid, from_nid);
	43	register_cpu_under_node(cpuid, to_nid);
	44	cpu->node_id = to_nid;
	45	}
	46
eda5867b	47	static int cpu_subsys_online(struct device *dev)
1da177e4	48	{
8a25a2fd	49	struct cpu *cpu = container_of(dev, struct cpu, dev);
0902a904 RW	50	int cpuid = dev->id;
0902a904 RW	51	int from_nid, to_nid;
6dedcca6	52	int ret;
1da177e4	53
0902a904	54	from_nid = cpu_to_node(cpuid);
c7991b0b	55	if (from_nid == NUMA_NO_NODE)
6dedcca6	56	return -ENODEV;
c7991b0b	57
33c3736e	58	ret = cpu_device_up(dev);
0902a904 RW	59	/*
	60	* When hot adding memory to memoryless node and enabling a cpu
	61	* on the node, node number of the cpu may internally change.
	62	*/
	63	to_nid = cpu_to_node(cpuid);
	64	if (from_nid != to_nid)
	65	change_cpu_under_node(cpu, from_nid, to_nid);
1da177e4	66
0902a904	67	return ret;
1da177e4 LT	68	}
1da177e4 LT	69
0902a904	70	static int cpu_subsys_offline(struct device *dev)
1da177e4	71	{
33c3736e	72	return cpu_device_down(dev);
1da177e4	73	}
1c4e2d70	74
76b67ed9	75	void unregister_cpu(struct cpu *cpu)
1da177e4	76	{
8a25a2fd	77	int logical_cpu = cpu->dev.id;
1da177e4	78
76b67ed9 KH	79	unregister_cpu_under_node(logical_cpu, cpu_to_node(logical_cpu));
76b67ed9 KH	80
8a25a2fd	81	device_unregister(&cpu->dev);
e37d05da	82	per_cpu(cpu_sys_devices, logical_cpu) = NULL;
1da177e4 LT	83	return;
1da177e4 LT	84	}
12633e80 NF	85
12633e80 NF	86	#ifdef CONFIG_ARCH_CPU_PROBE_RELEASE
8a25a2fd KS	87	static ssize_t cpu_probe_store(struct device *dev,
8a25a2fd KS	88	struct device_attribute *attr,
67fc233f	89	const char *buf,
12633e80 NF	90	size_t count)
12633e80 NF	91	{
574b851e TK	92	ssize_t cnt;
	93	int ret;
	94
	95	ret = lock_device_hotplug_sysfs();
	96	if (ret)
	97	return ret;
	98
	99	cnt = arch_cpu_probe(buf, count);
	100
	101	unlock_device_hotplug();
	102	return cnt;
12633e80 NF	103	}
12633e80 NF	104
8a25a2fd KS	105	static ssize_t cpu_release_store(struct device *dev,
8a25a2fd KS	106	struct device_attribute *attr,
67fc233f	107	const char *buf,
12633e80 NF	108	size_t count)
12633e80 NF	109	{
574b851e TK	110	ssize_t cnt;
	111	int ret;
	112
	113	ret = lock_device_hotplug_sysfs();
	114	if (ret)
	115	return ret;
	116
	117	cnt = arch_cpu_release(buf, count);
	118
	119	unlock_device_hotplug();
	120	return cnt;
12633e80 NF	121	}
12633e80 NF	122
8a25a2fd KS	123	static DEVICE_ATTR(probe, S_IWUSR, NULL, cpu_probe_store);
8a25a2fd KS	124	static DEVICE_ATTR(release, S_IWUSR, NULL, cpu_release_store);
12633e80	125	#endif /* CONFIG_ARCH_CPU_PROBE_RELEASE */
1da177e4 LT	126	#endif /* CONFIG_HOTPLUG_CPU */
1da177e4 LT	127
51be5606 VG	128	#ifdef CONFIG_KEXEC
	129	#include <linux/kexec.h>
	130
948b3edb JP	131	static ssize_t crash_notes_show(struct device *dev,
948b3edb JP	132	struct device_attribute *attr,
4a0b2b4d	133	char *buf)
51be5606	134	{
8a25a2fd	135	struct cpu *cpu = container_of(dev, struct cpu, dev);
51be5606 VG	136	unsigned long long addr;
	137	int cpunum;
	138
8a25a2fd	139	cpunum = cpu->dev.id;
51be5606 VG	140
	141	/*
	142	* Might be reading other cpu's data based on which cpu read thread
	143	* has been scheduled. But cpu data (memory) is allocated once during
	144	* boot up and this data does not change there after. Hence this
	145	* operation should be safe. No locking required.
	146	*/
3b034b0d	147	addr = per_cpu_ptr_to_phys(per_cpu_ptr(crash_notes, cpunum));
948b3edb JP	148
948b3edb JP	149	return sysfs_emit(buf, "%llx\n", addr);
51be5606	150	}
948b3edb	151	static DEVICE_ATTR_ADMIN_RO(crash_notes);
eca4549f	152
948b3edb	153	static ssize_t crash_notes_size_show(struct device *dev,
eca4549f ZY	154	struct device_attribute *attr,
	155	char *buf)
	156	{
948b3edb	157	return sysfs_emit(buf, "%zu\n", sizeof(note_buf_t));
eca4549f	158	}
948b3edb	159	static DEVICE_ATTR_ADMIN_RO(crash_notes_size);
c055da9f IM	160
	161	static struct attribute *crash_note_cpu_attrs[] = {
	162	&dev_attr_crash_notes.attr,
	163	&dev_attr_crash_notes_size.attr,
	164	NULL
	165	};
	166
5a576764	167	static const struct attribute_group crash_note_cpu_attr_group = {
c055da9f IM	168	.attrs = crash_note_cpu_attrs,
	169	};
	170	#endif
	171
	172	static const struct attribute_group *common_cpu_attr_groups[] = {
	173	#ifdef CONFIG_KEXEC
	174	&crash_note_cpu_attr_group,
51be5606	175	#endif
c055da9f IM	176	NULL
c055da9f IM	177	};
51be5606	178
1c4e2d70 IM	179	static const struct attribute_group *hotplugable_cpu_attr_groups[] = {
	180	#ifdef CONFIG_KEXEC
	181	&crash_note_cpu_attr_group,
1c4e2d70 IM	182	#endif
	183	NULL
	184	};
	185
9d1fe323 MT	186	/*
	187	* Print cpu online, possible, present, and system maps
	188	*/
265d2e2e AK	189
265d2e2e AK	190	struct cpu_attr {
8a25a2fd	191	struct device_attribute attr;
848e2391	192	const struct cpumask *const map;
265d2e2e AK	193	};
265d2e2e AK	194
8a25a2fd KS	195	static ssize_t show_cpus_attr(struct device *dev,
8a25a2fd KS	196	struct device_attribute *attr,
265d2e2e	197	char *buf)
9d1fe323	198	{
265d2e2e	199	struct cpu_attr *ca = container_of(attr, struct cpu_attr, attr);
9d1fe323	200
848e2391	201	return cpumap_print_to_pagebuf(true, buf, ca->map);
9d1fe323 MT	202	}
9d1fe323 MT	203
8a25a2fd KS	204	#define _CPU_ATTR(name, map) \
8a25a2fd KS	205	{ __ATTR(name, 0444, show_cpus_attr, NULL), map }
9d1fe323	206
8a25a2fd	207	/* Keep in sync with cpu_subsys_attrs */
265d2e2e	208	static struct cpu_attr cpu_attrs[] = {
848e2391 RV	209	_CPU_ATTR(online, &__cpu_online_mask),
	210	_CPU_ATTR(possible, &__cpu_possible_mask),
	211	_CPU_ATTR(present, &__cpu_present_mask),
265d2e2e	212	};
9d1fe323	213
e057d7ae MT	214	/*
	215	* Print values for NR_CPUS and offlined cpus
	216	*/
8a25a2fd KS	217	static ssize_t print_cpus_kernel_max(struct device *dev,
8a25a2fd KS	218	struct device_attribute attr, char buf)
e057d7ae	219	{
aa838896	220	return sysfs_emit(buf, "%d\n", NR_CPUS - 1);
e057d7ae	221	}
8a25a2fd	222	static DEVICE_ATTR(kernel_max, 0444, print_cpus_kernel_max, NULL);
e057d7ae MT	223
	224	/* arch-optional setting to enable display of offline cpus >= nr_cpu_ids */
	225	unsigned int total_cpus;
	226
8a25a2fd KS	227	static ssize_t print_cpus_offline(struct device *dev,
8a25a2fd KS	228	struct device_attribute attr, char buf)
e057d7ae	229	{
948b3edb	230	int len = 0;
e057d7ae MT	231	cpumask_var_t offline;
	232
	233	/* display offline cpus < nr_cpu_ids */
	234	if (!alloc_cpumask_var(&offline, GFP_KERNEL))
	235	return -ENOMEM;
cdc6e3d3	236	cpumask_andnot(offline, cpu_possible_mask, cpu_online_mask);
948b3edb	237	len += sysfs_emit_at(buf, len, "%*pbl", cpumask_pr_args(offline));
e057d7ae MT	238	free_cpumask_var(offline);
	239
	240	/* display offline cpus >= nr_cpu_ids */
	241	if (total_cpus && nr_cpu_ids < total_cpus) {
948b3edb	242	len += sysfs_emit_at(buf, len, ",");
e057d7ae MT	243
e057d7ae MT	244	if (nr_cpu_ids == total_cpus-1)
948b3edb	245	len += sysfs_emit_at(buf, len, "%u", nr_cpu_ids);
e057d7ae	246	else
948b3edb JP	247	len += sysfs_emit_at(buf, len, "%u-%d",
948b3edb JP	248	nr_cpu_ids, total_cpus - 1);
e057d7ae MT	249	}
e057d7ae MT	250
948b3edb JP	251	len += sysfs_emit_at(buf, len, "\n");
	252
	253	return len;
e057d7ae	254	}
8a25a2fd	255	static DEVICE_ATTR(offline, 0444, print_cpus_offline, NULL);
e057d7ae	256
59f30abe RR	257	static ssize_t print_cpus_isolated(struct device *dev,
	258	struct device_attribute attr, char buf)
	259	{
948b3edb	260	int len;
edb93821	261	cpumask_var_t isolated;
59f30abe	262
edb93821 FW	263	if (!alloc_cpumask_var(&isolated, GFP_KERNEL))
	264	return -ENOMEM;
	265
	266	cpumask_andnot(isolated, cpu_possible_mask,
04d4e665	267	housekeeping_cpumask(HK_TYPE_DOMAIN));
948b3edb	268	len = sysfs_emit(buf, "%*pbl\n", cpumask_pr_args(isolated));
edb93821 FW	269
edb93821 FW	270	free_cpumask_var(isolated);
59f30abe	271
948b3edb	272	return len;
59f30abe RR	273	}
	274	static DEVICE_ATTR(isolated, 0444, print_cpus_isolated, NULL);
	275
6570a9a1 RR	276	#ifdef CONFIG_NO_HZ_FULL
6570a9a1 RR	277	static ssize_t print_cpus_nohz_full(struct device *dev,
948b3edb	278	struct device_attribute attr, char buf)
6570a9a1	279	{
aa838896	280	return sysfs_emit(buf, "%*pbl\n", cpumask_pr_args(tick_nohz_full_mask));
6570a9a1 RR	281	}
	282	static DEVICE_ATTR(nohz_full, 0444, print_cpus_nohz_full, NULL);
	283	#endif
	284
2885e25c GKH	285	static void cpu_device_release(struct device *dev)
	286	{
	287	/*
	288	* This is an empty function to prevent the driver core from spitting a
	289	* warning at us. Yes, I know this is directly opposite of what the
	290	* documentation for the driver core and kobjects say, and the author
	291	* of this code has already been publically ridiculed for doing
	292	* something as foolish as this. However, at this point in time, it is
	293	* the only way to handle the issue of statically allocated cpu
	294	* devices. The different architectures will have their cpu device
	295	* code reworked to properly handle this in the near future, so this
	296	* function will then be changed to correctly free up the memory held
	297	* by the cpu device.
	298	*
	299	* Never copy this way of doing things, or you too will be made fun of
30a4840a	300	* on the linux-kernel list, you have been warned.
2885e25c GKH	301	*/
	302	}
	303
67bad2fd AB	304	#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
	305	static ssize_t print_cpu_modalias(struct device *dev,
	306	struct device_attribute *attr,
	307	char *buf)
	308	{
948b3edb	309	int len = 0;
67bad2fd AB	310	u32 i;
67bad2fd AB	311
948b3edb JP	312	len += sysfs_emit_at(buf, len,
	313	"cpu:type:" CPU_FEATURE_TYPEFMT ":feature:",
	314	CPU_FEATURE_TYPEVAL);
67bad2fd AB	315
	316	for (i = 0; i < MAX_CPU_FEATURES; i++)
	317	if (cpu_have_feature(i)) {
948b3edb	318	if (len + sizeof(",XXXX\n") >= PAGE_SIZE) {
67bad2fd AB	319	WARN(1, "CPU features overflow page\n");
	320	break;
	321	}
948b3edb	322	len += sysfs_emit_at(buf, len, ",%04X", i);
67bad2fd	323	}
948b3edb JP	324	len += sysfs_emit_at(buf, len, "\n");
948b3edb JP	325	return len;
67bad2fd	326	}
67bad2fd	327
2a81ada3	328	static int cpu_uevent(const struct device dev, struct kobj_uevent_env env)
67bad2fd AB	329	{
	330	char *buf = kzalloc(PAGE_SIZE, GFP_KERNEL);
	331	if (buf) {
	332	print_cpu_modalias(NULL, NULL, buf);
	333	add_uevent_var(env, "MODALIAS=%s", buf);
	334	kfree(buf);
	335	}
	336	return 0;
	337	}
	338	#endif
	339
2bc19066 GKH	340	struct bus_type cpu_subsys = {
	341	.name = "cpu",
	342	.dev_name = "cpu",
	343	.match = cpu_subsys_match,
	344	#ifdef CONFIG_HOTPLUG_CPU
	345	.online = cpu_subsys_online,
	346	.offline = cpu_subsys_offline,
	347	#endif
	348	#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
	349	.uevent = cpu_uevent,
	350	#endif
	351	};
	352	EXPORT_SYMBOL_GPL(cpu_subsys);
	353
1da177e4	354	/*
405ae7d3	355	* register_cpu - Setup a sysfs device for a CPU.
72486f1f SS	356	* @cpu - cpu->hotpluggable field set to 1 will generate a control file in
72486f1f SS	357	* sysfs for this CPU.
1da177e4 LT	358	* @num - CPU number to use when creating the device.
	359	*
	360	* Initialize and register the CPU device.
	361	*/
a83048eb	362	int register_cpu(struct cpu *cpu, int num)
1da177e4 LT	363	{
1da177e4 LT	364	int error;
76b67ed9	365
8a25a2fd	366	cpu->node_id = cpu_to_node(num);
29bb5d4f	367	memset(&cpu->dev, 0x00, sizeof(struct device));
8a25a2fd KS	368	cpu->dev.id = num;
8a25a2fd KS	369	cpu->dev.bus = &cpu_subsys;
2885e25c	370	cpu->dev.release = cpu_device_release;
0902a904	371	cpu->dev.offline_disabled = !cpu->hotpluggable;
1001b4d4	372	cpu->dev.offline = !cpu_online(num);
f86e4718	373	cpu->dev.of_node = of_get_cpu_node(num, NULL);
c055da9f	374	cpu->dev.groups = common_cpu_attr_groups;
1c4e2d70 IM	375	if (cpu->hotpluggable)
1c4e2d70 IM	376	cpu->dev.groups = hotplugable_cpu_attr_groups;
8a25a2fd	377	error = device_register(&cpu->dev);
3aaba245 AY	378	if (error) {
3aaba245 AY	379	put_device(&cpu->dev);
59fffa34	380	return error;
3aaba245	381	}
51be5606	382
59fffa34 AS	383	per_cpu(cpu_sys_devices, num) = &cpu->dev;
59fffa34 AS	384	register_cpu_under_node(num, cpu_to_node(num));
0759e80b RW	385	dev_pm_qos_expose_latency_limit(&cpu->dev,
0759e80b RW	386	PM_QOS_RESUME_LATENCY_NO_CONSTRAINT);
59fffa34 AS	387
59fffa34 AS	388	return 0;
1da177e4 LT	389	}
1da177e4 LT	390
e7deeb9d	391	struct device *get_cpu_device(unsigned int cpu)
ad74557a	392	{
e37d05da MT	393	if (cpu < nr_cpu_ids && cpu_possible(cpu))
e37d05da MT	394	return per_cpu(cpu_sys_devices, cpu);
ad74557a AR	395	else
	396	return NULL;
	397	}
8a25a2fd KS	398	EXPORT_SYMBOL_GPL(get_cpu_device);
8a25a2fd KS	399
3d52943b SH	400	static void device_create_release(struct device *dev)
	401	{
	402	kfree(dev);
	403	}
	404
fa548d79	405	__printf(4, 0)
3d52943b SH	406	static struct device *
	407	__cpu_device_create(struct device parent, void drvdata,
	408	const struct attribute_group **groups,
	409	const char *fmt, va_list args)
	410	{
	411	struct device *dev = NULL;
6b72cf12	412	int retval = -ENOMEM;
3d52943b SH	413
3d52943b SH	414	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
6b72cf12	415	if (!dev)
3d52943b	416	goto error;
3d52943b SH	417
	418	device_initialize(dev);
	419	dev->parent = parent;
	420	dev->groups = groups;
	421	dev->release = device_create_release;
85945c28	422	device_set_pm_not_required(dev);
3d52943b SH	423	dev_set_drvdata(dev, drvdata);
	424
	425	retval = kobject_set_name_vargs(&dev->kobj, fmt, args);
	426	if (retval)
	427	goto error;
	428
	429	retval = device_add(dev);
	430	if (retval)
	431	goto error;
	432
	433	return dev;
	434
	435	error:
	436	put_device(dev);
	437	return ERR_PTR(retval);
	438	}
	439
	440	struct device cpu_device_create(struct device parent, void *drvdata,
	441	const struct attribute_group **groups,
	442	const char *fmt, ...)
	443	{
	444	va_list vargs;
	445	struct device *dev;
	446
	447	va_start(vargs, fmt);
	448	dev = __cpu_device_create(parent, drvdata, groups, fmt, vargs);
	449	va_end(vargs);
	450	return dev;
	451	}
	452	EXPORT_SYMBOL_GPL(cpu_device_create);
	453
2b9c1f03	454	#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
67bad2fd	455	static DEVICE_ATTR(modalias, 0444, print_cpu_modalias, NULL);
fad12ac8 TR	456	#endif
fad12ac8 TR	457
8a25a2fd KS	458	static struct attribute *cpu_root_attrs[] = {
	459	#ifdef CONFIG_ARCH_CPU_PROBE_RELEASE
	460	&dev_attr_probe.attr,
	461	&dev_attr_release.attr,
	462	#endif
	463	&cpu_attrs[0].attr.attr,
	464	&cpu_attrs[1].attr.attr,
	465	&cpu_attrs[2].attr.attr,
	466	&dev_attr_kernel_max.attr,
	467	&dev_attr_offline.attr,
59f30abe	468	&dev_attr_isolated.attr,
6570a9a1 RR	469	#ifdef CONFIG_NO_HZ_FULL
	470	&dev_attr_nohz_full.attr,
	471	#endif
2b9c1f03	472	#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
fad12ac8 TR	473	&dev_attr_modalias.attr,
fad12ac8 TR	474	#endif
8a25a2fd KS	475	NULL
	476	};
	477
5a576764	478	static const struct attribute_group cpu_root_attr_group = {
8a25a2fd KS	479	.attrs = cpu_root_attrs,
	480	};
	481
	482	static const struct attribute_group *cpu_root_attr_groups[] = {
	483	&cpu_root_attr_group,
	484	NULL,
	485	};
1da177e4	486
e7deeb9d	487	bool cpu_is_hotpluggable(unsigned int cpu)
2987557f	488	{
7affca35	489	struct device *dev = get_cpu_device(cpu);
58d76682 JFG	490	return dev && container_of(dev, struct cpu, dev)->hotpluggable
58d76682 JFG	491	&& tick_nohz_cpu_hotpluggable(cpu);
2987557f JT	492	}
	493	EXPORT_SYMBOL_GPL(cpu_is_hotpluggable);
	494
9f13a1fd BH	495	#ifdef CONFIG_GENERIC_CPU_DEVICES
	496	static DEFINE_PER_CPU(struct cpu, cpu_devices);
	497	#endif
	498
	499	static void __init cpu_dev_register_generic(void)
	500	{
	501	#ifdef CONFIG_GENERIC_CPU_DEVICES
	502	int i;
	503
	504	for_each_possible_cpu(i) {
	505	if (register_cpu(&per_cpu(cpu_devices, i), i))
	506	panic("Failed to register CPU device");
	507	}
	508	#endif
	509	}
	510
87590ce6 TG	511	#ifdef CONFIG_GENERIC_CPU_VULNERABILITIES
	512
	513	ssize_t __weak cpu_show_meltdown(struct device *dev,
	514	struct device_attribute attr, char buf)
	515	{
aa838896	516	return sysfs_emit(buf, "Not affected\n");
87590ce6 TG	517	}
	518
	519	ssize_t __weak cpu_show_spectre_v1(struct device *dev,
	520	struct device_attribute attr, char buf)
	521	{
aa838896	522	return sysfs_emit(buf, "Not affected\n");
87590ce6 TG	523	}
	524
	525	ssize_t __weak cpu_show_spectre_v2(struct device *dev,
	526	struct device_attribute attr, char buf)
	527	{
aa838896	528	return sysfs_emit(buf, "Not affected\n");
87590ce6 TG	529	}
87590ce6 TG	530
c456442c KRW	531	ssize_t __weak cpu_show_spec_store_bypass(struct device *dev,
	532	struct device_attribute attr, char buf)
	533	{
aa838896	534	return sysfs_emit(buf, "Not affected\n");
c456442c KRW	535	}
c456442c KRW	536
17dbca11 AK	537	ssize_t __weak cpu_show_l1tf(struct device *dev,
	538	struct device_attribute attr, char buf)
	539	{
aa838896	540	return sysfs_emit(buf, "Not affected\n");
17dbca11 AK	541	}
17dbca11 AK	542
8a4b06d3 TG	543	ssize_t __weak cpu_show_mds(struct device *dev,
	544	struct device_attribute attr, char buf)
	545	{
aa838896	546	return sysfs_emit(buf, "Not affected\n");
8a4b06d3 TG	547	}
8a4b06d3 TG	548
6608b45a PG	549	ssize_t __weak cpu_show_tsx_async_abort(struct device *dev,
	550	struct device_attribute *attr,
	551	char *buf)
	552	{
aa838896	553	return sysfs_emit(buf, "Not affected\n");
6608b45a PG	554	}
6608b45a PG	555
db4d30fb	556	ssize_t __weak cpu_show_itlb_multihit(struct device *dev,
948b3edb	557	struct device_attribute attr, char buf)
db4d30fb	558	{
aa838896	559	return sysfs_emit(buf, "Not affected\n");
db4d30fb VT	560	}
db4d30fb VT	561
7e5b3c26 MG	562	ssize_t __weak cpu_show_srbds(struct device *dev,
	563	struct device_attribute attr, char buf)
	564	{
aa838896	565	return sysfs_emit(buf, "Not affected\n");
7e5b3c26 MG	566	}
7e5b3c26 MG	567
8d50cdf8 PG	568	ssize_t __weak cpu_show_mmio_stale_data(struct device *dev,
	569	struct device_attribute attr, char buf)
	570	{
	571	return sysfs_emit(buf, "Not affected\n");
	572	}
	573
6b80b59b AC	574	ssize_t __weak cpu_show_retbleed(struct device *dev,
	575	struct device_attribute attr, char buf)
	576	{
	577	return sysfs_emit(buf, "Not affected\n");
	578	}
	579
87590ce6 TG	580	static DEVICE_ATTR(meltdown, 0444, cpu_show_meltdown, NULL);
	581	static DEVICE_ATTR(spectre_v1, 0444, cpu_show_spectre_v1, NULL);
	582	static DEVICE_ATTR(spectre_v2, 0444, cpu_show_spectre_v2, NULL);
c456442c	583	static DEVICE_ATTR(spec_store_bypass, 0444, cpu_show_spec_store_bypass, NULL);
17dbca11	584	static DEVICE_ATTR(l1tf, 0444, cpu_show_l1tf, NULL);
8a4b06d3	585	static DEVICE_ATTR(mds, 0444, cpu_show_mds, NULL);
6608b45a	586	static DEVICE_ATTR(tsx_async_abort, 0444, cpu_show_tsx_async_abort, NULL);
db4d30fb	587	static DEVICE_ATTR(itlb_multihit, 0444, cpu_show_itlb_multihit, NULL);
7e5b3c26	588	static DEVICE_ATTR(srbds, 0444, cpu_show_srbds, NULL);
8d50cdf8	589	static DEVICE_ATTR(mmio_stale_data, 0444, cpu_show_mmio_stale_data, NULL);
6b80b59b	590	static DEVICE_ATTR(retbleed, 0444, cpu_show_retbleed, NULL);
87590ce6 TG	591
	592	static struct attribute *cpu_root_vulnerabilities_attrs[] = {
	593	&dev_attr_meltdown.attr,
	594	&dev_attr_spectre_v1.attr,
	595	&dev_attr_spectre_v2.attr,
c456442c	596	&dev_attr_spec_store_bypass.attr,
17dbca11	597	&dev_attr_l1tf.attr,
8a4b06d3	598	&dev_attr_mds.attr,
6608b45a	599	&dev_attr_tsx_async_abort.attr,
db4d30fb	600	&dev_attr_itlb_multihit.attr,
7e5b3c26	601	&dev_attr_srbds.attr,
8d50cdf8	602	&dev_attr_mmio_stale_data.attr,
6b80b59b	603	&dev_attr_retbleed.attr,
87590ce6 TG	604	NULL
	605	};
	606
	607	static const struct attribute_group cpu_root_vulnerabilities_group = {
	608	.name = "vulnerabilities",
	609	.attrs = cpu_root_vulnerabilities_attrs,
	610	};
	611
	612	static void __init cpu_register_vulnerabilities(void)
	613	{
8c99377e GKH	614	struct device *dev = bus_get_dev_root(&cpu_subsys);
	615
	616	if (dev) {
	617	if (sysfs_create_group(&dev->kobj, &cpu_root_vulnerabilities_group))
	618	pr_err("Unable to register CPU vulnerabilities\n");
	619	put_device(dev);
	620	}
87590ce6 TG	621	}
	622
	623	#else
	624	static inline void cpu_register_vulnerabilities(void) { }
	625	#endif
	626
024f7846	627	void __init cpu_dev_init(void)
1da177e4	628	{
024f7846 BH	629	if (subsys_system_register(&cpu_subsys, cpu_root_attr_groups))
024f7846 BH	630	panic("Failed to register CPU subsystem");
8a25a2fd	631
9f13a1fd	632	cpu_dev_register_generic();
87590ce6	633	cpu_register_vulnerabilities();
1da177e4	634	}