[linux-2.6-block.git] / fs / bcachefs / chardev.c

// SPDX-License-Identifier: GPL-2.0
#ifndef NO_BCACHEFS_CHARDEV

#include "bcachefs.h"
#include "bcachefs_ioctl.h"
#include "buckets.h"
#include "chardev.h"
#include "disk_accounting.h"
#include "fsck.h"
#include "journal.h"
#include "move.h"
#include "recovery_passes.h"
#include "replicas.h"
#include "sb-counters.h"
#include "super-io.h"
#include "thread_with_file.h"

#include <linux/cdev.h>
#include <linux/device.h>
#include <linux/fs.h>
#include <linux/ioctl.h>
#include <linux/major.h>
#include <linux/sched/task.h>
#include <linux/slab.h>
#include <linux/uaccess.h>

/* returns with ref on ca->ref */
static struct bch_dev *bch2_device_lookup(struct bch_fs *c, u64 dev,
					  unsigned flags)
{
	struct bch_dev *ca;

	if (flags & BCH_BY_INDEX) {
		if (dev >= c->sb.nr_devices)
			return ERR_PTR(-EINVAL);

		ca = bch2_dev_tryget_noerror(c, dev);
		if (!ca)
			return ERR_PTR(-EINVAL);
	} else {
		char *path;

		path = strndup_user((const char __user *)
				    (unsigned long) dev, PATH_MAX);
		if (IS_ERR(path))
			return ERR_CAST(path);

		ca = bch2_dev_lookup(c, path);
		kfree(path);
	}

	return ca;
}

#if 0
static long bch2_ioctl_assemble(struct bch_ioctl_assemble __user *user_arg)
{
	struct bch_ioctl_assemble arg;
	struct bch_fs *c;
	u64 *user_devs = NULL;
	char **devs = NULL;
	unsigned i;
	int ret = -EFAULT;

	if (copy_from_user(&arg, user_arg, sizeof(arg)))
		return -EFAULT;

	if (arg.flags || arg.pad)
		return -EINVAL;

	user_devs = kmalloc_array(arg.nr_devs, sizeof(u64), GFP_KERNEL);
	if (!user_devs)
		return -ENOMEM;

	devs = kcalloc(arg.nr_devs, sizeof(char *), GFP_KERNEL);

	if (copy_from_user(user_devs, user_arg->devs,
			   sizeof(u64) * arg.nr_devs))
		goto err;

	for (i = 0; i < arg.nr_devs; i++) {
		devs[i] = strndup_user((const char __user *)(unsigned long)
				       user_devs[i],
				       PATH_MAX);
		ret= PTR_ERR_OR_ZERO(devs[i]);
		if (ret)
			goto err;
	}

	c = bch2_fs_open(devs, arg.nr_devs, bch2_opts_empty());
	ret = PTR_ERR_OR_ZERO(c);
	if (!ret)
		closure_put(&c->cl);
err:
	if (devs)
		for (i = 0; i < arg.nr_devs; i++)
			kfree(devs[i]);
	kfree(devs);
	return ret;
}

static long bch2_ioctl_incremental(struct bch_ioctl_incremental __user *user_arg)
{
	struct bch_ioctl_incremental arg;
	const char *err;
	char *path;

	if (copy_from_user(&arg, user_arg, sizeof(arg)))
		return -EFAULT;

	if (arg.flags || arg.pad)
		return -EINVAL;

	path = strndup_user((const char __user *)(unsigned long) arg.dev, PATH_MAX);
	ret = PTR_ERR_OR_ZERO(path);
	if (ret)
		return ret;

	err = bch2_fs_open_incremental(path);
	kfree(path);

	if (err) {
		pr_err("Could not register bcachefs devices: %s", err);
		return -EINVAL;
	}

	return 0;
}
#endif

static long bch2_global_ioctl(unsigned cmd, void __user *arg)
{
	long ret;

	switch (cmd) {
#if 0
	case BCH_IOCTL_ASSEMBLE:
		return bch2_ioctl_assemble(arg);
	case BCH_IOCTL_INCREMENTAL:
		return bch2_ioctl_incremental(arg);
#endif
	case BCH_IOCTL_FSCK_OFFLINE: {
		ret = bch2_ioctl_fsck_offline(arg);
		break;
	}
	default:
		ret = -ENOTTY;
		break;
	}

	if (ret < 0)
		ret = bch2_err_class(ret);
	return ret;
}

static long bch2_ioctl_query_uuid(struct bch_fs *c,
			struct bch_ioctl_query_uuid __user *user_arg)
{
	return copy_to_user_errcode(&user_arg->uuid, &c->sb.user_uuid,
				    sizeof(c->sb.user_uuid));
}

#if 0
static long bch2_ioctl_start(struct bch_fs *c, struct bch_ioctl_start arg)
{
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	if (arg.flags || arg.pad)
		return -EINVAL;

	return bch2_fs_start(c);
}

static long bch2_ioctl_stop(struct bch_fs *c)
{
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	bch2_fs_stop(c);
	return 0;
}
#endif

static long bch2_ioctl_disk_add(struct bch_fs *c, struct bch_ioctl_disk arg)
{
	char *path;
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	if (arg.flags || arg.pad)
		return -EINVAL;

	path = strndup_user((const char __user *)(unsigned long) arg.dev, PATH_MAX);
	ret = PTR_ERR_OR_ZERO(path);
	if (ret)
		return ret;

	ret = bch2_dev_add(c, path);
	if (!IS_ERR(path))
		kfree(path);

	return ret;
}

static long bch2_ioctl_disk_remove(struct bch_fs *c, struct bch_ioctl_disk arg)
{
	struct bch_dev *ca;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	if ((arg.flags & ~(BCH_FORCE_IF_DATA_LOST|
			   BCH_FORCE_IF_METADATA_LOST|
			   BCH_FORCE_IF_DEGRADED|
			   BCH_BY_INDEX)) ||
	    arg.pad)
		return -EINVAL;

	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	if (IS_ERR(ca))
		return PTR_ERR(ca);

	return bch2_dev_remove(c, ca, arg.flags);
}

static long bch2_ioctl_disk_online(struct bch_fs *c, struct bch_ioctl_disk arg)
{
	char *path;
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	if (arg.flags || arg.pad)
		return -EINVAL;

	path = strndup_user((const char __user *)(unsigned long) arg.dev, PATH_MAX);
	ret = PTR_ERR_OR_ZERO(path);
	if (ret)
		return ret;

	ret = bch2_dev_online(c, path);
	kfree(path);
	return ret;
}

static long bch2_ioctl_disk_offline(struct bch_fs *c, struct bch_ioctl_disk arg)
{
	struct bch_dev *ca;
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	if ((arg.flags & ~(BCH_FORCE_IF_DATA_LOST|
			   BCH_FORCE_IF_METADATA_LOST|
			   BCH_FORCE_IF_DEGRADED|
			   BCH_BY_INDEX)) ||
	    arg.pad)
		return -EINVAL;

	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	if (IS_ERR(ca))
		return PTR_ERR(ca);

	ret = bch2_dev_offline(c, ca, arg.flags);
	bch2_dev_put(ca);
	return ret;
}

static long bch2_ioctl_disk_set_state(struct bch_fs *c,
			struct bch_ioctl_disk_set_state arg)
{
	struct bch_dev *ca;
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	if ((arg.flags & ~(BCH_FORCE_IF_DATA_LOST|
			   BCH_FORCE_IF_METADATA_LOST|
			   BCH_FORCE_IF_DEGRADED|
			   BCH_BY_INDEX)) ||
	    arg.pad[0] || arg.pad[1] || arg.pad[2] ||
	    arg.new_state >= BCH_MEMBER_STATE_NR)
		return -EINVAL;

	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	if (IS_ERR(ca))
		return PTR_ERR(ca);

	ret = bch2_dev_set_state(c, ca, arg.new_state, arg.flags);
	if (ret)
		bch_err(c, "Error setting device state: %s", bch2_err_str(ret));

	bch2_dev_put(ca);
	return ret;
}

struct bch_data_ctx {
	struct thread_with_file		thr;

	struct bch_fs			*c;
	struct bch_ioctl_data		arg;
	struct bch_move_stats		stats;
};

static int bch2_data_thread(void *arg)
{
	struct bch_data_ctx *ctx = container_of(arg, struct bch_data_ctx, thr);

	ctx->thr.ret = bch2_data_job(ctx->c, &ctx->stats, ctx->arg);
	if (ctx->thr.ret == -BCH_ERR_device_offline)
		ctx->stats.ret = BCH_IOCTL_DATA_EVENT_RET_device_offline;
	else {
		ctx->stats.ret = BCH_IOCTL_DATA_EVENT_RET_done;
		ctx->stats.data_type = (int) DATA_PROGRESS_DATA_TYPE_done;
	}
	enumerated_ref_put(&ctx->c->writes, BCH_WRITE_REF_ioctl_data);
	return 0;
}

static int bch2_data_job_release(struct inode *inode, struct file *file)
{
	struct bch_data_ctx *ctx = container_of(file->private_data, struct bch_data_ctx, thr);

	bch2_thread_with_file_exit(&ctx->thr);
	kfree(ctx);
	return 0;
}

static ssize_t bch2_data_job_read(struct file *file, char __user *buf,
				  size_t len, loff_t *ppos)
{
	struct bch_data_ctx *ctx = container_of(file->private_data, struct bch_data_ctx, thr);
	struct bch_fs *c = ctx->c;
	struct bch_ioctl_data_event e = {
		.type				= BCH_DATA_EVENT_PROGRESS,
		.ret				= ctx->stats.ret,
		.p.data_type			= ctx->stats.data_type,
		.p.btree_id			= ctx->stats.pos.btree,
		.p.pos				= ctx->stats.pos.pos,
		.p.sectors_done			= atomic64_read(&ctx->stats.sectors_seen),
		.p.sectors_error_corrected	= atomic64_read(&ctx->stats.sectors_error_corrected),
		.p.sectors_error_uncorrected	= atomic64_read(&ctx->stats.sectors_error_uncorrected),
	};

	if (ctx->arg.op == BCH_DATA_OP_scrub) {
		struct bch_dev *ca = bch2_dev_tryget(c, ctx->arg.scrub.dev);
		if (ca) {
			struct bch_dev_usage_full u;
			bch2_dev_usage_full_read_fast(ca, &u);
			for (unsigned i = BCH_DATA_btree; i < ARRAY_SIZE(u.d); i++)
				if (ctx->arg.scrub.data_types & BIT(i))
					e.p.sectors_total += u.d[i].sectors;
			bch2_dev_put(ca);
		}
	} else {
		e.p.sectors_total	= bch2_fs_usage_read_short(c).used;
	}

	if (len < sizeof(e))
		return -EINVAL;

	return copy_to_user_errcode(buf, &e, sizeof(e)) ?: sizeof(e);
}

static const struct file_operations bcachefs_data_ops = {
	.release	= bch2_data_job_release,
	.read		= bch2_data_job_read,
};

static long bch2_ioctl_data(struct bch_fs *c,
			    struct bch_ioctl_data arg)
{
	struct bch_data_ctx *ctx;
	int ret;

	if (!enumerated_ref_tryget(&c->writes, BCH_WRITE_REF_ioctl_data))
		return -EROFS;

	if (!capable(CAP_SYS_ADMIN)) {
		ret = -EPERM;
		goto put_ref;
	}

	if (arg.op >= BCH_DATA_OP_NR || arg.flags) {
		ret = -EINVAL;
		goto put_ref;
	}

	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
	if (!ctx) {
		ret = -ENOMEM;
		goto put_ref;
	}

	ctx->c = c;
	ctx->arg = arg;

	ret = bch2_run_thread_with_file(&ctx->thr,
			&bcachefs_data_ops,
			bch2_data_thread);
	if (ret < 0)
		goto cleanup;
	return ret;
cleanup:
	kfree(ctx);
put_ref:
	enumerated_ref_put(&c->writes, BCH_WRITE_REF_ioctl_data);
	return ret;
}

static noinline_for_stack long bch2_ioctl_fs_usage(struct bch_fs *c,
				struct bch_ioctl_fs_usage __user *user_arg)
{
	struct bch_ioctl_fs_usage arg = {};
	darray_char replicas = {};
	u32 replica_entries_bytes;
	int ret = 0;

	if (!test_bit(BCH_FS_started, &c->flags))
		return -EINVAL;

	if (get_user(replica_entries_bytes, &user_arg->replica_entries_bytes))
		return -EFAULT;

	ret   = bch2_fs_replicas_usage_read(c, &replicas) ?:
		(replica_entries_bytes < replicas.nr ? -ERANGE : 0) ?:
		copy_to_user_errcode(&user_arg->replicas, replicas.data, replicas.nr);
	if (ret)
		goto err;

	struct bch_fs_usage_short u = bch2_fs_usage_read_short(c);
	arg.capacity		= c->capacity;
	arg.used		= u.used;
	arg.online_reserved	= percpu_u64_get(c->online_reserved);
	arg.replica_entries_bytes = replicas.nr;

	for (unsigned i = 0; i < BCH_REPLICAS_MAX; i++) {
		struct disk_accounting_pos k;
		disk_accounting_key_init(k, persistent_reserved, .nr_replicas = i);

		bch2_accounting_mem_read(c,
					 disk_accounting_pos_to_bpos(&k),
					 &arg.persistent_reserved[i], 1);
	}

	ret = copy_to_user_errcode(user_arg, &arg, sizeof(arg));
err:
	darray_exit(&replicas);
	return ret;
}

static long bch2_ioctl_query_accounting(struct bch_fs *c,
			struct bch_ioctl_query_accounting __user *user_arg)
{
	struct bch_ioctl_query_accounting arg;
	darray_char accounting = {};
	int ret = 0;

	if (!test_bit(BCH_FS_started, &c->flags))
		return -EINVAL;

	ret   = copy_from_user_errcode(&arg, user_arg, sizeof(arg)) ?:
		bch2_fs_accounting_read(c, &accounting, arg.accounting_types_mask) ?:
		(arg.accounting_u64s * sizeof(u64) < accounting.nr ? -ERANGE : 0) ?:
		copy_to_user_errcode(&user_arg->accounting, accounting.data, accounting.nr);
	if (ret)
		goto err;

	arg.capacity		= c->capacity;
	arg.used		= bch2_fs_usage_read_short(c).used;
	arg.online_reserved	= percpu_u64_get(c->online_reserved);
	arg.accounting_u64s	= accounting.nr / sizeof(u64);

	ret = copy_to_user_errcode(user_arg, &arg, sizeof(arg));
err:
	darray_exit(&accounting);
	return ret;
}

/* obsolete, didn't allow for new data types: */
static noinline_for_stack long bch2_ioctl_dev_usage(struct bch_fs *c,
				 struct bch_ioctl_dev_usage __user *user_arg)
{
	struct bch_ioctl_dev_usage arg;
	struct bch_dev_usage_full src;
	struct bch_dev *ca;
	unsigned i;

	if (!test_bit(BCH_FS_started, &c->flags))
		return -EINVAL;

	if (copy_from_user(&arg, user_arg, sizeof(arg)))
		return -EFAULT;

	if ((arg.flags & ~BCH_BY_INDEX) ||
	    arg.pad[0] ||
	    arg.pad[1] ||
	    arg.pad[2])
		return -EINVAL;

	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	if (IS_ERR(ca))
		return PTR_ERR(ca);

	src = bch2_dev_usage_full_read(ca);

	arg.state		= ca->mi.state;
	arg.bucket_size		= ca->mi.bucket_size;
	arg.nr_buckets		= ca->mi.nbuckets - ca->mi.first_bucket;

	for (i = 0; i < ARRAY_SIZE(arg.d); i++) {
		arg.d[i].buckets	= src.d[i].buckets;
		arg.d[i].sectors	= src.d[i].sectors;
		arg.d[i].fragmented	= src.d[i].fragmented;
	}

	bch2_dev_put(ca);

	return copy_to_user_errcode(user_arg, &arg, sizeof(arg));
}

static long bch2_ioctl_dev_usage_v2(struct bch_fs *c,
				 struct bch_ioctl_dev_usage_v2 __user *user_arg)
{
	struct bch_ioctl_dev_usage_v2 arg;
	struct bch_dev_usage_full src;
	struct bch_dev *ca;
	int ret = 0;

	if (!test_bit(BCH_FS_started, &c->flags))
		return -EINVAL;

	if (copy_from_user(&arg, user_arg, sizeof(arg)))
		return -EFAULT;

	if ((arg.flags & ~BCH_BY_INDEX) ||
	    arg.pad[0] ||
	    arg.pad[1] ||
	    arg.pad[2])
		return -EINVAL;

	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	if (IS_ERR(ca))
		return PTR_ERR(ca);

	src = bch2_dev_usage_full_read(ca);

	arg.state		= ca->mi.state;
	arg.bucket_size		= ca->mi.bucket_size;
	arg.nr_data_types	= min(arg.nr_data_types, BCH_DATA_NR);
	arg.nr_buckets		= ca->mi.nbuckets - ca->mi.first_bucket;

	ret = copy_to_user_errcode(user_arg, &arg, sizeof(arg));
	if (ret)
		goto err;

	for (unsigned i = 0; i < arg.nr_data_types; i++) {
		struct bch_ioctl_dev_usage_type t = {
			.buckets	= src.d[i].buckets,
			.sectors	= src.d[i].sectors,
			.fragmented	= src.d[i].fragmented,
		};

		ret = copy_to_user_errcode(&user_arg->d[i], &t, sizeof(t));
		if (ret)
			goto err;
	}
err:
	bch2_dev_put(ca);
	return ret;
}

static long bch2_ioctl_read_super(struct bch_fs *c,
				  struct bch_ioctl_read_super arg)
{
	struct bch_dev *ca = NULL;
	struct bch_sb *sb;
	int ret = 0;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	if ((arg.flags & ~(BCH_BY_INDEX|BCH_READ_DEV)) ||
	    arg.pad)
		return -EINVAL;

	mutex_lock(&c->sb_lock);

	if (arg.flags & BCH_READ_DEV) {
		ca = bch2_device_lookup(c, arg.dev, arg.flags);
		ret = PTR_ERR_OR_ZERO(ca);
		if (ret)
			goto err_unlock;

		sb = ca->disk_sb.sb;
	} else {
		sb = c->disk_sb.sb;
	}

	if (vstruct_bytes(sb) > arg.size) {
		ret = -ERANGE;
		goto err;
	}

	ret = copy_to_user_errcode((void __user *)(unsigned long)arg.sb, sb,
				   vstruct_bytes(sb));
err:
	bch2_dev_put(ca);
err_unlock:
	mutex_unlock(&c->sb_lock);
	return ret;
}

static long bch2_ioctl_disk_get_idx(struct bch_fs *c,
				    struct bch_ioctl_disk_get_idx arg)
{
	dev_t dev = huge_decode_dev(arg.dev);

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	if (!dev)
		return -EINVAL;

	guard(rcu)();
	for_each_online_member_rcu(c, ca)
		if (ca->dev == dev)
			return ca->dev_idx;

	return bch_err_throw(c, ENOENT_dev_idx_not_found);
}

static long bch2_ioctl_disk_resize(struct bch_fs *c,
				   struct bch_ioctl_disk_resize arg)
{
	struct bch_dev *ca;
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	if ((arg.flags & ~BCH_BY_INDEX) ||
	    arg.pad)
		return -EINVAL;

	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	if (IS_ERR(ca))
		return PTR_ERR(ca);

	ret = bch2_dev_resize(c, ca, arg.nbuckets);

	bch2_dev_put(ca);
	return ret;
}

static long bch2_ioctl_disk_resize_journal(struct bch_fs *c,
				   struct bch_ioctl_disk_resize_journal arg)
{
	struct bch_dev *ca;
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	if ((arg.flags & ~BCH_BY_INDEX) ||
	    arg.pad)
		return -EINVAL;

	if (arg.nbuckets > U32_MAX)
		return -EINVAL;

	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	if (IS_ERR(ca))
		return PTR_ERR(ca);

	ret = bch2_set_nr_journal_buckets(c, ca, arg.nbuckets);

	bch2_dev_put(ca);
	return ret;
}

#define BCH_IOCTL(_name, _argtype)					\
do {									\
	_argtype i;							\
									\
	if (copy_from_user(&i, arg, sizeof(i)))				\
		return -EFAULT;						\
	ret = bch2_ioctl_##_name(c, i);					\
	goto out;							\
} while (0)

long bch2_fs_ioctl(struct bch_fs *c, unsigned cmd, void __user *arg)
{
	long ret;

	switch (cmd) {
	case BCH_IOCTL_QUERY_UUID:
		return bch2_ioctl_query_uuid(c, arg);
	case BCH_IOCTL_FS_USAGE:
		return bch2_ioctl_fs_usage(c, arg);
	case BCH_IOCTL_DEV_USAGE:
		return bch2_ioctl_dev_usage(c, arg);
	case BCH_IOCTL_DEV_USAGE_V2:
		return bch2_ioctl_dev_usage_v2(c, arg);
#if 0
	case BCH_IOCTL_START:
		BCH_IOCTL(start, struct bch_ioctl_start);
	case BCH_IOCTL_STOP:
		return bch2_ioctl_stop(c);
#endif
	case BCH_IOCTL_READ_SUPER:
		BCH_IOCTL(read_super, struct bch_ioctl_read_super);
	case BCH_IOCTL_DISK_GET_IDX:
		BCH_IOCTL(disk_get_idx, struct bch_ioctl_disk_get_idx);
	}

	if (!test_bit(BCH_FS_started, &c->flags))
		return -EINVAL;

	switch (cmd) {
	case BCH_IOCTL_DISK_ADD:
		BCH_IOCTL(disk_add, struct bch_ioctl_disk);
	case BCH_IOCTL_DISK_REMOVE:
		BCH_IOCTL(disk_remove, struct bch_ioctl_disk);
	case BCH_IOCTL_DISK_ONLINE:
		BCH_IOCTL(disk_online, struct bch_ioctl_disk);
	case BCH_IOCTL_DISK_OFFLINE:
		BCH_IOCTL(disk_offline, struct bch_ioctl_disk);
	case BCH_IOCTL_DISK_SET_STATE:
		BCH_IOCTL(disk_set_state, struct bch_ioctl_disk_set_state);
	case BCH_IOCTL_DATA:
		BCH_IOCTL(data, struct bch_ioctl_data);
	case BCH_IOCTL_DISK_RESIZE:
		BCH_IOCTL(disk_resize, struct bch_ioctl_disk_resize);
	case BCH_IOCTL_DISK_RESIZE_JOURNAL:
		BCH_IOCTL(disk_resize_journal, struct bch_ioctl_disk_resize_journal);
	case BCH_IOCTL_FSCK_ONLINE:
		BCH_IOCTL(fsck_online, struct bch_ioctl_fsck_online);
	case BCH_IOCTL_QUERY_ACCOUNTING:
		return bch2_ioctl_query_accounting(c, arg);
	case BCH_IOCTL_QUERY_COUNTERS:
		return bch2_ioctl_query_counters(c, arg);
	default:
		return -ENOTTY;
	}
out:
	if (ret < 0)
		ret = bch2_err_class(ret);
	return ret;
}

static DEFINE_IDR(bch_chardev_minor);

static long bch2_chardev_ioctl(struct file *filp, unsigned cmd, unsigned long v)
{
	unsigned minor = iminor(file_inode(filp));
	struct bch_fs *c = minor < U8_MAX ? idr_find(&bch_chardev_minor, minor) : NULL;
	void __user *arg = (void __user *) v;

	return c
		? bch2_fs_ioctl(c, cmd, arg)
		: bch2_global_ioctl(cmd, arg);
}

static const struct file_operations bch_chardev_fops = {
	.owner		= THIS_MODULE,
	.unlocked_ioctl = bch2_chardev_ioctl,
	.open		= nonseekable_open,
};

static int bch_chardev_major;
static const struct class bch_chardev_class = {
	.name = "bcachefs",
};
static struct device *bch_chardev;

void bch2_fs_chardev_exit(struct bch_fs *c)
{
	if (!IS_ERR_OR_NULL(c->chardev))
		device_unregister(c->chardev);
	if (c->minor >= 0)
		idr_remove(&bch_chardev_minor, c->minor);
}

int bch2_fs_chardev_init(struct bch_fs *c)
{
	c->minor = idr_alloc(&bch_chardev_minor, c, 0, 0, GFP_KERNEL);
	if (c->minor < 0)
		return c->minor;

	c->chardev = device_create(&bch_chardev_class, NULL,
				   MKDEV(bch_chardev_major, c->minor), c,
				   "bcachefs%u-ctl", c->minor);
	if (IS_ERR(c->chardev))
		return PTR_ERR(c->chardev);

	return 0;
}

void bch2_chardev_exit(void)
{
	device_destroy(&bch_chardev_class, MKDEV(bch_chardev_major, U8_MAX));
	class_unregister(&bch_chardev_class);
	if (bch_chardev_major > 0)
		unregister_chrdev(bch_chardev_major, "bcachefs");
}

int __init bch2_chardev_init(void)
{
	int ret;

	bch_chardev_major = register_chrdev(0, "bcachefs-ctl", &bch_chardev_fops);
	if (bch_chardev_major < 0)
		return bch_chardev_major;

	ret = class_register(&bch_chardev_class);
	if (ret)
		goto major_out;

	bch_chardev = device_create(&bch_chardev_class, NULL,
				    MKDEV(bch_chardev_major, U8_MAX),
				    NULL, "bcachefs-ctl");
	if (IS_ERR(bch_chardev)) {
		ret = PTR_ERR(bch_chardev);
		goto class_out;
	}

	return 0;

class_out:
	class_unregister(&bch_chardev_class);
major_out:
	unregister_chrdev(bch_chardev_major, "bcachefs-ctl");
	return ret;
}

#endif /* NO_BCACHEFS_CHARDEV */
Commit	Line	Data
	1	// SPDX-License-Identifier: GPL-2.0
	2	#ifndef NO_BCACHEFS_CHARDEV
	3
	4	#include "bcachefs.h"
	5	#include "bcachefs_ioctl.h"
	6	#include "buckets.h"
	7	#include "chardev.h"
	8	#include "disk_accounting.h"
	9	#include "fsck.h"
	10	#include "journal.h"
	11	#include "move.h"
	12	#include "recovery_passes.h"
	13	#include "replicas.h"
	14	#include "sb-counters.h"
	15	#include "super-io.h"
	16	#include "thread_with_file.h"
	17
	18	#include <linux/cdev.h>
	19	#include <linux/device.h>
	20	#include <linux/fs.h>
	21	#include <linux/ioctl.h>
	22	#include <linux/major.h>
	23	#include <linux/sched/task.h>
	24	#include <linux/slab.h>
	25	#include <linux/uaccess.h>
	26
	27	/* returns with ref on ca->ref */
	28	static struct bch_dev bch2_device_lookup(struct bch_fs c, u64 dev,
	29	unsigned flags)
	30	{
	31	struct bch_dev *ca;
	32
	33	if (flags & BCH_BY_INDEX) {
	34	if (dev >= c->sb.nr_devices)
	35	return ERR_PTR(-EINVAL);
	36
	37	ca = bch2_dev_tryget_noerror(c, dev);
	38	if (!ca)
	39	return ERR_PTR(-EINVAL);
	40	} else {
	41	char *path;
	42
	43	path = strndup_user((const char __user *)
	44	(unsigned long) dev, PATH_MAX);
	45	if (IS_ERR(path))
	46	return ERR_CAST(path);
	47
	48	ca = bch2_dev_lookup(c, path);
	49	kfree(path);
	50	}
	51
	52	return ca;
	53	}
	54
	55	#if 0
	56	static long bch2_ioctl_assemble(struct bch_ioctl_assemble __user *user_arg)
	57	{
	58	struct bch_ioctl_assemble arg;
	59	struct bch_fs *c;
	60	u64 *user_devs = NULL;
	61	char **devs = NULL;
	62	unsigned i;
	63	int ret = -EFAULT;
	64
	65	if (copy_from_user(&arg, user_arg, sizeof(arg)))
	66	return -EFAULT;
	67
	68	if (arg.flags \|\| arg.pad)
	69	return -EINVAL;
	70
	71	user_devs = kmalloc_array(arg.nr_devs, sizeof(u64), GFP_KERNEL);
	72	if (!user_devs)
	73	return -ENOMEM;
	74
	75	devs = kcalloc(arg.nr_devs, sizeof(char *), GFP_KERNEL);
	76
	77	if (copy_from_user(user_devs, user_arg->devs,
	78	sizeof(u64) * arg.nr_devs))
	79	goto err;
	80
	81	for (i = 0; i < arg.nr_devs; i++) {
	82	devs[i] = strndup_user((const char __user *)(unsigned long)
	83	user_devs[i],
	84	PATH_MAX);
	85	ret= PTR_ERR_OR_ZERO(devs[i]);
	86	if (ret)
	87	goto err;
	88	}
	89
	90	c = bch2_fs_open(devs, arg.nr_devs, bch2_opts_empty());
	91	ret = PTR_ERR_OR_ZERO(c);
	92	if (!ret)
	93	closure_put(&c->cl);
	94	err:
	95	if (devs)
	96	for (i = 0; i < arg.nr_devs; i++)
	97	kfree(devs[i]);
	98	kfree(devs);
	99	return ret;
	100	}
	101
	102	static long bch2_ioctl_incremental(struct bch_ioctl_incremental __user *user_arg)
	103	{
	104	struct bch_ioctl_incremental arg;
	105	const char *err;
	106	char *path;
	107
	108	if (copy_from_user(&arg, user_arg, sizeof(arg)))
	109	return -EFAULT;
	110
	111	if (arg.flags \|\| arg.pad)
	112	return -EINVAL;
	113
	114	path = strndup_user((const char __user *)(unsigned long) arg.dev, PATH_MAX);
	115	ret = PTR_ERR_OR_ZERO(path);
	116	if (ret)
	117	return ret;
	118
	119	err = bch2_fs_open_incremental(path);
	120	kfree(path);
	121
	122	if (err) {
	123	pr_err("Could not register bcachefs devices: %s", err);
	124	return -EINVAL;
	125	}
	126
	127	return 0;
	128	}
	129	#endif
	130
	131	static long bch2_global_ioctl(unsigned cmd, void __user *arg)
	132	{
	133	long ret;
	134
	135	switch (cmd) {
	136	#if 0
	137	case BCH_IOCTL_ASSEMBLE:
	138	return bch2_ioctl_assemble(arg);
	139	case BCH_IOCTL_INCREMENTAL:
	140	return bch2_ioctl_incremental(arg);
	141	#endif
	142	case BCH_IOCTL_FSCK_OFFLINE: {
	143	ret = bch2_ioctl_fsck_offline(arg);
	144	break;
	145	}
	146	default:
	147	ret = -ENOTTY;
	148	break;
	149	}
	150
	151	if (ret < 0)
	152	ret = bch2_err_class(ret);
	153	return ret;
	154	}
	155
	156	static long bch2_ioctl_query_uuid(struct bch_fs *c,
	157	struct bch_ioctl_query_uuid __user *user_arg)
	158	{
	159	return copy_to_user_errcode(&user_arg->uuid, &c->sb.user_uuid,
	160	sizeof(c->sb.user_uuid));
	161	}
	162
	163	#if 0
	164	static long bch2_ioctl_start(struct bch_fs *c, struct bch_ioctl_start arg)
	165	{
	166	if (!capable(CAP_SYS_ADMIN))
	167	return -EPERM;
	168
	169	if (arg.flags \|\| arg.pad)
	170	return -EINVAL;
	171
	172	return bch2_fs_start(c);
	173	}
	174
	175	static long bch2_ioctl_stop(struct bch_fs *c)
	176	{
	177	if (!capable(CAP_SYS_ADMIN))
	178	return -EPERM;
	179
	180	bch2_fs_stop(c);
	181	return 0;
	182	}
	183	#endif
	184
	185	static long bch2_ioctl_disk_add(struct bch_fs *c, struct bch_ioctl_disk arg)
	186	{
	187	char *path;
	188	int ret;
	189
	190	if (!capable(CAP_SYS_ADMIN))
	191	return -EPERM;
	192
	193	if (arg.flags \|\| arg.pad)
	194	return -EINVAL;
	195
	196	path = strndup_user((const char __user *)(unsigned long) arg.dev, PATH_MAX);
	197	ret = PTR_ERR_OR_ZERO(path);
	198	if (ret)
	199	return ret;
	200
	201	ret = bch2_dev_add(c, path);
	202	if (!IS_ERR(path))
	203	kfree(path);
	204
	205	return ret;
	206	}
	207
	208	static long bch2_ioctl_disk_remove(struct bch_fs *c, struct bch_ioctl_disk arg)
	209	{
	210	struct bch_dev *ca;
	211
	212	if (!capable(CAP_SYS_ADMIN))
	213	return -EPERM;
	214
	215	if ((arg.flags & ~(BCH_FORCE_IF_DATA_LOST\|
	216	BCH_FORCE_IF_METADATA_LOST\|
	217	BCH_FORCE_IF_DEGRADED\|
	218	BCH_BY_INDEX)) \|\|
	219	arg.pad)
	220	return -EINVAL;
	221
	222	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	223	if (IS_ERR(ca))
	224	return PTR_ERR(ca);
	225
	226	return bch2_dev_remove(c, ca, arg.flags);
	227	}
	228
	229	static long bch2_ioctl_disk_online(struct bch_fs *c, struct bch_ioctl_disk arg)
	230	{
	231	char *path;
	232	int ret;
	233
	234	if (!capable(CAP_SYS_ADMIN))
	235	return -EPERM;
	236
	237	if (arg.flags \|\| arg.pad)
	238	return -EINVAL;
	239
	240	path = strndup_user((const char __user *)(unsigned long) arg.dev, PATH_MAX);
	241	ret = PTR_ERR_OR_ZERO(path);
	242	if (ret)
	243	return ret;
	244
	245	ret = bch2_dev_online(c, path);
	246	kfree(path);
	247	return ret;
	248	}
	249
	250	static long bch2_ioctl_disk_offline(struct bch_fs *c, struct bch_ioctl_disk arg)
	251	{
	252	struct bch_dev *ca;
	253	int ret;
	254
	255	if (!capable(CAP_SYS_ADMIN))
	256	return -EPERM;
	257
	258	if ((arg.flags & ~(BCH_FORCE_IF_DATA_LOST\|
	259	BCH_FORCE_IF_METADATA_LOST\|
	260	BCH_FORCE_IF_DEGRADED\|
	261	BCH_BY_INDEX)) \|\|
	262	arg.pad)
	263	return -EINVAL;
	264
	265	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	266	if (IS_ERR(ca))
	267	return PTR_ERR(ca);
	268
	269	ret = bch2_dev_offline(c, ca, arg.flags);
	270	bch2_dev_put(ca);
	271	return ret;
	272	}
	273
	274	static long bch2_ioctl_disk_set_state(struct bch_fs *c,
	275	struct bch_ioctl_disk_set_state arg)
	276	{
	277	struct bch_dev *ca;
	278	int ret;
	279
	280	if (!capable(CAP_SYS_ADMIN))
	281	return -EPERM;
	282
	283	if ((arg.flags & ~(BCH_FORCE_IF_DATA_LOST\|
	284	BCH_FORCE_IF_METADATA_LOST\|
	285	BCH_FORCE_IF_DEGRADED\|
	286	BCH_BY_INDEX)) \|\|
	287	arg.pad[0] \|\| arg.pad[1] \|\| arg.pad[2] \|\|
	288	arg.new_state >= BCH_MEMBER_STATE_NR)
	289	return -EINVAL;
	290
	291	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	292	if (IS_ERR(ca))
	293	return PTR_ERR(ca);
	294
	295	ret = bch2_dev_set_state(c, ca, arg.new_state, arg.flags);
	296	if (ret)
	297	bch_err(c, "Error setting device state: %s", bch2_err_str(ret));
	298
	299	bch2_dev_put(ca);
	300	return ret;
	301	}
	302
	303	struct bch_data_ctx {
	304	struct thread_with_file thr;
	305
	306	struct bch_fs *c;
	307	struct bch_ioctl_data arg;
	308	struct bch_move_stats stats;
	309	};
	310
	311	static int bch2_data_thread(void *arg)
	312	{
	313	struct bch_data_ctx *ctx = container_of(arg, struct bch_data_ctx, thr);
	314
	315	ctx->thr.ret = bch2_data_job(ctx->c, &ctx->stats, ctx->arg);
	316	if (ctx->thr.ret == -BCH_ERR_device_offline)
	317	ctx->stats.ret = BCH_IOCTL_DATA_EVENT_RET_device_offline;
	318	else {
	319	ctx->stats.ret = BCH_IOCTL_DATA_EVENT_RET_done;
	320	ctx->stats.data_type = (int) DATA_PROGRESS_DATA_TYPE_done;
	321	}
	322	enumerated_ref_put(&ctx->c->writes, BCH_WRITE_REF_ioctl_data);
	323	return 0;
	324	}
	325
	326	static int bch2_data_job_release(struct inode inode, struct file file)
	327	{
	328	struct bch_data_ctx *ctx = container_of(file->private_data, struct bch_data_ctx, thr);
	329
	330	bch2_thread_with_file_exit(&ctx->thr);
	331	kfree(ctx);
	332	return 0;
	333	}
	334
	335	static ssize_t bch2_data_job_read(struct file file, char __user buf,
	336	size_t len, loff_t *ppos)
	337	{
	338	struct bch_data_ctx *ctx = container_of(file->private_data, struct bch_data_ctx, thr);
	339	struct bch_fs *c = ctx->c;
	340	struct bch_ioctl_data_event e = {
	341	.type = BCH_DATA_EVENT_PROGRESS,
	342	.ret = ctx->stats.ret,
	343	.p.data_type = ctx->stats.data_type,
	344	.p.btree_id = ctx->stats.pos.btree,
	345	.p.pos = ctx->stats.pos.pos,
	346	.p.sectors_done = atomic64_read(&ctx->stats.sectors_seen),
	347	.p.sectors_error_corrected = atomic64_read(&ctx->stats.sectors_error_corrected),
	348	.p.sectors_error_uncorrected = atomic64_read(&ctx->stats.sectors_error_uncorrected),
	349	};
	350
	351	if (ctx->arg.op == BCH_DATA_OP_scrub) {
	352	struct bch_dev *ca = bch2_dev_tryget(c, ctx->arg.scrub.dev);
	353	if (ca) {
	354	struct bch_dev_usage_full u;
	355	bch2_dev_usage_full_read_fast(ca, &u);
	356	for (unsigned i = BCH_DATA_btree; i < ARRAY_SIZE(u.d); i++)
	357	if (ctx->arg.scrub.data_types & BIT(i))
	358	e.p.sectors_total += u.d[i].sectors;
	359	bch2_dev_put(ca);
	360	}
	361	} else {
	362	e.p.sectors_total = bch2_fs_usage_read_short(c).used;
	363	}
	364
	365	if (len < sizeof(e))
	366	return -EINVAL;
	367
	368	return copy_to_user_errcode(buf, &e, sizeof(e)) ?: sizeof(e);
	369	}
	370
	371	static const struct file_operations bcachefs_data_ops = {
	372	.release = bch2_data_job_release,
	373	.read = bch2_data_job_read,
	374	};
	375
	376	static long bch2_ioctl_data(struct bch_fs *c,
	377	struct bch_ioctl_data arg)
	378	{
	379	struct bch_data_ctx *ctx;
	380	int ret;
	381
	382	if (!enumerated_ref_tryget(&c->writes, BCH_WRITE_REF_ioctl_data))
	383	return -EROFS;
	384
	385	if (!capable(CAP_SYS_ADMIN)) {
	386	ret = -EPERM;
	387	goto put_ref;
	388	}
	389
	390	if (arg.op >= BCH_DATA_OP_NR \|\| arg.flags) {
	391	ret = -EINVAL;
	392	goto put_ref;
	393	}
	394
	395	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
	396	if (!ctx) {
	397	ret = -ENOMEM;
	398	goto put_ref;
	399	}
	400
	401	ctx->c = c;
	402	ctx->arg = arg;
	403
	404	ret = bch2_run_thread_with_file(&ctx->thr,
	405	&bcachefs_data_ops,
	406	bch2_data_thread);
	407	if (ret < 0)
	408	goto cleanup;
	409	return ret;
	410	cleanup:
	411	kfree(ctx);
	412	put_ref:
	413	enumerated_ref_put(&c->writes, BCH_WRITE_REF_ioctl_data);
	414	return ret;
	415	}
	416
	417	static noinline_for_stack long bch2_ioctl_fs_usage(struct bch_fs *c,
	418	struct bch_ioctl_fs_usage __user *user_arg)
	419	{
	420	struct bch_ioctl_fs_usage arg = {};
	421	darray_char replicas = {};
	422	u32 replica_entries_bytes;
	423	int ret = 0;
	424
	425	if (!test_bit(BCH_FS_started, &c->flags))
	426	return -EINVAL;
	427
	428	if (get_user(replica_entries_bytes, &user_arg->replica_entries_bytes))
	429	return -EFAULT;
	430
	431	ret = bch2_fs_replicas_usage_read(c, &replicas) ?:
	432	(replica_entries_bytes < replicas.nr ? -ERANGE : 0) ?:
	433	copy_to_user_errcode(&user_arg->replicas, replicas.data, replicas.nr);
	434	if (ret)
	435	goto err;
	436
	437	struct bch_fs_usage_short u = bch2_fs_usage_read_short(c);
	438	arg.capacity = c->capacity;
	439	arg.used = u.used;
	440	arg.online_reserved = percpu_u64_get(c->online_reserved);
	441	arg.replica_entries_bytes = replicas.nr;
	442
	443	for (unsigned i = 0; i < BCH_REPLICAS_MAX; i++) {
	444	struct disk_accounting_pos k;
	445	disk_accounting_key_init(k, persistent_reserved, .nr_replicas = i);
	446
	447	bch2_accounting_mem_read(c,
	448	disk_accounting_pos_to_bpos(&k),
	449	&arg.persistent_reserved[i], 1);
	450	}
	451
	452	ret = copy_to_user_errcode(user_arg, &arg, sizeof(arg));
	453	err:
	454	darray_exit(&replicas);
	455	return ret;
	456	}
	457
	458	static long bch2_ioctl_query_accounting(struct bch_fs *c,
	459	struct bch_ioctl_query_accounting __user *user_arg)
	460	{
	461	struct bch_ioctl_query_accounting arg;
	462	darray_char accounting = {};
	463	int ret = 0;
	464
	465	if (!test_bit(BCH_FS_started, &c->flags))
	466	return -EINVAL;
	467
	468	ret = copy_from_user_errcode(&arg, user_arg, sizeof(arg)) ?:
	469	bch2_fs_accounting_read(c, &accounting, arg.accounting_types_mask) ?:
	470	(arg.accounting_u64s * sizeof(u64) < accounting.nr ? -ERANGE : 0) ?:
	471	copy_to_user_errcode(&user_arg->accounting, accounting.data, accounting.nr);
	472	if (ret)
	473	goto err;
	474
	475	arg.capacity = c->capacity;
	476	arg.used = bch2_fs_usage_read_short(c).used;
	477	arg.online_reserved = percpu_u64_get(c->online_reserved);
	478	arg.accounting_u64s = accounting.nr / sizeof(u64);
	479
	480	ret = copy_to_user_errcode(user_arg, &arg, sizeof(arg));
	481	err:
	482	darray_exit(&accounting);
	483	return ret;
	484	}
	485
	486	/* obsolete, didn't allow for new data types: */
	487	static noinline_for_stack long bch2_ioctl_dev_usage(struct bch_fs *c,
	488	struct bch_ioctl_dev_usage __user *user_arg)
	489	{
	490	struct bch_ioctl_dev_usage arg;
	491	struct bch_dev_usage_full src;
	492	struct bch_dev *ca;
	493	unsigned i;
	494
	495	if (!test_bit(BCH_FS_started, &c->flags))
	496	return -EINVAL;
	497
	498	if (copy_from_user(&arg, user_arg, sizeof(arg)))
	499	return -EFAULT;
	500
	501	if ((arg.flags & ~BCH_BY_INDEX) \|\|
	502	arg.pad[0] \|\|
	503	arg.pad[1] \|\|
	504	arg.pad[2])
	505	return -EINVAL;
	506
	507	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	508	if (IS_ERR(ca))
	509	return PTR_ERR(ca);
	510
	511	src = bch2_dev_usage_full_read(ca);
	512
	513	arg.state = ca->mi.state;
	514	arg.bucket_size = ca->mi.bucket_size;
	515	arg.nr_buckets = ca->mi.nbuckets - ca->mi.first_bucket;
	516
	517	for (i = 0; i < ARRAY_SIZE(arg.d); i++) {
	518	arg.d[i].buckets = src.d[i].buckets;
	519	arg.d[i].sectors = src.d[i].sectors;
	520	arg.d[i].fragmented = src.d[i].fragmented;
	521	}
	522
	523	bch2_dev_put(ca);
	524
	525	return copy_to_user_errcode(user_arg, &arg, sizeof(arg));
	526	}
	527
	528	static long bch2_ioctl_dev_usage_v2(struct bch_fs *c,
	529	struct bch_ioctl_dev_usage_v2 __user *user_arg)
	530	{
	531	struct bch_ioctl_dev_usage_v2 arg;
	532	struct bch_dev_usage_full src;
	533	struct bch_dev *ca;
	534	int ret = 0;
	535
	536	if (!test_bit(BCH_FS_started, &c->flags))
	537	return -EINVAL;
	538
	539	if (copy_from_user(&arg, user_arg, sizeof(arg)))
	540	return -EFAULT;
	541
	542	if ((arg.flags & ~BCH_BY_INDEX) \|\|
	543	arg.pad[0] \|\|
	544	arg.pad[1] \|\|
	545	arg.pad[2])
	546	return -EINVAL;
	547
	548	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	549	if (IS_ERR(ca))
	550	return PTR_ERR(ca);
	551
	552	src = bch2_dev_usage_full_read(ca);
	553
	554	arg.state = ca->mi.state;
	555	arg.bucket_size = ca->mi.bucket_size;
	556	arg.nr_data_types = min(arg.nr_data_types, BCH_DATA_NR);
	557	arg.nr_buckets = ca->mi.nbuckets - ca->mi.first_bucket;
	558
	559	ret = copy_to_user_errcode(user_arg, &arg, sizeof(arg));
	560	if (ret)
	561	goto err;
	562
	563	for (unsigned i = 0; i < arg.nr_data_types; i++) {
	564	struct bch_ioctl_dev_usage_type t = {
	565	.buckets = src.d[i].buckets,
	566	.sectors = src.d[i].sectors,
	567	.fragmented = src.d[i].fragmented,
	568	};
	569
	570	ret = copy_to_user_errcode(&user_arg->d[i], &t, sizeof(t));
	571	if (ret)
	572	goto err;
	573	}
	574	err:
	575	bch2_dev_put(ca);
	576	return ret;
	577	}
	578
	579	static long bch2_ioctl_read_super(struct bch_fs *c,
	580	struct bch_ioctl_read_super arg)
	581	{
	582	struct bch_dev *ca = NULL;
	583	struct bch_sb *sb;
	584	int ret = 0;
	585
	586	if (!capable(CAP_SYS_ADMIN))
	587	return -EPERM;
	588
	589	if ((arg.flags & ~(BCH_BY_INDEX\|BCH_READ_DEV)) \|\|
	590	arg.pad)
	591	return -EINVAL;
	592
	593	mutex_lock(&c->sb_lock);
	594
	595	if (arg.flags & BCH_READ_DEV) {
	596	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	597	ret = PTR_ERR_OR_ZERO(ca);
	598	if (ret)
	599	goto err_unlock;
	600
	601	sb = ca->disk_sb.sb;
	602	} else {
	603	sb = c->disk_sb.sb;
	604	}
	605
	606	if (vstruct_bytes(sb) > arg.size) {
	607	ret = -ERANGE;
	608	goto err;
	609	}
	610
	611	ret = copy_to_user_errcode((void __user *)(unsigned long)arg.sb, sb,
	612	vstruct_bytes(sb));
	613	err:
	614	bch2_dev_put(ca);
	615	err_unlock:
	616	mutex_unlock(&c->sb_lock);
	617	return ret;
	618	}
	619
	620	static long bch2_ioctl_disk_get_idx(struct bch_fs *c,
	621	struct bch_ioctl_disk_get_idx arg)
	622	{
	623	dev_t dev = huge_decode_dev(arg.dev);
	624
	625	if (!capable(CAP_SYS_ADMIN))
	626	return -EPERM;
	627
	628	if (!dev)
	629	return -EINVAL;
	630
	631	guard(rcu)();
	632	for_each_online_member_rcu(c, ca)
	633	if (ca->dev == dev)
	634	return ca->dev_idx;
	635
	636	return bch_err_throw(c, ENOENT_dev_idx_not_found);
	637	}
	638
	639	static long bch2_ioctl_disk_resize(struct bch_fs *c,
	640	struct bch_ioctl_disk_resize arg)
	641	{
	642	struct bch_dev *ca;
	643	int ret;
	644
	645	if (!capable(CAP_SYS_ADMIN))
	646	return -EPERM;
	647
	648	if ((arg.flags & ~BCH_BY_INDEX) \|\|
	649	arg.pad)
	650	return -EINVAL;
	651
	652	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	653	if (IS_ERR(ca))
	654	return PTR_ERR(ca);
	655
	656	ret = bch2_dev_resize(c, ca, arg.nbuckets);
	657
	658	bch2_dev_put(ca);
	659	return ret;
	660	}
	661
	662	static long bch2_ioctl_disk_resize_journal(struct bch_fs *c,
	663	struct bch_ioctl_disk_resize_journal arg)
	664	{
	665	struct bch_dev *ca;
	666	int ret;
	667
	668	if (!capable(CAP_SYS_ADMIN))
	669	return -EPERM;
	670
	671	if ((arg.flags & ~BCH_BY_INDEX) \|\|
	672	arg.pad)
	673	return -EINVAL;
	674
	675	if (arg.nbuckets > U32_MAX)
	676	return -EINVAL;
	677
	678	ca = bch2_device_lookup(c, arg.dev, arg.flags);
	679	if (IS_ERR(ca))
	680	return PTR_ERR(ca);
	681
	682	ret = bch2_set_nr_journal_buckets(c, ca, arg.nbuckets);
	683
	684	bch2_dev_put(ca);
	685	return ret;
	686	}
	687
	688	#define BCH_IOCTL(_name, _argtype) \
	689	do { \
	690	_argtype i; \
	691	\
	692	if (copy_from_user(&i, arg, sizeof(i))) \
	693	return -EFAULT; \
	694	ret = bch2_ioctl_##_name(c, i); \
	695	goto out; \
	696	} while (0)
	697
	698	long bch2_fs_ioctl(struct bch_fs c, unsigned cmd, void __user arg)
	699	{
	700	long ret;
	701
	702	switch (cmd) {
	703	case BCH_IOCTL_QUERY_UUID:
	704	return bch2_ioctl_query_uuid(c, arg);
	705	case BCH_IOCTL_FS_USAGE:
	706	return bch2_ioctl_fs_usage(c, arg);
	707	case BCH_IOCTL_DEV_USAGE:
	708	return bch2_ioctl_dev_usage(c, arg);
	709	case BCH_IOCTL_DEV_USAGE_V2:
	710	return bch2_ioctl_dev_usage_v2(c, arg);
	711	#if 0
	712	case BCH_IOCTL_START:
	713	BCH_IOCTL(start, struct bch_ioctl_start);
	714	case BCH_IOCTL_STOP:
	715	return bch2_ioctl_stop(c);
	716	#endif
	717	case BCH_IOCTL_READ_SUPER:
	718	BCH_IOCTL(read_super, struct bch_ioctl_read_super);
	719	case BCH_IOCTL_DISK_GET_IDX:
	720	BCH_IOCTL(disk_get_idx, struct bch_ioctl_disk_get_idx);
	721	}
	722
	723	if (!test_bit(BCH_FS_started, &c->flags))
	724	return -EINVAL;
	725
	726	switch (cmd) {
	727	case BCH_IOCTL_DISK_ADD:
	728	BCH_IOCTL(disk_add, struct bch_ioctl_disk);
	729	case BCH_IOCTL_DISK_REMOVE:
	730	BCH_IOCTL(disk_remove, struct bch_ioctl_disk);
	731	case BCH_IOCTL_DISK_ONLINE:
	732	BCH_IOCTL(disk_online, struct bch_ioctl_disk);
	733	case BCH_IOCTL_DISK_OFFLINE:
	734	BCH_IOCTL(disk_offline, struct bch_ioctl_disk);
	735	case BCH_IOCTL_DISK_SET_STATE:
	736	BCH_IOCTL(disk_set_state, struct bch_ioctl_disk_set_state);
	737	case BCH_IOCTL_DATA:
	738	BCH_IOCTL(data, struct bch_ioctl_data);
	739	case BCH_IOCTL_DISK_RESIZE:
	740	BCH_IOCTL(disk_resize, struct bch_ioctl_disk_resize);
	741	case BCH_IOCTL_DISK_RESIZE_JOURNAL:
	742	BCH_IOCTL(disk_resize_journal, struct bch_ioctl_disk_resize_journal);
	743	case BCH_IOCTL_FSCK_ONLINE:
	744	BCH_IOCTL(fsck_online, struct bch_ioctl_fsck_online);
	745	case BCH_IOCTL_QUERY_ACCOUNTING:
	746	return bch2_ioctl_query_accounting(c, arg);
	747	case BCH_IOCTL_QUERY_COUNTERS:
	748	return bch2_ioctl_query_counters(c, arg);
	749	default:
	750	return -ENOTTY;
	751	}
	752	out:
	753	if (ret < 0)
	754	ret = bch2_err_class(ret);
	755	return ret;
	756	}
	757
	758	static DEFINE_IDR(bch_chardev_minor);
	759
	760	static long bch2_chardev_ioctl(struct file *filp, unsigned cmd, unsigned long v)
	761	{
	762	unsigned minor = iminor(file_inode(filp));
	763	struct bch_fs *c = minor < U8_MAX ? idr_find(&bch_chardev_minor, minor) : NULL;
	764	void __user arg = (void __user ) v;
	765
	766	return c
	767	? bch2_fs_ioctl(c, cmd, arg)
	768	: bch2_global_ioctl(cmd, arg);
	769	}
	770
	771	static const struct file_operations bch_chardev_fops = {
	772	.owner = THIS_MODULE,
	773	.unlocked_ioctl = bch2_chardev_ioctl,
	774	.open = nonseekable_open,
	775	};
	776
	777	static int bch_chardev_major;
	778	static const struct class bch_chardev_class = {
	779	.name = "bcachefs",
	780	};
	781	static struct device *bch_chardev;
	782
	783	void bch2_fs_chardev_exit(struct bch_fs *c)
	784	{
	785	if (!IS_ERR_OR_NULL(c->chardev))
	786	device_unregister(c->chardev);
	787	if (c->minor >= 0)
	788	idr_remove(&bch_chardev_minor, c->minor);
	789	}
	790
	791	int bch2_fs_chardev_init(struct bch_fs *c)
	792	{
	793	c->minor = idr_alloc(&bch_chardev_minor, c, 0, 0, GFP_KERNEL);
	794	if (c->minor < 0)
	795	return c->minor;
	796
	797	c->chardev = device_create(&bch_chardev_class, NULL,
	798	MKDEV(bch_chardev_major, c->minor), c,
	799	"bcachefs%u-ctl", c->minor);
	800	if (IS_ERR(c->chardev))
	801	return PTR_ERR(c->chardev);
	802
	803	return 0;
	804	}
	805
	806	void bch2_chardev_exit(void)
	807	{
	808	device_destroy(&bch_chardev_class, MKDEV(bch_chardev_major, U8_MAX));
	809	class_unregister(&bch_chardev_class);
	810	if (bch_chardev_major > 0)
	811	unregister_chrdev(bch_chardev_major, "bcachefs");
	812	}
	813
	814	int __init bch2_chardev_init(void)
	815	{
	816	int ret;
	817
	818	bch_chardev_major = register_chrdev(0, "bcachefs-ctl", &bch_chardev_fops);
	819	if (bch_chardev_major < 0)
	820	return bch_chardev_major;
	821
	822	ret = class_register(&bch_chardev_class);
	823	if (ret)
	824	goto major_out;
	825
	826	bch_chardev = device_create(&bch_chardev_class, NULL,
	827	MKDEV(bch_chardev_major, U8_MAX),
	828	NULL, "bcachefs-ctl");
	829	if (IS_ERR(bch_chardev)) {
	830	ret = PTR_ERR(bch_chardev);
	831	goto class_out;
	832	}
	833
	834	return 0;
	835
	836	class_out:
	837	class_unregister(&bch_chardev_class);
	838	major_out:
	839	unregister_chrdev(bch_chardev_major, "bcachefs-ctl");
	840	return ret;
	841	}
	842
	843	#endif /* NO_BCACHEFS_CHARDEV */