[mirror_ubuntu-jammy-kernel.git] / fs / proc / inode.c

// SPDX-License-Identifier: GPL-2.0
/*
 *  linux/fs/proc/inode.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

#include <linux/cache.h>
#include <linux/time.h>
#include <linux/proc_fs.h>
#include <linux/kernel.h>
#include <linux/pid_namespace.h>
#include <linux/mm.h>
#include <linux/string.h>
#include <linux/stat.h>
#include <linux/completion.h>
#include <linux/poll.h>
#include <linux/printk.h>
#include <linux/file.h>
#include <linux/limits.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/sysctl.h>
#include <linux/seq_file.h>
#include <linux/slab.h>
#include <linux/mount.h>
#include <linux/magic.h>

#include <linux/uaccess.h>

#include "internal.h"

static void proc_evict_inode(struct inode *inode)
{
	struct proc_dir_entry *de;
	struct ctl_table_header *head;

	truncate_inode_pages_final(&inode->i_data);
	clear_inode(inode);

	/* Stop tracking associated processes */
	put_pid(PROC_I(inode)->pid);

	/* Let go of any associated proc directory entry */
	de = PDE(inode);
	if (de)
		pde_put(de);

	head = PROC_I(inode)->sysctl;
	if (head) {
		RCU_INIT_POINTER(PROC_I(inode)->sysctl, NULL);
		proc_sys_evict_inode(inode, head);
	}
}

static struct kmem_cache *proc_inode_cachep __ro_after_init;
static struct kmem_cache *pde_opener_cache __ro_after_init;

static struct inode *proc_alloc_inode(struct super_block *sb)
{
	struct proc_inode *ei;
	struct inode *inode;

	ei = kmem_cache_alloc(proc_inode_cachep, GFP_KERNEL);
	if (!ei)
		return NULL;
	ei->pid = NULL;
	ei->fd = 0;
	ei->op.proc_get_link = NULL;
	ei->pde = NULL;
	ei->sysctl = NULL;
	ei->sysctl_entry = NULL;
	ei->ns_ops = NULL;
	inode = &ei->vfs_inode;
	return inode;
}

static void proc_i_callback(struct rcu_head *head)
{
	struct inode *inode = container_of(head, struct inode, i_rcu);
	kmem_cache_free(proc_inode_cachep, PROC_I(inode));
}

static void proc_destroy_inode(struct inode *inode)
{
	call_rcu(&inode->i_rcu, proc_i_callback);
}

static void init_once(void *foo)
{
	struct proc_inode *ei = (struct proc_inode *) foo;

	inode_init_once(&ei->vfs_inode);
}

void __init proc_init_kmemcache(void)
{
	proc_inode_cachep = kmem_cache_create("proc_inode_cache",
					     sizeof(struct proc_inode),
					     0, (SLAB_RECLAIM_ACCOUNT|
						SLAB_MEM_SPREAD|SLAB_ACCOUNT|
						SLAB_PANIC),
					     init_once);
	pde_opener_cache =
		kmem_cache_create("pde_opener", sizeof(struct pde_opener), 0,
				  SLAB_ACCOUNT|SLAB_PANIC, NULL);
	proc_dir_entry_cache = kmem_cache_create_usercopy(
		"proc_dir_entry", SIZEOF_PDE, 0, SLAB_PANIC,
		offsetof(struct proc_dir_entry, inline_name),
		SIZEOF_PDE_INLINE_NAME, NULL);
	BUILD_BUG_ON(sizeof(struct proc_dir_entry) >= SIZEOF_PDE);
}

static int proc_show_options(struct seq_file *seq, struct dentry *root)
{
	struct super_block *sb = root->d_sb;
	struct pid_namespace *pid = sb->s_fs_info;

	if (!gid_eq(pid->pid_gid, GLOBAL_ROOT_GID))
		seq_printf(seq, ",gid=%u", from_kgid_munged(&init_user_ns, pid->pid_gid));
	if (pid->hide_pid != HIDEPID_OFF)
		seq_printf(seq, ",hidepid=%u", pid->hide_pid);

	return 0;
}

static const struct super_operations proc_sops = {
	.alloc_inode	= proc_alloc_inode,
	.destroy_inode	= proc_destroy_inode,
	.drop_inode	= generic_delete_inode,
	.evict_inode	= proc_evict_inode,
	.statfs		= simple_statfs,
	.remount_fs	= proc_remount,
	.show_options	= proc_show_options,
};

enum {BIAS = -1U<<31};

static inline int use_pde(struct proc_dir_entry *pde)
{
	return likely(atomic_inc_unless_negative(&pde->in_use));
}

static void unuse_pde(struct proc_dir_entry *pde)
{
	if (unlikely(atomic_dec_return(&pde->in_use) == BIAS))
		complete(pde->pde_unload_completion);
}

/* pde is locked on entry, unlocked on exit */
static void close_pdeo(struct proc_dir_entry *pde, struct pde_opener *pdeo)
{
	/*
	 * close() (proc_reg_release()) can't delete an entry and proceed:
	 * ->release hook needs to be available at the right moment.
	 *
	 * rmmod (remove_proc_entry() et al) can't delete an entry and proceed:
	 * "struct file" needs to be available at the right moment.
	 *
	 * Therefore, first process to enter this function does ->release() and
	 * signals its completion to the other process which does nothing.
	 */
	if (pdeo->closing) {
		/* somebody else is doing that, just wait */
		DECLARE_COMPLETION_ONSTACK(c);
		pdeo->c = &c;
		spin_unlock(&pde->pde_unload_lock);
		wait_for_completion(&c);
	} else {
		struct file *file;
		struct completion *c;

		pdeo->closing = true;
		spin_unlock(&pde->pde_unload_lock);
		file = pdeo->file;
		pde->proc_fops->release(file_inode(file), file);
		spin_lock(&pde->pde_unload_lock);
		/* After ->release. */
		list_del(&pdeo->lh);
		c = pdeo->c;
		spin_unlock(&pde->pde_unload_lock);
		if (unlikely(c))
			complete(c);
		kmem_cache_free(pde_opener_cache, pdeo);
	}
}

void proc_entry_rundown(struct proc_dir_entry *de)
{
	DECLARE_COMPLETION_ONSTACK(c);
	/* Wait until all existing callers into module are done. */
	de->pde_unload_completion = &c;
	if (atomic_add_return(BIAS, &de->in_use) != BIAS)
		wait_for_completion(&c);

	/* ->pde_openers list can't grow from now on. */

	spin_lock(&de->pde_unload_lock);
	while (!list_empty(&de->pde_openers)) {
		struct pde_opener *pdeo;
		pdeo = list_first_entry(&de->pde_openers, struct pde_opener, lh);
		close_pdeo(de, pdeo);
		spin_lock(&de->pde_unload_lock);
	}
	spin_unlock(&de->pde_unload_lock);
}

static loff_t proc_reg_llseek(struct file *file, loff_t offset, int whence)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	loff_t rv = -EINVAL;
	if (use_pde(pde)) {
		loff_t (*llseek)(struct file *, loff_t, int);
		llseek = pde->proc_fops->llseek;
		if (!llseek)
			llseek = default_llseek;
		rv = llseek(file, offset, whence);
		unuse_pde(pde);
	}
	return rv;
}

static ssize_t proc_reg_read(struct file *file, char __user *buf, size_t count, loff_t *ppos)
{
	ssize_t (*read)(struct file *, char __user *, size_t, loff_t *);
	struct proc_dir_entry *pde = PDE(file_inode(file));
	ssize_t rv = -EIO;
	if (use_pde(pde)) {
		read = pde->proc_fops->read;
		if (read)
			rv = read(file, buf, count, ppos);
		unuse_pde(pde);
	}
	return rv;
}

static ssize_t proc_reg_write(struct file *file, const char __user *buf, size_t count, loff_t *ppos)
{
	ssize_t (*write)(struct file *, const char __user *, size_t, loff_t *);
	struct proc_dir_entry *pde = PDE(file_inode(file));
	ssize_t rv = -EIO;
	if (use_pde(pde)) {
		write = pde->proc_fops->write;
		if (write)
			rv = write(file, buf, count, ppos);
		unuse_pde(pde);
	}
	return rv;
}

static __poll_t proc_reg_poll(struct file *file, struct poll_table_struct *pts)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	__poll_t rv = DEFAULT_POLLMASK;
	__poll_t (*poll)(struct file *, struct poll_table_struct *);
	if (use_pde(pde)) {
		poll = pde->proc_fops->poll;
		if (poll)
			rv = poll(file, pts);
		unuse_pde(pde);
	}
	return rv;
}

static long proc_reg_unlocked_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	long rv = -ENOTTY;
	long (*ioctl)(struct file *, unsigned int, unsigned long);
	if (use_pde(pde)) {
		ioctl = pde->proc_fops->unlocked_ioctl;
		if (ioctl)
			rv = ioctl(file, cmd, arg);
		unuse_pde(pde);
	}
	return rv;
}

#ifdef CONFIG_COMPAT
static long proc_reg_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	long rv = -ENOTTY;
	long (*compat_ioctl)(struct file *, unsigned int, unsigned long);
	if (use_pde(pde)) {
		compat_ioctl = pde->proc_fops->compat_ioctl;
		if (compat_ioctl)
			rv = compat_ioctl(file, cmd, arg);
		unuse_pde(pde);
	}
	return rv;
}
#endif

static int proc_reg_mmap(struct file *file, struct vm_area_struct *vma)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	int rv = -EIO;
	int (*mmap)(struct file *, struct vm_area_struct *);
	if (use_pde(pde)) {
		mmap = pde->proc_fops->mmap;
		if (mmap)
			rv = mmap(file, vma);
		unuse_pde(pde);
	}
	return rv;
}

static unsigned long
proc_reg_get_unmapped_area(struct file *file, unsigned long orig_addr,
			   unsigned long len, unsigned long pgoff,
			   unsigned long flags)
{
	struct proc_dir_entry *pde = PDE(file_inode(file));
	unsigned long rv = -EIO;

	if (use_pde(pde)) {
		typeof(proc_reg_get_unmapped_area) *get_area;

		get_area = pde->proc_fops->get_unmapped_area;
#ifdef CONFIG_MMU
		if (!get_area)
			get_area = current->mm->get_unmapped_area;
#endif

		if (get_area)
			rv = get_area(file, orig_addr, len, pgoff, flags);
		else
			rv = orig_addr;
		unuse_pde(pde);
	}
	return rv;
}

static int proc_reg_open(struct inode *inode, struct file *file)
{
	struct proc_dir_entry *pde = PDE(inode);
	int rv = 0;
	int (*open)(struct inode *, struct file *);
	int (*release)(struct inode *, struct file *);
	struct pde_opener *pdeo;

	/*
	 * Ensure that
	 * 1) PDE's ->release hook will be called no matter what
	 *    either normally by close()/->release, or forcefully by
	 *    rmmod/remove_proc_entry.
	 *
	 * 2) rmmod isn't blocked by opening file in /proc and sitting on
	 *    the descriptor (including "rmmod foo </proc/foo" scenario).
	 *
	 * Save every "struct file" with custom ->release hook.
	 */
	if (!use_pde(pde))
		return -ENOENT;

	release = pde->proc_fops->release;
	if (release) {
		pdeo = kmem_cache_alloc(pde_opener_cache, GFP_KERNEL);
		if (!pdeo) {
			rv = -ENOMEM;
			goto out_unuse;
		}
	}

	open = pde->proc_fops->open;
	if (open)
		rv = open(inode, file);

	if (release) {
		if (rv == 0) {
			/* To know what to release. */
			pdeo->file = file;
			pdeo->closing = false;
			pdeo->c = NULL;
			spin_lock(&pde->pde_unload_lock);
			list_add(&pdeo->lh, &pde->pde_openers);
			spin_unlock(&pde->pde_unload_lock);
		} else
			kmem_cache_free(pde_opener_cache, pdeo);
	}

out_unuse:
	unuse_pde(pde);
	return rv;
}

static int proc_reg_release(struct inode *inode, struct file *file)
{
	struct proc_dir_entry *pde = PDE(inode);
	struct pde_opener *pdeo;
	spin_lock(&pde->pde_unload_lock);
	list_for_each_entry(pdeo, &pde->pde_openers, lh) {
		if (pdeo->file == file) {
			close_pdeo(pde, pdeo);
			return 0;
		}
	}
	spin_unlock(&pde->pde_unload_lock);
	return 0;
}

static const struct file_operations proc_reg_file_ops = {
	.llseek		= proc_reg_llseek,
	.read		= proc_reg_read,
	.write		= proc_reg_write,
	.poll		= proc_reg_poll,
	.unlocked_ioctl	= proc_reg_unlocked_ioctl,
#ifdef CONFIG_COMPAT
	.compat_ioctl	= proc_reg_compat_ioctl,
#endif
	.mmap		= proc_reg_mmap,
	.get_unmapped_area = proc_reg_get_unmapped_area,
	.open		= proc_reg_open,
	.release	= proc_reg_release,
};

#ifdef CONFIG_COMPAT
static const struct file_operations proc_reg_file_ops_no_compat = {
	.llseek		= proc_reg_llseek,
	.read		= proc_reg_read,
	.write		= proc_reg_write,
	.poll		= proc_reg_poll,
	.unlocked_ioctl	= proc_reg_unlocked_ioctl,
	.mmap		= proc_reg_mmap,
	.get_unmapped_area = proc_reg_get_unmapped_area,
	.open		= proc_reg_open,
	.release	= proc_reg_release,
};
#endif

static void proc_put_link(void *p)
{
	unuse_pde(p);
}

static const char *proc_get_link(struct dentry *dentry,
				 struct inode *inode,
				 struct delayed_call *done)
{
	struct proc_dir_entry *pde = PDE(inode);
	if (!use_pde(pde))
		return ERR_PTR(-EINVAL);
	set_delayed_call(done, proc_put_link, pde);
	return pde->data;
}

const struct inode_operations proc_link_inode_operations = {
	.get_link	= proc_get_link,
};

struct inode *proc_get_inode(struct super_block *sb, struct proc_dir_entry *de)
{
	struct inode *inode = new_inode_pseudo(sb);

	if (inode) {
		inode->i_ino = de->low_ino;
		inode->i_mtime = inode->i_atime = inode->i_ctime = current_time(inode);
		PROC_I(inode)->pde = de;

		if (is_empty_pde(de)) {
			make_empty_dir_inode(inode);
			return inode;
		}
		if (de->mode) {
			inode->i_mode = de->mode;
			inode->i_uid = de->uid;
			inode->i_gid = de->gid;
		}
		if (de->size)
			inode->i_size = de->size;
		if (de->nlink)
			set_nlink(inode, de->nlink);
		WARN_ON(!de->proc_iops);
		inode->i_op = de->proc_iops;
		if (de->proc_fops) {
			if (S_ISREG(inode->i_mode)) {
#ifdef CONFIG_COMPAT
				if (!de->proc_fops->compat_ioctl)
					inode->i_fop =
						&proc_reg_file_ops_no_compat;
				else
#endif
					inode->i_fop = &proc_reg_file_ops;
			} else {
				inode->i_fop = de->proc_fops;
			}
		}
	} else
	       pde_put(de);
	return inode;
}

int proc_fill_super(struct super_block *s, void *data, int silent)
{
	struct pid_namespace *ns = get_pid_ns(s->s_fs_info);
	struct inode *root_inode;
	int ret;

	if (!proc_parse_options(data, ns))
		return -EINVAL;

	/* User space would break if executables or devices appear on proc */
	s->s_iflags |= SB_I_USERNS_VISIBLE | SB_I_NOEXEC | SB_I_NODEV;
	s->s_flags |= SB_NODIRATIME | SB_NOSUID | SB_NOEXEC;
	s->s_blocksize = 1024;
	s->s_blocksize_bits = 10;
	s->s_magic = PROC_SUPER_MAGIC;
	s->s_op = &proc_sops;
	s->s_time_gran = 1;

	/*
	 * procfs isn't actually a stacking filesystem; however, there is
	 * too much magic going on inside it to permit stacking things on
	 * top of it
	 */
	s->s_stack_depth = FILESYSTEM_MAX_STACK_DEPTH;
	
	/* procfs dentries and inodes don't require IO to create */
	s->s_shrink.seeks = 0;

	pde_get(&proc_root);
	root_inode = proc_get_inode(s, &proc_root);
	if (!root_inode) {
		pr_err("proc_fill_super: get root inode failed\n");
		return -ENOMEM;
	}

	s->s_root = d_make_root(root_inode);
	if (!s->s_root) {
		pr_err("proc_fill_super: allocate dentry failed\n");
		return -ENOMEM;
	}

	ret = proc_setup_self(s);
	if (ret) {
		return ret;
	}
	return proc_setup_thread_self(s);
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
1da177e4 LT	2	/*
	3	* linux/fs/proc/inode.c
	4	*
	5	* Copyright (C) 1991, 1992 Linus Torvalds
	6	*/
	7
efb1a57d	8	#include <linux/cache.h>
1da177e4 LT	9	#include <linux/time.h>
	10	#include <linux/proc_fs.h>
	11	#include <linux/kernel.h>
97412950	12	#include <linux/pid_namespace.h>
1da177e4 LT	13	#include <linux/mm.h>
	14	#include <linux/string.h>
	15	#include <linux/stat.h>
786d7e16	16	#include <linux/completion.h>
dd23aae4	17	#include <linux/poll.h>
87ebdc00	18	#include <linux/printk.h>
1da177e4 LT	19	#include <linux/file.h>
	20	#include <linux/limits.h>
	21	#include <linux/init.h>
	22	#include <linux/module.h>
9043476f	23	#include <linux/sysctl.h>
97412950	24	#include <linux/seq_file.h>
5a0e3ad6	25	#include <linux/slab.h>
97412950	26	#include <linux/mount.h>
303eb7e2	27	#include <linux/magic.h>
1da177e4	28
7c0f6ba6	29	#include <linux/uaccess.h>
1da177e4	30
fee781e6	31	#include "internal.h"
1da177e4	32
8267952b	33	static void proc_evict_inode(struct inode *inode)
1da177e4 LT	34	{
1da177e4 LT	35	struct proc_dir_entry *de;
dfef6dcd	36	struct ctl_table_header *head;
1da177e4	37
91b0abe3	38	truncate_inode_pages_final(&inode->i_data);
dbd5768f	39	clear_inode(inode);
fef26658	40
99f89551	41	/* Stop tracking associated processes */
13b41b09	42	put_pid(PROC_I(inode)->pid);
1da177e4 LT	43
1da177e4 LT	44	/* Let go of any associated proc directory entry */
6bee55f9	45	de = PDE(inode);
99b76233	46	if (de)
135d5655	47	pde_put(de);
d6cffbbe	48
dfef6dcd AV	49	head = PROC_I(inode)->sysctl;
dfef6dcd AV	50	if (head) {
1c44dbc8	51	RCU_INIT_POINTER(PROC_I(inode)->sysctl, NULL);
d6cffbbe	52	proc_sys_evict_inode(inode, head);
dfef6dcd	53	}
1da177e4 LT	54	}
1da177e4 LT	55
efb1a57d	56	static struct kmem_cache *proc_inode_cachep __ro_after_init;
195b8cf0	57	static struct kmem_cache *pde_opener_cache __ro_after_init;
1da177e4 LT	58
	59	static struct inode proc_alloc_inode(struct super_block sb)
	60	{
	61	struct proc_inode *ei;
	62	struct inode *inode;
	63
f245e1c1	64	ei = kmem_cache_alloc(proc_inode_cachep, GFP_KERNEL);
1da177e4 LT	65	if (!ei)
1da177e4 LT	66	return NULL;
13b41b09	67	ei->pid = NULL;
aed7a6c4	68	ei->fd = 0;
1da177e4 LT	69	ei->op.proc_get_link = NULL;
1da177e4 LT	70	ei->pde = NULL;
9043476f AV	71	ei->sysctl = NULL;
9043476f AV	72	ei->sysctl_entry = NULL;
3d3d35b1	73	ei->ns_ops = NULL;
1da177e4	74	inode = &ei->vfs_inode;
1da177e4 LT	75	return inode;
	76	}
	77
fa0d7e3d	78	static void proc_i_callback(struct rcu_head *head)
1da177e4	79	{
fa0d7e3d	80	struct inode *inode = container_of(head, struct inode, i_rcu);
1da177e4 LT	81	kmem_cache_free(proc_inode_cachep, PROC_I(inode));
	82	}
	83
fa0d7e3d NP	84	static void proc_destroy_inode(struct inode *inode)
	85	{
	86	call_rcu(&inode->i_rcu, proc_i_callback);
	87	}
	88
51cc5068	89	static void init_once(void *foo)
1da177e4 LT	90	{
	91	struct proc_inode ei = (struct proc_inode ) foo;
	92
a35afb83	93	inode_init_once(&ei->vfs_inode);
1da177e4	94	}
20c2df83	95
195b8cf0	96	void __init proc_init_kmemcache(void)
1da177e4 LT	97	{
	98	proc_inode_cachep = kmem_cache_create("proc_inode_cache",
	99	sizeof(struct proc_inode),
fffb60f9	100	0, (SLAB_RECLAIM_ACCOUNT\|
5d097056 VD	101	SLAB_MEM_SPREAD\|SLAB_ACCOUNT\|
5d097056 VD	102	SLAB_PANIC),
20c2df83	103	init_once);
195b8cf0 AD	104	pde_opener_cache =
195b8cf0 AD	105	kmem_cache_create("pde_opener", sizeof(struct pde_opener), 0,
2acddbe8	106	SLAB_ACCOUNT\|SLAB_PANIC, NULL);
b4884f23	107	proc_dir_entry_cache = kmem_cache_create_usercopy(
2d6e4e82 AD	108	"proc_dir_entry", SIZEOF_PDE, 0, SLAB_PANIC,
	109	offsetof(struct proc_dir_entry, inline_name),
	110	SIZEOF_PDE_INLINE_NAME, NULL);
	111	BUILD_BUG_ON(sizeof(struct proc_dir_entry) >= SIZEOF_PDE);
1da177e4 LT	112	}
1da177e4 LT	113
97412950 VK	114	static int proc_show_options(struct seq_file seq, struct dentry root)
97412950 VK	115	{
0499680a VK	116	struct super_block *sb = root->d_sb;
	117	struct pid_namespace *pid = sb->s_fs_info;
	118
dcb0f222 EB	119	if (!gid_eq(pid->pid_gid, GLOBAL_ROOT_GID))
dcb0f222 EB	120	seq_printf(seq, ",gid=%u", from_kgid_munged(&init_user_ns, pid->pid_gid));
796f571b	121	if (pid->hide_pid != HIDEPID_OFF)
0499680a VK	122	seq_printf(seq, ",hidepid=%u", pid->hide_pid);
0499680a VK	123
97412950 VK	124	return 0;
	125	}
	126
ee9b6d61	127	static const struct super_operations proc_sops = {
1da177e4 LT	128	.alloc_inode = proc_alloc_inode,
1da177e4 LT	129	.destroy_inode = proc_destroy_inode,
1da177e4	130	.drop_inode = generic_delete_inode,
8267952b	131	.evict_inode = proc_evict_inode,
1da177e4	132	.statfs = simple_statfs,
97412950 VK	133	.remount_fs = proc_remount,
97412950 VK	134	.show_options = proc_show_options,
1da177e4 LT	135	};
1da177e4 LT	136
866ad9a7 AV	137	enum {BIAS = -1U<<31};
	138
	139	static inline int use_pde(struct proc_dir_entry *pde)
	140	{
15b158b4	141	return likely(atomic_inc_unless_negative(&pde->in_use));
881adb85 AD	142	}
881adb85 AD	143
866ad9a7	144	static void unuse_pde(struct proc_dir_entry *pde)
881adb85	145	{
15b158b4	146	if (unlikely(atomic_dec_return(&pde->in_use) == BIAS))
05c0ae21	147	complete(pde->pde_unload_completion);
786d7e16 AD	148	}
786d7e16 AD	149
2f897424	150	/* pde is locked on entry, unlocked on exit */
ca469f35 AV	151	static void close_pdeo(struct proc_dir_entry pde, struct pde_opener pdeo)
ca469f35 AV	152	{
492b2da6 AD	153	/*
	154	* close() (proc_reg_release()) can't delete an entry and proceed:
	155	* ->release hook needs to be available at the right moment.
	156	*
	157	* rmmod (remove_proc_entry() et al) can't delete an entry and proceed:
	158	* "struct file" needs to be available at the right moment.
	159	*
	160	* Therefore, first process to enter this function does ->release() and
	161	* signals its completion to the other process which does nothing.
	162	*/
05c0ae21	163	if (pdeo->closing) {
ca469f35	164	/* somebody else is doing that, just wait */
05c0ae21 AV	165	DECLARE_COMPLETION_ONSTACK(c);
05c0ae21 AV	166	pdeo->c = &c;
ca469f35	167	spin_unlock(&pde->pde_unload_lock);
05c0ae21	168	wait_for_completion(&c);
ca469f35 AV	169	} else {
ca469f35 AV	170	struct file *file;
2f897424 AD	171	struct completion *c;
2f897424 AD	172
f5887c71	173	pdeo->closing = true;
ca469f35 AV	174	spin_unlock(&pde->pde_unload_lock);
	175	file = pdeo->file;
	176	pde->proc_fops->release(file_inode(file), file);
	177	spin_lock(&pde->pde_unload_lock);
492b2da6	178	/* After ->release. */
06a0c417	179	list_del(&pdeo->lh);
2f897424 AD	180	c = pdeo->c;
	181	spin_unlock(&pde->pde_unload_lock);
	182	if (unlikely(c))
	183	complete(c);
195b8cf0	184	kmem_cache_free(pde_opener_cache, pdeo);
05c0ae21	185	}
ca469f35 AV	186	}
ca469f35 AV	187
866ad9a7	188	void proc_entry_rundown(struct proc_dir_entry *de)
786d7e16	189	{
05c0ae21	190	DECLARE_COMPLETION_ONSTACK(c);
866ad9a7	191	/* Wait until all existing callers into module are done. */
05c0ae21 AV	192	de->pde_unload_completion = &c;
	193	if (atomic_add_return(BIAS, &de->in_use) != BIAS)
	194	wait_for_completion(&c);
786d7e16	195
492b2da6 AD	196	/* ->pde_openers list can't grow from now on. */
492b2da6 AD	197
05c0ae21	198	spin_lock(&de->pde_unload_lock);
866ad9a7 AV	199	while (!list_empty(&de->pde_openers)) {
866ad9a7 AV	200	struct pde_opener *pdeo;
866ad9a7	201	pdeo = list_first_entry(&de->pde_openers, struct pde_opener, lh);
ca469f35	202	close_pdeo(de, pdeo);
2f897424	203	spin_lock(&de->pde_unload_lock);
866ad9a7 AV	204	}
866ad9a7 AV	205	spin_unlock(&de->pde_unload_lock);
786d7e16 AD	206	}
786d7e16 AD	207
866ad9a7 AV	208	static loff_t proc_reg_llseek(struct file *file, loff_t offset, int whence)
	209	{
	210	struct proc_dir_entry *pde = PDE(file_inode(file));
	211	loff_t rv = -EINVAL;
	212	if (use_pde(pde)) {
	213	loff_t (llseek)(struct file , loff_t, int);
	214	llseek = pde->proc_fops->llseek;
	215	if (!llseek)
	216	llseek = default_llseek;
	217	rv = llseek(file, offset, whence);
	218	unuse_pde(pde);
	219	}
786d7e16 AD	220	return rv;
	221	}
	222
866ad9a7	223	static ssize_t proc_reg_read(struct file file, char __user buf, size_t count, loff_t *ppos)
786d7e16	224	{
866ad9a7	225	ssize_t (read)(struct file , char __user , size_t, loff_t );
496ad9aa	226	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16	227	ssize_t rv = -EIO;
866ad9a7 AV	228	if (use_pde(pde)) {
	229	read = pde->proc_fops->read;
	230	if (read)
	231	rv = read(file, buf, count, ppos);
	232	unuse_pde(pde);
786d7e16	233	}
866ad9a7 AV	234	return rv;
866ad9a7 AV	235	}
786d7e16	236
866ad9a7 AV	237	static ssize_t proc_reg_write(struct file file, const char __user buf, size_t count, loff_t *ppos)
	238	{
	239	ssize_t (write)(struct file , const char __user , size_t, loff_t );
	240	struct proc_dir_entry *pde = PDE(file_inode(file));
	241	ssize_t rv = -EIO;
	242	if (use_pde(pde)) {
	243	write = pde->proc_fops->write;
	244	if (write)
	245	rv = write(file, buf, count, ppos);
	246	unuse_pde(pde);
	247	}
786d7e16 AD	248	return rv;
	249	}
	250
076ccb76	251	static __poll_t proc_reg_poll(struct file file, struct poll_table_struct pts)
786d7e16	252	{
496ad9aa	253	struct proc_dir_entry *pde = PDE(file_inode(file));
e6c8adca	254	__poll_t rv = DEFAULT_POLLMASK;
a3f8683b	255	__poll_t (poll)(struct file , struct poll_table_struct *);
866ad9a7 AV	256	if (use_pde(pde)) {
	257	poll = pde->proc_fops->poll;
	258	if (poll)
	259	rv = poll(file, pts);
	260	unuse_pde(pde);
786d7e16	261	}
786d7e16 AD	262	return rv;
	263	}
	264
	265	static long proc_reg_unlocked_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
	266	{
496ad9aa	267	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16	268	long rv = -ENOTTY;
b19dd42f	269	long (ioctl)(struct file , unsigned int, unsigned long);
866ad9a7 AV	270	if (use_pde(pde)) {
	271	ioctl = pde->proc_fops->unlocked_ioctl;
	272	if (ioctl)
	273	rv = ioctl(file, cmd, arg);
	274	unuse_pde(pde);
786d7e16	275	}
786d7e16 AD	276	return rv;
	277	}
	278
	279	#ifdef CONFIG_COMPAT
	280	static long proc_reg_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
	281	{
496ad9aa	282	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16 AD	283	long rv = -ENOTTY;
786d7e16 AD	284	long (compat_ioctl)(struct file , unsigned int, unsigned long);
866ad9a7 AV	285	if (use_pde(pde)) {
	286	compat_ioctl = pde->proc_fops->compat_ioctl;
	287	if (compat_ioctl)
	288	rv = compat_ioctl(file, cmd, arg);
	289	unuse_pde(pde);
786d7e16	290	}
786d7e16 AD	291	return rv;
	292	}
	293	#endif
	294
	295	static int proc_reg_mmap(struct file file, struct vm_area_struct vma)
	296	{
496ad9aa	297	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16 AD	298	int rv = -EIO;
786d7e16 AD	299	int (mmap)(struct file , struct vm_area_struct *);
866ad9a7 AV	300	if (use_pde(pde)) {
	301	mmap = pde->proc_fops->mmap;
	302	if (mmap)
	303	rv = mmap(file, vma);
	304	unuse_pde(pde);
786d7e16	305	}
786d7e16 AD	306	return rv;
	307	}
	308
5721cf84 HD	309	static unsigned long
	310	proc_reg_get_unmapped_area(struct file *file, unsigned long orig_addr,
	311	unsigned long len, unsigned long pgoff,
	312	unsigned long flags)
c4fe2448 AD	313	{
c4fe2448 AD	314	struct proc_dir_entry *pde = PDE(file_inode(file));
2cbe3b0a	315	unsigned long rv = -EIO;
ae5758a1	316
c4fe2448	317	if (use_pde(pde)) {
ae5758a1 JB	318	typeof(proc_reg_get_unmapped_area) *get_area;
	319
	320	get_area = pde->proc_fops->get_unmapped_area;
fad1a86e	321	#ifdef CONFIG_MMU
ae5758a1 JB	322	if (!get_area)
ae5758a1 JB	323	get_area = current->mm->get_unmapped_area;
fad1a86e	324	#endif
ae5758a1	325
5721cf84 HD	326	if (get_area)
5721cf84 HD	327	rv = get_area(file, orig_addr, len, pgoff, flags);
ae5758a1 JB	328	else
ae5758a1 JB	329	rv = orig_addr;
c4fe2448 AD	330	unuse_pde(pde);
	331	}
	332	return rv;
	333	}
	334
786d7e16 AD	335	static int proc_reg_open(struct inode inode, struct file file)
	336	{
	337	struct proc_dir_entry *pde = PDE(inode);
	338	int rv = 0;
	339	int (open)(struct inode , struct file *);
881adb85 AD	340	int (release)(struct inode , struct file *);
	341	struct pde_opener *pdeo;
	342
	343	/*
492b2da6 AD	344	* Ensure that
	345	* 1) PDE's ->release hook will be called no matter what
	346	* either normally by close()/->release, or forcefully by
	347	* rmmod/remove_proc_entry.
	348	*
	349	* 2) rmmod isn't blocked by opening file in /proc and sitting on
	350	* the descriptor (including "rmmod foo </proc/foo" scenario).
881adb85	351	*
492b2da6	352	* Save every "struct file" with custom ->release hook.
881adb85	353	*/
e7a6e291	354	if (!use_pde(pde))
d2857e79	355	return -ENOENT;
e7a6e291	356
881adb85	357	release = pde->proc_fops->release;
e7a6e291	358	if (release) {
195b8cf0	359	pdeo = kmem_cache_alloc(pde_opener_cache, GFP_KERNEL);
e7a6e291 AD	360	if (!pdeo) {
	361	rv = -ENOMEM;
	362	goto out_unuse;
	363	}
	364	}
786d7e16	365
e7a6e291	366	open = pde->proc_fops->open;
786d7e16 AD	367	if (open)
	368	rv = open(inode, file);
	369
e7a6e291 AD	370	if (release) {
	371	if (rv == 0) {
	372	/* To know what to release. */
	373	pdeo->file = file;
	374	pdeo->closing = false;
	375	pdeo->c = NULL;
	376	spin_lock(&pde->pde_unload_lock);
	377	list_add(&pdeo->lh, &pde->pde_openers);
	378	spin_unlock(&pde->pde_unload_lock);
	379	} else
195b8cf0	380	kmem_cache_free(pde_opener_cache, pdeo);
e7a6e291	381	}
05c0ae21	382
e7a6e291	383	out_unuse:
05c0ae21	384	unuse_pde(pde);
786d7e16 AD	385	return rv;
	386	}
	387
	388	static int proc_reg_release(struct inode inode, struct file file)
	389	{
	390	struct proc_dir_entry *pde = PDE(inode);
881adb85	391	struct pde_opener *pdeo;
786d7e16	392	spin_lock(&pde->pde_unload_lock);
ca469f35 AV	393	list_for_each_entry(pdeo, &pde->pde_openers, lh) {
	394	if (pdeo->file == file) {
	395	close_pdeo(pde, pdeo);
2f897424	396	return 0;
ca469f35	397	}
881adb85	398	}
786d7e16	399	spin_unlock(&pde->pde_unload_lock);
ca469f35	400	return 0;
786d7e16 AD	401	}
	402
	403	static const struct file_operations proc_reg_file_ops = {
	404	.llseek = proc_reg_llseek,
	405	.read = proc_reg_read,
	406	.write = proc_reg_write,
	407	.poll = proc_reg_poll,
	408	.unlocked_ioctl = proc_reg_unlocked_ioctl,
	409	#ifdef CONFIG_COMPAT
	410	.compat_ioctl = proc_reg_compat_ioctl,
	411	#endif
	412	.mmap = proc_reg_mmap,
c4fe2448	413	.get_unmapped_area = proc_reg_get_unmapped_area,
786d7e16 AD	414	.open = proc_reg_open,
	415	.release = proc_reg_release,
	416	};
	417
778f3dd5 DM	418	#ifdef CONFIG_COMPAT
	419	static const struct file_operations proc_reg_file_ops_no_compat = {
	420	.llseek = proc_reg_llseek,
	421	.read = proc_reg_read,
	422	.write = proc_reg_write,
	423	.poll = proc_reg_poll,
	424	.unlocked_ioctl = proc_reg_unlocked_ioctl,
	425	.mmap = proc_reg_mmap,
c4fe2448	426	.get_unmapped_area = proc_reg_get_unmapped_area,
778f3dd5 DM	427	.open = proc_reg_open,
	428	.release = proc_reg_release,
	429	};
	430	#endif
	431
fceef393 AV	432	static void proc_put_link(void *p)
	433	{
	434	unuse_pde(p);
	435	}
	436
6b255391	437	static const char proc_get_link(struct dentry dentry,
fceef393 AV	438	struct inode *inode,
fceef393 AV	439	struct delayed_call *done)
7e0e953b	440	{
6b255391	441	struct proc_dir_entry *pde = PDE(inode);
15b158b4	442	if (!use_pde(pde))
7e0e953b	443	return ERR_PTR(-EINVAL);
fceef393	444	set_delayed_call(done, proc_put_link, pde);
680baacb	445	return pde->data;
7e0e953b AV	446	}
7e0e953b AV	447
7e0e953b	448	const struct inode_operations proc_link_inode_operations = {
6b255391	449	.get_link = proc_get_link,
7e0e953b AV	450	};
7e0e953b AV	451
6d1b6e4e	452	struct inode proc_get_inode(struct super_block sb, struct proc_dir_entry *de)
1da177e4	453	{
51f0885e	454	struct inode *inode = new_inode_pseudo(sb);
1da177e4	455
51f0885e LT	456	if (inode) {
51f0885e LT	457	inode->i_ino = de->low_ino;
078cd827	458	inode->i_mtime = inode->i_atime = inode->i_ctime = current_time(inode);
a1d4aebb	459	PROC_I(inode)->pde = de;
5e971dce	460
eb6d38d5 EB	461	if (is_empty_pde(de)) {
	462	make_empty_dir_inode(inode);
	463	return inode;
	464	}
5e971dce AD	465	if (de->mode) {
	466	inode->i_mode = de->mode;
	467	inode->i_uid = de->uid;
	468	inode->i_gid = de->gid;
	469	}
	470	if (de->size)
	471	inode->i_size = de->size;
	472	if (de->nlink)
bfe86848	473	set_nlink(inode, de->nlink);
b6cdc731 AV	474	WARN_ON(!de->proc_iops);
b6cdc731 AV	475	inode->i_op = de->proc_iops;
5e971dce AD	476	if (de->proc_fops) {
5e971dce AD	477	if (S_ISREG(inode->i_mode)) {
778f3dd5	478	#ifdef CONFIG_COMPAT
5e971dce AD	479	if (!de->proc_fops->compat_ioctl)
	480	inode->i_fop =
	481	&proc_reg_file_ops_no_compat;
	482	else
778f3dd5	483	#endif
5e971dce AD	484	inode->i_fop = &proc_reg_file_ops;
	485	} else {
	486	inode->i_fop = de->proc_fops;
778f3dd5	487	}
786d7e16	488	}
99b76233	489	} else
135d5655	490	pde_put(de);
1da177e4	491	return inode;
d3d009cb	492	}
1da177e4	493
e94591d0	494	int proc_fill_super(struct super_block s, void data, int silent)
1da177e4	495	{
e94591d0	496	struct pid_namespace *ns = get_pid_ns(s->s_fs_info);
87e0aab3	497	struct inode *root_inode;
0097875b	498	int ret;
87e0aab3	499
e94591d0 EB	500	if (!proc_parse_options(data, ns))
	501	return -EINVAL;
	502
a2982cc9 EB	503	/* User space would break if executables or devices appear on proc */
a2982cc9 EB	504	s->s_iflags \|= SB_I_USERNS_VISIBLE \| SB_I_NOEXEC \| SB_I_NODEV;
1751e8a6	505	s->s_flags \|= SB_NODIRATIME \| SB_NOSUID \| SB_NOEXEC;
1da177e4 LT	506	s->s_blocksize = 1024;
	507	s->s_blocksize_bits = 10;
	508	s->s_magic = PROC_SUPER_MAGIC;
	509	s->s_op = &proc_sops;
	510	s->s_time_gran = 1;
a867d734 LT	511
	512	/*
	513	* procfs isn't actually a stacking filesystem; however, there is
	514	* too much magic going on inside it to permit stacking things on
	515	* top of it
	516	*/
	517	s->s_stack_depth = FILESYSTEM_MAX_STACK_DEPTH;
1da177e4	518
4b85afbd JW	519	/* procfs dentries and inodes don't require IO to create */
	520	s->s_shrink.seeks = 0;
	521
135d5655	522	pde_get(&proc_root);
87e0aab3 MP	523	root_inode = proc_get_inode(s, &proc_root);
87e0aab3 MP	524	if (!root_inode) {
87ebdc00	525	pr_err("proc_fill_super: get root inode failed\n");
87e0aab3 MP	526	return -ENOMEM;
87e0aab3 MP	527	}
1da177e4	528
87e0aab3 MP	529	s->s_root = d_make_root(root_inode);
87e0aab3 MP	530	if (!s->s_root) {
87ebdc00	531	pr_err("proc_fill_super: allocate dentry failed\n");
87e0aab3 MP	532	return -ENOMEM;
	533	}
	534
0097875b EB	535	ret = proc_setup_self(s);
	536	if (ret) {
	537	return ret;
	538	}
	539	return proc_setup_thread_self(s);
1da177e4	540	}