[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);
}

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;
	
	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);
1da177e4 LT	107	}
1da177e4 LT	108
97412950 VK	109	static int proc_show_options(struct seq_file seq, struct dentry root)
97412950 VK	110	{
0499680a VK	111	struct super_block *sb = root->d_sb;
	112	struct pid_namespace *pid = sb->s_fs_info;
	113
dcb0f222 EB	114	if (!gid_eq(pid->pid_gid, GLOBAL_ROOT_GID))
dcb0f222 EB	115	seq_printf(seq, ",gid=%u", from_kgid_munged(&init_user_ns, pid->pid_gid));
796f571b	116	if (pid->hide_pid != HIDEPID_OFF)
0499680a VK	117	seq_printf(seq, ",hidepid=%u", pid->hide_pid);
0499680a VK	118
97412950 VK	119	return 0;
	120	}
	121
ee9b6d61	122	static const struct super_operations proc_sops = {
1da177e4 LT	123	.alloc_inode = proc_alloc_inode,
1da177e4 LT	124	.destroy_inode = proc_destroy_inode,
1da177e4	125	.drop_inode = generic_delete_inode,
8267952b	126	.evict_inode = proc_evict_inode,
1da177e4	127	.statfs = simple_statfs,
97412950 VK	128	.remount_fs = proc_remount,
97412950 VK	129	.show_options = proc_show_options,
1da177e4 LT	130	};
1da177e4 LT	131
866ad9a7 AV	132	enum {BIAS = -1U<<31};
	133
	134	static inline int use_pde(struct proc_dir_entry *pde)
	135	{
15b158b4	136	return likely(atomic_inc_unless_negative(&pde->in_use));
881adb85 AD	137	}
881adb85 AD	138
866ad9a7	139	static void unuse_pde(struct proc_dir_entry *pde)
881adb85	140	{
15b158b4	141	if (unlikely(atomic_dec_return(&pde->in_use) == BIAS))
05c0ae21	142	complete(pde->pde_unload_completion);
786d7e16 AD	143	}
786d7e16 AD	144
2f897424	145	/* pde is locked on entry, unlocked on exit */
ca469f35 AV	146	static void close_pdeo(struct proc_dir_entry pde, struct pde_opener pdeo)
ca469f35 AV	147	{
492b2da6 AD	148	/*
	149	* close() (proc_reg_release()) can't delete an entry and proceed:
	150	* ->release hook needs to be available at the right moment.
	151	*
	152	* rmmod (remove_proc_entry() et al) can't delete an entry and proceed:
	153	* "struct file" needs to be available at the right moment.
	154	*
	155	* Therefore, first process to enter this function does ->release() and
	156	* signals its completion to the other process which does nothing.
	157	*/
05c0ae21	158	if (pdeo->closing) {
ca469f35	159	/* somebody else is doing that, just wait */
05c0ae21 AV	160	DECLARE_COMPLETION_ONSTACK(c);
05c0ae21 AV	161	pdeo->c = &c;
ca469f35	162	spin_unlock(&pde->pde_unload_lock);
05c0ae21	163	wait_for_completion(&c);
ca469f35 AV	164	} else {
ca469f35 AV	165	struct file *file;
2f897424 AD	166	struct completion *c;
2f897424 AD	167
f5887c71	168	pdeo->closing = true;
ca469f35 AV	169	spin_unlock(&pde->pde_unload_lock);
	170	file = pdeo->file;
	171	pde->proc_fops->release(file_inode(file), file);
	172	spin_lock(&pde->pde_unload_lock);
492b2da6	173	/* After ->release. */
06a0c417	174	list_del(&pdeo->lh);
2f897424 AD	175	c = pdeo->c;
	176	spin_unlock(&pde->pde_unload_lock);
	177	if (unlikely(c))
	178	complete(c);
195b8cf0	179	kmem_cache_free(pde_opener_cache, pdeo);
05c0ae21	180	}
ca469f35 AV	181	}
ca469f35 AV	182
866ad9a7	183	void proc_entry_rundown(struct proc_dir_entry *de)
786d7e16	184	{
05c0ae21	185	DECLARE_COMPLETION_ONSTACK(c);
866ad9a7	186	/* Wait until all existing callers into module are done. */
05c0ae21 AV	187	de->pde_unload_completion = &c;
	188	if (atomic_add_return(BIAS, &de->in_use) != BIAS)
	189	wait_for_completion(&c);
786d7e16	190
492b2da6 AD	191	/* ->pde_openers list can't grow from now on. */
492b2da6 AD	192
05c0ae21	193	spin_lock(&de->pde_unload_lock);
866ad9a7 AV	194	while (!list_empty(&de->pde_openers)) {
866ad9a7 AV	195	struct pde_opener *pdeo;
866ad9a7	196	pdeo = list_first_entry(&de->pde_openers, struct pde_opener, lh);
ca469f35	197	close_pdeo(de, pdeo);
2f897424	198	spin_lock(&de->pde_unload_lock);
866ad9a7 AV	199	}
866ad9a7 AV	200	spin_unlock(&de->pde_unload_lock);
786d7e16 AD	201	}
786d7e16 AD	202
866ad9a7 AV	203	static loff_t proc_reg_llseek(struct file *file, loff_t offset, int whence)
	204	{
	205	struct proc_dir_entry *pde = PDE(file_inode(file));
	206	loff_t rv = -EINVAL;
	207	if (use_pde(pde)) {
	208	loff_t (llseek)(struct file , loff_t, int);
	209	llseek = pde->proc_fops->llseek;
	210	if (!llseek)
	211	llseek = default_llseek;
	212	rv = llseek(file, offset, whence);
	213	unuse_pde(pde);
	214	}
786d7e16 AD	215	return rv;
	216	}
	217
866ad9a7	218	static ssize_t proc_reg_read(struct file file, char __user buf, size_t count, loff_t *ppos)
786d7e16	219	{
866ad9a7	220	ssize_t (read)(struct file , char __user , size_t, loff_t );
496ad9aa	221	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16	222	ssize_t rv = -EIO;
866ad9a7 AV	223	if (use_pde(pde)) {
	224	read = pde->proc_fops->read;
	225	if (read)
	226	rv = read(file, buf, count, ppos);
	227	unuse_pde(pde);
786d7e16	228	}
866ad9a7 AV	229	return rv;
866ad9a7 AV	230	}
786d7e16	231
866ad9a7 AV	232	static ssize_t proc_reg_write(struct file file, const char __user buf, size_t count, loff_t *ppos)
	233	{
	234	ssize_t (write)(struct file , const char __user , size_t, loff_t );
	235	struct proc_dir_entry *pde = PDE(file_inode(file));
	236	ssize_t rv = -EIO;
	237	if (use_pde(pde)) {
	238	write = pde->proc_fops->write;
	239	if (write)
	240	rv = write(file, buf, count, ppos);
	241	unuse_pde(pde);
	242	}
786d7e16 AD	243	return rv;
	244	}
	245
076ccb76	246	static __poll_t proc_reg_poll(struct file file, struct poll_table_struct pts)
786d7e16	247	{
496ad9aa	248	struct proc_dir_entry *pde = PDE(file_inode(file));
e6c8adca	249	__poll_t rv = DEFAULT_POLLMASK;
a3f8683b	250	__poll_t (poll)(struct file , struct poll_table_struct *);
866ad9a7 AV	251	if (use_pde(pde)) {
	252	poll = pde->proc_fops->poll;
	253	if (poll)
	254	rv = poll(file, pts);
	255	unuse_pde(pde);
786d7e16	256	}
786d7e16 AD	257	return rv;
	258	}
	259
	260	static long proc_reg_unlocked_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
	261	{
496ad9aa	262	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16	263	long rv = -ENOTTY;
b19dd42f	264	long (ioctl)(struct file , unsigned int, unsigned long);
866ad9a7 AV	265	if (use_pde(pde)) {
	266	ioctl = pde->proc_fops->unlocked_ioctl;
	267	if (ioctl)
	268	rv = ioctl(file, cmd, arg);
	269	unuse_pde(pde);
786d7e16	270	}
786d7e16 AD	271	return rv;
	272	}
	273
	274	#ifdef CONFIG_COMPAT
	275	static long proc_reg_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
	276	{
496ad9aa	277	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16 AD	278	long rv = -ENOTTY;
786d7e16 AD	279	long (compat_ioctl)(struct file , unsigned int, unsigned long);
866ad9a7 AV	280	if (use_pde(pde)) {
	281	compat_ioctl = pde->proc_fops->compat_ioctl;
	282	if (compat_ioctl)
	283	rv = compat_ioctl(file, cmd, arg);
	284	unuse_pde(pde);
786d7e16	285	}
786d7e16 AD	286	return rv;
	287	}
	288	#endif
	289
	290	static int proc_reg_mmap(struct file file, struct vm_area_struct vma)
	291	{
496ad9aa	292	struct proc_dir_entry *pde = PDE(file_inode(file));
786d7e16 AD	293	int rv = -EIO;
786d7e16 AD	294	int (mmap)(struct file , struct vm_area_struct *);
866ad9a7 AV	295	if (use_pde(pde)) {
	296	mmap = pde->proc_fops->mmap;
	297	if (mmap)
	298	rv = mmap(file, vma);
	299	unuse_pde(pde);
786d7e16	300	}
786d7e16 AD	301	return rv;
	302	}
	303
5721cf84 HD	304	static unsigned long
	305	proc_reg_get_unmapped_area(struct file *file, unsigned long orig_addr,
	306	unsigned long len, unsigned long pgoff,
	307	unsigned long flags)
c4fe2448 AD	308	{
c4fe2448 AD	309	struct proc_dir_entry *pde = PDE(file_inode(file));
2cbe3b0a	310	unsigned long rv = -EIO;
ae5758a1	311
c4fe2448	312	if (use_pde(pde)) {
ae5758a1 JB	313	typeof(proc_reg_get_unmapped_area) *get_area;
	314
	315	get_area = pde->proc_fops->get_unmapped_area;
fad1a86e	316	#ifdef CONFIG_MMU
ae5758a1 JB	317	if (!get_area)
ae5758a1 JB	318	get_area = current->mm->get_unmapped_area;
fad1a86e	319	#endif
ae5758a1	320
5721cf84 HD	321	if (get_area)
5721cf84 HD	322	rv = get_area(file, orig_addr, len, pgoff, flags);
ae5758a1 JB	323	else
ae5758a1 JB	324	rv = orig_addr;
c4fe2448 AD	325	unuse_pde(pde);
	326	}
	327	return rv;
	328	}
	329
786d7e16 AD	330	static int proc_reg_open(struct inode inode, struct file file)
	331	{
	332	struct proc_dir_entry *pde = PDE(inode);
	333	int rv = 0;
	334	int (open)(struct inode , struct file *);
881adb85 AD	335	int (release)(struct inode , struct file *);
	336	struct pde_opener *pdeo;
	337
	338	/*
492b2da6 AD	339	* Ensure that
	340	* 1) PDE's ->release hook will be called no matter what
	341	* either normally by close()/->release, or forcefully by
	342	* rmmod/remove_proc_entry.
	343	*
	344	* 2) rmmod isn't blocked by opening file in /proc and sitting on
	345	* the descriptor (including "rmmod foo </proc/foo" scenario).
881adb85	346	*
492b2da6	347	* Save every "struct file" with custom ->release hook.
881adb85	348	*/
e7a6e291	349	if (!use_pde(pde))
d2857e79	350	return -ENOENT;
e7a6e291	351
881adb85	352	release = pde->proc_fops->release;
e7a6e291	353	if (release) {
195b8cf0	354	pdeo = kmem_cache_alloc(pde_opener_cache, GFP_KERNEL);
e7a6e291 AD	355	if (!pdeo) {
	356	rv = -ENOMEM;
	357	goto out_unuse;
	358	}
	359	}
786d7e16	360
e7a6e291	361	open = pde->proc_fops->open;
786d7e16 AD	362	if (open)
	363	rv = open(inode, file);
	364
e7a6e291 AD	365	if (release) {
	366	if (rv == 0) {
	367	/* To know what to release. */
	368	pdeo->file = file;
	369	pdeo->closing = false;
	370	pdeo->c = NULL;
	371	spin_lock(&pde->pde_unload_lock);
	372	list_add(&pdeo->lh, &pde->pde_openers);
	373	spin_unlock(&pde->pde_unload_lock);
	374	} else
195b8cf0	375	kmem_cache_free(pde_opener_cache, pdeo);
e7a6e291	376	}
05c0ae21	377
e7a6e291	378	out_unuse:
05c0ae21	379	unuse_pde(pde);
786d7e16 AD	380	return rv;
	381	}
	382
	383	static int proc_reg_release(struct inode inode, struct file file)
	384	{
	385	struct proc_dir_entry *pde = PDE(inode);
881adb85	386	struct pde_opener *pdeo;
786d7e16	387	spin_lock(&pde->pde_unload_lock);
ca469f35 AV	388	list_for_each_entry(pdeo, &pde->pde_openers, lh) {
	389	if (pdeo->file == file) {
	390	close_pdeo(pde, pdeo);
2f897424	391	return 0;
ca469f35	392	}
881adb85	393	}
786d7e16	394	spin_unlock(&pde->pde_unload_lock);
ca469f35	395	return 0;
786d7e16 AD	396	}
	397
	398	static const struct file_operations proc_reg_file_ops = {
	399	.llseek = proc_reg_llseek,
	400	.read = proc_reg_read,
	401	.write = proc_reg_write,
	402	.poll = proc_reg_poll,
	403	.unlocked_ioctl = proc_reg_unlocked_ioctl,
	404	#ifdef CONFIG_COMPAT
	405	.compat_ioctl = proc_reg_compat_ioctl,
	406	#endif
	407	.mmap = proc_reg_mmap,
c4fe2448	408	.get_unmapped_area = proc_reg_get_unmapped_area,
786d7e16 AD	409	.open = proc_reg_open,
	410	.release = proc_reg_release,
	411	};
	412
778f3dd5 DM	413	#ifdef CONFIG_COMPAT
	414	static const struct file_operations proc_reg_file_ops_no_compat = {
	415	.llseek = proc_reg_llseek,
	416	.read = proc_reg_read,
	417	.write = proc_reg_write,
	418	.poll = proc_reg_poll,
	419	.unlocked_ioctl = proc_reg_unlocked_ioctl,
	420	.mmap = proc_reg_mmap,
c4fe2448	421	.get_unmapped_area = proc_reg_get_unmapped_area,
778f3dd5 DM	422	.open = proc_reg_open,
	423	.release = proc_reg_release,
	424	};
	425	#endif
	426
fceef393 AV	427	static void proc_put_link(void *p)
	428	{
	429	unuse_pde(p);
	430	}
	431
6b255391	432	static const char proc_get_link(struct dentry dentry,
fceef393 AV	433	struct inode *inode,
fceef393 AV	434	struct delayed_call *done)
7e0e953b	435	{
6b255391	436	struct proc_dir_entry *pde = PDE(inode);
15b158b4	437	if (!use_pde(pde))
7e0e953b	438	return ERR_PTR(-EINVAL);
fceef393	439	set_delayed_call(done, proc_put_link, pde);
680baacb	440	return pde->data;
7e0e953b AV	441	}
7e0e953b AV	442
7e0e953b	443	const struct inode_operations proc_link_inode_operations = {
6b255391	444	.get_link = proc_get_link,
7e0e953b AV	445	};
7e0e953b AV	446
6d1b6e4e	447	struct inode proc_get_inode(struct super_block sb, struct proc_dir_entry *de)
1da177e4	448	{
51f0885e	449	struct inode *inode = new_inode_pseudo(sb);
1da177e4	450
51f0885e LT	451	if (inode) {
51f0885e LT	452	inode->i_ino = de->low_ino;
078cd827	453	inode->i_mtime = inode->i_atime = inode->i_ctime = current_time(inode);
a1d4aebb	454	PROC_I(inode)->pde = de;
5e971dce	455
eb6d38d5 EB	456	if (is_empty_pde(de)) {
	457	make_empty_dir_inode(inode);
	458	return inode;
	459	}
5e971dce AD	460	if (de->mode) {
	461	inode->i_mode = de->mode;
	462	inode->i_uid = de->uid;
	463	inode->i_gid = de->gid;
	464	}
	465	if (de->size)
	466	inode->i_size = de->size;
	467	if (de->nlink)
bfe86848	468	set_nlink(inode, de->nlink);
b6cdc731 AV	469	WARN_ON(!de->proc_iops);
b6cdc731 AV	470	inode->i_op = de->proc_iops;
5e971dce AD	471	if (de->proc_fops) {
5e971dce AD	472	if (S_ISREG(inode->i_mode)) {
778f3dd5	473	#ifdef CONFIG_COMPAT
5e971dce AD	474	if (!de->proc_fops->compat_ioctl)
	475	inode->i_fop =
	476	&proc_reg_file_ops_no_compat;
	477	else
778f3dd5	478	#endif
5e971dce AD	479	inode->i_fop = &proc_reg_file_ops;
	480	} else {
	481	inode->i_fop = de->proc_fops;
778f3dd5	482	}
786d7e16	483	}
99b76233	484	} else
135d5655	485	pde_put(de);
1da177e4	486	return inode;
d3d009cb	487	}
1da177e4	488
e94591d0	489	int proc_fill_super(struct super_block s, void data, int silent)
1da177e4	490	{
e94591d0	491	struct pid_namespace *ns = get_pid_ns(s->s_fs_info);
87e0aab3	492	struct inode *root_inode;
0097875b	493	int ret;
87e0aab3	494
e94591d0 EB	495	if (!proc_parse_options(data, ns))
	496	return -EINVAL;
	497
a2982cc9 EB	498	/* User space would break if executables or devices appear on proc */
a2982cc9 EB	499	s->s_iflags \|= SB_I_USERNS_VISIBLE \| SB_I_NOEXEC \| SB_I_NODEV;
1751e8a6	500	s->s_flags \|= SB_NODIRATIME \| SB_NOSUID \| SB_NOEXEC;
1da177e4 LT	501	s->s_blocksize = 1024;
	502	s->s_blocksize_bits = 10;
	503	s->s_magic = PROC_SUPER_MAGIC;
	504	s->s_op = &proc_sops;
	505	s->s_time_gran = 1;
a867d734 LT	506
	507	/*
	508	* procfs isn't actually a stacking filesystem; however, there is
	509	* too much magic going on inside it to permit stacking things on
	510	* top of it
	511	*/
	512	s->s_stack_depth = FILESYSTEM_MAX_STACK_DEPTH;
1da177e4	513
135d5655	514	pde_get(&proc_root);
87e0aab3 MP	515	root_inode = proc_get_inode(s, &proc_root);
87e0aab3 MP	516	if (!root_inode) {
87ebdc00	517	pr_err("proc_fill_super: get root inode failed\n");
87e0aab3 MP	518	return -ENOMEM;
87e0aab3 MP	519	}
1da177e4	520
87e0aab3 MP	521	s->s_root = d_make_root(root_inode);
87e0aab3 MP	522	if (!s->s_root) {
87ebdc00	523	pr_err("proc_fill_super: allocate dentry failed\n");
87e0aab3 MP	524	return -ENOMEM;
	525	}
	526
0097875b EB	527	ret = proc_setup_self(s);
	528	if (ret) {
	529	return ret;
	530	}
	531	return proc_setup_thread_self(s);
1da177e4	532	}