xref: /kernel/linux/linux-6.6/fs/devpts/inode.c (revision 62306a36) 
1// SPDX-License-Identifier: GPL-2.0-or-later
2/* -*- linux-c -*- --------------------------------------------------------- *
3 *
4 * linux/fs/devpts/inode.c
5 *
6 *  Copyright 1998-2004 H. Peter Anvin -- All Rights Reserved
7 *
8 * ------------------------------------------------------------------------- */
9
10#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
11
12#include <linux/module.h>
13#include <linux/init.h>
14#include <linux/fs.h>
15#include <linux/sched.h>
16#include <linux/namei.h>
17#include <linux/slab.h>
18#include <linux/mount.h>
19#include <linux/tty.h>
20#include <linux/mutex.h>
21#include <linux/magic.h>
22#include <linux/idr.h>
23#include <linux/devpts_fs.h>
24#include <linux/parser.h>
25#include <linux/fsnotify.h>
26#include <linux/seq_file.h>
27
28#define DEVPTS_DEFAULT_MODE 0600
29/*
30 * ptmx is a new node in /dev/pts and will be unused in legacy (single-
31 * instance) mode. To prevent surprises in user space, set permissions of
32 * ptmx to 0. Use 'chmod' or remount with '-o ptmxmode' to set meaningful
33 * permissions.
34 */
35#define DEVPTS_DEFAULT_PTMX_MODE 0000
36#define PTMX_MINOR	2
37
38/*
39 * sysctl support for setting limits on the number of Unix98 ptys allocated.
40 * Otherwise one can eat up all kernel memory by opening /dev/ptmx repeatedly.
41 */
42static int pty_limit = NR_UNIX98_PTY_DEFAULT;
43static int pty_reserve = NR_UNIX98_PTY_RESERVE;
44static int pty_limit_min;
45static int pty_limit_max = INT_MAX;
46static atomic_t pty_count = ATOMIC_INIT(0);
47
48static struct ctl_table pty_table[] = {
49	{
50		.procname	= "max",
51		.maxlen		= sizeof(int),
52		.mode		= 0644,
53		.data		= &pty_limit,
54		.proc_handler	= proc_dointvec_minmax,
55		.extra1		= &pty_limit_min,
56		.extra2		= &pty_limit_max,
57	}, {
58		.procname	= "reserve",
59		.maxlen		= sizeof(int),
60		.mode		= 0644,
61		.data		= &pty_reserve,
62		.proc_handler	= proc_dointvec_minmax,
63		.extra1		= &pty_limit_min,
64		.extra2		= &pty_limit_max,
65	}, {
66		.procname	= "nr",
67		.maxlen		= sizeof(int),
68		.mode		= 0444,
69		.data		= &pty_count,
70		.proc_handler	= proc_dointvec,
71	},
72	{}
73};
74
75struct pts_mount_opts {
76	int setuid;
77	int setgid;
78	kuid_t   uid;
79	kgid_t   gid;
80	umode_t mode;
81	umode_t ptmxmode;
82	int reserve;
83	int max;
84};
85
86enum {
87	Opt_uid, Opt_gid, Opt_mode, Opt_ptmxmode, Opt_newinstance,  Opt_max,
88	Opt_err
89};
90
91static const match_table_t tokens = {
92	{Opt_uid, "uid=%u"},
93	{Opt_gid, "gid=%u"},
94	{Opt_mode, "mode=%o"},
95	{Opt_ptmxmode, "ptmxmode=%o"},
96	{Opt_newinstance, "newinstance"},
97	{Opt_max, "max=%d"},
98	{Opt_err, NULL}
99};
100
101struct pts_fs_info {
102	struct ida allocated_ptys;
103	struct pts_mount_opts mount_opts;
104	struct super_block *sb;
105	struct dentry *ptmx_dentry;
106};
107
108static inline struct pts_fs_info *DEVPTS_SB(struct super_block *sb)
109{
110	return sb->s_fs_info;
111}
112
113static int devpts_ptmx_path(struct path *path)
114{
115	struct super_block *sb;
116	int err;
117
118	/* Is a devpts filesystem at "pts" in the same directory? */
119	err = path_pts(path);
120	if (err)
121		return err;
122
123	/* Is the path the root of a devpts filesystem? */
124	sb = path->mnt->mnt_sb;
125	if ((sb->s_magic != DEVPTS_SUPER_MAGIC) ||
126	    (path->mnt->mnt_root != sb->s_root))
127		return -ENODEV;
128
129	return 0;
130}
131
132/*
133 * Try to find a suitable devpts filesystem. We support the following
134 * scenarios:
135 * - The ptmx device node is located in the same directory as the devpts
136 *   mount where the pts device nodes are located.
137 *   This is e.g. the case when calling open on the /dev/pts/ptmx device
138 *   node when the devpts filesystem is mounted at /dev/pts.
139 * - The ptmx device node is located outside the devpts filesystem mount
140 *   where the pts device nodes are located. For example, the ptmx device
141 *   is a symlink, separate device node, or bind-mount.
142 *   A supported scenario is bind-mounting /dev/pts/ptmx to /dev/ptmx and
143 *   then calling open on /dev/ptmx. In this case a suitable pts
144 *   subdirectory can be found in the common parent directory /dev of the
145 *   devpts mount and the ptmx bind-mount, after resolving the /dev/ptmx
146 *   bind-mount.
147 *   If no suitable pts subdirectory can be found this function will fail.
148 *   This is e.g. the case when bind-mounting /dev/pts/ptmx to /ptmx.
149 */
150struct vfsmount *devpts_mntget(struct file *filp, struct pts_fs_info *fsi)
151{
152	struct path path;
153	int err = 0;
154
155	path = filp->f_path;
156	path_get(&path);
157
158	/* Walk upward while the start point is a bind mount of
159	 * a single file.
160	 */
161	while (path.mnt->mnt_root == path.dentry)
162		if (follow_up(&path) == 0)
163			break;
164
165	/* devpts_ptmx_path() finds a devpts fs or returns an error. */
166	if ((path.mnt->mnt_sb->s_magic != DEVPTS_SUPER_MAGIC) ||
167	    (DEVPTS_SB(path.mnt->mnt_sb) != fsi))
168		err = devpts_ptmx_path(&path);
169	dput(path.dentry);
170	if (!err) {
171		if (DEVPTS_SB(path.mnt->mnt_sb) == fsi)
172			return path.mnt;
173
174		err = -ENODEV;
175	}
176
177	mntput(path.mnt);
178	return ERR_PTR(err);
179}
180
181struct pts_fs_info *devpts_acquire(struct file *filp)
182{
183	struct pts_fs_info *result;
184	struct path path;
185	struct super_block *sb;
186
187	path = filp->f_path;
188	path_get(&path);
189
190	/* Has the devpts filesystem already been found? */
191	if (path.mnt->mnt_sb->s_magic != DEVPTS_SUPER_MAGIC) {
192		int err;
193
194		err = devpts_ptmx_path(&path);
195		if (err) {
196			result = ERR_PTR(err);
197			goto out;
198		}
199	}
200
201	/*
202	 * pty code needs to hold extra references in case of last /dev/tty close
203	 */
204	sb = path.mnt->mnt_sb;
205	atomic_inc(&sb->s_active);
206	result = DEVPTS_SB(sb);
207
208out:
209	path_put(&path);
210	return result;
211}
212
213void devpts_release(struct pts_fs_info *fsi)
214{
215	deactivate_super(fsi->sb);
216}
217
218#define PARSE_MOUNT	0
219#define PARSE_REMOUNT	1
220
221/*
222 * parse_mount_options():
223 *	Set @opts to mount options specified in @data. If an option is not
224 *	specified in @data, set it to its default value.
225 *
226 * Note: @data may be NULL (in which case all options are set to default).
227 */
228static int parse_mount_options(char *data, int op, struct pts_mount_opts *opts)
229{
230	char *p;
231	kuid_t uid;
232	kgid_t gid;
233
234	opts->setuid  = 0;
235	opts->setgid  = 0;
236	opts->uid     = GLOBAL_ROOT_UID;
237	opts->gid     = GLOBAL_ROOT_GID;
238	opts->mode    = DEVPTS_DEFAULT_MODE;
239	opts->ptmxmode = DEVPTS_DEFAULT_PTMX_MODE;
240	opts->max     = NR_UNIX98_PTY_MAX;
241
242	/* Only allow instances mounted from the initial mount
243	 * namespace to tap the reserve pool of ptys.
244	 */
245	if (op == PARSE_MOUNT)
246		opts->reserve =
247			(current->nsproxy->mnt_ns == init_task.nsproxy->mnt_ns);
248
249	while ((p = strsep(&data, ",")) != NULL) {
250		substring_t args[MAX_OPT_ARGS];
251		int token;
252		int option;
253
254		if (!*p)
255			continue;
256
257		token = match_token(p, tokens, args);
258		switch (token) {
259		case Opt_uid:
260			if (match_int(&args[0], &option))
261				return -EINVAL;
262			uid = make_kuid(current_user_ns(), option);
263			if (!uid_valid(uid))
264				return -EINVAL;
265			opts->uid = uid;
266			opts->setuid = 1;
267			break;
268		case Opt_gid:
269			if (match_int(&args[0], &option))
270				return -EINVAL;
271			gid = make_kgid(current_user_ns(), option);
272			if (!gid_valid(gid))
273				return -EINVAL;
274			opts->gid = gid;
275			opts->setgid = 1;
276			break;
277		case Opt_mode:
278			if (match_octal(&args[0], &option))
279				return -EINVAL;
280			opts->mode = option & S_IALLUGO;
281			break;
282		case Opt_ptmxmode:
283			if (match_octal(&args[0], &option))
284				return -EINVAL;
285			opts->ptmxmode = option & S_IALLUGO;
286			break;
287		case Opt_newinstance:
288			break;
289		case Opt_max:
290			if (match_int(&args[0], &option) ||
291			    option < 0 || option > NR_UNIX98_PTY_MAX)
292				return -EINVAL;
293			opts->max = option;
294			break;
295		default:
296			pr_err("called with bogus options\n");
297			return -EINVAL;
298		}
299	}
300
301	return 0;
302}
303
304static int mknod_ptmx(struct super_block *sb)
305{
306	int mode;
307	int rc = -ENOMEM;
308	struct dentry *dentry;
309	struct inode *inode;
310	struct dentry *root = sb->s_root;
311	struct pts_fs_info *fsi = DEVPTS_SB(sb);
312	struct pts_mount_opts *opts = &fsi->mount_opts;
313	kuid_t ptmx_uid = current_fsuid();
314	kgid_t ptmx_gid = current_fsgid();
315
316	inode_lock(d_inode(root));
317
318	/* If we have already created ptmx node, return */
319	if (fsi->ptmx_dentry) {
320		rc = 0;
321		goto out;
322	}
323
324	dentry = d_alloc_name(root, "ptmx");
325	if (!dentry) {
326		pr_err("Unable to alloc dentry for ptmx node\n");
327		goto out;
328	}
329
330	/*
331	 * Create a new 'ptmx' node in this mount of devpts.
332	 */
333	inode = new_inode(sb);
334	if (!inode) {
335		pr_err("Unable to alloc inode for ptmx node\n");
336		dput(dentry);
337		goto out;
338	}
339
340	inode->i_ino = 2;
341	inode->i_mtime = inode->i_atime = inode_set_ctime_current(inode);
342
343	mode = S_IFCHR|opts->ptmxmode;
344	init_special_inode(inode, mode, MKDEV(TTYAUX_MAJOR, 2));
345	inode->i_uid = ptmx_uid;
346	inode->i_gid = ptmx_gid;
347
348	d_add(dentry, inode);
349
350	fsi->ptmx_dentry = dentry;
351	rc = 0;
352out:
353	inode_unlock(d_inode(root));
354	return rc;
355}
356
357static void update_ptmx_mode(struct pts_fs_info *fsi)
358{
359	struct inode *inode;
360	if (fsi->ptmx_dentry) {
361		inode = d_inode(fsi->ptmx_dentry);
362		inode->i_mode = S_IFCHR|fsi->mount_opts.ptmxmode;
363	}
364}
365
366static int devpts_remount(struct super_block *sb, int *flags, char *data)
367{
368	int err;
369	struct pts_fs_info *fsi = DEVPTS_SB(sb);
370	struct pts_mount_opts *opts = &fsi->mount_opts;
371
372	err = parse_mount_options(data, PARSE_REMOUNT, opts);
373
374	/*
375	 * parse_mount_options() restores options to default values
376	 * before parsing and may have changed ptmxmode. So, update the
377	 * mode in the inode too. Bogus options don't fail the remount,
378	 * so do this even on error return.
379	 */
380	update_ptmx_mode(fsi);
381
382	return err;
383}
384
385static int devpts_show_options(struct seq_file *seq, struct dentry *root)
386{
387	struct pts_fs_info *fsi = DEVPTS_SB(root->d_sb);
388	struct pts_mount_opts *opts = &fsi->mount_opts;
389
390	if (opts->setuid)
391		seq_printf(seq, ",uid=%u",
392			   from_kuid_munged(&init_user_ns, opts->uid));
393	if (opts->setgid)
394		seq_printf(seq, ",gid=%u",
395			   from_kgid_munged(&init_user_ns, opts->gid));
396	seq_printf(seq, ",mode=%03o", opts->mode);
397	seq_printf(seq, ",ptmxmode=%03o", opts->ptmxmode);
398	if (opts->max < NR_UNIX98_PTY_MAX)
399		seq_printf(seq, ",max=%d", opts->max);
400
401	return 0;
402}
403
404static const struct super_operations devpts_sops = {
405	.statfs		= simple_statfs,
406	.remount_fs	= devpts_remount,
407	.show_options	= devpts_show_options,
408};
409
410static void *new_pts_fs_info(struct super_block *sb)
411{
412	struct pts_fs_info *fsi;
413
414	fsi = kzalloc(sizeof(struct pts_fs_info), GFP_KERNEL);
415	if (!fsi)
416		return NULL;
417
418	ida_init(&fsi->allocated_ptys);
419	fsi->mount_opts.mode = DEVPTS_DEFAULT_MODE;
420	fsi->mount_opts.ptmxmode = DEVPTS_DEFAULT_PTMX_MODE;
421	fsi->sb = sb;
422
423	return fsi;
424}
425
426static int
427devpts_fill_super(struct super_block *s, void *data, int silent)
428{
429	struct inode *inode;
430	int error;
431
432	s->s_iflags &= ~SB_I_NODEV;
433	s->s_blocksize = 1024;
434	s->s_blocksize_bits = 10;
435	s->s_magic = DEVPTS_SUPER_MAGIC;
436	s->s_op = &devpts_sops;
437	s->s_d_op = &simple_dentry_operations;
438	s->s_time_gran = 1;
439
440	error = -ENOMEM;
441	s->s_fs_info = new_pts_fs_info(s);
442	if (!s->s_fs_info)
443		goto fail;
444
445	error = parse_mount_options(data, PARSE_MOUNT, &DEVPTS_SB(s)->mount_opts);
446	if (error)
447		goto fail;
448
449	error = -ENOMEM;
450	inode = new_inode(s);
451	if (!inode)
452		goto fail;
453	inode->i_ino = 1;
454	inode->i_mtime = inode->i_atime = inode_set_ctime_current(inode);
455	inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO | S_IWUSR;
456	inode->i_op = &simple_dir_inode_operations;
457	inode->i_fop = &simple_dir_operations;
458	set_nlink(inode, 2);
459
460	s->s_root = d_make_root(inode);
461	if (!s->s_root) {
462		pr_err("get root dentry failed\n");
463		goto fail;
464	}
465
466	error = mknod_ptmx(s);
467	if (error)
468		goto fail_dput;
469
470	return 0;
471fail_dput:
472	dput(s->s_root);
473	s->s_root = NULL;
474fail:
475	return error;
476}
477
478/*
479 * devpts_mount()
480 *
481 *     Mount a new (private) instance of devpts.  PTYs created in this
482 *     instance are independent of the PTYs in other devpts instances.
483 */
484static struct dentry *devpts_mount(struct file_system_type *fs_type,
485	int flags, const char *dev_name, void *data)
486{
487	return mount_nodev(fs_type, flags, data, devpts_fill_super);
488}
489
490static void devpts_kill_sb(struct super_block *sb)
491{
492	struct pts_fs_info *fsi = DEVPTS_SB(sb);
493
494	if (fsi)
495		ida_destroy(&fsi->allocated_ptys);
496	kfree(fsi);
497	kill_litter_super(sb);
498}
499
500static struct file_system_type devpts_fs_type = {
501	.name		= "devpts",
502	.mount		= devpts_mount,
503	.kill_sb	= devpts_kill_sb,
504	.fs_flags	= FS_USERNS_MOUNT,
505};
506
507/*
508 * The normal naming convention is simply /dev/pts/<number>; this conforms
509 * to the System V naming convention
510 */
511
512int devpts_new_index(struct pts_fs_info *fsi)
513{
514	int index = -ENOSPC;
515
516	if (atomic_inc_return(&pty_count) >= (pty_limit -
517			  (fsi->mount_opts.reserve ? 0 : pty_reserve)))
518		goto out;
519
520	index = ida_alloc_max(&fsi->allocated_ptys, fsi->mount_opts.max - 1,
521			GFP_KERNEL);
522
523out:
524	if (index < 0)
525		atomic_dec(&pty_count);
526	return index;
527}
528
529void devpts_kill_index(struct pts_fs_info *fsi, int idx)
530{
531	ida_free(&fsi->allocated_ptys, idx);
532	atomic_dec(&pty_count);
533}
534
535/**
536 * devpts_pty_new -- create a new inode in /dev/pts/
537 * @fsi: Filesystem info for this instance.
538 * @index: used as a name of the node
539 * @priv: what's given back by devpts_get_priv
540 *
541 * The dentry for the created inode is returned.
542 * Remove it from /dev/pts/ with devpts_pty_kill().
543 */
544struct dentry *devpts_pty_new(struct pts_fs_info *fsi, int index, void *priv)
545{
546	struct dentry *dentry;
547	struct super_block *sb = fsi->sb;
548	struct inode *inode;
549	struct dentry *root;
550	struct pts_mount_opts *opts;
551	char s[12];
552
553	root = sb->s_root;
554	opts = &fsi->mount_opts;
555
556	inode = new_inode(sb);
557	if (!inode)
558		return ERR_PTR(-ENOMEM);
559
560	inode->i_ino = index + 3;
561	inode->i_uid = opts->setuid ? opts->uid : current_fsuid();
562	inode->i_gid = opts->setgid ? opts->gid : current_fsgid();
563	inode->i_mtime = inode->i_atime = inode_set_ctime_current(inode);
564	init_special_inode(inode, S_IFCHR|opts->mode, MKDEV(UNIX98_PTY_SLAVE_MAJOR, index));
565
566	sprintf(s, "%d", index);
567
568	dentry = d_alloc_name(root, s);
569	if (dentry) {
570		dentry->d_fsdata = priv;
571		d_add(dentry, inode);
572		fsnotify_create(d_inode(root), dentry);
573	} else {
574		iput(inode);
575		dentry = ERR_PTR(-ENOMEM);
576	}
577
578	return dentry;
579}
580
581/**
582 * devpts_get_priv -- get private data for a slave
583 * @dentry: dentry of the slave
584 *
585 * Returns whatever was passed as priv in devpts_pty_new for a given inode.
586 */
587void *devpts_get_priv(struct dentry *dentry)
588{
589	if (dentry->d_sb->s_magic != DEVPTS_SUPER_MAGIC)
590		return NULL;
591	return dentry->d_fsdata;
592}
593
594/**
595 * devpts_pty_kill -- remove inode form /dev/pts/
596 * @dentry: dentry of the slave to be removed
597 *
598 * This is an inverse operation of devpts_pty_new.
599 */
600void devpts_pty_kill(struct dentry *dentry)
601{
602	WARN_ON_ONCE(dentry->d_sb->s_magic != DEVPTS_SUPER_MAGIC);
603
604	dentry->d_fsdata = NULL;
605	drop_nlink(dentry->d_inode);
606	d_drop(dentry);
607	fsnotify_unlink(d_inode(dentry->d_parent), dentry);
608	dput(dentry);	/* d_alloc_name() in devpts_pty_new() */
609}
610
611static int __init init_devpts_fs(void)
612{
613	int err = register_filesystem(&devpts_fs_type);
614	if (!err) {
615		register_sysctl("kernel/pty", pty_table);
616	}
617	return err;
618}
619module_init(init_devpts_fs)
620