1 /* SPDX-License-Identifier: GPL-2.0-or-later */
2 /* Internal procfs definitions
3 *
4 * Copyright (C) 2004 Red Hat, Inc. All Rights Reserved.
5 * Written by David Howells (dhowells@redhat.com)
6 */
7
8 #include <linux/proc_fs.h>
9 #include <linux/proc_ns.h>
10 #include <linux/refcount.h>
11 #include <linux/spinlock.h>
12 #include <linux/atomic.h>
13 #include <linux/binfmts.h>
14 #include <linux/sched/coredump.h>
15 #include <linux/sched/task.h>
16 #include <linux/mm.h>
17
18 struct ctl_table_header;
19 struct mempolicy;
20
21 /*
22 * This is not completely implemented yet. The idea is to
23 * create an in-memory tree (like the actual /proc filesystem
24 * tree) of these proc_dir_entries, so that we can dynamically
25 * add new files to /proc.
26 *
27 * parent/subdir are used for the directory structure (every /proc file has a
28 * parent, but "subdir" is empty for all non-directory entries).
29 * subdir_node is used to build the rb tree "subdir" of the parent.
30 */
31 struct proc_dir_entry {
32 /*
33 * number of callers into module in progress;
34 * negative -> it's going away RSN
35 */
36 atomic_t in_use;
37 refcount_t refcnt;
38 struct list_head pde_openers; /* who did ->open, but not ->release */
39 /* protects ->pde_openers and all struct pde_opener instances */
40 spinlock_t pde_unload_lock;
41 struct completion *pde_unload_completion;
42 const struct inode_operations *proc_iops;
43 union {
44 const struct proc_ops *proc_ops;
45 const struct file_operations *proc_dir_ops;
46 };
47 const struct dentry_operations *proc_dops;
48 union {
49 const struct seq_operations *seq_ops;
50 int (*single_show)(struct seq_file *, void *);
51 };
52 proc_write_t write;
53 void *data;
54 unsigned int state_size;
55 unsigned int low_ino;
56 nlink_t nlink;
57 kuid_t uid;
58 kgid_t gid;
59 loff_t size;
60 struct proc_dir_entry *parent;
61 struct rb_root subdir;
62 struct rb_node subdir_node;
63 char *name;
64 umode_t mode;
65 u8 flags;
66 u8 namelen;
67 char inline_name[];
68 } __randomize_layout;
69
70 #define SIZEOF_PDE ( \
71 sizeof(struct proc_dir_entry) < 128 ? 128 : \
72 sizeof(struct proc_dir_entry) < 192 ? 192 : \
73 sizeof(struct proc_dir_entry) < 256 ? 256 : \
74 sizeof(struct proc_dir_entry) < 512 ? 512 : \
75 0)
76 #define SIZEOF_PDE_INLINE_NAME (SIZEOF_PDE - sizeof(struct proc_dir_entry))
77
pde_is_permanent(const struct proc_dir_entry * pde)78 static inline bool pde_is_permanent(const struct proc_dir_entry *pde)
79 {
80 return pde->flags & PROC_ENTRY_PERMANENT;
81 }
82
pde_make_permanent(struct proc_dir_entry * pde)83 static inline void pde_make_permanent(struct proc_dir_entry *pde)
84 {
85 pde->flags |= PROC_ENTRY_PERMANENT;
86 }
87
88 extern struct kmem_cache *proc_dir_entry_cache;
89 void pde_free(struct proc_dir_entry *pde);
90
91 union proc_op {
92 int (*proc_get_link)(struct dentry *, struct path *);
93 int (*proc_show)(struct seq_file *m,
94 struct pid_namespace *ns, struct pid *pid,
95 struct task_struct *task);
96 int lsmid;
97 };
98
99 struct proc_inode {
100 struct pid *pid;
101 unsigned int fd;
102 union proc_op op;
103 struct proc_dir_entry *pde;
104 struct ctl_table_header *sysctl;
105 struct ctl_table *sysctl_entry;
106 struct hlist_node sibling_inodes;
107 const struct proc_ns_operations *ns_ops;
108 struct inode vfs_inode;
109 } __randomize_layout;
110
111 /*
112 * General functions
113 */
PROC_I(const struct inode * inode)114 static inline struct proc_inode *PROC_I(const struct inode *inode)
115 {
116 return container_of(inode, struct proc_inode, vfs_inode);
117 }
118
PDE(const struct inode * inode)119 static inline struct proc_dir_entry *PDE(const struct inode *inode)
120 {
121 return PROC_I(inode)->pde;
122 }
123
proc_pid(const struct inode * inode)124 static inline struct pid *proc_pid(const struct inode *inode)
125 {
126 return PROC_I(inode)->pid;
127 }
128
get_proc_task(const struct inode * inode)129 static inline struct task_struct *get_proc_task(const struct inode *inode)
130 {
131 return get_pid_task(proc_pid(inode), PIDTYPE_PID);
132 }
133
134 void task_dump_owner(struct task_struct *task, umode_t mode,
135 kuid_t *ruid, kgid_t *rgid);
136
137 unsigned name_to_int(const struct qstr *qstr);
138 /*
139 * Offset of the first process in the /proc root directory..
140 */
141 #define FIRST_PROCESS_ENTRY 256
142
143 /* Worst case buffer size needed for holding an integer. */
144 #define PROC_NUMBUF 13
145
146 /**
147 * folio_precise_page_mapcount() - Number of mappings of this folio page.
148 * @folio: The folio.
149 * @page: The page.
150 *
151 * The number of present user page table entries that reference this page
152 * as tracked via the RMAP: either referenced directly (PTE) or as part of
153 * a larger area that covers this page (e.g., PMD).
154 *
155 * Use this function only for the calculation of existing statistics
156 * (USS, PSS, mapcount_max) and for debugging purposes (/proc/kpagecount).
157 *
158 * Do not add new users.
159 *
160 * Returns: The number of mappings of this folio page. 0 for
161 * folios that are not mapped to user space or are not tracked via the RMAP
162 * (e.g., shared zeropage).
163 */
folio_precise_page_mapcount(struct folio * folio,struct page * page)164 static inline int folio_precise_page_mapcount(struct folio *folio,
165 struct page *page)
166 {
167 int mapcount = atomic_read(&page->_mapcount) + 1;
168
169 if (page_mapcount_is_type(mapcount))
170 mapcount = 0;
171 if (folio_test_large(folio))
172 mapcount += folio_entire_mapcount(folio);
173
174 return mapcount;
175 }
176
177 /*
178 * array.c
179 */
180 extern const struct file_operations proc_tid_children_operations;
181
182 extern void proc_task_name(struct seq_file *m, struct task_struct *p,
183 bool escape);
184 extern int proc_tid_stat(struct seq_file *, struct pid_namespace *,
185 struct pid *, struct task_struct *);
186 extern int proc_tgid_stat(struct seq_file *, struct pid_namespace *,
187 struct pid *, struct task_struct *);
188 extern int proc_pid_status(struct seq_file *, struct pid_namespace *,
189 struct pid *, struct task_struct *);
190 extern int proc_pid_statm(struct seq_file *, struct pid_namespace *,
191 struct pid *, struct task_struct *);
192
193 /*
194 * base.c
195 */
196 extern const struct dentry_operations pid_dentry_operations;
197 extern int pid_getattr(struct mnt_idmap *, const struct path *,
198 struct kstat *, u32, unsigned int);
199 extern int proc_setattr(struct mnt_idmap *, struct dentry *,
200 struct iattr *);
201 extern void proc_pid_evict_inode(struct proc_inode *);
202 extern struct inode *proc_pid_make_inode(struct super_block *, struct task_struct *, umode_t);
203 extern void pid_update_inode(struct task_struct *, struct inode *);
204 extern int pid_delete_dentry(const struct dentry *);
205 extern int proc_pid_readdir(struct file *, struct dir_context *);
206 struct dentry *proc_pid_lookup(struct dentry *, unsigned int);
207 extern loff_t mem_lseek(struct file *, loff_t, int);
208
209 /* Lookups */
210 typedef struct dentry *instantiate_t(struct dentry *,
211 struct task_struct *, const void *);
212 bool proc_fill_cache(struct file *, struct dir_context *, const char *, unsigned int,
213 instantiate_t, struct task_struct *, const void *);
214
215 /*
216 * generic.c
217 */
218 struct proc_dir_entry *proc_create_reg(const char *name, umode_t mode,
219 struct proc_dir_entry **parent, void *data);
220 struct proc_dir_entry *proc_register(struct proc_dir_entry *dir,
221 struct proc_dir_entry *dp);
222 extern struct dentry *proc_lookup(struct inode *, struct dentry *, unsigned int);
223 struct dentry *proc_lookup_de(struct inode *, struct dentry *, struct proc_dir_entry *);
224 extern int proc_readdir(struct file *, struct dir_context *);
225 int proc_readdir_de(struct file *, struct dir_context *, struct proc_dir_entry *);
226
pde_get(struct proc_dir_entry * pde)227 static inline void pde_get(struct proc_dir_entry *pde)
228 {
229 refcount_inc(&pde->refcnt);
230 }
231 extern void pde_put(struct proc_dir_entry *);
232
is_empty_pde(const struct proc_dir_entry * pde)233 static inline bool is_empty_pde(const struct proc_dir_entry *pde)
234 {
235 return S_ISDIR(pde->mode) && !pde->proc_iops;
236 }
237 extern ssize_t proc_simple_write(struct file *, const char __user *, size_t, loff_t *);
238
239 /*
240 * inode.c
241 */
242 struct pde_opener {
243 struct list_head lh;
244 struct file *file;
245 bool closing;
246 struct completion *c;
247 } __randomize_layout;
248 extern const struct inode_operations proc_link_inode_operations;
249 extern const struct inode_operations proc_pid_link_inode_operations;
250 extern const struct super_operations proc_sops;
251
252 void proc_init_kmemcache(void);
253 void proc_invalidate_siblings_dcache(struct hlist_head *inodes, spinlock_t *lock);
254 void set_proc_pid_nlink(void);
255 extern struct inode *proc_get_inode(struct super_block *, struct proc_dir_entry *);
256 extern void proc_entry_rundown(struct proc_dir_entry *);
257
258 /*
259 * proc_namespaces.c
260 */
261 extern const struct inode_operations proc_ns_dir_inode_operations;
262 extern const struct file_operations proc_ns_dir_operations;
263
264 /*
265 * proc_net.c
266 */
267 extern const struct file_operations proc_net_operations;
268 extern const struct inode_operations proc_net_inode_operations;
269
270 #ifdef CONFIG_NET
271 extern int proc_net_init(void);
272 #else
proc_net_init(void)273 static inline int proc_net_init(void) { return 0; }
274 #endif
275
276 /*
277 * proc_self.c
278 */
279 extern int proc_setup_self(struct super_block *);
280
281 /*
282 * proc_thread_self.c
283 */
284 extern int proc_setup_thread_self(struct super_block *);
285 extern void proc_thread_self_init(void);
286
287 /*
288 * proc_sysctl.c
289 */
290 #ifdef CONFIG_PROC_SYSCTL
291 extern int proc_sys_init(void);
292 extern void proc_sys_evict_inode(struct inode *inode,
293 struct ctl_table_header *head);
294 #else
proc_sys_init(void)295 static inline void proc_sys_init(void) { }
proc_sys_evict_inode(struct inode * inode,struct ctl_table_header * head)296 static inline void proc_sys_evict_inode(struct inode *inode,
297 struct ctl_table_header *head) { }
298 #endif
299
300 /*
301 * proc_tty.c
302 */
303 #ifdef CONFIG_TTY
304 extern void proc_tty_init(void);
305 #else
proc_tty_init(void)306 static inline void proc_tty_init(void) {}
307 #endif
308
309 /*
310 * root.c
311 */
312 extern struct proc_dir_entry proc_root;
313
314 extern void proc_self_init(void);
315
316 /*
317 * task_[no]mmu.c
318 */
319 struct mem_size_stats;
320 struct proc_maps_private {
321 struct inode *inode;
322 struct task_struct *task;
323 struct mm_struct *mm;
324 struct vma_iterator iter;
325 #ifdef CONFIG_NUMA
326 struct mempolicy *task_mempolicy;
327 #endif
328 } __randomize_layout;
329
330 struct mm_struct *proc_mem_open(struct inode *inode, unsigned int mode);
331
332 extern const struct file_operations proc_pid_maps_operations;
333 extern const struct file_operations proc_pid_numa_maps_operations;
334 extern const struct file_operations proc_pid_smaps_operations;
335 extern const struct file_operations proc_pid_smaps_rollup_operations;
336 extern const struct file_operations proc_clear_refs_operations;
337 extern const struct file_operations proc_pagemap_operations;
338
339 extern unsigned long task_vsize(struct mm_struct *);
340 extern unsigned long task_statm(struct mm_struct *,
341 unsigned long *, unsigned long *,
342 unsigned long *, unsigned long *);
343 extern void task_mem(struct seq_file *, struct mm_struct *);
344
345 extern const struct dentry_operations proc_net_dentry_ops;
pde_force_lookup(struct proc_dir_entry * pde)346 static inline void pde_force_lookup(struct proc_dir_entry *pde)
347 {
348 /* /proc/net/ entries can be changed under us by setns(CLONE_NEWNET) */
349 pde->proc_dops = &proc_net_dentry_ops;
350 }
351
352 /*
353 * Add a new procfs dentry that can't serve as a mountpoint. That should
354 * encompass anything that is ephemeral and can just disappear while the
355 * process is still around.
356 */
proc_splice_unmountable(struct inode * inode,struct dentry * dentry,const struct dentry_operations * d_ops)357 static inline struct dentry *proc_splice_unmountable(struct inode *inode,
358 struct dentry *dentry, const struct dentry_operations *d_ops)
359 {
360 d_set_d_op(dentry, d_ops);
361 dont_mount(dentry);
362 return d_splice_alias(inode, dentry);
363 }
364