Commit | Line | Data |
---|---|---|
2874c5fd | 1 | /* SPDX-License-Identifier: GPL-2.0-or-later */ |
59d8053f | 2 | /* Internal procfs definitions |
1da177e4 LT |
3 | * |
4 | * Copyright (C) 2004 Red Hat, Inc. All Rights Reserved. | |
5 | * Written by David Howells (dhowells@redhat.com) | |
1da177e4 LT |
6 | */ |
7 | ||
8 | #include <linux/proc_fs.h> | |
59d8053f | 9 | #include <linux/proc_ns.h> |
9cdd83e3 | 10 | #include <linux/refcount.h> |
59d8053f DH |
11 | #include <linux/spinlock.h> |
12 | #include <linux/atomic.h> | |
e579d2c2 | 13 | #include <linux/binfmts.h> |
f7ccbae4 | 14 | #include <linux/sched/coredump.h> |
f719ff9b | 15 | #include <linux/sched/task.h> |
cdd9a571 | 16 | #include <linux/mm.h> |
1da177e4 | 17 | |
59d8053f DH |
18 | struct ctl_table_header; |
19 | struct mempolicy; | |
77b14db5 | 20 | |
59d8053f DH |
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 | * | |
710585d4 ND |
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. | |
59d8053f DH |
30 | */ |
31 | struct proc_dir_entry { | |
163cf548 AD |
32 | /* |
33 | * number of callers into module in progress; | |
34 | * negative -> it's going away RSN | |
35 | */ | |
36 | atomic_t in_use; | |
9cdd83e3 | 37 | refcount_t refcnt; |
163cf548 | 38 | struct list_head pde_openers; /* who did ->open, but not ->release */ |
53f63345 AD |
39 | /* protects ->pde_openers and all struct pde_opener instances */ |
40 | spinlock_t pde_unload_lock; | |
163cf548 AD |
41 | struct completion *pde_unload_completion; |
42 | const struct inode_operations *proc_iops; | |
d56c0d45 AD |
43 | union { |
44 | const struct proc_ops *proc_ops; | |
45 | const struct file_operations *proc_dir_ops; | |
46 | }; | |
1fde6f21 | 47 | const struct dentry_operations *proc_dops; |
3f3942ac CH |
48 | union { |
49 | const struct seq_operations *seq_ops; | |
50 | int (*single_show)(struct seq_file *, void *); | |
51 | }; | |
564def71 | 52 | proc_write_t write; |
163cf548 | 53 | void *data; |
44414d82 | 54 | unsigned int state_size; |
59d8053f | 55 | unsigned int low_ino; |
59d8053f DH |
56 | nlink_t nlink; |
57 | kuid_t uid; | |
58 | kgid_t gid; | |
59 | loff_t size; | |
710585d4 | 60 | struct proc_dir_entry *parent; |
4f113437 | 61 | struct rb_root subdir; |
710585d4 | 62 | struct rb_node subdir_node; |
b4884f23 | 63 | char *name; |
163cf548 | 64 | umode_t mode; |
d919b33d | 65 | u8 flags; |
59d8053f | 66 | u8 namelen; |
24074a35 | 67 | char inline_name[]; |
3859a271 | 68 | } __randomize_layout; |
1da177e4 | 69 | |
2d6e4e82 AD |
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)) | |
24074a35 | 77 | |
d919b33d AD |
78 | static inline bool pde_is_permanent(const struct proc_dir_entry *pde) |
79 | { | |
80 | return pde->flags & PROC_ENTRY_PERMANENT; | |
81 | } | |
82 | ||
ef1d6178 AD |
83 | static inline void pde_make_permanent(struct proc_dir_entry *pde) |
84 | { | |
85 | pde->flags |= PROC_ENTRY_PERMANENT; | |
86 | } | |
87 | ||
654b33ad YB |
88 | static inline bool pde_has_proc_read_iter(const struct proc_dir_entry *pde) |
89 | { | |
90 | return pde->flags & PROC_ENTRY_proc_read_iter; | |
91 | } | |
92 | ||
93 | static inline bool pde_has_proc_compat_ioctl(const struct proc_dir_entry *pde) | |
94 | { | |
95 | #ifdef CONFIG_COMPAT | |
96 | return pde->flags & PROC_ENTRY_proc_compat_ioctl; | |
97 | #else | |
98 | return false; | |
99 | #endif | |
100 | } | |
101 | ||
b4884f23 AD |
102 | extern struct kmem_cache *proc_dir_entry_cache; |
103 | void pde_free(struct proc_dir_entry *pde); | |
104 | ||
59d8053f DH |
105 | union proc_op { |
106 | int (*proc_get_link)(struct dentry *, struct path *); | |
59d8053f DH |
107 | int (*proc_show)(struct seq_file *m, |
108 | struct pid_namespace *ns, struct pid *pid, | |
109 | struct task_struct *task); | |
267c068e | 110 | int lsmid; |
59d8053f | 111 | }; |
99f89551 | 112 | |
59d8053f | 113 | struct proc_inode { |
f2beb798 | 114 | struct pid *pid; |
771187d6 | 115 | unsigned int fd; |
59d8053f DH |
116 | union proc_op op; |
117 | struct proc_dir_entry *pde; | |
118 | struct ctl_table_header *sysctl; | |
29e1095b | 119 | const struct ctl_table *sysctl_entry; |
0afa5ca8 | 120 | struct hlist_node sibling_inodes; |
3d3d35b1 | 121 | const struct proc_ns_operations *ns_ops; |
59d8053f | 122 | struct inode vfs_inode; |
3859a271 | 123 | } __randomize_layout; |
f2beb798 | 124 | |
c30480b9 DH |
125 | /* |
126 | * General functions | |
127 | */ | |
128 | static inline struct proc_inode *PROC_I(const struct inode *inode) | |
129 | { | |
130 | return container_of(inode, struct proc_inode, vfs_inode); | |
131 | } | |
132 | ||
133 | static inline struct proc_dir_entry *PDE(const struct inode *inode) | |
134 | { | |
135 | return PROC_I(inode)->pde; | |
136 | } | |
137 | ||
891ae71d | 138 | static inline struct pid *proc_pid(const struct inode *inode) |
99f89551 | 139 | { |
13b41b09 | 140 | return PROC_I(inode)->pid; |
99f89551 EB |
141 | } |
142 | ||
891ae71d | 143 | static inline struct task_struct *get_proc_task(const struct inode *inode) |
1da177e4 | 144 | { |
13b41b09 | 145 | return get_pid_task(proc_pid(inode), PIDTYPE_PID); |
1da177e4 LT |
146 | } |
147 | ||
c6eb50d2 | 148 | void task_dump_owner(struct task_struct *task, umode_t mode, |
68eb94f1 | 149 | kuid_t *ruid, kgid_t *rgid); |
faf60af1 | 150 | |
3ee2a199 | 151 | unsigned name_to_int(const struct qstr *qstr); |
1dd704b6 | 152 | /* |
59d8053f | 153 | * Offset of the first process in the /proc root directory.. |
1dd704b6 | 154 | */ |
59d8053f | 155 | #define FIRST_PROCESS_ENTRY 256 |
1dd704b6 | 156 | |
59d8053f DH |
157 | /* Worst case buffer size needed for holding an integer. */ |
158 | #define PROC_NUMBUF 13 | |
881adb85 | 159 | |
6dd55dd1 | 160 | #ifdef CONFIG_PAGE_MAPCOUNT |
cdd9a571 DH |
161 | /** |
162 | * folio_precise_page_mapcount() - Number of mappings of this folio page. | |
163 | * @folio: The folio. | |
164 | * @page: The page. | |
165 | * | |
166 | * The number of present user page table entries that reference this page | |
167 | * as tracked via the RMAP: either referenced directly (PTE) or as part of | |
168 | * a larger area that covers this page (e.g., PMD). | |
169 | * | |
170 | * Use this function only for the calculation of existing statistics | |
171 | * (USS, PSS, mapcount_max) and for debugging purposes (/proc/kpagecount). | |
172 | * | |
173 | * Do not add new users. | |
174 | * | |
175 | * Returns: The number of mappings of this folio page. 0 for | |
176 | * folios that are not mapped to user space or are not tracked via the RMAP | |
177 | * (e.g., shared zeropage). | |
178 | */ | |
179 | static inline int folio_precise_page_mapcount(struct folio *folio, | |
180 | struct page *page) | |
181 | { | |
182 | int mapcount = atomic_read(&page->_mapcount) + 1; | |
183 | ||
e880034c | 184 | if (page_mapcount_is_type(mapcount)) |
cdd9a571 DH |
185 | mapcount = 0; |
186 | if (folio_test_large(folio)) | |
187 | mapcount += folio_entire_mapcount(folio); | |
188 | ||
189 | return mapcount; | |
190 | } | |
6dd55dd1 DH |
191 | #else /* !CONFIG_PAGE_MAPCOUNT */ |
192 | static inline int folio_precise_page_mapcount(struct folio *folio, | |
193 | struct page *page) | |
194 | { | |
195 | BUILD_BUG(); | |
196 | } | |
197 | #endif /* CONFIG_PAGE_MAPCOUNT */ | |
cdd9a571 | 198 | |
ae4192b7 DH |
199 | /** |
200 | * folio_average_page_mapcount() - Average number of mappings per page in this | |
201 | * folio | |
202 | * @folio: The folio. | |
203 | * | |
204 | * The average number of user page table entries that reference each page in | |
205 | * this folio as tracked via the RMAP: either referenced directly (PTE) or | |
206 | * as part of a larger area that covers this page (e.g., PMD). | |
207 | * | |
208 | * The average is calculated by rounding to the nearest integer; however, | |
209 | * to avoid duplicated code in current callers, the average is at least | |
210 | * 1 if any page of the folio is mapped. | |
211 | * | |
212 | * Returns: The average number of mappings per page in this folio. | |
213 | */ | |
214 | static inline int folio_average_page_mapcount(struct folio *folio) | |
215 | { | |
216 | int mapcount, entire_mapcount, avg; | |
217 | ||
218 | if (!folio_test_large(folio)) | |
219 | return atomic_read(&folio->_mapcount) + 1; | |
220 | ||
221 | mapcount = folio_large_mapcount(folio); | |
222 | if (unlikely(mapcount <= 0)) | |
223 | return 0; | |
224 | entire_mapcount = folio_entire_mapcount(folio); | |
225 | if (mapcount <= entire_mapcount) | |
226 | return entire_mapcount; | |
227 | mapcount -= entire_mapcount; | |
228 | ||
229 | /* Round to closest integer ... */ | |
230 | avg = ((unsigned int)mapcount + folio_large_nr_pages(folio) / 2) >> folio_large_order(folio); | |
231 | /* ... but return at least 1. */ | |
232 | return max_t(int, avg + entire_mapcount, 1); | |
233 | } | |
59d8053f DH |
234 | /* |
235 | * array.c | |
236 | */ | |
237 | extern const struct file_operations proc_tid_children_operations; | |
3174c21b | 238 | |
88b72b31 TH |
239 | extern void proc_task_name(struct seq_file *m, struct task_struct *p, |
240 | bool escape); | |
59d8053f DH |
241 | extern int proc_tid_stat(struct seq_file *, struct pid_namespace *, |
242 | struct pid *, struct task_struct *); | |
243 | extern int proc_tgid_stat(struct seq_file *, struct pid_namespace *, | |
244 | struct pid *, struct task_struct *); | |
245 | extern int proc_pid_status(struct seq_file *, struct pid_namespace *, | |
246 | struct pid *, struct task_struct *); | |
247 | extern int proc_pid_statm(struct seq_file *, struct pid_namespace *, | |
248 | struct pid *, struct task_struct *); | |
249 | ||
250 | /* | |
251 | * base.c | |
252 | */ | |
253 | extern const struct dentry_operations pid_dentry_operations; | |
b74d24f7 | 254 | extern int pid_getattr(struct mnt_idmap *, const struct path *, |
549c7297 | 255 | struct kstat *, u32, unsigned int); |
c1632a0f | 256 | extern int proc_setattr(struct mnt_idmap *, struct dentry *, |
549c7297 | 257 | struct iattr *); |
7bc3e6e5 | 258 | extern void proc_pid_evict_inode(struct proc_inode *); |
db978da8 | 259 | extern struct inode *proc_pid_make_inode(struct super_block *, struct task_struct *, umode_t); |
1bbc5513 | 260 | extern void pid_update_inode(struct task_struct *, struct inode *); |
59d8053f | 261 | extern int pid_delete_dentry(const struct dentry *); |
f0c3b509 | 262 | extern int proc_pid_readdir(struct file *, struct dir_context *); |
867aaccf | 263 | struct dentry *proc_pid_lookup(struct dentry *, unsigned int); |
59d8053f | 264 | extern loff_t mem_lseek(struct file *, loff_t, int); |
866ad9a7 | 265 | |
59d8053f | 266 | /* Lookups */ |
0168b9e3 | 267 | typedef struct dentry *instantiate_t(struct dentry *, |
59d8053f | 268 | struct task_struct *, const void *); |
a4ef3895 | 269 | bool proc_fill_cache(struct file *, struct dir_context *, const char *, unsigned int, |
59d8053f | 270 | instantiate_t, struct task_struct *, const void *); |
3174c21b | 271 | |
59d8053f DH |
272 | /* |
273 | * generic.c | |
274 | */ | |
7aed53d1 CH |
275 | struct proc_dir_entry *proc_create_reg(const char *name, umode_t mode, |
276 | struct proc_dir_entry **parent, void *data); | |
61172eae CH |
277 | struct proc_dir_entry *proc_register(struct proc_dir_entry *dir, |
278 | struct proc_dir_entry *dp); | |
59d8053f | 279 | extern struct dentry *proc_lookup(struct inode *, struct dentry *, unsigned int); |
93ad5bc6 | 280 | struct dentry *proc_lookup_de(struct inode *, struct dentry *, struct proc_dir_entry *); |
f0c3b509 | 281 | extern int proc_readdir(struct file *, struct dir_context *); |
93ad5bc6 | 282 | int proc_readdir_de(struct file *, struct dir_context *, struct proc_dir_entry *); |
3174c21b | 283 | |
a9389683 | 284 | static inline void pde_get(struct proc_dir_entry *pde) |
135d5655 | 285 | { |
9cdd83e3 | 286 | refcount_inc(&pde->refcnt); |
135d5655 | 287 | } |
59d8053f | 288 | extern void pde_put(struct proc_dir_entry *); |
3174c21b | 289 | |
eb6d38d5 EB |
290 | static inline bool is_empty_pde(const struct proc_dir_entry *pde) |
291 | { | |
292 | return S_ISDIR(pde->mode) && !pde->proc_iops; | |
293 | } | |
564def71 | 294 | extern ssize_t proc_simple_write(struct file *, const char __user *, size_t, loff_t *); |
eb6d38d5 | 295 | |
3174c21b | 296 | /* |
59d8053f | 297 | * inode.c |
3174c21b | 298 | */ |
59d8053f | 299 | struct pde_opener { |
59d8053f | 300 | struct list_head lh; |
70a731c0 | 301 | struct file *file; |
f5887c71 | 302 | bool closing; |
59d8053f | 303 | struct completion *c; |
a9fabc3d | 304 | } __randomize_layout; |
7e0e953b | 305 | extern const struct inode_operations proc_link_inode_operations; |
59d8053f | 306 | extern const struct inode_operations proc_pid_link_inode_operations; |
60a3c3a5 | 307 | extern const struct super_operations proc_sops; |
6b4e306a | 308 | |
195b8cf0 | 309 | void proc_init_kmemcache(void); |
f90f3caf | 310 | void proc_invalidate_siblings_dcache(struct hlist_head *inodes, spinlock_t *lock); |
1270dd8d | 311 | void set_proc_pid_nlink(void); |
59d8053f | 312 | extern struct inode *proc_get_inode(struct super_block *, struct proc_dir_entry *); |
59d8053f | 313 | extern void proc_entry_rundown(struct proc_dir_entry *); |
6b4e306a | 314 | |
59d8053f DH |
315 | /* |
316 | * proc_namespaces.c | |
317 | */ | |
6b4e306a EB |
318 | extern const struct inode_operations proc_ns_dir_inode_operations; |
319 | extern const struct file_operations proc_ns_dir_operations; | |
320 | ||
59d8053f DH |
321 | /* |
322 | * proc_net.c | |
323 | */ | |
324 | extern const struct file_operations proc_net_operations; | |
325 | extern const struct inode_operations proc_net_inode_operations; | |
326 | ||
327 | #ifdef CONFIG_NET | |
328 | extern int proc_net_init(void); | |
329 | #else | |
330 | static inline int proc_net_init(void) { return 0; } | |
331 | #endif | |
332 | ||
333 | /* | |
334 | * proc_self.c | |
335 | */ | |
021ada7d | 336 | extern int proc_setup_self(struct super_block *); |
34db8aaf | 337 | |
0097875b EB |
338 | /* |
339 | * proc_thread_self.c | |
340 | */ | |
341 | extern int proc_setup_thread_self(struct super_block *); | |
342 | extern void proc_thread_self_init(void); | |
343 | ||
34db8aaf | 344 | /* |
59d8053f | 345 | * proc_sysctl.c |
34db8aaf | 346 | */ |
59d8053f DH |
347 | #ifdef CONFIG_PROC_SYSCTL |
348 | extern int proc_sys_init(void); | |
d6cffbbe KK |
349 | extern void proc_sys_evict_inode(struct inode *inode, |
350 | struct ctl_table_header *head); | |
59d8053f DH |
351 | #else |
352 | static inline void proc_sys_init(void) { } | |
d6cffbbe KK |
353 | static inline void proc_sys_evict_inode(struct inode *inode, |
354 | struct ctl_table_header *head) { } | |
59d8053f | 355 | #endif |
34db8aaf DH |
356 | |
357 | /* | |
358 | * proc_tty.c | |
359 | */ | |
360 | #ifdef CONFIG_TTY | |
361 | extern void proc_tty_init(void); | |
362 | #else | |
363 | static inline void proc_tty_init(void) {} | |
364 | #endif | |
59d8053f DH |
365 | |
366 | /* | |
367 | * root.c | |
368 | */ | |
369 | extern struct proc_dir_entry proc_root; | |
370 | ||
371 | extern void proc_self_init(void); | |
59d8053f DH |
372 | |
373 | /* | |
374 | * task_[no]mmu.c | |
375 | */ | |
493b0e9d | 376 | struct mem_size_stats; |
59d8053f | 377 | struct proc_maps_private { |
2c03376d | 378 | struct inode *inode; |
59d8053f | 379 | struct task_struct *task; |
29a40ace | 380 | struct mm_struct *mm; |
c4c84f06 | 381 | struct vma_iterator iter; |
59d8053f DH |
382 | #ifdef CONFIG_NUMA |
383 | struct mempolicy *task_mempolicy; | |
384 | #endif | |
3859a271 | 385 | } __randomize_layout; |
59d8053f | 386 | |
5381e169 ON |
387 | struct mm_struct *proc_mem_open(struct inode *inode, unsigned int mode); |
388 | ||
59d8053f | 389 | extern const struct file_operations proc_pid_maps_operations; |
59d8053f | 390 | extern const struct file_operations proc_pid_numa_maps_operations; |
59d8053f | 391 | extern const struct file_operations proc_pid_smaps_operations; |
493b0e9d | 392 | extern const struct file_operations proc_pid_smaps_rollup_operations; |
59d8053f DH |
393 | extern const struct file_operations proc_clear_refs_operations; |
394 | extern const struct file_operations proc_pagemap_operations; | |
395 | ||
396 | extern unsigned long task_vsize(struct mm_struct *); | |
397 | extern unsigned long task_statm(struct mm_struct *, | |
398 | unsigned long *, unsigned long *, | |
399 | unsigned long *, unsigned long *); | |
400 | extern void task_mem(struct seq_file *, struct mm_struct *); | |
c6c75ded AD |
401 | |
402 | extern const struct dentry_operations proc_net_dentry_ops; | |
403 | static inline void pde_force_lookup(struct proc_dir_entry *pde) | |
404 | { | |
405 | /* /proc/net/ entries can be changed under us by setns(CLONE_NEWNET) */ | |
406 | pde->proc_dops = &proc_net_dentry_ops; | |
407 | } | |
32a0a965 CB |
408 | |
409 | /* | |
410 | * Add a new procfs dentry that can't serve as a mountpoint. That should | |
411 | * encompass anything that is ephemeral and can just disappear while the | |
412 | * process is still around. | |
413 | */ | |
414 | static inline struct dentry *proc_splice_unmountable(struct inode *inode, | |
415 | struct dentry *dentry, const struct dentry_operations *d_ops) | |
416 | { | |
417 | d_set_d_op(dentry, d_ops); | |
418 | dont_mount(dentry); | |
419 | return d_splice_alias(inode, dentry); | |
420 | } |