Blame view
ipc/shm.c
30.4 KB
1da177e4c
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 |
/* * linux/ipc/shm.c * Copyright (C) 1992, 1993 Krishna Balasubramanian * Many improvements/fixes by Bruno Haible. * Replaced `struct shm_desc' by `struct vm_area_struct', July 1994. * Fixed the shm swap deallocation (shm_unuse()), August 1998 Andrea Arcangeli. * * /proc/sysvipc/shm support (c) 1999 Dragos Acostachioaie <dragos@iname.com> * BIGMEM support, Andrea Arcangeli <andrea@suse.de> * SMP thread shm, Jean-Luc Boyard <jean-luc.boyard@siemens.fr> * HIGHMEM support, Ingo Molnar <mingo@redhat.com> * Make shmmax, shmall, shmmni sysctl'able, Christoph Rohland <cr@sap.com> * Shared /dev/zero support, Kanoj Sarcar <kanoj@sgi.com> * Move the mm functionality over to mm/shmem.c, Christoph Rohland <cr@sap.com> * |
073115d6b
|
16 17 |
* support for audit of ipc object properties and permission changes * Dustin Kirkland <dustin.kirkland@us.ibm.com> |
4e9823111
|
18 19 20 21 |
* * namespaces support * OpenVZ, SWsoft Inc. * Pavel Emelianov <xemul@openvz.org> |
1da177e4c
|
22 |
*/ |
1da177e4c
|
23 24 25 26 27 28 29 |
#include <linux/slab.h> #include <linux/mm.h> #include <linux/hugetlb.h> #include <linux/shm.h> #include <linux/init.h> #include <linux/file.h> #include <linux/mman.h> |
1da177e4c
|
30 31 32 33 |
#include <linux/shmem_fs.h> #include <linux/security.h> #include <linux/syscalls.h> #include <linux/audit.h> |
c59ede7b7
|
34 |
#include <linux/capability.h> |
7d87e14c2
|
35 |
#include <linux/ptrace.h> |
19b4946ca
|
36 |
#include <linux/seq_file.h> |
3e148c799
|
37 |
#include <linux/rwsem.h> |
4e9823111
|
38 |
#include <linux/nsproxy.h> |
bc56bba8f
|
39 |
#include <linux/mount.h> |
ae5e1b22f
|
40 |
#include <linux/ipc_namespace.h> |
7d87e14c2
|
41 |
|
1da177e4c
|
42 43 44 |
#include <asm/uaccess.h> #include "util.h" |
bc56bba8f
|
45 46 47 48 49 50 51 52 |
struct shm_file_data { int id; struct ipc_namespace *ns; struct file *file; const struct vm_operations_struct *vm_ops; }; #define shm_file_data(file) (*((struct shm_file_data **)&(file)->private_data)) |
9a32144e9
|
53 |
static const struct file_operations shm_file_operations; |
f0f37e2f7
|
54 |
static const struct vm_operations_struct shm_vm_ops; |
1da177e4c
|
55 |
|
ed2ddbf88
|
56 |
#define shm_ids(ns) ((ns)->ids[IPC_SHM_IDS]) |
1da177e4c
|
57 |
|
4e9823111
|
58 59 |
#define shm_unlock(shp) \ ipc_unlock(&(shp)->shm_perm) |
1da177e4c
|
60 |
|
7748dbfaa
|
61 |
static int newseg(struct ipc_namespace *, struct ipc_params *); |
bc56bba8f
|
62 63 |
static void shm_open(struct vm_area_struct *vma); static void shm_close(struct vm_area_struct *vma); |
4e9823111
|
64 |
static void shm_destroy (struct ipc_namespace *ns, struct shmid_kernel *shp); |
1da177e4c
|
65 |
#ifdef CONFIG_PROC_FS |
19b4946ca
|
66 |
static int sysvipc_shm_proc_show(struct seq_file *s, void *it); |
1da177e4c
|
67 |
#endif |
ed2ddbf88
|
68 |
void shm_init_ns(struct ipc_namespace *ns) |
4e9823111
|
69 |
{ |
4e9823111
|
70 71 72 |
ns->shm_ctlmax = SHMMAX; ns->shm_ctlall = SHMALL; ns->shm_ctlmni = SHMMNI; |
b34a6b1da
|
73 |
ns->shm_rmid_forced = 0; |
4e9823111
|
74 |
ns->shm_tot = 0; |
e8148f758
|
75 |
ipc_init_ids(&shm_ids(ns)); |
4e9823111
|
76 |
} |
f4566f048
|
77 |
/* |
3e148c799
|
78 79 |
* Called with shm_ids.rw_mutex (writer) and the shp structure locked. * Only shm_ids.rw_mutex remains locked on exit. |
f4566f048
|
80 |
*/ |
01b8b07a5
|
81 |
static void do_shm_rmid(struct ipc_namespace *ns, struct kern_ipc_perm *ipcp) |
4e9823111
|
82 |
{ |
01b8b07a5
|
83 84 |
struct shmid_kernel *shp; shp = container_of(ipcp, struct shmid_kernel, shm_perm); |
4e9823111
|
85 86 87 88 89 90 91 92 |
if (shp->shm_nattch){ shp->shm_perm.mode |= SHM_DEST; /* Do not find it any more */ shp->shm_perm.key = IPC_PRIVATE; shm_unlock(shp); } else shm_destroy(ns, shp); } |
ae5e1b22f
|
93 |
#ifdef CONFIG_IPC_NS |
4e9823111
|
94 95 |
void shm_exit_ns(struct ipc_namespace *ns) { |
01b8b07a5
|
96 |
free_ipcs(ns, &shm_ids(ns), do_shm_rmid); |
7d6feeb28
|
97 |
idr_destroy(&ns->ids[IPC_SHM_IDS].ipcs_idr); |
4e9823111
|
98 |
} |
ae5e1b22f
|
99 |
#endif |
1da177e4c
|
100 |
|
140d0b210
|
101 |
static int __init ipc_ns_init(void) |
1da177e4c
|
102 |
{ |
ed2ddbf88
|
103 |
shm_init_ns(&init_ipc_ns); |
140d0b210
|
104 105 106 107 108 109 110 |
return 0; } pure_initcall(ipc_ns_init); void __init shm_init (void) { |
19b4946ca
|
111 |
ipc_init_proc_interface("sysvipc/shm", |
b79521807
|
112 113 114 115 116 117 118 |
#if BITS_PER_LONG <= 32 " key shmid perms size cpid lpid nattch uid gid cuid cgid atime dtime ctime rss swap ", #else " key shmid perms size cpid lpid nattch uid gid cuid cgid atime dtime ctime rss swap ", #endif |
4e9823111
|
119 |
IPC_SHM_IDS, sysvipc_shm_proc_show); |
1da177e4c
|
120 |
} |
3e148c799
|
121 |
/* |
3e148c799
|
122 |
* shm_lock_(check_) routines are called in the paths where the rw_mutex |
00c2bf85d
|
123 |
* is not necessarily held. |
3e148c799
|
124 |
*/ |
023a53557
|
125 |
static inline struct shmid_kernel *shm_lock(struct ipc_namespace *ns, int id) |
1da177e4c
|
126 |
{ |
03f02c765
|
127 |
struct kern_ipc_perm *ipcp = ipc_lock(&shm_ids(ns), id); |
b1ed88b47
|
128 129 |
if (IS_ERR(ipcp)) return (struct shmid_kernel *)ipcp; |
03f02c765
|
130 |
return container_of(ipcp, struct shmid_kernel, shm_perm); |
023a53557
|
131 |
} |
4c677e2ee
|
132 133 134 |
static inline void shm_lock_by_ptr(struct shmid_kernel *ipcp) { rcu_read_lock(); |
cf9d5d78d
|
135 |
ipc_lock_object(&ipcp->shm_perm); |
4c677e2ee
|
136 |
} |
023a53557
|
137 138 139 |
static inline struct shmid_kernel *shm_lock_check(struct ipc_namespace *ns, int id) { |
03f02c765
|
140 |
struct kern_ipc_perm *ipcp = ipc_lock_check(&shm_ids(ns), id); |
b1ed88b47
|
141 142 |
if (IS_ERR(ipcp)) return (struct shmid_kernel *)ipcp; |
03f02c765
|
143 |
return container_of(ipcp, struct shmid_kernel, shm_perm); |
1da177e4c
|
144 |
} |
7ca7e564e
|
145 |
static inline void shm_rmid(struct ipc_namespace *ns, struct shmid_kernel *s) |
1da177e4c
|
146 |
{ |
7ca7e564e
|
147 |
ipc_rmid(&shm_ids(ns), &s->shm_perm); |
1da177e4c
|
148 |
} |
1da177e4c
|
149 |
|
bc56bba8f
|
150 151 |
/* This is called by fork, once for every shm attach. */ static void shm_open(struct vm_area_struct *vma) |
4e9823111
|
152 |
{ |
bc56bba8f
|
153 154 |
struct file *file = vma->vm_file; struct shm_file_data *sfd = shm_file_data(file); |
1da177e4c
|
155 |
struct shmid_kernel *shp; |
bc56bba8f
|
156 |
shp = shm_lock(sfd->ns, sfd->id); |
023a53557
|
157 |
BUG_ON(IS_ERR(shp)); |
1da177e4c
|
158 |
shp->shm_atim = get_seconds(); |
b488893a3
|
159 |
shp->shm_lprid = task_tgid_vnr(current); |
1da177e4c
|
160 161 162 |
shp->shm_nattch++; shm_unlock(shp); } |
1da177e4c
|
163 164 165 |
/* * shm_destroy - free the struct shmid_kernel * |
f4566f048
|
166 |
* @ns: namespace |
1da177e4c
|
167 168 |
* @shp: struct to free * |
3e148c799
|
169 |
* It has to be called with shp and shm_ids.rw_mutex (writer) locked, |
1da177e4c
|
170 171 |
* but returns with shp unlocked and freed. */ |
4e9823111
|
172 |
static void shm_destroy(struct ipc_namespace *ns, struct shmid_kernel *shp) |
1da177e4c
|
173 |
{ |
4e9823111
|
174 |
ns->shm_tot -= (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT; |
7ca7e564e
|
175 |
shm_rmid(ns, shp); |
1da177e4c
|
176 177 178 |
shm_unlock(shp); if (!is_file_hugepages(shp->shm_file)) shmem_lock(shp->shm_file, 0, shp->mlock_user); |
353d5c30c
|
179 |
else if (shp->mlock_user) |
496ad9aa8
|
180 |
user_shm_unlock(file_inode(shp->shm_file)->i_size, |
1da177e4c
|
181 182 183 184 185 186 187 |
shp->mlock_user); fput (shp->shm_file); security_shm_free(shp); ipc_rcu_putref(shp); } /* |
b34a6b1da
|
188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 |
* shm_may_destroy - identifies whether shm segment should be destroyed now * * Returns true if and only if there are no active users of the segment and * one of the following is true: * * 1) shmctl(id, IPC_RMID, NULL) was called for this shp * * 2) sysctl kernel.shm_rmid_forced is set to 1. */ static bool shm_may_destroy(struct ipc_namespace *ns, struct shmid_kernel *shp) { return (shp->shm_nattch == 0) && (ns->shm_rmid_forced || (shp->shm_perm.mode & SHM_DEST)); } /* |
bc56bba8f
|
205 |
* remove the attach descriptor vma. |
1da177e4c
|
206 207 208 209 |
* free memory for segment if it is marked destroyed. * The descriptor has already been removed from the current->mm->mmap list * and will later be kfree()d. */ |
bc56bba8f
|
210 |
static void shm_close(struct vm_area_struct *vma) |
1da177e4c
|
211 |
{ |
bc56bba8f
|
212 213 |
struct file * file = vma->vm_file; struct shm_file_data *sfd = shm_file_data(file); |
1da177e4c
|
214 |
struct shmid_kernel *shp; |
bc56bba8f
|
215 |
struct ipc_namespace *ns = sfd->ns; |
4e9823111
|
216 |
|
3e148c799
|
217 |
down_write(&shm_ids(ns).rw_mutex); |
1da177e4c
|
218 |
/* remove from the list of attaches of the shm segment */ |
00c2bf85d
|
219 |
shp = shm_lock(ns, sfd->id); |
023a53557
|
220 |
BUG_ON(IS_ERR(shp)); |
b488893a3
|
221 |
shp->shm_lprid = task_tgid_vnr(current); |
1da177e4c
|
222 223 |
shp->shm_dtim = get_seconds(); shp->shm_nattch--; |
b34a6b1da
|
224 225 226 227 228 229 |
if (shm_may_destroy(ns, shp)) shm_destroy(ns, shp); else shm_unlock(shp); up_write(&shm_ids(ns).rw_mutex); } |
4c677e2ee
|
230 |
/* Called with ns->shm_ids(ns).rw_mutex locked */ |
b34a6b1da
|
231 232 233 |
static int shm_try_destroy_current(int id, void *p, void *data) { struct ipc_namespace *ns = data; |
4c677e2ee
|
234 235 |
struct kern_ipc_perm *ipcp = p; struct shmid_kernel *shp = container_of(ipcp, struct shmid_kernel, shm_perm); |
b34a6b1da
|
236 |
|
4c677e2ee
|
237 |
if (shp->shm_creator != current) |
5774ed014
|
238 |
return 0; |
5774ed014
|
239 240 241 242 243 244 245 246 247 248 249 250 |
/* * Mark it as orphaned to destroy the segment when * kernel.shm_rmid_forced is changed. * It is noop if the following shm_may_destroy() returns true. */ shp->shm_creator = NULL; /* * Don't even try to destroy it. If shm_rmid_forced=0 and IPC_RMID * is not set, it shouldn't be deleted here. */ |
4c677e2ee
|
251 |
if (!ns->shm_rmid_forced) |
b34a6b1da
|
252 |
return 0; |
b34a6b1da
|
253 |
|
4c677e2ee
|
254 255 |
if (shm_may_destroy(ns, shp)) { shm_lock_by_ptr(shp); |
b34a6b1da
|
256 |
shm_destroy(ns, shp); |
4c677e2ee
|
257 |
} |
b34a6b1da
|
258 259 |
return 0; } |
4c677e2ee
|
260 |
/* Called with ns->shm_ids(ns).rw_mutex locked */ |
b34a6b1da
|
261 262 263 |
static int shm_try_destroy_orphaned(int id, void *p, void *data) { struct ipc_namespace *ns = data; |
4c677e2ee
|
264 265 |
struct kern_ipc_perm *ipcp = p; struct shmid_kernel *shp = container_of(ipcp, struct shmid_kernel, shm_perm); |
b34a6b1da
|
266 267 268 269 |
/* * We want to destroy segments without users and with already * exit'ed originating process. |
4c677e2ee
|
270 271 |
* * As shp->* are changed under rw_mutex, it's safe to skip shp locking. |
b34a6b1da
|
272 |
*/ |
4c677e2ee
|
273 |
if (shp->shm_creator != NULL) |
b34a6b1da
|
274 |
return 0; |
b34a6b1da
|
275 |
|
4c677e2ee
|
276 277 |
if (shm_may_destroy(ns, shp)) { shm_lock_by_ptr(shp); |
4e9823111
|
278 |
shm_destroy(ns, shp); |
4c677e2ee
|
279 |
} |
b34a6b1da
|
280 281 282 283 284 285 |
return 0; } void shm_destroy_orphaned(struct ipc_namespace *ns) { down_write(&shm_ids(ns).rw_mutex); |
33a30ed4b
|
286 |
if (shm_ids(ns).in_use) |
4c677e2ee
|
287 |
idr_for_each(&shm_ids(ns).ipcs_idr, &shm_try_destroy_orphaned, ns); |
b34a6b1da
|
288 289 290 291 292 293 |
up_write(&shm_ids(ns).rw_mutex); } void exit_shm(struct task_struct *task) { |
4c677e2ee
|
294 |
struct ipc_namespace *ns = task->nsproxy->ipc_ns; |
b34a6b1da
|
295 |
|
298507d4d
|
296 297 |
if (shm_ids(ns).in_use == 0) return; |
b34a6b1da
|
298 299 |
/* Destroy all already created segments, but not mapped yet */ down_write(&shm_ids(ns).rw_mutex); |
33a30ed4b
|
300 |
if (shm_ids(ns).in_use) |
4c677e2ee
|
301 |
idr_for_each(&shm_ids(ns).ipcs_idr, &shm_try_destroy_current, ns); |
3e148c799
|
302 |
up_write(&shm_ids(ns).rw_mutex); |
1da177e4c
|
303 |
} |
d0217ac04
|
304 |
static int shm_fault(struct vm_area_struct *vma, struct vm_fault *vmf) |
bc56bba8f
|
305 306 307 |
{ struct file *file = vma->vm_file; struct shm_file_data *sfd = shm_file_data(file); |
d0217ac04
|
308 |
return sfd->vm_ops->fault(vma, vmf); |
bc56bba8f
|
309 310 311 |
} #ifdef CONFIG_NUMA |
d823e3e75
|
312 |
static int shm_set_policy(struct vm_area_struct *vma, struct mempolicy *new) |
bc56bba8f
|
313 314 315 316 317 318 319 320 |
{ struct file *file = vma->vm_file; struct shm_file_data *sfd = shm_file_data(file); int err = 0; if (sfd->vm_ops->set_policy) err = sfd->vm_ops->set_policy(vma, new); return err; } |
d823e3e75
|
321 322 |
static struct mempolicy *shm_get_policy(struct vm_area_struct *vma, unsigned long addr) |
bc56bba8f
|
323 324 325 326 327 328 329 |
{ struct file *file = vma->vm_file; struct shm_file_data *sfd = shm_file_data(file); struct mempolicy *pol = NULL; if (sfd->vm_ops->get_policy) pol = sfd->vm_ops->get_policy(vma, addr); |
52cd3b074
|
330 |
else if (vma->vm_policy) |
bc56bba8f
|
331 |
pol = vma->vm_policy; |
52cd3b074
|
332 |
|
bc56bba8f
|
333 334 335 |
return pol; } #endif |
1da177e4c
|
336 337 |
static int shm_mmap(struct file * file, struct vm_area_struct * vma) { |
bc56bba8f
|
338 |
struct shm_file_data *sfd = shm_file_data(file); |
b0e15190e
|
339 |
int ret; |
bc56bba8f
|
340 341 342 343 |
ret = sfd->file->f_op->mmap(sfd->file, vma); if (ret != 0) return ret; sfd->vm_ops = vma->vm_ops; |
2e92a3bae
|
344 |
#ifdef CONFIG_MMU |
54cb8821d
|
345 |
BUG_ON(!sfd->vm_ops->fault); |
2e92a3bae
|
346 |
#endif |
bc56bba8f
|
347 348 |
vma->vm_ops = &shm_vm_ops; shm_open(vma); |
b0e15190e
|
349 350 |
return ret; |
1da177e4c
|
351 |
} |
4e9823111
|
352 353 |
static int shm_release(struct inode *ino, struct file *file) { |
bc56bba8f
|
354 |
struct shm_file_data *sfd = shm_file_data(file); |
4e9823111
|
355 |
|
bc56bba8f
|
356 357 358 |
put_ipc_ns(sfd->ns); shm_file_data(file) = NULL; kfree(sfd); |
4e9823111
|
359 360 |
return 0; } |
02c24a821
|
361 |
static int shm_fsync(struct file *file, loff_t start, loff_t end, int datasync) |
516dffdcd
|
362 |
{ |
516dffdcd
|
363 |
struct shm_file_data *sfd = shm_file_data(file); |
516dffdcd
|
364 |
|
7ea808591
|
365 366 |
if (!sfd->file->f_op->fsync) return -EINVAL; |
02c24a821
|
367 |
return sfd->file->f_op->fsync(sfd->file, start, end, datasync); |
516dffdcd
|
368 |
} |
7d8a45695
|
369 370 371 372 373 374 375 376 377 |
static long shm_fallocate(struct file *file, int mode, loff_t offset, loff_t len) { struct shm_file_data *sfd = shm_file_data(file); if (!sfd->file->f_op->fallocate) return -EOPNOTSUPP; return sfd->file->f_op->fallocate(file, mode, offset, len); } |
bc56bba8f
|
378 379 380 381 382 |
static unsigned long shm_get_unmapped_area(struct file *file, unsigned long addr, unsigned long len, unsigned long pgoff, unsigned long flags) { struct shm_file_data *sfd = shm_file_data(file); |
c4caa7781
|
383 384 |
return sfd->file->f_op->get_unmapped_area(sfd->file, addr, len, pgoff, flags); |
bc56bba8f
|
385 |
} |
bc56bba8f
|
386 |
|
9a32144e9
|
387 |
static const struct file_operations shm_file_operations = { |
4e9823111
|
388 |
.mmap = shm_mmap, |
516dffdcd
|
389 |
.fsync = shm_fsync, |
4e9823111
|
390 |
.release = shm_release, |
ed5e5894b
|
391 392 393 |
#ifndef CONFIG_MMU .get_unmapped_area = shm_get_unmapped_area, #endif |
6038f373a
|
394 |
.llseek = noop_llseek, |
7d8a45695
|
395 |
.fallocate = shm_fallocate, |
c4caa7781
|
396 397 398 399 400 401 |
}; static const struct file_operations shm_file_operations_huge = { .mmap = shm_mmap, .fsync = shm_fsync, .release = shm_release, |
bc56bba8f
|
402 |
.get_unmapped_area = shm_get_unmapped_area, |
6038f373a
|
403 |
.llseek = noop_llseek, |
7d8a45695
|
404 |
.fallocate = shm_fallocate, |
1da177e4c
|
405 |
}; |
c4caa7781
|
406 407 408 409 |
int is_file_shm_hugepages(struct file *file) { return file->f_op == &shm_file_operations_huge; } |
f0f37e2f7
|
410 |
static const struct vm_operations_struct shm_vm_ops = { |
1da177e4c
|
411 412 |
.open = shm_open, /* callback for a new vm-area open */ .close = shm_close, /* callback for when the vm-area is released */ |
54cb8821d
|
413 |
.fault = shm_fault, |
bc56bba8f
|
414 415 416 |
#if defined(CONFIG_NUMA) .set_policy = shm_set_policy, .get_policy = shm_get_policy, |
1da177e4c
|
417 418 |
#endif }; |
f4566f048
|
419 420 421 422 423 |
/** * newseg - Create a new shared memory segment * @ns: namespace * @params: ptr to the structure that contains key, size and shmflg * |
3e148c799
|
424 |
* Called with shm_ids.rw_mutex held as a writer. |
f4566f048
|
425 |
*/ |
7748dbfaa
|
426 |
static int newseg(struct ipc_namespace *ns, struct ipc_params *params) |
1da177e4c
|
427 |
{ |
7748dbfaa
|
428 429 430 |
key_t key = params->key; int shmflg = params->flg; size_t size = params->u.size; |
1da177e4c
|
431 432 |
int error; struct shmid_kernel *shp; |
d69f3bad4
|
433 |
size_t numpages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT; |
1da177e4c
|
434 435 436 |
struct file * file; char name[13]; int id; |
ca16d140a
|
437 |
vm_flags_t acctflag = 0; |
1da177e4c
|
438 |
|
4e9823111
|
439 |
if (size < SHMMIN || size > ns->shm_ctlmax) |
1da177e4c
|
440 |
return -EINVAL; |
f66d45e99
|
441 |
if (ns->shm_tot + numpages > ns->shm_ctlall) |
1da177e4c
|
442 443 444 445 446 447 448 |
return -ENOSPC; shp = ipc_rcu_alloc(sizeof(*shp)); if (!shp) return -ENOMEM; shp->shm_perm.key = key; |
b33291c0b
|
449 |
shp->shm_perm.mode = (shmflg & S_IRWXUGO); |
1da177e4c
|
450 451 452 453 454 455 456 457 |
shp->mlock_user = NULL; shp->shm_perm.security = NULL; error = security_shm_alloc(shp); if (error) { ipc_rcu_putref(shp); return error; } |
9d66586f7
|
458 |
sprintf (name, "SYSV%08x", key); |
1da177e4c
|
459 |
if (shmflg & SHM_HUGETLB) { |
c103a4dc4
|
460 |
struct hstate *hs; |
091d0d55b
|
461 |
size_t hugesize; |
c103a4dc4
|
462 |
hs = hstate_sizelog((shmflg >> SHM_HUGE_SHIFT) & SHM_HUGE_MASK); |
091d0d55b
|
463 464 465 466 467 |
if (!hs) { error = -EINVAL; goto no_file; } hugesize = ALIGN(size, huge_page_size(hs)); |
af73e4d95
|
468 |
|
5a6fe1259
|
469 470 471 |
/* hugetlb_file_setup applies strict accounting */ if (shmflg & SHM_NORESERVE) acctflag = VM_NORESERVE; |
af73e4d95
|
472 |
file = hugetlb_file_setup(name, hugesize, acctflag, |
42d7395fe
|
473 474 |
&shp->mlock_user, HUGETLB_SHMFS_INODE, (shmflg >> SHM_HUGE_SHIFT) & SHM_HUGE_MASK); |
1da177e4c
|
475 |
} else { |
bf8f972d3
|
476 477 478 479 480 481 |
/* * Do not allow no accounting for OVERCOMMIT_NEVER, even * if it's asked for. */ if ((shmflg & SHM_NORESERVE) && sysctl_overcommit_memory != OVERCOMMIT_NEVER) |
fc8744adc
|
482 |
acctflag = VM_NORESERVE; |
bf8f972d3
|
483 |
file = shmem_file_setup(name, size, acctflag); |
1da177e4c
|
484 485 486 487 |
} error = PTR_ERR(file); if (IS_ERR(file)) goto no_file; |
48dea404e
|
488 |
id = ipc_addid(&shm_ids(ns), &shp->shm_perm, ns->shm_ctlmni); |
283bb7fad
|
489 490 |
if (id < 0) { error = id; |
1da177e4c
|
491 |
goto no_id; |
283bb7fad
|
492 |
} |
1da177e4c
|
493 |
|
b488893a3
|
494 |
shp->shm_cprid = task_tgid_vnr(current); |
1da177e4c
|
495 496 497 498 499 |
shp->shm_lprid = 0; shp->shm_atim = shp->shm_dtim = 0; shp->shm_ctim = get_seconds(); shp->shm_segsz = size; shp->shm_nattch = 0; |
1da177e4c
|
500 |
shp->shm_file = file; |
5774ed014
|
501 |
shp->shm_creator = current; |
dbfcd91f0
|
502 |
|
30475cc12
|
503 504 505 506 |
/* * shmid gets reported as "inode#" in /proc/pid/maps. * proc-ps tools use this. Changing this will break them. */ |
496ad9aa8
|
507 |
file_inode(file)->i_ino = shp->shm_perm.id; |
551110a94
|
508 |
|
4e9823111
|
509 |
ns->shm_tot += numpages; |
7ca7e564e
|
510 |
error = shp->shm_perm.id; |
dbfcd91f0
|
511 |
|
cf9d5d78d
|
512 |
ipc_unlock_object(&shp->shm_perm); |
dbfcd91f0
|
513 |
rcu_read_unlock(); |
7ca7e564e
|
514 |
return error; |
1da177e4c
|
515 516 |
no_id: |
2195d2818
|
517 |
if (is_file_hugepages(file) && shp->mlock_user) |
353d5c30c
|
518 |
user_shm_unlock(size, shp->mlock_user); |
1da177e4c
|
519 520 521 522 523 524 |
fput(file); no_file: security_shm_free(shp); ipc_rcu_putref(shp); return error; } |
f4566f048
|
525 |
/* |
3e148c799
|
526 |
* Called with shm_ids.rw_mutex and ipcp locked. |
f4566f048
|
527 |
*/ |
03f02c765
|
528 |
static inline int shm_security(struct kern_ipc_perm *ipcp, int shmflg) |
7748dbfaa
|
529 |
{ |
03f02c765
|
530 531 532 533 |
struct shmid_kernel *shp; shp = container_of(ipcp, struct shmid_kernel, shm_perm); return security_shm_associate(shp, shmflg); |
7748dbfaa
|
534 |
} |
f4566f048
|
535 |
/* |
3e148c799
|
536 |
* Called with shm_ids.rw_mutex and ipcp locked. |
f4566f048
|
537 |
*/ |
03f02c765
|
538 539 |
static inline int shm_more_checks(struct kern_ipc_perm *ipcp, struct ipc_params *params) |
7748dbfaa
|
540 |
{ |
03f02c765
|
541 542 543 544 |
struct shmid_kernel *shp; shp = container_of(ipcp, struct shmid_kernel, shm_perm); if (shp->shm_segsz < params->u.size) |
7748dbfaa
|
545 546 547 548 |
return -EINVAL; return 0; } |
d5460c997
|
549 |
SYSCALL_DEFINE3(shmget, key_t, key, size_t, size, int, shmflg) |
1da177e4c
|
550 |
{ |
4e9823111
|
551 |
struct ipc_namespace *ns; |
7748dbfaa
|
552 553 |
struct ipc_ops shm_ops; struct ipc_params shm_params; |
4e9823111
|
554 555 |
ns = current->nsproxy->ipc_ns; |
1da177e4c
|
556 |
|
7748dbfaa
|
557 558 559 |
shm_ops.getnew = newseg; shm_ops.associate = shm_security; shm_ops.more_checks = shm_more_checks; |
7ca7e564e
|
560 |
|
7748dbfaa
|
561 562 563 |
shm_params.key = key; shm_params.flg = shmflg; shm_params.u.size = size; |
1da177e4c
|
564 |
|
7748dbfaa
|
565 |
return ipcget(ns, &shm_ids(ns), &shm_ops, &shm_params); |
1da177e4c
|
566 567 568 569 570 571 572 573 574 575 |
} static inline unsigned long copy_shmid_to_user(void __user *buf, struct shmid64_ds *in, int version) { switch(version) { case IPC_64: return copy_to_user(buf, in, sizeof(*in)); case IPC_OLD: { struct shmid_ds out; |
3af54c9bd
|
576 |
memset(&out, 0, sizeof(out)); |
1da177e4c
|
577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 |
ipc64_perm_to_ipc_perm(&in->shm_perm, &out.shm_perm); out.shm_segsz = in->shm_segsz; out.shm_atime = in->shm_atime; out.shm_dtime = in->shm_dtime; out.shm_ctime = in->shm_ctime; out.shm_cpid = in->shm_cpid; out.shm_lpid = in->shm_lpid; out.shm_nattch = in->shm_nattch; return copy_to_user(buf, &out, sizeof(out)); } default: return -EINVAL; } } |
016d7132f
|
592 593 |
static inline unsigned long copy_shmid_from_user(struct shmid64_ds *out, void __user *buf, int version) |
1da177e4c
|
594 595 596 |
{ switch(version) { case IPC_64: |
016d7132f
|
597 |
if (copy_from_user(out, buf, sizeof(*out))) |
1da177e4c
|
598 |
return -EFAULT; |
1da177e4c
|
599 |
return 0; |
1da177e4c
|
600 601 602 603 604 605 |
case IPC_OLD: { struct shmid_ds tbuf_old; if (copy_from_user(&tbuf_old, buf, sizeof(tbuf_old))) return -EFAULT; |
016d7132f
|
606 607 608 |
out->shm_perm.uid = tbuf_old.shm_perm.uid; out->shm_perm.gid = tbuf_old.shm_perm.gid; out->shm_perm.mode = tbuf_old.shm_perm.mode; |
1da177e4c
|
609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 |
return 0; } default: return -EINVAL; } } static inline unsigned long copy_shminfo_to_user(void __user *buf, struct shminfo64 *in, int version) { switch(version) { case IPC_64: return copy_to_user(buf, in, sizeof(*in)); case IPC_OLD: { struct shminfo out; if(in->shmmax > INT_MAX) out.shmmax = INT_MAX; else out.shmmax = (int)in->shmmax; out.shmmin = in->shmmin; out.shmmni = in->shmmni; out.shmseg = in->shmseg; out.shmall = in->shmall; return copy_to_user(buf, &out, sizeof(out)); } default: return -EINVAL; } } |
f4566f048
|
642 |
/* |
b79521807
|
643 644 645 646 647 648 649 |
* Calculate and add used RSS and swap pages of a shm. * Called with shm_ids.rw_mutex held as a reader */ static void shm_add_rss_swap(struct shmid_kernel *shp, unsigned long *rss_add, unsigned long *swp_add) { struct inode *inode; |
496ad9aa8
|
650 |
inode = file_inode(shp->shm_file); |
b79521807
|
651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 |
if (is_file_hugepages(shp->shm_file)) { struct address_space *mapping = inode->i_mapping; struct hstate *h = hstate_file(shp->shm_file); *rss_add += pages_per_huge_page(h) * mapping->nrpages; } else { #ifdef CONFIG_SHMEM struct shmem_inode_info *info = SHMEM_I(inode); spin_lock(&info->lock); *rss_add += inode->i_mapping->nrpages; *swp_add += info->swapped; spin_unlock(&info->lock); #else *rss_add += inode->i_mapping->nrpages; #endif } } /* |
3e148c799
|
670 |
* Called with shm_ids.rw_mutex held as a reader |
f4566f048
|
671 |
*/ |
4e9823111
|
672 673 |
static void shm_get_stat(struct ipc_namespace *ns, unsigned long *rss, unsigned long *swp) |
1da177e4c
|
674 |
{ |
7ca7e564e
|
675 676 |
int next_id; int total, in_use; |
1da177e4c
|
677 678 679 |
*rss = 0; *swp = 0; |
7ca7e564e
|
680 681 682 |
in_use = shm_ids(ns).in_use; for (total = 0, next_id = 0; total < in_use; next_id++) { |
e562aebc6
|
683 |
struct kern_ipc_perm *ipc; |
1da177e4c
|
684 |
struct shmid_kernel *shp; |
1da177e4c
|
685 |
|
e562aebc6
|
686 687 |
ipc = idr_find(&shm_ids(ns).ipcs_idr, next_id); if (ipc == NULL) |
1da177e4c
|
688 |
continue; |
e562aebc6
|
689 |
shp = container_of(ipc, struct shmid_kernel, shm_perm); |
1da177e4c
|
690 |
|
b79521807
|
691 |
shm_add_rss_swap(shp, rss, swp); |
7ca7e564e
|
692 693 |
total++; |
1da177e4c
|
694 695 |
} } |
8d4cc8b5c
|
696 697 698 699 700 701 702 |
/* * This function handles some shmctl commands which require the rw_mutex * to be held in write mode. * NOTE: no locks must be held, the rw_mutex is taken inside this function. */ static int shmctl_down(struct ipc_namespace *ns, int shmid, int cmd, struct shmid_ds __user *buf, int version) |
1da177e4c
|
703 |
{ |
8d4cc8b5c
|
704 |
struct kern_ipc_perm *ipcp; |
016d7132f
|
705 |
struct shmid64_ds shmid64; |
1da177e4c
|
706 |
struct shmid_kernel *shp; |
8d4cc8b5c
|
707 708 709 |
int err; if (cmd == IPC_SET) { |
016d7132f
|
710 |
if (copy_shmid_from_user(&shmid64, buf, version)) |
8d4cc8b5c
|
711 712 |
return -EFAULT; } |
7b4cc5d84
|
713 714 |
down_write(&shm_ids(ns).rw_mutex); rcu_read_lock(); |
b0e77598f
|
715 716 |
ipcp = ipcctl_pre_down(ns, &shm_ids(ns), shmid, cmd, &shmid64.shm_perm, 0); |
7b4cc5d84
|
717 718 719 720 721 |
if (IS_ERR(ipcp)) { err = PTR_ERR(ipcp); /* the ipc lock is not held upon failure */ goto out_unlock1; } |
8d4cc8b5c
|
722 |
|
a5f75e7f2
|
723 |
shp = container_of(ipcp, struct shmid_kernel, shm_perm); |
8d4cc8b5c
|
724 725 726 |
err = security_shm_shmctl(shp, cmd); if (err) |
7b4cc5d84
|
727 |
goto out_unlock0; |
8d4cc8b5c
|
728 729 |
switch (cmd) { case IPC_RMID: |
7b4cc5d84
|
730 |
/* do_shm_rmid unlocks the ipc object and rcu */ |
8d4cc8b5c
|
731 732 733 |
do_shm_rmid(ns, ipcp); goto out_up; case IPC_SET: |
1efdb69b0
|
734 735 |
err = ipc_update_perm(&shmid64.shm_perm, ipcp); if (err) |
7b4cc5d84
|
736 |
goto out_unlock0; |
8d4cc8b5c
|
737 738 739 740 741 |
shp->shm_ctim = get_seconds(); break; default: err = -EINVAL; } |
7b4cc5d84
|
742 743 744 745 746 |
out_unlock0: ipc_unlock_object(&shp->shm_perm); out_unlock1: rcu_read_unlock(); |
8d4cc8b5c
|
747 748 749 750 |
out_up: up_write(&shm_ids(ns).rw_mutex); return err; } |
d5460c997
|
751 |
SYSCALL_DEFINE3(shmctl, int, shmid, int, cmd, struct shmid_ds __user *, buf) |
8d4cc8b5c
|
752 753 |
{ struct shmid_kernel *shp; |
1da177e4c
|
754 |
int err, version; |
4e9823111
|
755 |
struct ipc_namespace *ns; |
1da177e4c
|
756 757 758 759 760 761 762 |
if (cmd < 0 || shmid < 0) { err = -EINVAL; goto out; } version = ipc_parse_version(&cmd); |
4e9823111
|
763 |
ns = current->nsproxy->ipc_ns; |
1da177e4c
|
764 765 766 767 768 769 770 771 772 |
switch (cmd) { /* replace with proc interface ? */ case IPC_INFO: { struct shminfo64 shminfo; err = security_shm_shmctl(NULL, cmd); if (err) return err; |
e8148f758
|
773 |
memset(&shminfo, 0, sizeof(shminfo)); |
4e9823111
|
774 775 776 |
shminfo.shmmni = shminfo.shmseg = ns->shm_ctlmni; shminfo.shmmax = ns->shm_ctlmax; shminfo.shmall = ns->shm_ctlall; |
1da177e4c
|
777 778 779 780 |
shminfo.shmmin = SHMMIN; if(copy_shminfo_to_user (buf, &shminfo, version)) return -EFAULT; |
f4566f048
|
781 |
|
3e148c799
|
782 |
down_read(&shm_ids(ns).rw_mutex); |
7ca7e564e
|
783 |
err = ipc_get_maxid(&shm_ids(ns)); |
3e148c799
|
784 |
up_read(&shm_ids(ns).rw_mutex); |
f4566f048
|
785 |
|
1da177e4c
|
786 787 788 789 790 791 792 793 794 795 796 |
if(err<0) err = 0; goto out; } case SHM_INFO: { struct shm_info shm_info; err = security_shm_shmctl(NULL, cmd); if (err) return err; |
e8148f758
|
797 |
memset(&shm_info, 0, sizeof(shm_info)); |
3e148c799
|
798 |
down_read(&shm_ids(ns).rw_mutex); |
4e9823111
|
799 800 801 |
shm_info.used_ids = shm_ids(ns).in_use; shm_get_stat (ns, &shm_info.shm_rss, &shm_info.shm_swp); shm_info.shm_tot = ns->shm_tot; |
1da177e4c
|
802 803 |
shm_info.swap_attempts = 0; shm_info.swap_successes = 0; |
7ca7e564e
|
804 |
err = ipc_get_maxid(&shm_ids(ns)); |
3e148c799
|
805 |
up_read(&shm_ids(ns).rw_mutex); |
e8148f758
|
806 |
if (copy_to_user(buf, &shm_info, sizeof(shm_info))) { |
1da177e4c
|
807 808 809 810 811 812 813 814 815 816 817 818 |
err = -EFAULT; goto out; } err = err < 0 ? 0 : err; goto out; } case SHM_STAT: case IPC_STAT: { struct shmid64_ds tbuf; int result; |
023a53557
|
819 |
|
023a53557
|
820 821 822 823 824 825 |
if (cmd == SHM_STAT) { shp = shm_lock(ns, shmid); if (IS_ERR(shp)) { err = PTR_ERR(shp); goto out; } |
7ca7e564e
|
826 |
result = shp->shm_perm.id; |
1da177e4c
|
827 |
} else { |
023a53557
|
828 829 830 831 832 |
shp = shm_lock_check(ns, shmid); if (IS_ERR(shp)) { err = PTR_ERR(shp); goto out; } |
1da177e4c
|
833 834 |
result = 0; } |
e8148f758
|
835 |
err = -EACCES; |
b0e77598f
|
836 |
if (ipcperms(ns, &shp->shm_perm, S_IRUGO)) |
1da177e4c
|
837 838 839 840 |
goto out_unlock; err = security_shm_shmctl(shp, cmd); if (err) goto out_unlock; |
023a53557
|
841 |
memset(&tbuf, 0, sizeof(tbuf)); |
1da177e4c
|
842 843 844 845 846 847 848 |
kernel_to_ipc64_perm(&shp->shm_perm, &tbuf.shm_perm); tbuf.shm_segsz = shp->shm_segsz; tbuf.shm_atime = shp->shm_atim; tbuf.shm_dtime = shp->shm_dtim; tbuf.shm_ctime = shp->shm_ctim; tbuf.shm_cpid = shp->shm_cprid; tbuf.shm_lpid = shp->shm_lprid; |
bc56bba8f
|
849 |
tbuf.shm_nattch = shp->shm_nattch; |
1da177e4c
|
850 851 852 853 854 855 856 857 858 859 |
shm_unlock(shp); if(copy_shmid_to_user (buf, &tbuf, version)) err = -EFAULT; else err = result; goto out; } case SHM_LOCK: case SHM_UNLOCK: { |
85046579b
|
860 |
struct file *shm_file; |
89e004ea5
|
861 |
|
023a53557
|
862 863 864 |
shp = shm_lock_check(ns, shmid); if (IS_ERR(shp)) { err = PTR_ERR(shp); |
1da177e4c
|
865 866 |
goto out; } |
1da177e4c
|
867 |
|
a33e67510
|
868 |
audit_ipc_obj(&(shp->shm_perm)); |
073115d6b
|
869 |
|
b0e77598f
|
870 |
if (!ns_capable(ns->user_ns, CAP_IPC_LOCK)) { |
1efdb69b0
|
871 |
kuid_t euid = current_euid(); |
1da177e4c
|
872 |
err = -EPERM; |
1efdb69b0
|
873 874 |
if (!uid_eq(euid, shp->shm_perm.uid) && !uid_eq(euid, shp->shm_perm.cuid)) |
1da177e4c
|
875 |
goto out_unlock; |
f1eb1332b
|
876 |
if (cmd == SHM_LOCK && !rlimit(RLIMIT_MEMLOCK)) |
1da177e4c
|
877 878 879 880 881 882 |
goto out_unlock; } err = security_shm_shmctl(shp, cmd); if (err) goto out_unlock; |
85046579b
|
883 884 885 886 887 888 |
shm_file = shp->shm_file; if (is_file_hugepages(shm_file)) goto out_unlock; if (cmd == SHM_LOCK) { |
86a264abe
|
889 |
struct user_struct *user = current_user(); |
85046579b
|
890 891 892 893 |
err = shmem_lock(shm_file, 1, user); if (!err && !(shp->shm_perm.mode & SHM_LOCKED)) { shp->shm_perm.mode |= SHM_LOCKED; shp->mlock_user = user; |
1da177e4c
|
894 |
} |
85046579b
|
895 |
goto out_unlock; |
1da177e4c
|
896 |
} |
85046579b
|
897 898 899 900 901 902 903 904 |
/* SHM_UNLOCK */ if (!(shp->shm_perm.mode & SHM_LOCKED)) goto out_unlock; shmem_lock(shm_file, 0, shp->mlock_user); shp->shm_perm.mode &= ~SHM_LOCKED; shp->mlock_user = NULL; get_file(shm_file); |
1da177e4c
|
905 |
shm_unlock(shp); |
245132643
|
906 |
shmem_unlock_mapping(shm_file->f_mapping); |
85046579b
|
907 |
fput(shm_file); |
1da177e4c
|
908 909 910 |
goto out; } case IPC_RMID: |
1da177e4c
|
911 |
case IPC_SET: |
8d4cc8b5c
|
912 913 |
err = shmctl_down(ns, shmid, cmd, buf, version); return err; |
1da177e4c
|
914 |
default: |
8d4cc8b5c
|
915 |
return -EINVAL; |
1da177e4c
|
916 |
} |
1da177e4c
|
917 918 919 920 921 922 923 924 925 926 927 928 929 |
out_unlock: shm_unlock(shp); out: return err; } /* * Fix shmaddr, allocate descriptor, map shm, add attach descriptor to lists. * * NOTE! Despite the name, this is NOT a direct system call entrypoint. The * "raddr" thing points to kernel space, and there has to be a wrapper around * this. */ |
079a96ae3
|
930 931 |
long do_shmat(int shmid, char __user *shmaddr, int shmflg, ulong *raddr, unsigned long shmlba) |
1da177e4c
|
932 933 934 935 936 937 938 939 |
{ struct shmid_kernel *shp; unsigned long addr; unsigned long size; struct file * file; int err; unsigned long flags; unsigned long prot; |
1da177e4c
|
940 |
int acc_mode; |
4e9823111
|
941 |
struct ipc_namespace *ns; |
bc56bba8f
|
942 943 |
struct shm_file_data *sfd; struct path path; |
aeb5d7270
|
944 |
fmode_t f_mode; |
41badc15c
|
945 |
unsigned long populate = 0; |
1da177e4c
|
946 |
|
bc56bba8f
|
947 948 |
err = -EINVAL; if (shmid < 0) |
1da177e4c
|
949 |
goto out; |
bc56bba8f
|
950 |
else if ((addr = (ulong)shmaddr)) { |
079a96ae3
|
951 |
if (addr & (shmlba - 1)) { |
1da177e4c
|
952 |
if (shmflg & SHM_RND) |
079a96ae3
|
953 |
addr &= ~(shmlba - 1); /* round down */ |
1da177e4c
|
954 955 956 957 |
else #ifndef __ARCH_FORCE_SHMLBA if (addr & ~PAGE_MASK) #endif |
bc56bba8f
|
958 |
goto out; |
1da177e4c
|
959 960 961 962 |
} flags = MAP_SHARED | MAP_FIXED; } else { if ((shmflg & SHM_REMAP)) |
bc56bba8f
|
963 |
goto out; |
1da177e4c
|
964 965 966 967 968 969 |
flags = MAP_SHARED; } if (shmflg & SHM_RDONLY) { prot = PROT_READ; |
1da177e4c
|
970 |
acc_mode = S_IRUGO; |
bc56bba8f
|
971 |
f_mode = FMODE_READ; |
1da177e4c
|
972 973 |
} else { prot = PROT_READ | PROT_WRITE; |
1da177e4c
|
974 |
acc_mode = S_IRUGO | S_IWUGO; |
bc56bba8f
|
975 |
f_mode = FMODE_READ | FMODE_WRITE; |
1da177e4c
|
976 977 978 979 980 981 982 983 984 985 |
} if (shmflg & SHM_EXEC) { prot |= PROT_EXEC; acc_mode |= S_IXUGO; } /* * We cannot rely on the fs check since SYSV IPC does have an * additional creator id... */ |
4e9823111
|
986 |
ns = current->nsproxy->ipc_ns; |
023a53557
|
987 988 989 |
shp = shm_lock_check(ns, shmid); if (IS_ERR(shp)) { err = PTR_ERR(shp); |
1da177e4c
|
990 |
goto out; |
023a53557
|
991 |
} |
bc56bba8f
|
992 993 |
err = -EACCES; |
b0e77598f
|
994 |
if (ipcperms(ns, &shp->shm_perm, acc_mode)) |
bc56bba8f
|
995 |
goto out_unlock; |
1da177e4c
|
996 997 |
err = security_shm_shmat(shp, shmaddr, shmflg); |
bc56bba8f
|
998 999 |
if (err) goto out_unlock; |
2c48b9c45
|
1000 1001 |
path = shp->shm_file->f_path; path_get(&path); |
1da177e4c
|
1002 |
shp->shm_nattch++; |
bc56bba8f
|
1003 |
size = i_size_read(path.dentry->d_inode); |
1da177e4c
|
1004 |
shm_unlock(shp); |
bc56bba8f
|
1005 1006 1007 |
err = -ENOMEM; sfd = kzalloc(sizeof(*sfd), GFP_KERNEL); if (!sfd) |
ce8d2cdf3
|
1008 |
goto out_put_dentry; |
bc56bba8f
|
1009 |
|
2c48b9c45
|
1010 1011 |
file = alloc_file(&path, f_mode, is_file_hugepages(shp->shm_file) ? |
c4caa7781
|
1012 1013 |
&shm_file_operations_huge : &shm_file_operations); |
39b652527
|
1014 1015 |
err = PTR_ERR(file); if (IS_ERR(file)) |
bc56bba8f
|
1016 |
goto out_free; |
bc56bba8f
|
1017 |
file->private_data = sfd; |
bc56bba8f
|
1018 |
file->f_mapping = shp->shm_file->f_mapping; |
7ca7e564e
|
1019 |
sfd->id = shp->shm_perm.id; |
bc56bba8f
|
1020 1021 1022 |
sfd->ns = get_ipc_ns(ns); sfd->file = shp->shm_file; sfd->vm_ops = NULL; |
8b3ec6814
|
1023 1024 1025 |
err = security_mmap_file(file, prot, flags); if (err) goto out_fput; |
1da177e4c
|
1026 1027 |
down_write(¤t->mm->mmap_sem); if (addr && !(shmflg & SHM_REMAP)) { |
bc56bba8f
|
1028 |
err = -EINVAL; |
1da177e4c
|
1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 |
if (find_vma_intersection(current->mm, addr, addr + size)) goto invalid; /* * If shm segment goes below stack, make sure there is some * space left for the stack to grow (at least 4 pages). */ if (addr < current->mm->start_stack && addr > current->mm->start_stack - size - PAGE_SIZE * 5) goto invalid; } |
bebeb3d68
|
1040 1041 |
addr = do_mmap_pgoff(file, addr, size, prot, flags, 0, &populate); *raddr = addr; |
bc56bba8f
|
1042 |
err = 0; |
bebeb3d68
|
1043 1044 |
if (IS_ERR_VALUE(addr)) err = (long)addr; |
1da177e4c
|
1045 1046 |
invalid: up_write(¤t->mm->mmap_sem); |
bebeb3d68
|
1047 |
if (populate) |
41badc15c
|
1048 |
mm_populate(addr, populate); |
1da177e4c
|
1049 |
|
8b3ec6814
|
1050 |
out_fput: |
bc56bba8f
|
1051 1052 1053 |
fput(file); out_nattch: |
3e148c799
|
1054 |
down_write(&shm_ids(ns).rw_mutex); |
00c2bf85d
|
1055 |
shp = shm_lock(ns, shmid); |
023a53557
|
1056 |
BUG_ON(IS_ERR(shp)); |
1da177e4c
|
1057 |
shp->shm_nattch--; |
b34a6b1da
|
1058 |
if (shm_may_destroy(ns, shp)) |
4e9823111
|
1059 |
shm_destroy(ns, shp); |
1da177e4c
|
1060 1061 |
else shm_unlock(shp); |
3e148c799
|
1062 |
up_write(&shm_ids(ns).rw_mutex); |
1da177e4c
|
1063 |
|
1da177e4c
|
1064 1065 |
out: return err; |
bc56bba8f
|
1066 1067 1068 1069 1070 1071 1072 |
out_unlock: shm_unlock(shp); goto out; out_free: kfree(sfd); |
ce8d2cdf3
|
1073 |
out_put_dentry: |
2c48b9c45
|
1074 |
path_put(&path); |
bc56bba8f
|
1075 |
goto out_nattch; |
1da177e4c
|
1076 |
} |
d5460c997
|
1077 |
SYSCALL_DEFINE3(shmat, int, shmid, char __user *, shmaddr, int, shmflg) |
7d87e14c2
|
1078 1079 1080 |
{ unsigned long ret; long err; |
079a96ae3
|
1081 |
err = do_shmat(shmid, shmaddr, shmflg, &ret, SHMLBA); |
7d87e14c2
|
1082 1083 1084 1085 1086 |
if (err) return err; force_successful_syscall_return(); return (long)ret; } |
1da177e4c
|
1087 1088 1089 1090 |
/* * detach and kill segment if marked destroyed. * The work is done in shm_close. */ |
d5460c997
|
1091 |
SYSCALL_DEFINE1(shmdt, char __user *, shmaddr) |
1da177e4c
|
1092 1093 |
{ struct mm_struct *mm = current->mm; |
586c7e6a2
|
1094 |
struct vm_area_struct *vma; |
1da177e4c
|
1095 |
unsigned long addr = (unsigned long)shmaddr; |
1da177e4c
|
1096 |
int retval = -EINVAL; |
586c7e6a2
|
1097 1098 1099 1100 |
#ifdef CONFIG_MMU loff_t size = 0; struct vm_area_struct *next; #endif |
1da177e4c
|
1101 |
|
df1e2fb54
|
1102 1103 |
if (addr & ~PAGE_MASK) return retval; |
1da177e4c
|
1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 |
down_write(&mm->mmap_sem); /* * This function tries to be smart and unmap shm segments that * were modified by partial mlock or munmap calls: * - It first determines the size of the shm segment that should be * unmapped: It searches for a vma that is backed by shm and that * started at address shmaddr. It records it's size and then unmaps * it. * - Then it unmaps all shm vmas that started at shmaddr and that * are within the initially determined size. * Errors from do_munmap are ignored: the function only fails if * it's called with invalid parameters or if it's called to unmap * a part of a vma. Both calls in this function are for full vmas, * the parameters are directly copied from the vma itself and always * valid - therefore do_munmap cannot fail. (famous last words?) */ /* * If it had been mremap()'d, the starting address would not * match the usual checks anyway. So assume all vma's are * above the starting address given. */ vma = find_vma(mm, addr); |
8feae1311
|
1127 |
#ifdef CONFIG_MMU |
1da177e4c
|
1128 1129 1130 1131 1132 1133 1134 1135 |
while (vma) { next = vma->vm_next; /* * Check if the starting address would match, i.e. it's * a fragment created by mprotect() and/or munmap(), or it * otherwise it starts at this address with no hassles. */ |
bc56bba8f
|
1136 |
if ((vma->vm_ops == &shm_vm_ops) && |
1da177e4c
|
1137 |
(vma->vm_start - addr)/PAGE_SIZE == vma->vm_pgoff) { |
496ad9aa8
|
1138 |
size = file_inode(vma->vm_file)->i_size; |
1da177e4c
|
1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 |
do_munmap(mm, vma->vm_start, vma->vm_end - vma->vm_start); /* * We discovered the size of the shm segment, so * break out of here and fall through to the next * loop that uses the size information to stop * searching for matching vma's. */ retval = 0; vma = next; break; } vma = next; } /* * We need look no further than the maximum address a fragment * could possibly have landed at. Also cast things to loff_t to |
25985edce
|
1156 |
* prevent overflows and make comparisons vs. equal-width types. |
1da177e4c
|
1157 |
*/ |
8e36709d8
|
1158 |
size = PAGE_ALIGN(size); |
1da177e4c
|
1159 1160 1161 1162 |
while (vma && (loff_t)(vma->vm_end - addr) <= size) { next = vma->vm_next; /* finding a matching vma now does not alter retval */ |
bc56bba8f
|
1163 |
if ((vma->vm_ops == &shm_vm_ops) && |
1da177e4c
|
1164 1165 1166 1167 1168 |
(vma->vm_start - addr)/PAGE_SIZE == vma->vm_pgoff) do_munmap(mm, vma->vm_start, vma->vm_end - vma->vm_start); vma = next; } |
8feae1311
|
1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 |
#else /* CONFIG_MMU */ /* under NOMMU conditions, the exact address to be destroyed must be * given */ retval = -EINVAL; if (vma->vm_start == addr && vma->vm_ops == &shm_vm_ops) { do_munmap(mm, vma->vm_start, vma->vm_end - vma->vm_start); retval = 0; } #endif |
1da177e4c
|
1179 1180 1181 1182 1183 |
up_write(&mm->mmap_sem); return retval; } #ifdef CONFIG_PROC_FS |
19b4946ca
|
1184 |
static int sysvipc_shm_proc_show(struct seq_file *s, void *it) |
1da177e4c
|
1185 |
{ |
1efdb69b0
|
1186 |
struct user_namespace *user_ns = seq_user_ns(s); |
19b4946ca
|
1187 |
struct shmid_kernel *shp = it; |
b79521807
|
1188 1189 1190 |
unsigned long rss = 0, swp = 0; shm_add_rss_swap(shp, &rss, &swp); |
1da177e4c
|
1191 |
|
6c826818f
|
1192 1193 1194 1195 1196 |
#if BITS_PER_LONG <= 32 #define SIZE_SPEC "%10lu" #else #define SIZE_SPEC "%21lu" #endif |
1da177e4c
|
1197 |
|
6c826818f
|
1198 1199 |
return seq_printf(s, "%10d %10d %4o " SIZE_SPEC " %5u %5u " |
b79521807
|
1200 1201 1202 |
"%5lu %5u %5u %5u %5u %10lu %10lu %10lu " SIZE_SPEC " " SIZE_SPEC " ", |
19b4946ca
|
1203 |
shp->shm_perm.key, |
7ca7e564e
|
1204 |
shp->shm_perm.id, |
b33291c0b
|
1205 |
shp->shm_perm.mode, |
19b4946ca
|
1206 1207 1208 |
shp->shm_segsz, shp->shm_cprid, shp->shm_lprid, |
bc56bba8f
|
1209 |
shp->shm_nattch, |
1efdb69b0
|
1210 1211 1212 1213 |
from_kuid_munged(user_ns, shp->shm_perm.uid), from_kgid_munged(user_ns, shp->shm_perm.gid), from_kuid_munged(user_ns, shp->shm_perm.cuid), from_kgid_munged(user_ns, shp->shm_perm.cgid), |
19b4946ca
|
1214 1215 |
shp->shm_atim, shp->shm_dtim, |
b79521807
|
1216 1217 1218 |
shp->shm_ctim, rss * PAGE_SIZE, swp * PAGE_SIZE); |
1da177e4c
|
1219 1220 |
} #endif |