syscall.c 36.5 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12
/* Copyright (c) 2011-2014 PLUMgrid, http://plumgrid.com
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of version 2 of the GNU General Public
 * License as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
 * General Public License for more details.
 */
#include <linux/bpf.h>
13
#include <linux/bpf_trace.h>
14 15
#include <linux/syscalls.h>
#include <linux/slab.h>
16
#include <linux/sched/signal.h>
17 18
#include <linux/vmalloc.h>
#include <linux/mmzone.h>
19
#include <linux/anon_inodes.h>
20
#include <linux/file.h>
21 22
#include <linux/license.h>
#include <linux/filter.h>
23
#include <linux/version.h>
M
Mickaël Salaün 已提交
24
#include <linux/kernel.h>
M
Martin KaFai Lau 已提交
25
#include <linux/idr.h>
26 27 28
#include <linux/cred.h>
#include <linux/timekeeping.h>
#include <linux/ctype.h>
29

30 31 32 33 34 35 36
#define IS_FD_ARRAY(map) ((map)->map_type == BPF_MAP_TYPE_PROG_ARRAY || \
			   (map)->map_type == BPF_MAP_TYPE_PERF_EVENT_ARRAY || \
			   (map)->map_type == BPF_MAP_TYPE_CGROUP_ARRAY || \
			   (map)->map_type == BPF_MAP_TYPE_ARRAY_OF_MAPS)
#define IS_FD_HASH(map) ((map)->map_type == BPF_MAP_TYPE_HASH_OF_MAPS)
#define IS_FD_MAP(map) (IS_FD_ARRAY(map) || IS_FD_HASH(map))

37
DEFINE_PER_CPU(int, bpf_prog_active);
M
Martin KaFai Lau 已提交
38 39
static DEFINE_IDR(prog_idr);
static DEFINE_SPINLOCK(prog_idr_lock);
M
Martin KaFai Lau 已提交
40 41
static DEFINE_IDR(map_idr);
static DEFINE_SPINLOCK(map_idr_lock);
42

43 44
int sysctl_unprivileged_bpf_disabled __read_mostly;

45 46 47 48 49 50 51 52
static const struct bpf_map_ops * const bpf_map_types[] = {
#define BPF_PROG_TYPE(_id, _ops)
#define BPF_MAP_TYPE(_id, _ops) \
	[_id] = &_ops,
#include <linux/bpf_types.h>
#undef BPF_PROG_TYPE
#undef BPF_MAP_TYPE
};
53

54 55 56 57 58 59 60 61 62
/*
 * If we're handed a bigger struct than we know of, ensure all the unknown bits
 * are 0 - i.e. new user-space does not rely on any kernel feature extensions
 * we don't know about yet.
 *
 * There is a ToCToU between this function call and the following
 * copy_from_user() call. However, this is not a concern since this function is
 * meant to be a future-proofing of bits.
 */
63 64 65 66 67 68 69 70 71
static int check_uarg_tail_zero(void __user *uaddr,
				size_t expected_size,
				size_t actual_size)
{
	unsigned char __user *addr;
	unsigned char __user *end;
	unsigned char val;
	int err;

72 73 74 75 76 77
	if (unlikely(actual_size > PAGE_SIZE))	/* silly large */
		return -E2BIG;

	if (unlikely(!access_ok(VERIFY_READ, uaddr, actual_size)))
		return -EFAULT;

78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94
	if (actual_size <= expected_size)
		return 0;

	addr = uaddr + expected_size;
	end  = uaddr + actual_size;

	for (; addr < end; addr++) {
		err = get_user(val, addr);
		if (err)
			return err;
		if (val)
			return -E2BIG;
	}

	return 0;
}

95 96 97 98
static struct bpf_map *find_and_alloc_map(union bpf_attr *attr)
{
	struct bpf_map *map;

99 100 101
	if (attr->map_type >= ARRAY_SIZE(bpf_map_types) ||
	    !bpf_map_types[attr->map_type])
		return ERR_PTR(-EINVAL);
102

103 104 105 106 107 108
	map = bpf_map_types[attr->map_type]->map_alloc(attr);
	if (IS_ERR(map))
		return map;
	map->ops = bpf_map_types[attr->map_type];
	map->map_type = attr->map_type;
	return map;
109 110
}

111
void *bpf_map_area_alloc(size_t size, int numa_node)
112 113 114 115 116 117 118 119 120
{
	/* We definitely need __GFP_NORETRY, so OOM killer doesn't
	 * trigger under memory pressure as we really just want to
	 * fail instead.
	 */
	const gfp_t flags = __GFP_NOWARN | __GFP_NORETRY | __GFP_ZERO;
	void *area;

	if (size <= (PAGE_SIZE << PAGE_ALLOC_COSTLY_ORDER)) {
121
		area = kmalloc_node(size, GFP_USER | flags, numa_node);
122 123 124 125
		if (area != NULL)
			return area;
	}

126 127
	return __vmalloc_node_flags_caller(size, numa_node, GFP_KERNEL | flags,
					   __builtin_return_address(0));
128 129 130 131 132 133 134
}

void bpf_map_area_free(void *area)
{
	kvfree(area);
}

135 136 137 138 139 140 141 142 143 144 145 146 147
int bpf_map_precharge_memlock(u32 pages)
{
	struct user_struct *user = get_current_user();
	unsigned long memlock_limit, cur;

	memlock_limit = rlimit(RLIMIT_MEMLOCK) >> PAGE_SHIFT;
	cur = atomic_long_read(&user->locked_vm);
	free_uid(user);
	if (cur + pages > memlock_limit)
		return -EPERM;
	return 0;
}

148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173
static int bpf_map_charge_memlock(struct bpf_map *map)
{
	struct user_struct *user = get_current_user();
	unsigned long memlock_limit;

	memlock_limit = rlimit(RLIMIT_MEMLOCK) >> PAGE_SHIFT;

	atomic_long_add(map->pages, &user->locked_vm);

	if (atomic_long_read(&user->locked_vm) > memlock_limit) {
		atomic_long_sub(map->pages, &user->locked_vm);
		free_uid(user);
		return -EPERM;
	}
	map->user = user;
	return 0;
}

static void bpf_map_uncharge_memlock(struct bpf_map *map)
{
	struct user_struct *user = map->user;

	atomic_long_sub(map->pages, &user->locked_vm);
	free_uid(user);
}

M
Martin KaFai Lau 已提交
174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189
static int bpf_map_alloc_id(struct bpf_map *map)
{
	int id;

	spin_lock_bh(&map_idr_lock);
	id = idr_alloc_cyclic(&map_idr, map, 1, INT_MAX, GFP_ATOMIC);
	if (id > 0)
		map->id = id;
	spin_unlock_bh(&map_idr_lock);

	if (WARN_ON_ONCE(!id))
		return -ENOSPC;

	return id > 0 ? 0 : id;
}

M
Martin KaFai Lau 已提交
190
static void bpf_map_free_id(struct bpf_map *map, bool do_idr_lock)
M
Martin KaFai Lau 已提交
191
{
192 193
	unsigned long flags;

M
Martin KaFai Lau 已提交
194
	if (do_idr_lock)
195
		spin_lock_irqsave(&map_idr_lock, flags);
M
Martin KaFai Lau 已提交
196 197 198
	else
		__acquire(&map_idr_lock);

M
Martin KaFai Lau 已提交
199
	idr_remove(&map_idr, map->id);
M
Martin KaFai Lau 已提交
200 201

	if (do_idr_lock)
202
		spin_unlock_irqrestore(&map_idr_lock, flags);
M
Martin KaFai Lau 已提交
203 204
	else
		__release(&map_idr_lock);
M
Martin KaFai Lau 已提交
205 206
}

207 208 209 210 211
/* called from workqueue */
static void bpf_map_free_deferred(struct work_struct *work)
{
	struct bpf_map *map = container_of(work, struct bpf_map, work);

212
	bpf_map_uncharge_memlock(map);
213 214 215 216
	/* implementation dependent freeing */
	map->ops->map_free(map);
}

217 218 219 220 221 222 223 224
static void bpf_map_put_uref(struct bpf_map *map)
{
	if (atomic_dec_and_test(&map->usercnt)) {
		if (map->map_type == BPF_MAP_TYPE_PROG_ARRAY)
			bpf_fd_array_map_clear(map);
	}
}

225 226 227
/* decrement map refcnt and schedule it for freeing via workqueue
 * (unrelying map implementation ops->map_free() might sleep)
 */
M
Martin KaFai Lau 已提交
228
static void __bpf_map_put(struct bpf_map *map, bool do_idr_lock)
229 230
{
	if (atomic_dec_and_test(&map->refcnt)) {
231
		/* bpf_map_free_id() must be called first */
M
Martin KaFai Lau 已提交
232
		bpf_map_free_id(map, do_idr_lock);
233 234 235 236 237
		INIT_WORK(&map->work, bpf_map_free_deferred);
		schedule_work(&map->work);
	}
}

M
Martin KaFai Lau 已提交
238 239 240 241 242
void bpf_map_put(struct bpf_map *map)
{
	__bpf_map_put(map, true);
}

243
void bpf_map_put_with_uref(struct bpf_map *map)
244
{
245
	bpf_map_put_uref(map);
246
	bpf_map_put(map);
247 248 249 250
}

static int bpf_map_release(struct inode *inode, struct file *filp)
{
251 252 253 254 255 256
	struct bpf_map *map = filp->private_data;

	if (map->ops->map_release)
		map->ops->map_release(map, filp);

	bpf_map_put_with_uref(map);
257 258 259
	return 0;
}

260 261 262 263
#ifdef CONFIG_PROC_FS
static void bpf_map_show_fdinfo(struct seq_file *m, struct file *filp)
{
	const struct bpf_map *map = filp->private_data;
264 265
	const struct bpf_array *array;
	u32 owner_prog_type = 0;
266
	u32 owner_jited = 0;
267 268 269 270

	if (map->map_type == BPF_MAP_TYPE_PROG_ARRAY) {
		array = container_of(map, struct bpf_array, map);
		owner_prog_type = array->owner_prog_type;
271
		owner_jited = array->owner_jited;
272
	}
273 274 275 276 277

	seq_printf(m,
		   "map_type:\t%u\n"
		   "key_size:\t%u\n"
		   "value_size:\t%u\n"
278
		   "max_entries:\t%u\n"
279 280
		   "map_flags:\t%#x\n"
		   "memlock:\t%llu\n",
281 282 283
		   map->map_type,
		   map->key_size,
		   map->value_size,
284
		   map->max_entries,
285 286 287
		   map->map_flags,
		   map->pages * 1ULL << PAGE_SHIFT);

288
	if (owner_prog_type) {
289 290
		seq_printf(m, "owner_prog_type:\t%u\n",
			   owner_prog_type);
291 292 293
		seq_printf(m, "owner_jited:\t%u\n",
			   owner_jited);
	}
294 295 296
}
#endif

297
static const struct file_operations bpf_map_fops = {
298 299 300 301
#ifdef CONFIG_PROC_FS
	.show_fdinfo	= bpf_map_show_fdinfo,
#endif
	.release	= bpf_map_release,
302 303
};

304
int bpf_map_new_fd(struct bpf_map *map)
305 306 307 308 309
{
	return anon_inode_getfd("bpf-map", &bpf_map_fops, map,
				O_RDWR | O_CLOEXEC);
}

310 311 312 313 314 315 316 317
/* helper macro to check that unused fields 'union bpf_attr' are zero */
#define CHECK_ATTR(CMD) \
	memchr_inv((void *) &attr->CMD##_LAST_FIELD + \
		   sizeof(attr->CMD##_LAST_FIELD), 0, \
		   sizeof(*attr) - \
		   offsetof(union bpf_attr, CMD##_LAST_FIELD) - \
		   sizeof(attr->CMD##_LAST_FIELD)) != NULL

318 319 320 321 322 323 324
/* dst and src must have at least BPF_OBJ_NAME_LEN number of bytes.
 * Return 0 on success and < 0 on error.
 */
static int bpf_obj_name_cpy(char *dst, const char *src)
{
	const char *end = src + BPF_OBJ_NAME_LEN;

325 326
	memset(dst, 0, BPF_OBJ_NAME_LEN);

327 328 329 330 331 332 333 334 335 336 337 338 339 340
	/* Copy all isalnum() and '_' char */
	while (src < end && *src) {
		if (!isalnum(*src) && *src != '_')
			return -EINVAL;
		*dst++ = *src++;
	}

	/* No '\0' found in BPF_OBJ_NAME_LEN number of bytes */
	if (src == end)
		return -EINVAL;

	return 0;
}

341
#define BPF_MAP_CREATE_LAST_FIELD map_name
342 343 344
/* called via syscall */
static int map_create(union bpf_attr *attr)
{
345
	int numa_node = bpf_map_attr_numa_node(attr);
346 347 348 349 350 351 352
	struct bpf_map *map;
	int err;

	err = CHECK_ATTR(BPF_MAP_CREATE);
	if (err)
		return -EINVAL;

353
	if (numa_node != NUMA_NO_NODE &&
E
Eric Dumazet 已提交
354 355
	    ((unsigned int)numa_node >= nr_node_ids ||
	     !node_online(numa_node)))
356 357
		return -EINVAL;

358 359 360 361 362
	/* find map type and init map: hashtable vs rbtree vs bloom vs ... */
	map = find_and_alloc_map(attr);
	if (IS_ERR(map))
		return PTR_ERR(map);

363 364 365 366
	err = bpf_obj_name_cpy(map->name, attr->map_name);
	if (err)
		goto free_map_nouncharge;

367
	atomic_set(&map->refcnt, 1);
368
	atomic_set(&map->usercnt, 1);
369

370 371
	err = bpf_map_charge_memlock(map);
	if (err)
372
		goto free_map_nouncharge;
373

M
Martin KaFai Lau 已提交
374 375 376 377
	err = bpf_map_alloc_id(map);
	if (err)
		goto free_map;

378
	err = bpf_map_new_fd(map);
M
Martin KaFai Lau 已提交
379 380 381 382 383 384 385 386 387 388
	if (err < 0) {
		/* failed to allocate fd.
		 * bpf_map_put() is needed because the above
		 * bpf_map_alloc_id() has published the map
		 * to the userspace and the userspace may
		 * have refcnt-ed it through BPF_MAP_GET_FD_BY_ID.
		 */
		bpf_map_put(map);
		return err;
	}
389

390
	trace_bpf_map_create(map, err);
391 392 393
	return err;

free_map:
394 395
	bpf_map_uncharge_memlock(map);
free_map_nouncharge:
396 397 398 399
	map->ops->map_free(map);
	return err;
}

400 401 402
/* if error is returned, fd is released.
 * On success caller should complete fd access with matching fdput()
 */
403
struct bpf_map *__bpf_map_get(struct fd f)
404 405 406 407 408 409 410 411
{
	if (!f.file)
		return ERR_PTR(-EBADF);
	if (f.file->f_op != &bpf_map_fops) {
		fdput(f);
		return ERR_PTR(-EINVAL);
	}

412 413 414
	return f.file->private_data;
}

A
Alexei Starovoitov 已提交
415 416 417 418
/* prog's and map's refcnt limit */
#define BPF_MAX_REFCNT 32768

struct bpf_map *bpf_map_inc(struct bpf_map *map, bool uref)
419
{
A
Alexei Starovoitov 已提交
420 421 422 423
	if (atomic_inc_return(&map->refcnt) > BPF_MAX_REFCNT) {
		atomic_dec(&map->refcnt);
		return ERR_PTR(-EBUSY);
	}
424 425
	if (uref)
		atomic_inc(&map->usercnt);
A
Alexei Starovoitov 已提交
426
	return map;
427 428 429
}

struct bpf_map *bpf_map_get_with_uref(u32 ufd)
430 431 432 433 434 435 436 437
{
	struct fd f = fdget(ufd);
	struct bpf_map *map;

	map = __bpf_map_get(f);
	if (IS_ERR(map))
		return map;

A
Alexei Starovoitov 已提交
438
	map = bpf_map_inc(map, true);
439
	fdput(f);
440 441 442 443

	return map;
}

M
Martin KaFai Lau 已提交
444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
/* map_idr_lock should have been held */
static struct bpf_map *bpf_map_inc_not_zero(struct bpf_map *map,
					    bool uref)
{
	int refold;

	refold = __atomic_add_unless(&map->refcnt, 1, 0);

	if (refold >= BPF_MAX_REFCNT) {
		__bpf_map_put(map, false);
		return ERR_PTR(-EBUSY);
	}

	if (!refold)
		return ERR_PTR(-ENOENT);

	if (uref)
		atomic_inc(&map->usercnt);

	return map;
}

466 467 468 469 470
int __weak bpf_stackmap_copy(struct bpf_map *map, void *key, void *value)
{
	return -ENOTSUPP;
}

471 472 473 474 475
/* last field in 'union bpf_attr' used by this command */
#define BPF_MAP_LOOKUP_ELEM_LAST_FIELD value

static int map_lookup_elem(union bpf_attr *attr)
{
M
Mickaël Salaün 已提交
476 477
	void __user *ukey = u64_to_user_ptr(attr->key);
	void __user *uvalue = u64_to_user_ptr(attr->value);
478 479
	int ufd = attr->map_fd;
	struct bpf_map *map;
480
	void *key, *value, *ptr;
481
	u32 value_size;
482
	struct fd f;
483 484 485 486 487
	int err;

	if (CHECK_ATTR(BPF_MAP_LOOKUP_ELEM))
		return -EINVAL;

488
	f = fdget(ufd);
489
	map = __bpf_map_get(f);
490 491 492
	if (IS_ERR(map))
		return PTR_ERR(map);

A
Al Viro 已提交
493 494 495
	key = memdup_user(ukey, map->key_size);
	if (IS_ERR(key)) {
		err = PTR_ERR(key);
496
		goto err_put;
A
Al Viro 已提交
497
	}
498

499
	if (map->map_type == BPF_MAP_TYPE_PERCPU_HASH ||
500
	    map->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH ||
501 502
	    map->map_type == BPF_MAP_TYPE_PERCPU_ARRAY)
		value_size = round_up(map->value_size, 8) * num_possible_cpus();
503 504
	else if (IS_FD_MAP(map))
		value_size = sizeof(u32);
505 506 507
	else
		value_size = map->value_size;

508
	err = -ENOMEM;
509
	value = kmalloc(value_size, GFP_USER | __GFP_NOWARN);
510
	if (!value)
511 512
		goto free_key;

513 514
	if (map->map_type == BPF_MAP_TYPE_PERCPU_HASH ||
	    map->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH) {
515 516 517
		err = bpf_percpu_hash_copy(map, key, value);
	} else if (map->map_type == BPF_MAP_TYPE_PERCPU_ARRAY) {
		err = bpf_percpu_array_copy(map, key, value);
518 519
	} else if (map->map_type == BPF_MAP_TYPE_STACK_TRACE) {
		err = bpf_stackmap_copy(map, key, value);
520 521 522 523
	} else if (IS_FD_ARRAY(map)) {
		err = bpf_fd_array_map_lookup_elem(map, key, value);
	} else if (IS_FD_HASH(map)) {
		err = bpf_fd_htab_map_lookup_elem(map, key, value);
524 525 526 527 528 529 530 531
	} else {
		rcu_read_lock();
		ptr = map->ops->map_lookup_elem(map, key);
		if (ptr)
			memcpy(value, ptr, value_size);
		rcu_read_unlock();
		err = ptr ? 0 : -ENOENT;
	}
532

533
	if (err)
534
		goto free_value;
535 536

	err = -EFAULT;
537
	if (copy_to_user(uvalue, value, value_size) != 0)
538
		goto free_value;
539

540
	trace_bpf_map_lookup_elem(map, ufd, key, value);
541 542
	err = 0;

543 544
free_value:
	kfree(value);
545 546 547 548 549 550 551
free_key:
	kfree(key);
err_put:
	fdput(f);
	return err;
}

552
#define BPF_MAP_UPDATE_ELEM_LAST_FIELD flags
553 554 555

static int map_update_elem(union bpf_attr *attr)
{
M
Mickaël Salaün 已提交
556 557
	void __user *ukey = u64_to_user_ptr(attr->key);
	void __user *uvalue = u64_to_user_ptr(attr->value);
558 559 560
	int ufd = attr->map_fd;
	struct bpf_map *map;
	void *key, *value;
561
	u32 value_size;
562
	struct fd f;
563 564 565 566 567
	int err;

	if (CHECK_ATTR(BPF_MAP_UPDATE_ELEM))
		return -EINVAL;

568
	f = fdget(ufd);
569
	map = __bpf_map_get(f);
570 571 572
	if (IS_ERR(map))
		return PTR_ERR(map);

A
Al Viro 已提交
573 574 575
	key = memdup_user(ukey, map->key_size);
	if (IS_ERR(key)) {
		err = PTR_ERR(key);
576
		goto err_put;
A
Al Viro 已提交
577
	}
578

579
	if (map->map_type == BPF_MAP_TYPE_PERCPU_HASH ||
580
	    map->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH ||
581 582 583 584 585
	    map->map_type == BPF_MAP_TYPE_PERCPU_ARRAY)
		value_size = round_up(map->value_size, 8) * num_possible_cpus();
	else
		value_size = map->value_size;

586
	err = -ENOMEM;
587
	value = kmalloc(value_size, GFP_USER | __GFP_NOWARN);
588 589 590 591
	if (!value)
		goto free_key;

	err = -EFAULT;
592
	if (copy_from_user(value, uvalue, value_size) != 0)
593 594
		goto free_value;

595 596 597 598 599 600
	/* Need to create a kthread, thus must support schedule */
	if (map->map_type == BPF_MAP_TYPE_CPUMAP) {
		err = map->ops->map_update_elem(map, key, value, attr->flags);
		goto out;
	}

601 602 603 604 605
	/* must increment bpf_prog_active to avoid kprobe+bpf triggering from
	 * inside bpf map update or delete otherwise deadlocks are possible
	 */
	preempt_disable();
	__this_cpu_inc(bpf_prog_active);
606 607
	if (map->map_type == BPF_MAP_TYPE_PERCPU_HASH ||
	    map->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH) {
608 609 610
		err = bpf_percpu_hash_update(map, key, value, attr->flags);
	} else if (map->map_type == BPF_MAP_TYPE_PERCPU_ARRAY) {
		err = bpf_percpu_array_update(map, key, value, attr->flags);
611
	} else if (map->map_type == BPF_MAP_TYPE_PERF_EVENT_ARRAY ||
612
		   map->map_type == BPF_MAP_TYPE_PROG_ARRAY ||
613 614
		   map->map_type == BPF_MAP_TYPE_CGROUP_ARRAY ||
		   map->map_type == BPF_MAP_TYPE_ARRAY_OF_MAPS) {
615 616 617 618
		rcu_read_lock();
		err = bpf_fd_array_map_update_elem(map, f.file, key, value,
						   attr->flags);
		rcu_read_unlock();
M
Martin KaFai Lau 已提交
619 620 621 622 623
	} else if (map->map_type == BPF_MAP_TYPE_HASH_OF_MAPS) {
		rcu_read_lock();
		err = bpf_fd_htab_map_update_elem(map, f.file, key, value,
						  attr->flags);
		rcu_read_unlock();
624 625 626 627 628
	} else {
		rcu_read_lock();
		err = map->ops->map_update_elem(map, key, value, attr->flags);
		rcu_read_unlock();
	}
629 630
	__this_cpu_dec(bpf_prog_active);
	preempt_enable();
631
out:
632 633
	if (!err)
		trace_bpf_map_update_elem(map, ufd, key, value);
634 635 636 637 638 639 640 641 642 643 644 645 646
free_value:
	kfree(value);
free_key:
	kfree(key);
err_put:
	fdput(f);
	return err;
}

#define BPF_MAP_DELETE_ELEM_LAST_FIELD key

static int map_delete_elem(union bpf_attr *attr)
{
M
Mickaël Salaün 已提交
647
	void __user *ukey = u64_to_user_ptr(attr->key);
648 649
	int ufd = attr->map_fd;
	struct bpf_map *map;
650
	struct fd f;
651 652 653 654 655 656
	void *key;
	int err;

	if (CHECK_ATTR(BPF_MAP_DELETE_ELEM))
		return -EINVAL;

657
	f = fdget(ufd);
658
	map = __bpf_map_get(f);
659 660 661
	if (IS_ERR(map))
		return PTR_ERR(map);

A
Al Viro 已提交
662 663 664
	key = memdup_user(ukey, map->key_size);
	if (IS_ERR(key)) {
		err = PTR_ERR(key);
665
		goto err_put;
A
Al Viro 已提交
666
	}
667

668 669
	preempt_disable();
	__this_cpu_inc(bpf_prog_active);
670 671 672
	rcu_read_lock();
	err = map->ops->map_delete_elem(map, key);
	rcu_read_unlock();
673 674
	__this_cpu_dec(bpf_prog_active);
	preempt_enable();
675

676 677
	if (!err)
		trace_bpf_map_delete_elem(map, ufd, key);
678 679 680 681 682 683 684 685 686 687 688
	kfree(key);
err_put:
	fdput(f);
	return err;
}

/* last field in 'union bpf_attr' used by this command */
#define BPF_MAP_GET_NEXT_KEY_LAST_FIELD next_key

static int map_get_next_key(union bpf_attr *attr)
{
M
Mickaël Salaün 已提交
689 690
	void __user *ukey = u64_to_user_ptr(attr->key);
	void __user *unext_key = u64_to_user_ptr(attr->next_key);
691 692 693
	int ufd = attr->map_fd;
	struct bpf_map *map;
	void *key, *next_key;
694
	struct fd f;
695 696 697 698 699
	int err;

	if (CHECK_ATTR(BPF_MAP_GET_NEXT_KEY))
		return -EINVAL;

700
	f = fdget(ufd);
701
	map = __bpf_map_get(f);
702 703 704
	if (IS_ERR(map))
		return PTR_ERR(map);

705
	if (ukey) {
A
Al Viro 已提交
706 707 708
		key = memdup_user(ukey, map->key_size);
		if (IS_ERR(key)) {
			err = PTR_ERR(key);
709
			goto err_put;
A
Al Viro 已提交
710
		}
711 712 713
	} else {
		key = NULL;
	}
714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729

	err = -ENOMEM;
	next_key = kmalloc(map->key_size, GFP_USER);
	if (!next_key)
		goto free_key;

	rcu_read_lock();
	err = map->ops->map_get_next_key(map, key, next_key);
	rcu_read_unlock();
	if (err)
		goto free_next_key;

	err = -EFAULT;
	if (copy_to_user(unext_key, next_key, map->key_size) != 0)
		goto free_next_key;

730
	trace_bpf_map_next_key(map, ufd, key, next_key);
731 732 733 734 735 736 737 738 739 740 741
	err = 0;

free_next_key:
	kfree(next_key);
free_key:
	kfree(key);
err_put:
	fdput(f);
	return err;
}

742 743 744 745 746 747 748 749 750 751 752 753
static const struct bpf_prog_ops * const bpf_prog_types[] = {
#define BPF_PROG_TYPE(_id, _name) \
	[_id] = & _name ## _prog_ops,
#define BPF_MAP_TYPE(_id, _ops)
#include <linux/bpf_types.h>
#undef BPF_PROG_TYPE
#undef BPF_MAP_TYPE
};

static const struct bpf_verifier_ops * const bpf_verifier_ops[] = {
#define BPF_PROG_TYPE(_id, _name) \
	[_id] = & _name ## _verifier_ops,
754
#define BPF_MAP_TYPE(_id, _ops)
755 756
#include <linux/bpf_types.h>
#undef BPF_PROG_TYPE
757
#undef BPF_MAP_TYPE
758
};
759 760 761

static int find_prog_type(enum bpf_prog_type type, struct bpf_prog *prog)
{
762 763
	if (type >= ARRAY_SIZE(bpf_prog_types) || !bpf_prog_types[type])
		return -EINVAL;
764

765
	prog->aux->ops = bpf_prog_types[type];
766
	prog->aux->vops = bpf_verifier_ops[type];
767 768
	prog->type = type;
	return 0;
769 770 771 772 773 774 775 776 777 778 779 780 781
}

/* drop refcnt on maps used by eBPF program and free auxilary data */
static void free_used_maps(struct bpf_prog_aux *aux)
{
	int i;

	for (i = 0; i < aux->used_map_cnt; i++)
		bpf_map_put(aux->used_maps[i]);

	kfree(aux->used_maps);
}

782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803
int __bpf_prog_charge(struct user_struct *user, u32 pages)
{
	unsigned long memlock_limit = rlimit(RLIMIT_MEMLOCK) >> PAGE_SHIFT;
	unsigned long user_bufs;

	if (user) {
		user_bufs = atomic_long_add_return(pages, &user->locked_vm);
		if (user_bufs > memlock_limit) {
			atomic_long_sub(pages, &user->locked_vm);
			return -EPERM;
		}
	}

	return 0;
}

void __bpf_prog_uncharge(struct user_struct *user, u32 pages)
{
	if (user)
		atomic_long_sub(pages, &user->locked_vm);
}

804 805 806
static int bpf_prog_charge_memlock(struct bpf_prog *prog)
{
	struct user_struct *user = get_current_user();
807
	int ret;
808

809 810
	ret = __bpf_prog_charge(user, prog->pages);
	if (ret) {
811
		free_uid(user);
812
		return ret;
813
	}
814

815 816 817 818 819 820 821 822
	prog->aux->user = user;
	return 0;
}

static void bpf_prog_uncharge_memlock(struct bpf_prog *prog)
{
	struct user_struct *user = prog->aux->user;

823
	__bpf_prog_uncharge(user, prog->pages);
824 825 826
	free_uid(user);
}

M
Martin KaFai Lau 已提交
827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843
static int bpf_prog_alloc_id(struct bpf_prog *prog)
{
	int id;

	spin_lock_bh(&prog_idr_lock);
	id = idr_alloc_cyclic(&prog_idr, prog, 1, INT_MAX, GFP_ATOMIC);
	if (id > 0)
		prog->aux->id = id;
	spin_unlock_bh(&prog_idr_lock);

	/* id is in [1, INT_MAX) */
	if (WARN_ON_ONCE(!id))
		return -ENOSPC;

	return id > 0 ? 0 : id;
}

844
static void bpf_prog_free_id(struct bpf_prog *prog, bool do_idr_lock)
M
Martin KaFai Lau 已提交
845 846 847 848 849
{
	/* cBPF to eBPF migrations are currently not in the idr store. */
	if (!prog->aux->id)
		return;

850 851 852 853 854
	if (do_idr_lock)
		spin_lock_bh(&prog_idr_lock);
	else
		__acquire(&prog_idr_lock);

M
Martin KaFai Lau 已提交
855
	idr_remove(&prog_idr, prog->aux->id);
856 857 858 859 860

	if (do_idr_lock)
		spin_unlock_bh(&prog_idr_lock);
	else
		__release(&prog_idr_lock);
M
Martin KaFai Lau 已提交
861 862
}

863
static void __bpf_prog_put_rcu(struct rcu_head *rcu)
864 865 866 867
{
	struct bpf_prog_aux *aux = container_of(rcu, struct bpf_prog_aux, rcu);

	free_used_maps(aux);
868
	bpf_prog_uncharge_memlock(aux->prog);
869 870 871
	bpf_prog_free(aux->prog);
}

872
static void __bpf_prog_put(struct bpf_prog *prog, bool do_idr_lock)
873
{
874 875
	if (atomic_dec_and_test(&prog->aux->refcnt)) {
		trace_bpf_prog_put_rcu(prog);
876
		/* bpf_prog_free_id() must be called first */
877
		bpf_prog_free_id(prog, do_idr_lock);
878
		bpf_prog_kallsyms_del(prog);
879
		call_rcu(&prog->aux->rcu, __bpf_prog_put_rcu);
880
	}
881
}
882 883 884 885 886

void bpf_prog_put(struct bpf_prog *prog)
{
	__bpf_prog_put(prog, true);
}
887
EXPORT_SYMBOL_GPL(bpf_prog_put);
888 889 890 891 892

static int bpf_prog_release(struct inode *inode, struct file *filp)
{
	struct bpf_prog *prog = filp->private_data;

893
	bpf_prog_put(prog);
894 895 896
	return 0;
}

897 898 899 900
#ifdef CONFIG_PROC_FS
static void bpf_prog_show_fdinfo(struct seq_file *m, struct file *filp)
{
	const struct bpf_prog *prog = filp->private_data;
901
	char prog_tag[sizeof(prog->tag) * 2 + 1] = { };
902

903
	bin2hex(prog_tag, prog->tag, sizeof(prog->tag));
904 905 906
	seq_printf(m,
		   "prog_type:\t%u\n"
		   "prog_jited:\t%u\n"
907
		   "prog_tag:\t%s\n"
908 909 910
		   "memlock:\t%llu\n",
		   prog->type,
		   prog->jited,
911
		   prog_tag,
912 913 914 915
		   prog->pages * 1ULL << PAGE_SHIFT);
}
#endif

916
static const struct file_operations bpf_prog_fops = {
917 918 919 920
#ifdef CONFIG_PROC_FS
	.show_fdinfo	= bpf_prog_show_fdinfo,
#endif
	.release	= bpf_prog_release,
921 922
};

923
int bpf_prog_new_fd(struct bpf_prog *prog)
924 925 926 927 928
{
	return anon_inode_getfd("bpf-prog", &bpf_prog_fops, prog,
				O_RDWR | O_CLOEXEC);
}

929
static struct bpf_prog *____bpf_prog_get(struct fd f)
930 931 932 933 934 935 936 937
{
	if (!f.file)
		return ERR_PTR(-EBADF);
	if (f.file->f_op != &bpf_prog_fops) {
		fdput(f);
		return ERR_PTR(-EINVAL);
	}

938
	return f.file->private_data;
939 940
}

941
struct bpf_prog *bpf_prog_add(struct bpf_prog *prog, int i)
A
Alexei Starovoitov 已提交
942
{
943 944
	if (atomic_add_return(i, &prog->aux->refcnt) > BPF_MAX_REFCNT) {
		atomic_sub(i, &prog->aux->refcnt);
A
Alexei Starovoitov 已提交
945 946 947 948
		return ERR_PTR(-EBUSY);
	}
	return prog;
}
949 950
EXPORT_SYMBOL_GPL(bpf_prog_add);

951 952 953 954 955 956 957 958 959 960 961
void bpf_prog_sub(struct bpf_prog *prog, int i)
{
	/* Only to be used for undoing previous bpf_prog_add() in some
	 * error path. We still know that another entity in our call
	 * path holds a reference to the program, thus atomic_sub() can
	 * be safely used in such cases!
	 */
	WARN_ON(atomic_sub_return(i, &prog->aux->refcnt) == 0);
}
EXPORT_SYMBOL_GPL(bpf_prog_sub);

962 963 964 965
struct bpf_prog *bpf_prog_inc(struct bpf_prog *prog)
{
	return bpf_prog_add(prog, 1);
}
966
EXPORT_SYMBOL_GPL(bpf_prog_inc);
A
Alexei Starovoitov 已提交
967

968
/* prog_idr_lock should have been held */
969
struct bpf_prog *bpf_prog_inc_not_zero(struct bpf_prog *prog)
970 971 972 973 974 975 976 977 978 979 980 981 982 983 984
{
	int refold;

	refold = __atomic_add_unless(&prog->aux->refcnt, 1, 0);

	if (refold >= BPF_MAX_REFCNT) {
		__bpf_prog_put(prog, false);
		return ERR_PTR(-EBUSY);
	}

	if (!refold)
		return ERR_PTR(-ENOENT);

	return prog;
}
985
EXPORT_SYMBOL_GPL(bpf_prog_inc_not_zero);
986

987
static struct bpf_prog *__bpf_prog_get(u32 ufd, enum bpf_prog_type *type)
988 989 990 991
{
	struct fd f = fdget(ufd);
	struct bpf_prog *prog;

992
	prog = ____bpf_prog_get(f);
993 994
	if (IS_ERR(prog))
		return prog;
995 996 997 998
	if (type && prog->type != *type) {
		prog = ERR_PTR(-EINVAL);
		goto out;
	}
999

A
Alexei Starovoitov 已提交
1000
	prog = bpf_prog_inc(prog);
1001
out:
1002 1003 1004
	fdput(f);
	return prog;
}
1005 1006 1007 1008 1009 1010 1011 1012

struct bpf_prog *bpf_prog_get(u32 ufd)
{
	return __bpf_prog_get(ufd, NULL);
}

struct bpf_prog *bpf_prog_get_type(u32 ufd, enum bpf_prog_type type)
{
1013 1014 1015 1016 1017
	struct bpf_prog *prog = __bpf_prog_get(ufd, &type);

	if (!IS_ERR(prog))
		trace_bpf_prog_get_type(prog);
	return prog;
1018 1019
}
EXPORT_SYMBOL_GPL(bpf_prog_get_type);
1020 1021

/* last field in 'union bpf_attr' used by this command */
1022
#define	BPF_PROG_LOAD_LAST_FIELD prog_name
1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034

static int bpf_prog_load(union bpf_attr *attr)
{
	enum bpf_prog_type type = attr->prog_type;
	struct bpf_prog *prog;
	int err;
	char license[128];
	bool is_gpl;

	if (CHECK_ATTR(BPF_PROG_LOAD))
		return -EINVAL;

1035 1036 1037
	if (attr->prog_flags & ~BPF_F_STRICT_ALIGNMENT)
		return -EINVAL;

1038
	/* copy eBPF program license from user space */
M
Mickaël Salaün 已提交
1039
	if (strncpy_from_user(license, u64_to_user_ptr(attr->license),
1040 1041 1042 1043 1044 1045 1046
			      sizeof(license) - 1) < 0)
		return -EFAULT;
	license[sizeof(license) - 1] = 0;

	/* eBPF programs must be GPL compatible to use GPL-ed functions */
	is_gpl = license_is_gpl_compatible(license);

1047 1048
	if (attr->insn_cnt == 0 || attr->insn_cnt > BPF_MAXINSNS)
		return -E2BIG;
1049

1050 1051 1052 1053
	if (type == BPF_PROG_TYPE_KPROBE &&
	    attr->kern_version != LINUX_VERSION_CODE)
		return -EINVAL;

1054 1055 1056
	if (type != BPF_PROG_TYPE_SOCKET_FILTER &&
	    type != BPF_PROG_TYPE_CGROUP_SKB &&
	    !capable(CAP_SYS_ADMIN))
1057 1058
		return -EPERM;

1059 1060 1061 1062 1063
	/* plain bpf_prog allocation */
	prog = bpf_prog_alloc(bpf_prog_size(attr->insn_cnt), GFP_USER);
	if (!prog)
		return -ENOMEM;

1064 1065 1066 1067
	err = bpf_prog_charge_memlock(prog);
	if (err)
		goto free_prog_nouncharge;

1068 1069 1070
	prog->len = attr->insn_cnt;

	err = -EFAULT;
M
Mickaël Salaün 已提交
1071
	if (copy_from_user(prog->insns, u64_to_user_ptr(attr->insns),
1072
			   bpf_prog_insn_size(prog)) != 0)
1073 1074 1075
		goto free_prog;

	prog->orig_prog = NULL;
1076
	prog->jited = 0;
1077 1078

	atomic_set(&prog->aux->refcnt, 1);
1079
	prog->gpl_compatible = is_gpl ? 1 : 0;
1080 1081 1082 1083 1084 1085

	/* find program type: socket_filter vs tracing_filter */
	err = find_prog_type(type, prog);
	if (err < 0)
		goto free_prog;

1086 1087 1088 1089 1090
	prog->aux->load_time = ktime_get_boot_ns();
	err = bpf_obj_name_cpy(prog->aux->name, attr->prog_name);
	if (err)
		goto free_prog;

1091
	/* run eBPF verifier */
1092
	err = bpf_check(&prog, attr);
1093 1094 1095 1096
	if (err < 0)
		goto free_used_maps;

	/* eBPF program is ready to be JITed */
1097
	prog = bpf_prog_select_runtime(prog, &err);
1098 1099
	if (err < 0)
		goto free_used_maps;
1100

M
Martin KaFai Lau 已提交
1101 1102 1103 1104
	err = bpf_prog_alloc_id(prog);
	if (err)
		goto free_used_maps;

1105
	err = bpf_prog_new_fd(prog);
1106 1107 1108 1109 1110 1111 1112 1113 1114 1115
	if (err < 0) {
		/* failed to allocate fd.
		 * bpf_prog_put() is needed because the above
		 * bpf_prog_alloc_id() has published the prog
		 * to the userspace and the userspace may
		 * have refcnt-ed it through BPF_PROG_GET_FD_BY_ID.
		 */
		bpf_prog_put(prog);
		return err;
	}
1116

1117
	bpf_prog_kallsyms_add(prog);
1118
	trace_bpf_prog_load(prog, err);
1119 1120 1121 1122 1123
	return err;

free_used_maps:
	free_used_maps(prog->aux);
free_prog:
1124 1125
	bpf_prog_uncharge_memlock(prog);
free_prog_nouncharge:
1126 1127 1128 1129
	bpf_prog_free(prog);
	return err;
}

1130 1131 1132 1133 1134 1135 1136
#define BPF_OBJ_LAST_FIELD bpf_fd

static int bpf_obj_pin(const union bpf_attr *attr)
{
	if (CHECK_ATTR(BPF_OBJ))
		return -EINVAL;

M
Mickaël Salaün 已提交
1137
	return bpf_obj_pin_user(attr->bpf_fd, u64_to_user_ptr(attr->pathname));
1138 1139 1140 1141 1142 1143 1144
}

static int bpf_obj_get(const union bpf_attr *attr)
{
	if (CHECK_ATTR(BPF_OBJ) || attr->bpf_fd != 0)
		return -EINVAL;

M
Mickaël Salaün 已提交
1145
	return bpf_obj_get_user(u64_to_user_ptr(attr->pathname));
1146 1147
}

1148 1149
#ifdef CONFIG_CGROUP_BPF

1150
#define BPF_PROG_ATTACH_LAST_FIELD attach_flags
1151

1152
static int sockmap_get_from_fd(const union bpf_attr *attr, bool attach)
1153
{
1154
	struct bpf_prog *prog = NULL;
1155 1156 1157 1158 1159 1160 1161 1162 1163 1164
	int ufd = attr->target_fd;
	struct bpf_map *map;
	struct fd f;
	int err;

	f = fdget(ufd);
	map = __bpf_map_get(f);
	if (IS_ERR(map))
		return PTR_ERR(map);

1165 1166 1167 1168 1169 1170 1171
	if (attach) {
		prog = bpf_prog_get_type(attr->attach_bpf_fd,
					 BPF_PROG_TYPE_SK_SKB);
		if (IS_ERR(prog)) {
			fdput(f);
			return PTR_ERR(prog);
		}
1172 1173
	}

1174
	err = sock_map_prog(map, prog, attr->attach_type);
1175 1176
	if (err) {
		fdput(f);
1177 1178
		if (prog)
			bpf_prog_put(prog);
1179
		return err;
1180 1181 1182
	}

	fdput(f);
1183
	return 0;
1184
}
1185

1186 1187 1188
#define BPF_F_ATTACH_MASK \
	(BPF_F_ALLOW_OVERRIDE | BPF_F_ALLOW_MULTI)

1189 1190
static int bpf_prog_attach(const union bpf_attr *attr)
{
1191
	enum bpf_prog_type ptype;
1192 1193
	struct bpf_prog *prog;
	struct cgroup *cgrp;
1194
	int ret;
1195 1196 1197 1198 1199 1200 1201

	if (!capable(CAP_NET_ADMIN))
		return -EPERM;

	if (CHECK_ATTR(BPF_PROG_ATTACH))
		return -EINVAL;

1202
	if (attr->attach_flags & ~BPF_F_ATTACH_MASK)
1203 1204
		return -EINVAL;

1205 1206 1207
	switch (attr->attach_type) {
	case BPF_CGROUP_INET_INGRESS:
	case BPF_CGROUP_INET_EGRESS:
1208
		ptype = BPF_PROG_TYPE_CGROUP_SKB;
1209
		break;
1210 1211 1212
	case BPF_CGROUP_INET_SOCK_CREATE:
		ptype = BPF_PROG_TYPE_CGROUP_SOCK;
		break;
L
Lawrence Brakmo 已提交
1213 1214 1215
	case BPF_CGROUP_SOCK_OPS:
		ptype = BPF_PROG_TYPE_SOCK_OPS;
		break;
1216 1217
	case BPF_SK_SKB_STREAM_PARSER:
	case BPF_SK_SKB_STREAM_VERDICT:
1218
		return sockmap_get_from_fd(attr, true);
1219 1220 1221 1222
	default:
		return -EINVAL;
	}

1223 1224 1225 1226 1227 1228 1229 1230 1231 1232
	prog = bpf_prog_get_type(attr->attach_bpf_fd, ptype);
	if (IS_ERR(prog))
		return PTR_ERR(prog);

	cgrp = cgroup_get_from_fd(attr->target_fd);
	if (IS_ERR(cgrp)) {
		bpf_prog_put(prog);
		return PTR_ERR(cgrp);
	}

1233 1234
	ret = cgroup_bpf_attach(cgrp, prog, attr->attach_type,
				attr->attach_flags);
1235 1236
	if (ret)
		bpf_prog_put(prog);
1237 1238
	cgroup_put(cgrp);

1239
	return ret;
1240 1241 1242 1243 1244 1245
}

#define BPF_PROG_DETACH_LAST_FIELD attach_type

static int bpf_prog_detach(const union bpf_attr *attr)
{
1246 1247
	enum bpf_prog_type ptype;
	struct bpf_prog *prog;
1248
	struct cgroup *cgrp;
1249
	int ret;
1250 1251 1252 1253 1254 1255 1256 1257 1258 1259

	if (!capable(CAP_NET_ADMIN))
		return -EPERM;

	if (CHECK_ATTR(BPF_PROG_DETACH))
		return -EINVAL;

	switch (attr->attach_type) {
	case BPF_CGROUP_INET_INGRESS:
	case BPF_CGROUP_INET_EGRESS:
1260 1261
		ptype = BPF_PROG_TYPE_CGROUP_SKB;
		break;
1262
	case BPF_CGROUP_INET_SOCK_CREATE:
1263 1264
		ptype = BPF_PROG_TYPE_CGROUP_SOCK;
		break;
L
Lawrence Brakmo 已提交
1265
	case BPF_CGROUP_SOCK_OPS:
1266
		ptype = BPF_PROG_TYPE_SOCK_OPS;
1267
		break;
1268 1269
	case BPF_SK_SKB_STREAM_PARSER:
	case BPF_SK_SKB_STREAM_VERDICT:
1270
		return sockmap_get_from_fd(attr, false);
1271 1272 1273 1274
	default:
		return -EINVAL;
	}

1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286
	cgrp = cgroup_get_from_fd(attr->target_fd);
	if (IS_ERR(cgrp))
		return PTR_ERR(cgrp);

	prog = bpf_prog_get_type(attr->attach_bpf_fd, ptype);
	if (IS_ERR(prog))
		prog = NULL;

	ret = cgroup_bpf_detach(cgrp, prog, attr->attach_type, 0);
	if (prog)
		bpf_prog_put(prog);
	cgroup_put(cgrp);
1287
	return ret;
1288
}
L
Lawrence Brakmo 已提交
1289

1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320
#define BPF_PROG_QUERY_LAST_FIELD query.prog_cnt

static int bpf_prog_query(const union bpf_attr *attr,
			  union bpf_attr __user *uattr)
{
	struct cgroup *cgrp;
	int ret;

	if (!capable(CAP_NET_ADMIN))
		return -EPERM;
	if (CHECK_ATTR(BPF_PROG_QUERY))
		return -EINVAL;
	if (attr->query.query_flags & ~BPF_F_QUERY_EFFECTIVE)
		return -EINVAL;

	switch (attr->query.attach_type) {
	case BPF_CGROUP_INET_INGRESS:
	case BPF_CGROUP_INET_EGRESS:
	case BPF_CGROUP_INET_SOCK_CREATE:
	case BPF_CGROUP_SOCK_OPS:
		break;
	default:
		return -EINVAL;
	}
	cgrp = cgroup_get_from_fd(attr->query.target_fd);
	if (IS_ERR(cgrp))
		return PTR_ERR(cgrp);
	ret = cgroup_bpf_query(cgrp, attr, uattr);
	cgroup_put(cgrp);
	return ret;
}
1321 1322
#endif /* CONFIG_CGROUP_BPF */

1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344
#define BPF_PROG_TEST_RUN_LAST_FIELD test.duration

static int bpf_prog_test_run(const union bpf_attr *attr,
			     union bpf_attr __user *uattr)
{
	struct bpf_prog *prog;
	int ret = -ENOTSUPP;

	if (CHECK_ATTR(BPF_PROG_TEST_RUN))
		return -EINVAL;

	prog = bpf_prog_get(attr->test.prog_fd);
	if (IS_ERR(prog))
		return PTR_ERR(prog);

	if (prog->aux->ops->test_run)
		ret = prog->aux->ops->test_run(prog, attr, uattr);

	bpf_prog_put(prog);
	return ret;
}

1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372
#define BPF_OBJ_GET_NEXT_ID_LAST_FIELD next_id

static int bpf_obj_get_next_id(const union bpf_attr *attr,
			       union bpf_attr __user *uattr,
			       struct idr *idr,
			       spinlock_t *lock)
{
	u32 next_id = attr->start_id;
	int err = 0;

	if (CHECK_ATTR(BPF_OBJ_GET_NEXT_ID) || next_id >= INT_MAX)
		return -EINVAL;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	next_id++;
	spin_lock_bh(lock);
	if (!idr_get_next(idr, &next_id))
		err = -ENOENT;
	spin_unlock_bh(lock);

	if (!err)
		err = put_user(next_id, &uattr->next_id);

	return err;
}

1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404
#define BPF_PROG_GET_FD_BY_ID_LAST_FIELD prog_id

static int bpf_prog_get_fd_by_id(const union bpf_attr *attr)
{
	struct bpf_prog *prog;
	u32 id = attr->prog_id;
	int fd;

	if (CHECK_ATTR(BPF_PROG_GET_FD_BY_ID))
		return -EINVAL;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	spin_lock_bh(&prog_idr_lock);
	prog = idr_find(&prog_idr, id);
	if (prog)
		prog = bpf_prog_inc_not_zero(prog);
	else
		prog = ERR_PTR(-ENOENT);
	spin_unlock_bh(&prog_idr_lock);

	if (IS_ERR(prog))
		return PTR_ERR(prog);

	fd = bpf_prog_new_fd(prog);
	if (fd < 0)
		bpf_prog_put(prog);

	return fd;
}

M
Martin KaFai Lau 已提交
1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436
#define BPF_MAP_GET_FD_BY_ID_LAST_FIELD map_id

static int bpf_map_get_fd_by_id(const union bpf_attr *attr)
{
	struct bpf_map *map;
	u32 id = attr->map_id;
	int fd;

	if (CHECK_ATTR(BPF_MAP_GET_FD_BY_ID))
		return -EINVAL;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	spin_lock_bh(&map_idr_lock);
	map = idr_find(&map_idr, id);
	if (map)
		map = bpf_map_inc_not_zero(map, true);
	else
		map = ERR_PTR(-ENOENT);
	spin_unlock_bh(&map_idr_lock);

	if (IS_ERR(map))
		return PTR_ERR(map);

	fd = bpf_map_new_fd(map);
	if (fd < 0)
		bpf_map_put(map);

	return fd;
}

1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453
static int bpf_prog_get_info_by_fd(struct bpf_prog *prog,
				   const union bpf_attr *attr,
				   union bpf_attr __user *uattr)
{
	struct bpf_prog_info __user *uinfo = u64_to_user_ptr(attr->info.info);
	struct bpf_prog_info info = {};
	u32 info_len = attr->info.info_len;
	char __user *uinsns;
	u32 ulen;
	int err;

	err = check_uarg_tail_zero(uinfo, sizeof(info), info_len);
	if (err)
		return err;
	info_len = min_t(u32, sizeof(info), info_len);

	if (copy_from_user(&info, uinfo, info_len))
1454
		return -EFAULT;
1455 1456 1457

	info.type = prog->type;
	info.id = prog->aux->id;
1458 1459 1460
	info.load_time = prog->aux->load_time;
	info.created_by_uid = from_kuid_munged(current_user_ns(),
					       prog->aux->user->uid);
1461 1462

	memcpy(info.tag, prog->tag, sizeof(prog->tag));
1463 1464 1465 1466 1467 1468
	memcpy(info.name, prog->aux->name, sizeof(prog->aux->name));

	ulen = info.nr_map_ids;
	info.nr_map_ids = prog->aux->used_map_cnt;
	ulen = min_t(u32, info.nr_map_ids, ulen);
	if (ulen) {
1469
		u32 __user *user_map_ids = u64_to_user_ptr(info.map_ids);
1470 1471 1472 1473 1474 1475 1476
		u32 i;

		for (i = 0; i < ulen; i++)
			if (put_user(prog->aux->used_maps[i]->id,
				     &user_map_ids[i]))
				return -EFAULT;
	}
1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493

	if (!capable(CAP_SYS_ADMIN)) {
		info.jited_prog_len = 0;
		info.xlated_prog_len = 0;
		goto done;
	}

	ulen = info.jited_prog_len;
	info.jited_prog_len = prog->jited_len;
	if (info.jited_prog_len && ulen) {
		uinsns = u64_to_user_ptr(info.jited_prog_insns);
		ulen = min_t(u32, info.jited_prog_len, ulen);
		if (copy_to_user(uinsns, prog->bpf_func, ulen))
			return -EFAULT;
	}

	ulen = info.xlated_prog_len;
1494
	info.xlated_prog_len = bpf_prog_insn_size(prog);
1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529
	if (info.xlated_prog_len && ulen) {
		uinsns = u64_to_user_ptr(info.xlated_prog_insns);
		ulen = min_t(u32, info.xlated_prog_len, ulen);
		if (copy_to_user(uinsns, prog->insnsi, ulen))
			return -EFAULT;
	}

done:
	if (copy_to_user(uinfo, &info, info_len) ||
	    put_user(info_len, &uattr->info.info_len))
		return -EFAULT;

	return 0;
}

static int bpf_map_get_info_by_fd(struct bpf_map *map,
				  const union bpf_attr *attr,
				  union bpf_attr __user *uattr)
{
	struct bpf_map_info __user *uinfo = u64_to_user_ptr(attr->info.info);
	struct bpf_map_info info = {};
	u32 info_len = attr->info.info_len;
	int err;

	err = check_uarg_tail_zero(uinfo, sizeof(info), info_len);
	if (err)
		return err;
	info_len = min_t(u32, sizeof(info), info_len);

	info.type = map->map_type;
	info.id = map->id;
	info.key_size = map->key_size;
	info.value_size = map->value_size;
	info.max_entries = map->max_entries;
	info.map_flags = map->map_flags;
1530
	memcpy(info.name, map->name, sizeof(map->name));
1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567

	if (copy_to_user(uinfo, &info, info_len) ||
	    put_user(info_len, &uattr->info.info_len))
		return -EFAULT;

	return 0;
}

#define BPF_OBJ_GET_INFO_BY_FD_LAST_FIELD info.info

static int bpf_obj_get_info_by_fd(const union bpf_attr *attr,
				  union bpf_attr __user *uattr)
{
	int ufd = attr->info.bpf_fd;
	struct fd f;
	int err;

	if (CHECK_ATTR(BPF_OBJ_GET_INFO_BY_FD))
		return -EINVAL;

	f = fdget(ufd);
	if (!f.file)
		return -EBADFD;

	if (f.file->f_op == &bpf_prog_fops)
		err = bpf_prog_get_info_by_fd(f.file->private_data, attr,
					      uattr);
	else if (f.file->f_op == &bpf_map_fops)
		err = bpf_map_get_info_by_fd(f.file->private_data, attr,
					     uattr);
	else
		err = -EINVAL;

	fdput(f);
	return err;
}

1568 1569 1570 1571 1572
SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *, uattr, unsigned int, size)
{
	union bpf_attr attr = {};
	int err;

1573
	if (!capable(CAP_SYS_ADMIN) && sysctl_unprivileged_bpf_disabled)
1574 1575
		return -EPERM;

1576 1577 1578 1579
	err = check_uarg_tail_zero(uattr, sizeof(attr), size);
	if (err)
		return err;
	size = min_t(u32, size, sizeof(attr));
1580 1581 1582 1583 1584 1585 1586 1587 1588

	/* copy attributes from user space, may be less than sizeof(bpf_attr) */
	if (copy_from_user(&attr, uattr, size) != 0)
		return -EFAULT;

	switch (cmd) {
	case BPF_MAP_CREATE:
		err = map_create(&attr);
		break;
1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600
	case BPF_MAP_LOOKUP_ELEM:
		err = map_lookup_elem(&attr);
		break;
	case BPF_MAP_UPDATE_ELEM:
		err = map_update_elem(&attr);
		break;
	case BPF_MAP_DELETE_ELEM:
		err = map_delete_elem(&attr);
		break;
	case BPF_MAP_GET_NEXT_KEY:
		err = map_get_next_key(&attr);
		break;
1601 1602 1603
	case BPF_PROG_LOAD:
		err = bpf_prog_load(&attr);
		break;
1604 1605 1606 1607 1608 1609
	case BPF_OBJ_PIN:
		err = bpf_obj_pin(&attr);
		break;
	case BPF_OBJ_GET:
		err = bpf_obj_get(&attr);
		break;
1610 1611 1612 1613 1614 1615 1616
#ifdef CONFIG_CGROUP_BPF
	case BPF_PROG_ATTACH:
		err = bpf_prog_attach(&attr);
		break;
	case BPF_PROG_DETACH:
		err = bpf_prog_detach(&attr);
		break;
1617 1618 1619
	case BPF_PROG_QUERY:
		err = bpf_prog_query(&attr, uattr);
		break;
1620
#endif
1621 1622 1623
	case BPF_PROG_TEST_RUN:
		err = bpf_prog_test_run(&attr, uattr);
		break;
1624 1625 1626 1627 1628 1629 1630 1631
	case BPF_PROG_GET_NEXT_ID:
		err = bpf_obj_get_next_id(&attr, uattr,
					  &prog_idr, &prog_idr_lock);
		break;
	case BPF_MAP_GET_NEXT_ID:
		err = bpf_obj_get_next_id(&attr, uattr,
					  &map_idr, &map_idr_lock);
		break;
1632 1633 1634
	case BPF_PROG_GET_FD_BY_ID:
		err = bpf_prog_get_fd_by_id(&attr);
		break;
M
Martin KaFai Lau 已提交
1635 1636 1637
	case BPF_MAP_GET_FD_BY_ID:
		err = bpf_map_get_fd_by_id(&attr);
		break;
1638 1639 1640
	case BPF_OBJ_GET_INFO_BY_FD:
		err = bpf_obj_get_info_by_fd(&attr, uattr);
		break;
1641 1642 1643 1644 1645 1646 1647
	default:
		err = -EINVAL;
		break;
	}

	return err;
}