arraymap.c 8.7 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
/* 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>
#include <linux/err.h>
#include <linux/vmalloc.h>
#include <linux/slab.h>
#include <linux/mm.h>
17
#include <linux/filter.h>
18
#include <linux/perf_event.h>
19 20 21 22 23

/* Called from syscall */
static struct bpf_map *array_map_alloc(union bpf_attr *attr)
{
	struct bpf_array *array;
24
	u32 elem_size, array_size;
25 26 27 28 29 30

	/* check sanity of attributes */
	if (attr->max_entries == 0 || attr->key_size != 4 ||
	    attr->value_size == 0)
		return ERR_PTR(-EINVAL);

31 32 33 34 35 36
	if (attr->value_size >= 1 << (KMALLOC_SHIFT_MAX - 1))
		/* if value_size is bigger, the user space won't be able to
		 * access the elements.
		 */
		return ERR_PTR(-E2BIG);

37 38
	elem_size = round_up(attr->value_size, 8);

39 40
	/* check round_up into zero and u32 overflow */
	if (elem_size == 0 ||
41
	    attr->max_entries > (U32_MAX - PAGE_SIZE - sizeof(*array)) / elem_size)
42 43 44 45
		return ERR_PTR(-ENOMEM);

	array_size = sizeof(*array) + attr->max_entries * elem_size;

46
	/* allocate all map elements and zero-initialize them */
47
	array = kzalloc(array_size, GFP_USER | __GFP_NOWARN);
48
	if (!array) {
49
		array = vzalloc(array_size);
50 51 52 53 54 55 56 57
		if (!array)
			return ERR_PTR(-ENOMEM);
	}

	/* copy mandatory map attributes */
	array->map.key_size = attr->key_size;
	array->map.value_size = attr->value_size;
	array->map.max_entries = attr->max_entries;
58
	array->map.pages = round_up(array_size, PAGE_SIZE) >> PAGE_SHIFT;
59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110
	array->elem_size = elem_size;

	return &array->map;
}

/* Called from syscall or from eBPF program */
static void *array_map_lookup_elem(struct bpf_map *map, void *key)
{
	struct bpf_array *array = container_of(map, struct bpf_array, map);
	u32 index = *(u32 *)key;

	if (index >= array->map.max_entries)
		return NULL;

	return array->value + array->elem_size * index;
}

/* Called from syscall */
static int array_map_get_next_key(struct bpf_map *map, void *key, void *next_key)
{
	struct bpf_array *array = container_of(map, struct bpf_array, map);
	u32 index = *(u32 *)key;
	u32 *next = (u32 *)next_key;

	if (index >= array->map.max_entries) {
		*next = 0;
		return 0;
	}

	if (index == array->map.max_entries - 1)
		return -ENOENT;

	*next = index + 1;
	return 0;
}

/* Called from syscall or from eBPF program */
static int array_map_update_elem(struct bpf_map *map, void *key, void *value,
				 u64 map_flags)
{
	struct bpf_array *array = container_of(map, struct bpf_array, map);
	u32 index = *(u32 *)key;

	if (map_flags > BPF_EXIST)
		/* unknown flags */
		return -EINVAL;

	if (index >= array->map.max_entries)
		/* all elements were pre-allocated, cannot insert a new one */
		return -E2BIG;

	if (map_flags == BPF_NOEXIST)
111
		/* all elements already exist */
112 113
		return -EEXIST;

114
	memcpy(array->value + array->elem_size * index, value, map->value_size);
115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138
	return 0;
}

/* Called from syscall or from eBPF program */
static int array_map_delete_elem(struct bpf_map *map, void *key)
{
	return -EINVAL;
}

/* Called when map->refcnt goes to zero, either from workqueue or from syscall */
static void array_map_free(struct bpf_map *map)
{
	struct bpf_array *array = container_of(map, struct bpf_array, map);

	/* at this point bpf_prog->aux->refcnt == 0 and this map->refcnt == 0,
	 * so the programs (can be more than one that used this map) were
	 * disconnected from events. Wait for outstanding programs to complete
	 * and free the array
	 */
	synchronize_rcu();

	kvfree(array);
}

139
static const struct bpf_map_ops array_ops = {
140 141 142 143 144 145 146 147
	.map_alloc = array_map_alloc,
	.map_free = array_map_free,
	.map_get_next_key = array_map_get_next_key,
	.map_lookup_elem = array_map_lookup_elem,
	.map_update_elem = array_map_update_elem,
	.map_delete_elem = array_map_delete_elem,
};

148
static struct bpf_map_type_list array_type __read_mostly = {
149 150 151 152 153 154
	.ops = &array_ops,
	.type = BPF_MAP_TYPE_ARRAY,
};

static int __init register_array_map(void)
{
155
	bpf_register_map_type(&array_type);
156 157 158
	return 0;
}
late_initcall(register_array_map);
159

160
static struct bpf_map *fd_array_map_alloc(union bpf_attr *attr)
161
{
162
	/* only file descriptors can be stored in this type of map */
163 164 165 166 167
	if (attr->value_size != sizeof(u32))
		return ERR_PTR(-EINVAL);
	return array_map_alloc(attr);
}

168
static void fd_array_map_free(struct bpf_map *map)
169 170 171 172 173 174 175 176
{
	struct bpf_array *array = container_of(map, struct bpf_array, map);
	int i;

	synchronize_rcu();

	/* make sure it's empty */
	for (i = 0; i < array->map.max_entries; i++)
177
		BUG_ON(array->ptrs[i] != NULL);
178 179 180
	kvfree(array);
}

181
static void *fd_array_map_lookup_elem(struct bpf_map *map, void *key)
182 183 184 185 186
{
	return NULL;
}

/* only called from syscall */
187 188
static int fd_array_map_update_elem(struct bpf_map *map, void *key,
				    void *value, u64 map_flags)
189 190
{
	struct bpf_array *array = container_of(map, struct bpf_array, map);
191
	void *new_ptr, *old_ptr;
192 193 194 195 196 197 198 199 200
	u32 index = *(u32 *)key, ufd;

	if (map_flags != BPF_ANY)
		return -EINVAL;

	if (index >= array->map.max_entries)
		return -E2BIG;

	ufd = *(u32 *)value;
201 202 203
	new_ptr = map->ops->map_fd_get_ptr(map, ufd);
	if (IS_ERR(new_ptr))
		return PTR_ERR(new_ptr);
204

205 206 207
	old_ptr = xchg(array->ptrs + index, new_ptr);
	if (old_ptr)
		map->ops->map_fd_put_ptr(old_ptr);
208 209 210 211

	return 0;
}

212
static int fd_array_map_delete_elem(struct bpf_map *map, void *key)
213 214
{
	struct bpf_array *array = container_of(map, struct bpf_array, map);
215
	void *old_ptr;
216 217 218 219 220
	u32 index = *(u32 *)key;

	if (index >= array->map.max_entries)
		return -E2BIG;

221 222 223
	old_ptr = xchg(array->ptrs + index, NULL);
	if (old_ptr) {
		map->ops->map_fd_put_ptr(old_ptr);
224 225 226 227 228 229
		return 0;
	} else {
		return -ENOENT;
	}
}

230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250
static void *prog_fd_array_get_ptr(struct bpf_map *map, int fd)
{
	struct bpf_array *array = container_of(map, struct bpf_array, map);
	struct bpf_prog *prog = bpf_prog_get(fd);
	if (IS_ERR(prog))
		return prog;

	if (!bpf_prog_array_compatible(array, prog)) {
		bpf_prog_put(prog);
		return ERR_PTR(-EINVAL);
	}
	return prog;
}

static void prog_fd_array_put_ptr(void *ptr)
{
	struct bpf_prog *prog = ptr;

	bpf_prog_put_rcu(prog);
}

251
/* decrement refcnt of all bpf_progs that are stored in this map */
252
void bpf_fd_array_map_clear(struct bpf_map *map)
253 254 255 256 257
{
	struct bpf_array *array = container_of(map, struct bpf_array, map);
	int i;

	for (i = 0; i < array->map.max_entries; i++)
258
		fd_array_map_delete_elem(map, &i);
259 260 261
}

static const struct bpf_map_ops prog_array_ops = {
262 263
	.map_alloc = fd_array_map_alloc,
	.map_free = fd_array_map_free,
264
	.map_get_next_key = array_map_get_next_key,
265 266 267 268 269
	.map_lookup_elem = fd_array_map_lookup_elem,
	.map_update_elem = fd_array_map_update_elem,
	.map_delete_elem = fd_array_map_delete_elem,
	.map_fd_get_ptr = prog_fd_array_get_ptr,
	.map_fd_put_ptr = prog_fd_array_put_ptr,
270 271 272 273 274 275 276 277 278 279 280 281 282
};

static struct bpf_map_type_list prog_array_type __read_mostly = {
	.ops = &prog_array_ops,
	.type = BPF_MAP_TYPE_PROG_ARRAY,
};

static int __init register_prog_array_map(void)
{
	bpf_register_map_type(&prog_array_type);
	return 0;
}
late_initcall(register_prog_array_map);
283 284 285 286 287 288 289 290 291 292 293

static void perf_event_array_map_free(struct bpf_map *map)
{
	bpf_fd_array_map_clear(map);
	fd_array_map_free(map);
}

static void *perf_event_fd_array_get_ptr(struct bpf_map *map, int fd)
{
	struct perf_event *event;
	const struct perf_event_attr *attr;
294
	struct file *file;
295

296 297 298 299 300
	file = perf_event_get(fd);
	if (IS_ERR(file))
		return file;

	event = file->private_data;
301 302 303

	attr = perf_event_attrs(event);
	if (IS_ERR(attr))
304
		goto err;
305

306 307 308 309
	if (attr->inherit)
		goto err;

	if (attr->type == PERF_TYPE_RAW)
310
		return file;
311 312

	if (attr->type == PERF_TYPE_HARDWARE)
313
		return file;
314 315 316

	if (attr->type == PERF_TYPE_SOFTWARE &&
	    attr->config == PERF_COUNT_SW_BPF_OUTPUT)
317
		return file;
318
err:
319
	fput(file);
320
	return ERR_PTR(-EINVAL);
321 322 323 324
}

static void perf_event_fd_array_put_ptr(void *ptr)
{
325
	fput((struct file *)ptr);
326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349
}

static const struct bpf_map_ops perf_event_array_ops = {
	.map_alloc = fd_array_map_alloc,
	.map_free = perf_event_array_map_free,
	.map_get_next_key = array_map_get_next_key,
	.map_lookup_elem = fd_array_map_lookup_elem,
	.map_update_elem = fd_array_map_update_elem,
	.map_delete_elem = fd_array_map_delete_elem,
	.map_fd_get_ptr = perf_event_fd_array_get_ptr,
	.map_fd_put_ptr = perf_event_fd_array_put_ptr,
};

static struct bpf_map_type_list perf_event_array_type __read_mostly = {
	.ops = &perf_event_array_ops,
	.type = BPF_MAP_TYPE_PERF_EVENT_ARRAY,
};

static int __init register_perf_event_array_map(void)
{
	bpf_register_map_type(&perf_event_array_type);
	return 0;
}
late_initcall(register_perf_event_array_map);