idr.c 28.3 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8
/*
 * 2002-10-18  written by Jim Houston jim.houston@ccur.com
 *	Copyright (C) 2002 by Concurrent Computer Corporation
 *	Distributed under the GNU GPL license version 2.
 *
 * Modified by George Anzinger to reuse immediately and to use
 * find bit instructions.  Also removed _irq on spinlocks.
 *
N
Nadia Derbey 已提交
9 10
 * Modified by Nadia Derbey to make it RCU safe.
 *
11
 * Small id to pointer translation service.
L
Linus Torvalds 已提交
12
 *
13
 * It uses a radix tree like structure as a sparse array indexed
L
Linus Torvalds 已提交
14
 * by the id to obtain the pointer.  The bitmap makes allocating
15
 * a new id quick.
L
Linus Torvalds 已提交
16 17 18 19 20 21
 *
 * You call it to allocate an id (an int) an associate with that id a
 * pointer or what ever, we treat it as a (void *).  You can pass this
 * id to a user for him to pass back at a later time.  You then pass
 * that id to this code and it returns your pointer.

22
 * You can release ids at any time. When all ids are released, most of
23
 * the memory is returned (we keep MAX_IDR_FREE) in a local pool so we
24
 * don't need to go to the memory "store" during an id allocate, just
L
Linus Torvalds 已提交
25 26 27 28 29 30 31
 * so you don't need to be too concerned about locking and conflicts
 * with the slab allocator.
 */

#ifndef TEST                        // to test in user space...
#include <linux/slab.h>
#include <linux/init.h>
32
#include <linux/export.h>
L
Linus Torvalds 已提交
33
#endif
J
Jeff Mahoney 已提交
34
#include <linux/err.h>
L
Linus Torvalds 已提交
35 36
#include <linux/string.h>
#include <linux/idr.h>
37
#include <linux/spinlock.h>
38 39
#include <linux/percpu.h>
#include <linux/hardirq.h>
L
Linus Torvalds 已提交
40

41 42 43 44 45 46 47 48 49
#define MAX_IDR_SHIFT		(sizeof(int) * 8 - 1)
#define MAX_IDR_BIT		(1U << MAX_IDR_SHIFT)

/* Leave the possibility of an incomplete final layer */
#define MAX_IDR_LEVEL ((MAX_IDR_SHIFT + IDR_BITS - 1) / IDR_BITS)

/* Number of id_layer structs to leave in free list */
#define MAX_IDR_FREE (MAX_IDR_LEVEL * 2)

50
static struct kmem_cache *idr_layer_cache;
51 52
static DEFINE_PER_CPU(struct idr_layer *, idr_preload_head);
static DEFINE_PER_CPU(int, idr_preload_cnt);
53
static DEFINE_SPINLOCK(simple_ida_lock);
L
Linus Torvalds 已提交
54

T
Tejun Heo 已提交
55 56 57 58 59 60 61 62
/* the maximum ID which can be allocated given idr->layers */
static int idr_max(int layers)
{
	int bits = min_t(int, layers * IDR_BITS, MAX_IDR_SHIFT);

	return (1 << bits) - 1;
}

63
static struct idr_layer *get_from_free_list(struct idr *idp)
L
Linus Torvalds 已提交
64 65
{
	struct idr_layer *p;
66
	unsigned long flags;
L
Linus Torvalds 已提交
67

68
	spin_lock_irqsave(&idp->lock, flags);
L
Linus Torvalds 已提交
69 70 71 72 73
	if ((p = idp->id_free)) {
		idp->id_free = p->ary[0];
		idp->id_free_cnt--;
		p->ary[0] = NULL;
	}
74
	spin_unlock_irqrestore(&idp->lock, flags);
L
Linus Torvalds 已提交
75 76 77
	return(p);
}

78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121
/**
 * idr_layer_alloc - allocate a new idr_layer
 * @gfp_mask: allocation mask
 * @layer_idr: optional idr to allocate from
 *
 * If @layer_idr is %NULL, directly allocate one using @gfp_mask or fetch
 * one from the per-cpu preload buffer.  If @layer_idr is not %NULL, fetch
 * an idr_layer from @idr->id_free.
 *
 * @layer_idr is to maintain backward compatibility with the old alloc
 * interface - idr_pre_get() and idr_get_new*() - and will be removed
 * together with per-pool preload buffer.
 */
static struct idr_layer *idr_layer_alloc(gfp_t gfp_mask, struct idr *layer_idr)
{
	struct idr_layer *new;

	/* this is the old path, bypass to get_from_free_list() */
	if (layer_idr)
		return get_from_free_list(layer_idr);

	/* try to allocate directly from kmem_cache */
	new = kmem_cache_zalloc(idr_layer_cache, gfp_mask);
	if (new)
		return new;

	/*
	 * Try to fetch one from the per-cpu preload buffer if in process
	 * context.  See idr_preload() for details.
	 */
	if (in_interrupt())
		return NULL;

	preempt_disable();
	new = __this_cpu_read(idr_preload_head);
	if (new) {
		__this_cpu_write(idr_preload_head, new->ary[0]);
		__this_cpu_dec(idr_preload_cnt);
		new->ary[0] = NULL;
	}
	preempt_enable();
	return new;
}

N
Nadia Derbey 已提交
122 123 124 125 126 127 128 129 130 131 132 133 134
static void idr_layer_rcu_free(struct rcu_head *head)
{
	struct idr_layer *layer;

	layer = container_of(head, struct idr_layer, rcu_head);
	kmem_cache_free(idr_layer_cache, layer);
}

static inline void free_layer(struct idr_layer *p)
{
	call_rcu(&p->rcu_head, idr_layer_rcu_free);
}

S
Sonny Rao 已提交
135
/* only called when idp->lock is held */
136
static void __move_to_free_list(struct idr *idp, struct idr_layer *p)
S
Sonny Rao 已提交
137 138 139 140 141 142
{
	p->ary[0] = idp->id_free;
	idp->id_free = p;
	idp->id_free_cnt++;
}

143
static void move_to_free_list(struct idr *idp, struct idr_layer *p)
L
Linus Torvalds 已提交
144
{
145 146
	unsigned long flags;

L
Linus Torvalds 已提交
147 148 149
	/*
	 * Depends on the return element being zeroed.
	 */
150
	spin_lock_irqsave(&idp->lock, flags);
151
	__move_to_free_list(idp, p);
152
	spin_unlock_irqrestore(&idp->lock, flags);
L
Linus Torvalds 已提交
153 154
}

T
Tejun Heo 已提交
155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174
static void idr_mark_full(struct idr_layer **pa, int id)
{
	struct idr_layer *p = pa[0];
	int l = 0;

	__set_bit(id & IDR_MASK, &p->bitmap);
	/*
	 * If this layer is full mark the bit in the layer above to
	 * show that this part of the radix tree is full.  This may
	 * complete the layer above and require walking up the radix
	 * tree.
	 */
	while (p->bitmap == IDR_FULL) {
		if (!(p = pa[++l]))
			break;
		id = id >> IDR_BITS;
		__set_bit((id & IDR_MASK), &p->bitmap);
	}
}

L
Linus Torvalds 已提交
175
/**
176
 * idr_pre_get - reserve resources for idr allocation
L
Linus Torvalds 已提交
177 178 179
 * @idp:	idr handle
 * @gfp_mask:	memory allocation flags
 *
180 181 182 183
 * This function should be called prior to calling the idr_get_new* functions.
 * It preallocates enough memory to satisfy the worst possible allocation. The
 * caller should pass in GFP_KERNEL if possible.  This of course requires that
 * no spinning locks be held.
L
Linus Torvalds 已提交
184
 *
185 186
 * If the system is REALLY out of memory this function returns %0,
 * otherwise %1.
L
Linus Torvalds 已提交
187
 */
A
Al Viro 已提交
188
int idr_pre_get(struct idr *idp, gfp_t gfp_mask)
L
Linus Torvalds 已提交
189
{
190
	while (idp->id_free_cnt < MAX_IDR_FREE) {
L
Linus Torvalds 已提交
191
		struct idr_layer *new;
192
		new = kmem_cache_zalloc(idr_layer_cache, gfp_mask);
193
		if (new == NULL)
L
Linus Torvalds 已提交
194
			return (0);
195
		move_to_free_list(idp, new);
L
Linus Torvalds 已提交
196 197 198 199 200
	}
	return 1;
}
EXPORT_SYMBOL(idr_pre_get);

T
Tejun Heo 已提交
201 202 203 204 205 206
/**
 * sub_alloc - try to allocate an id without growing the tree depth
 * @idp: idr handle
 * @starting_id: id to start search at
 * @id: pointer to the allocated handle
 * @pa: idr_layer[MAX_IDR_LEVEL] used as backtrack buffer
207 208
 * @gfp_mask: allocation mask for idr_layer_alloc()
 * @layer_idr: optional idr passed to idr_layer_alloc()
T
Tejun Heo 已提交
209 210 211 212 213 214 215 216 217
 *
 * Allocate an id in range [@starting_id, INT_MAX] from @idp without
 * growing its depth.  Returns
 *
 *  the allocated id >= 0 if successful,
 *  -EAGAIN if the tree needs to grow for allocation to succeed,
 *  -ENOSPC if the id space is exhausted,
 *  -ENOMEM if more idr_layers need to be allocated.
 */
218 219
static int sub_alloc(struct idr *idp, int *starting_id, struct idr_layer **pa,
		     gfp_t gfp_mask, struct idr *layer_idr)
L
Linus Torvalds 已提交
220 221 222
{
	int n, m, sh;
	struct idr_layer *p, *new;
223
	int l, id, oid;
224
	unsigned long bm;
L
Linus Torvalds 已提交
225 226

	id = *starting_id;
227
 restart:
L
Linus Torvalds 已提交
228 229 230 231 232 233 234 235 236 237 238 239 240
	p = idp->top;
	l = idp->layers;
	pa[l--] = NULL;
	while (1) {
		/*
		 * We run around this while until we reach the leaf node...
		 */
		n = (id >> (IDR_BITS*l)) & IDR_MASK;
		bm = ~p->bitmap;
		m = find_next_bit(&bm, IDR_SIZE, n);
		if (m == IDR_SIZE) {
			/* no space available go back to previous layer. */
			l++;
241
			oid = id;
242
			id = (id | ((1 << (IDR_BITS * l)) - 1)) + 1;
243 244

			/* if already at the top layer, we need to grow */
245
			if (id >= 1 << (idp->layers * IDR_BITS)) {
L
Linus Torvalds 已提交
246
				*starting_id = id;
T
Tejun Heo 已提交
247
				return -EAGAIN;
L
Linus Torvalds 已提交
248
			}
249 250
			p = pa[l];
			BUG_ON(!p);
251 252 253 254 255 256 257 258 259

			/* If we need to go up one layer, continue the
			 * loop; otherwise, restart from the top.
			 */
			sh = IDR_BITS * (l + 1);
			if (oid >> sh == id >> sh)
				continue;
			else
				goto restart;
L
Linus Torvalds 已提交
260 261 262 263 264
		}
		if (m != n) {
			sh = IDR_BITS*l;
			id = ((id >> sh) ^ n ^ m) << sh;
		}
265
		if ((id >= MAX_IDR_BIT) || (id < 0))
T
Tejun Heo 已提交
266
			return -ENOSPC;
L
Linus Torvalds 已提交
267 268 269 270 271 272
		if (l == 0)
			break;
		/*
		 * Create the layer below if it is missing.
		 */
		if (!p->ary[m]) {
273
			new = idr_layer_alloc(gfp_mask, layer_idr);
274
			if (!new)
T
Tejun Heo 已提交
275
				return -ENOMEM;
276
			new->layer = l-1;
N
Nadia Derbey 已提交
277
			rcu_assign_pointer(p->ary[m], new);
L
Linus Torvalds 已提交
278 279 280 281 282
			p->count++;
		}
		pa[l--] = p;
		p = p->ary[m];
	}
T
Tejun Heo 已提交
283 284 285

	pa[l] = p;
	return id;
L
Linus Torvalds 已提交
286 287
}

T
Tejun Heo 已提交
288
static int idr_get_empty_slot(struct idr *idp, int starting_id,
289 290
			      struct idr_layer **pa, gfp_t gfp_mask,
			      struct idr *layer_idr)
L
Linus Torvalds 已提交
291 292 293
{
	struct idr_layer *p, *new;
	int layers, v, id;
294
	unsigned long flags;
295

L
Linus Torvalds 已提交
296 297 298 299 300
	id = starting_id;
build_up:
	p = idp->top;
	layers = idp->layers;
	if (unlikely(!p)) {
301
		if (!(p = idr_layer_alloc(gfp_mask, layer_idr)))
T
Tejun Heo 已提交
302
			return -ENOMEM;
303
		p->layer = 0;
L
Linus Torvalds 已提交
304 305 306 307 308 309
		layers = 1;
	}
	/*
	 * Add a new layer to the top of the tree if the requested
	 * id is larger than the currently allocated space.
	 */
T
Tejun Heo 已提交
310
	while (id > idr_max(layers)) {
L
Linus Torvalds 已提交
311
		layers++;
312 313 314 315 316 317
		if (!p->count) {
			/* special case: if the tree is currently empty,
			 * then we grow the tree by moving the top node
			 * upwards.
			 */
			p->layer++;
L
Linus Torvalds 已提交
318
			continue;
319
		}
320
		if (!(new = idr_layer_alloc(gfp_mask, layer_idr))) {
L
Linus Torvalds 已提交
321 322 323 324
			/*
			 * The allocation failed.  If we built part of
			 * the structure tear it down.
			 */
325
			spin_lock_irqsave(&idp->lock, flags);
L
Linus Torvalds 已提交
326 327 328 329
			for (new = p; p && p != idp->top; new = p) {
				p = p->ary[0];
				new->ary[0] = NULL;
				new->bitmap = new->count = 0;
330
				__move_to_free_list(idp, new);
L
Linus Torvalds 已提交
331
			}
332
			spin_unlock_irqrestore(&idp->lock, flags);
T
Tejun Heo 已提交
333
			return -ENOMEM;
L
Linus Torvalds 已提交
334 335 336
		}
		new->ary[0] = p;
		new->count = 1;
337
		new->layer = layers-1;
L
Linus Torvalds 已提交
338 339 340 341
		if (p->bitmap == IDR_FULL)
			__set_bit(0, &new->bitmap);
		p = new;
	}
N
Nadia Derbey 已提交
342
	rcu_assign_pointer(idp->top, p);
L
Linus Torvalds 已提交
343
	idp->layers = layers;
344
	v = sub_alloc(idp, &id, pa, gfp_mask, layer_idr);
T
Tejun Heo 已提交
345
	if (v == -EAGAIN)
L
Linus Torvalds 已提交
346 347 348 349
		goto build_up;
	return(v);
}

T
Tejun Heo 已提交
350 351 352 353 354
/*
 * @id and @pa are from a successful allocation from idr_get_empty_slot().
 * Install the user pointer @ptr and mark the slot full.
 */
static void idr_fill_slot(void *ptr, int id, struct idr_layer **pa)
T
Tejun Heo 已提交
355
{
T
Tejun Heo 已提交
356 357 358
	rcu_assign_pointer(pa[0]->ary[id & IDR_MASK], (struct idr_layer *)ptr);
	pa[0]->count++;
	idr_mark_full(pa, id);
T
Tejun Heo 已提交
359 360
}

L
Linus Torvalds 已提交
361
/**
362
 * idr_get_new_above - allocate new idr entry above or equal to a start id
L
Linus Torvalds 已提交
363
 * @idp: idr handle
364
 * @ptr: pointer you want associated with the id
N
Naohiro Aota 已提交
365
 * @starting_id: id to start search at
L
Linus Torvalds 已提交
366 367 368 369 370
 * @id: pointer to the allocated handle
 *
 * This is the allocate id function.  It should be called with any
 * required locks.
 *
371
 * If allocation from IDR's private freelist fails, idr_get_new_above() will
372
 * return %-EAGAIN.  The caller should retry the idr_pre_get() call to refill
373 374
 * IDR's preallocation and then retry the idr_get_new_above() call.
 *
375
 * If the idr is full idr_get_new_above() will return %-ENOSPC.
L
Linus Torvalds 已提交
376
 *
377
 * @id returns a value in the range @starting_id ... %0x7fffffff
L
Linus Torvalds 已提交
378 379 380
 */
int idr_get_new_above(struct idr *idp, void *ptr, int starting_id, int *id)
{
T
Tejun Heo 已提交
381
	struct idr_layer *pa[MAX_IDR_LEVEL + 1];
L
Linus Torvalds 已提交
382
	int rv;
383

384
	rv = idr_get_empty_slot(idp, starting_id, pa, 0, idp);
N
Nadia Derbey 已提交
385
	if (rv < 0)
T
Tejun Heo 已提交
386
		return rv == -ENOMEM ? -EAGAIN : rv;
T
Tejun Heo 已提交
387 388

	idr_fill_slot(ptr, rv, pa);
L
Linus Torvalds 已提交
389 390 391 392 393
	*id = rv;
	return 0;
}
EXPORT_SYMBOL(idr_get_new_above);

394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476
/**
 * idr_preload - preload for idr_alloc()
 * @gfp_mask: allocation mask to use for preloading
 *
 * Preload per-cpu layer buffer for idr_alloc().  Can only be used from
 * process context and each idr_preload() invocation should be matched with
 * idr_preload_end().  Note that preemption is disabled while preloaded.
 *
 * The first idr_alloc() in the preloaded section can be treated as if it
 * were invoked with @gfp_mask used for preloading.  This allows using more
 * permissive allocation masks for idrs protected by spinlocks.
 *
 * For example, if idr_alloc() below fails, the failure can be treated as
 * if idr_alloc() were called with GFP_KERNEL rather than GFP_NOWAIT.
 *
 *	idr_preload(GFP_KERNEL);
 *	spin_lock(lock);
 *
 *	id = idr_alloc(idr, ptr, start, end, GFP_NOWAIT);
 *
 *	spin_unlock(lock);
 *	idr_preload_end();
 *	if (id < 0)
 *		error;
 */
void idr_preload(gfp_t gfp_mask)
{
	/*
	 * Consuming preload buffer from non-process context breaks preload
	 * allocation guarantee.  Disallow usage from those contexts.
	 */
	WARN_ON_ONCE(in_interrupt());
	might_sleep_if(gfp_mask & __GFP_WAIT);

	preempt_disable();

	/*
	 * idr_alloc() is likely to succeed w/o full idr_layer buffer and
	 * return value from idr_alloc() needs to be checked for failure
	 * anyway.  Silently give up if allocation fails.  The caller can
	 * treat failures from idr_alloc() as if idr_alloc() were called
	 * with @gfp_mask which should be enough.
	 */
	while (__this_cpu_read(idr_preload_cnt) < MAX_IDR_FREE) {
		struct idr_layer *new;

		preempt_enable();
		new = kmem_cache_zalloc(idr_layer_cache, gfp_mask);
		preempt_disable();
		if (!new)
			break;

		/* link the new one to per-cpu preload list */
		new->ary[0] = __this_cpu_read(idr_preload_head);
		__this_cpu_write(idr_preload_head, new);
		__this_cpu_inc(idr_preload_cnt);
	}
}
EXPORT_SYMBOL(idr_preload);

/**
 * idr_alloc - allocate new idr entry
 * @idr: the (initialized) idr
 * @ptr: pointer to be associated with the new id
 * @start: the minimum id (inclusive)
 * @end: the maximum id (exclusive, <= 0 for max)
 * @gfp_mask: memory allocation flags
 *
 * Allocate an id in [start, end) and associate it with @ptr.  If no ID is
 * available in the specified range, returns -ENOSPC.  On memory allocation
 * failure, returns -ENOMEM.
 *
 * Note that @end is treated as max when <= 0.  This is to always allow
 * using @start + N as @end as long as N is inside integer range.
 *
 * The user is responsible for exclusively synchronizing all operations
 * which may modify @idr.  However, read-only accesses such as idr_find()
 * or iteration can be performed under RCU read lock provided the user
 * destroys @ptr in RCU-safe way after removal from idr.
 */
int idr_alloc(struct idr *idr, void *ptr, int start, int end, gfp_t gfp_mask)
{
	int max = end > 0 ? end - 1 : INT_MAX;	/* inclusive upper limit */
T
Tejun Heo 已提交
477
	struct idr_layer *pa[MAX_IDR_LEVEL + 1];
478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499
	int id;

	might_sleep_if(gfp_mask & __GFP_WAIT);

	/* sanity checks */
	if (WARN_ON_ONCE(start < 0))
		return -EINVAL;
	if (unlikely(max < start))
		return -ENOSPC;

	/* allocate id */
	id = idr_get_empty_slot(idr, start, pa, gfp_mask, NULL);
	if (unlikely(id < 0))
		return id;
	if (unlikely(id > max))
		return -ENOSPC;

	idr_fill_slot(ptr, id, pa);
	return id;
}
EXPORT_SYMBOL_GPL(idr_alloc);

L
Linus Torvalds 已提交
500 501
static void idr_remove_warning(int id)
{
N
Nadia Derbey 已提交
502 503
	printk(KERN_WARNING
		"idr_remove called for id=%d which is not allocated.\n", id);
L
Linus Torvalds 已提交
504 505 506 507 508 509
	dump_stack();
}

static void sub_remove(struct idr *idp, int shift, int id)
{
	struct idr_layer *p = idp->top;
T
Tejun Heo 已提交
510
	struct idr_layer **pa[MAX_IDR_LEVEL + 1];
L
Linus Torvalds 已提交
511
	struct idr_layer ***paa = &pa[0];
N
Nadia Derbey 已提交
512
	struct idr_layer *to_free;
L
Linus Torvalds 已提交
513 514 515 516 517 518 519 520 521 522 523 524 525 526 527
	int n;

	*paa = NULL;
	*++paa = &idp->top;

	while ((shift > 0) && p) {
		n = (id >> shift) & IDR_MASK;
		__clear_bit(n, &p->bitmap);
		*++paa = &p->ary[n];
		p = p->ary[n];
		shift -= IDR_BITS;
	}
	n = id & IDR_MASK;
	if (likely(p != NULL && test_bit(n, &p->bitmap))){
		__clear_bit(n, &p->bitmap);
N
Nadia Derbey 已提交
528 529
		rcu_assign_pointer(p->ary[n], NULL);
		to_free = NULL;
L
Linus Torvalds 已提交
530
		while(*paa && ! --((**paa)->count)){
N
Nadia Derbey 已提交
531 532 533
			if (to_free)
				free_layer(to_free);
			to_free = **paa;
L
Linus Torvalds 已提交
534 535
			**paa-- = NULL;
		}
536
		if (!*paa)
L
Linus Torvalds 已提交
537
			idp->layers = 0;
N
Nadia Derbey 已提交
538 539
		if (to_free)
			free_layer(to_free);
540
	} else
L
Linus Torvalds 已提交
541 542 543 544
		idr_remove_warning(id);
}

/**
545
 * idr_remove - remove the given id and free its slot
546 547
 * @idp: idr handle
 * @id: unique key
L
Linus Torvalds 已提交
548 549 550 551
 */
void idr_remove(struct idr *idp, int id)
{
	struct idr_layer *p;
N
Nadia Derbey 已提交
552
	struct idr_layer *to_free;
L
Linus Torvalds 已提交
553

554 555
	if (WARN_ON_ONCE(id < 0))
		return;
L
Linus Torvalds 已提交
556 557

	sub_remove(idp, (idp->layers - 1) * IDR_BITS, id);
558
	if (idp->top && idp->top->count == 1 && (idp->layers > 1) &&
N
Nadia Derbey 已提交
559 560 561 562 563 564 565 566
	    idp->top->ary[0]) {
		/*
		 * Single child at leftmost slot: we can shrink the tree.
		 * This level is not needed anymore since when layers are
		 * inserted, they are inserted at the top of the existing
		 * tree.
		 */
		to_free = idp->top;
L
Linus Torvalds 已提交
567
		p = idp->top->ary[0];
N
Nadia Derbey 已提交
568
		rcu_assign_pointer(idp->top, p);
L
Linus Torvalds 已提交
569
		--idp->layers;
N
Nadia Derbey 已提交
570 571
		to_free->bitmap = to_free->count = 0;
		free_layer(to_free);
L
Linus Torvalds 已提交
572
	}
573
	while (idp->id_free_cnt >= MAX_IDR_FREE) {
574
		p = get_from_free_list(idp);
N
Nadia Derbey 已提交
575 576 577 578 579
		/*
		 * Note: we don't call the rcu callback here, since the only
		 * layers that fall into the freelist are those that have been
		 * preallocated.
		 */
L
Linus Torvalds 已提交
580 581
		kmem_cache_free(idr_layer_cache, p);
	}
N
Nadia Derbey 已提交
582
	return;
L
Linus Torvalds 已提交
583 584 585
}
EXPORT_SYMBOL(idr_remove);

T
Tejun Heo 已提交
586
void __idr_remove_all(struct idr *idp)
K
Kristian Hoegsberg 已提交
587
{
588
	int n, id, max;
589
	int bt_mask;
K
Kristian Hoegsberg 已提交
590
	struct idr_layer *p;
T
Tejun Heo 已提交
591
	struct idr_layer *pa[MAX_IDR_LEVEL + 1];
K
Kristian Hoegsberg 已提交
592 593 594 595
	struct idr_layer **paa = &pa[0];

	n = idp->layers * IDR_BITS;
	p = idp->top;
596
	rcu_assign_pointer(idp->top, NULL);
T
Tejun Heo 已提交
597
	max = idr_max(idp->layers);
K
Kristian Hoegsberg 已提交
598 599

	id = 0;
T
Tejun Heo 已提交
600
	while (id >= 0 && id <= max) {
K
Kristian Hoegsberg 已提交
601 602 603 604 605 606
		while (n > IDR_BITS && p) {
			n -= IDR_BITS;
			*paa++ = p;
			p = p->ary[(id >> n) & IDR_MASK];
		}

607
		bt_mask = id;
K
Kristian Hoegsberg 已提交
608
		id += 1 << n;
609 610
		/* Get the highest bit that the above add changed from 0->1. */
		while (n < fls(id ^ bt_mask)) {
N
Nadia Derbey 已提交
611 612
			if (p)
				free_layer(p);
K
Kristian Hoegsberg 已提交
613 614 615 616 617 618
			n += IDR_BITS;
			p = *--paa;
		}
	}
	idp->layers = 0;
}
T
Tejun Heo 已提交
619
EXPORT_SYMBOL(__idr_remove_all);
K
Kristian Hoegsberg 已提交
620

A
Andrew Morton 已提交
621 622
/**
 * idr_destroy - release all cached layers within an idr tree
N
Naohiro Aota 已提交
623
 * @idp: idr handle
624 625 626 627 628 629 630 631 632
 *
 * Free all id mappings and all idp_layers.  After this function, @idp is
 * completely unused and can be freed / recycled.  The caller is
 * responsible for ensuring that no one else accesses @idp during or after
 * idr_destroy().
 *
 * A typical clean-up sequence for objects stored in an idr tree will use
 * idr_for_each() to free all objects, if necessay, then idr_destroy() to
 * free up the id mappings and cached idr_layers.
A
Andrew Morton 已提交
633 634 635
 */
void idr_destroy(struct idr *idp)
{
T
Tejun Heo 已提交
636
	__idr_remove_all(idp);
637

A
Andrew Morton 已提交
638
	while (idp->id_free_cnt) {
639
		struct idr_layer *p = get_from_free_list(idp);
A
Andrew Morton 已提交
640 641 642 643 644
		kmem_cache_free(idr_layer_cache, p);
	}
}
EXPORT_SYMBOL(idr_destroy);

L
Linus Torvalds 已提交
645 646 647 648 649 650 651 652 653
/**
 * idr_find - return pointer for given id
 * @idp: idr handle
 * @id: lookup key
 *
 * Return the pointer given the id it has been registered with.  A %NULL
 * return indicates that @id is not valid or you passed %NULL in
 * idr_get_new().
 *
N
Nadia Derbey 已提交
654 655
 * This function can be called under rcu_read_lock(), given that the leaf
 * pointers lifetimes are correctly managed.
L
Linus Torvalds 已提交
656 657 658 659 660 661
 */
void *idr_find(struct idr *idp, int id)
{
	int n;
	struct idr_layer *p;

662 663 664
	if (WARN_ON_ONCE(id < 0))
		return NULL;

665
	p = rcu_dereference_raw(idp->top);
666 667 668
	if (!p)
		return NULL;
	n = (p->layer+1) * IDR_BITS;
L
Linus Torvalds 已提交
669

T
Tejun Heo 已提交
670
	if (id > idr_max(p->layer + 1))
L
Linus Torvalds 已提交
671
		return NULL;
672
	BUG_ON(n == 0);
L
Linus Torvalds 已提交
673 674 675

	while (n > 0 && p) {
		n -= IDR_BITS;
676
		BUG_ON(n != p->layer*IDR_BITS);
677
		p = rcu_dereference_raw(p->ary[(id >> n) & IDR_MASK]);
L
Linus Torvalds 已提交
678 679 680 681 682
	}
	return((void *)p);
}
EXPORT_SYMBOL(idr_find);

K
Kristian Hoegsberg 已提交
683 684 685 686 687 688 689 690 691 692 693 694 695 696
/**
 * idr_for_each - iterate through all stored pointers
 * @idp: idr handle
 * @fn: function to be called for each pointer
 * @data: data passed back to callback function
 *
 * Iterate over the pointers registered with the given idr.  The
 * callback function will be called for each pointer currently
 * registered, passing the id, the pointer and the data pointer passed
 * to this function.  It is not safe to modify the idr tree while in
 * the callback, so functions such as idr_get_new and idr_remove are
 * not allowed.
 *
 * We check the return of @fn each time. If it returns anything other
697
 * than %0, we break out and return that value.
K
Kristian Hoegsberg 已提交
698 699 700 701 702 703 704 705
 *
 * The caller must serialize idr_for_each() vs idr_get_new() and idr_remove().
 */
int idr_for_each(struct idr *idp,
		 int (*fn)(int id, void *p, void *data), void *data)
{
	int n, id, max, error = 0;
	struct idr_layer *p;
T
Tejun Heo 已提交
706
	struct idr_layer *pa[MAX_IDR_LEVEL + 1];
K
Kristian Hoegsberg 已提交
707 708 709
	struct idr_layer **paa = &pa[0];

	n = idp->layers * IDR_BITS;
710
	p = rcu_dereference_raw(idp->top);
T
Tejun Heo 已提交
711
	max = idr_max(idp->layers);
K
Kristian Hoegsberg 已提交
712 713

	id = 0;
T
Tejun Heo 已提交
714
	while (id >= 0 && id <= max) {
K
Kristian Hoegsberg 已提交
715 716 717
		while (n > 0 && p) {
			n -= IDR_BITS;
			*paa++ = p;
718
			p = rcu_dereference_raw(p->ary[(id >> n) & IDR_MASK]);
K
Kristian Hoegsberg 已提交
719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737
		}

		if (p) {
			error = fn(id, (void *)p, data);
			if (error)
				break;
		}

		id += 1 << n;
		while (n < fls(id)) {
			n += IDR_BITS;
			p = *--paa;
		}
	}

	return error;
}
EXPORT_SYMBOL(idr_for_each);

K
KAMEZAWA Hiroyuki 已提交
738 739 740
/**
 * idr_get_next - lookup next object of id to given id.
 * @idp: idr handle
N
Naohiro Aota 已提交
741
 * @nextidp:  pointer to lookup key
K
KAMEZAWA Hiroyuki 已提交
742 743
 *
 * Returns pointer to registered object with id, which is next number to
744 745
 * given id. After being looked up, *@nextidp will be updated for the next
 * iteration.
746 747 748
 *
 * This function can be called under rcu_read_lock(), given that the leaf
 * pointers lifetimes are correctly managed.
K
KAMEZAWA Hiroyuki 已提交
749 750 751
 */
void *idr_get_next(struct idr *idp, int *nextidp)
{
T
Tejun Heo 已提交
752
	struct idr_layer *p, *pa[MAX_IDR_LEVEL + 1];
K
KAMEZAWA Hiroyuki 已提交
753 754 755 756 757
	struct idr_layer **paa = &pa[0];
	int id = *nextidp;
	int n, max;

	/* find first ent */
758
	p = rcu_dereference_raw(idp->top);
K
KAMEZAWA Hiroyuki 已提交
759 760
	if (!p)
		return NULL;
761
	n = (p->layer + 1) * IDR_BITS;
T
Tejun Heo 已提交
762
	max = idr_max(p->layer + 1);
K
KAMEZAWA Hiroyuki 已提交
763

T
Tejun Heo 已提交
764
	while (id >= 0 && id <= max) {
K
KAMEZAWA Hiroyuki 已提交
765 766 767
		while (n > 0 && p) {
			n -= IDR_BITS;
			*paa++ = p;
768
			p = rcu_dereference_raw(p->ary[(id >> n) & IDR_MASK]);
K
KAMEZAWA Hiroyuki 已提交
769 770 771 772 773 774 775
		}

		if (p) {
			*nextidp = id;
			return p;
		}

776 777 778 779 780 781 782 783
		/*
		 * Proceed to the next layer at the current level.  Unlike
		 * idr_for_each(), @id isn't guaranteed to be aligned to
		 * layer boundary at this point and adding 1 << n may
		 * incorrectly skip IDs.  Make sure we jump to the
		 * beginning of the next layer using round_up().
		 */
		id = round_up(id + 1, 1 << n);
K
KAMEZAWA Hiroyuki 已提交
784 785 786 787 788 789 790
		while (n < fls(id)) {
			n += IDR_BITS;
			p = *--paa;
		}
	}
	return NULL;
}
B
Ben Hutchings 已提交
791
EXPORT_SYMBOL(idr_get_next);
K
KAMEZAWA Hiroyuki 已提交
792 793


J
Jeff Mahoney 已提交
794 795 796 797 798 799 800
/**
 * idr_replace - replace pointer for given id
 * @idp: idr handle
 * @ptr: pointer you want associated with the id
 * @id: lookup key
 *
 * Replace the pointer registered with an id and return the old value.
801 802
 * A %-ENOENT return indicates that @id was not found.
 * A %-EINVAL return indicates that @id was not within valid constraints.
J
Jeff Mahoney 已提交
803
 *
N
Nadia Derbey 已提交
804
 * The caller must serialize with writers.
J
Jeff Mahoney 已提交
805 806 807 808 809 810
 */
void *idr_replace(struct idr *idp, void *ptr, int id)
{
	int n;
	struct idr_layer *p, *old_p;

811 812 813
	if (WARN_ON_ONCE(id < 0))
		return ERR_PTR(-EINVAL);

J
Jeff Mahoney 已提交
814
	p = idp->top;
815 816 817 818
	if (!p)
		return ERR_PTR(-EINVAL);

	n = (p->layer+1) * IDR_BITS;
J
Jeff Mahoney 已提交
819 820 821 822 823 824 825 826 827 828 829 830 831 832 833

	if (id >= (1 << n))
		return ERR_PTR(-EINVAL);

	n -= IDR_BITS;
	while ((n > 0) && p) {
		p = p->ary[(id >> n) & IDR_MASK];
		n -= IDR_BITS;
	}

	n = id & IDR_MASK;
	if (unlikely(p == NULL || !test_bit(n, &p->bitmap)))
		return ERR_PTR(-ENOENT);

	old_p = p->ary[n];
N
Nadia Derbey 已提交
834
	rcu_assign_pointer(p->ary[n], ptr);
J
Jeff Mahoney 已提交
835 836 837 838 839

	return old_p;
}
EXPORT_SYMBOL(idr_replace);

840
void __init idr_init_cache(void)
L
Linus Torvalds 已提交
841
{
842
	idr_layer_cache = kmem_cache_create("idr_layer_cache",
843
				sizeof(struct idr_layer), 0, SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
844 845 846 847 848 849 850 851 852 853 854 855 856 857 858
}

/**
 * idr_init - initialize idr handle
 * @idp:	idr handle
 *
 * This function is use to set up the handle (@idp) that you will pass
 * to the rest of the functions.
 */
void idr_init(struct idr *idp)
{
	memset(idp, 0, sizeof(struct idr));
	spin_lock_init(&idp->lock);
}
EXPORT_SYMBOL(idr_init);
859 860


861 862
/**
 * DOC: IDA description
863 864
 * IDA - IDR based ID allocator
 *
865
 * This is id allocator without id -> pointer translation.  Memory
866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897
 * usage is much lower than full blown idr because each id only
 * occupies a bit.  ida uses a custom leaf node which contains
 * IDA_BITMAP_BITS slots.
 *
 * 2007-04-25  written by Tejun Heo <htejun@gmail.com>
 */

static void free_bitmap(struct ida *ida, struct ida_bitmap *bitmap)
{
	unsigned long flags;

	if (!ida->free_bitmap) {
		spin_lock_irqsave(&ida->idr.lock, flags);
		if (!ida->free_bitmap) {
			ida->free_bitmap = bitmap;
			bitmap = NULL;
		}
		spin_unlock_irqrestore(&ida->idr.lock, flags);
	}

	kfree(bitmap);
}

/**
 * ida_pre_get - reserve resources for ida allocation
 * @ida:	ida handle
 * @gfp_mask:	memory allocation flag
 *
 * This function should be called prior to locking and calling the
 * following function.  It preallocates enough memory to satisfy the
 * worst possible allocation.
 *
898 899
 * If the system is REALLY out of memory this function returns %0,
 * otherwise %1.
900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924
 */
int ida_pre_get(struct ida *ida, gfp_t gfp_mask)
{
	/* allocate idr_layers */
	if (!idr_pre_get(&ida->idr, gfp_mask))
		return 0;

	/* allocate free_bitmap */
	if (!ida->free_bitmap) {
		struct ida_bitmap *bitmap;

		bitmap = kmalloc(sizeof(struct ida_bitmap), gfp_mask);
		if (!bitmap)
			return 0;

		free_bitmap(ida, bitmap);
	}

	return 1;
}
EXPORT_SYMBOL(ida_pre_get);

/**
 * ida_get_new_above - allocate new ID above or equal to a start id
 * @ida:	ida handle
N
Naohiro Aota 已提交
925
 * @starting_id: id to start search at
926 927
 * @p_id:	pointer to the allocated handle
 *
928 929
 * Allocate new ID above or equal to @starting_id.  It should be called
 * with any required locks.
930
 *
931
 * If memory is required, it will return %-EAGAIN, you should unlock
932
 * and go back to the ida_pre_get() call.  If the ida is full, it will
933
 * return %-ENOSPC.
934
 *
935
 * @p_id returns a value in the range @starting_id ... %0x7fffffff.
936 937 938
 */
int ida_get_new_above(struct ida *ida, int starting_id, int *p_id)
{
T
Tejun Heo 已提交
939
	struct idr_layer *pa[MAX_IDR_LEVEL + 1];
940 941 942 943 944 945 946 947
	struct ida_bitmap *bitmap;
	unsigned long flags;
	int idr_id = starting_id / IDA_BITMAP_BITS;
	int offset = starting_id % IDA_BITMAP_BITS;
	int t, id;

 restart:
	/* get vacant slot */
948
	t = idr_get_empty_slot(&ida->idr, idr_id, pa, 0, &ida->idr);
N
Nadia Derbey 已提交
949
	if (t < 0)
T
Tejun Heo 已提交
950
		return t == -ENOMEM ? -EAGAIN : t;
951

952
	if (t * IDA_BITMAP_BITS >= MAX_IDR_BIT)
953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970
		return -ENOSPC;

	if (t != idr_id)
		offset = 0;
	idr_id = t;

	/* if bitmap isn't there, create a new one */
	bitmap = (void *)pa[0]->ary[idr_id & IDR_MASK];
	if (!bitmap) {
		spin_lock_irqsave(&ida->idr.lock, flags);
		bitmap = ida->free_bitmap;
		ida->free_bitmap = NULL;
		spin_unlock_irqrestore(&ida->idr.lock, flags);

		if (!bitmap)
			return -EAGAIN;

		memset(bitmap, 0, sizeof(struct ida_bitmap));
N
Nadia Derbey 已提交
971 972
		rcu_assign_pointer(pa[0]->ary[idr_id & IDR_MASK],
				(void *)bitmap);
973 974 975 976 977 978 979 980 981 982 983 984 985
		pa[0]->count++;
	}

	/* lookup for empty slot */
	t = find_next_zero_bit(bitmap->bitmap, IDA_BITMAP_BITS, offset);
	if (t == IDA_BITMAP_BITS) {
		/* no empty slot after offset, continue to the next chunk */
		idr_id++;
		offset = 0;
		goto restart;
	}

	id = idr_id * IDA_BITMAP_BITS + t;
986
	if (id >= MAX_IDR_BIT)
987 988 989 990 991 992 993 994 995 996 997 998 999 1000
		return -ENOSPC;

	__set_bit(t, bitmap->bitmap);
	if (++bitmap->nr_busy == IDA_BITMAP_BITS)
		idr_mark_full(pa, idr_id);

	*p_id = id;

	/* Each leaf node can handle nearly a thousand slots and the
	 * whole idea of ida is to have small memory foot print.
	 * Throw away extra resources one by one after each successful
	 * allocation.
	 */
	if (ida->idr.id_free_cnt || ida->free_bitmap) {
1001
		struct idr_layer *p = get_from_free_list(&ida->idr);
1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059
		if (p)
			kmem_cache_free(idr_layer_cache, p);
	}

	return 0;
}
EXPORT_SYMBOL(ida_get_new_above);

/**
 * ida_remove - remove the given ID
 * @ida:	ida handle
 * @id:		ID to free
 */
void ida_remove(struct ida *ida, int id)
{
	struct idr_layer *p = ida->idr.top;
	int shift = (ida->idr.layers - 1) * IDR_BITS;
	int idr_id = id / IDA_BITMAP_BITS;
	int offset = id % IDA_BITMAP_BITS;
	int n;
	struct ida_bitmap *bitmap;

	/* clear full bits while looking up the leaf idr_layer */
	while ((shift > 0) && p) {
		n = (idr_id >> shift) & IDR_MASK;
		__clear_bit(n, &p->bitmap);
		p = p->ary[n];
		shift -= IDR_BITS;
	}

	if (p == NULL)
		goto err;

	n = idr_id & IDR_MASK;
	__clear_bit(n, &p->bitmap);

	bitmap = (void *)p->ary[n];
	if (!test_bit(offset, bitmap->bitmap))
		goto err;

	/* update bitmap and remove it if empty */
	__clear_bit(offset, bitmap->bitmap);
	if (--bitmap->nr_busy == 0) {
		__set_bit(n, &p->bitmap);	/* to please idr_remove() */
		idr_remove(&ida->idr, idr_id);
		free_bitmap(ida, bitmap);
	}

	return;

 err:
	printk(KERN_WARNING
	       "ida_remove called for id=%d which is not allocated.\n", id);
}
EXPORT_SYMBOL(ida_remove);

/**
 * ida_destroy - release all cached layers within an ida tree
N
Naohiro Aota 已提交
1060
 * @ida:		ida handle
1061 1062 1063 1064 1065 1066 1067 1068
 */
void ida_destroy(struct ida *ida)
{
	idr_destroy(&ida->idr);
	kfree(ida->free_bitmap);
}
EXPORT_SYMBOL(ida_destroy);

1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085
/**
 * ida_simple_get - get a new id.
 * @ida: the (initialized) ida.
 * @start: the minimum id (inclusive, < 0x8000000)
 * @end: the maximum id (exclusive, < 0x8000000 or 0)
 * @gfp_mask: memory allocation flags
 *
 * Allocates an id in the range start <= id < end, or returns -ENOSPC.
 * On memory allocation failure, returns -ENOMEM.
 *
 * Use ida_simple_remove() to get rid of an id.
 */
int ida_simple_get(struct ida *ida, unsigned int start, unsigned int end,
		   gfp_t gfp_mask)
{
	int ret, id;
	unsigned int max;
1086
	unsigned long flags;
1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101

	BUG_ON((int)start < 0);
	BUG_ON((int)end < 0);

	if (end == 0)
		max = 0x80000000;
	else {
		BUG_ON(end < start);
		max = end - 1;
	}

again:
	if (!ida_pre_get(ida, gfp_mask))
		return -ENOMEM;

1102
	spin_lock_irqsave(&simple_ida_lock, flags);
1103 1104 1105 1106 1107 1108 1109 1110 1111
	ret = ida_get_new_above(ida, start, &id);
	if (!ret) {
		if (id > max) {
			ida_remove(ida, id);
			ret = -ENOSPC;
		} else {
			ret = id;
		}
	}
1112
	spin_unlock_irqrestore(&simple_ida_lock, flags);
1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127

	if (unlikely(ret == -EAGAIN))
		goto again;

	return ret;
}
EXPORT_SYMBOL(ida_simple_get);

/**
 * ida_simple_remove - remove an allocated id.
 * @ida: the (initialized) ida.
 * @id: the id returned by ida_simple_get.
 */
void ida_simple_remove(struct ida *ida, unsigned int id)
{
1128 1129
	unsigned long flags;

1130
	BUG_ON((int)id < 0);
1131
	spin_lock_irqsave(&simple_ida_lock, flags);
1132
	ida_remove(ida, id);
1133
	spin_unlock_irqrestore(&simple_ida_lock, flags);
1134 1135 1136
}
EXPORT_SYMBOL(ida_simple_remove);

1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150
/**
 * ida_init - initialize ida handle
 * @ida:	ida handle
 *
 * This function is use to set up the handle (@ida) that you will pass
 * to the rest of the functions.
 */
void ida_init(struct ida *ida)
{
	memset(ida, 0, sizeof(struct ida));
	idr_init(&ida->idr);

}
EXPORT_SYMBOL(ida_init);