aboutsummaryrefslogtreecommitdiff
path: root/fs/file.c
blob: c8f1b0af8e0004c8ca901e8e6babb5877fd5cc21 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
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
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
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
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
/*
 *  linux/fs/file.c
 *
 *  Copyright (C) 1998-1999, Stephen Tweedie and Bill Hawes
 *
 *  Manage the dynamic fd arrays in the process files_struct.
 */

#include <linux/fs.h>
#include <linux/mm.h>
#include <linux/time.h>
#include <linux/slab.h>
#include <linux/vmalloc.h>
#include <linux/file.h>
#include <linux/bitops.h>
#include <linux/interrupt.h>
#include <linux/spinlock.h>
#include <linux/rcupdate.h>
#include <linux/workqueue.h>

struct fdtable_defer {
	spinlock_t lock;
	struct work_struct wq;
	struct timer_list timer;
	struct fdtable *next;
};

/*
 * We use this list to defer free fdtables that have vmalloced
 * sets/arrays. By keeping a per-cpu list, we avoid having to embed
 * the work_struct in fdtable itself which avoids a 64 byte (i386) increase in
 * this per-task structure.
 */
static DEFINE_PER_CPU(struct fdtable_defer, fdtable_defer_list);


/*
 * Allocate an fd array, using kmalloc or vmalloc.
 * Note: the array isn't cleared at allocation time.
 */
struct file ** alloc_fd_array(int num)
{
	struct file **new_fds;
	int size = num * sizeof(struct file *);

	if (size <= PAGE_SIZE)
		new_fds = (struct file **) kmalloc(size, GFP_KERNEL);
	else 
		new_fds = (struct file **) vmalloc(size);
	return new_fds;
}

void free_fd_array(struct file **array, int num)
{
	int size = num * sizeof(struct file *);

	if (!array) {
		printk (KERN_ERR "free_fd_array: array = 0 (num = %d)\n", num);
		return;
	}

	if (num <= NR_OPEN_DEFAULT) /* Don't free the embedded fd array! */
		return;
	else if (size <= PAGE_SIZE)
		kfree(array);
	else
		vfree(array);
}

static void __free_fdtable(struct fdtable *fdt)
{
	free_fdset(fdt->open_fds, fdt->max_fdset);
	free_fdset(fdt->close_on_exec, fdt->max_fdset);
	free_fd_array(fdt->fd, fdt->max_fds);
	kfree(fdt);
}

static void fdtable_timer(unsigned long data)
{
	struct fdtable_defer *fddef = (struct fdtable_defer *)data;

	spin_lock(&fddef->lock);
	/*
	 * If someone already emptied the queue return.
	 */
	if (!fddef->next)
		goto out;
	if (!schedule_work(&fddef->wq))
		mod_timer(&fddef->timer, 5);
out:
	spin_unlock(&fddef->lock);
}

static void free_fdtable_work(struct fdtable_defer *f)
{
	struct fdtable *fdt;

	spin_lock_bh(&f->lock);
	fdt = f->next;
	f->next = NULL;
	spin_unlock_bh(&f->lock);
	while(fdt) {
		struct fdtable *next = fdt->next;
		__free_fdtable(fdt);
		fdt = next;
	}
}

static void free_fdtable_rcu(struct rcu_head *rcu)
{
	struct fdtable *fdt = container_of(rcu, struct fdtable, rcu);
	int fdset_size, fdarray_size;
	struct fdtable_defer *fddef;

	BUG_ON(!fdt);
	fdset_size = fdt->max_fdset / 8;
	fdarray_size = fdt->max_fds * sizeof(struct file *);

	if (fdt->free_files) {
		/*
		 * The this fdtable was embedded in the files structure
		 * and the files structure itself was getting destroyed.
		 * It is now safe to free the files structure.
		 */
		kmem_cache_free(files_cachep, fdt->free_files);
		return;
	}
	if (fdt->max_fdset <= EMBEDDED_FD_SET_SIZE &&
		fdt->max_fds <= NR_OPEN_DEFAULT) {
		/*
		 * The fdtable was embedded
		 */
		return;
	}
	if (fdset_size <= PAGE_SIZE && fdarray_size <= PAGE_SIZE) {
		kfree(fdt->open_fds);
		kfree(fdt->close_on_exec);
		kfree(fdt->fd);
		kfree(fdt);
	} else {
		fddef = &get_cpu_var(fdtable_defer_list);
		spin_lock(&fddef->lock);
		fdt->next = fddef->next;
		fddef->next = fdt;
		/*
		 * vmallocs are handled from the workqueue context.
		 * If the per-cpu workqueue is running, then we
		 * defer work scheduling through a timer.
		 */
		if (!schedule_work(&fddef->wq))
			mod_timer(&fddef->timer, 5);
		spin_unlock(&fddef->lock);
		put_cpu_var(fdtable_defer_list);
	}
}

void free_fdtable(struct fdtable *fdt)
{
	if (fdt->free_files ||
		fdt->max_fdset > EMBEDDED_FD_SET_SIZE ||
		fdt->max_fds > NR_OPEN_DEFAULT)
		call_rcu(&fdt->rcu, free_fdtable_rcu);
}

/*
 * Expand the fdset in the files_struct.  Called with the files spinlock
 * held for write.
 */
static void copy_fdtable(struct fdtable *nfdt, struct fdtable *fdt)
{
	int i;
	int count;

	BUG_ON(nfdt->max_fdset < fdt->max_fdset);
	BUG_ON(nfdt->max_fds < fdt->max_fds);
	/* Copy the existing tables and install the new pointers */

	i = fdt->max_fdset / (sizeof(unsigned long) * 8);
	count = (nfdt->max_fdset - fdt->max_fdset) / 8;

	/*
	 * Don't copy the entire array if the current fdset is
	 * not yet initialised.
	 */
	if (i) {
		memcpy (nfdt->open_fds, fdt->open_fds,
						fdt->max_fdset/8);
		memcpy (nfdt->close_on_exec, fdt->close_on_exec,
						fdt->max_fdset/8);
		memset (&nfdt->open_fds->fds_bits[i], 0, count);
		memset (&nfdt->close_on_exec->fds_bits[i], 0, count);
	}

	/* Don't copy/clear the array if we are creating a new
	   fd array for fork() */
	if (fdt->max_fds) {
		memcpy(nfdt->fd, fdt->fd,
			fdt->max_fds * sizeof(struct file *));
		/* clear the remainder of the array */
		memset(&nfdt->fd[fdt->max_fds], 0,
		       (nfdt->max_fds - fdt->max_fds) *
					sizeof(struct file *));
	}
}

/*
 * Allocate an fdset array, using kmalloc or vmalloc.
 * Note: the array isn't cleared at allocation time.
 */
fd_set * alloc_fdset(int num)
{
	fd_set *new_fdset;
	int size = num / 8;

	if (size <= PAGE_SIZE)
		new_fdset = (fd_set *) kmalloc(size, GFP_KERNEL);
	else
		new_fdset = (fd_set *) vmalloc(size);
	return new_fdset;
}

void free_fdset(fd_set *array, int num)
{
	if (num <= EMBEDDED_FD_SET_SIZE) /* Don't free an embedded fdset */
		return;
	else if (num <= 8 * PAGE_SIZE)
		kfree(array);
	else
		vfree(array);
}

static struct fdtable *alloc_fdtable(int nr)
{
	struct fdtable *fdt = NULL;
	int nfds = 0;
  	fd_set *new_openset = NULL, *new_execset = NULL;
	struct file **new_fds;

	fdt = kzalloc(sizeof(*fdt), GFP_KERNEL);
	if (!fdt)
  		goto out;

	nfds = max_t(int, 8 * L1_CACHE_BYTES, roundup_pow_of_two(nfds));
	if (nfds > NR_OPEN)
		nfds = NR_OPEN;

  	new_openset = alloc_fdset(nfds);
  	new_execset = alloc_fdset(nfds);
  	if (!new_openset || !new_execset)
  		goto out;
	fdt->open_fds = new_openset;
	fdt->close_on_exec = new_execset;
	fdt->max_fdset = nfds;

	nfds = NR_OPEN_DEFAULT;
	/*
	 * Expand to the max in easy steps, and keep expanding it until
	 * we have enough for the requested fd array size.
	 */
	do {
#if NR_OPEN_DEFAULT < 256
		if (nfds < 256)
			nfds = 256;
		else
#endif
		if (nfds < (PAGE_SIZE / sizeof(struct file *)))
			nfds = PAGE_SIZE / sizeof(struct file *);
		else {
			nfds = nfds * 2;
			if (nfds > NR_OPEN)
				nfds = NR_OPEN;
  		}
	} while (nfds <= nr);
	new_fds = alloc_fd_array(nfds);
	if (!new_fds)
		goto out2;
	fdt->fd = new_fds;
	fdt->max_fds = nfds;
	fdt->free_files = NULL;
	return fdt;
out2:
	nfds = fdt->max_fdset;
out:
  	if (new_openset)
  		free_fdset(new_openset, nfds);
  	if (new_execset)
  		free_fdset(new_execset, nfds);
	kfree(fdt);
	return NULL;
}

/*
 * Expands the file descriptor table - it will allocate a new fdtable and
 * both fd array and fdset. It is expected to be called with the
 * files_lock held.
 */
static int expand_fdtable(struct files_struct *files, int nr)
	__releases(files->file_lock)
	__acquires(files->file_lock)
{
	int error = 0;
	struct fdtable *fdt;
	struct fdtable *nfdt = NULL;

	spin_unlock(&files->file_lock);
	nfdt = alloc_fdtable(nr);
	if (!nfdt) {
		error = -ENOMEM;
		spin_lock(&files->file_lock);
		goto out;
	}

	spin_lock(&files->file_lock);
	fdt = files_fdtable(files);
	/*
	 * Check again since another task may have expanded the
	 * fd table while we dropped the lock
	 */
	if (nr >= fdt->max_fds || nr >= fdt->max_fdset) {
		copy_fdtable(nfdt, fdt);
	} else {
		/* Somebody expanded while we dropped file_lock */
		spin_unlock(&files->file_lock);
		__free_fdtable(nfdt);
		spin_lock(&files->file_lock);
		goto out;
	}
	rcu_assign_pointer(files->fdt, nfdt);
	free_fdtable(fdt);
out:
	return error;
}

/*
 * Expand files.
 * Return <0 on error; 0 nothing done; 1 files expanded, we may have blocked.
 * Should be called with the files->file_lock spinlock held for write.
 */
int expand_files(struct files_struct *files, int nr)
{
	int err, expand = 0;
	struct fdtable *fdt;

	fdt = files_fdtable(files);
	if (nr >= fdt->max_fdset || nr >= fdt->max_fds) {
		if (fdt->max_fdset >= NR_OPEN ||
			fdt->max_fds >= NR_OPEN || nr >= NR_OPEN) {
			err = -EMFILE;
			goto out;
		}
		expand = 1;
		if ((err = expand_fdtable(files, nr)))
			goto out;
	}
	err = expand;
out:
	return err;
}

static void __devinit fdtable_defer_list_init(int cpu)
{
	struct fdtable_defer *fddef = &per_cpu(fdtable_defer_list, cpu);
	spin_lock_init(&fddef->lock);
	INIT_WORK(&fddef->wq, (void (*)(void *))free_fdtable_work, fddef);
	init_timer(&fddef->timer);
	fddef->timer.data = (unsigned long)fddef;
	fddef->timer.function = fdtable_timer;
	fddef->next = NULL;
}

void __init files_defer_init(void)
{
	int i;
	for_each_possible_cpu(i)
		fdtable_defer_list_init(i);
}