summaryrefslogtreecommitdiffstats
path: root/sys/kern/kern_alq.c
blob: 512f358b936531abaa2d13381811bce860edc32d (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
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
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
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
/*-
 * Copyright (c) 2002, Jeffrey Roberson <jeff@freebsd.org>
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice unmodified, this list of conditions, and the following
 *    disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");

#include "opt_mac.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/kthread.h>
#include <sys/lock.h>
#include <sys/mount.h>
#include <sys/mutex.h>
#include <sys/namei.h>
#include <sys/proc.h>
#include <sys/vnode.h>
#include <sys/alq.h>
#include <sys/malloc.h>
#include <sys/unistd.h>
#include <sys/fcntl.h>
#include <sys/eventhandler.h>

#include <security/mac/mac_framework.h>

/* Async. Logging Queue */
struct alq {
	int	aq_entmax;		/* Max entries */
	int	aq_entlen;		/* Entry length */
	char	*aq_entbuf;		/* Buffer for stored entries */
	int	aq_flags;		/* Queue flags */
	struct mtx	aq_mtx;		/* Queue lock */
	struct vnode	*aq_vp;		/* Open vnode handle */
	struct ucred	*aq_cred;	/* Credentials of the opening thread */
	struct ale	*aq_first;	/* First ent */
	struct ale	*aq_entfree;	/* First free ent */
	struct ale	*aq_entvalid;	/* First ent valid for writing */
	LIST_ENTRY(alq)	aq_act;		/* List of active queues */
	LIST_ENTRY(alq)	aq_link;	/* List of all queues */
};

#define	AQ_WANTED	0x0001		/* Wakeup sleeper when io is done */
#define	AQ_ACTIVE	0x0002		/* on the active list */
#define	AQ_FLUSHING	0x0004		/* doing IO */
#define	AQ_SHUTDOWN	0x0008		/* Queue no longer valid */

#define	ALQ_LOCK(alq)	mtx_lock_spin(&(alq)->aq_mtx)
#define	ALQ_UNLOCK(alq)	mtx_unlock_spin(&(alq)->aq_mtx)

static MALLOC_DEFINE(M_ALD, "ALD", "ALD");

/*
 * The ald_mtx protects the ald_queues list and the ald_active list.
 */
static struct mtx ald_mtx;
static LIST_HEAD(, alq) ald_queues;
static LIST_HEAD(, alq) ald_active;
static int ald_shutingdown = 0;
struct thread *ald_thread;
static struct proc *ald_proc;

#define	ALD_LOCK()	mtx_lock(&ald_mtx)
#define	ALD_UNLOCK()	mtx_unlock(&ald_mtx)

/* Daemon functions */
static int ald_add(struct alq *);
static int ald_rem(struct alq *);
static void ald_startup(void *);
static void ald_daemon(void);
static void ald_shutdown(void *, int);
static void ald_activate(struct alq *);
static void ald_deactivate(struct alq *);

/* Internal queue functions */
static void alq_shutdown(struct alq *);
static int alq_doio(struct alq *);


/*
 * Add a new queue to the global list.  Fail if we're shutting down.
 */
static int
ald_add(struct alq *alq)
{
	int error;

	error = 0;

	ALD_LOCK();
	if (ald_shutingdown) {
		error = EBUSY;
		goto done;
	}
	LIST_INSERT_HEAD(&ald_queues, alq, aq_link);
done:
	ALD_UNLOCK();
	return (error);
}

/*
 * Remove a queue from the global list unless we're shutting down.  If so,
 * the ald will take care of cleaning up it's resources.
 */
static int
ald_rem(struct alq *alq)
{
	int error;

	error = 0;

	ALD_LOCK();
	if (ald_shutingdown) {
		error = EBUSY;
		goto done;
	}
	LIST_REMOVE(alq, aq_link);
done:
	ALD_UNLOCK();
	return (error);
}

/*
 * Put a queue on the active list.  This will schedule it for writing.
 */
static void
ald_activate(struct alq *alq)
{
	LIST_INSERT_HEAD(&ald_active, alq, aq_act);
	wakeup(&ald_active);
}

static void
ald_deactivate(struct alq *alq)
{
	LIST_REMOVE(alq, aq_act);
	alq->aq_flags &= ~AQ_ACTIVE;
}

static void
ald_startup(void *unused)
{
	mtx_init(&ald_mtx, "ALDmtx", NULL, MTX_DEF|MTX_QUIET);
	LIST_INIT(&ald_queues);
	LIST_INIT(&ald_active);
}

static void
ald_daemon(void)
{
	int needwakeup;
	struct alq *alq;

	ald_thread = FIRST_THREAD_IN_PROC(ald_proc);

	EVENTHANDLER_REGISTER(shutdown_pre_sync, ald_shutdown, NULL,
	    SHUTDOWN_PRI_FIRST);

	ALD_LOCK();

	for (;;) {
		while ((alq = LIST_FIRST(&ald_active)) == NULL)
			msleep(&ald_active, &ald_mtx, PWAIT, "aldslp", 0);

		ALQ_LOCK(alq);
		ald_deactivate(alq);
		ALD_UNLOCK();
		needwakeup = alq_doio(alq);
		ALQ_UNLOCK(alq);
		if (needwakeup)
			wakeup(alq);
		ALD_LOCK();
	}
}

static void
ald_shutdown(void *arg, int howto)
{
	struct alq *alq;

	ALD_LOCK();
	ald_shutingdown = 1;

	while ((alq = LIST_FIRST(&ald_queues)) != NULL) {
		LIST_REMOVE(alq, aq_link);
		ALD_UNLOCK();
		alq_shutdown(alq);
		ALD_LOCK();
	}
	ALD_UNLOCK();
}

static void
alq_shutdown(struct alq *alq)
{
	ALQ_LOCK(alq);

	/* Stop any new writers. */
	alq->aq_flags |= AQ_SHUTDOWN;

	/* Drain IO */
	while (alq->aq_flags & (AQ_FLUSHING|AQ_ACTIVE)) {
		alq->aq_flags |= AQ_WANTED;
		msleep_spin(alq, &alq->aq_mtx, "aldclose", 0);
	}
	ALQ_UNLOCK(alq);

	vn_close(alq->aq_vp, FWRITE, alq->aq_cred,
	    curthread);
	crfree(alq->aq_cred);
}

/*
 * Flush all pending data to disk.  This operation will block.
 */
static int
alq_doio(struct alq *alq)
{
	struct thread *td;
	struct mount *mp;
	struct vnode *vp;
	struct uio auio;
	struct iovec aiov[2];
	struct ale *ale;
	struct ale *alstart;
	int totlen;
	int iov;
	int vfslocked;

	vp = alq->aq_vp;
	td = curthread;
	totlen = 0;
	iov = 0;

	alstart = ale = alq->aq_entvalid;
	alq->aq_entvalid = NULL;

	bzero(&aiov, sizeof(aiov));
	bzero(&auio, sizeof(auio));

	do {
		if (aiov[iov].iov_base == NULL)
			aiov[iov].iov_base = ale->ae_data;
		aiov[iov].iov_len += alq->aq_entlen;
		totlen += alq->aq_entlen;
		/* Check to see if we're wrapping the buffer */
		if (ale->ae_data + alq->aq_entlen != ale->ae_next->ae_data)
			iov++;
		ale->ae_flags &= ~AE_VALID;
		ale = ale->ae_next;
	} while (ale->ae_flags & AE_VALID);

	alq->aq_flags |= AQ_FLUSHING;
	ALQ_UNLOCK(alq);

	if (iov == 2 || aiov[iov].iov_base == NULL)
		iov--;

	auio.uio_iov = &aiov[0];
	auio.uio_offset = 0;
	auio.uio_segflg = UIO_SYSSPACE;
	auio.uio_rw = UIO_WRITE;
	auio.uio_iovcnt = iov + 1;
	auio.uio_resid = totlen;
	auio.uio_td = td;

	/*
	 * Do all of the junk required to write now.
	 */
	vfslocked = VFS_LOCK_GIANT(vp->v_mount);
	vn_start_write(vp, &mp, V_WAIT);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
	VOP_LEASE(vp, td, alq->aq_cred, LEASE_WRITE);
	/*
	 * XXX: VOP_WRITE error checks are ignored.
	 */
#ifdef MAC
	if (mac_vnode_check_write(alq->aq_cred, NOCRED, vp) == 0)
#endif
		VOP_WRITE(vp, &auio, IO_UNIT | IO_APPEND, alq->aq_cred);
	VOP_UNLOCK(vp, 0);
	vn_finished_write(mp);
	VFS_UNLOCK_GIANT(vfslocked);

	ALQ_LOCK(alq);
	alq->aq_flags &= ~AQ_FLUSHING;

	if (alq->aq_entfree == NULL)
		alq->aq_entfree = alstart;

	if (alq->aq_flags & AQ_WANTED) {
		alq->aq_flags &= ~AQ_WANTED;
		return (1);
	}

	return(0);
}

static struct kproc_desc ald_kp = {
        "ALQ Daemon",
        ald_daemon,
        &ald_proc
};

SYSINIT(aldthread, SI_SUB_KTHREAD_IDLE, SI_ORDER_ANY, kproc_start, &ald_kp);
SYSINIT(ald, SI_SUB_LOCK, SI_ORDER_ANY, ald_startup, NULL);


/* User visible queue functions */

/*
 * Create the queue data structure, allocate the buffer, and open the file.
 */
int
alq_open(struct alq **alqp, const char *file, struct ucred *cred, int cmode,
    int size, int count)
{
	struct thread *td;
	struct nameidata nd;
	struct ale *ale;
	struct ale *alp;
	struct alq *alq;
	char *bufp;
	int flags;
	int error;
	int i, vfslocked;

	*alqp = NULL;
	td = curthread;

	NDINIT(&nd, LOOKUP, NOFOLLOW | MPSAFE, UIO_SYSSPACE, file, td);
	flags = FWRITE | O_NOFOLLOW | O_CREAT;

	error = vn_open_cred(&nd, &flags, cmode, cred, NULL);
	if (error)
		return (error);

	vfslocked = NDHASGIANT(&nd);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	/* We just unlock so we hold a reference */
	VOP_UNLOCK(nd.ni_vp, 0);
	VFS_UNLOCK_GIANT(vfslocked);

	alq = malloc(sizeof(*alq), M_ALD, M_WAITOK|M_ZERO);
	alq->aq_entbuf = malloc(count * size, M_ALD, M_WAITOK|M_ZERO);
	alq->aq_first = malloc(sizeof(*ale) * count, M_ALD, M_WAITOK|M_ZERO);
	alq->aq_vp = nd.ni_vp;
	alq->aq_cred = crhold(cred);
	alq->aq_entmax = count;
	alq->aq_entlen = size;
	alq->aq_entfree = alq->aq_first;

	mtx_init(&alq->aq_mtx, "ALD Queue", NULL, MTX_SPIN|MTX_QUIET);

	bufp = alq->aq_entbuf;
	ale = alq->aq_first;
	alp = NULL;

	/* Match up entries with buffers */
	for (i = 0; i < count; i++) {
		if (alp)
			alp->ae_next = ale;
		ale->ae_data = bufp;
		alp = ale;
		ale++;
		bufp += size;
	}

	alp->ae_next = alq->aq_first;

	if ((error = ald_add(alq)) != 0)
		return (error);
	*alqp = alq;

	return (0);
}

/*
 * Copy a new entry into the queue.  If the operation would block either
 * wait or return an error depending on the value of waitok.
 */
int
alq_write(struct alq *alq, void *data, int waitok)
{
	struct ale *ale;

	if ((ale = alq_get(alq, waitok)) == NULL)
		return (EWOULDBLOCK);

	bcopy(data, ale->ae_data, alq->aq_entlen);
	alq_post(alq, ale);

	return (0);
}

struct ale *
alq_get(struct alq *alq, int waitok)
{
	struct ale *ale;
	struct ale *aln;

	ale = NULL;

	ALQ_LOCK(alq);

	/* Loop until we get an entry or we're shutting down */
	while ((alq->aq_flags & AQ_SHUTDOWN) == 0 && 
	    (ale = alq->aq_entfree) == NULL &&
	    (waitok & ALQ_WAITOK)) {
		alq->aq_flags |= AQ_WANTED;
		msleep_spin(alq, &alq->aq_mtx, "alqget", 0);
	}

	if (ale != NULL) {
		aln = ale->ae_next;
		if ((aln->ae_flags & AE_VALID) == 0)
			alq->aq_entfree = aln;
		else
			alq->aq_entfree = NULL;
	} else
		ALQ_UNLOCK(alq);


	return (ale);
}

void
alq_post(struct alq *alq, struct ale *ale)
{
	int activate;

	ale->ae_flags |= AE_VALID;

	if (alq->aq_entvalid == NULL)
		alq->aq_entvalid = ale;

	if ((alq->aq_flags & AQ_ACTIVE) == 0) {
		alq->aq_flags |= AQ_ACTIVE;
		activate = 1;
	} else
		activate = 0;

	ALQ_UNLOCK(alq);
	if (activate) {
		ALD_LOCK();
		ald_activate(alq);
		ALD_UNLOCK();
	}
}

void
alq_flush(struct alq *alq)
{
	int needwakeup = 0;

	ALD_LOCK();
	ALQ_LOCK(alq);
	if (alq->aq_flags & AQ_ACTIVE) {
		ald_deactivate(alq);
		ALD_UNLOCK();
		needwakeup = alq_doio(alq);
	} else
		ALD_UNLOCK();
	ALQ_UNLOCK(alq);

	if (needwakeup)
		wakeup(alq);
}

/*
 * Flush remaining data, close the file and free all resources.
 */
void
alq_close(struct alq *alq)
{
	/*
	 * If we're already shuting down someone else will flush and close
	 * the vnode.
	 */
	if (ald_rem(alq) != 0)
		return;

	/*
	 * Drain all pending IO.
	 */
	alq_shutdown(alq);

	mtx_destroy(&alq->aq_mtx);
	free(alq->aq_first, M_ALD);
	free(alq->aq_entbuf, M_ALD);
	free(alq, M_ALD);
}
OpenPOWER on IntegriCloud