scsi_lib.c 62.7 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1
2
3
4
5
6
7
8
9
10
/*
 *  scsi_lib.c Copyright (C) 1999 Eric Youngdale
 *
 *  SCSI queueing library.
 *      Initial versions: Eric Youngdale (eric@andante.org).
 *                        Based upon conversations with large numbers
 *                        of people at Linux Expo.
 */

#include <linux/bio.h>
11
#include <linux/bitops.h>
Linus Torvalds's avatar
Linus Torvalds committed
12
13
14
15
16
17
18
19
#include <linux/blkdev.h>
#include <linux/completion.h>
#include <linux/kernel.h>
#include <linux/mempool.h>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/pci.h>
#include <linux/delay.h>
20
#include <linux/hardirq.h>
Jens Axboe's avatar
Jens Axboe committed
21
#include <linux/scatterlist.h>
Linus Torvalds's avatar
Linus Torvalds committed
22
23

#include <scsi/scsi.h>
24
#include <scsi/scsi_cmnd.h>
Linus Torvalds's avatar
Linus Torvalds committed
25
26
27
28
29
30
31
32
33
34
#include <scsi/scsi_dbg.h>
#include <scsi/scsi_device.h>
#include <scsi/scsi_driver.h>
#include <scsi/scsi_eh.h>
#include <scsi/scsi_host.h>

#include "scsi_priv.h"
#include "scsi_logging.h"


35
#define SG_MEMPOOL_NR		ARRAY_SIZE(scsi_sg_pools)
36
#define SG_MEMPOOL_SIZE		2
Linus Torvalds's avatar
Linus Torvalds committed
37
38
39

struct scsi_host_sg_pool {
	size_t		size;
40
	char		*name;
41
	struct kmem_cache	*slab;
Linus Torvalds's avatar
Linus Torvalds committed
42
43
44
	mempool_t	*pool;
};

45
46
47
48
#define SP(x) { x, "sgpool-" __stringify(x) }
#if (SCSI_MAX_SG_SEGMENTS < 32)
#error SCSI_MAX_SG_SEGMENTS is too small (must be 32 or greater)
#endif
49
static struct scsi_host_sg_pool scsi_sg_pools[] = {
Linus Torvalds's avatar
Linus Torvalds committed
50
51
	SP(8),
	SP(16),
52
#if (SCSI_MAX_SG_SEGMENTS > 32)
53
	SP(32),
54
#if (SCSI_MAX_SG_SEGMENTS > 64)
55
56
	SP(64),
#if (SCSI_MAX_SG_SEGMENTS > 128)
Linus Torvalds's avatar
Linus Torvalds committed
57
	SP(128),
58
59
#if (SCSI_MAX_SG_SEGMENTS > 256)
#error SCSI_MAX_SG_SEGMENTS is too large (256 MAX)
60
61
62
#endif
#endif
#endif
63
64
#endif
	SP(SCSI_MAX_SG_SEGMENTS)
65
};
Linus Torvalds's avatar
Linus Torvalds committed
66
67
#undef SP

68
69
static struct kmem_cache *scsi_bidi_sdb_cache;

70
static void scsi_run_queue(struct request_queue *q);
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87

/*
 * Function:	scsi_unprep_request()
 *
 * Purpose:	Remove all preparation done for a request, including its
 *		associated scsi_cmnd, so that it can be requeued.
 *
 * Arguments:	req	- request to unprepare
 *
 * Lock status:	Assumed that no locks are held upon entry.
 *
 * Returns:	Nothing.
 */
static void scsi_unprep_request(struct request *req)
{
	struct scsi_cmnd *cmd = req->special;

88
	req->cmd_flags &= ~REQ_DONTPREP;
89
	req->special = NULL;
90
91
92

	scsi_put_command(cmd);
}
93

Linus Torvalds's avatar
Linus Torvalds committed
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
/*
 * Function:    scsi_queue_insert()
 *
 * Purpose:     Insert a command in the midlevel queue.
 *
 * Arguments:   cmd    - command that we are adding to queue.
 *              reason - why we are inserting command to queue.
 *
 * Lock status: Assumed that lock is not held upon entry.
 *
 * Returns:     Nothing.
 *
 * Notes:       We do this for one of two cases.  Either the host is busy
 *              and it cannot accept any more commands for the time being,
 *              or the device returned QUEUE_FULL and can accept no more
 *              commands.
 * Notes:       This could be called either from an interrupt context or a
 *              normal process context.
 */
int scsi_queue_insert(struct scsi_cmnd *cmd, int reason)
{
	struct Scsi_Host *host = cmd->device->host;
	struct scsi_device *device = cmd->device;
117
118
	struct request_queue *q = device->request_queue;
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
119
120
121
122
123

	SCSI_LOG_MLQUEUE(1,
		 printk("Inserting command %p into mlqueue\n", cmd));

	/*
124
	 * Set the appropriate busy bit for the device/host.
Linus Torvalds's avatar
Linus Torvalds committed
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
	 *
	 * If the host/device isn't busy, assume that something actually
	 * completed, and that we should be able to queue a command now.
	 *
	 * Note that the prior mid-layer assumption that any host could
	 * always queue at least one command is now broken.  The mid-layer
	 * will implement a user specifiable stall (see
	 * scsi_host.max_host_blocked and scsi_device.max_device_blocked)
	 * if a command is requeued with no other commands outstanding
	 * either for the device or for the host.
	 */
	if (reason == SCSI_MLQUEUE_HOST_BUSY)
		host->host_blocked = host->max_host_blocked;
	else if (reason == SCSI_MLQUEUE_DEVICE_BUSY)
		device->device_blocked = device->max_device_blocked;

	/*
	 * Decrement the counters, since these commands are no longer
	 * active on the host/device.
	 */
	scsi_device_unbusy(device);

	/*
148
149
	 * Requeue this command.  It will go before all other commands
	 * that are already in the queue.
Linus Torvalds's avatar
Linus Torvalds committed
150
151
152
153
	 *
	 * NOTE: there is magic here about the way the queue is plugged if
	 * we have no outstanding commands.
	 * 
154
	 * Although we *don't* plug the queue, we call the request
Linus Torvalds's avatar
Linus Torvalds committed
155
156
	 * function.  The SCSI request function detects the blocked condition
	 * and plugs the queue appropriately.
157
158
         */
	spin_lock_irqsave(q->queue_lock, flags);
159
	blk_requeue_request(q, cmd->request);
160
161
162
163
	spin_unlock_irqrestore(q->queue_lock, flags);

	scsi_run_queue(q);

Linus Torvalds's avatar
Linus Torvalds committed
164
165
166
	return 0;
}

167
/**
168
 * scsi_execute - insert request and wait for the result
169
170
171
172
173
174
175
176
 * @sdev:	scsi device
 * @cmd:	scsi command
 * @data_direction: data direction
 * @buffer:	data buffer
 * @bufflen:	len of buffer
 * @sense:	optional sense buffer
 * @timeout:	request timeout in seconds
 * @retries:	number of times to retry request
177
 * @flags:	or into request flags;
178
 *
179
 * returns the req->errors value which is the scsi_cmnd result
180
 * field.
181
 */
182
183
184
int scsi_execute(struct scsi_device *sdev, const unsigned char *cmd,
		 int data_direction, void *buffer, unsigned bufflen,
		 unsigned char *sense, int timeout, int retries, int flags)
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
{
	struct request *req;
	int write = (data_direction == DMA_TO_DEVICE);
	int ret = DRIVER_ERROR << 24;

	req = blk_get_request(sdev->request_queue, write, __GFP_WAIT);

	if (bufflen &&	blk_rq_map_kern(sdev->request_queue, req,
					buffer, bufflen, __GFP_WAIT))
		goto out;

	req->cmd_len = COMMAND_SIZE(cmd[0]);
	memcpy(req->cmd, cmd, req->cmd_len);
	req->sense = sense;
	req->sense_len = 0;
200
	req->retries = retries;
201
	req->timeout = timeout;
202
203
	req->cmd_type = REQ_TYPE_BLOCK_PC;
	req->cmd_flags |= flags | REQ_QUIET | REQ_PREEMPT;
204
205
206
207
208
209
210
211
212
213
214
215

	/*
	 * head injection *required* here otherwise quiesce won't work
	 */
	blk_execute_rq(req->q, NULL, req, 1);

	ret = req->errors;
 out:
	blk_put_request(req);

	return ret;
}
216
EXPORT_SYMBOL(scsi_execute);
217

218
219
220
221
222
223

int scsi_execute_req(struct scsi_device *sdev, const unsigned char *cmd,
		     int data_direction, void *buffer, unsigned bufflen,
		     struct scsi_sense_hdr *sshdr, int timeout, int retries)
{
	char *sense = NULL;
224
225
	int result;
	
226
	if (sshdr) {
227
		sense = kzalloc(SCSI_SENSE_BUFFERSIZE, GFP_NOIO);
228
229
230
		if (!sense)
			return DRIVER_ERROR << 24;
	}
231
	result = scsi_execute(sdev, cmd, data_direction, buffer, bufflen,
232
			      sense, timeout, retries, 0);
233
	if (sshdr)
234
		scsi_normalize_sense(sense, SCSI_SENSE_BUFFERSIZE, sshdr);
235
236
237
238
239
240

	kfree(sense);
	return result;
}
EXPORT_SYMBOL(scsi_execute_req);

241
242
243
244
245
246
struct scsi_io_context {
	void *data;
	void (*done)(void *data, char *sense, int result, int resid);
	char sense[SCSI_SENSE_BUFFERSIZE];
};

247
static struct kmem_cache *scsi_io_context_cache;
248

249
static void scsi_end_async(struct request *req, int uptodate)
250
251
252
253
254
255
{
	struct scsi_io_context *sioc = req->end_io_data;

	if (sioc->done)
		sioc->done(sioc->data, sioc->sense, req->errors, req->data_len);

256
	kmem_cache_free(scsi_io_context_cache, sioc);
257
258
259
260
261
262
263
264
265
266
267
268
	__blk_put_request(req->q, req);
}

static int scsi_merge_bio(struct request *rq, struct bio *bio)
{
	struct request_queue *q = rq->q;

	bio->bi_flags &= ~(1 << BIO_SEG_VALID);
	if (rq_data_dir(rq) == WRITE)
		bio->bi_rw |= (1 << BIO_RW);
	blk_queue_bounce(q, &bio);

NeilBrown's avatar
NeilBrown committed
269
	return blk_rq_append_bio(q, rq, bio);
270
271
}

272
static void scsi_bi_endio(struct bio *bio, int error)
273
274
275
276
277
278
279
{
	bio_put(bio);
}

/**
 * scsi_req_map_sg - map a scatterlist into a request
 * @rq:		request to fill
280
 * @sgl:	scatterlist
281
282
283
284
285
286
287
288
289
290
291
292
 * @nsegs:	number of elements
 * @bufflen:	len of buffer
 * @gfp:	memory allocation flags
 *
 * scsi_req_map_sg maps a scatterlist into a request so that the
 * request can be sent to the block layer. We do not trust the scatterlist
 * sent to use, as some ULDs use that struct to only organize the pages.
 */
static int scsi_req_map_sg(struct request *rq, struct scatterlist *sgl,
			   int nsegs, unsigned bufflen, gfp_t gfp)
{
	struct request_queue *q = rq->q;
293
	int nr_pages = (bufflen + sgl[0].offset + PAGE_SIZE - 1) >> PAGE_SHIFT;
294
	unsigned int data_len = bufflen, len, bytes, off;
Jens Axboe's avatar
Jens Axboe committed
295
	struct scatterlist *sg;
296
297
298
299
	struct page *page;
	struct bio *bio = NULL;
	int i, err, nr_vecs = 0;

Jens Axboe's avatar
Jens Axboe committed
300
	for_each_sg(sgl, sg, nsegs, i) {
301
		page = sg_page(sg);
Jens Axboe's avatar
Jens Axboe committed
302
303
		off = sg->offset;
		len = sg->length;
304

305
306
307
308
309
310
		while (len > 0 && data_len > 0) {
			/*
			 * sg sends a scatterlist that is larger than
			 * the data_len it wants transferred for certain
			 * IO sizes
			 */
311
			bytes = min_t(unsigned int, len, PAGE_SIZE - off);
312
			bytes = min(bytes, data_len);
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335

			if (!bio) {
				nr_vecs = min_t(int, BIO_MAX_PAGES, nr_pages);
				nr_pages -= nr_vecs;

				bio = bio_alloc(gfp, nr_vecs);
				if (!bio) {
					err = -ENOMEM;
					goto free_bios;
				}
				bio->bi_end_io = scsi_bi_endio;
			}

			if (bio_add_pc_page(q, bio, page, bytes, off) !=
			    bytes) {
				bio_put(bio);
				err = -EINVAL;
				goto free_bios;
			}

			if (bio->bi_vcnt >= nr_vecs) {
				err = scsi_merge_bio(rq, bio);
				if (err) {
336
					bio_endio(bio, 0);
337
338
339
340
341
342
343
					goto free_bios;
				}
				bio = NULL;
			}

			page++;
			len -= bytes;
344
			data_len -=bytes;
345
346
347
348
349
			off = 0;
		}
	}

	rq->buffer = rq->data = NULL;
350
	rq->data_len = bufflen;
351
352
353
354
355
356
357
358
	return 0;

free_bios:
	while ((bio = rq->bio) != NULL) {
		rq->bio = bio->bi_next;
		/*
		 * call endio instead of bio_put incase it was bounced
		 */
359
		bio_endio(bio, 0);
360
361
362
363
364
365
366
367
368
	}

	return err;
}

/**
 * scsi_execute_async - insert request
 * @sdev:	scsi device
 * @cmd:	scsi command
369
 * @cmd_len:	length of scsi cdb
370
 * @data_direction: DMA_TO_DEVICE, DMA_FROM_DEVICE, or DMA_NONE
371
372
373
374
375
 * @buffer:	data buffer (this can be a kernel buffer or scatterlist)
 * @bufflen:	len of buffer
 * @use_sg:	if buffer is a scatterlist this is the number of elements
 * @timeout:	request timeout in seconds
 * @retries:	number of times to retry request
376
377
378
379
 * @privdata:	data passed to done()
 * @done:	callback function when done
 * @gfp:	memory allocation flags
 */
380
int scsi_execute_async(struct scsi_device *sdev, const unsigned char *cmd,
381
		       int cmd_len, int data_direction, void *buffer, unsigned bufflen,
382
383
384
385
386
387
388
389
		       int use_sg, int timeout, int retries, void *privdata,
		       void (*done)(void *, char *, int, int), gfp_t gfp)
{
	struct request *req;
	struct scsi_io_context *sioc;
	int err = 0;
	int write = (data_direction == DMA_TO_DEVICE);

390
	sioc = kmem_cache_zalloc(scsi_io_context_cache, gfp);
391
392
393
394
395
396
	if (!sioc)
		return DRIVER_ERROR << 24;

	req = blk_get_request(sdev->request_queue, write, gfp);
	if (!req)
		goto free_sense;
397
398
	req->cmd_type = REQ_TYPE_BLOCK_PC;
	req->cmd_flags |= REQ_QUIET;
399
400
401
402
403
404
405
406
407

	if (use_sg)
		err = scsi_req_map_sg(req, buffer, use_sg, bufflen, gfp);
	else if (bufflen)
		err = blk_rq_map_kern(req->q, req, buffer, bufflen, gfp);

	if (err)
		goto free_req;

408
	req->cmd_len = cmd_len;
409
	memset(req->cmd, 0, BLK_MAX_CDB); /* ATAPI hates garbage after CDB */
410
411
412
413
	memcpy(req->cmd, cmd, req->cmd_len);
	req->sense = sioc->sense;
	req->sense_len = 0;
	req->timeout = timeout;
414
	req->retries = retries;
415
416
417
418
419
420
421
422
423
424
425
	req->end_io_data = sioc;

	sioc->data = privdata;
	sioc->done = done;

	blk_execute_rq_nowait(req->q, NULL, req, 1, scsi_end_async);
	return 0;

free_req:
	blk_put_request(req);
free_sense:
426
	kmem_cache_free(scsi_io_context_cache, sioc);
427
428
429
430
	return DRIVER_ERROR << 24;
}
EXPORT_SYMBOL_GPL(scsi_execute_async);

Linus Torvalds's avatar
Linus Torvalds committed
431
432
433
434
435
436
437
438
439
440
441
/*
 * Function:    scsi_init_cmd_errh()
 *
 * Purpose:     Initialize cmd fields related to error handling.
 *
 * Arguments:   cmd	- command that is ready to be queued.
 *
 * Notes:       This function has the job of initializing a number of
 *              fields related to error handling.   Typically this will
 *              be called once for each command, as required.
 */
442
static void scsi_init_cmd_errh(struct scsi_cmnd *cmd)
Linus Torvalds's avatar
Linus Torvalds committed
443
444
{
	cmd->serial_number = 0;
445
	scsi_set_resid(cmd, 0);
446
	memset(cmd->sense_buffer, 0, SCSI_SENSE_BUFFERSIZE);
Linus Torvalds's avatar
Linus Torvalds committed
447
	if (cmd->cmd_len == 0)
448
		cmd->cmd_len = scsi_command_size(cmd->cmnd);
Linus Torvalds's avatar
Linus Torvalds committed
449
450
451
452
453
454
455
456
457
}

void scsi_device_unbusy(struct scsi_device *sdev)
{
	struct Scsi_Host *shost = sdev->host;
	unsigned long flags;

	spin_lock_irqsave(shost->host_lock, flags);
	shost->host_busy--;
458
	if (unlikely(scsi_host_in_recovery(shost) &&
459
		     (shost->host_failed || shost->host_eh_scheduled)))
Linus Torvalds's avatar
Linus Torvalds committed
460
461
		scsi_eh_wakeup(shost);
	spin_unlock(shost->host_lock);
's avatar
committed
462
	spin_lock(sdev->request_queue->queue_lock);
Linus Torvalds's avatar
Linus Torvalds committed
463
	sdev->device_busy--;
's avatar
committed
464
	spin_unlock_irqrestore(sdev->request_queue->queue_lock, flags);
Linus Torvalds's avatar
Linus Torvalds committed
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
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
}

/*
 * Called for single_lun devices on IO completion. Clear starget_sdev_user,
 * and call blk_run_queue for all the scsi_devices on the target -
 * including current_sdev first.
 *
 * Called with *no* scsi locks held.
 */
static void scsi_single_lun_run(struct scsi_device *current_sdev)
{
	struct Scsi_Host *shost = current_sdev->host;
	struct scsi_device *sdev, *tmp;
	struct scsi_target *starget = scsi_target(current_sdev);
	unsigned long flags;

	spin_lock_irqsave(shost->host_lock, flags);
	starget->starget_sdev_user = NULL;
	spin_unlock_irqrestore(shost->host_lock, flags);

	/*
	 * Call blk_run_queue for all LUNs on the target, starting with
	 * current_sdev. We race with others (to set starget_sdev_user),
	 * but in most cases, we will be first. Ideally, each LU on the
	 * target would get some limited time or requests on the target.
	 */
	blk_run_queue(current_sdev->request_queue);

	spin_lock_irqsave(shost->host_lock, flags);
	if (starget->starget_sdev_user)
		goto out;
	list_for_each_entry_safe(sdev, tmp, &starget->devices,
			same_target_siblings) {
		if (sdev == current_sdev)
			continue;
		if (scsi_device_get(sdev))
			continue;

		spin_unlock_irqrestore(shost->host_lock, flags);
		blk_run_queue(sdev->request_queue);
		spin_lock_irqsave(shost->host_lock, flags);
	
		scsi_device_put(sdev);
	}
 out:
	spin_unlock_irqrestore(shost->host_lock, flags);
}

/*
 * Function:	scsi_run_queue()
 *
 * Purpose:	Select a proper request queue to serve next
 *
 * Arguments:	q	- last request's queue
 *
 * Returns:     Nothing
 *
 * Notes:	The previous command was completely finished, start
 *		a new one if possible.
 */
static void scsi_run_queue(struct request_queue *q)
{
	struct scsi_device *sdev = q->queuedata;
	struct Scsi_Host *shost = sdev->host;
	unsigned long flags;

531
	if (scsi_target(sdev)->single_lun)
Linus Torvalds's avatar
Linus Torvalds committed
532
533
534
535
536
537
538
		scsi_single_lun_run(sdev);

	spin_lock_irqsave(shost->host_lock, flags);
	while (!list_empty(&shost->starved_list) &&
	       !shost->host_blocked && !shost->host_self_blocked &&
		!((shost->can_queue > 0) &&
		  (shost->host_busy >= shost->can_queue))) {
539
540
541

		int flagset;

Linus Torvalds's avatar
Linus Torvalds committed
542
543
544
545
546
547
548
549
550
551
552
553
554
		/*
		 * As long as shost is accepting commands and we have
		 * starved queues, call blk_run_queue. scsi_request_fn
		 * drops the queue_lock and can add us back to the
		 * starved_list.
		 *
		 * host_lock protects the starved_list and starved_entry.
		 * scsi_request_fn must get the host_lock before checking
		 * or modifying starved_list or starved_entry.
		 */
		sdev = list_entry(shost->starved_list.next,
					  struct scsi_device, starved_entry);
		list_del_init(&sdev->starved_entry);
555
556
557
558
559
560
561
562
563
564
565
566
		spin_unlock(shost->host_lock);

		spin_lock(sdev->request_queue->queue_lock);
		flagset = test_bit(QUEUE_FLAG_REENTER, &q->queue_flags) &&
				!test_bit(QUEUE_FLAG_REENTER,
					&sdev->request_queue->queue_flags);
		if (flagset)
			queue_flag_set(QUEUE_FLAG_REENTER, sdev->request_queue);
		__blk_run_queue(sdev->request_queue);
		if (flagset)
			queue_flag_clear(QUEUE_FLAG_REENTER, sdev->request_queue);
		spin_unlock(sdev->request_queue->queue_lock);
567

568
		spin_lock(shost->host_lock);
Linus Torvalds's avatar
Linus Torvalds committed
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
		if (unlikely(!list_empty(&sdev->starved_entry)))
			/*
			 * sdev lost a race, and was put back on the
			 * starved list. This is unlikely but without this
			 * in theory we could loop forever.
			 */
			break;
	}
	spin_unlock_irqrestore(shost->host_lock, flags);

	blk_run_queue(q);
}

/*
 * Function:	scsi_requeue_command()
 *
 * Purpose:	Handle post-processing of completed commands.
 *
 * Arguments:	q	- queue to operate on
 *		cmd	- command that may need to be requeued.
 *
 * Returns:	Nothing
 *
 * Notes:	After command completion, there may be blocks left
 *		over which weren't finished by the previous command
 *		this can be for a number of reasons - the main one is
 *		I/O errors in the middle of the request, in which case
 *		we need to request the blocks that come after the bad
 *		sector.
598
 * Notes:	Upon return, cmd is a stale pointer.
Linus Torvalds's avatar
Linus Torvalds committed
599
600
601
 */
static void scsi_requeue_command(struct request_queue *q, struct scsi_cmnd *cmd)
{
602
	struct request *req = cmd->request;
603
604
	unsigned long flags;

605
	scsi_unprep_request(req);
606
	spin_lock_irqsave(q->queue_lock, flags);
607
	blk_requeue_request(q, req);
608
	spin_unlock_irqrestore(q->queue_lock, flags);
Linus Torvalds's avatar
Linus Torvalds committed
609
610
611
612
613
614

	scsi_run_queue(q);
}

void scsi_next_command(struct scsi_cmnd *cmd)
{
615
616
617
618
619
	struct scsi_device *sdev = cmd->device;
	struct request_queue *q = sdev->request_queue;

	/* need to hold a reference on the device before we let go of the cmd */
	get_device(&sdev->sdev_gendev);
Linus Torvalds's avatar
Linus Torvalds committed
620
621
622

	scsi_put_command(cmd);
	scsi_run_queue(q);
623
624
625

	/* ok to remove device now */
	put_device(&sdev->sdev_gendev);
Linus Torvalds's avatar
Linus Torvalds committed
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
}

void scsi_run_host_queues(struct Scsi_Host *shost)
{
	struct scsi_device *sdev;

	shost_for_each_device(sdev, shost)
		scsi_run_queue(sdev->request_queue);
}

/*
 * Function:    scsi_end_request()
 *
 * Purpose:     Post-processing of completed commands (usually invoked at end
 *		of upper level post-processing and scsi_io_completion).
 *
 * Arguments:   cmd	 - command that is complete.
643
 *              error    - 0 if I/O indicates success, < 0 for I/O error.
Linus Torvalds's avatar
Linus Torvalds committed
644
645
646
647
648
 *              bytes    - number of bytes of completed I/O
 *		requeue  - indicates whether we should requeue leftovers.
 *
 * Lock status: Assumed that lock is not held upon entry.
 *
649
 * Returns:     cmd if requeue required, NULL otherwise.
Linus Torvalds's avatar
Linus Torvalds committed
650
651
652
653
654
655
 *
 * Notes:       This is called for block device requests in order to
 *              mark some number of sectors as complete.
 * 
 *		We are guaranteeing that the request queue will be goosed
 *		at some point during this call.
656
 * Notes:	If cmd was requeued, upon return it will be a stale pointer.
Linus Torvalds's avatar
Linus Torvalds committed
657
 */
658
static struct scsi_cmnd *scsi_end_request(struct scsi_cmnd *cmd, int error,
Linus Torvalds's avatar
Linus Torvalds committed
659
660
					  int bytes, int requeue)
{
661
	struct request_queue *q = cmd->device->request_queue;
Linus Torvalds's avatar
Linus Torvalds committed
662
663
664
665
666
667
	struct request *req = cmd->request;

	/*
	 * If there are blocks left over at the end, set up the command
	 * to queue the remainder of them.
	 */
668
	if (blk_end_request(req, error, bytes)) {
Linus Torvalds's avatar
Linus Torvalds committed
669
670
671
672
673
674
		int leftover = (req->hard_nr_sectors << 9);

		if (blk_pc_request(req))
			leftover = req->data_len;

		/* kill remainder if no retrys */
675
676
		if (error && blk_noretry_request(req))
			blk_end_request(req, error, leftover);
Linus Torvalds's avatar
Linus Torvalds committed
677
		else {
678
			if (requeue) {
Linus Torvalds's avatar
Linus Torvalds committed
679
680
681
682
683
684
				/*
				 * Bleah.  Leftovers again.  Stick the
				 * leftovers in the front of the
				 * queue, and goose the queue again.
				 */
				scsi_requeue_command(q, cmd);
685
686
				cmd = NULL;
			}
Linus Torvalds's avatar
Linus Torvalds committed
687
688
689
690
691
692
693
694
695
696
697
698
			return cmd;
		}
	}

	/*
	 * This will goose the queue request function at the end, so we don't
	 * need to worry about launching another command.
	 */
	scsi_next_command(cmd);
	return NULL;
}

699
700
701
702
static inline unsigned int scsi_sgtable_index(unsigned short nents)
{
	unsigned int index;

703
704
705
	BUG_ON(nents > SCSI_MAX_SG_SEGMENTS);

	if (nents <= 8)
706
		index = 0;
707
708
	else
		index = get_count_order(nents) - 3;
Linus Torvalds's avatar
Linus Torvalds committed
709

710
711
712
	return index;
}

713
static void scsi_sg_free(struct scatterlist *sgl, unsigned int nents)
714
715
716
{
	struct scsi_host_sg_pool *sgp;

717
718
719
	sgp = scsi_sg_pools + scsi_sgtable_index(nents);
	mempool_free(sgl, sgp->pool);
}
720

721
722
723
static struct scatterlist *scsi_sg_alloc(unsigned int nents, gfp_t gfp_mask)
{
	struct scsi_host_sg_pool *sgp;
724

725
726
727
	sgp = scsi_sg_pools + scsi_sgtable_index(nents);
	return mempool_alloc(sgp->pool, gfp_mask);
}
728

729
730
static int scsi_alloc_sgtable(struct scsi_data_buffer *sdb, int nents,
			      gfp_t gfp_mask)
731
732
{
	int ret;
733

734
	BUG_ON(!nents);
735

736
737
	ret = __sg_alloc_table(&sdb->table, nents, SCSI_MAX_SG_SEGMENTS,
			       gfp_mask, scsi_sg_alloc);
738
	if (unlikely(ret))
739
		__sg_free_table(&sdb->table, SCSI_MAX_SG_SEGMENTS,
740
				scsi_sg_free);
741

742
	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
743
744
}

745
static void scsi_free_sgtable(struct scsi_data_buffer *sdb)
Linus Torvalds's avatar
Linus Torvalds committed
746
{
747
	__sg_free_table(&sdb->table, SCSI_MAX_SG_SEGMENTS, scsi_sg_free);
Linus Torvalds's avatar
Linus Torvalds committed
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
}

/*
 * Function:    scsi_release_buffers()
 *
 * Purpose:     Completion processing for block device I/O requests.
 *
 * Arguments:   cmd	- command that we are bailing.
 *
 * Lock status: Assumed that no lock is held upon entry.
 *
 * Returns:     Nothing
 *
 * Notes:       In the event that an upper level driver rejects a
 *		command, we must release resources allocated during
 *		the __init_io() function.  Primarily this would involve
 *		the scatter-gather table, and potentially any bounce
 *		buffers.
 */
767
void scsi_release_buffers(struct scsi_cmnd *cmd)
Linus Torvalds's avatar
Linus Torvalds committed
768
{
769
770
	if (cmd->sdb.table.nents)
		scsi_free_sgtable(&cmd->sdb);
Linus Torvalds's avatar
Linus Torvalds committed
771

772
	memset(&cmd->sdb, 0, sizeof(cmd->sdb));
773
774
775
776
777
778
779
780

	if (scsi_bidi_cmnd(cmd)) {
		struct scsi_data_buffer *bidi_sdb =
			cmd->request->next_rq->special;
		scsi_free_sgtable(bidi_sdb);
		kmem_cache_free(scsi_bidi_sdb_cache, bidi_sdb);
		cmd->request->next_rq->special = NULL;
	}
Linus Torvalds's avatar
Linus Torvalds committed
781
}
782
EXPORT_SYMBOL(scsi_release_buffers);
Linus Torvalds's avatar
Linus Torvalds committed
783

784
785
786
787
788
789
790
/*
 * Bidi commands Must be complete as a whole, both sides at once.
 * If part of the bytes were written and lld returned
 * scsi_in()->resid and/or scsi_out()->resid this information will be left
 * in req->data_len and req->next_rq->data_len. The upper-layer driver can
 * decide what to do with this information.
 */
791
static void scsi_end_bidi_request(struct scsi_cmnd *cmd)
792
{
793
794
795
796
797
798
799
800
801
802
	struct request *req = cmd->request;
	unsigned int dlen = req->data_len;
	unsigned int next_dlen = req->next_rq->data_len;

	req->data_len = scsi_out(cmd)->resid;
	req->next_rq->data_len = scsi_in(cmd)->resid;

	/* The req and req->next_rq have not been completed */
	BUG_ON(blk_end_bidi_request(req, 0, dlen, next_dlen));

803
804
805
806
807
808
809
810
811
	scsi_release_buffers(cmd);

	/*
	 * This will goose the queue request function at the end, so we don't
	 * need to worry about launching another command.
	 */
	scsi_next_command(cmd);
}

Linus Torvalds's avatar
Linus Torvalds committed
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
/*
 * Function:    scsi_io_completion()
 *
 * Purpose:     Completion processing for block device I/O requests.
 *
 * Arguments:   cmd   - command that is finished.
 *
 * Lock status: Assumed that no lock is held upon entry.
 *
 * Returns:     Nothing
 *
 * Notes:       This function is matched in terms of capabilities to
 *              the function that created the scatter-gather list.
 *              In other words, if there are no bounce buffers
 *              (the normal case for most drivers), we don't need
 *              the logic to deal with cleaning up afterwards.
 *
 *		We must do one of several things here:
 *
 *		a) Call scsi_end_request.  This will finish off the
 *		   specified number of sectors.  If we are done, the
 *		   command block will be released, and the queue
 *		   function will be goosed.  If we are not done, then
 *		   scsi_end_request will directly goose the queue.
 *
 *		b) We can just use scsi_requeue_command() here.  This would
 *		   be used if we just wanted to retry, for example.
 */
840
void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes)
Linus Torvalds's avatar
Linus Torvalds committed
841
842
{
	int result = cmd->result;
843
	int this_count = scsi_bufflen(cmd);
844
	struct request_queue *q = cmd->device->request_queue;
Linus Torvalds's avatar
Linus Torvalds committed
845
	struct request *req = cmd->request;
846
	int error = 0;
Linus Torvalds's avatar
Linus Torvalds committed
847
848
849
850
851
852
853
854
855
	struct scsi_sense_hdr sshdr;
	int sense_valid = 0;
	int sense_deferred = 0;

	if (result) {
		sense_valid = scsi_command_normalize_sense(cmd, &sshdr);
		if (sense_valid)
			sense_deferred = scsi_sense_is_deferred(&sshdr);
	}
856

Linus Torvalds's avatar
Linus Torvalds committed
857
858
859
860
861
862
863
864
865
866
867
868
869
870
	if (blk_pc_request(req)) { /* SG_IO ioctl from block level */
		req->errors = result;
		if (result) {
			if (sense_valid && req->sense) {
				/*
				 * SG_IO wants current and deferred errors
				 */
				int len = 8 + cmd->sense_buffer[7];

				if (len > SCSI_SENSE_BUFFERSIZE)
					len = SCSI_SENSE_BUFFERSIZE;
				memcpy(req->sense, cmd->sense_buffer,  len);
				req->sense_len = len;
			}
871
872
			if (!sense_deferred)
				error = -EIO;
873
		}
874
875
876
877
878
		if (scsi_bidi_cmnd(cmd)) {
			/* will also release_buffers */
			scsi_end_bidi_request(cmd);
			return;
		}
879
		req->data_len = scsi_get_resid(cmd);
Linus Torvalds's avatar
Linus Torvalds committed
880
881
	}

882
	BUG_ON(blk_bidi_rq(req)); /* bidi not support for !blk_pc_request yet */
883
884
	scsi_release_buffers(cmd);

Linus Torvalds's avatar
Linus Torvalds committed
885
886
887
888
	/*
	 * Next deal with any sectors which we were able to correctly
	 * handle.
	 */
889
890
891
892
893
894
895
896
	SCSI_LOG_HLCOMPLETE(1, printk("%ld sectors total, "
				      "%d bytes done.\n",
				      req->nr_sectors, good_bytes));

	/* A number of bytes were successfully read.  If there
	 * are leftovers and there is some kind of error
	 * (result != 0), retry the rest.
	 */
897
	if (scsi_end_request(cmd, error, good_bytes, result == 0) == NULL)
898
		return;
899
900
901

	/* good_bytes = 0, or (inclusive) there were leftovers and
	 * result = 0, so scsi_end_request couldn't retry.
Linus Torvalds's avatar
Linus Torvalds committed
902
903
904
905
906
	 */
	if (sense_valid && !sense_deferred) {
		switch (sshdr.sense_key) {
		case UNIT_ATTENTION:
			if (cmd->device->removable) {
907
				/* Detected disc change.  Set a bit
Linus Torvalds's avatar
Linus Torvalds committed
908
909
910
				 * and quietly refuse further access.
				 */
				cmd->device->changed = 1;
911
				scsi_end_request(cmd, -EIO, this_count, 1);
Linus Torvalds's avatar
Linus Torvalds committed
912
913
				return;
			} else {
914
915
916
917
918
				/* Must have been a power glitch, or a
				 * bus reset.  Could not have been a
				 * media change, so we just retry the
				 * request and see what happens.
				 */
Linus Torvalds's avatar
Linus Torvalds committed
919
920
921
922
923
				scsi_requeue_command(q, cmd);
				return;
			}
			break;
		case ILLEGAL_REQUEST:
924
925
926
927
928
929
930
931
			/* If we had an ILLEGAL REQUEST returned, then
			 * we may have performed an unsupported
			 * command.  The only thing this should be
			 * would be a ten byte read where only a six
			 * byte read was supported.  Also, on a system
			 * where READ CAPACITY failed, we may have
			 * read past the end of the disk.
			 */
932
933
			if ((cmd->device->use_10_for_rw &&
			    sshdr.asc == 0x20 && sshdr.ascq == 0x00) &&
Linus Torvalds's avatar
Linus Torvalds committed
934
935
936
			    (cmd->cmnd[0] == READ_10 ||
			     cmd->cmnd[0] == WRITE_10)) {
				cmd->device->use_10_for_rw = 0;
937
938
				/* This will cause a retry with a
				 * 6-byte command.
Linus Torvalds's avatar
Linus Torvalds committed
939
940
				 */
				scsi_requeue_command(q, cmd);
941
				return;
Linus Torvalds's avatar
Linus Torvalds committed
942
			} else {
943
				scsi_end_request(cmd, -EIO, this_count, 1);
Linus Torvalds's avatar
Linus Torvalds committed
944
945
946
947
				return;
			}
			break;
		case NOT_READY:
948
			/* If the device is in the process of becoming
949
			 * ready, or has a temporary blockage, retry.
Linus Torvalds's avatar
Linus Torvalds committed
950
			 */
951
952
953
954
955
956
957
958
959
960
961
962
963
964
			if (sshdr.asc == 0x04) {
				switch (sshdr.ascq) {
				case 0x01: /* becoming ready */
				case 0x04: /* format in progress */
				case 0x05: /* rebuild in progress */
				case 0x06: /* recalculation in progress */
				case 0x07: /* operation in progress */
				case 0x08: /* Long write in progress */
				case 0x09: /* self test in progress */
					scsi_requeue_command(q, cmd);
					return;
				default:
					break;
				}
Linus Torvalds's avatar
Linus Torvalds committed
965
			}
966
967
968
969
970
			if (!(req->cmd_flags & REQ_QUIET))
				scsi_cmd_print_sense_hdr(cmd,
							 "Device not ready",
							 &sshdr);

971
			scsi_end_request(cmd, -EIO, this_count, 1);
Linus Torvalds's avatar
Linus Torvalds committed
972
973
			return;
		case VOLUME_OVERFLOW:
974
			if (!(req->cmd_flags & REQ_QUIET)) {
975
				scmd_printk(KERN_INFO, cmd,
976
					    "Volume overflow, CDB: ");
977
				__scsi_print_command(cmd->cmnd);
978
979
				scsi_print_sense("", cmd);
			}
980
			/* See SSC3rXX or current. */
981
			scsi_end_request(cmd, -EIO, this_count, 1);
Linus Torvalds's avatar
Linus Torvalds committed
982
983
984
985
			return;
		default:
			break;
		}
986
	}
Linus Torvalds's avatar
Linus Torvalds committed
987
	if (host_byte(result) == DID_RESET) {
988
989
990
		/* Third party bus reset or reset for error recovery
		 * reasons.  Just retry the request and see what
		 * happens.
Linus Torvalds's avatar
Linus Torvalds committed
991
992
993
994
995
		 */
		scsi_requeue_command(q, cmd);
		return;
	}
	if (result) {
996
		if (!(req->cmd_flags & REQ_QUIET)) {
997
			scsi_print_result(cmd);
998
999
1000
			if (driver_byte(result) & DRIVER_SENSE)
				scsi_print_sense("", cmd);
		}
Linus Torvalds's avatar
Linus Torvalds committed
1001
	}
1002
	scsi_end_request(cmd, -EIO, this_count, !result);
Linus Torvalds's avatar
Linus Torvalds committed
1003
1004
}

1005
1006
static int scsi_init_sgtable(struct request *req, struct scsi_data_buffer *sdb,
			     gfp_t gfp_mask)
Linus Torvalds's avatar
Linus Torvalds committed
1007
{
1008
	int count;
Linus Torvalds's avatar
Linus Torvalds committed
1009
1010

	/*
1011
	 * If sg table allocation fails, requeue request later.
Linus Torvalds's avatar
Linus Torvalds committed
1012
	 */
1013
1014
	if (unlikely(scsi_alloc_sgtable(sdb, req->nr_phys_segments,
					gfp_mask))) {
Linus Torvalds's avatar
Linus Torvalds committed
1015
		return BLKPREP_DEFER;
Alan Stern's avatar
Alan Stern committed
1016
	}
Linus Torvalds's avatar
Linus Torvalds committed
1017

1018
	req->buffer = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
1019
1020
1021
1022
1023

	/* 
	 * Next, walk the list, and fill in the addresses and sizes of
	 * each segment.
	 */
1024
1025
1026
	count = blk_rq_map_sg(req->q, req, sdb->table.sgl);
	BUG_ON(count > sdb->table.nents);
	sdb->table.nents = count;
Tejun Heo's avatar
Tejun Heo committed
1027
1028
1029
1030
	if (blk_pc_request(req))
		sdb->length = req->data_len;
	else
		sdb->length = req->nr_sectors << 9;
1031
	return BLKPREP_OK;
Linus Torvalds's avatar
Linus Torvalds committed
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
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076

/*
 * Function:    scsi_init_io()
 *
 * Purpose:     SCSI I/O initialize function.
 *
 * Arguments:   cmd   - Command descriptor we wish to initialize
 *
 * Returns:     0 on success
 *		BLKPREP_DEFER if the failure is retryable
 *		BLKPREP_KILL if the failure is fatal
 */
int scsi_init_io(struct scsi_cmnd *cmd, gfp_t gfp_mask)
{
	int error = scsi_init_sgtable(cmd->request, &cmd->sdb, gfp_mask);
	if (error)
		goto err_exit;

	if (blk_bidi_rq(cmd->request)) {
		struct scsi_data_buffer *bidi_sdb = kmem_cache_zalloc(
			scsi_bidi_sdb_cache, GFP_ATOMIC);
		if (!bidi_sdb) {
			error = BLKPREP_DEFER;
			goto err_exit;
		}

		cmd->request->next_rq->special = bidi_sdb;
		error = scsi_init_sgtable(cmd->request->next_rq, bidi_sdb,
								    GFP_ATOMIC);
		if (error)
			goto err_exit;
	}

	return BLKPREP_OK ;

err_exit:
	scsi_release_buffers(cmd);
	if (error == BLKPREP_KILL)
		scsi_put_command(cmd);
	else /* BLKPREP_DEFER */
		scsi_unprep_request(cmd->request);

	return error;
}
1077
EXPORT_SYMBOL(scsi_init_io);
Linus Torvalds's avatar
Linus Torvalds committed
1078

1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
static struct scsi_cmnd *scsi_get_cmd_from_req(struct scsi_device *sdev,
		struct request *req)
{
	struct scsi_cmnd *cmd;

	if (!req->special) {
		cmd = scsi_get_command(sdev, GFP_ATOMIC);
		if (unlikely(!cmd))
			return NULL;
		req->special = cmd;
	} else {
		cmd = req->special;
	}

	/* pull a tag out of the request if we have one */
	cmd->tag = req->tag;
	cmd->request = req;

1097
1098
	cmd->cmnd = req->cmd;

1099
1100
1101
	return cmd;
}

1102
int scsi_setup_blk_pc_cmnd(struct scsi_device *sdev, struct request *req)
James Bottomley's avatar
James Bottomley committed
1103
{
1104
	struct scsi_cmnd *cmd;
1105
1106
1107
1108
	int ret = scsi_prep_state_check(sdev, req);

	if (ret != BLKPREP_OK)
		return ret;
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124

	cmd = scsi_get_cmd_from_req(sdev, req);
	if (unlikely(!cmd))
		return BLKPREP_DEFER;

	/*
	 * BLOCK_PC requests may transfer data, in which case they must
	 * a bio attached to them.  Or they might contain a SCSI command
	 * that does not transfer data, in which case they may optionally
	 * submit a request without an attached bio.
	 */
	if (req->bio) {
		int ret;

		BUG_ON(!req->nr_phys_segments);

1125
		ret = scsi_init_io(cmd, GFP_ATOMIC);
1126
1127
1128
1129
1130
1131
		if (unlikely(ret))
			return ret;
	} else {
		BUG_ON(req->data_len);
		BUG_ON(req->data);

1132
		memset(&cmd->sdb, 0, sizeof(cmd->sdb));
1133
1134
		req->buffer = NULL;
	}
James Bottomley's avatar
James Bottomley committed
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146

	cmd->cmd_len = req->cmd_len;
	if (!req->data_len)
		cmd->sc_data_direction = DMA_NONE;
	else if (rq_data_dir(req) == WRITE)
		cmd->sc_data_direction = DMA_TO_DEVICE;
	else
		cmd->sc_data_direction = DMA_FROM_DEVICE;
	
	cmd->transfersize = req->data_len;
	cmd->allowed = req->retries;
	cmd->timeout_per_command = req->timeout;
1147
	return BLKPREP_OK;
James Bottomley's avatar
James Bottomley committed
1148
}
1149
EXPORT_SYMBOL(scsi_setup_blk_pc_cmnd);
James Bottomley's avatar
James Bottomley committed
1150

1151
1152
1153
1154
1155
/*
 * Setup a REQ_TYPE_FS command.  These are simple read/write request
 * from filesystems that still need to be translated to SCSI CDBs from
 * the ULD.
 */
1156
int scsi_setup_fs_cmnd(struct scsi_device *sdev, struct request *req)
Linus Torvalds's avatar
Linus Torvalds committed
1157
1158
{
	struct scsi_cmnd *cmd;
1159
	int ret = scsi_prep_state_check(sdev, req);
Linus Torvalds's avatar
Linus Torvalds committed
1160

1161
1162
	if (ret != BLKPREP_OK)
		return ret;
Linus Torvalds's avatar
Linus Torvalds committed
1163
	/*
1164
	 * Filesystem requests must transfer data.
Linus Torvalds's avatar
Linus Torvalds committed
1165
	 */
1166
1167
1168
1169
1170
1171
	BUG_ON(!req->nr_phys_segments);

	cmd = scsi_get_cmd_from_req(sdev, req);
	if (unlikely(!cmd))
		return BLKPREP_DEFER;

1172
	memset(cmd->cmnd, 0, BLK_MAX_CDB);
1173
	return scsi_init_io(cmd, GFP_ATOMIC);
1174
}
1175
EXPORT_SYMBOL(scsi_setup_fs_cmnd);
1176

1177
int scsi_prep_state_check(struct scsi_device *sdev, struct request *req)
1178
1179
1180
{
	int ret = BLKPREP_OK;

Linus Torvalds's avatar
Linus Torvalds committed
1181
	/*
1182
1183
	 * If the device is not in running state we will reject some
	 * or all commands.
Linus Torvalds's avatar
Linus Torvalds committed
1184
	 */
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
	if (unlikely(sdev->sdev_state != SDEV_RUNNING)) {
		switch (sdev->sdev_state) {
		case SDEV_OFFLINE:
			/*
			 * If the device is offline we refuse to process any
			 * commands.  The device must be brought online
			 * before trying any recovery commands.
			 */
			sdev_printk(KERN_ERR, sdev,
				    "rejecting I/O to offline device\n");
			ret = BLKPREP_KILL;
			break;
		case SDEV_DEL:
			/*
			 * If the device is fully deleted, we refuse to
			 * process any commands as well.
			 */
1202
			sdev_printk(KERN_ERR, sdev,
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
				    "rejecting I/O to dead device\n");
			ret = BLKPREP_KILL;
			break;
		case SDEV_QUIESCE:
		case SDEV_BLOCK:
			/*
			 * If the devices is blocked we defer normal commands.
			 */
			if (!(req->cmd_flags & REQ_PREEMPT))
				ret = BLKPREP_DEFER;
			break;
		default:
			/*
			 * For any other not fully online state we only allow
			 * special commands.  In particular any user initiated
			 * command is not allowed.
			 */
			if (!(req->cmd_flags & REQ_PREEMPT))
				ret = BLKPREP_KILL;
			break;
Linus Torvalds's avatar
Linus Torvalds committed
1223
1224
		}
	}
1225
1226
1227
	return ret;
}
EXPORT_SYMBOL(scsi_prep_state_check);
Linus Torvalds's avatar
Linus Torvalds committed
1228

1229
1230
1231
int scsi_prep_return(struct request_queue *q, struct request *req, int ret)
{
	struct scsi_device *sdev = q->queuedata;
Linus Torvalds's avatar
Linus Torvalds committed
1232

1233
1234
1235
	switch (ret) {
	case BLKPREP_KILL:
		req->errors = DID_NO_CONNECT << 16;
1236
1237
1238
1239
1240
1241
1242
		/* release the command and kill it */
		if (req->special) {
			struct scsi_cmnd *cmd = req->special;
			scsi_release_buffers(cmd);
			scsi_put_command(cmd);
			req->special = NULL;
		}
1243
1244
		break;
	case BLKPREP_DEFER:
Linus Torvalds's avatar
Linus Torvalds committed
1245
		/*
1246
1247
1248
		 * If we defer, the elv_next_request() returns NULL, but the
		 * queue must be restarted, so we plug here if no returning
		 * command will automatically do that.
Linus Torvalds's avatar
Linus Torvalds committed
1249
		 */
1250
1251
1252
1253
1254
		if (sdev->device_busy == 0)
			blk_plug_device(q);
		break;
	default:
		req->cmd_flags |= REQ_DONTPREP;
Linus Torvalds's avatar
Linus Torvalds committed
1255
1256
	}

1257
	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
1258
}
1259
1260
EXPORT_SYMBOL(scsi_prep_return);

1261
int scsi_prep_fn(struct request_queue *q, struct request *req)
1262
1263
1264
1265
1266
1267
1268
1269
{
	struct scsi_device *sdev = q->queuedata;
	int ret = BLKPREP_KILL;

	if (req->cmd_type == REQ_TYPE_BLOCK_PC)
		ret = scsi_setup_blk_pc_cmnd(sdev, req);
	return scsi_prep_return(q, req, ret);
}
Linus Torvalds's avatar
Linus Torvalds committed
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287

/*
 * scsi_dev_queue_ready: if we can send requests to sdev, return 1 else
 * return 0.
 *
 * Called with the queue_lock held.
 */
static inline int scsi_dev_queue_ready(struct request_queue *q,
				  struct scsi_device *sdev)
{
	if (sdev->device_busy >= sdev->queue_depth)
		return 0;
	if (sdev->device_busy == 0 && sdev->device_blocked) {
		/*
		 * unblock after device_blocked iterates to zero
		 */
		if (--sdev->device_blocked == 0) {
			SCSI_LOG_MLQUEUE(3,
1288
1289
				   sdev_printk(KERN_INFO, sdev,
				   "unblocking device at zero depth\n"));
Linus Torvalds's avatar
Linus Torvalds committed
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
		} else {
			blk_plug_device(q);
			return 0;
		}
	}
	if (sdev->device_blocked)
		return 0;

	return 1;
}

/*
 * scsi_host_queue_ready: if we can send requests to shost, return 1 else
 * return 0. We must end up running the queue again whenever 0 is
 * returned, else IO can hang.
 *
 * Called with host_lock held.
 */
static inline int scsi_host_queue_ready(struct request_queue *q,
				   struct Scsi_Host *shost,
				   struct scsi_device *sdev)
{
1312
	if (scsi_host_in_recovery(shost))
Linus Torvalds's avatar
Linus Torvalds committed
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
		return 0;
	if (shost->host_busy == 0 && shost->host_blocked) {
		/*
		 * unblock after host_blocked iterates to zero
		 */
		if (--shost->host_blocked == 0) {
			SCSI_LOG_MLQUEUE(3,
				printk("scsi%d unblocking host at zero depth\n",
					shost->host_no));
		} else {
			blk_plug_device(q);
			return 0;
		}
	}
	if ((shost->can_queue > 0 && shost->host_busy >= shost->can_queue) ||
	    shost->host_blocked || shost->host_self_blocked) {
		if (list_empty(&sdev->starved_entry))
			list_add_tail(&sdev->starved_entry, &shost->starved_list);
		return 0;
	}

	/* We're OK to process the command, so we can't be starved */
	if (!list_empty(&sdev->starved_entry))
		list_del_init(&sdev->starved_entry);

	return 1;
}

/*
1342
 * Kill a request for a dead device
Linus Torvalds's avatar
Linus Torvalds committed
1343
 */
1344
static void scsi_kill_request(struct request *req, struct request_queue *q)
Linus Torvalds's avatar
Linus Torvalds committed
1345
{
1346
	struct scsi_cmnd *cmd = req->special;