scsi_lib.c 64.3 KB
Newer Older
L
Linus Torvalds 已提交
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>
J
James Bottomley 已提交
11
#include <linux/bitops.h>
L
Linus Torvalds 已提交
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>
J
Jens Axboe 已提交
21
#include <linux/scatterlist.h>
L
Linus Torvalds 已提交
22 23

#include <scsi/scsi.h>
24
#include <scsi/scsi_cmnd.h>
L
Linus Torvalds 已提交
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
L
Linus Torvalds 已提交
37 38 39

struct scsi_host_sg_pool {
	size_t		size;
40
	char		*name;
41
	struct kmem_cache	*slab;
L
Linus Torvalds 已提交
42 43 44
	mempool_t	*pool;
};

J
James Bottomley 已提交
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
A
Adrian Bunk 已提交
49
static struct scsi_host_sg_pool scsi_sg_pools[] = {
L
Linus Torvalds 已提交
50 51
	SP(8),
	SP(16),
F
FUJITA Tomonori 已提交
52
#if (SCSI_MAX_SG_SEGMENTS > 32)
J
James Bottomley 已提交
53
	SP(32),
F
FUJITA Tomonori 已提交
54
#if (SCSI_MAX_SG_SEGMENTS > 64)
J
James Bottomley 已提交
55 56
	SP(64),
#if (SCSI_MAX_SG_SEGMENTS > 128)
L
Linus Torvalds 已提交
57
	SP(128),
J
James Bottomley 已提交
58 59
#if (SCSI_MAX_SG_SEGMENTS > 256)
#error SCSI_MAX_SG_SEGMENTS is too large (256 MAX)
F
FUJITA Tomonori 已提交
60 61 62
#endif
#endif
#endif
J
James Bottomley 已提交
63 64
#endif
	SP(SCSI_MAX_SG_SEGMENTS)
65
};
L
Linus Torvalds 已提交
66 67
#undef SP

68
struct kmem_cache *scsi_sdb_cache;
69

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

L
Linus Torvalds 已提交
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;
L
Linus Torvalds 已提交
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.
L
Linus Torvalds 已提交
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.
L
Linus Torvalds 已提交
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
L
Linus Torvalds 已提交
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);
J
James Bottomley 已提交
159
	blk_requeue_request(q, cmd->request);
160 161 162 163
	spin_unlock_irqrestore(q->queue_lock, flags);

	scsi_run_queue(q);

L
Linus Torvalds 已提交
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

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

210 211 212 213 214 215 216 217 218
	/*
	 * Some devices (USB mass-storage in particular) may transfer
	 * garbage data together with a residue indicating that the data
	 * is invalid.  Prevent the garbage from being misinterpreted
	 * and prevent security leaks by zeroing out the excess data.
	 */
	if (unlikely(req->data_len > 0 && req->data_len <= bufflen))
		memset(buffer + (bufflen - req->data_len), 0, req->data_len);

219 220 221 222 223 224
	ret = req->errors;
 out:
	blk_put_request(req);

	return ret;
}
225
EXPORT_SYMBOL(scsi_execute);
226

227 228 229 230 231 232

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;
233 234
	int result;
	
235
	if (sshdr) {
J
Jes Sorensen 已提交
236
		sense = kzalloc(SCSI_SENSE_BUFFERSIZE, GFP_NOIO);
237 238 239
		if (!sense)
			return DRIVER_ERROR << 24;
	}
240
	result = scsi_execute(sdev, cmd, data_direction, buffer, bufflen,
J
Jes Sorensen 已提交
241
			      sense, timeout, retries, 0);
242
	if (sshdr)
243
		scsi_normalize_sense(sense, SCSI_SENSE_BUFFERSIZE, sshdr);
244 245 246 247 248 249

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

250 251 252 253 254 255
struct scsi_io_context {
	void *data;
	void (*done)(void *data, char *sense, int result, int resid);
	char sense[SCSI_SENSE_BUFFERSIZE];
};

256
static struct kmem_cache *scsi_io_context_cache;
257

258
static void scsi_end_async(struct request *req, int uptodate)
259 260 261 262 263 264
{
	struct scsi_io_context *sioc = req->end_io_data;

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

265
	kmem_cache_free(scsi_io_context_cache, sioc);
266 267 268 269 270 271 272 273 274 275 276 277
	__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);

N
NeilBrown 已提交
278
	return blk_rq_append_bio(q, rq, bio);
279 280
}

281
static void scsi_bi_endio(struct bio *bio, int error)
282 283 284 285 286 287 288
{
	bio_put(bio);
}

/**
 * scsi_req_map_sg - map a scatterlist into a request
 * @rq:		request to fill
289
 * @sgl:	scatterlist
290 291 292 293 294 295 296 297 298 299 300 301
 * @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;
302
	int nr_pages = (bufflen + sgl[0].offset + PAGE_SIZE - 1) >> PAGE_SHIFT;
303
	unsigned int data_len = bufflen, len, bytes, off;
J
Jens Axboe 已提交
304
	struct scatterlist *sg;
305 306 307 308
	struct page *page;
	struct bio *bio = NULL;
	int i, err, nr_vecs = 0;

J
Jens Axboe 已提交
309
	for_each_sg(sgl, sg, nsegs, i) {
J
Jens Axboe 已提交
310
		page = sg_page(sg);
J
Jens Axboe 已提交
311 312
		off = sg->offset;
		len = sg->length;
313

314 315 316 317 318 319
		while (len > 0 && data_len > 0) {
			/*
			 * sg sends a scatterlist that is larger than
			 * the data_len it wants transferred for certain
			 * IO sizes
			 */
320
			bytes = min_t(unsigned int, len, PAGE_SIZE - off);
321
			bytes = min(bytes, data_len);
322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344

			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) {
345
					bio_endio(bio, 0);
346 347 348 349 350 351 352
					goto free_bios;
				}
				bio = NULL;
			}

			page++;
			len -= bytes;
353
			data_len -=bytes;
354 355 356 357 358
			off = 0;
		}
	}

	rq->buffer = rq->data = NULL;
359
	rq->data_len = bufflen;
360 361 362 363 364 365 366 367
	return 0;

free_bios:
	while ((bio = rq->bio) != NULL) {
		rq->bio = bio->bi_next;
		/*
		 * call endio instead of bio_put incase it was bounced
		 */
368
		bio_endio(bio, 0);
369 370 371 372 373 374 375 376 377
	}

	return err;
}

/**
 * scsi_execute_async - insert request
 * @sdev:	scsi device
 * @cmd:	scsi command
378
 * @cmd_len:	length of scsi cdb
379
 * @data_direction: DMA_TO_DEVICE, DMA_FROM_DEVICE, or DMA_NONE
380 381 382 383 384
 * @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
385 386 387 388
 * @privdata:	data passed to done()
 * @done:	callback function when done
 * @gfp:	memory allocation flags
 */
389
int scsi_execute_async(struct scsi_device *sdev, const unsigned char *cmd,
390
		       int cmd_len, int data_direction, void *buffer, unsigned bufflen,
391 392 393 394 395 396 397 398
		       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);

399
	sioc = kmem_cache_zalloc(scsi_io_context_cache, gfp);
400 401 402 403 404 405
	if (!sioc)
		return DRIVER_ERROR << 24;

	req = blk_get_request(sdev->request_queue, write, gfp);
	if (!req)
		goto free_sense;
406 407
	req->cmd_type = REQ_TYPE_BLOCK_PC;
	req->cmd_flags |= REQ_QUIET;
408 409 410 411 412 413 414 415 416

	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;

417
	req->cmd_len = cmd_len;
418
	memset(req->cmd, 0, BLK_MAX_CDB); /* ATAPI hates garbage after CDB */
419 420 421 422
	memcpy(req->cmd, cmd, req->cmd_len);
	req->sense = sioc->sense;
	req->sense_len = 0;
	req->timeout = timeout;
423
	req->retries = retries;
424 425 426 427 428 429 430 431 432 433 434
	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:
435
	kmem_cache_free(scsi_io_context_cache, sioc);
436 437 438 439
	return DRIVER_ERROR << 24;
}
EXPORT_SYMBOL_GPL(scsi_execute_async);

L
Linus Torvalds 已提交
440 441 442 443 444 445 446 447 448 449 450
/*
 * 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.
 */
451
static void scsi_init_cmd_errh(struct scsi_cmnd *cmd)
L
Linus Torvalds 已提交
452 453
{
	cmd->serial_number = 0;
B
Boaz Harrosh 已提交
454
	scsi_set_resid(cmd, 0);
455
	memset(cmd->sense_buffer, 0, SCSI_SENSE_BUFFERSIZE);
L
Linus Torvalds 已提交
456
	if (cmd->cmd_len == 0)
457
		cmd->cmd_len = scsi_command_size(cmd->cmnd);
L
Linus Torvalds 已提交
458 459 460 461 462 463 464 465 466
}

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--;
467
	if (unlikely(scsi_host_in_recovery(shost) &&
468
		     (shost->host_failed || shost->host_eh_scheduled)))
L
Linus Torvalds 已提交
469 470
		scsi_eh_wakeup(shost);
	spin_unlock(shost->host_lock);
已提交
471
	spin_lock(sdev->request_queue->queue_lock);
L
Linus Torvalds 已提交
472
	sdev->device_busy--;
已提交
473
	spin_unlock_irqrestore(sdev->request_queue->queue_lock, flags);
L
Linus Torvalds 已提交
474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539
}

/*
 * 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;

540
	if (scsi_target(sdev)->single_lun)
L
Linus Torvalds 已提交
541 542 543 544 545 546 547
		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))) {
N
Nick Piggin 已提交
548 549 550

		int flagset;

L
Linus Torvalds 已提交
551 552 553 554 555 556 557 558 559 560 561 562 563
		/*
		 * 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);
N
Nick Piggin 已提交
564 565 566 567 568 569 570 571 572 573 574 575
		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);
576

N
Nick Piggin 已提交
577
		spin_lock(shost->host_lock);
L
Linus Torvalds 已提交
578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606
		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.
607
 * Notes:	Upon return, cmd is a stale pointer.
L
Linus Torvalds 已提交
608 609 610
 */
static void scsi_requeue_command(struct request_queue *q, struct scsi_cmnd *cmd)
{
611
	struct request *req = cmd->request;
612 613
	unsigned long flags;

614
	scsi_unprep_request(req);
615
	spin_lock_irqsave(q->queue_lock, flags);
616
	blk_requeue_request(q, req);
617
	spin_unlock_irqrestore(q->queue_lock, flags);
L
Linus Torvalds 已提交
618 619 620 621 622 623

	scsi_run_queue(q);
}

void scsi_next_command(struct scsi_cmnd *cmd)
{
624 625 626 627 628
	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);
L
Linus Torvalds 已提交
629 630 631

	scsi_put_command(cmd);
	scsi_run_queue(q);
632 633 634

	/* ok to remove device now */
	put_device(&sdev->sdev_gendev);
L
Linus Torvalds 已提交
635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651
}

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.
652
 *              error    - 0 if I/O indicates success, < 0 for I/O error.
L
Linus Torvalds 已提交
653 654 655 656 657
 *              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.
 *
658
 * Returns:     cmd if requeue required, NULL otherwise.
L
Linus Torvalds 已提交
659 660 661 662 663 664
 *
 * 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.
665
 * Notes:	If cmd was requeued, upon return it will be a stale pointer.
L
Linus Torvalds 已提交
666
 */
667
static struct scsi_cmnd *scsi_end_request(struct scsi_cmnd *cmd, int error,
L
Linus Torvalds 已提交
668 669
					  int bytes, int requeue)
{
670
	struct request_queue *q = cmd->device->request_queue;
L
Linus Torvalds 已提交
671 672 673 674 675 676
	struct request *req = cmd->request;

	/*
	 * If there are blocks left over at the end, set up the command
	 * to queue the remainder of them.
	 */
677
	if (blk_end_request(req, error, bytes)) {
L
Linus Torvalds 已提交
678 679 680 681 682 683
		int leftover = (req->hard_nr_sectors << 9);

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

		/* kill remainder if no retrys */
684 685
		if (error && blk_noretry_request(req))
			blk_end_request(req, error, leftover);
L
Linus Torvalds 已提交
686
		else {
687
			if (requeue) {
L
Linus Torvalds 已提交
688 689 690 691 692 693
				/*
				 * Bleah.  Leftovers again.  Stick the
				 * leftovers in the front of the
				 * queue, and goose the queue again.
				 */
				scsi_requeue_command(q, cmd);
694 695
				cmd = NULL;
			}
L
Linus Torvalds 已提交
696 697 698 699 700 701 702 703 704 705 706 707
			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;
}

708 709 710 711
static inline unsigned int scsi_sgtable_index(unsigned short nents)
{
	unsigned int index;

J
James Bottomley 已提交
712 713 714
	BUG_ON(nents > SCSI_MAX_SG_SEGMENTS);

	if (nents <= 8)
715
		index = 0;
J
James Bottomley 已提交
716 717
	else
		index = get_count_order(nents) - 3;
L
Linus Torvalds 已提交
718

719 720 721
	return index;
}

722
static void scsi_sg_free(struct scatterlist *sgl, unsigned int nents)
723 724 725
{
	struct scsi_host_sg_pool *sgp;

726 727 728
	sgp = scsi_sg_pools + scsi_sgtable_index(nents);
	mempool_free(sgl, sgp->pool);
}
729

730 731 732
static struct scatterlist *scsi_sg_alloc(unsigned int nents, gfp_t gfp_mask)
{
	struct scsi_host_sg_pool *sgp;
733

734 735 736
	sgp = scsi_sg_pools + scsi_sgtable_index(nents);
	return mempool_alloc(sgp->pool, gfp_mask);
}
737

B
Boaz Harrosh 已提交
738 739
static int scsi_alloc_sgtable(struct scsi_data_buffer *sdb, int nents,
			      gfp_t gfp_mask)
740 741
{
	int ret;
742

B
Boaz Harrosh 已提交
743
	BUG_ON(!nents);
744

B
Boaz Harrosh 已提交
745 746
	ret = __sg_alloc_table(&sdb->table, nents, SCSI_MAX_SG_SEGMENTS,
			       gfp_mask, scsi_sg_alloc);
747
	if (unlikely(ret))
B
Boaz Harrosh 已提交
748
		__sg_free_table(&sdb->table, SCSI_MAX_SG_SEGMENTS,
749
				scsi_sg_free);
J
Jens Axboe 已提交
750

751
	return ret;
L
Linus Torvalds 已提交
752 753
}

B
Boaz Harrosh 已提交
754
static void scsi_free_sgtable(struct scsi_data_buffer *sdb)
L
Linus Torvalds 已提交
755
{
B
Boaz Harrosh 已提交
756
	__sg_free_table(&sdb->table, SCSI_MAX_SG_SEGMENTS, scsi_sg_free);
L
Linus Torvalds 已提交
757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775
}

/*
 * 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.
 */
776
void scsi_release_buffers(struct scsi_cmnd *cmd)
L
Linus Torvalds 已提交
777
{
B
Boaz Harrosh 已提交
778 779
	if (cmd->sdb.table.nents)
		scsi_free_sgtable(&cmd->sdb);
L
Linus Torvalds 已提交
780

B
Boaz Harrosh 已提交
781
	memset(&cmd->sdb, 0, sizeof(cmd->sdb));
782 783 784 785 786

	if (scsi_bidi_cmnd(cmd)) {
		struct scsi_data_buffer *bidi_sdb =
			cmd->request->next_rq->special;
		scsi_free_sgtable(bidi_sdb);
787
		kmem_cache_free(scsi_sdb_cache, bidi_sdb);
788 789
		cmd->request->next_rq->special = NULL;
	}
790 791 792

	if (scsi_prot_sg_count(cmd))
		scsi_free_sgtable(cmd->prot_sdb);
L
Linus Torvalds 已提交
793
}
794
EXPORT_SYMBOL(scsi_release_buffers);
L
Linus Torvalds 已提交
795

796 797 798 799 800 801 802
/*
 * 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.
 */
803
static void scsi_end_bidi_request(struct scsi_cmnd *cmd)
804
{
805 806 807 808 809 810 811 812 813 814
	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));

815 816 817 818 819 820 821 822 823
	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);
}

L
Linus Torvalds 已提交
824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851
/*
 * 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.
 */
852
void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes)
L
Linus Torvalds 已提交
853 854
{
	int result = cmd->result;
B
Boaz Harrosh 已提交
855
	int this_count = scsi_bufflen(cmd);
856
	struct request_queue *q = cmd->device->request_queue;
L
Linus Torvalds 已提交
857
	struct request *req = cmd->request;
858
	int error = 0;
L
Linus Torvalds 已提交
859 860 861 862 863 864 865 866 867
	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);
	}
868

L
Linus Torvalds 已提交
869 870 871 872 873 874 875 876 877 878 879 880 881 882
	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;
			}
883 884
			if (!sense_deferred)
				error = -EIO;
885
		}
886 887 888 889 890
		if (scsi_bidi_cmnd(cmd)) {
			/* will also release_buffers */
			scsi_end_bidi_request(cmd);
			return;
		}
B
Boaz Harrosh 已提交
891
		req->data_len = scsi_get_resid(cmd);
L
Linus Torvalds 已提交
892 893
	}

894
	BUG_ON(blk_bidi_rq(req)); /* bidi not support for !blk_pc_request yet */
B
Boaz Harrosh 已提交
895 896
	scsi_release_buffers(cmd);

L
Linus Torvalds 已提交
897 898 899 900
	/*
	 * Next deal with any sectors which we were able to correctly
	 * handle.
	 */
901 902 903 904 905 906 907 908
	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.
	 */
909
	if (scsi_end_request(cmd, error, good_bytes, result == 0) == NULL)
910
		return;
911 912 913

	/* good_bytes = 0, or (inclusive) there were leftovers and
	 * result = 0, so scsi_end_request couldn't retry.
L
Linus Torvalds 已提交
914 915 916 917 918
	 */
	if (sense_valid && !sense_deferred) {
		switch (sshdr.sense_key) {
		case UNIT_ATTENTION:
			if (cmd->device->removable) {
919
				/* Detected disc change.  Set a bit
L
Linus Torvalds 已提交
920 921 922
				 * and quietly refuse further access.
				 */
				cmd->device->changed = 1;
923
				scsi_end_request(cmd, -EIO, this_count, 1);
L
Linus Torvalds 已提交
924 925
				return;
			} else {
926 927 928 929 930
				/* 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.
				 */
L
Linus Torvalds 已提交
931 932 933 934 935
				scsi_requeue_command(q, cmd);
				return;
			}
			break;
		case ILLEGAL_REQUEST:
936 937 938 939 940 941 942 943
			/* 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.
			 */
944 945
			if ((cmd->device->use_10_for_rw &&
			    sshdr.asc == 0x20 && sshdr.ascq == 0x00) &&
L
Linus Torvalds 已提交
946 947 948
			    (cmd->cmnd[0] == READ_10 ||
			     cmd->cmnd[0] == WRITE_10)) {
				cmd->device->use_10_for_rw = 0;
949 950
				/* This will cause a retry with a
				 * 6-byte command.
L
Linus Torvalds 已提交
951 952
				 */
				scsi_requeue_command(q, cmd);
953
				return;
L
Linus Torvalds 已提交
954
			} else {
955
				scsi_end_request(cmd, -EIO, this_count, 1);
L
Linus Torvalds 已提交
956 957 958 959
				return;
			}
			break;
		case NOT_READY:
960
			/* If the device is in the process of becoming
J
James Bottomley 已提交
961
			 * ready, or has a temporary blockage, retry.
L
Linus Torvalds 已提交
962
			 */
J
James Bottomley 已提交
963 964 965 966 967 968 969 970 971 972 973 974 975 976
			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;
				}
L
Linus Torvalds 已提交
977
			}
978 979 980 981 982
			if (!(req->cmd_flags & REQ_QUIET))
				scsi_cmd_print_sense_hdr(cmd,
							 "Device not ready",
							 &sshdr);

983
			scsi_end_request(cmd, -EIO, this_count, 1);
L
Linus Torvalds 已提交
984 985
			return;
		case VOLUME_OVERFLOW:
986
			if (!(req->cmd_flags & REQ_QUIET)) {
987
				scmd_printk(KERN_INFO, cmd,
988
					    "Volume overflow, CDB: ");
989
				__scsi_print_command(cmd->cmnd);
990 991
				scsi_print_sense("", cmd);
			}
992
			/* See SSC3rXX or current. */
993
			scsi_end_request(cmd, -EIO, this_count, 1);
L
Linus Torvalds 已提交
994 995 996 997
			return;
		default:
			break;
		}
998
	}
L
Linus Torvalds 已提交
999
	if (host_byte(result) == DID_RESET) {
1000 1001 1002
		/* Third party bus reset or reset for error recovery
		 * reasons.  Just retry the request and see what
		 * happens.
L
Linus Torvalds 已提交
1003 1004 1005 1006 1007
		 */
		scsi_requeue_command(q, cmd);
		return;
	}
	if (result) {
1008
		if (!(req->cmd_flags & REQ_QUIET)) {
1009
			scsi_print_result(cmd);
1010 1011 1012
			if (driver_byte(result) & DRIVER_SENSE)
				scsi_print_sense("", cmd);
		}
L
Linus Torvalds 已提交
1013
	}
1014
	scsi_end_request(cmd, -EIO, this_count, !result);
L
Linus Torvalds 已提交
1015 1016
}

1017 1018
static int scsi_init_sgtable(struct request *req, struct scsi_data_buffer *sdb,
			     gfp_t gfp_mask)
L
Linus Torvalds 已提交
1019
{
1020
	int count;
L
Linus Torvalds 已提交
1021 1022

	/*
C
Christoph Hellwig 已提交
1023
	 * If sg table allocation fails, requeue request later.
L
Linus Torvalds 已提交
1024
	 */
B
Boaz Harrosh 已提交
1025 1026
	if (unlikely(scsi_alloc_sgtable(sdb, req->nr_phys_segments,
					gfp_mask))) {
L
Linus Torvalds 已提交
1027
		return BLKPREP_DEFER;
A
Alan Stern 已提交
1028
	}
L
Linus Torvalds 已提交
1029

C
Christoph Hellwig 已提交
1030
	req->buffer = NULL;
L
Linus Torvalds 已提交
1031 1032 1033 1034 1035

	/* 
	 * Next, walk the list, and fill in the addresses and sizes of
	 * each segment.
	 */
B
Boaz Harrosh 已提交
1036 1037 1038
	count = blk_rq_map_sg(req->q, req, sdb->table.sgl);
	BUG_ON(count > sdb->table.nents);
	sdb->table.nents = count;
T
Tejun Heo 已提交
1039 1040 1041 1042
	if (blk_pc_request(req))
		sdb->length = req->data_len;
	else
		sdb->length = req->nr_sectors << 9;
1043
	return BLKPREP_OK;
L
Linus Torvalds 已提交
1044
}
1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064

/*
 * 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(
1065
			scsi_sdb_cache, GFP_ATOMIC);
1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077
		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;
	}

1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097
	if (blk_integrity_rq(cmd->request)) {
		struct scsi_data_buffer *prot_sdb = cmd->prot_sdb;
		int ivecs, count;

		BUG_ON(prot_sdb == NULL);
		ivecs = blk_rq_count_integrity_sg(cmd->request);

		if (scsi_alloc_sgtable(prot_sdb, ivecs, gfp_mask)) {
			error = BLKPREP_DEFER;
			goto err_exit;
		}

		count = blk_rq_map_integrity_sg(cmd->request,
						prot_sdb->table.sgl);
		BUG_ON(unlikely(count > ivecs));

		cmd->prot_sdb = prot_sdb;
		cmd->prot_sdb->table.nents = count;
	}

1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108
	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;
}
1109
EXPORT_SYMBOL(scsi_init_io);
L
Linus Torvalds 已提交
1110

C
Christoph Hellwig 已提交
1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128
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;

1129 1130
	cmd->cmnd = req->cmd;

C
Christoph Hellwig 已提交
1131 1132 1133
	return cmd;
}

1134
int scsi_setup_blk_pc_cmnd(struct scsi_device *sdev, struct request *req)
J
James Bottomley 已提交
1135
{
C
Christoph Hellwig 已提交
1136
	struct scsi_cmnd *cmd;
1137 1138 1139 1140
	int ret = scsi_prep_state_check(sdev, req);

	if (ret != BLKPREP_OK)
		return ret;
C
Christoph Hellwig 已提交
1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156

	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);

1157
		ret = scsi_init_io(cmd, GFP_ATOMIC);
C
Christoph Hellwig 已提交
1158 1159 1160 1161 1162 1163
		if (unlikely(ret))
			return ret;
	} else {
		BUG_ON(req->data_len);
		BUG_ON(req->data);

B
Boaz Harrosh 已提交
1164
		memset(&cmd->sdb, 0, sizeof(cmd->sdb));
C
Christoph Hellwig 已提交
1165 1166
		req->buffer = NULL;
	}
J
James Bottomley 已提交
1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178

	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;
C
Christoph Hellwig 已提交
1179
	return BLKPREP_OK;
J
James Bottomley 已提交
1180
}
1181
EXPORT_SYMBOL(scsi_setup_blk_pc_cmnd);
J
James Bottomley 已提交
1182

C
Christoph Hellwig 已提交
1183 1184 1185 1186 1187
/*
 * 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.
 */
1188
int scsi_setup_fs_cmnd(struct scsi_device *sdev, struct request *req)
L
Linus Torvalds 已提交
1189 1190
{
	struct scsi_cmnd *cmd;
1191
	int ret = scsi_prep_state_check(sdev, req);
L
Linus Torvalds 已提交
1192

1193 1194
	if (ret != BLKPREP_OK)
		return ret;
1195 1196 1197 1198 1199 1200 1201 1202

	if (unlikely(sdev->scsi_dh_data && sdev->scsi_dh_data->scsi_dh
			 && sdev->scsi_dh_data->scsi_dh->prep_fn)) {
		ret = sdev->scsi_dh_data->scsi_dh->prep_fn(sdev, req);
		if (ret != BLKPREP_OK)
			return ret;
	}

L
Linus Torvalds 已提交
1203
	/*
C
Christoph Hellwig 已提交
1204
	 * Filesystem requests must transfer data.
L
Linus Torvalds 已提交
1205
	 */
C
Christoph Hellwig 已提交
1206 1207 1208 1209 1210 1211
	BUG_ON(!req->nr_phys_segments);

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

1212
	memset(cmd->cmnd, 0, BLK_MAX_CDB);
1213
	return scsi_init_io(cmd, GFP_ATOMIC);
C
Christoph Hellwig 已提交
1214
}
1215
EXPORT_SYMBOL(scsi_setup_fs_cmnd);
C
Christoph Hellwig 已提交
1216

1217
int scsi_prep_state_check(struct scsi_device *sdev, struct request *req)
C
Christoph Hellwig 已提交
1218 1219 1220
{
	int ret = BLKPREP_OK;

L
Linus Torvalds 已提交
1221
	/*
C
Christoph Hellwig 已提交
1222 1223
	 * If the device is not in running state we will reject some
	 * or all commands.
L
Linus Torvalds 已提交
1224
	 */
C
Christoph Hellwig 已提交
1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241
	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.
			 */
1242
			sdev_printk(KERN_ERR, sdev,
C
Christoph Hellwig 已提交
1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262
				    "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;
L
Linus Torvalds 已提交
1263 1264
		}
	}
1265 1266 1267
	return ret;
}
EXPORT_SYMBOL(scsi_prep_state_check);
L
Linus Torvalds 已提交
1268

1269 1270 1271
int scsi_prep_return(struct request_queue *q, struct request *req, int ret)
{
	struct scsi_device *sdev = q->queuedata;
L
Linus Torvalds 已提交
1272

C
Christoph Hellwig 已提交
1273 1274 1275
	switch (ret) {
	case BLKPREP_KILL:
		req->errors = DID_NO_CONNECT << 16;
1276 1277 1278 1279 1280 1281 1282
		/* 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;
		}
C
Christoph Hellwig 已提交
1283 1284
		break;
	case BLKPREP_DEFER:
L
Linus Torvalds 已提交
1285
		/*
C
Christoph Hellwig 已提交
1286 1287 1288
		 * 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.
L
Linus Torvalds 已提交
1289
		 */
C
Christoph Hellwig 已提交
1290 1291 1292 1293 1294
		if (sdev->device_busy == 0)
			blk_plug_device(q);
		break;
	default:
		req->cmd_flags |= REQ_DONTPREP;
L
Linus Torvalds 已提交
1295 1296
	}

C
Christoph Hellwig 已提交
1297
	return ret;
L
Linus Torvalds 已提交
1298
}
1299 1300
EXPORT_SYMBOL(scsi_prep_return);

1301
int scsi_prep_fn(struct request_queue *q, struct request *req)
1302 1303 1304 1305 1306 1307 1308 1309
{
	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);
}
L
Linus Torvalds 已提交
1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327

/*
 * 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,
1328 1329
				   sdev_printk(KERN_INFO, sdev,
				   "unblocking device at zero depth\n"));
L
Linus Torvalds 已提交
1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351
		} 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)
{
1352
	if (scsi_host_in_recovery(shost))
L
Linus Torvalds 已提交
1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380
		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 {
			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;
}

/*
1381
 * Kill a request for a dead device
L
Linus Torvalds 已提交
1382
 */
1383
static void scsi_kill_request(struct request *req, struct request_queue *q)
L
Linus Torvalds 已提交
1384
{
1385
	struct scsi_cmnd *cmd = req->special;
1386 1387
	struct scsi_device *sdev = cmd->device;
	struct Scsi_Host *shost = sdev->host;
L
Linus Torvalds 已提交
1388

1389 1390
	blkdev_dequeue_request(req);

1391 1392 1393 1394
	if (unlikely(cmd == NULL)) {
		printk(KERN_CRIT "impossible request in %s.\n",
				 __FUNCTION__);
		BUG();
L
Linus Torvalds 已提交
1395
	}
1396 1397 1398 1399

	scsi_init_cmd_errh(cmd);
	cmd->result = DID_NO_CONNECT << 16;
	atomic_inc(&cmd->device->iorequest_cnt);
1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412

	/*
	 * SCSI request completion path will do scsi_device_unbusy(),
	 * bump busy counts.  To bump the counters, we need to dance
	 * with the locks as normal issue path does.
	 */
	sdev->device_busy++;
	spin_unlock(sdev->request_queue->queue_lock);
	spin_lock(shost->host_lock);
	shost->host_busy++;
	spin_unlock(shost->host_lock);
	spin_lock(sdev->request_queue->queue_lock);

1413
	__scsi_done(cmd);
L
Linus Torvalds 已提交
1414 1415
}

1416 1417 1418
static void scsi_softirq_done(struct request *rq)
{
	struct scsi_cmnd *cmd = rq->completion_data;
1419
	unsigned long wait_for = (cmd->allowed + 1) * cmd->timeout_per_command;
1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439
	int disposition;

	INIT_LIST_HEAD(&cmd->eh_entry);

	disposition = scsi_decide_disposition(cmd);
	if (disposition != SUCCESS &&
	    time_before(cmd->jiffies_at_alloc + wait_for, jiffies)) {
		sdev_printk(KERN_ERR, cmd->device,
			    "timing out command, waited %lus\n",
			    wait_for/HZ);
		disposition = SUCCESS;
	}
			
	scsi_log_completion(cmd, disposition);

	switch (disposition) {
		case SUCCESS:
			scsi_finish_command(cmd);
			break;
		case NEEDS_RETRY:
1440
			scsi_queue_insert(cmd, SCSI_MLQUEUE_EH_RETRY);
1441 1442 1443 1444 1445 1446 1447 1448 1449 1450
			break;
		case ADD_TO_MLQUEUE:
			scsi_queue_insert(cmd, SCSI_MLQUEUE_DEVICE_BUSY);
			break;
		default:
			if (!scsi_eh_scmd_add(cmd, 0))
				scsi_finish_command(cmd);
	}
}

L
Linus Torvalds 已提交
1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470
/*
 * Function:    scsi_request_fn()
 *
 * Purpose:     Main strategy routine for SCSI.
 *
 * Arguments:   q       - Pointer to actual queue.
 *
 * Returns:     Nothing
 *
 * Lock status: IO request lock assumed to be held when called.
 */
static void scsi_request_fn(struct request_queue *q)
{
	struct scsi_device *sdev = q->queuedata;
	struct Scsi_Host *shost;
	struct scsi_cmnd *cmd;
	struct request *req;

	if (!sdev) {
		printk("scsi: killing requests for dead queue\n");
1471 1472
		while ((req = elv_next_request(q)) != NULL)
			scsi_kill_request(req, q);
L
Linus Torvalds 已提交
1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496
		return;
	}

	if(!get_device(&sdev->sdev_gendev))
		/* We must be tearing the block queue down already */
		return;

	/*
	 * To start with, we keep looping until the queue is empty, or until
	 * the host is no longer able to accept any more requests.
	 */
	shost = sdev->host;
	while (!blk_queue_plugged(q)) {
		int rtn;
		/*
		 * get next queueable request.  We do this early to make sure
		 * that the request is fully prepared even if we cannot 
		 * accept it.
		 */
		req = elv_next_request(q);
		if (!req || !scsi_dev_queue_ready(q, sdev))
			break;

		if (unlikely(!scsi_device_online(sdev))) {
1497 1498
			sdev_printk(KERN_ERR, sdev,
				    "rejecting I/O to offline device\n");
1499
			scsi_kill_request(req, q);
L
Linus Torvalds 已提交
1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511
			continue;
		}


		/*
		 * Remove the request from the request list.
		 */
		if (!(blk_queue_tagged(q) && !blk_queue_start_tag(q, req)))
			blkdev_dequeue_request(req);
		sdev->device_busy++;

		spin_unlock(q->queue_lock);
1512 1513 1514 1515
		cmd = req->special;
		if (unlikely(cmd == NULL)) {
			printk(KERN_CRIT "impossible request in %s.\n"
					 "please mail a stack trace to "
1516
					 "linux-scsi@vger.kernel.org\n",
1517
					 __FUNCTION__);
1518
			blk_dump_rq_flags(req, "foo");
1519 1520
			BUG();
		}
L
Linus Torvalds 已提交
1521 1522
		spin_lock(shost->host_lock);

1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537
		/*
		 * We hit this when the driver is using a host wide
		 * tag map. For device level tag maps the queue_depth check
		 * in the device ready fn would prevent us from trying
		 * to allocate a tag. Since the map is a shared host resource
		 * we add the dev to the starved list so it eventually gets
		 * a run when a tag is freed.
		 */
		if (blk_queue_tagged(q) && (req->tag == -1)) {
			if (list_empty(&sdev->starved_entry))
				list_add_tail(&sdev->starved_entry,
					      &shost->starved_list);
			goto not_ready;
		}

L
Linus Torvalds 已提交
1538 1539
		if (!scsi_host_queue_ready(q, shost, sdev))
			goto not_ready;
1540
		if (scsi_target(sdev)->single_lun) {
L
Linus Torvalds 已提交
1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623
			if (scsi_target(sdev)->starget_sdev_user &&
			    scsi_target(sdev)->starget_sdev_user != sdev)
				goto not_ready;
			scsi_target(sdev)->starget_sdev_user = sdev;
		}
		shost->host_busy++;

		/*
		 * XXX(hch): This is rather suboptimal, scsi_dispatch_cmd will
		 *		take the lock again.
		 */
		spin_unlock_irq(shost->host_lock);

		/*
		 * Finally, initialize any error handling parameters, and set up
		 * the timers for timeouts.
		 */
		scsi_init_cmd_errh(cmd);

		/*
		 * Dispatch the command to the low-level driver.
		 */
		rtn = scsi_dispatch_cmd(cmd);
		spin_lock_irq(q->queue_lock);
		if(rtn) {
			/* we're refusing the command; because of
			 * the way locks get dropped, we need to 
			 * check here if plugging is required */
			if(sdev->device_busy == 0)
				blk_plug_device(q);

			break;
		}
	}

	goto out;

 not_ready:
	spin_unlock_irq(shost->host_lock);

	/*
	 * lock q, handle tag, requeue req, and decrement device_busy. We
	 * must return with queue_lock held.
	 *
	 * Decrementing device_busy without checking it is OK, as all such
	 * cases (host limits or settings) should run the queue at some
	 * later time.
	 */
	spin_lock_irq(q->queue_lock);
	blk_requeue_request(q, req);
	sdev->device_busy--;
	if(sdev->device_busy == 0)
		blk_plug_device(q);
 out:
	/* must be careful here...if we trigger the ->remove() function
	 * we cannot be holding the q lock */
	spin_unlock_irq(q->queue_lock);
	put_device(&sdev->sdev_gendev);
	spin_lock_irq(q->queue_lock);
}

u64 scsi_calculate_bounce_limit(struct Scsi_Host *shost)
{
	struct device *host_dev;
	u64 bounce_limit = 0xffffffff;

	if (shost->unchecked_isa_dma)
		return BLK_BOUNCE_ISA;
	/*
	 * Platforms with virtual-DMA translation
	 * hardware have no practical limit.
	 */
	if (!PCI_DMA_BUS_IS_PHYS)
		return BLK_BOUNCE_ANY;

	host_dev = scsi_get_device(shost);
	if (host_dev && host_dev->dma_mask)
		bounce_limit = *host_dev->dma_mask;

	return bounce_limit;
}
EXPORT_SYMBOL(scsi_calculate_bounce_limit);

1624 1625
struct request_queue *__scsi_alloc_queue(struct Scsi_Host *shost,
					 request_fn_proc *request_fn)
L
Linus Torvalds 已提交
1626 1627
{
	struct request_queue *q;
1628
	struct device *dev = shost->shost_gendev.parent;
L
Linus Torvalds 已提交
1629

1630
	q = blk_init_queue(request_fn, NULL);
L
Linus Torvalds 已提交
1631 1632 1633
	if (!q)
		return NULL;

1634 1635 1636
	/*
	 * this limit is imposed by hardware restrictions
	 */
L
Linus Torvalds 已提交
1637
	blk_queue_max_hw_segments(q, shost->sg_tablesize);
J
James Bottomley 已提交
1638
	blk_queue_max_phys_segments(q, SCSI_MAX_SG_CHAIN_SEGMENTS);
1639

L
Linus Torvalds 已提交
1640 1641 1642
	blk_queue_max_sectors(q, shost->max_sectors);
	blk_queue_bounce_limit(q, scsi_calculate_bounce_limit(shost));
	blk_queue_segment_boundary(q, shost->dma_boundary);
1643
	dma_set_seg_boundary(dev, shost->dma_boundary);
L
Linus Torvalds 已提交
1644

1645 1646
	blk_queue_max_segment_size(q, dma_get_max_seg_size(dev));

N
Nick Piggin 已提交
1647
	/* New queue, no concurrency on queue_flags */
L
Linus Torvalds 已提交
1648
	if (!shost->use_clustering)
N
Nick Piggin 已提交
1649
		queue_flag_clear_unlocked(QUEUE_FLAG_CLUSTER, q);
1650 1651 1652 1653 1654 1655 1656 1657

	/*
	 * set a reasonable default alignment on word boundaries: the
	 * host and device may alter it using
	 * blk_queue_update_dma_alignment() later.
	 */
	blk_queue_dma_alignment(q, 0x03);

L
Linus Torvalds 已提交
1658 1659
	return q;
}
1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673
EXPORT_SYMBOL(__scsi_alloc_queue);

struct request_queue *scsi_alloc_queue(struct scsi_device *sdev)
{
	struct request_queue *q;

	q = __scsi_alloc_queue(sdev->host, scsi_request_fn);
	if (!q)
		return NULL;

	blk_queue_prep_rq(q, scsi_prep_fn);
	blk_queue_softirq_done(q, scsi_softirq_done);
	return q;
}
L
Linus Torvalds 已提交
1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732

void scsi_free_queue(struct request_queue *q)
{
	blk_cleanup_queue(q);
}

/*
 * Function:    scsi_block_requests()
 *
 * Purpose:     Utility function used by low-level drivers to prevent further
 *		commands from being queued to the device.
 *
 * Arguments:   shost       - Host in question
 *
 * Returns:     Nothing
 *
 * Lock status: No locks are assumed held.
 *
 * Notes:       There is no timer nor any other means by which the requests
 *		get unblocked other than the low-level driver calling
 *		scsi_unblock_requests().
 */
void scsi_block_requests(struct Scsi_Host *shost)
{
	shost->host_self_blocked = 1;
}
EXPORT_SYMBOL(scsi_block_requests);

/*
 * Function:    scsi_unblock_requests()
 *
 * Purpose:     Utility function used by low-level drivers to allow further
 *		commands from being queued to the device.
 *
 * Arguments:   shost       - Host in question
 *
 * Returns:     Nothing
 *
 * Lock status: No locks are assumed held.
 *
 * Notes:       There is no timer nor any other means by which the requests
 *		get unblocked other than the low-level driver calling
 *		scsi_unblock_requests().
 *
 *		This is done as an API function so that changes to the
 *		internals of the scsi mid-layer won't require wholesale
 *		changes to drivers that use this feature.
 */
void scsi_unblock_requests(struct Scsi_Host *shost)
{
	shost->host_self_blocked = 0;
	scsi_run_host_queues(shost);
}
EXPORT_SYMBOL(scsi_unblock_requests);

int __init scsi_init_queue(void)
{
	int i;

1733 1734
	scsi_io_context_cache = kmem_cache_create("scsi_io_context",
					sizeof(struct scsi_io_context),
1735
					0, 0, NULL);
1736 1737 1738 1739 1740
	if (!scsi_io_context_cache) {
		printk(KERN_ERR "SCSI: can't init scsi io context cache\n");
		return -ENOMEM;
	}

1741 1742 1743 1744 1745
	scsi_sdb_cache = kmem_cache_create("scsi_data_buffer",
					   sizeof(struct scsi_data_buffer),
					   0, 0, NULL);
	if (!scsi_sdb_cache) {
		printk(KERN_ERR "SCSI: can't init scsi sdb cache\n");
1746
		goto cleanup_io_context;
1747 1748
	}

L
Linus Torvalds 已提交
1749 1750 1751 1752 1753
	for (i = 0; i < SG_MEMPOOL_NR; i++) {
		struct scsi_host_sg_pool *sgp = scsi_sg_pools + i;
		int size = sgp->size * sizeof(struct scatterlist);

		sgp->slab = kmem_cache_create(sgp->name, size, 0,
1754
				SLAB_HWCACHE_ALIGN, NULL);
L
Linus Torvalds 已提交
1755 1756 1757
		if (!sgp->slab) {
			printk(KERN_ERR "SCSI: can't init sg slab %s\n",
					sgp->name);
1758
			goto cleanup_sdb;
L
Linus Torvalds 已提交
1759 1760
		}

1761 1762
		sgp->pool = mempool_create_slab_pool(SG_MEMPOOL_SIZE,
						     sgp->slab);
L
Linus Torvalds 已提交
1763 1764 1765
		if (!sgp->pool) {
			printk(KERN_ERR "SCSI: can't init sg mempool %s\n",
					sgp->name);
1766
			goto cleanup_sdb;
L
Linus Torvalds 已提交
1767 1768 1769 1770
		}
	}

	return 0;
1771

1772
cleanup_sdb:
1773 1774 1775 1776 1777 1778 1779
	for (i = 0; i < SG_MEMPOOL_NR; i++) {
		struct scsi_host_sg_pool *sgp = scsi_sg_pools + i;
		if (sgp->pool)
			mempool_destroy(sgp->pool);
		if (sgp->slab)
			kmem_cache_destroy(sgp->slab);
	}
1780
	kmem_cache_destroy(scsi_sdb_cache);
1781 1782 1783 1784
cleanup_io_context:
	kmem_cache_destroy(scsi_io_context_cache);

	return -ENOMEM;
L
Linus Torvalds 已提交
1785 1786 1787 1788 1789 1790
}

void scsi_exit_queue(void)
{
	int i;

1791
	kmem_cache_destroy(scsi_io_context_cache);
1792
	kmem_cache_destroy(scsi_sdb_cache);
1793

L
Linus Torvalds 已提交
1794 1795 1796 1797 1798 1799
	for (i = 0; i < SG_MEMPOOL_NR; i++) {
		struct scsi_host_sg_pool *sgp = scsi_sg_pools + i;
		mempool_destroy(sgp->pool);
		kmem_cache_destroy(sgp->slab);
	}
}
1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811

/**
 *	scsi_mode_select - issue a mode select
 *	@sdev:	SCSI device to be queried
 *	@pf:	Page format bit (1 == standard, 0 == vendor specific)
 *	@sp:	Save page bit (0 == don't save, 1 == save)
 *	@modepage: mode page being requested
 *	@buffer: request buffer (may not be smaller than eight bytes)
 *	@len:	length of request buffer.
 *	@timeout: command timeout
 *	@retries: number of retries before failing
 *	@data: returns a structure abstracting the mode header data
1812
 *	@sshdr: place to put sense data (or NULL if no sense to be collected).
1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877
 *		must be SCSI_SENSE_BUFFERSIZE big.
 *
 *	Returns zero if successful; negative error number or scsi
 *	status on error
 *
 */
int
scsi_mode_select(struct scsi_device *sdev, int pf, int sp, int modepage,
		 unsigned char *buffer, int len, int timeout, int retries,
		 struct scsi_mode_data *data, struct scsi_sense_hdr *sshdr)
{
	unsigned char cmd[10];
	unsigned char *real_buffer;
	int ret;

	memset(cmd, 0, sizeof(cmd));
	cmd[1] = (pf ? 0x10 : 0) | (sp ? 0x01 : 0);

	if (sdev->use_10_for_ms) {
		if (len > 65535)
			return -EINVAL;
		real_buffer = kmalloc(8 + len, GFP_KERNEL);
		if (!real_buffer)
			return -ENOMEM;
		memcpy(real_buffer + 8, buffer, len);
		len += 8;
		real_buffer[0] = 0;
		real_buffer[1] = 0;
		real_buffer[2] = data->medium_type;
		real_buffer[3] = data->device_specific;
		real_buffer[4] = data->longlba ? 0x01 : 0;
		real_buffer[5] = 0;
		real_buffer[6] = data->block_descriptor_length >> 8;
		real_buffer[7] = data->block_descriptor_length;

		cmd[0] = MODE_SELECT_10;
		cmd[7] = len >> 8;
		cmd[8] = len;
	} else {
		if (len > 255 || data->block_descriptor_length > 255 ||
		    data->longlba)
			return -EINVAL;

		real_buffer = kmalloc(4 + len, GFP_KERNEL);
		if (!real_buffer)
			return -ENOMEM;
		memcpy(real_buffer + 4, buffer, len);
		len += 4;
		real_buffer[0] = 0;
		real_buffer[1] = data->medium_type;
		real_buffer[2] = data->device_specific;
		real_buffer[3] = data->block_descriptor_length;
		

		cmd[0] = MODE_SELECT;
		cmd[4] = len;
	}

	ret = scsi_execute_req(sdev, cmd, DMA_TO_DEVICE, real_buffer, len,
			       sshdr, timeout, retries);
	kfree(real_buffer);
	return ret;
}
EXPORT_SYMBOL_GPL(scsi_mode_select);

L
Linus Torvalds 已提交
1878
/**
1879
 *	scsi_mode_sense - issue a mode sense, falling back from 10 to six bytes if necessary.
1880
 *	@sdev:	SCSI device to be queried
L
Linus Torvalds 已提交
1881 1882 1883 1884 1885 1886 1887
 *	@dbd:	set if mode sense will allow block descriptors to be returned
 *	@modepage: mode page being requested
 *	@buffer: request buffer (may not be smaller than eight bytes)
 *	@len:	length of request buffer.
 *	@timeout: command timeout
 *	@retries: number of retries before failing
 *	@data: returns a structure abstracting the mode header data
1888
 *	@sshdr: place to put sense data (or NULL if no sense to be collected).
1889
 *		must be SCSI_SENSE_BUFFERSIZE big.
L
Linus Torvalds 已提交
1890 1891 1892 1893
 *
 *	Returns zero if unsuccessful, or the header offset (either 4
 *	or 8 depending on whether a six or ten byte command was
 *	issued) if successful.
1894
 */
L
Linus Torvalds 已提交
1895
int
1896
scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage,
L
Linus Torvalds 已提交
1897
		  unsigned char *buffer, int len, int timeout, int retries,
1898 1899
		  struct scsi_mode_data *data, struct scsi_sense_hdr *sshdr)
{
L
Linus Torvalds 已提交
1900 1901 1902
	unsigned char cmd[12];
	int use_10_for_ms;
	int header_length;
1903
	int result;
1904
	struct scsi_sense_hdr my_sshdr;
L
Linus Torvalds 已提交
1905 1906 1907 1908 1909 1910

	memset(data, 0, sizeof(*data));
	memset(&cmd[0], 0, 12);
	cmd[1] = dbd & 0x18;	/* allows DBD and LLBA bits */
	cmd[2] = modepage;

1911 1912 1913 1914
	/* caller might not be interested in sense, but we need it */
	if (!sshdr)
		sshdr = &my_sshdr;

L
Linus Torvalds 已提交
1915
 retry:
1916
	use_10_for_ms = sdev->use_10_for_ms;
L
Linus Torvalds 已提交
1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935

	if (use_10_for_ms) {
		if (len < 8)
			len = 8;

		cmd[0] = MODE_SENSE_10;
		cmd[8] = len;
		header_length = 8;
	} else {
		if (len < 4)
			len = 4;

		cmd[0] = MODE_SENSE;
		cmd[4] = len;
		header_length = 4;
	}

	memset(buffer, 0, len);

1936
	result = scsi_execute_req(sdev, cmd, DMA_FROM_DEVICE, buffer, len,
1937
				  sshdr, timeout, retries);
L
Linus Torvalds 已提交
1938 1939 1940 1941 1942 1943

	/* This code looks awful: what it's doing is making sure an
	 * ILLEGAL REQUEST sense return identifies the actual command
	 * byte as the problem.  MODE_SENSE commands can return
	 * ILLEGAL REQUEST if the code page isn't supported */

1944 1945
	if (use_10_for_ms && !scsi_status_is_good(result) &&
	    (driver_byte(result) & DRIVER_SENSE)) {
1946 1947 1948
		if (scsi_sense_valid(sshdr)) {
			if ((sshdr->sense_key == ILLEGAL_REQUEST) &&
			    (sshdr->asc == 0x20) && (sshdr->ascq == 0)) {
L
Linus Torvalds 已提交
1949 1950 1951
				/* 
				 * Invalid command operation code
				 */
1952
				sdev->use_10_for_ms = 0;
L
Linus Torvalds 已提交
1953 1954 1955 1956 1957
				goto retry;
			}
		}
	}

1958
	if(scsi_status_is_good(result)) {
1959 1960 1961 1962 1963 1964 1965 1966 1967 1968
		if (unlikely(buffer[0] == 0x86 && buffer[1] == 0x0b &&
			     (modepage == 6 || modepage == 8))) {
			/* Initio breakage? */
			header_length = 0;
			data->length = 13;
			data->medium_type = 0;
			data->device_specific = 0;
			data->longlba = 0;
			data->block_descriptor_length = 0;
		} else if(use_10_for_ms) {
L
Linus Torvalds 已提交
1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980
			data->length = buffer[0]*256 + buffer[1] + 2;
			data->medium_type = buffer[2];
			data->device_specific = buffer[3];
			data->longlba = buffer[4] & 0x01;
			data->block_descriptor_length = buffer[6]*256
				+ buffer[7];
		} else {
			data->length = buffer[0] + 1;
			data->medium_type = buffer[1];
			data->device_specific = buffer[2];
			data->block_descriptor_length = buffer[3];
		}
1981
		data->header_length = header_length;
L
Linus Torvalds 已提交
1982 1983
	}

1984
	return result;
L
Linus Torvalds 已提交
1985 1986 1987
}
EXPORT_SYMBOL(scsi_mode_sense);

1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
/**
 *	scsi_test_unit_ready - test if unit is ready
 *	@sdev:	scsi device to change the state of.
 *	@timeout: command timeout
 *	@retries: number of retries before failing
 *	@sshdr_external: Optional pointer to struct scsi_sense_hdr for
 *		returning sense. Make sure that this is cleared before passing
 *		in.
 *
 *	Returns zero if unsuccessful or an error if TUR failed.  For
 *	removable media, a return of NOT_READY or UNIT_ATTENTION is
 *	translated to success, with the ->changed flag updated.
 **/
L
Linus Torvalds 已提交
2001
int
2002 2003
scsi_test_unit_ready(struct scsi_device *sdev, int timeout, int retries,
		     struct scsi_sense_hdr *sshdr_external)
L
Linus Torvalds 已提交
2004 2005 2006 2007
{
	char cmd[] = {
		TEST_UNIT_READY, 0, 0, 0, 0, 0,
	};
2008
	struct scsi_sense_hdr *sshdr;
L
Linus Torvalds 已提交
2009
	int result;
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026

	if (!sshdr_external)
		sshdr = kzalloc(sizeof(*sshdr), GFP_KERNEL);
	else
		sshdr = sshdr_external;

	/* try to eat the UNIT_ATTENTION if there are enough retries */
	do {
		result = scsi_execute_req(sdev, cmd, DMA_NONE, NULL, 0, sshdr,
					  timeout, retries);
	} while ((driver_byte(result) & DRIVER_SENSE) &&
		 sshdr && sshdr->sense_key == UNIT_ATTENTION &&
		 --retries);

	if (!sshdr)
		/* could not allocate sense buffer, so can't process it */
		return result;
L
Linus Torvalds 已提交
2027

2028
	if ((driver_byte(result) & DRIVER_SENSE) && sdev->removable) {
L
Linus Torvalds 已提交
2029

2030 2031 2032
		if ((scsi_sense_valid(sshdr)) &&
		    ((sshdr->sense_key == UNIT_ATTENTION) ||
		     (sshdr->sense_key == NOT_READY))) {
L
Linus Torvalds 已提交
2033
			sdev->changed = 1;
2034
			result = 0;
L
Linus Torvalds 已提交
2035 2036
		}
	}
2037 2038
	if (!sshdr_external)
		kfree(sshdr);
L
Linus Torvalds 已提交
2039 2040 2041 2042 2043
	return result;
}
EXPORT_SYMBOL(scsi_test_unit_ready);

/**
2044
 *	scsi_device_set_state - Take the given device through the device state model.
L
Linus Torvalds 已提交
2045 2046 2047 2048 2049
 *	@sdev:	scsi device to change the state of.
 *	@state:	state to change to.
 *
 *	Returns zero if unsuccessful or an error if the requested 
 *	transition is illegal.
2050
 */
L
Linus Torvalds 已提交
2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113
int
scsi_device_set_state(struct scsi_device *sdev, enum scsi_device_state state)
{
	enum scsi_device_state oldstate = sdev->sdev_state;

	if (state == oldstate)
		return 0;

	switch (state) {
	case SDEV_CREATED:
		/* There are no legal states that come back to
		 * created.  This is the manually initialised start
		 * state */
		goto illegal;
			
	case SDEV_RUNNING:
		switch (oldstate) {
		case SDEV_CREATED:
		case SDEV_OFFLINE:
		case SDEV_QUIESCE:
		case SDEV_BLOCK:
			break;
		default:
			goto illegal;
		}
		break;

	case SDEV_QUIESCE:
		switch (oldstate) {
		case SDEV_RUNNING:
		case SDEV_OFFLINE:
			break;
		default:
			goto illegal;
		}
		break;

	case SDEV_OFFLINE:
		switch (oldstate) {
		case SDEV_CREATED:
		case SDEV_RUNNING:
		case SDEV_QUIESCE:
		case SDEV_BLOCK:
			break;
		default:
			goto illegal;
		}
		break;

	case SDEV_BLOCK:
		switch (oldstate) {
		case SDEV_CREATED:
		case SDEV_RUNNING:
			break;
		default:
			goto illegal;
		}
		break;

	case SDEV_CANCEL:
		switch (oldstate) {
		case SDEV_CREATED:
		case SDEV_RUNNING:
2114
		case SDEV_QUIESCE:
L
Linus Torvalds 已提交
2115 2116 2117 2118 2119 2120 2121 2122 2123 2124
		case SDEV_OFFLINE:
		case SDEV_BLOCK:
			break;
		default:
			goto illegal;
		}
		break;

	case SDEV_DEL:
		switch (oldstate) {
2125 2126 2127
		case SDEV_CREATED:
		case SDEV_RUNNING:
		case SDEV_OFFLINE:
L
Linus Torvalds 已提交
2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140
		case SDEV_CANCEL:
			break;
		default:
			goto illegal;
		}
		break;

	}
	sdev->sdev_state = state;
	return 0;

 illegal:
	SCSI_LOG_ERROR_RECOVERY(1, 
2141 2142 2143 2144
				sdev_printk(KERN_ERR, sdev,
					    "Illegal state transition %s->%s\n",
					    scsi_device_state_name(oldstate),
					    scsi_device_state_name(state))
L
Linus Torvalds 已提交
2145 2146 2147 2148 2149
				);
	return -EINVAL;
}
EXPORT_SYMBOL(scsi_device_set_state);

2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222
/**
 * 	sdev_evt_emit - emit a single SCSI device uevent
 *	@sdev: associated SCSI device
 *	@evt: event to emit
 *
 *	Send a single uevent (scsi_event) to the associated scsi_device.
 */
static void scsi_evt_emit(struct scsi_device *sdev, struct scsi_event *evt)
{
	int idx = 0;
	char *envp[3];

	switch (evt->evt_type) {
	case SDEV_EVT_MEDIA_CHANGE:
		envp[idx++] = "SDEV_MEDIA_CHANGE=1";
		break;

	default:
		/* do nothing */
		break;
	}

	envp[idx++] = NULL;

	kobject_uevent_env(&sdev->sdev_gendev.kobj, KOBJ_CHANGE, envp);
}

/**
 * 	sdev_evt_thread - send a uevent for each scsi event
 *	@work: work struct for scsi_device
 *
 *	Dispatch queued events to their associated scsi_device kobjects
 *	as uevents.
 */
void scsi_evt_thread(struct work_struct *work)
{
	struct scsi_device *sdev;
	LIST_HEAD(event_list);

	sdev = container_of(work, struct scsi_device, event_work);

	while (1) {
		struct scsi_event *evt;
		struct list_head *this, *tmp;
		unsigned long flags;

		spin_lock_irqsave(&sdev->list_lock, flags);
		list_splice_init(&sdev->event_list, &event_list);
		spin_unlock_irqrestore(&sdev->list_lock, flags);

		if (list_empty(&event_list))
			break;

		list_for_each_safe(this, tmp, &event_list) {
			evt = list_entry(this, struct scsi_event, node);
			list_del(&evt->node);
			scsi_evt_emit(sdev, evt);
			kfree(evt);
		}
	}
}

/**
 * 	sdev_evt_send - send asserted event to uevent thread
 *	@sdev: scsi_device event occurred on
 *	@evt: event to send
 *
 *	Assert scsi device event asynchronously.
 */
void sdev_evt_send(struct scsi_device *sdev, struct scsi_event *evt)
{
	unsigned long flags;

2223 2224 2225 2226
#if 0
	/* FIXME: currently this check eliminates all media change events
	 * for polled devices.  Need to update to discriminate between AN
	 * and polled events */
2227 2228 2229 2230
	if (!test_bit(evt->evt_type, sdev->supported_events)) {
		kfree(evt);
		return;
	}
2231
#endif
2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290

	spin_lock_irqsave(&sdev->list_lock, flags);
	list_add_tail(&evt->node, &sdev->event_list);
	schedule_work(&sdev->event_work);
	spin_unlock_irqrestore(&sdev->list_lock, flags);
}
EXPORT_SYMBOL_GPL(sdev_evt_send);

/**
 * 	sdev_evt_alloc - allocate a new scsi event
 *	@evt_type: type of event to allocate
 *	@gfpflags: GFP flags for allocation
 *
 *	Allocates and returns a new scsi_event.
 */
struct scsi_event *sdev_evt_alloc(enum scsi_device_event evt_type,
				  gfp_t gfpflags)
{
	struct scsi_event *evt = kzalloc(sizeof(struct scsi_event), gfpflags);
	if (!evt)
		return NULL;

	evt->evt_type = evt_type;
	INIT_LIST_HEAD(&evt->node);

	/* evt_type-specific initialization, if any */
	switch (evt_type) {
	case SDEV_EVT_MEDIA_CHANGE:
	default:
		/* do nothing */
		break;
	}

	return evt;
}
EXPORT_SYMBOL_GPL(sdev_evt_alloc);

/**
 * 	sdev_evt_send_simple - send asserted event to uevent thread
 *	@sdev: scsi_device event occurred on
 *	@evt_type: type of event to send
 *	@gfpflags: GFP flags for allocation
 *
 *	Assert scsi device event asynchronously, given an event type.
 */
void sdev_evt_send_simple(struct scsi_device *sdev,
			  enum scsi_device_event evt_type, gfp_t gfpflags)
{
	struct scsi_event *evt = sdev_evt_alloc(evt_type, gfpflags);
	if (!evt) {
		sdev_printk(KERN_ERR, sdev, "event %d eaten due to OOM\n",
			    evt_type);
		return;
	}

	sdev_evt_send(sdev, evt);
}
EXPORT_SYMBOL_GPL(sdev_evt_send_simple);

L
Linus Torvalds 已提交
2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304
/**
 *	scsi_device_quiesce - Block user issued commands.
 *	@sdev:	scsi device to quiesce.
 *
 *	This works by trying to transition to the SDEV_QUIESCE state
 *	(which must be a legal transition).  When the device is in this
 *	state, only special requests will be accepted, all others will
 *	be deferred.  Since special requests may also be requeued requests,
 *	a successful return doesn't guarantee the device will be 
 *	totally quiescent.
 *
 *	Must be called with user context, may sleep.
 *
 *	Returns zero if unsuccessful or an error if not.
2305
 */
L
Linus Torvalds 已提交
2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329
int
scsi_device_quiesce(struct scsi_device *sdev)
{
	int err = scsi_device_set_state(sdev, SDEV_QUIESCE);
	if (err)
		return err;

	scsi_run_queue(sdev->request_queue);
	while (sdev->device_busy) {
		msleep_interruptible(200);
		scsi_run_queue(sdev->request_queue);
	}
	return 0;
}
EXPORT_SYMBOL(scsi_device_quiesce);

/**
 *	scsi_device_resume - Restart user issued commands to a quiesced device.
 *	@sdev:	scsi device to resume.
 *
 *	Moves the device from quiesced back to running and restarts the
 *	queues.
 *
 *	Must be called with user context, may sleep.
2330
 */
L
Linus Torvalds 已提交
2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366
void
scsi_device_resume(struct scsi_device *sdev)
{
	if(scsi_device_set_state(sdev, SDEV_RUNNING))
		return;
	scsi_run_queue(sdev->request_queue);
}
EXPORT_SYMBOL(scsi_device_resume);

static void
device_quiesce_fn(struct scsi_device *sdev, void *data)
{
	scsi_device_quiesce(sdev);
}

void
scsi_target_quiesce(struct scsi_target *starget)
{
	starget_for_each_device(starget, NULL, device_quiesce_fn);
}
EXPORT_SYMBOL(scsi_target_quiesce);

static void
device_resume_fn(struct scsi_device *sdev, void *data)
{
	scsi_device_resume(sdev);
}

void
scsi_target_resume(struct scsi_target *starget)
{
	starget_for_each_device(starget, NULL, device_resume_fn);
}
EXPORT_SYMBOL(scsi_target_resume);

/**
2367
 * scsi_internal_device_block - internal function to put a device temporarily into the SDEV_BLOCK state
L
Linus Torvalds 已提交
2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381
 * @sdev:	device to block
 *
 * Block request made by scsi lld's to temporarily stop all
 * scsi commands on the specified device.  Called from interrupt
 * or normal process context.
 *
 * Returns zero if successful or error if not
 *
 * Notes:       
 *	This routine transitions the device to the SDEV_BLOCK state
 *	(which must be a legal transition).  When the device is in this
 *	state, all commands are deferred until the scsi lld reenables
 *	the device with scsi_device_unblock or device_block_tmo fires.
 *	This routine assumes the host_lock is held on entry.
2382
 */
L
Linus Torvalds 已提交
2383 2384 2385
int
scsi_internal_device_block(struct scsi_device *sdev)
{
2386
	struct request_queue *q = sdev->request_queue;
L
Linus Torvalds 已提交
2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421
	unsigned long flags;
	int err = 0;

	err = scsi_device_set_state(sdev, SDEV_BLOCK);
	if (err)
		return err;

	/* 
	 * The device has transitioned to SDEV_BLOCK.  Stop the
	 * block layer from calling the midlayer with this device's
	 * request queue. 
	 */
	spin_lock_irqsave(q->queue_lock, flags);
	blk_stop_queue(q);
	spin_unlock_irqrestore(q->queue_lock, flags);

	return 0;
}
EXPORT_SYMBOL_GPL(scsi_internal_device_block);
 
/**
 * scsi_internal_device_unblock - resume a device after a block request
 * @sdev:	device to resume
 *
 * Called by scsi lld's or the midlayer to restart the device queue
 * for the previously suspended scsi device.  Called from interrupt or
 * normal process context.
 *
 * Returns zero if successful or error if not.
 *
 * Notes:       
 *	This routine transitions the device to the SDEV_RUNNING state
 *	(which must be a legal transition) allowing the midlayer to
 *	goose the queue for this device.  This routine assumes the 
 *	host_lock is held upon entry.
2422
 */
L
Linus Torvalds 已提交
2423 2424 2425
int
scsi_internal_device_unblock(struct scsi_device *sdev)
{
2426
	struct request_queue *q = sdev->request_queue; 
L
Linus Torvalds 已提交
2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496
	int err;
	unsigned long flags;
	
	/* 
	 * Try to transition the scsi device to SDEV_RUNNING
	 * and goose the device queue if successful.  
	 */
	err = scsi_device_set_state(sdev, SDEV_RUNNING);
	if (err)
		return err;

	spin_lock_irqsave(q->queue_lock, flags);
	blk_start_queue(q);
	spin_unlock_irqrestore(q->queue_lock, flags);

	return 0;
}
EXPORT_SYMBOL_GPL(scsi_internal_device_unblock);

static void
device_block(struct scsi_device *sdev, void *data)
{
	scsi_internal_device_block(sdev);
}

static int
target_block(struct device *dev, void *data)
{
	if (scsi_is_target_device(dev))
		starget_for_each_device(to_scsi_target(dev), NULL,
					device_block);
	return 0;
}

void
scsi_target_block(struct device *dev)
{
	if (scsi_is_target_device(dev))
		starget_for_each_device(to_scsi_target(dev), NULL,
					device_block);
	else
		device_for_each_child(dev, NULL, target_block);
}
EXPORT_SYMBOL_GPL(scsi_target_block);

static void
device_unblock(struct scsi_device *sdev, void *data)
{
	scsi_internal_device_unblock(sdev);
}

static int
target_unblock(struct device *dev, void *data)
{
	if (scsi_is_target_device(dev))
		starget_for_each_device(to_scsi_target(dev), NULL,
					device_unblock);
	return 0;
}

void
scsi_target_unblock(struct device *dev)
{
	if (scsi_is_target_device(dev))
		starget_for_each_device(to_scsi_target(dev), NULL,
					device_unblock);
	else
		device_for_each_child(dev, NULL, target_unblock);
}
EXPORT_SYMBOL_GPL(scsi_target_unblock);
2497 2498 2499

/**
 * scsi_kmap_atomic_sg - find and atomically map an sg-elemnt
2500
 * @sgl:	scatter-gather list
2501 2502 2503 2504 2505 2506
 * @sg_count:	number of segments in sg
 * @offset:	offset in bytes into sg, on return offset into the mapped area
 * @len:	bytes to map, on return number of bytes mapped
 *
 * Returns virtual address of the start of the mapped page
 */
J
Jens Axboe 已提交
2507
void *scsi_kmap_atomic_sg(struct scatterlist *sgl, int sg_count,
2508 2509 2510 2511
			  size_t *offset, size_t *len)
{
	int i;
	size_t sg_len = 0, len_complete = 0;
J
Jens Axboe 已提交
2512
	struct scatterlist *sg;
2513 2514
	struct page *page;

2515 2516
	WARN_ON(!irqs_disabled());

J
Jens Axboe 已提交
2517
	for_each_sg(sgl, sg, sg_count, i) {
2518
		len_complete = sg_len; /* Complete sg-entries */
J
Jens Axboe 已提交
2519
		sg_len += sg->length;
2520 2521 2522 2523 2524
		if (sg_len > *offset)
			break;
	}

	if (unlikely(i == sg_count)) {
2525 2526
		printk(KERN_ERR "%s: Bytes in sg: %zu, requested offset %zu, "
			"elements %d\n",
2527 2528 2529 2530 2531 2532
		       __FUNCTION__, sg_len, *offset, sg_count);
		WARN_ON(1);
		return NULL;
	}

	/* Offset starting from the beginning of first page in this sg-entry */
J
Jens Axboe 已提交
2533
	*offset = *offset - len_complete + sg->offset;
2534 2535

	/* Assumption: contiguous pages can be accessed as "page + i" */
J
Jens Axboe 已提交
2536
	page = nth_page(sg_page(sg), (*offset >> PAGE_SHIFT));
2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548
	*offset &= ~PAGE_MASK;

	/* Bytes in this sg-entry from *offset to the end of the page */
	sg_len = PAGE_SIZE - *offset;
	if (*len > sg_len)
		*len = sg_len;

	return kmap_atomic(page, KM_BIO_SRC_IRQ);
}
EXPORT_SYMBOL(scsi_kmap_atomic_sg);

/**
2549
 * scsi_kunmap_atomic_sg - atomically unmap a virtual address, previously mapped with scsi_kmap_atomic_sg
2550 2551 2552 2553 2554 2555 2556
 * @virt:	virtual address to be unmapped
 */
void scsi_kunmap_atomic_sg(void *virt)
{
	kunmap_atomic(virt, KM_BIO_SRC_IRQ);
}
EXPORT_SYMBOL(scsi_kunmap_atomic_sg);