scsi_lib.c 64.9 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
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

94 95 96 97 98
/**
 * __scsi_queue_insert - private queue insertion
 * @cmd: The SCSI command being requeued
 * @reason:  The reason for the requeue
 * @unbusy: Whether the queue should be unbusied
Linus Torvalds's avatar
Linus Torvalds committed
99
 *
100 101 102 103 104
 * This is a private queue insertion.  The public interface
 * scsi_queue_insert() always assumes the queue should be unbusied
 * because it's always called before the completion.  This function is
 * for a requeue after completion, which should only occur in this
 * file.
Linus Torvalds's avatar
Linus Torvalds committed
105
 */
106
static int __scsi_queue_insert(struct scsi_cmnd *cmd, int reason, int unbusy)
Linus Torvalds's avatar
Linus Torvalds committed
107 108 109
{
	struct Scsi_Host *host = cmd->device->host;
	struct scsi_device *device = cmd->device;
110
	struct scsi_target *starget = scsi_target(device);
111 112
	struct request_queue *q = device->request_queue;
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
113 114 115 116 117

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

	/*
118
	 * Set the appropriate busy bit for the device/host.
Linus Torvalds's avatar
Linus Torvalds committed
119 120 121 122 123 124 125 126 127 128 129
	 *
	 * 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.
	 */
130 131
	switch (reason) {
	case SCSI_MLQUEUE_HOST_BUSY:
Linus Torvalds's avatar
Linus Torvalds committed
132
		host->host_blocked = host->max_host_blocked;
133 134
		break;
	case SCSI_MLQUEUE_DEVICE_BUSY:
Linus Torvalds's avatar
Linus Torvalds committed
135
		device->device_blocked = device->max_device_blocked;
136 137 138 139 140
		break;
	case SCSI_MLQUEUE_TARGET_BUSY:
		starget->target_blocked = starget->max_target_blocked;
		break;
	}
Linus Torvalds's avatar
Linus Torvalds committed
141 142 143 144 145

	/*
	 * Decrement the counters, since these commands are no longer
	 * active on the host/device.
	 */
146 147
	if (unbusy)
		scsi_device_unbusy(device);
Linus Torvalds's avatar
Linus Torvalds committed
148 149

	/*
150 151
	 * Requeue this command.  It will go before all other commands
	 * that are already in the queue.
Linus Torvalds's avatar
Linus Torvalds committed
152 153 154 155
	 *
	 * NOTE: there is magic here about the way the queue is plugged if
	 * we have no outstanding commands.
	 * 
156
	 * Although we *don't* plug the queue, we call the request
Linus Torvalds's avatar
Linus Torvalds committed
157 158
	 * function.  The SCSI request function detects the blocked condition
	 * and plugs the queue appropriately.
159 160
         */
	spin_lock_irqsave(q->queue_lock, flags);
161
	blk_requeue_request(q, cmd->request);
162 163 164 165
	spin_unlock_irqrestore(q->queue_lock, flags);

	scsi_run_queue(q);

Linus Torvalds's avatar
Linus Torvalds committed
166 167 168
	return 0;
}

169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
/*
 * 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)
{
	return __scsi_queue_insert(cmd, reason, 1);
}
192
/**
193
 * scsi_execute - insert request and wait for the result
194 195 196 197 198 199 200 201
 * @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
202
 * @flags:	or into request flags;
203
 * @resid:	optional residual length
204
 *
205
 * returns the req->errors value which is the scsi_cmnd result
206
 * field.
207
 */
208 209
int scsi_execute(struct scsi_device *sdev, const unsigned char *cmd,
		 int data_direction, void *buffer, unsigned bufflen,
210 211
		 unsigned char *sense, int timeout, int retries, int flags,
		 int *resid)
212 213 214 215 216 217 218 219 220 221 222 223 224 225 226
{
	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;
227
	req->retries = retries;
228
	req->timeout = timeout;
229 230
	req->cmd_type = REQ_TYPE_BLOCK_PC;
	req->cmd_flags |= flags | REQ_QUIET | REQ_PREEMPT;
231 232 233 234 235 236

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

237 238 239 240 241 242 243 244 245
	/*
	 * 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);

246 247
	if (resid)
		*resid = req->data_len;
248 249 250 251 252 253
	ret = req->errors;
 out:
	blk_put_request(req);

	return ret;
}
254
EXPORT_SYMBOL(scsi_execute);
255

256 257 258

int scsi_execute_req(struct scsi_device *sdev, const unsigned char *cmd,
		     int data_direction, void *buffer, unsigned bufflen,
259 260
		     struct scsi_sense_hdr *sshdr, int timeout, int retries,
		     int *resid)
261 262
{
	char *sense = NULL;
263 264
	int result;
	
265
	if (sshdr) {
266
		sense = kzalloc(SCSI_SENSE_BUFFERSIZE, GFP_NOIO);
267 268 269
		if (!sense)
			return DRIVER_ERROR << 24;
	}
270
	result = scsi_execute(sdev, cmd, data_direction, buffer, bufflen,
271
			      sense, timeout, retries, 0, resid);
272
	if (sshdr)
273
		scsi_normalize_sense(sense, SCSI_SENSE_BUFFERSIZE, sshdr);
274 275 276 277 278 279

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

Linus Torvalds's avatar
Linus Torvalds committed
280 281 282 283 284 285 286 287 288 289 290
/*
 * 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.
 */
291
static void scsi_init_cmd_errh(struct scsi_cmnd *cmd)
Linus Torvalds's avatar
Linus Torvalds committed
292 293
{
	cmd->serial_number = 0;
294
	scsi_set_resid(cmd, 0);
295
	memset(cmd->sense_buffer, 0, SCSI_SENSE_BUFFERSIZE);
Linus Torvalds's avatar
Linus Torvalds committed
296
	if (cmd->cmd_len == 0)
297
		cmd->cmd_len = scsi_command_size(cmd->cmnd);
Linus Torvalds's avatar
Linus Torvalds committed
298 299 300 301 302
}

void scsi_device_unbusy(struct scsi_device *sdev)
{
	struct Scsi_Host *shost = sdev->host;
303
	struct scsi_target *starget = scsi_target(sdev);
Linus Torvalds's avatar
Linus Torvalds committed
304 305 306 307
	unsigned long flags;

	spin_lock_irqsave(shost->host_lock, flags);
	shost->host_busy--;
308
	starget->target_busy--;
309
	if (unlikely(scsi_host_in_recovery(shost) &&
310
		     (shost->host_failed || shost->host_eh_scheduled)))
Linus Torvalds's avatar
Linus Torvalds committed
311 312
		scsi_eh_wakeup(shost);
	spin_unlock(shost->host_lock);
's avatar
committed
313
	spin_lock(sdev->request_queue->queue_lock);
Linus Torvalds's avatar
Linus Torvalds committed
314
	sdev->device_busy--;
's avatar
committed
315
	spin_unlock_irqrestore(sdev->request_queue->queue_lock, flags);
Linus Torvalds's avatar
Linus Torvalds committed
316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363
}

/*
 * 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);
}

364 365 366 367 368 369 370 371
static inline int scsi_device_is_busy(struct scsi_device *sdev)
{
	if (sdev->device_busy >= sdev->queue_depth || sdev->device_blocked)
		return 1;

	return 0;
}

372 373 374 375 376 377 378
static inline int scsi_target_is_busy(struct scsi_target *starget)
{
	return ((starget->can_queue > 0 &&
		 starget->target_busy >= starget->can_queue) ||
		 starget->target_blocked);
}

379 380 381 382 383 384 385 386 387
static inline int scsi_host_is_busy(struct Scsi_Host *shost)
{
	if ((shost->can_queue > 0 && shost->host_busy >= shost->can_queue) ||
	    shost->host_blocked || shost->host_self_blocked)
		return 1;

	return 0;
}

Linus Torvalds's avatar
Linus Torvalds committed
388 389 390 391 392 393 394 395 396 397 398 399 400 401
/*
 * 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)
{
402
	struct scsi_device *sdev = q->queuedata;
Linus Torvalds's avatar
Linus Torvalds committed
403
	struct Scsi_Host *shost = sdev->host;
404
	LIST_HEAD(starved_list);
Linus Torvalds's avatar
Linus Torvalds committed
405 406
	unsigned long flags;

407
	if (scsi_target(sdev)->single_lun)
Linus Torvalds's avatar
Linus Torvalds committed
408 409 410
		scsi_single_lun_run(sdev);

	spin_lock_irqsave(shost->host_lock, flags);
411 412 413
	list_splice_init(&shost->starved_list, &starved_list);

	while (!list_empty(&starved_list)) {
414 415
		int flagset;

Linus Torvalds's avatar
Linus Torvalds committed
416 417 418 419 420 421 422 423 424 425
		/*
		 * 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.
		 */
426
		if (scsi_host_is_busy(shost))
427 428
			break;

429 430 431
		sdev = list_entry(starved_list.next,
				  struct scsi_device, starved_entry);
		list_del_init(&sdev->starved_entry);
432 433 434 435 436 437
		if (scsi_target_is_busy(scsi_target(sdev))) {
			list_move_tail(&sdev->starved_entry,
				       &shost->starved_list);
			continue;
		}

438 439 440 441 442 443 444 445 446 447 448 449
		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);
450

451
		spin_lock(shost->host_lock);
Linus Torvalds's avatar
Linus Torvalds committed
452
	}
453 454
	/* put any unprocessed entries back */
	list_splice(&starved_list, &shost->starved_list);
Linus Torvalds's avatar
Linus Torvalds committed
455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
	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.
476
 * Notes:	Upon return, cmd is a stale pointer.
Linus Torvalds's avatar
Linus Torvalds committed
477 478 479
 */
static void scsi_requeue_command(struct request_queue *q, struct scsi_cmnd *cmd)
{
480
	struct request *req = cmd->request;
481 482 483
	unsigned long flags;

	spin_lock_irqsave(q->queue_lock, flags);
484
	scsi_unprep_request(req);
485
	blk_requeue_request(q, req);
486
	spin_unlock_irqrestore(q->queue_lock, flags);
Linus Torvalds's avatar
Linus Torvalds committed
487 488 489 490 491 492

	scsi_run_queue(q);
}

void scsi_next_command(struct scsi_cmnd *cmd)
{
493 494 495 496 497
	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
498 499 500

	scsi_put_command(cmd);
	scsi_run_queue(q);
501 502 503

	/* ok to remove device now */
	put_device(&sdev->sdev_gendev);
Linus Torvalds's avatar
Linus Torvalds committed
504 505 506 507 508 509 510 511 512 513
}

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);
}

514 515
static void __scsi_release_buffers(struct scsi_cmnd *, int);

Linus Torvalds's avatar
Linus Torvalds committed
516 517 518 519 520 521 522
/*
 * 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.
523
 *              error    - 0 if I/O indicates success, < 0 for I/O error.
Linus Torvalds's avatar
Linus Torvalds committed
524 525 526 527 528
 *              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.
 *
529
 * Returns:     cmd if requeue required, NULL otherwise.
Linus Torvalds's avatar
Linus Torvalds committed
530 531 532 533 534 535
 *
 * 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.
536
 * Notes:	If cmd was requeued, upon return it will be a stale pointer.
Linus Torvalds's avatar
Linus Torvalds committed
537
 */
538
static struct scsi_cmnd *scsi_end_request(struct scsi_cmnd *cmd, int error,
Linus Torvalds's avatar
Linus Torvalds committed
539 540
					  int bytes, int requeue)
{
541
	struct request_queue *q = cmd->device->request_queue;
Linus Torvalds's avatar
Linus Torvalds committed
542 543 544 545 546 547
	struct request *req = cmd->request;

	/*
	 * If there are blocks left over at the end, set up the command
	 * to queue the remainder of them.
	 */
548
	if (blk_end_request(req, error, bytes)) {
Linus Torvalds's avatar
Linus Torvalds committed
549 550 551 552 553 554
		int leftover = (req->hard_nr_sectors << 9);

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

		/* kill remainder if no retrys */
555
		if (error && scsi_noretry_cmd(cmd))
556
			blk_end_request(req, error, leftover);
Linus Torvalds's avatar
Linus Torvalds committed
557
		else {
558
			if (requeue) {
Linus Torvalds's avatar
Linus Torvalds committed
559 560 561 562 563
				/*
				 * Bleah.  Leftovers again.  Stick the
				 * leftovers in the front of the
				 * queue, and goose the queue again.
				 */
564
				scsi_release_buffers(cmd);
Linus Torvalds's avatar
Linus Torvalds committed
565
				scsi_requeue_command(q, cmd);
566 567
				cmd = NULL;
			}
Linus Torvalds's avatar
Linus Torvalds committed
568 569 570 571 572 573 574 575
			return cmd;
		}
	}

	/*
	 * This will goose the queue request function at the end, so we don't
	 * need to worry about launching another command.
	 */
576
	__scsi_release_buffers(cmd, 0);
Linus Torvalds's avatar
Linus Torvalds committed
577 578 579 580
	scsi_next_command(cmd);
	return NULL;
}

581 582 583 584
static inline unsigned int scsi_sgtable_index(unsigned short nents)
{
	unsigned int index;

585 586 587
	BUG_ON(nents > SCSI_MAX_SG_SEGMENTS);

	if (nents <= 8)
588
		index = 0;
589 590
	else
		index = get_count_order(nents) - 3;
Linus Torvalds's avatar
Linus Torvalds committed
591

592 593 594
	return index;
}

595
static void scsi_sg_free(struct scatterlist *sgl, unsigned int nents)
596 597 598
{
	struct scsi_host_sg_pool *sgp;

599 600 601
	sgp = scsi_sg_pools + scsi_sgtable_index(nents);
	mempool_free(sgl, sgp->pool);
}
602

603 604 605
static struct scatterlist *scsi_sg_alloc(unsigned int nents, gfp_t gfp_mask)
{
	struct scsi_host_sg_pool *sgp;
606

607 608 609
	sgp = scsi_sg_pools + scsi_sgtable_index(nents);
	return mempool_alloc(sgp->pool, gfp_mask);
}
610

611 612
static int scsi_alloc_sgtable(struct scsi_data_buffer *sdb, int nents,
			      gfp_t gfp_mask)
613 614
{
	int ret;
615

616
	BUG_ON(!nents);
617

618 619
	ret = __sg_alloc_table(&sdb->table, nents, SCSI_MAX_SG_SEGMENTS,
			       gfp_mask, scsi_sg_alloc);
620
	if (unlikely(ret))
621
		__sg_free_table(&sdb->table, SCSI_MAX_SG_SEGMENTS,
622
				scsi_sg_free);
623

624
	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
625 626
}

627
static void scsi_free_sgtable(struct scsi_data_buffer *sdb)
Linus Torvalds's avatar
Linus Torvalds committed
628
{
629
	__sg_free_table(&sdb->table, SCSI_MAX_SG_SEGMENTS, scsi_sg_free);
Linus Torvalds's avatar
Linus Torvalds committed
630 631
}

632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651
static void __scsi_release_buffers(struct scsi_cmnd *cmd, int do_bidi_check)
{

	if (cmd->sdb.table.nents)
		scsi_free_sgtable(&cmd->sdb);

	memset(&cmd->sdb, 0, sizeof(cmd->sdb));

	if (do_bidi_check && scsi_bidi_cmnd(cmd)) {
		struct scsi_data_buffer *bidi_sdb =
			cmd->request->next_rq->special;
		scsi_free_sgtable(bidi_sdb);
		kmem_cache_free(scsi_sdb_cache, bidi_sdb);
		cmd->request->next_rq->special = NULL;
	}

	if (scsi_prot_sg_count(cmd))
		scsi_free_sgtable(cmd->prot_sdb);
}

Linus Torvalds's avatar
Linus Torvalds committed
652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668
/*
 * 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.
 */
669
void scsi_release_buffers(struct scsi_cmnd *cmd)
Linus Torvalds's avatar
Linus Torvalds committed
670
{
671
	__scsi_release_buffers(cmd, 1);
Linus Torvalds's avatar
Linus Torvalds committed
672
}
673
EXPORT_SYMBOL(scsi_release_buffers);
Linus Torvalds's avatar
Linus Torvalds committed
674

675 676 677 678 679 680 681
/*
 * 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.
 */
682
static void scsi_end_bidi_request(struct scsi_cmnd *cmd)
683
{
684 685 686 687 688 689 690 691 692 693
	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));

694 695 696 697 698 699 700 701 702
	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
703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719
/*
 * 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.
 *
720 721 722 723 724
 *		We must 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 we have to
 *		figure out what to do next:
Linus Torvalds's avatar
Linus Torvalds committed
725
 *
726 727 728 729 730
 *		a) We can call scsi_requeue_command().  The request
 *		   will be unprepared and put back on the queue.  Then
 *		   a new command will be created for it.  This should
 *		   be used if we made forward progress, or if we want
 *		   to switch from READ(10) to READ(6) for example.
Linus Torvalds's avatar
Linus Torvalds committed
731
 *
732 733 734 735 736 737
 *		b) We can call scsi_queue_insert().  The request will
 *		   be put back on the queue and retried using the same
 *		   command as before, possibly after a delay.
 *
 *		c) We can call blk_end_request() with -EIO to fail
 *		   the remainder of the request.
Linus Torvalds's avatar
Linus Torvalds committed
738
 */
739
void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes)
Linus Torvalds's avatar
Linus Torvalds committed
740 741
{
	int result = cmd->result;
742
	int this_count;
743
	struct request_queue *q = cmd->device->request_queue;
Linus Torvalds's avatar
Linus Torvalds committed
744
	struct request *req = cmd->request;
745
	int error = 0;
Linus Torvalds's avatar
Linus Torvalds committed
746 747 748
	struct scsi_sense_hdr sshdr;
	int sense_valid = 0;
	int sense_deferred = 0;
749 750 751
	enum {ACTION_FAIL, ACTION_REPREP, ACTION_RETRY,
	      ACTION_DELAYED_RETRY} action;
	char *description = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
752 753 754 755 756 757

	if (result) {
		sense_valid = scsi_command_normalize_sense(cmd, &sshdr);
		if (sense_valid)
			sense_deferred = scsi_sense_is_deferred(&sshdr);
	}
758

Linus Torvalds's avatar
Linus Torvalds committed
759 760 761 762 763 764 765 766 767 768 769 770 771 772
	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;
			}
773 774
			if (!sense_deferred)
				error = -EIO;
775
		}
776 777 778 779 780
		if (scsi_bidi_cmnd(cmd)) {
			/* will also release_buffers */
			scsi_end_bidi_request(cmd);
			return;
		}
781
		req->data_len = scsi_get_resid(cmd);
Linus Torvalds's avatar
Linus Torvalds committed
782 783
	}

784
	BUG_ON(blk_bidi_rq(req)); /* bidi not support for !blk_pc_request yet */
785

Linus Torvalds's avatar
Linus Torvalds committed
786 787 788 789
	/*
	 * Next deal with any sectors which we were able to correctly
	 * handle.
	 */
790 791 792 793
	SCSI_LOG_HLCOMPLETE(1, printk("%ld sectors total, "
				      "%d bytes done.\n",
				      req->nr_sectors, good_bytes));

794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809
	/*
	 * Recovered errors need reporting, but they're always treated
	 * as success, so fiddle the result code here.  For BLOCK_PC
	 * we already took a copy of the original into rq->errors which
	 * is what gets returned to the user
	 */
	if (sense_valid && sshdr.sense_key == RECOVERED_ERROR) {
		if (!(req->cmd_flags & REQ_QUIET))
			scsi_print_sense("", cmd);
		result = 0;
		/* BLOCK_PC may have set error */
		error = 0;
	}

	/*
	 * A number of bytes were successfully read.  If there
810 811 812
	 * are leftovers and there is some kind of error
	 * (result != 0), retry the rest.
	 */
813
	if (scsi_end_request(cmd, error, good_bytes, result == 0) == NULL)
814
		return;
815
	this_count = blk_rq_bytes(req);
816

817 818
	error = -EIO;

819 820 821 822 823 824 825
	if (host_byte(result) == DID_RESET) {
		/* Third party bus reset or reset for error recovery
		 * reasons.  Just retry the command and see what
		 * happens.
		 */
		action = ACTION_RETRY;
	} else if (sense_valid && !sense_deferred) {
Linus Torvalds's avatar
Linus Torvalds committed
826 827 828
		switch (sshdr.sense_key) {
		case UNIT_ATTENTION:
			if (cmd->device->removable) {
829
				/* Detected disc change.  Set a bit
Linus Torvalds's avatar
Linus Torvalds committed
830 831 832
				 * and quietly refuse further access.
				 */
				cmd->device->changed = 1;
833 834
				description = "Media Changed";
				action = ACTION_FAIL;
Linus Torvalds's avatar
Linus Torvalds committed
835
			} else {
836 837 838
				/* Must have been a power glitch, or a
				 * bus reset.  Could not have been a
				 * media change, so we just retry the
839
				 * command and see what happens.
840
				 */
841
				action = ACTION_RETRY;
Linus Torvalds's avatar
Linus Torvalds committed
842 843 844
			}
			break;
		case ILLEGAL_REQUEST:
845 846 847 848 849 850 851 852
			/* 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.
			 */
853 854
			if ((cmd->device->use_10_for_rw &&
			    sshdr.asc == 0x20 && sshdr.ascq == 0x00) &&
Linus Torvalds's avatar
Linus Torvalds committed
855 856
			    (cmd->cmnd[0] == READ_10 ||
			     cmd->cmnd[0] == WRITE_10)) {
857
				/* This will issue a new 6-byte command. */
Linus Torvalds's avatar
Linus Torvalds committed
858
				cmd->device->use_10_for_rw = 0;
859
				action = ACTION_REPREP;
860 861 862 863
			} else if (sshdr.asc == 0x10) /* DIX */ {
				description = "Host Data Integrity Failure";
				action = ACTION_FAIL;
				error = -EILSEQ;
864 865 866
			} else
				action = ACTION_FAIL;
			break;
867
		case ABORTED_COMMAND:
868
			action = ACTION_FAIL;
869
			if (sshdr.asc == 0x10) { /* DIF */
870 871
				description = "Target Data Integrity Failure";
				error = -EILSEQ;
872
			}
Linus Torvalds's avatar
Linus Torvalds committed
873 874
			break;
		case NOT_READY:
875
			/* If the device is in the process of becoming
876
			 * ready, or has a temporary blockage, retry.
Linus Torvalds's avatar
Linus Torvalds committed
877
			 */
878 879 880 881 882 883 884 885 886
			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 */
887
					action = ACTION_DELAYED_RETRY;
888
					break;
889 890 891 892
				default:
					description = "Device not ready";
					action = ACTION_FAIL;
					break;
893
				}
894 895 896
			} else {
				description = "Device not ready";
				action = ACTION_FAIL;
Linus Torvalds's avatar
Linus Torvalds committed
897
			}
898
			break;
Linus Torvalds's avatar
Linus Torvalds committed
899
		case VOLUME_OVERFLOW:
900
			/* See SSC3rXX or current. */
901 902
			action = ACTION_FAIL;
			break;
Linus Torvalds's avatar
Linus Torvalds committed
903
		default:
904 905
			description = "Unhandled sense code";
			action = ACTION_FAIL;
Linus Torvalds's avatar
Linus Torvalds committed
906 907
			break;
		}
908 909 910
	} else {
		description = "Unhandled error code";
		action = ACTION_FAIL;
911
	}
912 913 914 915

	switch (action) {
	case ACTION_FAIL:
		/* Give up and fail the remainder of the request */
916
		scsi_release_buffers(cmd);
917
		if (!(req->cmd_flags & REQ_QUIET)) {
918
			if (description)
919
				scmd_printk(KERN_INFO, cmd, "%s\n",
920
					    description);
921
			scsi_print_result(cmd);
922 923 924
			if (driver_byte(result) & DRIVER_SENSE)
				scsi_print_sense("", cmd);
		}
925
		blk_end_request_all(req, -EIO);
926 927 928 929 930 931
		scsi_next_command(cmd);
		break;
	case ACTION_REPREP:
		/* Unprep the request and put it back at the head of the queue.
		 * A new command will be prepared and issued.
		 */
932
		scsi_release_buffers(cmd);
933 934 935 936
		scsi_requeue_command(q, cmd);
		break;
	case ACTION_RETRY:
		/* Retry the same command immediately */
937
		__scsi_queue_insert(cmd, SCSI_MLQUEUE_EH_RETRY, 0);
938 939 940
		break;
	case ACTION_DELAYED_RETRY:
		/* Retry the same command after a delay */
941
		__scsi_queue_insert(cmd, SCSI_MLQUEUE_DEVICE_BUSY, 0);
942
		break;
Linus Torvalds's avatar
Linus Torvalds committed
943 944 945
	}
}

946 947
static int scsi_init_sgtable(struct request *req, struct scsi_data_buffer *sdb,
			     gfp_t gfp_mask)
Linus Torvalds's avatar
Linus Torvalds committed
948
{
949
	int count;
Linus Torvalds's avatar
Linus Torvalds committed
950 951

	/*
952
	 * If sg table allocation fails, requeue request later.
Linus Torvalds's avatar
Linus Torvalds committed
953
	 */
954 955
	if (unlikely(scsi_alloc_sgtable(sdb, req->nr_phys_segments,
					gfp_mask))) {
Linus Torvalds's avatar
Linus Torvalds committed
956
		return BLKPREP_DEFER;
Alan Stern's avatar
Alan Stern committed
957
	}
Linus Torvalds's avatar
Linus Torvalds committed
958

959
	req->buffer = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
960 961 962 963 964

	/* 
	 * Next, walk the list, and fill in the addresses and sizes of
	 * each segment.
	 */
965 966 967
	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
968 969 970 971
	if (blk_pc_request(req))
		sdb->length = req->data_len;
	else
		sdb->length = req->nr_sectors << 9;
972
	return BLKPREP_OK;
Linus Torvalds's avatar
Linus Torvalds committed
973
}
974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993

/*
 * 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(
994
			scsi_sdb_cache, GFP_ATOMIC);
995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006
		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;
	}

1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026
	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;
	}

1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037
	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;
}
1038
EXPORT_SYMBOL(scsi_init_io);
Linus Torvalds's avatar
Linus Torvalds committed
1039

1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057
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;

1058 1059
	cmd->cmnd = req->cmd;

1060 1061 1062
	return cmd;
}

1063
int scsi_setup_blk_pc_cmnd(struct scsi_device *sdev, struct request *req)
James Bottomley's avatar
James Bottomley committed
1064
{
1065
	struct scsi_cmnd *cmd;
1066 1067 1068 1069
	int ret = scsi_prep_state_check(sdev, req);

	if (ret != BLKPREP_OK)
		return ret;
1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085

	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);

1086
		ret = scsi_init_io(cmd, GFP_ATOMIC);
1087 1088 1089 1090 1091
		if (unlikely(ret))
			return ret;
	} else {
		BUG_ON(req->data_len);

1092
		memset(&cmd->sdb, 0, sizeof(cmd->sdb));
1093 1094
		req->buffer = NULL;
	}
James Bottomley's avatar
James Bottomley committed
1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105

	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;
1106
	return BLKPREP_OK;
James Bottomley's avatar
James Bottomley committed
1107
}
1108
EXPORT_SYMBOL(scsi_setup_blk_pc_cmnd);
James Bottomley's avatar
James Bottomley committed
1109

1110 1111 1112 1113 1114
/*
 * 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.
 */
1115
int scsi_setup_fs_cmnd(struct scsi_device *sdev, struct request *req)
Linus Torvalds's avatar