libata-eh.c 108 KB
Newer Older
1 2 3
/*
 *  libata-eh.c - libata error handling
 *
4
 *  Maintained by:  Tejun Heo <tj@kernel.org>
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
 *    		    Please ALWAYS copy linux-ide@vger.kernel.org
 *		    on emails.
 *
 *  Copyright 2006 Tejun Heo <htejun@gmail.com>
 *
 *
 *  This program is free software; you can redistribute it and/or
 *  modify it under the terms of the GNU General Public License as
 *  published by the Free Software Foundation; either version 2, or
 *  (at your option) any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 *  General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; see the file COPYING.  If not, write to
 *  the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139,
 *  USA.
 *
 *
 *  libata documentation is available via 'make {ps|pdf}docs',
 *  as Documentation/DocBook/libata.*
 *
 *  Hardware documentation available from http://www.t13.org/ and
 *  http://www.sata-io.org/
 *
 */

#include <linux/kernel.h>
36
#include <linux/blkdev.h>
37
#include <linux/export.h>
38
#include <linux/pci.h>
39 40 41 42 43
#include <scsi/scsi.h>
#include <scsi/scsi_host.h>
#include <scsi/scsi_eh.h>
#include <scsi/scsi_device.h>
#include <scsi/scsi_cmnd.h>
44
#include <scsi/scsi_dbg.h>
45
#include "../scsi/scsi_transport_api.h"
46 47 48 49 50

#include <linux/libata.h>

#include "libata.h"

51
enum {
52
	/* speed down verdicts */
53 54 55
	ATA_EH_SPDN_NCQ_OFF		= (1 << 0),
	ATA_EH_SPDN_SPEED_DOWN		= (1 << 1),
	ATA_EH_SPDN_FALLBACK_TO_PIO	= (1 << 2),
56
	ATA_EH_SPDN_KEEP_ERRORS		= (1 << 3),
57 58 59

	/* error flags */
	ATA_EFLAG_IS_IO			= (1 << 0),
60
	ATA_EFLAG_DUBIOUS_XFER		= (1 << 1),
61
	ATA_EFLAG_OLD_ER                = (1 << 31),
62 63 64 65 66 67

	/* error categories */
	ATA_ECAT_NONE			= 0,
	ATA_ECAT_ATA_BUS		= 1,
	ATA_ECAT_TOUT_HSM		= 2,
	ATA_ECAT_UNK_DEV		= 3,
68 69 70 71 72
	ATA_ECAT_DUBIOUS_NONE		= 4,
	ATA_ECAT_DUBIOUS_ATA_BUS	= 5,
	ATA_ECAT_DUBIOUS_TOUT_HSM	= 6,
	ATA_ECAT_DUBIOUS_UNK_DEV	= 7,
	ATA_ECAT_NR			= 8,
73

74 75
	ATA_EH_CMD_DFL_TIMEOUT		=  5000,

76 77 78
	/* always put at least this amount of time between resets */
	ATA_EH_RESET_COOL_DOWN		=  5000,

79 80 81 82 83 84 85
	/* Waiting in ->prereset can never be reliable.  It's
	 * sometimes nice to wait there but it can't be depended upon;
	 * otherwise, we wouldn't be resetting.  Just give it enough
	 * time for most drives to spin up.
	 */
	ATA_EH_PRERESET_TIMEOUT		= 10000,
	ATA_EH_FASTDRAIN_INTERVAL	=  3000,
86 87

	ATA_EH_UA_TRIES			= 5,
88 89 90 91

	/* probe speed down parameters, see ata_eh_schedule_probe() */
	ATA_EH_PROBE_TRIAL_INTERVAL	= 60000,	/* 1 min */
	ATA_EH_PROBE_TRIALS		= 2,
92 93 94 95 96 97 98 99 100
};

/* The following table determines how we sequence resets.  Each entry
 * represents timeout for that try.  The first try can be soft or
 * hardreset.  All others are hardreset if available.  In most cases
 * the first reset w/ 10sec timeout should succeed.  Following entries
 * are mostly for error handling, hotplug and retarded devices.
 */
static const unsigned long ata_eh_reset_timeouts[] = {
101 102 103 104
	10000,	/* most drives spin up by 10sec */
	10000,	/* > 99% working drives spin up before 20sec */
	35000,	/* give > 30 secs of idleness for retarded devices */
	 5000,	/* and sweet one last chance */
105
	ULONG_MAX, /* > 1 min has elapsed, give up */
106 107
};

108 109 110 111 112 113 114
static const unsigned long ata_eh_identify_timeouts[] = {
	 5000,	/* covers > 99% of successes and not too boring on failures */
	10000,  /* combined time till here is enough even for media access */
	30000,	/* for true idiots */
	ULONG_MAX,
};

115 116 117 118 119 120 121
static const unsigned long ata_eh_flush_timeouts[] = {
	15000,	/* be generous with flush */
	15000,  /* ditto */
	30000,	/* and even more generous */
	ULONG_MAX,
};

122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158
static const unsigned long ata_eh_other_timeouts[] = {
	 5000,	/* same rationale as identify timeout */
	10000,	/* ditto */
	/* but no merciful 30sec for other commands, it just isn't worth it */
	ULONG_MAX,
};

struct ata_eh_cmd_timeout_ent {
	const u8		*commands;
	const unsigned long	*timeouts;
};

/* The following table determines timeouts to use for EH internal
 * commands.  Each table entry is a command class and matches the
 * commands the entry applies to and the timeout table to use.
 *
 * On the retry after a command timed out, the next timeout value from
 * the table is used.  If the table doesn't contain further entries,
 * the last value is used.
 *
 * ehc->cmd_timeout_idx keeps track of which timeout to use per
 * command class, so if SET_FEATURES times out on the first try, the
 * next try will use the second timeout value only for that class.
 */
#define CMDS(cmds...)	(const u8 []){ cmds, 0 }
static const struct ata_eh_cmd_timeout_ent
ata_eh_cmd_timeout_table[ATA_EH_CMD_TIMEOUT_TABLE_SIZE] = {
	{ .commands = CMDS(ATA_CMD_ID_ATA, ATA_CMD_ID_ATAPI),
	  .timeouts = ata_eh_identify_timeouts, },
	{ .commands = CMDS(ATA_CMD_READ_NATIVE_MAX, ATA_CMD_READ_NATIVE_MAX_EXT),
	  .timeouts = ata_eh_other_timeouts, },
	{ .commands = CMDS(ATA_CMD_SET_MAX, ATA_CMD_SET_MAX_EXT),
	  .timeouts = ata_eh_other_timeouts, },
	{ .commands = CMDS(ATA_CMD_SET_FEATURES),
	  .timeouts = ata_eh_other_timeouts, },
	{ .commands = CMDS(ATA_CMD_INIT_DEV_PARAMS),
	  .timeouts = ata_eh_other_timeouts, },
159 160
	{ .commands = CMDS(ATA_CMD_FLUSH, ATA_CMD_FLUSH_EXT),
	  .timeouts = ata_eh_flush_timeouts },
161 162 163
};
#undef CMDS

164
static void __ata_port_freeze(struct ata_port *ap);
165
#ifdef CONFIG_PM
166 167
static void ata_eh_handle_port_suspend(struct ata_port *ap);
static void ata_eh_handle_port_resume(struct ata_port *ap);
168 169 170 171 172 173 174
#else /* CONFIG_PM */
static void ata_eh_handle_port_suspend(struct ata_port *ap)
{ }

static void ata_eh_handle_port_resume(struct ata_port *ap)
{ }
#endif /* CONFIG_PM */
175

176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240
static void __ata_ehi_pushv_desc(struct ata_eh_info *ehi, const char *fmt,
				 va_list args)
{
	ehi->desc_len += vscnprintf(ehi->desc + ehi->desc_len,
				     ATA_EH_DESC_LEN - ehi->desc_len,
				     fmt, args);
}

/**
 *	__ata_ehi_push_desc - push error description without adding separator
 *	@ehi: target EHI
 *	@fmt: printf format string
 *
 *	Format string according to @fmt and append it to @ehi->desc.
 *
 *	LOCKING:
 *	spin_lock_irqsave(host lock)
 */
void __ata_ehi_push_desc(struct ata_eh_info *ehi, const char *fmt, ...)
{
	va_list args;

	va_start(args, fmt);
	__ata_ehi_pushv_desc(ehi, fmt, args);
	va_end(args);
}

/**
 *	ata_ehi_push_desc - push error description with separator
 *	@ehi: target EHI
 *	@fmt: printf format string
 *
 *	Format string according to @fmt and append it to @ehi->desc.
 *	If @ehi->desc is not empty, ", " is added in-between.
 *
 *	LOCKING:
 *	spin_lock_irqsave(host lock)
 */
void ata_ehi_push_desc(struct ata_eh_info *ehi, const char *fmt, ...)
{
	va_list args;

	if (ehi->desc_len)
		__ata_ehi_push_desc(ehi, ", ");

	va_start(args, fmt);
	__ata_ehi_pushv_desc(ehi, fmt, args);
	va_end(args);
}

/**
 *	ata_ehi_clear_desc - clean error description
 *	@ehi: target EHI
 *
 *	Clear @ehi->desc.
 *
 *	LOCKING:
 *	spin_lock_irqsave(host lock)
 */
void ata_ehi_clear_desc(struct ata_eh_info *ehi)
{
	ehi->desc[0] = '\0';
	ehi->desc_len = 0;
}

241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302
/**
 *	ata_port_desc - append port description
 *	@ap: target ATA port
 *	@fmt: printf format string
 *
 *	Format string according to @fmt and append it to port
 *	description.  If port description is not empty, " " is added
 *	in-between.  This function is to be used while initializing
 *	ata_host.  The description is printed on host registration.
 *
 *	LOCKING:
 *	None.
 */
void ata_port_desc(struct ata_port *ap, const char *fmt, ...)
{
	va_list args;

	WARN_ON(!(ap->pflags & ATA_PFLAG_INITIALIZING));

	if (ap->link.eh_info.desc_len)
		__ata_ehi_push_desc(&ap->link.eh_info, " ");

	va_start(args, fmt);
	__ata_ehi_pushv_desc(&ap->link.eh_info, fmt, args);
	va_end(args);
}

#ifdef CONFIG_PCI

/**
 *	ata_port_pbar_desc - append PCI BAR description
 *	@ap: target ATA port
 *	@bar: target PCI BAR
 *	@offset: offset into PCI BAR
 *	@name: name of the area
 *
 *	If @offset is negative, this function formats a string which
 *	contains the name, address, size and type of the BAR and
 *	appends it to the port description.  If @offset is zero or
 *	positive, only name and offsetted address is appended.
 *
 *	LOCKING:
 *	None.
 */
void ata_port_pbar_desc(struct ata_port *ap, int bar, ssize_t offset,
			const char *name)
{
	struct pci_dev *pdev = to_pci_dev(ap->host->dev);
	char *type = "";
	unsigned long long start, len;

	if (pci_resource_flags(pdev, bar) & IORESOURCE_MEM)
		type = "m";
	else if (pci_resource_flags(pdev, bar) & IORESOURCE_IO)
		type = "i";

	start = (unsigned long long)pci_resource_start(pdev, bar);
	len = (unsigned long long)pci_resource_len(pdev, bar);

	if (offset < 0)
		ata_port_desc(ap, "%s %s%llu@0x%llx", name, type, len, start);
	else
303 304
		ata_port_desc(ap, "%s 0x%llx", name,
				start + (unsigned long long)offset);
305 306 307 308
}

#endif /* CONFIG_PCI */

309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375
static int ata_lookup_timeout_table(u8 cmd)
{
	int i;

	for (i = 0; i < ATA_EH_CMD_TIMEOUT_TABLE_SIZE; i++) {
		const u8 *cur;

		for (cur = ata_eh_cmd_timeout_table[i].commands; *cur; cur++)
			if (*cur == cmd)
				return i;
	}

	return -1;
}

/**
 *	ata_internal_cmd_timeout - determine timeout for an internal command
 *	@dev: target device
 *	@cmd: internal command to be issued
 *
 *	Determine timeout for internal command @cmd for @dev.
 *
 *	LOCKING:
 *	EH context.
 *
 *	RETURNS:
 *	Determined timeout.
 */
unsigned long ata_internal_cmd_timeout(struct ata_device *dev, u8 cmd)
{
	struct ata_eh_context *ehc = &dev->link->eh_context;
	int ent = ata_lookup_timeout_table(cmd);
	int idx;

	if (ent < 0)
		return ATA_EH_CMD_DFL_TIMEOUT;

	idx = ehc->cmd_timeout_idx[dev->devno][ent];
	return ata_eh_cmd_timeout_table[ent].timeouts[idx];
}

/**
 *	ata_internal_cmd_timed_out - notification for internal command timeout
 *	@dev: target device
 *	@cmd: internal command which timed out
 *
 *	Notify EH that internal command @cmd for @dev timed out.  This
 *	function should be called only for commands whose timeouts are
 *	determined using ata_internal_cmd_timeout().
 *
 *	LOCKING:
 *	EH context.
 */
void ata_internal_cmd_timed_out(struct ata_device *dev, u8 cmd)
{
	struct ata_eh_context *ehc = &dev->link->eh_context;
	int ent = ata_lookup_timeout_table(cmd);
	int idx;

	if (ent < 0)
		return;

	idx = ehc->cmd_timeout_idx[dev->devno][ent];
	if (ata_eh_cmd_timeout_table[ent].timeouts[idx + 1] != ULONG_MAX)
		ehc->cmd_timeout_idx[dev->devno][ent]++;
}

376
static void ata_ering_record(struct ata_ering *ering, unsigned int eflags,
377 378 379 380 381 382 383 384 385 386
			     unsigned int err_mask)
{
	struct ata_ering_entry *ent;

	WARN_ON(!err_mask);

	ering->cursor++;
	ering->cursor %= ATA_ERING_SIZE;

	ent = &ering->ring[ering->cursor];
387
	ent->eflags = eflags;
388 389 390 391
	ent->err_mask = err_mask;
	ent->timestamp = get_jiffies_64();
}

392 393 394 395 396 397 398 399 400
static struct ata_ering_entry *ata_ering_top(struct ata_ering *ering)
{
	struct ata_ering_entry *ent = &ering->ring[ering->cursor];

	if (ent->err_mask)
		return ent;
	return NULL;
}

401 402 403
int ata_ering_map(struct ata_ering *ering,
		  int (*map_fn)(struct ata_ering_entry *, void *),
		  void *arg)
404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421
{
	int idx, rc = 0;
	struct ata_ering_entry *ent;

	idx = ering->cursor;
	do {
		ent = &ering->ring[idx];
		if (!ent->err_mask)
			break;
		rc = map_fn(ent, arg);
		if (rc)
			break;
		idx = (idx - 1 + ATA_ERING_SIZE) % ATA_ERING_SIZE;
	} while (idx != ering->cursor);

	return rc;
}

422
static int ata_ering_clear_cb(struct ata_ering_entry *ent, void *void_arg)
423 424 425 426 427 428 429 430 431 432
{
	ent->eflags |= ATA_EFLAG_OLD_ER;
	return 0;
}

static void ata_ering_clear(struct ata_ering *ering)
{
	ata_ering_map(ering, ata_ering_clear_cb, NULL);
}

433 434
static unsigned int ata_eh_dev_action(struct ata_device *dev)
{
435
	struct ata_eh_context *ehc = &dev->link->eh_context;
436 437 438 439

	return ehc->i.action | ehc->i.dev_action[dev->devno];
}

440
static void ata_eh_clear_action(struct ata_link *link, struct ata_device *dev,
441 442
				struct ata_eh_info *ehi, unsigned int action)
{
443
	struct ata_device *tdev;
444 445 446

	if (!dev) {
		ehi->action &= ~action;
Tejun Heo's avatar
Tejun Heo committed
447
		ata_for_each_dev(tdev, link, ALL)
448
			ehi->dev_action[tdev->devno] &= ~action;
449 450 451 452 453 454
	} else {
		/* doesn't make sense for port-wide EH actions */
		WARN_ON(!(action & ATA_EH_PERDEV_MASK));

		/* break ehi->action into ehi->dev_action */
		if (ehi->action & action) {
Tejun Heo's avatar
Tejun Heo committed
455
			ata_for_each_dev(tdev, link, ALL)
456 457
				ehi->dev_action[tdev->devno] |=
					ehi->action & action;
458 459 460 461 462 463 464 465
			ehi->action &= ~action;
		}

		/* turn off the specified per-dev action */
		ehi->dev_action[dev->devno] &= ~action;
	}
}

466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500
/**
 *	ata_eh_acquire - acquire EH ownership
 *	@ap: ATA port to acquire EH ownership for
 *
 *	Acquire EH ownership for @ap.  This is the basic exclusion
 *	mechanism for ports sharing a host.  Only one port hanging off
 *	the same host can claim the ownership of EH.
 *
 *	LOCKING:
 *	EH context.
 */
void ata_eh_acquire(struct ata_port *ap)
{
	mutex_lock(&ap->host->eh_mutex);
	WARN_ON_ONCE(ap->host->eh_owner);
	ap->host->eh_owner = current;
}

/**
 *	ata_eh_release - release EH ownership
 *	@ap: ATA port to release EH ownership for
 *
 *	Release EH ownership for @ap if the caller.  The caller must
 *	have acquired EH ownership using ata_eh_acquire() previously.
 *
 *	LOCKING:
 *	EH context.
 */
void ata_eh_release(struct ata_port *ap)
{
	WARN_ON_ONCE(ap->host->eh_owner != current);
	ap->host->eh_owner = NULL;
	mutex_unlock(&ap->host->eh_mutex);
}

501 502 503 504 505 506 507 508 509 510 511
/**
 *	ata_scsi_timed_out - SCSI layer time out callback
 *	@cmd: timed out SCSI command
 *
 *	Handles SCSI layer timeout.  We race with normal completion of
 *	the qc for @cmd.  If the qc is already gone, we lose and let
 *	the scsi command finish (EH_HANDLED).  Otherwise, the qc has
 *	timed out and EH should be invoked.  Prevent ata_qc_complete()
 *	from finishing it by setting EH_SCHEDULED and return
 *	EH_NOT_HANDLED.
 *
512 513
 *	TODO: kill this function once old EH is gone.
 *
514 515 516 517 518 519
 *	LOCKING:
 *	Called from timer context
 *
 *	RETURNS:
 *	EH_HANDLED or EH_NOT_HANDLED
 */
520
enum blk_eh_timer_return ata_scsi_timed_out(struct scsi_cmnd *cmd)
521 522
{
	struct Scsi_Host *host = cmd->device->host;
523
	struct ata_port *ap = ata_shost_to_port(host);
524 525
	unsigned long flags;
	struct ata_queued_cmd *qc;
526
	enum blk_eh_timer_return ret;
527 528 529

	DPRINTK("ENTER\n");

530
	if (ap->ops->error_handler) {
531
		ret = BLK_EH_NOT_HANDLED;
532 533 534
		goto out;
	}

535
	ret = BLK_EH_HANDLED;
536
	spin_lock_irqsave(ap->lock, flags);
537
	qc = ata_qc_from_tag(ap, ap->link.active_tag);
538 539 540 541
	if (qc) {
		WARN_ON(qc->scsicmd != cmd);
		qc->flags |= ATA_QCFLAG_EH_SCHEDULED;
		qc->err_mask |= AC_ERR_TIMEOUT;
542
		ret = BLK_EH_NOT_HANDLED;
543
	}
544
	spin_unlock_irqrestore(ap->lock, flags);
545

546
 out:
547 548 549 550
	DPRINTK("EXIT, ret=%d\n", ret);
	return ret;
}

551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575
static void ata_eh_unload(struct ata_port *ap)
{
	struct ata_link *link;
	struct ata_device *dev;
	unsigned long flags;

	/* Restore SControl IPM and SPD for the next driver and
	 * disable attached devices.
	 */
	ata_for_each_link(link, ap, PMP_FIRST) {
		sata_scr_write(link, SCR_CONTROL, link->saved_scontrol & 0xff0);
		ata_for_each_dev(dev, link, ALL)
			ata_dev_disable(dev);
	}

	/* freeze and set UNLOADED */
	spin_lock_irqsave(ap->lock, flags);

	ata_port_freeze(ap);			/* won't be thawed */
	ap->pflags &= ~ATA_PFLAG_EH_PENDING;	/* clear pending from freeze */
	ap->pflags |= ATA_PFLAG_UNLOADED;

	spin_unlock_irqrestore(ap->lock, flags);
}

576 577 578 579 580 581 582 583 584 585 586 587
/**
 *	ata_scsi_error - SCSI layer error handler callback
 *	@host: SCSI host on which error occurred
 *
 *	Handles SCSI-layer-thrown error events.
 *
 *	LOCKING:
 *	Inherited from SCSI layer (none, can sleep)
 *
 *	RETURNS:
 *	Zero.
 */
588
void ata_scsi_error(struct Scsi_Host *host)
589
{
590
	struct ata_port *ap = ata_shost_to_port(host);
591
	unsigned long flags;
James Bottomley's avatar
James Bottomley committed
592
	LIST_HEAD(eh_work_q);
593 594 595

	DPRINTK("ENTER\n");

James Bottomley's avatar
James Bottomley committed
596 597 598 599
	spin_lock_irqsave(host->host_lock, flags);
	list_splice_init(&host->eh_cmd_q, &eh_work_q);
	spin_unlock_irqrestore(host->host_lock, flags);

600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627
	ata_scsi_cmd_error_handler(host, ap, &eh_work_q);

	/* If we timed raced normal completion and there is nothing to
	   recover nr_timedout == 0 why exactly are we doing error recovery ? */
	ata_scsi_port_error_handler(host, ap);

	/* finish or retry handled scmd's and clean up */
	WARN_ON(host->host_failed || !list_empty(&eh_work_q));

	DPRINTK("EXIT\n");
}

/**
 * ata_scsi_cmd_error_handler - error callback for a list of commands
 * @host:	scsi host containing the port
 * @ap:		ATA port within the host
 * @eh_work_q:	list of commands to process
 *
 * process the given list of commands and return those finished to the
 * ap->eh_done_q.  This function is the first part of the libata error
 * handler which processes a given list of failed commands.
 */
void ata_scsi_cmd_error_handler(struct Scsi_Host *host, struct ata_port *ap,
				struct list_head *eh_work_q)
{
	int i;
	unsigned long flags;

628 629
	/* make sure sff pio task is not running */
	ata_sff_flush_pio_task(ap);
630

Jeff Garzik's avatar
Jeff Garzik committed
631
	/* synchronize with host lock and sort out timeouts */
632 633 634

	/* For new EH, all qcs are finished in one of three ways -
	 * normal completion, error completion, and SCSI timeout.
635
	 * Both completions can race against SCSI timeout.  When normal
636 637 638 639 640 641 642 643 644 645 646 647 648
	 * completion wins, the qc never reaches EH.  When error
	 * completion wins, the qc has ATA_QCFLAG_FAILED set.
	 *
	 * When SCSI timeout wins, things are a bit more complex.
	 * Normal or error completion can occur after the timeout but
	 * before this point.  In such cases, both types of
	 * completions are honored.  A scmd is determined to have
	 * timed out iff its associated qc is active and not failed.
	 */
	if (ap->ops->error_handler) {
		struct scsi_cmnd *scmd, *tmp;
		int nr_timedout = 0;

649
		spin_lock_irqsave(ap->lock, flags);
650

651 652
		/* This must occur under the ap->lock as we don't want
		   a polled recovery to race the real interrupt handler
653

654 655
		   The lost_interrupt handler checks for any completed but
		   non-notified command and completes much like an IRQ handler.
656

657 658 659 660 661
		   We then fall into the error recovery code which will treat
		   this as if normal completion won the race */

		if (ap->ops->lost_interrupt)
			ap->ops->lost_interrupt(ap);
662

663
		list_for_each_entry_safe(scmd, tmp, eh_work_q, eh_entry) {
664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699
			struct ata_queued_cmd *qc;

			for (i = 0; i < ATA_MAX_QUEUE; i++) {
				qc = __ata_qc_from_tag(ap, i);
				if (qc->flags & ATA_QCFLAG_ACTIVE &&
				    qc->scsicmd == scmd)
					break;
			}

			if (i < ATA_MAX_QUEUE) {
				/* the scmd has an associated qc */
				if (!(qc->flags & ATA_QCFLAG_FAILED)) {
					/* which hasn't failed yet, timeout */
					qc->err_mask |= AC_ERR_TIMEOUT;
					qc->flags |= ATA_QCFLAG_FAILED;
					nr_timedout++;
				}
			} else {
				/* Normal completion occurred after
				 * SCSI timeout but before this point.
				 * Successfully complete it.
				 */
				scmd->retries = scmd->allowed;
				scsi_eh_finish_cmd(scmd, &ap->eh_done_q);
			}
		}

		/* If we have timed out qcs.  They belong to EH from
		 * this point but the state of the controller is
		 * unknown.  Freeze the port to make sure the IRQ
		 * handler doesn't diddle with those qcs.  This must
		 * be done atomically w.r.t. setting QCFLAG_FAILED.
		 */
		if (nr_timedout)
			__ata_port_freeze(ap);

700
		spin_unlock_irqrestore(ap->lock, flags);
701 702 703

		/* initialize eh_tries */
		ap->eh_tries = ATA_EH_MAX_TRIES;
704
	} else
705
		spin_unlock_wait(ap->lock);
706

707 708 709 710 711 712 713 714 715 716 717 718 719 720
}
EXPORT_SYMBOL(ata_scsi_cmd_error_handler);

/**
 * ata_scsi_port_error_handler - recover the port after the commands
 * @host:	SCSI host containing the port
 * @ap:		the ATA port
 *
 * Handle the recovery of the port @ap after all the commands
 * have been recovered.
 */
void ata_scsi_port_error_handler(struct Scsi_Host *host, struct ata_port *ap)
{
	unsigned long flags;
721 722 723

	/* invoke error handler */
	if (ap->ops->error_handler) {
724 725
		struct ata_link *link;

726 727 728
		/* acquire EH ownership */
		ata_eh_acquire(ap);
 repeat:
729 730 731
		/* kill fast drain timer */
		del_timer_sync(&ap->fastdrain_timer);

732 733 734
		/* process port resume request */
		ata_eh_handle_port_resume(ap);

735
		/* fetch & clear EH info */
736
		spin_lock_irqsave(ap->lock, flags);
737

Tejun Heo's avatar
Tejun Heo committed
738
		ata_for_each_link(link, ap, HOST_FIRST) {
739 740 741
			struct ata_eh_context *ehc = &link->eh_context;
			struct ata_device *dev;

742 743 744
			memset(&link->eh_context, 0, sizeof(link->eh_context));
			link->eh_context.i = link->eh_info;
			memset(&link->eh_info, 0, sizeof(link->eh_info));
745

Tejun Heo's avatar
Tejun Heo committed
746
			ata_for_each_dev(dev, link, ENABLED) {
747 748 749 750 751 752
				int devno = dev->devno;

				ehc->saved_xfer_mode[devno] = dev->xfer_mode;
				if (ata_ncq_enabled(dev))
					ehc->saved_ncq_enabled |= 1 << devno;
			}
753
		}
754

755 756
		ap->pflags |= ATA_PFLAG_EH_IN_PROGRESS;
		ap->pflags &= ~ATA_PFLAG_EH_PENDING;
757
		ap->excl_link = NULL;	/* don't maintain exclusion over EH */
758

759
		spin_unlock_irqrestore(ap->lock, flags);
760

761 762
		/* invoke EH, skip if unloading or suspended */
		if (!(ap->pflags & (ATA_PFLAG_UNLOADING | ATA_PFLAG_SUSPENDED)))
763
			ap->ops->error_handler(ap);
764 765 766 767 768
		else {
			/* if unloading, commence suicide */
			if ((ap->pflags & ATA_PFLAG_UNLOADING) &&
			    !(ap->pflags & ATA_PFLAG_UNLOADED))
				ata_eh_unload(ap);
769
			ata_eh_finish(ap);
770
		}
771

772 773 774
		/* process port suspend request */
		ata_eh_handle_port_suspend(ap);

775
		/* Exception might have happened after ->error_handler
776 777 778
		 * recovered the port but before this point.  Repeat
		 * EH in such case.
		 */
779
		spin_lock_irqsave(ap->lock, flags);
780

781
		if (ap->pflags & ATA_PFLAG_EH_PENDING) {
782
			if (--ap->eh_tries) {
783
				spin_unlock_irqrestore(ap->lock, flags);
784 785
				goto repeat;
			}
786 787 788
			ata_port_err(ap,
				     "EH pending after %d tries, giving up\n",
				     ATA_EH_MAX_TRIES);
789
			ap->pflags &= ~ATA_PFLAG_EH_PENDING;
790 791
		}

792
		/* this run is complete, make sure EH info is clear */
Tejun Heo's avatar
Tejun Heo committed
793
		ata_for_each_link(link, ap, HOST_FIRST)
794
			memset(&link->eh_info, 0, sizeof(link->eh_info));
795

796 797 798
		/* end eh (clear host_eh_scheduled) while holding
		 * ap->lock such that if exception occurs after this
		 * point but before EH completion, SCSI midlayer will
799 800
		 * re-initiate EH.
		 */
801
		ap->ops->end_eh(ap);
802

803
		spin_unlock_irqrestore(ap->lock, flags);
804
		ata_eh_release(ap);
805
	} else {
806
		WARN_ON(ata_qc_from_tag(ap, ap->link.active_tag) == NULL);
807 808
		ap->ops->eng_timeout(ap);
	}
809 810 811

	scsi_eh_flush_done_q(&ap->eh_done_q);

812
	/* clean up */
813
	spin_lock_irqsave(ap->lock, flags);
814

815
	if (ap->pflags & ATA_PFLAG_LOADING)
816
		ap->pflags &= ~ATA_PFLAG_LOADING;
817
	else if (ap->pflags & ATA_PFLAG_SCSI_HOTPLUG)
818
		schedule_delayed_work(&ap->hotplug_task, 0);
819 820

	if (ap->pflags & ATA_PFLAG_RECOVERED)
821
		ata_port_info(ap, "EH complete\n");
822

823
	ap->pflags &= ~(ATA_PFLAG_SCSI_HOTPLUG | ATA_PFLAG_RECOVERED);
824

825
	/* tell wait_eh that we're done */
826
	ap->pflags &= ~ATA_PFLAG_EH_IN_PROGRESS;
827 828
	wake_up_all(&ap->eh_wait_q);

829
	spin_unlock_irqrestore(ap->lock, flags);
830
}
831
EXPORT_SYMBOL_GPL(ata_scsi_port_error_handler);
832

833 834 835 836 837 838 839 840 841 842 843 844 845 846 847
/**
 *	ata_port_wait_eh - Wait for the currently pending EH to complete
 *	@ap: Port to wait EH for
 *
 *	Wait until the currently pending EH is complete.
 *
 *	LOCKING:
 *	Kernel thread context (may sleep).
 */
void ata_port_wait_eh(struct ata_port *ap)
{
	unsigned long flags;
	DEFINE_WAIT(wait);

 retry:
848
	spin_lock_irqsave(ap->lock, flags);
849

850
	while (ap->pflags & (ATA_PFLAG_EH_PENDING | ATA_PFLAG_EH_IN_PROGRESS)) {
851
		prepare_to_wait(&ap->eh_wait_q, &wait, TASK_UNINTERRUPTIBLE);
852
		spin_unlock_irqrestore(ap->lock, flags);
853
		schedule();
854
		spin_lock_irqsave(ap->lock, flags);
855
	}
856
	finish_wait(&ap->eh_wait_q, &wait);
857

858
	spin_unlock_irqrestore(ap->lock, flags);
859 860

	/* make sure SCSI EH is complete */
Jeff Garzik's avatar
Jeff Garzik committed
861
	if (scsi_host_in_recovery(ap->scsi_host)) {
862
		ata_msleep(ap, 10);
863 864 865
		goto retry;
	}
}
866
EXPORT_SYMBOL_GPL(ata_port_wait_eh);
867

868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911
static int ata_eh_nr_in_flight(struct ata_port *ap)
{
	unsigned int tag;
	int nr = 0;

	/* count only non-internal commands */
	for (tag = 0; tag < ATA_MAX_QUEUE - 1; tag++)
		if (ata_qc_from_tag(ap, tag))
			nr++;

	return nr;
}

void ata_eh_fastdrain_timerfn(unsigned long arg)
{
	struct ata_port *ap = (void *)arg;
	unsigned long flags;
	int cnt;

	spin_lock_irqsave(ap->lock, flags);

	cnt = ata_eh_nr_in_flight(ap);

	/* are we done? */
	if (!cnt)
		goto out_unlock;

	if (cnt == ap->fastdrain_cnt) {
		unsigned int tag;

		/* No progress during the last interval, tag all
		 * in-flight qcs as timed out and freeze the port.
		 */
		for (tag = 0; tag < ATA_MAX_QUEUE - 1; tag++) {
			struct ata_queued_cmd *qc = ata_qc_from_tag(ap, tag);
			if (qc)
				qc->err_mask |= AC_ERR_TIMEOUT;
		}

		ata_port_freeze(ap);
	} else {
		/* some qcs have finished, give it another chance */
		ap->fastdrain_cnt = cnt;
		ap->fastdrain_timer.expires =
912
			ata_deadline(jiffies, ATA_EH_FASTDRAIN_INTERVAL);
913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951
		add_timer(&ap->fastdrain_timer);
	}

 out_unlock:
	spin_unlock_irqrestore(ap->lock, flags);
}

/**
 *	ata_eh_set_pending - set ATA_PFLAG_EH_PENDING and activate fast drain
 *	@ap: target ATA port
 *	@fastdrain: activate fast drain
 *
 *	Set ATA_PFLAG_EH_PENDING and activate fast drain if @fastdrain
 *	is non-zero and EH wasn't pending before.  Fast drain ensures
 *	that EH kicks in in timely manner.
 *
 *	LOCKING:
 *	spin_lock_irqsave(host lock)
 */
static void ata_eh_set_pending(struct ata_port *ap, int fastdrain)
{
	int cnt;

	/* already scheduled? */
	if (ap->pflags & ATA_PFLAG_EH_PENDING)
		return;

	ap->pflags |= ATA_PFLAG_EH_PENDING;

	if (!fastdrain)
		return;

	/* do we have in-flight qcs? */
	cnt = ata_eh_nr_in_flight(ap);
	if (!cnt)
		return;

	/* activate fast drain */
	ap->fastdrain_cnt = cnt;
952 953
	ap->fastdrain_timer.expires =
		ata_deadline(jiffies, ATA_EH_FASTDRAIN_INTERVAL);
954 955 956
	add_timer(&ap->fastdrain_timer);
}

957 958 959 960 961 962 963 964
/**
 *	ata_qc_schedule_eh - schedule qc for error handling
 *	@qc: command to schedule error handling for
 *
 *	Schedule error handling for @qc.  EH will kick in as soon as
 *	other commands are drained.
 *
 *	LOCKING:
Jeff Garzik's avatar
Jeff Garzik committed
965
 *	spin_lock_irqsave(host lock)
966 967 968 969
 */
void ata_qc_schedule_eh(struct ata_queued_cmd *qc)
{
	struct ata_port *ap = qc->ap;
970 971
	struct request_queue *q = qc->scsicmd->device->request_queue;
	unsigned long flags;
972 973 974 975

	WARN_ON(!ap->ops->error_handler);

	qc->flags |= ATA_QCFLAG_FAILED;
976
	ata_eh_set_pending(ap, 1);
977 978 979 980 981 982

	/* The following will fail if timeout has already expired.
	 * ata_scsi_error() takes care of such scmds on EH entry.
	 * Note that ATA_QCFLAG_FAILED is unconditionally set after
	 * this function completes.
	 */
983
	spin_lock_irqsave(q->queue_lock, flags);
984
	blk_abort_request(qc->scsicmd->request);
985
	spin_unlock_irqrestore(q->queue_lock, flags);
986 987
}

988
/**
989 990
 * ata_std_sched_eh - non-libsas ata_ports issue eh with this common routine
 * @ap: ATA port to schedule EH for
991
 *
992
 *	LOCKING: inherited from ata_port_schedule_eh
Jeff Garzik's avatar
Jeff Garzik committed
993
 *	spin_lock_irqsave(host lock)
994
 */
995
void ata_std_sched_eh(struct ata_port *ap)
996 997 998
{
	WARN_ON(!ap->ops->error_handler);

999 1000 1001
	if (ap->pflags & ATA_PFLAG_INITIALIZING)
		return;

1002
	ata_eh_set_pending(ap, 1);
Jeff Garzik's avatar
Jeff Garzik committed
1003
	scsi_schedule_eh(ap->scsi_host);
1004 1005 1006

	DPRINTK("port EH scheduled\n");
}
1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044
EXPORT_SYMBOL_GPL(ata_std_sched_eh);

/**
 * ata_std_end_eh - non-libsas ata_ports complete eh with this common routine
 * @ap: ATA port to end EH for
 *
 * In the libata object model there is a 1:1 mapping of ata_port to
 * shost, so host fields can be directly manipulated under ap->lock, in
 * the libsas case we need to hold a lock at the ha->level to coordinate
 * these events.
 *
 *	LOCKING:
 *	spin_lock_irqsave(host lock)
 */
void ata_std_end_eh(struct ata_port *ap)
{
	struct Scsi_Host *host = ap->scsi_host;

	host->host_eh_scheduled = 0;
}
EXPORT_SYMBOL(ata_std_end_eh);


/**
 *	ata_port_schedule_eh - schedule error handling without a qc
 *	@ap: ATA port to schedule EH for
 *
 *	Schedule error handling for @ap.  EH will kick in as soon as
 *	all commands are drained.
 *
 *	LOCKING:
 *	spin_lock_irqsave(host lock)
 */
void ata_port_schedule_eh(struct ata_port *ap)
{
	/* see: ata_std_sched_eh, unless you know better */
	ap->ops->sched_eh(ap);
}
1045

1046
static int ata_do_link_abort(struct ata_port *ap, struct ata_link *link)
1047 1048 1049 1050 1051
{
	int tag, nr_aborted = 0;

	WARN_ON(!ap->ops->error_handler);

1052 1053 1054
	/* we're gonna abort all commands, no need for fast drain */
	ata_eh_set_pending(ap, 0);

1055 1056 1057
	for (tag = 0; tag < ATA_MAX_QUEUE; tag++) {
		struct ata_queued_cmd *qc = ata_qc_from_tag(ap, tag);

1058
		if (qc && (!link || qc->dev->link == link)) {
1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070
			qc->flags |= ATA_QCFLAG_FAILED;
			ata_qc_complete(qc);
			nr_aborted++;
		}
	}

	if (!nr_aborted)
		ata_port_schedule_eh(ap);

	return nr_aborted;
}

1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104
/**
 *	ata_link_abort - abort all qc's on the link
 *	@link: ATA link to abort qc's for
 *
 *	Abort all active qc's active on @link and schedule EH.
 *
 *	LOCKING:
 *	spin_lock_irqsave(host lock)
 *
 *	RETURNS:
 *	Number of aborted qc's.
 */
int ata_link_abort(struct ata_link *link)
{
	return ata_do_link_abort(link->ap, link);
}

/**
 *	ata_port_abort - abort all qc's on the port
 *	@ap: ATA port to abort qc's for
 *
 *	Abort all active qc's of @ap and schedule EH.
 *
 *	LOCKING:
 *	spin_lock_irqsave(host_set lock)
 *
 *	RETURNS:
 *	Number of aborted qc's.
 */
int ata_port_abort(struct ata_port *ap)
{
	return ata_do_link_abort(ap, NULL);
}

1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120
/**
 *	__ata_port_freeze - freeze port
 *	@ap: ATA port to freeze
 *
 *	This function is called when HSM violation or some other
 *	condition disrupts normal operation of the port.  Frozen port
 *	is not allowed to perform any operation until the port is
 *	thawed, which usually follows a successful reset.
 *
 *	ap->ops->freeze() callback can be used for freezing the port
 *	hardware-wise (e.g. mask interrupt and stop DMA engine).  If a
 *	port cannot be frozen hardware-wise, the interrupt handler
 *	must ack and clear interrupts unconditionally while the port
 *	is frozen.
 *
 *	LOCKING:
Jeff Garzik's avatar
Jeff Garzik committed
1121
 *	spin_lock_irqsave(host lock)
1122 1123 1124 1125 1126 1127 1128 1129
 */
static void __ata_port_freeze(struct ata_port *ap)
{
	WARN_ON(!ap->ops->error_handler);

	if (ap->ops->freeze)
		ap->ops->freeze(ap);

1130
	ap->pflags |= ATA_PFLAG_FROZEN;
1131

1132
	DPRINTK("ata%u port frozen\n", ap->print_id);
1133 1134 1135 1136 1137 1138
}

/**
 *	ata_port_freeze - abort & freeze port
 *	@ap: ATA port to freeze
 *
1139 1140 1141
 *	Abort and freeze @ap.  The freeze operation must be called
 *	first, because some hardware requires special operations
 *	before the taskfile registers are accessible.
1142 1143
 *
 *	LOCKING:
Jeff Garzik's avatar
Jeff Garzik committed
1144
 *	spin_lock_irqsave(host lock)
1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155
 *
 *	RETURNS:
 *	Number of aborted commands.
 */
int ata_port_freeze(struct ata_port *ap)
{
	int nr_aborted;

	WARN_ON(!ap->ops->error_handler);

	__ata_port_freeze(ap);
1156
	nr_aborted = ata_port_abort(ap);
1157 1158 1159 1160

	return nr_aborted;
}

1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185
/**
 *	sata_async_notification - SATA async notification handler
 *	@ap: ATA port where async notification is received
 *
 *	Handler to be called when async notification via SDB FIS is
 *	received.  This function schedules EH if necessary.
 *
 *	LOCKING:
 *	spin_lock_irqsave(host lock)
 *
 *	RETURNS:
 *	1 if EH is scheduled, 0 otherwise.
 */
int sata_async_notification(struct ata_port *ap)
{
	u32 sntf;
	int rc;

	if (!(ap->flags & ATA_FLAG_AN))
		return 0;

	rc = sata_scr_read(&ap->link, SCR_NOTIFICATION, &sntf);
	if (rc == 0)
		sata_scr_write(&ap->link, SCR_NOTIFICATION, sntf);

1186
	if (!sata_pmp_attached(ap) || rc) {
1187
		/* PMP is not attached or SNTF is not available */
1188
		if (!sata_pmp_attached(ap)) {
1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212
			/* PMP is not attached.  Check whether ATAPI
			 * AN is configured.  If so, notify media
			 * change.
			 */
			struct ata_device *dev = ap->link.device;

			if ((dev->class == ATA_DEV_ATAPI) &&
			    (dev->flags & ATA_DFLAG_AN))
				ata_scsi_media_change_notify(dev);
			return 0;
		} else {
			/* PMP is attached but SNTF is not available.
			 * ATAPI async media change notification is
			 * not used.  The PMP must be reporting PHY
			 * status change, schedule EH.
			 */
			ata_port_schedule_eh(ap);
			return 1;
		}
	} else {
		/* PMP is attached and SNTF is available */
		struct ata_link *link;

		/* check and notify ATAPI AN */
Tejun Heo's avatar
Tejun Heo committed
1213
		ata_for_each_link(link, ap, EDGE) {
1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233
			if (!(sntf & (1 << link->pmp)))
				continue;

			if ((link->device->class == ATA_DEV_ATAPI) &&
			    (link->device->flags & ATA_DFLAG_AN))
				ata_scsi_media_change_notify(link->device);
		}

		/* If PMP is reporting that PHY status of some
		 * downstream ports has changed, schedule EH.
		 */
		if (sntf & (1 << SATA_PMP_CTRL_PORT)) {
			ata_port_schedule_eh(ap);
			return 1;
		}

		return 0;
	}
}

1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249
/**
 *	ata_eh_freeze_port - EH helper to freeze port
 *	@ap: ATA port to freeze
 *
 *	Freeze @ap.
 *
 *	LOCKING:
 *	None.
 */
void ata_eh_freeze_port(struct ata_port *ap)
{
	unsigned long flags;

	if (!ap->ops->error_handler)
		return;

1250
	spin_lock_irqsave(ap->lock, flags);
1251
	__ata_port_freeze(ap);
1252
	spin_unlock_irqrestore(ap->lock, flags);
1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270
}

/**
 *	ata_port_thaw_port - EH helper to thaw port
 *	@ap: ATA port to thaw
 *
 *	Thaw frozen port @ap.
 *
 *	LOCKING:
 *	None.
 */
void ata_eh_thaw_port(struct ata_port *ap)
{
	unsigned long flags;

	if (!ap->ops->error_handler)
		return;

1271
	spin_lock_irqsave(ap->lock, flags);
1272

1273
	ap->pflags &= ~ATA_PFLAG_FROZEN;
1274 1275 1276 1277

	if (ap->ops->thaw)
		ap->ops->thaw(ap);

1278
	spin_unlock_irqrestore(ap->lock, flags);
1279

1280
	DPRINTK("ata%u port thawed\n", ap->print_id);
1281 1282
}

1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293
static void ata_eh_scsidone(struct scsi_cmnd *scmd)
{
	/* nada */
}

static void __ata_eh_qc_complete(struct ata_queued_cmd *qc)
{
	struct ata_port *ap = qc->ap;
	struct scsi_cmnd *scmd = qc->scsicmd;
	unsigned long flags;

1294
	spin_lock_irqsave(ap->lock, flags);
1295 1296 1297
	qc->scsidone = ata_eh_scsidone;
	__ata_qc_complete(qc);
	WARN_ON(ata_tag_valid(qc->tag));
1298
	spin_unlock_irqrestore(ap->lock, flags);
1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324

	scsi_eh_finish_cmd(scmd, &ap->eh_done_q);
}

/**
 *	ata_eh_qc_complete - Complete an active ATA command from EH
 *	@qc: Command to complete
 *
 *	Indicate to the mid and upper layers that an ATA command has
 *	completed.  To be used from EH.
 */
void ata_eh_qc_complete(struct ata_queued_cmd *qc)
{
	struct scsi_cmnd *scmd = qc->scsicmd;
	scmd->retries = scmd->allowed;
	__ata_eh_qc_complete(qc);
}

/**
 *	ata_eh_qc_retry - Tell midlayer to retry an ATA command after EH
 *	@qc: Command to retry
 *
 *	Indicate to the mid and upper layers that an ATA command
 *	should be retried.  To be used from EH.
 *
 *	SCSI midlayer limits the number of retries to scmd->allowed.
1325
 *	scmd->allowed is incremented for commands which get retried
1326 1327 1328 1329 1330
 *	due to unrelated failures (qc->err_mask is zero).
 */
void ata_eh_qc_retry(struct ata_queued_cmd *qc)
{
	struct scsi_cmnd *scmd = qc->scsicmd;
1331 1332
	if (!qc->err_mask)
		scmd->allowed++;
1333 1334
	__ata_eh_qc_complete(qc);
}
1335

1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350
/**
 *	ata_dev_disable - disable ATA device
 *	@dev: ATA device to disable
 *
 *	Disable @dev.
 *
 *	Locking:
 *	EH context.
 */
void ata_dev_disable(struct ata_device *dev)
{
	if (!ata_dev_enabled(dev))
		return;

	if (ata_msg_drv(dev->link->ap))
1351
		ata_dev_warn(dev, "disabled\n");
1352 1353 1354
	ata_acpi_on_disable(dev);
	ata_down_xfermask_limit(dev, ATA_DNXFER_FORCE_PIO0 | ATA_DNXFER_QUIET);
	dev->class++;
1355 1356 1357 1358 1359

	/* From now till the next successful probe, ering is used to
	 * track probe failures.  Clear accumulated device error info.
	 */
	ata_ering_clear(&dev->ering);
1360 1361
}

1362 1363 1364 1365 1366 1367 1368 1369 1370
/**
 *	ata_eh_detach_dev - detach ATA device
 *	@dev: ATA device to detach
 *
 *	Detach @dev.
 *
 *	LOCKING:
 *	None.
 */
1371
void ata_eh_detach_dev(struct ata_device *dev)
1372
{
1373 1374
	struct ata_link *link = dev->link;
	struct ata_port *ap = link->ap;
1375
	struct ata_eh_context *ehc = &link->eh_context;
1376 1377 1378 1379
	unsigned long flags;

	ata_dev_disable(dev);

1380
	spin_lock_irqsave(ap->lock, flags);
1381 1382 1383 1384 1385

	dev->flags &= ~ATA_DFLAG_DETACH;

	if (ata_scsi_offline_dev(dev)) {
		dev->flags |= ATA_DFLAG_DETACHED;
1386
		ap->pflags |= ATA_PFLAG_SCSI_HOTPLUG;
1387 1388
	}

1389
	/* clear per-dev EH info */
1390 1391
	ata_eh_clear_action(link, dev, &link->eh_info, ATA_EH_PERDEV_MASK);
	ata_eh_clear_action(link, dev, &link->eh_context.i, ATA_EH_PERDEV_MASK);
1392 1393
	ehc->saved_xfer_mode[dev->devno] = 0;
	ehc->saved_ncq_enabled &= ~(1 << dev->devno);
1394

1395
	spin_unlock_irqrestore(ap->lock, flags);
1396 1397
}

1398 1399
/**
 *	ata_eh_about_to_do - about to perform eh_action
1400
 *	@link: target ATA link
1401
 *	@dev: target ATA dev for per-dev action (can be NULL)
1402 1403 1404
 *	@action: action about to be performed
 *
 *	Called just before performing EH actions to clear related bits
1405 1406
 *	in @link->eh_info such that eh actions are not unnecessarily
 *	repeated.
1407 1408 1409 1410
 *
 *	LOCKING:
 *	None.
 */
1411 1412
void ata_eh_about_to_do(struct ata_link *link, struct ata_device *dev,
			unsigned int action)
1413
{
1414 1415 1416
	struct ata_port *ap = link->ap;
	struct ata_eh_info *ehi = &link->eh_info;
	struct ata_eh_context *ehc = &link->eh_context;
1417 1418
	unsigned long flags;

1419
	spin_lock_irqsave(ap->lock, flags);
1420

1421
	ata_eh_clear_action(link, dev, ehi, action);
1422

1423 1424 1425 1426
	/* About to take EH action, set RECOVERED.  Ignore actions on
	 * slave links as master will do them again.
	 */
	if (!(ehc->i.flags & ATA_EHI_QUIET) && link != ap->slave_link)
1427 1428
		ap->pflags |= ATA_PFLAG_RECOVERED;

1429
	spin_unlock_irqrestore(ap->lock, flags);
1430 1431
}

1432 1433
/**
 *	ata_eh_done - EH action complete
1434
*	@ap: target ATA port
1435 1436 1437 1438
 *	@dev: target ATA dev for per-dev action (can be NULL)
 *	@action: action just completed
 *
 *	Called right after performing EH actions to clear related bits
1439
 *	in @link->eh_context.
1440 1441 1442 1443
 *
 *	LOCKING:
 *	None.
 */
1444 1445
void ata_eh_done(struct ata_link *link, struct ata_device *dev,
		 unsigned int action)
1446
{
1447
	struct ata_eh_context *ehc = &link->eh_context;
1448

1449
	ata_eh_clear_action(link, dev, &ehc->i, action);
1450 1451
}

1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465
/**
 *	ata_err_string - convert err_mask to descriptive string
 *	@err_mask: error mask to convert to string
 *
 *	Convert @err_mask to descriptive string.  Errors are
 *	prioritized according to severity and only the most severe
 *	error is reported.
 *
 *	LOCKING:
 *	None.
 *
 *	RETURNS:
 *	Descriptive string for @err_mask
 */
1466
static const char *ata_err_string(unsigned int err_mask)
1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486
{
	if (err_mask & AC_ERR_HOST_BUS)
		return "host bus error";
	if (err_mask & AC_ERR_ATA_BUS)
		return "ATA bus error";
	if (err_mask & AC_ERR_TIMEOUT)
		return "timeout";
	if (err_mask & AC_ERR_HSM)
		return "HSM violation";
	if (err_mask & AC_ERR_SYSTEM)
		return "internal error";
	if (err_mask & AC_ERR_MEDIA)
		return "media error";
	if (err_mask & AC_ERR_INVALID)
		return "invalid argument";
	if (err_mask & AC_ERR_DEV)
		return "device error";
	return "unknown error";
}

1487 1488 1489
/**
 *	ata_read_log_page - read a specific log page
 *	@dev: target device
1490
 *	@log: log to read
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502
 *	@page: page to read
 *	@buf: buffer to store read page
 *	@sectors: number of sectors to read
 *
 *	Read log page using READ_LOG_EXT command.
 *
 *	LOCKING:
 *	Kernel thread context (may sleep).
 *
 *	RETURNS:
 *	0 on success, AC_ERR_* mask otherwise.
 */
1503 1504
unsigned int ata_read_log_page(struct ata_device *dev, u8 log,
			       u8 page, void *buf, unsigned int sectors)
1505 1506 1507 1508
{
	struct ata_taskfile tf;
	unsigned int err_mask;

1509
	DPRINTK("read log page - log 0x%x, page 0x%x\n", log, page);
1510 1511 1512

	ata_tf_init(dev, &tf);
	tf.command = ATA_CMD_READ_LOG_EXT;
1513 1514
	tf.lbal = log;
	tf.lbam = page;
1515 1516 1517 1518 1519 1520
	tf.nsect = sectors;
	tf.hob_nsect = sectors >> 8;
	tf.flags |= ATA_TFLAG_ISADDR | ATA_TFLAG_LBA48 | ATA_TFLAG_DEVICE;
	tf.protocol = ATA_PROT_PIO;

	err_mask = ata_exec_internal(dev, &tf, NULL, DMA_FROM_DEVICE,
1521
				     buf, sectors * ATA_SECT_SIZE, 0);
1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544

	DPRINTK("EXIT, err_mask=%x\n", err_mask);
	return err_mask;
}

/**
 *	ata_eh_read_log_10h - Read log page 10h for NCQ error details
 *	@dev: Device to read log page 10h from
 *	@tag: Resulting tag of the failed command
 *	@tf: Resulting taskfile registers of the failed command
 *
 *	Read log page 10h to obtain NCQ error details and clear error
 *	condition.
 *
 *	LOCKING:
 *	Kernel thread context (may sleep).
 *
 *	RETURNS:
 *	0 on success, -errno otherwise.
 */
static int ata_eh_read_log_10h(struct ata_device *dev,
			       int *tag, struct ata_taskfile *tf)
{
1545
	u8 *buf = dev->link->ap->sector_buf;
1546 1547 1548 1549
	unsigned int err_mask;
	u8 csum;
	int i;

1550
	err_mask = ata_read_log_page(dev, ATA_LOG_SATA_NCQ, 0, buf, 1);
1551 1552 1553 1554 1555 1556 1557
	if (err_mask)
		return -EIO;

	csum = 0;
	for (i = 0; i < ATA_SECT_SIZE; i++)
		csum += buf[i];
	if (csum)
1558 1559
		ata_dev_warn(dev, "invalid checksum 0x%x on log page 10h\n",
			     csum);
1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580

	if (buf[0] & 0x80)
		return -ENOENT;

	*tag = buf[0] & 0x1f;

	tf->command = buf[2];
	tf->feature = buf[3];
	tf->lbal = buf[4];
	tf->lbam = buf[5];
	tf->lbah = buf[6];
	tf->device = buf[7];
	tf->hob_lbal = buf[8];
	tf->hob_lbam = buf[9];
	tf->hob_lbah = buf[10];
	tf->nsect = buf[12];
	tf->hob_nsect = buf[13];

	return 0;
}

1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593
/**
 *	atapi_eh_tur - perform ATAPI TEST_UNIT_READY
 *	@dev: target ATAPI device
 *	@r_sense_key: out parameter for sense_key
 *
 *	Perform ATAPI TEST_UNIT_READY.
 *
 *	LOCKING:
 *	EH context (may sleep).
 *
 *	RETURNS:
 *	0 on success, AC_ERR_* mask on failure.
 */
1594
unsigned int atapi_eh_tur(struct ata_device *dev, u8 *r_sense_key)
1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611
{
	u8 cdb[ATAPI_CDB_LEN] = { TEST_UNIT_READY, 0, 0, 0, 0, 0 };
	struct ata_taskfile tf;
	unsigned int err_mask;

	ata_tf_init(dev, &tf);

	tf.flags |= ATA_TFLAG_ISADDR | ATA_TFLAG_DEVICE;
	tf.command = ATA_CMD_PACKET;
	tf.protocol = ATAPI_PROT_NODATA;

	err_mask = ata_exec_internal(dev, &tf, cdb, DMA_NONE, NULL, 0, 0);
	if (err_mask == AC_ERR_DEV)
		*r_sense_key = tf.feature >> 4;
	return err_mask;
}

1612 1613 1614 1615
/**
 *	atapi_eh_request_sense - perform ATAPI REQUEST_SENSE
 *	@dev: device to perform REQUEST_SENSE to
 *	@sense_buf: result sense data buffer (SCSI_SENSE_BUFFERSIZE bytes long)
1616
 *	@dfl_sense_key: default sense key to use
1617 1618 1619 1620 1621 1622 1623 1624 1625 1626
 *
 *	Perform ATAPI REQUEST_SENSE after the device reported CHECK
 *	SENSE.  This function is EH helper.
 *
 *	LOCKING:
 *	Kernel thread context (may sleep).
 *
 *	RETURNS:
 *	0 on success, AC_ERR_* mask on failure
 */
1627
unsigned int atapi_eh_request_sense(struct ata_device *dev,
1628
					   u8 *sense_buf, u8 dfl_sense_key)
1629
{
1630 1631
	u8 cdb[ATAPI_CDB_LEN] =
		{ REQUEST_SENSE, 0, 0, 0, SCSI_SENSE_BUFFERSIZE, 0 };
1632
	struct ata_port *ap = dev->link->ap;
1633 1634 1635 1636 1637 1638 1639
	struct ata_taskfile tf;

	DPRINTK("ATAPI request sense\n");

	/* FIXME: is this needed? */
	memset(sense_buf, 0, SCSI_SENSE_BUFFERSIZE);

1640 1641 1642
	/* initialize sense_buf with the error register,
	 * for the case where they are -not- overwritten
	 */
1643
	sense_buf[0] = 0x70;
1644
	sense_buf[2] = dfl_sense_key;
1645

1646
	/* some devices time out if garbage left in tf */
1647
	ata_tf_init(dev, &tf);
1648 1649 1650 1651 1652 1653

	tf.flags |= ATA_TFLAG_ISADDR | ATA_TFLAG_DEVICE;
	tf.command = ATA_CMD_PACKET;

	/* is it pointless to prefer PIO for "safety reasons"? */
	if (ap->flags & ATA_FLAG_PIO_DMA) {
1654
		tf.protocol = ATAPI_PROT_DMA;
1655 1656
		tf.feature |= ATAPI_PKT_DMA;
	} else {
1657
		tf.protocol = ATAPI_PROT_PIO;
1658 1659
		tf.lbam = SCSI_SENSE_BUFFERSIZE;
		tf.lbah = 0;
1660 1661 1662
	}

	return ata_exec_internal(dev, &tf, cdb, DMA_FROM_DEVICE,
1663
				 sense_buf, SCSI_SENSE_BUFFERSIZE, 0);
1664 1665 1666 1667
}

/**
 *	ata_eh_analyze_serror - analyze SError for a failed port
1668
 *	@link: ATA link to analyze SError for
1669 1670 1671 1672 1673 1674 1675
 *
 *	Analyze SError if available and further determine cause of
 *	failure.
 *
 *	LOCKING:
 *	None.
 */
1676
static void ata_eh_analyze_serror(struct ata_link *link)
1677
{
1678
	struct ata_eh_context *ehc = &link->eh_context;
1679 1680
	u32 serror = ehc->i.serror;
	unsigned int err_mask = 0, action = 0;
1681
	u32 hotplug_mask;
1682

1683
	if (serror & (SERR_PERSISTENT | SERR_DATA)) {
1684
		err_mask |= AC_ERR_ATA_BUS;
Tejun Heo's avatar
Tejun Heo committed
1685
		action |= ATA_EH_RESET;
1686 1687 1688
	}
	if (serror & SERR_PROTOCOL) {
		err_mask |= AC_ERR_HSM;
Tejun Heo's avatar
Tejun Heo committed
1689
		action |= ATA_EH_RESET;
1690 1691 1692
	}
	if (serror & SERR_INTERNAL) {
		err_mask |= AC_ERR_SYSTEM;
Tejun Heo's avatar
Tejun Heo committed
1693
		action |= ATA_EH_RESET;
1694
	}
1695 1696 1697 1698 1699 1700

	/* Determine whether a hotplug event has occurred.  Both
	 * SError.N/X are considered hotplug events for enabled or
	 * host links.  For disabled PMP links, only N bit is
	 * considered as X bit is left at 1 for link plugging.
	 */
1701
	if (link->lpm_policy > ATA_LPM_MAX_POWER)
1702 1703
		hotplug_mask = 0;	/* hotplug doesn't work w/ LPM */
	else if (!(link->flags & ATA_LFLAG_DISABLED) || ata_is_host_link(link))
1704 1705 1706 1707 1708
		hotplug_mask = SERR_PHYRDY_CHG | SERR_DEV_XCHG;
	else
		hotplug_mask = SERR_PHYRDY_CHG;

	if (serror & hotplug_mask)
1709
		ata_ehi_hotplugged(&ehc->i);
1710 1711 1712 1713 1714

	ehc->i.err_mask |= err_mask;
	ehc->i.action |= action;
}

1715 1716
/**
 *	ata_eh_analyze_ncq_error - analyze NCQ error
1717
 *	@link: ATA link to analyze NCQ error for
1718 1719 1720 1721 1722 1723 1724 1725 1726
 *
 *	Read log page 10h, determine the offending qc and acquire
 *	error status TF.  For NCQ device errors, all LLDDs have to do
 *	is setting AC_ERR_DEV in ehi->err_mask.  This function takes
 *	care of the rest.
 *
 *	LOCKING:
 *	Kernel thread context (may sleep).
 */
1727
void ata_eh_analyze_ncq_error(struct ata_link *link)
1728
{
1729 1730 1731
	struct ata_port *ap = link->ap;
	struct ata_eh_context *ehc = &link->eh_context;
	struct ata_device *dev = link->device;
1732 1733 1734 1735 1736
	struct ata_queued_cmd *qc;
	struct ata_taskfile tf;
	int tag, rc;

	/* if frozen, we can't do much */
1737
	if (ap->pflags & ATA_PFLAG_FROZEN)
1738 1739 1740
		return;

	/* is it NCQ device error? */
1741
	if (!link->sactive || !(ehc->i.err_mask & AC_ERR_DEV))
1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755
		return;

	/* has LLDD analyzed already? */
	for (tag = 0; tag < ATA_MAX_QUEUE; tag++) {
		qc = __ata_qc_from_tag(ap, tag);

		if (!(qc->flags & ATA_QCFLAG_FAILED))
			continue;

		if (qc->err_mask)
			return;
	}

	/* okay, this error is ours */
1756
	memset(&tf, 0, sizeof(tf));
1757 1758
	rc = ata_eh_read_log_10h(dev, &tag, &tf);
	if (rc) {
1759 1760
		ata_link_err(link, "failed to read log page 10h (errno=%d)\n",
			     rc);
1761 1762 1763
		return;
	}

1764
	if (!(link->sactive & (1 << tag))) {