libata-eh.c

/*
 *  libata-eh.c - libata error handling
 *
 *  Maintained by:  Jeff Garzik <jgarzik@pobox.com>
 *    		    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/config.h>
#include <linux/kernel.h>
#include <scsi/scsi.h>
#include <scsi/scsi_host.h>
#include <scsi/scsi_eh.h>
#include <scsi/scsi_device.h>
#include <scsi/scsi_cmnd.h>

#include <linux/libata.h>

#include "libata.h"

static void __ata_port_freeze(struct ata_port *ap);

/**
 *	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.
 *
 *	TODO: kill this function once old EH is gone.
 *
 *	LOCKING:
 *	Called from timer context
 *
 *	RETURNS:
 *	EH_HANDLED or EH_NOT_HANDLED
 */
enum scsi_eh_timer_return ata_scsi_timed_out(struct scsi_cmnd *cmd)
{
	struct Scsi_Host *host = cmd->device->host;
	struct ata_port *ap = ata_shost_to_port(host);
	unsigned long flags;
	struct ata_queued_cmd *qc;
	enum scsi_eh_timer_return ret;

	DPRINTK("ENTER\n");

	if (ap->ops->error_handler) {
		ret = EH_NOT_HANDLED;
		goto out;
	}

	ret = EH_HANDLED;
	spin_lock_irqsave(&ap->host_set->lock, flags);
	qc = ata_qc_from_tag(ap, ap->active_tag);
	if (qc) {
		WARN_ON(qc->scsicmd != cmd);
		qc->flags |= ATA_QCFLAG_EH_SCHEDULED;
		qc->err_mask |= AC_ERR_TIMEOUT;
		ret = EH_NOT_HANDLED;
	}
	spin_unlock_irqrestore(&ap->host_set->lock, flags);

 out:
	DPRINTK("EXIT, ret=%d\n", ret);
	return ret;
}

/**
 *	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.
 */
void ata_scsi_error(struct Scsi_Host *host)
{
	struct ata_port *ap = ata_shost_to_port(host);
	spinlock_t *hs_lock = &ap->host_set->lock;
	int i, repeat_cnt = ATA_EH_MAX_REPEAT;
	unsigned long flags;

	DPRINTK("ENTER\n");

	/* synchronize with port task */
	ata_port_flush_task(ap);

	/* synchronize with host_set lock and sort out timeouts */

	/* For new EH, all qcs are finished in one of three ways -
	 * normal completion, error completion, and SCSI timeout.
	 * Both cmpletions can race against SCSI timeout.  When normal
	 * 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;

		spin_lock_irqsave(hs_lock, flags);

		list_for_each_entry_safe(scmd, tmp, &host->eh_cmd_q, eh_entry) {
			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);

		spin_unlock_irqrestore(hs_lock, flags);
	} else
		spin_unlock_wait(hs_lock);

 repeat:
	/* invoke error handler */
	if (ap->ops->error_handler) {
		/* clear EH pending */
		spin_lock_irqsave(hs_lock, flags);
		ap->flags &= ~ATA_FLAG_EH_PENDING;
		spin_unlock_irqrestore(hs_lock, flags);

		/* invoke EH */
		ap->ops->error_handler(ap);

		/* Exception might have happend after ->error_handler
		 * recovered the port but before this point.  Repeat
		 * EH in such case.
		 */
		spin_lock_irqsave(hs_lock, flags);

		if (ap->flags & ATA_FLAG_EH_PENDING) {
			if (--repeat_cnt) {
				ata_port_printk(ap, KERN_INFO,
					"EH pending after completion, "
					"repeating EH (cnt=%d)\n", repeat_cnt);
				spin_unlock_irqrestore(hs_lock, flags);
				goto repeat;
			}
			ata_port_printk(ap, KERN_ERR, "EH pending after %d "
					"tries, giving up\n", ATA_EH_MAX_REPEAT);
		}

		/* Clear host_eh_scheduled while holding hs_lock such
		 * that if exception occurs after this point but
		 * before EH completion, SCSI midlayer will
		 * re-initiate EH.
		 */
		host->host_eh_scheduled = 0;

		spin_unlock_irqrestore(hs_lock, flags);
	} else {
		WARN_ON(ata_qc_from_tag(ap, ap->active_tag) == NULL);
		ap->ops->eng_timeout(ap);
	}

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

	scsi_eh_flush_done_q(&ap->eh_done_q);

	/* clean up */
	spin_lock_irqsave(hs_lock, flags);

	if (ap->flags & ATA_FLAG_RECOVERED)
		ata_port_printk(ap, KERN_INFO, "EH complete\n");
	ap->flags &= ~ATA_FLAG_RECOVERED;

	spin_unlock_irqrestore(hs_lock, flags);

	DPRINTK("EXIT\n");
}

/**
 *	ata_qc_timeout - Handle timeout of queued command
 *	@qc: Command that timed out
 *
 *	Some part of the kernel (currently, only the SCSI layer)
 *	has noticed that the active command on port @ap has not
 *	completed after a specified length of time.  Handle this
 *	condition by disabling DMA (if necessary) and completing
 *	transactions, with error if necessary.
 *
 *	This also handles the case of the "lost interrupt", where
 *	for some reason (possibly hardware bug, possibly driver bug)
 *	an interrupt was not delivered to the driver, even though the
 *	transaction completed successfully.
 *
 *	TODO: kill this function once old EH is gone.
 *
 *	LOCKING:
 *	Inherited from SCSI layer (none, can sleep)
 */
static void ata_qc_timeout(struct ata_queued_cmd *qc)
{
	struct ata_port *ap = qc->ap;
	struct ata_host_set *host_set = ap->host_set;
	u8 host_stat = 0, drv_stat;
	unsigned long flags;

	DPRINTK("ENTER\n");

	ap->hsm_task_state = HSM_ST_IDLE;

	spin_lock_irqsave(&host_set->lock, flags);

	switch (qc->tf.protocol) {

	case ATA_PROT_DMA:
	case ATA_PROT_ATAPI_DMA:
		host_stat = ap->ops->bmdma_status(ap);

		/* before we do anything else, clear DMA-Start bit */
		ap->ops->bmdma_stop(qc);

		/* fall through */

	default:
		ata_altstatus(ap);
		drv_stat = ata_chk_status(ap);

		/* ack bmdma irq events */
		ap->ops->irq_clear(ap);

		ata_dev_printk(qc->dev, KERN_ERR, "command 0x%x timeout, "
			       "stat 0x%x host_stat 0x%x\n",
			       qc->tf.command, drv_stat, host_stat);

		/* complete taskfile transaction */
		qc->err_mask |= ac_err_mask(drv_stat);
		break;
	}

	spin_unlock_irqrestore(&host_set->lock, flags);

	ata_eh_qc_complete(qc);

	DPRINTK("EXIT\n");
}

/**
 *	ata_eng_timeout - Handle timeout of queued command
 *	@ap: Port on which timed-out command is active
 *
 *	Some part of the kernel (currently, only the SCSI layer)
 *	has noticed that the active command on port @ap has not
 *	completed after a specified length of time.  Handle this
 *	condition by disabling DMA (if necessary) and completing
 *	transactions, with error if necessary.
 *
 *	This also handles the case of the "lost interrupt", where
 *	for some reason (possibly hardware bug, possibly driver bug)
 *	an interrupt was not delivered to the driver, even though the
 *	transaction completed successfully.
 *
 *	TODO: kill this function once old EH is gone.
 *
 *	LOCKING:
 *	Inherited from SCSI layer (none, can sleep)
 */
void ata_eng_timeout(struct ata_port *ap)
{
	DPRINTK("ENTER\n");

	ata_qc_timeout(ata_qc_from_tag(ap, ap->active_tag));

	DPRINTK("EXIT\n");
}

/**
 *	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:
 *	spin_lock_irqsave(host_set lock)
 */
void ata_qc_schedule_eh(struct ata_queued_cmd *qc)
{
	struct ata_port *ap = qc->ap;

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

	qc->flags |= ATA_QCFLAG_FAILED;
	qc->ap->flags |= ATA_FLAG_EH_PENDING;

	/* 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.
	 */
	scsi_req_abort_cmd(qc->scsicmd);
}

/**
 *	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_set lock)
 */
void ata_port_schedule_eh(struct ata_port *ap)
{
	WARN_ON(!ap->ops->error_handler);

	ap->flags |= ATA_FLAG_EH_PENDING;
	ata_schedule_scsi_eh(ap->host);

	DPRINTK("port EH scheduled\n");
}

/**
 *	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)
{
	int tag, nr_aborted = 0;

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

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

		if (qc) {
			qc->flags |= ATA_QCFLAG_FAILED;
			ata_qc_complete(qc);
			nr_aborted++;
		}
	}

	if (!nr_aborted)
		ata_port_schedule_eh(ap);

	return nr_aborted;
}

/**
 *	__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:
 *	spin_lock_irqsave(host_set lock)
 */
static void __ata_port_freeze(struct ata_port *ap)
{
	WARN_ON(!ap->ops->error_handler);

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

	ap->flags |= ATA_FLAG_FROZEN;

	DPRINTK("ata%u port frozen\n", ap->id);
}

/**
 *	ata_port_freeze - abort & freeze port
 *	@ap: ATA port to freeze
 *
 *	Abort and freeze @ap.
 *
 *	LOCKING:
 *	spin_lock_irqsave(host_set lock)
 *
 *	RETURNS:
 *	Number of aborted commands.
 */
int ata_port_freeze(struct ata_port *ap)
{
	int nr_aborted;

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

	nr_aborted = ata_port_abort(ap);
	__ata_port_freeze(ap);

	return nr_aborted;
}

/**
 *	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;

	spin_lock_irqsave(&ap->host_set->lock, flags);
	__ata_port_freeze(ap);
	spin_unlock_irqrestore(&ap->host_set->lock, flags);
}

/**
 *	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;

	spin_lock_irqsave(&ap->host_set->lock, flags);

	ap->flags &= ~ATA_FLAG_FROZEN;

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

	spin_unlock_irqrestore(&ap->host_set->lock, flags);

	DPRINTK("ata%u port thawed\n", ap->id);
}

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;

	spin_lock_irqsave(&ap->host_set->lock, flags);
	qc->scsidone = ata_eh_scsidone;
	__ata_qc_complete(qc);
	WARN_ON(ata_tag_valid(qc->tag));
	spin_unlock_irqrestore(&ap->host_set->lock, flags);

	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.
 *	scmd->retries is decremented for commands which get retried
 *	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;
	if (!qc->err_mask && scmd->retries)
		scmd->retries--;
	__ata_eh_qc_complete(qc);
}