Merge branch 'for-jeff' of git://htj.dyndns.org/libata-tj into tejun-merge

drivers/scsi/ahci.c

@@ -56,12 +56,15 @@ enum {
 	AHCI_MAX_SG		= 168, /* hardware max is 64K */
 	AHCI_DMA_BOUNDARY	= 0xffffffff,
 	AHCI_USE_CLUSTERING	= 0,
-	AHCI_CMD_SLOT_SZ	= 32 * 32,
+	AHCI_MAX_CMDS		= 32,
+	AHCI_CMD_SZ		= 32,
+	AHCI_CMD_SLOT_SZ	= AHCI_MAX_CMDS * AHCI_CMD_SZ,
 	AHCI_RX_FIS_SZ		= 256,
-	AHCI_CMD_TBL_HDR	= 0x80,
 	AHCI_CMD_TBL_CDB	= 0x40,
-	AHCI_CMD_TBL_SZ		= AHCI_CMD_TBL_HDR + (AHCI_MAX_SG * 16),
-	AHCI_PORT_PRIV_DMA_SZ	= AHCI_CMD_SLOT_SZ + AHCI_CMD_TBL_SZ +
+	AHCI_CMD_TBL_HDR_SZ	= 0x80,
+	AHCI_CMD_TBL_SZ		= AHCI_CMD_TBL_HDR_SZ + (AHCI_MAX_SG * 16),
+	AHCI_CMD_TBL_AR_SZ	= AHCI_CMD_TBL_SZ * AHCI_MAX_CMDS,
+	AHCI_PORT_PRIV_DMA_SZ	= AHCI_CMD_SLOT_SZ + AHCI_CMD_TBL_AR_SZ +
 				  AHCI_RX_FIS_SZ,
 	AHCI_IRQ_ON_SG		= (1 << 31),
 	AHCI_CMD_ATAPI		= (1 << 5),
@@ -71,6 +74,7 @@ enum {
 	AHCI_CMD_CLR_BUSY	= (1 << 10),
 
 	RX_FIS_D2H_REG		= 0x40,	/* offset of D2H Register FIS data */
+	RX_FIS_UNK		= 0x60, /* offset of Unknown FIS data */
 
 	board_ahci		= 0,
 	board_ahci_vt8251	= 1,
@@ -88,8 +92,9 @@ enum {
 	HOST_AHCI_EN		= (1 << 31), /* AHCI enabled */
 
 	/* HOST_CAP bits */
-	HOST_CAP_64		= (1 << 31), /* PCI DAC (64-bit DMA) support */
 	HOST_CAP_CLO		= (1 << 24), /* Command List Override support */
+	HOST_CAP_NCQ		= (1 << 30), /* Native Command Queueing */
+	HOST_CAP_64		= (1 << 31), /* PCI DAC (64-bit DMA) support */
 
 	/* registers for each SATA port */
 	PORT_LST_ADDR		= 0x00, /* command list DMA addr */
@@ -128,15 +133,16 @@ enum {
 	PORT_IRQ_PIOS_FIS	= (1 << 1), /* PIO Setup FIS rx'd */
 	PORT_IRQ_D2H_REG_FIS	= (1 << 0), /* D2H Register FIS rx'd */
 
-	PORT_IRQ_FATAL		= PORT_IRQ_TF_ERR |
-				  PORT_IRQ_HBUS_ERR |
-				  PORT_IRQ_HBUS_DATA_ERR |
-				  PORT_IRQ_IF_ERR,
-	DEF_PORT_IRQ		= PORT_IRQ_FATAL | PORT_IRQ_PHYRDY |
-				  PORT_IRQ_CONNECT | PORT_IRQ_SG_DONE |
-				  PORT_IRQ_UNK_FIS | PORT_IRQ_SDB_FIS |
-				  PORT_IRQ_DMAS_FIS | PORT_IRQ_PIOS_FIS |
-				  PORT_IRQ_D2H_REG_FIS,
+	PORT_IRQ_FREEZE		= PORT_IRQ_HBUS_ERR |
+				  PORT_IRQ_IF_ERR |
+				  PORT_IRQ_CONNECT |
+				  PORT_IRQ_UNK_FIS,
+	PORT_IRQ_ERROR		= PORT_IRQ_FREEZE |
+				  PORT_IRQ_TF_ERR |
+				  PORT_IRQ_HBUS_DATA_ERR,
+	DEF_PORT_IRQ		= PORT_IRQ_ERROR | PORT_IRQ_SG_DONE |
+				  PORT_IRQ_SDB_FIS | PORT_IRQ_DMAS_FIS |
+				  PORT_IRQ_PIOS_FIS | PORT_IRQ_D2H_REG_FIS,
 
 	/* PORT_CMD bits */
 	PORT_CMD_ATAPI		= (1 << 24), /* Device is ATAPI */
@@ -185,7 +191,6 @@ struct ahci_port_priv {
 	dma_addr_t		cmd_slot_dma;
 	void			*cmd_tbl;
 	dma_addr_t		cmd_tbl_dma;
-	struct ahci_sg		*cmd_tbl_sg;
 	void			*rx_fis;
 	dma_addr_t		rx_fis_dma;
 };
@@ -197,13 +202,15 @@ static unsigned int ahci_qc_issue(struct ata_queued_cmd *qc);
 static irqreturn_t ahci_interrupt (int irq, void *dev_instance, struct pt_regs *regs);
 static int ahci_probe_reset(struct ata_port *ap, unsigned int *classes);
 static void ahci_irq_clear(struct ata_port *ap);
-static void ahci_eng_timeout(struct ata_port *ap);
 static int ahci_port_start(struct ata_port *ap);
 static void ahci_port_stop(struct ata_port *ap);
 static void ahci_tf_read(struct ata_port *ap, struct ata_taskfile *tf);
 static void ahci_qc_prep(struct ata_queued_cmd *qc);
 static u8 ahci_check_status(struct ata_port *ap);
-static inline int ahci_host_intr(struct ata_port *ap, struct ata_queued_cmd *qc);
+static void ahci_freeze(struct ata_port *ap);
+static void ahci_thaw(struct ata_port *ap);
+static void ahci_error_handler(struct ata_port *ap);
+static void ahci_post_internal_cmd(struct ata_queued_cmd *qc);
 static void ahci_remove_one (struct pci_dev *pdev);
 
 static struct scsi_host_template ahci_sht = {
@@ -211,7 +218,8 @@ static struct scsi_host_template ahci_sht = {
 	.name			= DRV_NAME,
 	.ioctl			= ata_scsi_ioctl,
 	.queuecommand		= ata_scsi_queuecmd,
-	.can_queue		= ATA_DEF_QUEUE,
+	.change_queue_depth	= ata_scsi_change_queue_depth,
+	.can_queue		= AHCI_MAX_CMDS - 1,
 	.this_id		= ATA_SHT_THIS_ID,
 	.sg_tablesize		= AHCI_MAX_SG,
 	.cmd_per_lun		= ATA_SHT_CMD_PER_LUN,
@@ -237,14 +245,18 @@ static const struct ata_port_operations ahci_ops = {
 	.qc_prep		= ahci_qc_prep,
 	.qc_issue		= ahci_qc_issue,
 
-	.eng_timeout		= ahci_eng_timeout,
-
 	.irq_handler		= ahci_interrupt,
 	.irq_clear		= ahci_irq_clear,
 
 	.scr_read		= ahci_scr_read,
 	.scr_write		= ahci_scr_write,
 
+	.freeze			= ahci_freeze,
+	.thaw			= ahci_thaw,
+
+	.error_handler		= ahci_error_handler,
+	.post_internal_cmd	= ahci_post_internal_cmd,
+
 	.port_start		= ahci_port_start,
 	.port_stop		= ahci_port_stop,
 };
@@ -390,8 +402,6 @@ static int ahci_port_start(struct ata_port *ap)
 	pp->cmd_tbl = mem;
 	pp->cmd_tbl_dma = mem_dma;
 
-	pp->cmd_tbl_sg = mem + AHCI_CMD_TBL_HDR;
-
 	ap->private_data = pp;
 
 	if (hpriv->cap & HOST_CAP_64)
@@ -524,12 +534,17 @@ static unsigned int ahci_dev_classify(struct ata_port *ap)
 	return ata_dev_classify(&tf);
 }
 
-static void ahci_fill_cmd_slot(struct ahci_port_priv *pp, u32 opts)
+static void ahci_fill_cmd_slot(struct ahci_port_priv *pp, unsigned int tag,
+			       u32 opts)
 {
-	pp->cmd_slot[0].opts = cpu_to_le32(opts);
-	pp->cmd_slot[0].status = 0;
-	pp->cmd_slot[0].tbl_addr = cpu_to_le32(pp->cmd_tbl_dma & 0xffffffff);
-	pp->cmd_slot[0].tbl_addr_hi = cpu_to_le32((pp->cmd_tbl_dma >> 16) >> 16);
+	dma_addr_t cmd_tbl_dma;
+
+	cmd_tbl_dma = pp->cmd_tbl_dma + tag * AHCI_CMD_TBL_SZ;
+
+	pp->cmd_slot[tag].opts = cpu_to_le32(opts);
+	pp->cmd_slot[tag].status = 0;
+	pp->cmd_slot[tag].tbl_addr = cpu_to_le32(cmd_tbl_dma & 0xffffffff);
+	pp->cmd_slot[tag].tbl_addr_hi = cpu_to_le32((cmd_tbl_dma >> 16) >> 16);
 }
 
 static int ahci_clo(struct ata_port *ap)
@@ -567,7 +582,7 @@ static int ahci_softreset(struct ata_port *ap, unsigned int *class)
 
 	DPRINTK("ENTER\n");
 
-	if (!sata_dev_present(ap)) {
+	if (ata_port_offline(ap)) {
 		DPRINTK("PHY reports no device\n");
 		*class = ATA_DEV_NONE;
 		return 0;
@@ -597,11 +612,12 @@ static int ahci_softreset(struct ata_port *ap, unsigned int *class)
 	/* restart engine */
 	ahci_start_engine(ap);
 
-	ata_tf_init(ap, &tf, 0);
+	ata_tf_init(ap->device, &tf);
 	fis = pp->cmd_tbl;
 
 	/* issue the first D2H Register FIS */
-	ahci_fill_cmd_slot(pp, cmd_fis_len | AHCI_CMD_RESET | AHCI_CMD_CLR_BUSY);
+	ahci_fill_cmd_slot(pp, 0,
+			   cmd_fis_len | AHCI_CMD_RESET | AHCI_CMD_CLR_BUSY);
 
 	tf.ctl |= ATA_SRST;
 	ata_tf_to_fis(&tf, fis, 0);
@@ -620,7 +636,7 @@ static int ahci_softreset(struct ata_port *ap, unsigned int *class)
 	msleep(1);
 
 	/* issue the second D2H Register FIS */
-	ahci_fill_cmd_slot(pp, cmd_fis_len);
+	ahci_fill_cmd_slot(pp, 0, cmd_fis_len);
 
 	tf.ctl &= ~ATA_SRST;
 	ata_tf_to_fis(&tf, fis, 0);
@@ -640,7 +656,7 @@ static int ahci_softreset(struct ata_port *ap, unsigned int *class)
 	msleep(150);
 
 	*class = ATA_DEV_NONE;
-	if (sata_dev_present(ap)) {
+	if (ata_port_online(ap)) {
 		if (ata_busy_sleep(ap, ATA_TMOUT_BOOT_QUICK, ATA_TMOUT_BOOT)) {
 			rc = -EIO;
 			reason = "device not ready";
@@ -655,8 +671,7 @@ static int ahci_softreset(struct ata_port *ap, unsigned int *class)
  fail_restart:
 	ahci_start_engine(ap);
  fail:
-	printk(KERN_ERR "ata%u: softreset failed (%s)\n",
-	       ap->id, reason);
+	ata_port_printk(ap, KERN_ERR, "softreset failed (%s)\n", reason);
 	return rc;
 }
 
@@ -670,7 +685,7 @@ static int ahci_hardreset(struct ata_port *ap, unsigned int *class)
 	rc = sata_std_hardreset(ap, class);
 	ahci_start_engine(ap);
 
-	if (rc == 0)
+	if (rc == 0 && ata_port_online(ap))
 		*class = ahci_dev_classify(ap);
 	if (*class == ATA_DEV_UNKNOWN)
 		*class = ATA_DEV_NONE;
@@ -726,9 +741,8 @@ static void ahci_tf_read(struct ata_port *ap, struct ata_taskfile *tf)
 	ata_tf_from_fis(d2h_fis, tf);
 }
 
-static unsigned int ahci_fill_sg(struct ata_queued_cmd *qc)
+static unsigned int ahci_fill_sg(struct ata_queued_cmd *qc, void *cmd_tbl)
 {
-	struct ahci_port_priv *pp = qc->ap->private_data;
 	struct scatterlist *sg;
 	struct ahci_sg *ahci_sg;
 	unsigned int n_sg = 0;
@@ -738,7 +752,7 @@ static unsigned int ahci_fill_sg(struct ata_queued_cmd *qc)
 	/*
 	 * Next, the S/G list.
 	 */
-	ahci_sg = pp->cmd_tbl_sg;
+	ahci_sg = cmd_tbl + AHCI_CMD_TBL_HDR_SZ;
 	ata_for_each_sg(sg, qc) {
 		dma_addr_t addr = sg_dma_address(sg);
 		u32 sg_len = sg_dma_len(sg);
@@ -759,6 +773,7 @@ static void ahci_qc_prep(struct ata_queued_cmd *qc)
 	struct ata_port *ap = qc->ap;
 	struct ahci_port_priv *pp = ap->private_data;
 	int is_atapi = is_atapi_taskfile(&qc->tf);
+	void *cmd_tbl;
 	u32 opts;
 	const u32 cmd_fis_len = 5; /* five dwords */
 	unsigned int n_elem;
@@ -767,16 +782,17 @@ static void ahci_qc_prep(struct ata_queued_cmd *qc)
 	 * Fill in command table information.  First, the header,
 	 * a SATA Register - Host to Device command FIS.
 	 */
-	ata_tf_to_fis(&qc->tf, pp->cmd_tbl, 0);
+	cmd_tbl = pp->cmd_tbl + qc->tag * AHCI_CMD_TBL_SZ;
+
+	ata_tf_to_fis(&qc->tf, cmd_tbl, 0);
 	if (is_atapi) {
-		memset(pp->cmd_tbl + AHCI_CMD_TBL_CDB, 0, 32);
-		memcpy(pp->cmd_tbl + AHCI_CMD_TBL_CDB, qc->cdb,
-		       qc->dev->cdb_len);
+		memset(cmd_tbl + AHCI_CMD_TBL_CDB, 0, 32);
+		memcpy(cmd_tbl + AHCI_CMD_TBL_CDB, qc->cdb, qc->dev->cdb_len);
 	}
 
 	n_elem = 0;
 	if (qc->flags & ATA_QCFLAG_DMAMAP)
-		n_elem = ahci_fill_sg(qc);
+		n_elem = ahci_fill_sg(qc, cmd_tbl);
 
 	/*
 	 * Fill in command slot information.
@@ -787,112 +803,123 @@ static void ahci_qc_prep(struct ata_queued_cmd *qc)
 	if (is_atapi)
 		opts |= AHCI_CMD_ATAPI | AHCI_CMD_PREFETCH;
 
-	ahci_fill_cmd_slot(pp, opts);
+	ahci_fill_cmd_slot(pp, qc->tag, opts);
 }
 
-static void ahci_restart_port(struct ata_port *ap, u32 irq_stat)
+static void ahci_error_intr(struct ata_port *ap, u32 irq_stat)
 {
-	void __iomem *mmio = ap->host_set->mmio_base;
-	void __iomem *port_mmio = ahci_port_base(mmio, ap->port_no);
-	u32 tmp;
+	struct ahci_port_priv *pp = ap->private_data;
+	struct ata_eh_info *ehi = &ap->eh_info;
+	unsigned int err_mask = 0, action = 0;
+	struct ata_queued_cmd *qc;
+	u32 serror;
 
-	if ((ap->device[0].class != ATA_DEV_ATAPI) ||
-	    ((irq_stat & PORT_IRQ_TF_ERR) == 0))
-		printk(KERN_WARNING "ata%u: port reset, "
-		       "p_is %x is %x pis %x cmd %x tf %x ss %x se %x\n",
-			ap->id,
-			irq_stat,
-			readl(mmio + HOST_IRQ_STAT),
-			readl(port_mmio + PORT_IRQ_STAT),
-			readl(port_mmio + PORT_CMD),
-			readl(port_mmio + PORT_TFDATA),
-			readl(port_mmio + PORT_SCR_STAT),
-			readl(port_mmio + PORT_SCR_ERR));
-
-	/* stop DMA */
-	ahci_stop_engine(ap);
+	ata_ehi_clear_desc(ehi);
 
-	/* clear SATA phy error, if any */
-	tmp = readl(port_mmio + PORT_SCR_ERR);
-	writel(tmp, port_mmio + PORT_SCR_ERR);
+	/* AHCI needs SError cleared; otherwise, it might lock up */
+	serror = ahci_scr_read(ap, SCR_ERROR);
+	ahci_scr_write(ap, SCR_ERROR, serror);
 
-	/* if DRQ/BSY is set, device needs to be reset.
-	 * if so, issue COMRESET
-	 */
-	tmp = readl(port_mmio + PORT_TFDATA);
-	if (tmp & (ATA_BUSY | ATA_DRQ)) {
-		writel(0x301, port_mmio + PORT_SCR_CTL);
-		readl(port_mmio + PORT_SCR_CTL); /* flush */
-		udelay(10);
-		writel(0x300, port_mmio + PORT_SCR_CTL);
-		readl(port_mmio + PORT_SCR_CTL); /* flush */
+	/* analyze @irq_stat */
+	ata_ehi_push_desc(ehi, "irq_stat 0x%08x", irq_stat);
+
+	if (irq_stat & PORT_IRQ_TF_ERR)
+		err_mask |= AC_ERR_DEV;
+
+	if (irq_stat & (PORT_IRQ_HBUS_ERR | PORT_IRQ_HBUS_DATA_ERR)) {
+		err_mask |= AC_ERR_HOST_BUS;
+		action |= ATA_EH_SOFTRESET;
 	}
 
-	/* re-start DMA */
-	ahci_start_engine(ap);
-}
+	if (irq_stat & PORT_IRQ_IF_ERR) {
+		err_mask |= AC_ERR_ATA_BUS;
+		action |= ATA_EH_SOFTRESET;
+		ata_ehi_push_desc(ehi, ", interface fatal error");
+	}
 
-static void ahci_eng_timeout(struct ata_port *ap)
-{
-	struct ata_host_set *host_set = ap->host_set;
-	void __iomem *mmio = host_set->mmio_base;
-	void __iomem *port_mmio = ahci_port_base(mmio, ap->port_no);
-	struct ata_queued_cmd *qc;
-	unsigned long flags;
+	if (irq_stat & (PORT_IRQ_CONNECT | PORT_IRQ_PHYRDY)) {
+		err_mask |= AC_ERR_ATA_BUS;
+		action |= ATA_EH_SOFTRESET;
+		ata_ehi_push_desc(ehi, ", %s", irq_stat & PORT_IRQ_CONNECT ?
+			"connection status changed" : "PHY RDY changed");
+	}
 
-	printk(KERN_WARNING "ata%u: handling error/timeout\n", ap->id);
+	if (irq_stat & PORT_IRQ_UNK_FIS) {
+		u32 *unk = (u32 *)(pp->rx_fis + RX_FIS_UNK);
 
-	spin_lock_irqsave(&host_set->lock, flags);
+		err_mask |= AC_ERR_HSM;
+		action |= ATA_EH_SOFTRESET;
+		ata_ehi_push_desc(ehi, ", unknown FIS %08x %08x %08x %08x",
+				  unk[0], unk[1], unk[2], unk[3]);
+	}
 
-	ahci_restart_port(ap, readl(port_mmio + PORT_IRQ_STAT));
-	qc = ata_qc_from_tag(ap, ap->active_tag);
-	qc->err_mask |= AC_ERR_TIMEOUT;
+	/* okay, let's hand over to EH */
+	ehi->serror |= serror;
+	ehi->action |= action;
 
-	spin_unlock_irqrestore(&host_set->lock, flags);
+	qc = ata_qc_from_tag(ap, ap->active_tag);
+	if (qc)
+		qc->err_mask |= err_mask;
+	else
+		ehi->err_mask |= err_mask;
 
-	ata_eh_qc_complete(qc);
+	if (irq_stat & PORT_IRQ_FREEZE)
+		ata_port_freeze(ap);
+	else
+		ata_port_abort(ap);
 }
 
-static inline int ahci_host_intr(struct ata_port *ap, struct ata_queued_cmd *qc)
+static void ahci_host_intr(struct ata_port *ap)
 {
 	void __iomem *mmio = ap->host_set->mmio_base;
 	void __iomem *port_mmio = ahci_port_base(mmio, ap->port_no);
-	u32 status, serr, ci;
-
-	serr = readl(port_mmio + PORT_SCR_ERR);
-	writel(serr, port_mmio + PORT_SCR_ERR);
+	struct ata_eh_info *ehi = &ap->eh_info;
+	u32 status, qc_active;
+	int rc;
 
 	status = readl(port_mmio + PORT_IRQ_STAT);
 	writel(status, port_mmio + PORT_IRQ_STAT);
 
-	ci = readl(port_mmio + PORT_CMD_ISSUE);
-	if (likely((ci & 0x1) == 0)) {
-		if (qc) {
-			WARN_ON(qc->err_mask);
-			ata_qc_complete(qc);
-			qc = NULL;
-		}
+	if (unlikely(status & PORT_IRQ_ERROR)) {
+		ahci_error_intr(ap, status);
+		return;
 	}
 
-	if (status & PORT_IRQ_FATAL) {
-		unsigned int err_mask;
-		if (status & PORT_IRQ_TF_ERR)
-			err_mask = AC_ERR_DEV;
-		else if (status & PORT_IRQ_IF_ERR)
-			err_mask = AC_ERR_ATA_BUS;
-		else
-			err_mask = AC_ERR_HOST_BUS;
-
-		/* command processing has stopped due to error; restart */
-		ahci_restart_port(ap, status);
-
-		if (qc) {
-			qc->err_mask |= err_mask;
-			ata_qc_complete(qc);
-		}
+	if (ap->sactive)
+		qc_active = readl(port_mmio + PORT_SCR_ACT);
+	else
+		qc_active = readl(port_mmio + PORT_CMD_ISSUE);
+
+	rc = ata_qc_complete_multiple(ap, qc_active, NULL);
+	if (rc > 0)
+		return;
+	if (rc < 0) {
+		ehi->err_mask |= AC_ERR_HSM;
+		ehi->action |= ATA_EH_SOFTRESET;
+		ata_port_freeze(ap);
+		return;
+	}
+
+	/* hmmm... a spurious interupt */
+
+	/* some devices send D2H reg with I bit set during NCQ command phase */
+	if (ap->sactive && status & PORT_IRQ_D2H_REG_FIS)
+		return;
+
+	/* ignore interim PIO setup fis interrupts */
+	if (ata_tag_valid(ap->active_tag)) {
+		struct ata_queued_cmd *qc =
+			ata_qc_from_tag(ap, ap->active_tag);
+
+		if (qc && qc->tf.protocol == ATA_PROT_PIO &&
+		    (status & PORT_IRQ_PIOS_FIS))
+			return;
 	}
 
-	return 1;
+	if (ata_ratelimit())
+		ata_port_printk(ap, KERN_INFO, "spurious interrupt "
+				"(irq_stat 0x%x active_tag %d sactive 0x%x)\n",
+				status, ap->active_tag, ap->sactive);
 }
 
 static void ahci_irq_clear(struct ata_port *ap)
@@ -900,7 +927,7 @@ static void ahci_irq_clear(struct ata_port *ap)
 	/* TODO */
 }
 
-static irqreturn_t ahci_interrupt (int irq, void *dev_instance, struct pt_regs *regs)
+static irqreturn_t ahci_interrupt(int irq, void *dev_instance, struct pt_regs *regs)
 {
 	struct ata_host_set *host_set = dev_instance;
 	struct ahci_host_priv *hpriv;
@@ -929,14 +956,7 @@ static irqreturn_t ahci_interrupt (int irq, void *dev_instance, struct pt_regs *
 
 		ap = host_set->ports[i];
 		if (ap) {
-			struct ata_queued_cmd *qc;
-			qc = ata_qc_from_tag(ap, ap->active_tag);
-			if (!ahci_host_intr(ap, qc))
-				if (ata_ratelimit())
-					dev_printk(KERN_WARNING, host_set->dev,
-					  "unhandled interrupt on port %u\n",
-					  i);
-
+			ahci_host_intr(ap);
 			VPRINTK("port %u\n", i);
 		} else {
 			VPRINTK("port %u (no irq)\n", i);
@@ -953,7 +973,7 @@ static irqreturn_t ahci_interrupt (int irq, void *dev_instance, struct pt_regs *
 		handled = 1;
 	}
 
-        spin_unlock(&host_set->lock);
+	spin_unlock(&host_set->lock);
 
 	VPRINTK("EXIT\n");
 
@@ -965,12 +985,64 @@ static unsigned int ahci_qc_issue(struct ata_queued_cmd *qc)
 	struct ata_port *ap = qc->ap;
 	void __iomem *port_mmio = (void __iomem *) ap->ioaddr.cmd_addr;
 
-	writel(1, port_mmio + PORT_CMD_ISSUE);
+	if (qc->tf.protocol == ATA_PROT_NCQ)
+		writel(1 << qc->tag, port_mmio + PORT_SCR_ACT);
+	writel(1 << qc->tag, port_mmio + PORT_CMD_ISSUE);
 	readl(port_mmio + PORT_CMD_ISSUE);	/* flush */
 
 	return 0;
 }
 
+static void ahci_freeze(struct ata_port *ap)
+{
+	void __iomem *mmio = ap->host_set->mmio_base;
+	void __iomem *port_mmio = ahci_port_base(mmio, ap->port_no);
+
+	/* turn IRQ off */
+	writel(0, port_mmio + PORT_IRQ_MASK);
+}
+
+static void ahci_thaw(struct ata_port *ap)
+{
+	void __iomem *mmio = ap->host_set->mmio_base;
+	void __iomem *port_mmio = ahci_port_base(mmio, ap->port_no);
+	u32 tmp;
+
+	/* clear IRQ */
+	tmp = readl(port_mmio + PORT_IRQ_STAT);
+	writel(tmp, port_mmio + PORT_IRQ_STAT);
+	writel(1 << ap->id, mmio + HOST_IRQ_STAT);
+
+	/* turn IRQ back on */
+	writel(DEF_PORT_IRQ, port_mmio + PORT_IRQ_MASK);
+}
+
+static void ahci_error_handler(struct ata_port *ap)
+{
+	if (!(ap->flags & ATA_FLAG_FROZEN)) {
+		/* restart engine */
+		ahci_stop_engine(ap);
+		ahci_start_engine(ap);
+	}
+
+	/* perform recovery */
+	ata_do_eh(ap, ahci_softreset, ahci_hardreset, ahci_postreset);
+}
+
+static void ahci_post_internal_cmd(struct ata_queued_cmd *qc)
+{
+	struct ata_port *ap = qc->ap;
+
+	if (qc->flags & ATA_QCFLAG_FAILED)
+		qc->err_mask |= AC_ERR_OTHER;
+
+	if (qc->err_mask) {
+		/* make DMA engine forget about the failed command */
+		ahci_stop_engine(ap);
+		ahci_start_engine(ap);
+	}
+}
+
 static void ahci_setup_port(struct ata_ioports *port, unsigned long base,
 			    unsigned int port_idx)
 {
@@ -1115,9 +1187,6 @@ static int ahci_host_init(struct ata_probe_ent *probe_ent)
 			writel(tmp, port_mmio + PORT_IRQ_STAT);
 
 		writel(1 << i, mmio + HOST_IRQ_STAT);
-
-		/* set irq mask (enables interrupts) */
-		writel(DEF_PORT_IRQ, port_mmio + PORT_IRQ_MASK);
 	}
 
 	tmp = readl(mmio + HOST_CTL);
@@ -1215,6 +1284,8 @@ static int ahci_init_one (struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	VPRINTK("ENTER\n");
 
+	WARN_ON(ATA_MAX_QUEUE > AHCI_MAX_CMDS);
+
 	if (!printed_version++)
 		dev_printk(KERN_DEBUG, &pdev->dev, "version " DRV_VERSION "\n");
 
@@ -1282,6 +1353,9 @@ static int ahci_init_one (struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (rc)
 		goto err_out_hpriv;
 
+	if (hpriv->cap & HOST_CAP_NCQ)
+		probe_ent->host_flags |= ATA_FLAG_NCQ;
+
 	ahci_print_info(probe_ent);
 
 	/* FIXME: check ata_device_add return value */

drivers/scsi/ata_piix.c

@@ -243,7 +243,10 @@ static const struct ata_port_operations piix_pata_ops = {
 	.qc_prep		= ata_qc_prep,
 	.qc_issue		= ata_qc_issue_prot,
 
-	.eng_timeout		= ata_eng_timeout,
+	.freeze			= ata_bmdma_freeze,
+	.thaw			= ata_bmdma_thaw,
+	.error_handler		= ata_bmdma_error_handler,
+	.post_internal_cmd	= ata_bmdma_post_internal_cmd,
 
 	.irq_handler		= ata_interrupt,
 	.irq_clear		= ata_bmdma_irq_clear,
@@ -271,7 +274,10 @@ static const struct ata_port_operations piix_sata_ops = {
 	.qc_prep		= ata_qc_prep,
 	.qc_issue		= ata_qc_issue_prot,
 
-	.eng_timeout		= ata_eng_timeout,
+	.freeze			= ata_bmdma_freeze,
+	.thaw			= ata_bmdma_thaw,
+	.error_handler		= ata_bmdma_error_handler,
+	.post_internal_cmd	= ata_bmdma_post_internal_cmd,
 
 	.irq_handler		= ata_interrupt,
 	.irq_clear		= ata_bmdma_irq_clear,
@@ -484,7 +490,7 @@ static int piix_pata_probe_reset(struct ata_port *ap, unsigned int *classes)
 	struct pci_dev *pdev = to_pci_dev(ap->host_set->dev);
 
 	if (!pci_test_config_bits(pdev, &piix_enable_bits[ap->hard_port_no])) {
-		printk(KERN_INFO "ata%u: port disabled. ignoring.\n", ap->id);
+		ata_port_printk(ap, KERN_INFO, "port disabled. ignoring.\n");
 		return 0;
 	}
 
@@ -565,7 +571,7 @@ static unsigned int piix_sata_probe (struct ata_port *ap)
 static int piix_sata_probe_reset(struct ata_port *ap, unsigned int *classes)
 {
 	if (!piix_sata_probe(ap)) {
-		printk(KERN_INFO "ata%u: SATA port has no device.\n", ap->id);
+		ata_port_printk(ap, KERN_INFO, "SATA port has no device.\n");
 		return 0;
 	}
 

drivers/scsi/libata-bmdma.c

@@ -652,6 +652,149 @@ void ata_bmdma_stop(struct ata_queued_cmd *qc)
 	ata_altstatus(ap);        /* dummy read */
 }
 
+/**
+ *	ata_bmdma_freeze - Freeze BMDMA controller port
+ *	@ap: port to freeze
+ *
+ *	Freeze BMDMA controller port.
+ *
+ *	LOCKING:
+ *	Inherited from caller.
+ */
+void ata_bmdma_freeze(struct ata_port *ap)
+{
+	struct ata_ioports *ioaddr = &ap->ioaddr;
+
+	ap->ctl |= ATA_NIEN;
+	ap->last_ctl = ap->ctl;
+
+	if (ap->flags & ATA_FLAG_MMIO)
+		writeb(ap->ctl, (void __iomem *)ioaddr->ctl_addr);
+	else
+		outb(ap->ctl, ioaddr->ctl_addr);
+}
+
+/**
+ *	ata_bmdma_thaw - Thaw BMDMA controller port
+ *	@ap: port to thaw
+ *
+ *	Thaw BMDMA controller port.
+ *
+ *	LOCKING:
+ *	Inherited from caller.
+ */
+void ata_bmdma_thaw(struct ata_port *ap)
+{
+	/* clear & re-enable interrupts */
+	ata_chk_status(ap);
+	ap->ops->irq_clear(ap);
+	if (ap->ioaddr.ctl_addr)	/* FIXME: hack. create a hook instead */
+		ata_irq_on(ap);
+}
+
+/**
+ *	ata_bmdma_drive_eh - Perform EH with given methods for BMDMA controller
+ *	@ap: port to handle error for
+ *	@softreset: softreset method (can be NULL)
+ *	@hardreset: hardreset method (can be NULL)
+ *	@postreset: postreset method (can be NULL)
+ *
+ *	Handle error for ATA BMDMA controller.  It can handle both
+ *	PATA and SATA controllers.  Many controllers should be able to
+ *	use this EH as-is or with some added handling before and
+ *	after.
+ *
+ *	This function is intended to be used for constructing
+ *	->error_handler callback by low level drivers.
+ *
+ *	LOCKING:
+ *	Kernel thread context (may sleep)
+ */
+void ata_bmdma_drive_eh(struct ata_port *ap, ata_reset_fn_t softreset,
+			ata_reset_fn_t hardreset, ata_postreset_fn_t postreset)
+{
+	struct ata_host_set *host_set = ap->host_set;
+	struct ata_eh_context *ehc = &ap->eh_context;
+	struct ata_queued_cmd *qc;
+	unsigned long flags;
+	int thaw = 0;
+
+	qc = __ata_qc_from_tag(ap, ap->active_tag);
+	if (qc && !(qc->flags & ATA_QCFLAG_FAILED))
+		qc = NULL;
+
+	/* reset PIO HSM and stop DMA engine */
+	spin_lock_irqsave(&host_set->lock, flags);
+
+	ap->hsm_task_state = HSM_ST_IDLE;
+
+	if (qc && (qc->tf.protocol == ATA_PROT_DMA ||
+		   qc->tf.protocol == ATA_PROT_ATAPI_DMA)) {
+		u8 host_stat;
+
+		host_stat = ata_bmdma_status(ap);
+
+		ata_ehi_push_desc(&ehc->i, "BMDMA stat 0x%x", host_stat);
+
+		/* BMDMA controllers indicate host bus error by
+		 * setting DMA_ERR bit and timing out.  As it wasn't
+		 * really a timeout event, adjust error mask and
+		 * cancel frozen state.
+		 */
+		if (qc->err_mask == AC_ERR_TIMEOUT && host_stat & ATA_DMA_ERR) {
+			qc->err_mask = AC_ERR_HOST_BUS;
+			thaw = 1;
+		}
+
+		ap->ops->bmdma_stop(qc);
+	}
+
+	ata_altstatus(ap);
+	ata_chk_status(ap);
+	ap->ops->irq_clear(ap);
+
+	spin_unlock_irqrestore(&host_set->lock, flags);
+
+	if (thaw)
+		ata_eh_thaw_port(ap);
+
+	/* PIO and DMA engines have been stopped, perform recovery */
+	ata_do_eh(ap, softreset, hardreset, postreset);
+}
+
+/**
+ *	ata_bmdma_error_handler - Stock error handler for BMDMA controller
+ *	@ap: port to handle error for
+ *
+ *	Stock error handler for BMDMA controller.
+ *
+ *	LOCKING:
+ *	Kernel thread context (may sleep)
+ */
+void ata_bmdma_error_handler(struct ata_port *ap)
+{
+	ata_reset_fn_t hardreset;
+
+	hardreset = NULL;
+	if (sata_scr_valid(ap))
+		hardreset = sata_std_hardreset;
+
+	ata_bmdma_drive_eh(ap, ata_std_softreset, hardreset, ata_std_postreset);
+}
+
+/**
+ *	ata_bmdma_post_internal_cmd - Stock post_internal_cmd for
+ *				      BMDMA controller
+ *	@qc: internal command to clean up
+ *
+ *	LOCKING:
+ *	Kernel thread context (may sleep)
+ */
+void ata_bmdma_post_internal_cmd(struct ata_queued_cmd *qc)
+{
+	ata_bmdma_stop(qc);
+}
+
 #ifdef CONFIG_PCI
 static struct ata_probe_ent *
 ata_probe_ent_alloc(struct device *dev, const struct ata_port_info *port)

drivers/scsi/libata-eh.c

@@ -44,6 +44,53 @@
 
 #include "libata.h"
 
+static void __ata_port_freeze(struct ata_port *ap);
+
+static void ata_ering_record(struct ata_ering *ering, int is_io,
+			     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];
+	ent->is_io = is_io;
+	ent->err_mask = err_mask;
+	ent->timestamp = get_jiffies_64();
+}
+
+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 NULL;
+	return ent;
+}
+
+static int ata_ering_map(struct ata_ering *ering,
+			 int (*map_fn)(struct ata_ering_entry *, void *),
+			 void *arg)
+{
+	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;
+}
+
 /**
  *	ata_scsi_timed_out - SCSI layer time out callback
  *	@cmd: timed out SCSI command
@@ -55,6 +102,8 @@
  *	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
  *
@@ -67,10 +116,16 @@ enum scsi_eh_timer_return ata_scsi_timed_out(struct scsi_cmnd *cmd)
 	struct ata_port *ap = ata_shost_to_port(host);
 	unsigned long flags;
 	struct ata_queued_cmd *qc;
-	enum scsi_eh_timer_return ret = EH_HANDLED;
+	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) {
@@ -81,6 +136,7 @@ enum scsi_eh_timer_return ata_scsi_timed_out(struct scsi_cmnd *cmd)
 	}
 	spin_unlock_irqrestore(&ap->host_set->lock, flags);
 
+ out:
 	DPRINTK("EXIT, ret=%d\n", ret);
 	return ret;
 }
@@ -100,21 +156,141 @@ enum scsi_eh_timer_return ata_scsi_timed_out(struct scsi_cmnd *cmd)
 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 IRQ handler and port task */
-	spin_unlock_wait(&ap->host_set->lock);
+	/* synchronize with port task */
 	ata_port_flush_task(ap);
 
-	WARN_ON(ata_qc_from_tag(ap, ap->active_tag) == NULL);
+	/* 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) {
+		/* fetch & clear EH info */
+		spin_lock_irqsave(hs_lock, flags);
 
-	ap->ops->eng_timeout(ap);
+		memset(&ap->eh_context, 0, sizeof(ap->eh_context));
+		ap->eh_context.i = ap->eh_info;
+		memset(&ap->eh_info, 0, sizeof(ap->eh_info));
 
+		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);
+		}
+
+		/* this run is complete, make sure EH info is clear */
+		memset(&ap->eh_info, 0, sizeof(ap->eh_info));
+
+		/* 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");
 }
 
@@ -133,6 +309,8 @@ void ata_scsi_error(struct Scsi_Host *host)
  *	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)
  */
@@ -167,11 +345,12 @@ static void ata_qc_timeout(struct ata_queued_cmd *qc)
 		/* ack bmdma irq events */
 		ap->ops->irq_clear(ap);
 
-		printk(KERN_ERR "ata%u: command 0x%x timeout, stat 0x%x host_stat 0x%x\n",
-		       ap->id, qc->tf.command, drv_stat, host_stat);
+		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);
+		qc->err_mask |= AC_ERR_TIMEOUT;
 		break;
 	}
 
@@ -197,6 +376,8 @@ static void ata_qc_timeout(struct ata_queued_cmd *qc)
  *	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)
  */
@@ -209,6 +390,190 @@ void ata_eng_timeout(struct ata_port *ap)
 	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 */
@@ -261,3 +626,933 @@ void ata_eh_qc_retry(struct ata_queued_cmd *qc)
 		scmd->retries--;
 	__ata_eh_qc_complete(qc);
 }
+
+/**
+ *	ata_eh_about_to_do - about to perform eh_action
+ *	@ap: target ATA port
+ *	@action: action about to be performed
+ *
+ *	Called just before performing EH actions to clear related bits
+ *	in @ap->eh_info such that eh actions are not unnecessarily
+ *	repeated.
+ *
+ *	LOCKING:
+ *	None.
+ */
+static void ata_eh_about_to_do(struct ata_port *ap, unsigned int action)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&ap->host_set->lock, flags);
+	ap->eh_info.action &= ~action;
+	ap->flags |= ATA_FLAG_RECOVERED;
+	spin_unlock_irqrestore(&ap->host_set->lock, flags);
+}
+
+/**
+ *	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
+ */
+static const char * ata_err_string(unsigned int err_mask)
+{
+	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";
+}
+
+/**
+ *	ata_read_log_page - read a specific log page
+ *	@dev: target device
+ *	@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.
+ */
+static unsigned int ata_read_log_page(struct ata_device *dev,
+				      u8 page, void *buf, unsigned int sectors)
+{
+	struct ata_taskfile tf;
+	unsigned int err_mask;
+
+	DPRINTK("read log page - page %d\n", page);
+
+	ata_tf_init(dev, &tf);
+	tf.command = ATA_CMD_READ_LOG_EXT;
+	tf.lbal = page;
+	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,
+				     buf, sectors * ATA_SECT_SIZE);
+
+	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)
+{
+	u8 *buf = dev->ap->sector_buf;
+	unsigned int err_mask;
+	u8 csum;
+	int i;
+
+	err_mask = ata_read_log_page(dev, ATA_LOG_SATA_NCQ, buf, 1);
+	if (err_mask)
+		return -EIO;
+
+	csum = 0;
+	for (i = 0; i < ATA_SECT_SIZE; i++)
+		csum += buf[i];
+	if (csum)
+		ata_dev_printk(dev, KERN_WARNING,
+			       "invalid checksum 0x%x on log page 10h\n", csum);
+
+	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;
+}
+
+/**
+ *	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)
+ *
+ *	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
+ */
+static unsigned int atapi_eh_request_sense(struct ata_device *dev,
+					   unsigned char *sense_buf)
+{
+	struct ata_port *ap = dev->ap;
+	struct ata_taskfile tf;
+	u8 cdb[ATAPI_CDB_LEN];
+
+	DPRINTK("ATAPI request sense\n");
+
+	ata_tf_init(dev, &tf);
+
+	/* FIXME: is this needed? */
+	memset(sense_buf, 0, SCSI_SENSE_BUFFERSIZE);
+
+	/* XXX: why tf_read here? */
+	ap->ops->tf_read(ap, &tf);
+
+	/* fill these in, for the case where they are -not- overwritten */
+	sense_buf[0] = 0x70;
+	sense_buf[2] = tf.feature >> 4;
+
+	memset(cdb, 0, ATAPI_CDB_LEN);
+	cdb[0] = REQUEST_SENSE;
+	cdb[4] = SCSI_SENSE_BUFFERSIZE;
+
+	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) {
+		tf.protocol = ATA_PROT_ATAPI_DMA;
+		tf.feature |= ATAPI_PKT_DMA;
+	} else {
+		tf.protocol = ATA_PROT_ATAPI;
+		tf.lbam = (8 * 1024) & 0xff;
+		tf.lbah = (8 * 1024) >> 8;
+	}
+
+	return ata_exec_internal(dev, &tf, cdb, DMA_FROM_DEVICE,
+				 sense_buf, SCSI_SENSE_BUFFERSIZE);
+}
+
+/**
+ *	ata_eh_analyze_serror - analyze SError for a failed port
+ *	@ap: ATA port to analyze SError for
+ *
+ *	Analyze SError if available and further determine cause of
+ *	failure.
+ *
+ *	LOCKING:
+ *	None.
+ */
+static void ata_eh_analyze_serror(struct ata_port *ap)
+{
+	struct ata_eh_context *ehc = &ap->eh_context;
+	u32 serror = ehc->i.serror;
+	unsigned int err_mask = 0, action = 0;
+
+	if (serror & SERR_PERSISTENT) {
+		err_mask |= AC_ERR_ATA_BUS;
+		action |= ATA_EH_HARDRESET;
+	}
+	if (serror &
+	    (SERR_DATA_RECOVERED | SERR_COMM_RECOVERED | SERR_DATA)) {
+		err_mask |= AC_ERR_ATA_BUS;
+		action |= ATA_EH_SOFTRESET;
+	}
+	if (serror & SERR_PROTOCOL) {
+		err_mask |= AC_ERR_HSM;
+		action |= ATA_EH_SOFTRESET;
+	}
+	if (serror & SERR_INTERNAL) {
+		err_mask |= AC_ERR_SYSTEM;
+		action |= ATA_EH_SOFTRESET;
+	}
+	if (serror & (SERR_PHYRDY_CHG | SERR_DEV_XCHG)) {
+		err_mask |= AC_ERR_ATA_BUS;
+		action |= ATA_EH_HARDRESET;
+	}
+
+	ehc->i.err_mask |= err_mask;
+	ehc->i.action |= action;
+}
+
+/**
+ *	ata_eh_analyze_ncq_error - analyze NCQ error
+ *	@ap: ATA port to analyze NCQ error for
+ *
+ *	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).
+ */
+static void ata_eh_analyze_ncq_error(struct ata_port *ap)
+{
+	struct ata_eh_context *ehc = &ap->eh_context;
+	struct ata_device *dev = ap->device;
+	struct ata_queued_cmd *qc;
+	struct ata_taskfile tf;
+	int tag, rc;
+
+	/* if frozen, we can't do much */
+	if (ap->flags & ATA_FLAG_FROZEN)
+		return;
+
+	/* is it NCQ device error? */
+	if (!ap->sactive || !(ehc->i.err_mask & AC_ERR_DEV))
+		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 */
+	rc = ata_eh_read_log_10h(dev, &tag, &tf);
+	if (rc) {
+		ata_port_printk(ap, KERN_ERR, "failed to read log page 10h "
+				"(errno=%d)\n", rc);
+		return;
+	}
+
+	if (!(ap->sactive & (1 << tag))) {
+		ata_port_printk(ap, KERN_ERR, "log page 10h reported "
+				"inactive tag %d\n", tag);
+		return;
+	}
+
+	/* we've got the perpetrator, condemn it */
+	qc = __ata_qc_from_tag(ap, tag);
+	memcpy(&qc->result_tf, &tf, sizeof(tf));
+	qc->err_mask |= AC_ERR_DEV;
+	ehc->i.err_mask &= ~AC_ERR_DEV;
+}
+
+/**
+ *	ata_eh_analyze_tf - analyze taskfile of a failed qc
+ *	@qc: qc to analyze
+ *	@tf: Taskfile registers to analyze
+ *
+ *	Analyze taskfile of @qc and further determine cause of
+ *	failure.  This function also requests ATAPI sense data if
+ *	avaliable.
+ *
+ *	LOCKING:
+ *	Kernel thread context (may sleep).
+ *
+ *	RETURNS:
+ *	Determined recovery action
+ */
+static unsigned int ata_eh_analyze_tf(struct ata_queued_cmd *qc,
+				      const struct ata_taskfile *tf)
+{
+	unsigned int tmp, action = 0;
+	u8 stat = tf->command, err = tf->feature;
+
+	if ((stat & (ATA_BUSY | ATA_DRQ | ATA_DRDY)) != ATA_DRDY) {
+		qc->err_mask |= AC_ERR_HSM;
+		return ATA_EH_SOFTRESET;
+	}
+
+	if (!(qc->err_mask & AC_ERR_DEV))
+		return 0;
+
+	switch (qc->dev->class) {
+	case ATA_DEV_ATA:
+		if (err & ATA_ICRC)
+			qc->err_mask |= AC_ERR_ATA_BUS;
+		if (err & ATA_UNC)
+			qc->err_mask |= AC_ERR_MEDIA;
+		if (err & ATA_IDNF)
+			qc->err_mask |= AC_ERR_INVALID;
+		break;
+
+	case ATA_DEV_ATAPI:
+		tmp = atapi_eh_request_sense(qc->dev,
+					     qc->scsicmd->sense_buffer);
+		if (!tmp) {
+			/* ATA_QCFLAG_SENSE_VALID is used to tell
+			 * atapi_qc_complete() that sense data is
+			 * already valid.
+			 *
+			 * TODO: interpret sense data and set
+			 * appropriate err_mask.
+			 */
+			qc->flags |= ATA_QCFLAG_SENSE_VALID;
+		} else
+			qc->err_mask |= tmp;
+	}
+
+	if (qc->err_mask & (AC_ERR_HSM | AC_ERR_TIMEOUT | AC_ERR_ATA_BUS))
+		action |= ATA_EH_SOFTRESET;
+
+	return action;
+}
+
+static int ata_eh_categorize_ering_entry(struct ata_ering_entry *ent)
+{
+	if (ent->err_mask & (AC_ERR_ATA_BUS | AC_ERR_TIMEOUT))
+		return 1;
+
+	if (ent->is_io) {
+		if (ent->err_mask & AC_ERR_HSM)
+			return 1;
+		if ((ent->err_mask &
+		     (AC_ERR_DEV|AC_ERR_MEDIA|AC_ERR_INVALID)) == AC_ERR_DEV)
+			return 2;
+	}
+
+	return 0;
+}
+
+struct speed_down_needed_arg {
+	u64 since;
+	int nr_errors[3];
+};
+
+static int speed_down_needed_cb(struct ata_ering_entry *ent, void *void_arg)
+{
+	struct speed_down_needed_arg *arg = void_arg;
+
+	if (ent->timestamp < arg->since)
+		return -1;
+
+	arg->nr_errors[ata_eh_categorize_ering_entry(ent)]++;
+	return 0;
+}
+
+/**
+ *	ata_eh_speed_down_needed - Determine wheter speed down is necessary
+ *	@dev: Device of interest
+ *
+ *	This function examines error ring of @dev and determines
+ *	whether speed down is necessary.  Speed down is necessary if
+ *	there have been more than 3 of Cat-1 errors or 10 of Cat-2
+ *	errors during last 15 minutes.
+ *
+ *	Cat-1 errors are ATA_BUS, TIMEOUT for any command and HSM
+ *	violation for known supported commands.
+ *
+ *	Cat-2 errors are unclassified DEV error for known supported
+ *	command.
+ *
+ *	LOCKING:
+ *	Inherited from caller.
+ *
+ *	RETURNS:
+ *	1 if speed down is necessary, 0 otherwise
+ */
+static int ata_eh_speed_down_needed(struct ata_device *dev)
+{
+	const u64 interval = 15LLU * 60 * HZ;
+	static const int err_limits[3] = { -1, 3, 10 };
+	struct speed_down_needed_arg arg;
+	struct ata_ering_entry *ent;
+	int err_cat;
+	u64 j64;
+
+	ent = ata_ering_top(&dev->ering);
+	if (!ent)
+		return 0;
+
+	err_cat = ata_eh_categorize_ering_entry(ent);
+	if (err_cat == 0)
+		return 0;
+
+	memset(&arg, 0, sizeof(arg));
+
+	j64 = get_jiffies_64();
+	if (j64 >= interval)
+		arg.since = j64 - interval;
+	else
+		arg.since = 0;
+
+	ata_ering_map(&dev->ering, speed_down_needed_cb, &arg);
+
+	return arg.nr_errors[err_cat] > err_limits[err_cat];
+}
+
+/**
+ *	ata_eh_speed_down - record error and speed down if necessary
+ *	@dev: Failed device
+ *	@is_io: Did the device fail during normal IO?
+ *	@err_mask: err_mask of the error
+ *
+ *	Record error and examine error history to determine whether
+ *	adjusting transmission speed is necessary.  It also sets
+ *	transmission limits appropriately if such adjustment is
+ *	necessary.
+ *
+ *	LOCKING:
+ *	Kernel thread context (may sleep).
+ *
+ *	RETURNS:
+ *	0 on success, -errno otherwise
+ */
+static int ata_eh_speed_down(struct ata_device *dev, int is_io,
+			     unsigned int err_mask)
+{
+	if (!err_mask)
+		return 0;
+
+	/* record error and determine whether speed down is necessary */
+	ata_ering_record(&dev->ering, is_io, err_mask);
+
+	if (!ata_eh_speed_down_needed(dev))
+		return 0;
+
+	/* speed down SATA link speed if possible */
+	if (sata_down_spd_limit(dev->ap) == 0)
+		return ATA_EH_HARDRESET;
+
+	/* lower transfer mode */
+	if (ata_down_xfermask_limit(dev, 0) == 0)
+		return ATA_EH_SOFTRESET;
+
+	ata_dev_printk(dev, KERN_ERR,
+		       "speed down requested but no transfer mode left\n");
+	return 0;
+}
+
+/**
+ *	ata_eh_autopsy - analyze error and determine recovery action
+ *	@ap: ATA port to perform autopsy on
+ *
+ *	Analyze why @ap failed and determine which recovery action is
+ *	needed.  This function also sets more detailed AC_ERR_* values
+ *	and fills sense data for ATAPI CHECK SENSE.
+ *
+ *	LOCKING:
+ *	Kernel thread context (may sleep).
+ */
+static void ata_eh_autopsy(struct ata_port *ap)
+{
+	struct ata_eh_context *ehc = &ap->eh_context;
+	unsigned int action = ehc->i.action;
+	struct ata_device *failed_dev = NULL;
+	unsigned int all_err_mask = 0;
+	int tag, is_io = 0;
+	u32 serror;
+	int rc;
+
+	DPRINTK("ENTER\n");
+
+	/* obtain and analyze SError */
+	rc = sata_scr_read(ap, SCR_ERROR, &serror);
+	if (rc == 0) {
+		ehc->i.serror |= serror;
+		ata_eh_analyze_serror(ap);
+	} else if (rc != -EOPNOTSUPP)
+		action |= ATA_EH_HARDRESET;
+
+	/* analyze NCQ failure */
+	ata_eh_analyze_ncq_error(ap);
+
+	/* any real error trumps AC_ERR_OTHER */
+	if (ehc->i.err_mask & ~AC_ERR_OTHER)
+		ehc->i.err_mask &= ~AC_ERR_OTHER;
+
+	all_err_mask |= ehc->i.err_mask;
+
+	for (tag = 0; tag < ATA_MAX_QUEUE; tag++) {
+		struct ata_queued_cmd *qc = __ata_qc_from_tag(ap, tag);
+
+		if (!(qc->flags & ATA_QCFLAG_FAILED))
+			continue;
+
+		/* inherit upper level err_mask */
+		qc->err_mask |= ehc->i.err_mask;
+
+		if (qc->err_mask & AC_ERR_TIMEOUT)
+			action |= ATA_EH_SOFTRESET;
+
+		/* analyze TF */
+		action |= ata_eh_analyze_tf(qc, &qc->result_tf);
+
+		/* DEV errors are probably spurious in case of ATA_BUS error */
+		if (qc->err_mask & AC_ERR_ATA_BUS)
+			qc->err_mask &= ~(AC_ERR_DEV | AC_ERR_MEDIA |
+					  AC_ERR_INVALID);
+
+		/* any real error trumps unknown error */
+		if (qc->err_mask & ~AC_ERR_OTHER)
+			qc->err_mask &= ~AC_ERR_OTHER;
+
+		/* SENSE_VALID trumps dev/unknown error and revalidation */
+		if (qc->flags & ATA_QCFLAG_SENSE_VALID) {
+			qc->err_mask &= ~(AC_ERR_DEV | AC_ERR_OTHER);
+			action &= ~ATA_EH_REVALIDATE;
+		}
+
+		/* accumulate error info */
+		failed_dev = qc->dev;
+		all_err_mask |= qc->err_mask;
+		if (qc->flags & ATA_QCFLAG_IO)
+			is_io = 1;
+	}
+
+	/* speed down iff command was in progress */
+	if (failed_dev)
+		action |= ata_eh_speed_down(failed_dev, is_io, all_err_mask);
+
+	if (all_err_mask)
+		action |= ATA_EH_REVALIDATE;
+
+	ehc->i.dev = failed_dev;
+	ehc->i.action = action;
+
+	DPRINTK("EXIT\n");
+}
+
+/**
+ *	ata_eh_report - report error handling to user
+ *	@ap: ATA port EH is going on
+ *
+ *	Report EH to user.
+ *
+ *	LOCKING:
+ *	None.
+ */
+static void ata_eh_report(struct ata_port *ap)
+{
+	struct ata_eh_context *ehc = &ap->eh_context;
+	const char *frozen, *desc;
+	int tag, nr_failed = 0;
+
+	desc = NULL;
+	if (ehc->i.desc[0] != '\0')
+		desc = ehc->i.desc;
+
+	for (tag = 0; tag < ATA_MAX_QUEUE; tag++) {
+		struct ata_queued_cmd *qc = __ata_qc_from_tag(ap, tag);
+
+		if (!(qc->flags & ATA_QCFLAG_FAILED))
+			continue;
+		if (qc->flags & ATA_QCFLAG_SENSE_VALID && !qc->err_mask)
+			continue;
+
+		nr_failed++;
+	}
+
+	if (!nr_failed && !ehc->i.err_mask)
+		return;
+
+	frozen = "";
+	if (ap->flags & ATA_FLAG_FROZEN)
+		frozen = " frozen";
+
+	if (ehc->i.dev) {
+		ata_dev_printk(ehc->i.dev, KERN_ERR, "exception Emask 0x%x "
+			       "SAct 0x%x SErr 0x%x action 0x%x%s\n",
+			       ehc->i.err_mask, ap->sactive, ehc->i.serror,
+			       ehc->i.action, frozen);
+		if (desc)
+			ata_dev_printk(ehc->i.dev, KERN_ERR, "(%s)\n", desc);
+	} else {
+		ata_port_printk(ap, KERN_ERR, "exception Emask 0x%x "
+				"SAct 0x%x SErr 0x%x action 0x%x%s\n",
+				ehc->i.err_mask, ap->sactive, ehc->i.serror,
+				ehc->i.action, frozen);
+		if (desc)
+			ata_port_printk(ap, KERN_ERR, "(%s)\n", desc);
+	}
+
+	for (tag = 0; tag < ATA_MAX_QUEUE; tag++) {
+		struct ata_queued_cmd *qc = __ata_qc_from_tag(ap, tag);
+
+		if (!(qc->flags & ATA_QCFLAG_FAILED) || !qc->err_mask)
+			continue;
+
+		ata_dev_printk(qc->dev, KERN_ERR, "tag %d cmd 0x%x "
+			       "Emask 0x%x stat 0x%x err 0x%x (%s)\n",
+			       qc->tag, qc->tf.command, qc->err_mask,
+			       qc->result_tf.command, qc->result_tf.feature,
+			       ata_err_string(qc->err_mask));
+	}
+}
+
+static int ata_eh_reset(struct ata_port *ap, ata_reset_fn_t softreset,
+			ata_reset_fn_t hardreset, ata_postreset_fn_t postreset)
+{
+	struct ata_eh_context *ehc = &ap->eh_context;
+	unsigned int classes[ATA_MAX_DEVICES];
+	int tries = ATA_EH_RESET_TRIES;
+	ata_reset_fn_t reset;
+	int rc;
+
+	if (softreset && (!hardreset || (!sata_set_spd_needed(ap) &&
+					 !(ehc->i.action & ATA_EH_HARDRESET))))
+		reset = softreset;
+	else
+		reset = hardreset;
+
+ retry:
+	ata_port_printk(ap, KERN_INFO, "%s resetting port\n",
+			reset == softreset ? "soft" : "hard");
+
+	/* reset */
+	ata_eh_about_to_do(ap, ATA_EH_RESET_MASK);
+	ehc->i.flags |= ATA_EHI_DID_RESET;
+
+	rc = ata_do_reset(ap, reset, classes);
+
+	if (rc && --tries) {
+		ata_port_printk(ap, KERN_WARNING,
+				"%sreset failed, retrying in 5 secs\n",
+				reset == softreset ? "soft" : "hard");
+		ssleep(5);
+
+		if (reset == hardreset)
+			sata_down_spd_limit(ap);
+		if (hardreset)
+			reset = hardreset;
+		goto retry;
+	}
+
+	if (rc == 0) {
+		if (postreset)
+			postreset(ap, classes);
+
+		/* reset successful, schedule revalidation */
+		ehc->i.dev = NULL;
+		ehc->i.action &= ~ATA_EH_RESET_MASK;
+		ehc->i.action |= ATA_EH_REVALIDATE;
+	}
+
+	return rc;
+}
+
+static int ata_eh_revalidate(struct ata_port *ap,
+			     struct ata_device **r_failed_dev)
+{
+	struct ata_eh_context *ehc = &ap->eh_context;
+	struct ata_device *dev;
+	int i, rc = 0;
+
+	DPRINTK("ENTER\n");
+
+	for (i = 0; i < ATA_MAX_DEVICES; i++) {
+		dev = &ap->device[i];
+
+		if (ehc->i.action & ATA_EH_REVALIDATE && ata_dev_enabled(dev) &&
+		    (!ehc->i.dev || ehc->i.dev == dev)) {
+			if (ata_port_offline(ap)) {
+				rc = -EIO;
+				break;
+			}
+
+			ata_eh_about_to_do(ap, ATA_EH_REVALIDATE);
+			rc = ata_dev_revalidate(dev,
+					ehc->i.flags & ATA_EHI_DID_RESET);
+			if (rc)
+				break;
+
+			ehc->i.action &= ~ATA_EH_REVALIDATE;
+		}
+	}
+
+	if (rc)
+		*r_failed_dev = dev;
+
+	DPRINTK("EXIT\n");
+	return rc;
+}
+
+static int ata_port_nr_enabled(struct ata_port *ap)
+{
+	int i, cnt = 0;
+
+	for (i = 0; i < ATA_MAX_DEVICES; i++)
+		if (ata_dev_enabled(&ap->device[i]))
+			cnt++;
+	return cnt;
+}
+
+/**
+ *	ata_eh_recover - recover host port after error
+ *	@ap: host port to recover
+ *	@softreset: softreset method (can be NULL)
+ *	@hardreset: hardreset method (can be NULL)
+ *	@postreset: postreset method (can be NULL)
+ *
+ *	This is the alpha and omega, eum and yang, heart and soul of
+ *	libata exception handling.  On entry, actions required to
+ *	recover each devices are recorded in eh_context.  This
+ *	function executes all the operations with appropriate retrials
+ *	and fallbacks to resurrect failed devices.
+ *
+ *	LOCKING:
+ *	Kernel thread context (may sleep).
+ *
+ *	RETURNS:
+ *	0 on success, -errno on failure.
+ */
+static int ata_eh_recover(struct ata_port *ap, ata_reset_fn_t softreset,
+			  ata_reset_fn_t hardreset,
+			  ata_postreset_fn_t postreset)
+{
+	struct ata_eh_context *ehc = &ap->eh_context;
+	struct ata_device *dev;
+	int down_xfermask, i, rc;
+
+	DPRINTK("ENTER\n");
+
+	/* prep for recovery */
+	for (i = 0; i < ATA_MAX_DEVICES; i++) {
+		dev = &ap->device[i];
+
+		ehc->tries[dev->devno] = ATA_EH_DEV_TRIES;
+	}
+
+ retry:
+	down_xfermask = 0;
+	rc = 0;
+
+	/* skip EH if possible. */
+	if (!ata_port_nr_enabled(ap) && !(ap->flags & ATA_FLAG_FROZEN))
+		ehc->i.action = 0;
+
+	/* reset */
+	if (ehc->i.action & ATA_EH_RESET_MASK) {
+		ata_eh_freeze_port(ap);
+
+		rc = ata_eh_reset(ap, softreset, hardreset, postreset);
+		if (rc) {
+			ata_port_printk(ap, KERN_ERR,
+					"reset failed, giving up\n");
+			goto out;
+		}
+
+		ata_eh_thaw_port(ap);
+	}
+
+	/* revalidate existing devices */
+	rc = ata_eh_revalidate(ap, &dev);
+	if (rc)
+		goto dev_fail;
+
+	/* configure transfer mode if the port has been reset */
+	if (ehc->i.flags & ATA_EHI_DID_RESET) {
+		rc = ata_set_mode(ap, &dev);
+		if (rc) {
+			down_xfermask = 1;
+			goto dev_fail;
+		}
+	}
+
+	goto out;
+
+ dev_fail:
+	switch (rc) {
+	case -ENODEV:
+	case -EINVAL:
+		ehc->tries[dev->devno] = 0;
+		break;
+	case -EIO:
+		sata_down_spd_limit(ap);
+	default:
+		ehc->tries[dev->devno]--;
+		if (down_xfermask &&
+		    ata_down_xfermask_limit(dev, ehc->tries[dev->devno] == 1))
+			ehc->tries[dev->devno] = 0;
+	}
+
+	/* disable device if it has used up all its chances */
+	if (ata_dev_enabled(dev) && !ehc->tries[dev->devno])
+		ata_dev_disable(dev);
+
+	/* soft didn't work?  be haaaaard */
+	if (ehc->i.flags & ATA_EHI_DID_RESET)
+		ehc->i.action |= ATA_EH_HARDRESET;
+	else
+		ehc->i.action |= ATA_EH_SOFTRESET;
+
+	if (ata_port_nr_enabled(ap)) {
+		ata_port_printk(ap, KERN_WARNING, "failed to recover some "
+				"devices, retrying in 5 secs\n");
+		ssleep(5);
+	} else {
+		/* no device left, repeat fast */
+		msleep(500);
+	}
+
+	goto retry;
+
+ out:
+	if (rc) {
+		for (i = 0; i < ATA_MAX_DEVICES; i++)
+			ata_dev_disable(&ap->device[i]);
+	}
+
+	DPRINTK("EXIT, rc=%d\n", rc);
+	return rc;
+}
+
+/**
+ *	ata_eh_finish - finish up EH
+ *	@ap: host port to finish EH for
+ *
+ *	Recovery is complete.  Clean up EH states and retry or finish
+ *	failed qcs.
+ *
+ *	LOCKING:
+ *	None.
+ */
+static void ata_eh_finish(struct ata_port *ap)
+{
+	int tag;
+
+	/* retry or finish qcs */
+	for (tag = 0; tag < ATA_MAX_QUEUE; tag++) {
+		struct ata_queued_cmd *qc = __ata_qc_from_tag(ap, tag);
+
+		if (!(qc->flags & ATA_QCFLAG_FAILED))
+			continue;
+
+		if (qc->err_mask) {
+			/* FIXME: Once EH migration is complete,
+			 * generate sense data in this function,
+			 * considering both err_mask and tf.
+			 */
+			if (qc->err_mask & AC_ERR_INVALID)
+				ata_eh_qc_complete(qc);
+			else
+				ata_eh_qc_retry(qc);
+		} else {
+			if (qc->flags & ATA_QCFLAG_SENSE_VALID) {
+				ata_eh_qc_complete(qc);
+			} else {
+				/* feed zero TF to sense generation */
+				memset(&qc->result_tf, 0, sizeof(qc->result_tf));
+				ata_eh_qc_retry(qc);
+			}
+		}
+	}
+}
+
+/**
+ *	ata_do_eh - do standard error handling
+ *	@ap: host port to handle error for
+ *	@softreset: softreset method (can be NULL)
+ *	@hardreset: hardreset method (can be NULL)
+ *	@postreset: postreset method (can be NULL)
+ *
+ *	Perform standard error handling sequence.
+ *
+ *	LOCKING:
+ *	Kernel thread context (may sleep).
+ */
+void ata_do_eh(struct ata_port *ap, ata_reset_fn_t softreset,
+	       ata_reset_fn_t hardreset, ata_postreset_fn_t postreset)
+{
+	ata_eh_autopsy(ap);
+	ata_eh_report(ap);
+	ata_eh_recover(ap, softreset, hardreset, postreset);
+	ata_eh_finish(ap);
+}

drivers/scsi/libata-scsi.c

@@ -41,6 +41,7 @@
 #include <scsi/scsi_eh.h>
 #include <scsi/scsi_device.h>
 #include <scsi/scsi_request.h>
+#include <scsi/scsi_tcq.h>
 #include <scsi/scsi_transport.h>
 #include <linux/libata.h>
 #include <linux/hdreg.h>
@@ -302,7 +303,6 @@ int ata_scsi_ioctl(struct scsi_device *scsidev, int cmd, void __user *arg)
 
 /**
  *	ata_scsi_qc_new - acquire new ata_queued_cmd reference
- *	@ap: ATA port to which the new command is attached
  *	@dev: ATA device to which the new command is attached
  *	@cmd: SCSI command that originated this ATA command
  *	@done: SCSI command completion function
@@ -321,14 +321,13 @@ int ata_scsi_ioctl(struct scsi_device *scsidev, int cmd, void __user *arg)
  *	RETURNS:
  *	Command allocated, or %NULL if none available.
  */
-struct ata_queued_cmd *ata_scsi_qc_new(struct ata_port *ap,
-				       struct ata_device *dev,
+struct ata_queued_cmd *ata_scsi_qc_new(struct ata_device *dev,
 				       struct scsi_cmnd *cmd,
 				       void (*done)(struct scsi_cmnd *))
 {
 	struct ata_queued_cmd *qc;
 
-	qc = ata_qc_new_init(ap, dev);
+	qc = ata_qc_new_init(dev);
 	if (qc) {
 		qc->scsicmd = cmd;
 		qc->scsidone = done;
@@ -398,7 +397,7 @@ int ata_scsi_device_resume(struct scsi_device *sdev)
 	struct ata_port *ap = ata_shost_to_port(sdev->host);
 	struct ata_device *dev = &ap->device[sdev->id];
 
-	return ata_device_resume(ap, dev);
+	return ata_device_resume(dev);
 }
 
 int ata_scsi_device_suspend(struct scsi_device *sdev, pm_message_t state)
@@ -406,7 +405,7 @@ int ata_scsi_device_suspend(struct scsi_device *sdev, pm_message_t state)
 	struct ata_port *ap = ata_shost_to_port(sdev->host);
 	struct ata_device *dev = &ap->device[sdev->id];
 
-	return ata_device_suspend(ap, dev, state);
+	return ata_device_suspend(dev, state);
 }
 
 /**
@@ -417,6 +416,7 @@ int ata_scsi_device_suspend(struct scsi_device *sdev, pm_message_t state)
  *	@sk: the sense key we'll fill out
  *	@asc: the additional sense code we'll fill out
  *	@ascq: the additional sense code qualifier we'll fill out
+ *	@verbose: be verbose
  *
  *	Converts an ATA error into a SCSI error.  Fill out pointers to
  *	SK, ASC, and ASCQ bytes for later use in fixed or descriptor
@@ -426,7 +426,7 @@ int ata_scsi_device_suspend(struct scsi_device *sdev, pm_message_t state)
  *	spin_lock_irqsave(host_set lock)
  */
 void ata_to_sense_error(unsigned id, u8 drv_stat, u8 drv_err, u8 *sk, u8 *asc,
-			u8 *ascq)
+			u8 *ascq, int verbose)
 {
 	int i;
 
@@ -491,8 +491,9 @@ void ata_to_sense_error(unsigned id, u8 drv_stat, u8 drv_err, u8 *sk, u8 *asc,
 			}
 		}
 		/* No immediate match */
-		printk(KERN_WARNING "ata%u: no sense translation for "
-		       "error 0x%02x\n", id, drv_err);
+		if (verbose)
+			printk(KERN_WARNING "ata%u: no sense translation for "
+			       "error 0x%02x\n", id, drv_err);
 	}
 
 	/* Fall back to interpreting status bits */
@@ -505,8 +506,9 @@ void ata_to_sense_error(unsigned id, u8 drv_stat, u8 drv_err, u8 *sk, u8 *asc,
 		}
 	}
 	/* No error?  Undecoded? */
-	printk(KERN_WARNING "ata%u: no sense translation for status: 0x%02x\n",
-	       id, drv_stat);
+	if (verbose)
+		printk(KERN_WARNING "ata%u: no sense translation for "
+		       "status: 0x%02x\n", id, drv_stat);
 
 	/* We need a sensible error return here, which is tricky, and one
 	   that won't cause people to do things like return a disk wrongly */
@@ -515,9 +517,10 @@ void ata_to_sense_error(unsigned id, u8 drv_stat, u8 drv_err, u8 *sk, u8 *asc,
 	*ascq = 0x00;
 
  translate_done:
-	printk(KERN_ERR "ata%u: translated ATA stat/err 0x%02x/%02x to "
-	       "SCSI SK/ASC/ASCQ 0x%x/%02x/%02x\n", id, drv_stat, drv_err,
-	       *sk, *asc, *ascq);
+	if (verbose)
+		printk(KERN_ERR "ata%u: translated ATA stat/err 0x%02x/%02x "
+		       "to SCSI SK/ASC/ASCQ 0x%x/%02x/%02x\n",
+		       id, drv_stat, drv_err, *sk, *asc, *ascq);
 	return;
 }
 
@@ -537,9 +540,10 @@ void ata_to_sense_error(unsigned id, u8 drv_stat, u8 drv_err, u8 *sk, u8 *asc,
 void ata_gen_ata_desc_sense(struct ata_queued_cmd *qc)
 {
 	struct scsi_cmnd *cmd = qc->scsicmd;
-	struct ata_taskfile *tf = &qc->tf;
+	struct ata_taskfile *tf = &qc->result_tf;
 	unsigned char *sb = cmd->sense_buffer;
 	unsigned char *desc = sb + 8;
+	int verbose = qc->ap->ops->error_handler == NULL;
 
 	memset(sb, 0, SCSI_SENSE_BUFFERSIZE);
 
@@ -552,7 +556,7 @@ void ata_gen_ata_desc_sense(struct ata_queued_cmd *qc)
 	if (qc->err_mask ||
 	    tf->command & (ATA_BUSY | ATA_DF | ATA_ERR | ATA_DRQ)) {
 		ata_to_sense_error(qc->ap->id, tf->command, tf->feature,
-				   &sb[1], &sb[2], &sb[3]);
+				   &sb[1], &sb[2], &sb[3], verbose);
 		sb[1] &= 0x0f;
 	}
 
@@ -608,8 +612,9 @@ void ata_gen_ata_desc_sense(struct ata_queued_cmd *qc)
 void ata_gen_fixed_sense(struct ata_queued_cmd *qc)
 {
 	struct scsi_cmnd *cmd = qc->scsicmd;
-	struct ata_taskfile *tf = &qc->tf;
+	struct ata_taskfile *tf = &qc->result_tf;
 	unsigned char *sb = cmd->sense_buffer;
+	int verbose = qc->ap->ops->error_handler == NULL;
 
 	memset(sb, 0, SCSI_SENSE_BUFFERSIZE);
 
@@ -622,7 +627,7 @@ void ata_gen_fixed_sense(struct ata_queued_cmd *qc)
 	if (qc->err_mask ||
 	    tf->command & (ATA_BUSY | ATA_DF | ATA_ERR | ATA_DRQ)) {
 		ata_to_sense_error(qc->ap->id, tf->command, tf->feature,
-				   &sb[2], &sb[12], &sb[13]);
+				   &sb[2], &sb[12], &sb[13], verbose);
 		sb[2] &= 0x0f;
 	}
 
@@ -680,6 +685,14 @@ static void ata_scsi_dev_config(struct scsi_device *sdev,
 		request_queue_t *q = sdev->request_queue;
 		blk_queue_max_hw_segments(q, q->max_hw_segments - 1);
 	}
+
+	if (dev->flags & ATA_DFLAG_NCQ) {
+		int depth;
+
+		depth = min(sdev->host->can_queue, ata_id_queue_depth(dev->id));
+		depth = min(ATA_MAX_QUEUE - 1, depth);
+		scsi_adjust_queue_depth(sdev, MSG_SIMPLE_TAG, depth);
+	}
 }
 
 /**
@@ -713,6 +726,43 @@ int ata_scsi_slave_config(struct scsi_device *sdev)
 	return 0;	/* scsi layer doesn't check return value, sigh */
 }
 
+/**
+ *	ata_scsi_change_queue_depth - SCSI callback for queue depth config
+ *	@sdev: SCSI device to configure queue depth for
+ *	@queue_depth: new queue depth
+ *
+ *	This is libata standard hostt->change_queue_depth callback.
+ *	SCSI will call into this callback when user tries to set queue
+ *	depth via sysfs.
+ *
+ *	LOCKING:
+ *	SCSI layer (we don't care)
+ *
+ *	RETURNS:
+ *	Newly configured queue depth.
+ */
+int ata_scsi_change_queue_depth(struct scsi_device *sdev, int queue_depth)
+{
+	struct ata_port *ap = ata_shost_to_port(sdev->host);
+	struct ata_device *dev;
+	int max_depth;
+
+	if (queue_depth < 1)
+		return sdev->queue_depth;
+
+	dev = ata_scsi_find_dev(ap, sdev);
+	if (!dev || !ata_dev_enabled(dev))
+		return sdev->queue_depth;
+
+	max_depth = min(sdev->host->can_queue, ata_id_queue_depth(dev->id));
+	max_depth = min(ATA_MAX_QUEUE - 1, max_depth);
+	if (queue_depth > max_depth)
+		queue_depth = max_depth;
+
+	scsi_adjust_queue_depth(sdev, MSG_SIMPLE_TAG, queue_depth);
+	return queue_depth;
+}
+
 /**
  *	ata_scsi_start_stop_xlat - Translate SCSI START STOP UNIT command
  *	@qc: Storage for translated ATA taskfile
@@ -748,7 +798,7 @@ static unsigned int ata_scsi_start_stop_xlat(struct ata_queued_cmd *qc,
 		tf->nsect = 1;	/* 1 sector, lba=0 */
 
 		if (qc->dev->flags & ATA_DFLAG_LBA) {
-			qc->tf.flags |= ATA_TFLAG_LBA;
+			tf->flags |= ATA_TFLAG_LBA;
 
 			tf->lbah = 0x0;
 			tf->lbam = 0x0;
@@ -1099,7 +1149,36 @@ static unsigned int ata_scsi_rw_xlat(struct ata_queued_cmd *qc, const u8 *scsicm
 		 */
 		goto nothing_to_do;
 
-	if (dev->flags & ATA_DFLAG_LBA) {
+	if ((dev->flags & (ATA_DFLAG_PIO | ATA_DFLAG_NCQ)) == ATA_DFLAG_NCQ) {
+		/* yay, NCQ */
+		if (!lba_48_ok(block, n_block))
+			goto out_of_range;
+
+		tf->protocol = ATA_PROT_NCQ;
+		tf->flags |= ATA_TFLAG_LBA | ATA_TFLAG_LBA48;
+
+		if (tf->flags & ATA_TFLAG_WRITE)
+			tf->command = ATA_CMD_FPDMA_WRITE;
+		else
+			tf->command = ATA_CMD_FPDMA_READ;
+
+		qc->nsect = n_block;
+
+		tf->nsect = qc->tag << 3;
+		tf->hob_feature = (n_block >> 8) & 0xff;
+		tf->feature = n_block & 0xff;
+
+		tf->hob_lbah = (block >> 40) & 0xff;
+		tf->hob_lbam = (block >> 32) & 0xff;
+		tf->hob_lbal = (block >> 24) & 0xff;
+		tf->lbah = (block >> 16) & 0xff;
+		tf->lbam = (block >> 8) & 0xff;
+		tf->lbal = block & 0xff;
+
+		tf->device = 1 << 6;
+		if (tf->flags & ATA_TFLAG_FUA)
+			tf->device |= 1 << 7;
+	} else if (dev->flags & ATA_DFLAG_LBA) {
 		tf->flags |= ATA_TFLAG_LBA;
 
 		if (lba_28_ok(block, n_block)) {
@@ -1199,14 +1278,11 @@ static void ata_scsi_qc_complete(struct ata_queued_cmd *qc)
 	 */
 	if (((cdb[0] == ATA_16) || (cdb[0] == ATA_12)) &&
  	    ((cdb[2] & 0x20) || need_sense)) {
-		qc->ap->ops->tf_read(qc->ap, &qc->tf);
  		ata_gen_ata_desc_sense(qc);
 	} else {
 		if (!need_sense) {
 			cmd->result = SAM_STAT_GOOD;
 		} else {
-			qc->ap->ops->tf_read(qc->ap, &qc->tf);
-
 			/* TODO: decide which descriptor format to use
 			 * for 48b LBA devices and call that here
 			 * instead of the fixed desc, which is only
@@ -1217,19 +1293,49 @@ static void ata_scsi_qc_complete(struct ata_queued_cmd *qc)
 		}
 	}
 
-	if (need_sense) {
-		/* The ata_gen_..._sense routines fill in tf */
-		ata_dump_status(qc->ap->id, &qc->tf);
-	}
+	if (need_sense && !qc->ap->ops->error_handler)
+		ata_dump_status(qc->ap->id, &qc->result_tf);
 
 	qc->scsidone(cmd);
 
 	ata_qc_free(qc);
 }
 
+/**
+ *	ata_scmd_need_defer - Check whether we need to defer scmd
+ *	@dev: ATA device to which the command is addressed
+ *	@is_io: Is the command IO (and thus possibly NCQ)?
+ *
+ *	NCQ and non-NCQ commands cannot run together.  As upper layer
+ *	only knows the queue depth, we are responsible for maintaining
+ *	exclusion.  This function checks whether a new command can be
+ *	issued to @dev.
+ *
+ *	LOCKING:
+ *	spin_lock_irqsave(host_set lock)
+ *
+ *	RETURNS:
+ *	1 if deferring is needed, 0 otherwise.
+ */
+static int ata_scmd_need_defer(struct ata_device *dev, int is_io)
+{
+	struct ata_port *ap = dev->ap;
+
+	if (!(dev->flags & ATA_DFLAG_NCQ))
+		return 0;
+
+	if (is_io) {
+		if (!ata_tag_valid(ap->active_tag))
+			return 0;
+	} else {
+		if (!ata_tag_valid(ap->active_tag) && !ap->sactive)
+			return 0;
+	}
+	return 1;
+}
+
 /**
  *	ata_scsi_translate - Translate then issue SCSI command to ATA device
- *	@ap: ATA port to which the command is addressed
  *	@dev: ATA device to which the command is addressed
  *	@cmd: SCSI command to execute
  *	@done: SCSI command completion function
@@ -1250,19 +1356,25 @@ static void ata_scsi_qc_complete(struct ata_queued_cmd *qc)
  *
  *	LOCKING:
  *	spin_lock_irqsave(host_set lock)
+ *
+ *	RETURNS:
+ *	0 on success, SCSI_ML_QUEUE_DEVICE_BUSY if the command
+ *	needs to be deferred.
  */
-
-static void ata_scsi_translate(struct ata_port *ap, struct ata_device *dev,
-			      struct scsi_cmnd *cmd,
+static int ata_scsi_translate(struct ata_device *dev, struct scsi_cmnd *cmd,
 			      void (*done)(struct scsi_cmnd *),
 			      ata_xlat_func_t xlat_func)
 {
 	struct ata_queued_cmd *qc;
 	u8 *scsicmd = cmd->cmnd;
+	int is_io = xlat_func == ata_scsi_rw_xlat;
 
 	VPRINTK("ENTER\n");
 
-	qc = ata_scsi_qc_new(ap, dev, cmd, done);
+	if (unlikely(ata_scmd_need_defer(dev, is_io)))
+		goto defer;
+
+	qc = ata_scsi_qc_new(dev, cmd, done);
 	if (!qc)
 		goto err_mem;
 
@@ -1270,8 +1382,8 @@ static void ata_scsi_translate(struct ata_port *ap, struct ata_device *dev,
 	if (cmd->sc_data_direction == DMA_FROM_DEVICE ||
 	    cmd->sc_data_direction == DMA_TO_DEVICE) {
 		if (unlikely(cmd->request_bufflen < 1)) {
-			printk(KERN_WARNING "ata%u(%u): WARNING: zero len r/w req\n",
-			       ap->id, dev->devno);
+			ata_dev_printk(dev, KERN_WARNING,
+				       "WARNING: zero len r/w req\n");
 			goto err_did;
 		}
 
@@ -1293,13 +1405,13 @@ static void ata_scsi_translate(struct ata_port *ap, struct ata_device *dev,
 	ata_qc_issue(qc);
 
 	VPRINTK("EXIT\n");
-	return;
+	return 0;
 
 early_finish:
         ata_qc_free(qc);
 	done(cmd);
 	DPRINTK("EXIT - early finish (good or error)\n");
-	return;
+	return 0;
 
 err_did:
 	ata_qc_free(qc);
@@ -1307,7 +1419,11 @@ err_mem:
 	cmd->result = (DID_ERROR << 16);
 	done(cmd);
 	DPRINTK("EXIT - internal\n");
-	return;
+	return 0;
+
+defer:
+	DPRINTK("EXIT - defer\n");
+	return SCSI_MLQUEUE_DEVICE_BUSY;
 }
 
 /**
@@ -2004,7 +2120,6 @@ static void atapi_sense_complete(struct ata_queued_cmd *qc)
 		 * a sense descriptors, since that's only
 		 * correct for ATA, not ATAPI
 		 */
-		qc->ap->ops->tf_read(qc->ap, &qc->tf);
 		ata_gen_ata_desc_sense(qc);
 	}
 
@@ -2070,6 +2185,26 @@ static void atapi_qc_complete(struct ata_queued_cmd *qc)
 
 	VPRINTK("ENTER, err_mask 0x%X\n", err_mask);
 
+	/* handle completion from new EH */
+	if (unlikely(qc->ap->ops->error_handler &&
+		     (err_mask || qc->flags & ATA_QCFLAG_SENSE_VALID))) {
+
+		if (!(qc->flags & ATA_QCFLAG_SENSE_VALID)) {
+			/* FIXME: not quite right; we don't want the
+			 * translation of taskfile registers into a
+			 * sense descriptors, since that's only
+			 * correct for ATA, not ATAPI
+			 */
+			ata_gen_ata_desc_sense(qc);
+		}
+
+		qc->scsicmd->result = SAM_STAT_CHECK_CONDITION;
+		qc->scsidone(cmd);
+		ata_qc_free(qc);
+		return;
+	}
+
+	/* successful completion or old EH failure path */
 	if (unlikely(err_mask & AC_ERR_DEV)) {
 		cmd->result = SAM_STAT_CHECK_CONDITION;
 		atapi_request_sense(qc);
@@ -2080,7 +2215,6 @@ static void atapi_qc_complete(struct ata_queued_cmd *qc)
 		 * a sense descriptors, since that's only
 		 * correct for ATA, not ATAPI
 		 */
-		qc->ap->ops->tf_read(qc->ap, &qc->tf);
 		ata_gen_ata_desc_sense(qc);
 	} else {
 		u8 *scsicmd = cmd->cmnd;
@@ -2211,8 +2345,9 @@ ata_scsi_find_dev(struct ata_port *ap, const struct scsi_device *scsidev)
 
 	if (!atapi_enabled || (ap->flags & ATA_FLAG_NO_ATAPI)) {
 		if (unlikely(dev->class == ATA_DEV_ATAPI)) {
-			printk(KERN_WARNING "ata%u(%u): WARNING: ATAPI is %s, device ignored.\n",
-			       ap->id, dev->devno, atapi_enabled ? "not supported with this driver" : "disabled");
+			ata_dev_printk(dev, KERN_WARNING,
+				"WARNING: ATAPI is %s, device ignored.\n",
+				atapi_enabled ? "not supported with this driver" : "disabled");
 			return NULL;
 		}
 	}
@@ -2361,6 +2496,9 @@ ata_scsi_pass_thru(struct ata_queued_cmd *qc, const u8 *scsicmd)
 	 */
 	qc->nsect = cmd->bufflen / ATA_SECT_SIZE;
 
+	/* request result TF */
+	qc->flags |= ATA_QCFLAG_RESULT_TF;
+
 	return 0;
 
  invalid_fld:
@@ -2437,19 +2575,24 @@ static inline void ata_scsi_dump_cdb(struct ata_port *ap,
 #endif
 }
 
-static inline void __ata_scsi_queuecmd(struct scsi_cmnd *cmd, void (*done)(struct scsi_cmnd *),
-				       struct ata_port *ap, struct ata_device *dev)
+static inline int __ata_scsi_queuecmd(struct scsi_cmnd *cmd,
+				      void (*done)(struct scsi_cmnd *),
+				      struct ata_device *dev)
 {
+	int rc = 0;
+
 	if (dev->class == ATA_DEV_ATA) {
 		ata_xlat_func_t xlat_func = ata_get_xlat_func(dev,
 							      cmd->cmnd[0]);
 
 		if (xlat_func)
-			ata_scsi_translate(ap, dev, cmd, done, xlat_func);
+			rc = ata_scsi_translate(dev, cmd, done, xlat_func);
 		else
-			ata_scsi_simulate(ap, dev, cmd, done);
+			ata_scsi_simulate(dev, cmd, done);
 	} else
-		ata_scsi_translate(ap, dev, cmd, done, atapi_xlat);
+		rc = ata_scsi_translate(dev, cmd, done, atapi_xlat);
+
+	return rc;
 }
 
 /**
@@ -2468,15 +2611,16 @@ static inline void __ata_scsi_queuecmd(struct scsi_cmnd *cmd, void (*done)(struc
  *	Releases scsi-layer-held lock, and obtains host_set lock.
  *
  *	RETURNS:
- *	Zero.
+ *	Return value from __ata_scsi_queuecmd() if @cmd can be queued,
+ *	0 otherwise.
  */
-
 int ata_scsi_queuecmd(struct scsi_cmnd *cmd, void (*done)(struct scsi_cmnd *))
 {
 	struct ata_port *ap;
 	struct ata_device *dev;
 	struct scsi_device *scsidev = cmd->device;
 	struct Scsi_Host *shost = scsidev->host;
+	int rc = 0;
 
 	ap = ata_shost_to_port(shost);
 
@@ -2487,7 +2631,7 @@ int ata_scsi_queuecmd(struct scsi_cmnd *cmd, void (*done)(struct scsi_cmnd *))
 
 	dev = ata_scsi_find_dev(ap, scsidev);
 	if (likely(dev))
-		__ata_scsi_queuecmd(cmd, done, ap, dev);
+		rc = __ata_scsi_queuecmd(cmd, done, dev);
 	else {
 		cmd->result = (DID_BAD_TARGET << 16);
 		done(cmd);
@@ -2495,12 +2639,11 @@ int ata_scsi_queuecmd(struct scsi_cmnd *cmd, void (*done)(struct scsi_cmnd *))
 
 	spin_unlock(&ap->host_set->lock);
 	spin_lock(shost->host_lock);
-	return 0;
+	return rc;
 }
 
 /**
  *	ata_scsi_simulate - simulate SCSI command on ATA device
- *	@ap: port the device is connected to
  *	@dev: the target device
  *	@cmd: SCSI command being sent to device.
  *	@done: SCSI command completion function.
@@ -2512,14 +2655,12 @@ int ata_scsi_queuecmd(struct scsi_cmnd *cmd, void (*done)(struct scsi_cmnd *))
  *	spin_lock_irqsave(host_set lock)
  */
 
-void ata_scsi_simulate(struct ata_port *ap, struct ata_device *dev,
-		      struct scsi_cmnd *cmd,
+void ata_scsi_simulate(struct ata_device *dev, struct scsi_cmnd *cmd,
 		      void (*done)(struct scsi_cmnd *))
 {
 	struct ata_scsi_args args;
 	const u8 *scsicmd = cmd->cmnd;
 
-	args.ap = ap;
 	args.dev = dev;
 	args.id = dev->id;
 	args.cmd = cmd;
@@ -2605,3 +2746,26 @@ void ata_scsi_scan_host(struct ata_port *ap)
 	}
 }
 
+/**
+ *	ata_schedule_scsi_eh - schedule EH for SCSI host
+ *	@shost:	SCSI host to invoke error handling on.
+ *
+ *	Schedule SCSI EH without scmd.  This is a hack.
+ *
+ *	LOCKING:
+ *	spin_lock_irqsave(host_set lock)
+ **/
+void ata_schedule_scsi_eh(struct Scsi_Host *shost)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(shost->host_lock, flags);
+
+	if (scsi_host_set_state(shost, SHOST_RECOVERY) == 0 ||
+	    scsi_host_set_state(shost, SHOST_CANCEL_RECOVERY) == 0) {
+		shost->host_eh_scheduled++;
+		scsi_eh_wakeup(shost);
+	}
+
+	spin_unlock_irqrestore(shost->host_lock, flags);
+}

drivers/scsi/libata.h

@@ -32,7 +32,6 @@
 #define DRV_VERSION	"1.30"	/* must be exactly four chars */
 
 struct ata_scsi_args {
-	struct ata_port		*ap;
 	struct ata_device	*dev;
 	u16			*id;
 	struct scsi_cmnd	*cmd;
@@ -43,23 +42,22 @@ struct ata_scsi_args {
 extern int atapi_enabled;
 extern int atapi_dmadir;
 extern int libata_fua;
-extern struct ata_queued_cmd *ata_qc_new_init(struct ata_port *ap,
-				      struct ata_device *dev);
+extern struct ata_queued_cmd *ata_qc_new_init(struct ata_device *dev);
 extern int ata_rwcmd_protocol(struct ata_queued_cmd *qc);
-extern void ata_dev_disable(struct ata_port *ap, struct ata_device *dev);
+extern void ata_dev_disable(struct ata_device *dev);
 extern void ata_port_flush_task(struct ata_port *ap);
-extern unsigned ata_exec_internal(struct ata_port *ap, struct ata_device *dev,
+extern unsigned ata_exec_internal(struct ata_device *dev,
 				  struct ata_taskfile *tf, const u8 *cdb,
 				  int dma_dir, void *buf, unsigned int buflen);
-extern int ata_down_sata_spd_limit(struct ata_port *ap);
-extern int ata_set_sata_spd_needed(struct ata_port *ap);
-extern int ata_down_xfermask_limit(struct ata_port *ap, struct ata_device *dev,
-				   int force_pio0);
+extern int sata_down_spd_limit(struct ata_port *ap);
+extern int sata_set_spd_needed(struct ata_port *ap);
+extern int ata_down_xfermask_limit(struct ata_device *dev, int force_pio0);
 extern int ata_set_mode(struct ata_port *ap, struct ata_device **r_failed_dev);
 extern int ata_do_reset(struct ata_port *ap, ata_reset_fn_t reset,
-			ata_postreset_fn_t postreset, unsigned int *classes);
+			unsigned int *classes);
 extern void ata_qc_free(struct ata_queued_cmd *qc);
 extern void ata_qc_issue(struct ata_queued_cmd *qc);
+extern void __ata_qc_complete(struct ata_queued_cmd *qc);
 extern int ata_check_atapi_dma(struct ata_queued_cmd *qc);
 extern void ata_dev_select(struct ata_port *ap, unsigned int device,
                            unsigned int wait, unsigned int can_sleep);
@@ -100,9 +98,11 @@ extern void ata_scsi_set_sense(struct scsi_cmnd *cmd,
 extern void ata_scsi_rbuf_fill(struct ata_scsi_args *args,
                         unsigned int (*actor) (struct ata_scsi_args *args,
                                            u8 *rbuf, unsigned int buflen));
+extern void ata_schedule_scsi_eh(struct Scsi_Host *shost);
 
 /* libata-eh.c */
 extern enum scsi_eh_timer_return ata_scsi_timed_out(struct scsi_cmnd *cmd);
 extern void ata_scsi_error(struct Scsi_Host *host);
+extern void ata_qc_schedule_eh(struct ata_queued_cmd *qc);
 
 #endif /* __LIBATA_H__ */

drivers/scsi/pdc_adma.c

@@ -455,13 +455,13 @@ static inline unsigned int adma_intr_pkt(struct ata_host_set *host_set)
 			continue;
 		handled = 1;
 		adma_enter_reg_mode(ap);
-		if (ap->flags & (ATA_FLAG_DISABLED | ATA_FLAG_NOINTR))
+		if (ap->flags & ATA_FLAG_DISABLED)
 			continue;
 		pp = ap->private_data;
 		if (!pp || pp->state != adma_state_pkt)
 			continue;
 		qc = ata_qc_from_tag(ap, ap->active_tag);
-		if (qc && (!(qc->tf.ctl & ATA_NIEN))) {
+		if (qc && (!(qc->tf.flags & ATA_TFLAG_POLLING))) {
 			if ((status & (aPERR | aPSD | aUIRQ)))
 				qc->err_mask |= AC_ERR_OTHER;
 			else if (pp->pkt[0] != cDONE)
@@ -480,13 +480,13 @@ static inline unsigned int adma_intr_mmio(struct ata_host_set *host_set)
 	for (port_no = 0; port_no < host_set->n_ports; ++port_no) {
 		struct ata_port *ap;
 		ap = host_set->ports[port_no];
-		if (ap && (!(ap->flags & (ATA_FLAG_DISABLED | ATA_FLAG_NOINTR)))) {
+		if (ap && (!(ap->flags & ATA_FLAG_DISABLED))) {
 			struct ata_queued_cmd *qc;
 			struct adma_port_priv *pp = ap->private_data;
 			if (!pp || pp->state != adma_state_mmio)
 				continue;
 			qc = ata_qc_from_tag(ap, ap->active_tag);
-			if (qc && (!(qc->tf.ctl & ATA_NIEN))) {
+			if (qc && (!(qc->tf.flags & ATA_TFLAG_POLLING))) {
 
 				/* check main status, clearing INTRQ */
 				u8 status = ata_check_status(ap);

drivers/scsi/sata_mv.c

@@ -87,7 +87,7 @@ enum {
 	MV_FLAG_IRQ_COALESCE	= (1 << 29),  /* IRQ coalescing capability */
 	MV_COMMON_FLAGS		= (ATA_FLAG_SATA | ATA_FLAG_NO_LEGACY |
 				   ATA_FLAG_SATA_RESET | ATA_FLAG_MMIO |
-				   ATA_FLAG_NO_ATAPI),
+				   ATA_FLAG_PIO_POLLING),
 	MV_6XXX_FLAGS		= MV_FLAG_IRQ_COALESCE,
 
 	CRQB_FLAG_READ		= (1 << 0),
@@ -680,7 +680,7 @@ static void mv_stop_dma(struct ata_port *ap)
 	}
 
 	if (EDMA_EN & reg) {
-		printk(KERN_ERR "ata%u: Unable to stop eDMA\n", ap->id);
+		ata_port_printk(ap, KERN_ERR, "Unable to stop eDMA\n");
 		/* FIXME: Consider doing a reset here to recover */
 	}
 }
@@ -1309,8 +1309,8 @@ static void mv_err_intr(struct ata_port *ap)
 	edma_err_cause = readl(port_mmio + EDMA_ERR_IRQ_CAUSE_OFS);
 
 	if (EDMA_ERR_SERR & edma_err_cause) {
-		serr = scr_read(ap, SCR_ERROR);
-		scr_write_flush(ap, SCR_ERROR, serr);
+		sata_scr_read(ap, SCR_ERROR, &serr);
+		sata_scr_write_flush(ap, SCR_ERROR, serr);
 	}
 	if (EDMA_ERR_SELF_DIS & edma_err_cause) {
 		struct mv_port_priv *pp	= ap->private_data;
@@ -1396,7 +1396,7 @@ static void mv_host_intr(struct ata_host_set *host_set, u32 relevant,
 			}
 		}
 
-		if (ap->flags & (ATA_FLAG_DISABLED | ATA_FLAG_NOINTR))
+		if (ap && (ap->flags & ATA_FLAG_DISABLED))
 			continue;
 
 		err_mask = ac_err_mask(ata_status);
@@ -1417,7 +1417,7 @@ static void mv_host_intr(struct ata_host_set *host_set, u32 relevant,
 				VPRINTK("port %u IRQ found for qc, "
 					"ata_status 0x%x\n", port,ata_status);
 				/* mark qc status appropriately */
-				if (!(qc->tf.ctl & ATA_NIEN)) {
+				if (!(qc->tf.flags & ATA_TFLAG_POLLING)) {
 					qc->err_mask |= err_mask;
 					ata_qc_complete(qc);
 				}
@@ -1934,15 +1934,16 @@ static void __mv_phy_reset(struct ata_port *ap, int can_sleep)
 
 	/* Issue COMRESET via SControl */
 comreset_retry:
-	scr_write_flush(ap, SCR_CONTROL, 0x301);
+	sata_scr_write_flush(ap, SCR_CONTROL, 0x301);
 	__msleep(1, can_sleep);
 
-	scr_write_flush(ap, SCR_CONTROL, 0x300);
+	sata_scr_write_flush(ap, SCR_CONTROL, 0x300);
 	__msleep(20, can_sleep);
 
 	timeout = jiffies + msecs_to_jiffies(200);
 	do {
-		sstatus = scr_read(ap, SCR_STATUS) & 0x3;
+		sata_scr_read(ap, SCR_STATUS, &sstatus);
+		sstatus &= 0x3;
 		if ((sstatus == 3) || (sstatus == 0))
 			break;
 
@@ -1959,11 +1960,12 @@ comreset_retry:
 		"SCtrl 0x%08x\n", mv_scr_read(ap, SCR_STATUS),
 		mv_scr_read(ap, SCR_ERROR), mv_scr_read(ap, SCR_CONTROL));
 
-	if (sata_dev_present(ap)) {
+	if (ata_port_online(ap)) {
 		ata_port_probe(ap);
 	} else {
-		printk(KERN_INFO "ata%u: no device found (phy stat %08x)\n",
-		       ap->id, scr_read(ap, SCR_STATUS));
+		sata_scr_read(ap, SCR_STATUS, &sstatus);
+		ata_port_printk(ap, KERN_INFO,
+				"no device found (phy stat %08x)\n", sstatus);
 		ata_port_disable(ap);
 		return;
 	}
@@ -2021,7 +2023,7 @@ static void mv_eng_timeout(struct ata_port *ap)
 {
 	struct ata_queued_cmd *qc;
 
-	printk(KERN_ERR "ata%u: Entering mv_eng_timeout\n",ap->id);
+	ata_port_printk(ap, KERN_ERR, "Entering mv_eng_timeout\n");
 	DPRINTK("All regs @ start of eng_timeout\n");
 	mv_dump_all_regs(ap->host_set->mmio_base, ap->port_no,
 			 to_pci_dev(ap->host_set->dev));

drivers/scsi/sata_nv.c

@@ -279,11 +279,11 @@ static irqreturn_t nv_interrupt (int irq, void *dev_instance,
 
 		ap = host_set->ports[i];
 		if (ap &&
-		    !(ap->flags & (ATA_FLAG_DISABLED | ATA_FLAG_NOINTR))) {
+		    !(ap->flags & ATA_FLAG_DISABLED)) {
 			struct ata_queued_cmd *qc;
 
 			qc = ata_qc_from_tag(ap, ap->active_tag);
-			if (qc && (!(qc->tf.ctl & ATA_NIEN)))
+			if (qc && (!(qc->tf.flags & ATA_TFLAG_POLLING)))
 				handled += ata_host_intr(ap, qc);
 			else
 				// No request pending?  Clear interrupt status

drivers/scsi/sata_promise.c

@@ -76,7 +76,8 @@ enum {
 	PDC_RESET		= (1 << 11), /* HDMA reset */
 
 	PDC_COMMON_FLAGS	= ATA_FLAG_NO_LEGACY | ATA_FLAG_SRST |
-				  ATA_FLAG_MMIO | ATA_FLAG_NO_ATAPI,
+				  ATA_FLAG_MMIO | ATA_FLAG_NO_ATAPI |
+				  ATA_FLAG_PIO_POLLING,
 };
 
 
@@ -435,7 +436,7 @@ static void pdc_eng_timeout(struct ata_port *ap)
 	switch (qc->tf.protocol) {
 	case ATA_PROT_DMA:
 	case ATA_PROT_NODATA:
-		printk(KERN_ERR "ata%u: command timeout\n", ap->id);
+		ata_port_printk(ap, KERN_ERR, "command timeout\n");
 		drv_stat = ata_wait_idle(ap);
 		qc->err_mask |= __ac_err_mask(drv_stat);
 		break;
@@ -443,8 +444,9 @@ static void pdc_eng_timeout(struct ata_port *ap)
 	default:
 		drv_stat = ata_busy_wait(ap, ATA_BUSY | ATA_DRQ, 1000);
 
-		printk(KERN_ERR "ata%u: unknown timeout, cmd 0x%x stat 0x%x\n",
-		       ap->id, qc->tf.command, drv_stat);
+		ata_port_printk(ap, KERN_ERR,
+				"unknown timeout, cmd 0x%x stat 0x%x\n",
+				qc->tf.command, drv_stat);
 
 		qc->err_mask |= ac_err_mask(drv_stat);
 		break;
@@ -533,11 +535,11 @@ static irqreturn_t pdc_interrupt (int irq, void *dev_instance, struct pt_regs *r
 		ap = host_set->ports[i];
 		tmp = mask & (1 << (i + 1));
 		if (tmp && ap &&
-		    !(ap->flags & (ATA_FLAG_DISABLED | ATA_FLAG_NOINTR))) {
+		    !(ap->flags & ATA_FLAG_DISABLED)) {
 			struct ata_queued_cmd *qc;
 
 			qc = ata_qc_from_tag(ap, ap->active_tag);
-			if (qc && (!(qc->tf.ctl & ATA_NIEN)))
+			if (qc && (!(qc->tf.flags & ATA_TFLAG_POLLING)))
 				handled += pdc_host_intr(ap, qc);
 		}
 	}

drivers/scsi/sata_qstor.c

@@ -175,7 +175,7 @@ static const struct ata_port_info qs_port_info[] = {
 		.host_flags	= ATA_FLAG_SATA | ATA_FLAG_NO_LEGACY |
 				  ATA_FLAG_SATA_RESET |
 				  //FIXME ATA_FLAG_SRST |
-				  ATA_FLAG_MMIO,
+				  ATA_FLAG_MMIO | ATA_FLAG_PIO_POLLING,
 		.pio_mask	= 0x10, /* pio4 */
 		.udma_mask	= 0x7f, /* udma0-6 */
 		.port_ops	= &qs_ata_ops,
@@ -394,14 +394,13 @@ static inline unsigned int qs_intr_pkt(struct ata_host_set *host_set)
 			DPRINTK("SFF=%08x%08x: sCHAN=%u sHST=%d sDST=%02x\n",
 					sff1, sff0, port_no, sHST, sDST);
 			handled = 1;
-			if (ap && !(ap->flags &
-				    (ATA_FLAG_DISABLED|ATA_FLAG_NOINTR))) {
+			if (ap && !(ap->flags & ATA_FLAG_DISABLED)) {
 				struct ata_queued_cmd *qc;
 				struct qs_port_priv *pp = ap->private_data;
 				if (!pp || pp->state != qs_state_pkt)
 					continue;
 				qc = ata_qc_from_tag(ap, ap->active_tag);
-				if (qc && (!(qc->tf.ctl & ATA_NIEN))) {
+				if (qc && (!(qc->tf.flags & ATA_TFLAG_POLLING))) {
 					switch (sHST) {
 					case 0: /* successful CPB */
 					case 3: /* device error */
@@ -428,13 +427,13 @@ static inline unsigned int qs_intr_mmio(struct ata_host_set *host_set)
 		struct ata_port *ap;
 		ap = host_set->ports[port_no];
 		if (ap &&
-		    !(ap->flags & (ATA_FLAG_DISABLED | ATA_FLAG_NOINTR))) {
+		    !(ap->flags & ATA_FLAG_DISABLED)) {
 			struct ata_queued_cmd *qc;
 			struct qs_port_priv *pp = ap->private_data;
 			if (!pp || pp->state != qs_state_mmio)
 				continue;
 			qc = ata_qc_from_tag(ap, ap->active_tag);
-			if (qc && (!(qc->tf.ctl & ATA_NIEN))) {
+			if (qc && (!(qc->tf.flags & ATA_TFLAG_POLLING))) {
 
 				/* check main status, clearing INTRQ */
 				u8 status = ata_check_status(ap);

drivers/scsi/sata_sil.c

@@ -96,6 +96,8 @@ static void sil_dev_config(struct ata_port *ap, struct ata_device *dev);
 static u32 sil_scr_read (struct ata_port *ap, unsigned int sc_reg);
 static void sil_scr_write (struct ata_port *ap, unsigned int sc_reg, u32 val);
 static void sil_post_set_mode (struct ata_port *ap);
+static void sil_freeze(struct ata_port *ap);
+static void sil_thaw(struct ata_port *ap);
 
 
 static const struct pci_device_id sil_pci_tbl[] = {
@@ -174,7 +176,10 @@ static const struct ata_port_operations sil_ops = {
 	.bmdma_status		= ata_bmdma_status,
 	.qc_prep		= ata_qc_prep,
 	.qc_issue		= ata_qc_issue_prot,
-	.eng_timeout		= ata_eng_timeout,
+	.freeze			= sil_freeze,
+	.thaw			= sil_thaw,
+	.error_handler		= ata_bmdma_error_handler,
+	.post_internal_cmd	= ata_bmdma_post_internal_cmd,
 	.irq_handler		= ata_interrupt,
 	.irq_clear		= ata_bmdma_irq_clear,
 	.scr_read		= sil_scr_read,
@@ -314,6 +319,33 @@ static void sil_scr_write (struct ata_port *ap, unsigned int sc_reg, u32 val)
 		writel(val, mmio);
 }
 
+static void sil_freeze(struct ata_port *ap)
+{
+	void __iomem *mmio_base = ap->host_set->mmio_base;
+	u32 tmp;
+
+	/* plug IRQ */
+	tmp = readl(mmio_base + SIL_SYSCFG);
+	tmp |= SIL_MASK_IDE0_INT << ap->port_no;
+	writel(tmp, mmio_base + SIL_SYSCFG);
+	readl(mmio_base + SIL_SYSCFG);	/* flush */
+}
+
+static void sil_thaw(struct ata_port *ap)
+{
+	void __iomem *mmio_base = ap->host_set->mmio_base;
+	u32 tmp;
+
+	/* clear IRQ */
+	ata_chk_status(ap);
+	ata_bmdma_irq_clear(ap);
+
+	/* turn on IRQ */
+	tmp = readl(mmio_base + SIL_SYSCFG);
+	tmp &= ~(SIL_MASK_IDE0_INT << ap->port_no);
+	writel(tmp, mmio_base + SIL_SYSCFG);
+}
+
 /**
  *	sil_dev_config - Apply device/host-specific errata fixups
  *	@ap: Port containing device to be examined
@@ -360,16 +392,16 @@ static void sil_dev_config(struct ata_port *ap, struct ata_device *dev)
 	if (slow_down ||
 	    ((ap->flags & SIL_FLAG_MOD15WRITE) &&
 	     (quirks & SIL_QUIRK_MOD15WRITE))) {
-		printk(KERN_INFO "ata%u(%u): applying Seagate errata fix (mod15write workaround)\n",
-		       ap->id, dev->devno);
+		ata_dev_printk(dev, KERN_INFO, "applying Seagate errata fix "
+			       "(mod15write workaround)\n");
 		dev->max_sectors = 15;
 		return;
 	}
 
 	/* limit to udma5 */
 	if (quirks & SIL_QUIRK_UDMA5MAX) {
-		printk(KERN_INFO "ata%u(%u): applying Maxtor errata fix %s\n",
-		       ap->id, dev->devno, model_num);
+		ata_dev_printk(dev, KERN_INFO,
+			       "applying Maxtor errata fix %s\n", model_num);
 		dev->udma_mask &= ATA_UDMA5;
 		return;
 	}
@@ -384,7 +416,7 @@ static int sil_init_one (struct pci_dev *pdev, const struct pci_device_id *ent)
 	int rc;
 	unsigned int i;
 	int pci_dev_busy = 0;
-	u32 tmp, irq_mask;
+	u32 tmp;
 	u8 cls;
 
 	if (!printed_version++)
@@ -474,24 +506,11 @@ static int sil_init_one (struct pci_dev *pdev, const struct pci_device_id *ent)
 	}
 
 	if (ent->driver_data == sil_3114) {
-		irq_mask = SIL_MASK_4PORT;
-
 		/* flip the magic "make 4 ports work" bit */
 		tmp = readl(mmio_base + sil_port[2].bmdma);
 		if ((tmp & SIL_INTR_STEERING) == 0)
 			writel(tmp | SIL_INTR_STEERING,
 			       mmio_base + sil_port[2].bmdma);
-
-	} else {
-		irq_mask = SIL_MASK_2PORT;
-	}
-
-	/* make sure IDE0/1/2/3 interrupts are not masked */
-	tmp = readl(mmio_base + SIL_SYSCFG);
-	if (tmp & irq_mask) {
-		tmp &= ~irq_mask;
-		writel(tmp, mmio_base + SIL_SYSCFG);
-		readl(mmio_base + SIL_SYSCFG);	/* flush */
 	}
 
 	/* mask all SATA phy-related interrupts */

drivers/scsi/sata_sil24.c

@@ -156,6 +156,9 @@ enum {
 	PORT_IRQ_HANDSHAKE	= (1 << 10), /* handshake error threshold */
 	PORT_IRQ_SDB_NOTIFY	= (1 << 11), /* SDB notify received */
 
+	DEF_PORT_IRQ		= PORT_IRQ_COMPLETE | PORT_IRQ_ERROR |
+				  PORT_IRQ_DEV_XCHG | PORT_IRQ_UNK_FIS,
+
 	/* bits[27:16] are unmasked (raw) */
 	PORT_IRQ_RAW_SHIFT	= 16,
 	PORT_IRQ_MASKED_MASK	= 0x7ff,
@@ -213,6 +216,8 @@ enum {
 	SGE_DRD			= (1 << 29), /* discard data read (/dev/null)
 						data address ignored */
 
+	SIL24_MAX_CMDS		= 31,
+
 	/* board id */
 	BID_SIL3124		= 0,
 	BID_SIL3132		= 1,
@@ -220,7 +225,8 @@ enum {
 
 	/* host flags */
 	SIL24_COMMON_FLAGS	= ATA_FLAG_SATA | ATA_FLAG_NO_LEGACY |
-				  ATA_FLAG_MMIO | ATA_FLAG_PIO_DMA,
+				  ATA_FLAG_MMIO | ATA_FLAG_PIO_DMA |
+				  ATA_FLAG_NCQ,
 	SIL24_FLAG_PCIX_IRQ_WOC	= (1 << 24), /* IRQ loss errata on PCI-X */
 
 	IRQ_STAT_4PORTS		= 0xf,
@@ -242,6 +248,58 @@ union sil24_cmd_block {
 	struct sil24_atapi_block atapi;
 };
 
+static struct sil24_cerr_info {
+	unsigned int err_mask, action;
+	const char *desc;
+} sil24_cerr_db[] = {
+	[0]			= { AC_ERR_DEV, ATA_EH_REVALIDATE,
+				    "device error" },
+	[PORT_CERR_DEV]		= { AC_ERR_DEV, ATA_EH_REVALIDATE,
+				    "device error via D2H FIS" },
+	[PORT_CERR_SDB]		= { AC_ERR_DEV, ATA_EH_REVALIDATE,
+				    "device error via SDB FIS" },
+	[PORT_CERR_DATA]	= { AC_ERR_ATA_BUS, ATA_EH_SOFTRESET,
+				    "error in data FIS" },
+	[PORT_CERR_SEND]	= { AC_ERR_ATA_BUS, ATA_EH_SOFTRESET,
+				    "failed to transmit command FIS" },
+	[PORT_CERR_INCONSISTENT] = { AC_ERR_HSM, ATA_EH_SOFTRESET,
+				     "protocol mismatch" },
+	[PORT_CERR_DIRECTION]	= { AC_ERR_HSM, ATA_EH_SOFTRESET,
+				    "data directon mismatch" },
+	[PORT_CERR_UNDERRUN]	= { AC_ERR_HSM, ATA_EH_SOFTRESET,
+				    "ran out of SGEs while writing" },
+	[PORT_CERR_OVERRUN]	= { AC_ERR_HSM, ATA_EH_SOFTRESET,
+				    "ran out of SGEs while reading" },
+	[PORT_CERR_PKT_PROT]	= { AC_ERR_HSM, ATA_EH_SOFTRESET,
+				    "invalid data directon for ATAPI CDB" },
+	[PORT_CERR_SGT_BOUNDARY] = { AC_ERR_SYSTEM, ATA_EH_SOFTRESET,
+				     "SGT no on qword boundary" },
+	[PORT_CERR_SGT_TGTABRT]	= { AC_ERR_HOST_BUS, ATA_EH_SOFTRESET,
+				    "PCI target abort while fetching SGT" },
+	[PORT_CERR_SGT_MSTABRT]	= { AC_ERR_HOST_BUS, ATA_EH_SOFTRESET,
+				    "PCI master abort while fetching SGT" },
+	[PORT_CERR_SGT_PCIPERR]	= { AC_ERR_HOST_BUS, ATA_EH_SOFTRESET,
+				    "PCI parity error while fetching SGT" },
+	[PORT_CERR_CMD_BOUNDARY] = { AC_ERR_SYSTEM, ATA_EH_SOFTRESET,
+				     "PRB not on qword boundary" },
+	[PORT_CERR_CMD_TGTABRT]	= { AC_ERR_HOST_BUS, ATA_EH_SOFTRESET,
+				    "PCI target abort while fetching PRB" },
+	[PORT_CERR_CMD_MSTABRT]	= { AC_ERR_HOST_BUS, ATA_EH_SOFTRESET,
+				    "PCI master abort while fetching PRB" },
+	[PORT_CERR_CMD_PCIPERR]	= { AC_ERR_HOST_BUS, ATA_EH_SOFTRESET,
+				    "PCI parity error while fetching PRB" },
+	[PORT_CERR_XFR_UNDEF]	= { AC_ERR_HOST_BUS, ATA_EH_SOFTRESET,
+				    "undefined error while transferring data" },
+	[PORT_CERR_XFR_TGTABRT]	= { AC_ERR_HOST_BUS, ATA_EH_SOFTRESET,
+				    "PCI target abort while transferring data" },
+	[PORT_CERR_XFR_MSTABRT]	= { AC_ERR_HOST_BUS, ATA_EH_SOFTRESET,
+				    "PCI master abort while transferring data" },
+	[PORT_CERR_XFR_PCIPERR]	= { AC_ERR_HOST_BUS, ATA_EH_SOFTRESET,
+				    "PCI parity error while transferring data" },
+	[PORT_CERR_SENDSERVICE]	= { AC_ERR_HSM, ATA_EH_SOFTRESET,
+				    "FIS received while sending service FIS" },
+};
+
 /*
  * ap->private_data
  *
@@ -269,8 +327,11 @@ static int sil24_probe_reset(struct ata_port *ap, unsigned int *classes);
 static void sil24_qc_prep(struct ata_queued_cmd *qc);
 static unsigned int sil24_qc_issue(struct ata_queued_cmd *qc);
 static void sil24_irq_clear(struct ata_port *ap);
-static void sil24_eng_timeout(struct ata_port *ap);
 static irqreturn_t sil24_interrupt(int irq, void *dev_instance, struct pt_regs *regs);
+static void sil24_freeze(struct ata_port *ap);
+static void sil24_thaw(struct ata_port *ap);
+static void sil24_error_handler(struct ata_port *ap);
+static void sil24_post_internal_cmd(struct ata_queued_cmd *qc);
 static int sil24_port_start(struct ata_port *ap);
 static void sil24_port_stop(struct ata_port *ap);
 static void sil24_host_stop(struct ata_host_set *host_set);
@@ -297,7 +358,8 @@ static struct scsi_host_template sil24_sht = {
 	.name			= DRV_NAME,
 	.ioctl			= ata_scsi_ioctl,
 	.queuecommand		= ata_scsi_queuecmd,
-	.can_queue		= ATA_DEF_QUEUE,
+	.change_queue_depth	= ata_scsi_change_queue_depth,
+	.can_queue		= SIL24_MAX_CMDS,
 	.this_id		= ATA_SHT_THIS_ID,
 	.sg_tablesize		= LIBATA_MAX_PRD,
 	.cmd_per_lun		= ATA_SHT_CMD_PER_LUN,
@@ -325,14 +387,17 @@ static const struct ata_port_operations sil24_ops = {
 	.qc_prep		= sil24_qc_prep,
 	.qc_issue		= sil24_qc_issue,
 
-	.eng_timeout		= sil24_eng_timeout,
-
 	.irq_handler		= sil24_interrupt,
 	.irq_clear		= sil24_irq_clear,
 
 	.scr_read		= sil24_scr_read,
 	.scr_write		= sil24_scr_write,
 
+	.freeze			= sil24_freeze,
+	.thaw			= sil24_thaw,
+	.error_handler		= sil24_error_handler,
+	.post_internal_cmd	= sil24_post_internal_cmd,
+
 	.port_start		= sil24_port_start,
 	.port_stop		= sil24_port_stop,
 	.host_stop		= sil24_host_stop,
@@ -376,6 +441,13 @@ static struct ata_port_info sil24_port_info[] = {
 	},
 };
 
+static int sil24_tag(int tag)
+{
+	if (unlikely(ata_tag_internal(tag)))
+		return 0;
+	return tag;
+}
+
 static void sil24_dev_config(struct ata_port *ap, struct ata_device *dev)
 {
 	void __iomem *port = (void __iomem *)ap->ioaddr.cmd_addr;
@@ -459,21 +531,17 @@ static int sil24_softreset(struct ata_port *ap, unsigned int *class)
 	struct sil24_port_priv *pp = ap->private_data;
 	struct sil24_prb *prb = &pp->cmd_block[0].ata.prb;
 	dma_addr_t paddr = pp->cmd_block_dma;
-	u32 mask, irq_enable, irq_stat;
+	u32 mask, irq_stat;
 	const char *reason;
 
 	DPRINTK("ENTER\n");
 
-	if (!sata_dev_present(ap)) {
+	if (ata_port_offline(ap)) {
 		DPRINTK("PHY reports no device\n");
 		*class = ATA_DEV_NONE;
 		goto out;
 	}
 
-	/* temporarily turn off IRQs during SRST */
-	irq_enable = readl(port + PORT_IRQ_ENABLE_SET);
-	writel(irq_enable, port + PORT_IRQ_ENABLE_CLR);
-
 	/* put the port into known state */
 	if (sil24_init_port(ap)) {
 		reason ="port not ready";
@@ -494,9 +562,6 @@ static int sil24_softreset(struct ata_port *ap, unsigned int *class)
 	writel(irq_stat, port + PORT_IRQ_STAT); /* clear IRQs */
 	irq_stat >>= PORT_IRQ_RAW_SHIFT;
 
-	/* restore IRQs */
-	writel(irq_enable, port + PORT_IRQ_ENABLE_SET);
-
 	if (!(irq_stat & PORT_IRQ_COMPLETE)) {
 		if (irq_stat & PORT_IRQ_ERROR)
 			reason = "SRST command error";
@@ -516,7 +581,7 @@ static int sil24_softreset(struct ata_port *ap, unsigned int *class)
 	return 0;
 
  err:
-	printk(KERN_ERR "ata%u: softreset failed (%s)\n", ap->id, reason);
+	ata_port_printk(ap, KERN_ERR, "softreset failed (%s)\n", reason);
 	return -EIO;
 }
 
@@ -528,10 +593,10 @@ static int sil24_hardreset(struct ata_port *ap, unsigned int *class)
 	u32 tmp;
 
 	/* sil24 does the right thing(tm) without any protection */
-	ata_set_sata_spd(ap);
+	sata_set_spd(ap);
 
 	tout_msec = 100;
-	if (sata_dev_present(ap))
+	if (ata_port_online(ap))
 		tout_msec = 5000;
 
 	writel(PORT_CS_DEV_RST, port + PORT_CTRL_STAT);
@@ -544,7 +609,7 @@ static int sil24_hardreset(struct ata_port *ap, unsigned int *class)
 	msleep(100);
 
 	if (tmp & PORT_CS_DEV_RST) {
-		if (!sata_dev_present(ap))
+		if (ata_port_offline(ap))
 			return 0;
 		reason = "link not ready";
 		goto err;
@@ -561,7 +626,7 @@ static int sil24_hardreset(struct ata_port *ap, unsigned int *class)
 	return 0;
 
  err:
-	printk(KERN_ERR "ata%u: hardreset failed (%s)\n", ap->id, reason);
+	ata_port_printk(ap, KERN_ERR, "hardreset failed (%s)\n", reason);
 	return -EIO;
 }
 
@@ -595,14 +660,17 @@ static void sil24_qc_prep(struct ata_queued_cmd *qc)
 {
 	struct ata_port *ap = qc->ap;
 	struct sil24_port_priv *pp = ap->private_data;
-	union sil24_cmd_block *cb = pp->cmd_block + qc->tag;
+	union sil24_cmd_block *cb;
 	struct sil24_prb *prb;
 	struct sil24_sge *sge;
 	u16 ctrl = 0;
 
+	cb = &pp->cmd_block[sil24_tag(qc->tag)];
+
 	switch (qc->tf.protocol) {
 	case ATA_PROT_PIO:
 	case ATA_PROT_DMA:
+	case ATA_PROT_NCQ:
 	case ATA_PROT_NODATA:
 		prb = &cb->ata.prb;
 		sge = cb->ata.sge;
@@ -640,12 +708,17 @@ static void sil24_qc_prep(struct ata_queued_cmd *qc)
 static unsigned int sil24_qc_issue(struct ata_queued_cmd *qc)
 {
 	struct ata_port *ap = qc->ap;
-	void __iomem *port = (void __iomem *)ap->ioaddr.cmd_addr;
 	struct sil24_port_priv *pp = ap->private_data;
-	dma_addr_t paddr = pp->cmd_block_dma + qc->tag * sizeof(*pp->cmd_block);
+	void __iomem *port = (void __iomem *)ap->ioaddr.cmd_addr;
+	unsigned int tag = sil24_tag(qc->tag);
+	dma_addr_t paddr;
+	void __iomem *activate;
 
-	writel((u32)paddr, port + PORT_CMD_ACTIVATE);
-	writel((u64)paddr >> 32, port + PORT_CMD_ACTIVATE + 4);
+	paddr = pp->cmd_block_dma + tag * sizeof(*pp->cmd_block);
+	activate = port + PORT_CMD_ACTIVATE + tag * 8;
+
+	writel((u32)paddr, activate);
+	writel((u64)paddr >> 32, activate + 4);
 
 	return 0;
 }
@@ -655,166 +728,141 @@ static void sil24_irq_clear(struct ata_port *ap)
 	/* unused */
 }
 
-static int __sil24_restart_controller(void __iomem *port)
+static void sil24_freeze(struct ata_port *ap)
 {
-	u32 tmp;
-	int cnt;
-
-	writel(PORT_CS_INIT, port + PORT_CTRL_STAT);
-
-	/* Max ~10ms */
-	for (cnt = 0; cnt < 10000; cnt++) {
-		tmp = readl(port + PORT_CTRL_STAT);
-		if (tmp & PORT_CS_RDY)
-			return 0;
-		udelay(1);
-	}
+	void __iomem *port = (void __iomem *)ap->ioaddr.cmd_addr;
 
-	return -1;
+	/* Port-wide IRQ mask in HOST_CTRL doesn't really work, clear
+	 * PORT_IRQ_ENABLE instead.
+	 */
+	writel(0xffff, port + PORT_IRQ_ENABLE_CLR);
 }
 
-static void sil24_restart_controller(struct ata_port *ap)
+static void sil24_thaw(struct ata_port *ap)
 {
-	if (__sil24_restart_controller((void __iomem *)ap->ioaddr.cmd_addr))
-		printk(KERN_ERR DRV_NAME
-		       " ata%u: failed to restart controller\n", ap->id);
+	void __iomem *port = (void __iomem *)ap->ioaddr.cmd_addr;
+	u32 tmp;
+
+	/* clear IRQ */
+	tmp = readl(port + PORT_IRQ_STAT);
+	writel(tmp, port + PORT_IRQ_STAT);
+
+	/* turn IRQ back on */
+	writel(DEF_PORT_IRQ, port + PORT_IRQ_ENABLE_SET);
 }
 
-static int __sil24_reset_controller(void __iomem *port)
+static void sil24_error_intr(struct ata_port *ap)
 {
-	int cnt;
-	u32 tmp;
+	void __iomem *port = (void __iomem *)ap->ioaddr.cmd_addr;
+	struct ata_eh_info *ehi = &ap->eh_info;
+	int freeze = 0;
+	u32 irq_stat;
 
-	/* Reset controller state.  Is this correct? */
-	writel(PORT_CS_DEV_RST, port + PORT_CTRL_STAT);
-	readl(port + PORT_CTRL_STAT);	/* sync */
+	/* on error, we need to clear IRQ explicitly */
+	irq_stat = readl(port + PORT_IRQ_STAT);
+	writel(irq_stat, port + PORT_IRQ_STAT);
 
-	/* Max ~100ms */
-	for (cnt = 0; cnt < 1000; cnt++) {
-		udelay(100);
-		tmp = readl(port + PORT_CTRL_STAT);
-		if (!(tmp & PORT_CS_DEV_RST))
-			break;
-	}
+	/* first, analyze and record host port events */
+	ata_ehi_clear_desc(ehi);
 
-	if (tmp & PORT_CS_DEV_RST)
-		return -1;
+	ata_ehi_push_desc(ehi, "irq_stat 0x%08x", irq_stat);
 
-	if (tmp & PORT_CS_RDY)
-		return 0;
+	if (irq_stat & PORT_IRQ_DEV_XCHG) {
+		ehi->err_mask |= AC_ERR_ATA_BUS;
+		/* sil24 doesn't recover very well from phy
+		 * disconnection with a softreset.  Force hardreset.
+		 */
+		ehi->action |= ATA_EH_HARDRESET;
+		ata_ehi_push_desc(ehi, ", device_exchanged");
+		freeze = 1;
+	}
 
-	return __sil24_restart_controller(port);
-}
+	if (irq_stat & PORT_IRQ_UNK_FIS) {
+		ehi->err_mask |= AC_ERR_HSM;
+		ehi->action |= ATA_EH_SOFTRESET;
+		ata_ehi_push_desc(ehi , ", unknown FIS");
+		freeze = 1;
+	}
 
-static void sil24_reset_controller(struct ata_port *ap)
-{
-	printk(KERN_NOTICE DRV_NAME
-	       " ata%u: resetting controller...\n", ap->id);
-	if (__sil24_reset_controller((void __iomem *)ap->ioaddr.cmd_addr))
-                printk(KERN_ERR DRV_NAME
-                       " ata%u: failed to reset controller\n", ap->id);
-}
+	/* deal with command error */
+	if (irq_stat & PORT_IRQ_ERROR) {
+		struct sil24_cerr_info *ci = NULL;
+		unsigned int err_mask = 0, action = 0;
+		struct ata_queued_cmd *qc;
+		u32 cerr;
+
+		/* analyze CMD_ERR */
+		cerr = readl(port + PORT_CMD_ERR);
+		if (cerr < ARRAY_SIZE(sil24_cerr_db))
+			ci = &sil24_cerr_db[cerr];
+
+		if (ci && ci->desc) {
+			err_mask |= ci->err_mask;
+			action |= ci->action;
+			ata_ehi_push_desc(ehi, ", %s", ci->desc);
+		} else {
+			err_mask |= AC_ERR_OTHER;
+			action |= ATA_EH_SOFTRESET;
+			ata_ehi_push_desc(ehi, ", unknown command error %d",
+					  cerr);
+		}
 
-static void sil24_eng_timeout(struct ata_port *ap)
-{
-	struct ata_queued_cmd *qc;
+		/* record error info */
+		qc = ata_qc_from_tag(ap, ap->active_tag);
+		if (qc) {
+			sil24_update_tf(ap);
+			qc->err_mask |= err_mask;
+		} else
+			ehi->err_mask |= err_mask;
 
-	qc = ata_qc_from_tag(ap, ap->active_tag);
+		ehi->action |= action;
+	}
 
-	printk(KERN_ERR "ata%u: command timeout\n", ap->id);
-	qc->err_mask |= AC_ERR_TIMEOUT;
-	ata_eh_qc_complete(qc);
+	/* freeze or abort */
+	if (freeze)
+		ata_port_freeze(ap);
+	else
+		ata_port_abort(ap);
+}
 
-	sil24_reset_controller(ap);
+static void sil24_finish_qc(struct ata_queued_cmd *qc)
+{
+	if (qc->flags & ATA_QCFLAG_RESULT_TF)
+		sil24_update_tf(qc->ap);
 }
 
-static void sil24_error_intr(struct ata_port *ap, u32 slot_stat)
+static inline void sil24_host_intr(struct ata_port *ap)
 {
-	struct ata_queued_cmd *qc = ata_qc_from_tag(ap, ap->active_tag);
-	struct sil24_port_priv *pp = ap->private_data;
 	void __iomem *port = (void __iomem *)ap->ioaddr.cmd_addr;
-	u32 irq_stat, cmd_err, sstatus, serror;
-	unsigned int err_mask;
+	u32 slot_stat, qc_active;
+	int rc;
 
-	irq_stat = readl(port + PORT_IRQ_STAT);
-	writel(irq_stat, port + PORT_IRQ_STAT);		/* clear irq */
+	slot_stat = readl(port + PORT_SLOT_STAT);
 
-	if (!(irq_stat & PORT_IRQ_ERROR)) {
-		/* ignore non-completion, non-error irqs for now */
-		printk(KERN_WARNING DRV_NAME
-		       "ata%u: non-error exception irq (irq_stat %x)\n",
-		       ap->id, irq_stat);
+	if (unlikely(slot_stat & HOST_SSTAT_ATTN)) {
+		sil24_error_intr(ap);
 		return;
 	}
 
-	cmd_err = readl(port + PORT_CMD_ERR);
-	sstatus = readl(port + PORT_SSTATUS);
-	serror = readl(port + PORT_SERROR);
-	if (serror)
-		writel(serror, port + PORT_SERROR);
+	if (ap->flags & SIL24_FLAG_PCIX_IRQ_WOC)
+		writel(PORT_IRQ_COMPLETE, port + PORT_IRQ_STAT);
 
-	/*
-	 * Don't log ATAPI device errors.  They're supposed to happen
-	 * and any serious errors will be logged using sense data by
-	 * the SCSI layer.
-	 */
-	if (ap->device[0].class != ATA_DEV_ATAPI || cmd_err > PORT_CERR_SDB)
-		printk("ata%u: error interrupt on port%d\n"
-		       "  stat=0x%x irq=0x%x cmd_err=%d sstatus=0x%x serror=0x%x\n",
-		       ap->id, ap->port_no, slot_stat, irq_stat, cmd_err, sstatus, serror);
-
-	if (cmd_err == PORT_CERR_DEV || cmd_err == PORT_CERR_SDB) {
-		/*
-		 * Device is reporting error, tf registers are valid.
-		 */
-		sil24_update_tf(ap);
-		err_mask = ac_err_mask(pp->tf.command);
-		sil24_restart_controller(ap);
-	} else {
-		/*
-		 * Other errors.  libata currently doesn't have any
-		 * mechanism to report these errors.  Just turn on
-		 * ATA_ERR.
-		 */
-		err_mask = AC_ERR_OTHER;
-		sil24_reset_controller(ap);
+	qc_active = slot_stat & ~HOST_SSTAT_ATTN;
+	rc = ata_qc_complete_multiple(ap, qc_active, sil24_finish_qc);
+	if (rc > 0)
+		return;
+	if (rc < 0) {
+		struct ata_eh_info *ehi = &ap->eh_info;
+		ehi->err_mask |= AC_ERR_HSM;
+		ehi->action |= ATA_EH_SOFTRESET;
+		ata_port_freeze(ap);
+		return;
 	}
 
-	if (qc) {
-		qc->err_mask |= err_mask;
-		ata_qc_complete(qc);
-	}
-}
-
-static inline void sil24_host_intr(struct ata_port *ap)
-{
-	struct ata_queued_cmd *qc = ata_qc_from_tag(ap, ap->active_tag);
-	void __iomem *port = (void __iomem *)ap->ioaddr.cmd_addr;
-	u32 slot_stat;
-
-	slot_stat = readl(port + PORT_SLOT_STAT);
-	if (!(slot_stat & HOST_SSTAT_ATTN)) {
-		struct sil24_port_priv *pp = ap->private_data;
-
-		if (ap->flags & SIL24_FLAG_PCIX_IRQ_WOC)
-			writel(PORT_IRQ_COMPLETE, port + PORT_IRQ_STAT);
-
-		/*
-		 * !HOST_SSAT_ATTN guarantees successful completion,
-		 * so reading back tf registers is unnecessary for
-		 * most commands.  TODO: read tf registers for
-		 * commands which require these values on successful
-		 * completion (EXECUTE DEVICE DIAGNOSTIC, CHECK POWER,
-		 * DEVICE RESET and READ PORT MULTIPLIER (any more?).
-		 */
-		sil24_update_tf(ap);
-
-		if (qc) {
-			qc->err_mask |= ac_err_mask(pp->tf.command);
-			ata_qc_complete(qc);
-		}
-	} else
-		sil24_error_intr(ap, slot_stat);
+	if (ata_ratelimit())
+		ata_port_printk(ap, KERN_INFO, "spurious interrupt "
+			"(slot_stat 0x%x active_tag %d sactive 0x%x)\n",
+			slot_stat, ap->active_tag, ap->sactive);
 }
 
 static irqreturn_t sil24_interrupt(int irq, void *dev_instance, struct pt_regs *regs)
@@ -854,9 +902,34 @@ static irqreturn_t sil24_interrupt(int irq, void *dev_instance, struct pt_regs *
 	return IRQ_RETVAL(handled);
 }
 
+static void sil24_error_handler(struct ata_port *ap)
+{
+	struct ata_eh_context *ehc = &ap->eh_context;
+
+	if (sil24_init_port(ap)) {
+		ata_eh_freeze_port(ap);
+		ehc->i.action |= ATA_EH_HARDRESET;
+	}
+
+	/* perform recovery */
+	ata_do_eh(ap, sil24_softreset, sil24_hardreset, ata_std_postreset);
+}
+
+static void sil24_post_internal_cmd(struct ata_queued_cmd *qc)
+{
+	struct ata_port *ap = qc->ap;
+
+	if (qc->flags & ATA_QCFLAG_FAILED)
+		qc->err_mask |= AC_ERR_OTHER;
+
+	/* make DMA engine forget about the failed command */
+	if (qc->err_mask)
+		sil24_init_port(ap);
+}
+
 static inline void sil24_cblk_free(struct sil24_port_priv *pp, struct device *dev)
 {
-	const size_t cb_size = sizeof(*pp->cmd_block);
+	const size_t cb_size = sizeof(*pp->cmd_block) * SIL24_MAX_CMDS;
 
 	dma_free_coherent(dev, cb_size, pp->cmd_block, pp->cmd_block_dma);
 }
@@ -866,7 +939,7 @@ static int sil24_port_start(struct ata_port *ap)
 	struct device *dev = ap->host_set->dev;
 	struct sil24_port_priv *pp;
 	union sil24_cmd_block *cb;
-	size_t cb_size = sizeof(*cb);
+	size_t cb_size = sizeof(*cb) * SIL24_MAX_CMDS;
 	dma_addr_t cb_dma;
 	int rc = -ENOMEM;
 
@@ -1066,15 +1139,6 @@ static int sil24_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 		/* Always use 64bit activation */
 		writel(PORT_CS_32BIT_ACTV, port + PORT_CTRL_CLR);
 
-		/* Configure interrupts */
-		writel(0xffff, port + PORT_IRQ_ENABLE_CLR);
-		writel(PORT_IRQ_COMPLETE | PORT_IRQ_ERROR |
-		       PORT_IRQ_SDB_NOTIFY, port + PORT_IRQ_ENABLE_SET);
-
-		/* Clear interrupts */
-		writel(0x0fff0fff, port + PORT_IRQ_STAT);
-		writel(PORT_CS_IRQ_WOC, port + PORT_CTRL_CLR);
-
 		/* Clear port multiplier enable and resume bits */
 		writel(PORT_CS_PM_EN | PORT_CS_RESUME, port + PORT_CTRL_CLR);
 	}

drivers/scsi/sata_sx4.c

@@ -218,7 +218,7 @@ static const struct ata_port_info pdc_port_info[] = {
 		.sht		= &pdc_sata_sht,
 		.host_flags	= ATA_FLAG_SATA | ATA_FLAG_NO_LEGACY |
 				  ATA_FLAG_SRST | ATA_FLAG_MMIO |
-				  ATA_FLAG_NO_ATAPI,
+				  ATA_FLAG_PIO_POLLING,
 		.pio_mask	= 0x1f, /* pio0-4 */
 		.mwdma_mask	= 0x07, /* mwdma0-2 */
 		.udma_mask	= 0x7f, /* udma0-6 ; FIXME */
@@ -833,11 +833,11 @@ static irqreturn_t pdc20621_interrupt (int irq, void *dev_instance, struct pt_re
 		tmp = mask & (1 << i);
 		VPRINTK("seq %u, port_no %u, ap %p, tmp %x\n", i, port_no, ap, tmp);
 		if (tmp && ap &&
-		    !(ap->flags & (ATA_FLAG_DISABLED | ATA_FLAG_NOINTR))) {
+		    !(ap->flags & ATA_FLAG_DISABLED)) {
 			struct ata_queued_cmd *qc;
 
 			qc = ata_qc_from_tag(ap, ap->active_tag);
-			if (qc && (!(qc->tf.ctl & ATA_NIEN)))
+			if (qc && (!(qc->tf.flags & ATA_TFLAG_POLLING)))
 				handled += pdc20621_host_intr(ap, qc, (i > 4),
 							      mmio_base);
 		}
@@ -868,15 +868,16 @@ static void pdc_eng_timeout(struct ata_port *ap)
 	switch (qc->tf.protocol) {
 	case ATA_PROT_DMA:
 	case ATA_PROT_NODATA:
-		printk(KERN_ERR "ata%u: command timeout\n", ap->id);
+		ata_port_printk(ap, KERN_ERR, "command timeout\n");
 		qc->err_mask |= __ac_err_mask(ata_wait_idle(ap));
 		break;
 
 	default:
 		drv_stat = ata_busy_wait(ap, ATA_BUSY | ATA_DRQ, 1000);
 
-		printk(KERN_ERR "ata%u: unknown timeout, cmd 0x%x stat 0x%x\n",
-		       ap->id, qc->tf.command, drv_stat);
+		ata_port_printk(ap, KERN_ERR,
+				"unknown timeout, cmd 0x%x stat 0x%x\n",
+				qc->tf.command, drv_stat);
 
 		qc->err_mask |= ac_err_mask(drv_stat);
 		break;

drivers/scsi/sata_vsc.c

@@ -221,14 +221,21 @@ static irqreturn_t vsc_sata_interrupt (int irq, void *dev_instance,
 
 			ap = host_set->ports[i];
 
-			if (ap && !(ap->flags &
-				    (ATA_FLAG_DISABLED|ATA_FLAG_NOINTR))) {
+			if (is_vsc_sata_int_err(i, int_status)) {
+				u32 err_status;
+				printk(KERN_DEBUG "%s: ignoring interrupt(s)\n", __FUNCTION__);
+				err_status = ap ? vsc_sata_scr_read(ap, SCR_ERROR) : 0;
+				vsc_sata_scr_write(ap, SCR_ERROR, err_status);
+				handled++;
+			}
+
+			if (ap && !(ap->flags & ATA_FLAG_DISABLED)) {
 				struct ata_queued_cmd *qc;
 
 				qc = ata_qc_from_tag(ap, ap->active_tag);
-				if (qc && (!(qc->tf.ctl & ATA_NIEN))) {
+				if (qc && (!(qc->tf.flags & ATA_TFLAG_POLLING)))
 					handled += ata_host_intr(ap, qc);
-				} else if (is_vsc_sata_int_err(i, int_status)) {
+				else if (is_vsc_sata_int_err(i, int_status)) {
 					/*
 					 * On some chips (i.e. Intel 31244), an error
 					 * interrupt will sneak in at initialization

drivers/scsi/scsi.c

@@ -719,6 +719,24 @@ void scsi_init_cmd_from_req(struct scsi_cmnd *cmd, struct scsi_request *sreq)
  */
 static DEFINE_PER_CPU(struct list_head, scsi_done_q);
 
+/**
+ * scsi_req_abort_cmd -- Request command recovery for the specified command
+ * cmd: pointer to the SCSI command of interest
+ *
+ * This function requests that SCSI Core start recovery for the
+ * command by deleting the timer and adding the command to the eh
+ * queue.  It can be called by either LLDDs or SCSI Core.  LLDDs who
+ * implement their own error recovery MAY ignore the timeout event if
+ * they generated scsi_req_abort_cmd.
+ */
+void scsi_req_abort_cmd(struct scsi_cmnd *cmd)
+{
+	if (!scsi_delete_timer(cmd))
+		return;
+	scsi_times_out(cmd);
+}
+EXPORT_SYMBOL(scsi_req_abort_cmd);
+
 /**
  * scsi_done - Enqueue the finished SCSI command into the done queue.
  * @cmd: The SCSI Command for which a low-level device driver (LLDD) gives

drivers/scsi/scsi_error.c

@@ -56,6 +56,7 @@ void scsi_eh_wakeup(struct Scsi_Host *shost)
 				printk("Waking error handler thread\n"));
 	}
 }
+EXPORT_SYMBOL_GPL(scsi_eh_wakeup);
 
 /**
  * scsi_eh_scmd_add - add scsi cmd to error handling.
@@ -1517,7 +1518,7 @@ int scsi_error_handler(void *data)
 	 */
 	set_current_state(TASK_INTERRUPTIBLE);
 	while (!kthread_should_stop()) {
-		if (shost->host_failed == 0 ||
+		if ((shost->host_failed == 0 && shost->host_eh_scheduled == 0) ||
 		    shost->host_failed != shost->host_busy) {
 			SCSI_LOG_ERROR_RECOVERY(1,
 				printk("Error handler scsi_eh_%d sleeping\n",

drivers/scsi/scsi_lib.c

@@ -566,7 +566,7 @@ void scsi_device_unbusy(struct scsi_device *sdev)
 	spin_lock_irqsave(shost->host_lock, flags);
 	shost->host_busy--;
 	if (unlikely(scsi_host_in_recovery(shost) &&
-		     shost->host_failed))
+		     (shost->host_failed || shost->host_eh_scheduled)))
 		scsi_eh_wakeup(shost);
 	spin_unlock(shost->host_lock);
 	spin_lock(sdev->request_queue->queue_lock);

drivers/scsi/scsi_priv.h

@@ -63,7 +63,6 @@ extern int scsi_delete_timer(struct scsi_cmnd *);
 extern void scsi_times_out(struct scsi_cmnd *cmd);
 extern int scsi_error_handler(void *host);
 extern int scsi_decide_disposition(struct scsi_cmnd *cmd);
-extern void scsi_eh_wakeup(struct Scsi_Host *shost);
 extern int scsi_eh_scmd_add(struct scsi_cmnd *, int);
 
 /* scsi_lib.c */

include/linux/ata.h

@@ -97,6 +97,9 @@ enum {
 	ATA_DRQ			= (1 << 3),	/* data request i/o */
 	ATA_ERR			= (1 << 0),	/* have an error */
 	ATA_SRST		= (1 << 2),	/* software reset */
+	ATA_ICRC		= (1 << 7),	/* interface CRC error */
+	ATA_UNC			= (1 << 6),	/* uncorrectable media error */
+	ATA_IDNF		= (1 << 4),	/* ID not found */
 	ATA_ABORTED		= (1 << 2),	/* command aborted */
 
 	/* ATA command block registers */
@@ -130,6 +133,8 @@ enum {
 	ATA_CMD_WRITE		= 0xCA,
 	ATA_CMD_WRITE_EXT	= 0x35,
 	ATA_CMD_WRITE_FUA_EXT	= 0x3D,
+	ATA_CMD_FPDMA_READ	= 0x60,
+	ATA_CMD_FPDMA_WRITE	= 0x61,
 	ATA_CMD_PIO_READ	= 0x20,
 	ATA_CMD_PIO_READ_EXT	= 0x24,
 	ATA_CMD_PIO_WRITE	= 0x30,
@@ -148,6 +153,10 @@ enum {
 	ATA_CMD_INIT_DEV_PARAMS	= 0x91,
 	ATA_CMD_READ_NATIVE_MAX	= 0xF8,
 	ATA_CMD_READ_NATIVE_MAX_EXT = 0x27,
+	ATA_CMD_READ_LOG_EXT	= 0x2f,
+
+	/* READ_LOG_EXT pages */
+	ATA_LOG_SATA_NCQ	= 0x10,
 
 	/* SETFEATURES stuff */
 	SETFEATURES_XFER	= 0x03,
@@ -192,6 +201,16 @@ enum {
 	SCR_ACTIVE		= 3,
 	SCR_NOTIFICATION	= 4,
 
+	/* SError bits */
+	SERR_DATA_RECOVERED	= (1 << 0), /* recovered data error */
+	SERR_COMM_RECOVERED	= (1 << 1), /* recovered comm failure */
+	SERR_DATA		= (1 << 8), /* unrecovered data error */
+	SERR_PERSISTENT		= (1 << 9), /* persistent data/comm error */
+	SERR_PROTOCOL		= (1 << 10), /* protocol violation */
+	SERR_INTERNAL		= (1 << 11), /* host internal error */
+	SERR_PHYRDY_CHG		= (1 << 16), /* PHY RDY changed */
+	SERR_DEV_XCHG		= (1 << 26), /* device exchanged */
+
 	/* struct ata_taskfile flags */
 	ATA_TFLAG_LBA48		= (1 << 0), /* enable 48-bit LBA and "HOB" */
 	ATA_TFLAG_ISADDR	= (1 << 1), /* enable r/w to nsect/lba regs */
@@ -199,6 +218,7 @@ enum {
 	ATA_TFLAG_WRITE		= (1 << 3), /* data dir: host->dev==1 (write) */
 	ATA_TFLAG_LBA		= (1 << 4), /* enable LBA */
 	ATA_TFLAG_FUA		= (1 << 5), /* enable FUA */
+	ATA_TFLAG_POLLING	= (1 << 6), /* set nIEN to 1 and use polling */
 };
 
 enum ata_tf_protocols {
@@ -207,6 +227,7 @@ enum ata_tf_protocols {
 	ATA_PROT_NODATA,	/* no data */
 	ATA_PROT_PIO,		/* PIO single sector */
 	ATA_PROT_DMA,		/* DMA */
+	ATA_PROT_NCQ,		/* NCQ */
 	ATA_PROT_ATAPI,		/* packet command, PIO data xfer*/
 	ATA_PROT_ATAPI_NODATA,	/* packet command, no data */
 	ATA_PROT_ATAPI_DMA,	/* packet command with special DMA sauce */
@@ -262,6 +283,8 @@ struct ata_taskfile {
 #define ata_id_has_pm(id)	((id)[82] & (1 << 3))
 #define ata_id_has_lba(id)	((id)[49] & (1 << 9))
 #define ata_id_has_dma(id)	((id)[49] & (1 << 8))
+#define ata_id_has_ncq(id)	((id)[76] & (1 << 8))
+#define ata_id_queue_depth(id)	(((id)[75] & 0x1f) + 1)
 #define ata_id_removeable(id)	((id)[0] & (1 << 7))
 #define ata_id_has_dword_io(id)	((id)[50] & (1 << 0))
 #define ata_id_u32(id,n)	\
@@ -272,6 +295,8 @@ struct ata_taskfile {
 	  ((u64) (id)[(n) + 1] << 16) |	\
 	  ((u64) (id)[(n) + 0]) )
 
+#define ata_id_cdb_intr(id)	(((id)[0] & 0x60) == 0x20)
+
 static inline unsigned int ata_id_major_version(const u16 *id)
 {
 	unsigned int mver;
@@ -311,6 +336,15 @@ static inline int is_atapi_taskfile(const struct ata_taskfile *tf)
 	       (tf->protocol == ATA_PROT_ATAPI_DMA);
 }
 
+static inline int is_multi_taskfile(struct ata_taskfile *tf)
+{
+	return (tf->command == ATA_CMD_READ_MULTI) ||
+	       (tf->command == ATA_CMD_WRITE_MULTI) ||
+	       (tf->command == ATA_CMD_READ_MULTI_EXT) ||
+	       (tf->command == ATA_CMD_WRITE_MULTI_EXT) ||
+	       (tf->command == ATA_CMD_WRITE_MULTI_FUA_EXT);
+}
+
 static inline int ata_ok(u8 status)
 {
 	return ((status & (ATA_BUSY | ATA_DRDY | ATA_DF | ATA_DRQ | ATA_ERR))

include/linux/libata.h

@@ -108,7 +108,9 @@ enum {
 	LIBATA_MAX_PRD		= ATA_MAX_PRD / 2,
 	ATA_MAX_PORTS		= 8,
 	ATA_DEF_QUEUE		= 1,
-	ATA_MAX_QUEUE		= 1,
+	/* tag ATA_MAX_QUEUE - 1 is reserved for internal commands */
+	ATA_MAX_QUEUE		= 32,
+	ATA_TAG_INTERNAL	= ATA_MAX_QUEUE - 1,
 	ATA_MAX_SECTORS		= 200,	/* FIXME */
 	ATA_MAX_BUS		= 2,
 	ATA_DEF_BUSY_WAIT	= 10000,
@@ -122,6 +124,8 @@ enum {
 	/* struct ata_device stuff */
 	ATA_DFLAG_LBA		= (1 << 0), /* device supports LBA */
 	ATA_DFLAG_LBA48		= (1 << 1), /* device supports LBA48 */
+	ATA_DFLAG_CDB_INTR	= (1 << 2), /* device asserts INTRQ when ready for CDB */
+	ATA_DFLAG_NCQ		= (1 << 3), /* device supports NCQ */
 	ATA_DFLAG_CFG_MASK	= (1 << 8) - 1,
 
 	ATA_DFLAG_PIO		= (1 << 8), /* device currently in PIO mode */
@@ -145,14 +149,19 @@ enum {
 	ATA_FLAG_PIO_DMA	= (1 << 7), /* PIO cmds via DMA */
 	ATA_FLAG_PIO_LBA48	= (1 << 8), /* Host DMA engine is LBA28 only */
 	ATA_FLAG_IRQ_MASK	= (1 << 9), /* Mask IRQ in PIO xfers */
+	ATA_FLAG_PIO_POLLING	= (1 << 10), /* use polling PIO if LLD
+					      * doesn't handle PIO interrupts */
+	ATA_FLAG_NCQ		= (1 << 11), /* host supports NCQ */
 
-	ATA_FLAG_NOINTR		= (1 << 16), /* FIXME: Remove this once
-					      * proper HSM is in place. */
-	ATA_FLAG_DEBUGMSG	= (1 << 17),
-	ATA_FLAG_FLUSH_PORT_TASK = (1 << 18), /* flush port task */
+	ATA_FLAG_DEBUGMSG	= (1 << 14),
+	ATA_FLAG_FLUSH_PORT_TASK = (1 << 15), /* flush port task */
 
-	ATA_FLAG_DISABLED	= (1 << 19), /* port is disabled, ignore it */
-	ATA_FLAG_SUSPENDED	= (1 << 20), /* port is suspended */
+	ATA_FLAG_EH_PENDING	= (1 << 16), /* EH pending */
+	ATA_FLAG_FROZEN		= (1 << 17), /* port is frozen */
+	ATA_FLAG_RECOVERED	= (1 << 18), /* recovery action performed */
+
+	ATA_FLAG_DISABLED	= (1 << 22), /* port is disabled, ignore it */
+	ATA_FLAG_SUSPENDED	= (1 << 23), /* port is suspended (power) */
 
 	/* bits 24:31 of ap->flags are reserved for LLDD specific flags */
 
@@ -162,17 +171,18 @@ enum {
 	ATA_QCFLAG_SINGLE	= (1 << 2), /* no s/g, just a single buffer */
 	ATA_QCFLAG_DMAMAP	= ATA_QCFLAG_SG | ATA_QCFLAG_SINGLE,
 	ATA_QCFLAG_IO		= (1 << 3), /* standard IO command */
-	ATA_QCFLAG_EH_SCHEDULED = (1 << 4), /* EH scheduled */
+	ATA_QCFLAG_RESULT_TF	= (1 << 4), /* result TF requested */
+
+	ATA_QCFLAG_FAILED	= (1 << 16), /* cmd failed and is owned by EH */
+	ATA_QCFLAG_SENSE_VALID	= (1 << 17), /* sense data valid */
+	ATA_QCFLAG_EH_SCHEDULED = (1 << 18), /* EH scheduled (obsolete) */
 
 	/* host set flags */
 	ATA_HOST_SIMPLEX	= (1 << 0),	/* Host is simplex, one DMA channel per host_set only */
 	
 	/* various lengths of time */
-	ATA_TMOUT_PIO		= 30 * HZ,
 	ATA_TMOUT_BOOT		= 30 * HZ,	/* heuristic */
 	ATA_TMOUT_BOOT_QUICK	= 7 * HZ,	/* heuristic */
-	ATA_TMOUT_CDB		= 30 * HZ,
-	ATA_TMOUT_CDB_QUICK	= 5 * HZ,
 	ATA_TMOUT_INTERNAL	= 30 * HZ,
 	ATA_TMOUT_INTERNAL_QUICK = 5 * HZ,
 
@@ -216,19 +226,39 @@ enum {
 	ATA_PORT_PRIMARY	= (1 << 0),
 	ATA_PORT_SECONDARY	= (1 << 1),
 
+	/* ering size */
+	ATA_ERING_SIZE		= 32,
+
+	/* desc_len for ata_eh_info and context */
+	ATA_EH_DESC_LEN		= 80,
+
+	/* reset / recovery action types */
+	ATA_EH_REVALIDATE	= (1 << 0),
+	ATA_EH_SOFTRESET	= (1 << 1),
+	ATA_EH_HARDRESET	= (1 << 2),
+
+	ATA_EH_RESET_MASK	= ATA_EH_SOFTRESET | ATA_EH_HARDRESET,
+
+	/* ata_eh_info->flags */
+	ATA_EHI_DID_RESET	= (1 << 0), /* already reset this port */
+
+	/* max repeat if error condition is still set after ->error_handler */
+	ATA_EH_MAX_REPEAT	= 5,
+
 	/* how hard are we gonna try to probe/recover devices */
 	ATA_PROBE_MAX_TRIES	= 3,
+	ATA_EH_RESET_TRIES	= 3,
+	ATA_EH_DEV_TRIES	= 3,
 };
 
 enum hsm_task_states {
-	HSM_ST_UNKNOWN,
-	HSM_ST_IDLE,
-	HSM_ST_POLL,
-	HSM_ST_TMOUT,
-	HSM_ST,
-	HSM_ST_LAST,
-	HSM_ST_LAST_POLL,
-	HSM_ST_ERR,
+	HSM_ST_UNKNOWN,		/* state unknown */
+	HSM_ST_IDLE,		/* no command on going */
+	HSM_ST,			/* (waiting the device to) transfer data */
+	HSM_ST_LAST,		/* (waiting the device to) complete command */
+	HSM_ST_ERR,		/* error */
+	HSM_ST_FIRST,		/* (waiting the device to)
+				   write CDB or first data block */
 };
 
 enum ata_completion_errors {
@@ -343,7 +373,7 @@ struct ata_queued_cmd {
 	struct scatterlist	*__sg;
 
 	unsigned int		err_mask;
-
+	struct ata_taskfile	result_tf;
 	ata_qc_cb_t		complete_fn;
 
 	void			*private_data;
@@ -355,12 +385,24 @@ struct ata_host_stats {
 	unsigned long		rw_reqbuf;
 };
 
+struct ata_ering_entry {
+	int			is_io;
+	unsigned int		err_mask;
+	u64			timestamp;
+};
+
+struct ata_ering {
+	int			cursor;
+	struct ata_ering_entry	ring[ATA_ERING_SIZE];
+};
+
 struct ata_device {
+	struct ata_port		*ap;
 	u64			n_sectors;	/* size of device, if ATA */
 	unsigned long		flags;		/* ATA_DFLAG_xxx */
 	unsigned int		class;		/* ATA_DEV_xxx */
 	unsigned int		devno;		/* 0 or 1 */
-	u16			*id;		/* IDENTIFY xxx DEVICE data */
+	u16			id[ATA_ID_WORDS]; /* IDENTIFY xxx DEVICE data */
 	u8			pio_mode;
 	u8			dma_mode;
 	u8			xfer_mode;
@@ -380,6 +422,24 @@ struct ata_device {
 	u16			cylinders;	/* Number of cylinders */
 	u16			heads;		/* Number of heads */
 	u16			sectors;	/* Number of sectors per track */
+
+	/* error history */
+	struct ata_ering	ering;
+};
+
+struct ata_eh_info {
+	struct ata_device	*dev;		/* offending device */
+	u32			serror;		/* SError from LLDD */
+	unsigned int		err_mask;	/* port-wide err_mask */
+	unsigned int		action;		/* ATA_EH_* action mask */
+	unsigned int		flags;		/* ATA_EHI_* flags */
+	char			desc[ATA_EH_DESC_LEN];
+	int			desc_len;
+};
+
+struct ata_eh_context {
+	struct ata_eh_info	i;
+	int			tries[ATA_MAX_DEVICES];
 };
 
 struct ata_port {
@@ -406,11 +466,19 @@ struct ata_port {
 	unsigned int		cbl;	/* cable type; ATA_CBL_xxx */
 	unsigned int		sata_spd_limit;	/* SATA PHY speed limit */
 
+	/* record runtime error info, protected by host_set lock */
+	struct ata_eh_info	eh_info;
+	/* EH context owned by EH */
+	struct ata_eh_context	eh_context;
+
 	struct ata_device	device[ATA_MAX_DEVICES];
 
 	struct ata_queued_cmd	qcmd[ATA_MAX_QUEUE];
-	unsigned long		qactive;
+	unsigned long		qc_allocated;
+	unsigned int		qc_active;
+
 	unsigned int		active_tag;
+	u32			sactive;
 
 	struct ata_host_stats	stats;
 	struct ata_host_set	*host_set;
@@ -419,12 +487,13 @@ struct ata_port {
 	struct work_struct	port_task;
 
 	unsigned int		hsm_task_state;
-	unsigned long		pio_task_timeout;
 
 	u32			msg_enable;
 	struct list_head	eh_done_q;
 
 	void			*private_data;
+
+	u8			sector_buf[ATA_SECT_SIZE]; /* owned by EH */
 };
 
 struct ata_port_operations {
@@ -458,7 +527,15 @@ struct ata_port_operations {
 	void (*qc_prep) (struct ata_queued_cmd *qc);
 	unsigned int (*qc_issue) (struct ata_queued_cmd *qc);
 
-	void (*eng_timeout) (struct ata_port *ap);
+	/* Error handlers.  ->error_handler overrides ->eng_timeout and
+	 * indicates that new-style EH is in place.
+	 */
+	void (*eng_timeout) (struct ata_port *ap); /* obsolete */
+
+	void (*freeze) (struct ata_port *ap);
+	void (*thaw) (struct ata_port *ap);
+	void (*error_handler) (struct ata_port *ap);
+	void (*post_internal_cmd) (struct ata_queued_cmd *qc);
 
 	irqreturn_t (*irq_handler)(int, void *, struct pt_regs *);
 	void (*irq_clear) (struct ata_port *);
@@ -504,7 +581,7 @@ extern void ata_port_probe(struct ata_port *);
 extern void __sata_phy_reset(struct ata_port *ap);
 extern void sata_phy_reset(struct ata_port *ap);
 extern void ata_bus_reset(struct ata_port *ap);
-extern int ata_set_sata_spd(struct ata_port *ap);
+extern int sata_set_spd(struct ata_port *ap);
 extern int ata_drive_probe_reset(struct ata_port *ap,
 			ata_probeinit_fn_t probeinit,
 			ata_reset_fn_t softreset, ata_reset_fn_t hardreset,
@@ -513,8 +590,7 @@ extern void ata_std_probeinit(struct ata_port *ap);
 extern int ata_std_softreset(struct ata_port *ap, unsigned int *classes);
 extern int sata_std_hardreset(struct ata_port *ap, unsigned int *class);
 extern void ata_std_postreset(struct ata_port *ap, unsigned int *classes);
-extern int ata_dev_revalidate(struct ata_port *ap, struct ata_device *dev,
-			      int post_reset);
+extern int ata_dev_revalidate(struct ata_device *dev, int post_reset);
 extern void ata_port_disable(struct ata_port *);
 extern void ata_std_ports(struct ata_ioports *ioaddr);
 #ifdef CONFIG_PCI
@@ -530,14 +606,18 @@ extern void ata_host_set_remove(struct ata_host_set *host_set);
 extern int ata_scsi_detect(struct scsi_host_template *sht);
 extern int ata_scsi_ioctl(struct scsi_device *dev, int cmd, void __user *arg);
 extern int ata_scsi_queuecmd(struct scsi_cmnd *cmd, void (*done)(struct scsi_cmnd *));
-extern void ata_eh_qc_complete(struct ata_queued_cmd *qc);
-extern void ata_eh_qc_retry(struct ata_queued_cmd *qc);
 extern int ata_scsi_release(struct Scsi_Host *host);
 extern unsigned int ata_host_intr(struct ata_port *ap, struct ata_queued_cmd *qc);
+extern int sata_scr_valid(struct ata_port *ap);
+extern int sata_scr_read(struct ata_port *ap, int reg, u32 *val);
+extern int sata_scr_write(struct ata_port *ap, int reg, u32 val);
+extern int sata_scr_write_flush(struct ata_port *ap, int reg, u32 val);
+extern int ata_port_online(struct ata_port *ap);
+extern int ata_port_offline(struct ata_port *ap);
 extern int ata_scsi_device_resume(struct scsi_device *);
 extern int ata_scsi_device_suspend(struct scsi_device *, pm_message_t state);
-extern int ata_device_resume(struct ata_port *, struct ata_device *);
-extern int ata_device_suspend(struct ata_port *, struct ata_device *, pm_message_t state);
+extern int ata_device_resume(struct ata_device *);
+extern int ata_device_suspend(struct ata_device *, pm_message_t state);
 extern int ata_ratelimit(void);
 extern unsigned int ata_busy_sleep(struct ata_port *ap,
 				   unsigned long timeout_pat,
@@ -582,16 +662,26 @@ extern void ata_bmdma_start (struct ata_queued_cmd *qc);
 extern void ata_bmdma_stop(struct ata_queued_cmd *qc);
 extern u8   ata_bmdma_status(struct ata_port *ap);
 extern void ata_bmdma_irq_clear(struct ata_port *ap);
-extern void __ata_qc_complete(struct ata_queued_cmd *qc);
-extern void ata_scsi_simulate(struct ata_port *ap, struct ata_device *dev,
-			      struct scsi_cmnd *cmd,
+extern void ata_bmdma_freeze(struct ata_port *ap);
+extern void ata_bmdma_thaw(struct ata_port *ap);
+extern void ata_bmdma_drive_eh(struct ata_port *ap,
+			       ata_reset_fn_t softreset,
+			       ata_reset_fn_t hardreset,
+			       ata_postreset_fn_t postreset);
+extern void ata_bmdma_error_handler(struct ata_port *ap);
+extern void ata_bmdma_post_internal_cmd(struct ata_queued_cmd *qc);
+extern void ata_qc_complete(struct ata_queued_cmd *qc);
+extern int ata_qc_complete_multiple(struct ata_port *ap, u32 qc_active,
+				    void (*finish_qc)(struct ata_queued_cmd *));
+extern void ata_scsi_simulate(struct ata_device *dev, struct scsi_cmnd *cmd,
 			      void (*done)(struct scsi_cmnd *));
 extern int ata_std_bios_param(struct scsi_device *sdev,
 			      struct block_device *bdev,
 			      sector_t capacity, int geom[]);
 extern int ata_scsi_slave_config(struct scsi_device *sdev);
-extern struct ata_device *ata_dev_pair(struct ata_port *ap, 
-				       struct ata_device *adev);
+extern int ata_scsi_change_queue_depth(struct scsi_device *sdev,
+				       int queue_depth);
+extern struct ata_device *ata_dev_pair(struct ata_device *adev);
 
 /*
  * Timing helpers
@@ -641,10 +731,46 @@ extern unsigned long ata_pci_default_filter(const struct ata_port *, struct ata_
  * EH
  */
 extern void ata_eng_timeout(struct ata_port *ap);
+
+extern void ata_port_schedule_eh(struct ata_port *ap);
+extern int ata_port_abort(struct ata_port *ap);
+extern int ata_port_freeze(struct ata_port *ap);
+
+extern void ata_eh_freeze_port(struct ata_port *ap);
+extern void ata_eh_thaw_port(struct ata_port *ap);
+
 extern void ata_eh_qc_complete(struct ata_queued_cmd *qc);
 extern void ata_eh_qc_retry(struct ata_queued_cmd *qc);
 
+extern void ata_do_eh(struct ata_port *ap, ata_reset_fn_t softreset,
+		      ata_reset_fn_t hardreset, ata_postreset_fn_t postreset);
+
+/*
+ * printk helpers
+ */
+#define ata_port_printk(ap, lv, fmt, args...) \
+	printk(lv"ata%u: "fmt, (ap)->id , ##args)
+
+#define ata_dev_printk(dev, lv, fmt, args...) \
+	printk(lv"ata%u.%02u: "fmt, (dev)->ap->id, (dev)->devno , ##args)
+
+/*
+ * ata_eh_info helpers
+ */
+#define ata_ehi_push_desc(ehi, fmt, args...) do { \
+	(ehi)->desc_len += scnprintf((ehi)->desc + (ehi)->desc_len, \
+				     ATA_EH_DESC_LEN - (ehi)->desc_len, \
+				     fmt , ##args); \
+} while (0)
+
+#define ata_ehi_clear_desc(ehi) do { \
+	(ehi)->desc[0] = '\0'; \
+	(ehi)->desc_len = 0; \
+} while (0)
 
+/*
+ * qc helpers
+ */
 static inline int
 ata_sg_is_last(struct scatterlist *sg, struct ata_queued_cmd *qc)
 {
@@ -687,6 +813,11 @@ static inline unsigned int ata_tag_valid(unsigned int tag)
 	return (tag < ATA_MAX_QUEUE) ? 1 : 0;
 }
 
+static inline unsigned int ata_tag_internal(unsigned int tag)
+{
+	return tag == ATA_MAX_QUEUE - 1;
+}
+
 static inline unsigned int ata_class_enabled(unsigned int class)
 {
 	return class == ATA_DEV_ATA || class == ATA_DEV_ATAPI;
@@ -795,20 +926,35 @@ static inline void ata_qc_set_polling(struct ata_queued_cmd *qc)
 	qc->tf.ctl |= ATA_NIEN;
 }
 
-static inline struct ata_queued_cmd *ata_qc_from_tag (struct ata_port *ap,
-						      unsigned int tag)
+static inline struct ata_queued_cmd *__ata_qc_from_tag(struct ata_port *ap,
+						       unsigned int tag)
 {
 	if (likely(ata_tag_valid(tag)))
 		return &ap->qcmd[tag];
 	return NULL;
 }
 
-static inline void ata_tf_init(struct ata_port *ap, struct ata_taskfile *tf, unsigned int device)
+static inline struct ata_queued_cmd *ata_qc_from_tag(struct ata_port *ap,
+						     unsigned int tag)
+{
+	struct ata_queued_cmd *qc = __ata_qc_from_tag(ap, tag);
+
+	if (unlikely(!qc) || !ap->ops->error_handler)
+		return qc;
+
+	if ((qc->flags & (ATA_QCFLAG_ACTIVE |
+			  ATA_QCFLAG_FAILED)) == ATA_QCFLAG_ACTIVE)
+		return qc;
+
+	return NULL;
+}
+
+static inline void ata_tf_init(struct ata_device *dev, struct ata_taskfile *tf)
 {
 	memset(tf, 0, sizeof(*tf));
 
-	tf->ctl = ap->ctl;
-	if (device == 0)
+	tf->ctl = dev->ap->ctl;
+	if (dev->devno == 0)
 		tf->device = ATA_DEVICE_OBS;
 	else
 		tf->device = ATA_DEVICE_OBS | ATA_DEV1;
@@ -823,26 +969,11 @@ static inline void ata_qc_reinit(struct ata_queued_cmd *qc)
 	qc->nbytes = qc->curbytes = 0;
 	qc->err_mask = 0;
 
-	ata_tf_init(qc->ap, &qc->tf, qc->dev->devno);
-}
+	ata_tf_init(qc->dev, &qc->tf);
 
-/**
- *	ata_qc_complete - Complete an active ATA command
- *	@qc: Command to complete
- *	@err_mask: ATA Status register contents
- *
- *	Indicate to the mid and upper layers that an ATA
- *	command has completed, with either an ok or not-ok status.
- *
- *	LOCKING:
- *	spin_lock_irqsave(host_set lock)
- */
-static inline void ata_qc_complete(struct ata_queued_cmd *qc)
-{
-	if (unlikely(qc->flags & ATA_QCFLAG_EH_SCHEDULED))
-		return;
-
-	__ata_qc_complete(qc);
+	/* init result_tf such that it indicates normal completion */
+	qc->result_tf.command = ATA_DRDY;
+	qc->result_tf.feature = 0;
 }
 
 /**
@@ -921,28 +1052,6 @@ static inline u8 ata_irq_ack(struct ata_port *ap, unsigned int chk_drq)
 	return status;
 }
 
-static inline u32 scr_read(struct ata_port *ap, unsigned int reg)
-{
-	return ap->ops->scr_read(ap, reg);
-}
-
-static inline void scr_write(struct ata_port *ap, unsigned int reg, u32 val)
-{
-	ap->ops->scr_write(ap, reg, val);
-}
-
-static inline void scr_write_flush(struct ata_port *ap, unsigned int reg,
-				   u32 val)
-{
-	ap->ops->scr_write(ap, reg, val);
-	(void) ap->ops->scr_read(ap, reg);
-}
-
-static inline unsigned int sata_dev_present(struct ata_port *ap)
-{
-	return ((scr_read(ap, SCR_STATUS) & 0xf) == 0x3) ? 1 : 0;
-}
-
 static inline int ata_try_flush_cache(const struct ata_device *dev)
 {
 	return ata_id_wcache_enabled(dev->id) ||

include/scsi/scsi_cmnd.h

@@ -151,5 +151,6 @@ extern struct scsi_cmnd *scsi_get_command(struct scsi_device *, gfp_t);
 extern void scsi_put_command(struct scsi_cmnd *);
 extern void scsi_io_completion(struct scsi_cmnd *, unsigned int, unsigned int);
 extern void scsi_finish_command(struct scsi_cmnd *cmd);
+extern void scsi_req_abort_cmd(struct scsi_cmnd *cmd);
 
 #endif /* _SCSI_SCSI_CMND_H */

include/scsi/scsi_eh.h

@@ -35,6 +35,7 @@ static inline int scsi_sense_valid(struct scsi_sense_hdr *sshdr)
 }
 
 
+extern void scsi_eh_wakeup(struct Scsi_Host *shost);
 extern void scsi_eh_finish_cmd(struct scsi_cmnd *scmd,
 			       struct list_head *done_q);
 extern void scsi_eh_flush_done_q(struct list_head *done_q);

include/scsi/scsi_host.h

@@ -472,6 +472,7 @@ struct Scsi_Host {
 	 */
 	unsigned int host_busy;		   /* commands actually active on low-level */
 	unsigned int host_failed;	   /* commands that failed. */
+	unsigned int host_eh_scheduled;    /* EH scheduled without command */
     
 	unsigned short host_no;  /* Used for IOCTL_GET_IDLUN, /proc/scsi et al. */
 	int resetting; /* if set, it means that last_reset is a valid value */