[PATCH] libata: Fix the HSM error_mask mapping (was: Re: libata-tj and SMART)

Fix the HSM error_mask mapping.

Changes:
- Better mapping in ac_err_mask()
- In HSM_ST_FIRST ans HSM_ST state, check ATA_ERR|ATA_DF and map it to AC_ERR_DEV instead of AC_ERR_HSM.
- In HSM_ST_FIRST and HSM_ST state, map DRQ=1 ERR=1 to AC_ERR_HSM.
- For PIO data in and DRQ=1 ERR=1, add check after the junk data block is read.

Signed-off-by: Albert Lee <albertcc@tw.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

drivers/scsi/libata-core.c

@@ -4009,9 +4009,15 @@ fsm_start:
 		poll_next = (qc->tf.flags & ATA_TFLAG_POLLING);
 
 		/* check device status */
-		if (unlikely((status & (ATA_BUSY | ATA_DRQ)) != ATA_DRQ)) {
-			/* Wrong status. Let EH handle this */
-			qc->err_mask |= AC_ERR_HSM;
+		if (unlikely((status & ATA_DRQ) == 0)) {
+			/* handle BSY=0, DRQ=0 as error */
+			if (likely(status & (ATA_ERR | ATA_DF)))
+				/* device stops HSM for abort/error */
+				qc->err_mask |= AC_ERR_DEV;
+			else
+				/* HSM violation. Let EH handle this */
+				qc->err_mask |= AC_ERR_HSM;
+
 			ap->hsm_task_state = HSM_ST_ERR;
 			goto fsm_start;
 		}
@@ -4025,7 +4031,7 @@ fsm_start:
 		if (unlikely(status & (ATA_ERR | ATA_DF))) {
 			printk(KERN_WARNING "ata%d: DRQ=1 with device error, dev_stat 0x%X\n",
 			       ap->id, status);
-			qc->err_mask |= AC_ERR_DEV;
+			qc->err_mask |= AC_ERR_HSM;
 			ap->hsm_task_state = HSM_ST_ERR;
 			goto fsm_start;
 		}
@@ -4067,7 +4073,9 @@ fsm_start:
 		if (qc->tf.protocol == ATA_PROT_ATAPI) {
 			/* ATAPI PIO protocol */
 			if ((status & ATA_DRQ) == 0) {
-				/* no more data to transfer */
+				/* No more data to transfer or device error.
+				 * Device error will be tagged in HSM_ST_LAST.
+				 */
 				ap->hsm_task_state = HSM_ST_LAST;
 				goto fsm_start;
 			}
@@ -4081,7 +4089,7 @@ fsm_start:
 			if (unlikely(status & (ATA_ERR | ATA_DF))) {
 				printk(KERN_WARNING "ata%d: DRQ=1 with device error, dev_stat 0x%X\n",
 				       ap->id, status);
-				qc->err_mask |= AC_ERR_DEV;
+				qc->err_mask |= AC_ERR_HSM;
 				ap->hsm_task_state = HSM_ST_ERR;
 				goto fsm_start;
 			}
@@ -4096,7 +4104,13 @@ fsm_start:
 			/* ATA PIO protocol */
 			if (unlikely((status & ATA_DRQ) == 0)) {
 				/* handle BSY=0, DRQ=0 as error */
-				qc->err_mask |= AC_ERR_HSM;
+				if (likely(status & (ATA_ERR | ATA_DF)))
+					/* device stops HSM for abort/error */
+					qc->err_mask |= AC_ERR_DEV;
+				else
+					/* HSM violation. Let EH handle this */
+					qc->err_mask |= AC_ERR_HSM;
+
 				ap->hsm_task_state = HSM_ST_ERR;
 				goto fsm_start;
 			}
@@ -4121,6 +4135,9 @@ fsm_start:
 					status = ata_wait_idle(ap);
 				}
 
+				if (status & (ATA_BUSY | ATA_DRQ))
+					qc->err_mask |= AC_ERR_HSM;
+
 				/* ata_pio_sectors() might change the
 				 * state to HSM_ST_LAST. so, the state
 				 * is changed after ata_pio_sectors().

include/linux/libata.h

@@ -1062,7 +1062,7 @@ static inline int ata_try_flush_cache(const struct ata_device *dev)
 
 static inline unsigned int ac_err_mask(u8 status)
 {
-	if (status & ATA_BUSY)
+	if (status & (ATA_BUSY | ATA_DRQ))
 		return AC_ERR_HSM;
 	if (status & (ATA_ERR | ATA_DF))
 		return AC_ERR_DEV;