Skip to contentSkip to conversation

ALUA device handler update, part II

PATCH Superseded Hannes Reinecke 3 participants 38 messages Jan 12, 2016 – Jan 13, 2016 SCSI

v2 of 8 Superseded Latest is v8 →

lore
Share & tools
lore
#0 lore raw

v2 cover ALUA device handler update, part II

(resending as I failed to include linux-scsi :-( )

Hi all,

as promised here is now the second part of my ALUA device handler update. This contains a major rework of the ALUA device handler as execution is moved onto a workqueue. This has the advantage that we avoid having to do multiple calls to the same LUN (as happens frequently when failing over a LUN with several paths) and finally retries are handled correctly. As some arrays are only capable of handling one STPG at a time I've added a second hardware handler parameter which then uses a singlethreaded workqueue, thereby effectively synchronize STPG handling. Thanks to Bart for this suggestion.

As usual, comments and reviews are welcome.

Changes to v1:
- Include reviews from hch
- Switch to hardware handler parameter instead of module option

Hannes Reinecke (22):
scsi_dh_alua: Pass buffer as function argument
scsi_dh_alua: separate out alua_stpg()
scsi_dh_alua: Make stpg synchronous
scsi_dh_alua: call alua_rtpg() if stpg fails
scsi_dh_alua: switch to scsi_execute_req_flags()
scsi_dh_alua: allocate RTPG buffer separately
scsi_dh_alua: Use separate alua_port_group structure
scsi_dh_alua: use unique device id
scsi_dh_alua: simplify alua_initialize()
revert commit a8e5a2d593cb ("[SCSI] scsi_dh_alua: ALUA handler attach
should succeed while TPG is transitioning")
scsi_dh_alua: move optimize_stpg evaluation
scsi_dh_alua: remove 'rel_port' from alua_dh_data structure
scsi_dh_alua: Use workqueue for RTPG
scsi_dh_alua: Allow workqueue to run synchronously
scsi_dh_alua: Recheck state on unit attention
scsi_dh_alua: update all port states
scsi_dh_alua: Send TEST UNIT READY to poll for transitioning
scsi_dh: add 'rescan' callback
scsi: Add 'access_state' attribute
scsi_dh_alua: use common definitions for ALUA state
scsi_dh_alua: update 'access_state' field
scsi_dh_alua: Update version to 2.0

 drivers/scsi/device_handler/scsi_dh_alua.c | 983 ++++++++++++++++++++---------
 drivers/scsi/scsi_lib.c                    |   1 +
 drivers/scsi/scsi_scan.c                   |   9 +-
 drivers/scsi/scsi_sysfs.c                  |  49 ++
 include/scsi/scsi_device.h                 |   1 +
 include/scsi/scsi_dh.h                     |   2 +
 include/scsi/scsi_proto.h                  |  13 +
 7 files changed, 747 insertions(+), 311 deletions(-)
Signature · 1 line
1.8.5.6
#1 lore raw

PATCH v2 1/22 scsi_dh_alua: Pass buffer as function argument

Pass in the buffer as a function argument for submit_rtpg().

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +7 -6
index 5a328bf..df71e85 100644
@@ -135,12 +135,13 @@ static struct request *get_alua_req(struct scsi_device *sdev,
  * submit_rtpg - Issue a REPORT TARGET GROUP STATES command
  * @sdev: sdev the command should be sent to
  */
-static unsigned submit_rtpg(struct scsi_device *sdev, struct alua_dh_data *h)
+static unsigned submit_rtpg(struct scsi_device *sdev, unsigned char *buff,
+			    int bufflen, unsigned char *sense, int flags)
 {
 	struct request *rq;
 	int err = 0;
 
-	rq = get_alua_req(sdev, h->buff, h->bufflen, READ);
+	rq = get_alua_req(sdev, buff, bufflen, READ);
 	if (!rq) {
 		err = DRIVER_BUSY << 24;
 		goto done;
@@ -148,14 +149,14 @@ static unsigned submit_rtpg(struct scsi_device *sdev, struct alua_dh_data *h)
 
 	/* Prepare the command. */
 	rq->cmd[0] = MAINTENANCE_IN;
-	if (!(h->flags & ALUA_RTPG_EXT_HDR_UNSUPP))
+	if (!(flags & ALUA_RTPG_EXT_HDR_UNSUPP))
 		rq->cmd[1] = MI_REPORT_TARGET_PGS | MI_EXT_HDR_PARAM_FMT;
 	else
 		rq->cmd[1] = MI_REPORT_TARGET_PGS;
-	put_unaligned_be32(h->bufflen, &rq->cmd[6]);
+	put_unaligned_be32(bufflen, &rq->cmd[6]);
 	rq->cmd_len = COMMAND_SIZE(MAINTENANCE_IN);
 
-	rq->sense = h->sense;
+	rq->sense = sense;
 	memset(rq->sense, 0, SCSI_SENSE_BUFFERSIZE);
 	rq->sense_len = 0;
 
@@ -446,7 +447,7 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 		expiry = round_jiffies_up(jiffies + h->transition_tmo * HZ);
 
  retry:
-	retval = submit_rtpg(sdev, h);
+	retval = submit_rtpg(sdev, h->buff, h->bufflen, h->sense, h->flags);
 	if (retval) {
 		if (!scsi_normalize_sense(h->sense, SCSI_SENSE_BUFFERSIZE,
 					  &sense_hdr)) {
Signature · 1 line
1.8.5.6
#2 lore raw

PATCH v2 2/22 scsi_dh_alua: separate out alua_stpg()

In reply to Hannes Reinecke

Separate out SET TARGET PORT GROUP functionality into a separate function alua_stpg().

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 95 +++++++++++++++++++-----------
 1 file changed, 61 insertions(+), 34 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +61 -34
index df71e85..7c66e4a 100644
@@ -579,6 +579,65 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 }
 
 /*
+ * alua_stpg - Issue a SET TARGET GROUP STATES command
+ *
+ * Issue a SET TARGET GROUP STATES command and evaluate the
+ * response. Returns SCSI_DH_RETRY if the target port group
+ * state is found to be in 'transitioning'.
+ * If SCSI_DH_OK is returned the passed-in 'fn' function
+ * this function will take care of executing 'fn'.
+ * Otherwise 'fn' should be executed by the caller with the
+ * returned error code.
+ */
+static unsigned alua_stpg(struct scsi_device *sdev, struct alua_dh_data *h,
+			  activate_complete fn, void *data)
+{
+	int err = SCSI_DH_OK;
+	int stpg = 0;
+
+	if (!(h->tpgs & TPGS_MODE_EXPLICIT))
+		/* Only implicit ALUA supported */
+		goto out;
+
+	switch (h->state) {
+	case TPGS_STATE_NONOPTIMIZED:
+		stpg = 1;
+		if ((h->flags & ALUA_OPTIMIZE_STPG) &&
+		    !h->pref &&
+		    (h->tpgs & TPGS_MODE_IMPLICIT))
+			stpg = 0;
+		break;
+	case TPGS_STATE_STANDBY:
+	case TPGS_STATE_UNAVAILABLE:
+		stpg = 1;
+		break;
+	case TPGS_STATE_OFFLINE:
+		err = SCSI_DH_IO;
+		break;
+	case TPGS_STATE_TRANSITIONING:
+		err = SCSI_DH_RETRY;
+		break;
+	default:
+		break;
+	}
+
+	if (stpg) {
+		h->callback_fn = fn;
+		h->callback_data = data;
+		err = submit_stpg(h);
+		if (err != SCSI_DH_OK)
+			h->callback_fn = h->callback_data = NULL;
+		else
+			fn = NULL;
+	}
+out:
+	if (fn)
+		fn(data, err);
+
+	return err;
+}
+
+/*
  * alua_initialize - Initialize ALUA state
  * @sdev: the device to be initialized
  *
@@ -655,7 +714,6 @@ static int alua_activate(struct scsi_device *sdev,
 {
 	struct alua_dh_data *h = sdev->handler_data;
 	int err = SCSI_DH_OK;
-	int stpg = 0;
 
 	err = alua_rtpg(sdev, h, 1);
 	if (err != SCSI_DH_OK)
@@ -664,41 +722,10 @@ static int alua_activate(struct scsi_device *sdev,
 	if (optimize_stpg)
 		h->flags |= ALUA_OPTIMIZE_STPG;
 
-	if (h->tpgs & TPGS_MODE_EXPLICIT) {
-		switch (h->state) {
-		case TPGS_STATE_NONOPTIMIZED:
-			stpg = 1;
-			if ((h->flags & ALUA_OPTIMIZE_STPG) &&
-			    (!h->pref) &&
-			    (h->tpgs & TPGS_MODE_IMPLICIT))
-				stpg = 0;
-			break;
-		case TPGS_STATE_STANDBY:
-		case TPGS_STATE_UNAVAILABLE:
-			stpg = 1;
-			break;
-		case TPGS_STATE_OFFLINE:
-			err = SCSI_DH_IO;
-			break;
-		case TPGS_STATE_TRANSITIONING:
-			err = SCSI_DH_RETRY;
-			break;
-		default:
-			break;
-		}
-	}
-
-	if (stpg) {
-		h->callback_fn = fn;
-		h->callback_data = data;
-		err = submit_stpg(h);
-		if (err == SCSI_DH_OK)
-			return 0;
-		h->callback_fn = h->callback_data = NULL;
-	}
+	err = alua_stpg(sdev, h, fn, data);
 
 out:
-	if (fn)
+	if (err != SCSI_DH_OK && fn)
 		fn(data, err);
 	return 0;
 }
Signature · 1 line
1.8.5.6
#3 lore raw

PATCH v2 3/22 scsi_dh_alua: Make stpg synchronous

In reply to Hannes Reinecke

The 'activate_complete' function needs to be executed after stpg has finished, so we can as well execute stpg synchronously and call the function directly.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 156 ++++++++++-------------------
 1 file changed, 55 insertions(+), 101 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +55 -101
index 7c66e4a..609691f 100644
@@ -169,81 +169,28 @@ done:
 }
 
 /*
- * stpg_endio - Evaluate SET TARGET GROUP STATES
- * @sdev: the device to be evaluated
- * @state: the new target group state
- *
- * Evaluate a SET TARGET GROUP STATES command response.
- */
-static void stpg_endio(struct request *req, int error)
-{
-	struct alua_dh_data *h = req->end_io_data;
-	struct scsi_sense_hdr sense_hdr;
-	unsigned err = SCSI_DH_OK;
-
-	if (host_byte(req->errors) != DID_OK ||
-	    msg_byte(req->errors) != COMMAND_COMPLETE) {
-		err = SCSI_DH_IO;
-		goto done;
-	}
-
-	if (scsi_normalize_sense(h->sense, SCSI_SENSE_BUFFERSIZE,
-				 &sense_hdr)) {
-		if (sense_hdr.sense_key == NOT_READY &&
-		    sense_hdr.asc == 0x04 && sense_hdr.ascq == 0x0a) {
-			/* ALUA state transition already in progress */
-			err = SCSI_DH_OK;
-			goto done;
-		}
-		if (sense_hdr.sense_key == UNIT_ATTENTION) {
-			err = SCSI_DH_RETRY;
-			goto done;
-		}
-		sdev_printk(KERN_INFO, h->sdev, "%s: stpg failed\n",
-			    ALUA_DH_NAME);
-		scsi_print_sense_hdr(h->sdev, ALUA_DH_NAME, &sense_hdr);
-		err = SCSI_DH_IO;
-	} else if (error)
-		err = SCSI_DH_IO;
-
-	if (err == SCSI_DH_OK) {
-		h->state = TPGS_STATE_OPTIMIZED;
-		sdev_printk(KERN_INFO, h->sdev,
-			    "%s: port group %02x switched to state %c\n",
-			    ALUA_DH_NAME, h->group_id,
-			    print_alua_state(h->state));
-	}
-done:
-	req->end_io_data = NULL;
-	__blk_put_request(req->q, req);
-	if (h->callback_fn) {
-		h->callback_fn(h->callback_data, err);
-		h->callback_fn = h->callback_data = NULL;
-	}
-	return;
-}
-
-/*
  * submit_stpg - Issue a SET TARGET GROUP STATES command
  *
  * Currently we're only setting the current target port group state
  * to 'active/optimized' and let the array firmware figure out
  * the states of the remaining groups.
  */
-static unsigned submit_stpg(struct alua_dh_data *h)
+static unsigned submit_stpg(struct scsi_device *sdev, int group_id,
+			    unsigned char *sense)
 {
 	struct request *rq;
+	unsigned char stpg_data[8];
 	int stpg_len = 8;
-	struct scsi_device *sdev = h->sdev;
+	int err = 0;
 
 	/* Prepare the data buffer */
-	memset(h->buff, 0, stpg_len);
-	h->buff[4] = TPGS_STATE_OPTIMIZED & 0x0f;
-	put_unaligned_be16(h->group_id, &h->buff[6]);
+	memset(stpg_data, 0, stpg_len);
+	stpg_data[4] = TPGS_STATE_OPTIMIZED & 0x0f;
+	put_unaligned_be16(group_id, &stpg_data[6]);
 
-	rq = get_alua_req(sdev, h->buff, stpg_len, WRITE);
+	rq = get_alua_req(sdev, stpg_data, stpg_len, WRITE);
 	if (!rq)
-		return SCSI_DH_RES_TEMP_UNAVAIL;
+		return DRIVER_BUSY << 24;
 
 	/* Prepare the command. */
 	rq->cmd[0] = MAINTENANCE_OUT;
@@ -251,13 +198,17 @@ static unsigned submit_stpg(struct alua_dh_data *h)
 	put_unaligned_be32(stpg_len, &rq->cmd[6]);
 	rq->cmd_len = COMMAND_SIZE(MAINTENANCE_OUT);
 
-	rq->sense = h->sense;
+	rq->sense = sense;
 	memset(rq->sense, 0, SCSI_SENSE_BUFFERSIZE);
 	rq->sense_len = 0;
-	rq->end_io_data = h;
 
-	blk_execute_rq_nowait(rq->q, NULL, rq, 1, stpg_endio);
-	return SCSI_DH_OK;
+	blk_execute_rq(rq->q, NULL, rq, 1);
+	if (rq->errors)
+		err = rq->errors;
+
+	blk_put_request(rq);
+
+	return err;
 }
 
 /*
@@ -582,59 +533,63 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
  * alua_stpg - Issue a SET TARGET GROUP STATES command
  *
  * Issue a SET TARGET GROUP STATES command and evaluate the
- * response. Returns SCSI_DH_RETRY if the target port group
- * state is found to be in 'transitioning'.
- * If SCSI_DH_OK is returned the passed-in 'fn' function
- * this function will take care of executing 'fn'.
- * Otherwise 'fn' should be executed by the caller with the
- * returned error code.
+ * response. Returns SCSI_DH_RETRY per default to trigger
+ * a re-evaluation of the target group state or SCSI_DH_OK
+ * if no further action needs to be taken.
  */
-static unsigned alua_stpg(struct scsi_device *sdev, struct alua_dh_data *h,
-			  activate_complete fn, void *data)
+static unsigned alua_stpg(struct scsi_device *sdev, struct alua_dh_data *h)
 {
-	int err = SCSI_DH_OK;
-	int stpg = 0;
-
-	if (!(h->tpgs & TPGS_MODE_EXPLICIT))
-		/* Only implicit ALUA supported */
-		goto out;
+	int retval;
+	struct scsi_sense_hdr sense_hdr;
 
+	if (!(h->tpgs & TPGS_MODE_EXPLICIT)) {
+		/* Only implicit ALUA supported, retry */
+		return SCSI_DH_RETRY;
+	}
 	switch (h->state) {
+	case TPGS_STATE_OPTIMIZED:
+		return SCSI_DH_OK;
 	case TPGS_STATE_NONOPTIMIZED:
-		stpg = 1;
 		if ((h->flags & ALUA_OPTIMIZE_STPG) &&
 		    !h->pref &&
 		    (h->tpgs & TPGS_MODE_IMPLICIT))
-			stpg = 0;
+			return SCSI_DH_OK;
 		break;
 	case TPGS_STATE_STANDBY:
 	case TPGS_STATE_UNAVAILABLE:
-		stpg = 1;
 		break;
 	case TPGS_STATE_OFFLINE:
-		err = SCSI_DH_IO;
+		return SCSI_DH_IO;
 		break;
 	case TPGS_STATE_TRANSITIONING:
-		err = SCSI_DH_RETRY;
 		break;
 	default:
+		sdev_printk(KERN_INFO, sdev,
+			    "%s: stpg failed, unhandled TPGS state %d",
+			    ALUA_DH_NAME, h->state);
+		return SCSI_DH_NOSYS;
 		break;
 	}
+	/* Set state to transitioning */
+	h->state = TPGS_STATE_TRANSITIONING;
+	retval = submit_stpg(sdev, h->group_id, h->sense);
 
-	if (stpg) {
-		h->callback_fn = fn;
-		h->callback_data = data;
-		err = submit_stpg(h);
-		if (err != SCSI_DH_OK)
-			h->callback_fn = h->callback_data = NULL;
-		else
-			fn = NULL;
+	if (retval) {
+		if (!scsi_normalize_sense(h->sense, SCSI_SENSE_BUFFERSIZE,
+					  &sense_hdr)) {
+			sdev_printk(KERN_INFO, sdev,
+				    "%s: stpg failed, result %d",
+				    ALUA_DH_NAME, retval);
+			if (driver_byte(retval) == DRIVER_BUSY)
+				return SCSI_DH_DEV_TEMP_BUSY;
+		} else {
+			sdev_printk(KERN_INFO, h->sdev, "%s: stpg failed\n",
+				    ALUA_DH_NAME);
+			scsi_print_sense_hdr(sdev, ALUA_DH_NAME, &sense_hdr);
+		}
 	}
-out:
-	if (fn)
-		fn(data, err);
-
-	return err;
+	/* Retry RTPG */
+	return SCSI_DH_RETRY;
 }
 
 /*
@@ -722,10 +677,9 @@ static int alua_activate(struct scsi_device *sdev,
 	if (optimize_stpg)
 		h->flags |= ALUA_OPTIMIZE_STPG;
 
-	err = alua_stpg(sdev, h, fn, data);
-
+	err = alua_stpg(sdev, h);
 out:
-	if (err != SCSI_DH_OK && fn)
+	if (fn)
 		fn(data, err);
 	return 0;
 }
Signature · 1 line
1.8.5.6
#4 lore raw

PATCH v2 4/22 scsi_dh_alua: call alua_rtpg() if stpg fails

In reply to Hannes Reinecke

If the call to SET TARGET PORT GROUPS fails we have no idea what state the array is left in, so we need to issue a call to REPORT TARGET PORT GROUPS in these cases.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/scsi/device_handler/scsi_dh_alua.c +2 -0
index 609691f..ba3d23e 100644
@@ -678,6 +678,8 @@ static int alua_activate(struct scsi_device *sdev,
 		h->flags |= ALUA_OPTIMIZE_STPG;
 
 	err = alua_stpg(sdev, h);
+	if (err == SCSI_DH_RETRY)
+		err = alua_rtpg(sdev, h, 1);
 out:
 	if (fn)
 		fn(data, err);
Signature · 1 line
1.8.5.6
#5 lore raw

PATCH v2 5/22 scsi_dh_alua: switch to scsi_execute_req_flags()

In reply to Hannes Reinecke

All commands are issued synchronously, so no need to open-code scsi_execute_req_flags() anymore. And we can get rid of the static sense code structure element. scsi_execute_req_flags() will be setting REQ_QUIET and REQ_PREEMPT, but that is perfectly fine as we're evaluating and logging any errors ourselves and we really need to send the command even if the device is quiesced.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 125 +++++++++--------------------
 1 file changed, 36 insertions(+), 89 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +36 -89
index ba3d23e..063c03a 100644
@@ -75,7 +75,6 @@ struct alua_dh_data {
 	unsigned char		*buff;
 	int			bufflen;
 	unsigned char		transition_tmo;
-	unsigned char		sense[SCSI_SENSE_BUFFERSIZE];
 	struct scsi_device	*sdev;
 	activate_complete	callback_fn;
 	void			*callback_data;
@@ -101,71 +100,30 @@ static int realloc_buffer(struct alua_dh_data *h, unsigned len)
 	return 0;
 }
 
-static struct request *get_alua_req(struct scsi_device *sdev,
-				    void *buffer, unsigned buflen, int rw)
-{
-	struct request *rq;
-	struct request_queue *q = sdev->request_queue;
-
-	rq = blk_get_request(q, rw, GFP_NOIO);
-
-	if (IS_ERR(rq)) {
-		sdev_printk(KERN_INFO, sdev,
-			    "%s: blk_get_request failed\n", __func__);
-		return NULL;
-	}
-	blk_rq_set_block_pc(rq);
-
-	if (buflen && blk_rq_map_kern(q, rq, buffer, buflen, GFP_NOIO)) {
-		blk_put_request(rq);
-		sdev_printk(KERN_INFO, sdev,
-			    "%s: blk_rq_map_kern failed\n", __func__);
-		return NULL;
-	}
-
-	rq->cmd_flags |= REQ_FAILFAST_DEV | REQ_FAILFAST_TRANSPORT |
-			 REQ_FAILFAST_DRIVER;
-	rq->retries = ALUA_FAILOVER_RETRIES;
-	rq->timeout = ALUA_FAILOVER_TIMEOUT * HZ;
-
-	return rq;
-}
-
 /*
  * submit_rtpg - Issue a REPORT TARGET GROUP STATES command
  * @sdev: sdev the command should be sent to
  */
-static unsigned submit_rtpg(struct scsi_device *sdev, unsigned char *buff,
-			    int bufflen, unsigned char *sense, int flags)
+static int submit_rtpg(struct scsi_device *sdev, unsigned char *buff,
+		       int bufflen, struct scsi_sense_hdr *sshdr, int flags)
 {
-	struct request *rq;
-	int err = 0;
-
-	rq = get_alua_req(sdev, buff, bufflen, READ);
-	if (!rq) {
-		err = DRIVER_BUSY << 24;
-		goto done;
-	}
+	u8 cdb[COMMAND_SIZE(MAINTENANCE_IN)];
+	int req_flags = REQ_FAILFAST_DEV | REQ_FAILFAST_TRANSPORT |
+		REQ_FAILFAST_DRIVER;
 
 	/* Prepare the command. */
-	rq->cmd[0] = MAINTENANCE_IN;
+	memset(cdb, 0x0, COMMAND_SIZE(MAINTENANCE_IN));
+	cdb[0] = MAINTENANCE_IN;
 	if (!(flags & ALUA_RTPG_EXT_HDR_UNSUPP))
-		rq->cmd[1] = MI_REPORT_TARGET_PGS | MI_EXT_HDR_PARAM_FMT;
+		cdb[1] = MI_REPORT_TARGET_PGS | MI_EXT_HDR_PARAM_FMT;
 	else
-		rq->cmd[1] = MI_REPORT_TARGET_PGS;
-	put_unaligned_be32(bufflen, &rq->cmd[6]);
-	rq->cmd_len = COMMAND_SIZE(MAINTENANCE_IN);
-
-	rq->sense = sense;
-	memset(rq->sense, 0, SCSI_SENSE_BUFFERSIZE);
-	rq->sense_len = 0;
-
-	blk_execute_rq(rq->q, NULL, rq, 1);
-	if (rq->errors)
-		err = rq->errors;
-	blk_put_request(rq);
-done:
-	return err;
+		cdb[1] = MI_REPORT_TARGET_PGS;
+	put_unaligned_be32(bufflen, &cdb[6]);
+
+	return scsi_execute_req_flags(sdev, cdb, DMA_FROM_DEVICE,
+				      buff, bufflen, sshdr,
+				      ALUA_FAILOVER_TIMEOUT * HZ,
+				      ALUA_FAILOVER_RETRIES, NULL, req_flags);
 }
 
 /*
@@ -175,40 +133,30 @@ done:
  * to 'active/optimized' and let the array firmware figure out
  * the states of the remaining groups.
  */
-static unsigned submit_stpg(struct scsi_device *sdev, int group_id,
-			    unsigned char *sense)
+static int submit_stpg(struct scsi_device *sdev, int group_id,
+		       struct scsi_sense_hdr *sshdr)
 {
-	struct request *rq;
+	u8 cdb[COMMAND_SIZE(MAINTENANCE_OUT)];
 	unsigned char stpg_data[8];
 	int stpg_len = 8;
-	int err = 0;
+	int req_flags = REQ_FAILFAST_DEV | REQ_FAILFAST_TRANSPORT |
+		REQ_FAILFAST_DRIVER;
 
 	/* Prepare the data buffer */
 	memset(stpg_data, 0, stpg_len);
 	stpg_data[4] = TPGS_STATE_OPTIMIZED & 0x0f;
 	put_unaligned_be16(group_id, &stpg_data[6]);
 
-	rq = get_alua_req(sdev, stpg_data, stpg_len, WRITE);
-	if (!rq)
-		return DRIVER_BUSY << 24;
-
 	/* Prepare the command. */
-	rq->cmd[0] = MAINTENANCE_OUT;
-	rq->cmd[1] = MO_SET_TARGET_PGS;
-	put_unaligned_be32(stpg_len, &rq->cmd[6]);
-	rq->cmd_len = COMMAND_SIZE(MAINTENANCE_OUT);
-
-	rq->sense = sense;
-	memset(rq->sense, 0, SCSI_SENSE_BUFFERSIZE);
-	rq->sense_len = 0;
-
-	blk_execute_rq(rq->q, NULL, rq, 1);
-	if (rq->errors)
-		err = rq->errors;
-
-	blk_put_request(rq);
-
-	return err;
+	memset(cdb, 0x0, COMMAND_SIZE(MAINTENANCE_OUT));
+	cdb[0] = MAINTENANCE_OUT;
+	cdb[1] = MO_SET_TARGET_PGS;
+	put_unaligned_be32(stpg_len, &cdb[6]);
+
+	return scsi_execute_req_flags(sdev, cdb, DMA_TO_DEVICE,
+				      stpg_data, stpg_len,
+				      sshdr, ALUA_FAILOVER_TIMEOUT * HZ,
+				      ALUA_FAILOVER_RETRIES, NULL, req_flags);
 }
 
 /*
@@ -398,14 +346,14 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 		expiry = round_jiffies_up(jiffies + h->transition_tmo * HZ);
 
  retry:
-	retval = submit_rtpg(sdev, h->buff, h->bufflen, h->sense, h->flags);
+	retval = submit_rtpg(sdev, h->buff, h->bufflen, &sense_hdr, h->flags);
+
 	if (retval) {
-		if (!scsi_normalize_sense(h->sense, SCSI_SENSE_BUFFERSIZE,
-					  &sense_hdr)) {
+		if (!scsi_sense_valid(&sense_hdr)) {
 			sdev_printk(KERN_INFO, sdev,
 				    "%s: rtpg failed, result %d\n",
 				    ALUA_DH_NAME, retval);
-			if (driver_byte(retval) == DRIVER_BUSY)
+			if (driver_byte(retval) == DRIVER_ERROR)
 				return SCSI_DH_DEV_TEMP_BUSY;
 			return SCSI_DH_IO;
 		}
@@ -572,15 +520,14 @@ static unsigned alua_stpg(struct scsi_device *sdev, struct alua_dh_data *h)
 	}
 	/* Set state to transitioning */
 	h->state = TPGS_STATE_TRANSITIONING;
-	retval = submit_stpg(sdev, h->group_id, h->sense);
+	retval = submit_stpg(sdev, h->group_id, &sense_hdr);
 
 	if (retval) {
-		if (!scsi_normalize_sense(h->sense, SCSI_SENSE_BUFFERSIZE,
-					  &sense_hdr)) {
+		if (!scsi_sense_valid(&sense_hdr)) {
 			sdev_printk(KERN_INFO, sdev,
 				    "%s: stpg failed, result %d",
 				    ALUA_DH_NAME, retval);
-			if (driver_byte(retval) == DRIVER_BUSY)
+			if (driver_byte(retval) == DRIVER_ERROR)
 				return SCSI_DH_DEV_TEMP_BUSY;
 		} else {
 			sdev_printk(KERN_INFO, h->sdev, "%s: stpg failed\n",
Signature · 1 line
1.8.5.6
#6 lore raw

Re: [PATCHv2 05/22] scsi_dh_alua: switch to scsi_execute_req_flags()

reproduce: wget https://git.kernel.org/cgit/linux/kernel/git/wfg/lkp-tests.git/plain/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross # save the attached .config to linux build tree…

Hi Hannes,

[auto build test WARNING on scsi/for-next]
[also build test WARNING on next-20160112]
[cannot apply to v4.4]
[if your patch is applied to the wrong git tree, please drop us a note to help improving the system]

url:    https://github.com/0day-ci/linux/commits/Hannes-Reinecke/ALUA-device-handler-update-part-II/20160112-234244
base:   https://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi.git for-next
config: s390-allyesconfig (attached as .config)

reproduce:
wget https://git.kernel.org/cgit/linux/kernel/git/wfg/lkp-tests.git/plain/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=s390

All warnings (new ones prefixed by >>):

drivers/scsi/device_handler/scsi_dh_alua.c: In function 'alua_rtpg':

>> drivers/scsi/device_handler/scsi_dh_alua.c:478:1: warning: 'alua_rtpg' uses dynamic stack allocation
    }
    ^
   drivers/scsi/device_handler/scsi_dh_alua.c: In function 'alua_activate':
   drivers/scsi/device_handler/scsi_dh_alua.c:634:1: warning: 'alua_activate' uses dynamic stack allocation
    }
    ^

vim +/alua_rtpg +478 drivers/scsi/device_handler/scsi_dh_alua.c

a8e5a2d5 Stewart, Sean   2013-10-15  462  			err = SCSI_DH_OK;
a8e5a2d5 Stewart, Sean   2013-10-15  463  		}
a8e5a2d5 Stewart, Sean   2013-10-15  464  
a8e5a2d5 Stewart, Sean   2013-10-15  465  		/* Transitioning time exceeded, set port to standby */
69723d17 Hannes Reinecke 2010-09-24  466  		h->state = TPGS_STATE_STANDBY;
057ea7c9 Hannes Reinecke 2008-07-17  467  		break;
057ea7c9 Hannes Reinecke 2008-07-17  468  	case TPGS_STATE_OFFLINE:
e47f8976 Bart Van Assche 2012-08-24  469  		/* Path unusable */
057ea7c9 Hannes Reinecke 2008-07-17  470  		err = SCSI_DH_DEV_OFFLINED;
057ea7c9 Hannes Reinecke 2008-07-17  471  		break;
057ea7c9 Hannes Reinecke 2008-07-17  472  	default:
057ea7c9 Hannes Reinecke 2008-07-17  473  		/* Useable path if active */
057ea7c9 Hannes Reinecke 2008-07-17  474  		err = SCSI_DH_OK;
69723d17 Hannes Reinecke 2010-09-24  475  		break;
057ea7c9 Hannes Reinecke 2008-07-17  476  	}
057ea7c9 Hannes Reinecke 2008-07-17  477  	return err;
057ea7c9 Hannes Reinecke 2008-07-17 @478  }
057ea7c9 Hannes Reinecke 2008-07-17  479  
057ea7c9 Hannes Reinecke 2008-07-17  480  /*
d8af10dc Hannes Reinecke 2016-01-12  481   * alua_stpg - Issue a SET TARGET GROUP STATES command
d8af10dc Hannes Reinecke 2016-01-12  482   *
d8af10dc Hannes Reinecke 2016-01-12  483   * Issue a SET TARGET GROUP STATES command and evaluate the
0d223456 Hannes Reinecke 2016-01-12  484   * response. Returns SCSI_DH_RETRY per default to trigger
0d223456 Hannes Reinecke 2016-01-12  485   * a re-evaluation of the target group state or SCSI_DH_OK
0d223456 Hannes Reinecke 2016-01-12  486   * if no further action needs to be taken.

:::::: The code at line 478 was first introduced by commit
:::::: 057ea7c9683c3d684128cced796f03c179ecf1c2 [SCSI] scsi_dh: add generic SPC-3 alua handler

:::::: TO: Hannes Reinecke <[email protected]>
:::::: CC: James Bottomley <[email protected]>

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

#7 lore raw

PATCH v2 6/22 scsi_dh_alua: allocate RTPG buffer separately

In reply to Hannes Reinecke

The RTPG buffer will only evaluated within alua_rtpg(), so we can allocate it locally there and avoid having to put it into the global structure.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 57 ++++++++++++------------------
 1 file changed, 23 insertions(+), 34 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +23 -34
index 063c03a..49dea42 100644
@@ -56,7 +56,7 @@
 #define TPGS_MODE_IMPLICIT		0x1
 #define TPGS_MODE_EXPLICIT		0x2
 
-#define ALUA_INQUIRY_SIZE		36
+#define ALUA_RTPG_SIZE			128
 #define ALUA_FAILOVER_TIMEOUT		60
 #define ALUA_FAILOVER_RETRIES		5
 
@@ -71,9 +71,6 @@ struct alua_dh_data {
 	int			state;
 	int			pref;
 	unsigned		flags; /* used for optimizing STPG */
-	unsigned char		inq[ALUA_INQUIRY_SIZE];
-	unsigned char		*buff;
-	int			bufflen;
 	unsigned char		transition_tmo;
 	struct scsi_device	*sdev;
 	activate_complete	callback_fn;
@@ -85,21 +82,6 @@ struct alua_dh_data {
 
 static char print_alua_state(int);
 
-static int realloc_buffer(struct alua_dh_data *h, unsigned len)
-{
-	if (h->buff && h->buff != h->inq)
-		kfree(h->buff);
-
-	h->buff = kmalloc(len, GFP_NOIO);
-	if (!h->buff) {
-		h->buff = h->inq;
-		h->bufflen = ALUA_INQUIRY_SIZE;
-		return 1;
-	}
-	h->bufflen = len;
-	return 0;
-}
-
 /*
  * submit_rtpg - Issue a REPORT TARGET GROUP STATES command
  * @sdev: sdev the command should be sent to
@@ -333,8 +315,8 @@ static int alua_check_sense(struct scsi_device *sdev,
 static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_for_transition)
 {
 	struct scsi_sense_hdr sense_hdr;
-	int len, k, off, valid_states = 0;
-	unsigned char *ucp;
+	int len, k, off, valid_states = 0, bufflen = ALUA_RTPG_SIZE;
+	unsigned char *ucp, *buff;
 	unsigned err, retval;
 	unsigned long expiry, interval = 0;
 	unsigned int tpg_desc_tbl_off;
@@ -345,14 +327,19 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 	else
 		expiry = round_jiffies_up(jiffies + h->transition_tmo * HZ);
 
+	buff = kzalloc(bufflen, GFP_KERNEL);
+	if (!buff)
+		return SCSI_DH_DEV_TEMP_BUSY;
+
  retry:
-	retval = submit_rtpg(sdev, h->buff, h->bufflen, &sense_hdr, h->flags);
+	retval = submit_rtpg(sdev, buff, bufflen, &sense_hdr, h->flags);
 
 	if (retval) {
 		if (!scsi_sense_valid(&sense_hdr)) {
 			sdev_printk(KERN_INFO, sdev,
 				    "%s: rtpg failed, result %d\n",
 				    ALUA_DH_NAME, retval);
+			kfree(buff);
 			if (driver_byte(retval) == DRIVER_ERROR)
 				return SCSI_DH_DEV_TEMP_BUSY;
 			return SCSI_DH_IO;
@@ -390,14 +377,18 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 		sdev_printk(KERN_ERR, sdev, "%s: rtpg failed\n",
 			    ALUA_DH_NAME);
 		scsi_print_sense_hdr(sdev, ALUA_DH_NAME, &sense_hdr);
+		kfree(buff);
 		return SCSI_DH_IO;
 	}
 
-	len = get_unaligned_be32(&h->buff[0]) + 4;
+	len = get_unaligned_be32(&buff[0]) + 4;
 
-	if (len > h->bufflen) {
+	if (len > bufflen) {
 		/* Resubmit with the correct length */
-		if (realloc_buffer(h, len)) {
+		kfree(buff);
+		bufflen = len;
+		buff = kmalloc(bufflen, GFP_KERNEL);
+		if (!buff) {
 			sdev_printk(KERN_WARNING, sdev,
 				    "%s: kmalloc buffer failed\n",__func__);
 			/* Temporary failure, bypass */
@@ -407,24 +398,25 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 	}
 
 	orig_transition_tmo = h->transition_tmo;
-	if ((h->buff[4] & RTPG_FMT_MASK) == RTPG_FMT_EXT_HDR && h->buff[5] != 0)
-		h->transition_tmo = h->buff[5];
+	if ((buff[4] & RTPG_FMT_MASK) == RTPG_FMT_EXT_HDR && buff[5] != 0)
+		h->transition_tmo = buff[5];
 	else
 		h->transition_tmo = ALUA_FAILOVER_TIMEOUT;
 
-	if (wait_for_transition && (orig_transition_tmo != h->transition_tmo)) {
+	if (wait_for_transition &&
+	    (orig_transition_tmo != h->transition_tmo)) {
 		sdev_printk(KERN_INFO, sdev,
 			    "%s: transition timeout set to %d seconds\n",
 			    ALUA_DH_NAME, h->transition_tmo);
 		expiry = jiffies + h->transition_tmo * HZ;
 	}
 
-	if ((h->buff[4] & RTPG_FMT_MASK) == RTPG_FMT_EXT_HDR)
+	if ((buff[4] & RTPG_FMT_MASK) == RTPG_FMT_EXT_HDR)
 		tpg_desc_tbl_off = 8;
 	else
 		tpg_desc_tbl_off = 4;
 
-	for (k = tpg_desc_tbl_off, ucp = h->buff + tpg_desc_tbl_off;
+	for (k = tpg_desc_tbl_off, ucp = buff + tpg_desc_tbl_off;
 	     k < len;
 	     k += off, ucp += off) {
 
@@ -474,6 +466,7 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 		err = SCSI_DH_OK;
 		break;
 	}
+	kfree(buff);
 	return err;
 }
 
@@ -672,8 +665,6 @@ static int alua_bus_attach(struct scsi_device *sdev)
 	h->state = TPGS_STATE_OPTIMIZED;
 	h->group_id = -1;
 	h->rel_port = -1;
-	h->buff = h->inq;
-	h->bufflen = ALUA_INQUIRY_SIZE;
 	h->sdev = sdev;
 
 	err = alua_initialize(sdev, h);
@@ -695,8 +686,6 @@ static void alua_bus_detach(struct scsi_device *sdev)
 {
 	struct alua_dh_data *h = sdev->handler_data;
 
-	if (h->buff && h->inq != h->buff)
-		kfree(h->buff);
 	sdev->handler_data = NULL;
 	kfree(h);
 }
Signature · 1 line
1.8.5.6
#8 lore raw

PATCH v2 7/22 scsi_dh_alua: Use separate alua_port_group structure

In reply to Hannes Reinecke

The port group needs to be a separate structure as several LUNs might belong to the same group.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 183 ++++++++++++++++++++---------
 include/scsi/scsi_dh.h                     |   1 +
 2 files changed, 130 insertions(+), 54 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +129 -54
index 49dea42..14b4ca6 100644
@@ -64,14 +64,24 @@
 #define ALUA_OPTIMIZE_STPG		1
 #define ALUA_RTPG_EXT_HDR_UNSUPP	2
 
-struct alua_dh_data {
+static LIST_HEAD(port_group_list);
+static DEFINE_SPINLOCK(port_group_lock);
+
+struct alua_port_group {
+	struct kref		kref;
+	struct list_head	node;
 	int			group_id;
-	int			rel_port;
 	int			tpgs;
 	int			state;
 	int			pref;
 	unsigned		flags; /* used for optimizing STPG */
 	unsigned char		transition_tmo;
+};
+
+struct alua_dh_data {
+	struct alua_port_group	*pg;
+	int			group_id;
+	int			rel_port;
 	struct scsi_device	*sdev;
 	activate_complete	callback_fn;
 	void			*callback_data;
@@ -82,6 +92,17 @@ struct alua_dh_data {
 
 static char print_alua_state(int);
 
+static void release_port_group(struct kref *kref)
+{
+	struct alua_port_group *pg;
+
+	pg = container_of(kref, struct alua_port_group, kref);
+	spin_lock(&port_group_lock);
+	list_del(&pg->node);
+	spin_unlock(&port_group_lock);
+	kfree(pg);
+}
+
 /*
  * submit_rtpg - Issue a REPORT TARGET GROUP STATES command
  * @sdev: sdev the command should be sent to
@@ -142,6 +163,35 @@ static int submit_stpg(struct scsi_device *sdev, int group_id,
 }
 
 /*
+ * alua_alloc_pg - Allocate a new port_group structure
+ * @sdev: scsi device
+ * @h: alua device_handler data
+ * @group_id: port group id
+ *
+ * Allocate a new port_group structure for a given
+ * device.
+ */
+struct alua_port_group *alua_alloc_pg(struct scsi_device *sdev,
+				      int group_id, int tpgs)
+{
+	struct alua_port_group *pg;
+
+	pg = kzalloc(sizeof(struct alua_port_group), GFP_KERNEL);
+	if (!pg)
+		return NULL;
+
+	pg->group_id = group_id;
+	pg->tpgs = tpgs;
+	pg->state = TPGS_STATE_OPTIMIZED;
+	kref_init(&pg->kref);
+	spin_lock(&port_group_lock);
+	list_add(&pg->node, &port_group_list);
+	spin_unlock(&port_group_lock);
+
+	return pg;
+}
+
+/*
  * alua_check_tpgs - Evaluate TPGS setting
  * @sdev: device to be checked
  *
@@ -216,7 +266,6 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h)
 			    ALUA_DH_NAME);
 		return SCSI_DH_DEV_UNSUPP;
 	}
-	h->state = TPGS_STATE_OPTIMIZED;
 	h->group_id = group_id;
 
 	sdev_printk(KERN_INFO, sdev,
@@ -312,7 +361,7 @@ static int alua_check_sense(struct scsi_device *sdev,
  * Returns SCSI_DH_DEV_OFFLINED if the path is
  * found to be unusable.
  */
-static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_for_transition)
+static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg, int wait_for_transition)
 {
 	struct scsi_sense_hdr sense_hdr;
 	int len, k, off, valid_states = 0, bufflen = ALUA_RTPG_SIZE;
@@ -322,17 +371,17 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 	unsigned int tpg_desc_tbl_off;
 	unsigned char orig_transition_tmo;
 
-	if (!h->transition_tmo)
+	if (!pg->transition_tmo)
 		expiry = round_jiffies_up(jiffies + ALUA_FAILOVER_TIMEOUT * HZ);
 	else
-		expiry = round_jiffies_up(jiffies + h->transition_tmo * HZ);
+		expiry = round_jiffies_up(jiffies + pg->transition_tmo * HZ);
 
 	buff = kzalloc(bufflen, GFP_KERNEL);
 	if (!buff)
 		return SCSI_DH_DEV_TEMP_BUSY;
 
  retry:
-	retval = submit_rtpg(sdev, buff, bufflen, &sense_hdr, h->flags);
+	retval = submit_rtpg(sdev, buff, bufflen, &sense_hdr, pg->flags);
 
 	if (retval) {
 		if (!scsi_sense_valid(&sense_hdr)) {
@@ -353,10 +402,10 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 		 * The retry without rtpg_ext_hdr_req set
 		 * handles this.
 		 */
-		if (!(h->flags & ALUA_RTPG_EXT_HDR_UNSUPP) &&
+		if (!(pg->flags & ALUA_RTPG_EXT_HDR_UNSUPP) &&
 		    sense_hdr.sense_key == ILLEGAL_REQUEST &&
 		    sense_hdr.asc == 0x24 && sense_hdr.ascq == 0) {
-			h->flags |= ALUA_RTPG_EXT_HDR_UNSUPP;
+			pg->flags |= ALUA_RTPG_EXT_HDR_UNSUPP;
 			goto retry;
 		}
 		/*
@@ -397,18 +446,18 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 		goto retry;
 	}
 
-	orig_transition_tmo = h->transition_tmo;
+	orig_transition_tmo = pg->transition_tmo;
 	if ((buff[4] & RTPG_FMT_MASK) == RTPG_FMT_EXT_HDR && buff[5] != 0)
-		h->transition_tmo = buff[5];
+		pg->transition_tmo = buff[5];
 	else
-		h->transition_tmo = ALUA_FAILOVER_TIMEOUT;
+		pg->transition_tmo = ALUA_FAILOVER_TIMEOUT;
 
 	if (wait_for_transition &&
-	    (orig_transition_tmo != h->transition_tmo)) {
+	    (orig_transition_tmo != pg->transition_tmo)) {
 		sdev_printk(KERN_INFO, sdev,
 			    "%s: transition timeout set to %d seconds\n",
-			    ALUA_DH_NAME, h->transition_tmo);
-		expiry = jiffies + h->transition_tmo * HZ;
+			    ALUA_DH_NAME, pg->transition_tmo);
+		expiry = jiffies + pg->transition_tmo * HZ;
 	}
 
 	if ((buff[4] & RTPG_FMT_MASK) == RTPG_FMT_EXT_HDR)
@@ -420,9 +469,9 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 	     k < len;
 	     k += off, ucp += off) {
 
-		if (h->group_id == get_unaligned_be16(&ucp[2])) {
-			h->state = ucp[0] & 0x0f;
-			h->pref = ucp[0] >> 7;
+		if (pg->group_id == get_unaligned_be16(&ucp[2])) {
+			pg->state = ucp[0] & 0x0f;
+			pg->pref = ucp[0] >> 7;
 			valid_states = ucp[1];
 		}
 		off = 8 + (ucp[7] * 4);
@@ -430,8 +479,8 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 
 	sdev_printk(KERN_INFO, sdev,
 		    "%s: port group %02x state %c %s supports %c%c%c%c%c%c%c\n",
-		    ALUA_DH_NAME, h->group_id, print_alua_state(h->state),
-		    h->pref ? "preferred" : "non-preferred",
+		    ALUA_DH_NAME, pg->group_id, print_alua_state(pg->state),
+		    pg->pref ? "preferred" : "non-preferred",
 		    valid_states&TPGS_SUPPORT_TRANSITION?'T':'t',
 		    valid_states&TPGS_SUPPORT_OFFLINE?'O':'o',
 		    valid_states&TPGS_SUPPORT_LBA_DEPENDENT?'L':'l',
@@ -440,7 +489,7 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 		    valid_states&TPGS_SUPPORT_NONOPTIMIZED?'N':'n',
 		    valid_states&TPGS_SUPPORT_OPTIMIZED?'A':'a');
 
-	switch (h->state) {
+	switch (pg->state) {
 	case TPGS_STATE_TRANSITIONING:
 		if (wait_for_transition) {
 			if (time_before(jiffies, expiry)) {
@@ -455,7 +504,7 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
 		}
 
 		/* Transitioning time exceeded, set port to standby */
-		h->state = TPGS_STATE_STANDBY;
+		pg->state = TPGS_STATE_STANDBY;
 		break;
 	case TPGS_STATE_OFFLINE:
 		/* Path unusable */
@@ -478,22 +527,22 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_dh_data *h, int wait_
  * a re-evaluation of the target group state or SCSI_DH_OK
  * if no further action needs to be taken.
  */
-static unsigned alua_stpg(struct scsi_device *sdev, struct alua_dh_data *h)
+static unsigned alua_stpg(struct scsi_device *sdev, struct alua_port_group *pg)
 {
 	int retval;
 	struct scsi_sense_hdr sense_hdr;
 
-	if (!(h->tpgs & TPGS_MODE_EXPLICIT)) {
+	if (!(pg->tpgs & TPGS_MODE_EXPLICIT)) {
 		/* Only implicit ALUA supported, retry */
 		return SCSI_DH_RETRY;
 	}
-	switch (h->state) {
+	switch (pg->state) {
 	case TPGS_STATE_OPTIMIZED:
 		return SCSI_DH_OK;
 	case TPGS_STATE_NONOPTIMIZED:
-		if ((h->flags & ALUA_OPTIMIZE_STPG) &&
-		    !h->pref &&
-		    (h->tpgs & TPGS_MODE_IMPLICIT))
+		if ((pg->flags & ALUA_OPTIMIZE_STPG) &&
+		    !pg->pref &&
+		    (pg->tpgs & TPGS_MODE_IMPLICIT))
 			return SCSI_DH_OK;
 		break;
 	case TPGS_STATE_STANDBY:
@@ -507,13 +556,13 @@ static unsigned alua_stpg(struct scsi_device *sdev, struct alua_dh_data *h)
 	default:
 		sdev_printk(KERN_INFO, sdev,
 			    "%s: stpg failed, unhandled TPGS state %d",
-			    ALUA_DH_NAME, h->state);
+			    ALUA_DH_NAME, pg->state);
 		return SCSI_DH_NOSYS;
 		break;
 	}
 	/* Set state to transitioning */
-	h->state = TPGS_STATE_TRANSITIONING;
-	retval = submit_stpg(sdev, h->group_id, &sense_hdr);
+	pg->state = TPGS_STATE_TRANSITIONING;
+	retval = submit_stpg(sdev, pg->group_id, &sense_hdr);
 
 	if (retval) {
 		if (!scsi_sense_valid(&sense_hdr)) {
@@ -523,7 +572,7 @@ static unsigned alua_stpg(struct scsi_device *sdev, struct alua_dh_data *h)
 			if (driver_byte(retval) == DRIVER_ERROR)
 				return SCSI_DH_DEV_TEMP_BUSY;
 		} else {
-			sdev_printk(KERN_INFO, h->sdev, "%s: stpg failed\n",
+			sdev_printk(KERN_INFO, sdev, "%s: stpg failed\n",
 				    ALUA_DH_NAME);
 			scsi_print_sense_hdr(sdev, ALUA_DH_NAME, &sense_hdr);
 		}
@@ -541,20 +590,24 @@ static unsigned alua_stpg(struct scsi_device *sdev, struct alua_dh_data *h)
  */
 static int alua_initialize(struct scsi_device *sdev, struct alua_dh_data *h)
 {
-	int err = SCSI_DH_DEV_UNSUPP;
+	int err = SCSI_DH_DEV_UNSUPP, tpgs;
 
-	h->tpgs = alua_check_tpgs(sdev);
-	if (h->tpgs == TPGS_MODE_NONE)
+	tpgs = alua_check_tpgs(sdev);
+	if (tpgs == TPGS_MODE_NONE)
 		goto out;
 
 	err = alua_check_vpd(sdev, h);
 	if (err != SCSI_DH_OK)
 		goto out;
 
-	err = alua_rtpg(sdev, h, 0);
-	if (err != SCSI_DH_OK)
+	h->pg = alua_alloc_pg(sdev, h->group_id, tpgs);
+	if (!h->pg) {
+		err = SCSI_DH_NOMEM;
 		goto out;
-
+	}
+	kref_get(&h->pg->kref);
+	err = alua_rtpg(sdev, h->pg, 0);
+	kref_put(&h->pg->kref, release_port_group);
 out:
 	return err;
 }
@@ -570,6 +623,7 @@ out:
 static int alua_set_params(struct scsi_device *sdev, const char *params)
 {
 	struct alua_dh_data *h = sdev->handler_data;
+	struct alua_port_group *pg = NULL;
 	unsigned int optimize = 0, argc;
 	const char *p = params;
 	int result = SCSI_DH_OK;
@@ -582,10 +636,14 @@ static int alua_set_params(struct scsi_device *sdev, const char *params)
 	if ((sscanf(p, "%u", &optimize) != 1) || (optimize > 1))
 		return -EINVAL;
 
+	pg = h->pg;
+	if (!pg)
+		return -ENXIO;
+
 	if (optimize)
-		h->flags |= ALUA_OPTIMIZE_STPG;
+		pg->flags |= ALUA_OPTIMIZE_STPG;
 	else
-		h->flags &= ~ALUA_OPTIMIZE_STPG;
+		pg->flags |= ~ALUA_OPTIMIZE_STPG;
 
 	return result;
 }
@@ -610,16 +668,23 @@ static int alua_activate(struct scsi_device *sdev,
 	struct alua_dh_data *h = sdev->handler_data;
 	int err = SCSI_DH_OK;
 
-	err = alua_rtpg(sdev, h, 1);
-	if (err != SCSI_DH_OK)
+	if (!h->pg)
 		goto out;
 
+	kref_get(&h->pg->kref);
+
 	if (optimize_stpg)
-		h->flags |= ALUA_OPTIMIZE_STPG;
+		h->pg->flags |= ALUA_OPTIMIZE_STPG;
 
-	err = alua_stpg(sdev, h);
+	err = alua_rtpg(sdev, h->pg, 1);
+	if (err != SCSI_DH_OK) {
+		kref_put(&h->pg->kref, release_port_group);
+		goto out;
+	}
+	err = alua_stpg(sdev, h->pg);
 	if (err == SCSI_DH_RETRY)
-		err = alua_rtpg(sdev, h, 1);
+		err = alua_rtpg(sdev, h->pg, 1);
+	kref_put(&h->pg->kref, release_port_group);
 out:
 	if (fn)
 		fn(data, err);
@@ -635,13 +700,19 @@ out:
 static int alua_prep_fn(struct scsi_device *sdev, struct request *req)
 {
 	struct alua_dh_data *h = sdev->handler_data;
+	int state;
 	int ret = BLKPREP_OK;
 
-	if (h->state == TPGS_STATE_TRANSITIONING)
+	if (!h->pg)
+		return ret;
+	kref_get(&h->pg->kref);
+	state = h->pg->state;
+	kref_put(&h->pg->kref, release_port_group);
+	if (state == TPGS_STATE_TRANSITIONING)
 		ret = BLKPREP_DEFER;
-	else if (h->state != TPGS_STATE_OPTIMIZED &&
-		 h->state != TPGS_STATE_NONOPTIMIZED &&
-		 h->state != TPGS_STATE_LBA_DEPENDENT) {
+	else if (state != TPGS_STATE_OPTIMIZED &&
+		 state != TPGS_STATE_NONOPTIMIZED &&
+		 state != TPGS_STATE_LBA_DEPENDENT) {
 		ret = BLKPREP_KILL;
 		req->cmd_flags |= REQ_QUIET;
 	}
@@ -656,18 +727,18 @@ static int alua_prep_fn(struct scsi_device *sdev, struct request *req)
 static int alua_bus_attach(struct scsi_device *sdev)
 {
 	struct alua_dh_data *h;
-	int err;
+	int err, ret = -EINVAL;
 
 	h = kzalloc(sizeof(*h) , GFP_KERNEL);
 	if (!h)
 		return -ENOMEM;
-	h->tpgs = TPGS_MODE_UNINITIALIZED;
-	h->state = TPGS_STATE_OPTIMIZED;
-	h->group_id = -1;
+	h->pg = NULL;
 	h->rel_port = -1;
 	h->sdev = sdev;
 
 	err = alua_initialize(sdev, h);
+	if (err == SCSI_DH_NOMEM)
+		ret = -ENOMEM;
 	if (err != SCSI_DH_OK && err != SCSI_DH_DEV_OFFLINED)
 		goto failed;
 
@@ -675,7 +746,7 @@ static int alua_bus_attach(struct scsi_device *sdev)
 	return 0;
 failed:
 	kfree(h);
-	return -EINVAL;
+	return ret;
 }
 
 /*
@@ -686,6 +757,10 @@ static void alua_bus_detach(struct scsi_device *sdev)
 {
 	struct alua_dh_data *h = sdev->handler_data;
 
+	if (h->pg) {
+		kref_put(&h->pg->kref, release_port_group);
+		h->pg = NULL;
+	}
 	sdev->handler_data = NULL;
 	kfree(h);
 }
include/scsi/scsi_dh.h +1 -0
index 85d7317..7e184c6 100644
@@ -52,6 +52,7 @@ enum {
 	SCSI_DH_TIMED_OUT,
 	SCSI_DH_RES_TEMP_UNAVAIL,
 	SCSI_DH_DEV_OFFLINED,
+	SCSI_DH_NOMEM,
 	SCSI_DH_NOSYS,
 	SCSI_DH_DRIVER_MAX,
 };
Signature · 1 line
1.8.5.6
#9 lore raw

PATCH v2 8/22 scsi_dh_alua: use unique device id

In reply to Hannes Reinecke

Use scsi_vpd_lun_id() to assign a unique device identification to the alua port group structure.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 55 +++++++++++++++++++++++++++---
 1 file changed, 51 insertions(+), 4 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +51 -4
index 14b4ca6..41fd5e8 100644
@@ -70,6 +70,8 @@ static DEFINE_SPINLOCK(port_group_lock);
 struct alua_port_group {
 	struct kref		kref;
 	struct list_head	node;
+	unsigned char		device_id_str[256];
+	int			device_id_len;
 	int			group_id;
 	int			tpgs;
 	int			state;
@@ -162,6 +164,25 @@ static int submit_stpg(struct scsi_device *sdev, int group_id,
 				      ALUA_FAILOVER_RETRIES, NULL, req_flags);
 }
 
+struct alua_port_group *alua_lookup_pg(char *id_str, size_t id_size,
+				       int group_id)
+{
+	struct alua_port_group *pg;
+
+	list_for_each_entry(pg, &port_group_list, node) {
+		if (pg->group_id != group_id)
+			continue;
+		if (pg->device_id_len != id_size)
+			continue;
+		if (strncmp(pg->device_id_str, id_str, id_size))
+			continue;
+		kref_get(&pg->kref);
+		return pg;
+	}
+
+	return NULL;
+}
+
 /*
  * alua_alloc_pg - Allocate a new port_group structure
  * @sdev: scsi device
@@ -174,17 +195,38 @@ static int submit_stpg(struct scsi_device *sdev, int group_id,
 struct alua_port_group *alua_alloc_pg(struct scsi_device *sdev,
 				      int group_id, int tpgs)
 {
-	struct alua_port_group *pg;
+	struct alua_port_group *pg, *tmp_pg;
 
 	pg = kzalloc(sizeof(struct alua_port_group), GFP_KERNEL);
 	if (!pg)
-		return NULL;
+		return ERR_PTR(-ENOMEM);
 
+	pg->device_id_len = scsi_vpd_lun_id(sdev, pg->device_id_str,
+					    sizeof(pg->device_id_str));
+	if (pg->device_id_len <= 0) {
+		/*
+		 * Internal error: TPGS supported but no device
+		 * identifcation found. Disable ALUA support.
+		 */
+		kfree(pg);
+		sdev_printk(KERN_INFO, sdev,
+			    "%s: No device descriptors found\n",
+			    ALUA_DH_NAME);
+		return ERR_PTR(-ENXIO);
+	}
 	pg->group_id = group_id;
 	pg->tpgs = tpgs;
 	pg->state = TPGS_STATE_OPTIMIZED;
 	kref_init(&pg->kref);
+
 	spin_lock(&port_group_lock);
+	tmp_pg = alua_lookup_pg(pg->device_id_str, pg->device_id_len, group_id);
+	if (tmp_pg) {
+		spin_unlock(&port_group_lock);
+		kfree(pg);
+		return tmp_pg;
+	}
+
 	list_add(&pg->node, &port_group_list);
 	spin_unlock(&port_group_lock);
 
@@ -601,10 +643,15 @@ static int alua_initialize(struct scsi_device *sdev, struct alua_dh_data *h)
 		goto out;
 
 	h->pg = alua_alloc_pg(sdev, h->group_id, tpgs);
-	if (!h->pg) {
-		err = SCSI_DH_NOMEM;
+	if (PTR_ERR(h->pg)) {
+		if (PTR_ERR(h->pg) == -ENOMEM)
+			err = SCSI_DH_NOMEM;
 		goto out;
 	}
+	sdev_printk(KERN_INFO, sdev,
+		    "%s: device %s port group %02x rel port %02x\n",
+		    ALUA_DH_NAME, h->pg->device_id_str,
+		    h->group_id, h->rel_port);
 	kref_get(&h->pg->kref);
 	err = alua_rtpg(sdev, h->pg, 0);
 	kref_put(&h->pg->kref, release_port_group);
Signature · 1 line
1.8.5.6
#10 lore raw

Re: [PATCHv2 08/22] scsi_dh_alua: use unique device id

drivers/scsi/device_handler/scsi_dh_alua.c static int submit_stpg(struct scsi_device *sdev, int group_id, · line 168

Quoted 2 lines
drivers/scsi/device_handler/scsi_dh_alua.c +2 -0
+struct alua_port_group *alua_lookup_pg(char *id_str, size_t id_size,
+				       int group_id)

Not too important, but I would have preferred alua_find_get_pg here.

drivers/scsi/device_handler/scsi_dh_alua.c static int submit_stpg(struct scsi_device *sdev, int group_id, · line 179

Quoted 11 lines
drivers/scsi/device_handler/scsi_dh_alua.c +11 -0
+{
+	struct alua_port_group *pg;
+
+	list_for_each_entry(pg, &port_group_list, node) {
+		if (pg->group_id != group_id)
+			continue;
+		if (pg->device_id_len != id_size)
+			continue;
+		if (strncmp(pg->device_id_str, id_str, id_size))
+			continue;
+		kref_get(&pg->kref);

This needs to be a kref_get_unless_zero

#11 lore raw

PATCH v2 9/22 scsi_dh_alua: simplify alua_initialize()

In reply to Hannes Reinecke

Rework alua_check_vpd() to use scsi_vpd_get_tpg() and move the port group selection into the function, too. With that we can simplify alua_initialize() to just call alua_check_tpgs() and alua_check_vpd();

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 42 ++++++++++++------------------
 1 file changed, 17 insertions(+), 25 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +17 -25
index 41fd5e8..c4ee225 100644
@@ -92,6 +92,7 @@ struct alua_dh_data {
 #define ALUA_POLICY_SWITCH_CURRENT	0
 #define ALUA_POLICY_SWITCH_ALL		1
 
+static int alua_rtpg(struct scsi_device *, struct alua_port_group *, int);
 static char print_alua_state(int);
 
 static void release_port_group(struct kref *kref)
@@ -292,7 +293,8 @@ static int alua_check_tpgs(struct scsi_device *sdev)
  * Extract the relative target port and the target port group
  * descriptor from the list of identificators.
  */
-static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h)
+static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h,
+			  int tpgs)
 {
 	int rel_port = -1, group_id;
 
@@ -308,13 +310,21 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h)
 			    ALUA_DH_NAME);
 		return SCSI_DH_DEV_UNSUPP;
 	}
-	h->group_id = group_id;
+
+	h->pg = alua_alloc_pg(sdev, group_id, tpgs);
+	if (PTR_ERR(h->pg)) {
+		if (PTR_ERR(h->pg) == -ENOMEM)
+			return SCSI_DH_NOMEM;
+		return SCSI_DH_DEV_UNSUPP;
+	}
+	h->rel_port = rel_port;
 
 	sdev_printk(KERN_INFO, sdev,
-		    "%s: port group %02x rel port %02x\n",
-		    ALUA_DH_NAME, h->group_id, h->rel_port);
+		    "%s: device %s port group %02x rel port %02x\n",
+		    ALUA_DH_NAME, h->pg->device_id_str,
+		    h->group_id, h->rel_port);
 
-	return 0;
+	return alua_rtpg(sdev, h->pg, 0);
 }
 
 static char print_alua_state(int state)
@@ -635,27 +645,9 @@ static int alua_initialize(struct scsi_device *sdev, struct alua_dh_data *h)
 	int err = SCSI_DH_DEV_UNSUPP, tpgs;
 
 	tpgs = alua_check_tpgs(sdev);
-	if (tpgs == TPGS_MODE_NONE)
-		goto out;
-
-	err = alua_check_vpd(sdev, h);
-	if (err != SCSI_DH_OK)
-		goto out;
+	if (tpgs != TPGS_MODE_NONE)
+		err = alua_check_vpd(sdev, h, tpgs);
 
-	h->pg = alua_alloc_pg(sdev, h->group_id, tpgs);
-	if (PTR_ERR(h->pg)) {
-		if (PTR_ERR(h->pg) == -ENOMEM)
-			err = SCSI_DH_NOMEM;
-		goto out;
-	}
-	sdev_printk(KERN_INFO, sdev,
-		    "%s: device %s port group %02x rel port %02x\n",
-		    ALUA_DH_NAME, h->pg->device_id_str,
-		    h->group_id, h->rel_port);
-	kref_get(&h->pg->kref);
-	err = alua_rtpg(sdev, h->pg, 0);
-	kref_put(&h->pg->kref, release_port_group);
-out:
 	return err;
 }
 /*
Signature · 1 line
1.8.5.6
#12 lore raw

PATCH v2 10/22 revert commit a8e5a2d593cb ("[SCSI] scsi_dh_alua: ALUA handler attach should succeed while TPG is transitioning")

In reply to Hannes Reinecke

This reverts commit a8e5a2d593cbfccf530c3382c2c328d2edaa7b66

Obsoleted by the next patch.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 31 ++++++++++++------------------
 1 file changed, 12 insertions(+), 19 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +12 -19
index c4ee225..2762aba 100644
@@ -92,7 +92,7 @@ struct alua_dh_data {
 #define ALUA_POLICY_SWITCH_CURRENT	0
 #define ALUA_POLICY_SWITCH_ALL		1
 
-static int alua_rtpg(struct scsi_device *, struct alua_port_group *, int);
+static int alua_rtpg(struct scsi_device *, struct alua_port_group *);
 static char print_alua_state(int);
 
 static void release_port_group(struct kref *kref)
@@ -324,7 +324,7 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h,
 		    ALUA_DH_NAME, h->pg->device_id_str,
 		    h->group_id, h->rel_port);
 
-	return alua_rtpg(sdev, h->pg, 0);
+	return alua_rtpg(sdev, h->pg);
 }
 
 static char print_alua_state(int state)
@@ -407,13 +407,12 @@ static int alua_check_sense(struct scsi_device *sdev,
 /*
  * alua_rtpg - Evaluate REPORT TARGET GROUP STATES
  * @sdev: the device to be evaluated.
- * @wait_for_transition: if nonzero, wait ALUA_FAILOVER_TIMEOUT seconds for device to exit transitioning state
  *
  * Evaluate the Target Port Group State.
  * Returns SCSI_DH_DEV_OFFLINED if the path is
  * found to be unusable.
  */
-static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg, int wait_for_transition)
+static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 {
 	struct scsi_sense_hdr sense_hdr;
 	int len, k, off, valid_states = 0, bufflen = ALUA_RTPG_SIZE;
@@ -504,8 +503,7 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg, int w
 	else
 		pg->transition_tmo = ALUA_FAILOVER_TIMEOUT;
 
-	if (wait_for_transition &&
-	    (orig_transition_tmo != pg->transition_tmo)) {
+	if (orig_transition_tmo != pg->transition_tmo) {
 		sdev_printk(KERN_INFO, sdev,
 			    "%s: transition timeout set to %d seconds\n",
 			    ALUA_DH_NAME, pg->transition_tmo);
@@ -543,19 +541,14 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg, int w
 
 	switch (pg->state) {
 	case TPGS_STATE_TRANSITIONING:
-		if (wait_for_transition) {
-			if (time_before(jiffies, expiry)) {
-				/* State transition, retry */
-				interval += 2000;
-				msleep(interval);
-				goto retry;
-			}
-			err = SCSI_DH_RETRY;
-		} else {
-			err = SCSI_DH_OK;
+		if (time_before(jiffies, expiry)) {
+			/* State transition, retry */
+			interval += 2000;
+			msleep(interval);
+			goto retry;
 		}
-
 		/* Transitioning time exceeded, set port to standby */
+		err = SCSI_DH_RETRY;
 		pg->state = TPGS_STATE_STANDBY;
 		break;
 	case TPGS_STATE_OFFLINE:
@@ -715,14 +708,14 @@ static int alua_activate(struct scsi_device *sdev,
 	if (optimize_stpg)
 		h->pg->flags |= ALUA_OPTIMIZE_STPG;
 
-	err = alua_rtpg(sdev, h->pg, 1);
+	err = alua_rtpg(sdev, h->pg);
 	if (err != SCSI_DH_OK) {
 		kref_put(&h->pg->kref, release_port_group);
 		goto out;
 	}
 	err = alua_stpg(sdev, h->pg);
 	if (err == SCSI_DH_RETRY)
-		err = alua_rtpg(sdev, h->pg, 1);
+		err = alua_rtpg(sdev, h->pg);
 	kref_put(&h->pg->kref, release_port_group);
 out:
 	if (fn)
Signature · 1 line
1.8.5.6
#13 lore raw

PATCH v2 11/22 scsi_dh_alua: move optimize_stpg evaluation

In reply to Hannes Reinecke

When the optimize_stpg module option is set we should just set it once during port_group allocation. Doing so allows us to override it later with device specific settings.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +6 -7
index 2762aba..7bd5108 100644
@@ -64,6 +64,10 @@
 #define ALUA_OPTIMIZE_STPG		1
 #define ALUA_RTPG_EXT_HDR_UNSUPP	2
 
+static uint optimize_stpg;
+module_param(optimize_stpg, uint, S_IRUGO|S_IWUSR);
+MODULE_PARM_DESC(optimize_stpg, "Allow use of a non-optimized path, rather than sending a STPG, when implicit TPGS is supported (0=No,1=Yes). Default is 0.");
+
 static LIST_HEAD(port_group_list);
 static DEFINE_SPINLOCK(port_group_lock);
 
@@ -218,6 +222,8 @@ struct alua_port_group *alua_alloc_pg(struct scsi_device *sdev,
 	pg->group_id = group_id;
 	pg->tpgs = tpgs;
 	pg->state = TPGS_STATE_OPTIMIZED;
+	if (optimize_stpg)
+		pg->flags |= ALUA_OPTIMIZE_STPG;
 	kref_init(&pg->kref);
 
 	spin_lock(&port_group_lock);
@@ -680,10 +686,6 @@ static int alua_set_params(struct scsi_device *sdev, const char *params)
 	return result;
 }
 
-static uint optimize_stpg;
-module_param(optimize_stpg, uint, S_IRUGO|S_IWUSR);
-MODULE_PARM_DESC(optimize_stpg, "Allow use of a non-optimized path, rather than sending a STPG, when implicit TPGS is supported (0=No,1=Yes). Default is 0.");
-
 /*
  * alua_activate - activate a path
  * @sdev: device on the path to be activated
@@ -705,9 +707,6 @@ static int alua_activate(struct scsi_device *sdev,
 
 	kref_get(&h->pg->kref);
 
-	if (optimize_stpg)
-		h->pg->flags |= ALUA_OPTIMIZE_STPG;
-
 	err = alua_rtpg(sdev, h->pg);
 	if (err != SCSI_DH_OK) {
 		kref_put(&h->pg->kref, release_port_group);
Signature · 1 line
1.8.5.6
Reviewed-by on v2 11/22
#14 lore raw

Re: [PATCHv2 11/22] scsi_dh_alua: move optimize_stpg evaluation

Quoted 3 lines · Hannes Reinecke

When the optimize_stpg module option is set we should just set it once during port_group allocation. Doing so allows us to override it later with device specific settings.

Looks fine,

#15 lore raw

PATCH v2 12/22 scsi_dh_alua: remove 'rel_port' from alua_dh_data structure

In reply to Hannes Reinecke

The 'relative port' field is not used, and might get stale when the port group changes. So remove the field altogether.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +1 -5
index 7bd5108..c9d028c 100644
@@ -87,7 +87,6 @@ struct alua_port_group {
 struct alua_dh_data {
 	struct alua_port_group	*pg;
 	int			group_id;
-	int			rel_port;
 	struct scsi_device	*sdev;
 	activate_complete	callback_fn;
 	void			*callback_data;
@@ -323,12 +322,10 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h,
 			return SCSI_DH_NOMEM;
 		return SCSI_DH_DEV_UNSUPP;
 	}
-	h->rel_port = rel_port;
-
 	sdev_printk(KERN_INFO, sdev,
 		    "%s: device %s port group %02x rel port %02x\n",
 		    ALUA_DH_NAME, h->pg->device_id_str,
-		    h->group_id, h->rel_port);
+		    h->group_id, rel_port);
 
 	return alua_rtpg(sdev, h->pg);
 }
@@ -764,7 +761,6 @@ static int alua_bus_attach(struct scsi_device *sdev)
 	if (!h)
 		return -ENOMEM;
 	h->pg = NULL;
-	h->rel_port = -1;
 	h->sdev = sdev;
 
 	err = alua_initialize(sdev, h);
Signature · 1 line
1.8.5.6
#17 lore raw

PATCH v2 13/22 scsi_dh_alua: Use workqueue for RTPG

In reply to Hannes Reinecke

The current ALUA device_handler has two drawbacks:
- We're sending a 'SET TARGET PORT GROUP' command to every LUN, disregarding the fact that several LUNs might be in a port group and will be automatically switched whenever _any_ LUN within that port group receives the command.
- Whenever a LUN is in 'transitioning' mode we cannot block I/O to that LUN, instead the controller has to abort the command. This leads to increased traffic across the wire and heavy load on the controller during switchover.

With this patch the RTPG handling is moved to a per-portgroup workqueue. This reduces the number of 'REPORT TARGET PORT GROUP' and 'SET TARGET PORT GROUPS' sent to the controller as we're sending them now per port group, and not per device as previously. It also allows us to block I/O to any LUN / port group found to be in 'transitioning' ALUA mode, as the workqueue item will be requeued until the controller moves out of transitioning.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 308 ++++++++++++++++++++++++-----
 1 file changed, 257 insertions(+), 51 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +257 -51
index c9d028c..9d56d6a 100644
@@ -59,10 +59,15 @@
 #define ALUA_RTPG_SIZE			128
 #define ALUA_FAILOVER_TIMEOUT		60
 #define ALUA_FAILOVER_RETRIES		5
+#define ALUA_RTPG_DELAY_MSECS		5
 
 /* device handler flags */
-#define ALUA_OPTIMIZE_STPG		1
-#define ALUA_RTPG_EXT_HDR_UNSUPP	2
+#define ALUA_OPTIMIZE_STPG		0x01
+#define ALUA_RTPG_EXT_HDR_UNSUPP	0x02
+/* State machine flags */
+#define ALUA_PG_RUN_RTPG		0x10
+#define ALUA_PG_RUN_STPG		0x20
+#define ALUA_PG_RUNNING			0x40
 
 static uint optimize_stpg;
 module_param(optimize_stpg, uint, S_IRUGO|S_IWUSR);
@@ -70,6 +75,7 @@ MODULE_PARM_DESC(optimize_stpg, "Allow use of a non-optimized path, rather than
 
 static LIST_HEAD(port_group_list);
 static DEFINE_SPINLOCK(port_group_lock);
+static struct workqueue_struct *kaluad_wq;
 
 struct alua_port_group {
 	struct kref		kref;
@@ -82,12 +88,25 @@ struct alua_port_group {
 	int			pref;
 	unsigned		flags; /* used for optimizing STPG */
 	unsigned char		transition_tmo;
+	unsigned long		expiry;
+	unsigned long		interval;
+	struct delayed_work	rtpg_work;
+	spinlock_t		lock;
+	struct list_head	rtpg_list;
+	struct scsi_device	*rtpg_sdev;
 };
 
 struct alua_dh_data {
 	struct alua_port_group	*pg;
 	int			group_id;
+	spinlock_t		pg_lock;
 	struct scsi_device	*sdev;
+	int			init_error;
+	struct mutex		init_mutex;
+};
+
+struct alua_queue_data {
+	struct list_head	entry;
 	activate_complete	callback_fn;
 	void			*callback_data;
 };
@@ -95,8 +114,10 @@ struct alua_dh_data {
 #define ALUA_POLICY_SWITCH_CURRENT	0
 #define ALUA_POLICY_SWITCH_ALL		1
 
-static int alua_rtpg(struct scsi_device *, struct alua_port_group *);
-static char print_alua_state(int);
+static void alua_rtpg_work(struct work_struct *work);
+static void alua_rtpg_queue(struct alua_port_group *pg,
+			    struct scsi_device *sdev,
+			    struct alua_queue_data *qdata);
 
 static void release_port_group(struct kref *kref)
 {
@@ -106,6 +127,7 @@ static void release_port_group(struct kref *kref)
 	spin_lock(&port_group_lock);
 	list_del(&pg->node);
 	spin_unlock(&port_group_lock);
+	WARN_ON(pg->rtpg_sdev);
 	kfree(pg);
 }
 
@@ -180,7 +202,8 @@ struct alua_port_group *alua_lookup_pg(char *id_str, size_t id_size,
 			continue;
 		if (strncmp(pg->device_id_str, id_str, id_size))
 			continue;
-		kref_get(&pg->kref);
+		if (!kref_get_unless_zero(&pg->kref))
+			continue;
 		return pg;
 	}
 
@@ -224,6 +247,10 @@ struct alua_port_group *alua_alloc_pg(struct scsi_device *sdev,
 	if (optimize_stpg)
 		pg->flags |= ALUA_OPTIMIZE_STPG;
 	kref_init(&pg->kref);
+	INIT_DELAYED_WORK(&pg->rtpg_work, alua_rtpg_work);
+	INIT_LIST_HEAD(&pg->rtpg_list);
+	INIT_LIST_HEAD(&pg->node);
+	spin_lock_init(&pg->lock);
 
 	spin_lock(&port_group_lock);
 	tmp_pg = alua_lookup_pg(pg->device_id_str, pg->device_id_len, group_id);
@@ -302,6 +329,8 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h,
 			  int tpgs)
 {
 	int rel_port = -1, group_id;
+	struct alua_port_group *pg, *old_pg = NULL;
+	bool pg_updated = false;
 
 	group_id = scsi_vpd_tpg_id(sdev, &rel_port);
 	if (group_id < 0) {
@@ -316,18 +345,42 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h,
 		return SCSI_DH_DEV_UNSUPP;
 	}
 
-	h->pg = alua_alloc_pg(sdev, group_id, tpgs);
-	if (PTR_ERR(h->pg)) {
+	pg = alua_alloc_pg(sdev, group_id, tpgs);
+	if (PTR_ERR(pg)) {
 		if (PTR_ERR(h->pg) == -ENOMEM)
 			return SCSI_DH_NOMEM;
 		return SCSI_DH_DEV_UNSUPP;
 	}
 	sdev_printk(KERN_INFO, sdev,
 		    "%s: device %s port group %02x rel port %02x\n",
-		    ALUA_DH_NAME, h->pg->device_id_str,
-		    h->group_id, rel_port);
+		    ALUA_DH_NAME, pg->device_id_str, group_id, rel_port);
 
-	return alua_rtpg(sdev, h->pg);
+	/* Check for existing port_group references */
+	spin_lock(&h->pg_lock);
+	if (h->pg) {
+		old_pg = pg;
+		/* port_group has changed. Update to new port group */
+		if (h->pg != pg) {
+			old_pg = h->pg;
+			rcu_assign_pointer(h->pg, pg);
+			pg_updated = true;
+		}
+	} else {
+		rcu_assign_pointer(h->pg, pg);
+		pg_updated = true;
+	}
+	alua_rtpg_queue(h->pg, sdev, NULL);
+	spin_unlock(&h->pg_lock);
+
+	if (pg_updated)
+		synchronize_rcu();
+	if (old_pg) {
+		if (old_pg->rtpg_sdev)
+			flush_delayed_work(&old_pg->rtpg_work);
+		kref_put(&old_pg->kref, release_port_group);
+	}
+
+	return SCSI_DH_OK;
 }
 
 static char print_alua_state(int state)
@@ -421,14 +474,17 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 	int len, k, off, valid_states = 0, bufflen = ALUA_RTPG_SIZE;
 	unsigned char *ucp, *buff;
 	unsigned err, retval;
-	unsigned long expiry, interval = 0;
 	unsigned int tpg_desc_tbl_off;
 	unsigned char orig_transition_tmo;
 
-	if (!pg->transition_tmo)
-		expiry = round_jiffies_up(jiffies + ALUA_FAILOVER_TIMEOUT * HZ);
-	else
-		expiry = round_jiffies_up(jiffies + pg->transition_tmo * HZ);
+	if (!pg->expiry) {
+		unsigned long transition_tmo = ALUA_FAILOVER_TIMEOUT * HZ;
+
+		if (pg->transition_tmo)
+			transition_tmo = pg->transition_tmo * HZ;
+
+		pg->expiry = round_jiffies_up(jiffies + transition_tmo);
+	}
 
 	buff = kzalloc(bufflen, GFP_KERNEL);
 	if (!buff)
@@ -471,16 +527,18 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 			err = SCSI_DH_RETRY;
 		else if (sense_hdr.sense_key == UNIT_ATTENTION)
 			err = SCSI_DH_RETRY;
-		if (err == SCSI_DH_RETRY && time_before(jiffies, expiry)) {
+		if (err == SCSI_DH_RETRY &&
+		    pg->expiry != 0 && time_before(jiffies, pg->expiry)) {
 			sdev_printk(KERN_ERR, sdev, "%s: rtpg retry\n",
 				    ALUA_DH_NAME);
 			scsi_print_sense_hdr(sdev, ALUA_DH_NAME, &sense_hdr);
-			goto retry;
+			return err;
 		}
 		sdev_printk(KERN_ERR, sdev, "%s: rtpg failed\n",
 			    ALUA_DH_NAME);
 		scsi_print_sense_hdr(sdev, ALUA_DH_NAME, &sense_hdr);
 		kfree(buff);
+		pg->expiry = 0;
 		return SCSI_DH_IO;
 	}
 
@@ -495,6 +553,7 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 			sdev_printk(KERN_WARNING, sdev,
 				    "%s: kmalloc buffer failed\n",__func__);
 			/* Temporary failure, bypass */
+			pg->expiry = 0;
 			return SCSI_DH_DEV_TEMP_BUSY;
 		}
 		goto retry;
@@ -510,7 +569,7 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 		sdev_printk(KERN_INFO, sdev,
 			    "%s: transition timeout set to %d seconds\n",
 			    ALUA_DH_NAME, pg->transition_tmo);
-		expiry = jiffies + pg->transition_tmo * HZ;
+		pg->expiry = jiffies + pg->transition_tmo * HZ;
 	}
 
 	if ((buff[4] & RTPG_FMT_MASK) == RTPG_FMT_EXT_HDR)
@@ -544,23 +603,26 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 
 	switch (pg->state) {
 	case TPGS_STATE_TRANSITIONING:
-		if (time_before(jiffies, expiry)) {
+		if (time_before(jiffies, pg->expiry)) {
 			/* State transition, retry */
-			interval += 2000;
-			msleep(interval);
-			goto retry;
+			pg->interval = 2;
+			err = SCSI_DH_RETRY;
+		} else {
+			/* Transitioning time exceeded, set port to standby */
+			err = SCSI_DH_IO;
+			pg->state = TPGS_STATE_STANDBY;
+			pg->expiry = 0;
 		}
-		/* Transitioning time exceeded, set port to standby */
-		err = SCSI_DH_RETRY;
-		pg->state = TPGS_STATE_STANDBY;
 		break;
 	case TPGS_STATE_OFFLINE:
 		/* Path unusable */
 		err = SCSI_DH_DEV_OFFLINED;
+		pg->expiry = 0;
 		break;
 	default:
 		/* Useable path if active */
 		err = SCSI_DH_OK;
+		pg->expiry = 0;
 		break;
 	}
 	kfree(buff);
@@ -629,6 +691,106 @@ static unsigned alua_stpg(struct scsi_device *sdev, struct alua_port_group *pg)
 	return SCSI_DH_RETRY;
 }
 
+static void alua_rtpg_work(struct work_struct *work)
+{
+	struct alua_port_group *pg =
+		container_of(work, struct alua_port_group, rtpg_work.work);
+	struct scsi_device *sdev;
+	LIST_HEAD(qdata_list);
+	int err = SCSI_DH_OK;
+	struct alua_queue_data *qdata, *tmp;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pg->lock, flags);
+	sdev = pg->rtpg_sdev;
+	if (!sdev) {
+		WARN_ON(pg->flags & ALUA_PG_RUN_RTPG ||
+			pg->flags & ALUA_PG_RUN_STPG);
+		spin_unlock_irqrestore(&pg->lock, flags);
+		return;
+	}
+	pg->flags |= ALUA_PG_RUNNING;
+	if (pg->flags & ALUA_PG_RUN_RTPG) {
+		spin_unlock_irqrestore(&pg->lock, flags);
+		err = alua_rtpg(sdev, pg);
+		spin_lock_irqsave(&pg->lock, flags);
+		if (err == SCSI_DH_RETRY) {
+			pg->flags &= ~ALUA_PG_RUNNING;
+			spin_unlock_irqrestore(&pg->lock, flags);
+			queue_delayed_work(kaluad_wq, &pg->rtpg_work,
+					   pg->interval * HZ);
+			return;
+		}
+		pg->flags &= ~ALUA_PG_RUN_RTPG;
+		if (err != SCSI_DH_OK)
+			pg->flags &= ~ALUA_PG_RUN_STPG;
+	}
+	if (pg->flags & ALUA_PG_RUN_STPG) {
+		spin_unlock_irqrestore(&pg->lock, flags);
+		err = alua_stpg(sdev, pg);
+		spin_lock_irqsave(&pg->lock, flags);
+		pg->flags &= ~ALUA_PG_RUN_STPG;
+		if (err == SCSI_DH_RETRY) {
+			pg->flags |= ALUA_PG_RUN_RTPG;
+			pg->interval = 0;
+			pg->flags &= ~ALUA_PG_RUNNING;
+			spin_unlock_irqrestore(&pg->lock, flags);
+			queue_delayed_work(kaluad_wq, &pg->rtpg_work,
+					   pg->interval * HZ);
+			return;
+		}
+	}
+
+	list_splice_init(&pg->rtpg_list, &qdata_list);
+	pg->rtpg_sdev = NULL;
+	spin_unlock_irqrestore(&pg->lock, flags);
+
+	list_for_each_entry_safe(qdata, tmp, &qdata_list, entry) {
+		list_del(&qdata->entry);
+		if (qdata->callback_fn)
+			qdata->callback_fn(qdata->callback_data, err);
+		kfree(qdata);
+	}
+	spin_lock_irqsave(&pg->lock, flags);
+	pg->flags &= ~ALUA_PG_RUNNING;
+	spin_unlock_irqrestore(&pg->lock, flags);
+	scsi_device_put(sdev);
+	kref_put(&pg->kref, release_port_group);
+}
+
+static void alua_rtpg_queue(struct alua_port_group *pg,
+			    struct scsi_device *sdev,
+			    struct alua_queue_data *qdata)
+{
+	int start_queue = 0;
+	unsigned long flags;
+
+	if (!pg)
+		return;
+
+	spin_lock_irqsave(&pg->lock, flags);
+	if (qdata) {
+		list_add_tail(&qdata->entry, &pg->rtpg_list);
+		pg->flags |= ALUA_PG_RUN_STPG;
+	}
+	if (pg->rtpg_sdev == NULL) {
+		pg->interval = 0;
+		pg->flags |= ALUA_PG_RUN_RTPG;
+		kref_get(&pg->kref);
+		pg->rtpg_sdev = sdev;
+		scsi_device_get(sdev);
+		start_queue = 1;
+	}
+	spin_unlock_irqrestore(&pg->lock, flags);
+
+	if (start_queue &&
+	    !queue_delayed_work(kaluad_wq, &pg->rtpg_work,
+				msecs_to_jiffies(ALUA_RTPG_DELAY_MSECS))) {
+		scsi_device_put(sdev);
+		kref_put(&pg->kref, release_port_group);
+	}
+}
+
 /*
  * alua_initialize - Initialize ALUA state
  * @sdev: the device to be initialized
@@ -640,10 +802,12 @@ static int alua_initialize(struct scsi_device *sdev, struct alua_dh_data *h)
 {
 	int err = SCSI_DH_DEV_UNSUPP, tpgs;
 
+	mutex_lock(&h->init_mutex);
 	tpgs = alua_check_tpgs(sdev);
 	if (tpgs != TPGS_MODE_NONE)
 		err = alua_check_vpd(sdev, h, tpgs);
-
+	h->init_error = err;
+	mutex_unlock(&h->init_mutex);
 	return err;
 }
 /*
@@ -662,6 +826,7 @@ static int alua_set_params(struct scsi_device *sdev, const char *params)
 	unsigned int optimize = 0, argc;
 	const char *p = params;
 	int result = SCSI_DH_OK;
+	unsigned long flags;
 
 	if ((sscanf(params, "%u", &argc) != 1) || (argc != 1))
 		return -EINVAL;
@@ -671,14 +836,19 @@ static int alua_set_params(struct scsi_device *sdev, const char *params)
 	if ((sscanf(p, "%u", &optimize) != 1) || (optimize > 1))
 		return -EINVAL;
 
-	pg = h->pg;
-	if (!pg)
+	rcu_read_lock();
+	pg = rcu_dereference(h->pg);
+	if (!pg) {
+		rcu_read_unlock();
 		return -ENXIO;
-
+	}
+	spin_lock_irqsave(&pg->lock, flags);
 	if (optimize)
 		pg->flags |= ALUA_OPTIMIZE_STPG;
 	else
 		pg->flags |= ~ALUA_OPTIMIZE_STPG;
+	spin_unlock_irqrestore(&pg->lock, flags);
+	rcu_read_unlock();
 
 	return result;
 }
@@ -698,21 +868,33 @@ static int alua_activate(struct scsi_device *sdev,
 {
 	struct alua_dh_data *h = sdev->handler_data;
 	int err = SCSI_DH_OK;
+	struct alua_queue_data *qdata;
+	struct alua_port_group *pg;
 
-	if (!h->pg)
+	qdata = kzalloc(sizeof(*qdata), GFP_KERNEL);
+	if (!qdata) {
+		err = SCSI_DH_RES_TEMP_UNAVAIL;
 		goto out;
-
-	kref_get(&h->pg->kref);
-
-	err = alua_rtpg(sdev, h->pg);
-	if (err != SCSI_DH_OK) {
-		kref_put(&h->pg->kref, release_port_group);
+	}
+	qdata->callback_fn = fn;
+	qdata->callback_data = data;
+
+	mutex_lock(&h->init_mutex);
+	rcu_read_lock();
+	pg = rcu_dereference(h->pg);
+	if (!pg || !kref_get_unless_zero(&pg->kref)) {
+		rcu_read_unlock();
+		kfree(qdata);
+		err = h->init_error;
+		mutex_unlock(&h->init_mutex);
 		goto out;
 	}
-	err = alua_stpg(sdev, h->pg);
-	if (err == SCSI_DH_RETRY)
-		err = alua_rtpg(sdev, h->pg);
-	kref_put(&h->pg->kref, release_port_group);
+	mutex_unlock(&h->init_mutex);
+	fn = NULL;
+	rcu_read_unlock();
+
+	alua_rtpg_queue(pg, sdev, qdata);
+	kref_put(&pg->kref, release_port_group);
 out:
 	if (fn)
 		fn(data, err);
@@ -728,14 +910,19 @@ out:
 static int alua_prep_fn(struct scsi_device *sdev, struct request *req)
 {
 	struct alua_dh_data *h = sdev->handler_data;
-	int state;
+	struct alua_port_group *pg;
+	int state = TPGS_STATE_OPTIMIZED;
 	int ret = BLKPREP_OK;
 
-	if (!h->pg)
-		return ret;
-	kref_get(&h->pg->kref);
-	state = h->pg->state;
-	kref_put(&h->pg->kref, release_port_group);
+	rcu_read_lock();
+	pg = rcu_dereference(h->pg);
+	if (pg) {
+		state = pg->state;
+		/* Defer I/O while rtpg_work is active */
+		if (pg->rtpg_sdev)
+			state = TPGS_STATE_TRANSITIONING;
+	}
+	rcu_read_unlock();
 	if (state == TPGS_STATE_TRANSITIONING)
 		ret = BLKPREP_DEFER;
 	else if (state != TPGS_STATE_OPTIMIZED &&
@@ -760,9 +947,12 @@ static int alua_bus_attach(struct scsi_device *sdev)
 	h = kzalloc(sizeof(*h) , GFP_KERNEL);
 	if (!h)
 		return -ENOMEM;
-	h->pg = NULL;
+	spin_lock_init(&h->pg_lock);
+	rcu_assign_pointer(h->pg, NULL);
+	h->init_error = SCSI_DH_OK;
 	h->sdev = sdev;
 
+	mutex_init(&h->init_mutex);
 	err = alua_initialize(sdev, h);
 	if (err == SCSI_DH_NOMEM)
 		ret = -ENOMEM;
@@ -783,10 +973,18 @@ failed:
 static void alua_bus_detach(struct scsi_device *sdev)
 {
 	struct alua_dh_data *h = sdev->handler_data;
+	struct alua_port_group *pg;
 
-	if (h->pg) {
-		kref_put(&h->pg->kref, release_port_group);
-		h->pg = NULL;
+	spin_lock(&h->pg_lock);
+	pg = h->pg;
+	rcu_assign_pointer(h->pg, NULL);
+	h->sdev = NULL;
+	spin_unlock(&h->pg_lock);
+	if (pg) {
+		synchronize_rcu();
+		if (pg->rtpg_sdev)
+			flush_delayed_work(&pg->rtpg_work);
+		kref_put(&pg->kref, release_port_group);
 	}
 	sdev->handler_data = NULL;
 	kfree(h);
@@ -807,16 +1005,24 @@ static int __init alua_init(void)
 {
 	int r;
 
+	kaluad_wq = alloc_workqueue("kaluad", WQ_MEM_RECLAIM, 0);
+	if (!kaluad_wq) {
+		/* Temporary failure, bypass */
+		return SCSI_DH_DEV_TEMP_BUSY;
+	}
 	r = scsi_register_device_handler(&alua_dh);
-	if (r != 0)
+	if (r != 0) {
 		printk(KERN_ERR "%s: Failed to register scsi device handler",
 			ALUA_DH_NAME);
+		destroy_workqueue(kaluad_wq);
+	}
 	return r;
 }
 
 static void __exit alua_exit(void)
 {
 	scsi_unregister_device_handler(&alua_dh);
+	destroy_workqueue(kaluad_wq);
 }
 
 module_init(alua_init);
Signature · 1 line
1.8.5.6
#18 lore raw

Re: [PATCHv2 13/22] scsi_dh_alua: Use workqueue for RTPG

reproduce:

Hi Hannes,

[auto build test WARNING on scsi/for-next]
[also build test WARNING on next-20160112]
[cannot apply to v4.4]
[if your patch is applied to the wrong git tree, please drop us a note to help improving the system]

url:    https://github.com/0day-ci/linux/commits/Hannes-Reinecke/ALUA-device-handler-update-part-II/20160112-234244
base:   https://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi.git for-next
reproduce:
        # apt-get install sparse
        make ARCH=x86_64 allmodconfig
        make C=1 CF=-D__CHECK_ENDIAN__

sparse warnings: (new ones prefixed by >>)

include/linux/compiler.h:228:8: sparse: attribute 'no_sanitize_address': unknown attribute

>> drivers/scsi/device_handler/scsi_dh_alua.c:840:14: sparse: incompatible types in comparison expression (different address spaces)

drivers/scsi/device_handler/scsi_dh_alua.c:884:14: sparse: incompatible types in comparison expression (different address spaces) drivers/scsi/device_handler/scsi_dh_alua.c:918:14: sparse: incompatible types in comparison expression (different address spaces)

vim +840 drivers/scsi/device_handler/scsi_dh_alua.c

824 struct alua_dh_data *h = sdev->handler_data;
825 struct alua_port_group *pg = NULL;
826 unsigned int optimize = 0, argc;
827 const char *p = params;
828 int result = SCSI_DH_OK;
829 unsigned long flags;
830
831 if ((sscanf(params, "%u", &argc) != 1) || (argc != 1))
832 return -EINVAL;
833
834 while (*p++)
835 ;
836 if ((sscanf(p, "%u", &optimize) != 1) || (optimize > 1))
837 return -EINVAL;
838
839 rcu_read_lock();

 > 840		pg = rcu_dereference(h->pg);

841 if (!pg) {
842 rcu_read_unlock();
843 return -ENXIO;
844 }
845 spin_lock_irqsave(&pg->lock, flags);
846 if (optimize)
847 pg->flags |= ALUA_OPTIMIZE_STPG;
848 else

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

#19 lore raw

Re: [PATCHv2 13/22] scsi_dh_alua: Use workqueue for RTPG

In reply to Hannes Reinecke

drivers/scsi/device_handler/scsi_dh_alua.c struct alua_port_group *alua_lookup_pg(char *id_str, size_t id_size, · line 206

Quoted 3 lines
drivers/scsi/device_handler/scsi_dh_alua.c +2 -1
-		kref_get(&pg->kref);
+		if (!kref_get_unless_zero(&pg->kref))
+			continue;

As pointed out earlier this should be done from the start.

Quoted 14 lines
drivers/scsi/device_handler/scsi_dh_alua.c +14 -0
+	/* Check for existing port_group references */
+	spin_lock(&h->pg_lock);
+	if (h->pg) {
+		old_pg = pg;
+		/* port_group has changed. Update to new port group */
+		if (h->pg != pg) {
+			old_pg = h->pg;
+			rcu_assign_pointer(h->pg, pg);
+			pg_updated = true;
+		}
+	} else {
+		rcu_assign_pointer(h->pg, pg);
+		pg_updated = true;
+	}

drivers/scsi/device_handler/scsi_dh_alua.c static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h, · line 381

Quoted 10 lines
drivers/scsi/device_handler/scsi_dh_alua.c +10 -0
+	alua_rtpg_queue(h->pg, sdev, NULL);
+	spin_unlock(&h->pg_lock);
+
+	if (pg_updated)
+		synchronize_rcu();
+	if (old_pg) {
+		if (old_pg->rtpg_sdev)
+			flush_delayed_work(&old_pg->rtpg_work);
+		kref_put(&old_pg->kref, release_port_group);
+	}

The synchronize_rcu() needs to be done in release_port_group, or even better be replaced by doing a kfree_rcu there instead of a kfree.

And unless I'm mistaken the flush_delayed_work should probably be done in release_port_group as well.

#20 lore raw

Re: [PATCHv2 13/22] scsi_dh_alua: Use workqueue for RTPG

Quoted 6 lines · Christoph Hellwig

drivers/scsi/device_handler/scsi_dh_alua.c struct alua_port_group *alua_lookup_pg(char *id_str, size_t id_size, · line 206

Quoted 3 lines
drivers/scsi/device_handler/scsi_dh_alua.c +2 -1
-		kref_get(&pg->kref);
+		if (!kref_get_unless_zero(&pg->kref))
+			continue;

As pointed out earlier this should be done from the start.

Yep.

Quoted 19 lines
Quoted 19 lines
drivers/scsi/device_handler/scsi_dh_alua.c +19 -0
+	/* Check for existing port_group references */
+	spin_lock(&h->pg_lock);
+	if (h->pg) {
+		old_pg = pg;
+		/* port_group has changed. Update to new port group */
+		if (h->pg != pg) {
+			old_pg = h->pg;
+			rcu_assign_pointer(h->pg, pg);
+			pg_updated = true;
+		}
+	} else {
+		rcu_assign_pointer(h->pg, pg);
+		pg_updated = true;
+	}
+	alua_rtpg_queue(h->pg, sdev, NULL);
+	spin_unlock(&h->pg_lock);
+
+	if (pg_updated)
+		synchronize_rcu();
Quoted 9 lines

drivers/scsi/device_handler/scsi_dh_alua.c static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h, · line 381

Quoted 5 lines
drivers/scsi/device_handler/scsi_dh_alua.c +5 -0
+	if (old_pg) {
+		if (old_pg->rtpg_sdev)
+			flush_delayed_work(&old_pg->rtpg_work);
+		kref_put(&old_pg->kref, release_port_group);
+	}

The synchronize_rcu() needs to be done in release_port_group, or even better be replaced by doing a kfree_rcu there instead of a kfree.

Point is that we don't necessarily have an old_pg to call release_port_group() on, but we still need to call synchronize_rcu() to inform everyong that h->pg now has a new value.
So while I could do that it would end with a mess of if-clauses here.

Quoted 3 lines

And unless I'm mistaken the flush_delayed_work should probably be done in release_port_group as well.

_Actually_ we only need to call flush_delayed_work if sdev == rtgp_sdev. Otherwise the workqueue item is running off a different device and won't be affected.

Cheers,

Hannes

Signature · 9 lines
Dr. Hannes Reinecke		   Teamlead Storage & Networking
[email protected]			               +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#21 lore raw

Re: [PATCHv2 13/22] scsi_dh_alua: Use workqueue for RTPG

Quoted 42 lines · Hannes Reinecke
Quoted 6 lines · Christoph Hellwig

drivers/scsi/device_handler/scsi_dh_alua.c struct alua_port_group *alua_lookup_pg(char *id_str, size_t id_size, · line 206

Quoted 3 lines
drivers/scsi/device_handler/scsi_dh_alua.c +2 -1
-        kref_get(&pg->kref);
+        if (!kref_get_unless_zero(&pg->kref))
+            continue;

As pointed out earlier this should be done from the start.

Yep.

Quoted 25 lines
Quoted 14 lines
drivers/scsi/device_handler/scsi_dh_alua.c +14 -0
+    /* Check for existing port_group references */
+    spin_lock(&h->pg_lock);
+    if (h->pg) {
+        old_pg = pg;
+        /* port_group has changed. Update to new port group */
+        if (h->pg != pg) {
+            old_pg = h->pg;
+            rcu_assign_pointer(h->pg, pg);
+            pg_updated = true;
+        }
+    } else {
+        rcu_assign_pointer(h->pg, pg);
+        pg_updated = true;
+    }

drivers/scsi/device_handler/scsi_dh_alua.c static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h, · line 381

Quoted 10 lines
drivers/scsi/device_handler/scsi_dh_alua.c +10 -0
+    alua_rtpg_queue(h->pg, sdev, NULL);
+    spin_unlock(&h->pg_lock);
+
+    if (pg_updated)
+        synchronize_rcu();
+    if (old_pg) {
+        if (old_pg->rtpg_sdev)
+            flush_delayed_work(&old_pg->rtpg_work);
+        kref_put(&old_pg->kref, release_port_group);
+    }
Quoted 3 lines

The synchronize_rcu() needs to be done in release_port_group, or even better be replaced by doing a kfree_rcu there instead of a kfree.

Point is that we don't necessarily have an old_pg to call release_port_group() on, but we still need to call synchronize_rcu() to inform everyong that h->pg now has a new value.
So while I could do that it would end with a mess of if-clauses here.

Quoted 7 lines
Quoted 3 lines

And unless I'm mistaken the flush_delayed_work should probably be done in release_port_group as well.

_Actually_ we only need to call flush_delayed_work if sdev == rtgp_sdev. Otherwise the workqueue item is running off a different device and won't be affected.

Hmm. Well, not quite. We run into flush_delayed_work() only if the port group changes or upon bus detach. For all other callers rtpg_sdev() should already be NULL.
But looking at the call sites we can indeed move the flush_delayed_work() into the release function.

Cheers,

Hannes

Signature · 9 lines
Dr. Hannes Reinecke		   Teamlead Storage & Networking
[email protected]			               +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#22 lore raw

Re: [PATCHv2 13/22] scsi_dh_alua: Use workqueue for RTPG

Quoted 9 lines · Hannes Reinecke
Quoted 4 lines

_Actually_ we only need to call flush_delayed_work if sdev == rtgp_sdev. Otherwise the workqueue item is running off a different device and won't be affected.

Hmm. Well, not quite. We run into flush_delayed_work() only if the port group changes or upon bus detach.
For all other callers rtpg_sdev() should already be NULL.
But looking at the call sites we can indeed move the flush_delayed_work() into the release function.

Looks like the earlier you made sense after all :)

#23 lore raw

Re: [PATCHv2 13/22] scsi_dh_alua: Use workqueue for RTPG

In reply to Hannes Reinecke
Quoted 28 lines · Hannes Reinecke
Quoted 19 lines
Quoted 19 lines
drivers/scsi/device_handler/scsi_dh_alua.c +19 -0
+	/* Check for existing port_group references */
+	spin_lock(&h->pg_lock);
+	if (h->pg) {
+		old_pg = pg;
+		/* port_group has changed. Update to new port group */
+		if (h->pg != pg) {
+			old_pg = h->pg;
+			rcu_assign_pointer(h->pg, pg);
+			pg_updated = true;
+		}
+	} else {
+		rcu_assign_pointer(h->pg, pg);
+		pg_updated = true;
+	}
+	alua_rtpg_queue(h->pg, sdev, NULL);
+	spin_unlock(&h->pg_lock);
+
+	if (pg_updated)
+		synchronize_rcu();
Quoted 9 lines

drivers/scsi/device_handler/scsi_dh_alua.c static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h, · line 381

Quoted 5 lines
drivers/scsi/device_handler/scsi_dh_alua.c +5 -0
+	if (old_pg) {
+		if (old_pg->rtpg_sdev)
+			flush_delayed_work(&old_pg->rtpg_work);
+		kref_put(&old_pg->kref, release_port_group);
+	}

The synchronize_rcu() needs to be done in release_port_group, or even better be replaced by doing a kfree_rcu there instead of a kfree.

Quoted 3 lines

Point is that we don't necessarily have an old_pg to call release_port_group() on, but we still need to call synchronize_rcu() to inform everyong that h->pg now has a new value.

No, you don't. synchronize_rcu() just waits for all previously started RCU grace periods to finish. If old_pg doesn't get freed you don't need to wait for the grace period. synchronize_rcu does not notify anyone about anything, it just waits.

Quoted 6 lines
Quoted 3 lines

And unless I'm mistaken the flush_delayed_work should probably be done in release_port_group as well.

_Actually_ we only need to call flush_delayed_work if sdev == rtgp_sdev. Otherwise the workqueue item is running off a different device and won't be affected.

I'm actually pretty sure I suggested that in an earlier iteratation, but you said that's not the case. So please only run it if sdev == rtgp_sdev only then.

#24 lore raw

Re: [PATCHv2 13/22] scsi_dh_alua: Use workqueue for RTPG

Quoted 29 lines · Christoph Hellwig
Quoted 28 lines · Hannes Reinecke
Quoted 28 lines
Quoted 14 lines
drivers/scsi/device_handler/scsi_dh_alua.c +14 -0
+	/* Check for existing port_group references */
+	spin_lock(&h->pg_lock);
+	if (h->pg) {
+		old_pg = pg;
+		/* port_group has changed. Update to new port group */
+		if (h->pg != pg) {
+			old_pg = h->pg;
+			rcu_assign_pointer(h->pg, pg);
+			pg_updated = true;
+		}
+	} else {
+		rcu_assign_pointer(h->pg, pg);
+		pg_updated = true;
+	}

drivers/scsi/device_handler/scsi_dh_alua.c static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h, · line 381

Quoted 10 lines
drivers/scsi/device_handler/scsi_dh_alua.c +10 -0
+	alua_rtpg_queue(h->pg, sdev, NULL);
+	spin_unlock(&h->pg_lock);
+
+	if (pg_updated)
+		synchronize_rcu();
+	if (old_pg) {
+		if (old_pg->rtpg_sdev)
+			flush_delayed_work(&old_pg->rtpg_work);
+		kref_put(&old_pg->kref, release_port_group);
+	}

The synchronize_rcu() needs to be done in release_port_group, or even better be replaced by doing a kfree_rcu there instead of a kfree.

Quoted 9 lines
Quoted 3 lines

Point is that we don't necessarily have an old_pg to call release_port_group() on, but we still need to call synchronize_rcu() to inform everyong that h->pg now has a new value.

No, you don't. synchronize_rcu() just waits for all previously started RCU grace periods to finish. If old_pg doesn't get freed you don't need to wait for the grace period. synchronize_rcu does not notify anyone about anything, it just waits.

Ah. Right. Good point. Of course we don't need to call it if there's nothing to be freed.

Will be folding it into the next round of patches, or just resending this one if the remaining patches meets with approval.

Cheers,

Hannes

Signature · 9 lines
Dr. Hannes Reinecke		   Teamlead Storage & Networking
[email protected]			               +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#25 lore raw

Re: [PATCHv2 13/22] scsi_dh_alua: Use workqueue for RTPG

In reply to Hannes Reinecke
reproduce: wget https://git.kernel.org/cgit/linux/kernel/git/wfg/lkp-tests.git/plain/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross # save the attached .config to linux build tree…

Hi Hannes,

[auto build test WARNING on scsi/for-next]
[also build test WARNING on next-20160112]
[cannot apply to v4.4]
[if your patch is applied to the wrong git tree, please drop us a note to help improving the system]

url:    https://github.com/0day-ci/linux/commits/Hannes-Reinecke/ALUA-device-handler-update-part-II/20160112-234244
base:   https://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi.git for-next
config: s390-allyesconfig (attached as .config)

reproduce:
wget https://git.kernel.org/cgit/linux/kernel/git/wfg/lkp-tests.git/plain/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=s390

All warnings (new ones prefixed by >>):

drivers/scsi/device_handler/scsi_dh_alua.c: In function 'alua_rtpg':
drivers/scsi/device_handler/scsi_dh_alua.c:630:1: warning: 'alua_rtpg' uses dynamic stack allocation

    }
    ^
   drivers/scsi/device_handler/scsi_dh_alua.c: In function 'alua_rtpg_work':
>> drivers/scsi/device_handler/scsi_dh_alua.c:759:1: warning: 'alua_rtpg_work' uses dynamic stack allocation
    }
    ^

vim +/alua_rtpg_work +759 drivers/scsi/device_handler/scsi_dh_alua.c

624 err = SCSI_DH_OK;
625 pg->expiry = 0;
626 break;
627 }
628 kfree(buff);
629 return err;

 > 630	}
   631	
   632	/*
   633	 * alua_stpg - Issue a SET TARGET GROUP STATES command
   634	 *
   635	 * Issue a SET TARGET GROUP STATES command and evaluate the
   636	 * response. Returns SCSI_DH_RETRY per default to trigger
   637	 * a re-evaluation of the target group state or SCSI_DH_OK
   638	 * if no further action needs to be taken.
   639	 */
   640	static unsigned alua_stpg(struct scsi_device *sdev, struct alua_port_group *pg)
   641	{
   642		int retval;
   643		struct scsi_sense_hdr sense_hdr;
   644	
   645		if (!(pg->tpgs & TPGS_MODE_EXPLICIT)) {
   646			/* Only implicit ALUA supported, retry */
   647			return SCSI_DH_RETRY;
   648		}
   649		switch (pg->state) {
   650		case TPGS_STATE_OPTIMIZED:
   651			return SCSI_DH_OK;
   652		case TPGS_STATE_NONOPTIMIZED:
   653			if ((pg->flags & ALUA_OPTIMIZE_STPG) &&
   654			    !pg->pref &&
   655			    (pg->tpgs & TPGS_MODE_IMPLICIT))
   656				return SCSI_DH_OK;
   657			break;
   658		case TPGS_STATE_STANDBY:
   659		case TPGS_STATE_UNAVAILABLE:
   660			break;
   661		case TPGS_STATE_OFFLINE:
   662			return SCSI_DH_IO;
   663			break;
   664		case TPGS_STATE_TRANSITIONING:
   665			break;
   666		default:
   667			sdev_printk(KERN_INFO, sdev,
   668				    "%s: stpg failed, unhandled TPGS state %d",
   669				    ALUA_DH_NAME, pg->state);
   670			return SCSI_DH_NOSYS;
   671			break;
   672		}
   673		/* Set state to transitioning */
   674		pg->state = TPGS_STATE_TRANSITIONING;
   675		retval = submit_stpg(sdev, pg->group_id, &sense_hdr);
   676	
   677		if (retval) {
   678			if (!scsi_sense_valid(&sense_hdr)) {
   679				sdev_printk(KERN_INFO, sdev,
   680					    "%s: stpg failed, result %d",
   681					    ALUA_DH_NAME, retval);
   682				if (driver_byte(retval) == DRIVER_ERROR)
   683					return SCSI_DH_DEV_TEMP_BUSY;
   684			} else {
   685				sdev_printk(KERN_INFO, sdev, "%s: stpg failed\n",
   686					    ALUA_DH_NAME);
   687				scsi_print_sense_hdr(sdev, ALUA_DH_NAME, &sense_hdr);
   688			}
   689		}
   690		/* Retry RTPG */
   691		return SCSI_DH_RETRY;
   692	}
   693	
   694	static void alua_rtpg_work(struct work_struct *work)
   695	{
   696		struct alua_port_group *pg =
   697			container_of(work, struct alua_port_group, rtpg_work.work);
   698		struct scsi_device *sdev;
   699		LIST_HEAD(qdata_list);
   700		int err = SCSI_DH_OK;
   701		struct alua_queue_data *qdata, *tmp;
   702		unsigned long flags;
   703	
   704		spin_lock_irqsave(&pg->lock, flags);
   705		sdev = pg->rtpg_sdev;
   706		if (!sdev) {
   707			WARN_ON(pg->flags & ALUA_PG_RUN_RTPG ||
   708				pg->flags & ALUA_PG_RUN_STPG);
   709			spin_unlock_irqrestore(&pg->lock, flags);
   710			return;
   711		}
   712		pg->flags |= ALUA_PG_RUNNING;
   713		if (pg->flags & ALUA_PG_RUN_RTPG) {
   714			spin_unlock_irqrestore(&pg->lock, flags);
   715			err = alua_rtpg(sdev, pg);
   716			spin_lock_irqsave(&pg->lock, flags);
   717			if (err == SCSI_DH_RETRY) {
   718				pg->flags &= ~ALUA_PG_RUNNING;
   719				spin_unlock_irqrestore(&pg->lock, flags);
   720				queue_delayed_work(kaluad_wq, &pg->rtpg_work,
   721						   pg->interval * HZ);
   722				return;
   723			}
   724			pg->flags &= ~ALUA_PG_RUN_RTPG;
   725			if (err != SCSI_DH_OK)
   726				pg->flags &= ~ALUA_PG_RUN_STPG;
   727		}
   728		if (pg->flags & ALUA_PG_RUN_STPG) {
   729			spin_unlock_irqrestore(&pg->lock, flags);
   730			err = alua_stpg(sdev, pg);
   731			spin_lock_irqsave(&pg->lock, flags);
   732			pg->flags &= ~ALUA_PG_RUN_STPG;
   733			if (err == SCSI_DH_RETRY) {
   734				pg->flags |= ALUA_PG_RUN_RTPG;
   735				pg->interval = 0;
   736				pg->flags &= ~ALUA_PG_RUNNING;
   737				spin_unlock_irqrestore(&pg->lock, flags);
   738				queue_delayed_work(kaluad_wq, &pg->rtpg_work,
   739						   pg->interval * HZ);
   740				return;
   741			}
   742		}
   743	
   744		list_splice_init(&pg->rtpg_list, &qdata_list);
   745		pg->rtpg_sdev = NULL;
   746		spin_unlock_irqrestore(&pg->lock, flags);
   747	
   748		list_for_each_entry_safe(qdata, tmp, &qdata_list, entry) {
   749			list_del(&qdata->entry);
   750			if (qdata->callback_fn)
   751				qdata->callback_fn(qdata->callback_data, err);
   752			kfree(qdata);
   753		}
   754		spin_lock_irqsave(&pg->lock, flags);
   755		pg->flags &= ~ALUA_PG_RUNNING;
   756		spin_unlock_irqrestore(&pg->lock, flags);
   757		scsi_device_put(sdev);
   758		kref_put(&pg->kref, release_port_group);
 > 759	}

760
761 static void alua_rtpg_queue(struct alua_port_group *pg,
762 struct scsi_device *sdev,

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

#26 lore raw

PATCH v2 14/22 scsi_dh_alua: Allow workqueue to run synchronously

In reply to Hannes Reinecke

Some arrays may only capable of handling one STPG at a time, so this patch adds a singlethreaded workqueue for STPGs to be submitted synchronously.
Synchronous STPG can be activated by setting the second module option to '1'. The corresponding entry in multipath.conf would be:
hardware_handler "3 alua 1 1"

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 41 ++++++++++++++++++++++++------
 1 file changed, 33 insertions(+), 8 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +33 -8
index 9d56d6a..eb12074 100644
@@ -64,6 +64,7 @@
 /* device handler flags */
 #define ALUA_OPTIMIZE_STPG		0x01
 #define ALUA_RTPG_EXT_HDR_UNSUPP	0x02
+#define ALUA_SYNC_STPG			0x04
 /* State machine flags */
 #define ALUA_PG_RUN_RTPG		0x10
 #define ALUA_PG_RUN_STPG		0x20
@@ -76,6 +77,7 @@ MODULE_PARM_DESC(optimize_stpg, "Allow use of a non-optimized path, rather than
 static LIST_HEAD(port_group_list);
 static DEFINE_SPINLOCK(port_group_lock);
 static struct workqueue_struct *kaluad_wq;
+static struct workqueue_struct *kaluad_sync_wq;
 
 struct alua_port_group {
 	struct kref		kref;
@@ -700,6 +702,7 @@ static void alua_rtpg_work(struct work_struct *work)
 	int err = SCSI_DH_OK;
 	struct alua_queue_data *qdata, *tmp;
 	unsigned long flags;
+	struct workqueue_struct *alua_wq = kaluad_wq;
 
 	spin_lock_irqsave(&pg->lock, flags);
 	sdev = pg->rtpg_sdev;
@@ -709,6 +712,8 @@ static void alua_rtpg_work(struct work_struct *work)
 		spin_unlock_irqrestore(&pg->lock, flags);
 		return;
 	}
+	if (pg->flags & ALUA_SYNC_STPG)
+		alua_wq = kaluad_sync_wq;
 	pg->flags |= ALUA_PG_RUNNING;
 	if (pg->flags & ALUA_PG_RUN_RTPG) {
 		spin_unlock_irqrestore(&pg->lock, flags);
@@ -717,7 +722,7 @@ static void alua_rtpg_work(struct work_struct *work)
 		if (err == SCSI_DH_RETRY) {
 			pg->flags &= ~ALUA_PG_RUNNING;
 			spin_unlock_irqrestore(&pg->lock, flags);
-			queue_delayed_work(kaluad_wq, &pg->rtpg_work,
+			queue_delayed_work(alua_wq, &pg->rtpg_work,
 					   pg->interval * HZ);
 			return;
 		}
@@ -735,7 +740,7 @@ static void alua_rtpg_work(struct work_struct *work)
 			pg->interval = 0;
 			pg->flags &= ~ALUA_PG_RUNNING;
 			spin_unlock_irqrestore(&pg->lock, flags);
-			queue_delayed_work(kaluad_wq, &pg->rtpg_work,
+			queue_delayed_work(alua_wq, &pg->rtpg_work,
 					   pg->interval * HZ);
 			return;
 		}
@@ -764,6 +769,7 @@ static void alua_rtpg_queue(struct alua_port_group *pg,
 {
 	int start_queue = 0;
 	unsigned long flags;
+	struct workqueue_struct *alua_wq = kaluad_wq;
 
 	if (!pg)
 		return;
@@ -781,10 +787,12 @@ static void alua_rtpg_queue(struct alua_port_group *pg,
 		scsi_device_get(sdev);
 		start_queue = 1;
 	}
+	if (pg->flags & ALUA_SYNC_STPG)
+		alua_wq = kaluad_sync_wq;
 	spin_unlock_irqrestore(&pg->lock, flags);
 
 	if (start_queue &&
-	    !queue_delayed_work(kaluad_wq, &pg->rtpg_work,
+	    !queue_delayed_work(alua_wq, &pg->rtpg_work,
 				msecs_to_jiffies(ALUA_RTPG_DELAY_MSECS))) {
 		scsi_device_put(sdev);
 		kref_put(&pg->kref, release_port_group);
@@ -811,30 +819,36 @@ static int alua_initialize(struct scsi_device *sdev, struct alua_dh_data *h)
 	return err;
 }
 /*
- * alua_set_params - set/unset the optimize flag
+ * alua_set_params - set/unset the optimize and sync flag
  * @sdev: device on the path to be activated
  * params - parameters in the following format
  *      "no_of_params\0param1\0param2\0param3\0...\0"
  * For example, to set the flag pass the following parameters
  * from multipath.conf
- *     hardware_handler        "2 alua 1"
+ *     hardware_handler        "3 alua 1 1"
  */
 static int alua_set_params(struct scsi_device *sdev, const char *params)
 {
 	struct alua_dh_data *h = sdev->handler_data;
 	struct alua_port_group *pg = NULL;
-	unsigned int optimize = 0, argc;
+	unsigned int optimize = 0, sync = 0, argc;
 	const char *p = params;
 	int result = SCSI_DH_OK;
 	unsigned long flags;
 
-	if ((sscanf(params, "%u", &argc) != 1) || (argc != 1))
+	if ((sscanf(params, "%u", &argc) != 1) || (argc > 2))
 		return -EINVAL;
 
 	while (*p++)
 		;
 	if ((sscanf(p, "%u", &optimize) != 1) || (optimize > 1))
 		return -EINVAL;
+	if (argc == 2) {
+		while (*p++)
+			;
+		if ((sscanf(p, "%u", &sync) != 1) || (sync > 1))
+			return -EINVAL;
+	}
 
 	rcu_read_lock();
 	pg = rcu_dereference(h->pg);
@@ -846,7 +860,11 @@ static int alua_set_params(struct scsi_device *sdev, const char *params)
 	if (optimize)
 		pg->flags |= ALUA_OPTIMIZE_STPG;
 	else
-		pg->flags |= ~ALUA_OPTIMIZE_STPG;
+		pg->flags &= ~ALUA_OPTIMIZE_STPG;
+	if (sync)
+		pg->flags |= ALUA_SYNC_STPG;
+	else
+		pg->flags &= ~ALUA_SYNC_STPG;
 	spin_unlock_irqrestore(&pg->lock, flags);
 	rcu_read_unlock();
 
@@ -1010,10 +1028,16 @@ static int __init alua_init(void)
 		/* Temporary failure, bypass */
 		return SCSI_DH_DEV_TEMP_BUSY;
 	}
+	kaluad_sync_wq = create_workqueue("kaluad_sync");
+	if (!kaluad_sync_wq) {
+		destroy_workqueue(kaluad_wq);
+		return SCSI_DH_DEV_TEMP_BUSY;
+	}
 	r = scsi_register_device_handler(&alua_dh);
 	if (r != 0) {
 		printk(KERN_ERR "%s: Failed to register scsi device handler",
 			ALUA_DH_NAME);
+		destroy_workqueue(kaluad_sync_wq);
 		destroy_workqueue(kaluad_wq);
 	}
 	return r;
@@ -1022,6 +1046,7 @@ static int __init alua_init(void)
 static void __exit alua_exit(void)
 {
 	scsi_unregister_device_handler(&alua_dh);
+	destroy_workqueue(kaluad_sync_wq);
 	destroy_workqueue(kaluad_wq);
 }
 
Signature · 1 line
1.8.5.6
#27 lore raw

Re: [PATCHv2 14/22] scsi_dh_alua: Allow workqueue to run synchronously

Quoted 6 lines · Hannes Reinecke

Some arrays may only capable of handling one STPG at a time, so this patch adds a singlethreaded workqueue for STPGs to be submitted synchronously.
Synchronous STPG can be activated by setting the second module option to '1'. The corresponding entry in multipath.conf would be:
hardware_handler "3 alua 1 1"

No new hardware_handle options please. The SCSI way to handle hardware tweaks is by a devlist entry. The devlist entries ensure setups work out of the box, but if needed still can be updated dynamically.

#28 lore raw

PATCH v2 15/22 scsi_dh_alua: Recheck state on unit attention

In reply to Hannes Reinecke

When we receive a unit attention code of 'ALUA state changed' we should recheck the state, as it might be due to an implicit ALUA state transition. This allows us to return NEEDS_RETRY instead of ADD_TO_MLQUEUE, allowing to terminate the retries after a certain time.
At the same time a workqueue item might already be queued, which should be started immediately to avoid any delays.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 57 ++++++++++++++++++++++++------
 1 file changed, 46 insertions(+), 11 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +46 -11
index eb12074..dd7438b 100644
@@ -119,7 +119,8 @@ struct alua_queue_data {
 static void alua_rtpg_work(struct work_struct *work);
 static void alua_rtpg_queue(struct alua_port_group *pg,
 			    struct scsi_device *sdev,
-			    struct alua_queue_data *qdata);
+			    struct alua_queue_data *qdata, bool force);
+static void alua_check(struct scsi_device *sdev, bool force);
 
 static void release_port_group(struct kref *kref)
 {
@@ -371,7 +372,7 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h,
 		rcu_assign_pointer(h->pg, pg);
 		pg_updated = true;
 	}
-	alua_rtpg_queue(h->pg, sdev, NULL);
+	alua_rtpg_queue(h->pg, sdev, NULL, true);
 	spin_unlock(&h->pg_lock);
 
 	if (pg_updated)
@@ -412,18 +413,24 @@ static int alua_check_sense(struct scsi_device *sdev,
 {
 	switch (sense_hdr->sense_key) {
 	case NOT_READY:
-		if (sense_hdr->asc == 0x04 && sense_hdr->ascq == 0x0a)
+		if (sense_hdr->asc == 0x04 && sense_hdr->ascq == 0x0a) {
 			/*
 			 * LUN Not Accessible - ALUA state transition
 			 */
-			return ADD_TO_MLQUEUE;
+			alua_check(sdev, false);
+			return NEEDS_RETRY;
+		}
 		break;
 	case UNIT_ATTENTION:
-		if (sense_hdr->asc == 0x29 && sense_hdr->ascq == 0x00)
+		if (sense_hdr->asc == 0x29 && sense_hdr->ascq == 0x00) {
 			/*
-			 * Power On, Reset, or Bus Device Reset, just retry.
+			 * Power On, Reset, or Bus Device Reset.
+			 * Might have obscured a state transition,
+			 * so schedule a recheck.
 			 */
+			alua_check(sdev, true);
 			return ADD_TO_MLQUEUE;
+		}
 		if (sense_hdr->asc == 0x29 && sense_hdr->ascq == 0x04)
 			/*
 			 * Device internal reset
@@ -434,16 +441,20 @@ static int alua_check_sense(struct scsi_device *sdev,
 			 * Mode Parameters Changed
 			 */
 			return ADD_TO_MLQUEUE;
-		if (sense_hdr->asc == 0x2a && sense_hdr->ascq == 0x06)
+		if (sense_hdr->asc == 0x2a && sense_hdr->ascq == 0x06) {
 			/*
 			 * ALUA state changed
 			 */
+			alua_check(sdev, true);
 			return ADD_TO_MLQUEUE;
-		if (sense_hdr->asc == 0x2a && sense_hdr->ascq == 0x07)
+		}
+		if (sense_hdr->asc == 0x2a && sense_hdr->ascq == 0x07) {
 			/*
 			 * Implicit ALUA state transition failed
 			 */
+			alua_check(sdev, true);
 			return ADD_TO_MLQUEUE;
+		}
 		if (sense_hdr->asc == 0x3f && sense_hdr->ascq == 0x03)
 			/*
 			 * Inquiry data has changed
@@ -765,7 +776,7 @@ static void alua_rtpg_work(struct work_struct *work)
 
 static void alua_rtpg_queue(struct alua_port_group *pg,
 			    struct scsi_device *sdev,
-			    struct alua_queue_data *qdata)
+			    struct alua_queue_data *qdata, bool force)
 {
 	int start_queue = 0;
 	unsigned long flags;
@@ -786,7 +797,9 @@ static void alua_rtpg_queue(struct alua_port_group *pg,
 		pg->rtpg_sdev = sdev;
 		scsi_device_get(sdev);
 		start_queue = 1;
-	}
+	} else if (!(pg->flags & ALUA_PG_RUN_RTPG) && force)
+		start_queue = 1;
+
 	if (pg->flags & ALUA_SYNC_STPG)
 		alua_wq = kaluad_sync_wq;
 	spin_unlock_irqrestore(&pg->lock, flags);
@@ -911,7 +924,7 @@ static int alua_activate(struct scsi_device *sdev,
 	fn = NULL;
 	rcu_read_unlock();
 
-	alua_rtpg_queue(pg, sdev, qdata);
+	alua_rtpg_queue(pg, sdev, qdata, true);
 	kref_put(&pg->kref, release_port_group);
 out:
 	if (fn)
@@ -920,6 +933,28 @@ out:
 }
 
 /*
+ * alua_check - check path status
+ * @sdev: device on the path to be checked
+ *
+ * Check the device status
+ */
+static void alua_check(struct scsi_device *sdev, bool force)
+{
+	struct alua_dh_data *h = sdev->handler_data;
+	struct alua_port_group *pg;
+
+	rcu_read_lock();
+	pg = rcu_dereference(h->pg);
+	if (!pg || !kref_get_unless_zero(&pg->kref)) {
+		rcu_read_unlock();
+		return;
+	}
+	rcu_read_unlock();
+	alua_rtpg_queue(pg, sdev, NULL, force);
+	kref_put(&pg->kref, release_port_group);
+}
+
+/*
  * alua_prep_fn - request callback
  *
  * Fail I/O to all paths not in state
Signature · 1 line
1.8.5.6
#30 lore raw

PATCH v2 16/22 scsi_dh_alua: update all port states

In reply to Hannes Reinecke

When we read in the target port group state we should be updating all affected port groups, otherwise we risk running out of sync.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 31 +++++++++++++++++++++---------
 1 file changed, 22 insertions(+), 9 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +22 -9
index dd7438b..6fefe9f 100644
@@ -484,11 +484,13 @@ static int alua_check_sense(struct scsi_device *sdev,
 static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 {
 	struct scsi_sense_hdr sense_hdr;
+	struct alua_port_group *tmp_pg;
 	int len, k, off, valid_states = 0, bufflen = ALUA_RTPG_SIZE;
-	unsigned char *ucp, *buff;
+	unsigned char *desc, *buff;
 	unsigned err, retval;
 	unsigned int tpg_desc_tbl_off;
 	unsigned char orig_transition_tmo;
+	unsigned long flags;
 
 	if (!pg->expiry) {
 		unsigned long transition_tmo = ALUA_FAILOVER_TIMEOUT * HZ;
@@ -590,16 +592,27 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 	else
 		tpg_desc_tbl_off = 4;
 
-	for (k = tpg_desc_tbl_off, ucp = buff + tpg_desc_tbl_off;
+	for (k = tpg_desc_tbl_off, desc = buff + tpg_desc_tbl_off;
 	     k < len;
-	     k += off, ucp += off) {
-
-		if (pg->group_id == get_unaligned_be16(&ucp[2])) {
-			pg->state = ucp[0] & 0x0f;
-			pg->pref = ucp[0] >> 7;
-			valid_states = ucp[1];
+	     k += off, desc += off) {
+		u16 group_id = get_unaligned_be16(&desc[2]);
+
+		spin_lock_irqsave(&port_group_lock, flags);
+		list_for_each_entry(tmp_pg, &port_group_list, node) {
+			if (tmp_pg->group_id != group_id)
+				continue;
+			if (tmp_pg->device_id_len != pg->device_id_len)
+				continue;
+			if (strncmp(tmp_pg->device_id_str, pg->device_id_str,
+				    tmp_pg->device_id_len))
+				continue;
+			tmp_pg->state = desc[0] & 0x0f;
+			tmp_pg->pref = desc[0] >> 7;
+			if (tmp_pg == pg)
+				valid_states = desc[1];
 		}
-		off = 8 + (ucp[7] * 4);
+		spin_unlock_irqrestore(&port_group_lock, flags);
+		off = 8 + (desc[7] * 4);
 	}
 
 	sdev_printk(KERN_INFO, sdev,
Signature · 1 line
1.8.5.6
#31 lore raw

PATCH v2 17/22 scsi_dh_alua: Send TEST UNIT READY to poll for transitioning

In reply to Hannes Reinecke

Sending a 'REPORT TARGET PORT GROUP' command is a costly operation, as the array has to gather information about all ports. So instead of using RTPG to poll for a status update when a port is in transitioning we should be sending a TEST UNIT READY, and wait for the sense code to report success.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 37 ++++++++++++++++++++++++++++++
 1 file changed, 37 insertions(+)
drivers/scsi/device_handler/scsi_dh_alua.c +37 -0
index 6fefe9f..9df2203 100644
@@ -474,6 +474,30 @@ static int alua_check_sense(struct scsi_device *sdev,
 }
 
 /*
+ * alua_tur - Send a TEST UNIT READY
+ * @sdev: device to which the TEST UNIT READY command should be send
+ *
+ * Send a TEST UNIT READY to @sdev to figure out the device state
+ * Returns SCSI_DH_RETRY if the sense code is NOT READY/ALUA TRANSITIONING,
+ * SCSI_DH_OK if no error occurred, and SCSI_DH_IO otherwise.
+ */
+static int alua_tur(struct scsi_device *sdev)
+{
+	struct scsi_sense_hdr sense_hdr;
+	int retval;
+
+	retval = scsi_test_unit_ready(sdev, ALUA_FAILOVER_TIMEOUT * HZ,
+				      ALUA_FAILOVER_RETRIES, &sense_hdr);
+	if (sense_hdr.sense_key == NOT_READY &&
+	    sense_hdr.asc == 0x04 && sense_hdr.ascq == 0x0a)
+		return SCSI_DH_RETRY;
+	else if (retval)
+		return SCSI_DH_IO;
+	else
+		return SCSI_DH_OK;
+}
+
+/*
  * alua_rtpg - Evaluate REPORT TARGET GROUP STATES
  * @sdev: the device to be evaluated.
  *
@@ -740,7 +764,20 @@ static void alua_rtpg_work(struct work_struct *work)
 		alua_wq = kaluad_sync_wq;
 	pg->flags |= ALUA_PG_RUNNING;
 	if (pg->flags & ALUA_PG_RUN_RTPG) {
+		int state = pg->state;
+
 		spin_unlock_irqrestore(&pg->lock, flags);
+		if (state == TPGS_STATE_TRANSITIONING) {
+			if (alua_tur(sdev) == SCSI_DH_RETRY) {
+				spin_lock_irqsave(&pg->lock, flags);
+				pg->flags &= ~ALUA_PG_RUNNING;
+				spin_unlock_irqrestore(&pg->lock, flags);
+				queue_delayed_work(alua_wq, &pg->rtpg_work,
+						   pg->interval * HZ);
+				return;
+			}
+			/* Send RTPG on failure or if TUR indicates SUCCESS */
+		}
 		err = alua_rtpg(sdev, pg);
 		spin_lock_irqsave(&pg->lock, flags);
 		if (err == SCSI_DH_RETRY) {
Signature · 1 line
1.8.5.6
#32 lore raw

PATCH v2 18/22 scsi_dh: add 'rescan' callback

In reply to Hannes Reinecke

If a device needs to be rescanned the device_handler might need to be rechecked, too.
So add a 'rescan' callback to the device handler and call it upon scsi_rescan_device().

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 8 ++++++++
 drivers/scsi/scsi_lib.c                    | 1 +
 drivers/scsi/scsi_scan.c                   | 8 +++++++-
 include/scsi/scsi_dh.h                     | 1 +
 4 files changed, 17 insertions(+), 1 deletion(-)
drivers/scsi/device_handler/scsi_dh_alua.c +8 -0
index 9df2203..ffc419e 100644
@@ -1038,6 +1038,13 @@ static int alua_prep_fn(struct scsi_device *sdev, struct request *req)
 
 }
 
+static void alua_rescan(struct scsi_device *sdev)
+{
+	struct alua_dh_data *h = sdev->handler_data;
+
+	alua_initialize(sdev, h);
+}
+
 /*
  * alua_bus_attach - Attach device handler
  * @sdev: device to be attached to
@@ -1101,6 +1108,7 @@ static struct scsi_device_handler alua_dh = {
 	.prep_fn = alua_prep_fn,
 	.check_sense = alua_check_sense,
 	.activate = alua_activate,
+	.rescan = alua_rescan,
 	.set_params = alua_set_params,
 };
 
drivers/scsi/scsi_lib.c +1 -0
index fa6b2c4..d46193a 100644
@@ -2699,6 +2699,7 @@ static void scsi_evt_emit(struct scsi_device *sdev, struct scsi_event *evt)
 		envp[idx++] = "SDEV_MEDIA_CHANGE=1";
 		break;
 	case SDEV_EVT_INQUIRY_CHANGE_REPORTED:
+		scsi_rescan_device(&sdev->sdev_gendev);
 		envp[idx++] = "SDEV_UA=INQUIRY_DATA_HAS_CHANGED";
 		break;
 	case SDEV_EVT_CAPACITY_CHANGE_REPORTED:
drivers/scsi/scsi_scan.c +7 -1
index a1c195d..05dc1aa 100644
@@ -43,6 +43,7 @@
 #include <scsi/scsi_devinfo.h>
 #include <scsi/scsi_host.h>
 #include <scsi/scsi_transport.h>
+#include <scsi/scsi_dh.h>
 #include <scsi/scsi_eh.h>
 
 #include "scsi_priv.h"
@@ -1516,9 +1517,14 @@ EXPORT_SYMBOL(scsi_add_device);
 
 void scsi_rescan_device(struct device *dev)
 {
+	struct scsi_device *sdev = to_scsi_device(dev);
+
 	device_lock(dev);
 
-	scsi_attach_vpd(to_scsi_device(dev));
+	scsi_attach_vpd(sdev);
+
+	if (sdev->handler && sdev->handler->rescan)
+		sdev->handler->rescan(sdev);
 
 	if (dev->driver && try_module_get(dev->driver->owner)) {
 		struct scsi_driver *drv = to_scsi_driver(dev->driver);
include/scsi/scsi_dh.h +1 -0
index 7e184c6..c7bba2b 100644
@@ -71,6 +71,7 @@ struct scsi_device_handler {
 	int (*activate)(struct scsi_device *, activate_complete, void *);
 	int (*prep_fn)(struct scsi_device *, struct request *);
 	int (*set_params)(struct scsi_device *, const char *);
+	void (*rescan)(struct scsi_device *);
 };
 
 #ifdef CONFIG_SCSI_DH
Signature · 1 line
1.8.5.6
#33 lore raw

PATCH v2 19/22 scsi: Add 'access_state' attribute

In reply to Hannes Reinecke

Add an 'access_state' attribute to struct scsi_device to display the asymmetric LUN access state.

---
 drivers/scsi/scsi_scan.c   |  1 +
 drivers/scsi/scsi_sysfs.c  | 49 ++++++++++++++++++++++++++++++++++++++++++++++
 include/scsi/scsi_device.h |  1 +
 include/scsi/scsi_proto.h  | 13 ++++++++++++
 4 files changed, 64 insertions(+)
drivers/scsi/scsi_scan.c +1 -0
index 05dc1aa..9fc1531 100644
@@ -231,6 +231,7 @@ static struct scsi_device *scsi_alloc_sdev(struct scsi_target *starget,
 	sdev->lun = lun;
 	sdev->channel = starget->channel;
 	sdev->sdev_state = SDEV_CREATED;
+	sdev->access_state = SCSI_ACCESS_STATE_UNKNOWN;
 	INIT_LIST_HEAD(&sdev->siblings);
 	INIT_LIST_HEAD(&sdev->same_target_siblings);
 	INIT_LIST_HEAD(&sdev->cmd_list);
drivers/scsi/scsi_sysfs.c +49 -0

67 lines not on this page · raw message

include/scsi/scsi_device.h +1 -0
index f63a167..5dd00e9f 100644
@@ -200,6 +200,7 @@ struct scsi_device {
 	struct scsi_device_handler *handler;
 	void			*handler_data;
 
+	enum scsi_access_state	access_state;
 	enum scsi_device_state sdev_state;
 	unsigned long		sdev_data[0];
 } __attribute__((aligned(sizeof(unsigned long))));
include/scsi/scsi_proto.h +13 -0
index a9fbf1b..80e85e7 100644
@@ -277,5 +277,18 @@ struct scsi_lun {
 	__u8 scsi_lun[8];
 };
 
+/* SPC asymmetric access states */
+enum scsi_access_state {
+	SCSI_ACCESS_STATE_OPTIMAL = 0,
+	SCSI_ACCESS_STATE_ACTIVE,
+	SCSI_ACCESS_STATE_STANDBY,
+	SCSI_ACCESS_STATE_UNAVAILABLE,
+	SCSI_ACCESS_STATE_LBA,
+	SCSI_ACCESS_STATE_OFFLINE = 0xe,
+	SCSI_ACCESS_STATE_TRANSITIONING = 0xf,
+	SCSI_ACCESS_STATE_UNKNOWN = 0x70,
+};
+#define SCSI_ACCESS_STATE_MASK 0x0f
+#define SCSI_ACCESS_STATE_PREFERRED 0x80
 
 #endif /* _SCSI_PROTO_H_ */
Signature · 1 line
1.8.5.6
#34 lore raw

PATCH v2 20/22 scsi_dh_alua: use common definitions for ALUA state

In reply to Hannes Reinecke

scsi_proto.h now contains definitions for the ALUA state, so we don't have to carry them in the device handler.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 62 +++++++++++++-----------------
 1 file changed, 27 insertions(+), 35 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +27 -35
index ffc419e..e283c43 100644
@@ -31,14 +31,6 @@
 #define ALUA_DH_NAME "alua"
 #define ALUA_DH_VER "1.3"
 
-#define TPGS_STATE_OPTIMIZED		0x0
-#define TPGS_STATE_NONOPTIMIZED		0x1
-#define TPGS_STATE_STANDBY		0x2
-#define TPGS_STATE_UNAVAILABLE		0x3
-#define TPGS_STATE_LBA_DEPENDENT	0x4
-#define TPGS_STATE_OFFLINE		0xe
-#define TPGS_STATE_TRANSITIONING	0xf
-
 #define TPGS_SUPPORT_NONE		0x00
 #define TPGS_SUPPORT_OPTIMIZED		0x01
 #define TPGS_SUPPORT_NONOPTIMIZED	0x02
@@ -178,7 +170,7 @@ static int submit_stpg(struct scsi_device *sdev, int group_id,
 
 	/* Prepare the data buffer */
 	memset(stpg_data, 0, stpg_len);
-	stpg_data[4] = TPGS_STATE_OPTIMIZED & 0x0f;
+	stpg_data[4] = SCSI_ACCESS_STATE_OPTIMAL & 0x0f;
 	put_unaligned_be16(group_id, &stpg_data[6]);
 
 	/* Prepare the command. */
@@ -246,7 +238,7 @@ struct alua_port_group *alua_alloc_pg(struct scsi_device *sdev,
 	}
 	pg->group_id = group_id;
 	pg->tpgs = tpgs;
-	pg->state = TPGS_STATE_OPTIMIZED;
+	pg->state = SCSI_ACCESS_STATE_OPTIMAL;
 	if (optimize_stpg)
 		pg->flags |= ALUA_OPTIMIZE_STPG;
 	kref_init(&pg->kref);
@@ -386,22 +378,22 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h,
 	return SCSI_DH_OK;
 }
 
-static char print_alua_state(int state)
+static char print_alua_state(enum scsi_access_state state)
 {
 	switch (state) {
-	case TPGS_STATE_OPTIMIZED:
+	case SCSI_ACCESS_STATE_OPTIMAL:
 		return 'A';
-	case TPGS_STATE_NONOPTIMIZED:
+	case SCSI_ACCESS_STATE_ACTIVE:
 		return 'N';
-	case TPGS_STATE_STANDBY:
+	case SCSI_ACCESS_STATE_STANDBY:
 		return 'S';
-	case TPGS_STATE_UNAVAILABLE:
+	case SCSI_ACCESS_STATE_UNAVAILABLE:
 		return 'U';
-	case TPGS_STATE_LBA_DEPENDENT:
+	case SCSI_ACCESS_STATE_LBA:
 		return 'L';
-	case TPGS_STATE_OFFLINE:
+	case SCSI_ACCESS_STATE_OFFLINE:
 		return 'O';
-	case TPGS_STATE_TRANSITIONING:
+	case SCSI_ACCESS_STATE_TRANSITIONING:
 		return 'T';
 	default:
 		return 'X';
@@ -652,7 +644,7 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 		    valid_states&TPGS_SUPPORT_OPTIMIZED?'A':'a');
 
 	switch (pg->state) {
-	case TPGS_STATE_TRANSITIONING:
+	case SCSI_ACCESS_STATE_TRANSITIONING:
 		if (time_before(jiffies, pg->expiry)) {
 			/* State transition, retry */
 			pg->interval = 2;
@@ -660,11 +652,11 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 		} else {
 			/* Transitioning time exceeded, set port to standby */
 			err = SCSI_DH_IO;
-			pg->state = TPGS_STATE_STANDBY;
+			pg->state = SCSI_ACCESS_STATE_STANDBY;
 			pg->expiry = 0;
 		}
 		break;
-	case TPGS_STATE_OFFLINE:
+	case SCSI_ACCESS_STATE_OFFLINE:
 		/* Path unusable */
 		err = SCSI_DH_DEV_OFFLINED;
 		pg->expiry = 0;
@@ -697,21 +689,21 @@ static unsigned alua_stpg(struct scsi_device *sdev, struct alua_port_group *pg)
 		return SCSI_DH_RETRY;
 	}
 	switch (pg->state) {
-	case TPGS_STATE_OPTIMIZED:
+	case SCSI_ACCESS_STATE_OPTIMAL:
 		return SCSI_DH_OK;
-	case TPGS_STATE_NONOPTIMIZED:
+	case SCSI_ACCESS_STATE_ACTIVE:
 		if ((pg->flags & ALUA_OPTIMIZE_STPG) &&
 		    !pg->pref &&
 		    (pg->tpgs & TPGS_MODE_IMPLICIT))
 			return SCSI_DH_OK;
 		break;
-	case TPGS_STATE_STANDBY:
-	case TPGS_STATE_UNAVAILABLE:
+	case SCSI_ACCESS_STATE_STANDBY:
+	case SCSI_ACCESS_STATE_UNAVAILABLE:
 		break;
-	case TPGS_STATE_OFFLINE:
+	case SCSI_ACCESS_STATE_OFFLINE:
 		return SCSI_DH_IO;
 		break;
-	case TPGS_STATE_TRANSITIONING:
+	case SCSI_ACCESS_STATE_TRANSITIONING:
 		break;
 	default:
 		sdev_printk(KERN_INFO, sdev,
@@ -721,7 +713,7 @@ static unsigned alua_stpg(struct scsi_device *sdev, struct alua_port_group *pg)
 		break;
 	}
 	/* Set state to transitioning */
-	pg->state = TPGS_STATE_TRANSITIONING;
+	pg->state = SCSI_ACCESS_STATE_TRANSITIONING;
 	retval = submit_stpg(sdev, pg->group_id, &sense_hdr);
 
 	if (retval) {
@@ -767,7 +759,7 @@ static void alua_rtpg_work(struct work_struct *work)
 		int state = pg->state;
 
 		spin_unlock_irqrestore(&pg->lock, flags);
-		if (state == TPGS_STATE_TRANSITIONING) {
+		if (state == SCSI_ACCESS_STATE_TRANSITIONING) {
 			if (alua_tur(sdev) == SCSI_DH_RETRY) {
 				spin_lock_irqsave(&pg->lock, flags);
 				pg->flags &= ~ALUA_PG_RUNNING;
@@ -1014,7 +1006,7 @@ static int alua_prep_fn(struct scsi_device *sdev, struct request *req)
 {
 	struct alua_dh_data *h = sdev->handler_data;
 	struct alua_port_group *pg;
-	int state = TPGS_STATE_OPTIMIZED;
+	enum scsi_access_state state = SCSI_ACCESS_STATE_OPTIMAL;
 	int ret = BLKPREP_OK;
 
 	rcu_read_lock();
@@ -1023,14 +1015,14 @@ static int alua_prep_fn(struct scsi_device *sdev, struct request *req)
 		state = pg->state;
 		/* Defer I/O while rtpg_work is active */
 		if (pg->rtpg_sdev)
-			state = TPGS_STATE_TRANSITIONING;
+			state = SCSI_ACCESS_STATE_TRANSITIONING;
 	}
 	rcu_read_unlock();
-	if (state == TPGS_STATE_TRANSITIONING)
+	if (state == SCSI_ACCESS_STATE_TRANSITIONING)
 		ret = BLKPREP_DEFER;
-	else if (state != TPGS_STATE_OPTIMIZED &&
-		 state != TPGS_STATE_NONOPTIMIZED &&
-		 state != TPGS_STATE_LBA_DEPENDENT) {
+	else if (state != SCSI_ACCESS_STATE_OPTIMAL &&
+		 state != SCSI_ACCESS_STATE_ACTIVE &&
+		 state != SCSI_ACCESS_STATE_LBA) {
 		ret = BLKPREP_KILL;
 		req->cmd_flags |= REQ_QUIET;
 	}
Signature · 1 line
1.8.5.6
#35 lore raw

PATCH v2 21/22 scsi_dh_alua: update 'access_state' field

In reply to Hannes Reinecke

Track attached SCSI devices and update the 'access_state' field whenever an ALUA state change has been detected.

---
 drivers/scsi/device_handler/scsi_dh_alua.c | 29 +++++++++++++++++++++++++++--
 1 file changed, 27 insertions(+), 2 deletions(-)
drivers/scsi/device_handler/scsi_dh_alua.c +27 -2
index e283c43..93dc425 100644
@@ -24,6 +24,7 @@
 #include <linux/module.h>
 #include <asm/unaligned.h>
 #include <scsi/scsi.h>
+#include <scsi/scsi_proto.h>
 #include <scsi/scsi_dbg.h>
 #include <scsi/scsi_eh.h>
 #include <scsi/scsi_dh.h>
@@ -74,6 +75,7 @@ static struct workqueue_struct *kaluad_sync_wq;
 struct alua_port_group {
 	struct kref		kref;
 	struct list_head	node;
+	struct list_head	dh_list;
 	unsigned char		device_id_str[256];
 	int			device_id_len;
 	int			group_id;
@@ -91,6 +93,7 @@ struct alua_port_group {
 };
 
 struct alua_dh_data {
+	struct list_head	node;
 	struct alua_port_group	*pg;
 	int			group_id;
 	spinlock_t		pg_lock;
@@ -245,6 +248,7 @@ struct alua_port_group *alua_alloc_pg(struct scsi_device *sdev,
 	INIT_DELAYED_WORK(&pg->rtpg_work, alua_rtpg_work);
 	INIT_LIST_HEAD(&pg->rtpg_list);
 	INIT_LIST_HEAD(&pg->node);
+	INIT_LIST_HEAD(&pg->dh_list);
 	spin_lock_init(&pg->lock);
 
 	spin_lock(&port_group_lock);
@@ -326,6 +330,7 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h,
 	int rel_port = -1, group_id;
 	struct alua_port_group *pg, *old_pg = NULL;
 	bool pg_updated = false;
+	unsigned long flags;
 
 	group_id = scsi_vpd_tpg_id(sdev, &rel_port);
 	if (group_id < 0) {
@@ -357,6 +362,9 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h,
 		/* port_group has changed. Update to new port group */
 		if (h->pg != pg) {
 			old_pg = h->pg;
+			spin_lock_irqsave(&old_pg->lock, flags);
+			list_del_rcu(&h->node);
+			spin_unlock_irqrestore(&old_pg->lock, flags);
 			rcu_assign_pointer(h->pg, pg);
 			pg_updated = true;
 		}
@@ -367,8 +375,12 @@ static int alua_check_vpd(struct scsi_device *sdev, struct alua_dh_data *h,
 	alua_rtpg_queue(h->pg, sdev, NULL, true);
 	spin_unlock(&h->pg_lock);
 
-	if (pg_updated)
+	if (pg_updated) {
+		spin_lock_irqsave(&pg->lock, flags);
+		list_add_rcu(&h->node, &pg->dh_list);
+		spin_unlock_irqrestore(&pg->lock, flags);
 		synchronize_rcu();
+	}
 	if (old_pg) {
 		if (old_pg->rtpg_sdev)
 			flush_delayed_work(&old_pg->rtpg_work);
@@ -615,6 +627,8 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 
 		spin_lock_irqsave(&port_group_lock, flags);
 		list_for_each_entry(tmp_pg, &port_group_list, node) {
+			struct alua_dh_data *h;
+
 			if (tmp_pg->group_id != group_id)
 				continue;
 			if (tmp_pg->device_id_len != pg->device_id_len)
@@ -624,6 +638,13 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg)
 				continue;
 			tmp_pg->state = desc[0] & 0x0f;
 			tmp_pg->pref = desc[0] >> 7;
+			rcu_read_lock();
+			list_for_each_entry_rcu(h, &tmp_pg->dh_list, node) {
+				/* h->sdev should always be valid */
+				BUG_ON(!h->sdev);
+				h->sdev->access_state = desc[0];
+			}
+			rcu_read_unlock();
 			if (tmp_pg == pg)
 				valid_states = desc[1];
 		}
@@ -1053,6 +1074,7 @@ static int alua_bus_attach(struct scsi_device *sdev)
 	rcu_assign_pointer(h->pg, NULL);
 	h->init_error = SCSI_DH_OK;
 	h->sdev = sdev;
+	INIT_LIST_HEAD(&h->node);
 
 	mutex_init(&h->init_mutex);
 	err = alua_initialize(sdev, h);
@@ -1080,14 +1102,17 @@ static void alua_bus_detach(struct scsi_device *sdev)
 	spin_lock(&h->pg_lock);
 	pg = h->pg;
 	rcu_assign_pointer(h->pg, NULL);
-	h->sdev = NULL;
 	spin_unlock(&h->pg_lock);
 	if (pg) {
+		spin_lock(&pg->lock);
+		list_del_rcu(&h->node);
+		spin_unlock(&pg->lock);
 		synchronize_rcu();
 		if (pg->rtpg_sdev)
 			flush_delayed_work(&pg->rtpg_work);
 		kref_put(&pg->kref, release_port_group);
 	}
+	h->sdev = NULL;
 	sdev->handler_data = NULL;
 	kfree(h);
 }
Signature · 1 line
1.8.5.6
#36 lore raw

PATCH v2 22/22 scsi_dh_alua: Update version to 2.0

In reply to Hannes Reinecke
---
 drivers/scsi/device_handler/scsi_dh_alua.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/device_handler/scsi_dh_alua.c +1 -1
index 93dc425..47ca96c 100644
@@ -30,7 +30,7 @@
 #include <scsi/scsi_dh.h>
 
 #define ALUA_DH_NAME "alua"
-#define ALUA_DH_VER "1.3"
+#define ALUA_DH_VER "2.0"
 
 #define TPGS_SUPPORT_NONE		0x00
 #define TPGS_SUPPORT_OPTIMIZED		0x01
Signature · 1 line
1.8.5.6