* [PATCH v1 0/3] Fixes for problems found during testing
@ 2016-10-11 22:58 Michael Cyr
2016-10-11 22:58 ` [PATCH v1 1/3] Synchronization of cmds during termination conditions Michael Cyr
` (3 more replies)
0 siblings, 4 replies; 15+ messages in thread
From: Michael Cyr @ 2016-10-11 22:58 UTC (permalink / raw)
To: nab, joe
Cc: hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen, Michael Cyr
The first one seems bigger than it is, because I needed to rearrange the
order of some functions.
Michael Cyr (3):
Synchronization of cmds during termination conditions
Properly handle if target_submit_cmd/tmr fails
Return correct partition name/number to client
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 1076 ++++++++++++++----------------
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h | 5 +-
2 files changed, 488 insertions(+), 593 deletions(-)
--
2.5.0
^ permalink raw reply [flat|nested] 15+ messages in thread
* [PATCH v1 1/3] Synchronization of cmds during termination conditions
2016-10-11 22:58 [PATCH v1 0/3] Fixes for problems found during testing Michael Cyr
@ 2016-10-11 22:58 ` Michael Cyr
2016-10-12 6:20 ` Greg KH
2016-10-12 14:09 ` Bryant G. Ly
2016-10-11 22:58 ` [PATCH v1 2/3] Properly handle if target_submit_cmd/tmr fails Michael Cyr
` (2 subsequent siblings)
3 siblings, 2 replies; 15+ messages in thread
From: Michael Cyr @ 2016-10-11 22:58 UTC (permalink / raw)
To: nab, joe
Cc: hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen, Michael Cyr
Signed-off-by: Michael Cyr <mikecyr@us.ibm.com>
---
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 1082 +++++++++++++-----------------
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h | 5 +-
2 files changed, 486 insertions(+), 601 deletions(-)
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
index 642b739..34f758c 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
@@ -22,7 +22,7 @@
*
****************************************************************************/
-#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/module.h>
#include <linux/kernel.h>
@@ -81,7 +81,7 @@ static void ibmvscsis_determine_resid(struct se_cmd *se_cmd,
}
} else if (se_cmd->se_cmd_flags & SCF_OVERFLOW_BIT) {
if (se_cmd->data_direction == DMA_TO_DEVICE) {
- /* residual data from an overflow write */
+ /* residual data from an overflow write */
rsp->flags = SRP_RSP_FLAG_DOOVER;
rsp->data_out_res_cnt = cpu_to_be32(residual_count);
} else if (se_cmd->data_direction == DMA_FROM_DEVICE) {
@@ -101,7 +101,7 @@ static void ibmvscsis_determine_resid(struct se_cmd *se_cmd,
* and the function returns TRUE.
*
* EXECUTION ENVIRONMENT:
- * Interrupt or Process environment
+ * Interrupt or Process environment
*/
static bool connection_broken(struct scsi_info *vscsi)
{
@@ -324,7 +324,7 @@ static struct viosrp_crq *ibmvscsis_cmd_q_dequeue(uint mask,
}
/**
- * ibmvscsis_send_init_message() - send initialize message to the client
+ * ibmvscsis_send_init_message() - send initialize message to the client
* @vscsi: Pointer to our adapter structure
* @format: Which Init Message format to send
*
@@ -382,13 +382,13 @@ static long ibmvscsis_check_init_msg(struct scsi_info *vscsi, uint *format)
vscsi->cmd_q.base_addr);
if (crq) {
*format = (uint)(crq->format);
- rc = ERROR;
+ rc = ERROR;
crq->valid = INVALIDATE_CMD_RESP_EL;
dma_rmb();
}
} else {
*format = (uint)(crq->format);
- rc = ERROR;
+ rc = ERROR;
crq->valid = INVALIDATE_CMD_RESP_EL;
dma_rmb();
}
@@ -397,166 +397,6 @@ static long ibmvscsis_check_init_msg(struct scsi_info *vscsi, uint *format)
}
/**
- * ibmvscsis_establish_new_q() - Establish new CRQ queue
- * @vscsi: Pointer to our adapter structure
- * @new_state: New state being established after resetting the queue
- *
- * Must be called with interrupt lock held.
- */
-static long ibmvscsis_establish_new_q(struct scsi_info *vscsi, uint new_state)
-{
- long rc = ADAPT_SUCCESS;
- uint format;
-
- vscsi->flags &= PRESERVE_FLAG_FIELDS;
- vscsi->rsp_q_timer.timer_pops = 0;
- vscsi->debit = 0;
- vscsi->credit = 0;
-
- rc = vio_enable_interrupts(vscsi->dma_dev);
- if (rc) {
- pr_warn("reset_queue: failed to enable interrupts, rc %ld\n",
- rc);
- return rc;
- }
-
- rc = ibmvscsis_check_init_msg(vscsi, &format);
- if (rc) {
- dev_err(&vscsi->dev, "reset_queue: check_init_msg failed, rc %ld\n",
- rc);
- return rc;
- }
-
- if (format == UNUSED_FORMAT && new_state == WAIT_CONNECTION) {
- rc = ibmvscsis_send_init_message(vscsi, INIT_MSG);
- switch (rc) {
- case H_SUCCESS:
- case H_DROPPED:
- case H_CLOSED:
- rc = ADAPT_SUCCESS;
- break;
-
- case H_PARAMETER:
- case H_HARDWARE:
- break;
-
- default:
- vscsi->state = UNDEFINED;
- rc = H_HARDWARE;
- break;
- }
- }
-
- return rc;
-}
-
-/**
- * ibmvscsis_reset_queue() - Reset CRQ Queue
- * @vscsi: Pointer to our adapter structure
- * @new_state: New state to establish after resetting the queue
- *
- * This function calls h_free_q and then calls h_reg_q and does all
- * of the bookkeeping to get us back to where we can communicate.
- *
- * Actually, we don't always call h_free_crq. A problem was discovered
- * where one partition would close and reopen his queue, which would
- * cause his partner to get a transport event, which would cause him to
- * close and reopen his queue, which would cause the original partition
- * to get a transport event, etc., etc. To prevent this, we don't
- * actually close our queue if the client initiated the reset, (i.e.
- * either we got a transport event or we have detected that the client's
- * queue is gone)
- *
- * EXECUTION ENVIRONMENT:
- * Process environment, called with interrupt lock held
- */
-static void ibmvscsis_reset_queue(struct scsi_info *vscsi, uint new_state)
-{
- int bytes;
- long rc = ADAPT_SUCCESS;
-
- pr_debug("reset_queue: flags 0x%x\n", vscsi->flags);
-
- /* don't reset, the client did it for us */
- if (vscsi->flags & (CLIENT_FAILED | TRANS_EVENT)) {
- vscsi->flags &= PRESERVE_FLAG_FIELDS;
- vscsi->rsp_q_timer.timer_pops = 0;
- vscsi->debit = 0;
- vscsi->credit = 0;
- vscsi->state = new_state;
- vio_enable_interrupts(vscsi->dma_dev);
- } else {
- rc = ibmvscsis_free_command_q(vscsi);
- if (rc == ADAPT_SUCCESS) {
- vscsi->state = new_state;
-
- bytes = vscsi->cmd_q.size * PAGE_SIZE;
- rc = h_reg_crq(vscsi->dds.unit_id,
- vscsi->cmd_q.crq_token, bytes);
- if (rc == H_CLOSED || rc == H_SUCCESS) {
- rc = ibmvscsis_establish_new_q(vscsi,
- new_state);
- }
-
- if (rc != ADAPT_SUCCESS) {
- pr_debug("reset_queue: reg_crq rc %ld\n", rc);
-
- vscsi->state = ERR_DISCONNECTED;
- vscsi->flags |= RESPONSE_Q_DOWN;
- ibmvscsis_free_command_q(vscsi);
- }
- } else {
- vscsi->state = ERR_DISCONNECTED;
- vscsi->flags |= RESPONSE_Q_DOWN;
- }
- }
-}
-
-/**
- * ibmvscsis_free_cmd_resources() - Free command resources
- * @vscsi: Pointer to our adapter structure
- * @cmd: Command which is not longer in use
- *
- * Must be called with interrupt lock held.
- */
-static void ibmvscsis_free_cmd_resources(struct scsi_info *vscsi,
- struct ibmvscsis_cmd *cmd)
-{
- struct iu_entry *iue = cmd->iue;
-
- switch (cmd->type) {
- case TASK_MANAGEMENT:
- case SCSI_CDB:
- /*
- * When the queue goes down this value is cleared, so it
- * cannot be cleared in this general purpose function.
- */
- if (vscsi->debit)
- vscsi->debit -= 1;
- break;
- case ADAPTER_MAD:
- vscsi->flags &= ~PROCESSING_MAD;
- break;
- case UNSET_TYPE:
- break;
- default:
- dev_err(&vscsi->dev, "free_cmd_resources unknown type %d\n",
- cmd->type);
- break;
- }
-
- cmd->iue = NULL;
- list_add_tail(&cmd->list, &vscsi->free_cmd);
- srp_iu_put(iue);
-
- if (list_empty(&vscsi->active_q) && list_empty(&vscsi->schedule_q) &&
- list_empty(&vscsi->waiting_rsp) && (vscsi->flags & WAIT_FOR_IDLE)) {
- vscsi->flags &= ~WAIT_FOR_IDLE;
- complete(&vscsi->wait_idle);
- }
-}
-
-/**
* ibmvscsis_disconnect() - Helper function to disconnect
* @work: Pointer to work_struct, gives access to our adapter structure
*
@@ -575,7 +415,6 @@ static void ibmvscsis_disconnect(struct work_struct *work)
proc_work);
u16 new_state;
bool wait_idle = false;
- long rc = ADAPT_SUCCESS;
spin_lock_bh(&vscsi->intr_lock);
new_state = vscsi->new_state;
@@ -589,7 +428,7 @@ static void ibmvscsis_disconnect(struct work_struct *work)
* should transitition to the new state
*/
switch (vscsi->state) {
- /* Should never be called while in this state. */
+ /* Should never be called while in this state. */
case NO_QUEUE:
/*
* Can never transition from this state;
@@ -628,30 +467,24 @@ static void ibmvscsis_disconnect(struct work_struct *work)
vscsi->state = new_state;
break;
- /*
- * If this is a transition into an error state.
- * a client is attempting to establish a connection
- * and has violated the RPA protocol.
- * There can be nothing pending on the adapter although
- * there can be requests in the command queue.
- */
case WAIT_ENABLED:
- case PART_UP_WAIT_ENAB:
switch (new_state) {
- case ERR_DISCONNECT:
- vscsi->flags |= RESPONSE_Q_DOWN;
+ case UNCONFIGURING:
vscsi->state = new_state;
+ vscsi->flags |= RESPONSE_Q_DOWN;
vscsi->flags &= ~(SCHEDULE_DISCONNECT |
DISCONNECT_SCHEDULED);
- ibmvscsis_free_command_q(vscsi);
- break;
- case ERR_DISCONNECT_RECONNECT:
- ibmvscsis_reset_queue(vscsi, WAIT_ENABLED);
+ dma_rmb();
+ if (vscsi->flags & CFG_SLEEPING) {
+ vscsi->flags &= ~CFG_SLEEPING;
+ complete(&vscsi->unconfig);
+ }
break;
/* should never happen */
+ case ERR_DISCONNECT:
+ case ERR_DISCONNECT_RECONNECT:
case WAIT_IDLE:
- rc = ERROR;
dev_err(&vscsi->dev, "disconnect: invalid state %d for WAIT_IDLE\n",
vscsi->state);
break;
@@ -660,6 +493,13 @@ static void ibmvscsis_disconnect(struct work_struct *work)
case WAIT_IDLE:
switch (new_state) {
+ case UNCONFIGURING:
+ vscsi->flags |= RESPONSE_Q_DOWN;
+ vscsi->state = new_state;
+ vscsi->flags &= ~(SCHEDULE_DISCONNECT |
+ DISCONNECT_SCHEDULED);
+ ibmvscsis_free_command_q(vscsi);
+ break;
case ERR_DISCONNECT:
case ERR_DISCONNECT_RECONNECT:
vscsi->state = new_state;
@@ -764,45 +604,349 @@ static void ibmvscsis_post_disconnect(struct scsi_info *vscsi, uint new_state,
else
state = vscsi->state;
- switch (state) {
- case NO_QUEUE:
- case UNCONFIGURING:
- break;
+ switch (state) {
+ case NO_QUEUE:
+ case UNCONFIGURING:
+ break;
+
+ case ERR_DISCONNECTED:
+ case ERR_DISCONNECT:
+ case UNDEFINED:
+ if (new_state == UNCONFIGURING)
+ vscsi->new_state = new_state;
+ break;
+
+ case ERR_DISCONNECT_RECONNECT:
+ switch (new_state) {
+ case UNCONFIGURING:
+ case ERR_DISCONNECT:
+ vscsi->new_state = new_state;
+ break;
+ default:
+ break;
+ }
+ break;
+
+ case WAIT_ENABLED:
+ case WAIT_IDLE:
+ case WAIT_CONNECTION:
+ case CONNECTED:
+ case SRP_PROCESSING:
+ vscsi->new_state = new_state;
+ break;
+
+ default:
+ break;
+ }
+ }
+
+ pr_debug("Leaving post_disconnect: flags 0x%x, new_state 0x%x\n",
+ vscsi->flags, vscsi->new_state);
+}
+
+/**
+ * ibmvscsis_handle_init_compl_msg() - Respond to an Init Complete Message
+ * @vscsi: Pointer to our adapter structure
+ *
+ * Must be called with interrupt lock held.
+ */
+static long ibmvscsis_handle_init_compl_msg(struct scsi_info *vscsi)
+{
+ long rc = ADAPT_SUCCESS;
+
+ switch (vscsi->state) {
+ case NO_QUEUE:
+ case ERR_DISCONNECT:
+ case ERR_DISCONNECT_RECONNECT:
+ case ERR_DISCONNECTED:
+ case UNCONFIGURING:
+ case UNDEFINED:
+ rc = ERROR;
+ break;
+
+ case WAIT_CONNECTION:
+ vscsi->state = CONNECTED;
+ break;
+
+ case WAIT_IDLE:
+ case SRP_PROCESSING:
+ case CONNECTED:
+ case WAIT_ENABLED:
+ default:
+ rc = ERROR;
+ dev_err(&vscsi->dev, "init_msg: invalid state %d to get init compl msg\n",
+ vscsi->state);
+ ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
+ break;
+ }
+
+ return rc;
+}
+
+/**
+ * ibmvscsis_handle_init_msg() - Respond to an Init Message
+ * @vscsi: Pointer to our adapter structure
+ *
+ * Must be called with interrupt lock held.
+ */
+static long ibmvscsis_handle_init_msg(struct scsi_info *vscsi)
+{
+ long rc = ADAPT_SUCCESS;
+
+ switch (vscsi->state) {
+ case WAIT_CONNECTION:
+ rc = ibmvscsis_send_init_message(vscsi, INIT_COMPLETE_MSG);
+ switch (rc) {
+ case H_SUCCESS:
+ vscsi->state = CONNECTED;
+ break;
+
+ case H_PARAMETER:
+ dev_err(&vscsi->dev, "init_msg: failed to send, rc %ld\n",
+ rc);
+ ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT, 0);
+ break;
+
+ case H_DROPPED:
+ dev_err(&vscsi->dev, "init_msg: failed to send, rc %ld\n",
+ rc);
+ rc = ERROR;
+ ibmvscsis_post_disconnect(vscsi,
+ ERR_DISCONNECT_RECONNECT, 0);
+ break;
+
+ case H_CLOSED:
+ pr_warn("init_msg: failed to send, rc %ld\n", rc);
+ rc = 0;
+ break;
+ }
+ break;
+
+ case UNDEFINED:
+ rc = ERROR;
+ break;
+
+ case UNCONFIGURING:
+ break;
+
+ case WAIT_ENABLED:
+ case CONNECTED:
+ case SRP_PROCESSING:
+ case WAIT_IDLE:
+ case NO_QUEUE:
+ case ERR_DISCONNECT:
+ case ERR_DISCONNECT_RECONNECT:
+ case ERR_DISCONNECTED:
+ default:
+ rc = ERROR;
+ dev_err(&vscsi->dev, "init_msg: invalid state %d to get init msg\n",
+ vscsi->state);
+ ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
+ break;
+ }
+
+ return rc;
+}
+
+/**
+ * ibmvscsis_init_msg() - Respond to an init message
+ * @vscsi: Pointer to our adapter structure
+ * @crq: Pointer to CRQ element containing the Init Message
+ *
+ * EXECUTION ENVIRONMENT:
+ * Interrupt, interrupt lock held
+ */
+static long ibmvscsis_init_msg(struct scsi_info *vscsi, struct viosrp_crq *crq)
+{
+ long rc = ADAPT_SUCCESS;
+
+ pr_debug("init_msg: state 0x%hx\n", vscsi->state);
+
+ rc = h_vioctl(vscsi->dds.unit_id, H_GET_PARTNER_INFO,
+ (u64)vscsi->map_ioba | ((u64)PAGE_SIZE << 32), 0, 0, 0,
+ 0);
+ if (rc == H_SUCCESS) {
+ vscsi->client_data.partition_number =
+ be64_to_cpu(*(u64 *)vscsi->map_buf);
+ pr_debug("init_msg, part num %d\n",
+ vscsi->client_data.partition_number);
+ } else {
+ pr_debug("init_msg h_vioctl rc %ld\n", rc);
+ rc = ADAPT_SUCCESS;
+ }
+
+ if (crq->format == INIT_MSG) {
+ rc = ibmvscsis_handle_init_msg(vscsi);
+ } else if (crq->format == INIT_COMPLETE_MSG) {
+ rc = ibmvscsis_handle_init_compl_msg(vscsi);
+ } else {
+ rc = ERROR;
+ dev_err(&vscsi->dev, "init_msg: invalid format %d\n",
+ (uint)crq->format);
+ ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
+ }
+
+ return rc;
+}
+
+/**
+ * ibmvscsis_establish_new_q() - Establish new CRQ queue
+ * @vscsi: Pointer to our adapter structure
+ *
+ * Must be called with interrupt lock held.
+ */
+static long ibmvscsis_establish_new_q(struct scsi_info *vscsi)
+{
+ long rc = ADAPT_SUCCESS;
+ uint format;
+
+ vscsi->flags &= PRESERVE_FLAG_FIELDS;
+ vscsi->rsp_q_timer.timer_pops = 0;
+ vscsi->debit = 0;
+ vscsi->credit = 0;
+
+ rc = vio_enable_interrupts(vscsi->dma_dev);
+ if (rc) {
+ pr_warn("establish_new_q: failed to enable interrupts, rc %ld\n",
+ rc);
+ return rc;
+ }
+
+ rc = ibmvscsis_check_init_msg(vscsi, &format);
+ if (rc) {
+ dev_err(&vscsi->dev, "establish_new_q: check_init_msg failed, rc %ld\n",
+ rc);
+ return rc;
+ }
+
+ if (format == UNUSED_FORMAT) {
+ rc = ibmvscsis_send_init_message(vscsi, INIT_MSG);
+ switch (rc) {
+ case H_SUCCESS:
+ case H_DROPPED:
+ case H_CLOSED:
+ rc = ADAPT_SUCCESS;
+ break;
+
+ case H_PARAMETER:
+ case H_HARDWARE:
+ break;
+
+ default:
+ vscsi->state = UNDEFINED;
+ rc = H_HARDWARE;
+ break;
+ }
+ }
+ else if (format == INIT_MSG) {
+ rc = ibmvscsis_handle_init_msg(vscsi);
+ }
+
+ return rc;
+}
+
+/**
+ * ibmvscsis_reset_queue() - Reset CRQ Queue
+ * @vscsi: Pointer to our adapter structure
+ *
+ * This function calls h_free_q and then calls h_reg_q and does all
+ * of the bookkeeping to get us back to where we can communicate.
+ *
+ * Actually, we don't always call h_free_crq. A problem was discovered
+ * where one partition would close and reopen his queue, which would
+ * cause his partner to get a transport event, which would cause him to
+ * close and reopen his queue, which would cause the original partition
+ * to get a transport event, etc., etc. To prevent this, we don't
+ * actually close our queue if the client initiated the reset, (i.e.
+ * either we got a transport event or we have detected that the client's
+ * queue is gone)
+ *
+ * EXECUTION ENVIRONMENT:
+ * Process environment, called with interrupt lock held
+ */
+static void ibmvscsis_reset_queue(struct scsi_info *vscsi)
+{
+ int bytes;
+ long rc = ADAPT_SUCCESS;
+
+ pr_debug("reset_queue: flags 0x%x\n", vscsi->flags);
+
+ /* don't reset, the client did it for us */
+ if (vscsi->flags & (CLIENT_FAILED | TRANS_EVENT)) {
+ vscsi->flags &= PRESERVE_FLAG_FIELDS;
+ vscsi->rsp_q_timer.timer_pops = 0;
+ vscsi->debit = 0;
+ vscsi->credit = 0;
+ vscsi->state = WAIT_CONNECTION;
+ vio_enable_interrupts(vscsi->dma_dev);
+ } else {
+ rc = ibmvscsis_free_command_q(vscsi);
+ if (rc == ADAPT_SUCCESS) {
+ vscsi->state = WAIT_CONNECTION;
+
+ bytes = vscsi->cmd_q.size * PAGE_SIZE;
+ rc = h_reg_crq(vscsi->dds.unit_id,
+ vscsi->cmd_q.crq_token, bytes);
+ if (rc == H_CLOSED || rc == H_SUCCESS) {
+ rc = ibmvscsis_establish_new_q(vscsi);
+ }
- case ERR_DISCONNECTED:
- case ERR_DISCONNECT:
- case UNDEFINED:
- if (new_state == UNCONFIGURING)
- vscsi->new_state = new_state;
- break;
+ if (rc != ADAPT_SUCCESS) {
+ pr_debug("reset_queue: reg_crq rc %ld\n", rc);
- case ERR_DISCONNECT_RECONNECT:
- switch (new_state) {
- case UNCONFIGURING:
- case ERR_DISCONNECT:
- vscsi->new_state = new_state;
- break;
- default:
- break;
+ vscsi->state = ERR_DISCONNECTED;
+ vscsi->flags |= RESPONSE_Q_DOWN;
+ ibmvscsis_free_command_q(vscsi);
}
- break;
+ } else {
+ vscsi->state = ERR_DISCONNECTED;
+ vscsi->flags |= RESPONSE_Q_DOWN;
+ }
+ }
+}
- case WAIT_ENABLED:
- case PART_UP_WAIT_ENAB:
- case WAIT_IDLE:
- case WAIT_CONNECTION:
- case CONNECTED:
- case SRP_PROCESSING:
- vscsi->new_state = new_state;
- break;
+/**
+ * ibmvscsis_free_cmd_resources() - Free command resources
+ * @vscsi: Pointer to our adapter structure
+ * @cmd: Command which is not longer in use
+ *
+ * Must be called with interrupt lock held.
+ */
+static void ibmvscsis_free_cmd_resources(struct scsi_info *vscsi,
+ struct ibmvscsis_cmd *cmd)
+{
+ struct iu_entry *iue = cmd->iue;
- default:
- break;
- }
+ switch (cmd->type) {
+ case TASK_MANAGEMENT:
+ case SCSI_CDB:
+ /*
+ * When the queue goes down this value is cleared, so it
+ * cannot be cleared in this general purpose function.
+ */
+ if (vscsi->debit)
+ vscsi->debit -= 1;
+ break;
+ case ADAPTER_MAD:
+ vscsi->flags &= ~PROCESSING_MAD;
+ break;
+ case UNSET_TYPE:
+ break;
+ default:
+ dev_err(&vscsi->dev, "free_cmd_resources unknown type %d\n",
+ cmd->type);
+ break;
}
- pr_debug("Leaving post_disconnect: flags 0x%x, new_state 0x%x\n",
- vscsi->flags, vscsi->new_state);
+ cmd->iue = NULL;
+ list_add_tail(&cmd->list, &vscsi->free_cmd);
+ srp_iu_put(iue);
+
+ if (list_empty(&vscsi->active_q) && list_empty(&vscsi->schedule_q) &&
+ list_empty(&vscsi->waiting_rsp) && (vscsi->flags & WAIT_FOR_IDLE)) {
+ vscsi->flags &= ~WAIT_FOR_IDLE;
+ complete(&vscsi->wait_idle);
+ }
}
/**
@@ -863,10 +1007,6 @@ static long ibmvscsis_trans_event(struct scsi_info *vscsi,
TRANS_EVENT));
break;
- case PART_UP_WAIT_ENAB:
- vscsi->state = WAIT_ENABLED;
- break;
-
case SRP_PROCESSING:
if ((vscsi->debit > 0) ||
!list_empty(&vscsi->schedule_q) ||
@@ -895,7 +1035,7 @@ static long ibmvscsis_trans_event(struct scsi_info *vscsi,
}
}
- rc = vscsi->flags & SCHEDULE_DISCONNECT;
+ rc = vscsi->flags & SCHEDULE_DISCONNECT;
pr_debug("Leaving trans_event: flags 0x%x, state 0x%hx, rc %ld\n",
vscsi->flags, vscsi->state, rc);
@@ -1066,16 +1206,28 @@ static void ibmvscsis_adapter_idle(struct scsi_info *vscsi)
free_qs = true;
switch (vscsi->state) {
+ case UNCONFIGURING:
+ ibmvscsis_free_command_q(vscsi);
+ dma_rmb();
+ isync();
+ if (vscsi->flags & CFG_SLEEPING) {
+ vscsi->flags &= ~CFG_SLEEPING;
+ complete(&vscsi->unconfig);
+ }
+ break;
case ERR_DISCONNECT_RECONNECT:
- ibmvscsis_reset_queue(vscsi, WAIT_CONNECTION);
+ ibmvscsis_reset_queue(vscsi);
pr_debug("adapter_idle, disc_rec: flags 0x%x\n", vscsi->flags);
break;
case ERR_DISCONNECT:
ibmvscsis_free_command_q(vscsi);
- vscsi->flags &= ~DISCONNECT_SCHEDULED;
+ vscsi->flags &= ~(SCHEDULE_DISCONNECT | DISCONNECT_SCHEDULED);
vscsi->flags |= RESPONSE_Q_DOWN;
- vscsi->state = ERR_DISCONNECTED;
+ if (vscsi->tport.enabled)
+ vscsi->state = ERR_DISCONNECTED;
+ else
+ vscsi->state = WAIT_ENABLED;
pr_debug("adapter_idle, disc: flags 0x%x, state 0x%hx\n",
vscsi->flags, vscsi->state);
break;
@@ -1220,7 +1372,7 @@ static long ibmvscsis_copy_crq_packet(struct scsi_info *vscsi,
* @iue: Information Unit containing the Adapter Info MAD request
*
* EXECUTION ENVIRONMENT:
- * Interrupt adpater lock is held
+ * Interrupt adapter lock is held
*/
static long ibmvscsis_adapter_info(struct scsi_info *vscsi,
struct iu_entry *iue)
@@ -1620,8 +1772,8 @@ static void ibmvscsis_send_messages(struct scsi_info *vscsi)
be64_to_cpu(msg_hi),
be64_to_cpu(cmd->rsp.tag));
- pr_debug("send_messages: tag 0x%llx, rc %ld\n",
- be64_to_cpu(cmd->rsp.tag), rc);
+ pr_debug("send_messages: cmd %p, tag 0x%llx, rc %ld\n",
+ cmd, be64_to_cpu(cmd->rsp.tag), rc);
/* if all ok free up the command element resources */
if (rc == H_SUCCESS) {
@@ -1691,7 +1843,7 @@ static void ibmvscsis_send_mad_resp(struct scsi_info *vscsi,
* @crq: Pointer to the CRQ entry containing the MAD request
*
* EXECUTION ENVIRONMENT:
- * Interrupt called with adapter lock held
+ * Interrupt, called with adapter lock held
*/
static long ibmvscsis_mad(struct scsi_info *vscsi, struct viosrp_crq *crq)
{
@@ -1864,7 +2016,7 @@ static long ibmvscsis_srp_login_rej(struct scsi_info *vscsi,
break;
case H_PERMISSION:
if (connection_broken(vscsi))
- flag_bits = RESPONSE_Q_DOWN | CLIENT_FAILED;
+ flag_bits = RESPONSE_Q_DOWN | CLIENT_FAILED;
dev_err(&vscsi->dev, "login_rej: error copying to client, rc %ld\n",
rc);
ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT,
@@ -2083,254 +2235,104 @@ static void ibmvscsis_srp_cmd(struct scsi_info *vscsi, struct viosrp_crq *crq)
}
if (vscsi->state == SRP_PROCESSING) {
- switch (srp->opcode) {
- case SRP_LOGIN_REQ:
- rc = ibmvscsis_srp_login(vscsi, cmd, crq);
- break;
-
- case SRP_TSK_MGMT:
- tsk = &vio_iu(iue)->srp.tsk_mgmt;
- pr_debug("tsk_mgmt tag: %llu (0x%llx)\n", tsk->tag,
- tsk->tag);
- cmd->rsp.tag = tsk->tag;
- vscsi->debit += 1;
- cmd->type = TASK_MANAGEMENT;
- list_add_tail(&cmd->list, &vscsi->schedule_q);
- queue_work(vscsi->work_q, &cmd->work);
- break;
-
- case SRP_CMD:
- pr_debug("srp_cmd tag: %llu (0x%llx)\n", srp->tag,
- srp->tag);
- cmd->rsp.tag = srp->tag;
- vscsi->debit += 1;
- cmd->type = SCSI_CDB;
- /*
- * We want to keep track of work waiting for
- * the workqueue.
- */
- list_add_tail(&cmd->list, &vscsi->schedule_q);
- queue_work(vscsi->work_q, &cmd->work);
- break;
-
- case SRP_I_LOGOUT:
- rc = ibmvscsis_srp_i_logout(vscsi, cmd, crq);
- break;
-
- case SRP_CRED_RSP:
- case SRP_AER_RSP:
- default:
- ibmvscsis_free_cmd_resources(vscsi, cmd);
- dev_err(&vscsi->dev, "invalid srp cmd, opcode %d\n",
- (uint)srp->opcode);
- ibmvscsis_post_disconnect(vscsi,
- ERR_DISCONNECT_RECONNECT, 0);
- break;
- }
- } else if (srp->opcode == SRP_LOGIN_REQ && vscsi->state == CONNECTED) {
- rc = ibmvscsis_srp_login(vscsi, cmd, crq);
- } else {
- ibmvscsis_free_cmd_resources(vscsi, cmd);
- dev_err(&vscsi->dev, "Invalid state %d to handle srp cmd\n",
- vscsi->state);
- ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
- }
-}
-
-/**
- * ibmvscsis_ping_response() - Respond to a ping request
- * @vscsi: Pointer to our adapter structure
- *
- * Let the client know that the server is alive and waiting on
- * its native I/O stack.
- * If any type of error occurs from the call to queue a ping
- * response then the client is either not accepting or receiving
- * interrupts. Disconnect with an error.
- *
- * EXECUTION ENVIRONMENT:
- * Interrupt, interrupt lock held
- */
-static long ibmvscsis_ping_response(struct scsi_info *vscsi)
-{
- struct viosrp_crq *crq;
- u64 buffer[2] = { 0, 0 };
- long rc;
-
- crq = (struct viosrp_crq *)&buffer;
- crq->valid = VALID_CMD_RESP_EL;
- crq->format = (u8)MESSAGE_IN_CRQ;
- crq->status = PING_RESPONSE;
-
- rc = h_send_crq(vscsi->dds.unit_id, cpu_to_be64(buffer[MSG_HI]),
- cpu_to_be64(buffer[MSG_LOW]));
-
- switch (rc) {
- case H_SUCCESS:
- break;
- case H_CLOSED:
- vscsi->flags |= CLIENT_FAILED;
- case H_DROPPED:
- vscsi->flags |= RESPONSE_Q_DOWN;
- case H_REMOTE_PARM:
- dev_err(&vscsi->dev, "ping_response: h_send_crq failed, rc %ld\n",
- rc);
- ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
- break;
- default:
- dev_err(&vscsi->dev, "ping_response: h_send_crq returned unknown rc %ld\n",
- rc);
- ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT, 0);
- break;
- }
-
- return rc;
-}
-
-/**
- * ibmvscsis_handle_init_compl_msg() - Respond to an Init Complete Message
- * @vscsi: Pointer to our adapter structure
- *
- * Must be called with interrupt lock held.
- */
-static long ibmvscsis_handle_init_compl_msg(struct scsi_info *vscsi)
-{
- long rc = ADAPT_SUCCESS;
-
- switch (vscsi->state) {
- case NO_QUEUE:
- case ERR_DISCONNECT:
- case ERR_DISCONNECT_RECONNECT:
- case ERR_DISCONNECTED:
- case UNCONFIGURING:
- case UNDEFINED:
- rc = ERROR;
- break;
-
- case WAIT_CONNECTION:
- vscsi->state = CONNECTED;
- break;
-
- case WAIT_IDLE:
- case SRP_PROCESSING:
- case CONNECTED:
- case WAIT_ENABLED:
- case PART_UP_WAIT_ENAB:
- default:
- rc = ERROR;
- dev_err(&vscsi->dev, "init_msg: invalid state %d to get init compl msg\n",
- vscsi->state);
- ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
- break;
- }
-
- return rc;
-}
-
-/**
- * ibmvscsis_handle_init_msg() - Respond to an Init Message
- * @vscsi: Pointer to our adapter structure
- *
- * Must be called with interrupt lock held.
- */
-static long ibmvscsis_handle_init_msg(struct scsi_info *vscsi)
-{
- long rc = ADAPT_SUCCESS;
+ switch (srp->opcode) {
+ case SRP_LOGIN_REQ:
+ rc = ibmvscsis_srp_login(vscsi, cmd, crq);
+ break;
- switch (vscsi->state) {
- case WAIT_ENABLED:
- vscsi->state = PART_UP_WAIT_ENAB;
- break;
+ case SRP_TSK_MGMT:
+ tsk = &vio_iu(iue)->srp.tsk_mgmt;
+ pr_debug("tsk_mgmt tag: %llu (0x%llx)\n", tsk->tag,
+ tsk->tag);
+ cmd->rsp.tag = tsk->tag;
+ vscsi->debit += 1;
+ cmd->type = TASK_MANAGEMENT;
+ list_add_tail(&cmd->list, &vscsi->schedule_q);
+ queue_work(vscsi->work_q, &cmd->work);
+ break;
- case WAIT_CONNECTION:
- rc = ibmvscsis_send_init_message(vscsi, INIT_COMPLETE_MSG);
- switch (rc) {
- case H_SUCCESS:
- vscsi->state = CONNECTED;
+ case SRP_CMD:
+ pr_debug("srp_cmd tag: %llu (0x%llx)\n", srp->tag,
+ srp->tag);
+ cmd->rsp.tag = srp->tag;
+ vscsi->debit += 1;
+ cmd->type = SCSI_CDB;
+ /*
+ * We want to keep track of work waiting for
+ * the workqueue.
+ */
+ list_add_tail(&cmd->list, &vscsi->schedule_q);
+ queue_work(vscsi->work_q, &cmd->work);
break;
- case H_PARAMETER:
- dev_err(&vscsi->dev, "init_msg: failed to send, rc %ld\n",
- rc);
- ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT, 0);
+ case SRP_I_LOGOUT:
+ rc = ibmvscsis_srp_i_logout(vscsi, cmd, crq);
break;
- case H_DROPPED:
- dev_err(&vscsi->dev, "init_msg: failed to send, rc %ld\n",
- rc);
- rc = ERROR;
+ case SRP_CRED_RSP:
+ case SRP_AER_RSP:
+ default:
+ ibmvscsis_free_cmd_resources(vscsi, cmd);
+ dev_err(&vscsi->dev, "invalid srp cmd, opcode %d\n",
+ (uint)srp->opcode);
ibmvscsis_post_disconnect(vscsi,
ERR_DISCONNECT_RECONNECT, 0);
break;
-
- case H_CLOSED:
- pr_warn("init_msg: failed to send, rc %ld\n", rc);
- rc = 0;
- break;
}
- break;
-
- case UNDEFINED:
- rc = ERROR;
- break;
-
- case UNCONFIGURING:
- break;
-
- case PART_UP_WAIT_ENAB:
- case CONNECTED:
- case SRP_PROCESSING:
- case WAIT_IDLE:
- case NO_QUEUE:
- case ERR_DISCONNECT:
- case ERR_DISCONNECT_RECONNECT:
- case ERR_DISCONNECTED:
- default:
- rc = ERROR;
- dev_err(&vscsi->dev, "init_msg: invalid state %d to get init msg\n",
+ } else if (srp->opcode == SRP_LOGIN_REQ && vscsi->state == CONNECTED) {
+ rc = ibmvscsis_srp_login(vscsi, cmd, crq);
+ } else {
+ ibmvscsis_free_cmd_resources(vscsi, cmd);
+ dev_err(&vscsi->dev, "Invalid state %d to handle srp cmd\n",
vscsi->state);
ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
- break;
}
-
- return rc;
}
/**
- * ibmvscsis_init_msg() - Respond to an init message
+ * ibmvscsis_ping_response() - Respond to a ping request
* @vscsi: Pointer to our adapter structure
- * @crq: Pointer to CRQ element containing the Init Message
+ *
+ * Let the client know that the server is alive and waiting on
+ * its native I/O stack.
+ * If any type of error occurs from the call to queue a ping
+ * response then the client is either not accepting or receiving
+ * interrupts. Disconnect with an error.
*
* EXECUTION ENVIRONMENT:
* Interrupt, interrupt lock held
*/
-static long ibmvscsis_init_msg(struct scsi_info *vscsi, struct viosrp_crq *crq)
+static long ibmvscsis_ping_response(struct scsi_info *vscsi)
{
- long rc = ADAPT_SUCCESS;
+ struct viosrp_crq *crq;
+ u64 buffer[2] = { 0, 0 };
+ long rc;
- pr_debug("init_msg: state 0x%hx\n", vscsi->state);
+ crq = (struct viosrp_crq *)&buffer;
+ crq->valid = VALID_CMD_RESP_EL;
+ crq->format = (u8)MESSAGE_IN_CRQ;
+ crq->status = PING_RESPONSE;
- rc = h_vioctl(vscsi->dds.unit_id, H_GET_PARTNER_INFO,
- (u64)vscsi->map_ioba | ((u64)PAGE_SIZE << 32), 0, 0, 0,
- 0);
- if (rc == H_SUCCESS) {
- vscsi->client_data.partition_number =
- be64_to_cpu(*(u64 *)vscsi->map_buf);
- pr_debug("init_msg, part num %d\n",
- vscsi->client_data.partition_number);
- } else {
- pr_debug("init_msg h_vioctl rc %ld\n", rc);
- rc = ADAPT_SUCCESS;
- }
+ rc = h_send_crq(vscsi->dds.unit_id, cpu_to_be64(buffer[MSG_HI]),
+ cpu_to_be64(buffer[MSG_LOW]));
- if (crq->format == INIT_MSG) {
- rc = ibmvscsis_handle_init_msg(vscsi);
- } else if (crq->format == INIT_COMPLETE_MSG) {
- rc = ibmvscsis_handle_init_compl_msg(vscsi);
- } else {
- rc = ERROR;
- dev_err(&vscsi->dev, "init_msg: invalid format %d\n",
- (uint)crq->format);
+ switch (rc) {
+ case H_SUCCESS:
+ break;
+ case H_CLOSED:
+ vscsi->flags |= CLIENT_FAILED;
+ case H_DROPPED:
+ vscsi->flags |= RESPONSE_Q_DOWN;
+ case H_REMOTE_PARM:
+ dev_err(&vscsi->dev, "ping_response: h_send_crq failed, rc %ld\n",
+ rc);
ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
+ break;
+ default:
+ dev_err(&vscsi->dev, "ping_response: h_send_crq returned unknown rc %ld\n",
+ rc);
+ ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT, 0);
+ break;
}
return rc;
@@ -2391,7 +2393,7 @@ static long ibmvscsis_parse_command(struct scsi_info *vscsi,
break;
case VALID_TRANS_EVENT:
- rc = ibmvscsis_trans_event(vscsi, crq);
+ rc = ibmvscsis_trans_event(vscsi, crq);
break;
case VALID_INIT_MSG:
@@ -2786,36 +2788,6 @@ static irqreturn_t ibmvscsis_interrupt(int dummy, void *data)
}
/**
- * ibmvscsis_check_q() - Helper function to Check Init Message Valid
- * @vscsi: Pointer to our adapter structure
- *
- * Checks if a initialize message was queued by the initiatior
- * while the timing window was open. This function is called from
- * probe after the CRQ is created and interrupts are enabled.
- * It would only be used by adapters who wait for some event before
- * completing the init handshake with the client. For ibmvscsi, this
- * event is waiting for the port to be enabled.
- *
- * EXECUTION ENVIRONMENT:
- * Process level only, interrupt lock held
- */
-static long ibmvscsis_check_q(struct scsi_info *vscsi)
-{
- uint format;
- long rc;
-
- rc = ibmvscsis_check_init_msg(vscsi, &format);
- if (rc)
- ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
- else if (format == UNUSED_FORMAT)
- vscsi->state = WAIT_ENABLED;
- else
- vscsi->state = PART_UP_WAIT_ENAB;
-
- return rc;
-}
-
-/**
* ibmvscsis_enable_change_state() - Set new state based on enabled status
* @vscsi: Pointer to our adapter structure
*
@@ -2826,77 +2798,19 @@ static long ibmvscsis_check_q(struct scsi_info *vscsi)
*/
static long ibmvscsis_enable_change_state(struct scsi_info *vscsi)
{
+ int bytes;
long rc = ADAPT_SUCCESS;
-handle_state_change:
- switch (vscsi->state) {
- case WAIT_ENABLED:
- rc = ibmvscsis_send_init_message(vscsi, INIT_MSG);
- switch (rc) {
- case H_SUCCESS:
- case H_DROPPED:
- case H_CLOSED:
- vscsi->state = WAIT_CONNECTION;
- rc = ADAPT_SUCCESS;
- break;
-
- case H_PARAMETER:
- break;
-
- case H_HARDWARE:
- break;
-
- default:
- vscsi->state = UNDEFINED;
- rc = H_HARDWARE;
- break;
- }
- break;
- case PART_UP_WAIT_ENAB:
- rc = ibmvscsis_send_init_message(vscsi, INIT_COMPLETE_MSG);
- switch (rc) {
- case H_SUCCESS:
- vscsi->state = CONNECTED;
- rc = ADAPT_SUCCESS;
- break;
-
- case H_DROPPED:
- case H_CLOSED:
- vscsi->state = WAIT_ENABLED;
- goto handle_state_change;
-
- case H_PARAMETER:
- break;
-
- case H_HARDWARE:
- break;
-
- default:
- rc = H_HARDWARE;
- break;
- }
- break;
-
- case WAIT_CONNECTION:
- case WAIT_IDLE:
- case SRP_PROCESSING:
- case CONNECTED:
- rc = ADAPT_SUCCESS;
- break;
- /* should not be able to get here */
- case UNCONFIGURING:
- rc = ERROR;
- vscsi->state = UNDEFINED;
- break;
+ bytes = vscsi->cmd_q.size * PAGE_SIZE;
+ rc = h_reg_crq(vscsi->dds.unit_id, vscsi->cmd_q.crq_token, bytes);
+ if (rc == H_CLOSED || rc == H_SUCCESS) {
+ vscsi->state = WAIT_CONNECTION;
+ rc = ibmvscsis_establish_new_q(vscsi);
+ }
- /* driver should never allow this to happen */
- case ERR_DISCONNECT:
- case ERR_DISCONNECT_RECONNECT:
- default:
- dev_err(&vscsi->dev, "in invalid state %d during enable_change_state\n",
- vscsi->state);
- rc = ADAPT_SUCCESS;
- break;
+ if (rc != ADAPT_SUCCESS) {
+ vscsi->state = ERR_DISCONNECTED;
+ vscsi->flags |= RESPONSE_Q_DOWN;
}
return rc;
@@ -2916,7 +2830,6 @@ handle_state_change:
*/
static long ibmvscsis_create_command_q(struct scsi_info *vscsi, int num_cmds)
{
- long rc = 0;
int pages;
struct vio_dev *vdev = vscsi->dma_dev;
@@ -2940,22 +2853,7 @@ static long ibmvscsis_create_command_q(struct scsi_info *vscsi, int num_cmds)
return -ENOMEM;
}
- rc = h_reg_crq(vscsi->dds.unit_id, vscsi->cmd_q.crq_token, PAGE_SIZE);
- if (rc) {
- if (rc == H_CLOSED) {
- vscsi->state = WAIT_ENABLED;
- rc = 0;
- } else {
- dma_unmap_single(&vdev->dev, vscsi->cmd_q.crq_token,
- PAGE_SIZE, DMA_BIDIRECTIONAL);
- free_page((unsigned long)vscsi->cmd_q.base_addr);
- rc = -ENODEV;
- }
- } else {
- vscsi->state = WAIT_ENABLED;
- }
-
- return rc;
+ return 0;
}
/**
@@ -3270,7 +3168,7 @@ static void ibmvscsis_handle_crq(unsigned long data)
/*
* if we are in a path where we are waiting for all pending commands
* to complete because we received a transport event and anything in
- * the command queue is for a new connection, do nothing
+ * the command queue is for a new connection, do nothing
*/
if (TARGET_STOP(vscsi)) {
vio_enable_interrupts(vscsi->dma_dev);
@@ -3314,7 +3212,7 @@ cmd_work:
* everything but transport events on the queue
*
* need to decrement the queue index so we can
- * look at the elment again
+ * look at the element again
*/
if (vscsi->cmd_q.index)
vscsi->cmd_q.index -= 1;
@@ -3469,6 +3367,7 @@ static int ibmvscsis_probe(struct vio_dev *vdev,
(unsigned long)vscsi);
init_completion(&vscsi->wait_idle);
+ init_completion(&vscsi->unconfig);
snprintf(wq_name, 24, "ibmvscsis%s", dev_name(&vdev->dev));
vscsi->work_q = create_workqueue(wq_name);
@@ -3485,31 +3384,12 @@ static int ibmvscsis_probe(struct vio_dev *vdev,
goto destroy_WQ;
}
- spin_lock_bh(&vscsi->intr_lock);
- vio_enable_interrupts(vdev);
- if (rc) {
- dev_err(&vscsi->dev, "enabling interrupts failed, rc %d\n", rc);
- rc = -ENODEV;
- spin_unlock_bh(&vscsi->intr_lock);
- goto free_irq;
- }
-
- if (ibmvscsis_check_q(vscsi)) {
- rc = ERROR;
- dev_err(&vscsi->dev, "probe: check_q failed, rc %d\n", rc);
- spin_unlock_bh(&vscsi->intr_lock);
- goto disable_interrupt;
- }
- spin_unlock_bh(&vscsi->intr_lock);
+ vscsi->state = WAIT_ENABLED;
dev_set_drvdata(&vdev->dev, vscsi);
return 0;
-disable_interrupt:
- vio_disable_interrupts(vdev);
-free_irq:
- free_irq(vdev->irq, vscsi);
destroy_WQ:
destroy_workqueue(vscsi->work_q);
unmap_buf:
@@ -3543,10 +3423,11 @@ static int ibmvscsis_remove(struct vio_dev *vdev)
pr_debug("remove (%s)\n", dev_name(&vscsi->dma_dev->dev));
- /*
- * TBD: Need to handle if there are commands on the waiting_rsp q
- * Actually, can there still be cmds outstanding to tcm?
- */
+ spin_lock_bh(&vscsi->intr_lock);
+ ibmvscsis_post_disconnect(vscsi, UNCONFIGURING, 0);
+ vscsi->flags |= CFG_SLEEPING;
+ spin_unlock_bh(&vscsi->intr_lock);
+ wait_for_completion(&vscsi->unconfig);
vio_disable_interrupts(vdev);
free_irq(vdev->irq, vscsi);
@@ -3555,7 +3436,6 @@ static int ibmvscsis_remove(struct vio_dev *vdev)
DMA_BIDIRECTIONAL);
kfree(vscsi->map_buf);
tasklet_kill(&vscsi->work_task);
- ibmvscsis_unregister_command_q(vscsi);
ibmvscsis_destroy_command_q(vscsi);
ibmvscsis_freetimer(vscsi);
ibmvscsis_free_cmds(vscsi);
@@ -3903,18 +3783,22 @@ static ssize_t ibmvscsis_tpg_enable_store(struct config_item *item,
}
if (tmp) {
- tport->enabled = true;
spin_lock_bh(&vscsi->intr_lock);
+ tport->enabled = true;
lrc = ibmvscsis_enable_change_state(vscsi);
if (lrc)
pr_err("enable_change_state failed, rc %ld state %d\n",
lrc, vscsi->state);
spin_unlock_bh(&vscsi->intr_lock);
} else {
+ spin_lock_bh(&vscsi->intr_lock);
tport->enabled = false;
+ /* This simulates the server going down */
+ ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT, 0);
+ spin_unlock_bh(&vscsi->intr_lock);
}
- pr_debug("tpg_enable_store, state %d\n", vscsi->state);
+ pr_debug("tpg_enable_store, tmp %ld, state %d\n", tmp, vscsi->state);
return count;
}
@@ -3983,10 +3867,10 @@ static struct attribute *ibmvscsis_dev_attrs[] = {
ATTRIBUTE_GROUPS(ibmvscsis_dev);
static struct class ibmvscsis_class = {
- .name = "ibmvscsis",
- .dev_release = ibmvscsis_dev_release,
- .class_attrs = ibmvscsis_class_attrs,
- .dev_groups = ibmvscsis_dev_groups,
+ .name = "ibmvscsis",
+ .dev_release = ibmvscsis_dev_release,
+ .class_attrs = ibmvscsis_class_attrs,
+ .dev_groups = ibmvscsis_dev_groups,
};
static struct vio_device_id ibmvscsis_device_table[] = {
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h
index 981a0c9..98b0ca7 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h
@@ -204,8 +204,6 @@ struct scsi_info {
struct list_head waiting_rsp;
#define NO_QUEUE 0x00
#define WAIT_ENABLED 0X01
- /* driver has received an initialize command */
-#define PART_UP_WAIT_ENAB 0x02
#define WAIT_CONNECTION 0x04
/* have established a connection */
#define CONNECTED 0x08
@@ -259,6 +257,8 @@ struct scsi_info {
#define SCHEDULE_DISCONNECT 0x00400
/* disconnect handler is scheduled */
#define DISCONNECT_SCHEDULED 0x00800
+ /* remove function is sleeping */
+#define CFG_SLEEPING 0x01000
u32 flags;
/* adapter lock */
spinlock_t intr_lock;
@@ -287,6 +287,7 @@ struct scsi_info {
struct workqueue_struct *work_q;
struct completion wait_idle;
+ struct completion unconfig;
struct device dev;
struct vio_dev *dma_dev;
struct srp_target target;
--
2.5.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH v1 2/3] Properly handle if target_submit_cmd/tmr fails
2016-10-11 22:58 [PATCH v1 0/3] Fixes for problems found during testing Michael Cyr
2016-10-11 22:58 ` [PATCH v1 1/3] Synchronization of cmds during termination conditions Michael Cyr
@ 2016-10-11 22:58 ` Michael Cyr
2016-10-11 22:58 ` [PATCH v1 3/3] Return correct partition name/number to client Michael Cyr
2016-10-13 16:02 ` [PATCH v2 0/6] Fixes for ibmvscsis driver Michael Cyr
3 siblings, 0 replies; 15+ messages in thread
From: Michael Cyr @ 2016-10-11 22:58 UTC (permalink / raw)
To: nab, joe
Cc: hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen, Michael Cyr
Signed-off-by: Michael Cyr <mikecyr@us.ibm.com>
---
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
index 34f758c..5c4b979 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
@@ -2561,6 +2561,10 @@ static void ibmvscsis_parse_cmd(struct scsi_info *vscsi,
data_len, attr, dir, 0);
if (rc) {
dev_err(&vscsi->dev, "target_submit_cmd failed, rc %d\n", rc);
+ spin_lock_bh(&vscsi->intr_lock);
+ list_del(&cmd->list);
+ ibmvscsis_free_cmd_resources(vscsi, cmd);
+ spin_unlock_bh(&vscsi->intr_lock);
goto fail;
}
return;
@@ -2640,6 +2644,9 @@ static void ibmvscsis_parse_task(struct scsi_info *vscsi,
if (rc) {
dev_err(&vscsi->dev, "target_submit_tmr failed, rc %d\n",
rc);
+ spin_lock_bh(&vscsi->intr_lock);
+ list_del(&cmd->list);
+ spin_unlock_bh(&vscsi->intr_lock);
cmd->se_cmd.se_tmr_req->response =
TMR_FUNCTION_REJECTED;
}
--
2.5.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH v1 3/3] Return correct partition name/number to client
2016-10-11 22:58 [PATCH v1 0/3] Fixes for problems found during testing Michael Cyr
2016-10-11 22:58 ` [PATCH v1 1/3] Synchronization of cmds during termination conditions Michael Cyr
2016-10-11 22:58 ` [PATCH v1 2/3] Properly handle if target_submit_cmd/tmr fails Michael Cyr
@ 2016-10-11 22:58 ` Michael Cyr
2016-10-13 16:02 ` [PATCH v2 0/6] Fixes for ibmvscsis driver Michael Cyr
3 siblings, 0 replies; 15+ messages in thread
From: Michael Cyr @ 2016-10-11 22:58 UTC (permalink / raw)
To: nab, joe
Cc: hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen, Michael Cyr
Signed-off-by: Michael Cyr <mikecyr@us.ibm.com>
---
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
index 5c4b979..4dbbe4b 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
@@ -3298,6 +3298,9 @@ static int ibmvscsis_probe(struct vio_dev *vdev,
strncat(vscsi->eye, vdev->name, MAX_EYE);
vscsi->dds.unit_id = vdev->unit_address;
+ strncpy(vscsi->dds.partition_name, partition_name,
+ sizeof(vscsi->dds.partition_name));
+ vscsi->dds.partition_num = partition_number;
spin_lock_bh(&ibmvscsis_dev_lock);
list_add_tail(&vscsi->list, &ibmvscsis_dev_list);
@@ -3496,7 +3499,7 @@ static int ibmvscsis_get_system_info(void)
num = of_get_property(rootdn, "ibm,partition-no", NULL);
if (num)
- partition_number = *num;
+ partition_number = of_read_number(num, 1);
of_node_put(rootdn);
--
2.5.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* Re: [PATCH v1 1/3] Synchronization of cmds during termination conditions
2016-10-11 22:58 ` [PATCH v1 1/3] Synchronization of cmds during termination conditions Michael Cyr
@ 2016-10-12 6:20 ` Greg KH
2016-10-12 14:09 ` Bryant G. Ly
1 sibling, 0 replies; 15+ messages in thread
From: Greg KH @ 2016-10-12 6:20 UTC (permalink / raw)
To: Michael Cyr
Cc: nab, joe, hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, seroyer, linux-scsi, target-devel,
martin.petersen
On Tue, Oct 11, 2016 at 05:58:03PM -0500, Michael Cyr wrote:
> Signed-off-by: Michael Cyr <mikecyr@us.ibm.com>
I almost never accept patches with no changelog information. For a
patch that changes 1082 lines in a non-trivial way, you would think that
you could explain why you are doing this work...
Also, check the prefix of your subject, it give no clue as to where in
the kernel your patches touch :(
thanks,
greg k-h
^ permalink raw reply [flat|nested] 15+ messages in thread
* Re: [PATCH v1 1/3] Synchronization of cmds during termination conditions
2016-10-11 22:58 ` [PATCH v1 1/3] Synchronization of cmds during termination conditions Michael Cyr
2016-10-12 6:20 ` Greg KH
@ 2016-10-12 14:09 ` Bryant G. Ly
1 sibling, 0 replies; 15+ messages in thread
From: Bryant G. Ly @ 2016-10-12 14:09 UTC (permalink / raw)
To: Michael Cyr, nab, joe
Cc: hch, James.Bottomley, tyreld, brking, akpm, bart.vanassche,
gregkh, seroyer, linux-scsi, target-devel, martin.petersen
On 10/11/16 5:58 PM, Michael Cyr wrote:
> Signed-off-by: Michael Cyr <mikecyr@us.ibm.com>
> ---
> drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 1082 +++++++++++++-----------------
> drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h | 5 +-
> 2 files changed, 486 insertions(+), 601 deletions(-)
>
>
I would make the first patch reorganization and styling fixes, no actual
code changes.
Then for each of the commits make sure you amend them to give more
description
as to what you are doing in each specific patch.
i.e. First patch:
ibmvscsis: Re-organization of commands/styling
Need to move functions around in order for the proceeding patches to work.
Patch 2/3 and 3/3 look fine to me after you change the title to include
"ibmvscsis:" in it.
-Bryant
^ permalink raw reply [flat|nested] 15+ messages in thread
* [PATCH v2 0/6] Fixes for ibmvscsis driver
2016-10-11 22:58 [PATCH v1 0/3] Fixes for problems found during testing Michael Cyr
` (2 preceding siblings ...)
2016-10-11 22:58 ` [PATCH v1 3/3] Return correct partition name/number to client Michael Cyr
@ 2016-10-13 16:02 ` Michael Cyr
2016-10-13 16:02 ` [PATCH v2 1/6] ibmvscsis: Rearrange functions for future patches Michael Cyr
` (7 more replies)
3 siblings, 8 replies; 15+ messages in thread
From: Michael Cyr @ 2016-10-13 16:02 UTC (permalink / raw)
To: nab, joe
Cc: hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen, Michael Cyr
Various fixes and cleanups for the ibmvscsis driver.
The first is a sort of prequel to the second, which is the primary
change (and the biggest). The rest are fairly small fixes.
Michael Cyr (6):
ibmvscsis: Rearrange functions for future patches
ibmvscsis: Synchronize cmds at tpg_enable_store time
ibmvscsis: Synchronize cmds at remove time
ibmvscsis: Clean up properly if target_submit_cmd/tmr fails
ibmvscsis: Return correct partition name/# to client
ibmvscsis: Issues from Dan Carpenter/Smatch
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 1096 +++++++++++++-----------------
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h | 5 +-
2 files changed, 494 insertions(+), 607 deletions(-)
--
2.5.0
^ permalink raw reply [flat|nested] 15+ messages in thread
* [PATCH v2 1/6] ibmvscsis: Rearrange functions for future patches
2016-10-13 16:02 ` [PATCH v2 0/6] Fixes for ibmvscsis driver Michael Cyr
@ 2016-10-13 16:02 ` Michael Cyr
2016-10-13 16:02 ` [PATCH v2 2/6] ibmvscsis: Synchronize cmds at tpg_enable_store time Michael Cyr
` (6 subsequent siblings)
7 siblings, 0 replies; 15+ messages in thread
From: Michael Cyr @ 2016-10-13 16:02 UTC (permalink / raw)
To: nab, joe
Cc: hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen, Michael Cyr
This patch reorders functions in a manner necessary for a follow-on
patch. It also makes some minor styling changes (mostly removing extra
spaces) and fixes some typos.
There are no code changes in this patch, with one exception: due to the
reordering of the functions, I needed to explicitly declare a function
at the top of the file. However, this will be removed in the next patch,
since the code requiring the predeclaration will be removed.
Signed-off-by: Michael Cyr <mikecyr@us.ibm.com>
Signed-off-by: Bryant G. Ly <bryantly@linux.vnet.ibm.com>
---
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 796 ++++++++++++++++---------------
1 file changed, 399 insertions(+), 397 deletions(-)
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
index 642b739..01a430c 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
@@ -22,7 +22,7 @@
*
****************************************************************************/
-#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/module.h>
#include <linux/kernel.h>
@@ -61,6 +61,8 @@ static long ibmvscsis_parse_command(struct scsi_info *vscsi,
static void ibmvscsis_adapter_idle(struct scsi_info *vscsi);
+static void ibmvscsis_reset_queue(struct scsi_info *vscsi, uint new_state);
+
static void ibmvscsis_determine_resid(struct se_cmd *se_cmd,
struct srp_rsp *rsp)
{
@@ -81,7 +83,7 @@ static void ibmvscsis_determine_resid(struct se_cmd *se_cmd,
}
} else if (se_cmd->se_cmd_flags & SCF_OVERFLOW_BIT) {
if (se_cmd->data_direction == DMA_TO_DEVICE) {
- /* residual data from an overflow write */
+ /* residual data from an overflow write */
rsp->flags = SRP_RSP_FLAG_DOOVER;
rsp->data_out_res_cnt = cpu_to_be32(residual_count);
} else if (se_cmd->data_direction == DMA_FROM_DEVICE) {
@@ -101,7 +103,7 @@ static void ibmvscsis_determine_resid(struct se_cmd *se_cmd,
* and the function returns TRUE.
*
* EXECUTION ENVIRONMENT:
- * Interrupt or Process environment
+ * Interrupt or Process environment
*/
static bool connection_broken(struct scsi_info *vscsi)
{
@@ -324,7 +326,7 @@ static struct viosrp_crq *ibmvscsis_cmd_q_dequeue(uint mask,
}
/**
- * ibmvscsis_send_init_message() - send initialize message to the client
+ * ibmvscsis_send_init_message() - send initialize message to the client
* @vscsi: Pointer to our adapter structure
* @format: Which Init Message format to send
*
@@ -382,13 +384,13 @@ static long ibmvscsis_check_init_msg(struct scsi_info *vscsi, uint *format)
vscsi->cmd_q.base_addr);
if (crq) {
*format = (uint)(crq->format);
- rc = ERROR;
+ rc = ERROR;
crq->valid = INVALIDATE_CMD_RESP_EL;
dma_rmb();
}
} else {
*format = (uint)(crq->format);
- rc = ERROR;
+ rc = ERROR;
crq->valid = INVALIDATE_CMD_RESP_EL;
dma_rmb();
}
@@ -397,166 +399,6 @@ static long ibmvscsis_check_init_msg(struct scsi_info *vscsi, uint *format)
}
/**
- * ibmvscsis_establish_new_q() - Establish new CRQ queue
- * @vscsi: Pointer to our adapter structure
- * @new_state: New state being established after resetting the queue
- *
- * Must be called with interrupt lock held.
- */
-static long ibmvscsis_establish_new_q(struct scsi_info *vscsi, uint new_state)
-{
- long rc = ADAPT_SUCCESS;
- uint format;
-
- vscsi->flags &= PRESERVE_FLAG_FIELDS;
- vscsi->rsp_q_timer.timer_pops = 0;
- vscsi->debit = 0;
- vscsi->credit = 0;
-
- rc = vio_enable_interrupts(vscsi->dma_dev);
- if (rc) {
- pr_warn("reset_queue: failed to enable interrupts, rc %ld\n",
- rc);
- return rc;
- }
-
- rc = ibmvscsis_check_init_msg(vscsi, &format);
- if (rc) {
- dev_err(&vscsi->dev, "reset_queue: check_init_msg failed, rc %ld\n",
- rc);
- return rc;
- }
-
- if (format == UNUSED_FORMAT && new_state == WAIT_CONNECTION) {
- rc = ibmvscsis_send_init_message(vscsi, INIT_MSG);
- switch (rc) {
- case H_SUCCESS:
- case H_DROPPED:
- case H_CLOSED:
- rc = ADAPT_SUCCESS;
- break;
-
- case H_PARAMETER:
- case H_HARDWARE:
- break;
-
- default:
- vscsi->state = UNDEFINED;
- rc = H_HARDWARE;
- break;
- }
- }
-
- return rc;
-}
-
-/**
- * ibmvscsis_reset_queue() - Reset CRQ Queue
- * @vscsi: Pointer to our adapter structure
- * @new_state: New state to establish after resetting the queue
- *
- * This function calls h_free_q and then calls h_reg_q and does all
- * of the bookkeeping to get us back to where we can communicate.
- *
- * Actually, we don't always call h_free_crq. A problem was discovered
- * where one partition would close and reopen his queue, which would
- * cause his partner to get a transport event, which would cause him to
- * close and reopen his queue, which would cause the original partition
- * to get a transport event, etc., etc. To prevent this, we don't
- * actually close our queue if the client initiated the reset, (i.e.
- * either we got a transport event or we have detected that the client's
- * queue is gone)
- *
- * EXECUTION ENVIRONMENT:
- * Process environment, called with interrupt lock held
- */
-static void ibmvscsis_reset_queue(struct scsi_info *vscsi, uint new_state)
-{
- int bytes;
- long rc = ADAPT_SUCCESS;
-
- pr_debug("reset_queue: flags 0x%x\n", vscsi->flags);
-
- /* don't reset, the client did it for us */
- if (vscsi->flags & (CLIENT_FAILED | TRANS_EVENT)) {
- vscsi->flags &= PRESERVE_FLAG_FIELDS;
- vscsi->rsp_q_timer.timer_pops = 0;
- vscsi->debit = 0;
- vscsi->credit = 0;
- vscsi->state = new_state;
- vio_enable_interrupts(vscsi->dma_dev);
- } else {
- rc = ibmvscsis_free_command_q(vscsi);
- if (rc == ADAPT_SUCCESS) {
- vscsi->state = new_state;
-
- bytes = vscsi->cmd_q.size * PAGE_SIZE;
- rc = h_reg_crq(vscsi->dds.unit_id,
- vscsi->cmd_q.crq_token, bytes);
- if (rc == H_CLOSED || rc == H_SUCCESS) {
- rc = ibmvscsis_establish_new_q(vscsi,
- new_state);
- }
-
- if (rc != ADAPT_SUCCESS) {
- pr_debug("reset_queue: reg_crq rc %ld\n", rc);
-
- vscsi->state = ERR_DISCONNECTED;
- vscsi->flags |= RESPONSE_Q_DOWN;
- ibmvscsis_free_command_q(vscsi);
- }
- } else {
- vscsi->state = ERR_DISCONNECTED;
- vscsi->flags |= RESPONSE_Q_DOWN;
- }
- }
-}
-
-/**
- * ibmvscsis_free_cmd_resources() - Free command resources
- * @vscsi: Pointer to our adapter structure
- * @cmd: Command which is not longer in use
- *
- * Must be called with interrupt lock held.
- */
-static void ibmvscsis_free_cmd_resources(struct scsi_info *vscsi,
- struct ibmvscsis_cmd *cmd)
-{
- struct iu_entry *iue = cmd->iue;
-
- switch (cmd->type) {
- case TASK_MANAGEMENT:
- case SCSI_CDB:
- /*
- * When the queue goes down this value is cleared, so it
- * cannot be cleared in this general purpose function.
- */
- if (vscsi->debit)
- vscsi->debit -= 1;
- break;
- case ADAPTER_MAD:
- vscsi->flags &= ~PROCESSING_MAD;
- break;
- case UNSET_TYPE:
- break;
- default:
- dev_err(&vscsi->dev, "free_cmd_resources unknown type %d\n",
- cmd->type);
- break;
- }
-
- cmd->iue = NULL;
- list_add_tail(&cmd->list, &vscsi->free_cmd);
- srp_iu_put(iue);
-
- if (list_empty(&vscsi->active_q) && list_empty(&vscsi->schedule_q) &&
- list_empty(&vscsi->waiting_rsp) && (vscsi->flags & WAIT_FOR_IDLE)) {
- vscsi->flags &= ~WAIT_FOR_IDLE;
- complete(&vscsi->wait_idle);
- }
-}
-
-/**
* ibmvscsis_disconnect() - Helper function to disconnect
* @work: Pointer to work_struct, gives access to our adapter structure
*
@@ -589,7 +431,7 @@ static void ibmvscsis_disconnect(struct work_struct *work)
* should transitition to the new state
*/
switch (vscsi->state) {
- /* Should never be called while in this state. */
+ /* Should never be called while in this state. */
case NO_QUEUE:
/*
* Can never transition from this state;
@@ -725,84 +567,394 @@ static void ibmvscsis_disconnect(struct work_struct *work)
* the new state (if the one passed in is more "severe" than the
* previous one).
*
- * PRECONDITION:
- * interrupt lock is held
+ * PRECONDITION:
+ * interrupt lock is held
+ */
+static void ibmvscsis_post_disconnect(struct scsi_info *vscsi, uint new_state,
+ uint flag_bits)
+{
+ uint state;
+
+ /* check the validity of the new state */
+ switch (new_state) {
+ case UNCONFIGURING:
+ case ERR_DISCONNECT:
+ case ERR_DISCONNECT_RECONNECT:
+ case WAIT_IDLE:
+ break;
+
+ default:
+ dev_err(&vscsi->dev, "post_disconnect: Invalid new state %d\n",
+ new_state);
+ return;
+ }
+
+ vscsi->flags |= flag_bits;
+
+ pr_debug("post_disconnect: new_state 0x%x, flag_bits 0x%x, vscsi->flags 0x%x, state %hx\n",
+ new_state, flag_bits, vscsi->flags, vscsi->state);
+
+ if (!(vscsi->flags & (DISCONNECT_SCHEDULED | SCHEDULE_DISCONNECT))) {
+ vscsi->flags |= SCHEDULE_DISCONNECT;
+ vscsi->new_state = new_state;
+
+ INIT_WORK(&vscsi->proc_work, ibmvscsis_disconnect);
+ (void)queue_work(vscsi->work_q, &vscsi->proc_work);
+ } else {
+ if (vscsi->new_state)
+ state = vscsi->new_state;
+ else
+ state = vscsi->state;
+
+ switch (state) {
+ case NO_QUEUE:
+ case UNCONFIGURING:
+ break;
+
+ case ERR_DISCONNECTED:
+ case ERR_DISCONNECT:
+ case UNDEFINED:
+ if (new_state == UNCONFIGURING)
+ vscsi->new_state = new_state;
+ break;
+
+ case ERR_DISCONNECT_RECONNECT:
+ switch (new_state) {
+ case UNCONFIGURING:
+ case ERR_DISCONNECT:
+ vscsi->new_state = new_state;
+ break;
+ default:
+ break;
+ }
+ break;
+
+ case WAIT_ENABLED:
+ case PART_UP_WAIT_ENAB:
+ case WAIT_IDLE:
+ case WAIT_CONNECTION:
+ case CONNECTED:
+ case SRP_PROCESSING:
+ vscsi->new_state = new_state;
+ break;
+
+ default:
+ break;
+ }
+ }
+
+ pr_debug("Leaving post_disconnect: flags 0x%x, new_state 0x%x\n",
+ vscsi->flags, vscsi->new_state);
+}
+
+/**
+ * ibmvscsis_handle_init_compl_msg() - Respond to an Init Complete Message
+ * @vscsi: Pointer to our adapter structure
+ *
+ * Must be called with interrupt lock held.
+ */
+static long ibmvscsis_handle_init_compl_msg(struct scsi_info *vscsi)
+{
+ long rc = ADAPT_SUCCESS;
+
+ switch (vscsi->state) {
+ case NO_QUEUE:
+ case ERR_DISCONNECT:
+ case ERR_DISCONNECT_RECONNECT:
+ case ERR_DISCONNECTED:
+ case UNCONFIGURING:
+ case UNDEFINED:
+ rc = ERROR;
+ break;
+
+ case WAIT_CONNECTION:
+ vscsi->state = CONNECTED;
+ break;
+
+ case WAIT_IDLE:
+ case SRP_PROCESSING:
+ case CONNECTED:
+ case WAIT_ENABLED:
+ case PART_UP_WAIT_ENAB:
+ default:
+ rc = ERROR;
+ dev_err(&vscsi->dev, "init_msg: invalid state %d to get init compl msg\n",
+ vscsi->state);
+ ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
+ break;
+ }
+
+ return rc;
+}
+
+/**
+ * ibmvscsis_handle_init_msg() - Respond to an Init Message
+ * @vscsi: Pointer to our adapter structure
+ *
+ * Must be called with interrupt lock held.
+ */
+static long ibmvscsis_handle_init_msg(struct scsi_info *vscsi)
+{
+ long rc = ADAPT_SUCCESS;
+
+ switch (vscsi->state) {
+ case WAIT_ENABLED:
+ vscsi->state = PART_UP_WAIT_ENAB;
+ break;
+
+ case WAIT_CONNECTION:
+ rc = ibmvscsis_send_init_message(vscsi, INIT_COMPLETE_MSG);
+ switch (rc) {
+ case H_SUCCESS:
+ vscsi->state = CONNECTED;
+ break;
+
+ case H_PARAMETER:
+ dev_err(&vscsi->dev, "init_msg: failed to send, rc %ld\n",
+ rc);
+ ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT, 0);
+ break;
+
+ case H_DROPPED:
+ dev_err(&vscsi->dev, "init_msg: failed to send, rc %ld\n",
+ rc);
+ rc = ERROR;
+ ibmvscsis_post_disconnect(vscsi,
+ ERR_DISCONNECT_RECONNECT, 0);
+ break;
+
+ case H_CLOSED:
+ pr_warn("init_msg: failed to send, rc %ld\n", rc);
+ rc = 0;
+ break;
+ }
+ break;
+
+ case UNDEFINED:
+ rc = ERROR;
+ break;
+
+ case UNCONFIGURING:
+ break;
+
+ case PART_UP_WAIT_ENAB:
+ case CONNECTED:
+ case SRP_PROCESSING:
+ case WAIT_IDLE:
+ case NO_QUEUE:
+ case ERR_DISCONNECT:
+ case ERR_DISCONNECT_RECONNECT:
+ case ERR_DISCONNECTED:
+ default:
+ rc = ERROR;
+ dev_err(&vscsi->dev, "init_msg: invalid state %d to get init msg\n",
+ vscsi->state);
+ ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
+ break;
+ }
+
+ return rc;
+}
+
+/**
+ * ibmvscsis_init_msg() - Respond to an init message
+ * @vscsi: Pointer to our adapter structure
+ * @crq: Pointer to CRQ element containing the Init Message
+ *
+ * EXECUTION ENVIRONMENT:
+ * Interrupt, interrupt lock held
+ */
+static long ibmvscsis_init_msg(struct scsi_info *vscsi, struct viosrp_crq *crq)
+{
+ long rc = ADAPT_SUCCESS;
+
+ pr_debug("init_msg: state 0x%hx\n", vscsi->state);
+
+ rc = h_vioctl(vscsi->dds.unit_id, H_GET_PARTNER_INFO,
+ (u64)vscsi->map_ioba | ((u64)PAGE_SIZE << 32), 0, 0, 0,
+ 0);
+ if (rc == H_SUCCESS) {
+ vscsi->client_data.partition_number =
+ be64_to_cpu(*(u64 *)vscsi->map_buf);
+ pr_debug("init_msg, part num %d\n",
+ vscsi->client_data.partition_number);
+ } else {
+ pr_debug("init_msg h_vioctl rc %ld\n", rc);
+ rc = ADAPT_SUCCESS;
+ }
+
+ if (crq->format == INIT_MSG) {
+ rc = ibmvscsis_handle_init_msg(vscsi);
+ } else if (crq->format == INIT_COMPLETE_MSG) {
+ rc = ibmvscsis_handle_init_compl_msg(vscsi);
+ } else {
+ rc = ERROR;
+ dev_err(&vscsi->dev, "init_msg: invalid format %d\n",
+ (uint)crq->format);
+ ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
+ }
+
+ return rc;
+}
+
+/**
+ * ibmvscsis_establish_new_q() - Establish new CRQ queue
+ * @vscsi: Pointer to our adapter structure
+ * @new_state: New state being established after resetting the queue
+ *
+ * Must be called with interrupt lock held.
+ */
+static long ibmvscsis_establish_new_q(struct scsi_info *vscsi, uint new_state)
+{
+ long rc = ADAPT_SUCCESS;
+ uint format;
+
+ vscsi->flags &= PRESERVE_FLAG_FIELDS;
+ vscsi->rsp_q_timer.timer_pops = 0;
+ vscsi->debit = 0;
+ vscsi->credit = 0;
+
+ rc = vio_enable_interrupts(vscsi->dma_dev);
+ if (rc) {
+ pr_warn("reset_queue: failed to enable interrupts, rc %ld\n",
+ rc);
+ return rc;
+ }
+
+ rc = ibmvscsis_check_init_msg(vscsi, &format);
+ if (rc) {
+ dev_err(&vscsi->dev, "reset_queue: check_init_msg failed, rc %ld\n",
+ rc);
+ return rc;
+ }
+
+ if (format == UNUSED_FORMAT && new_state == WAIT_CONNECTION) {
+ rc = ibmvscsis_send_init_message(vscsi, INIT_MSG);
+ switch (rc) {
+ case H_SUCCESS:
+ case H_DROPPED:
+ case H_CLOSED:
+ rc = ADAPT_SUCCESS;
+ break;
+
+ case H_PARAMETER:
+ case H_HARDWARE:
+ break;
+
+ default:
+ vscsi->state = UNDEFINED;
+ rc = H_HARDWARE;
+ break;
+ }
+ }
+
+ return rc;
+}
+
+/**
+ * ibmvscsis_reset_queue() - Reset CRQ Queue
+ * @vscsi: Pointer to our adapter structure
+ * @new_state: New state to establish after resetting the queue
+ *
+ * This function calls h_free_q and then calls h_reg_q and does all
+ * of the bookkeeping to get us back to where we can communicate.
+ *
+ * Actually, we don't always call h_free_crq. A problem was discovered
+ * where one partition would close and reopen his queue, which would
+ * cause his partner to get a transport event, which would cause him to
+ * close and reopen his queue, which would cause the original partition
+ * to get a transport event, etc., etc. To prevent this, we don't
+ * actually close our queue if the client initiated the reset, (i.e.
+ * either we got a transport event or we have detected that the client's
+ * queue is gone)
+ *
+ * EXECUTION ENVIRONMENT:
+ * Process environment, called with interrupt lock held
+ */
+static void ibmvscsis_reset_queue(struct scsi_info *vscsi, uint new_state)
+{
+ int bytes;
+ long rc = ADAPT_SUCCESS;
+
+ pr_debug("reset_queue: flags 0x%x\n", vscsi->flags);
+
+ /* don't reset, the client did it for us */
+ if (vscsi->flags & (CLIENT_FAILED | TRANS_EVENT)) {
+ vscsi->flags &= PRESERVE_FLAG_FIELDS;
+ vscsi->rsp_q_timer.timer_pops = 0;
+ vscsi->debit = 0;
+ vscsi->credit = 0;
+ vscsi->state = new_state;
+ vio_enable_interrupts(vscsi->dma_dev);
+ } else {
+ rc = ibmvscsis_free_command_q(vscsi);
+ if (rc == ADAPT_SUCCESS) {
+ vscsi->state = new_state;
+
+ bytes = vscsi->cmd_q.size * PAGE_SIZE;
+ rc = h_reg_crq(vscsi->dds.unit_id,
+ vscsi->cmd_q.crq_token, bytes);
+ if (rc == H_CLOSED || rc == H_SUCCESS) {
+ rc = ibmvscsis_establish_new_q(vscsi,
+ new_state);
+ }
+
+ if (rc != ADAPT_SUCCESS) {
+ pr_debug("reset_queue: reg_crq rc %ld\n", rc);
+
+ vscsi->state = ERR_DISCONNECTED;
+ vscsi->flags |= RESPONSE_Q_DOWN;
+ ibmvscsis_free_command_q(vscsi);
+ }
+ } else {
+ vscsi->state = ERR_DISCONNECTED;
+ vscsi->flags |= RESPONSE_Q_DOWN;
+ }
+ }
+}
+
+/**
+ * ibmvscsis_free_cmd_resources() - Free command resources
+ * @vscsi: Pointer to our adapter structure
+ * @cmd: Command which is not longer in use
+ *
+ * Must be called with interrupt lock held.
*/
-static void ibmvscsis_post_disconnect(struct scsi_info *vscsi, uint new_state,
- uint flag_bits)
+static void ibmvscsis_free_cmd_resources(struct scsi_info *vscsi,
+ struct ibmvscsis_cmd *cmd)
{
- uint state;
+ struct iu_entry *iue = cmd->iue;
- /* check the validity of the new state */
- switch (new_state) {
- case UNCONFIGURING:
- case ERR_DISCONNECT:
- case ERR_DISCONNECT_RECONNECT:
- case WAIT_IDLE:
+ switch (cmd->type) {
+ case TASK_MANAGEMENT:
+ case SCSI_CDB:
+ /*
+ * When the queue goes down this value is cleared, so it
+ * cannot be cleared in this general purpose function.
+ */
+ if (vscsi->debit)
+ vscsi->debit -= 1;
+ break;
+ case ADAPTER_MAD:
+ vscsi->flags &= ~PROCESSING_MAD;
+ break;
+ case UNSET_TYPE:
break;
-
default:
- dev_err(&vscsi->dev, "post_disconnect: Invalid new state %d\n",
- new_state);
- return;
+ dev_err(&vscsi->dev, "free_cmd_resources unknown type %d\n",
+ cmd->type);
+ break;
}
- vscsi->flags |= flag_bits;
-
- pr_debug("post_disconnect: new_state 0x%x, flag_bits 0x%x, vscsi->flags 0x%x, state %hx\n",
- new_state, flag_bits, vscsi->flags, vscsi->state);
-
- if (!(vscsi->flags & (DISCONNECT_SCHEDULED | SCHEDULE_DISCONNECT))) {
- vscsi->flags |= SCHEDULE_DISCONNECT;
- vscsi->new_state = new_state;
-
- INIT_WORK(&vscsi->proc_work, ibmvscsis_disconnect);
- (void)queue_work(vscsi->work_q, &vscsi->proc_work);
- } else {
- if (vscsi->new_state)
- state = vscsi->new_state;
- else
- state = vscsi->state;
-
- switch (state) {
- case NO_QUEUE:
- case UNCONFIGURING:
- break;
-
- case ERR_DISCONNECTED:
- case ERR_DISCONNECT:
- case UNDEFINED:
- if (new_state == UNCONFIGURING)
- vscsi->new_state = new_state;
- break;
-
- case ERR_DISCONNECT_RECONNECT:
- switch (new_state) {
- case UNCONFIGURING:
- case ERR_DISCONNECT:
- vscsi->new_state = new_state;
- break;
- default:
- break;
- }
- break;
-
- case WAIT_ENABLED:
- case PART_UP_WAIT_ENAB:
- case WAIT_IDLE:
- case WAIT_CONNECTION:
- case CONNECTED:
- case SRP_PROCESSING:
- vscsi->new_state = new_state;
- break;
+ cmd->iue = NULL;
+ list_add_tail(&cmd->list, &vscsi->free_cmd);
+ srp_iu_put(iue);
- default:
- break;
- }
+ if (list_empty(&vscsi->active_q) && list_empty(&vscsi->schedule_q) &&
+ list_empty(&vscsi->waiting_rsp) && (vscsi->flags & WAIT_FOR_IDLE)) {
+ vscsi->flags &= ~WAIT_FOR_IDLE;
+ complete(&vscsi->wait_idle);
}
-
- pr_debug("Leaving post_disconnect: flags 0x%x, new_state 0x%x\n",
- vscsi->flags, vscsi->new_state);
}
/**
@@ -895,7 +1047,7 @@ static long ibmvscsis_trans_event(struct scsi_info *vscsi,
}
}
- rc = vscsi->flags & SCHEDULE_DISCONNECT;
+ rc = vscsi->flags & SCHEDULE_DISCONNECT;
pr_debug("Leaving trans_event: flags 0x%x, state 0x%hx, rc %ld\n",
vscsi->flags, vscsi->state, rc);
@@ -1220,7 +1372,7 @@ static long ibmvscsis_copy_crq_packet(struct scsi_info *vscsi,
* @iue: Information Unit containing the Adapter Info MAD request
*
* EXECUTION ENVIRONMENT:
- * Interrupt adpater lock is held
+ * Interrupt adapter lock is held
*/
static long ibmvscsis_adapter_info(struct scsi_info *vscsi,
struct iu_entry *iue)
@@ -1691,7 +1843,7 @@ static void ibmvscsis_send_mad_resp(struct scsi_info *vscsi,
* @crq: Pointer to the CRQ entry containing the MAD request
*
* EXECUTION ENVIRONMENT:
- * Interrupt called with adapter lock held
+ * Interrupt, called with adapter lock held
*/
static long ibmvscsis_mad(struct scsi_info *vscsi, struct viosrp_crq *crq)
{
@@ -1864,7 +2016,7 @@ static long ibmvscsis_srp_login_rej(struct scsi_info *vscsi,
break;
case H_PERMISSION:
if (connection_broken(vscsi))
- flag_bits = RESPONSE_Q_DOWN | CLIENT_FAILED;
+ flag_bits = RESPONSE_Q_DOWN | CLIENT_FAILED;
dev_err(&vscsi->dev, "login_rej: error copying to client, rc %ld\n",
rc);
ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT,
@@ -2187,156 +2339,6 @@ static long ibmvscsis_ping_response(struct scsi_info *vscsi)
}
/**
- * ibmvscsis_handle_init_compl_msg() - Respond to an Init Complete Message
- * @vscsi: Pointer to our adapter structure
- *
- * Must be called with interrupt lock held.
- */
-static long ibmvscsis_handle_init_compl_msg(struct scsi_info *vscsi)
-{
- long rc = ADAPT_SUCCESS;
-
- switch (vscsi->state) {
- case NO_QUEUE:
- case ERR_DISCONNECT:
- case ERR_DISCONNECT_RECONNECT:
- case ERR_DISCONNECTED:
- case UNCONFIGURING:
- case UNDEFINED:
- rc = ERROR;
- break;
-
- case WAIT_CONNECTION:
- vscsi->state = CONNECTED;
- break;
-
- case WAIT_IDLE:
- case SRP_PROCESSING:
- case CONNECTED:
- case WAIT_ENABLED:
- case PART_UP_WAIT_ENAB:
- default:
- rc = ERROR;
- dev_err(&vscsi->dev, "init_msg: invalid state %d to get init compl msg\n",
- vscsi->state);
- ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
- break;
- }
-
- return rc;
-}
-
-/**
- * ibmvscsis_handle_init_msg() - Respond to an Init Message
- * @vscsi: Pointer to our adapter structure
- *
- * Must be called with interrupt lock held.
- */
-static long ibmvscsis_handle_init_msg(struct scsi_info *vscsi)
-{
- long rc = ADAPT_SUCCESS;
-
- switch (vscsi->state) {
- case WAIT_ENABLED:
- vscsi->state = PART_UP_WAIT_ENAB;
- break;
-
- case WAIT_CONNECTION:
- rc = ibmvscsis_send_init_message(vscsi, INIT_COMPLETE_MSG);
- switch (rc) {
- case H_SUCCESS:
- vscsi->state = CONNECTED;
- break;
-
- case H_PARAMETER:
- dev_err(&vscsi->dev, "init_msg: failed to send, rc %ld\n",
- rc);
- ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT, 0);
- break;
-
- case H_DROPPED:
- dev_err(&vscsi->dev, "init_msg: failed to send, rc %ld\n",
- rc);
- rc = ERROR;
- ibmvscsis_post_disconnect(vscsi,
- ERR_DISCONNECT_RECONNECT, 0);
- break;
-
- case H_CLOSED:
- pr_warn("init_msg: failed to send, rc %ld\n", rc);
- rc = 0;
- break;
- }
- break;
-
- case UNDEFINED:
- rc = ERROR;
- break;
-
- case UNCONFIGURING:
- break;
-
- case PART_UP_WAIT_ENAB:
- case CONNECTED:
- case SRP_PROCESSING:
- case WAIT_IDLE:
- case NO_QUEUE:
- case ERR_DISCONNECT:
- case ERR_DISCONNECT_RECONNECT:
- case ERR_DISCONNECTED:
- default:
- rc = ERROR;
- dev_err(&vscsi->dev, "init_msg: invalid state %d to get init msg\n",
- vscsi->state);
- ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
- break;
- }
-
- return rc;
-}
-
-/**
- * ibmvscsis_init_msg() - Respond to an init message
- * @vscsi: Pointer to our adapter structure
- * @crq: Pointer to CRQ element containing the Init Message
- *
- * EXECUTION ENVIRONMENT:
- * Interrupt, interrupt lock held
- */
-static long ibmvscsis_init_msg(struct scsi_info *vscsi, struct viosrp_crq *crq)
-{
- long rc = ADAPT_SUCCESS;
-
- pr_debug("init_msg: state 0x%hx\n", vscsi->state);
-
- rc = h_vioctl(vscsi->dds.unit_id, H_GET_PARTNER_INFO,
- (u64)vscsi->map_ioba | ((u64)PAGE_SIZE << 32), 0, 0, 0,
- 0);
- if (rc == H_SUCCESS) {
- vscsi->client_data.partition_number =
- be64_to_cpu(*(u64 *)vscsi->map_buf);
- pr_debug("init_msg, part num %d\n",
- vscsi->client_data.partition_number);
- } else {
- pr_debug("init_msg h_vioctl rc %ld\n", rc);
- rc = ADAPT_SUCCESS;
- }
-
- if (crq->format == INIT_MSG) {
- rc = ibmvscsis_handle_init_msg(vscsi);
- } else if (crq->format == INIT_COMPLETE_MSG) {
- rc = ibmvscsis_handle_init_compl_msg(vscsi);
- } else {
- rc = ERROR;
- dev_err(&vscsi->dev, "init_msg: invalid format %d\n",
- (uint)crq->format);
- ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
- }
-
- return rc;
-}
-
-/**
* ibmvscsis_parse_command() - Parse an element taken from the cmd rsp queue.
* @vscsi: Pointer to our adapter structure
* @crq: Pointer to CRQ element containing the SRP request
@@ -2391,7 +2393,7 @@ static long ibmvscsis_parse_command(struct scsi_info *vscsi,
break;
case VALID_TRANS_EVENT:
- rc = ibmvscsis_trans_event(vscsi, crq);
+ rc = ibmvscsis_trans_event(vscsi, crq);
break;
case VALID_INIT_MSG:
@@ -3270,7 +3272,7 @@ static void ibmvscsis_handle_crq(unsigned long data)
/*
* if we are in a path where we are waiting for all pending commands
* to complete because we received a transport event and anything in
- * the command queue is for a new connection, do nothing
+ * the command queue is for a new connection, do nothing
*/
if (TARGET_STOP(vscsi)) {
vio_enable_interrupts(vscsi->dma_dev);
@@ -3314,7 +3316,7 @@ cmd_work:
* everything but transport events on the queue
*
* need to decrement the queue index so we can
- * look at the elment again
+ * look at the element again
*/
if (vscsi->cmd_q.index)
vscsi->cmd_q.index -= 1;
@@ -3983,10 +3985,10 @@ static struct attribute *ibmvscsis_dev_attrs[] = {
ATTRIBUTE_GROUPS(ibmvscsis_dev);
static struct class ibmvscsis_class = {
- .name = "ibmvscsis",
- .dev_release = ibmvscsis_dev_release,
- .class_attrs = ibmvscsis_class_attrs,
- .dev_groups = ibmvscsis_dev_groups,
+ .name = "ibmvscsis",
+ .dev_release = ibmvscsis_dev_release,
+ .class_attrs = ibmvscsis_class_attrs,
+ .dev_groups = ibmvscsis_dev_groups,
};
static struct vio_device_id ibmvscsis_device_table[] = {
--
2.5.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH v2 2/6] ibmvscsis: Synchronize cmds at tpg_enable_store time
2016-10-13 16:02 ` [PATCH v2 0/6] Fixes for ibmvscsis driver Michael Cyr
2016-10-13 16:02 ` [PATCH v2 1/6] ibmvscsis: Rearrange functions for future patches Michael Cyr
@ 2016-10-13 16:02 ` Michael Cyr
2016-10-13 16:02 ` [PATCH v2 3/6] ibmvscsis: Synchronize cmds at remove time Michael Cyr
` (5 subsequent siblings)
7 siblings, 0 replies; 15+ messages in thread
From: Michael Cyr @ 2016-10-13 16:02 UTC (permalink / raw)
To: nab, joe
Cc: hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen, Michael Cyr
This patch changes the way the IBM vSCSI server driver manages its
Command/Response Queue (CRQ). We used to register the CRQ with phyp at
probe time. Now we wait until tpg_enable_store. Similarly, when
tpg_enable_store is called to "disable" (i.e. the stored value is 0),
we unregister the queue with phyp.
One consquence to this is that we have no need for the PART_UP_WAIT_ENAB
state, since we can't get an Init Message from the client in our CRQ if
we're waiting to be enabled, since we haven't registered the queue yet.
Signed-off-by: Michael Cyr <mikecyr@us.ibm.com>
Signed-off-by: Bryant G. Ly <bryantly@linux.vnet.ibm.com>
---
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 224 ++++++-------------------------
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h | 2 -
2 files changed, 38 insertions(+), 188 deletions(-)
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
index 01a430c..2ce1d73 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
@@ -61,8 +61,6 @@ static long ibmvscsis_parse_command(struct scsi_info *vscsi,
static void ibmvscsis_adapter_idle(struct scsi_info *vscsi);
-static void ibmvscsis_reset_queue(struct scsi_info *vscsi, uint new_state);
-
static void ibmvscsis_determine_resid(struct se_cmd *se_cmd,
struct srp_rsp *rsp)
{
@@ -417,7 +415,6 @@ static void ibmvscsis_disconnect(struct work_struct *work)
proc_work);
u16 new_state;
bool wait_idle = false;
- long rc = ADAPT_SUCCESS;
spin_lock_bh(&vscsi->intr_lock);
new_state = vscsi->new_state;
@@ -470,30 +467,12 @@ static void ibmvscsis_disconnect(struct work_struct *work)
vscsi->state = new_state;
break;
- /*
- * If this is a transition into an error state.
- * a client is attempting to establish a connection
- * and has violated the RPA protocol.
- * There can be nothing pending on the adapter although
- * there can be requests in the command queue.
- */
case WAIT_ENABLED:
- case PART_UP_WAIT_ENAB:
switch (new_state) {
+ /* should never happen */
case ERR_DISCONNECT:
- vscsi->flags |= RESPONSE_Q_DOWN;
- vscsi->state = new_state;
- vscsi->flags &= ~(SCHEDULE_DISCONNECT |
- DISCONNECT_SCHEDULED);
- ibmvscsis_free_command_q(vscsi);
- break;
case ERR_DISCONNECT_RECONNECT:
- ibmvscsis_reset_queue(vscsi, WAIT_ENABLED);
- break;
-
- /* should never happen */
case WAIT_IDLE:
- rc = ERROR;
dev_err(&vscsi->dev, "disconnect: invalid state %d for WAIT_IDLE\n",
vscsi->state);
break;
@@ -630,7 +609,6 @@ static void ibmvscsis_post_disconnect(struct scsi_info *vscsi, uint new_state,
break;
case WAIT_ENABLED:
- case PART_UP_WAIT_ENAB:
case WAIT_IDLE:
case WAIT_CONNECTION:
case CONNECTED:
@@ -675,7 +653,6 @@ static long ibmvscsis_handle_init_compl_msg(struct scsi_info *vscsi)
case SRP_PROCESSING:
case CONNECTED:
case WAIT_ENABLED:
- case PART_UP_WAIT_ENAB:
default:
rc = ERROR;
dev_err(&vscsi->dev, "init_msg: invalid state %d to get init compl msg\n",
@@ -698,10 +675,6 @@ static long ibmvscsis_handle_init_msg(struct scsi_info *vscsi)
long rc = ADAPT_SUCCESS;
switch (vscsi->state) {
- case WAIT_ENABLED:
- vscsi->state = PART_UP_WAIT_ENAB;
- break;
-
case WAIT_CONNECTION:
rc = ibmvscsis_send_init_message(vscsi, INIT_COMPLETE_MSG);
switch (rc) {
@@ -737,7 +710,7 @@ static long ibmvscsis_handle_init_msg(struct scsi_info *vscsi)
case UNCONFIGURING:
break;
- case PART_UP_WAIT_ENAB:
+ case WAIT_ENABLED:
case CONNECTED:
case SRP_PROCESSING:
case WAIT_IDLE:
@@ -800,11 +773,10 @@ static long ibmvscsis_init_msg(struct scsi_info *vscsi, struct viosrp_crq *crq)
/**
* ibmvscsis_establish_new_q() - Establish new CRQ queue
* @vscsi: Pointer to our adapter structure
- * @new_state: New state being established after resetting the queue
*
* Must be called with interrupt lock held.
*/
-static long ibmvscsis_establish_new_q(struct scsi_info *vscsi, uint new_state)
+static long ibmvscsis_establish_new_q(struct scsi_info *vscsi)
{
long rc = ADAPT_SUCCESS;
uint format;
@@ -816,19 +788,19 @@ static long ibmvscsis_establish_new_q(struct scsi_info *vscsi, uint new_state)
rc = vio_enable_interrupts(vscsi->dma_dev);
if (rc) {
- pr_warn("reset_queue: failed to enable interrupts, rc %ld\n",
+ pr_warn("establish_new_q: failed to enable interrupts, rc %ld\n",
rc);
return rc;
}
rc = ibmvscsis_check_init_msg(vscsi, &format);
if (rc) {
- dev_err(&vscsi->dev, "reset_queue: check_init_msg failed, rc %ld\n",
+ dev_err(&vscsi->dev, "establish_new_q: check_init_msg failed, rc %ld\n",
rc);
return rc;
}
- if (format == UNUSED_FORMAT && new_state == WAIT_CONNECTION) {
+ if (format == UNUSED_FORMAT) {
rc = ibmvscsis_send_init_message(vscsi, INIT_MSG);
switch (rc) {
case H_SUCCESS:
@@ -846,6 +818,8 @@ static long ibmvscsis_establish_new_q(struct scsi_info *vscsi, uint new_state)
rc = H_HARDWARE;
break;
}
+ } else if (format == INIT_MSG) {
+ rc = ibmvscsis_handle_init_msg(vscsi);
}
return rc;
@@ -854,7 +828,6 @@ static long ibmvscsis_establish_new_q(struct scsi_info *vscsi, uint new_state)
/**
* ibmvscsis_reset_queue() - Reset CRQ Queue
* @vscsi: Pointer to our adapter structure
- * @new_state: New state to establish after resetting the queue
*
* This function calls h_free_q and then calls h_reg_q and does all
* of the bookkeeping to get us back to where we can communicate.
@@ -871,7 +844,7 @@ static long ibmvscsis_establish_new_q(struct scsi_info *vscsi, uint new_state)
* EXECUTION ENVIRONMENT:
* Process environment, called with interrupt lock held
*/
-static void ibmvscsis_reset_queue(struct scsi_info *vscsi, uint new_state)
+static void ibmvscsis_reset_queue(struct scsi_info *vscsi)
{
int bytes;
long rc = ADAPT_SUCCESS;
@@ -884,19 +857,18 @@ static void ibmvscsis_reset_queue(struct scsi_info *vscsi, uint new_state)
vscsi->rsp_q_timer.timer_pops = 0;
vscsi->debit = 0;
vscsi->credit = 0;
- vscsi->state = new_state;
+ vscsi->state = WAIT_CONNECTION;
vio_enable_interrupts(vscsi->dma_dev);
} else {
rc = ibmvscsis_free_command_q(vscsi);
if (rc == ADAPT_SUCCESS) {
- vscsi->state = new_state;
+ vscsi->state = WAIT_CONNECTION;
bytes = vscsi->cmd_q.size * PAGE_SIZE;
rc = h_reg_crq(vscsi->dds.unit_id,
vscsi->cmd_q.crq_token, bytes);
if (rc == H_CLOSED || rc == H_SUCCESS) {
- rc = ibmvscsis_establish_new_q(vscsi,
- new_state);
+ rc = ibmvscsis_establish_new_q(vscsi);
}
if (rc != ADAPT_SUCCESS) {
@@ -1015,10 +987,6 @@ static long ibmvscsis_trans_event(struct scsi_info *vscsi,
TRANS_EVENT));
break;
- case PART_UP_WAIT_ENAB:
- vscsi->state = WAIT_ENABLED;
- break;
-
case SRP_PROCESSING:
if ((vscsi->debit > 0) ||
!list_empty(&vscsi->schedule_q) ||
@@ -1219,15 +1187,18 @@ static void ibmvscsis_adapter_idle(struct scsi_info *vscsi)
switch (vscsi->state) {
case ERR_DISCONNECT_RECONNECT:
- ibmvscsis_reset_queue(vscsi, WAIT_CONNECTION);
+ ibmvscsis_reset_queue(vscsi);
pr_debug("adapter_idle, disc_rec: flags 0x%x\n", vscsi->flags);
break;
case ERR_DISCONNECT:
ibmvscsis_free_command_q(vscsi);
- vscsi->flags &= ~DISCONNECT_SCHEDULED;
+ vscsi->flags &= ~(SCHEDULE_DISCONNECT | DISCONNECT_SCHEDULED);
vscsi->flags |= RESPONSE_Q_DOWN;
- vscsi->state = ERR_DISCONNECTED;
+ if (vscsi->tport.enabled)
+ vscsi->state = ERR_DISCONNECTED;
+ else
+ vscsi->state = WAIT_ENABLED;
pr_debug("adapter_idle, disc: flags 0x%x, state 0x%hx\n",
vscsi->flags, vscsi->state);
break;
@@ -1772,8 +1743,8 @@ static void ibmvscsis_send_messages(struct scsi_info *vscsi)
be64_to_cpu(msg_hi),
be64_to_cpu(cmd->rsp.tag));
- pr_debug("send_messages: tag 0x%llx, rc %ld\n",
- be64_to_cpu(cmd->rsp.tag), rc);
+ pr_debug("send_messages: cmd %p, tag 0x%llx, rc %ld\n",
+ cmd, be64_to_cpu(cmd->rsp.tag), rc);
/* if all ok free up the command element resources */
if (rc == H_SUCCESS) {
@@ -2788,36 +2759,6 @@ static irqreturn_t ibmvscsis_interrupt(int dummy, void *data)
}
/**
- * ibmvscsis_check_q() - Helper function to Check Init Message Valid
- * @vscsi: Pointer to our adapter structure
- *
- * Checks if a initialize message was queued by the initiatior
- * while the timing window was open. This function is called from
- * probe after the CRQ is created and interrupts are enabled.
- * It would only be used by adapters who wait for some event before
- * completing the init handshake with the client. For ibmvscsi, this
- * event is waiting for the port to be enabled.
- *
- * EXECUTION ENVIRONMENT:
- * Process level only, interrupt lock held
- */
-static long ibmvscsis_check_q(struct scsi_info *vscsi)
-{
- uint format;
- long rc;
-
- rc = ibmvscsis_check_init_msg(vscsi, &format);
- if (rc)
- ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT_RECONNECT, 0);
- else if (format == UNUSED_FORMAT)
- vscsi->state = WAIT_ENABLED;
- else
- vscsi->state = PART_UP_WAIT_ENAB;
-
- return rc;
-}
-
-/**
* ibmvscsis_enable_change_state() - Set new state based on enabled status
* @vscsi: Pointer to our adapter structure
*
@@ -2828,77 +2769,19 @@ static long ibmvscsis_check_q(struct scsi_info *vscsi)
*/
static long ibmvscsis_enable_change_state(struct scsi_info *vscsi)
{
+ int bytes;
long rc = ADAPT_SUCCESS;
-handle_state_change:
- switch (vscsi->state) {
- case WAIT_ENABLED:
- rc = ibmvscsis_send_init_message(vscsi, INIT_MSG);
- switch (rc) {
- case H_SUCCESS:
- case H_DROPPED:
- case H_CLOSED:
- vscsi->state = WAIT_CONNECTION;
- rc = ADAPT_SUCCESS;
- break;
-
- case H_PARAMETER:
- break;
-
- case H_HARDWARE:
- break;
-
- default:
- vscsi->state = UNDEFINED;
- rc = H_HARDWARE;
- break;
- }
- break;
- case PART_UP_WAIT_ENAB:
- rc = ibmvscsis_send_init_message(vscsi, INIT_COMPLETE_MSG);
- switch (rc) {
- case H_SUCCESS:
- vscsi->state = CONNECTED;
- rc = ADAPT_SUCCESS;
- break;
-
- case H_DROPPED:
- case H_CLOSED:
- vscsi->state = WAIT_ENABLED;
- goto handle_state_change;
-
- case H_PARAMETER:
- break;
-
- case H_HARDWARE:
- break;
-
- default:
- rc = H_HARDWARE;
- break;
- }
- break;
-
- case WAIT_CONNECTION:
- case WAIT_IDLE:
- case SRP_PROCESSING:
- case CONNECTED:
- rc = ADAPT_SUCCESS;
- break;
- /* should not be able to get here */
- case UNCONFIGURING:
- rc = ERROR;
- vscsi->state = UNDEFINED;
- break;
+ bytes = vscsi->cmd_q.size * PAGE_SIZE;
+ rc = h_reg_crq(vscsi->dds.unit_id, vscsi->cmd_q.crq_token, bytes);
+ if (rc == H_CLOSED || rc == H_SUCCESS) {
+ vscsi->state = WAIT_CONNECTION;
+ rc = ibmvscsis_establish_new_q(vscsi);
+ }
- /* driver should never allow this to happen */
- case ERR_DISCONNECT:
- case ERR_DISCONNECT_RECONNECT:
- default:
- dev_err(&vscsi->dev, "in invalid state %d during enable_change_state\n",
- vscsi->state);
- rc = ADAPT_SUCCESS;
- break;
+ if (rc != ADAPT_SUCCESS) {
+ vscsi->state = ERR_DISCONNECTED;
+ vscsi->flags |= RESPONSE_Q_DOWN;
}
return rc;
@@ -2918,7 +2801,6 @@ handle_state_change:
*/
static long ibmvscsis_create_command_q(struct scsi_info *vscsi, int num_cmds)
{
- long rc = 0;
int pages;
struct vio_dev *vdev = vscsi->dma_dev;
@@ -2942,22 +2824,7 @@ static long ibmvscsis_create_command_q(struct scsi_info *vscsi, int num_cmds)
return -ENOMEM;
}
- rc = h_reg_crq(vscsi->dds.unit_id, vscsi->cmd_q.crq_token, PAGE_SIZE);
- if (rc) {
- if (rc == H_CLOSED) {
- vscsi->state = WAIT_ENABLED;
- rc = 0;
- } else {
- dma_unmap_single(&vdev->dev, vscsi->cmd_q.crq_token,
- PAGE_SIZE, DMA_BIDIRECTIONAL);
- free_page((unsigned long)vscsi->cmd_q.base_addr);
- rc = -ENODEV;
- }
- } else {
- vscsi->state = WAIT_ENABLED;
- }
-
- return rc;
+ return 0;
}
/**
@@ -3487,31 +3354,12 @@ static int ibmvscsis_probe(struct vio_dev *vdev,
goto destroy_WQ;
}
- spin_lock_bh(&vscsi->intr_lock);
- vio_enable_interrupts(vdev);
- if (rc) {
- dev_err(&vscsi->dev, "enabling interrupts failed, rc %d\n", rc);
- rc = -ENODEV;
- spin_unlock_bh(&vscsi->intr_lock);
- goto free_irq;
- }
-
- if (ibmvscsis_check_q(vscsi)) {
- rc = ERROR;
- dev_err(&vscsi->dev, "probe: check_q failed, rc %d\n", rc);
- spin_unlock_bh(&vscsi->intr_lock);
- goto disable_interrupt;
- }
- spin_unlock_bh(&vscsi->intr_lock);
+ vscsi->state = WAIT_ENABLED;
dev_set_drvdata(&vdev->dev, vscsi);
return 0;
-disable_interrupt:
- vio_disable_interrupts(vdev);
-free_irq:
- free_irq(vdev->irq, vscsi);
destroy_WQ:
destroy_workqueue(vscsi->work_q);
unmap_buf:
@@ -3905,18 +3753,22 @@ static ssize_t ibmvscsis_tpg_enable_store(struct config_item *item,
}
if (tmp) {
- tport->enabled = true;
spin_lock_bh(&vscsi->intr_lock);
+ tport->enabled = true;
lrc = ibmvscsis_enable_change_state(vscsi);
if (lrc)
pr_err("enable_change_state failed, rc %ld state %d\n",
lrc, vscsi->state);
spin_unlock_bh(&vscsi->intr_lock);
} else {
+ spin_lock_bh(&vscsi->intr_lock);
tport->enabled = false;
+ /* This simulates the server going down */
+ ibmvscsis_post_disconnect(vscsi, ERR_DISCONNECT, 0);
+ spin_unlock_bh(&vscsi->intr_lock);
}
- pr_debug("tpg_enable_store, state %d\n", vscsi->state);
+ pr_debug("tpg_enable_store, tmp %ld, state %d\n", tmp, vscsi->state);
return count;
}
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h
index 981a0c9..17e0ef4 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h
@@ -204,8 +204,6 @@ struct scsi_info {
struct list_head waiting_rsp;
#define NO_QUEUE 0x00
#define WAIT_ENABLED 0X01
- /* driver has received an initialize command */
-#define PART_UP_WAIT_ENAB 0x02
#define WAIT_CONNECTION 0x04
/* have established a connection */
#define CONNECTED 0x08
--
2.5.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH v2 3/6] ibmvscsis: Synchronize cmds at remove time
2016-10-13 16:02 ` [PATCH v2 0/6] Fixes for ibmvscsis driver Michael Cyr
2016-10-13 16:02 ` [PATCH v2 1/6] ibmvscsis: Rearrange functions for future patches Michael Cyr
2016-10-13 16:02 ` [PATCH v2 2/6] ibmvscsis: Synchronize cmds at tpg_enable_store time Michael Cyr
@ 2016-10-13 16:02 ` Michael Cyr
2016-10-13 16:02 ` [PATCH v2 4/6] ibmvscsis: Clean up properly if target_submit_cmd/tmr fails Michael Cyr
` (4 subsequent siblings)
7 siblings, 0 replies; 15+ messages in thread
From: Michael Cyr @ 2016-10-13 16:02 UTC (permalink / raw)
To: nab, joe
Cc: hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen, Michael Cyr
This patch adds code to disconnect from the client, which will make sure
any outstanding commands have been completed, before continuing on with
the remove operation.
Signed-off-by: Michael Cyr <mikecyr@us.ibm.com>
Signed-off-by: Bryant G. Ly <bryantly@linux.vnet.ibm.com>
---
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 39 ++++++++++++++++++++++++++++----
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h | 3 +++
2 files changed, 37 insertions(+), 5 deletions(-)
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
index 2ce1d73..41af435 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
@@ -469,6 +469,18 @@ static void ibmvscsis_disconnect(struct work_struct *work)
case WAIT_ENABLED:
switch (new_state) {
+ case UNCONFIGURING:
+ vscsi->state = new_state;
+ vscsi->flags |= RESPONSE_Q_DOWN;
+ vscsi->flags &= ~(SCHEDULE_DISCONNECT |
+ DISCONNECT_SCHEDULED);
+ dma_rmb();
+ if (vscsi->flags & CFG_SLEEPING) {
+ vscsi->flags &= ~CFG_SLEEPING;
+ complete(&vscsi->unconfig);
+ }
+ break;
+
/* should never happen */
case ERR_DISCONNECT:
case ERR_DISCONNECT_RECONNECT:
@@ -481,6 +493,13 @@ static void ibmvscsis_disconnect(struct work_struct *work)
case WAIT_IDLE:
switch (new_state) {
+ case UNCONFIGURING:
+ vscsi->flags |= RESPONSE_Q_DOWN;
+ vscsi->state = new_state;
+ vscsi->flags &= ~(SCHEDULE_DISCONNECT |
+ DISCONNECT_SCHEDULED);
+ ibmvscsis_free_command_q(vscsi);
+ break;
case ERR_DISCONNECT:
case ERR_DISCONNECT_RECONNECT:
vscsi->state = new_state;
@@ -1186,6 +1205,15 @@ static void ibmvscsis_adapter_idle(struct scsi_info *vscsi)
free_qs = true;
switch (vscsi->state) {
+ case UNCONFIGURING:
+ ibmvscsis_free_command_q(vscsi);
+ dma_rmb();
+ isync();
+ if (vscsi->flags & CFG_SLEEPING) {
+ vscsi->flags &= ~CFG_SLEEPING;
+ complete(&vscsi->unconfig);
+ }
+ break;
case ERR_DISCONNECT_RECONNECT:
ibmvscsis_reset_queue(vscsi);
pr_debug("adapter_idle, disc_rec: flags 0x%x\n", vscsi->flags);
@@ -3338,6 +3366,7 @@ static int ibmvscsis_probe(struct vio_dev *vdev,
(unsigned long)vscsi);
init_completion(&vscsi->wait_idle);
+ init_completion(&vscsi->unconfig);
snprintf(wq_name, 24, "ibmvscsis%s", dev_name(&vdev->dev));
vscsi->work_q = create_workqueue(wq_name);
@@ -3393,10 +3422,11 @@ static int ibmvscsis_remove(struct vio_dev *vdev)
pr_debug("remove (%s)\n", dev_name(&vscsi->dma_dev->dev));
- /*
- * TBD: Need to handle if there are commands on the waiting_rsp q
- * Actually, can there still be cmds outstanding to tcm?
- */
+ spin_lock_bh(&vscsi->intr_lock);
+ ibmvscsis_post_disconnect(vscsi, UNCONFIGURING, 0);
+ vscsi->flags |= CFG_SLEEPING;
+ spin_unlock_bh(&vscsi->intr_lock);
+ wait_for_completion(&vscsi->unconfig);
vio_disable_interrupts(vdev);
free_irq(vdev->irq, vscsi);
@@ -3405,7 +3435,6 @@ static int ibmvscsis_remove(struct vio_dev *vdev)
DMA_BIDIRECTIONAL);
kfree(vscsi->map_buf);
tasklet_kill(&vscsi->work_task);
- ibmvscsis_unregister_command_q(vscsi);
ibmvscsis_destroy_command_q(vscsi);
ibmvscsis_freetimer(vscsi);
ibmvscsis_free_cmds(vscsi);
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h
index 17e0ef4..98b0ca7 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h
@@ -257,6 +257,8 @@ struct scsi_info {
#define SCHEDULE_DISCONNECT 0x00400
/* disconnect handler is scheduled */
#define DISCONNECT_SCHEDULED 0x00800
+ /* remove function is sleeping */
+#define CFG_SLEEPING 0x01000
u32 flags;
/* adapter lock */
spinlock_t intr_lock;
@@ -285,6 +287,7 @@ struct scsi_info {
struct workqueue_struct *work_q;
struct completion wait_idle;
+ struct completion unconfig;
struct device dev;
struct vio_dev *dma_dev;
struct srp_target target;
--
2.5.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH v2 4/6] ibmvscsis: Clean up properly if target_submit_cmd/tmr fails
2016-10-13 16:02 ` [PATCH v2 0/6] Fixes for ibmvscsis driver Michael Cyr
` (2 preceding siblings ...)
2016-10-13 16:02 ` [PATCH v2 3/6] ibmvscsis: Synchronize cmds at remove time Michael Cyr
@ 2016-10-13 16:02 ` Michael Cyr
2016-10-13 16:02 ` [PATCH v2 5/6] ibmvscsis: Return correct partition name/# to client Michael Cyr
` (3 subsequent siblings)
7 siblings, 0 replies; 15+ messages in thread
From: Michael Cyr @ 2016-10-13 16:02 UTC (permalink / raw)
To: nab, joe
Cc: hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen, Michael Cyr
Signed-off-by: Michael Cyr <mikecyr@us.ibm.com>
Signed-off-by: Bryant G. Ly <bryantly@linux.vnet.ibm.com>
---
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
index 41af435..cd9f5c7 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
@@ -2560,6 +2560,10 @@ static void ibmvscsis_parse_cmd(struct scsi_info *vscsi,
data_len, attr, dir, 0);
if (rc) {
dev_err(&vscsi->dev, "target_submit_cmd failed, rc %d\n", rc);
+ spin_lock_bh(&vscsi->intr_lock);
+ list_del(&cmd->list);
+ ibmvscsis_free_cmd_resources(vscsi, cmd);
+ spin_unlock_bh(&vscsi->intr_lock);
goto fail;
}
return;
@@ -2639,6 +2643,9 @@ static void ibmvscsis_parse_task(struct scsi_info *vscsi,
if (rc) {
dev_err(&vscsi->dev, "target_submit_tmr failed, rc %d\n",
rc);
+ spin_lock_bh(&vscsi->intr_lock);
+ list_del(&cmd->list);
+ spin_unlock_bh(&vscsi->intr_lock);
cmd->se_cmd.se_tmr_req->response =
TMR_FUNCTION_REJECTED;
}
--
2.5.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH v2 5/6] ibmvscsis: Return correct partition name/# to client
2016-10-13 16:02 ` [PATCH v2 0/6] Fixes for ibmvscsis driver Michael Cyr
` (3 preceding siblings ...)
2016-10-13 16:02 ` [PATCH v2 4/6] ibmvscsis: Clean up properly if target_submit_cmd/tmr fails Michael Cyr
@ 2016-10-13 16:02 ` Michael Cyr
2016-10-13 16:02 ` [PATCH v2 6/6] ibmvscsis: Issues from Dan Carpenter/Smatch Michael Cyr
` (2 subsequent siblings)
7 siblings, 0 replies; 15+ messages in thread
From: Michael Cyr @ 2016-10-13 16:02 UTC (permalink / raw)
To: nab, joe
Cc: hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen, Michael Cyr
Signed-off-by: Michael Cyr <mikecyr@us.ibm.com>
Signed-off-by: Bryant G. Ly <bryantly@linux.vnet.ibm.com>
---
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
index cd9f5c7..fe220a1 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
@@ -3297,6 +3297,9 @@ static int ibmvscsis_probe(struct vio_dev *vdev,
strncat(vscsi->eye, vdev->name, MAX_EYE);
vscsi->dds.unit_id = vdev->unit_address;
+ strncpy(vscsi->dds.partition_name, partition_name,
+ sizeof(vscsi->dds.partition_name));
+ vscsi->dds.partition_num = partition_number;
spin_lock_bh(&ibmvscsis_dev_lock);
list_add_tail(&vscsi->list, &ibmvscsis_dev_list);
@@ -3495,7 +3498,7 @@ static int ibmvscsis_get_system_info(void)
num = of_get_property(rootdn, "ibm,partition-no", NULL);
if (num)
- partition_number = *num;
+ partition_number = of_read_number(num, 1);
of_node_put(rootdn);
--
2.5.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH v2 6/6] ibmvscsis: Issues from Dan Carpenter/Smatch
2016-10-13 16:02 ` [PATCH v2 0/6] Fixes for ibmvscsis driver Michael Cyr
` (4 preceding siblings ...)
2016-10-13 16:02 ` [PATCH v2 5/6] ibmvscsis: Return correct partition name/# to client Michael Cyr
@ 2016-10-13 16:02 ` Michael Cyr
[not found] ` <8fcdcfd1eedecae72a868ab45a872e06@linux.vnet.ibm.com>
2016-10-19 1:21 ` Martin K. Petersen
7 siblings, 0 replies; 15+ messages in thread
From: Michael Cyr @ 2016-10-13 16:02 UTC (permalink / raw)
To: nab, joe
Cc: hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen, Michael Cyr
Signed-off-by: Michael Cyr <mikecyr@us.ibm.com>
Signed-off-by: Bryant G. Ly <bryantly@linux.vnet.ibm.com>
---
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 13 +++----------
1 file changed, 3 insertions(+), 10 deletions(-)
diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
index fe220a1..c9fa356 100644
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
@@ -1896,14 +1896,7 @@ static long ibmvscsis_mad(struct scsi_info *vscsi, struct viosrp_crq *crq)
pr_debug("mad: type %d\n", be32_to_cpu(mad->type));
- if (be16_to_cpu(mad->length) < 0) {
- dev_err(&vscsi->dev, "mad: length is < 0\n");
- ibmvscsis_post_disconnect(vscsi,
- ERR_DISCONNECT_RECONNECT, 0);
- rc = SRP_VIOLATION;
- } else {
- rc = ibmvscsis_process_mad(vscsi, iue);
- }
+ rc = ibmvscsis_process_mad(vscsi, iue);
pr_debug("mad: status %hd, rc %ld\n", be16_to_cpu(mad->status),
rc);
@@ -2523,7 +2516,6 @@ static void ibmvscsis_parse_cmd(struct scsi_info *vscsi,
dev_err(&vscsi->dev, "0x%llx: parsing SRP descriptor table failed.\n",
srp->tag);
goto fail;
- return;
}
cmd->rsp.sol_not = srp->sol_not;
@@ -3282,7 +3274,8 @@ static int ibmvscsis_probe(struct vio_dev *vdev,
INIT_LIST_HEAD(&vscsi->waiting_rsp);
INIT_LIST_HEAD(&vscsi->active_q);
- snprintf(vscsi->tport.tport_name, 256, "%s", dev_name(&vdev->dev));
+ snprintf(vscsi->tport.tport_name, IBMVSCSIS_NAMELEN, "%s",
+ dev_name(&vdev->dev));
pr_debug("probe tport_name: %s\n", vscsi->tport.tport_name);
--
2.5.0
^ permalink raw reply related [flat|nested] 15+ messages in thread
* Re: [PATCH v2 0/6] Fixes for ibmvscsis driver
[not found] ` <8fcdcfd1eedecae72a868ab45a872e06@linux.vnet.ibm.com>
@ 2016-10-17 16:15 ` Tyrel Datwyler
0 siblings, 0 replies; 15+ messages in thread
From: Tyrel Datwyler @ 2016-10-17 16:15 UTC (permalink / raw)
To: Steven Royer, Michael Cyr
Cc: nab, joe, hch, bryantly, James.Bottomley, brking, akpm,
bart.vanassche, gregkh, linux-scsi, target-devel,
martin.petersen
On 10/14/2016 03:25 PM, Steven Royer wrote:
> On 2016-10-13 11:02, Michael Cyr wrote:
>> Various fixes and cleanups for the ibmvscsis driver.
>> The first is a sort of prequel to the second, which is the primary
>> change (and the biggest). The rest are fairly small fixes.
>>
>> Michael Cyr (6):
>> ibmvscsis: Rearrange functions for future patches
>> ibmvscsis: Synchronize cmds at tpg_enable_store time
>> ibmvscsis: Synchronize cmds at remove time
>> ibmvscsis: Clean up properly if target_submit_cmd/tmr fails
>> ibmvscsis: Return correct partition name/# to client
>> ibmvscsis: Issues from Dan Carpenter/Smatch
>>
>> drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 1096
>> +++++++++++++-----------------
>> drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.h | 5 +-
>> 2 files changed, 494 insertions(+), 607 deletions(-)
>
> I've tested with this patch. Seems to fix cleanup issues, in particular
> after a target is removed or at target reboot time.
> Signed-off-by: Steven Royer <seroyer@linux.vnet.ibm.com>
>
I think the "Tested-by" tag is more appropriate seeing as you tested the
patch.
-Tyrel
^ permalink raw reply [flat|nested] 15+ messages in thread
* Re: [PATCH v2 0/6] Fixes for ibmvscsis driver
2016-10-13 16:02 ` [PATCH v2 0/6] Fixes for ibmvscsis driver Michael Cyr
` (6 preceding siblings ...)
[not found] ` <8fcdcfd1eedecae72a868ab45a872e06@linux.vnet.ibm.com>
@ 2016-10-19 1:21 ` Martin K. Petersen
7 siblings, 0 replies; 15+ messages in thread
From: Martin K. Petersen @ 2016-10-19 1:21 UTC (permalink / raw)
To: Michael Cyr
Cc: nab, joe, hch, bryantly, James.Bottomley, tyreld, brking, akpm,
bart.vanassche, gregkh, seroyer, linux-scsi, target-devel,
martin.petersen
>>>>> "Michael" == Michael Cyr <mikecyr@us.ibm.com> writes:
Michael> Various fixes and cleanups for the ibmvscsis driver. The first
Michael> is a sort of prequel to the second, which is the primary change
Michael> (and the biggest). The rest are fairly small fixes.
Applied to 4.10/scsi-queue.
--
Martin K. Petersen Oracle Linux Engineering
^ permalink raw reply [flat|nested] 15+ messages in thread
end of thread, other threads:[~2016-10-19 1:22 UTC | newest]
Thread overview: 15+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-10-11 22:58 [PATCH v1 0/3] Fixes for problems found during testing Michael Cyr
2016-10-11 22:58 ` [PATCH v1 1/3] Synchronization of cmds during termination conditions Michael Cyr
2016-10-12 6:20 ` Greg KH
2016-10-12 14:09 ` Bryant G. Ly
2016-10-11 22:58 ` [PATCH v1 2/3] Properly handle if target_submit_cmd/tmr fails Michael Cyr
2016-10-11 22:58 ` [PATCH v1 3/3] Return correct partition name/number to client Michael Cyr
2016-10-13 16:02 ` [PATCH v2 0/6] Fixes for ibmvscsis driver Michael Cyr
2016-10-13 16:02 ` [PATCH v2 1/6] ibmvscsis: Rearrange functions for future patches Michael Cyr
2016-10-13 16:02 ` [PATCH v2 2/6] ibmvscsis: Synchronize cmds at tpg_enable_store time Michael Cyr
2016-10-13 16:02 ` [PATCH v2 3/6] ibmvscsis: Synchronize cmds at remove time Michael Cyr
2016-10-13 16:02 ` [PATCH v2 4/6] ibmvscsis: Clean up properly if target_submit_cmd/tmr fails Michael Cyr
2016-10-13 16:02 ` [PATCH v2 5/6] ibmvscsis: Return correct partition name/# to client Michael Cyr
2016-10-13 16:02 ` [PATCH v2 6/6] ibmvscsis: Issues from Dan Carpenter/Smatch Michael Cyr
[not found] ` <8fcdcfd1eedecae72a868ab45a872e06@linux.vnet.ibm.com>
2016-10-17 16:15 ` [PATCH v2 0/6] Fixes for ibmvscsis driver Tyrel Datwyler
2016-10-19 1:21 ` Martin K. Petersen
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.