[PATCH 0/7] mpt3sas: Enhancements and bug fixes

[PATCH 0/7] mpt3sas: Enhancements and bug fixes

[Suganath Prabu S]

This patch set has below bug fixes and enhancements.

Dump IOC system interface register set when IOC fails to
transit to READY state during initialization.

During controller reset, cancel the running work along
with the queued works. This avoids processing the expired event after
the controller reset.

Before returning the target/lun reset with failure status
to SML, poll all the reply descriptor pools looking for the reply of
timed out SCSI command. Due to interrupt latency issues it is possible
that interrupt for timed out command might have been delayed. So poll
all the reply descriptor pools looking for the reply of timed out SCSI
command, if the timed out command 's reply is found then return
success status other return failure status.

Memset the config command reply buffer before issuing the
config request.

Suganath Prabu S (7):
  mpt3sas: Memset config_cmds.reply buffer with zeros
  mpt3sas: Dump system registers for debugging.
  mpt3sas: Cancel the running work during host reset.
  mpt3sas: Rename and export interrupt mask/unmask fn's.
  mpt3sas: Added support functions to find target and luns.
  mpt3sas: Postprocessing of target and LUN reset.
  mpt3sas: Update driver version to 35.100.00.00

 drivers/scsi/mpt3sas/compile.sh       |   6 +-
 drivers/scsi/mpt3sas/mpt3sas_base.c   |  50 ++++--
 drivers/scsi/mpt3sas/mpt3sas_base.h   |  21 ++-
 drivers/scsi/mpt3sas/mpt3sas_config.c |   2 +-
 drivers/scsi/mpt3sas/mpt3sas_ctl.c    |   6 +-
 drivers/scsi/mpt3sas/mpt3sas_scsih.c  | 245 +++++++++++++++++++++++---
 6 files changed, 282 insertions(+), 48 deletions(-)

-- 
2.26.2

[PATCH 1/7] mpt3sas: Memset config_cmds.reply buffer with zeros

[Suganath Prabu S]

Currently config_cmds.reply buffer is not memset to zero before
posting config page request message.
In some cases, for the current config request, the previous config
reply is getting processed and we will observe PageType mismatch
between request to reply buffer. It will be difficult to debug this
type of issue and it confuses by thinking that HBA Firmware itself
posted the wrong config reply.
So it is better to memset the config_cmds.reply buffer with zeros
before issuing the config request.

Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
---
 drivers/scsi/mpt3sas/mpt3sas_config.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/scsi/mpt3sas/mpt3sas_config.c b/drivers/scsi/mpt3sas/mpt3sas_config.c
index 62ddf53..17c7abf 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_config.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_config.c
@@ -372,7 +372,7 @@ _config_request(struct MPT3SAS_ADAPTER *ioc, Mpi2ConfigRequest_t
 	}
 
 	r = 0;
-	memset(mpi_reply, 0, sizeof(Mpi2ConfigReply_t));
+	memset(ioc->config_cmds.reply, 0, sizeof(Mpi2ConfigReply_t));
 	ioc->config_cmds.status = MPT3_CMD_PENDING;
 	config_request = mpt3sas_base_get_msg_frame(ioc, smid);
 	ioc->config_cmds.smid = smid;
-- 
2.26.2

[PATCH 2/7] mpt3sas: Dump system registers for debugging.

[Suganath Prabu S]

When controller fails to transition to READY state during driver
probe, dump the system interface register set. This will give
snapshot of the firmware status for debugging driver load issues.

Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
---
 drivers/scsi/mpt3sas/mpt3sas_base.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c
index 96b78fd..b9d8f08 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.c
@@ -5617,6 +5617,23 @@ _base_wait_on_iocstate(struct MPT3SAS_ADAPTER *ioc, u32 ioc_state, int timeout)
 	return current_state;
 }
 
+/**
+ * _base_dump_reg_set -	This function will print hexdump of register set.
+ * @ioc: per adapter object
+ *
+ * Returns nothing.
+ */
+static inline void
+_base_dump_reg_set(struct MPT3SAS_ADAPTER *ioc)
+{
+	unsigned int i, sz = 256;
+	u32 __iomem *reg = (u32 __iomem *)ioc->chip;
+
+	ioc_info(ioc, "System Register set:\n");
+	for (i = 0; i < (sz / sizeof(u32)); i++)
+		pr_info("%08x: %08x\n", (i * 4), readl(&reg[i]));
+}
+
 /**
  * _base_wait_for_doorbell_int - waiting for controller interrupt(generated by
  * a write to the doorbell)
@@ -6795,6 +6812,7 @@ _base_diag_reset(struct MPT3SAS_ADAPTER *ioc)
 		if (count++ > 20) {
 			ioc_info(ioc,
 			    "Stop writing magic sequence after 20 retries\n");
+			_base_dump_reg_set(ioc);
 			goto out;
 		}
 
@@ -6823,6 +6841,7 @@ _base_diag_reset(struct MPT3SAS_ADAPTER *ioc)
 		if (host_diagnostic == 0xFFFFFFFF) {
 			ioc_info(ioc,
 			    "Invalid host diagnostic register value\n");
+			_base_dump_reg_set(ioc);
 			goto out;
 		}
 		if (!(host_diagnostic & MPI2_DIAG_RESET_ADAPTER))
@@ -6857,6 +6876,7 @@ _base_diag_reset(struct MPT3SAS_ADAPTER *ioc)
 	if (ioc_state) {
 		ioc_err(ioc, "%s: failed going to ready state (ioc_state=0x%x)\n",
 			__func__, ioc_state);
+		_base_dump_reg_set(ioc);
 		goto out;
 	}
 
-- 
2.26.2

[PATCH 3/7] mpt3sas: Cancel the running work during host reset.

[Suganath Prabu S]

Currently during host reset time driver is cancelling only those Firmware
event works which are pending in Firmware event workqueue. It is not
cancelling the work which is running. With this patch driver cancels the
running work also.

Issue Description:

Even though it is not recommended to issue back to back host reset without
any delay, but if someone issues back to back host reset then we observe
that target devices gets unregistered and re-register with SML.
And if OS drive is behind the HBA then when it get unregistered, than
file-system goes into read-only mode. Normally during host reset driver
marks the target device as responding (if they are accessible) and add the
event 'MPT3SAS_REMOVE_UNRESPONDING_DEVICES' to remove the non-responding
devices through FW worker thread. while processing this event driver
unregistered the non-responding devices and clears the responding flag for
all the devices.

The reason why target devices are getting unregistered during successive host
resets is that during the host reset driver has to cleanup all the
outstanding FW event work (both queued one and the currently processing one)
but actually driver is cleaning only the queued events. So if
MPT3SAS_REMOVE_UNRESPONDING_DEVICES event is currently under process then
this event is not getting cleaned up, so at end of all successive host
reset this same event is getting processed more than once. And after the
event got processed for the first time, all the target devices responding
flag is cleared, so when the same is processed for the second time it see
that responding flag is zero, so driver unregistered all the target drives
even-though drives are responding. If driver cleanups the current running
work along with pending work, this type of behavior won't be observed.

Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
---
 drivers/scsi/mpt3sas/mpt3sas_base.h  |  4 ++++
 drivers/scsi/mpt3sas/mpt3sas_scsih.c | 17 ++++++++++++-----
 2 files changed, 16 insertions(+), 5 deletions(-)

diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.h b/drivers/scsi/mpt3sas/mpt3sas_base.h
index 4fca393..4ed704c 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.h
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.h
@@ -1036,6 +1036,8 @@ typedef void (*MPT3SAS_FLUSH_RUNNING_CMDS)(struct MPT3SAS_ADAPTER *ioc);
  * @firmware_event_thread: ""
  * @fw_event_lock:
  * @fw_event_list: list of fw events
+ * @current_evet: current processing firmware event
+ * @fw_event_cleanup: set to one while cleaning up the fw events
  * @aen_event_read_flag: event log was read
  * @broadcast_aen_busy: broadcast aen waiting to be serviced
  * @shost_recovery: host reset in progress
@@ -1217,6 +1219,8 @@ struct MPT3SAS_ADAPTER {
 	struct workqueue_struct	*firmware_event_thread;
 	spinlock_t	fw_event_lock;
 	struct list_head fw_event_list;
+	struct fw_event_work	*current_event;
+	u8		fw_events_cleanup;
 
 	 /* misc flags */
 	int		aen_event_read_flag;
diff --git a/drivers/scsi/mpt3sas/mpt3sas_scsih.c b/drivers/scsi/mpt3sas/mpt3sas_scsih.c
index 08fc4b3..66b29d4 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_scsih.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_scsih.c
@@ -3323,11 +3323,13 @@ _scsih_fw_event_cleanup_queue(struct MPT3SAS_ADAPTER *ioc)
 {
 	struct fw_event_work *fw_event;
 
-	if (list_empty(&ioc->fw_event_list) ||
+	if ((list_empty(&ioc->fw_event_list) && !ioc->current_event) ||
 	     !ioc->firmware_event_thread || in_interrupt())
 		return;
 
-	while ((fw_event = dequeue_next_fw_event(ioc))) {
+	ioc->fw_events_cleanup = 1;
+	while ((fw_event = dequeue_next_fw_event(ioc)) ||
+	     (fw_event = ioc->current_event)) {
 		/*
 		 * Wait on the fw_event to complete. If this returns 1, then
 		 * the event was never executed, and we need a put for the
@@ -3341,6 +3343,7 @@ _scsih_fw_event_cleanup_queue(struct MPT3SAS_ADAPTER *ioc)
 
 		fw_event_work_put(fw_event);
 	}
+	ioc->fw_events_cleanup = 0;
 }
 
 /**
@@ -9421,11 +9424,13 @@ mpt3sas_scsih_reset_done_handler(struct MPT3SAS_ADAPTER *ioc)
 static void
 _mpt3sas_fw_work(struct MPT3SAS_ADAPTER *ioc, struct fw_event_work *fw_event)
 {
+	ioc->current_event = fw_event;
 	_scsih_fw_event_del_from_list(ioc, fw_event);
 
 	/* the queue is being flushed so ignore this event */
 	if (ioc->remove_host || ioc->pci_error_recovery) {
 		fw_event_work_put(fw_event);
+		ioc->current_event = NULL;
 		return;
 	}
 
@@ -9439,10 +9444,10 @@ _mpt3sas_fw_work(struct MPT3SAS_ADAPTER *ioc, struct fw_event_work *fw_event)
 		while (scsi_host_in_recovery(ioc->shost) ||
 					 ioc->shost_recovery) {
 			/*
-			 * If we're unloading, bail. Otherwise, this can become
-			 * an infinite loop.
+			 * If we're unloading or cancelling the work, bail.
+			 * Otherwise, this can become an infinite loop.
 			 */
-			if (ioc->remove_host)
+			if (ioc->remove_host || ioc->fw_events_cleanup)
 				goto out;
 			ssleep(1);
 		}
@@ -9503,11 +9508,13 @@ _mpt3sas_fw_work(struct MPT3SAS_ADAPTER *ioc, struct fw_event_work *fw_event)
 		break;
 	case MPI2_EVENT_PCIE_TOPOLOGY_CHANGE_LIST:
 		_scsih_pcie_topology_change_event(ioc, fw_event);
+		ioc->current_event = NULL;
 			return;
 	break;
 	}
 out:
 	fw_event_work_put(fw_event);
+	ioc->current_event = NULL;
 }
 
 /**
-- 
2.26.2

[PATCH 4/7] mpt3sas: Rename and export interrupt mask/unmask fn's.

[Suganath Prabu S]

Rename Function _base_unmask_interrupts() to
mpt3sas_base_unmask_interrupts() and
_base_mask_interrupts() to mpt3sas_base_mask_interrupts()
Also added function declarion to mpt3sas_base.h

Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
---
 drivers/scsi/mpt3sas/mpt3sas_base.c | 24 +++++++++++-------------
 drivers/scsi/mpt3sas/mpt3sas_base.h |  2 ++
 2 files changed, 13 insertions(+), 13 deletions(-)

diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c
index b9d8f08..2ffdbbb 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.c
@@ -129,8 +129,6 @@ _base_wait_on_iocstate(struct MPT3SAS_ADAPTER *ioc,
 static int
 _base_get_ioc_facts(struct MPT3SAS_ADAPTER *ioc);
 static void
-_base_mask_interrupts(struct MPT3SAS_ADAPTER *ioc);
-static void
 _base_clear_outstanding_commands(struct MPT3SAS_ADAPTER *ioc);
 
 /**
@@ -680,7 +678,7 @@ _base_fault_reset_work(struct work_struct *work)
 			ioc->shost_recovery = 1;
 			spin_unlock_irqrestore(
 			    &ioc->ioc_reset_in_progress_lock, flags);
-			_base_mask_interrupts(ioc);
+			mpt3sas_base_mask_interrupts(ioc);
 			_base_clear_outstanding_commands(ioc);
 		}
 
@@ -1465,13 +1463,13 @@ _base_get_cb_idx(struct MPT3SAS_ADAPTER *ioc, u16 smid)
 }
 
 /**
- * _base_mask_interrupts - disable interrupts
+ * mpt3sas_base_mask_interrupts - disable interrupts
  * @ioc: per adapter object
  *
  * Disabling ResetIRQ, Reply and Doorbell Interrupts
  */
-static void
-_base_mask_interrupts(struct MPT3SAS_ADAPTER *ioc)
+void
+mpt3sas_base_mask_interrupts(struct MPT3SAS_ADAPTER *ioc)
 {
 	u32 him_register;
 
@@ -1483,13 +1481,13 @@ _base_mask_interrupts(struct MPT3SAS_ADAPTER *ioc)
 }
 
 /**
- * _base_unmask_interrupts - enable interrupts
+ * mpt3sas_base_unmask_interrupts - enable interrupts
  * @ioc: per adapter object
  *
  * Enabling only Reply Interrupts
  */
-static void
-_base_unmask_interrupts(struct MPT3SAS_ADAPTER *ioc)
+void
+mpt3sas_base_unmask_interrupts(struct MPT3SAS_ADAPTER *ioc)
 {
 	u32 him_register;
 
@@ -3371,7 +3369,7 @@ mpt3sas_base_map_resources(struct MPT3SAS_ADAPTER *ioc)
 		goto out_fail;
 	}
 
-	_base_mask_interrupts(ioc);
+	mpt3sas_base_mask_interrupts(ioc);
 
 	r = _base_get_ioc_facts(ioc);
 	if (r) {
@@ -7119,7 +7117,7 @@ _base_make_ioc_operational(struct MPT3SAS_ADAPTER *ioc)
 
  skip_init_reply_post_host_index:
 
-	_base_unmask_interrupts(ioc);
+	mpt3sas_base_unmask_interrupts(ioc);
 
 	if (ioc->hba_mpi_version_belonged != MPI2_VERSION) {
 		r = _base_display_fwpkg_version(ioc);
@@ -7168,7 +7166,7 @@ mpt3sas_base_free_resources(struct MPT3SAS_ADAPTER *ioc)
 	/* synchronizing freeing resource with pci_access_mutex lock */
 	mutex_lock(&ioc->pci_access_mutex);
 	if (ioc->chip_phys && ioc->chip) {
-		_base_mask_interrupts(ioc);
+		mpt3sas_base_mask_interrupts(ioc);
 		ioc->shost_recovery = 1;
 		_base_make_ioc_ready(ioc, SOFT_RESET);
 		ioc->shost_recovery = 0;
@@ -7734,7 +7732,7 @@ mpt3sas_base_hard_reset_handler(struct MPT3SAS_ADAPTER *ioc,
 	}
 	_base_pre_reset_handler(ioc);
 	mpt3sas_wait_for_commands_to_complete(ioc);
-	_base_mask_interrupts(ioc);
+	mpt3sas_base_mask_interrupts(ioc);
 	r = _base_make_ioc_ready(ioc, type);
 	if (r)
 		goto out;
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.h b/drivers/scsi/mpt3sas/mpt3sas_base.h
index 4ed704c..9da7265 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.h
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.h
@@ -1529,6 +1529,8 @@ __le32 mpt3sas_base_get_sense_buffer_dma(struct MPT3SAS_ADAPTER *ioc,
 void *mpt3sas_base_get_pcie_sgl(struct MPT3SAS_ADAPTER *ioc, u16 smid);
 dma_addr_t mpt3sas_base_get_pcie_sgl_dma(struct MPT3SAS_ADAPTER *ioc, u16 smid);
 void mpt3sas_base_sync_reply_irqs(struct MPT3SAS_ADAPTER *ioc);
+void mpt3sas_base_mask_interrupts(struct MPT3SAS_ADAPTER *ioc);
+void mpt3sas_base_unmask_interrupts(struct MPT3SAS_ADAPTER *ioc);
 
 void mpt3sas_base_put_smid_fast_path(struct MPT3SAS_ADAPTER *ioc, u16 smid,
 	u16 handle);
-- 
2.26.2

[PATCH 5/7] mpt3sas: Added helper functions to check any cmd is outstanding on Target and LUN.

[Suganath Prabu S]

* Added helper functions to check whether any SCSI command  is
outstanding on particular Target, LUN device,
* Added addition function parameters 'channel', 'id' to function
mpt3sas_scsih_issue_tm().

Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
---
 drivers/scsi/mpt3sas/mpt3sas_base.h  |  9 +--
 drivers/scsi/mpt3sas/mpt3sas_ctl.c   |  6 +-
 drivers/scsi/mpt3sas/mpt3sas_scsih.c | 92 ++++++++++++++++++++++++----
 3 files changed, 88 insertions(+), 19 deletions(-)

diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.h b/drivers/scsi/mpt3sas/mpt3sas_base.h
index 9da7265..fd61deb 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.h
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.h
@@ -1610,11 +1610,12 @@ void mpt3sas_scsih_clear_outstanding_scsi_tm_commands(
 	struct MPT3SAS_ADAPTER *ioc);
 void mpt3sas_scsih_reset_done_handler(struct MPT3SAS_ADAPTER *ioc);
 
-int mpt3sas_scsih_issue_tm(struct MPT3SAS_ADAPTER *ioc, u16 handle, u64 lun,
-	u8 type, u16 smid_task, u16 msix_task, u8 timeout, u8 tr_method);
+int mpt3sas_scsih_issue_tm(struct MPT3SAS_ADAPTER *ioc, u16 handle,
+	uint channel, uint id, u64 lun, u8 type, u16 smid_task,
+	u16 msix_task, u8 timeout, u8 tr_method);
 int mpt3sas_scsih_issue_locked_tm(struct MPT3SAS_ADAPTER *ioc, u16 handle,
-	u64 lun, u8 type, u16 smid_task, u16 msix_task,
-	u8 timeout, u8 tr_method);
+	uint channel, uint id, u64 lun, u8 type, u16 smid_task,
+	u16 msix_task, u8 timeout, u8 tr_method);
 
 void mpt3sas_scsih_set_tm_flag(struct MPT3SAS_ADAPTER *ioc, u16 handle);
 void mpt3sas_scsih_clear_tm_flag(struct MPT3SAS_ADAPTER *ioc, u16 handle);
diff --git a/drivers/scsi/mpt3sas/mpt3sas_ctl.c b/drivers/scsi/mpt3sas/mpt3sas_ctl.c
index 62e5528..98d6b9c 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_ctl.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_ctl.c
@@ -1109,13 +1109,15 @@ _ctl_do_mpt_command(struct MPT3SAS_ADAPTER *ioc, struct mpt3_ioctl_command karg,
 			    pcie_device->device_info))))
 				mpt3sas_scsih_issue_locked_tm(ioc,
 				  le16_to_cpu(mpi_request->FunctionDependent1),
-				  0, MPI2_SCSITASKMGMT_TASKTYPE_TARGET_RESET, 0,
+				  0, 0, 0,
+				  MPI2_SCSITASKMGMT_TASKTYPE_TARGET_RESET, 0,
 				  0, pcie_device->reset_timeout,
 			MPI26_SCSITASKMGMT_MSGFLAGS_PROTOCOL_LVL_RST_PCIE);
 			else
 				mpt3sas_scsih_issue_locked_tm(ioc,
 				  le16_to_cpu(mpi_request->FunctionDependent1),
-				  0, MPI2_SCSITASKMGMT_TASKTYPE_TARGET_RESET, 0,
+				  0, 0, 0,
+				  MPI2_SCSITASKMGMT_TASKTYPE_TARGET_RESET, 0,
 				  0, 30, MPI2_SCSITASKMGMT_MSGFLAGS_LINK_RESET);
 		} else
 			mpt3sas_base_hard_reset_handler(ioc, FORCE_BIG_HAMMER);
diff --git a/drivers/scsi/mpt3sas/mpt3sas_scsih.c b/drivers/scsi/mpt3sas/mpt3sas_scsih.c
index 66b29d4..ce12253 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_scsih.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_scsih.c
@@ -1512,6 +1512,66 @@ _scsih_is_nvme_pciescsi_device(u32 device_info)
 		return 0;
 }
 
+/**
+ * _scsih_scsi_lookup_find_by_target - search for matching channel:id
+ * @ioc: per adapter object
+ * @id: target id
+ * @channel: channel
+ * Context: This function will acquire ioc->scsi_lookup_lock.
+ *
+ * This will search for a matching channel:id in the scsi_lookup array,
+ * returning 1 if found.
+ */
+static u8
+_scsih_scsi_lookup_find_by_target(struct MPT3SAS_ADAPTER *ioc, int id,
+	int channel)
+{
+	int smid;
+	struct scsi_cmnd *scmd;
+
+	for (smid = 1;
+	     smid <= ioc->shost->can_queue; smid++) {
+		scmd = mpt3sas_scsih_scsi_lookup_get(ioc, smid);
+		if (!scmd)
+			continue;
+		if (scmd->device->id == id &&
+		    scmd->device->channel == channel)
+			return 1;
+	}
+	return 0;
+}
+
+/**
+ * _scsih_scsi_lookup_find_by_lun - search for matching channel:id:lun
+ * @ioc: per adapter object
+ * @id: target id
+ * @lun: lun number
+ * @channel: channel
+ * Context: This function will acquire ioc->scsi_lookup_lock.
+ *
+ * This will search for a matching channel:id:lun in the scsi_lookup array,
+ * returning 1 if found.
+ */
+static u8
+_scsih_scsi_lookup_find_by_lun(struct MPT3SAS_ADAPTER *ioc, int id,
+	unsigned int lun, int channel)
+{
+	int smid;
+	struct scsi_cmnd *scmd;
+
+	for (smid = 1; smid <= ioc->shost->can_queue; smid++) {
+
+		scmd = mpt3sas_scsih_scsi_lookup_get(ioc, smid);
+		if (!scmd)
+			continue;
+		if (scmd->device->id == id &&
+		    scmd->device->channel == channel &&
+		    scmd->device->lun == lun)
+			return 1;
+	}
+	return 0;
+}
+
 /**
  * mpt3sas_scsih_scsi_lookup_get - returns scmd entry
  * @ioc: per adapter object
@@ -2704,6 +2764,8 @@ mpt3sas_scsih_clear_tm_flag(struct MPT3SAS_ADAPTER *ioc, u16 handle)
  * mpt3sas_scsih_issue_tm - main routine for sending tm requests
  * @ioc: per adapter struct
  * @handle: device handle
+ * @channel: the channel assigned by the OS
+ * @id: the id assigned by the OS
  * @lun: lun number
  * @type: MPI2_SCSITASKMGMT_TASKTYPE__XXX (defined in mpi2_init.h)
  * @smid_task: smid assigned to the task
@@ -2720,8 +2782,9 @@ mpt3sas_scsih_clear_tm_flag(struct MPT3SAS_ADAPTER *ioc, u16 handle)
  * Return: SUCCESS or FAILED.
  */
 int
-mpt3sas_scsih_issue_tm(struct MPT3SAS_ADAPTER *ioc, u16 handle, u64 lun,
-	u8 type, u16 smid_task, u16 msix_task, u8 timeout, u8 tr_method)
+mpt3sas_scsih_issue_tm(struct MPT3SAS_ADAPTER *ioc, u16 handle, uint channel,
+	uint id, u64 lun, u8 type, u16 smid_task, u16 msix_task,
+	u8 timeout, u8 tr_method)
 {
 	Mpi2SCSITaskManagementRequest_t *mpi_request;
 	Mpi2SCSITaskManagementReply_t *mpi_reply;
@@ -2826,14 +2889,14 @@ out:
 }
 
 int mpt3sas_scsih_issue_locked_tm(struct MPT3SAS_ADAPTER *ioc, u16 handle,
-		u64 lun, u8 type, u16 smid_task, u16 msix_task,
-		u8 timeout, u8 tr_method)
+		uint channel, uint id, u64 lun, u8 type, u16 smid_task,
+		u16 msix_task, u8 timeout, u8 tr_method)
 {
 	int ret;
 
 	mutex_lock(&ioc->tm_cmds.mutex);
-	ret = mpt3sas_scsih_issue_tm(ioc, handle, lun, type, smid_task,
-			msix_task, timeout, tr_method);
+	ret = mpt3sas_scsih_issue_tm(ioc, handle, channel, id, lun, type,
+			smid_task, msix_task, timeout, tr_method);
 	mutex_unlock(&ioc->tm_cmds.mutex);
 
 	return ret;
@@ -2980,7 +3043,8 @@ scsih_abort(struct scsi_cmnd *scmd)
 	if (pcie_device && (!ioc->tm_custom_handling) &&
 	    (!(mpt3sas_scsih_is_pcie_scsi_device(pcie_device->device_info))))
 		timeout = ioc->nvme_abort_timeout;
-	r = mpt3sas_scsih_issue_locked_tm(ioc, handle, scmd->device->lun,
+	r = mpt3sas_scsih_issue_locked_tm(ioc, handle, scmd->device->channel,
+		scmd->device->id, scmd->device->lun,
 		MPI2_SCSITASKMGMT_TASKTYPE_ABORT_TASK,
 		st->smid, st->msix_io, timeout, 0);
 	/* Command must be cleared after abort */
@@ -3056,7 +3120,8 @@ scsih_dev_reset(struct scsi_cmnd *scmd)
 	} else
 		tr_method = MPI2_SCSITASKMGMT_MSGFLAGS_LINK_RESET;
 
-	r = mpt3sas_scsih_issue_locked_tm(ioc, handle, scmd->device->lun,
+	r = mpt3sas_scsih_issue_locked_tm(ioc, handle, scmd->device->channel,
+		scmd->device->id, scmd->device->lun,
 		MPI2_SCSITASKMGMT_TASKTYPE_LOGICAL_UNIT_RESET, 0, 0,
 		tr_timeout, tr_method);
 	/* Check for busy commands after reset */
@@ -3134,7 +3199,8 @@ scsih_target_reset(struct scsi_cmnd *scmd)
 		tr_method = MPI26_SCSITASKMGMT_MSGFLAGS_PROTOCOL_LVL_RST_PCIE;
 	} else
 		tr_method = MPI2_SCSITASKMGMT_MSGFLAGS_LINK_RESET;
-	r = mpt3sas_scsih_issue_locked_tm(ioc, handle, 0,
+	r = mpt3sas_scsih_issue_locked_tm(ioc, handle, scmd->device->channel,
+		scmd->device->id, 0,
 		MPI2_SCSITASKMGMT_TASKTYPE_TARGET_RESET, 0, 0,
 	    tr_timeout, tr_method);
 	/* Check for busy commands after reset */
@@ -7530,7 +7596,7 @@ _scsih_sas_broadcast_primitive_event(struct MPT3SAS_ADAPTER *ioc,
 			goto out;
 
 		spin_unlock_irqrestore(&ioc->scsi_lookup_lock, flags);
-		r = mpt3sas_scsih_issue_tm(ioc, handle, lun,
+		r = mpt3sas_scsih_issue_tm(ioc, handle, 0, 0, lun,
 			MPI2_SCSITASKMGMT_TASKTYPE_QUERY_TASK, st->smid,
 			st->msix_io, 30, 0);
 		if (r == FAILED) {
@@ -7571,9 +7637,9 @@ _scsih_sas_broadcast_primitive_event(struct MPT3SAS_ADAPTER *ioc,
 		if (ioc->shost_recovery)
 			goto out_no_lock;
 
-		r = mpt3sas_scsih_issue_tm(ioc, handle, sdev->lun,
-			MPI2_SCSITASKMGMT_TASKTYPE_ABORT_TASK, st->smid,
-			st->msix_io, 30, 0);
+		r = mpt3sas_scsih_issue_tm(ioc, handle, sdev->channel, sdev->id,
+			sdev->lun, MPI2_SCSITASKMGMT_TASKTYPE_ABORT_TASK,
+			st->smid, st->msix_io, 30, 0);
 		if (r == FAILED || st->cb_idx != 0xFF) {
 			sdev_printk(KERN_WARNING, sdev,
 			    "mpt3sas_scsih_issue_tm: ABORT_TASK: FAILED : "
-- 
2.26.2

[PATCH 6/7] mpt3sas: Postprocessing of target and LUN reset.

[Suganath Prabu S]

If driver has not received the interrupt for the aborted SCSI command
before processing the TM reply, Driver polls all the reply descriptor
pools looking for the reply for the aborted SCSI command before marking
TM as FAILED. If it finds the reply then mark the TM as SUCCESS
otherwise mark as FAILED.
scsih_tm_cmd_map_status() -> Checks whether TM has aborted
the timed out SCSI command or not, if TM has aborted the IO
then returns SUCCESS else returns FAILED.
scsih_tm_post_processing() -> Post proceesing of target &
LUN reset

Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
---
 drivers/scsi/mpt3sas/mpt3sas_base.c  |   6 +-
 drivers/scsi/mpt3sas/mpt3sas_base.h  |   2 +-
 drivers/scsi/mpt3sas/mpt3sas_scsih.c | 136 ++++++++++++++++++++++++++-
 3 files changed, 139 insertions(+), 5 deletions(-)

diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c
index 2ffdbbb..3606676 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.c
@@ -1784,12 +1784,14 @@ _base_is_controller_msix_enabled(struct MPT3SAS_ADAPTER *ioc)
 /**
  * mpt3sas_base_sync_reply_irqs - flush pending MSIX interrupts
  * @ioc: per adapter object
+ * @poll: poll over reply descriptor pools incase interrupt for
+ *		timed-out SCSI command got delayed
  * Context: non ISR conext
  *
  * Called when a Task Management request has completed.
  */
 void
-mpt3sas_base_sync_reply_irqs(struct MPT3SAS_ADAPTER *ioc)
+mpt3sas_base_sync_reply_irqs(struct MPT3SAS_ADAPTER *ioc, u8 poll)
 {
 	struct adapter_reply_queue *reply_q;
 
@@ -1819,6 +1821,8 @@ mpt3sas_base_sync_reply_irqs(struct MPT3SAS_ADAPTER *ioc)
 		}
 		synchronize_irq(pci_irq_vector(ioc->pdev, reply_q->msix_index));
 	}
+	if (poll)
+		_base_process_reply_queue(reply_q);
 }
 
 /**
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.h b/drivers/scsi/mpt3sas/mpt3sas_base.h
index fd61deb..feb8328 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.h
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.h
@@ -1528,7 +1528,7 @@ __le32 mpt3sas_base_get_sense_buffer_dma(struct MPT3SAS_ADAPTER *ioc,
 	u16 smid);
 void *mpt3sas_base_get_pcie_sgl(struct MPT3SAS_ADAPTER *ioc, u16 smid);
 dma_addr_t mpt3sas_base_get_pcie_sgl_dma(struct MPT3SAS_ADAPTER *ioc, u16 smid);
-void mpt3sas_base_sync_reply_irqs(struct MPT3SAS_ADAPTER *ioc);
+void mpt3sas_base_sync_reply_irqs(struct MPT3SAS_ADAPTER *ioc, u8 poll);
 void mpt3sas_base_mask_interrupts(struct MPT3SAS_ADAPTER *ioc);
 void mpt3sas_base_unmask_interrupts(struct MPT3SAS_ADAPTER *ioc);
 
diff --git a/drivers/scsi/mpt3sas/mpt3sas_scsih.c b/drivers/scsi/mpt3sas/mpt3sas_scsih.c
index ce12253..67270e0 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_scsih.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_scsih.c
@@ -2760,6 +2760,96 @@ mpt3sas_scsih_clear_tm_flag(struct MPT3SAS_ADAPTER *ioc, u16 handle)
 	}
 }
 
+/**
+ * scsih_tm_cmd_map_status - map the target reset & LUN reset TM status
+ * @ioc - per adapter object
+ * @channel - the channel assigned by the OS
+ * @id: the id assigned by the OS
+ * @lun: lun number
+ * @type: MPI2_SCSITASKMGMT_TASKTYPE__XXX (defined in mpi2_init.h)
+ * @smid_task: smid assigned to the task
+ *
+ * Look whether TM has aborted the timed out SCSI command, if
+ * TM has aborted the IO then return SUCCESS else return FAILED.
+ */
+int
+scsih_tm_cmd_map_status(struct MPT3SAS_ADAPTER *ioc, uint channel,
+	uint id, uint lun, u8 type, u16 smid_task)
+{
+
+	if (smid_task <= ioc->shost->can_queue) {
+		switch (type) {
+		case MPI2_SCSITASKMGMT_TASKTYPE_TARGET_RESET:
+			if (!(_scsih_scsi_lookup_find_by_target(ioc,
+			    id, channel)))
+				return SUCCESS;
+			break;
+		case MPI2_SCSITASKMGMT_TASKTYPE_ABRT_TASK_SET:
+		case MPI2_SCSITASKMGMT_TASKTYPE_LOGICAL_UNIT_RESET:
+			if (!(_scsih_scsi_lookup_find_by_lun(ioc, id,
+			    lun, channel)))
+				return SUCCESS;
+			break;
+		default:
+			return SUCCESS;
+		}
+	} else if (smid_task == ioc->scsih_cmds.smid) {
+		if ((ioc->scsih_cmds.status & MPT3_CMD_COMPLETE) ||
+		    (ioc->scsih_cmds.status & MPT3_CMD_NOT_USED))
+			return SUCCESS;
+	} else if (smid_task == ioc->ctl_cmds.smid) {
+		if ((ioc->ctl_cmds.status & MPT3_CMD_COMPLETE) ||
+		    (ioc->ctl_cmds.status & MPT3_CMD_NOT_USED))
+			return SUCCESS;
+	}
+
+	return FAILED;
+}
+
+/**
+ * scsih_tm_post_processing - post proceesing of target & LUN reset
+ * @ioc - per adapter object
+ * @handle: device handle
+ * @channel - the channel assigned by the OS
+ * @id: the id assigned by the OS
+ * @lun: lun number
+ * @type: MPI2_SCSITASKMGMT_TASKTYPE__XXX (defined in mpi2_init.h)
+ * @smid_task: smid assigned to the task
+ *
+ * Post processing of target & LUN reset. Due to interrupt latency
+ * issue it possible that interrupt for aborted IO might not
+ * received yet. So before returning failure status poll the
+ * reply descriptor pools for the reply of timed out SCSI command.
+ * Return FAILED status if reply for timed out is not received
+ * otherwise return SUCCESS.
+ */
+int
+scsih_tm_post_processing(struct MPT3SAS_ADAPTER *ioc, u16 handle,
+	uint channel, uint id, uint lun, u8 type, u16 smid_task)
+{
+	int rc;
+
+	rc = scsih_tm_cmd_map_status(ioc, channel, id, lun, type, smid_task);
+	if (rc == SUCCESS)
+		return rc;
+
+	ioc_info(ioc,
+	    "Poll ReplyDescriptor queues for completion of"
+	    " smid(%d), task_type(0x%02x), handle(0x%04x)\n",
+	    smid_task, type, handle);
+
+	/*
+	 * Due to interrupt latency issues, driver may receive interrupt for
+	 * TM first and then for aborted SCSI IO command. So, poll all the
+	 * ReplyDescriptor pools before returning the FAILED status to SML.
+	 */
+	mpt3sas_base_mask_interrupts(ioc);
+	mpt3sas_base_sync_reply_irqs(ioc, 1);
+	mpt3sas_base_unmask_interrupts(ioc);
+
+	return scsih_tm_cmd_map_status(ioc, channel, id, lun, type, smid_task);
+}
+
 /**
  * mpt3sas_scsih_issue_tm - main routine for sending tm requests
  * @ioc: per adapter struct
@@ -2788,6 +2878,7 @@ mpt3sas_scsih_issue_tm(struct MPT3SAS_ADAPTER *ioc, u16 handle, uint channel,
 {
 	Mpi2SCSITaskManagementRequest_t *mpi_request;
 	Mpi2SCSITaskManagementReply_t *mpi_reply;
+	Mpi25SCSIIORequest_t *request;
 	u16 smid = 0;
 	u32 ioc_state;
 	int rc;
@@ -2843,7 +2934,9 @@ mpt3sas_scsih_issue_tm(struct MPT3SAS_ADAPTER *ioc, u16 handle, uint channel,
 	mpi_request->Function = MPI2_FUNCTION_SCSI_TASK_MGMT;
 	mpi_request->DevHandle = cpu_to_le16(handle);
 	mpi_request->TaskType = type;
-	mpi_request->MsgFlags = tr_method;
+	if (type == MPI2_SCSITASKMGMT_TASKTYPE_ABORT_TASK ||
+	    type == MPI2_SCSITASKMGMT_TASKTYPE_QUERY_TASK)
+		mpi_request->MsgFlags = tr_method;
 	mpi_request->TaskMID = cpu_to_le16(smid_task);
 	int_to_scsilun(lun, (struct scsi_lun *)mpi_request->LUN);
 	mpt3sas_scsih_set_tm_flag(ioc, handle);
@@ -2863,7 +2956,7 @@ mpt3sas_scsih_issue_tm(struct MPT3SAS_ADAPTER *ioc, u16 handle, uint channel,
 	}
 
 	/* sync IRQs in case those were busy during flush. */
-	mpt3sas_base_sync_reply_irqs(ioc);
+	mpt3sas_base_sync_reply_irqs(ioc, 0);
 
 	if (ioc->tm_cmds.status & MPT3_CMD_REPLY_VALID) {
 		mpt3sas_trigger_master(ioc, MASTER_TRIGGER_TASK_MANAGMENT);
@@ -2880,7 +2973,44 @@ mpt3sas_scsih_issue_tm(struct MPT3SAS_ADAPTER *ioc, u16 handle, uint channel,
 				    sizeof(Mpi2SCSITaskManagementRequest_t)/4);
 		}
 	}
-	rc = SUCCESS;
+
+	switch (type) {
+	case MPI2_SCSITASKMGMT_TASKTYPE_ABORT_TASK:
+		rc = SUCCESS;
+		/*
+		 * If DevHandle filed in smid_task's entry of request pool
+		 * doesn't matches with device handle on which this task abort
+		 * TM is received then it means that TM has successfully
+		 * aborted the timed out command. Since smid_task's entry in
+		 * request pool will be memset to zero once the timed out
+		 * command is returned to the SML. If the command is not
+		 * aborted then smid_task’s entry won’t be cleared and it
+		 * will have same DevHandle value on which this task abort TM
+		 * is received and driver will return the TM status as FAILED.
+		 */
+		request = mpt3sas_base_get_msg_frame(ioc, smid_task);
+		if (le16_to_cpu(request->DevHandle) != handle)
+			break;
+
+		ioc_info(ioc, "Task abort tm failed: handle(0x%04x),"
+		    "timeout(%d) tr_method(0x%x) smid(%d) msix_index(%d)\n",
+		    handle, timeout, tr_method, smid_task, msix_task);
+		rc = FAILED;
+		break;
+
+	case MPI2_SCSITASKMGMT_TASKTYPE_TARGET_RESET:
+	case MPI2_SCSITASKMGMT_TASKTYPE_ABRT_TASK_SET:
+	case MPI2_SCSITASKMGMT_TASKTYPE_LOGICAL_UNIT_RESET:
+		rc = scsih_tm_post_processing(ioc, handle, channel, id, lun,
+		    type, smid_task);
+		break;
+	case MPI2_SCSITASKMGMT_TASKTYPE_QUERY_TASK:
+		rc = SUCCESS;
+		break;
+	default:
+		rc = FAILED;
+		break;
+	}
 
 out:
 	mpt3sas_scsih_clear_tm_flag(ioc, handle);
-- 
2.26.2

[PATCH 7/7] mpt3sas: Update driver version to 35.100.00.00

[Suganath Prabu S]

Updated driver version to 35.100.00.00

Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
---
 drivers/scsi/mpt3sas/compile.sh     | 6 +++---
 drivers/scsi/mpt3sas/mpt3sas_base.h | 4 ++--
 2 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/scsi/mpt3sas/compile.sh b/drivers/scsi/mpt3sas/compile.sh
index f0bcf6b..d8ac38f 100755
--- a/drivers/scsi/mpt3sas/compile.sh
+++ b/drivers/scsi/mpt3sas/compile.sh
@@ -48,9 +48,9 @@ fi
 # Set SPARSE to non-zero value to enable sparse checking
 #	kernel="2.6.18-8.el5"
 #	kernel="2.6.18-53.el5"
-#	kernel="2.6.27.15-2-default"
-	kernel=`uname -r`
-SPARSE=0
+	kernel="5.8.0-rc1+"
+#	kernel=`uname -r`
+SPARSE=1
 
 blacklist_enclosure
 rm -fr output.log
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.h b/drivers/scsi/mpt3sas/mpt3sas_base.h
index feb8328..b11459a 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.h
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.h
@@ -76,8 +76,8 @@
 #define MPT3SAS_DRIVER_NAME		"mpt3sas"
 #define MPT3SAS_AUTHOR "Avago Technologies <MPT-FusionLinux.pdl@avagotech.com>"
 #define MPT3SAS_DESCRIPTION	"LSI MPT Fusion SAS 3.0 Device Driver"
-#define MPT3SAS_DRIVER_VERSION		"34.100.00.00"
-#define MPT3SAS_MAJOR_VERSION		34
+#define MPT3SAS_DRIVER_VERSION		"35.100.00.00"
+#define MPT3SAS_MAJOR_VERSION		35
 #define MPT3SAS_MINOR_VERSION		100
 #define MPT3SAS_BUILD_VERSION		0
 #define MPT3SAS_RELEASE_VERSION	00
-- 
2.26.2

Re: [PATCH 0/7] mpt3sas: Enhancements and bug fixes

[Martin K. Petersen]

Suganath,

> This patch set has below bug fixes and enhancements.

Applied to my staging tree. You'll get a formal merge message once 5.10
opens.

A few asks for the next submission:

 - Please revisit Documentation/process/submitting-patches.rst section 2
   about how to write commit messages.

 - Please don't include outbox driver files in the submitted patches.

 - Please build with C=1/W=1. Lots of warnings in this code.

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: [PATCH 0/7] mpt3sas: Enhancements and bug fixes

[Martin K. Petersen]

On Thu, 30 Jul 2020 13:33:42 +0530, Suganath Prabu S wrote:

> This patch set has below bug fixes and enhancements.
> 
> Dump IOC system interface register set when IOC fails to
> transit to READY state during initialization.
> 
> During controller reset, cancel the running work along
> with the queued works. This avoids processing the expired event after
> the controller reset.
> 
> [...]

Applied to 5.10/scsi-queue, thanks!

[1/7] scsi: mpt3sas: Memset config_cmds.reply buffer with zeros
      https://git.kernel.org/mkp/scsi/c/01348b76fa82
[2/7] scsi: mpt3sas: Dump system registers for debugging
      https://git.kernel.org/mkp/scsi/c/7f3cca0f1f54
[3/7] scsi: mpt3sas: Cancel the running work during host reset
      https://git.kernel.org/mkp/scsi/c/b5f025c3f702
[4/7] scsi: mpt3sas: Rename and export interrupt mask/unmask functions
      https://git.kernel.org/mkp/scsi/c/87a1cbcfe55b
[5/7] scsi: mpt3sas: Add functions to check if any cmd is outstanding on Target and LUN
      https://git.kernel.org/mkp/scsi/c/844b4f05661e
[6/7] scsi: mpt3sas: Postprocessing of target and LUN reset
      https://git.kernel.org/mkp/scsi/c/caa596babea7
[7/7] scsi: mpt3sas: Update driver version to 35.100.00.00
      https://git.kernel.org/mkp/scsi/c/d2101b4ee8f8

-- 
Martin K. Petersen	Oracle Linux Engineering