The logic for sg_seg_cnt is a bit convoluted. This patch tries to clean
up a couple of areas, especially around the +2 and +1 logic.

This patch:

- Cleans up the lpfc_sg_seg_cnt attribute to specify a real minimum
  rather than making the minimum be whatever the default is.

- Removes the hardcoding of +2 (for the number of elements we use in a
  sgl for cmd iu and rsp iu) and +1 (an additional entry to compensate
  for nvme's reduction of io size based on a possible partial page)
  logic in sg list initialization. In the case where the +1 logic is
  referenced in host and target io checks, use the values set in the
  transport template as that value was properly set.

There can certainly be more done in this area and it will be addressed
in combined host/target driver effort.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

During driver unload, the driver may crash due to NULL pointers.  The
NULL pointers were due to the driver not protecting itself sufficiently
during some of the teardown paths.  Additionally, the driver was not
waiting for and cleanup up nvme io resources. As such, the driver wasn't
making the callbacks to the transport, stalling the transports
association teardown.

This patch waits for io clean up before tearding down and adds checks
for possible NULL pointers.

Cc: <stable@vger.kernel.org> # 4.12+
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

When the driver is unloading, the nvme transport could be in the process
of submitting new requests, will send abort requests to terminate
associations, or may make LS-related requests.  The driver's abort and
request entry points currently is ignorant of the unloading state and is
starting the requests even though the infrastructure to complete them
continues to teardown.

Change the entry points for new requests to check whether unloading and
if so, reject the requests. Abort routines check unloading, and if so,
noop the request. An abort is noop'd as the teardown paths are already
aborting/terminating the io outstanding at the time the teardown
initiated.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The driver's interaction with the host nvme transport has been incorrect
for a while. The driver did not wait for the unregister callbacks
(waited only 5 jiffies). Thus the driver may remove objects that may be
referenced by subsequent abort commands from the transport, and the
actual unregister callback was effectively a noop. This was especially
problematic if the driver was unloaded.

The driver now waits for the unregister callbacks, as it should, before
continuing with teardown.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The driver currently registers any remote port that has NVME support.
It should only be registering target ports.

Register only target ports.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

During RSCN storms, the driver does not rediscover some targets.  The
driver marks some RSCN as to be handled after the ones it's working
on. The driver missed processing some deferred RSCN.

Move where the driver checks for deferred RSCNs and initiate deferred
RSCN handling if the flag was set. Also revise nport state within the
RSCN confirm routine. Add some state data to a possible debug print to
aid future debugging.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

pt2pt ndlp ref count prematurely goes to 0. There was reference removed
that should only be removed if connected to a switch, not if in
point-to-point mode.

Add a mode check before the reference remove.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The current default for async hw receive queues is 1, which presents
issues under heavy load as number of queues influence the available
async receive buffer limits.

Raise the default to the either the current hw limit (16) or the number
of hw qs configured (io channel value).

Revise the attribute definition for mrq to better reflect what we do for
hw queues. E.g. 0 means default to optimal (# of cpus), non-zero
specifies a specific limit. Before this change, mrq=0 meant target mode
was disabled. As 0 now has a different meaning, rework the if tests to
use the better nvmet_support check.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Display for lpfc/fnX/iDiag/queInfo isn't formatted perfectly.  Corrected
the format strings for the queue info debug messages.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The driver does not respond to PLOGI from the direct attach target.  The
driver uses incorrect S_ID in CONFIG_LINK, after FLOGI completion

Correct by issuing CONFIG_LINK with the correct S_ID after receiving the
PLOGI from the target
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Raise the maximum NVME sg list size allowed to 256 elements.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Performing an LS abort results in the following message being seen:
  0603 Invalid CQ subtype 6: 00000300 22000002 ffff0016 d0050000
and the associated exchange is not properly freed.

The code did not recognize the exchange type that was aborted, thus it
was not properly handled.

Correct by adding the NVME LS ELS type to the exchange types that are
recognized.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

In test cases where an instance of the driver is detached and
reattached, the driver will crash on reattachment. There is a compound
if statement that will skip over the bar setup if the pci_resource_start
call is not successful. The driver erroneously returns success to its
bar setup in this scenario even though the bars aren't properly
configured.

Rework the offending code segment for proper initialization steps.  If
the pci_resource_start call fails, -ENOMEM is now returned.

Sample stack:

rport-5:0-10: blocked FC remote port time out: removing rport
BUG: unable to handle kernel NULL pointer dereference at           (null)
... lpfc_sli4_wait_bmbx_ready+0x32/0x70 [lpfc]
...
...  RIP: 0010:...  ... lpfc_sli4_wait_bmbx_ready+0x32/0x70 [lpfc]
 Call Trace:
  ... lpfc_sli4_post_sync_mbox+0x106/0x4d0 [lpfc]
  ... ? __alloc_pages_nodemask+0x176/0x420
  ... ? __kmalloc+0x2e/0x230
  ... lpfc_sli_issue_mbox_s4+0x533/0x720 [lpfc]
  ... ? mempool_alloc+0x69/0x170
  ... ? dma_generic_alloc_coherent+0x8f/0x140
  ... lpfc_sli_issue_mbox+0xf/0x20 [lpfc]
  ... lpfc_sli4_driver_resource_setup+0xa6f/0x1130 [lpfc]
  ... ? lpfc_pci_probe_one+0x23e/0x16f0 [lpfc]
  ... lpfc_pci_probe_one+0x445/0x16f0 [lpfc]
  ... local_pci_probe+0x45/0xa0
  ... work_for_cpu_fn+0x14/0x20
  ... process_one_work+0x17a/0x440

Cc: <stable@vger.kernel.org> # 4.12+
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

XRI_ABORTED_CQE completions were not being handled in the fast path.
They were being queued and deferred to the lpfc worker thread for
processing. This is an artifact of the driver design prior to moving
queue processing out of the isr and into a workq element. Now that queue
processing is already in a deferred context, remove this artifact and
process them directly.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Hardware queues are a fast staging area to push commands into the
adapter.  The adapter should drain them extremely quickly. However,
under heavy io load, the host cpu is pushing commands faster than the
drain rate of the adapter causing the driver to resource busy commands.

Enlarge the hardware queue (wq & cq) to support a larger number of queue
entries (4x the prior size) before backpressure. Enlarging the queue
requires larger contiguous buffers (16k) per logical page for the
hardware. This changed calling sequences that were expecting 4K page
sizes that now must pass a parameter with the page sizes. It also
required use of a new version of an adapter command that can vary the
page size values.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

When the HBA is connected to a private loop, the driver reports FLOGI
loop-open failure as functional error. This is an expected condition.

Mark loop-open failure as a warning instead of error.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

This patch calls the new nvme transport routine for dev_loss_tmo
whenever the SCSI fc transport calls the lldd to make a dynamic
change to a remote ports dev_loss_tmo.
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

System crashed due to a hard lockup at lpfc_els_timeout_handler+0x128.

The els ring's txcmplq list is corrupted: the last element in the list
does not point back the the head causing a loop. Issue is the els
processing path for sli4 hbas are using the hbalock instead of the
ring_lock for removing elements from the txcmplq list.

Use the adapter SLI_REV to determine which lock should be used for
removing iocbqs from the els rings txcmplq.

note: the future refactoring will address this so that we don't have
this ugly type-based lock code.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NEwan D. Milne <emilne@redhat.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

drivers/scsi/lpfc/lpfc_debugfs.c:5460:22-29: WARNING: kzalloc should be used for phba -> nvmeio_trc, instead of kmalloc/memset
drivers/scsi/lpfc/lpfc_debugfs.c:2230:20-27: WARNING: kzalloc should be used for phba -> nvmeio_trc, instead of kmalloc/memset

 Use kzalloc rather than kmalloc followed by memset with 0

Generated by: scripts/coccinelle/api/alloc/kzalloc-simple.cocci
Signed-off-by: NVasyl Gomonovych <gomonovych@gmail.com>
Acked-by: NJames Smart <james.smart@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.

Cc: James Smart <james.smart@broadcom.com>
Cc: Dick Kennedy <dick.kennedy@broadcom.com>
Cc: "James E.J. Bottomley" <jejb@linux.vnet.ibm.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
Cc: linux-scsi@vger.kernel.org
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>

The ! has higher precedence than the & operation.  I've added
parenthesis so this works as intended.

Fixes: 952c303b ("scsi: lpfc: Ensure io aborts interlocked with the target.")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Change version to 11.4.0.4
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The internal cfg flag is actually smaller, by 1 (for a partial page
sge), than the sg list maintained by the driver. Thus the check on sg
segments errored out when it shouldn't have

Ensure the check is +1

Note: having a value that is less than what it really is is bogus.
Correcting it now would be a significant rework. Add this item to the
list to be refactored in the merge with efct.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

When running NVME io as a NVME host, if the driver is unloaded there
would be oops in lpfc_sli4_issue_wqe.

When unloading, controllers are torn down and the transport initiates
set_property commands to reset the controller and issues aborts to
terminate existing io.  The drivers nvme abort and fcp io submit
routines needed to recognize the driver is unloading and fail the new
requests. It didn't, resulting in the oops.

Revise the ls and fcp io submit routines to detect the unloading state
and properly handle their cleanup.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Support RDP and Multiple Frames

If the remote Nport is not logged in, the driver would not populate all
the descriptors in the RDP response payload. Doing so would create a
payload length that requires multiple frames due to exceeding the
default rx buffer size without an explicit login. Currently FC-LS
explicitly states the RDP response must be a single frame sequence.
Thus we did not violate the standard.

Recently, a modification to FC-LS was accepted which allows multi-frame
sequences and all vendors have indicated they are interoperable with the
change. As such, extend RDP support with the additional fields and send
a multi-frame sequence.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Before releasing nvme io back to the io stack for possible retry on
other paths, ensure the io termination is interlocked with the target
device by ensuring the entire ABTS-LS protocol is complete.

Additionally, FC-NVME ABTS-LS protocol does not use RRQ. Remove RRQ
behavior from ABTS-LS.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Firmware update fails with: status x17 add_status x56 on the final write

If multiple DMA buffers are used for the download, some firmware revs
have difficulty with signatures and crcs split across the dma buffer
boundaries.  Resolve by making all writes be a single 4k page in length.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The driver is seeing a NULL pointer in lpfc_nvme_fcp_io_submit.  This
was ultimately due to a transport AER being sent on a terminated
controller, thus some of the values were not set. In case we're in a
system without a corrected transport and in case a race condition occurs
where we enter the routine as the teardown is happening in a separate
thread, validate the parameters before starting the io.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The initial implementation of NVME didn't merge with NPIV support.  As
such, there are several issues if NPIV is used with NVME. For now,
ensure that if NVME is enabled then NPIV is not enabled.

Support for NPIV with NVME will be added in the near future.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

if nvmet targetport registration fails, the driver encounters a NULL
pointer oops in lpfc_hb_timeout_handler.

To fix: if registration fails, ensure nvmet_support is cleared on the
port structure.

Also enhanced the log message on failure.

Cc: <stable@vger.kernel.org> # 4.12+
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The descriptions for lpfc_xri_split and lpfc_enable_fc4_type were
poor. Revise for better understanding:

  lpfc_xri_split - Percentage of FCP XRI resources versus NVME
  lpfc_enable_fc4_type - Enable FC4 Protocol support - FCP / NVME
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Always set ctxp->state to LPFC_NVMET_STE_ABORT if ABORT op gets called
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

There are several log messages that report abnormal terminations that by
default are marked warn. These are typically the result of failures due
to invalid controller state or abort completions. They are all natural
when a controller resets.

Unfortunately, as they are logged by default, it makes the admin very
concerned.

Convert the messages to Info.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The driver is encountering  oops in lpfc_sli_calc_ring.

The driver is setting hba_wqidx for FCP based on the policy in use for
NVME. The two may not be the same.  Change to set the wqidx based on the
FCP policy.

Cc: <stable@vger.kernel.org> # 4.12+
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Under heavy target nvme load duration, the lpfc irq handler is
encountering cpu lockup warnings.

Convert the driver to a shortened ISR handler which identifies the
interrupting condition then schedules a workq thread to process the
completion queue the interrupt was for. This moves all the real work
into the workq element.

As nvmet_fc upcalls are no longer in ISR context, don't set the feature
flags
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Need to make ktime samples more accurate

If ktime is turned on in the middle of an IO, the max calculation could
be misleading. Base sampling on the start time of the IO as opposed to
ktime_on.

Make ISR ktime timestamps be from when CQE is read instead of EQE.
Added additional sanity checks when deciding whether to accept an IO
sample or not.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Local Reject/Invalid RPI errors seen during discovery.

Temporary RPI cleanup was occurring regardless of SLI rev. It's only
necessary on SLI-4.

Adjust the test for whether cleanup is necessary.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Warning messages when NVME_TARGET_FC not defined on ppc builds

The lpfc_nvmet_replenish_context() function is only meaningful when NVME
target mode enabled. Surround the function body with ifdefs for target
mode enablement.
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reported-by: NStephen Rothwell <sfr@canb.auug.org.au>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

In a link bounce scenario, a condition can occur where the discovery
engine swaps an ndlp structure (address change for an nport). While the
swap was successfully executed by the discovery engine, the driver did
not properly detect a change in the ndlp bound to the nvme rport.  This
error resulted in the nvme host transport issuing an IO to the correct
nvme rport, but the lpfc driver addressed a ndlp with an NLP_UNUSED
status and failed the io. This resulting it it looking like there were
missing namespaces and applications failed due to io errors.

To fix, in lpfc_nvme_register_rport, rework the "rebind" case to break
the nvme rport<->ndlp association when the ndlp already has an
nrport. Then rebind the rport to the correct ndlp data and backpointers.

[mkp: typo]
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The driver crashes when attempting to use a freed ndpl pointer.

The pci_remove_one handler runs on a separate kernel thread. The order
of the removal is starting by freeing all of the ndlps and then
disabling interrupts. In between these two events the driver can still
receive an ELS and process it. When it tries to use the ndlp pointer
will be NULL

Change the order of the pci_remove_one vs disable interrupts so that
interrupts are disabled before the ndlp's are freed.

Cc: <stable@vger.kernel.org> # 4.12+
Signed-off-by: NDick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: NJames Smart <james.smart@broadcom.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>