scsi: qedf: Remove redundant variable op

The variable 'op' is assigned a value and is never read. The variable is
not used and is redundant, remove it.

Link: https://lore.kernel.org/r/20220517092518.93159-1-colin.i.king@gmail.com
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hisi_sas: Fix memory ordering in hisi_sas_task_deliver()

The memories for the slot should be observed to be written prior to
observing the slot as ready.

Prior to commit 26fc0ea74fcb ("scsi: libsas: Drop SAS_TASK_AT_INITIATOR"),
we had a spin_lock() + spin_unlock() immediately before marking the slot as
ready. The spin_unlock() - with release semantics - caused the slot memory
to be observed to be written.

Now that the spin_lock() + spin_unlock() is gone, use a smp_wmb().

Link: https://lore.kernel.org/r/1652774661-12935-1-git-send-email-john.garry@huawei.com
Fixes: 26fc0ea74fcb ("scsi: libsas: Drop SAS_TASK_AT_INITIATOR")
Reported-by: Yihang Li <liyihang6@hisilicon.com>
Tested-by: Yihang Li <liyihang6@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: fnic: Replace DMA mask of 64 bits with 47 bits

Cisco VIC supports only 47 bits.  If the host sends DMA addresses that are
greater than 47 bits, it causes work queue (WQ) errors in the VIC.

Link: https://lore.kernel.org/r/20220513205605.81788-1-kartilak@cisco.com
Tested-by: Karan Tilak Kumar <kartilak@cisco.com>
Co-developed-by: Dhanraj Jhawar <djhawar@cisco.com>
Signed-off-by: Dhanraj Jhawar <djhawar@cisco.com>
Co-developed-by: Sesidhar Baddela <sebaddel@cisco.com>
Signed-off-by: Sesidhar Baddela <sebaddel@cisco.com>
Signed-off-by: Arulprabhu Ponnusamy <arulponn@cisco.com>
Signed-off-by: Karan Tilak Kumar <kartilak@cisco.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Add target device related sysfs attributes

Add sysfs attributes for exposing target device details such as SAS
address, firmware device handle, and persistent ID for the
controller-attached devices and RAID volumes.

Link: https://lore.kernel.org/r/20220517115310.13062-3-sreekanth.reddy@broadcom.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Add shost related sysfs attributes

Add shost related sysfs attributes to display the controller's firmware
version, queue depth, number of requests, and number of reply queues. Also
add an attribute to set & get the logging_level.

Link: https://lore.kernel.org/r/20220517115310.13062-2-sreekanth.reddy@broadcom.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: elx: efct: Remove redundant memset() statement

As memset() of bmbx is immediately followed by a memcpy() where bmbx is the
destination, the memset() is redundant.

Link: https://lore.kernel.org/r/20220505143703.45441-1-harshit.m.mogalapalli@oracle.com
Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: Remove redundant memset() statement

As memset() of scmd->sense_buffer is immediately followed by a memcpy()
where scmd->sense_buffer is the destination. The memset() is redundant.

Link: https://lore.kernel.org/r/20220505143214.44908-1-harshit.m.mogalapalli@oracle.com
Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Return error if dma_alloc_coherent() fails

Return -ENOMEM instead of success if dma_alloc_coherent() fails.

Link: https://lore.kernel.org/r/YnOmMGHqCOtUCYQ1@kili
Fixes: 43ca11005098 ("scsi: mpi3mr: Add support for PEL commands")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hisi_sas: Fix rescan after deleting a disk

Removing an ATA device via sysfs means that the device may not be found
through re-scanning:

root@ubuntu:/home/john# lsscsi
[0:0:0:0] disk SanDisk LT0200MO P404 /dev/sda
[0:0:1:0] disk ATA HGST HUS724040AL A8B0 /dev/sdb
[0:0:8:0] enclosu 12G SAS Expander RevB -
root@ubuntu:/home/john# echo 1 > /sys/block/sdb/device/delete
root@ubuntu:/home/john# echo "- - -" > /sys/class/scsi_host/host0/scan
root@ubuntu:/home/john# lsscsi
[0:0:0:0] disk SanDisk LT0200MO P404 /dev/sda
[0:0:8:0] enclosu 12G SAS Expander RevB -
root@ubuntu:/home/john#

The problem is that the rescan of the device may conflict with the device
in being re-initialized, as follows:

 - In the rescan we call hisi_sas_slave_alloc() in store_scan() ->
   sas_user_scan() -> [__]scsi_scan_target() -> scsi_probe_and_add_lunc()
   -> scsi_alloc_sdev() -> hisi_sas_slave_alloc() -> hisi_sas_init_device()
   In hisi_sas_init_device() we issue an IT nexus reset for ATA devices

 - That IT nexus causes the remote PHY to go down and this triggers a bcast
   event

 - In parallel libsas processes the bcast event, finds that the phy is down
   and marks the device as gone

The hard reset issued in hisi_sas_init_device() is unncessary - as
described in the code comment - so remove it. Also set dev status as
HISI_SAS_DEV_NORMAL as the hisi_sas_init_device() call.

Link: https://lore.kernel.org/r/1652354134-171343-4-git-send-email-john.garry@huawei.com
Fixes: 36c6b7613ef1 ("scsi: hisi_sas: Initialise devices in .slave_alloc callback")
Tested-by: Yihang Li <liyihang6@hisilicon.com>
Reviewed-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hisi_sas: Use sas_ata_wait_after_reset() in IT nexus reset

We have seen errors like this when a SATA device is probed:

[524.566298] hisi_sas_v3_hw 0000L74:02.0: erroneous completion iptt=4096 ...
[524.582827] sas: TMF task open reject failed 500e004aaaaaaaa00

Since commit 21c7e972475e ("scsi: hisi_sas: Disable SATA disk phy for
severe I_T nexus reset failure"), we issue an ATA softreset to disks after
a phy reset to ensure that they are in sound working order. If the
softreset is issued before the remote phy has come back up then the
softreset will fail (errors as above). Remedy this by waiting for the phy
to come back up after the reset.

Link: https://lore.kernel.org/r/1652354134-171343-3-git-send-email-john.garry@huawei.com
Tested-by: Yihang Li <liyihang6@hisilicon.com>
Reviewed-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Refactor sas_ata_hard_reset()

Create function sas_ata_wait_after_reset() from sas_ata_hard_reset() as
some LLDDs may want to check for a remote ATA phy is up after reset.

Link: https://lore.kernel.org/r/1652354134-171343-2-git-send-email-john.garry@huawei.com
Tested-by: Yihang Li <liyihang6@hisilicon.com>
Reviewed-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpt3sas: Update driver version to 42.100.00.00

Update driver version to 42.100.00.00.

Link: https://lore.kernel.org/r/20220511072621.30657-2-sreekanth.reddy@broadcom.com
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpt3sas: Fix junk chars displayed while printing ChipName

Terminate string after copying 16 bytes of ChipName data from Manufacturing
Page0 to prevent %s from printing junk characters.

Link: https://lore.kernel.org/r/20220511072621.30657-1-sreekanth.reddy@broadcom.com
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ipr: Use kobj_to_dev()

Use kobj_to_dev() instead of open-coding it.

Link: https://lore.kernel.org/r/20220510105113.1351891-1-chi.minghao@zte.com.cn
Reported-by: Zeal Robot <zealci@zte.com.cn>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Minghao Chi <chi.minghao@zte.com.cn>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Fix a NULL vs IS_ERR() bug in mpi3mr_bsg_init()

The bsg_setup_queue() function does not return NULL.  It returns error
pointers.  Fix the check accordingly.

Link: https://lore.kernel.org/r/YnUf7RQl+A3tigWh@kili
Fixes: 4268fa751365 ("scsi: mpi3mr: Add bsg device support")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: bnx2fc: Avoid using get_cpu() in bnx2fc_cmd_alloc()

Using get_cpu() leads to disabling preemption and in this context it is not
possible to acquire the following spinlock_t on PREEMPT_RT because it
becomes a sleeping lock.

Commit 0ea5c27583e1 ("[SCSI] bnx2fc: common free list for cleanup
commands") says that it is using get_cpu() as a fix in case the CPU is
preempted.  While this might be true, the important part is that it is now
using the same CPU for locking and unlocking while previously it always
relied on smp_processor_id().  The date structure itself is protected with
a lock so it does not rely on CPU-local access.

Replace get_cpu() with raw_smp_processor_id() to obtain the current CPU
number which is used as an index for the per-CPU resource.

Link: https://lore.kernel.org/r/20220506105758.283887-5-bigeasy@linutronix.de
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libfc: Remove get_cpu() semantics in fc_exch_em_alloc()

The get_cpu() in fc_exch_em_alloc() was introduced in commit f018b73af6db
("[SCSI] libfc, libfcoe, fcoe: use smp_processor_id() only when preempt
disabled") for no other reason than to simply use smp_processor_id()
without getting a warning, because everything is done with the pool->lock
held anyway.  However, get_cpu(), by disabling preemption, does not play
well with PREEMPT_RT, particularly when acquiring a regular (and thus
sleepable) spinlock.

Therefore remove the get_cpu() and just use the unstable value as we will
have CPU locality guarantees next by taking the lock.  The window of
migration, as noted by Sebastian, is small and even if it happens the
result is correct.

Link: https://lore.kernel.org/r/20211117025956.79616-2-dave@stgolabs.net
Link: https://lore.kernel.org/r/20220506105758.283887-4-bigeasy@linutronix.de
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: fcoe: Use per-CPU API to update per-CPU statistics

The per-CPU statistics (struct fc_stats) is updated by getting a stable
per-CPU pointer via get_cpu() + per_cpu_ptr() and then performing the
increment. This can be optimized by using this_cpu_*() which will do
whatever is needed on the architecture to perform the update safe and
efficient.  The read out of the individual value (fc_get_host_stats())
should be done by using READ_ONCE() instead of a plain-C access. The
difference is that READ_ONCE() will always perform a single access while
the plain-C access can be split by the compiler into two loads if it
appears beneficial.  The usage of u64 has the side-effect that it is also
64bit wide on 32bit architectures and the read is always split into two
loads. The can lead to strange values if the read happens during an update
which alters both 32bit parts of the 64bit value. This can be circumvented
by either using a 32bit variables on 32bit architecures or extending the
statistics with a sequence counter.

Use this_cpu_*() API to update the statistics and READ_ONCE() to read it.

Link: https://lore.kernel.org/r/20220506105758.283887-3-bigeasy@linutronix.de
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: fcoe: Add a local_lock to fcoe_percpu

fcoe_get_paged_crc_eof() relies on the caller having preemption disabled to
ensure the per-CPU fcoe_percpu context remains valid throughout the
call. This is done by either holding spinlocks (such as bnx2fc_global_lock
or qedf_global_lock) or the get_cpu() from fcoe_alloc_paged_crc_eof(). This
last one breaks PREEMPT_RT semantics as there can be memory allocation and
end up sleeping in atomic contexts.

Introduce a local_lock_t to struct fcoe_percpu that will keep the non-RT
case the same, mapping to preempt_disable/enable, while RT will use a
per-CPU spinlock allowing the region to be preemptible but still maintain
CPU locality. The other users of fcoe_percpu are already safe in this
regard and do not require local_lock()ing.

Link: https://lore.kernel.org/r/20211117025956.79616-3-dave@stgolabs.net
Link: https://lore.kernel.org/r/20220506105758.283887-2-bigeasy@linutronix.de
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Davidlohr Bueso <dave@stgolabs.net>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: target: iscsi: Rename iscsi_session to iscsit_session

The structure iscsi_session naming is used by the iSCSI initiator
driver. Rename the target session to iscsit_session to have more readable
code.

Link: https://lore.kernel.org/r/20220428092939.36768-3-mgurtovoy@nvidia.com
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: target: iscsi: Rename iscsi_conn to iscsit_conn

The structure iscsi_conn naming is used by the iSCSI initiator
driver. Rename the target conn to iscsit_conn to have more readable code.

Link: https://lore.kernel.org/r/20220428092939.36768-2-mgurtovoy@nvidia.com
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: target: iscsi: Rename iscsi_cmd to iscsit_cmd

The structure iscsi_cmd naming is used by the iSCSI initiator driver.
Rename the target cmd to iscsit_cmd to have more readable code.

Link: https://lore.kernel.org/r/20220428092939.36768-1-mgurtovoy@nvidia.com
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Return I/Os to an unrecoverable HBA with DID_ERROR

Complete all new I/O requests issued to an unrecoverable controller with
DID_ERROR status instead of returning the I/O requests with
SCSI_MLQUEUE_HOST_BUSY. This will prevent the infinite retries of the new
I/Os when a controller is in an unrecoverable state.

Link: https://lore.kernel.org/r/20220505184808.24049-1-sreekanth.reddy@broadcom.com
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Hidden drives not removed during soft reset

If any drive is missing during reset, the driver checks whether the device
is exposed to the OS. If it is, then it removes the device from the OS and
its own internal list. For hidden devices, even if they are found as
missing during reset, the driver is not removing them from its internal
list.

Modify driver to remove hidden devices from the driver's target device list
if they are missing during soft reset.

Link: https://lore.kernel.org/r/20220505184808.24049-2-sreekanth.reddy@broadcom.com
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Increase I/O timeout value to 60s

Set each SCSI device's default I/O timeout and default error handling I/O
timeout to 60s.

Link: https://lore.kernel.org/r/20220505184808.24049-3-sreekanth.reddy@broadcom.com
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Update lpfc version to 14.2.0.3

Update lpfc version to 14.2.0.3

Link: https://lore.kernel.org/r/20220506035519.50908-13-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Use sg_dma_address() and sg_dma_len() macros for NVMe I/O

NVMe I/O problems may be seen on IOMMU enabled platforms. Adapter I/Os
failing with transfer length mismatches.

The sg list processing routine for NVMe I/O is accessing the sg entry
directly for the length and address fields. On some IOMMU platforms,
contigous mappings are compressed to the first sg entry with the sum of the
lengths set to the sg entry dma_length field. The length fields are left
for later use by the unmap call. As such, the driver didn't see the actual
dma_length value, just the first entries length value.  Drivers are to use
the sg_dma_length() and sg_dma_address() macros to reference the sg
entry. The macros select the proper length field (dma_length or length) to
reference.

Fix the offending code to use the sg_dma_xxx macros.

Link: https://lore.kernel.org/r/20220506035519.50908-12-jsmart2021@gmail.com
Tested-by: Jerry Snitselaar <jsnitsel@redhat.com>
Reviewed-by: Jerry Snitselaar <jsnitsel@redhat.com>
Co-developed-by: Nigel Kirkland <nkirkland2304@gmail.com>
Signed-off-by: Nigel Kirkland <nkirkland2304@gmail.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Alter FPIN stat accounting logic

When configuring CMF management based on signals instead of FPINs, FPIN
alarm and warning statistics are not tracked.

Change the behavior so that FPIN alarms and warnings are always tracked
regardless of the configured mode.

Similar changes are made in the CMF signal stat accounting logic.  Upon
receipt of a signal, only track signaled alarms and warnings. FPIN stats
should not be incremented upon receipt of a signal.

Link: https://lore.kernel.org/r/20220506035519.50908-11-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Rework FDMI initialization after link up

After a link up, it's possible for the switch to change FDMI support (e.g.
FDMI1 vs FDMI2 vs SmartSAN).  If the switch reverts to FDMI1, then the
revert is currently not detected.

Additionally, when NPIV is configured, it's possible the physical port's
RHBA is unprocessed by the switch before reciept of an NPIV port issued
RPRT.  This causes some switches vendors to reject the NPIV's RPRT.

Fix by reinitializing base FDMI mode on link up, and defer FDMI vport RPRT
submission until after confirming physical port's RHBA is completed.

Link: https://lore.kernel.org/r/20220506035519.50908-10-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Change VMID registration to be based on fabric parameters

Currently, VMID registration is configured via module parameters.  This
could lead to VMID compatibility issues if two ports are connected to
different brands of switches, as the two brands implement VMID differently.

Make logical changes so that VMID registration is based on common service
parameters from FLOGI_ACC with fabric rather than module parameters.

Link: https://lore.kernel.org/r/20220506035519.50908-9-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Decrement outstanding gidft_inp counter if lpfc_err_lost_link()

During large NPIV port testing, it was sometimes seen that not all vports
would log back in to the target device.

There are instances when the fabric is slow to respond to a spam of GID_PT
requests and as a result the SLI PORT may abort the GID_PT request because
the fabric takes so long.  lpfc_cmpl_ct_cmd_gid_pt() would enter the
lpfc_err_lost_link() logic and attempt to lpfc_els_flush_rscn(), which is
fine, but forgets to decrement the gidft_inp counter.  This results in a
vport->gidft_inp never reaching 0 and never restarting discovery again.

Decrement vport->gidft_inp if lpfc_err_lost_link() is true for both
lpfc_cmpl_ct_cmd_gid_pt() and lpfc_cmpl_ct_cmd_gid_ft().

Increase logging info during RSCN timeout and lpfc_err_lost_link() events.

Link: https://lore.kernel.org/r/20220506035519.50908-8-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Use list_for_each_entry_safe() in rscn_recovery_check()

In GID_PT mode with lpfc_ns_query=1, a race condition between iterating the
vport->fc_nodes list in lpfc_rscn_recovery_check() and cleanup of an ndlp
can trigger a crash while processing the RSCN of another initiator from the
same zone.

During iteration of the vport->fc_nodes list, an ndlp is cleaned up and
released. lpfc_dequeue_node() is called from lpfc_cleanup_node() leading to
a bad ndlp dereference in lpfc_rscn_recovery_check().

Change list_for_each_entry() to list_for_each_entry_safe() in
lpfc_rscn_recovery_check() to protect against removal of an initiator ndlp,
while walking the vport->fc_nodes list.

Link: https://lore.kernel.org/r/20220506035519.50908-7-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Fix dmabuf ptr assignment in lpfc_ct_reject_event()

Upon driver receipt of a CT cmd for type = 0xFA (Management Server) and
subtype = 0x11 (Fabric Device Management Interface), the driver is
responding with garbage CT cmd data when it should send a properly formed
RJT.

The __lpfc_prep_xmit_seq64_s4() routine was using the wrong buffer for the
reject.

Fix by converting the routine to use the buffer specified in the bde within
the wqe rather than the ill-set bmp element.

Link: https://lore.kernel.org/r/20220506035519.50908-6-jsmart2021@gmail.com
Fixes: 61910d6a5243 ("scsi: lpfc: SLI path split: Refactor CT paths")
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Inhibit aborts if external loopback plug is inserted

After running a short external loopback test, when the external loopback is
removed and a normal cable inserted that is directly connected to a target
device, the system oops in the llpfc_set_rrq_active() routine.

When the loopback was inserted an FLOGI was transmit. As we're looped back,
we receive the FLOGI request. The FLOGI is ABTS'd as we recognize the same
wppn thus understand it's a loopback. However, as the ABTS sends address
information the port is not set to (fffffe), the ABTS is dropped on the
wire. A short 1 frame loopback test is run and completes before the ABTS
times out. The looback is unplugged and the new cable plugged in, and the
an FLOGI to the new device occurs and completes. Due to a mixup in ref
counting the completion of the new FLOGI releases the fabric ndlp. Then the
original ABTS completes and references the released ndlp generating the
oops.

Correct by no-op'ing the ABTS when in loopback mode (it will be dropped
anyway). Added a flag to track the mode to recognize when it should be
no-op'd.

Link: https://lore.kernel.org/r/20220506035519.50908-5-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Fix ndlp put following a LOGO completion

During testing with repeated asynchronous resets of the target, an issue
was found when the driver issues a LOGO to disconnect its login and recover
all exchanges. The LOGO command takes a node reference but neglects to
remove it, keeping the node reference count artifically high.

Add a call to lpfc_nlp_put() to lpfc_nlp_logo_unreg() and move the mempool
free call to the routine exit along with the needed put.  This is always
safe as this will not be the last reference removed as lpfc_unreg_rpi()
ensures there is an additional reference on the ndlp.

Link: https://lore.kernel.org/r/20220506035519.50908-4-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Fill in missing ndlp kref puts in error paths

Code review, following every lpfc_nlp_get() call vs calls during error
handling, discovered cases of missing put calls.

Correct by adding ndlp kref puts in the respective error paths.

Also added comments to several of the error paths to record relationships
to reference counts.

Link: https://lore.kernel.org/r/20220506035519.50908-3-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Fix element offset in __lpfc_sli_release_iocbq_s4()

The prior commit that moved from iocb elements to explicit wqe elements
missed a name change.

Correct __lpfc_sli_release_iocbq_s4() to reference wqe rather than iocb.

Link: https://lore.kernel.org/r/20220506035519.50908-2-jsmart2021@gmail.com
Fixes: a680a9298e7b ("scsi: lpfc: SLI path split: Refactor lpfc_iocbq")
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: ufshpb: Clean up ufshpb_suspend()/resume()

ufshpb_resume() is only called when the HPB state is HPB_SUSPEND, so the
check statement for "ufshpb_get_state(hpb) != HPB_PRESENT" is useless.

Link: https://lore.kernel.org/r/20220505134707.35929-7-huobean@gmail.com
Reviewed-by: Keoseong Park <keosung.park@samsung.com>
Signed-off-by: Bean Huo <beanhuo@micron.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: ufshpb: Add handing of device reset regions in HPB device mode

In UFS HPB Spec JESD220-3A,

"5.8. Active and inactive information upon power cycle
...
When the device is powered off by the host, the device may restore L2P map
data upon power up or build from the host's HPB READ command. In case
device powered up and lost HPB information, device can signal to the host
through HPB Sense data, by setting HPB Operation as '2' which will inform
the host that device reset HPB information."

Therefore, for HPB device control mode, if the UFS device is reset via the
RST_N pin, the active region information in the device will be reset. If
the host side receives this notification from the device side, it is
recommended to inactivate all active regions in the host's HPB cache.

Link: https://lore.kernel.org/r/20220505134707.35929-6-huobean@gmail.com
Reviewed-by: Keoseong Park <keosung.park@samsung.com>
Reviewed-by: Daejun Park <daejun7.park@samsung.com>
Signed-off-by: Bean Huo <beanhuo@micron.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: ufshpb: Change sysfs node hpb_stats/rb_* prefix to start with rcmd_*

According to the documentation of the sysfs nodes rb_noti_cnt,
rb_active_cnt and rb_inactive_cnt, these are all related to HPB
recommendation in UPIU response packet. 'rcmd' (recommendation) should be
the correct abbreviation.

Change the sysfs documentation about these sysfs nodes to highlight what
they mean under different HPB control modes.

Link: https://lore.kernel.org/r/20220505134707.35929-5-huobean@gmail.com
Reviewed-by: Keoseong Park <keosung.park@samsung.com>
Signed-off-by: Bean Huo <beanhuo@micron.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: ufshpb: Clean up the handler when device resets HPB information

"When the device is powered off by the host, the device may restore L2P map
data upon power up or build from the host's HPB READ command. In case
device powered up and lost HPB information, device can signal to the host
through HPB Sense data, by setting HPB Operation as '2' which will inform
the host that device reset HPB information."

Clean up the handler and make the intent of this handler more readable, no
functional change.

Link: https://lore.kernel.org/r/20220505134707.35929-4-huobean@gmail.com
Reviewed-by: Keoseong Park <keosung.park@samsung.com>
Reviewed-by: Daejun Park <daejun7.park@samsung.com>
Signed-off-by: Bean Huo <beanhuo@micron.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: ufshpb: Remove enum initialization value

If the first enumerator has no initializer, the value of the corresponding
constant is zero.

Link: https://lore.kernel.org/r/20220505134707.35929-3-huobean@gmail.com
Reviewed-by: Keoseong Park <keosung.park@samsung.com>
Signed-off-by: Bean Huo <beanhuo@micron.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: ufshpb: Merge ufshpb_reset() and ufshpb_reset_host()

There is no functional change in this patch, just merge ufshpb_reset() and
ufshpb_reset_host() into one function ufshpb_toggle_state().

Link: https://lore.kernel.org/r/20220505134707.35929-2-huobean@gmail.com
Reviewed-by: Keoseong Park <keosung.park@samsung.com>
Signed-off-by: Bean Huo <beanhuo@micron.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: qcom: Enable RPM_AUTOSUSPEND for runtime PM

In order to allow the block devices to enter autosuspend mode during
runtime, thereby allowing the ufshcd host driver to also runtime suspend,
let's make use of the RPM_AUTOSUSPEND flag.

Without this flag, userspace needs to enable the autosuspend feature of
the block devices through sysfs.

Link: https://lore.kernel.org/r/20220504084212.11605-6-manivannan.sadhasivam@linaro.org
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: core: Remove redundant wmb() in ufshcd_send_command()

The wmb() inside ufshcd_send_command() is added to make sure that the
doorbell is committed immediately. This leads to couple of expectations:

 1. The doorbell write should complete before the function return.

 2. The doorbell write should not cross the function boundary.

2nd expectation is fullfilled by the Linux memory model as there is a
guarantee that the critical section won't cross the unlock (release)
operation.

1st expectation is not really needed here as there is no following read/
write that depends on the doorbell to be complete implicitly. Even if the
doorbell write is in a CPUs Write Buffer (WB), wmb() won't flush it. And
there is no real need of a WB flush here as well.

So let's get rid of the wmb() that seems redundant.

Link: https://lore.kernel.org/r/20220504084212.11605-5-manivannan.sadhasivam@linaro.org
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Acked-by: Bean Huo <beanhuo@micron.com>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: qcom: Add a readl() to make sure ref_clk gets enabled

In ufs_qcom_dev_ref_clk_ctrl(), it was noted that the ref_clk needs to be
stable for at least 1us. Even though there is wmb() to make sure the write
gets "completed", there is no guarantee that the write actually reached the
UFS device. There is a good chance that the write could be stored in a
Write Buffer (WB). In that case, even though the CPU waits for 1us, the
ref_clk might not be stable for that period.

So lets do a readl() to make sure that the previous write has reached the
UFS device before udelay().

Also, the wmb() after writel_relaxed() is not really needed. Both writel()
and readl() are ordered on all architectures and the CPU won't speculate
instructions after readl() due to the in-built control dependency with read
value on weakly ordered architectures. So it can be safely removed.

Link: https://lore.kernel.org/r/20220504084212.11605-4-manivannan.sadhasivam@linaro.org
Fixes: f06fcc7155dc ("scsi: ufs-qcom: add QUniPro hardware support and power optimizations")
Cc: stable@vger.kernel.org
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: qcom: Simplify handling of devm_phy_get()

There is no need to call devm_phy_get() if ACPI is used, so skip it.  The
host->generic_phy pointer should already be NULL due to the kzalloc(), so
no need to set it NULL again.

While at it, also remove the comment that has no relationship with
devm_phy_get().

Link: https://lore.kernel.org/r/20220504084212.11605-3-manivannan.sadhasivam@linaro.org
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: qcom: Fix acquiring the optional reset control line

On Qcom UFS platforms, the reset control line seems to be optional (for
SoCs like MSM8996 and probably for others too). The current logic tries to
mimic the devm_reset_control_get_optional() API but it also continues the
probe if there is an error with the declared reset line in DT/ACPI.

In an ideal case, if the reset line is not declared in DT/ACPI, the probe
should continue. But if there is problem in acquiring the declared reset
line (like EPROBE_DEFER) it should fail and return the appropriate error
code.

Link: https://lore.kernel.org/r/20220504084212.11605-2-manivannan.sadhasivam@linaro.org
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hisi_sas: Undo RPM resume for failed notify phy event for v3 HW

If we fail to notify the phy up event then undo the RPM resume, as the phy
up notify event handling pairs with that RPM resume.

Link: https://lore.kernel.org/r/1651839939-101188-1-git-send-email-john.garry@huawei.com
Reported-by: Yihang Li <liyihang6@hisilicon.com>
Tested-by: Yihang Li <liyihang6@hisilicon.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Update driver version to 8.0.0.69.0

Link: https://lore.kernel.org/r/20220429211641.642010-9-sumit.saxena@broadcom.com
Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Add support for NVMe passthrough

Add support for management applications to send an MPI3 Encapsulated NVMe
passthru command to the NVMe devices attached to an Avenger controller.
Since the NVMe drives are exposed as SCSI devices by the controller, the
standard NVMe applications cannot be used to interact with the drives and
the command sets supported are also limited by the controller firmware.
Special handling is required for MPI3 Encapsulated NVMe passthru commands
for PRP/SGL setup in the commands.

Link: https://lore.kernel.org/r/20220429211641.642010-8-sumit.saxena@broadcom.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Expose adapter state to sysfs

Link: https://lore.kernel.org/r/20220429211641.642010-7-sumit.saxena@broadcom.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Add support for PEL commands

Implement driver support for management applications to enable persistent
event log (PEL) notifications. Upon receipt of events, the driver will
increment a sysfs variable named event_counter. The management application
will poll for event_counter value changes and signal the application about
events.

Link: https://lore.kernel.org/r/20220429211641.642010-6-sumit.saxena@broadcom.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Add support for MPT commands

There are certain management commands which require firmware intervention.
These commands are termed MPT commands. Add support for them.

Link: https://lore.kernel.org/r/20220429211641.642010-5-sumit.saxena@broadcom.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Move data structures/definitions from MPI headers to uapi header

This patch moves the data structures/definitions which are used by
userspace applications from MPI headers to uapi/scsi/scsi_bsg_mpi3mr.h

Link: https://lore.kernel.org/r/20220429211641.642010-4-sumit.saxena@broadcom.com
Reported by: Stephen Rothwell <sfr@canb.auug.org.au>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Add support for driver commands

There are certain bsg commands which need to be completed by the driver
without involving firmware. These requests are termed driver commands. Add
support for these.

Link: https://lore.kernel.org/r/20220429211641.642010-3-sumit.saxena@broadcom.com
Reported by: Stephen Rothwell <sfr@canb.auug.org.au>
Reported-by: kernel test robot <lkp@intel.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Add bsg device support

Create bsg device per controller for controller management purposes.

bsg device nodes will be named /dev/bsg/mpi3mrctl0, /dev/bsg/mpi3mrctl1,
etc.

Link: https://lore.kernel.org/r/20220429211641.642010-2-sumit.saxena@broadcom.com
Reported-by: kernel test robot <lkp@intel.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: sr: Add memory allocation failure handling for get_capabilities()

The function get_capabilities() has the possibility of failing to allocate
the transfer buffer but it does not currently handle this. This may lead to
exceptions when accessing the buffer.

Add error handling when memory allocation fails.

Link: https://lore.kernel.org/r/20220427025647.298358-1-lienze@kylinos.cn
Signed-off-by: Enze Li <lienze@kylinos.cn>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: target: tcmu: Fix possible data corruption

When tcmu_vma_fault() gets a page successfully, before the current context
completes page fault procedure, find_free_blocks() may run and call
unmap_mapping_range() to unmap the page. Assume that when
find_free_blocks() initially completes and the previous page fault
procedure starts to run again and completes, then one truncated page has
been mapped to userspace. But note that tcmu_vma_fault() has gotten a
refcount for the page so any other subsystem won't be able to use the page
unless the userspace address is unmapped later.

If another command subsequently runs and needs to extend dbi_thresh it may
reuse the corresponding slot for the previous page in data_bitmap. Then
though we'll allocate new page for this slot in data_area, no page fault
will happen because we have a valid map and the real request's data will be
lost.

Filesystem implementations will also run into this issue but they usually
lock the page when vm_operations_struct->fault gets a page and unlock the
page after finish_fault() completes. For truncate filesystems lock pages in
truncate_inode_pages() to protect against racing wrt. page faults.

To fix this possible data corruption scenario we can apply a method similar
to the filesystems.  For pages that are to be freed, tcmu_blocks_release()
locks and unlocks. Make tcmu_vma_fault() also lock found page under
cmdr_lock. At the same time, since tcmu_vma_fault() gets an extra page
refcount, tcmu_blocks_release() won't free pages if pages are in page fault
procedure, which means it is safe to call tcmu_blocks_release() before
unmap_mapping_range().

With these changes tcmu_blocks_release() will wait for all page faults to
be completed before calling unmap_mapping_range(). And later, if
unmap_mapping_range() is called, it will ensure stale mappings are removed.

Link: https://lore.kernel.org/r/20220421023735.9018-1-xiaoguang.wang@linux.alibaba.com
Reviewed-by: Bodo Stroesser <bostroesser@gmail.com>
Signed-off-by: Xiaoguang Wang <xiaoguang.wang@linux.alibaba.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Remove redundant lpfc_sli_prep_wqe() call

Prior patch added a call to lpfc_sli_prep_wqe() prior to
lpfc_sli_issue_iocb().  This call should not have been added as prep_wqe is
called within the issue_iocb routine. So it's called twice now.

Remove the redundant prep call.

Link: https://lore.kernel.org/r/20220427222223.57920-1-jsmart2021@gmail.com
Fixes: 31a59f75702f ("scsi: lpfc: SLI path split: Refactor Abort paths")
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Fix additional reference counting in lpfc_bsg_rport_els()

Code inspection has found an additional reference is taken in
lpfc_bsg_rport_els(). Results in the ndlp not being freed thus is leaked.

Fix by removing the redundant refcount taken before WQE submission.

Link: https://lore.kernel.org/r/20220427222158.57867-1-jsmart2021@gmail.com
Co-developed-by: Nigel Kirkland <nigel.kirkland@broadcom.com>
Signed-off-by: Nigel Kirkland <nigel.kirkland@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: sd: Reorganize DIF/DIX code to avoid calling revalidate twice

During device discovery we ended up calling revalidate twice and thus
requested the same parameters multiple times. This was originally necessary
due to the request_queue and gendisk needing to be instantiated to
configure the block integrity profile.

Since this dependency no longer exists, reorganize the integrity probing
code so it can be run once at the end of discovery and drop the superfluous
revalidate call. Postponing the registration step involves splitting
sd_read_protection() into two functions, one to read the device protection
type and one to configure the mode of operation.

As part of this cleanup, make the printing code a bit less verbose.

Link: https://lore.kernel.org/r/20220302053559.32147-14-martin.petersen@oracle.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: sd: Optimal I/O size should be a multiple of reported granularity

Commit a83da8a4509d ("scsi: sd: Optimal I/O size should be a multiple of
physical block size") validated the reported optimal I/O size against the
physical block size to overcome problems with devices reporting nonsensical
transfer sizes.

However, some devices claim conformity to older SCSI versions that predate
the physical block size being reported. Other devices do not report a
physical block size at all. We need to be able to validate the optimal I/O
size on those devices as well.

Many devices report an OPTIMAL TRANSFER LENGTH GRANULARITY in the same VPD
page as the OPTIMAL TRANSFER LENGTH. Use this value to validate the optimal
I/O size. Also check that the reported granularity is a multiple of the
physical block size, if supported.

Link: https://lore.kernel.org/r/33fb522e-4f61-1b76-914f-c9e6a3553c9b@gmail.com
Link: https://lore.kernel.org/r/20220302053559.32147-9-martin.petersen@oracle.com
Reported-by: Bernhard Sulzer <micraft.b@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: sd: Switch to using scsi_device VPD pages

Use the VPD pages already provided by the SCSI midlayer. No need to request
them individually in the SCSI disk driver.

Link: https://lore.kernel.org/r/20220302053559.32147-8-martin.petersen@oracle.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: sd: Use cached ATA Information VPD page

Since the ATA Information VPD is now cached at device discovery time it is
no longer necessary to request this page when we configure WRITE SAME.
Instead use the cached information to determine if this disk sits behind a
SCSI-ATA translation layer.

Link: https://lore.kernel.org/r/20220302053559.32147-7-martin.petersen@oracle.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: Do not truncate INQUIRY data on modern devices

Low-level device drivers have had the ability to limit the size of an
INQUIRY for many years. This made sense for a wide variety of legacy
devices. However, we are unnecessarily truncating the INQUIRY response for
many modern devices. This prevents us from consulting fields beyond the
first 36 bytes.

If a device reports that it supports a larger INQUIRY response, and the
device also reports that it implements SPC-4 or newer, allow the larger
INQUIRY to proceed.

Link: https://lore.kernel.org/r/20220302053559.32147-4-martin.petersen@oracle.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: Cache VPD pages b0, b1, b2

The SCSI disk driver consults VPD pages b0 (Block Limits), b1 (Block Device
Characteristics), and b2 (Logical Block Provisioning). Instead of having
sd.c request these pages every revalidate cycle, cache them along with the
other commonly used VPDs.

Link: https://lore.kernel.org/r/20220302053559.32147-6-martin.petersen@oracle.com
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: Pick suitable allocation length in scsi_report_opcode()

Some devices hang when a buffer size larger than expected is passed in the
ALLOCATION LENGTH field. For REPORT SUPPORTED OPERATION CODES we currently
only request a single command descriptor at a time and therefore the actual
size of the command is known ahead of time. Limit the ALLOCATION LENGTH to
the header size plus the command length of the opcode we are asking about.

Link: https://lore.kernel.org/r/20220302053559.32147-5-martin.petersen@oracle.com
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: Query VPD size before getting full page

We currently default to 255 bytes when fetching VPD pages during discovery.
However, we have had a few devices that are known to wedge if the requested
buffer exceeds a certain size. See commit af73623f5f10 ("[SCSI] sd: Reduce
buffer size for vpd request") which works around one example of this
problem in the SCSI disk driver.

With commit d188b0675b21 ("scsi: core: Add sysfs attributes for VPD pages
0h and 89h") we now risk triggering the same issue in the generic midlayer
code.

The problem with the ATA VPD page in particular is that the SCSI portion of
the page is trailed by 512 bytes of verbatim ATA Identify Device
information.  However, not all controllers actually provide the additional
512 bytes and will lock up if one asks for more than the 64 bytes
containing the SCSI protocol fields.

Instead of picking a new, somewhat arbitrary, number of bytes for the VPD
buffer size, start fetching the 4-byte header for each page. The header
contains the size of the page as far as the device is concerned. We can use
the reported size to specify the correct allocation length when
subsequently fetching the full page.

The header validation is done by a new helper function scsi_get_vpd_size()
and both scsi_get_vpd_page() and scsi_get_vpd_buf() now rely on this to
query the page size.

In addition, scsi_get_vpd_page() is simplified to mirror the logic in
scsi_get_vpd_page(). This involves removing the Supported VPD Pages lookup
prior to attempting to query a page. There does not appear any evidence,
even in the oldest SCSI specs, that this step is required. We already rely
on scsi_get_vpd_page() throughout the stack and this function never
consulted the Supported VPD Pages. Since this has not caused any problems
it should be safe to remove the precondition from scsi_get_vpd_page().

Instrumented runs also revealed that the Supported VPD Pages lookup had
little effect since the device page index often was larger than the
supplied buffer size. As a result, inquiries frequently bypassed the index
check and went through the "If we ran off the end of the buffer, give us
the benefit of the doubt" code path which assumed the page was present
despite not being listed. The revised code takes both the page size
reported by the device as well as the size of the buffer provided by the
scsi_get_vpd_page() caller into account.

Link: https://lore.kernel.org/r/20220302053559.32147-3-martin.petersen@oracle.com
Fixes: d188b0675b21 ("scsi: core: Add sysfs attributes for VPD pages 0h and 89h")
Reported-by: Maciej W. Rozycki <macro@orcam.me.uk>
Tested-by: Maciej W. Rozycki <macro@orcam.me.uk>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpt3sas: Use cached ATA Information VPD page

We now cache VPD page 0x89 (ATA Information) so there is no need to request
it from the hardware. Make mpt3sas use the cached page.

Link: https://lore.kernel.org/r/20220302053559.32147-2-martin.petersen@oracle.com
Cc: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Fix resource leak in lpfc_sli4_send_seq_to_ulp()

If no handler is found in lpfc_complete_unsol_iocb() to match the rctl of a
received frame, the frame is dropped and resources are leaked.

Fix by returning resources when discarding an unhandled frame type.  Update
lpfc_fc_frame_check() handling of NOP basic link service.

Link: https://lore.kernel.org/r/20220426181419.9154-1-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Remove unnecessary null ndlp check in lpfc_sli_prep_wqe()

Smatch had the following warning:

drivers/scsi/lpfc/lpfc_sli.c:22305 lpfc_sli_prep_wqe() error: we previously assumed 'ndlp' could be null (see line 22298)

Remove the unnecessary null check.

Link: https://lore.kernel.org/r/20220426181315.8990-1-jsmart2021@gmail.com
Fixes: d51cf5bd926c ("scsi: lpfc: Fix field overload in lpfc_iocbq data structure")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: Use pm_runtime_resume_and_get() instead of pm_runtime_get_sync()

Using pm_runtime_resume_and_get() to replace pm_runtime_get_sync() and
pm_runtime_put_noidle(). This change is just to simplify the code, no
actual functional changes.

Link: https://lore.kernel.org/r/20220420090353.2588804-1-chi.minghao@zte.com.cn
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Minghao Chi <chi.minghao@zte.com.cn>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid: Fix error check return value of register_chrdev()

If major equals 0, register_chrdev() returns an error code when it fails.
This function dynamically allocates a major and returns its number on
success, so we should use "< 0" to check it instead of "!".

Link: https://lore.kernel.org/r/20220418105755.2558828-1-lv.ruyi@zte.com.cn
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Lv Ruyi <lv.ruyi@zte.com.cn>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: edif: Remove unneeded variable

Fix the following coccicheck warning:

drivers/scsi/qla2xxx/qla_edif.c:660:11-15: Unneeded variable: "rval".

Return "0" on line 761.

Link: https://lore.kernel.org/r/20220426074334.9281-1-guozhengkui@vivo.com
Signed-off-by: Guo Zhengkui <guozhengkui@vivo.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: dc395x: Fix a missing check on list iterator

The bug is here:

p->target_id, p->target_lun);

The list iterator 'p' will point to a bogus position containing HEAD if the
list is empty or no element is found. This case must be checked before any
use of the iterator, otherwise it will lead to an invalid memory access.

To fix this bug, add a check. Use a new variable 'iter' as the list
iterator, and use the original variable 'p' as a dedicated pointer to point
to the found element.

Link: https://lore.kernel.org/r/20220414040231.2662-1-xiam0nd.tong@gmail.com
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Cc: stable@vger.kernel.org
Signed-off-by: Xiaomeng Tong <xiam0nd.tong@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qedf: Remove an unneeded NULL check on list iterator

The list iterator 'fcport' is always non-NULL so it doesn't need to be
checked. Thus just remove the unnecessary NULL check. Also remove the
unnecessary initializer because the list iterator is always initialized.
And adjust the position of blank lines.

Link: https://lore.kernel.org/r/20220405004055.24312-1-xiam0nd.tong@gmail.com
Signed-off-by: Xiaomeng Tong <xiam0nd.tong@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: core: Remove duplicate include in ufshcd

Fix following checkincludes warning:

drivers/scsi/ufs/ufshcd.c: linux/nls.h is included more than once.

The include is in line 14. Remove the duplicate.

Link: https://lore.kernel.org/r/20220426104509.621394-1-wanjiabing@vivo.com
Reviewed-by: Avri Altman <avri.altman@wdc.com>
Signed-off-by: Wan Jiabing <wanjiabing@vivo.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: core: Exclude UECxx from SFR dump list

Some devices may return invalid or zeroed data during an UIC error
condition. In addition, reading these SFRs will clear them. This means the
subsequent error handling will not be able to see them and therefore no
error handling will be scheduled.

Skip reading these SFRs in ufshcd_dump_regs().

Link: https://lore.kernel.org/r/1648689845-33521-1-git-send-email-kwmad.kim@samsung.com
Fixes: d67247566450 ("scsi: ufs: Use explicit access size in ufshcd_dump_regs")
Signed-off-by: Kiwoong Kim <kwmad.kim@samsung.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: Refine how we set tag_set NUMA node

For SCSI hosts which enable host_tagset the NUMA node returned from
blk_mq_hw_queue_to_node() is NUMA_NO_NODE always. Then, since in
scsi_mq_setup_tags() the default we choose for the tag_set NUMA node is
NUMA_NO_NODE, we always evaluate the NUMA node as NUMA_NO_NODE in functions
like blk_mq_alloc_rq_map().

The reason we get NUMA_NO_NODE from blk_mq_hw_queue_to_node() is that the
hctx_idx passed is BLK_MQ_NO_HCTX_IDX - so we can't match against a (HW)
queue mapping index.

Improve this by defaulting the tag_set NUMA node to the same NUMA node of
the SCSI host DMA dev.

Link: https://lore.kernel.org/r/1648640315-21419-1-git-send-email-john.garry@huawei.com
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Remove unneeded variable

Remove unneeded variable 'rc' used to store return value.

Link: https://lore.kernel.org/r/20220419065750.2573861-1-deng.changcheng@zte.com.cn
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Changcheng Deng <deng.changcheng@zte.com.cn>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: fnic: Remove unneeded flush_workqueue()

All work currently pending will be done first by calling
destroy_workqueue(). There is no need to flush it explicitly.

Link: https://lore.kernel.org/r/20220424065406.3228528-1-ran.jianping@zte.com.cn
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: ran jianping <ran.jianping@zte.com.cn>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Remove unneeded flush_workqueue()

All work currently pending will be done first by calling
destroy_workqueue(). There is no need to flush it explicitly.

Link: https://lore.kernel.org/r/20220424062413.3220315-1-ran.jianping@zte.com.cn
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: ran jianping <ran.jianping@zte.com.cn>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: bfa: Remove unneeded flush_workqueue()

All work currently pending will be done first by calling
destroy_workqueue(). There is no need to flush it explicitly.

Link: https://lore.kernel.org/r/20220424061845.3218774-1-ran.jianping@zte.com.cn
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: ran jianping <ran.jianping@zte.com.cn>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qedf: Remove unnecessary code

Buffer 'msg' is memset() but never actually used. Drop it.

Link: https://lore.kernel.org/r/1650536693-13089-1-git-send-email-baihaowen@meizu.com
Signed-off-by: Haowen Bai <baihaowen@meizu.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla4xxx: Drop redundant memset()

The region set by the call to memset() is immediately overwritten by the
subsequent call to memcpy(). Drop redundant memset().

Link: https://lore.kernel.org/r/1650533091-28815-1-git-send-email-baihaowen@meizu.com
Signed-off-by: Haowen Bai <baihaowen@meizu.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ipr: Directly return instead of using local ret variable

Fixes coccinelle warning:

./drivers/scsi/ipr.c:10095:13-15: Unneeded variable: "rc". Return "IRQ_HANDLED" on line 10104

Link: https://lore.kernel.org/r/1648433103-24308-1-git-send-email-baihaowen@meizu.com
Signed-off-by: Haowen Bai <baihaowen@meizu.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: elx: efct: Remove unnecessary memset() in efct_io()

io->sgl is allocated by kzalloc(). The memory is set to zero.
It is unnecessary to call memset again.

Link: https://lore.kernel.org/r/20220318145230.1031-1-wanjiabing@vivo.com
Signed-off-by: Wan Jiabing <wanjiabing@vivo.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: message: fusion: Remove unused variable retval

The following warning showed up when compiling with W=1.

drivers/message/fusion/mptctl.c: In function ‘mptctl_hp_hostinfo’:
drivers/message/fusion/mptctl.c:2337:8: warning: variable ‘retval’ set but not used [-Wunused-but-set-variable]
  int   retval;

Fixing by removing the variable.

Link: https://lore.kernel.org/r/20220317030325.30526-1-alexander.vorwerk@stud.uni-goettingen.de
Signed-off-by: Alexander Vorwerk <alexander.vorwerk@stud.uni-goettingen.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aha1542: Remove unneeded semicolon

Fixes coccicheck warning:

drivers/scsi/aha1542.c:553:2-3: Unneeded semicolon
drivers/scsi/aha1542.c:582:2-3: Unneeded semicolon
drivers/scsi/aha1542.c:605:2-3: Unneeded semicolon
drivers/scsi/aha1542.c:306:2-3: Unneeded semicolon
drivers/scsi/aha1542.c:348:3-4: Unneeded semicolon
drivers/scsi/aha1542.c:412:2-3: Unneeded semicolon
drivers/scsi/aha1542.c:640:2-3: Unneeded semicolon
drivers/scsi/aha1542.c:658:2-3: Unneeded semicolon
drivers/scsi/aha1542.c:677:2-3: Unneeded semicolon
drivers/scsi/aha1542.c:538:2-3: Unneeded semicolon

Link: https://lore.kernel.org/r/1647309865-15620-1-git-send-email-baihaowen@meizu.com
Signed-off-by: Haowen Bai <baihaowen@meizu.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: bfa: Remove redundant NULL check

Fix the following warning reported by coccicheck:

drivers/scsi/bfa/bfad_debugfs.c:375:2-7: WARNING: NULL check before some freeing functions is not needed.

Link: https://lore.kernel.org/r/1647309434-13936-1-git-send-email-baihaowen@meizu.com
Signed-off-by: Haowen Bai <baihaowen@meizu.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: fnic: Remove redundant NULL check

Fix the following warning reported by coccicheck:

drivers/scsi/fnic/fnic_debugfs.c:90:2-7: WARNING: NULL check before some freeing functions is not needed.

Link: https://lore.kernel.org/r/1647309219-12772-1-git-send-email-baihaowen@meizu.com
Signed-off-by: Haowen Bai <baihaowen@meizu.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mac53c94: Fix warning comparing pointer to 0

Fix the following coccicheck warning:

drivers/scsi/mac53c94.c:237:12-13: WARNING comparing pointer to 0

Link: https://lore.kernel.org/r/1647244711-31575-1-git-send-email-baihaowen@meizu.com
Signed-off-by: Haowen Bai <baihaowen@meizu.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Fix undefined behavior due to shift overflowing the constant

Fix:

  drivers/scsi/aacraid/commsup.c: In function ‘aac_handle_sa_aif’:
  drivers/scsi/aacraid/commsup.c:1983:2: error: case label does not reduce to an integer constant
    case SA_AIF_BPCFG_CHANGE:
    ^~~~

See https://lore.kernel.org/r/YkwQ6%2BtIH8GQpuct@zn.tnic for the gory
details as to why it triggers with older gccs only.

Link: https://lore.kernel.org/r/20220405151517.29753-2-bp@alien8.de
Cc: Adaptec OEM Raid Solutions <aacraid@microsemi.com>
Cc: "James E.J. Bottomley" <jejb@linux.ibm.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
Cc: linux-scsi@vger.kernel.org
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: scsi_debug: Add gap zone support

Add the 'zone_cap_mb' kernel module parameter. This parameter defines the
zone capacity. The zone capacity must be less than or equal to the zone
size.

Report that sequential write zones and gap zones are paired in the Zoned
Block Device Characteristics VPD page (page B6h).

This patch has been tested as follows:

modprobe scsi_debug delay=0 sector_size=512 dev_size_mb=128 zbc=host-managed zone_nr_conv=16 zone_size_mb=4 zone_cap_mb=3
modprobe brd rd_nr=1 rd_size=$((1<<20))
mkfs.f2fs -m /dev/ram0 -c /dev/${scsi_debug_dev}
mount /dev/ram0 /mnt
 # Run a fio job that uses /mnt

Link: https://lore.kernel.org/r/20220421183023.3462291-10-bvanassche@acm.org
Cc: Douglas Gilbert <dgilbert@interlog.com>
Acked-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
[ bvanassche: Switched to reporting a constant zone starting LBA granularity ]
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: scsi_debug: Rename zone type constants

Rename the scsi_debug zone type constants to prevent a conflict with the
ZBC_ZONE_TYPE_GAP constant from include/scsi/scsi_proto.h.

Link: https://lore.kernel.org/r/20220421183023.3462291-9-bvanassche@acm.org
Cc: Douglas Gilbert <dgilbert@interlog.com>
Acked-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
[ bvanassche: Extracted these changes from a larger patch ]
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: scsi_debug: Fix a typo

Change a single occurrence of "nad" into "and".

Link: https://lore.kernel.org/r/20220421183023.3462291-8-bvanassche@acm.org
Cc: Douglas Gilbert <dgilbert@interlog.com>
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Acked-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: sd: sd_zbc: Hide gap zones

ZBC-2 allows host-managed disks to report gap zones. This allow zoned disks
to report an offset between data zone starts that is a power of two even if
the number of logical blocks with data per zone is not a power of two.

Another new feature in ZBC-2 is support for constant zone starting LBA
offsets. For zoned disks that report a constant zone starting LBA offset,
hide the gap zones from the block layer. Report the offset between data
zone starts as zone size and report the number of logical blocks with data
per zone as the zone capacity.

Link: https://lore.kernel.org/r/20220421183023.3462291-7-bvanassche@acm.org
Acked-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
[ bvanassche: Reworked this patch ]
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: sd: sd_zbc: Return early in sd_zbc_check_zoned_characteristics()

Return early in sd_zbc_check_zoned_characteristics() for host-aware
disks. This patch does not change any functionality but makes a later patch
easier to read.

Link: https://lore.kernel.org/r/20220421183023.3462291-6-bvanassche@acm.org
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Acked-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
[ bvanassche: extracted this change from a larger patch ]
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: sd: sd_zbc: Introduce struct zoned_disk_info

Deriving the meaning of the nr_zones, rev_nr_zones, zone_blocks and
rev_zone_blocks member variables requires careful analysis of the source
code. Make the meaning of these member variables easier to understand by
introducing struct zoned_disk_info.

Link: https://lore.kernel.org/r/20220421183023.3462291-5-bvanassche@acm.org
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Acked-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>