scsi: qla2xxx: Fix spelling mistake "tranport" -> "transport"

There is a spelling mistake in a ql_dbg message. Fix it.

Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Link: https://lore.kernel.org/r/20230828213101.758609-1-colin.i.king@gmail.com
Acked-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Merge branch '6.5/scsi-fixes' into 6.6/scsi-staging

Pull in the fixes tree for a commit that missed 6.5. Also resolve a
trivial merge conflict in fnic.

* 6.5/scsi-fixes: (36 commits)
  scsi: storvsc: Handle additional SRB status values
  scsi: snic: Fix double free in snic_tgt_create()
  scsi: core: raid_class: Remove raid_component_add()
  scsi: ufs: ufs-qcom: Clear qunipro_g4_sel for HW major version > 5
  scsi: ufs: mcq: Fix the search/wrap around logic
  scsi: qedf: Fix firmware halt over suspend and resume
  scsi: qedi: Fix firmware halt over suspend and resume
  scsi: qedi: Fix potential deadlock on &qedi_percpu->p_work_lock
  scsi: lpfc: Remove reftag check in DIF paths
  scsi: ufs: renesas: Fix private allocation
  scsi: snic: Fix possible memory leak if device_add() fails
  scsi: core: Fix possible memory leak if device_add() fails
  scsi: core: Fix legacy /proc parsing buffer overflow
  scsi: 53c700: Check that command slot is not NULL
  scsi: fnic: Replace return codes in fnic_clean_pending_aborts()
  scsi: storvsc: Fix handling of virtual Fibre Channel timeouts
  scsi: pm80xx: Fix error return code in pm8001_pci_probe()
  scsi: zfcp: Defer fc_rport blocking until after ADISC response
  scsi: storvsc: Limit max_sectors for virtual Fibre Channel devices
  scsi: sg: Fix checking return value of blk_get_queue()
  ...

Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: fnic: Replace sgreset tag with max_tag_id

sgreset is issued with a SCSI command pointer. The device reset code
assumes that it was issued on a hardware queue, and calls block multiqueue
layer. However, the assumption is broken, and there is no hardware queue
associated with the sgreset, and this leads to a crash due to a null
pointer exception.

Fix the code to use the max_tag_id as a tag which does not overlap with the
other tags issued by mid layer.

Tested by running FC traffic for a few minutes, and by issuing sgreset on
the device in parallel.  Without the fix, the crash is observed right away.
With this fix, no crash is observed.

Reviewed-by: Sesidhar Baddela <sebaddel@cisco.com>
Tested-by: Karan Tilak Kumar <kartilak@cisco.com>
Signed-off-by: Karan Tilak Kumar <kartilak@cisco.com>
Link: https://lore.kernel.org/r/20230817182146.229059-1-kartilak@cisco.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: storvsc: Handle additional SRB status values

Testing of virtual Fibre Channel devices under Hyper-V has shown additional
SRB status values being returned for various error cases.  Because these
SRB status values are not recognized by storvsc, the I/O operations are not
flagged as an error. Requests are treated as if they completed normally but
with zero data transferred, which can cause a flood of retries.

Add definitions for these SRB status values and handle them like other
error statuses from the Hyper-V host.

Signed-off-by: Michael Kelley <mikelley@microsoft.com>
Link: https://lore.kernel.org/r/1692984084-95105-1-git-send-email-mikelley@microsoft.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Merge patch series "qla2xxx driver misc features"

Nilesh Javali <njavali@marvell.com> says:

Martin,

Please apply the qla2xxx driver miscellaneous features and bug fixes
to the scsi tree at your earliest convenience.

Link: https://lore.kernel.org/r/20230821130045.34850-1-njavali@marvell.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Remove unused variables in qla24xx_build_scsi_type_6_iocbs()

Sparse warning reported,

drivers/scsi/qla2xxx/qla_iocb.c: In function 'qla24xx_build_scsi_type_6_iocbs':
>> drivers/scsi/qla2xxx/qla_iocb.c:594:29: warning: variable 'ha' set but not used [-Wunused-but-set-variable]
     594 |         struct qla_hw_data *ha;
         |                             ^~

Remove unused variables 'vha' and 'ha'.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202308230757.VKMIztAB-lkp@intel.com/
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230825070017.46066-1-njavali@marvell.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Fix nvme_fc_rcv_ls_req() undefined error

The kernel robot reported below build error,

>> ERROR: modpost: "nvme_fc_rcv_ls_req" [drivers/scsi/qla2xxx/qla2xxx.ko] undefined!

Use CONFIG_NVME_FC enabled check to fix the build error.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202308021445.txlNq7UC-lkp@intel.com/
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230824151521.35261-1-njavali@marvell.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Merge patch series "smartpqi updates"

Don Brace <don.brace@microchip.com> says:

cat smartpqi_6.6_cover_letter
These patches are based on Martin Petersen's 6.6/scsi-queue tree
  https://git.kernel.org/pub/scm/linux/kernel/git/mkp/scsi.git
  6.6/scsi-queue

The biggest functional change to smartpqi is the addition of an abort
handler. Some customers were complaining about I/O stalls to all
devices when only one device is reset. Adding an abort handler helps
to prevent I/O stalls to all devices.

All of the reset of the patches are small changes to logging messages,
MACRO and variable name changes, and one minor change for LUN
assignments.

This set of changes consists of:

 * smartpqi-add-abort-handler
   When a device reset occurs, the SML pauses I/O to all devices presented
   by a controller instance causing some performance issues.
   To only affect device with a problematic request, we added an abort handler.
   The abort handler is implemented by using a device reset, but I/O to the
   other devices is no longer affected.
 * smartpqi-refactor-rename-MACRO-to-clarify-purpose
   The MACRO SOP_RC_INCORRECT_LOGICAL_UNIT was used to check for a condition
   where a TMF was sent an incorrect LUN. We renamed this MACRO to
   SOP_TMF_INCORRECT_LOGICAL_UNIT for clarity.
 * smartpqi-refactor-rename-pciinfo-to-pci_info
   Change the pciinfo variable to pci_info to make more readable code.
   No functional changes.
 * smartpqi-simplify-lun_number-assignment
   We simplified the conditional expression used to populate LUN numbers
   for requests.
 * smartpqi-enhance-shutdown-notification
   Clarify controller cache flush errors. We added in more precise information
   to the cache flush informational message.
   No functional changes.
 * smartpqi-enhance-controller-offline-notification
   The driver can offline a controller for multiple reasons. We added
   a description of why these rare offline actions are taken. And a function
   to provide the specific details of the shutdown.
 * smartpqi-enhance-error-messages
   We added host:bus:target:lun to messages emitted in our reset/abort handlers.
   No functional changes.
 * smartpqi-change-driver-version-to-2.1.24-046

Link: https://lore.kernel.org/r/20230824155812.789913-1-don.brace@microchip.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: smartpqi: Change driver version to 2.1.24-046

Reviewed-by: Gerry Morong <gerry.morong@microchip.com>
Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Link: https://lore.kernel.org/r/20230824155812.789913-9-don.brace@microchip.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: smartpqi: Enhance error messages

Add more detail to some TMF messages.

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Mahesh Rajashekhara <Mahesh.Rajashekhara@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Link: https://lore.kernel.org/r/20230824155812.789913-8-don.brace@microchip.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: smartpqi: Enhance controller offline notification

Add a description for the reason the controller has been taken off-line.

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Signed-off-by: David Strahan <David.Strahan@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Link: https://lore.kernel.org/r/20230824155812.789913-7-don.brace@microchip.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: smartpqi: Enhance shutdown notification

Provide more detailed information about cache flush errors during shutdown.

Reviewed-by: Mahesh Rajashekhara <mahesh.rajashekhara@microchip.com>
Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: David Strahan <David.Strahan@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Link: https://lore.kernel.org/r/20230824155812.789913-6-don.brace@microchip.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: smartpqi: Simplify lun_number assignment

Simplify lun_number assignment. lun_number assignment is only required for
non-AIO requests.

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: David Strahan <David.Strahan@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Link: https://lore.kernel.org/r/20230824155812.789913-5-don.brace@microchip.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: smartpqi: Rename pciinfo to pci_info

Make pci device structure names consistent and readable.

Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Signed-off-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Link: https://lore.kernel.org/r/20230824155812.789913-4-don.brace@microchip.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: smartpqi: Rename MACRO to clarify purpose

Rename SOP_RC_INCORRECT_LOGICAL_UNIT to SOP_TMF_INCORRECT_LOGICAL_UNIT to
clarify the intended purpose.

Reviewed-by: Mahesh Rajashekhara <mahesh.rajashekhara@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Signed-off-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Link: https://lore.kernel.org/r/20230824155812.789913-3-don.brace@microchip.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: smartpqi: Add abort handler

Implement aborts as resets.

Avoid I/O stalls across all devices attached to a controller when device
I/O requests time out.

Reviewed-by: Mahesh Rajashekhara <mahesh.rajashekhara@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Signed-off-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Link: https://lore.kernel.org/r/20230824155812.789913-2-don.brace@microchip.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: snic: Fix double free in snic_tgt_create()

Commit 41320b18a0e0 ("scsi: snic: Fix possible memory leak if device_add()
fails") fixed the memory leak caused by dev_set_name() when device_add()
failed. However, it did not consider that 'tgt' has already been released
when put_device(&tgt->dev) is called. Remove kfree(tgt) in the error path
to avoid double free of 'tgt' and move put_device(&tgt->dev) after the
removed kfree(tgt) to avoid a use-after-free.

Fixes: 41320b18a0e0 ("scsi: snic: Fix possible memory leak if device_add() fails")
Signed-off-by: Zhu Wang <wangzhu9@huawei.com>
Link: https://lore.kernel.org/r/20230819083941.164365-1-wangzhu9@huawei.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: sd: Remove the number of forward declarations

Move the sd_pm_ops and sd_template data structures to just above init_sd()
such that the number of forward function declarations can be reduced.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20230823210628.523244-1-bvanassche@acm.org
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: st: Add third party poweron reset handling

Many tape devices will automatically rewind following a poweron/reset.
This can result in data loss as other operations in the driver can write to
the tape when the position is unknown. E.g. MTEOM can write a filemark at
the beginning of the tape. This patch adds code to detect poweron/reset
unit attentions and prevents the driver from writing to the tape when the
position could be unknown.

Customer reported problem description:

We have experienced an issue with the SCSI tape driver (st) which has led
to data loss for us on two separate occasions in production, as well as in
a third case in which we were able to reproduce the failure in our test
environment.

The tape device involved is an Amazon Tape Gateway, a virtual tape library
(VTL) appliance which presents as a series of iSCSI targets (multiple tape
drives and a changer) and is backed by storage in Amazon S3. The problem is
a general one and not limited to any particular SCSI transport or tape
device, though the nature of both iSCSI and the VTL make data loss somewhat
more likely with this combination than with a physical tape drive.

The observed behavior occurs when an error causes the VTL tape gateway
process (on the appliance) to crash and restart. This interrupts the iSCSI
TCP connections and, when it occurs during a write, causes the write to
fail with EIO. However, we then find that the virtual tape in question is
now completely blank. We raised this issue with AWS support, thinking this
must be a bug in the VTL appliance, but that turns out not to be the case.

Per AWS support, when the gateway crashes in this manner, its notion of the
current tape position is reset to the beginning of the tape. It also sets a
unit attention condition, such that the next request results in a CHECK
CONDITION status with sense key UNIT ATTENTION and asc/ascq indicating a
device reset. According to their logs the next command being sent is WRITE
FILEMARK, which results in writing an FM at the beginning of the tape,
effectively discarding its contents.

In fact, once the write fails with EIO, our software attempts to recover by
rewinding and repositioning the tape, then resuming operation. If this
fails, it attempts to rewind and reposition again, write a marker at the
end of the tape, and then unmount. It does not under any circumstances
write either data or filemarks without having successfully positioned the
tape to a known point.

What actually happens is that, since the last operation was a write, the
kernel executes an implied MTWEOF operation (which translate to a Write
Filemarks command) before the rewind that was actually requested. This
seems not entirely unreasonable, provided the tape position is known.
However, once this request fails (due to the unit attention condition), our
next rewind attempt also triggers an implied MTWEOF, which does _not_ fail
(the unit attention condition persists only until the initiator has been
notified); this is the command that unexpectedly erases the tape.

Our analysis is that the st driver is in fact completely ignoring the UNIT
ATTENTION and associated reset notification from the device. This is not a
condition that can be detected in the transport or mid-layer, as it occurs
entirely within the target and is reported only via the UNIT ATTENTION
sense key. The upper driver (i.e. st) needs to detect this indication and
reset its internal model of the device to an unknown state.

Suggested-by: Jeffrey Hutzelman <jhutz@cmu.edu>
Signed-off-by: John Meneghini <jmeneghi@redhat.com>
Link: https://lore.kernel.org/r/20230822181413.1210647-1-jmeneghi@redhat.com
Acked-by: Kai Mäkisara <kai.makisara@kolumbus.fi>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: Report error list information in debugfs

Provide information in debugfs about SCSI error handling to make it easier
to debug the SCSI error handler. Additionally, report the maximum number of
retries in debugfs (.allowed).

Reviewed-by: John Garry <john.g.garry@oracle.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Cc: Mike Christie <michael.christie@oracle.com>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20230822163811.219569-1-bvanassche@acm.org
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: Improve type safety of scsi_rescan_device()

Most callers of scsi_rescan_device() have the scsi_device pointer readily
available. Pass a struct scsi_device pointer to scsi_rescan_device()
instead of a struct device pointer. This change prevents that a pointer to
another struct device would be passed accidentally to scsi_rescan_device().

Remove the scsi_rescan_device() declaration from the scsi_priv.h header
file since it duplicates the declaration in <scsi/scsi_host.h>.

Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: John Garry <john.g.garry@oracle.com>
Cc: Mike Christie <michael.christie@oracle.com>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20230822153043.4046244-1-bvanassche@acm.org
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qedi: Remove unused declarations

These declarations were never implemented, remove them.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20230822143338.19120-1-yuehaibing@huawei.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qedf: Remove unused declaration

This declaration was never implemented, remove it.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20230822143338.19120-1-yuehaibing@huawei.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: xen-scsifront: shost_priv() can never return NULL

There is no need to check whether shost_priv() returns a non-NULL value, as
the pointer returned is just an offset to the passed in parameter.

While at it replace an open coded shost_priv() instance.

Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/20230822064817.27257-1-jgross@suse.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: raid_class: Remove raid_component_add()

The raid_component_add() function was added to the kernel tree via patch
"[SCSI] embryonic RAID class" (2005). Remove this function since it never
has had any callers in the Linux kernel. And also raid_component_release()
is only used in raid_component_add(), so it is also removed.

Signed-off-by: Zhu Wang <wangzhu9@huawei.com>
Link: https://lore.kernel.org/r/20230822015254.184270-1-wangzhu9@huawei.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Fixes: 04b5b5cb0136 ("scsi: core: Fix possible memory leak if device_add() fails")
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Merge patch series "libsas: Some tidy-up"

John Garry <john.g.garry@oracle.com> says:

This series tidies-up libsas a bit, including:
- delete structure(s) with only one member
- delete structure members which are only ever set
- delete structure members which are never set and code which relies on
  that member being set

This conflicts with the following series:
https://lore.kernel.org/linux-scsi/20230809132249.37948-1-yuehaibing@huawei.com/

Any conflict should be trivial to resolve.

Based on mkp-scsi staging at a18e81d17a7e ("scsi: ufs: ufs-pci: Add support for QEMU")

Link: https://lore.kernel.org/r/20230815115156.343535-1-john.g.garry@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Merge patch series "Returning FIS on success for CDL"

Igor Pylypiv <ipylypiv@google.com> says:

This patch series plumbs libata's request for a result taskfile
(ATA_QCFLAG_RESULT_TF) through libsas to pm80xx LLDD. Other libsas LLDDs
can start using the newly added return_fis_on_success as well, if needed.

For Command Duration Limits policy 0xD (command completes without an
error) libata needs FIS in order to detect the ATA_SENSE bit and read
the Sense Data for Successful NCQ Commands log (0Fh). pm80xx HBAs do
not return FIS on success by default, hence, the driver is updated to
set the RETFIS bit (Return FIS on good completion) when requested by
libsas.

Link: https://lore.kernel.org/r/20230819213040.1101044-1-ipylypiv@google.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Delete sas_ata_task.retry_count

Since libsas was introduced in commit 2908d778ab3e ("[SCSI] aic94xx: new
driver"), sas_ata_task.retry_count is never set, so delete it and the
reference in asd_build_ata_ascb().

Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230815115156.343535-11-john.g.garry@oracle.com
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Delete sas_ata_task.stp_affil_pol

Since libsas was introduced in commit 2908d778ab3e ("[SCSI] aic94xx: new
driver"), sas_ata_task.stp_affil_pol is never set, so delete it and the
reference in asd_build_ata_ascb().

Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230815115156.343535-10-john.g.garry@oracle.com
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Delete sas_ata_task.set_affil_pol

Since libsas was introduced in commit 2908d778ab3e ("[SCSI] aic94xx: new
driver"), sas_ata_task.set_affil_pol is never set, so delete it and the
reference in asd_build_ata_ascb().

Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230815115156.343535-9-john.g.garry@oracle.com
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Delete sas_ssp_task.task_prio

Since libsas was introduced in commit 2908d778ab3e ("[SCSI] aic94xx: new
driver"), sas_ssp_task.task_prio is never set, so delete it and any
references which depend on it being set (all of them).

Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230815115156.343535-8-john.g.garry@oracle.com
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Delete sas_ssp_task.enable_first_burst

Since libsas was introduced in commit 2908d778ab3e ("[SCSI] aic94xx: new
driver"), sas_ssp_task.enable_first_burst is never set, so delete it and
any references.

Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230815115156.343535-7-john.g.garry@oracle.com
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Delete sas_ssp_task.retry_count

Since libsas was introduced in commit 2908d778ab3e ("[SCSI] aic94xx: new
driver"), sas_ssp_task.retry_count is only ever set, so delete it.

The aic94xx driver also had its own retry_count definition in struct scb
sub-structs, which may have caused a mix-up.

Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230815115156.343535-6-john.g.garry@oracle.com
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Delete struct scsi_core

Since commit 79855d178557 ("libsas: remove task_collector mode"), struct
scsi_core only contains a reference to the shost. struct scsi_core is only
used in sas_ha_struct.core, so delete scsi_core and replace with a
reference to the shost there.

Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230815115156.343535-5-john.g.garry@oracle.com
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Delete enum sas_phy_type

enum sas_phy_type is used for asd_sas_phy.type, which is only ever set, so
delete this member and the enum.

Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230815115156.343535-4-john.g.garry@oracle.com
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Delete enum sas_class

enum sas_class prob would have been useful if function sas_show_class() was
ever implemented, which it wasn't.

enum sas_class is used as asd_sas_port.class and asd_sas_phy.class, which
are only ever set, so delete these members and the enum.

Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230815115156.343535-3-john.g.garry@oracle.com
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Delete sas_ha_struct.lldd_module

Since libsas was introduced in commit 2908d778ab3e ("[SCSI] aic94xx: new
driver"), sas_ha_struct.lldd_module has only ever been set, so remove it.

Struct scsi_host_template already has a reference to the LLD driver
module as to stop the driver being removed unexpectedly.

Signed-off-by: John Garry <john.g.garry@oracle.com>
Link: https://lore.kernel.org/r/20230815115156.343535-2-john.g.garry@oracle.com
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Update version to 10.02.09.100-k

Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230821130045.34850-10-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Revert "scsi: qla2xxx: Fix buffer overrun"

Revert due to Get PLOGI Template failed.
This reverts commit b68710a8094fdffe8dd4f7a82c82649f479bb453.

Cc: stable@vger.kernel.org
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230821130045.34850-9-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Fix smatch warn for qla_init_iocb_limit()

Fix indentation for warning reported by smatch:

drivers/scsi/qla2xxx/qla_init.c:4199 qla_init_iocb_limit() warn: inconsistent indenting

Fixes: efa74a62aaa2 ("scsi: qla2xxx: Adjust IOCB resource on qpair create")
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230821130045.34850-8-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Remove unsupported ql2xenabledif option

User accidently passed module parameter ql2xenabledif=1 which is
unsupported. However, driver still initialized which lead to guard tag
errors during device discovery.

Remove unsupported ql2xenabledif=1 option and validate the user input.

Cc: stable@vger.kernel.org
Signed-off-by: Manish Rangankar <mrangankar@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230821130045.34850-7-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Error code did not return to upper layer

TMF was returned with an error code. The error code was not preserved to be
returned to upper layer. Instead, the error code from the Marker was
returned.

Preserve error code from TMF and return it to upper layer.

Cc: stable@vger.kernel.org
Fixes: da7c21b72aa8 ("scsi: qla2xxx: Fix command flush during TMF")
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230821130045.34850-6-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Add logs for SFP temperature monitoring

Add logs for SFP Temperature Alert async event to check if laser is
enabled/disabled.

Signed-off-by: Bikash Hazarika <bhazarika@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230821130045.34850-5-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Fix firmware resource tracking

The storage was not draining I/Os and the work load was not spread out
across different CPUs evenly. This led to firmware resource counters
getting overrun on the busy CPU. This overrun prevented error recovery from
happening in a timely manner.

By switching the counter to atomic, it allows the count to be little more
accurate to prevent the overrun.

Cc: stable@vger.kernel.org
Fixes: da7c21b72aa8 ("scsi: qla2xxx: Fix command flush during TMF")
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230821130045.34850-4-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Flush mailbox commands on chip reset

Fix race condition between Interrupt thread and Chip reset thread in trying
to flush the same mailbox. With the race condition, the "ha->mbx_intr_comp"
will get an extra complete() call. The extra complete call create erroneous
mailbox timeout condition when the next mailbox is sent where the mailbox
call does not wait for interrupt to arrive. Instead, it advances without
waiting.

Add lock protection around the check for mailbox completion.

Cc: stable@vger.kernel.org
Fixes: b2000805a975 ("scsi: qla2xxx: Flush mailbox commands on chip reset")
Signed-off-by: Quinn Tran <quinn.tran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230821130045.34850-3-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Add Unsolicited LS Request and Response Support for NVMe

Introduce infrastructure in the driver to support the processing of
unsolicited LS (Link Service) requests. This will involve the utilization
of a new pass-up of unsolicited FC-NVMe request IOCB interface. Unsolicited
requests will be submitted to the NVMe transport layer through
nvme_fc_rcv_ls_req(). Any received LS responses, which are sent using
xmt_ls_rsp(), will be forwarded to the firmware through the existing
Pass-Through IOCB interface, responsible for sending FC-NVMe Link Service
requests and responses.

Signed-off-by: Manish Rangankar <mrangankar@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230821130045.34850-2-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Allow 32-byte CDBs

System crashes when a 32-byte CDB was sent to a non T10 PI disk:

[  177.143279]  ? qla2xxx_dif_start_scsi_mq+0xcd8/0xce0 [qla2xxx]
[  177.149165]  ? internal_add_timer+0x42/0x70
[  177.153372]  qla2xxx_mqueuecommand+0x207/0x2b0 [qla2xxx]
[  177.158730]  scsi_queue_rq+0x2b7/0xc00
[  177.162501]  blk_mq_dispatch_rq_list+0x3ea/0x7e0

Current code attempted to use CRC IOCB to send the command but failed.
Instead, type 6 IOCB should be used to send the I/O.

Clone existing type 6 IOCB code with addition of MQ support to allow
32-byte CDBs to go through.

Signed-off-by: Quinn Tran <qutran@marvell.com>
Cc: Laurence Oberman <loberman@redhat.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230817063132.21900-3-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Move resource to allow code reuse

dsd_list contains a list of dsd buffer resources allocated during traffic
time. It resides in the qla_hw_data location where some of the code is not
reusable.

Move this list to qpair to allow reuse by either single queue or multi
queue adapter / code.

Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230817063132.21900-2-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: ufs-qcom: Clear qunipro_g4_sel for HW major version > 5

The qunipro_g4_sel clear is also needed for new platforms with major
version > 5. Fix the version check to take this into account.

Fixes: 9c02aa24bf40 ("scsi: ufs: ufs-qcom: Clear qunipro_g4_sel for HW version major 5")
Acked-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Nitin Rawat <quic_nitirawa@quicinc.com>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://lore.kernel.org/r/20230821-topic-sm8x50-upstream-ufs-major-5-plus-v2-1-f42a4b712e58@linaro.org
Reviewed-by: "Bao D. Nguyen" <quic_nguyenb@quicinc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: target: Fix write perf due to unneeded throttling

The write back throttling (WBT) code checks if REQ_SYNC | REQ_IDLE is set
to determine if a write is O_DIRECT vs buffered. If the bits are not set
then it assumes it's a buffered write and will throttle LIO if we hit
certain metrics. LIO itself is not using the buffer cache and is doing
direct I/O, so this has us set the direct bits so we are not throttled.

When the initiator application is doing direct I/O this can greatly improve
performance. It depends on the backend device but we have seen where the
WBT code is throttling writes to only 20K IOPs with 4K I/Os when the device
can support 100K+.

Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20230817192902.346791-1-michael.christie@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: mcq: Fix the search/wrap around logic

The search and wrap around logic in the ufshcd_mcq_sqe_search() function
does not work correctly when the hwq's queue depth is not a power of two
number. Correct it so that any queue depth with a positive integer value
within the supported range would work.

Signed-off-by: "Bao D. Nguyen" <quic_nguyenb@quicinc.com>
Link: https://lore.kernel.org/r/ff49c15be205135ed3ec186f3086694c02867dbd.1692149603.git.quic_nguyenb@quicinc.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Fixes: 8d7290348992 ("scsi: ufs: mcq: Add supporting functions for MCQ abort")
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Do not abuse UUID APIs and LPFC_COMPRESS_VMID_SIZE

The lpfc_vmid_host_uuid is not defined as uuid_t and its usage is not the
same as for uuid_t operations (like exporting or importing).  Hence replace
call to uuid_is_null() by respective memchr_inv() without abusing casting.

With that, replace LPFC_COMPRESS_VMID_SIZE with plain number and respective
sizeof() to make code robust to changes in the future, if any.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20230818155452.875781-1-andriy.shevchenko@linux.intel.com
Reviewed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: pm8001: Remove unused declarations

Commit 4fcf812ca392 ("[SCSI] libsas: export sas_alloc_task()") removed
these implementations but not the declarations.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20230818124700.49724-1-yuehaibing@huawei.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: fcoe: Fix potential deadlock on &fip->ctlr_lock

There is a long call chain that &fip->ctlr_lock is acquired by isr
fnic_isr_msix_wq_copy() under hard IRQ context. Thus other process context
code acquiring the lock should disable IRQ, otherwise deadlock could happen
if the IRQ preempts the execution while the lock is held in process context
on the same CPU.

[ISR]
fnic_isr_msix_wq_copy()
 -> fnic_wq_copy_cmpl_handler()
 -> fnic_fcpio_cmpl_handler()
 -> fnic_fcpio_flogi_reg_cmpl_handler()
 -> fnic_flush_tx()
 -> fnic_send_frame()
 -> fcoe_ctlr_els_send()
 -> spin_lock_bh(&fip->ctlr_lock)

[Process Context]
1. fcoe_ctlr_timer_work()
 -> fcoe_ctlr_flogi_send()
 -> spin_lock_bh(&fip->ctlr_lock)

2. fcoe_ctlr_recv_work()
 -> fcoe_ctlr_recv_handler()
 -> fcoe_ctlr_recv_els()
 -> fcoe_ctlr_announce()
 -> spin_lock_bh(&fip->ctlr_lock)

3. fcoe_ctlr_recv_work()
 -> fcoe_ctlr_recv_handler()
 -> fcoe_ctlr_recv_els()
 -> fcoe_ctlr_flogi_retry()
 -> spin_lock_bh(&fip->ctlr_lock)

4. -> fcoe_xmit()
 -> fcoe_ctlr_els_send()
 -> spin_lock_bh(&fip->ctlr_lock)

spin_lock_bh() is not enough since fnic_isr_msix_wq_copy() is a
hardirq.

These flaws were found by an experimental static analysis tool I am
developing for irq-related deadlock.

The patch fix the potential deadlocks by spin_lock_irqsave() to disable
hard irq.

Fixes: 794d98e77f59 ("[SCSI] libfcoe: retry rejected FLOGI to another FCF if possible")
Signed-off-by: Chengfeng Ye <dg573847474@gmail.com>
Link: https://lore.kernel.org/r/20230817074708.7509-1-dg573847474@gmail.com
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: elx: sli4: Remove code duplication

In the function sli_xmit_bls_rsp64_wqe(), the 'if' and 'else' conditions
evaluates the same expression and give the same output. Also, params->s_id
shall not be equal to U32_MAX. Remove the unused code.

This fixes coccinelle warning such as:
drivers/scsi/elx/libefc_sli/sli4.c:2320:2-4: WARNING: possible
condition with no effect (if == else)

Signed-off-by: Rajeshwar R Shinde <coolrrsh@gmail.com>
Link: https://lore.kernel.org/r/20230817114301.17601-1-coolrrsh@gmail.com
Reviewed-by: Ram Vegesna <ram.vegesna@broadcom.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: bfa: Replace one-element array with flexible-array member in struct fc_rscn_pl_s

One-element and zero-length arrays are deprecated. So, replace one-element
array in struct fc_rscn_pl_s with flexible-array member.

This results in no differences in binary output.

Link: https://github.com/KSPP/linux/issues/339
Signed-off-by: "Gustavo A. R. Silva" <gustavoars@kernel.org>
Link: https://lore.kernel.org/r/ZN0VTpDBOSVHGayb@work
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla2xxx: Remove unused declarations

These declarations are not used anymore, remove them.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20230816130842.16684-1-yuehaibing@huawei.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: pmcraid: Use pci_dev_id() to simplify the code

PCI core API pci_dev_id() can be used to get the BDF number for a PCI
device. We don't need to compose it manually. Use pci_dev_id() to simplify
the code a little bit.

Signed-off-by: Zheng Zengkai <zhengzengkai@huawei.com>
Link: https://lore.kernel.org/r/20230811111310.32364-1-zhengzengkai@huawei.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: pm80xx: Set RETFIS when requested by libsas

By default PM80xx HBAs return FIS only when a drive reports an error.
The RETFIS bit forces the controller to populate FIS even when a drive
reports no error.

Signed-off-by: Igor Pylypiv <ipylypiv@google.com>
Link: https://lore.kernel.org/r/20230819213040.1101044-3-ipylypiv@google.com
Reviewed-by: Niklas Cassel <niklas.cassel@wdc.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Add return_fis_on_success to sas_ata_task

Set return_fis_on_success when libata requests result taskfile.

For Command Duration Limits policy 0xD (command completes without
an error) libata needs FIS in order to detect the ATA_SENSE bit and
read the Sense Data for Successful NCQ Commands log (0Fh).

Signed-off-by: Igor Pylypiv <ipylypiv@google.com>
Link: https://lore.kernel.org/r/20230819213040.1101044-2-ipylypiv@google.com
Reviewed-by: Niklas Cassel <niklas.cassel@wdc.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid: Use pci_dev_id() to simplify the code

PCI core API pci_dev_id() can be used to get the BDF number for a PCI
device. We don't need to compose it manually. Use pci_dev_id() to simplify
the code a little bit.

Signed-off-by: Jialin Zhang <zhangjialin11@huawei.com>
Link: https://lore.kernel.org/r/20230815025419.3523236-4-zhangjialin11@huawei.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: Use pci_dev_id() to simplify the code

PCI core API pci_dev_id() can be used to get the BDF number for a PCI
device. We don't need to compose it manually. Use pci_dev_id() to simplify
the code a little bit.

Signed-off-by: Jialin Zhang <zhangjialin11@huawei.com>
Link: https://lore.kernel.org/r/20230815025419.3523236-3-zhangjialin11@huawei.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mvumi: Use pci_dev_id() to simplify the code

PCI core API pci_dev_id() can be used to get the BDF number for a PCI
device. We don't need to compose it mannally. Use pci_dev_id() to simplify
the code a little bit.

Signed-off-by: Jialin Zhang <zhangjialin11@huawei.com>
Link: https://lore.kernel.org/r/20230815025419.3523236-2-zhangjialin11@huawei.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: host: Convert to dev_err_probe() in ufshcd_pltfrm_init()

Convert ufshcd_pltfrm_init() over to use dev_err_probe() to avoid the
following log message on bootup due to an -EPROBE_DEFER return code:

    ufshcd-qcom 1d84000.ufs: Initialization failed

While this line is changed, let's also go ahead and add the error code to
the message as well.

Signed-off-by: Brian Masney <bmasney@redhat.com>
Link: https://lore.kernel.org/r/20230814184352.200531-3-bmasney@redhat.com
Reviewed-by: "Bao D. Nguyen" <quic_nguyenb@quicinc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: core: Convert to dev_err_probe() in ufshcd_variant_hba_init()

Convert ufshcd_variant_hba_init() over to use dev_err_probe() to avoid log
messages like the following on bootup:

    ufshcd-qcom 1d84000.ufs: ufshcd_variant_hba_init: variant qcom init
        failed err -517

Signed-off-by: Brian Masney <bmasney@redhat.com>
Link: https://lore.kernel.org/r/20230814184352.200531-2-bmasney@redhat.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: Use 32-bit hostnum in scsi_host_lookup()

Change scsi_host_lookup() hostnum argument type from unsigned short to
unsigned int to match the type used everywhere else.

Fixes: 6d49f63b415c ("[SCSI] Make host_no an unsigned int")
Signed-off-by: Tony Battersby <tonyb@cybernetics.com>
Link: https://lore.kernel.org/r/a02497e7-c12b-ef15-47fc-3f0a0b00ffce@cybernetics.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: isci: Return result of sas_register_ha()

To properly manage possible failure of sas_register_ha() in
isci_register_sas_ha(), return its result instead of zero

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Signed-off-by: Artem Chernyshev <artem.chernyshev@red-soft.ru>
Link: https://lore.kernel.org/r/20230813202336.240874-1-artem.chernyshev@red-soft.ru
Reviewed-by: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: gvp11: Remove unused gvp11_setup() function

This function has no declaration, which causes a warning:

drivers/scsi/gvp11.c:53:6: error: no previous prototype for 'gvp11_setup' [-Werror=missing-prototypes]

Since there is also no caller, just remove the function.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20230810141947.1236730-12-arnd@kernel.org
Reviewed-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Palmer Dabbelt <palmer@rivosinc.com> # RISC-V
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qlogicpti: Mark qlogicpti_info() static

The qlogicpti_info() function is only used in this file and should be
static to avoid a warning:

drivers/scsi/qlogicpti.c:846:13: error: no previous prototype for 'qlogicpti_info' [-Werror=missing-prototypes]

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20230810141947.1236730-8-arnd@kernel.org
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Acked-by: Palmer Dabbelt <palmer@rivosinc.com> # RISC-V
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ppa: Add a module parameter for the transfer mode

I have an Iomega Z100P2 zip drive, but it does not work with my StarTech
PEX1P2 AX99100 PCIe parallel port, which evidently does not support 16-bit
or 32-bit EPP. Currently the only way to tell the PPA driver to use 8-bit
EPP is to write 'mode=3' to /proc/scsi/ppa/*, but the driver doesn't
actually distinguish between the three EPP modes and still tries to use
16-bit or 32-bit EPP. And even if writing to that file did make the driver
use 8-bit EPP, it still wouldn't do me any good because by the time that
file exists, the drive has already failed to initialize.

Add a new parameter /sys/module/ppa/mode to set the transfer mode before
initializing the drive. This parameter replaces the use of
CONFIG_SCSI_IZIP_EPP16 in the PPA driver.

At the same time, default to 8-bit EPP. 16-bit and 32-bit EPP are not
necessary for the drive to function, nor are they part of the IEEE 1284
standard, so the driver should not assume that they are available.

Signed-off-by: Alex Henrie <alexhenrie24@gmail.com>
Link: https://lore.kernel.org/r/20230807155856.362864-2-alexhenrie24@gmail.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ppa: Fix compilation with PPA_DEBUG=1

Fix a regression introduced in February 2003 in Linux 2.5.61 by a patch
from Alan Cox titled "fix ppa for new scsi".[1]

Link: https://lore.kernel.org/lkml/E18jn1B-0005gQ-00@the-village.bc.nu/
Signed-off-by: Alex Henrie <alexhenrie24@gmail.com>
Link: https://lore.kernel.org/r/20230807155856.362864-1-alexhenrie24@gmail.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: arcmsr: Add __init and __exit for arcmsr_module_{init,exit}()

Add __init and __exit for arcmsr_module_{init,exit}().

Signed-off-by: Xiang Yang <xiangyang3@huawei.com>
Link: https://lore.kernel.org/r/20230804022913.1917023-1-xiangyang3@huawei.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: Remove unused extern declarations

These functions have never been implemented since the beginning of git
history.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20230809142107.42756-1-yuehaibing@huawei.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: Remove unused declarations

Commit 042ebd293b86 ("scsi: libsas: kill useless ha_event and do some
cleanup") removed sas_hae_reset() but not its declaration.  Commit
2908d778ab3e ("[SCSI] aic94xx: new driver") declared but never implemented
other functions.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20230809132249.37948-1-yuehaibing@huawei.com
Reviewed-by: John Garry <john.g.garry@oracle.com>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: ufs-pci: Add support for QEMU

To ensure that the PCI based QEMU UFS device properly works with Linux,
register the device ID (0x0013) and vendor ID (0x1b36) of QEMU UFS device.

QEMU UFS will enable testing of the UFS driver inside a virtual machine on
systems without UFS host controller. It can also be used to preemptively
implement and test new features before the real device is created.

The new QEMU UFS device can be found at:

https://lore.kernel.org/qemu-devel/20230727155239.GA979354@fedora

Signed-off-by: Jeuk Kim <jeuk20.kim@samsung.com>
Link: https://lore.kernel.org/r/20230807013726epcms2p1c604cb8e98680aebebb7cc5ab2d580f5@epcms2p1
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Merge patch series "mpi3mr: Few Enhancements and minor fixes"

Ranjan Kumar <ranjan.kumar@broadcom.com> says:

Few Enhancements and minor fixes of mpi3mr driver.

Link: https://lore.kernel.org/r/20230804104248.118924-1-ranjan.kumar@broadcom.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Update driver version to 8.5.0.0.0

Update driver version to 8.5.0.0.0

Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
Link: https://lore.kernel.org/r/20230804104248.118924-7-ranjan.kumar@broadcom.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Enhance handling of devices removed after controller reset

Mark all of the devices that are exposed to the OS prior to a controller
reset and not detected by the controller after the reset as removed devices
and the I/Os to those devices are unblocked (and returned with
DID_NO_CONNECT) prior to removing the devices one after the other.

Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
Link: https://lore.kernel.org/r/20230804104248.118924-6-ranjan.kumar@broadcom.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: WRITE SAME implementation

Enhance driver to divert the WRITE SAME commands that are issued with
UNMAP=1 and NDOB=1 and with the transfer length greater than the max WRITE
SAME length specified by the firmware for the particular drive to the
controller firmware.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202307280034.DXU5pTVV-lkp@intel.com/
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
Link: https://lore.kernel.org/r/20230804104248.118924-5-ranjan.kumar@broadcom.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Add support for more than 1MB I/O

Enhance the driver to get the maximum data length per I/O request from IOC
Facts data and report that to the upper layers.  If the IOC facts data is
not reported then a default I/O size of 1MB is reported to the OS.

Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
Link: https://lore.kernel.org/r/20230804104248.118924-4-ranjan.kumar@broadcom.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Update MPI Headers to version 3.00.28

Updated MPI Headers to version 3.00.28.

Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
Link: https://lore.kernel.org/r/20230804104248.118924-3-ranjan.kumar@broadcom.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpi3mr: Invoke soft reset upon TSU or event ack time out

When a timestamp update or an event acknowledgment command times out, the
driver invokes the soft reset handler to recover the controller while
holding a mutex lock. The soft reset handler also tries to acquire the same
mutex to send initialization commands to the controller which leads to a
deadlock scenario.

To resolve the issue the driver will check thestatus and if this indicates
the controller is operational, the driver will issue a diagnostic fault
reset and exit out of the command processing function. If the controller is
already faulted or asynchronously reset, then the driver will just exit the
command processing function.

Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
Link: https://lore.kernel.org/r/20230804104248.118924-2-ranjan.kumar@broadcom.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qedf: Fix firmware halt over suspend and resume

While performing certain power-off sequences, PCI drivers are called to
suspend and resume their underlying devices through PCI PM (power
management) interface. However the hardware does not support PCI PM
suspend/resume operations so system wide suspend/resume leads to bad MFW
(management firmware) state which causes various follow-up errors in driver
when communicating with the device/firmware.

To fix this driver implements PCI PM suspend handler to indicate
unsupported operation to the PCI subsystem explicitly, thus avoiding system
to go into suspended/standby mode.

Fixes: 61d8658b4a43 ("scsi: qedf: Add QLogic FastLinQ offload FCoE driver framework.")
Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230807093725.46829-1-njavali@marvell.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qedi: Fix firmware halt over suspend and resume

While performing certain power-off sequences, PCI drivers are called to
suspend and resume their underlying devices through PCI PM (power
management) interface. However the hardware does not support PCI PM
suspend/resume operations so system wide suspend/resume leads to bad MFW
(management firmware) state which causes various follow-up errors in driver
when communicating with the device/firmware.

To fix this driver implements PCI PM suspend handler to indicate
unsupported operation to the PCI subsystem explicitly, thus avoiding system
to go into suspended/standby mode.

Fixes: ace7f46ba5fd ("scsi: qedi: Add QLogic FastLinQ offload iSCSI driver framework.")
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Link: https://lore.kernel.org/r/20230807093725.46829-2-njavali@marvell.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qedi: Fix potential deadlock on &qedi_percpu->p_work_lock

As &qedi_percpu->p_work_lock is acquired by hard IRQ qedi_msix_handler(),
other acquisitions of the same lock under process context should disable
IRQ, otherwise deadlock could happen if the IRQ preempts the execution
while the lock is held in process context on the same CPU.

qedi_cpu_offline() is one such function which acquires the lock in process
context.

[Deadlock Scenario]
qedi_cpu_offline()
    ->spin_lock(&p->p_work_lock)
        <irq>
        ->qedi_msix_handler()
        ->edi_process_completions()
        ->spin_lock_irqsave(&p->p_work_lock, flags); (deadlock here)

This flaw was found by an experimental static analysis tool I am developing
for IRQ-related deadlocks.

The tentative patch fix the potential deadlock by spin_lock_irqsave()
under process context.

Signed-off-by: Chengfeng Ye <dg573847474@gmail.com>
Link: https://lore.kernel.org/r/20230726125655.4197-1-dg573847474@gmail.com
Acked-by: Manish Rangankar <mrangankar@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Remove reftag check in DIF paths

When preparing protection DIF I/O for DMA, the driver obtains reference
tags from scsi_prot_ref_tag().  Previously, there was a wrong assumption
that an all 0xffffffff value meant error and thus the driver failed the
I/O.  This patch removes the evaluation code and accepts whatever the upper
layer returns.

Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Link: https://lore.kernel.org/r/20230803211932.155745-1-justintee8345@gmail.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: renesas: Fix private allocation

Should use devm_kzalloc() for struct ufs_renesas_priv because the
.initialized should be false as default.

Fixes: d69520288efd ("scsi: ufs: ufs-renesas: Add support for Renesas R-Car UFS controller")
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Link: https://lore.kernel.org/r/20230803081812.1446282-1-yoshihiro.shimoda.uh@renesas.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: snic: Fix possible memory leak if device_add() fails

If device_add() returns error, the name allocated by dev_set_name() needs
be freed. As the comment of device_add() says, put_device() should be used
to give up the reference in the error path. So fix this by calling
put_device(), then the name can be freed in kobject_cleanp().

Fixes: c8806b6c9e82 ("snic: driver for Cisco SCSI HBA")
Signed-off-by: Zhu Wang <wangzhu9@huawei.com>
Acked-by: Narsimhulu Musini <nmusini@cisco.com>
Link: https://lore.kernel.org/r/20230801111421.63651-1-wangzhu9@huawei.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: Fix possible memory leak if device_add() fails

If device_add() returns error, the name allocated by dev_set_name() needs
be freed. As the comment of device_add() says, put_device() should be used
to decrease the reference count in the error path. So fix this by calling
put_device(), then the name can be freed in kobject_cleanp().

Fixes: ee959b00c335 ("SCSI: convert struct class_device to struct device")
Signed-off-by: Zhu Wang <wangzhu9@huawei.com>
Link: https://lore.kernel.org/r/20230803020230.226903-1-wangzhu9@huawei.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Modify when a node should be put in device recovery mode during RSCN

Only nodes whose state is at least past a PLOGI issue and strictly less
than a PRLI issue should be put into device recovery mode upon RSCN
receipt.  Previously, the allowance of LOGO and PRLI completion states did
not make sense because those nodes should be allowed to flow through and
marked as NPort dissappeared as is normally done.  A follow up RSCN GID_FT
would recover those nodes in such cases.

Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Link: https://lore.kernel.org/r/20230804195546.157839-1-justintee8345@gmail.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: qcom: Make struct ufs_qcom_bw_table static const

ufs_qcom_bw_table is not modified anywhere. So make it static const so that
it can be placed in read-only memory.

Reported-by: Bart Van Assche <bvanassche@acm.org>
Closes: https://lore.kernel.org/linux-scsi/43cd0057-c6d8-bc92-08f4-d767336d2cfe@acm.org/
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Link: https://lore.kernel.org/r/20230802040154.10652-1-manivannan.sadhasivam@linaro.org
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: core: Fix the build for gcc 9 and before

gcc compilers before version 10 cannot do constant-folding for sub-byte
bitfields. This makes the compiler layout tests fail. Hence skip the layout
checks for gcc 9 and before.

Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Naresh Kamboju <naresh.kamboju@linaro.org>
Cc: Nathan Chancellor <nathan@kernel.org>
Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Closes: https://lore.kernel.org/linux-scsi/CA+G9fYur8UJoUyTLJFVEJPh-15TJ7kbdD2q8xVz8a3fLjkxxVw@mail.gmail.com/
Suggested-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20230801232204.1481902-1-bvanassche@acm.org
Tested-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: ufs-qcom: Check host controller state

Commit 52a518019ca1 ("scsi: ufs: core: Fix missing clk change notification
on host reset") added UFS clock scaling notification to
ufshcd_host_reset_and_restore(). This invokes hibern8 enter and exit on
Qualcomm platform which fails because controller is in reset state.

Fix this by checking the Host controller state before sending hibern8
command.

__ufshcd_wl_resume()
ufshcd_reset_and_restore()
ufshcd_host_reset_and_restore()
ufshcd_scale_clks()
ufshcd_vops_clk_scale_notify()
ufs_qcom_clk_scale_notify()
ufshcd_uic_hibern8_enter()

Fixes: 52a518019ca1 ("scsi: ufs: core: Fix missing clk change notification on host reset")
Co-developed-by: Manish Pandey <quic_mapa@quicinc.com>
Signed-off-by: Manish Pandey <quic_mapa@quicinc.com>
Signed-off-by: Nitin Rawat <quic_nitirawa@quicinc.com>
Link: https://lore.kernel.org/r/20230726134140.7180-3-quic_nitirawa@quicinc.com
Acked-by: Manivannan Sadhasivam <mani@kernel.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: core: Export ufshcd_is_hba_active()

Export ufshcd_is_hba_active() to allow driver modules to check the state of
the host controller.

Signed-off-by: Nitin Rawat <quic_nitirawa@quicinc.com>
Link: https://lore.kernel.org/r/20230726134140.7180-2-quic_nitirawa@quicinc.com
Acked-by: Manivannan Sadhasivam <mani@kernel.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: core: Fix legacy /proc parsing buffer overflow

(lightly modified commit message mostly by Linus Torvalds)

The parsing code for /proc/scsi/scsi is disgusting and broken.  We should
have just used 'sscanf()' or something simple like that, but the logic may
actually predate our kernel sscanf library routine for all I know.  It
certainly predates both git and BK histories.

And we can't change it to be something sane like that now, because the
string matching at the start is done case-insensitively, and the separator
parsing between numbers isn't done at all, so *any* separator will work,
including a possible terminating NUL character.

This interface is root-only, and entirely for legacy use, so there is
absolutely no point in trying to tighten up the parsing.  Because any
separator has traditionally worked, it's entirely possible that people have
used random characters rather than the suggested space.

So don't bother to try to pretty it up, and let's just make a minimal patch
that can be back-ported and we can forget about this whole sorry thing for
another two decades.

Just make it at least not read past the end of the supplied data.

Link: https://lore.kernel.org/linux-scsi/b570f5fe-cb7c-863a-6ed9-f6774c219b88@cybernetics.com/
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Martin K Petersen <martin.petersen@oracle.com>
Cc: James Bottomley <jejb@linux.ibm.com>
Cc: Willy Tarreau <w@1wt.eu>
Cc: stable@kernel.org
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Tony Battersby <tonyb@cybernetics.com>
Signed-off-by: Martin K Petersen <martin.petersen@oracle.com>

Merge patch series "Multiple cleanup patches for the UFS driver"

Bart Van Assche <bvanassche@acm.org> says:

Hi Martin,

This patch includes the following changes, none of which should change the
functionality of the UFS host controller driver:
- Improve the kernel-doc headers further.
- Fix multiple W=2 compiler warnings.
- Simplify ufshcd_abort_all().
- Simplify the code for creating and parsing UFS Transport Protocol (UTP)
  headers.

Please consider this patch series for the next merge window.

Thanks,

Bart.

Link: https://lore.kernel.org/r/20230727194457.3152309-1-bvanassche@acm.org
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: Simplify response header parsing

Make the code that parses UTP transfer request headers easier to read by
using u8 instead of __be32 where appropriate.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20230727194457.3152309-13-bvanassche@acm.org
Reviewed-by: Avri Altman <avri.altman@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: Simplify transfer request header initialization

Make the code that initializes UTP transfer request headers easier to read
by using bitfields instead of __le32 where appropriate.

Cc: "Bao D. Nguyen" <quic_nguyenb@quicinc.com>
Cc: Eric Biggers <ebiggers@google.com>
Cc: Avri Altman <avri.altman@wdc.com>
Cc: Bean Huo <beanhuo@micron.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20230727194457.3152309-12-bvanassche@acm.org
Reviewed-by: Avri Altman <avri.altman@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: Remove a member variable

Remove the 'response' member variable because no code reads its value.
Additionally, move the ufs_query_req and ufs_query_res data structure
definitions into include/ufs/ufshcd.h because these data structures are
related to the UFS host controller driver.

Reviewed-by: Avri Altman <avri.altman@wdc.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20230727194457.3152309-11-bvanassche@acm.org
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: Simplify ufshcd_abort_all()

Unify the MCQ and legacy code paths. This patch reworks code introduced by
commit ab248643d3d6 ("scsi: ufs: core: Add error handling for MCQ mode").

Cc: "Bao D. Nguyen" <quic_nguyenb@quicinc.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20230727194457.3152309-10-bvanassche@acm.org
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>