As a performance enhancement, make the completion queue interrupts managed.

In addition, in commit bf0beec0 ("blk-mq: drain I/O when all CPUs in a
hctx are offline"), CPU hotplug for MQ devices using managed interrupts is
made safe. So expose HW queues to blk-mq to take advantage of this.

Flag Scsi_host.host_tagset is also set to ensure that the HBA is not sent
more commands than it can handle. However the driver still does not use
request tag for IPTT as there are many HW bugs means that special rules
apply for IPTT allocation.

Link: https://lore.kernel.org/r/1606905417-183214-6-git-send-email-john.garry@huawei.comSigned-off-by: NJohn Garry <john.garry@huawei.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

This reverts commit 1a0e1943.

Commit b3c6a599 ("block: Fix a lockdep complaint triggered by
request queue flushing") has been reverted and commit fb01a293 has
been introduced in its place. Consequently, it is now safe to
reinstate the megaraid_sas tagset changes that led to boot problems in
5.10.
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The in_interrupt() check in sr_init_command() is a leftover from the
past, pre v2.3.16 era to be exact. Back then the ioctl() was served by
`sr' itself and sector size changes by CDROMREADMODE2 (as noted in the
comment) were accounted within sr's data structures which allowed a
"lazy" reset so it could be skipped on the next request and reset back
to the default value once the device node was closed or before a command
from the blockqueue was issued.

This does not work like that anymore. The CDROMREADMODE2 is served by
cdrom's mmc_ioctl() function which may change the sector size but the
`sr' driver does not learn about it and so its ->sector_size is not
updated.
The ioctl() resets the changed sector size back to 2048.
sr_read_sector() also resets the sector size back to the default once it
is done.

Remove the conditional sector size update from sr_init_command() and
sr_release() because it is not needed.

Link: https://lkml.kernel.org/r/20201204164803.ovwurzs3257em2rp@linutronix.deSigned-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

sr_read_sector() is hardly used since v2.3.16. Its only purpose is to
check if it is a XA medium via sr_is_xa(). This check is only enabled if
the module parameter `xa_test' is enabled.

Change the sector size back to 2048 if it was changed. With this change,
there is no lazy sector size changing left.
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

This reverts commit 3b8c72d0.

Dexuan reported a regression where StorVSC fails to probe a device (and
where, consequently, the VM may fail to boot).  The root-cause analysis led
to a long-standing race condition that is exposed by the validation /commit
in question.  Let's put the new validation aside until a proper solution
for that race condition is in place.

Link: https://lore.kernel.org/r/20201211131404.21359-1-parri.andrea@gmail.com
Fixes: 3b8c72d0 ("scsi: storvsc: Validate length of incoming packet in storvsc_on_channel_callback()")
Cc: Dexuan Cui <decui@microsoft.com>
Cc: "James E.J. Bottomley" <jejb@linux.ibm.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
Cc: linux-scsi@vger.kernel.org
Signed-off-by: NAndrea Parri (Microsoft) <parri.andrea@gmail.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

SCSI_CXGB4_ISCSI selects CHELSIO_T4. The latter depends on TLS || TLS=n, so
since 'select' does not check dependencies of the selected symbol,
SCSI_CXGB4_ISCSI should also depend on TLS || TLS=n.

This prevents the following kconfig warning and restricts SCSI_CXGB4_ISCSI
to 'm' whenever TLS=m.

WARNING: unmet direct dependencies detected for CHELSIO_T4
  Depends on [m]: NETDEVICES [=y] && ETHERNET [=y] && NET_VENDOR_CHELSIO [=y] && PCI [=y] && (IPV6 [=y] || IPV6 [=y]=n) && (TLS [=m] || TLS [=m]=n)
  Selected by [y]:
  - SCSI_CXGB4_ISCSI [=y] && SCSI_LOWLEVEL [=y] && SCSI [=y] && PCI [=y] && INET [=y] && (IPV6 [=y] || IPV6 [=y]=n) && ETHERNET [=y]

Link: https://lore.kernel.org/r/20201208220505.24488-1-rdunlap@infradead.org
Fixes: 7b36b6e0 ("[SCSI] cxgb4i v5: iscsi driver")
Cc: Karen Xie <kxie@chelsio.com>
Cc: linux-scsi@vger.kernel.org
Cc: "James E.J. Bottomley" <jejb@linux.ibm.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
Signed-off-by: NRandy Dunlap <rdunlap@infradead.org>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

More and more statements are being added to ufshcd_vops_device_reset() and
this function is being called from multiple locations in the driver.
Un-inline the function to allow the compiler to make better decisions.

Link: https://lore.kernel.org/r/20201208135635.15326-3-stanley.chu@mediatek.comReviewed-by: NBean Huo <beanhuo@micron.com>
Signed-off-by: NStanley Chu <stanley.chu@mediatek.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

UFS 3.1 specification mentions that the WriteBooster flags listed below
will be set to their default values, i.e. disabled, after power cycle or
any type of reset event. Thus we need to reset the flag variables kept in
struct hba to align with the device status and ensure that
WriteBooster-related functions are configured properly after device reset.

Without this fix, WriteBooster will not be enabled successfully after by
ufshcd_wb_ctrl() after device reset because hba->wb_enabled remains true.

Flags required to be reset to default values:

 - fWriteBoosterEn: hba->wb_enabled

 - fWriteBoosterBufferFlushEn: hba->wb_buf_flush_enabled

 - fWriteBoosterBufferFlushDuringHibernate: No variable mapped

Link: https://lore.kernel.org/r/20201208135635.15326-2-stanley.chu@mediatek.com
Fixes: 3d17b9b5 ("scsi: ufs: Add write booster feature support")
Reviewed-by: NBean Huo <beanhuo@micron.com>
Signed-off-by: NStanley Chu <stanley.chu@mediatek.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

When the kernel is compiled with allmodconfig, the following error is
reported:

In file included from drivers/scsi/ufs/ufs-mediatek-trace.h:36:0,
                 from drivers/scsi/ufs/ufs-mediatek.c:28:
./include/trace/define_trace.h:95:42: fatal error: ./ufs-mediatek-trace.h: No such file or directory
 #include TRACE_INCLUDE(TRACE_INCLUDE_FILE)

The comment in include/trace/define_trace.h specifies that:
TRACE_INCLUDE_PATH: Note, the path is relative to define_trace.h, not the
file including it. Full path names for out of tree modules must be used.

So without "CFLAGS_ufs-mediatek.o := -I$(src)", the current directory "."
is "include/trace/", the relative path of ufs-mediatek-trace.h is
"../../drivers/scsi/ufs/".

Link: https://lore.kernel.org/r/20201209063144.1840-2-thunder.leizhen@huawei.com
Fixes: ca1bb061 ("scsi: ufs-mediatek: Introduce event_notify implementation")
Reviewed-by: NStanley Chu <stanley.chu@mediatek.com>
Signed-off-by: NZhen Lei <thunder.leizhen@huawei.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The "trigger_flags" variable needs to be signed for the error checking to
work.

Link: https://lore.kernel.org/r/X9DZH37bYPHwSQRP@mwanda
Fixes: aec93e8e ("scsi: mpt3sas: Add persistent trigger pages support")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Instead of submitting all SCSI commands submitted with scsi_execute() to a
SCSI device if rpm_status != RPM_ACTIVE, only submit RQF_PM (power
management requests) if rpm_status != RPM_ACTIVE. This patch makes the SCSI
core handle the runtime power management status (rpm_status) as it should
be handled.

Link: https://lore.kernel.org/r/20201209052951.16136-7-bvanassche@acm.org
Cc: Can Guo <cang@codeaurora.org>
Cc: Stanley Chu <stanley.chu@mediatek.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Martin Kepplinger <martin.kepplinger@puri.sm>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NJens Axboe <axboe@kernel.dk>
Reviewed-by: NCan Guo <cang@codeaurora.org>
Signed-off-by: NBart Van Assche <bvanassche@acm.org>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Disable runtime power management during domain validation. Since a later
patch removes RQF_PREEMPT, set RQF_PM for domain validation commands such
that these are executed in the quiesced SCSI device state.

Link: https://lore.kernel.org/r/20201209052951.16136-6-bvanassche@acm.org
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: James Bottomley <James.Bottomley@HansenPartnership.com>
Cc: Woody Suwalski <terraluna977@gmail.com>
Cc: Can Guo <cang@codeaurora.org>
Cc: Stanley Chu <stanley.chu@mediatek.com>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Stan Johnson <userm57@yahoo.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NJens Axboe <axboe@kernel.dk>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NBart Van Assche <bvanassche@acm.org>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Enable runtime PM auto-suspend by default for Intel host controllers.

Link: https://lore.kernel.org/r/20201207083120.26732-5-adrian.hunter@intel.comSigned-off-by: NAdrian Hunter <adrian.hunter@intel.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Intel controllers can end up in an unrecoverable state after a hibernate
exit error unless a full reset and restore is done before anything else.
Force that to happen.

Link: https://lore.kernel.org/r/20201207083120.26732-4-adrian.hunter@intel.comSigned-off-by: NAdrian Hunter <adrian.hunter@intel.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The expectation for suspend-to-disk is that devices will be powered-off, so
the UFS device should be put in PowerDown mode. If spm_lvl is not 5, then
that will not happen. Change the pm callbacks to force spm_lvl 5 for
suspend-to-disk poweroff.

Link: https://lore.kernel.org/r/20201207083120.26732-3-adrian.hunter@intel.comSigned-off-by: NAdrian Hunter <adrian.hunter@intel.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Currently, ufshcd-pci is the only UFS driver with support for
suspend-to-disk PM callbacks (i.e. freeze/thaw/restore/poweroff). These
callbacks are set by the macro SET_SYSTEM_SLEEP_PM_OPS to the same
functions as system suspend/resume. That will work with spm_lvl 5 because
spm_lvl 5 will result in a full restore for the ->restore() callback.  In
the absence of a full restore, the host controller registers will have
values set up by the restore kernel (the kernel that boots and loads the
restore image) which are not necessarily the same. However it turns out,
the only registers that sometimes need restore are the base address
registers. This has gone un-noticed because, depending on IOMMU settings,
the kernel can end up allocating the same addresses every time.

For Intel controllers, an spm_lvl other than 5 can be used, so to support
S4 (suspend-to-disk) with spm_lvl other than 5, restore the base address
registers.

Link: https://lore.kernel.org/r/20201207083120.26732-2-adrian.hunter@intel.comSigned-off-by: NAdrian Hunter <adrian.hunter@intel.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

For some devices which need extra delay after VCC power down, VCC shall be
kept always-on in some MediaTek UFS platforms to ensure the stability of
such devices because the extra delay may not be enough in those platforms.

Link: https://lore.kernel.org/r/20201207054955.24366-3-stanley.chu@mediatek.comReviewed-by: NAndy Teng <andy.teng@mediatek.com>
Signed-off-by: NStanley Chu <stanley.chu@mediatek.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Update driver version to 36.100.00.00

Link: https://lore.kernel.org/r/20201126094311.8686-9-suganath-prabu.subramani@broadcom.comSigned-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Introduce a flag "always_on" in struct ufs_vreg to allow vendors to keep
the regulator always-on.

Link: https://lore.kernel.org/r/20201207054955.24366-2-stanley.chu@mediatek.comReviewed-by: NAndy Teng <andy.teng@mediatek.com>
Signed-off-by: NStanley Chu <stanley.chu@mediatek.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

If a firmware update adds support for the trigger pages, then the driver
should handle this by writing the existing trigger data from the driver's
internal data structure to the corresponding trigger pages in NVRAM.

Also handle the case where the trigger page capability is no longer present
after a firmware downgrade.

Link: https://lore.kernel.org/r/20201126094311.8686-8-suganath-prabu.subramani@broadcom.comSigned-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

If RPMB is not provisioned, we may see RPMB failure after UFS
suspend/resume.  Inject request_sense to clear uac in ufshcd reset flow.

Link: https://lore.kernel.org/r/20201201041402.3860525-1-jaegeuk@kernel.orgReported-by: Nkernel test robot <lkp@intel.com>
Reviewed-by: NStanley Chu <stanley.chu@mediatek.com>
Signed-off-by: NRandall Huang <huangrandall@google.com>
Signed-off-by: NLeo Liou <leoliou@google.com>
Signed-off-by: NJaegeuk Kim <jaegeuk@google.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

This page is used to store information about MPI (IOC Status & LogInfo)
triggers.

 Driver Persistent Trigger Page-4 format:
 -------------------------------------------------------
 | 31       24 23           16 15         8 7          0|  Byte
 -------------------------------------------------------
 | PageType   | PageNumber    | Reserved  | PageVersion |  0x00
 --------------------------------------------------------
 | Reserved   | ExtPageType   |      ExtPageLength      |  0x04
 --------------------------------------------------------
 |          Reserved          | NumMpiTriggerEntries    |  0x08
 --------------------------------------------------------
 |             MPITriggerEntry[0]                       |  0x0C
 --------------------------------------------------------
 |               …                                      |
 --------------------------------------------------------
 |            MPITriggerEntry[19]                       |  0xA4
 --------------------------------------------------------

NumMpiTriggerEntries:

This field indicates number of MPI (IOC Status & LogInfo) trigger entries
stored in this page. Currently driver is supporting a maximum of 20-MPI
trigger entries.

MPITriggerEntry:

 -----------------------------------------------------
 | 31                    16 15                     0 |
 -----------------------------------------------------
 |        Reserved         |      IOCStatus          |
 -----------------------------------------------------
 |                   IOCLogInfo                      |
 -----------------------------------------------------

IOCStatus  => Status value from the IOC
IOCLogInfo => Specific value that supplements the IOCStatus.

Link: https://lore.kernel.org/r/20201126094311.8686-7-suganath-prabu.subramani@broadcom.comReported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Change dev_err() print message from "dme-reset" to "dme_enable" in function
ufshcd_dme_enable().

Link: https://lore.kernel.org/r/20201207190137.6858-3-huobean@gmail.comAcked-by: NAlim Akhtar <alim.akhtar@samsung.com>
Acked-by: NAvri Altman <avri.altman@wdc.com>
Signed-off-by: NBean Huo <beanhuo@micron.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Trigger Page3 is used to store information about SCSI Sense triggers:

 Persistent Trigger Page-3
 ------------------------------------------------------------------
 | 31         24 23            16 15              8 7            0|    Byte
 ------------------------------------------------------------------
 | PageType      | PageNumber     | Reserved       | PageVersion  |    0x00
 ------------------------------------------------------------------
 | Reserved      | ExtPageType    |          ExtPageLen           |    0x04
 ------------------------------------------------------------------
 | Reserved      | NumScsiSense   |         TriggerEntries        |    0x08
 ------------------------------------------------------------------
 |               ScsiSenseTriggerEntry[0]                         |    0x0C
 ------------------------------------------------------------------
 |                    …        …                                  |
 ------------------------------------------------------------------
 |               ScsiSenseTriggerEntry[19]                        |    0x58
 ------------------------------------------------------------------

NumScsiSenseTriggerEntries:

This field indicates number of SCSI Sense trigger entries stored in this
page. Currently driver is supporting a maximum of 20-SCSI Sense trigger
entries.

ScsiSenseTriggerEntry:

 -----------------------------------------------
 | 31      24 23       16 15       8 7       0 |
 -----------------------------------------------
 | Reserved   | SenseKey  |    ASC   |   ASCQ  |
 -----------------------------------------------

 ASCQ     => Additional Sense Code Qualifier
 ASC      => Additional Sense Code
 SenseKey => Sense Key values

 ASCQ     => Additional Sense Code Qualifier
 ASC      => Additional Sense Code
 SenseKey => Sense Key values

Link: https://lore.kernel.org/r/20201126094311.8686-6-suganath-prabu.subramani@broadcom.comReported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

POWER_DESC_MAX_SIZE is unused, remove it.

Link: https://lore.kernel.org/r/20201207190137.6858-2-huobean@gmail.comAcked-by: NAvri Altman <avri.altman@wdc.com>
Acked-by: NAlim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: NBean Huo <beanhuo@micron.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Trigger Page2 is used to store information about Event triggers:

  31     24 23        16 15      8 7         0   Byte
 -----------------------------------------------
 |PageType  |PageNumber  |Reserved |PageVersion| 0x00
 -----------------------------------------------
 |Reserved  |ExtPageType |   ExtPageLength     | 0x04
 -----------------------------------------------
 |     Reserved          | NumMPIEventTriggers | 0x08
 -----------------------------------------------
 |                 MPIEventTriggerEntries      | 0x0C
 |                                             | 0xFC
 -----------------------------------------------

Number of MPI Event Trigger Entries currently stored in this page.  If this
is set to zero, there are no valid MPI-Event-Trigger entries available in
this page.

MPIEventTriggerEntry:

 - MPIEventCode [15:00]
   MPI Event code specified in MPI-Spec

 - MPIEventCodeSpecific [16:31]
   For Event Code “MPI2_EVENT_LOG_ENTRY_ADDED (0x0021)”,
   this field specifies the Log-Entry-Qualifier.
   For all other Event Codes, this field is reserved and not used

Maximum of 20-event trigger entries can be stored in this page.

Link: https://lore.kernel.org/r/20201126094311.8686-5-suganath-prabu.subramani@broadcom.comReported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Trigger Page 1 is used to store information about Master triggers.  Below
are the Master trigger conditions:

  Bit[3]  Trigger condition for Device Removal event
  Bit[2]  Trigger condition for TM command issued by driver
  Bit[1]  Trigger condition for Adapter reset issued by driver
  Bit[0]  Trigger condition for IOC Fault state

During driver load, if Master trigger type bit is enabled in the Persistent
Trigger Page0, then read the Persistent Trigger Page1 and update the IOC
instance's diag_trigger_master.MasterData with Persistent Trigger Page1's
MasterTriggerFlags.

Link: https://lore.kernel.org/r/20201126094311.8686-4-suganath-prabu.subramani@broadcom.comReported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The user can set trigger values in order to collect the IOC's host trace
buffer automatically upon detecting certain conditions. However, the
trigger values that the user sets are not persistent across system reboot
or reload of the driver.

In order to make the user trigger settings persistent, these trigger values
need to be saved in the IOC's NVRAM pages:

 - Driver Persistent Trigger Page 0:
     This page is used to store list of trigger types that are enabled

 - Driver Persistent Trigger Page 1:
     This page stores the list of Master triggers that are enabled

 - Driver Persistent Trigger Page 2:
     This page stores the list of MPI Event Triggers that are enabled

 - Driver Persistent Trigger Page 3:
     This page stores the list of SCSI Sense Triggers that are enabled

 - Driver Persistent Trigger Page 4:
     This page stores the list of IOCStatus-LogInfo Triggers that are
     enabled.

Whenever user configures triggers, the driver persists the values in the
corresponding trigger pages. When the driver is subsequently reloaded, the
driver reads the values from the trigger pages and configures the triggers
accordingly.

During firmware upload operation, if the newer firmware supports the
trigger page feature, then driver persists the configured diag trigger
values to NVRAM.

Link: https://lore.kernel.org/r/20201126094311.8686-3-suganath-prabu.subramani@broadcom.comReported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The controller time currently gets updated with host time during driver
load or when a controller reset is issued. I.e. when host issues the
IOCInit request message to the HBA firmware. This IOCInit message has a
field named 'TimeStamp' with which the host updates the controller time.

Sometimes controller time drifts with respect to the host and it is
difficult to correlate host logs with controller logs. Issuing a controller
reset to sync the time would impact in-flight I/O and is not a viable
option.

Instead the driver now sends an IO_UNIT_CONTROL Request to sync the time
periodically. This is done from the watchdog thread which gets invoked
every second.

The time synchronization interval is specified in the 'TimeSyncInterval'
field in Manufacturing Page11 by the controller:

    TimeSyncInterval - 8 bits
	bits  0-6: Time stamp Synchronization interval value
	bit	7: Time stamp Synchronization interval unit,
		   (if this bit is one then Timestamp Synchronization
		   interval value is specified in terms of hours else
		   Timestamp Synchronization interval value is
		   specified in terms of minutes).

The driver keeps track of the timer using IOC's timestamp_update_count
field. This field value gets incremented whenever the watchdog thread gets
invoked. And whenever this field value is greater than or equal to the Time
Stamp Synchronization interval value, the driver sends the IO_UNIT_CONTROL
Request message to controller to update the time and then it resets the
timestamp_update_count field to zero.

Link: https://lore.kernel.org/r/20201126094311.8686-2-suganath-prabu.subramani@broadcom.comReported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Link: https://lore.kernel.org/r/20201202132312.19966-16-njavali@marvell.comReviewed-by: NHimanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: NNilesh Javali <njavali@marvell.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Due to a bug in the older scan logic, when a once lost device re-appeared,
it was not discovered. Fix this by resetting login_retry counter upon
device discovery.

This is applicable only for 4G and older HBAs.

Link: https://lore.kernel.org/r/20201202132312.19966-15-njavali@marvell.comReviewed-by: NHimanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: NArun Easi <aeasi@marvell.com>
Signed-off-by: NNilesh Javali <njavali@marvell.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Driver unload with I/Os in flight causes server to crash.  Complete I/O
with DID_IMM_RETRY if fcport undergoing deletion.

CPU: 44 PID: 35008 Comm: qla2xxx_4_dpc Kdump: loaded Tainted: G
OE  X   5.3.18-22-default #1 SLE15-SP2 (unreleased)
Hardware name: HPE ProLiant DL380 Gen10/ProLiant DL380 Gen10, BIOS U30 07/16/2020
RIP: 0010:dma_direct_unmap_sg+0x24/0x60
Code: 4c 8b 04 24 eb b9 0f 1f 44 00 00 85 d2 7e 4e 41 57
      4d 89 c7 41 56 41 89 ce 41 55 49 89 fd 41 54 41 89 d4 55 31 ed 53 48 89
      f3 <8b> 53 18 48 8b 73 10 4d 89 f8 44 89 f1 4c 89 ef 83 c5 01 e8 44 ff
RSP: 0018:ffffc0c661037d88 EFLAGS: 00010046
RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000002
RDX: 000000000000001d RSI: 0000000000000000 RDI: ffff9a51ee53b0b0
RBP: 0000000000000000 R08: 0000000000000000 R09: ffff9a51ee53b0b0
R10: ffffc0c646463dc8 R11: ffff9a4a067087c8 R12: 000000000000001d
R13: ffff9a51ee53b0b0 R14: 0000000000000002 R15: 0000000000000000
FS:  0000000000000000(0000) GS:ffff9a523f800000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000018 CR3: 000000043740a004 CR4: 00000000007606e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
PKRU: 55555554
Call Trace:
qla2xxx_qpair_sp_free_dma+0x20d/0x3c0 [qla2xxx]
qla2xxx_qpair_sp_compl+0x35/0x90 [qla2xxx]
__qla2x00_abort_all_cmds+0x180/0x390 [qla2xxx]
? qla24xx_process_purex_list+0x100/0x100 [qla2xxx]
qla2x00_abort_all_cmds+0x5e/0x80 [qla2xxx]
qla2x00_do_dpc+0x317/0xa30 [qla2xxx]
kthread+0x10d/0x130
? kthread_park+0xa0/0xa0
ret_from_fork+0x35/0x40

Link: https://lore.kernel.org/r/20201202132312.19966-14-njavali@marvell.comReviewed-by: NHimanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: NSaurav Kashyap <skashyap@marvell.com>
Signed-off-by: NNilesh Javali <njavali@marvell.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The call trace was because workqueue was allocated without any flags, added
WQ_MEM_RECLAIM as flag when allocating.

kernel: workqueue: WQ_MEM_RECLAIM
kblockd:blk_mq_run_work_fn is flushing !WQ_MEM_RECLAIM qla2xxx_wq:0x0
kernel: WARNING: CPU: 0 PID: 2475 at
kernel/workqueue.c:2593 check_flush_dependency+0x110/0x130
kernel: CPU: 0 PID: 2475 Comm: kworker/0:1H Kdump:
loaded Tainted: G           OE    --------- -  - 4.18.0-193.el8.x86_64 #1
kernel: Hardware name: HPE ProLiant XL170r Gen10/ProLiant XL170r Gen10, BIOS U38 05/21/2019
kernel: Workqueue: kblockd blk_mq_run_work_fn
kernel: RIP: 0010:check_flush_dependency+0x110/0x130
kernel: Code: ff ff 48 8b 50 18 48 8d 8b b0 00 00 00 49 89 e8 48 81 c6 b0 00 00 00 48 c7 c7 00 1e e9
	95 c6 05 dc 9a 2f 01 01 e8 1a 42 fe ff <0f> 0b e9 0a ff ff ff 80 3d ca 9a 2f 01 0 0 75 95 e9 41 ff ff ff 90
kernel: RSP: 0018:ffffa40f48b2baf8 EFLAGS: 00010282
kernel: RAX: 0000000000000000 RBX: ffff946795282600 RCX: 0000000000000000
kernel: RDX: 000000000000005f RSI: ffffffff96a1af7f RDI: 0000000000000246
kernel: RBP: 0000000000000000 R08: ffffffff96a1af20 R09: 0000000000029480
kernel: R10: 00080c89bb3e7462 R11: 00000000000009ab R12: ffff946773628000
kernel: R13: 0000000000000282 R14: 0000000000000246 R15: ffffa40f48b2bb40
kernel: FS: 	0000000000000000(0000) 	GS:ffff94679fa00000(0000) knlGS:0000000000000000
kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
kernel: CR2: 00005570c4b60110 CR3: 000000029140a005 CR4: 00000000007606f0
kernel: DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
kernel: DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
kernel: PKRU: 55555554
kernel: Call Trace:
kernel: flush_workqueue+0x13a/0x440
kernel: qla2x00_wait_for_sess_deletion+0x1d6/0x200 [qla2xxx]
kernel: ? finish_wait+0x80/0x80
kernel: qla2xxx_disable_port+0x2b/0x30 [qla2xxx]
kernel: qla2x00_process_vendor_specific+0x1dc9/0x2d20 [qla2xxx]
kernel: ? blk_rq_map_sg+0x195/0x570
kernel: qla24xx_bsg_request+0x1a3/0xf90 [qla2xxx]

Link: https://lore.kernel.org/r/20201202132312.19966-13-njavali@marvell.comReviewed-by: NHimanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: NSaurav Kashyap <skashyap@marvell.com>
Signed-off-by: NNilesh Javali <njavali@marvell.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Flash update failed due to missing endian conversion in FLT region access
as well as in checksum computation.

Link: https://lore.kernel.org/r/20201202132312.19966-12-njavali@marvell.comReviewed-by: NHimanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: NArun Easi <aeasi@marvell.com>
Signed-off-by: NNilesh Javali <njavali@marvell.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Call trace observed while shutting down the adapter ports (LINK DOWN).
Handle aborts correctly.

localhost kernel: INFO: task nvme:44209 blocked for more than 120 seconds.
localhost kernel: "echo 0 >/proc/sys/kernel/hung_task_timeout_secs" disables this message.
localhost kernel: nvme            D ffff88b45fb5acc0     0 44209 1 0x00000080
localhost kernel: Call Trace:
localhost kernel: [<ffffffffbd187169>] schedule+0x29/0x70
localhost kernel: [<ffffffffbd184c51>] schedule_timeout+0x221/0x2d0
localhost kernel: [<ffffffffbcad7229>] ? ttwu_do_wakeup+0x19/0xe0
localhost kernel: [<ffffffffbcad735f>] ? ttwu_do_activate+0x6f/0x80
localhost kernel: [<ffffffffbcada830>] ? try_to_wake_up+0x190/0x390
localhost kernel: [<ffffffffbd18751d>] wait_for_completion+0xfd/0x140
localhost kernel: [<ffffffffbcadaaf0>] ? wake_up_state+0x20/0x20
localhost kernel: [<ffffffffbcabe3da>] flush_work+0x10a/0x1b0
localhost kernel: [<ffffffffbcabb0f0>] ? move_linked_works+0x90/0x90
localhost kernel: [<ffffffffbcabe6cf>] flush_delayed_work+0x3f/0x50
localhost kernel: [<ffffffffc0452767>] nvme_fc_init_ctrl+0x657/0x6a0 [nvme_fc]
localhost kernel: [<ffffffffc045293a>] nvme_fc_create_ctrl+0x18a/0x210 [nvme_fc]
localhost kernel: [<ffffffffc028962f>] nvmf_dev_write+0x98f/0xb35 [nvme_fabrics]
localhost kernel: [<ffffffffbcd08927>] ? security_file_permission+0x27/0xa0
localhost kernel: [<ffffffffbcc4db50>] vfs_write+0xc0/0x1f0
localhost kernel: [<ffffffffbcc4e92f>] SyS_write+0x7f/0xf0
localhost kernel: [<ffffffffbd193f92>] system_call_fastpath+0x25/0x2a

Link: https://lore.kernel.org/r/20201202132312.19966-11-njavali@marvell.comReviewed-by: NHimanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: NSaurav Kashyap <skashyap@marvell.com>
Signed-off-by: NNilesh Javali <njavali@marvell.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

FC-NVMe target discovery failed when initiator wwpn < target wwpn in an N2N
(Direct Attach) config, where the driver was stuck on FCP PRLI mode and
failed to retry with NVMe PRLI.

Link: https://lore.kernel.org/r/20201202132312.19966-10-njavali@marvell.com
Fixes: 84ed362a ("scsi: qla2xxx: Dual FCP-NVMe target port support”)
Fixes: 983f1276 ("scsi: qla2xxx: Retry PLOGI on FC-NVMe PRLI failure”)
Signed-off-by: NQuinn Tran <qutran@marvell.com>
Signed-off-by: NNilesh Javali <njavali@marvell.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Some fields are not correctly byte swapped causing failure during
initialization. As probe() returns failure, HBAs will not be claimed when
this happens.

qla2xxx [0007:01:00.0]-ffff:3: Secure Flash Update in FW: Supported
qla2xxx [0007:01:00.0]-ffff:3: SCM in FW: Supported
qla2xxx [0007:01:00.0]-00d2:3: Init Firmware **** FAILED ****.
qla2xxx [0007:01:00.0]-00d6:3: Failed to initialize adapter - Adapter flags 2.
qla2xxx 0007:01:00.1: enabling device (0140 -> 0142)
qla2xxx [0007:01:00.1]-011c: : MSI-X vector count: 128.
qla2xxx [0007:01:00.1]-001d: : Found an ISP2289 irq 18 iobase 0xd000080080004000.
qla2xxx 0007:01:00.1: Using 64-bit direct DMA at offset 800000000000000
BUG: Bad page state in process insmod  pfn:67118 page:f00000000168bd40
count:-1 mapcount:0 mapping: (null) index:0x0
page flags: 0x3ffff800000000() page dumped because: nonzero _count
Modules linked in: qla2xxx(OE+) nvme_fc nvme_fabrics
	nvme_core scsi_transport_fc scsi_tgt nls_utf8 isofs ip6t_rpfilter
	ipt_REJECT nf_reject_ipv4 ip6t_REJECT nf_reject_ipv6 xt_conntrack ip_set
	nfnetlink ebtable_nat ebtable_broute bridge stp llc ip6table_nat
	nf_conntrack_ipv6 nf_defrag_ipv6 nf_nat_ipv6 ip6table_mangle
	ip6table_security ip6table_raw iptable_nat nf_conntrack_ipv4
	nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack iptable_mangle
	iptable_security iptable_raw ebtable_filter ebtables ip6table_filter
	ip6_tables iptable_filter nx_crypto ses enclosure scsi_transport_sas
	pseries_rng sg ip_tables xfs libcrc32c sr_mod cdrom sd_mod crc_t10dif
	crct10dif_generic crct10dif_common usb_storage ipr libata tg3 ptp
	pps_core dm_mirror dm_region_hash dm_log dm_mod
CPU: 32 PID: 8560 Comm: insmod Kdump: loaded Tainted: G
	OE  ------------   3.10.0-957.el7.ppc64 #1
Call Trace:
[c0000006dd7caa70] [c00000000001cca8] .show_stack+0x88/0x330 (unreliable)
[c0000006dd7cab30] [c000000000ac3d88] .dump_stack+0x28/0x3c
[c0000006dd7caba0] [c00000000029e48c] .bad_page+0x15c/0x1c0
[c0000006dd7cac40] [c00000000029f938] .get_page_from_freelist+0x11e8/0x1ea0
[c0000006dd7caf40] [c0000000002a1d30] .__alloc_pages_nodemask+0x1c0/0xc70
[c0000006dd7cb140] [c00000000002ba0c] .__dma_direct_alloc_coherent+0x8c/0x170
[c0000006dd7cb1e0] [d000000010a94688] .qla2x00_mem_alloc+0x10f8/0x1370 [qla2xxx]
[c0000006dd7cb2d0] [d000000010a9c790] .qla2x00_probe_one+0xb60/0x22e0 [qla2xxx]
[c0000006dd7cb540] [c0000000005de764] .pci_device_probe+0x204/0x300
[c0000006dd7cb600] [c0000000006ca61c] .driver_probe_device+0x2cc/0x6f0
[c0000006dd7cb6b0] [c0000000006cabec] .__driver_attach+0x10c/0x110
[c0000006dd7cb740] [c0000000006c5f04] .bus_for_each_dev+0x94/0x100
[c0000006dd7cb7e0] [c0000000006c94f4] .driver_attach+0x34/0x50
[c0000006dd7cb860] [c0000000006c8f58] .bus_add_driver+0x298/0x3b0
[c0000006dd7cb900] [c0000000006cb6e0] .driver_register+0xb0/0x1a0
[c0000006dd7cb980] [c0000000005dc474] .__pci_register_driver+0xc4/0xf0
[c0000006dd7cba10] [d000000010b94e20] .qla2x00_module_init+0x2a8/0x328 [qla2xxx]
[c0000006dd7cbaa0] [c00000000000c130] .do_one_initcall+0x130/0x2e0
[c0000006dd7cbb50] [c0000000001b2e8c] .load_module+0x1afc/0x2340
[c0000006dd7cbd40] [c0000000001b3920] .SyS_finit_module+0xd0/0x130
[c0000006dd7cbe30] [c00000000000a284] 	system_call+0x38/0xfc

Link: https://lore.kernel.org/r/20201202132312.19966-9-njavali@marvell.com
Fixes: 9f2475fe ("scsi: qla2xxx: SAN congestion management implementation")
Fixes: cf3c54fb ("scsi: qla2xxx: Add SLER and PI control support”)
Reviewed-by: NHimanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: NArun Easi <aeasi@marvell.com>
Signed-off-by: NNilesh Javali <njavali@marvell.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Crash stack:
	[576544.715489] Unable to handle kernel paging request for data at address 0xd00000000f970000
	[576544.715497] Faulting instruction address: 0xd00000000f880f64
	[576544.715503] Oops: Kernel access of bad area, sig: 11 [#1]
	[576544.715506] SMP NR_CPUS=2048 NUMA pSeries
	:
	[576544.715703] NIP [d00000000f880f64] .qla27xx_fwdt_template_valid+0x94/0x100 [qla2xxx]
	[576544.715722] LR [d00000000f7952dc] .qla24xx_load_risc_flash+0x2fc/0x590 [qla2xxx]
	[576544.715726] Call Trace:
	[576544.715731] [c0000004d0ffb000] [c0000006fe02c350] 0xc0000006fe02c350 (unreliable)
	[576544.715750] [c0000004d0ffb080] [d00000000f7952dc] .qla24xx_load_risc_flash+0x2fc/0x590 [qla2xxx]
	[576544.715770] [c0000004d0ffb170] [d00000000f7aa034] .qla81xx_load_risc+0x84/0x1a0 [qla2xxx]
	[576544.715789] [c0000004d0ffb210] [d00000000f79f7c8] .qla2x00_setup_chip+0xc8/0x910 [qla2xxx]
	[576544.715808] [c0000004d0ffb300] [d00000000f7a631c] .qla2x00_initialize_adapter+0x4dc/0xb00 [qla2xxx]
	[576544.715826] [c0000004d0ffb3e0] [d00000000f78ce28] .qla2x00_probe_one+0xf08/0x2200 [qla2xxx]

Link: https://lore.kernel.org/r/20201202132312.19966-8-njavali@marvell.com
Fixes: f73cb695 ("[SCSI] qla2xxx: Add support for ISP2071.")
Cc: stable@vger.kernel.org
Reviewed-by: NHimanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: NArun Easi <aeasi@marvell.com>
Signed-off-by: NNilesh Javali <njavali@marvell.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Fix compile time errors reported on PPC systems,

qla_gbl.h:991:20: error: inlining failed in call to always_inline
     ‘qla_nvme_abort_set_option’: function body not available

Link: https://lore.kernel.org/r/20201202132312.19966-7-njavali@marvell.comSigned-off-by: NArun Easi <aeasi@marvell.com>
Signed-off-by: NNilesh Javali <njavali@marvell.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

NVMe commands can come only after successful addition of rport and NVMe
connect, and rport is only registered after FW started bit is set. Remove
the redundant check.

Link: https://lore.kernel.org/r/20201202132312.19966-6-njavali@marvell.comReviewed-by: NHimanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: NSaurav Kashyap <skashyap@marvell.com>
Signed-off-by: NNilesh Javali <njavali@marvell.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>