scsi: megaraid_sas: add print in device removal path

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: enhance debug logs in OCR context

Add additional logging from driver in OCR context.
Add debug logs for partial completion of IOs is iodone context.

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: set residual bytes count during IO completion

Fixing issue of not setting residual bytes correctly.

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: raid 1 write performance for large io

Avoid Host side PCI bandwidth bottleneck and hint FW to do Write
buffering using RaidFlag MR_RAID_FLAGS_IO_SUB_TYPE_LDIO_BW_LIMIT.  Once
IO is landed in FW with MR_RAID_FLAGS_IO_SUB_TYPE_LDIO_BW_LIMIT, it will
do single DMA from host and buffer the Write operation. On back end, FW
will DMA same buffer to the Mirror and Data Arm.  This will improve
large block IO performance which bottleneck due to Host side PCI
bandwidth limitation.

Consistent ~4000MB T.P for 256K Block size is expected performance
numbers.  IOPS for small Block size should be on par with Disk
performance.  (E.g 42 SAS Disk in JBOD mode gives 3700MB T.P.  Same
Drive used in R1 WT mode, should give ~1800MB T.P)

Using this patch 24 R1 VDs (HDD) gives below performance for Sequential
Write.  Without this patch, we cannot reach above 3200MB (Throughput is
in MB.)

Block Size    50% 256K and 50% 4K          100% 256K
4K                 3100                        2030
8K                 3140                        2740
16K                3140                        3140
32K                3400                        3240
64K                3500                        3700
128K               3870                        3870
256K               3920                        3920

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: NVME fast path io support

This patch provide true fast path IO support.  Driver creates PRP for
NVME drives and send Fast Path for performance.  Certain h/w requirement
needs to be taken care in driver.

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: NVME interface target prop added

This patch fetch true values of NVME property from FW using New DCMD
interface MR_DCMD_DEV_GET_TARGET_PROP

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: NVME Interface detection and prop settings

Adding detection logic for NVME device attached behind Ventura
controller.  Driver set HostPageSize in IOC_INIT frame to inform about
page size for NVME devices.  Firmware reports NVME page size to the
driver.  PD INFO DCMD provide new interface type NVME_PD. Driver set
property of NVME device.

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: change issue_dcmd to return void from int

With the changes to remove checks for a valid request descriptor,
issue_dcmd will now always return DCMD_SUCCESS. This patch changes
return type of issue_dcmd to void and change all callers appropriately.

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: megasas_get_request_descriptor always return valid desc

No functional change. Code clean up. Removing error code which is not
valid scenario.  In megasas_get_request_descriptor we can remove the
error handling which is not required.  With fusion controllers, if there
is a valid message frame available, we are guaranteed to get a
corresponding request descriptor.

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: Use DID_REQUEUE

Moving to use DID_REQUEUE return type for reliable unconditional
retries.  Driver wants unconditional re-queue, so replace DID_RESET with
DID_REQUEUE

Discussed below -
https://www.spinics.net/lists/linux-scsi/msg102848.html

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: RAID map is accessed for SYS PDs when use_seqnum_jbod_fp is not set

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: Refactor MEGASAS_IS_LOGICAL macro using sdev

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: 32 bit descriptor fire cmd optimization

No functional change. Code refactor.

megasas_fire_cmd_fusion can always use 32 bit descriptor write for
ventura. No need to pass extra flag.  Only IOC INIT required 64 bit
Descriptor write.

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: raid 1 fast path code optimize

No functional change. Code refactor.

Remove function megasas_fpio_to_ldio as we never require to convert fpio
to ldio because of frame unavailability.  Grab extra frame of raid 1
write fast path before it creates first frame as Fast Path.  Removed
is_raid_1_fp_write flag as raid 1 write fast path command is decided
using r1_alt_dev_handle only.  Move resetting megasas_cmd_fusion fields
at common function megasas_return_cmd_fusion.

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: cpu select rework.

No functional change. Code refactor.

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Revert "scsi: megaraid_sas: Enable or Disable Fast path based on the PCI Threshold Bandwidth"

This reverts commit "3e5eadb1a881" ("scsi: megaraid_sas: Enable or
Disable Fast path based on the PCI Threshold Bandwidth")

This patch was aimed to increase performance of R1 Write operation for
large IO size.  Since this method used timer approach, it turn on/off
fast path did not work as expected.  Patch 0013 describes new algorithm
and performance number.

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qedi: Fix possible memory leak in qedi_iscsi_update_conn()

'conn_info' is malloced in qedi_iscsi_update_conn() and should be freed
before leaving from the error handling cases, otherwise it will cause
memory leak.

Fixes: ace7f46ba5fd ("scsi: qedi: Add QLogic FastLinQ offload iSCSI driver framework.")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Acked-by: Manish Rangankar <Manish.Rangankar@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: dump hw regs on link failures

Dump host state, power info and host/vendor specific registers
on link failures. This provides useful info to debug the failures.

Signed-off-by: Venkat Gopalakrishnan <venkatg@codeaurora.org>
Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs-qcom: dump additional testbus registers

Change testbus default config, dump additional testbus registers along
with other debug vendor specific registers. These additional info are
useful in debugging link related failures.

Signed-off-by: Venkat Gopalakrishnan <venkatg@codeaurora.org>
Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: kick start clock scaling only after device detection

UFS clock scaling might start kicking in even before the device
is running at the fastest interface speed which is undesirable.
This change moves the clock scaling kick start only after the
device is detected and running at the fastest interface speed.

Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: don't suspend clock scaling during clock gating

Currently we are suspending clock scaling during clock gating which doesn't
allow us to have clock gating timeout lower than clock scaling polling
window. If clock gating timeout is smaller than the clock scaling polling
window then we will mostly suspend the clock scaling before clock scaling
polling window expires and we might get stuck in same state (scaled down
or scaled up) for quite a long time. And for this reason, we have clock
gating timeout (150ms) greater than clock scaling polling window (100ms).

We would like to have aggressive clock gating timeout even lower than the
clock scaling polling window hence this change is decoupling the clock
scaling suspend/resume from clock gate/ungate. We will not suspend the
clock scaling as part of clock gating instead clock scaling context will
schedule scaling suspend work if there are no more pending transfer
requests.

Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: add host state prints in failure cases

Whenever some UFS failure occurs the driver prints the UFS
registers in order to help with analysis of the failure.
However this may not be sufficient in some cases, so having
the host controller state as it is represented and managed in
the driver will contribute to analysis efforts.
Added prints of various fields in the hba struct which may be
of interest.

Signed-off-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: add load based scaling of UFS gear

UFS driver's load based clock scaling feature scales down the ufs related
clocks in order to allow low power modes of chipsets. UniPro 1.6 supports
maximum gear up to HS-G3 (High Speed Gear3) and some of the chipsets
low power modes may not be allowed in HS-G3 hence this change adds support
to scale gear between HS-G3 and HS-G1 based on same existing load based
clock scaling logic.

Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: reduce printout for aborted requests

Details printed for each request that is aborted can overload the
target as there can be several requests that are aborted at once.
This change will print full request details only for the first
aborted request since the last link reset, and minimal details
for other subsequent requests.

Signed-off-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: skip request abort task when previous aborts failed

On certain error conditions request abort task itself might fail
when aborting a request. In such case, subsequent request aborts
should skip issuing the abort task as it is expected to fail as well,
and device reset handler will be called next.

Signed-off-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: avoid open-coded upper_32_bits

Shifting a dma_addr_t right by 32 bits causes a compile-time warning
when that type is only 32 bit wide:

drivers/scsi/aacraid/src.c: In function 'aac_src_start_adapter':
drivers/scsi/aacraid/src.c:414:29: error: right shift count >= width of type [-Werror=shift-count-overflow]

This changes the driver to use the predefined macros consistently,
including one correct but open-coded upper_32_bits() instance.

Fixes: d1ef4da8487f ("scsi: aacraid: added support for init_struct_8")
Fixes: 423400e64d37 ("scsi: aacraid: Include HBA direct interface")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: rcode is unsigned and should be signed int

aac_fib_send can return -ve error returns and hence rcode should be
signed. Currently the rcode >= 0 check is always true and -ve errors are
not being checked.

Thanks to Dan Carpenter for spotting my original broken fix to this
issue.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: pm8001: switch to pci_irq_alloc_vectors

Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Jack Wang <jinpu.wang@profitbricks.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: remove tsk_mgmt_response and it_nexus_response transport methods

They are never called and just dispatch to methods of the same names in
the FC and SRP transport classes that are never implemented.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: remove eh_timed_out methods in the transport template

Instead define the timeout behavior purely based on the host_template
eh_timed_out method and wire up the existing transport implementations
in the host templates.  This also clears up the confusion that the
transport template method overrides the host template one, so some
drivers have to re-override the transport template one.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tyrel Datwyler <tyreld@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: libsas: remove sas_scsi_timed_out

EH_NOT_HANDLED is the default case if no eh_timed_out method is
provided, so there is no need to supply it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mvumi: remove fake transport template

These days we can specify an eh_timed_out handler in the host_template,
so don't have a transport_template definition just for it.

[mkp: fixed typo]

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: update version

Update the driver version to 50740

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Change Driver Version Prefix

Change the aacraid driver prefix from 1.2-1 to 1.2.1

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Update copyrights

Added new copyright messages

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Retrieve HBA host information ioctl

Added a new ioctl interface to retrieve the host device information.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Added ioctl to trigger IOP/IWBR reset

Added a new ioctl interface to trigger an IOP or IWBR reset from ioctl.
Primary used by management utility to trigger resets.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Added new IWBR reset

Added a new IWBR soft reset type, reworked the IOP reset interface for
a bit.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: VPD 83 type3 support

This patch adds support to retrieve the unique identifier data (VPD page
83 type3) for Logical drives created on SmartIOC 2000 products. In
addition  added a sysfs device structure to expose the id information.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Added support to abort cmd and reset lun

Added task management command support to abort any timed out commands
in case of a eh_abort call and to reset lun's in case of eh_reset call.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Add task management functionality

Added support to send out task management commands.

[mkp: removed // fibsize... ]

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Include HBA direct interface

Added support to send direct pasthru srb commands from management utilty
to the  controller.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Added support for hotplug

Added support for drive hotplug add and removal

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Added support to set QD of attached drives

Added support to set qd of drives in slave_configure.This only works for
HBA1000 attached drives.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Retrieve Queue Depth from Adapter FW

Retrieved queue depth from fw and saved it for future use.
Only applicable for HBA1000 drives.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Added support for periodic wellness sync

This patch adds a new functions that periodically sync the time of host
to the adapter. In addition also informs the adapter that the driver is
alive and kicking. Only applicable to the HBA1000 and SMARTIOC2000.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Reworked aac_command_thread

Reworked aac_command_thread into aac_process_events

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Added support for read medium error

This patch processes Raw IO read medium errors.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Added support for response path

This patch enables the driver to actually process the I/O, or srb replies
from adapter. In addition to any HBA1000 or SmartIOC2000 adapter events.

Signed-off-by: Raghava Aditya Renukunta <raghavaaditya.renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Process Error for response I/O

Make sure that the driver processes error conditions even in the fast
response path for response from the adapter.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Reworked scsi command submission path

Moved the READ and WRITE switch cases to the top. Added a  default
case to the switch case and replaced duplicate scsi result value with a
macro.

The idea is that since most of scsi commands we care about performance
wise are read or write, we need to process them first.

Internally the compiler (GCC) converts a switch case into either a jump
table or a bunch of if else conditions, so placing the often used read,
write cases at the top is an effort in optimization.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Retrieve and update the device types

This patch adds support to retrieve the type of each adapter connected
device. Applicable to HBA1000 and SmartIOC2000 products

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Added sa firmware support

sa_firmware adds the capability to differentiate the new SmartIOC family
of adapters from the series 8 and below.

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: added support for init_struct_8

This  patch lays the groundwork for supporting the new HBA-1000 controller
family.A new INIT structure INIT_STRUCT_8 has been added which allows for a
variable size for MSI-x vectors among other things,  and is used for both
Series-8, HBA-1000 and SmartIOC-2000.

Signed-off-by: Raghava Aditya Renukunta <raghavaaditya.renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Added aacraid.h include guard

Added aacraid.h include guard

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: aacraid: Remove duplicate irq management code

Removed duplicate code that for acquiring and releasing irqs

Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: Dave Carroll <David.Carroll@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: scsi_debug: Add OPTIMAL TRANSFER LENGTH GRANULARITY option.

[mkp: whitespace fixes]

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

scsi: mpt3sas: Updating driver version to v15.100.00.00

Updated driver version to "15.100.00.00"

Signed-off-by: Chaitra P B <chaitra.basappa@broadcom.com>
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpt3sas: Fix Firmware fault state 0x2100 during heavy 4K RR FIO stress test.

Due existence of loop in the IO path our HBA will receive heavy IOs and
also as driver is not updating the Reply Post Host Index frequently, So
there will be a high chance that our Firmware unable to find any free
entry in the Reply Post Descriptor Queue (i.e. Queue overflow occurs)
and can observe 0x2100 firmware fault.  So to fix this, we have defined
a thresh hold value. After continuously processing this thresh hold
number of reply descriptors driver will update the Reply Descriptor Host
Index so that this thresh hold number of reply descriptors entries will
be freed and these entries will be available for firmware and we won't
observe this Firmware fault. We have defined this threshold value as
1/3rd of the hba queue depth.

Signed-off-by: Chaitra P B <chaitra.basappa@broadcom.com>
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpt3sas: Fix for Crusader to achieve product targets with SAS devices.

Small glitch/degraded performance in Crusader is improved with SAS
drives by removing unnecessary spinlocks while clearing scsi command in
drivers internal lookup table.

Signed-off-by: Chaitra P B <chaitra.basappa@broadcom.com>
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: mpt3sas: Added print to notify cable running at a degraded speed.

Driver processes the event MPI26_EVENT_ACTIVE_CABLE_DEGRADED when a
cable is present and is running at a degraded speed (below the SAS3 12
Gb/s rate). Prints added to inform the user that the cable is not
running at optimal speed.

Signed-off-by: Chaitra P B <chaitra.basappa@broadcom.com>
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hpsa: remove coalescing settings for ioaccel2

- Setting coalescing has a significant negative impact on low
  queue-depth performance.
- Does not help high queue-depth performance.

Signed-off-by: Don Brace <don.brace@microsemi.com>
Reviewed-by: Scott Benesh <scott.benesh@microsemi.com>
Reviewed-by: Scott Teel <scott.teel@microsemi.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microsemi.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: storvsc: remove bogus code to transfer struct scatterlist

Remove a piece of code in storvsc_queuecommand that tries to pass the
physical address of the kernel struct scatterlist pointer to the host.

Fortunately the code can't ever be reached anyway.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: atari_scsi: Reset DMA during bus reset only under ST-DMA lock

The atari_scsi driver should not access Falcon DMA chip registers unless
it has acquired exclusive access to that chip. If the driver doesn't
have exclusive access then there's no need for a DMA reset as there are
no scsi commands in progress.

Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ncr5380: Improve target selection robustness

Handle timeout or bus phase change errors that could occur when sending
the IDENTIFY message.

Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ncr5380: Resolve various static checker warnings

Avoid various warnings from "make C=1" by annotating a couple of
unlock-then-lock sequences, replacing a zero with NULL and correcting
some type casts.

Also avoid a warning from "make W=1" by adding braces.

Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ncr5380: Reduce #include files

The NCR5380 wrapper drivers don't export symbols or declarations and
don't actually need separate header files. Most of these header files
were removed already; only sun3_scsi.h and g_NCR5380.h remain.

Move the remaining definitions to the corresponding .c files to improve
readability and proximity. The #defines which influence the #included
core driver are no longer mixed up with unrelated #defines and #includes.

Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ncr5380: Clean up dead code and redundant macro usage

Remove dead code inside #if 0 conditionals.

Remove the #ifdef __KERNEL__ test, since NCR5380.h has no definitions
that relate to userspace code.

Remove two redundant macro definitions which were overlooked in
commit e9db3198e08b ("sun3_scsi: Adopt NCR5380.c core driver").

Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ncr5380: Shorten host info string by removing unused option macros

The DIFFERENTIAL and PARITY option macros are unused: no supported
hardware uses differential signalling and the core driver never
implemented parity checking. These options just waste space in the host
info string.

While we are here, fix a typo in the NCR5380_info() kernel-doc comment.

Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hisi_sas: decrease running_req in hisi_sas_slot_task_free()

There is an issue that hisi_sas_dev.running_req is not
decremented properly for internal abort and TMF.

To resolve, only decrease running_req in hisi_sas_slot_task_free()

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hisi_sas: fix probe ordering problem

There is a potential probe issue in how we trigger the hw initialisation.
Although we use 1s timer to delay hw initialisation, there is still a
potential that sas_register_ha() is not be finished before we start
the PHY init from hw->hw_init().
To avoid this issue, initialise the hw after sas_register_ha() in the
same probe context.
Note: it is not necessary to use 1s timer now (modified v2 hw only).

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hisi_sas: downgrade internal abort exit print

Downgrade the exit print in hisi_sas_internal_task_abort()
to dbg level, as info is not required.

Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Xiang Chen <chenxiang66@hisilicon.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hisi_sas: modify hard reset for directed-attached disk

Correctly set registers in v2 for root PHY hardreset for directly
attached disk.

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hisi_sas: downgrade refclk message

The message to inform that the controller has no refclk
is currently at warning level, which is unnecessary, so
downgrade to debug.

Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Xiang Chen <chenxiang66@hisilicon.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hisi_sas: modify some values of ITCT table

Set SMP connection timeout and continue AWT timer;
Clear ITCT table when dev gone.

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: hisi_sas: workaround v2 hw SATA IO timeout issue

The v2 SAS controller needs more time to detect channel idle
and send setup link request than SATA disk does, so it is
difficult for the SAS controller to setup an STP link. Therefore
it may cause some IO timeouts.

We need to periodically configure the SAS controller so it
doesn't receive STP setup requests from SATA disks for a while,
so IO can be sent during this period.

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Fix lpfc_wwn_set return code check

When I reversed the patch to re-add the lpfc_soft_wwn parameter feature,
it re-added code that had a long-standing bug. (that's what I get I
guess :)

As Dan Carpenter pointed out - error checks looked at wrong polarity.  0
is success, -errno is failure. Updated checks.

Signed-off-by: James Smart <james.smart@broadcom.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: sd: Cleaned up comment references to @sdp argument explanation.

In sd.c there are two comment references to 'struct scsi_device *sdp' as
an argument.  One of the references has a typo and the other should be a
reference to 'struct device *dev' instead.

Fixed by correcting the typo in the first and changing the explanation
in the second.

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

scsi: be2iscsi: Use GFP_ATOMIC under spin lock

A spin lock is taken here so we should use GFP_ATOMIC.

Fixes: 987132167f4b ("scsi: be2iscsi: Fix for crash in beiscsi_eh_device_reset")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Reviewed-by: Jitendra Bhivare <jitendra.bhivare@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: lpfc: Replace BUG() with BUG_ON()

Replace BUG() with BUG_ON() using coccinelle.

Signed-off-by: Shyam Saini <mayhs11saini@gmail.com>
Acked-by: James Smart <james.smart@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: qla4xxx: remove two unused MSI-X related #defines

Spotted while preparing qla2xxx changes as the symbols exist in both
drivers (sigh..).

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: cxlflash: Cancel scheduled workers before stopping AFU

When processing an AFU asynchronous interrupt, if the action results in an
operation that requires off level processing (a link reset for example),
the worker thread is scheduled. In the meantime a reset event (i.e.: EEH)
could unmap the AFU to recover. This results in an Oops when the worker
thread tries to access the AFU mapping.

[c000000f17e03b90] d000000007cd5978 cxlflash_worker_thread+0x268/0x550
[c000000f17e03c40] c00000000011883c process_one_work+0x1dc/0x680
[c000000f17e03ce0] c000000000118e80 worker_thread+0x1a0/0x520
[c000000f17e03d80] c000000000126174 kthread+0xf4/0x100
[c000000f17e03e30] c00000000000a47c ret_from_kernel_thread+0x5c/0xe0

In an effort to avoid this, a mapcount was introduced in
commit b45cdbaf9f7f ("cxlflash: Resolve oops in wait_port_offline")
but due to the race condition described above, this solution is incomplete.

In order to fully resolve this problem and to simplify things, this commit
removes the mapcount solution. Instead, the scheduled worker thread is
cancelled after interrupts have been disabled and prior to the mapping
being freed.

Fixes: b45cdbaf9f7f ("cxlflash: Resolve oops in wait_port_offline")
Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: cxlflash: Cleanup prints

The usage of prints within the cxlflash driver is inconsistent. This
hinders debug and makes the driver source and log output appear sloppy.

The following cleanups help unify the prints within cxlflash:
 - move all prints to dev-* where possible
 - transition all hex prints to lowercase
 - standardize variable prints in debug output
 - derive pointers in a consistent manner
 - change int to bool where appropriate
 - remove superfluous data from prints and print statements that do not
   make sense

Signed-off-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: cxlflash: Support SQ Command Mode

The SISLite specification outlines a new queuing model to improve
over the MMIO-based IOARRIN model that exists today. This new model
uses a submission queue that exists in host memory and is shared with
the device. Each entry in the queue is an IOARCB that describes a
transfer request. When requests are submitted, IOARCBs ('current'
position tracked in host software) are populated and the submission
queue tail pointer is then updated via MMIO to make the device aware
of the requests.

Signed-off-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: cxlflash: Refactor context reset to share reset logic

As staging for supporting hardware with different context reset
registers but a similar reset procedure, refactor the existing context
reset routine to move the reset logic to a common routine. This will
allow hardware with a different reset register to leverage existing
code.

Signed-off-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: fix arguments order some trace calls

Colin Ian King reported that with commit 7ff5ab473633 ("scsi: ufs: add
tracing support") static analysis is reporting that we may have swapped
arguments on calls to:

    trace_ufshcd_runtime_resume,
    trace_ufshcd_runtime_suspend,
    trace_ufshcd_system_suspend,
    trace_ufshcd_system_resume,
and trace_ufshcd_init

Where:
    hba->uic_link_state is passed to dev_state
    hba->curr_dev_pwr_mode is passed to link_state

This wasn't intentional so it's a bug. This change fixed this bug.

Reported-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Acked-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: vmw_pvscsi: switch to pci_alloc_irq_vectors

And simplify the interrupt handler by splitting the INTx case that needs
to deal with shared interrupts into a separate helper.

[mkp: typo fixage]

Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Jim Gill <jgill@vmware.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: driver version upgrade

Upgrade driver version.

Signed-off-by: Sasikumar Chandrasekaran <sasikumar.pc@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: Implement the PD Map support for SAS3.5 Generic Megaraid Controllers

Update Linux driver to use new pdTargetId field for JBOD target ID

Signed-off-by: Sasikumar Chandrasekaran <sasikumar.pc@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: ldio_outstanding variable is not decremented in completion path

ldio outstanding variable needs to be decremented in io completion path for
iMR dual queue depth

Signed-off-by: Sasikumar Chandrasekaran <sasikumar.pc@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: Enable or Disable Fast path based on the PCI Threshold Bandwidth

Large SEQ IO workload should sent as non fast path commands

Signed-off-by: Sasikumar Chandrasekaran <sasikumar.pc@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: Add the Support for SAS3.5 Generic Megaraid Controllers Capabilities

The Megaraid driver has to support the SAS3.5 Generic Megaraid Controllers Firmware functionality.

Signed-off-by: Sasikumar Chandrasekaran <sasikumar.pc@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: Dynamic Raid Map Changes for SAS3.5 Generic Megaraid Controllers

SAS3.5 Generic Megaraid Controllers FW will support new dynamic RaidMap to have different
sizes for different number of supported VDs.

Signed-off-by: Sasikumar Chandrasekaran <sasikumar.pc@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: SAS3.5 Generic Megaraid Controllers Fast Path for RAID 1/10 Writes

To improve RAID 1/10 Write performance, OS drivers need to issue the
required Write IOs as Fast Path IOs (after the appropriate checks
allowing Fast Path to be used) to the appropriate physical drives
(translated from the OS logical IO) and wait for all Write IOs to complete.

Design: A write IO on RAID volume will be examined if it can be sent in
Fast Path based on IO size and starting LBA and ending LBA falling on to
a Physical Drive boundary. If the underlying RAID volume is a RAID 1/10,
driver issues two fast path write IOs one for each corresponding physical
drive after computing the corresponding start LBA for each physical drive.
Both write IOs will have the same payload and are posted to HW such that
replies land in the same reply queue.

If there are no resources available for sending two IOs, driver will send
the original IO from SCSI layer to RAID volume through the Firmware.

Based on PCI bandwidth and write payload, every second this feature is
enabled/disabled.

When both IOs are completed by HW, the resources will be released
and SCSI IO completion handler will be called.

Signed-off-by: Sasikumar Chandrasekaran <sasikumar.pc@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: SAS3.5 Generic Megaraid Controllers Stream Detection and IO Coalescing

Detect sequential Write IOs and pass the hint that it is part of sequential
stream to help HBA Firmware do the Full Stripe Writes. For read IOs on
certain RAID volumes like Read Ahead volumes,this will help driver to
send it to Firmware even if the IOs can potentially be sent to
hardware directly (called fast path) bypassing firmware.

Design: 8 streams are maintained per RAID volume as per the combined
firmware/driver design. When there is no stream detected the LRU stream
is used for next potential stream and LRU/MRU map is updated to make this
as MRU stream. Every time a stream is detected the MRU map
is updated to make the current stream as MRU stream.

Signed-off-by: Sasikumar Chandrasekaran <sasikumar.pc@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: EEDP Escape Mode Support for SAS3.5 Generic Megaraid Controllers

An UNMAP command on a PI formatted device will leave the Logical Block Application
Tag and Logical Block Reference Tag as all F's (for those LBAs that are unmapped).
To avoid IO errors if those LBAs are subsequently read before they are written with
valid tag fields, the MPI SCSI IO requests need to set the EEDPFlags element EEDP
Escape Mode field, Bits [7:6] appropriately.  A value of 2 should be set to disable
all PI checks if the Logical Block Application Tag is 0xFFFF for PI types 1 and 2.
A value of 3 should be set to disable all PI checks if the Logical Block Application
Tag is 0xFFFF and the Logical Block Reference Tag is 0xFFFFFFFF for PI type 3.

Signed-off-by: Sasikumar Chandrasekaran <sasikumar.pc@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: 128 MSIX Support

SAS3.5 Generic Megaraid based Controllers will have the support for 128 MSI-X vectors,
resulting in the need to support 128 reply queues

Signed-off-by: Sasikumar Chandrasekaran <sasikumar.pc@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: Add new pci device Ids for SAS3.5 Generic Megaraid Controllers

This patch contains new pci device ids for SAS3.5 Generic Megaraid Controllers

Signed-off-by: Sasikumar Chandrasekaran <sasikumar.pc@broadcom.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: remove useless acpi functions in the header file

commit f1bc1e4c44b1 ("ata: acpi: rework the ata acpi bind support")
removed scsi_register_acpi_bus_type() and
scsi_unregister_acpi_bus_type(), but forgot to remove them in the header
file, do it now.

Signed-off-by: Hanjun Guo <hanjun.guo@linaro.org>
Reviewed-by: John Garry <john.garry@huawei.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: refactor device descriptor reading

Pull device descriptor reading out of ufs quirk so it can be used also
for other purposes.

Revamp the fixup setup:

1. Rename ufs_device_info to ufs_dev_desc as very similar name
   ufs_dev_info is already in use.

2. Make the handlers static as they are not used out of the ufshdc.c
   file.

[mkp: applied by hand]

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Reviewed-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>