Below Functions are added in various paths to support NVMe drive
addition.

_scsih_pcie_add_device
_scsih_pcie_device_add
_scsih_pcie_device_init_add
_scsih_check_pcie_access_status
_scsih_pcie_check_device

mpt3sas_get_pdev_by_handle

mpt3sas_config_get_pcie_device_pg0
mpt3sas_config_get_pcie_device_pg2
Signed-off-by: Chaitra P B <chaitra.basappa@broadcom.com>
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

* Mpt3sas driver uses the NVMe Encapsulated Request message to send an
  NVMe command to an NVMe device attached to the IOC.

* Normal I/O commands like reads and writes are passed to the
  controller as SCSI commands and the controller has the ability to
  translate the commands to NVMe equivalent.

* This encapsulated NVMe command is used by applications to send
  direct NVMe commands to NVMe drives.
Signed-off-by: Chaitra P B <chaitra.basappa@broadcom.com>
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

* Added support for translating the SGLs associated with incoming
commands either to IEE SGL or NVMe PRPs for NVMe devices.

* The hardware translation of IEEE SGL to NVMe PRPs has limitations
and if a command cannot be translated by hardware then it will go to
firmware and the firmware needs to translate it. This will have a
performance impact. To avoid that, the driver proactively checks
whether the translation will be done in hardware or not. If not, then
driver translates.

[mkp: clarified commit message]
Signed-off-by: Chaitra P B <chaitra.basappa@broadcom.com>
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

1) Added support for probing pcie device and adding NVMe drives to SML
and driver's internal list pcie_device_list.

2) Added support for determing NVMe as boot device.

3) Added nvme device support for call back functions scan_finished
target_alloc,slave_alloc,target destroy and slave destroy.

 a) During scan, pcie devices are probed and added to SML to drivers
internal list.

 b) target_alloc & slave alloc API's allocates resources for
(MPT3SAS_TARGET & MPT3SAS_DEVICE) private datas and holds information
like handle, target_id etc.

 c) slave_destroy & target_destroy are called when driver unregisters
or removes device. Also frees allocated resources and info.
Signed-off-by: Chaitra P B <chaitra.basappa@broadcom.com>
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Trivial fix to spelling mistake in dev_info message
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

When running multipath on a VM if all available paths go down the driver
can schedule large amounts of storvsc_remove_lun work items to the same
lun. In response to the failing paths typically storvsc responds by
taking host->scan_mutex and issuing a TUR per lun. If there has been
heavy IO to the failed device all the failed IOs are returned from the
host. A remove lun work item is issued per failed IO. If the outstanding
TURs have not been completed in a timely manner the scan_mutex is never
released or released too late. Consequently the many remove lun work
items are not completed as scsi_remove_device also tries to take
host->scan_mutex.  This results in dragging the VM down and sometimes
completely.

This patch only allows one remove lun to be issued to a particular lun
while it is an instantiated member of the scsi stack.
Signed-off-by: Cathy Avery <cavery@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Long Li <longli@microsoft.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megasas_alloc_cmds is to alloc cmd_list of instance instead of fusion,
and fusion is useless in this function. Just remove it.
Signed-off-by: Yisheng Xie <xieyisheng1@huawei.com>
Acked-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Updated comment. We are keeping track of maximum number of retries per
command via retries/allowed in struct scsi_cmnd. Corrected comment
positioning.

[mkp: applied by hand]
Signed-off-by: Petros Koutoupis <petros@petroskoutoupis.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

The scsi_debug driver incorrectly suggests there is an error with the
SCSI WRITE SAME command when the number_of_logical_blocks is greater
than 1. It will also suggest there is an error when NDOB
(no data-out buffer) is set and the number_of_logical_blocks is
greater than 0. Both are valid, fix.
Signed-off-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Trivial fix to spelling mistakes in 'iscsi_get_host_stats'.
Signed-off-by: Arvind Yadav <arvind.yadav.cs@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

The 5380 drivers only support asynchronous transfers and the 5380
controllers only have narrow busses. Hence, the core driver will reject
any SDTR and WDTR messages from target devices. Don't log this, it's
expected behaviour. Also, fix the off-by-one array indices in the
arguments to scmd_printk().
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

The cxlflash driver tracks process IDs alongside contexts to validate
context ownership. Currently, the process IDs are derived by directly
accessing values from the 'current' task pointer. While this method of
access is fine for the current process, it is incorrect when the parent
process ID is needed as the access requires serialization.

To address the incorrect issue and provide a consistent means of
deriving the process ID within the cxlflash driver, use the task
accessors defined linux/sched.h.
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>

Currently, all adapters that cxlflash supports must have WWPN VPD
keywords to complete configuration. This was required as cards with
external FC ports needed to be programmed with WWPNs.

Newer supported cards do not have an external FC interface and therefore
do not require WWPN. To support backwards compatibility, these devices
have included 'dummy' WWPN VPD with WWPN values of zero. This however
places a dependency that all future cards have WWPN VPD, which may not
always be the case.

Allow for cards to not have WWPN, designating which cards are expected
to have it in order to configure properly.
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>

The existing write same routine within the cxlflash driver uses a
statically defined value for the maximum write same transfer length.
While this is close to the value reflected by the original device that
was supported by cxlflash, newer devices are capable of much larger
lengths. Supporting what the device is capable of offers substantial
performance improvement as the scrub routine within cxlflash operates on
'chunk size' units (256MB with a 4K sector size).

Instead of a #define, use the write same maximum length that is stored
in the block layer in units of 512 byte sectors. This value is initially
determined from the block limits VPD page during device discovery and
can also be manipulated from sysfs. As a general cleanup, designate the
timeout used when executing the write same command as constant.
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>

Add support for servicing AXI errors handling. We do a host controller
reset for such errors.
Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Add support for processing AXI bus fatal errors.

If AXI bus fatal error happen, do controller reset to recover.
Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

The CQ event is handled in tasklet context, and it could be delayed if
the system loading is high.

It is possible to run into some problems when executing a host reset
when cq_tasklet_vx_hw() is being executed.

So, prior to host reset, execute tasklet_kill() to ensure that all CQ
tasklets are complete.

Besides, as the function hisi_sas_wait_tasklets_done() is added to do
tasklet_kill(), this patch refactors some code where tasklet_kill() is
used.
Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Use completion to wait on ITCT CLR interrupt finishing before
processing other things when freeing a device.

This is safer than the pre-existing process of polling the register.
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Add init code of hisi_hba->rst_work for v3 hw. Because v3 hw also need
it to recover controller when some hw errors occurs.
Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Realise get_events() to add DFX feature for v3 hw. Just like v2 hw, We
support the following errors:
- loss_of_dword_sync_count
- invalid_dword_count
- phy_reset_problem_count
- running_disparity_error_count
Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Add initialization of register CON_CFG_DRIVER for v3 hw, to limit
number of the times of setup connection.
Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

We should check register PHY_STATE when getting the bitmap of a
wideport, as, if the PHY is not ready, the value of register
PHY_PORT_NUM_MA is not valid.

V2 hw has done this check, and v3 hw should do this check too.
Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

The code to print AXI errors in v2 hw driver is repetitive.

This patch condenses the code by looping an array of errors.

Also, a formatting error in one_bit_ecc_errors[] and
multi_bit_ecc_errors[] is fixed.
Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

The function hisi_sas_slot_task_free() is used to free the slot and do
tidy-up of LLDD resources. The LLDD generally should know the state of
a slot and decide when to free it, and it should only be done once.

For some scenarios, we really don't know the state, like when TMF
timeout. In this case, we check task->lldd_task before calling
hisi_sas_slot_task_free().

However, we may miss some scenarios when we should also check
task->lldd_task, and it is not SMP safe to check task->lldd_task as we
don't protect it within spin lock.

This patch is to fix this risk of freeing slot twice, as follows:

  1. Check task->lldd_task in the hisi_sas_slot_task_free(), and give
     up freeing of this time if task->lldd_task is NULL.

  2. Set slot->buf to NULL after it is freed.
Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

This patch adds a NULL check of task->lldd_task before freeing the
slot in SMP path.

This is to guard against the scenario of the slot being freed during
the from the preceding internal abort.
Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

When a PHY_FUNC_LINK_RESET is issued, we need to fill the transport
identify_frame to SAS controller before the PHYs are enabled.

Without this, we may find that if a PHY which belonged to a wideport
before the reset may generate a new port id.
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Currently the size of memory we allocate for SATA breakpoint buffer is
incorrect.

The breakpoint memory size should be as follows: 32 (NCQ tags) * 128 *
2048 (max #devs) = 8MB

Currently we only allocate 0.5MB, but get away with it as we never
have SATA device index > 128 typically.

To conserve precious DMA memory (8MB may not be even available), limit
the number of devices per HBA to 1024, which means 4MB of memory
required for SATA breakpoint.

The 1024 device limit applied to all HW versions. For v3 hw, we need
to configure this value.
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

When adding/removing slots from device list, we need to lock this
operation with hisi_hba lock for safety.

This patch adds missing instances of this.
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

We used spin_lock() to grab hisi_hba.lock in two places where
spin_lock_irqsave() should be used, as hisi_hba.lock can be taken in
interrupt context.

This patch is to fix this.
Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

When an internal abort times out in hisi_sas_internal_task_abort(),
goto the exit label in and not go through the other task status
checks.
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

We already relocated hisi_sas_get_ncq_tag() into common file main.c,
so delete get_ncq_tag_v3_hw() and use hisi_sas_get_ncq_tag() instead.
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Add generated scsi_devinfo_tbl.c into .gitignore.

Fixes: 345e2960 ("scsi: scsi: Export blacklist flags to sysfs")
Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

The latest MegaRAID Firmware (for Invader series) has support for 64bit
DMA for both streaming and consistent DMA buffers.  All Ventura series
controller FW always support 64 bit consistent DMA.  Also, on a few
architectures 32bit DMA is not supported.

Current driver always prefers 32bit for consistent DMA and 64bit for
streaming DMA.  This behavior was unintentional and carried forwarded
from legacy controller FW. Need to enhance the driver to support 64bit
consistent DMA buffers based on the firmware capability.

Below is the DMA setting strategy in driver with this patch.  For
Ventura series, always try to set 64bit DMA mask. If it fails fall back
to 32bit DMA mask.  For Invader series and earlier generation
controllers, first try to set to 32bit consistent DMA mask irrespective
of FW capability. This is needed to ensure firmware downgrades do not
break. If 32bit DMA setting fails, check FW capability and try seting to
64bit DMA mask.

There are certain restrictions in the hardware for having all sense
buffers and all reply descriptors to be in the same 4GB memory region.
This limitation is h/w dependent and can not be changed in firmware.
This limitation needs to be taken care in driver while allocating the
buffers.  There was a discussion regarding this - find details at below
link.  https://www.spinics.net/lists/linux-scsi/msg108251.htmlSigned-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Driver load fails if memory allocation for request frame pool fails due
to the higher queue_depth requirement. The driver now allows dynamically
reducing queue_depth if memory allocations fail rather than failing
load.  With this, there is no need to limit queue_depth to 1K.
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

In certain cases, the host memory is limited and with FW supporting
higher queue depths there are increasing chances of IO request frame
allocation failures that we are seeing. In case of request frame
allocation failures, retry allocation with reduced queue depth (in steps
of 64) to continue to configure the controller with a reduced
performance rather than failing load.
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

cmd->frame->dcmd.opcode will be valid only for MFI_CMD_DCMD
IOCTL frames. Currently driver check for cmd->frame->dcmd.opcode without
checking cmd type. Ensure we check dcmd opcode only for MFI_CMD_DCMD
commands. Separate handling of MFI_CMD_SMP/STP commands from
MFI_CMD_DCMD in completion path.
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Older firmware version unconditionally pulls 4k frame for IOC INIT MFA
frame.  But driver allocates 1k or 4k max_chain_frame_sz based on FW
capability.  During boot time, this results in DMA read errors.
Workaround fix in driver by allocating separate ioc_init frame of 4k
size to support older firmware.
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Cc: stable@vger.kernel.org
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

Driver needs to send current host time to firmware during init.
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: megaraid_sas: Move controller memory allocations and DMA mask settings from probe to megasas_init_fw
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>