In megasas_get_seq_num, the status of the DCMD fired to FW is not
returned, it always returns success.  We could end up registering AEN
request with incorrect sequence number if the DCMD failed.  Return the
DCMD status back to caller.  This was discovered during code review and
very rare to see issue in field to see AEN request failed bt FW.
Signed-off-by: NShivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Commit b9637d14 ("scsi: megaraid_sas: Resize MFA frame used for IOC
INIT to 4k") increased the size of IOC INIT frame to 4k.  Need to use
updated size when memsetting init_frame.
Signed-off-by: NShivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Memory allocated for IOC_INIT command and stream detection array are not
zero'd before using. Use kzalloc instead of kmalloc to zero out the
memory allocated.
Signed-off-by: NSumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: NShivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

This patch does not change any functionality but makes the SCSI core
source code slightly easier to read.
Signed-off-by: NBart Van Assche <bart.vanassche@wdc.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The delay for the rescan worker needs to 10 seconds, missed the HZ in
there.

Fixes: a1367e4a (scsi: aacraid: Reschedule host scan in case of failure)
Signed-off-by: NRaghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The correct lun count needs to be divided by 24, missed it in the
previous patch set.

Fixes: 4b000227 (scsi: aacraid: Create helper functions to get lun info)
Signed-off-by: NRaghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

For hardware only supporting active/optimized there's no point in ever
re-issuing RTPG as the only new state we can possibly read is
active/optimized.  This avoid spurious errors during path failover on
such arrays.
Signed-off-by: NHannes Reinecke <hare@suse.com>
Reviewed-by: NBart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Use the ARRAY_SIZE macro on array __pciids to determine size of the
array.  Improvement suggested by coccinelle.
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The initialization of d is redundant as this value is never read and it
is overwritten inside the subsequent for-loop.  Remove this redundant
assignment.

Cleans up clang warning:
drivers/scsi/qla2xxx/qla_gs.c:3985:29: warning: Value stored to 'd'
during its initialization is never read
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Use dma_zalloc_coherent instead of dma_alloc_coherent followed by memset
0.

Generated-by: scripts/coccinelle/api/alloc/kzalloc-simple.cocci
Suggested-by: NLuis R. Rodriguez <mcgrof@kernel.org>
Signed-off-by: NHimanshu Jha <himanshujha199640@gmail.com>
Acked-by: NManish Rangankar <Manish.Rangankar@cavium.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Use dma_zalloc_coherent instead of dma_alloc_coherent followed by memset
0.

Generated-by: scripts/coccinelle/api/alloc/kzalloc-simple.cocci
Suggested-by: NLuis R. Rodriguez <mcgrof@kernel.org>
Signed-off-by: NHimanshu Jha <himanshujha199640@gmail.com>
Acked-by: NChad Dupuis <chad.dupuis@cavium.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

According to ATA protocol, SET MAX commands belong to different frame
types. So judge features field of SET MAX commands to decide which
frame type they belongs to.
Signed-off-by: NXiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: NJohn Garry <john.garry@huawei.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

There are two other values for SET MAX feature field according to ata
protocol. So definite them.
Signed-off-by: NXiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: NJohn Garry <john.garry@huawei.com>
Acked-by: NTejun Heo <tj@kernel.org>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

If CONFIG_SCSI_SMARTPQI=y then don't build this driver as a module.
Signed-off-by: NSteffen Weber <steffen.weber@gmail.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Add resp_write_scat() function to support decoding WRITE SCATTERED
(16 and 32). Also weave resp_write_scat() into the cdb decoding
logic.
Signed-off-by: NDouglas Gilbert <dgilbert@interlog.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Reviewer suggested using the ARRAY_SIZE macro. That reduced one of the subtle
inter-dependencies in the parser's tables.

It is important that commands which simulate media access, indicate this in the
flags for that command. The flag to do that was FF_DIRECT_IO. On reflection
FF_MEDIA_IO seems a more accurate description.
Signed-off-by: NDouglas Gilbert <dgilbert@interlog.com>
Reviewed-by: NBart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

WRITE SCATTERED needs to take several "bites" out of the data-out buffer.
Expand the do_device_access() function to take a sg_skip argument.
Signed-off-by: NDouglas Gilbert <dgilbert@interlog.com>
Reviewed-by: NBart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Various cdb masks incorrectly assumed the GROUP NUMBER field
was 5 bits long. It is actually 6 bits long. Correct.
Also fix mask failure (in same byte) to allow DLD0 in READ(16)
and WRITE(16).
Signed-off-by: NDouglas Gilbert <dgilbert@interlog.com>
Reviewed-by: NBart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Some of my development tools tend to add spaces (my preference) rather
than tabs (kernel convention). Running unexpand to clean these spaces
up found more of them than checkpatch.pl did. Then checkpatch.pl
complained about other style violations in those newly tabbed lines.
Signed-off-by: NDouglas Gilbert <dgilbert@interlog.com>
Reviewed-by: NBart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Performance improvement using block layer tag.

Curent driver gets scsiio tracker and free smid from link list and array
based tracking managed by driver.  Accessing list in main io path is
performance pentaly because of protection using spinlock
"scsi_lookup_lock".

In this patch:

1. Driver removes all link list access from main io path and
   use scmd->request->tag to get free smid.

2. Instead of holding 'struct scsiio_tracker' in its own pool
   driver can embed it into the scsi command.

Driver provides cmd_size in scsi_host_template, so that struct
scsiio_tracker is preallocated by scsi mid layer for each scsi command.
Signed-off-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Use 'host_busy' instead of counting outstanding commands by hand.
Signed-off-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Move the check for outstanding commands out of the function allowing us
to simplify the overall code.
Signed-off-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

No functional change. Code optimization.

One can simply check 'target_busy' or 'device_busy' when figuring out if
there are outstanding commands; no need to painstakingly count them by
hand.

[mkp: tweaked patch description]
Signed-off-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

ioctl passthrough commands require a SCSIIO smid, but cannot easily
integrate with the block layer. But the driver already has reserved some
SCSIIO smids and we're only ever allowing one ioctl command at a time we
can use the first reserved smid for ioctl commands.
Signed-off-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

When attempting a command abort we should check the command status prior
to sending the abort; the command might've been completed already.
Signed-off-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Abstract accesses to the scsi_lookup array by introducing
mpt3sas_get_st_from_smid().
Signed-off-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Just a wrapper around the scsi lookup array and only used in one place,
so open-code it.
Signed-off-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

No functional change.
Signed-off-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Use 'list_splice_init()' instead of hand-crafted function.  No
functional change.
Signed-off-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

No functional change
Signed-off-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NSuganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

As staging to support future accelerator transports, add a shim layer
such that the underlying services the cxlflash driver requires can be
conditional upon the accelerator infrastructure.
Signed-off-by: NMatthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: NUma Krishnan <ukrishn@linux.vnet.ibm.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Adapter context creation can return either NULL or an error pointer.
Updating the check condition to reflect this.
Signed-off-by: NUma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: NMatthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The CXL-specific work structure used to request the number of interrupts
currently resides as a nested member of both the context information and
hardware queue structures. It is used to cache values (specifically the
number of interrupts) required by the CXL layer when starting a context.

To facilitate staging that will ultimately allow the cxlflash core to
become agnostic of the underlying accelerator transport, remove these
embedded work structures.
Signed-off-by: NMatthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: NUma Krishnan <ukrishn@linux.vnet.ibm.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

The number of interrupts a user requests during a context attach is
presently stored within the CXL work ioctl structure that is nested
alongside the per context metadata. Keeping this data in a structure
that is tied to a particular hardware implementation (CXL) will only
complicate matters when supporting newer accelerator transports.

Instead of relying upon the number of interrupts being cached within
a CXL-specific structure, explicitly cache the value within the context
information structure.
Signed-off-by: NMatthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: NUma Krishnan <ukrishn@linux.vnet.ibm.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Convert cxl-specific pointers to generic cookies to facilitate future
enhancements.
Signed-off-by: NUma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: NMatthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

In the event of a command failure, cxlflash returns the failure to the upper
layers to process. After processing the error, when the command is queued
again, the private command structure will not be zeroed and the ioasc could be
stale. Per the SISLite specification, the AFU only sets the ioasc in the
presence of a failure. Thus, even though the original command succeeds the
second time, the command is considered a failure due to stale ioasc. This
cycle repeats indefinitely and can cause a hang or IO failure.

To fix the issue, clear the ioasc before queuing any command.

[mkp: added Cc: stable per request]

Fixes: 479ad8e9 ("scsi: cxlflash: Remove zeroing of private command data")
Cc: <stable@vger.kernel.org>
Signed-off-by: NUma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: NMatthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

There are two places queuing the disco event DISCE_REVALIDATE_DOMAIN.
One is in sas_porte_broadcast_rcvd() and uses sas_chain_event() to queue
the event. The other is in sas_enable_revalidation() and uses
sas_queue_event() to queue the event. We have diffrent work queues for
event and discovery now, so the DISCE_REVALIDATE_DOMAIN event may be
processed in both event queue and discovery queue.

Now since we do synchronous event handling, we cannot do it in discovery
queue, so have to trigger a fake broadcast event to re-trigger the
revalidation from event queue.
Signed-off-by: NJason Yan <yanaijie@huawei.com>
CC: John Garry <john.garry@huawei.com>
CC: Johannes Thumshirn <jthumshirn@suse.de>
CC: Ewan Milne <emilne@redhat.com>
CC: Christoph Hellwig <hch@lst.de>
CC: Tomas Henzl <thenzl@redhat.com>
CC: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

In commit 87c8331f ("[SCSI] libsas: prevent domain rediscovery
competing with ata error handling") introduced disco mutex to prevent
rediscovery competing with ata error handling and put the whole
revalidation in the mutex. But the rphy add/remove needs to wait for the
error handling which also grabs the disco mutex. This may leads to dead
lock.So the probe and destruct event were introduce to do the rphy
add/remove asynchronously and out of the lock.

The asynchronously processed workers makes the whole discovery process
not atomic, the other events may interrupt the process. For example,
if a loss of signal event inserted before the probe event, the
sas_deform_port() is called and the port will be deleted.

And sas_port_delete() may run before the destruct event, but the
port-x:x is the top parent of end device or expander. This leads to
a kernel WARNING such as:

[   82.042979] sysfs group 'power' not found for kobject 'phy-1:0:22'
[   82.042983] ------------[ cut here ]------------
[   82.042986] WARNING: CPU: 54 PID: 1714 at fs/sysfs/group.c:237
sysfs_remove_group+0x94/0xa0
[   82.043059] Call trace:
[   82.043082] [<ffff0000082e7624>] sysfs_remove_group+0x94/0xa0
[   82.043085] [<ffff00000864e320>] dpm_sysfs_remove+0x60/0x70
[   82.043086] [<ffff00000863ee10>] device_del+0x138/0x308
[   82.043089] [<ffff00000869a2d0>] sas_phy_delete+0x38/0x60
[   82.043091] [<ffff00000869a86c>] do_sas_phy_delete+0x6c/0x80
[   82.043093] [<ffff00000863dc20>] device_for_each_child+0x58/0xa0
[   82.043095] [<ffff000008696f80>] sas_remove_children+0x40/0x50
[   82.043100] [<ffff00000869d1bc>] sas_destruct_devices+0x64/0xa0
[   82.043102] [<ffff0000080e93bc>] process_one_work+0x1fc/0x4b0
[   82.043104] [<ffff0000080e96c0>] worker_thread+0x50/0x490
[   82.043105] [<ffff0000080f0364>] kthread+0xfc/0x128
[   82.043107] [<ffff0000080836c0>] ret_from_fork+0x10/0x50

Make probe and destruct a direct call in the disco and revalidate function,
but put them outside the lock. The whole discovery or revalidate won't
be interrupted by other events. And the DISCE_PROBE and DISCE_DESTRUCT
event are deleted as a result of the direct call.

Introduce a new list to destruct the sas_port and put the port delete after
the destruct. This makes sure the right order of destroying the sysfs
kobject and fix the warning above.

In sas_ex_revalidate_domain() have a loop to find all broadcasted
device, and sometimes we have a chance to find the same expander twice.
Because the sas_port will be deleted at the end of the whole revalidate
process, sas_port with the same name cannot be added before this.
Otherwise the sysfs will complain of creating duplicate filename. Since
the LLDD will send broadcast for every device change, we can only
process one expander's revalidation.

[mkp: kbuild test robot warning]
Signed-off-by: NJason Yan <yanaijie@huawei.com>
CC: John Garry <john.garry@huawei.com>
CC: Johannes Thumshirn <jthumshirn@suse.de>
CC: Ewan Milne <emilne@redhat.com>
CC: Christoph Hellwig <hch@lst.de>
CC: Tomas Henzl <thenzl@redhat.com>
CC: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Now we are processing sas event and discover event in different
workqueues.  It's safe to wait the discover event done in the sas event
work. Use flush_workqueue() to insure the disco and revalidate events
processed synchronously so that the whole discover and revalidate
process will not be interrupted by other events.
Signed-off-by: NJason Yan <yanaijie@huawei.com>
CC: John Garry <john.garry@huawei.com>
CC: Johannes Thumshirn <jthumshirn@suse.de>
CC: Ewan Milne <emilne@redhat.com>
CC: Christoph Hellwig <hch@lst.de>
CC: Tomas Henzl <thenzl@redhat.com>
CC: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>

Now all libsas works are queued to scsi host workqueue, include sas
event work post by LLDD and sas discovery work, and a sas hotplug flow
may be divided into several works, e.g libsas receive a
PORTE_BYTES_DMAED event, currently we process it as following steps:

sas_form_port  --- run in work in shost workq
	sas_discover_domain  --- run in another work in shost workq
		...
		sas_probe_devices  --- run in new work in shost workq
We found during hot-add a device, libsas may need run several
works in same workqueue to add device in system, the process is
not atomic, it may interrupt by other sas event works, like
PHYE_LOSS_OF_SIGNAL.

This patch is preparation of execute libsas sas event in sync. We need
to use different workqueue to run sas event and disco event. Otherwise
the work will be blocked for waiting another chained work in the same
workqueue.
Signed-off-by: NYijing Wang <wangyijing@huawei.com>
CC: John Garry <john.garry@huawei.com>
CC: Johannes Thumshirn <jthumshirn@suse.de>
CC: Ewan Milne <emilne@redhat.com>
CC: Christoph Hellwig <hch@lst.de>
CC: Tomas Henzl <thenzl@redhat.com>
CC: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: NJason Yan <yanaijie@huawei.com>
Reviewed-by: NHannes Reinecke <hare@suse.com>
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>