In the case of I/O requests that fail at submit time because of a
pending reset condition, the host lock for SATA/STP devices must be
managed for any SCSI-initiated I/O before sas_task_abort is called.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

When a TMF times-out, the request is set back to "aborting".
Requests in the "aborting" state must be terminated when
LUN and device resets occur.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Converting the all CAPS data types to lower case.
Reported-by: NChristoph Hellwig <hch@infradead.org>
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

These routines are just stubs, re-add them when / if they are needed.  Also
cleanup remote_device_stopped.
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

* mark needlessly global routines static
* delete unused functions
* move kernel-doc blocks from header files to source
* reorder some functions to delete declarations
* more default handler cleanups phy
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Delete some macros that are longer to type than the open coded operation
that they perform.

scic_sds_phy_get_base_state_machine
scic_sds_phy_get_starting_substate_machine
scic_sds_port_get_base_state_machine
scic_sds_port_get_ready_substate_machine
scic_sds_remote_device_get_base_state_machine
scic_sds_remote_device_get_ready_substate_machine
scic_sds_remote_node_context_set_remote_node_index
scic_sds_controller_get_base_state_machine

Also performs some collateral cleanups like killing casts that assume
structure member ordering, and consolidating a lot of duplicated default
handler code (the primary callers of the *_get_base_state_machine macros) via
a helper.
Reported-by: NChristoph Hellwig <hch@infradead.org>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Just move isci_pci_driver below the function definitions and delete the
declarations.  A couple other whitespace fixups, and unused symbol
deletions.
Reported-by: NChristoph Hellwig <hch@infradead.org>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Removed isci_event_* calls and call those functions directly.
Reported-by: NChristoph Hellwig <hch@infradead.org>
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Use min_t to address:
drivers/scsi/isci/probe_roms.c: In function ‘isci_get_efi_var’:
drivers/scsi/isci/probe_roms.c:241: warning: comparison of distinct pointer types lacks a cast
Reported-by: NDavid Milburn <dmilburn@redhat.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

The original apc mode definition is the correct one, the fix from commit
4711ba10 "isci: fix oem parameter initialization and mode detection" was based
on a typo from a specification update.
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Exposing the user config parameters through the kernel module parameters.
The kernel module params will have the default values set and we will no
longer pulling the default values for user params from the core.
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

1/ Since commit 858d4aa7 "isci: Move firmware loading to per PCI device" we have
   been silently falling back to built-in defaults for the parameter settings by
   skipping the call to scic_oem_parameters_set().

2/ The afe parameters from the firmware were not being honored

3/ The latest oem parameter definition flips the mode_type values which are
   now 0: for APC 1: for MPC.  For APC we need to make sure all the phys
   default to the same address otherwise strict_wide_ports will cause duplicate
   domains.

4/ Fix up the driver announcement to indicate the source of the
   parameters.

5/ Fix up the sas addresses to be unique per controller (in the fallback case)
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Updating the EFI variable OEM parameter retrieval after examining the EFI
variable exported via sysfs.
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Added fixups for the OROM parsing code after testing with BIOS OROM
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Since the data structure for oem from orom/efi/firmware is the same as what
the core uses, we can just do a direct copy instead of assignment.
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

These are the finalized values that the driver can expect to see in
production.
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

1/ add OEM paramater support for mode_type (MPC vs APC)
2/ add OEM parameter support for max_number_concurrent_device_spin_up
3/ cleanup scic_sds_controller_start_next_phy

todo: hook up the amp control afe parameters into the afe init code
Signed-off-by: NHenryk Dembkowski <henryk.dembkowski@intel.com>
Signed-off-by: NJacek Danecki <Jacek.Danecki@intel.com>
[cleaned up scic_sds_controller_start_next_phy]
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Adding EFI variable retrieving for OEM parameters. Still need GUID and
variable name.

Also updated the data struct for oem parameters and hex file for firmware
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
[fix CONFIG_EFI=n compile error]
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

We need to scan the OROM for signature and grab the OEM parameters. We
also need to do the same for EFI. If all fails then we resort to user
binary blob, and if that fails then we go to the defaults.

Share the format with the create_fw utility so that all possible sources
of the parameters are in-sync.
Signed-off-by: NDave Jiang <dave.jiang@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Don't assume the hardware is in a known state at init.
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

A usage of "FALSE" leaked in as well as some checkpatch escapes.
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

In the case where submitted I/Os fail with the status code
SCI_FAILURE_REMOTE_DEVICE_RESET_REQUIRED, the execute function now waits
until scic_lock is cleared before calling the helper function
"isci_request_signal_device_reset" which sets the flag for the pending
reset condition on the I/O.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

A domain_device has the same lifetime as its related scsi_target.  The
scsi_target is reference counted based on outstanding commands,
therefore it is safe to assume that if we have a valid sas_task that the
->dev pointer is also valid.

The asd_sas_port of a domain_device has the same lifetime as the driver
so it can also never be NULL as long as the sas_task is valid and the
driver is loaded.

This also cleans up isci_task_complete_for_upper_layer(), renames it to
isci_task_refuse() and notices that the isci_completion_selection
parameter was set to isci_perform_normal_io_completion by all callers.
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Allow each controller to be identified via sysfs.

# cat /sys/class/scsi_host/host13/isci_id
1
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Make sure all pending I/O including any in the libsas error handler
process is cleaned-up.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

In the case of I/O requests being failed because of a required device
reset condition, set the response and status to indicate an I/O failure.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Since libsas takes the domain device sata_dev.ap->lock before submitting
a task, error completions in the submit path for SATA devices must
unlock/relock when completing the sas_task back to libsas.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

The request may be in the "aborted" or the "completed" state when
performing a task management operation on it.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

In the case where a SAS or SATA LUN reset TMF is built a NULL pointer
dereference occurred because of the (unused) callback data pointer.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>
Signed-off-by: NJacek Danecki <Jacek.Danecki@intel.com>

Added a request "dead" state for use when a termination wait times-out.

isci_terminate_pending_requests now detaches the device's pending list
and terminates each entry on the detached list.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Since the request structure contains a pointer to the completion to be
used if the request is being aborted or terminated, there is no reason
to pass the completion as a pointer to isci_terminate_request_core().
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NJacek Danecki <Jacek.Danecki@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Made sure the device ready check accounts for all states.
Moved the aborted task check into the loop of pulling task requests
off of the submitted list.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NJacek Danecki <Jacek.Danecki@intel.com>
[remove host and device starting state checks]
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

The pointer to the core representation of a request is marked NULL at
completion, but we need to save the i/o tag for task management.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NJacek Danecki <Jacek.Danecki@intel.com>
[revise changelog]
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

If there is a pending device reset, the I/O is used to accomplish the reset by setting the
RESET bit in the task status, and then putting the task into the error handler
path using sas abort task.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NJacek Danecki <Jacek.Danecki@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Corrected use of the request state_lock in the completion callback.

In the case where an abort (or reset) thread is trying to terminate an
I/O request, it sets the request state to "aborting" (or "terminating")
if the state is still "starting".  One of the bugs was to never set the
state to "completed".  Another was to not correctly recognize the
situation where the I/O had completed but the sas_task was still pending
callback to task_done - this was typically a problem in the LUN and
device reset cases.

It is now possible that we leave isci_task_abort_task() with
request->io_request_completion pointing to localy allocated
aborted_io_completion struct. It may result in a system crash.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NMaciej Trela <Maciej.Trela@intel.com>
Signed-off-by: NJacek Danecki <Jacek.Danecki@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Changes to move management of the reqs_in_process entry for the request here.
Made changes to note when the task is already in the abort path and
cannot be completed through callbacks.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NJacek Danecki <Jacek.Danecki@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

In the condition where outstanding I/Os are being cleaned from the device
requests in process list, the cleanup function needs to check that the
request is actually a sas-task and not a task management function.
Signed-off-by: NJeff Skirvin <jeffrey.d.skirvin@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Reported-by: NJames Bottomley <James.Bottomley@suse.de>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

The remote_device_lock is currently used to protect a controller global
resource (RNCs), but the remote_device_lock is per-port.
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Until we synchronize against device removal this limits the damage of
use after free bugs to the driver's own objects.  Unless we implement
reference counting we need to ensure at least a subset of a remote
device is valid at all times.  We follow the lead of other libsas
drivers that also preallocate devices.

This also enforces maximum remote device accounting at the lldd layer,
but the core may still run out of RNC's before we hit this limit.
Signed-off-by: NDan Williams <dan.j.williams@intel.com>