Per the recent discussions at the Linux Plumbers Conference, when
enabling MSI or MSI-X, generate a test interrupt to verify the
interrupt routing is working properly.  If the test interrupt fails,
fall back to MSI first, and if that fails as well, to INTx.  If the
interrupt test fails with INTx, log an error and fail the PCI probe.

Also changed the use of spin_(lock|unlock) to the _irq(save|restore)
variants in the interrupt handlers because with multi-message MSI-X,
both interrupt handlers can now run in parallel.
Signed-off-by: NJames Smart <James.Smart@emulex.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

- Avoid polling HBA Error Attention when HBA's PCI channel is offline
  due to PCI EEH

- Fix handling of RSCN with non-zero event qualifiers

- Remove unnecessary sleeps during HBA initialization which slow down
  driver load

- Fix internal and external loopback on FCoE HBAs

- Fix incorrect decrement of cmd_pending count in lpfc_queuecomand
  error path

- Fix reporting of port busy events to management application

- Rename lpfc_adjust_queue_depth() to lpfc_rampdown_queue_depth() for
  consistency with its partner lpfc_rampup_queue_depth()

- Delete redundant lpfc_cmd->start_time = jiffies assignment in
  lpfc_queuecommand()

- Fix handling for ELS, mailbox and heartbeat time outs in the worker
  thread by removing unnecessary checking of the work_port_events
  flags.

- Fix NULL pointer dereference in lpfc_prep_els_iocb

- In lpfc_device_recov_npr_node(), move clearing of NLP_NPR_2B_DISC
  flag after call to lpfc_cancel_retry_delay_tmo() to keep
  targets-in-discovery count correct

- Remove lpfc_probe_one()'s call to scsi_scan_host() which could cause
  concurrent SCSI scans to step on each other
Signed-off-by: NJames Smart <James.Smart@emulex.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

The phba hbq_get address was setup before the HBA reported support for
Interrupt Enable Block (INB).  This caused loopback tests to fail
after all HBQs had been used once.  Moved setting of hba_get until
after config port mailbox command completes.
Signed-off-by: NJames Smart <James.Smart@emulex.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

- Added FC_REG_VPORTRSCN_EVENT to lpfc_nl.h

- Added code to provide option ROM version from HBA and via sysfs

- Added support for HPS bit in config port mailbox command to tell HBA
  that host group pointers are in host memory.

- Bugfix for Extended Error Handling (EEH) support on IBM PowerPC P6
  platform with MSI enabled
Signed-off-by: NJames Smart <James.Smart@emulex.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

- Fix lock up on rmmod with vports defined by having
  lpfc_pci_remove_one() invoke fc_vport_terminate() to remove all the
  vports before invoking fc_remove_host() for the physical port

- Fix echotest failure when NPIV is enabled

- Add the vport_disable function to the physical port's transport
  template to make the vport disable attribute visible

- Set the vport state to DISABLE on create if the disable flag is
  true

- Call lpfc_alloc_sysfs_attr() for vports so that statistical data
  collection works on them

- Support setting a vport's symbolic name via sysfs by writing to
  /sys/class/fc_vport/vportX/symbolic_name

- Fix create vport fails when link is down or in loop mode.  Should be
  able to be create vports any time NPIV is enabled

- Fix slow vport deletes when deleting multiple vports at once
Signed-off-by: NJames Smart <James.Smart@emulex.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

In the IOCB completion handler, always check if the node is valid
before accessing the node object.  Added lpfc_initialize_node() to
initialize nodes.
Signed-off-by: NJames Smart <James.Smart@emulex.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Add kernel-doc function headers to lpfc_scsi.c
Signed-off-by: NJames Smart <James.Smart@emulex.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Implement lpfc_pci_suspend_one() and lpfc_pci_resume_one() and
register them in the pci_driver table.
Signed-off-by: NJames Smart <James.Smart@emulex.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Rework RSCN netlink event to send entire RSCN payload

Also replaces (SCSI_NL_VID_TYPE_PCI | PCI_VENDOR_ID_EMULEX) with
LPFC_NL_VENDOR_ID
Signed-off-by: NJames Smart <James.Smart@emulex.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

This patch changes scsi_probe_lun() to retry INQUIRY if the device has
not actually sent back any INQUIRY data,

This enables the Thecus N2050 storage device to work better. The
firmware on that device starts up strangely; it sends no data in
response to the initial INQUIRY, and it sends the INQUIRY information
in response to the followup REQUEST SENSE. But after that it works
better, so retrying the INQUIRY is enough to get it going.
Signed-off-by: NFUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

scsi_execute() and scsi_execute_req() discard the residual length
information. Some callers need it. This adds residual argument
(optional) to scsi_execute and scsi_execute_req.
Signed-off-by: NFUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

In order to ensure the VIOS sees a consistent command buffer, we
need to add a memory barrier after building the command buffer
but before sending the command.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Adds a delay prior to retrying a failed NPIV login. This fixes
a scenario if the backing fibre channel adapter is getting reset
due to an EEH event, NPIV login will fail. Currently, ibmvfc
retries three times very quickly, resets the CRQ and tries one
more time. If the adapter is getting reset due to EEH, this isn't
enough time. This adds a delay prior to retrying a failed NPIV
login and also increments the number of retries.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

cxgb3i does not offload the processing of the header,
but it will always process the padding. This patch
adds a padding offload flag to detect when the LLD
supports this.

The patch also modifies the header processing so that
we do not try to read/bypass the header dugest in the
skb. cxgb3i will not include it with the header like
with other offload cards.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Mgmt setup used to not fail so we did not have to check
the return value. Now with cxgb3i it can so this has us
pass up a error.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

We do not need to allocate a itt for data_out, so this
passes the opcode to the alloc_pdu callout.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

bnx2i and cxgb3i need to encode LLD info in the itt so that
the firmware/hardware can process the pdu. This patch allows
the LLDs to encode info in the task->hdr->itt that they
setup in the alloc_pdu callout (any resources that are allocated
can be freed with the pdu in the cleanup_task callout). If
the LLD encodes info in the itt they should implement a
parse_pdu_itt callout. If parse_pdu_itt is not implemented
libiscsi will do the right thing for the LLD.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

This hooks iscsi_tcp into the libiscsi_tcp module and removes
code that is now in libiscsi_tcp.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

As explained in the previous mails, cxgb3i needs iscsi_tcp's
r2t/data_out and data_in procesing so this just moves functions
that both drivers want to use to a new module libiscsi_tcp. The
next patch will hook iscsi_tcp in.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Add iscsi_tcp prefix to most functions. Some are not changed
becuase they are going to move in the next patch.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

This modifies the login buffer allocation to use __get_free_pages.
It will allow drivers that want to send this data with zero copy
operations to easily line things up on page boundaries.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

libiscsi's iscsi_prep_data_out_pdu now handles what
iscsi_tcp's helpers were so we can remove iscsi_tcp's helpers.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

This converts iscsi_tcp to the new api and modifies how
it handles r2ts.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

cxgb3i offloads data transfers. It does not offload the entire scsi/iscsi
procssing like qla4xxx and it does not offload the iscsi sequence
processing like how bnx2i does. cxgb3i relies on iscsi_tcp for the
seqeunce handling so this changes how we transfer unsolicitied data by
adding a common r2t struct and helpers.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

cxgb3i is unlike qla4xxx and bnx2i in that it does not offload entire
scsi commands or iscsi sequences. Instead it only offloads the transfer
of a ISCSI DATA_IN pdu's data, the digests and padding. This patch fixes up the
iscsi tcp recv path so that it exports its skb recv processing so
cxgb3i and other drivers can call them. All they have to do is pass
the function the skb with the hdr or data pdu header and this function
will do the rest.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Since we're trying to eliminate struct scsi_device timeout, the tape
driver has to be updated to use the block queue timeout instead.  The
tape use of scsi_device timeout looks to be self consistent, so I don't
think this necessarily fixes any bug, but it has to be done to allow me
to remove the timeout parameter from struct scsi_device.
Acked-by: NKai Makisara <Kai.Makisara@kolumbus.fi>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

On the Quadra 900 and 950 there are two ESP chips sharing one IRQ. Because
the shared IRQ is edge-triggered, we must make sure that an IRQ transition
from one chip doesn't go unnoticed when the shared IRQ is already active
due to the other. This patch prevents interrupts getting lost so that both
SCSI busses may be used simultaneously.
Signed-off-by: NFinn Thain <fthain@telegraphics.com.au>
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Fix asm constraints and arguments so as not to transfer an odd byte when
there may be more words to transfer. The bug would probably also cause
exceptions sometimes by transferring one too many bytes.
Signed-off-by: NFinn Thain <fthain@telegraphics.com.au>
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

The following patch changes the handling of bus reset when issued from a
vport. In the bus reset code, an extra check is made to make sure that the lip
reset is not done before resetting the targets if the bus reset came from a
vport.
Signed-off-by: NAnirban Chakraborty <anirban.chakraborty@qlogic.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

This patch (as1142b) consolidates a lot of repetitious code in
scsi_io_completion().  It also fixes a few comments.  Most
importantly, however, it clearly distinguishes among the three sorts
of retries that can be done when a command fails to complete:

	Unprepare the request and resubmit it, so that a new
	command will be created for it.

	Requeue the request directly so that it will be retried
	immediately using the same command.

	Requeue the request so that it will be retried following
	a short delay.

	Complete the remainder of the request with an I/O error.

[jejb: Updates
     1. For several error conditions, we would now print the sense twice
        in slightly different ways, so unify the location of sense
        printing.
     2. I added more descriptions to actual failure conditions for
        better debugging
     3. according to spec, ABORTED_COMMAND is supposed to be retried
        (except on DIF failure).  Our old behaviour of erroring it looks
        to be a bug.
     4. I'd prefer not to default initialise the action variable because
        that ensures that every leg of the error handler has an
        associated action and the compiler will warn if someone later
        accidentally misses one or removes one.
]
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Bump driver version
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

The virtual fibre channel stack can return a failure response for a command
indicating the port login has been invalidated without sending the client
an async event. Add code to handle this response and initiate a PLOGI.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

The CRQs used by the ibmvfc driver are read and written by both
the client and the server. Therefore, we need to mark them volatile
so that we do not cache their contents when handling an interrupt.
This fixes a problem which can surface as occasional command timeouts.
No commands were actually timing out, but due to accessing cached data
for the CRQ in the interrupt handler, the interrupt was not processing
all command completions as it should.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Fixes an oops that can occur in the interrupt handler
if we get a lot of async events.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

[jejb: fixed whitespace damage]
Signed-off-by: NBabu Moger <Babu.Moger@lsi.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

drivers/scsi/qla2xxx/qla_os.c: In function 'qla2x00_probe_one':
drivers/scsi/qla2xxx/qla_os.c:1582: warning: 'mem_only' is used uninitialized in this function
Reported-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NAnirban Chakraborty <anirban.chakraborty@qlogic.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Following changes have been made:

1. Outstanding commands are based on a request queue, scsi_qla_host
   does not maintain it anymore.

2. start_scsi is accessed via isp_ops struct instead of direct
   invocation.

3. Interrupt registrations are done using response queue instead of
   device id.
Signed-off-by: NAnirban Chakraborty <anirban.chakraborty@qlogic.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Following changes have been made to the qla2xxx FC driver in
preparation for the multi- queue and future SR IOV hardware.

1. scsi_qla_host structure has been changed to contain scsi host
   specific data only.

2. A new structure, qla_hw_data is created to contain HBA specific
   hardware data.

3. Request and response IO specific data strucures are created.

4. The global list of fcports for the hba is not maintained anymore,
   instead a fcport list is construted on per scsi_qla_host.
Signed-of-by: NAnirban Chakraborty <anirban.chakraborty@qlogic.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

This patch improves handling of TASK ABORTED status by Linux SCSI
mid-layer. Currently, command returned with this status considered
failed and returned to upper layers. It leads to additional error
recovery load on file systems and block layer, which sometimes can
cause undesired side effects, like I/O errors and file systems
corruptions. See http://lkml.org/lkml/2008/11/1/38, for instance.

From other side, TASK ABORTED status is returned by SCSI target if the
corresponding command was aborted by another initiator and the target
has TAS bit set in the control mode page. So, in the majority of cases
commands with TASK ABORTED status should be simply retried. In other
cases, maybe_retry path will not retry if no retries are allowed.

This patch implement suggestion by James Bottomley from
http://marc.info/?l=linux-scsi&m=121932916906009&w=2.
Signed-off-by: NVladislav Bolkhovitin <vst@vlnb.net>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

When the mode select sent to the controller fails with the retryable
error, it is better to retry the mode_select from the hardware handler
itself, instead of propagating the failure to dm-multipath.
Signed-off-by: NChandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>