=======================================================
[ INFO: possible circular locking dependency detected ]
2.6.31-39.x.20090917-s390xdefault #1
-------------------------------------------------------
kslowcrw/83 is trying to acquire lock:
 (&adapter->scan_work){+.+.+.}, at: [<0000000000169c5c>] __cancel_work_timer+0x64/0x3d4

but task is already holding lock:
 (&zfcp_data.config_mutex){+.+.+.}, at: [<00000000004671ea>] zfcp_ccw_remove+0x66/0x384

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #1 (&zfcp_data.config_mutex){+.+.+.}:
       [<0000000000189962>] __lock_acquire+0xe26/0x1834
       [<000000000018a4b6>] lock_acquire+0x146/0x178
       [<000000000058cb5a>] mutex_lock_nested+0x82/0x3ec
       [<0000000000477170>] zfcp_fc_scan_ports+0x3ec/0x728
       [<0000000000168e34>] worker_thread+0x278/0x3a8
       [<000000000016ff08>] kthread+0x9c/0xa4
       [<0000000000109ebe>] kernel_thread_starter+0x6/0xc
       [<0000000000109eb8>] kernel_thread_starter+0x0/0xc

-> #0 (&adapter->scan_work){+.+.+.}:
       [<0000000000189e60>] __lock_acquire+0x1324/0x1834
       [<000000000018a4b6>] lock_acquire+0x146/0x178
       [<0000000000169c9a>] __cancel_work_timer+0xa2/0x3d4
       [<0000000000465cb2>] zfcp_adapter_dequeue+0x32/0x14c
       [<00000000004673e4>] zfcp_ccw_remove+0x260/0x384
       [<00000000004250f6>] ccw_device_remove+0x42/0x1ac
       [<00000000003cb6be>] __device_release_driver+0x9a/0x10c
       [<00000000003cb856>] device_release_driver+0x3a/0x4c
       [<00000000003ca94c>] bus_remove_device+0xcc/0x114
       [<00000000003c8506>] device_del+0x162/0x21c
       [<0000000000425ff2>] ccw_device_unregister+0x5e/0x7c
       [<000000000042607e>] io_subchannel_remove+0x6e/0x9c
       [<000000000041ff9a>] css_remove+0x3e/0x7c
       [<00000000003cb6be>] __device_release_driver+0x9a/0x10c
       [<00000000003cb856>] device_release_driver+0x3a/0x4c
       [<00000000003ca94c>] bus_remove_device+0xcc/0x114
       [<00000000003c8506>] device_del+0x162/0x21c
       [<00000000003c85e8>] device_unregister+0x28/0x38
       [<0000000000420152>] css_sch_device_unregister+0x46/0x58
       [<00000000004276a6>] io_subchannel_sch_event+0x28e/0x794
       [<0000000000420442>] css_evaluate_known_subchannel+0x46/0xd0
       [<0000000000420ebc>] slow_eval_known_fn+0x88/0xa0
       [<00000000003caffa>] bus_for_each_dev+0x7e/0xd0
       [<000000000042188c>] for_each_subchannel_staged+0x6c/0xd4
       [<0000000000421a00>] css_slow_path_func+0x54/0xd8
       [<0000000000168e34>] worker_thread+0x278/0x3a8
       [<000000000016ff08>] kthread+0x9c/0xa4
       [<0000000000109ebe>] kernel_thread_starter+0x6/0xc
       [<0000000000109eb8>] kernel_thread_starter+0x0/0xc

cancel_work_sync is called while holding the config_mutex. But the
work that is being cancelled or flushed also uses the config_mutex.
Fix the resulting deadlock possibility by calling cancel_work_sync
earlier without holding the mutex. The best place to do is is after
offlining the device.  No new port scan work will be scheduled for the
offline device, so this is a safe place to call cancel_work_sync.
Reviewed-by: NFelix Beck <felix.beck@de.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

With the change for delaying the allocation of zfcp_adapter, the
initial device parameter function has to first call
ccw_device_set_online which allocates the zfcp_adapter structure.
Change this and adapt the cfdc part accordingly.
Reviewed-by: NFelix Beck <felix.beck@de.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

dev_set_name tries to allocate memory, so check the return value for
allocation failures. After dev_set_name succeeds, call device_register
as next step to be able to use put_device during error handling.
Reviewed-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Don't use kfree directly after device registration started.
Signed-off-by: NSebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

The config semaphore is only used as a mutex, so replace it with a
simple mutex.
Reviewed-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

So far, zfcp allocated all resources required for FCP
adapters/subchannels when the device was discovered in the ccw_probe
callback. If there are lots of unused FCP subchannels attached to a
system, this is a waste of resources. To alleviate this, defer the
resource allocation to the first call to ccw_set_online. To avoid
disruptions during possible following calls to ccw_set_offline and
then ccw_set_online, keep the adapter resources until the device is
finally being removed via ccw_remove. While doing this, also manage
the zfcp erp thread together with all other adapter resources in
zfcp_adapter_enqueue/dequeue.
Reviewed-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Switch the creation of the zfcp erp thread from the deprecated
kernel_thread API to the kthread API. This allows also the removal of
some flags in zfcp since the kthread API handles thread creation and
shutdown internally. To allow the usage of the kthread_stop function,
replace the erp ready semaphore with a waitqueue for waiting until erp
actions arrive on the ready queue.
Reviewed-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

The fc_rport structure reserves a reference where a LLD can put
information required in a situation where the fc transport class is
triggering LLD callbacks. The zfcp driver was using this variable
directly which is discouraged. This patch solves this issue by making
this reference unnecessary.  In addition the dev_loss_tmo callback is
removed, it is not required: zfcp does not access the fc_rport after
calling fc_remote_port_delete.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Update the Fibre Channel related code to use the zfcp_fc prefix.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Extract independent data structures and introduce common _setup and
_destroy routines for QDIO and Fibre Channel related data structures
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Change the dbf data and functions to use the zfcp_dbf prefix
throughout the code. Also change the calls to dbf to use zfcp_dbf
instead of zfcp_adapter.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Don't let the erp wait for gid_pn requests to complete. Instead, queue
the gid_pn work, exit erp and let the finished gid_pn work trigger a
new port reopen.
Reviewed-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

The zfcp_adapter structure was growing over time to a size of almost
one memory page. To reduce the size of the data structure and to
seperate different layers, put all qdio related data in the new
zfcp_qdio data structure.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Remove the global driver work queue and replace it with a workqueue
local to the adapter. The usage of this workqueue makes this the
correct place for the structure. In addition multiple adapters won't
block each other due to the serialization of the queued work.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Remove the special case for NO_QTCB requests and optimize the
mempool and cache processing for fsfreqs. Especially use seperate
mempools for the zfcp_fsf_req and zfcp_qtcb structs.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

An adapter shutdown implicitly closes all open ports. Make sure to
mark all WKA ports as offline, not only the directory server. Also
make sure that no pending wka port work is running when the adapter is
being removed.
Reviewed-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

The struct zfcp_adapter includes everything related to the debug
traces. This introduces dependences between the definitions in
zfcp_def.h and zfcp_dbf.h. Move all debug related data structures to a
new data structure to break those dependencies and manage the debug
data in zfcp_dbf.[hc].
Reviewed-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

In certain error scenarios ports, rports are getting attached,
validated and removed from the systems environment. Depending on the
layer this occurs asynchronously. This patch fixes the few races
which existed and ensures all references and cross references are
cleared at the time they're invalid. In addition fc transports
actions are only scheduled when required.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Provide the ability to do fibre channel requests from the userspace to
our zfcp driver.  Patch builds upon extension to the fibre channel
tranport class by James Smart and Seokmann Ju.  See here
http://marc.info/?l=linux-scsi&m=123808882309133&w=2Signed-off-by: NSven Schuetz <sven@linux.vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

The zfcp_port might have been removed, while the FC fast_io_fail timer
is still running and could trigger the terminate_rport_io callback.
Set the pointer to the zfcp_port to NULL and check accordingly
before using it.
Reviewed-by: NMartin Petermann <martin@linux.vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

The nameserver port might be in state online when the adapter is
offlined. On adapter reactivation the nameserver port is not
re-opened due to the PORT_ONLINE status. This results in an
unsuccessful recovery. In forcing the nameserver port status
to offline on all adapter offline events this issue is prevented.

Waiting for the reference count to drop to zero in
zfcp_wka_port_offline is not required, so remove it.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

When running the scsi_scan from the zfcp workqueue and the target
device does not respond, the zfcp workqueue can block until the
scsi_scan hits a timeout. Move the work to the scsi host workqueue,
since this one is also used for the scan from the SCSI midlayer.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

A scheduled work might still be pending, running while the adapter is
in progress to get dequeued from the system. This can lead to an
invalid pointer dereference (Oops).  Once the adpater is set online
again, ensure the nameserver environment is initialized to the
appropriate values again.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Ensure the refcounting is correct even if we were not able to
schedule a work. In addition we have to make sure no scheduled
work is pending while we're dequeing the adapter from the
systems environment.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Use the I/O blocking mechanism in the FC transport class to allow
faster failovers for multipathing:
- Call fc_remote_port_delete early to set the rport to BLOCKED.
- Check the rport status in queuecommand with fc_remote_portchkready
  to no longer accept new I/O for this port and fail the I/O with the
  appropriate scsi_cmnd result.
- Implement the terminate_rport_io handler to abort all pending I/O
  requests
- Return SCSI commands with DID_TRANSPORT_DISRUPTED while erp is
  running.
- When updating the remote port status, check for late changes and
  update the remote ports status accordingly.
Acked-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

The current number based id ERP logging is replaced by a string
based tag version. The benefit is an easier location of the code in
question and the removal of the lengthy array referencing the
individual messages.
The string (7 bytes) based version does not use more space since those
bytes were "used" anyway due to the alignment of the structure.
The encoding of the 7 byte string is as follows
        [0-1] = filename
        [2-5] = task/function
        [6]   = section
Due to the character of this string (fixed length) a string
termination is not required here.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Issue ELS ADISC requests from workqueue. This allows the link test
request to be sent when the request queue is full due to I/O load for
other remote ports. It also simplifies request queue locking,
zfcp_fsf_send_fcp_command_task is now the only function that has
interrupts disabled from the caller. This is also a prereq for the FC
passthrough support that issues ELS requests from userspace.
Acked-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Provide measurement data for the utilisation of the QDIO outbound queue.
The additional value allows to calculate an average queue utilisation
by looking at the deltas per time unit. Needed for capacity planning.
It is up to user space to handle wrap-arounds of the 64 bit value.

The new counter neatly complements the existing counter for queue full
conditions. That is why, both statistics counter have been integrated.
Signed-off-by: NMartin Peschke <mpeschke@linux.vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

PORT_PHYS_CLOSING is only set and cleared, but not actually used
for status checking.

PORT_INVALID_WWPN is set when the GID_PN request does not return
a d_id for a remote port, e.g. when a remote port has been
unplugged. For this case, the d_id is zero. In the erp we can
check the d_id and use the normal escalation procedure that gives
up after three retries and remove the special case.

PORT_NO_WWPN is unused: Each port in the remote port list has a
valid wwpn. The WKA ports are now tracked outside the port
list. Remove the PORT_NO_WWPN flag, since this is no longer set
for any port.
Acked-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

The information from the kernel parameter is only needed during init.
Keep the three pieces (busid, wwpn and lun) local to simplify the
global zfcp_data structures. While at it, also remove the unused
loglevel variable and give the module parameter variable a better
name.
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Acked-by: NFelix Beck <felix@linux.vnet.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

After the latest changes, the list of FCP devices is only used to
lookup the adapter for requests from the actcli tool. Change this to
use the lookup function in the cio layer. Now we can remove the
adapter list and have one place less to use the global config_lock.
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Acked-by: NFelix Beck <felix@linux.vnet.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Synchronize the registration and de-registration with the SCSI layer
at CCW registration, de-registration. Before we registered with the
SCSI layer on adapter activation. This way the reg and de-reg process
is in balance.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NMartin Schwidefsky <schwidefsky@de.ibm.com>

Acked-by: NCornelia Huck <cornelia.huck@de.ibm.com>
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NKay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NHeiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: NMartin Schwidefsky <schwidefsky@de.ibm.com>

Register zfcp with the new /proc/service_level interface to report the
FCP microcode level. When the adapter goes offline or a channel path
disappears, zfcp unregisters, since the microcode version might change
and zfcp does not know about it.
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NMartin Schwidefsky <schwidefsky@de.ibm.com>

Fix leftover from last typecast patch:

drivers/s390/scsi/zfcp_aux.c: In function ‘zfcp_port_enqueue’:
drivers/s390/scsi/zfcp_aux.c:629: warning: format ‘%016llx’ expects
type ‘long long unsigned int’, but argument 3 has type ‘u64’
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Convert most s390 users setting bus_id to dev_set_name().
css and ccw busses are deferred since they need some special
treatment.
Signed-off-by: NCornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NMartin Schwidefsky <schwidefsky@de.ibm.com>

Due to the character of a scheduled work we cannot guarantee the
LUN register to be finished before an initial device tries to use it.
Therefor we have to wait for PENDING_SCSI_WORK flag to be cleared
before proceeding.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Each adapter reopen trigger automatically a scan_port task which
is waiting for the ERP to be finished before further processing.
Since the initial device setup enqueues adapter, port and LUN which
are individual ERP actions, this process would start after
everything is done. Unfortunately the port_reopen requires another
scheduled work to be finished which is queued after the automatic
scan_port -> deadlock !

This fix creates an own work queue for ERP based nameserver requests.
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Reduce the size of zfcp data structures by removing unused and
redundant members. scsi_lun is only the mangled version of the
fcp_lun. So, remove the redundant field and use the fcp_lun instead.

Since the queue lock and the pci_batch indicator are only used in the
request queue, move them from the common queue struct to the adapter
struct.
Signed-off-by: NChristof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: NSwen Schillig <swen@vnet.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>