linux-stable.git/drivers/infiniband/ulp/iser, branch linux-3.17.y

IB/iser: Bump version to 1.4.1

2014-09-22T16:46:42+00:00

Signed-off-by: Roi Dayan 
Signed-off-by: Or Gerlitz 
Signed-off-by: Roland Dreier

IB/iser: Allow bind only when connection state is UP

2014-09-22T16:46:42+00:00

We need to fail the bind operation if the iser connection state != UP
(started teardown) and this should be done under the state lock.

Signed-off-by: Sagi Grimberg 
Signed-off-by: Or Gerlitz 
Signed-off-by: Roland Dreier

IB/iser: Fix RX/TX CQ resource leak on error flow

2014-09-22T16:46:42+00:00

When failing to allocate TX CQ we already allocated RX CQ, so we need to make
sure we release it. Also, when failing to register notification to the RX CQ
we currently leak both RX and TX CQs of the current index, fix that too.

Signed-off-by: Roi Dayan 
Signed-off-by: Sagi Grimberg 
Signed-off-by: Or Gerlitz 
Signed-off-by: Roland Dreier

IB/iser: Clarify a duplicate counters check

2014-08-01T22:10:05+00:00

This is to prevent someone from thinking that this code section is
redundant.

Signed-off-by: Ariel Nahum 
Signed-off-by: Roi Dayan 
Signed-off-by: Or Gerlitz 
Signed-off-by: Roland Dreier

IB/iser: Replace connection waitqueue with completion object

2014-08-01T22:10:05+00:00

Instead of waiting for events and condition changes of the iser
connection state, we wait for explicit completion of connection
establishment and teardown.

Separate connection establishment wait object from the teardown object
to avoid a situation where racing connection establishment and
teardown may concurrently wakeup each other.

ep_poll will wait for up_completion invoked by
iser_connected_handler() and iser release worker will wait for
flush_completion before releasing the connection.

Bound the completion wait with a 30 seconds timeout for cases where
iscsid (the user space iscsi daemon) is too slow or gone.

Signed-off-by: Ariel Nahum 
Signed-off-by: Sagi Grimberg 
Signed-off-by: Or Gerlitz 
Signed-off-by: Roland Dreier

IB/iser: Protect iser state machine with a mutex

2014-08-01T22:10:05+00:00

The iser connection state lookups and transitions are not fully protected.

Some transitions are protected with a spinlock, and in some cases the
state is accessed unprotected due to specific assumptions of the flow.

Introduce a new mutex to protect the connection state access. We use a
mutex since we need to also include a scheduling operations executed
under the state lock.

Each state transition/condition and its corresponding action will be
protected with the state mutex.

The rdma_cm events handler acquires the mutex when handling connection
events. Since iser connection state can transition to DOWN
concurrently during connection establishment, we bailout from
addr/route resolution events when the state is not PENDING.

This addresses a scenario where ep_poll retries expire during CMA
connection establishment. In this case ep_disconnect is invoked while
CMA events keep coming (address/route resolution, connected, etc...).

Signed-off-by: Ariel Nahum 
Signed-off-by: Sagi Grimberg 
Signed-off-by: Or Gerlitz 
Signed-off-by: Roland Dreier

IB/iser: Remove redundant return code in iser_free_ib_conn_res()

2014-08-01T22:10:04+00:00

Make it void.

Signed-off-by: Sagi Grimberg 
Signed-off-by: Or Gerlitz 
Signed-off-by: Roland Dreier

IB/iser: Seperate iser_conn and iscsi_endpoint storage space

2014-08-01T22:10:04+00:00

iser connection needs asynchronous cleanup completions which are
triggered in ep_disconnect.  As a result we are keeping the
corresponding iscsi_endpoint structure hanging for no good reason. In
order to avoid that, we seperate iser_conn from iscsi_endpoint storage
space to have their destruction being independent.

iscsi_endpoint will be destroyed at ep_disconnect stage, while the
iser connection will wait for asynchronous completions to be released
in an orderly fashion.

Signed-off-by: Ariel Nahum 
Signed-off-by: Roi Dayan 
Signed-off-by: Sagi Grimberg 
Signed-off-by: Or Gerlitz 
Signed-off-by: Roland Dreier

IB/iser: Fix responder resources advertisement

2014-08-01T22:10:04+00:00

The iser initiator is the RDMA responder so it should publish to the
target the max inflight rdma read requests its local HCA can handle in
responder_resources (max_qp_rd_atom).

The iser target should take the min of that and its local HCA max
inflight oustanding rdma read requests (max_qp_init_rd_atom).

We keep initiator_depth set to 1 in order to compat with old targets.

Signed-off-by: Sagi Grimberg 
Signed-off-by: Or Gerlitz 
Signed-off-by: Roland Dreier

IB/iser: Add TIMEWAIT_EXIT event handling

2014-08-01T22:10:04+00:00

In case the DISCONNECTED event is not delivered after rdma_disconnect
is called, the CM waits TIMEWAIT seconds and delivers the
TIMEWAIT_EXIT local event. We use this as the notification needed to
continue in the teardown and release sequence.

Signed-off-by: Ariel Nahum 
Signed-off-by: Roi Dayan 
Reviewed-by: Sagi Grimberg 
Signed-off-by: Or Gerlitz 
Signed-off-by: Roland Dreier