Remove fastreg page list allocation as the page vector
is now private to the provider. Instead of constructing
the page list and fast_req work request, call ib_map_mr_sg
and construct ib_reg_wr.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Doug Ledford <dledford@redhat.com>
The block layer can reliably guarantee that SG lists won't
contain gaps (page unaligned) if a driver set the queue
virt_boundary.
With this setting the block layer will:
- refuse merges if bios are not aligned to the virtual boundary
- split bios/requests that are not aligned to the virtual boundary
- or, bounce buffer SG_IOs that are not aligned to the virtual boundary
Since iser is working in 4K page size, set the virt_boundary to
4K pages. With this setting, we can now safely remove the bounce
buffering logic in iser.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Doug Ledford <dledford@redhat.com>
This patch split up struct ib_send_wr so that all non-trivial verbs
use their own structure which embedds struct ib_send_wr. This dramaticly
shrinks the size of a WR for most common operations:
sizeof(struct ib_send_wr) (old): 96
sizeof(struct ib_send_wr): 48
sizeof(struct ib_rdma_wr): 64
sizeof(struct ib_atomic_wr): 96
sizeof(struct ib_ud_wr): 88
sizeof(struct ib_fast_reg_wr): 88
sizeof(struct ib_bind_mw_wr): 96
sizeof(struct ib_sig_handover_wr): 80
And with Sagi's pending MR rework the fast registration WR will also be
down to a reasonable size:
sizeof(struct ib_fastreg_wr): 64
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com> [srp, srpt]
Reviewed-by: Chuck Lever <chuck.lever@oracle.com> [sunrpc]
Tested-by: Haggai Eran <haggaie@mellanox.com>
Tested-by: Sagi Grimberg <sagig@mellanox.com>
Tested-by: Steve Wise <swise@opengridcomputing.com>
This module parameter forces memory registration even for
a continuous memory region. It is true by default as sending
an all-physical rkey with remote permissions might be insecure.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Chaning of send work requests benefits performance by
reducing the send queue lock contention (acquired in
ib_post_send) and saves us HW doorbells which is posted
only once.
Currently, in normal IO flows iser does not chain the CDB send
work request with the registration work request. Also in PI
flows, signature work requests are not chained as well.
Lets chain those and post only once.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
iser support up to 512KB data transfer in a single scsi command.
This means that larger IOs will split to different request. While
iser can easily saturate FDR/EDR wires, some arrays are fine tuned
for 1MB (or larger) IO sizes, hence add an option to support larger
transfers (up to 8MB) if the device allows it.
Given that a few target implementations don't support data transfers
of more than 512KB by default and the fact that larger IO sizes require
more resources, we introduce a module parameter to determine the
maximum number of 512B sectors in a single scsi command.
Users that are interested in larger transfers can change this value given
that the target supports larger transfers.
At the moment, iser works in 4K pages granularity, In a later stage
we will get it to work with system page size instead.
IO operations that consists of N pages will need a page vector
of size N+1 in case the first SG element contains an offset. Given
that some devices allocates memory regions in powers of 2, this
means that allocating a region with N+1 pages, will result in
region resources allocation of the next power of 2. Since we don't
want that to happen, in case we are in the limit of IO size supported
and the first SG element has an offset, we align the SG list using a
bounce buffer (which is OK given that this is not likely to happen a lot).
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Hard coded for now. This will allow to allocate different
sized MRs depending on the IO size needed (and device
capabilities).
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
iser_reg_rdma_mem_[fastreg|fmr] share a lot of code, and
logically do the same thing other than the buffer registration
method itself (iser_fast_reg_mr vs. iser_fast_reg_fmr).
The DIF logic is not implemented in the FMR flow as there is no
existing device that supports FMRs and Signature feature.
This patch unifies the flow in a single routine iser_reg_rdma_mem
and just split to fmr/frwr for the buffer registration itself.
Also, for symmetry reasons, unify iser_unreg_rdma_mem (which will
call the relevant device specific unreg routine).
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Adir Lev <adirl@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
As for fmrs we will hold a single registration descriptor
as no need for multiple like in the frwr mode (descriptor
for each task). This change helps unifying the duplicate
registration code paths.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Adir Lev <adirl@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
This will allow us to unify the memory registration code path between
the various methods which vary by the device capabilities. This change
will make it easier and less intrusive to remove fmr_pools from the
code when we'd want to.
The reason we use a single descriptor is to avoid taking a
redundant spinlock when working with FMRs.
We also change the signature of iser_reg_page_vec to make it match
iser_fast_reg_mr (and the future indirect registration method).
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Adir Lev <adirl@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Instead of having it a part of the connection structure,
have it be under a dedicated (embedded) structure in the
connection. A logical separation of the registration pool
and the connection structure.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Adir Lev <adirl@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Move all the per-device function pointers to an easy
extensible iser_reg_ops structure that contains all
the iser registration operations.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Avoid struct names without iser_ prefix.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Have fast_reg_descriptor hold struct iser_reg_resources
(mr, frpl, valid flag). This will be useful when the
actual buffer registration routines will be passed with
the needed registration resources (i.e. iser_reg_resources)
without being aware of their nature (i.e. data or protection).
In order to achieve this, we remove reg_indicators flags container
and place specific flags (mr_valid) within iser_reg_resources struct.
We also place the sig_mr_valid and sig_protcted flags in iser_pi_context.
This patch also modifies iser_fast_reg_mr to receive the
reg_resources instead of the fast_reg_descriptor and a data/protection
indicator.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Adir Lev <adirl@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
If iser_initialize_task_headers() routine failed before
dma mapping, we should not attempt to unmap in cleanup_task().
Fixes: 7414dde0a6 (IB/iser: Fix race between iser connection ...)
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
In some rare cases, IO operations may be not aligned to page
boundaries. This prevents iser from performing fast memory
registration. In order to overcome that iser uses a bounce
buffer to carry the transaction. We basically allocate a buffer
in the size of the transaction and perform a copy.
The buffer allocation using kmalloc is too restrictive since it
requires higher order (atomic) allocations for large transactions
(which may result in memory exhaustion fairly fast for some workloads).
We rewrite the bounce buffer code path to allocate scattered pages
and perform a copy between the transaction sg and the bounce sg.
Reported-by: Alex Lyakas <alex@zadarastorage.com>
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
No need to keep lkey, va, len variables, we can keep
them as struct ib_sge. This will help when we change the
memory registration logic.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Adir Lev <adirl@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Instead of open-coding connection fastreg pool get/put,
we introduce iser_reg_desc[get|put] helpers.
We aren't setting these static as this will be a per-device
routine later on. Also, cleanup iser_unreg_rdma_mem_fastreg
a bit.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Adir Lev <adirl@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
This struct members other than struct iser_mem_reg are unused,
so remove it altogether.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Adir Lev <adirl@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
As memory registration/de-registration methods, lets
move them to their natural location. While we're at it,
make iser_reg_page_vec routine static.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
No need to keep two iser_data_buf structures just in case we use
mem copy. We can avoid that just by adding a pointer to the original
sg. So keep only two iser_data_buf per command (data and protection)
and pass the relevant data_buf to bounce buffer routine.
This patch does not change any functionality.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Adir Lev <adirl@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
We always unmap SGs with the same direction instead of unmapping
with the direction the mapping was done, fix that.
Fixes: 9a8b08fad2 ("IB/iser: Generalize iser_unmap_task_data and [...]")
Signed-off-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
And fix a checkpatch warning.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
No reason to settle with four, can use the min between device max comp
vectors and number of cores.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
It is enough to check mem_h pointer assignment, mem_h == NULL will
indicate that buffer is not registered using mr.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Fix a regression was introduced in commit 6df5a128f0 ("IB/iser:
Suppress scsi command send completions").
The sig_count was wrongly set to be static variable, thus it is
possible that we won't reach to (sig_count % ISER_SIGNAL_BATCH) == 0
condition (due to races) and the send queue will be overflowed.
Instead keep sig_count per connection. We don't need it to be atomic
as we are safe under the iscsi session frwd_lock taken by libiscsi on
the queuecommand path.
Fixes: 6df5a128f0 ("IB/iser: Suppress scsi command send completions")
Signed-off-by: Max Gurtovoy <maxg@mellanox.com>
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Re-adjust max CQEs per CQ and max send_wr per QP according
to the resource limits supported by underlying hardware.
Signed-off-by: Minh Tran <minhduc.tran@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Acked-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Update the driver version and add Sagi Grimberg as maintainer
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Singal completion of every 32 scsi commands and suppress all the rest.
We don't do anything upon getting the completion so no need to "just
consume" it. Cleanup of scsi command is done in cleanup_task callback.
Still keep dataout and control send completions as we may need to
cleanup there. This helps reducing the amount of interrupts/completions
in the IO path.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Poll in batch of 16. Since we don't want it on the stack, keep under
iser completion context (iser_comp).
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Avoid post_send counting (atomic) in the IO path just to keep track of
how many completions we need to consume. Use a beacon post to indicate
that all prior posts completed.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
This will solve a possible condition where we might miss TX completion
(flush error) during session teardown. Since we are using a single
CQ, we don't need to actively drain the TX CQ, instead just wait for
flush_completion (when counters reach zero) and remove iser_poll_for_flush_errors().
This patch might introduce a minor performance regression on its own,
but the next patches will enhance performance using a single CQ for RX
and TX.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Introduce iser_comp which centralizes all iser completion related
items and is referenced by iser_device and each ib_conn.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
iscsi daemon is in user-space, thus we can't rely on it to be invoked
at connection teardown (if not running or does not receive CPU time).
This patch addresses the issue by re-structuring iSER connection
teardown logic and CM events handling.
The CM events will dictate the RDMA resources destruction (ib_conn)
and iser_conn is kept around as long as iscsi_conn is left around
allowing iscsi/iser callbacks to continue after RDMA transport was
destroyed.
This patch introduces a separation in logic when handling CM events:
- DISCONNECTED_HANDLER, ADDR_CHANGED
This events indicate the start of teardown process.
Actions:
1. Terminate the connection: rdma_disconnect (send DREQ/DREP)
2. Notify iSCSI of connection failure
3. Change state to TERMINATING
4. Poll for all flush errors to be consumed
- TIMEWAIT_EXIT, DEVICE_REMOVAL
These events indicate the final stage of termination process and
we can free RDMA related resources.
Actions:
1. Call disconnected handler (we are not guaranteed that DISCONNECTED
event was invoked in the past)
2. Cleanup RDMA related resources
3. For DEVICE_REMOVAL return non-zero rc from cma_handler to
implicitly destroy the cm_id (Can't rely on user-space, make sure
we have forward progress)
We replace flush_completion (indicate all flushes were consumed) with
ib_completion (rdma resources were cleaned up).
The iser_conn_release_work will wait for teardown completions:
- conn_stop was completed (tasks were cleaned-up) - stop_completion
- RDMA resources were destroyed - ib_completion
And then will continue to free iser connection representation (iser_conn).
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Ariel Nahum <arieln@mellanox.com>
Signed-off-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Structure that describes the RDMA relates connection objects. Static
member of iser_conn.
This patch does not change any functionality
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Two reasons why we choose to do this:
1. No point today calling struct iser_conn by another name ib_conn
2. In the next patches we will restructure iser control plane representation
- struct iser_conn: connection logical representation
- struct ib_conn: connection RDMA layout representation
This patch does not change any functionality.
Signed-off-by: Ariel Nahum <arieln@mellanox.com>
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
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 <arieln@mellanox.com>
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
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 <arieln@mellanox.com>
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Replace struct sockaddr_in with struct sockaddr which supports both
IPv4 and IPv6, and print using the %pIS format directive.
Signed-off-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
iSER relies on refcounting to manage iser connections establishment
and teardown.
Following commit 39ff05dbbb ("IB/iser: Enhance disconnection logic
for multi-pathing"), iser connection maintain 3 references:
- iscsi_endpoint (at creation stage)
- cma_id (at connection request stage)
- iscsi_conn (at bind stage)
We can avoid taking explicit refcounts by correctly serializing iser
teardown flows (graceful and non-graceful).
Our approach is to trigger a scheduled work to handle ordered teardown
by gracefully waiting for 2 cleanup stages to complete:
1. Cleanup of live pending tasks indicated by iscsi_conn_stop completion
2. Flush errors processing
Each completed stage will notify a waiting worker thread when it is
done to allow teardwon continuation.
Since iSCSI connection establishment may trigger endpoint disconnect
without a successful endpoint connect, we rely on the iscsi <-> iser
binding (.conn_bind) to learn about the teardown policy we should take
wrt cleanup stages.
Since all cleanup worker threads are scheduled (release_wq) in
.ep_disconnect it is safe to assume that when module_exit is called,
all cleanup workers are already scheduled. Thus proper module unload
shall flush all scheduled works before allowing safe exit, to
guarantee no resources got left behind.
Signed-off-by: Ariel Nahum <arieln@mellanox.com>
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Update Mellanox copyrights for 2014 on the iser initiator driver.
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
Sagi Grimberg