Use the new CQ abstraction to simplify completions in the iSER
initiator.
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Nicer this way.
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

We'll need it later with the new CQ abstraction. also switch
login bufs to void pointers.
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

This also moves recv completion handling from hardirq context into
softirq context.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Remove struct rdma_iu and instead allocate the struct ib_rdma_wr array
early and fill out directly.  This allows us to chain the WRs, and thus
archives both less lock contention on the HCA workqueue as well as much
simpler error handling.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NBart Van Assche <bart.vanassche@sandisk.com>

This adds an abstraction that allows ULPs to simply pass a completion
object and completion callback with each submitted WR and let the RDMA
core handle the nitty gritty details of how to handle completion
interrupts and poll the CQ.

In detail there is a new ib_cqe structure which just contains the
completion callback, and which can be used to get at the containing
object using container_of.  It is pointed to by the WR and WC as an
alternative to the wr_id field, similar to how many ULPs already use
the field to store a pointer using casts.

A driver using the new completion callbacks allocates it's CQs using
the new ib_create_cq API, which in addition to the number of CQEs and
the completion vectors also takes a mode on how we poll for CQEs.
Three modes are available: direct for drivers that never take CQ
interrupts and just poll for them, softirq to poll from softirq context
using the to be renamed blk-iopoll infrastructure which takes care of
rearming and budgeting, or a workqueue for consumer who want to be
called from user context.

Thanks a lot to Sagi Grimberg who helped reviewing the API, wrote
the current version of the workqueue code because my two previous
attempts sucked too much and converted the iSER initiator to the new
API.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

This patch changes the !blk-mq path to the same defaults as the blk-mq
I/O path by always enabling block tagging, and always using host wide
tags.  We've had blk-mq available for a few releases so bugs with
this mode should have been ironed out, and this ensures we get better
coverage of over tagging setup over different configs.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NJens Axboe <axboe@kernel.dk>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Odin.com>

The new fast registration API does not reuqire a page vector
so we can't avoid allocating it.
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Tested-by: NBart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

No callers left, remove it.
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Tested-by: NBart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Instead of constructing a page list, call ib_map_mr_sg
and post a new ib_reg_wr. srp_map_finish_fr now returns
the number of sg elements registered.

Remove srp_finish_mapping since no one is calling it.
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Tested-by: NBart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

This is a preparation patch for the new registration API
conversion. It splits srp_map_sg per registration strategy
(srp_map_sg[fmr|fr|dma]. On its own it adds some code duplication,
but it makes the API switch easier to comprehend.
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Tested-by: NBart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

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: NSagi Grimberg <sagig@mellanox.com>
Acked-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

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: NSagi Grimberg <sagig@mellanox.com>
Acked-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Add support for network namespaces in the ib_cma module. This is
accomplished by:

1. Adding network namespace parameter for rdma_create_id. This parameter is
   used to populate the network namespace field in rdma_id_private.
   rdma_create_id keeps a reference on the network namespace.
2. Using the network namespace from the rdma_id instead of init_net inside
   of ib_cma, when listening on an ID and when looking for an ID for an
   incoming request.
3. Decrementing the reference count for the appropriate network namespace
   when calling rdma_destroy_id.

In order to preserve the current behavior init_net is passed when calling
from other modules.
Signed-off-by: NGuy Shapiro <guysh@mellanox.com>
Signed-off-by: NHaggai Eran <haggaie@mellanox.com>
Signed-off-by: NYotam Kenneth <yotamke@mellanox.com>
Signed-off-by: NShachar Raindel <raindel@mellanox.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

iser is perfectly capable supporting SG clustering as it translates
the SG list to a page vector. Enabling SG clustering can dramatically
reduce the number of SG elements, which doesn't make much of a difference
at this point, but with arbitrary SG list support, reducing the
number of SG elements can benefit greatly as as it would reduce
the length of the HW descriptors array.
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NDoug 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: NSagi Grimberg <sagig@mellanox.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Detected this by compiling with W=1.
Signed-off-by: NBart Van Assche <bart.vanassche@sandisk.com>
Cc: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Detected this by compiling with W=1.
Signed-off-by: NBart Van Assche <bart.vanassche@sandisk.com>
Cc: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Adding an ability to query the IB cache by a netdev and get the
attributes of a GID. These parameters are necessary in order to
successfully resolve the required GID (when the netdevice is known)
and get the Ethernet L2 attributes from a GID.
Signed-off-by: NMatan Barak <matanb@mellanox.com>
Reviewed-By: NDevesh Sharma <devesh.sharma@avagotech.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Just fix a typo in the code comment.
Signed-off-by: NGeliang Tang <geliangtang@163.com>
Acked-by: NSagi Grimberg <sagig@mellanox.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

This also allows to remove the target-specific old configfs macros, and
gets rid of the target_core_fabric_configfs.h header which only had one
function declaration left that could be moved to a better place.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NNicholas Bellinger <nab@linux-iscsi.org>
Acked-by: NNicholas Bellinger <nab@linux-iscsi.org>
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

When we leave the multicast group on expiration of a neighbor we
do not free the mcast structure. This results in a memory leak
that causes ib_dealloc_pd to fail and print a WARN_ON message
and backtrace.

Fixes: bd99b2e0 (IB/ipoib: Expire sendonly multicast joins)
Signed-off-by: NChristoph Lameter <cl@linux.com>
Tested-by: NSagi Grimberg <sagig@mellanox.com>
Signed-off-by: NDoug 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: NChristoph 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: NHaggai Eran <haggaie@mellanox.com>
Tested-by: NSagi Grimberg <sagig@mellanox.com>
Tested-by: NSteve Wise <swise@opengridcomputing.com>

When performing sendonly joins, we queue the packets that trigger
the join until the join completes.  This may take on the order of
hundreds of milliseconds.  It is easy to have many more than three
packets come in during that time.  Expand the maximum queue depth
in order to try and prevent dropped packets during the time it
takes to join the multicast group.
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Since IPoIB should, as much as possible, emulate how multicast
sends work on Ethernet for regular TCP/IP apps, there should be
no requirement to subscribe to a multicast group before your
sends are properly sent.  However, due to the difference in how
multicast is handled on InfiniBand, we must join the appropriate
multicast group before we can send to it.  Previously we tried
not to trigger the auto-create feature of the subnet manager when
doing this because we didn't have tracking of these sendonly
groups and the auto-creation might never get undone.  The previous
patch added timing to these sendonly joins and allows us to
leave them after a reasonable idle expiration time.  So supply
all of the information needed to auto-create group.
Signed-off-by: NDoug Ledford <dledford@redhat.com>

On neighbor expiration, check to see if the neighbor was actually a
sendonly multicast join, and if so, leave the multicast group as we
expire the neighbor.
Signed-off-by: NChristoph Lameter <cl@linux.com>
Signed-off-by: NDoug Ledford <dledford@redhat.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: NSagi Grimberg <sagig@mellanox.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Given that supporting zcopy immediate data for all IOs requires
iser driver to use its own buffer allocations, we settle with
avoiding data copy for IOs with data length of up to 8K (which
is more latency sensitive anyway).

This trims IO write latency by up to 3us and increase IOPs
by up to 40% by saving CPU time doing sg_copy_from_buffer
(8K IO size is the obvious winner here).
Signed-off-by: NJenny Derzhavetz <jennyf@mellanox.com>
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

iser target batches post recv operations to avoid
the overhead of acquiring the recv queue lock and
posting a HW doorbell for each command.

We change it to be per command in order to support
zcopy immediate data for IOs that fits in the 8K
transfer boundary (in the next patch).

(Fix minor patch fuzz due to ib_mr removal - nab)
Signed-off-by: NJenny Derzhavetz <jennyf@mellanox.com>
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

Instead of handing a connection to the iscsi stack
for processing right after accepting (rdma_accept) we only hand
the connection to the iscsi core after we reached to a connected
state (ESTABLISHED CM event). This will prevent two error scenrios:

1. race between rdma connection teardown and iscsi login sequence
   reported by Nic in: (ce9a9fc2 "iser-target: Fix REJECT CM event
   use-after-free OOPs")

2. target stack shutdown sequence race with constant login attempts by
   multiple initiators.

We address this by maintaining two queues at the isert_np level:
- accepted: connections that were accepted but have not reached
  connected state (might get rejected, unreachable or error).
- pending: connections in connected state, but have yet to handed
  to the iscsi core for login processing. iser connections are promoted
  to the pending queue only from the accepted queue.

This way the iscsi core now will only handle functional iser connections
and once we shutdown the target stack, we look for any stales that
got left behind so we can safely release them.
Signed-off-by: NJenny Derzhavetz <jennyf@mellanox.com>
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Cc: <stable@vger.kernel.org> # v3.10+
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

These are always referenced from np-> so no need
for the prefix.
Signed-off-by: NJenny Derzhavetz <jennyf@mellanox.com>
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

Signed-off-by: NJenny Derzhavetz <jennyf@mellanox.com>
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

The iscsi target core teardown sequence calls wait_conn for
all active commands to finish gracefully by:
- move the queue-pair to error state
- drain all the completions
- wait for the core to finish handling all session commands

However, when tearing down a session while there are sequenced
commands that are still waiting for unsolicited data outs, we can
block forever as these are missing an extra reference put.

We basically need the equivalent of iscsit_free_queue_reqs_for_conn()
which is called after wait_conn has returned. Address this by an
explicit walk on conn_cmd_list and put the extra reference.
Signed-off-by: NJenny Derzhavetz <jennyf@mellanox.com>
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Cc: <stable@vger.kernel.org> # v3.10+
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

As documented in iscsit_sequence_cmd:
/*
 * Existing callers for iscsit_sequence_cmd() will silently
 * ignore commands with CMDSN_LOWER_THAN_EXP, so force this
 * return for CMDSN_MAXCMDSN_OVERRUN as well..
 */

We need to silently finish a command when it's in ISTATE_REMOVE.
This fixes an teardown hang we were seeing where a mis-behaved
initiator (triggered by allocation error injections) sent us a
cmdsn which was lower than expected.
Signed-off-by: NJenny Derzhavetz <jennyf@mellanox.com>
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Cc: <stable@vger.kernel.org> # v3.10+
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

We expect send only joins to fail, it just means there are no listeners
for the group. The correct thing to do is silently drop the packet
at source.

Eg avahi will full join 224.0.0.251 which causes a send only IGMP packet
to 224.0.0.22, and then a warning level kmessage like this:

 ib0: sendonly multicast join failed for ff12:401b:ffff:0000:0000:0000:0000:0016, status -22

If there is no IP router listening to IGMP.
Signed-off-by: NJason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Even though we don't expect the group to be created by the SM we
sill need to provide all the parameters to force the SM to validate
they are correct.
Signed-off-by: NJason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

srp_destroy_qp is designed to indicate we are safe to continue with
freeing the channel resources by modifying the qp error state,
posting a dummy wr on the queue-pair and waiting for it to flush.
This also holds for the channel registration pool as we are unmapping
the memory region when handling a scsi response. Destroying the
channel registration pool before we make sure we processed all the
inflight IO might introduce a use-after-free of the registration pool.

This use-after-free is demonstrated in the stack trace below where
srp is trying to unmap a used FMR after the fmr_pool was already destroyed.

general protection fault: 0000 [#1] SMP
RIP: 0010:[<ffffffff8151121b>]  [<ffffffff8151121b>] _raw_spin_lock_irqsave+0x1b/0x50
Call Trace:
 [<ffffffffa055d88a>] ib_fmr_pool_unmap+0x1a/0xb0 [ib_core]
 [<ffffffffa06c00ed>] srp_unmap_data.isra.28+0x17d/0x250 [ib_srp]
 [<ffffffffa06c01eb>] srp_free_req+0x2b/0x60 [ib_srp]
 [<ffffffffa06c0c94>] srp_recv_completion+0x174/0x580 [ib_srp]
 [<ffffffffa04580fe>] mlx4_eq_int+0x4de/0xe50 [mlx4_core]
 [<ffffffffa0458b00>] mlx4_msi_x_interrupt+0x10/0x20 [mlx4_core]
 [<ffffffff810abc45>] handle_irq_event_percpu+0x35/0x1b0
 [<ffffffff810abdf2>] handle_irq_event+0x32/0x50
 [<ffffffff810ae5cf>] handle_edge_irq+0x6f/0x120
 [<ffffffff8100455a>] handle_irq+0x1a/0x30
 [<ffffffff8151b475>] do_IRQ+0x45/0xb0
 [<ffffffff8151162d>] common_interrupt+0x6d/0x6d
 [<ffffffff813e4d2f>] cpuidle_enter_state+0x4f/0xc0
 [<ffffffff813e4e6c>] cpuidle_idle_call+0xcc/0x210
 [<ffffffff8100b9ea>] arch_cpu_idle+0xa/0x30
 [<ffffffff810ab1e1>] cpu_startup_entry+0xe1/0x270
 [<ffffffff81030b3a>] start_secondary+0x21a/0x2c0
Reported-by: NEliott Kespi <eliottk@mellanox.com>
Signed-off-by: NSagi Grimberg <sagig@mellanox.com>
Signed-off-by: NBart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

The majority of callers never check the return value, and even if they
did, they can't do anything about a failure.

All possible failure cases represent a bug in the caller, so just
WARN_ON inside the function instead.

This fixes a few random errors:
 net/rd/iw.c infinite loops while it fails. (racing with EBUSY?)

This also lays the ground work to get rid of error return from the
drivers. Most drivers do not error, the few that do are broken since
it cannot be handled.

Since uverbs can legitimately make use of EBUSY, open code the
check.
Signed-off-by: NJason Gunthorpe <jgunthorpe@obsidianresearch.com>
Reviewed-by: NChuck Lever <chuck.lever@oracle.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

The SRP initiator only needs this if the insecure register_always=N
performance optimization is enabled, or if FRWR/FMR is not supported
in the driver.

Do not create an all physical MR unless it is needed to support
either of those modes. Default register_always to true so the out of
the box configuration does not create an insecure all physical MR.
Signed-off-by: NJason Gunthorpe <jgunthorpe@obsidianresearch.com>
[bvanassche: reworked and rebased this patch]
Signed-off-by: NBart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Instead of always using the global rkey for the indirect data
buffer descriptor, register that descriptor with the HCA if
the kernel module parameter register_always has been set to Y.
Signed-off-by: NBart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>