Commit Graph

1157 Commits

Author SHA1 Message Date
Jin Yu
d76951c7ba nvme_rdma: fix the recvs_to_post.first to NULL
nvme_rdma_qpair_submit_recvs is not judged in
nvme_rdma_poll_group_process_completions path.
If we do not clean the recvs_to_post.first we
may get the wrong current_num_recvs when the rc
is non-zero and call it again.

Change-Id: If0046e711525dcfcb419132a01fed7a09db13ba0
Signed-off-by: Jin Yu <jin.yu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3163
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-07-06 07:20:33 +00:00
Jin Yu
05805e54a0 nvme:disconnected state then destroying state
Put the destroying state after the disconnected state.
Because nvme_transport_ctrlr_disconnect_qpair will modify the state
of qpair to disconnected, and in the path of rdma, it will postpone
the deletion of qpair until the release of pg by judging the
destroying state. So qpair is not deleted.

Change-Id: Ica606905cddf67d0ffda14bd48cc5f4e424f01ee
Signed-off-by: Jin Yu <jin.yu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3136
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-07-06 07:20:26 +00:00
Jin Yu
19228a0602 nvme_rdma:fix current_num_sends to current_num_recvs
Change-Id: I1a3067165c06db3fe7d7fd1c1ec149e845100b27
Signed-off-by: Jin Yu <jin.yu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3162
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-07-06 07:20:26 +00:00
Simon A. F. Lund
ae6c40a4a5 lib/nvme: fallback to NVM when then given command_set is not supported
When the user requests a non-default command set configuration, via the
probe_cb() supplied to spdk_nvme_probe(), which is not supported, then
the controller initialization will not proceed.

This patch changes that behavior into falling back to the NVM command
set and continue with the controller initialization. It is done by
assigning the NVM command set to opts.command_set such that the user
knows in attach_cb() with which command_set the controller is
configured/enabled.

The fallback is needed since the user does not have access to the
controller capabilities register. The strategy left for the user is thus
to try. However, this is an issue, as the user only has one attempt, as
subsequent calls to spdk_nvme_probe() will not trigger probe_cb() for
the controllers whose initialization did not proceed.

Signed-off-by: Simon A. F. Lund <simon.lund@samsung.com>
Change-Id: Ia414628fcd7d56956649647775462e62d98c0a90
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2931
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2020-07-03 07:31:31 +00:00
Jim Harris
7ff9609f3d nvme: add security send/receive quirk for Intel 900P SSDs
These SSDs set the oacs.security bit but do not actually
support OPAL.  So do not set the controller flag indicating
SECURITY_SEND_RECV support in this case.

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I7fcfeafcc8d9439a1c53c60a1aea1801923a2ce5

Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3156
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2020-07-03 07:31:20 +00:00
Jim Harris
16364fdc14 nvme: copy 8 bytes at a time into virtualized CMB
QEMU won't emulate wide (>8 byte) instructions on an MMIO
region.  We tried to fix this previously but it wasn't sufficient.
So this version adds an explicit copy function that will only
copy 8 bytes at a time from the source cmd pointer into the
SQ cmd point in the virtualized CMB.

Fixes issue #1458.

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I804005a2e8cf3f7365c1608674f82a3f843e59ba

Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3135
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-07-02 07:21:38 +00:00
Alexey Marchuk
e762508854 nvme_rdma: Add check for keyed SGL length
The length of a keyed SGL data block is limited by 3 bytes.
Add a check to fail requests which length exceeds 3 bytes.
In other case we can send an incorrectly formed SGL request with
an invalid or zero length.

Fixes issue #1450

Change-Id: I77cdaff5fbf4be5754a3ac6008b8ccd532ac5905
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3056
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2020-07-02 07:21:31 +00:00
Ziye Yang
cb98b2ab3e nvme/tcp: Report that we have free entries if send_ack is set.
Without this patch, we will face the following warning code
when compiled with (--with-uring --enable-debug) while testing
big I/O size: 256KB, e.g.,

"nvme_qpair.c: 474:nvme_qpair_resubmit_requests: *ERROR*:
Unable to resubmit as many requests as we completed"

The reason is because the nvme_tcp_request structure is not freed yet
if send_ack is not set, so there will be no entries when there
are other requests submit again.

And this patch can mitigate such issue.

Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I4c7616fbd3c82a883b4e9facd257a1a4f66e876d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3123
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-07-01 07:51:17 +00:00
Seth Howell
203ed4f673 lib/nvme: report rdma_connect errors up the stack.
This will allow applications to discern specific connect
behavior and make choices relative to it.

Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I46182c285367ceb8a72511defe4508b3592b4572
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3095
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-06-29 09:19:09 +00:00
Alexey Marchuk
8421f83973 rdma: Fix qpair desctruction in error flow
rdma_qp may not be initialized when qpair is not fully
created. When such a qpair is being destroyed we may pass
a NULL pointer to spdk_rdma_qp_disconnect or spdk_rdma_qp_destroy
and hit an assert. This patch fixes this problem for NVMEoF
target and initiator.

Change-Id: I84787dc1b1211293c2a19f59d47727eaecd9d5a1
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3050
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2020-06-29 09:18:52 +00:00
Michael Haeuptle
89013903fe NVME: Fixes stuck IOs during hot remove (#1451)
When a NVMe device is hot removed, subsequent calls to
nvme_qpair_submit_request can fail with ENXIO.
The failure path handling for ENXIO did not free the request which
exhausts the qpair's free_req list eventually and all IOs are stuck
going forward.

This fix adds the same cleanup handling to nvme_qpair_submit_request
for this error case as it is done in _nvme_qpair_submit_request.

Signed-off-by: Michael Haeuptle <michael.haeuptle@hpe.com>
Change-Id: I5677d53965bdbd6d339c013483cdf42ce782099a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3018
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2020-06-29 09:18:29 +00:00
Ziye Yang
ceb07eb8f4 nvme/tcp: Fix send_cb and recv pdu function contention when there is R2T.
When using uring socket, we see following assert

nvme_tcp.c:1018: nvme_tcp_capsule_resp_hdr_handle:
Assertion `tcp_req->state == NVME_TCP_REQ_ACTIVE' failed.
Detailed info is in

https://ci.spdk.io/results/autotest-per-patch/builds/19205/archive/nvmf-tcp-vg-autotest/build.log

We face this issue, because there is also code execution ordering
between "sending callback function" and "pdu receving function".
We did not find it in physical machine testing, but finding it
in vagrant machine in CI.

Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I5eb241d564c0fc42ce0601b7c85999a2550f0de3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3046
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2020-06-29 09:18:13 +00:00
Ziye Yang
2ac8d3ff5e nvme/tcp: Allocate send_pdu with DMA allocated memory.
Purpose: It will be used to leverage the uring acceleration
later when we use io_uring_prep_write_fixed.

Because for using the Registered buffers feature in I/O uring,
we currently can register all the huge memories. And if we
allocate send_pdus in DMA memory, we can leverage such feature.

Change-Id: Id0ba5f7fe43202027c0378e9cbe74d861aad21e5
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3002
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2020-06-24 08:22:17 +00:00
Shuhei Matsumoto
465b2f8a6b nvme/rdma: Inline nvme_rdma_req_put()
nvme_rdma_req_complete() and nvme_rdma_req_put() are called in a
row except a single case. Move clearing completion_flags and req
of rdma_req from nvme_rdma_req_put() to nvme_rdma_req_complete(),
and then inline nvme_rdma_req_put() because nvme_rdma_req_put()
does only insert now. To do this, change the type of the second
parameter of nvme_rdma_req_complete() from struct nvme_request
to struct spdk_nvme_rdma_req.

For the exceptional case that only nvme_rdma_req_put() is called,
change nvme_rdma_req_init() to clear rdma_req->req if returned with
error.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ibf7e6d245f3a48fb895cd9e6d92596ef833f26d1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2876
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
2020-06-24 08:19:43 +00:00
Shuhei Matsumoto
a57aeac1fe nvme/rdma: Dequeue request from outstanding list before calling completion
Each request has a callback context as cb_arg, and the callback to
nvme_complete_request() for the completed request may reuse the context
to the new request.

On the other hand, RDMA transport dequeues rdma_req from
rqpair->outstanding_reqs after calling nvme_complete_request() for
the request pointed by rdma_req.

Hence while nvme_complete_request() is executed, rqpair->outstanding_reqs
may have two requests which has the same callback context, the
completed request and the new submitted request.

The upcoming patch will search all requests whose cb_arg matches to
abort them. In the above case, the search may find two requests by
mistake.

To avoid such error, move dequeueing rdma_req from rqpair->outstanding_reqs
before calling nvme_request_complete().

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ia183733f4a4cd4f85de17514ef3a884693910a05
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2863
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
2020-06-24 08:19:43 +00:00
Shuhei Matsumoto
299fce881a nvme/pcie: Dequeue request from outstanding list before calling completion
Each request has a callback context as cb_arg, and the callback to
nvme_complete_request() for the completed request may reuse the context
to the new request.

On the other hand, nvme_pcie_qpair_complete_tracker() dequeues tr from
pqpair->outstanding_tr after calling nvme_complete_request() for the
request pointed by tr.

Hence while nvme_complete_request() is executed, pqpair->outstanding_tr
may have two requests which has the same callback context, the
completed request and the new submitted request.

The upcoming patch will search all requests whose cb_arg matches to
abort them. In the above case, the search may find two requests by
mistake.

To avoid such error, change nvme_pcie_qpair_complete_tracker() to
dequeue tr from pqpair->outstanding_tr before calling nvme_complete_request().

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ie9efc200d06d02d8ee2be3cb8e9fd64591bc210d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2861
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
2020-06-24 08:19:43 +00:00
Shuhei Matsumoto
aacac0b433 lib/nvme: Get ctrlr simply from req->qpair->ctrlr at spdk_nvme_ctrlr_cmd_abort()
req->qpair and qpair->ctrlr are set at their initialization, and
hence we don't have to store ctrlr to req->user_buffer but we can
get ctrlr simply from req->qpair->ctrlr.

Besides, save a local variable by overwriting qpair when it is NULL.
This is also a small clean up.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I1cca763fbaae55079b38f7f16fcf779360232797
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2890
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
2020-06-24 08:19:43 +00:00
Shuhei Matsumoto
43f3a8d99f lib/nvme: Factor out submit abort request into a helper function
To reuse the code for the upcoming new function
spdk_nvme_ctrlr_cmd_abort_ext(), factor out the operation to submit
an abort request into an new helper function
_nvme_ctrlr_submit_abort_request().

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Id00bed8fb8dc15b77e3c7f19105c1623588faa53
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2494
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
2020-06-24 08:19:43 +00:00
Shuhei Matsumoto
5322f30787 lib/nvme: Fix aborting queued abort requests at controller reset or destruct
ctrlr->outstanding_aborts is counted only for submitted abort requests.
However ctrlr->outstanding_aborts had been decremented for queued
abort requests by mistake.

Subsequent patches will use parent-children for abort requests but
nvme_free_request() is not aware of such relationship.

Queued abort requests had not been canceled or aborted when controller
was destructed. Retry submitting queued abort requests had been
repeated recursively and had caused stack overflow.

This patch fixes all.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I8ce0ae51ddd5ed3e1e8ac86329c8bdb7a9236b2f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2555
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2020-06-24 08:19:43 +00:00
Shuhei Matsumoto
5cbc1d5cae lib/nvme: Factor out retry queued abort into a helper function
To reuse the code for the upcoming new function
spdk_nvme_ctrlr_cmd_abort_ext(), factor out the operation to retry
queued abort into an new helper function nvme_ctrlr_retry_queued_abort().

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I56bdfefe2c54990c9fd0463303b3d65ccf6e8942
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2493
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2020-06-24 08:19:43 +00:00
Shuhei Matsumoto
9d4751a550 lib/nvme: Fix Abort Command Limit to process as 0's based value
SPDK NVMe driver had processed ACL as 1's based value by mistake,
and SPDK NVMe-oF target sets ACL to 0. Hence If NVMe driver connects
to SPDK NVMe-oF transport, spdk_nvme_ctrlr_cmd_abort() always queued
abort request.

Fix this bug to process ACL as 0's based value in
spdk_nvme_ctrlr_cmd_abort(). Besides, initialize ACL explicitly to
0 in spdk_nvmf_ctrlr_identify_ctrlr() for clarification.

Signed-off-by: Michael Haeuptle <michael.haeuptle@hpe.com>
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Id4f3a469776cdab88bcc6f41e7893885a7b78d8c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2513
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
2020-06-24 08:19:43 +00:00
Shuhei Matsumoto
10c4193363 lib/nvme: Set the parent to failure when submission of any children failed
When one of the children is failed to submit, if any children is
already submitted, the function can return success to wait for those children
to complete, but the parent should be set to failure.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I2ea53856ee58da991bceca0058d1e1f55d42af37
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2492
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
2020-06-24 08:19:43 +00:00
Shuhei Matsumoto
b3d0c369a8 lib/nvme: Fix wrong completion status when retrying queued abort command
next->cpl should have been set if nvme_ctrlr_submit_admin_request()
failed in spdk_nvme_ctrlr_cmd_abort_cpl().

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I4114294e398b18148d8a9ad798a656a7686c5a85
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2491
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
2020-06-24 08:19:43 +00:00
Ziye Yang
3a1f5364d2 nvme/tcp: Fix nvme_tcp_req free conflict between cmd sending and incoming pdu receiving
This patch tries to solve the out of order
call back handling for cmd sending and the incoming pdu handling.

Normally, the cmd call back will be called before
receving the next PDU from the target if the application
uses the sync manner.

With the uring implementation, after sending the
cmd to the target, we may have the following scenerio:
(1) Firstly receive the incoming pdu(e.g., CapsuleResp pdu, C2hdata pdu)
due to the group polling read event.
(2) Secondly execute the callback function related with NVMe command sending.

This means that the data from the initiator is really sent out to the target,
and the target receives, then sending back the data to the initiator. But the
uring io_uring_cqe event is not handled, thus if we execute
(1) first, it will clean the data structures related with nvme_tcp_req, and the
nvme_tcp_req will be used for other purpose. Then causes wrong behaviour like
the following:

"Rand Write test failed at QD=128 because fio hangs with the following error:
nvme_tcp.c: 971:nvme_tcp_capsule_resp_hdr_handle: *ERROR*:
no tcp_req is found with cid=66 for tqpair=0x7f23d8001710".

And this patch can address this issue.

Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I5043aaa8adf5033d93dedac15f633f0850e0b9f2
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2818
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-06-22 07:47:31 +00:00
Alexey Marchuk
268aacb24a rdma: Add new API spdk_rdma_qp_accept
This API is a wrapper for rdma_accept which allows
to remove spdk_rdma_qp_init_attr::initiator_side.

Change-Id: Iba2be5e74e537c498fb11c939c922b2bbda95309
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2908
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2020-06-18 07:28:04 +00:00
Alexey Marchuk
5f6ef934df mk/nvme/nvmf: Add deps for RDMA libs
These dependencies were removed in patches that added
RDMA provider. It was incorrect change since it causes
SEGFAULT when SPDK is built with shared libraries

Change-Id: I15f4ff86a75b3d080e1c7c89d75af4959c4ed989
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2900
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Seth Howell <seth.howell@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2020-06-18 07:28:04 +00:00
Ziye Yang
18450e8b82 nvme: solve the spdk_nvme_connect compatibilty issue.
This is used to make spdk_nvme_connect can support
the old library for compatibility.

Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I49d92fb473c3cbabd8e1240785b920480202eee9
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1998
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2020-06-05 09:04:11 +00:00
Ziye Yang
05cd548cf1 nvme: Fix the field initialization of struct spdk_nvme_ctrlr_opts
Purpose: Make the initilaization in order.

Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I66962073a02b6a4c2fc79ac343cdf5310075dd63
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2766
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2020-06-05 09:04:11 +00:00
Tomasz Kulasek
550d4867a0 lib/nvme: update io msg producers on ns manage
Change-Id: I20adb92ae4e13e775b5e70617c705afd32e16c9e
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/681
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2020-06-05 09:01:44 +00:00
Tomasz Kulasek
11aca20442 nvme/cuse: add spdk_nvme_cuse_update api
spdk_nvme_cuse_update_namespaces() API allows application
to repopulate namespaces on NVMe CUSE devices.

Change-Id: I5375b13bcc91b17cd26b4dc6f02d83dc33826542
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/658
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2020-06-05 09:01:44 +00:00
Seth Howell
1039254319 nvme/rdma: add cq resizing.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I6350d76b8c1e778c18e693b2dfbb10dd36b3e3d8
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1927
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-06-04 07:20:16 +00:00
Seth Howell
67b0dcfe29 nvme_rdma: add tracking for rdma objects in qpair.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I0b45aed21dc649888bb9d93c5937fb553f35eb27
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2568
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-06-04 07:20:16 +00:00
Seth Howell
8bef6f0bdf lib/nvme: rdma poll group with shared cq.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: Ifde29f633f09cccbebfdcde5ab2f96d9590449f1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1167
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-06-04 07:20:16 +00:00
Maciej Szwed
a2606d4b5d all: Fix warnings found by gcc 9.3.0 and 9.3.1
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: Ie7c5b7b1c2d59a1f927eb5602f4ea5a003eff287
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2433
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2020-06-02 09:51:57 +00:00
Seth Howell
6f97efb784 test/make: add leaf type changes check to check_so_deps.
This is an important aspect of ABI versioning.

There was a change to the spdk_bdev_opts struct
which is accessible from the public API that results
in an ABI change.

There was also a change in the spdk_accel_module_if struct that
will affect new modules so its major version needs to be revved.

There was also a change in the publically accessible rmda_hooks
structure which forced an ABI change in the NVMe library.

Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I5cc6886fe01b4adc2836b6e15995471f0361dc29
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2663
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2020-06-01 13:07:30 +00:00
Seth Howell
1a9c19a954 lib/nvme: remove spdk prefix from internal headers.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: Iccde5860b83217163428ff504cba87a1cf209720
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2444
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
2020-06-01 13:07:30 +00:00
Seth Howell
6d18ea425b lib/nvme: force qpair disconnect before aborting rdma requests.
This is needed for shared completion queues which can still give us
successful completions on aborted requests if the qpair hasn't been
disconnected.

Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I85cf1a81ef563d8c02d684b09d2f7ad5008e38cb
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1961
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-06-01 09:22:05 +00:00
Seth Howell
b4e060b560 lib/nvme: check that req is not null in RDMA.
When a request has been aborted, it's possible to get a completion
for an rdma request but the rdma_req->req object has already been
cleared to NULL.

Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I5f7b1b96ff4be8c436aae9a7e2a7c9927d04e627
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1960
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2020-06-01 09:22:05 +00:00
Tomasz Kulasek
c62887ad0b nvme/cuse: add dynamic nvme cuse devices update
This patch adds a posibility to update cuse devices after cuse
device registration.

During NVMe controller life cycle there's a posibility to create
and delete namespaces with Namespace Manage feature from NVMe
spec 1.4.
To guarantee consistency, after changing the number of namespaces
NVMe CUSE devices must be repopulated after Namespace Manage
operation.

This patch introduces cuse_nvme_ctrl_update() entry point to allow
repopulate NVMe CUSE devices.

Change-Id: I478faae040027965321af585abfc34eaf35c908c
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/650
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2020-06-01 09:21:48 +00:00
Tomasz Kulasek
a7567709a6 lib/nvme: implement SPDK_NVME_DATA_HOST_TO_CONTROLLER in cuse
Change-Id: I07bc1771eec284f90d887a0a806d315fbc989428
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1218
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2020-06-01 09:21:48 +00:00
Tomasz Kulasek
8da21f69d3 lib/nvme: rework cuse admin command
For now only controller to host transmition is implemented
for CUSE.

This patch separates cuse_nvme_admin_cmd_send() entry point
to be used to implement another transmition directions.


Change-Id: Ic9013a30c16cf71957c8b411ee00a43c7aa8bbb6
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1674
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
2020-06-01 09:21:48 +00:00
Shuhei Matsumoto
f21f51bd81 lib/nvme: Remove inclusion of SPDK event library
Remove inclusion of spdk/event.h and spdk_internal/event.h from
SPDK NVMe library. Their dependency had been removed before.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ide3a0902b1cebb9c9033ade45d7488622e38696c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2688
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2020-06-01 09:20:41 +00:00
Changpeng Liu
3b6622046e nvme/pcie: use 16 bits width access API for PCI_COMMAND register
PCI_COMMAND register with offset 0x4 is 16 bits width.

Change-Id: I86483e58a3ac680efaae6ff7edaadde8efc289f6
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2589
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2020-05-29 14:59:24 +00:00
Tomasz Kulasek
c84b3cccdd lib/nvme: fix do not use cuse session on fail
Change-Id: I9946829183c43286da4be6bda887a4d13d93e398
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2612
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2020-05-29 14:58:03 +00:00
Jim Harris
89e47f6014 nvme: create netlink socket during nvme_driver_init
This helps ensure thread safety on creation of the
netlink socket, when probe is called from multiple
threads at once.  It is also a lot more clean - we just
create it once, rather than checking every time probe
is called to see if it has to be created.

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I528cedc3ff44de6ea8ecaf6d2389226502ba408e

Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2681
Community-CI: Broadcom CI
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2020-05-29 09:26:05 +00:00
Jim Harris
18f79f2449 nvme: add mutex to nvme_driver_init
This will allow spdk_nvme_probe and variants to be
called from multiple threads in parallel.

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I534db605c9e192b943afe973981b7b503d8b7e34

Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2680
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2020-05-29 09:26:05 +00:00
Seth Howell
684b3a49f0 lib/nvme: split request resubmission into function.
This will need to be done separately for poll groups.

Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I0e432493bdb02e13fe5c73a8a09911cef573307b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1664
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2020-05-28 07:13:44 +00:00
Seth Howell
63732d8880 lib/nvme: split cq completion processing to its own function.
This helps create a separation between processing a qpair and processing
a completion queue which can be shared across multiple qpairs.

Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I111dd16ec4327854f232988a96891a65813f00e1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1166
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2020-05-28 07:13:44 +00:00
zkhatami88
fe3fab26bf nvme/rdma: Using hooks in reg mr
Signed-off-by: zkhatami88 <z.khatami88@gmail.com>
Change-Id: I9493fe82b5b758c0092d20ef18b79d652fefed85
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1905
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2020-05-28 07:11:39 +00:00
GangCao
34430edda7 nvme: add the ctrlr traddr in the error message
In the case, there are several ctrlr used, it's better to
have the traddr to indicate which ctrlr has the issue to
shutdown.

Change-Id: Ie564bb70566ba5822938efc99125d063f7b4ae4a
Signed-off-by: GangCao <gang.cao@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2588
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Xiaodong Liu <xiaodong.liu@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2020-05-27 07:36:58 +00:00