Some memory alloc in nvme_allocate_request_user_copy, and submit
through nvme_qpair_submit_request, if nvme ctrlr is failed or
qpair state not meet the requirements, submit will return -ENXIO,
and call nvme_free_request(), but it will not free
req->payload.contig_or_cb_arg, those memory only gets freed when the
request is actually completed, through nvme_user_copy_cmd_complete().
Let's fix this by add check when submit failed.
Fixes issue #2832
Change-Id: I54f0fc60dbb53ced9f52da7d89017be13db2eee1
Signed-off-by: Fengnan Chang <changfengnan@bytedance.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15985
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Xiaodong Liu <xiaodong.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
rc to -ENXIO and goto error, make all error handle in one place,
so it's easy to add more check in later patch.
Change-Id: I13edeef75bbf6c52e18d6b94b78c2e560012bfee
Signed-off-by: Fengnan Chang <changfengnan@bytedance.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16004
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Xiaodong Liu <xiaodong.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
The spdk_nvme_ctrlr_opts now supports a transport_tos option
that allows setting of the 'type of service' value in the IPv4 header.
This is needed to support lossless RoCE setups.
Note: Only RDMA is supported at this point.
Change-Id: I21825fc197c60f539a7d2d651a970ea380d8b56d
Signed-off-by: Michael Haeuptle <michael.haeuptle@hpe.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15908
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Add spdk_nvme_cpl_get_status_type_string() to return ASCII
string for the type of an error.
Append a dummy entry to return "RESERVED" for unknown types.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Ibc07132ee067f146ac149884c6344f313bfcbfff
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15835
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
spdk_nvme_cpl_get_status_string() will be used to count and display
NVMe specific errors via JSON-RPC. This patch is a preparation.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Ia96890172d752d2906549e3033c0b26eef9c20bd
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15834
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Found with misspell-fixer.
Signed-off-by: Michal Berger <michal.berger@intel.com>
Change-Id: If062df0189d92e4fb2da3f055fb981909780dc04
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15207
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Use the deprecation API to annotate and log the deprecation of
spdk_nvme_ctrlr_prepare_for_reset() using the tag
"nvme_ctrlr_prepare_for_reset".
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Change-Id: I98fd840aa9acc028a49bb47daf4ab7e88f1eb818
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15756
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
In the following patches, nvme_rdma_poll_group_set_cq() will
touch not only CQ but also SRQ and receive WR objects.
All these resources are of a poller.
Hence for clarification, rename nvme_rdma_poll_group_set_cq()
by nvme_rdma_qpair_set_poller().
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Ic59ba5a45833e39b1b2647c000c8b953f1031d6b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14910
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@nvidia.com>
Factor out reset failed recvs operation into a helper function
nvme_rdma_reset_failed_recvs(). This will make the following
patches simpler.
For send operation, this change is not required yet, but in future
we may support something like shared SQ. Hence, we do this change
for send operation too.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Signed-off-by: Denis Nagorny <denisn@nvidia.com>
Signed-off-by: Evgeniy Kochetov <evgeniik@nvidia.com>
Change-Id: Ib44acebe63e97e5a60ea6fa701b49278c7f44b45
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14171
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@nvidia.com>
In the following patches, poll group will have rsps objects and to share
the code between poll group and qpair, option for creation will be used.
As a preparation, merge nvme_rdma_alloc_rsps() and
nvme_rdma_register_rsps() into nvme_rdma_create_rsps(). For consistency,
merge nvme_rdma_alloc_reqs() and nvme_rdma_register_reqs() into
nvme_rdma_create_reqs().
Update unit tests accordingly.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Signed-off-by: Denis Nagorny <denisn@nvidia.com>
Signed-off-by: Evgeniy Kochetov <evgeniik@nvidia.com>
Change-Id: I92ec9e642043da601b38b890089eaa96c3ad870a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14170
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@nvidia.com>
When SRQ is supported, recv objects will be allocated by poll group
and qpair will associated and use them. In this case, we do not want
qpair to allocate and free recv objects. When connection is established,
it will be decided if SRQ is used or not. Hence, defer recv objects
allocation until connection is established.
Send objects are not affected directly by SRQ, but
nvme_rdma_register_reqs() no longer does any registration and deferring
send objects allocation makes the code more consistent. Hence, defer
send objects allocation until connection is established too.
Even after this patch, we rely on nvme_rdma_ctrlr_delete_io_qpair()
to free resources completely.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Signed-off-by: Denis Nagorny <denisn@nvidia.com>
Signed-off-by: Evgeniy Kochetov <evgeniik@nvidia.com>
Change-Id: Ic151fad01009d92a7fc809a730e6e9dff1a365f3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14169
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@nvidia.com>
Response objects will be in poll group when SRQ is enabled. But we want
to share the code to allocate and register response objects between SRQ
is enabled or disabled. To do it cleanly, move
nvme_rdma_qpair_submit_recvs() from nvme_rdma_register_rsps() to
nvme_rdma_connect_established(). A few clean up of error handling are
done in this patch. Unregistration will be done when qpair is
disconnected.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Signed-off-by: Denis Nagorny <denisn@nvidia.com>
Signed-off-by: Evgeniy Kochetov <evgeniik@nvidia.com>
Change-Id: I38dc5a6cb84a6bf56c01d5fb7f2cf3d3b63918e0
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14168
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@nvidia.com>
This will make the following patches simpler.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Id3d7c025525b35c1c2b96027430789a8d8f2697b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14422
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@nvidia.com>
Inline nvme_rdma_post_recv() into the callers.
We do not have any similar helper function for posting send WR.
This will make the following patches simpler and will be reasonable.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Ia95a4b350942d20bdb65e84f7575c2dcf67c149b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14421
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Extract and inline the conditional nvme_rdma_qpair_submit_sends()
and nvme_rdma_qpair_submit_recvs() calls.
This will cralify the logic and make the following patches simpler.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Ibe217c6f4fb2880af1add8c0429f92b4de107da8
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14420
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@nvidia.com>
When SRQ is supported, rsp array will be in either qpair or poller.
To make this difference transparent, rdma_req caches rdma_rsp and
rdma_rsp caches recv_wr directly instead of caching indecies.
Additionally, do a very small clean up together.
spdk_rdma_get_translation() gets a translation for a single entry
of a rsps array. It is more intuitive to use rsp.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Signed-off-by: Denis Nagorny <denisn@nvidia.com>
Signed-off-by: Evgeniy Kochetov <evgeniik@nvidia.com>
Change-Id: I61c9d6981227dc69d3e306cf51e08ea1318fac4b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/13602
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@nvidia.com>
Factor out processing recv completion and send completion into helper
functions to make the following patches simpler.
Additionally, invert if condition to check if both send and recv are
completed to make the following patches simpler.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Idcd951adc7b42594e33e195e82122f6fe55bc4aa
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14419
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@nvidia.com>
nvme_tcp_read_pdu in a loop
nvme_tcp_read_pdu itself has a loop in it that runs until no more data
is available, so the extra loop does nothing.
Change-Id: I1471018e396c43187d1f06bd18ce8a6846a71c94
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15139
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Also add unit tests that explicitly test this
condition. They fail without the nvme driver changes
in this patch.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Iaa369be341eb4eba394f248990e56dce001d3940
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15579
Reviewed-by: Mariusz Barczak <mariusz.barczak@intel.com>
Reviewed-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
NVMe controllers can be marked as removed even if we cannot receive
uevents (e.g. by the VMD driver), so we should process them regardless
of hotplug_fd.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Iaaf13a136929200e824f7a6dd3b5584998801630
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15547
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tom Nabarro <tom.nabarro@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
We should always build all function that are part of the API, even if
some of the libraries they depend on are missing. In that case, they
can return an error instead.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I72b450b3a1d62e222bd843e45be547d926414775
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15414
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@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
per Intel policy to include file commit date using git cmd
below. The policy does not apply to non-Intel (C) notices.
git log --follow -C90% --format=%ad --date default <file> | tail -1
and then pull just the 4 digit year from the result.
Intel copyrights were not added to files where Intel either had
no contribution ot the contribution lacked substance (ie license
header updates, formatting changes, etc). Contribution date used
"--follow -C95%" to get the most accurate date.
Note that several files in this patch didn't end the license/(c)
block with a blank comment line so these were added as the vast
majority of files do have this last blank line. Simply there for
consistency.
Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: Id5b7ce4f658fe87132f14139ead58d6e285c04d4
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15192
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Mellanox Build Bot
Additionally, print the string representation of the ctrlr state, as it
makes debugging init failures much easier.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I572ef3d6f7d5bbd52039a8872733578c92be4c4a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15305
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
We do not support Soft RoCE anymore. Remove a workaround for Soft RoCE's
bug that we amy receive a completion without error status after qpair is
disconnected/destroyed. Then add a assert to check if rdma_req->req is
not NULL. This will simplify the code and the following patches.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I80c349053adc0f79679eaf8a5d7265d555d3c2b0
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14909
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The following patches will support SRQ and SRQ will be per poller.
We will need SRQ in nvme_rdma_cq_process_completions().
It is not possible to identify poller if poll_group is passed to
nvme_rdma_cq_process_completions().
Based on these thoughts, add poll_group pointer to poller and pass
poller to nvme_rdma_cq_process_completions() instead of poll_group.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Signed-off-by: Denis Nagorny <denisn@nvidia.com>
Signed-off-by: Evgeniy Kochetov <evgeniik@nvidia.com>
Change-Id: I322a7a0cc08bdcc8e87e720ad65dd8f0b6ae9112
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14282
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@nvidia.com>
NVMe-RDMA target has a helper function get_rdma_qpair_from_wc() and
uses it to identify a qpair from a WC.
NVMe-RDMA initiator has a similar function
nvme_rdma_poll_group_get_qpair_by_id().
NVMe-RDMA initiator will support SRQ in the following patches, and
it will want to identify a qpair from a WC.
get_rdma_qpair_from_wc() of NVMe-RDMA target uses wc->qp_num internally
anyway.
However, the upcoming custom transport for RDMA will have to use other
variables of WC.
Hence, it will be convenient to pass WC instead of qp_num if we consider
future enhancements.
Based on these thoughts, for NVMe-RDMA initiator rename
nvme_rdma_poll_group_get_qpair_by_id() by get_rdma_qpair_from_wc().
remove unnecessary declaration, and pass WC instead of qp_num.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Signed-off-by: Denis Nagorny <denisn@nvidia.com>
Signed-off-by: Evgeniy Kochetov <evgeniik@nvidia.com>
Change-Id: I01ead4730207e2c6ac53b83f151bd5f977a11465
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14279
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@nvidia.com>
Poller will have more shared resources when SRQ is supported.
This is a preparation.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Signed-off-by: Denis Nagorny <denisn@nvidia.com>
Signed-off-by: Evgeniy Kochetov <evgeniik@nvidia.com>
Change-Id: Ic3d1cb93dde3f53653a9536a103e5518cebd58e1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14173
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@nvidia.com>
nvme_rdma_ctrlr_disconnect_qpair() does not poll the qpair until it is
actually disconnected if it is in a poll group even if its async mode
is disabled. Hence, spdk_nvme_ctrlr_free_io_qpair() removes the qpair
from a poll group when it is being disconnected.
On the other hand, I/O qpair is destroyed after it is actually
disconnected.
When SRQ is enabled and used, a SRQ is destroyed if the corresponding
poller does not have any I/O qpair after an I/O qpair is removed from
the poller.
In particular, if we use spdk_nvme_ctrlr_free_io_qpair(), a SRQ is
destroyed before the corresponding I/O qpairs are destroyed.
Destroying a SRQ failed because it is still referenced by I/O qpairs.
This bug was found when running the SPDK NVMe perf tool with SRQ.
The reason was we had nvme_rdma_poll_group_process_completions() to call
disconnected_qpair_cb after the qpair is actually disconnected.
However, it is ensured that nvme_rdma_poll_group_process_completions()
calls disconnected_qpair_cb for any disconnected qpair.
Hence, remove a check if qpair->poll_group is not NULL from
nvme_rdma_ctrlr_disconnect_qpair() and update the comment.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I0fde0d827eec3280e1cc5a0fce34d163a6069bc4
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14908
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@nvidia.com>
With RDMA, the admin poller can experience a remote disconnect when
processing completions. The admin qpair will be disconnected to handle
this. The disconnect code path will manually complete queued aborts.
However, the completion callback for the abort will attempt to resubmit
other queued aborts from the queue, which will result in a very large
stack and can eventually cause a segfault.
The fix is to not resubmit queued aborts if the admin qpair is in any
kind of failed state.
Change-Id: I4a6f959232c8a1bd30c87ca50459014e556cbaa0
Signed-off-by: Vasuki Manikarnike <vasuki.manikarnike@hpe.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15114
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
A loop inside 'nvme_tcp_qpair_process_completions' makes
'max_completions' actually behaving like a minimum:
do {
rc = nvme_tcp_read_pdu(tqpair, &reaped);
[...]
} while (reaped < max_completions);
Before this change 'max_completion' constraint, in its true sense,
was actually not respected and a loop inside 'nvme_tcp_read_pdu'
could be executed indefinitely as long as a recv state changed.
To prevent this behavior, max_completion must be passed to
'nvme_tcp_read_pdu' and used as an additional exit condition.
Signed-off-by: Szulik, Maciej <maciej.szulik@intel.com>
Change-Id: I28da962f4a62f08ddb51915b5d0dae9611a82dee
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15136
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
The client vfio_user library doesn't require this flag as
it is totally owned in SPDK, so remove it.
Change-Id: I8f7b1df18017ceac24dbb8a0417871f25f6bee0d
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/13895
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
An example:
There are 3 c2h data PDUs for one read request. Data digest is
enabled, accel_poller is enabled. The first PDU will be offload
to accel_poller. Then the others will use CPU to calc the crc32c.
If the last PDU is calc done and the first PDU is not calc down,
SPDK will direct success the read request, and free some objects.
When accel_poller calc down, it will find the request is freed,
and abort the SPDK.
Disable multi c2hs async process to prevent this situation.
Signed-off-by: MengjinWu <mengjin.wu@intel.com>
Change-Id: I03c9e5b30622bbe84523c0836aa93cfed672896
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14079
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
psh len is not the same with header len.
Add an assert in nvme_tcp.c to prevent this happen again.
Signed-off-by: MengjinWu <mengjin.wu@intel.com>
Change-Id: Ibc250752bedf3da8994f79c51fb01577a222d364
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14521
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This "if" is of no use here.
The state machine has the "NVME_TCP_PDU_RECV_STATE_AWAIT_PDU_CH"
state means the pdu does not receive enough length of header.
Signed-off-by: MengjinWu <mengjin.wu@intel.com>
Change-Id: Id50943f77b570fd337e2bb4e3b45281018d159e4
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14504
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
These functions used to allocate resources
using calloc/spdk_zmalloc depending on the
g_nvme_hooks pointer. Later these functions
were refactored to always use spdk_zmalloc,
so they became simple wrappers of spdk_zmalloc
and spdk_free. There is no sense to use them,
call spdk memory API directly.
Signed-off-by: Aleksey Marchuk <alexeymar@nvidia.com>
Change-Id: I3b514b20e2128beb5d2397881d3de00111a8a3bc
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14429
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Since now cmds and rsps buffers are allocated
from huge pages, there are already registered
MR for this memory. In that way we can avoid
registering 2 additional MRs per qpair, just
perform memory translation to get lkey.
Signed-off-by: Aleksey Marchuk <alexeymar@nvidia.com>
Change-Id: I2cb39a15e5d224698c293ac18af00a909840eaa8
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14428
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
state change function do not need to use swtich to do some work.
Do memset in state machine.
Signed-off-by: MengjinWu <mengjin.wu@intel.com>
Change-Id: Ie66454d8f31860f403171f20858a6b4a24e3c76f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14502
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: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Xiaodong Liu <xiaodong.liu@intel.com>
Adding `psk` field to `spdk_nvme_ctrlr_opts`
Adding `psk` parameter to `bdev_nvme_attach_controller` RPC
Change-Id: Ie6f0d8b04ce472e6153934e985c026acded6cdfc
Signed-off-by: Boris Glimcher <Boris.Glimcher@emc.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14046
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
For non-fabric controllers, the corresponding I/O qpairs are simply
re-enabled at controller reset.
This had a issue when I/O qpairs span multiple threads and poll group
is used.
spdk_nvme_ctrlr_reconnect_poll_async() calls
nvme_transport_ctrlr_connect_qpair() with qpair->async being false.
Then nvme_transport_ctrlr_connect_qpair() calls
spdk_nvme_poll_group_process_completions() until the qpair is connected.
spdk_nvme_poll_group_process_completions() may poll other qpairs.
This may cause I/O to complete on a wrong thread.
For PCIe controller, spdk_nvme_poll_group_process_completions() calls
spdk_nvme_qpair_process_completions() simply for each qpair.
Hence change nvme_transport_ctrlr_connect_qpair() to call
spdk_nvme_qpair_process_completions() if the controller is non-fabrics.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Ieb270c2fb154124021ef6d25577b817d05e5ca9e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14295
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
By the previous patches, a qpair is destroyed after it is actually
disconnected.
But after the qpair is destroyed, it is checked if drained by using
rqpair->current_num_sends and rqpair->current_num_recvs.
However, if the qpair is the last of a poller of a poll group,
CQ is destroyed before checking if the qpair is drained.
If CQ is destroyed, at least rqpair->current_num_recvs is not updated,
and we may get one second timeout.
This should be avoided.
Hence, destroy the qpair after it is disconnected and drained.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Ibd6c83e8a3e7b6e11e9b45cee42669da6d42a621
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14278
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
If the being disconnected qpair is the last of a poller of a poll group,
CQ is destroyed and the poller is released before the qpair is actually
disconnected.
This patch destroy CQ and release the poller after the qpair is actually
disconnected.
One exception is when spdk_nvme_ctrlr_free_io_qpair() is called to a
connected qpair. In this case, the qpair is removed from a poll group
before the qpair is actually disconnected. In this case, destroy CQ and
release the poller when the qpair is removed from the poll group.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Idf266bbb6dbb40f04ae6313db724fabf80865763
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14253
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
We have two cases to call nvme_rdma_poll_group_put_poller().
For consistency, make the two cases the same sequence.
This will make the next patch easier. The next patch will release
poller from poll group when qpair is actually disconnected as
possible as we can.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I4178113d5277240e287e83a57e97cf32fd0f7457
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14252
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: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Hyper-V NVMe SSD controllers require admin queue
size to be even multiples of a page. Add quirk to
adjust the admin queue size if user overrides the
default value to something other than an even
multiple.
As part of this change, set the quirks earlier
when constructing a pcie controller, so that the
quirks value can be used in the generic
nvme_ctrlr_construct() function.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I417cd3cdc7e3ba512ec412f4876b0e0b7432341c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14220
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@nvidia.com>
Better not to cache a value especially for there's an error return.
Signed-off-by: yidong0635 <dongx.yi@intel.com>
Change-Id: I3b243a66f4db9af34bc2ea01bafdac33004be128
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/13650
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
This was correct back when we only supported PCIe, but doesn't
in the newfangled world of fabrics and vfio-user.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I565edd2dab1eff862844585df8c25da508e4816d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14136
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Jacek Kalwas <jacek.kalwas@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
When we specify source address for admin and I/O qpairs,
rdma_resolve_addr() succeeded only for admin qpair and failed for
following all I/O qpairs because rdma_resolve_addr() returned
-EADDRINUSE.
To reuse source address among multiple qpairs, set the REUSEADDR option
for each CM ID before executing rdma_resolve_addr() if source address
is specified.
We may miss something. Even if rdma_set_option() fails, execute
rdma_resolve_addr().
Fixes issue #2604
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: If03f82d4499cf83c0e428a62e91c9d9e6aad28e0
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14229
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
