Separate bdev_nvme_reset() and _bdev_nvme_reset() by making
bdev_nvme_reset() call _bdev_nvme_reset_start(), and then
remove the ctx parameter from _bdev_nvme_reset().
This clarifies the next patch and reduces the size of the next patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I76b0f2f5b83445845a313203e594dca0be150bc3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6949
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>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
We need to pass bio only when the reset request came from the upper
layer. Previously, bdev_nvme_reset() called _bdev_nvme_reset() and
_bdev_nvme_reset() had the second parameter, ctx, to pass bio to
for_each_channel(). However, it was not clean.
Let's separate bdev_nvme_reset() which processes a reset request
came from the upper layer and _bdev_nvme_reset() which proesses
an internal reset request by adding a helper function
_bdev_nvme_reset_start().
The next patch will remove the second parameter, ctx, from
_bdev_nvme_reset() for clarification and further refactorings.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I2bd00e3abb56fe8d540b8af55743ee92ca9cd526
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6948
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>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
The recent refactoring removed the destruct poller and change the
reset processing to destruct ctrlr after its completion by conditionally
sending message.
But differentiating callback function is difficult if we reset multiple
ctrlrs.
If nvme_bdev_ctrlr->destruct is set, any new reset cannot start. So we can
use an new variable and always execute the callback function.
Add an new variable pending_destruct to struct nvme_bdev_ctrlr, and set
pending_destruct if ctrlr->ref is zero and ctrlr->destruct is true, and
then start destruct ctrlr if ctrlr->destruct_after_reset is set after
clearing pending resets.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I9f34c42a40c5a5da54611e7871aef8c58117a56a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6714
Community-CI: Broadcom CI
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: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Purpose: To setup an accelerated function callback
for created spdk_nvme_poll_group. In this patch,
we just create the interface. The real usage of this
call back will be provided in the other patch.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I0d936aa4eba4dbfcc0137942156b9f2919eb5b78
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6758
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@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>
Previously, not nvme_bdev_ctrlr but spdk_nvme_ctrlr pointer had been
passed to spdk_get_io_channel() in nvme_rpc_io_cmd_bdev_nvme().
There was no unit test for bdev_nvme_get_io_qpair(), and so add
unit test for it to guard us from degradation.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I00fdc14ea6467162c5fb90d04389883d7c409a17
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6920
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Add mutex per nvme_bdev_ctrlr and replace g_bdev_nvme_mutex by it
for possible cases. These are just drop-in-replacement.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ib1d5eebd65e514cfe96ad60a9004d9bace7b4f6b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6693
Community-CI: Broadcom CI
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@mellanox.com>
This will make us easier to maintain the operation to delete
nvme_bdev_ctrlr and its trids. The added unit test cases guard us
from degradation.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I400d4092020e89bacaebc7be045a456b8760ed8d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6688
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
By adding the second parameter, hotplug, factor out the common part
of remove_cb() and bdev_nvme_delete() into a helper function
_bdev_nvme_delete().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I5becbc7a7a94e852510e6a2745afb9c533d21713
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6692
Community-CI: Broadcom CI
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@mellanox.com>
nvme_bdev_ctrlr is valid while populating namespaces of the
nvme_bdev_ctrlr regardless of the result. nvme_bdev_ns holds
nvme_bdev_ctrlr by its pointer, ctrlr. Hence it is not necessary for
nvme_ctrlr_populate_namespaces_done() to use nvme_bdev_ctrlr_get_by_name()
to get nvme_bdev_ctrlr.
To simplify the code, add nvme_bdev_ctrlr parameter to
nvme_ctrlr_populate_namespaces_done() and remove the
nvme_bdev_ctrlr_get_by_name() call from nvme_ctrlr_populate_namespaces_done().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ic2013842fad16ea0de49c18b5b853d147c23a6b2
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6621
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
bdev_nvme_no_pi_readv() had not been used for bdev_nvme_no_pi_readv() yet.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I9bf815b36ef0daa2f1248f618a9c7e2f1ea62346
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6603
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
In nvme_bdev_ctrlr_create, calloc will be called with a zero size
allocation request if the number of namespaces is zero. The behaviour
is implementation defined if the size of the space requested is zero -
calloc will either return a pointer that mustn't be dereferenced, or
NULL. If NULL is returned, the nvme_bdev_ctrlr_create will fail.
Only call calloc if there are a non-zero number of namespaces.
Otherwise, leave the namespaces pointer with a NULL value. All
references to namespaces[] are either known to be safe, or occur
in the context of looping through the namespaces which will be
skipped if the count is zero. The exception to this is in
vbdev_opal_create, where an assert has been added to match
equivalent code in bdev_ocssd_create_bdev.
Tested by running unit tests on a system that returns a null pointer
for a zero size allocation.
Signed-off-by: Nick Connolly <nick.connolly@mayadata.io>
Change-Id: I058b0683fd9b3a20bf90e54db93ca48b9bb4e40e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6551
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: sunshihao <sunshihao@huawei.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
When multipath is supported, nvme_bdev will be got via bdev_subsystem.
To make such change transparent, add a helper function
nvme_bdev_ns_to_bdev() and use it for some cases.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ia5ad0b87178bc739ba0a75789f9f26b860241a04
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5801
Community-CI: Broadcom CI
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: Aleksey Marchuk <alexeymar@mellanox.com>
By the recent changes, the linked list nvme_ns->bdevs has only a
single bdev at the maximum. Hence replace it by the pointer
nvme_ns->bdev, and remove the linked list pointer nvme_bdev->tailq.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ib976e15bb128ba8479070b58e5f4c43fb9dcf479
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6230
Community-CI: Broadcom CI
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@mellanox.com>
By the last changes, not only standard namespace but also ocssd
namespace has only one nvme_bdev, and standard namespace processes
only the head of nvme_ns->bdevs.
This patch changes the common and standard namespace specific
part to process only the head of nvme_ns->bdevs.
The following patch will replace the linked list nvme_ns->bdevs
by the pointer nvme_ns->bdev.
Add a particular error case that nvme_bdev is failed to create even
if ctrlr has one namespace. If ctrlr has one namespace but the
corresponding bdev is failed to create, nvme_ns->populated should
be false and hence nvme_ns->bdevs should not be accessed. However
the code had not assumed such case.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I5495882fad8c8a012305177179a46d4373ba75f5
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5800
Community-CI: Broadcom CI
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@mellanox.com>
Function bdev_nvme_create accpets new parameter - ctrlr
opts which is passed and filled by RPC handler. That
will allow us to add config parameters for other ctrlr
options with minimal changes.
Change-Id: I96ac1b21e7a3816c652765cddade75423eb843ca
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6023
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>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Track whether we attached a controller while
polling the probe ctx. If we didn't when the probe
ctx is done, then it means the controller failed to
attach and we need to free the ctx.
Fixes issue #1723.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Ia7f040a073e4c824c29c0ed493f8391b69f94174
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5818
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>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
When multipath is supported, nvme_bdev_ns will be got via bdev_subsystem.
To make such change invisible, add a helper function
nvme_bdev_to_bdev_ns() and use it in bdev_nvme_io_type_supported()
and bdev_nvme_dump_info_json().
Inline the function and locate it in common.h to use for ocssd_bdev
too.
ctrlr can be got from ns using spdk_nvme_ns_get_ctrlr().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: If85c86737e8194b8e34ed62df04a3968443f23bb
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5795
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Fctor out getting nvme_ns and qpair pointers in _bdev_nvme_submit_request()
into a helper function bdev_nvme_find_io_path().
bdev_nvme_find_io_path() will be used for ocssd_bdev, and hence
locate it in common.h, inline it because it is used in I/O paths.
ocssd_bdev needs not spdk_nvme_ns but nvme_bdev_ns pointer in I/O paths,
and bdev_nvme_find_io_path() returns nvme_bdev_ns.
Besides, move inclusion of likely.h from bdev_nvme.c and bdev_ocssd.c
to common.h.
The next patch will apply bdev_nvme_find_io_path to ocssd_bdev.
By the following patches, bdev_nvme_find_io_path() will take ANA
state into consideration.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I3817c9f56606021ebea90fdfbcf0656df9faba82
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5528
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>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
When we do detect new devices, we would like to get them
attached as quickly as possible. Controller initialization
requires a non-trivial number of admin commands, and when
using async probe, it means that after we have detected
a new device, it will take many iterations of calling
spdk_nvme_probe_poll_async() before the controller is
fully attached.
So when we are actively probing a probe_ctx, create
a new poller that is solely responsible for probing
that context at a much higher frequency.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I72fbe3faef2d72608edb163bd87907902d7c3adc
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5646
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Object structure will be different after introducing subsystem.
Let's inline nvme_bdev_attach_bdev_to_ns() and
nvme_bdev_detach_bdev_from_ns() as a preparation.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I93fd43db231780b9e86ebaec1a9fb40f8511dab9
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5756
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>
Factor out the common operation of nvme_ctrlr_depopulate_namespace_done()
and nvme_bdev_attach_bdev_to_ns() into a helper function nvme_bdev_ns_detach().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I8efa8bc2c2d455d0fd9e0865ff85fd265d14ee06
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5613
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Separate reference count of nvme_bdev_ctrlr between nvme_bdev_ctrlr
and nvme_bdev_ns.
Set ctrlr->ref to 1 when creating ctrlr, increment ctrlr->ref when
populating ns, decrement ctrlr->ref when destructing ctrlr or
when ns->ref becomes 0, and destruct ctrlr actually when ctrlr->ref is 0.
Set ns->ref to 1 when populating ns, increment ns->ref when
adding bdev to ns, decrement ns->ref when depopulating ns or removing
bdev from ns, and decrement ns->ctrlr->ref when ns->ref becomes 0.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I7810384d97a174d8f55d316e5cdf2a9ef4a11432
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5608
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
For further simplification, merge decrementing reference count of
nvme_bdev_ctrlr into nvme_bdev_ctrlr_destruct().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I604039c3f38a60b316ae465d4649e9eb11bfb6cc
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5573
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
destruct_poller had been used to destruct ctrlr after completing reset, but
we can remove destruct_poller and change reset processing to destruct ctrlr
after its completion by itself.
spdk_io_device_unregister() may fail spdk_for_each_channel(). Hence call
nvme_bdev_ctrlr_do_destruct() as the completion function of spdk_for_each_channel().
The first idea was to always run destruct_poller at nvme_bdev_ctrlr_destruct(),
but this patch will be simpler and more intuitive.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I43a400bdb67ab015d707fb9679693bd3d5bfb070
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5607
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: <dongx.yi@intel.com>
The following patches will need to update nvme_bdev_ns at
nvme_ctrlr_depopulate_namespace_done(). nvme_bdev_ctrlr can be
got from nvme_bdev_ns, and so this change will cause no issue.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I7cbfe0d936a3a757a8792fdf842e1228b2df05e6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4531
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Move nvme_bdev_attach_bdev_to_ns() into nvme_bdev_create() and
change the return type of nvme_bdev_create() to int.
These refactorings are necessary to share a single nvme_bdev among
multiple nvme_bdev_ns.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I81c702bfec15f14c6b4a6588d1d3dfa28a9e3e6d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5650
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This will make us a little easier to understand nvme_bdev and clarify
the difference between nvme_bdev and ocssd_bdev.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: If6e93bcc315612ec88c452cf9ae87a3db8ec7855
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5649
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>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This is a preparation for the next patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I20f7b03e15fe8cc74b22ff46040a4eba9cc63231
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5648
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>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
We do not know if reconnecting I/O qpair succeeded or failed now.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I030c3d8553efac7878da0fe358a624a502dd8656
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5699
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Add an new variable is_failed to struct nvme_bdev_ctrlr_trid, and
set it to true when starting failover or when failover failed, or
set it to false when initializing or failover succeeded.
Then add an new failover trid before all failed trids.
The test log showed that many failover failed because new trid was
added after failed trids.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I220839245c6414b5d3ef69a2fa1b97904d88d8bd
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5698
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
We do not know failover is attempted from which path to which path.
Get these information by NOTICELOG.
Failover or reset is serialized by the flag ctrlr->resetting and
ctrlr->failover_in_progress.
Hence it is enough to add such information only to start.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I97400c048b39dc1b70e6aeb71643b5ed1ca23e72
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5695
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
bdev_nvme_submit_request() calls bdev_nvme_get_buf_cb() directly
when the buffer is already allocated. This is for performance improvement
but will be helpful for the upcoming multipath feature to retry
read I/O submission because buffer should be allocated only once.
By reading the code a little more critically, bdev_nvme_submit_request()
can call bdev_nvme_readv() directly in this case, and do in this patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I005c70deabf7fbfe84b7029e9ef103ae122605e1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5558
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
nvme_bdev_ctrlr pointer is already cached to nvme_io_channel.
Hence let's remove it from the helper functions, bdev_nvme_create_qpair()
and bdev_nvme_destroy_qpair().
The following minor fixes are done together in this patch.
Fix the error paths of bdev_nvme_create_qpair() because
bdev_nvme_destroy_qpair() cannot be used as is in this case.
When calling bdev_ocssd_destroy_io_channel(), check if nvme_ch->ocssd_ch
is not NULL instead of spdk_nvme_ctrlr_is_ocssd_supported().
Cache nvme_bdev_ctrlr to nvme_io_channel just before calling
bdev_nvme_create_qpair() because the pointer is more associated
with qpair.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ic96b7d0fa27dd1f32d573eecdc9d4eacc5593bde
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5511
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Factor out each RPC configuration from bdev_nvme_config_json() and
bdev_ocssd_namespace_config_json() into a helper function, respectively.
These changes will make us easier to introduce subsystem into the
NVMe bdev module.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I297f3fb47893d45463e742834c58d2b1213a4b01
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5468
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
When a device is removed, we should use the remove_cb
that was specified when the device was originally probed
and attached, if one was set.
Also add a new spdk_nvme_ctrlr_set_remove_cb API. This
can be used for cases where a different remove_ctx is
desired than was specified for the probe call. This
also enables setting a remove_cb when using connect APIs
which do not have a way currently to provide a remove_cb.
This also requires fixing the bdev nvme module, which
was depending on the previously errant behavior.
Fixes issue #1715.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Id848b39040099ff7a21fe57ea6b194a8c25ae015
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5510
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
It is recommended to read CSTS when there is a timeout.
If CSTS.CFS (Controller Fatal Status) is set, we should
reset the controller.
But if an admin command on a fabrics controller times
out, reading CSTS submits another fabrics command that
could also timeout. Even worse, we are recursively
polling the admin queue for completions in this case.
Fixes issue #1716.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I23d31f6302375c52eba6f4370748d622fbd25ca7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5513
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>
This was missed when this option was added recently in
46a348597 (rpc: add keep alive parameter in bdev nvme).
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I38fa2d4fc7c995bb0cea89b7b8cb51b2ce72aaa0
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5512
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Add keep alive timetout parameter in bdev_nvme_set_options.
NVMe bdev can set this value especially when we test with
bdevperf.
Fix github issue: #1690
Change-Id: I255c935671b74cdb615a8d393e7d7e84524f3c23
Signed-off-by: Jin Yu <jin.yu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5306
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
In other places, nvme_ns->populated has been used to check an namespace
is active for an existing nvme_bdev_ctrlr. We do not need to call
spdk_nvme_ctrlr_is_active_ns() again. Follow it in this patch.
Besides, factor out the namespace comparison into a small helper
function bdev_nvme_compare_ns(). It will be used to check if two
controllers have the common namespace for ANA multipath.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I8aad2f58aae1fd45ca0dd981dded1c406f215ac0
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5359
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Previously, we separated two cases, creating an new nvme_bdev_ctrlr
and adding an new path to an existing nvme_bdev_ctrlr as early as
possible.
However, we can merge creating an new connection between these two
cases. In this patch, connect_attach_cb() calls bdev_nvme_add_trid()
and then calls spdk_nvme_detach() if the nvme_bdev_ctrlr whose name is
ctx->base_name already exists.
This will simplify ANA multipath feature.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I2eec92dbf96e979690a1087c5c4fc306a29f15de
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5329
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Move bdev_nvme_add_trid() up in a file to locate it before
connect_attach_cb(). The next patch will change bdev_nvme_add_trid()
to be called by connect_attach_cb() to merge the connect operation
for failover case.
bdev_nvme_add_trid() will be associated more closely with
connect_attach_cb() and bdev_nvme_create(). To make it clear,
move bdev_nvme_remove_trid() down before bdev_nvme_delete() too.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ia5eb491baea57d715aadc07f206b94884334efaf
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5328
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>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Even if we use TAILQ_FOREACH_SAFE(), it is not safe to lock and unlock
while linked list is parsed.
To fix such erroneous implementation, use SPDK message.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I79a126560ccd847149d4d06bb3d2111e43060ff7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5321
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Factor out the common operation to destruct nvme_bdev_ctrlr further
into an new helper function _nvme_bdev_ctrlr_destruct() from
remove_cb(), bdev_nvme_delete(), and bdev_nvme_library_fini().
The next patch will use SPDK message to fix the race condition
by locking/unlocking during linked list parse.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: If5392bebcf830fac6165aa424b06aec0c9274dc8
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5320
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Currently only nvme bdev module implements this interface. Bdev module
context (in this case spdk_nvme_ctrlr opaque handle) allows for nvme
interface usage for additional management.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I6302c9229d5f7f294a3c1472d9e8dc1519637ffb
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4924
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: Aleksey Marchuk <alexeymar@mellanox.com>
When we support ANA multipath, I/O path will get nvme_bdev_ns from
nvme_bdev_ctrlr->namespaces, and nvme_bdev_ctrlr will be got from
nvme_io_channel.
Hence this patch adds a pointer to nvme_bdev_ctrlr to struct
nvme_io_channel.
As a side effect, nvme_io_channel->ctrlr is helpful to simplify
admin command submit functions, and do it in this patch.
For ocssd_bdev, we will do the similar change later.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I6976d1543ccda37dde4b81b2124d88f0d05ed7eb
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5055
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
When we support ANA multipath, we will use spdk_nvme_ns and
spdk_nvme_qpair pointers for I/O submit functions after checking
ANA state of spdk_nvme_ns.
It will be efficient and easy to understand if we unify such check.
The next patch will provide a helper function to do such check.
This patch changes the parameters of I/O submit functions as a
preparation.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ic741caa1d7e6df844b875beea776e79d4eb8d6fc
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5204
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
bdev_nvme_flush() just completes with success now and uses only
bio parameter now, but matching parameters with other similar I/O
APIs will be good for future enhancement and following patches.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I64be668fa781396ef68b6b506b22cb153d7b6850
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5203
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Using bdev_io->bdev will be more straightforward than using
bdev_io->bdev->ctxt and &nbdev->disk.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ie5854ab063b7b21710e9f899dd93340624a8cb49
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5202
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This and the next change will enable us to pass spdk_nvme_ns and
spdk_nvme_qpair as parameters to I/O passthrough APIs. When we support
ANA multipath, ANA state of spdk_nvme_ns will be checked. These changes
follow that.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Id38a0e14501ff6402f9b151eaf9e44864079412b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5200
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
POSIX defines PRId64/PRIu64/PRIx64 for printing 64-bit values in a
portable way. Replace a few references to %ld to remove the assumption
about the size of a long. Similarly, use %z with size_t arguments.
Where the value being printed is an unsigned 64-bit value, use PRIu64
instead of %ld.
Signed-off-by: Nick Connolly <nick.connolly@mayadata.io>
Change-Id: I0ec4c0cb5e7a8f4b94a1ee6272c8df9ee7f06aa6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5141
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Following a previous patch, cache pointers to spdk_nvme_ctrlr and
spdk_nvme_ns in bdev_nvme_dump_info_json().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I769d112f815e29262d52faee5ec17ccb914a7154
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5104
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
bdev_nvme_failover() updates trid of spdk_nvme_ctrlr using
spdk_nvme_ctrlr_set_trid(). Hence trid of spdk_nvme_ctrlr is always
up-to-date.
Hence bdev_nvme_dump_info_json() can query trid simply by calling
spdk_nvme_ctrlr_get_transport_id(). The queried trid is a const
parameter. Hence add const to nvme_bdev_dump_trid_json().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Id04881f9e7596b910bdf26c4f14f1ebcb3cdbbee
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5103
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
bdev_nvme_io_type_supported() queries NVMe driver to return if the
specified IO type is supported. To clarify it, cache pointers to
struct spdk_nvme_ctrlr and spdk_nvme_ns.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I0956506e720722216e03cb808f61737ada617329
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5102
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
In the following patches, nvme_bdev_subsystem will be added and nvme_bdev_ns
will be added to not nvme_bdev but nvme_bdev_subsystem.
To do it easier, factor out creating an NVMe bdev into an new helper
function nvme_bdev_create().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ifa902912ba1ea979153b04ef4ac9fae10c09b7e2
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5018
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
bdev_nvme_reset() has been called only by a single caller now.
Hence pass nvme_ch to bdev_nvme_reset() to queue I/O easily.
This will improve readability and clarify the logic.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Idc22cb499fcf8455a2b04d44688f999344abc352
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4992
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Factor out the common operation between internal calls and bdev I/O
calls of bdev_nvme_reset() into an new helper function
_bdev_nvme_reset().
This is to improve readability and to clean-up the operation only
for bdev I/O calls in the next patch.
Replace bdev_nvme_reset() by _bdev_nvme_reset() for internal calls.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I7580a8fb40d5a1d694d5f913cde770ca95a0b0c8
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4991
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
_bdev_nvme_reset_ctrlr will figure out what it really does.
_bdev_nvme_reset will be used by an new helper function in the
next patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I890f962529a663ea412bcd416ef9ba31fadfbf61
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4990
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
bdev_nvme_delete() had called nvme_bdev_ctrlr_get_by_name() and
then called remove_cb(). But nvme_bdev_ctrlr_get_by_name() had not
been guarded by lock and remove_cb() parsed the list again by using
spdk_nvme_ctrlr as a key.
These were not safe and efficient.
This patch refines bdev_nvme_delete() to inline remove_cb() into
it and acquire mutex before calling nvme_bdev_ctrlr_get_by_name().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ibf39859809efa6f22899353c53b7b1eef58aa470
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4842
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Add the second parameter remove and include path removal into
bdev_nvme_failover(). Additionally, bdev_nvme_failover() returns
only two return values, 0 or -EAGAIN, and hence change
bdev_nvme_remove_trid() to return the return value of
bdev_nvme_failover() directly.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ife560f886e32ecec23691e01e87069502e27cc0c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4867
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Checking if destruct is false and setting destruct to true are
separated by mutex in remove_cb() and bdev_nvme_library_fini().
It was possible that multiple threads called nvme_bdev_ctrlr_destruct()
because the caller could call nvme_bdev_ctrlr_destruct() before
setting destruct to true after knowing it was false.
This patch ensures that nvme_bdev_ctrlr_destruct() is called only once.
Set ctrlr->destruct to true without releasing mutex after checking
if ctrlr->destruct is false.
If destruct is set to true before calling nvme_ctrlr_depopulate_namespaces(),
nvme_ctrlr_depopulate_namespace_done() may call nvme_bdev_ctrlr_destruct()
because it is likely that reference count is zero and destruct is true.
In this case, remove_cb() or bdev_nvme_library_fini() cannot call
nvme_bdev_destruct() after returning from nvme_ctrlr_depopulate_namespaces().
On the other hand, if a controller has no namespace,
nvme_ctrlr_depopulate_namespaces() does nothing and
nvme_ctrlr_depopulate_namespace_done() is not called. Hence
remove_cb() or bdev_nvme_library_fini() has to call nvme_bdev_destruct()
after returning nvme_ctrlr_depopulate_namespaces().
To unify both cases, initialize reference count to one as a sentinel value
and remove_cb() and bdev_nvme_library_fini() decrement reference count
and then calls nvme_bdev_ctrlr_destruct() if reference count is zero
after returning from nvme_ctrlr_depopulate_namespaces().
Additionally, add assert to check if reference count is not negative
to find bug in future.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I8a617b5aa4d0a9faff832e63c2ed4b353341dd6b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4817
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
subnqn will be used to manage subsystem, and so add it to the output
of bdev_nvme_dump_info_json().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Id28023754fa19d14aa9297f4e6676d235f024e0f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4974
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
To improve readability and make the following patches easier,
factor out the common operation of bdev_nvme_create_cb() and
_bdev_nvme_reset_create_qpair() into bdev_nvme_create_qpair(),
and factor out the main operation of bdev_nvme_destroy_cb()
into bdev_nvme_destroy_qpair(). bdev_nvme_destroy_qpair() is used
in the error path of bdev_nvme_create_qpair().
Updating g_opts.io_queue_requests again is no harm and so is done
in bdev_nvme_create_qpair().
Use the term nvme_ch in bdev_nvme_create_cb() and bdev_nvme_destroy_cb()
for clarification.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ie0a4251023037958d122fbb5d7ee72d42238b3ed
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4981
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>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
This change will make the next patch smaller and its review easier.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I2784fa74dce98676958180bbf80a9a4f938454db
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5101
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>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
This change removes the necessity to get nvme_bdev_ctrlr by calling
nvme_bdev_ctrlr_get_by_name() and simplifies the code.
bdev_nvme_reset() checks if nvme_bdev_ctrlr->destruct is false first.
Hence we will not degrade stability.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I7a3baab666644c3063af404905284c3e16bc62ab
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4843
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Calling spdk_nvme_ns_get_data(ns) on an inactive namespace will crash.
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: Ib8cfa84e785a4a04dc3b1826c700a22eeec51e3c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5013
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>
We have two functions, remove_cb() and bdev_nvme_library_fini() to
depopulate all existing namespaces.
Factoring out the operation into a helper function
nvme_ctrlr_depopulate_namespaces() will improve the readability.
Besides, change the line after return value for two functions to
adjust format.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ibda921bdc94a50c14cca0ece462f76197efc5ccb
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4642
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Swap the resetting case and the non-resetting case. Unlock mutex
before calling spdk_bdev_io_complete(). Use curr_trid instead of
tmp_trid to clarify the meaning of the stored value.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: If3b520a16459aba498ac80ed887b9223ac5ca3d8
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4523
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: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Consolidating error paths will be helpful to create and associate
with the creating controller. Fix a couple of memory leaks in error
paths together.
Then rename create_ctrlr() by nvme_bdev_ctrlr_create() to follow
naming conventions and clarify what type of controller is created.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I3f9073756198c0b0bc65b67ce84a41233297eac0
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4509
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
There is nothing left here, so remove it.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ib947d42bc577dbebb4650b1be885e05a80f8f8cf
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4541
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Anil Veerabhadrappa <anil.veerabhadrappa@broadcom.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Community-CI: Broadcom CI
This patch removes the string from register component.
Removed are all instances in libs or hardcoded in apps.
Starting with this patch literal passed to register,
serves as name for the flag.
All instances of SPDK_LOG_* were replaced with just *
in lowercase.
No actual name change for flags occur in this patch.
Affected are SPDK_LOG_REGISTER_COMPONENT() and
SPDK_*LOG() macros.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I002b232fde57ecf9c6777726b181fc0341f1bb17
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4495
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Mellanox Build Bot
Reviewed-by: Anil Veerabhadrappa <anil.veerabhadrappa@broadcom.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Broadcom CI
This clean up the code a little more. Use the term failover instead
of multipath in the error message.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I6b7fb8fe50b19b5639169c03f5242b9d6f14128b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4390
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
bdev_nvme_add_trid() is used once in the same file.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I738b7e959d13d419e090a019da6e7f1f09aa7109
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4388
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
PCIe failover is not supported yet and it is not possible that the
controller is included in g_skipped_nvme_ctrlrs for failover cases.
We can consolidate the code to create failover path in bdev_nvme_create().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ide456707a0b66a4b26eaf66392b9b53c3e659d9e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4387
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This simpifies the code a little without any degradation.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I6006266e250f2d083aa12424809d04fac6a9f8f0
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4385
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
_bdev_nvme_submit_request() already has a pointer to nvme_io_channel.
Passing nvme_io_channel simply to I/O APIs will be reasonable
simplification.
Two admin APIs, bdev_nvme_abort() and bdev_nvme_admin_passthru()
get the associated thread to the spdk_io_channel and set it to
nvme_bdev_io. These APIs get spdk_io_channal from nvme_io_channel using
the container_of macro. These APIs are not performance sensitive.
Using the contaier_of macro will be better than extracting the operation
from these APIs because we can keep the changes local to these APIs.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I67985075d70551e4fe3bb84dfee15f1e4e882451
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4384
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
When NVMe bdev module supports multipath, nvme_bdev will have multiple
nvme_bdev_ns, and I/O APIs need only nvme_bdev_ns. So change the parameter
from nvme_bdev to nvme_bdev_ns for I/O APIs.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I1e71abbab856112fdd6a333a89ba5f02f4e63ccd
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4382
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This is a preparation to pass nvme_bdev_ns instead of nvme_bdev
to I/O APIs.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I095478bc3ad305ba940fca101b74e74c7e8e74ba
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4381
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Struct nvme_bdev has a pointer to nvme_bdev_ctrlr and a pointer to
nvme_bdev_ns now, but nvme_bdev_ctrlr can be accessed via
nvme_bdev->nvme_ns->ctrlr, and so nvme_bdev->nvme_bdev_ctrlr is
duplicated.
The upcoming multipath feature will need struct nvme_bdev to have
multiple nvme_bdev_ns pointers.
Hence remove nvme_bdev_ctrlr from struct nvme_bdev and get a pointer
to nvme_bdev_ctrlr from nvme_bdev->nvme_ns->ctrlr.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ie5b785cb26e535c7186fe630c1f8b7d598b0ba46
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4288
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The bdev layer nicely handles the case where we call this
function with the buffers already present - it just
immediately calls the get_buf_cb. But this adds extra
overhead in the case where the buffer is already present.
Since nvme has no alignment restrictions, we can just
check the iovs directly and avoid the extra call to
spdk_bdev_io_get_buf when possible.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I66df0fde574a35e995a3432999d75bdbf9b27212
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4317
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
write instead of writev
Call spdk_nvme_ns_cmd_write_with_md if there is only one iov element.
Use spdk_nvme_ns_cmd_writev_with_md only if there is more than one.
This is about a 15% improvement in I/Ops.
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: I1a99b5507d37cde04e81feada65b14554ad01a17
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4319
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
readv API
It's cheaper to call spdk_nvme_ns_cmd_read_with_md than it is to call
spdk_nvme_ns_cmd_readv_with_md, so do a quick check of the iovcnt and
use the best one for the job.
This is about a 15% improvement in I/Ops.
Change-Id: I82e6677d1ac47abf9919f95e651e7a7595c5e9a3
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4316
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This will allow us to further develop the ecosystem for multipath
failover support.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I24a8cf13e60e6cc0d5b6374da33c8a4e5b6c499a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3069
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>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This allows us to avoid creating a separate rpc just for multipath TRIDs.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I4e83167eaf16e50a72efbd513333a4d09c52be61
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2884
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@mellanox.com>
When we fail to process admin completions on a controller
attempt to failover to a previously registered trid
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I547bd010f4b339b2af7f2b33027cddad4b4926bc
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3045
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
We are about to instrument some checks in this function
that will access the nvme_bdev_ctrlr struct directly so
pass that directly since the ctrlr is still accessible from
the nvme_bdev_ctrlr struct.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I4e0d0d6c433149368e84275426362f8283903027
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3340
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: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This is part of a larger series enabling failover at the bdev
layer for NVMe.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: Ic5c128244699c1a47275145ca7e41aa5f1366259
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3044
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@mellanox.com>