Commit Graph

296 Commits

Author SHA1 Message Date
John Barnard
8e8084903e nvmf: Move target opts to transport opts (part 1)
- Move most of the target opts from nvmf_tgt to nvmf_transport.
  - Update transport create functions to pass in transport opts.
  - When transport opts are NULL in transport create function, use
    target opts. (for backward compatiblity)
  - Part 1 of 2 patches. Part 2 (to follow after part 1 accepted)
    will allow independent creation of transport with specific opts
    while maintaining backward compatibility with current apps and
    rpc configuration that still use the add listener method to
    create a transport.

Change-Id: I0e27447c4a98e0b6a6c590541404b4e4be879b47
Signed-off-by: John Barnard <john.barnard@broadcom.com>
Reviewed-on: https://review.gerrithub.io/423329
Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com>
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>
2018-08-27 20:43:53 +00:00
Chen Wang
6fa48bbf62 lib: fix typos in the lib directory
Change-Id: Idcb60b79d2902bb316facc6f60e0a81e5cf847ed
Signed-off-by: Chen Wang <chenx.wang@intel.com>
Reviewed-on: https://review.gerrithub.io/423372
Reviewed-by: GangCao <gang.cao@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2018-08-24 17:15:12 +00:00
Ben Walker
008ec0bd91 nvmf: Store thread in controller structure
The admin queue pair may get disconnected before
the controller is entirely destroyed and can't
be relied on to obtain the correct thread.

Change-Id: I5e80ef286693d53a161134610dd8354c458f8390
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/422134
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: qun wan <qun.wan@intel.com>
2018-08-16 03:30:24 +00:00
Ziye Yang
4c4cba9a95 nvmf: simplify the qpair_mask handling.
We should not use mutex, but use the spdk_send_msg policy,
then we can let only one thread to handle that and
eliminates the segement fault issue.

Now in the code, the qpair_mask is handled by the same
thread, e.g., the thread which owns the admin qpair of
the ctrlr.

Change-Id: I609fd4d49f5ecc85bc47bf9c23afbb507900be7c
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/420827
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@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>
2018-08-03 03:38:34 +00:00
Seth Howell
4bee4e03b6 nvmf: free AER resourcess before disconnecting qpair
It is necessary to free the AER without sending a completion to ensure
that the host does not attempt to send an additional AER upon receiving
the first completion.

Change-Id: I2b3f8f286d6396019d8ace97d2376547705b8d9d
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/420661
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2018-07-27 20:50:36 +00:00
Philipp Skadorov
4bfb557d80 nvmf/rdma: recover qp from fatal errors
RDMA QP is attempted to recover after IBV_EVENT_QP_FATAL event
is received from IBV asynchronous event API.

RDMA QP is put into ERROR state and is not processing any inbound
requests. The outstanding requests are only allowed to COMPLETED
and FREE states, no outbound transfers are performed.

IBV_EVENT_QP_LAST_WQE_REACHED or IBV_EVENT_SQ_DRAINED event is
expected to follow IBV_EVENT_QP_FATAL, giving a go to draining of
all outstanding requests and freeing the associated resources.

The requests executed by block layer are gracefully allowed to
complete, but no outbound transfers are made.

Note, outstanding requests can not be reliably completed through
polling the CQ, as WC's with failure status might not have all
the fields valid. The failed WC's are dropped and the outstanding
requests are fetched from the appropriate state's linked list.

QP recovery is triggered when there is no more outstanding requests.
If QP recovery is completed succesfully, the RDMA QP is put back into
ACTIVE state, the QP disconnect is triggered otherwise.

Change-Id: I45ee7feea067f80ccc6402518990014d691fbda3
Signed-off-by: Philipp Skadorov <philipp.skadorov@wdc.com>
Reviewed-on: https://review.gerrithub.io/416879
Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2018-07-18 20:58:50 +00:00
Daniel Verkamp
d4ef57c072 nvme_spec: add VWC flush broadcast bits (TP 4035)
Also report that the NVMe-oF target does not support Flush with the
broadcast NSID.

Change-Id: Iad761a6f81eb099940463a67074a6dccb8871b70
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/416445
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2018-06-26 22:51:37 +00:00
Daniel Verkamp
f519fbbc81 nvmf: fix SCT for fabrics command-specific errors
Some of the uses of SPDK_NVMF_FABRIC_SC_* were neglecting to set SCT to
indicate that these status codes are from the command-specific set.

Change-Id: Ief2a77e06d69e5925f8d8b1ec89ff2612250dcf0
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/416430
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-06-22 16:25:24 +00:00
Ben Walker
7d87273f60 nvmf: Reimplement abort handling
Now that the qpairs are stored on the poll groups,
reimplement abort handling to search each one. This
doesn't add new functionality compared to before - it
still only aborts AER commands.

Change-Id: I73651cf412b10f91a14be68bd9b8c03d8f22626e
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/415318
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2018-06-20 22:07:24 +00:00
Ben Walker
65aa57b007 nvmf: Keep a list of qpairs per poll group
Remove the list of qpairs per controller.

Change-Id: I6aaeed43c64b38b60c5a609123b3542bb7130a90
Reviewed-on: https://review.gerrithub.io/415317
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2018-06-20 22:07:24 +00:00
Ben Walker
84523c0439 nvmf: Remove abort implementation
This will be re-added later in the series, implemented
in a different way.

Change-Id: I1f9df22025eb37f7f58ce1547862f61906ea557c
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/415315
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-06-20 22:07:24 +00:00
Ben Walker
a5279b6403 nvmf: Move spdk_nvmf_ctrlr_disconnect to nvmf.c
This is only code movement. Nothing changed. This is
preparing for a rewrite of the qpair disconnect path.

Change-Id: Ic782c090293a5b2e4baaba9c222479dbb2b64fec
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/415312
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2018-06-15 19:11:29 +00:00
Daniel Verkamp
eb387189c2 nvmf: don't change NN while ctrlrs exist
This was partially fixed in commit 1e481d0438 ("nvmf: Do not allow NN
to change while connections present"), but we did not handle the case
where the user asked to add a NS with a NSID outside the current NN.

This patch reworks the logic (again) to be more straightforward and
hopefully more obviously correct.

Some confusion between max_allowed_nsid and max_nsid is also clarified;
if max_allowed_nsid is set, then max_nsid == max_allowed_nsid at all
times, so we don't need the extra logic when calculating NN in
spdk_nvmf_ctrlr_identify_ctrlr().

Change-Id: If531baf1114e03441ff3e1e1be098071702d9056
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/414894
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2018-06-13 21:52:38 +00:00
Ben Walker
a83f91c29a thread: Replace #include of io_channel.h with thread.h
Change-Id: I6babd4cf990bf19b510db88bdfb0ca81e29d9252
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/414700
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Madhu Pai <mpai@netapp.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2018-06-12 15:24:07 +00:00
Ben Walker
f4f839ecd3 nvmf: Eliminate num_qpairs counter in controller struct
The qpair mask can be used instead.

Change-Id: I9063e656c7c44663dc96adfd5ef52a5a1065d2ba
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/413860
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2018-06-07 16:07:38 +00:00
Ben Walker
f08cea7169 nvmf: Perform QID validation using a bit mask
Instead of scanning a list of all qpairs, use a bit
mask to determine if the requested QID is unique.

This is not for performance reasons, but because
eventually the ctrlr's list of qpairs is going to
need to go away.

Change-Id: Ic25ee60e4f9cd9d596815719760d5be892f29d0c
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/413286
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2018-06-07 16:07:38 +00:00
Ben Walker
4d58c1328a nvmf: Move QID error checks to common path
QIDs need to be validated for both I/O queues and admin
queues, so move the checking to a common path.

Change-Id: Icb2009b3fe4d70f146b741c046812d9df62cc2fb
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/413859
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2018-06-06 17:10:57 +00:00
Ben Walker
b95fc6fc64 nvmf: Return the correct error for out-of-range QID
This one is actually a bit tough to deduce in the specification.
The NVMe-oF spec says that QID errors detected in the RDMA
transport shall return an RDMA-specific error indicating
the problem. However, our code doesn't detect the error in
RDMA-specific code, and it isn't clear if the language is
a "must" or a "should".

The NVMe specification does clearly indicate what error
to return on invalid QID in response to a Create
I/O Queue Pair command. For now, return that while
we game plan whether we need to call into the RDMA
transport to correctly report this error.

Change-Id: I7faf37bad9b9202bc50a906214a51c17e4808fc0
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/413858
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2018-06-06 17:10:57 +00:00
Ben Walker
72800826ec nvmf: Quiesce I/O before closing spdk_nvmf_qpairs
Change-Id: Iff58943f3f9db39b6fce1437a0cfe7380991daa6
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/412078
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2018-06-04 17:20:34 +00:00
Daniel Verkamp
40b6f761b2 nvmf: allow abort of AER requests
These are trivial to abort: since we just hold on to aer_req until an
event is triggered, we just need to set aer_req back to NULL and
complete the request.

Change-Id: I74ffe7a227fcaf816c0c584fe5c82a940475687f
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/412881
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2018-05-31 06:03:50 +00:00
Daniel Verkamp
85a822292d nvmf: refactor abort handling
Combine request lookup and abort into a single operation. Keeping them
separate would result in duplicating a lot of logic for finding the
proper list from which to remove aborted requests.

This is still a no-op for now, but it paves the way for aborting
requests that are still queued in software.

Change-Id: If8f268521f2c9f93b413261d87e9f39e539813aa
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/412880
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2018-05-31 05:50:54 +00:00
Ben Walker
f14189b9a6 nvmf: Add states to spdk_nvmf_qpair
Change-Id: Ie37e0173f9c1b6fc766a154191df7d0bfa71042b
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/412077
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-05-30 17:38:26 +00:00
Ben Walker
4a8b3adb44 nvmf: Simplify qpair disconnect code path
This path works for disconnect events on qpairs at run time.
Disconnects in response to killing the target have
not been worked out yet.

This path does not currently wait for outstanding I/O to
complete.

Change-Id: I8e476c8444b460c18e51601fb950b9132d12f67d
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/412076
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-05-30 17:38:26 +00:00
Ben Walker
1e481d0438 nvmf: Do not allow NN to change while connections present
Per the NVMe specification, NN cannot change while there
are connections present. There was originally a check
for this that was removed in commit 763ab88 to match
the behavior in the Linux kernel. However, after a
discussion with the NVMe specification committee, SPDK
was originally correct.

Change-Id: I42414d1ee0c8c83f3335d8790edbf65d813c5c74
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/412544
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-05-30 15:16:48 +00:00
Daniel Verkamp
a35b1eb6ab nvmf: fix Identify Namespace for inactive NSIDs
The NVMe spec says that Identify Namespace should return a zero filled
data structure for namespaces that aren't active, rather than failing
the command with a status code of Invalid Namespace or Format.

Change-Id: Ia9156477d8701694f8b295a67e1669fd09e2cb62
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/412872
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2018-05-30 15:14:04 +00:00
Ben Walker
99c8c6d8e1 nvme: Allow users to request which I/O command set they'd like to use
As of NVMe 1.3b, there is only one command set. But pipe
this through the driver per-spec anyway.

Change-Id: I4faf8596f5ce638e5e2a500b424e00ceb6e89edc
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/412102
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2018-05-24 22:42:38 +00:00
Daniel Verkamp
35efde744a nvmf: send abort request to the qpair group thread
Abort is not currently implemented, since spdk_nvmf_qpair_get_request()
always returns NULL, but this will allow it to be implemented in a
thread-safe way.

Change-Id: I6dfd1ee50848deed0f4a2667aad5a811d8dd4ca7
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/410723
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2018-05-24 22:39:23 +00:00
Ziye Yang
cbbec6dd16 nvmf: make each io qpair recycled by its own thread
This patch selects part of commit:
adfa9f6dc7

In the previous patch, the failed reason is the logic
that admin qpair actively destroys the ioqpair when
the admin qpair is recevied early. For this part,
we need to investigate it later since for multiple core
usage in our current handling case, we possibly have
admin qpair firstly to enter the destruct phase even if
the initiator sends the I/O qpair deletion first. Thus,
the ioqpair deletion can be recycled twice.

And this patch is selecting the part of the logic:
only to make each ioqpair recycled by its own thread.

Change-Id: Ic9cc3fbdf56fa8a75cb5d12a16ce6c3129220bca
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/409363
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2018-05-03 21:33:15 +00:00
Daniel Verkamp
87ec451b07 Revert "nvmf: disconnect other connections if admin qpair disconnects."
This seems to be causing a new intermittent test failure with a call
stack that looks like:

 # nvme disconnect -n nqn.2016-06.io.spdk:cnode1
 transport.c:175:23: runtime error: member access within misaligned
 address 0x000000004321 for type 'const struct spdk_nvmf_transport_ops',
 which requires 8 byte alignment
 0x000000004321: note: pointer points here
 <memory cannot be printed>
 #0 0x4e014b in spdk_nvmf_transport_qpair_fini
 #1 0x4f14fe in nvmf_qpair_fini
 #2 0x502a95 in _spdk_reactor_msg_passed

This reverts commit adfa9f6dc7.

Change-Id: I4bab9d8b5294e241c960683df14acbf457bf3391
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/409251
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2018-04-27 16:38:46 +00:00
Ziye Yang
adfa9f6dc7 nvmf: disconnect other connections if admin qpair disconnects.
Also this patch make each io qpair recycled by its own thread.

Change-Id: I386e411f5d4cee4629054e4cce8d1f4401ec49c8
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/407743
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
2018-04-26 17:04:09 +00:00
Ziye Yang
27d47b9a10 nvmf: make the qpair disconnnect in the right order.
Reason: Initiator sends the qpair deletion in wrong order.
The correct order should be all io qpairs deletion and then
the admin qpair deletion. However, nvme perf does not follow
this since we did not catch ctrlr + c. If we catch ctrlr + c,
we need to use the spdk app framework, or other methods. We should
support this in another patch.

Morever, to prevent the incorrect behavior causing the coredump
of target, we need also consider such case in our NVMe-oF target
design. And this patch, can solve this issue, it will defer the
admin qpair deletion when there are still active io qpairs.

Change-Id: Iec9b88c1d6254f36963c92402ebfe8bd99abaea5
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.gerrithub.io/407771
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2018-04-17 01:58:36 -04:00
Ziye Yang
b332897a04 nvmf: add the assert in spdk_nvmf_ctrlr_disconnect
Change-Id: I41c9946e6c30c3eda1abf3669270a9b9533e3a50
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.gerrithub.io/407612
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2018-04-16 19:23:18 -04:00
Daniel Verkamp
61d379fd7e nvmf: move Command Effects log page to ctrlr.c
No code change - the existing code is moved as-is.

This is the only place it is used, and it is conceptually part of the
virtual controller implementation.

Change-Id: I863f5931225c4d78e515341445e91b29a73da64d
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/406251
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-04-04 15:10:30 -04:00
Daniel Verkamp
1ec2c81bbc nvmf: report support for multiple ports and controllers
The NVMe-oF subsystems that our target exposes may have multiple
subsystem ports and multiple controllers; report these in the Identify
Controller data structure Controller Multi-Path I/O Capabilities field.

Change-Id: I5f1d4ce91698fe6abd281fdc022b6e3b37efd3ef
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/406244
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-04-04 14:55:48 -04:00
Daniel Verkamp
3165d0c070 nvmf: report support for NS Attribute notification
Fixes: 26541489ef ("nvmf: implement Get Log Page - Changed NS List")
Fixes: 763ab88884 ("nvmf: add Namespace attribute notice support")
Change-Id: Id015e4d96fe98a7f6c845821395463e341b7066c
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/406247
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-04-04 14:55:48 -04:00
Daniel Verkamp
26541489ef nvmf: implement Get Log Page - Changed NS List
Change-Id: I6219ace8deba0bfc3e3c7f6a4e00c0098584d5ee
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/405122
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
2018-04-03 17:29:56 -04:00
Changpeng Liu
763ab88884 nvmf: add Namespace attribute notice support
Users can use RPC to add/remove a namespace to/from
existing NVMe controller, SPDK NVMeoF target will
generate an asynchronous event as an indication to
host when asynchronous event request is available.

While here, we also set the event with invalid log
identifier, so that the host doesn't need to clear
the event. Users can use Set Feature to disable
such event.

Change-Id: I93c4d752f552d3c86c53e80877aa61c093e167cc
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/398759
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-04-03 17:29:56 -04:00
Ziye Yang
6d4e6ffd08 nvmf: add the assert in ctrlr_delete_qpair
Change-Id: I4c23f83ebaf80f1b62fb4c1144ff864bdfd6efbe
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/406014
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-04-02 19:57:05 -04:00
Daniel Verkamp
d2e7daa435 nvmf: add stubs for all mandatory Get/Set Features
Most of these don't actually do anything useful yet, but they are all
required by the NVMe 1.3 specification.

The features are also rearranged in numeric order to make it easier to
match up the cases in the switch statements to the specification.

Change-Id: I5f70ecb0bb38b9e1d8287d32fc2e03bca5f1f8a6
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/403903
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-03-30 20:18:54 -04:00
Daniel Verkamp
c5008b37a0 nvmf: report support for Command Effects log page
Change-Id: Ia382110fbbadea51d3b454e07e775afc9e286eae
Fixes: c5cd53cce0 ("Support for Commands Supported and Effects Log Page")
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/405821
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-03-30 16:26:57 -04:00
Daniel Verkamp
ed6e83f958 nvmf: implement Firmware Slot Information log page
Change-Id: I93a4cae46dad7ec42fd82fa4917fd27e3b14c46a
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/405827
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-03-30 16:26:10 -04:00
Daniel Verkamp
9a43cbd490 nvmf: mask off reserved bits in AER configuration
Set Features - Asynchronous Event Configuration has reserved bits in the
CDW11 value, which we shouldn't allow the host to set.  Explicitly set
them to 0 in the Set Features handler to avoid propagating them to the
Get Features return value.

Change-Id: I73ed87d6d310e585c073b1db3b875147b4727007
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/403902
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-03-23 03:45:23 -04:00
Daniel Verkamp
4fce1a5fa6 nvmf: factor out generic Get Features handler
Most Get Features commands can be handled by just setting CDW0 to a
particular value and returning success.

Get Features - Host Identifier needs special handling, so it isn't
converted.

Change-Id: I8a3e2d5659a9ecbb3904589912ccf44d0d25b715
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/403901
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-03-23 03:45:23 -04:00
Daniel Verkamp
0f56183728 nvmf: use Set/Get Features structure types
No functionality change, just a cleanup.  This will help with future
simplification patches.

Change-Id: Id2efd851ed0add664e516725ab98b5e1f8cafff5
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/403900
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-03-23 03:45:23 -04:00
Daniel Verkamp
a05f88ff34 nvmf: retrieve namespace UUID from bdev
A new optional parameter is added to the NVMe-oF target namespace
options to set the Namespace UUID, which will be reported via the
Identify command Namespace Identification Descriptor list.

Change-Id: I53c6d6fa6475d3395d412d196b42000715e9ed81
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/400894
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2018-03-20 17:18:44 -04:00
Daniel Verkamp
b86ae8534a nvmf: increase emulated NVMe version to 1.3
Now that the NVMe-oF target emulated controller supports Identify with
CNS = 03h (Namespace Identification Descriptor list), I believe we
support all of the mandatory parts of NVMe 1.3, so let's increase the
reported VS/VER fields to 1.3.

Change-Id: Iee1052bf9ee1a0954f47770c7e5151886e63f11c
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/403891
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-03-16 14:15:47 -04:00
Changpeng Liu
5d841040ab nvmf: remove the duplicated line for aerl setting
Change-Id: I22c11cdd5ab0ffdf72b27223c0c9cd7282537613
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/402667
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-03-08 11:21:42 -05:00
param
c5cd53cce0 Support for Commands Supported and Effects Log Page
Change-Id: If28662585887b613e3510a8fd8414caeca9ba99b
Signed-off-by: param <kumaraparamesh92@gmail.com>
Reviewed-on: https://review.gerrithub.io/401314
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-03-02 13:45:35 -05:00
Daniel Verkamp
5c2952ab9d nvmf: implement NS Identification Descriptor list
Currently, this just reports the EUI64 and NGUID types, which are
already available in the Identify Namespace data, but this sets up the
framework for reporting a Namespace UUID (which doesn't have a
corresponding field in Identify Namespace) in the future.

Change-Id: I758e6d402ff874d65eca3f1db98d92c1a3f7a11f
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/400893
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2018-02-27 13:01:28 -05:00
Ziye Yang
7346be69e7 nvmf: Make the ctrlr create/remove in subsystem in an asynchronous way
Ctrlrs list maintanined by the subsystem structure should be operated
by the thread which creates the subsystem. And this will make the
operations correct.

Change-Id: I7f881a77b1846658b3acd4270b74f86816e87803
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/401541
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-02-27 12:35:08 -05:00
Ziye Yang
6d4c78eaf4 lib/nvmf: reorder the order of the functions
Purpose: To make the nvmf subsystem management correct
in next patch.

Change-Id: Ie5433f3a7d47f05c5949dfc723aba73465623f88
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/401540
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-02-27 12:35:08 -05:00
Daniel Verkamp
1023ca7b46 nvmf: allow configuration of ns NGUID and EUI64
Add optional parameters to namespace creation to let the user pick the
namespace globally unique identifier and EUI-64.

Change-Id: Ia3eebaf22f8a64733a00a83f90cafb4977c2d07a
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/399531
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
2018-02-20 12:01:47 -05:00
Daniel Verkamp
250d342bc1 nvmf: pass an options struct for ns creation
This will allow more parameters to be added to
spdk_nvmf_subsystem_add_ns() without breaking API/ABI compatibility
later.

Change-Id: I6b2f58f1a2d5fcd4c754830cbd4713dc461a31fc
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/399519
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2018-02-16 13:06:11 -05:00
Ziye Yang
30ef8cac7d nvmf: Solve the coredump issue when client conduct nvme disconnect
This patch is used to solve
https://github.com/spdk/spdk/issues/235

With multiple core bining for NVMe-oF target,
the qpairs which belonging to the same ctrlr may
be scheduled to different cores. Thus there is
resource contention to access the struct spdk_nvmf_ctrlr.

And we put the thread info in polling group. Morever,
we introduce an admin_qpair in ctrlr. Since admin_qpair will
always be created at first and freed at last, to reference
this pointer is safe.

Change-Id: I12ac26f9e65b4ed8e48687750046455af0e3be1d
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/398904
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-02-15 13:19:08 -05:00
Ziye Yang
114a91fdea nvmf: extract some codes in functions.
This patch is used to solve
https://github.com/spdk/spdk/issues/235
in the later patch.

Change-Id: I8ef6d3effbff7a6b27e82c397969687d0000542b
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/399349
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2018-02-14 18:09:45 -05:00
Daniel Verkamp
d13d21e651 nvmf: remove spdk_nvmf_qpair::type field
The type of a queue is purely a function of its queue ID.

Add a helper function, spdk_nvmf_qpair_is_admin_queue(), to make the
logic more obvious (akin to nvme_qpair_is_admin_queue() in the NVMe
library).

Change-Id: I7cf5a82d1e3dc50834cc2ff03f5f88c5719c6952
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/398663
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2018-02-07 12:21:54 -05:00
Ben Walker
7358fb6f85 nvmf: Hot plug notifications now use the subsystem pause functionality
Use the general purpose mechanism to pause a subsystem when a bdev
is hot removed.

Change-Id: I3cd4fb15140d6682707b7840c6a9c3f63fe48e55
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/394259
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2018-01-17 17:28:13 -05:00
Ben Walker
823b565b5f nvmf: No longer route fabrics/admin commands to a single thread
These commands can now pause a subsystem if they need to operate
on it. We don't currently implement any of the NVMe commands
that would need to pause, so this patch is simpler than most
would expect.

Change-Id: I25bfdf8e7577cda2bb0ce248d2889447032b9b4c
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/394121
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2018-01-17 17:28:13 -05:00
Cunyin Chang
bdcb0d709a nvmf: add support of hotplug for nvmf.
Change-Id: Iebd5b75e3525e77bf256f5b7f52aa2504d7a68c3
Signed-off-by: Cunyin Chang <cunyin.chang@intel.com>
Reviewed-on: https://review.gerrithub.io/390549
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
2017-12-15 16:14:02 -05:00
Ben Walker
8b79ef3372 nvmf: Remove poll group from controller
Now rely entirely on the user to create and poll
the poll groups.

Change-Id: I66baaa2d0f493390a055a32e6c902f5e2f574534
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/385954
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-12-12 15:01:17 -05:00
Daniel Verkamp
bedc405f35 nvmf: add duplicate QID check in I/O Connect
Disallow multiple I/O queues for the same controller with the same queue
ID.  NVMe-oF 1.0 ECN 001 specifies that this condition should return a
status value of Command Sequence Error.

Change-Id: I41126ddec388a985c403025e099ab15da5d3987c
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/390662
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2017-12-08 14:08:19 -05:00
Daniel Verkamp
ea1c15791f log: rename SPDK_TRACE_* to SPDK_LOG_*
Disambiguate the log components from the trace functionality
(include/spdk/trace.h).

The internal spdk_trace_flag structure and related functions will be
renamed in a later commit - this is just a find and replace on
SPDK_TRACE_* and SPDK_LOG_REGISTER_TRACE_FLAG().

Change-Id: I617bd5a9fbe35ffb44ae6020b292658c094a0ad6
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/376421
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2017-12-07 12:23:19 -05:00
Ben Walker
d0733ff3b7 nvmf: qpairs now point to their poll group instead of thread
There is a 1:1 correspondence between the two, but it is much
more useful to have a pointer to the poll group.

Change-Id: I8494b0573a9b2cce5819545548d0c65a18bdbe86
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/389640
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
2017-12-04 16:02:16 -05:00
Ben Walker
eaaddf3d48 nvmf: Poll groups now automatically register pollers
This simplifies the public API and requirements for
user applications.

Change-Id: Ibb0d25a7838a0fa683f39e79cb4fef78adf6aee8
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/388040
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
2017-11-28 15:29:35 -05:00
Daniel Verkamp
5323a02613 nvmf: move fabrics command processing to ctrlr.c
This is the final set of command processing functions (admin and I/O
commands were already handled in ctrlr.c), and it allows us to clean up
the ctrlr.c API some more.

Change-Id: Ic73c56c44d7fd64bfaa72b6f5f6f0984b4dfa053
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/378018
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
2017-09-19 17:14:52 -04:00
Daniel Verkamp
4ff0eba871 nvmf: merge connect handling into single function
Combine nvmf_process_connect() into spdk_nvmf_ctrlr_connect() to
simplify the logic and keep it all together in one function.

This also allows us to move the invalid connect helper function and
macros back out of the header into a static function in one file.

Change-Id: Ia3bd80d0309392520d51bf8f5830d3f23332c5e3
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/378016
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
2017-09-19 17:14:52 -04:00
Daniel Verkamp
f4a4ddd8a1 nvmf: add subsystem ctrlr management functions
This moves the subsystem->ctrlrs list management fully into the
subsystem code, which will help simplify thread safety considerations
once we start adding locks.

Change-Id: Ibc118923f1bd520f1e524cde5d45ccfcc69aee1e
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/376025
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
2017-09-19 17:14:52 -04:00
Ben Walker
21ea290192 nvmf: cntlid is now only unique within a subsystem
Previously we made cntlid globally unique as part of
a strategy for scaling connections that never panned
out. Now, we have a new strategy and don't need cntlid
to be globally unique, so relax the restrictions
and simplify the code.

Change-Id: I167772f5e7d37183715bf9967b0102529144bb2b
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/376250
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2017-09-19 17:14:52 -04:00
Daniel Verkamp
0caab4e1da nvmf: add Write Zeroes support
Change-Id: I743f5e4d1c24ad5ef9f1fef4c2678e347b179a9f
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/377260
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-09-14 17:41:52 -04:00
Ben Walker
496ba0d273 nvmf: Temporarily remove hot plug support
The current hot plug support assumes only one thread
can access a subsystem at a time, but now that we're
changing that this will need to be reworked. In the
interest of making it easier to change the threading
model, remove hot plug support temporarily. We'll add
it back in once the threading model changes are done.

Change-Id: I15b75b402b85aa62f5ba864a64cde1de3cdb4ba3
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/376417
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-09-13 12:54:49 -04:00
Daniel Verkamp
a7b62cc4cb nbd, nvmf: fix clang 4.0 packed member warnings
clang 4.0 introduced a new warning, -Waddress-of-packed-member, which
triggers on a couple of spots in NBD and NVMf. Fix them up to silence
the warnings.

Change-Id: I134618f93528ea9a3d08050c34056670a58abdab
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/377441
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-09-07 12:15:45 -04:00
Daniel Verkamp
1e714cfe22 nvmf: factor out invalid connect response helpers
These were duplicated in ctrlr.c and request.c; pull them into
nvmf_internal.h so there's only one copy.

Change-Id: I00d499dd17689e907c182d01e61bde075d217af8
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/376020
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ziye Yang <optimistyzy@gmail.com>
2017-09-06 13:03:20 -04:00
Ben Walker
7c6ca97834 nvmf: Move all struct definitions to nvmf_internal.h
This is not a public API, so simplify the number
of internal header files where important types
are defined.

Change-Id: I115d0497d37e3cfe399c3a5b2546d20aa4fe24b4
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/376249
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-09-01 12:43:26 -04:00
Ben Walker
c1535ca0af nvmf: Poll groups can now span transports
We are still creating one poll group per controller,
so this isn't particularly useful just yet.

Change-Id: I65c54385bdba587d4b1098629727877970d39277
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/376241
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-09-01 12:43:26 -04:00
Ben Walker
d5ce9cff63 nvmf: Transport polling now done by poll group
Instead of polling each individual qpair, polling is now
done by poll group. This allows transports to use more
efficient polling schemes in the future.

The RDMA transport as of this patch still just loops
over each qpair in the group and polls it individually,
so this patch results in no performance change yet.

Change-Id: I0f63f0dbbc5fd43c1e0d9729b10b37c2cb0d9881
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/376239
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-30 15:07:22 -04:00
Ben Walker
0ab300f872 nvmf: Send all fabrics and admin commands to master core
This prepares us to fan out I/O qpairs to other cores

Change-Id: I3e9a60226cddf3ccd26b7ad121775ee36f07e6a9
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/375480
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-30 15:07:22 -04:00
Ben Walker
ff4f68a55f nvmf: Rename spdk_nvmf_ctrlr_gent_cntlid to spdk_nvmf_tgt_gen_cntlid
Also, move it to the appropriate compilation unit. Further,
remove use of g_nvmf_tgt.

Shift the function to a new compilation unit as well.

Change-Id: I1a43ff366532b450f00aed54a290fb9eed9bf453
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/375455
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-29 14:59:33 -04:00
Ben Walker
81c3400a1e nvmf: Remove use of g_nvmf_tgt in spdk_nvmf_ctrlr_identify_ctrlr
The target is accessible through the subsystem.

Change-Id: I69616df48185dbfcbe7c90ca6948457a412c6666
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/375454
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2017-08-29 14:59:33 -04:00
Daniel Verkamp
d92f0f75ca log: rename SPDK_TRACELOG to SPDK_DEBUGLOG
This matches the name to the behavior and prepares for addition of a new
log macro for "info" log level.

Change-Id: I94ccd49face4309d3368e399528776ab140748c4
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/375833
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-29 13:25:58 -04:00
Daniel Verkamp
afb3f73869 nvmf: always zero admin ctrlr-to-host data buffers
For admin commands with controller to host transfers, make sure no
uninitialized data is returned to the host.  This allows us to remove a
few other memset()s in per-command handlers.

Change-Id: I00ef42945a118b6e7a0b68ab21c59fdcd21d0ee2
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/376024
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-29 13:14:51 -04:00
Daniel Verkamp
12ab1e26df nvmf: report SPDK version as firmware version
Use the version #defines from spdk/version.h to build a firmware version
string for the virtual NVMe-oF controller model.

Change-Id: I18285cfa75c8915a7db46d4d79c0447392bd5d8c
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/375252
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-23 13:55:42 -04:00
Ben Walker
f1015fbda6 nvmf: ctrlr_discovery.c no longer uses g_nvmf_tgt
Change-Id: I19d1866fb90ffd3e3543ccc66a098571ade3e3de
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/375203
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-23 11:06:47 -04:00
Ben Walker
70bc390c82 nvmf: spdk_nvmf_find_subsystem now takes a tgt parameter
The user can now specify which target they want to
search for the subsystem. Also, change the name to
spdk_nvmf_tgt_find_subsystem and put it in the correct
compilation unit.

Change-Id: I7c085959814c14d8400a0ba2572103b0814a4d0e
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/374879
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-23 11:06:47 -04:00
Ben Walker
02b640bfdf nvmf: Add a pointer to the target to the subsystem
This eliminates a couple of references to g_nvmf_tgt
and sets the stage for eliminating that global entirely
in the future.

Change-Id: I068d0874cc8ba122be780e8dbd55bb1efabe10b7
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/374876
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-22 11:29:03 -04:00
Ben Walker
d868cd0895 nvmf: Change spdk_nvmf_tgt_init to take an options struct
The list of options is going to continue to grow, so avoid
constant API churn by adding an options struct.

Change-Id: Ie9e7248281726d4aee42b3519fcf7535ba01ee34
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/374872
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-22 11:29:03 -04:00
Daniel Verkamp
14451d76fe nvmf: encapsulate namespace data into a struct
Add a new struct spdk_nvmf_ns to hold namespace information, and add
accessor functions so that it can be moved out of the public API in
future patches.

Change-Id: Id926aaa3cc279649057afc65e08024628edbc435
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/374740
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-18 17:50:58 -04:00
Daniel Verkamp
81062f6049 nvmf: disallow I/O connect on Discovery controller
Discovery controllers only have an admin queue.

Change-Id: I5d5a018a7df40164838a4b03f7f106a96db7e33b
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/374736
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-18 17:50:58 -04:00
Daniel Verkamp
5b4b66bab9 nvmf: move admin processing to ctrlr.c
Now that the discovery controller is using the common admin command
functions, move all of them into the common ctrlr.c file.

This also eliminates the subsystem ops, which are now just direct calls.

Change-Id: I0a25a61e0ad8742d3d76a3cacd46db4701fc7d63
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/374733
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-18 17:50:58 -04:00
Daniel Verkamp
bf6caa753d nvmf: refactor identify into a common function
Also remove the stored controller data from spdk_nvmf_ctrlr and instead
generate up-to-date information every time the command is called.

Change-Id: Icdcc7c826eff25711c6700d12089a86a8dedee54
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/374559
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-17 17:58:49 -04:00
Daniel Verkamp
89d35cefd5 nvmf: refactor get_log_page into a common function
Both regular NVM controllers and discovery controllers implement the Get
Log Page command; combine the implementations into one in ctrlr.c.

Change-Id: I7fabf40ec52d8738263ac152afe9cd7773ff7fbd
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/374555
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-17 17:58:49 -04:00
Daniel Verkamp
2d1a2926c5 nvmf: refactor controller creation
Simplify the initialization of controller identify data and properties.
This is mostly common between discovery and NVM subsystems.

Change-Id: Id8074d5260213e32892fd77845884a80071c9d88
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/374534
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-17 17:58:49 -04:00
Ben Walker
baa936a173 nvmf: Replace transport controller with poll_group
For now, this is a name change of controller to poll_group
in the transport layer. Over time, the poll_group will
become a more general concept than a controller, allowing
for qpairs to be spread across cores.

Change-Id: Ia92a2934541ad336f462f73175d53aaaf021f67b
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/371775
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2017-08-11 19:44:35 -04:00
Ben Walker
a4e28342a8 nvmf: Add wrappers for transport calls
Instead of scattering direct calls to the function
callbacks throughout the code, add some wrappers.
This will make some later refactoring marginally
easier.

Change-Id: If735089967e3ce828dcff68f2430e7810bf2f123
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/371749
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
2017-08-08 15:46:28 -04:00
Ben Walker
31d033f944 nvmf: Transports are no longer global
Create one transport per nvmf target. Today, there is just
one global nvmf target, but this paves the way for multiple.

Change-Id: Iaa1f8c5e7b3c1e87621ef2a636c68c2dd8fd929e
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/371748
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
2017-08-08 14:09:45 -04:00
Ben Walker
1d304bc5d8 nvmf: Rename spdk_nvmf_conn to spdk_nvmf_qpair
Match the terminology used in the NVMe-oF specification,
which is queue pair. For the RDMA transport, this maps to
an RDMA queue pair, but may map to other things for other
transports. It still is logically a "connection" from
a networking sense.

Change-Id: Ic43a5398e63ac85c93a8e0417e4b0d2905bf2dfc
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/371747
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2017-08-08 13:46:48 -04:00
Ben Walker
03788f93df nvmf: Rename spdk_nvmf_session to spdk_nvmf_ctrlr
This is just a rename - the functionality hasn't changed.
Use the same terminology as the specification (which is controller)
so those familiar with the specification can more easily
approach the code base.

This is still conceptually equivalent to a "session" in the
networking sense.

Change-Id: I388b56df62d19560224c4adc2a03c71eae6fed0d
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/371746
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2017-08-08 13:46:48 -04:00