Previously, the R2T was sent and if an H2C arrived prior
to seeing the R2T ack, it was processed anyway. Serialize
this process.
In practice, if the H2C arrives with a correctly functioning
initiator, that means the R2T already made it to the initiator.
But because the PDU hasn't been released yet, immediately processing the
PDU requires an extra PDU associated with the request. Basically, making
this change halves the worst-case number of PDUs required per
connection.
In the current sock layer implementations, it's not actually possible
for the R2T send ack to occur after that H2C arrives. But with the
upcoming addition of MSG_ZEROCOPY and other sock implementations, it's
best to fix this now.
Change-Id: Ifefaf48fcf2ff1dcc75e1686bbb9229b7ae3c219
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479906
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This function was only called from one spot.
Change-Id: I856f564d3ef6c6157be7a32a2cd812c702516a8d
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482003
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This seems like a more descriptive name
Change-Id: Ia616865b3fb36d8f9ccc5fb2ca6185bdd8543cf8
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482002
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
With our target design, there's no advantage to sending
multiple R2T PDUs per nvme command. This patch starts by
setting up the math so that at most 1 R2T PDU is required
per request. This can be guaranteed because the maximum
data transfer size (MDTS) is pre-negotiated in NVMe-oF
to a reasonable size at start up.
It then proceeds to simplify all of the logic around mapping
requests to PDUs. It turns out that the mapping is now always
1:1. There are two additional cases where there is no request
object at all but a PDU is still needed - the connection response
and termination request. Put an extra PDU on the queue object
for that purpose.
This is a major simplification.
Change-Id: I8d41f9bf95e70c354ece8fb786793624bec757ea
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479905
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
We can always accept up to the maximum I/O size in an H2C,
so eliminate the #define.
Change-Id: I349dab5f9b6ec482a7c580b1396e03c8d30a250b
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482278
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
The resources allocated to a queue pair do not need to be directly
correlated to the queue size requested by the initiator in NVMe-oF, as
long as enough resources are present. The RDMA transport, for instance,
does complex pooling of the resources behind the scenes when using a
shared receive queue.
Simplify the resource allocation for a TCP qpair to just always allocate
the max allowed queue size right away. This is a configurable parameter,
so system administrators can adjust for their needs. The initiator may
then request a queue size less than or equal to that, which will only be
enforced by queue depth counting and not impact the actual number of
resources allocated on the target.
This change relies on the MaxC2HSize being equal to the Maximum Data
Transfer Size (MDTS) reported. That is the default configuration, but
MDTS is configurable. Changing the MDTS with this patch to a value
larger than 128k will cause the target to break. This is addressed in
the next patch in this series.
Change-Id: Ibd4723785c6a4d8d444f9b7bbfa89f98de2320f5
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479733
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
These values do not need to be negative.
Change-Id: Id9f798cf1c9da354448f9c6fbb90e599f877bb32
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482277
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
By releasing the just-completed PDU prior to calling the callback,
for flows that immediately submit another PDU inside the callback,
the just-released PDU can be immediately reused. This reduces the number
of PDUs required in the pool to continue forward progress to half of the
previous value, while also making it more CPU cache friendly.
Change-Id: I8031b8f9f57ac05f261d96433d9899fe5e31d318
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479904
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Or Gerlitz <gerlitz.or@gmail.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This was calling a callback for another function which
attempted to release the request. The code only worked because
in the r2t case the cb_arg was set to NULL, and that makes
the request free function do nothing.
Change-Id: Id9ec30ceb0eaa41deb67aa995da5d6f786d9b9f0
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479903
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Or Gerlitz <gerlitz.or@gmail.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This wasn't actually used. Every PDU only had a single reference.
Change-Id: I8adaa7edeca5fe175aa853c156df741170d76c10
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479902
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This would allow to respond for add listener rpc request even
when there are async calls in transport specific function.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I94a9f45b7ba9e8d46a60ae3785953cea12554732
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479511
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Anil Veerabhadrappa <anil.veerabhadrappa@broadcom.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Make common code as part of successful return.
In rdma check if already listening first.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: Ib0c87ac11db7daff00dc4042c9e0ab20eb7ffd0f
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/478721
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Purpose: With this patch,
(1)We can support using different sock implementations in
one application together.
(2)For one IP address managed by kernel, we can use different method
to listen/connect, e.g., posix, or uring. With this patch, we can
designate the specified sock implementation if impl_name is not NULL
and valid. Otherwise, spdk_sock_listen/connect will try to use the sock
implementations in the list by order if impl_name is NULL.
Without this patch, the app will always use the same type of sock implementation
if the order is fixed. For example, if we have posix and uring together,
the first one will always be uring.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: Ic49563f5025085471d356798e522ff7ab748f586
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/478140
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
The trtype should be stored as both an enum and string. This is intended to
help pave the way for pluggable NVMe-oF transports.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I6af658d7a17c405e191ff401b80ab704c65497e7
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/478744
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
It is faster for the kernel to pin memory in hugepages, so allocate
the pdu pool from hugepages. This will help more
with upcoming changes to leverage MSG_ZEROCOPY.
Change-Id: I9ce581acca9c6edb71bd8119258966e3b405db77
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/475801
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Or Gerlitz <gerlitz.or@gmail.com>
All transitions to the EXITING state go through the disconnect function
now
Change-Id: Ia55816351b2998bfef26130b6ffdc4a1010567a1
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/470533
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This function can't actually return NULL. It aborts if we get
our math wrong.
Change-Id: Iaf77112addc3c14c70755a56043c5dba3427890d
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/478911
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
We can use spdk_min to get the copy_len in spdk_nvmf_tcp_send_c2h_term_req.
It confirms copy_len it's not larger than SPDK_NVME_TCP_TERM_REQ_ERROR_DATA_MAX_SIZE
Signed-off-by: dongx.yi <dongx.yi@intel.com>
Change-Id: Id343928e1911e4ab77fca7463f3f0cc55889db30
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479118
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Minor optimisation done by code analysis, both cmd and dif are
overridden in TCP_REQUEST_STATE_NEW.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I6bae4ddae175035d029c0693f7e4351b95a296ab
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/478604
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
It's not wrong, just to keep consistency with other functions.
So remove these.
Signed-off-by: dongx.yi <dongx.yi@intel.com>
Change-Id: I833211ea8ee6c6b02c874ea340a3f936a0c4c00f
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/478684
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
It will be the expected behavior when the error message will
printed if we use asynchrounous I/O. And the real
error message for not getting the tcp_req is located in
spdk_nvmf_tcp_capsule_cmd_hdr_handle.
Change-Id: I1a608fbd3a04050eacb6cb68eafd50e5128925ab
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/477872
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: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
TCP_REQUEST_STATE_NEW is already set in spdk_nvmf_tcp_req_get.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: Ia835f3763cd74ef9b504901c719d9954317f49af
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/476164
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This eliminates the flushing logic, simplifying the tcp
transport.
This also happens to greatly improve performance, especially
on random read tests. The batching done in spdk_sock_writev_async seems
to be more effectively than the previous batching logic in the tcp
transport.
Change-Id: Id980ac6073e380dc75f95df3f69cb224f50fb01b
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/470532
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
It is valuable to have more detail status instead
SPDK_NVME_SC_INTERNAL_DEVICE_ERROR.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: Ifd003b490a7ae9af017645c97636ceaf2f93d4b0
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/476634
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Destroy in poll_group_add results in heap-use-after-free because
upper layer calls qpair_fini in case poll_group_add returns
error.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I3e921a21b7ab5f7c15c80bc5919cb97cbda0b5d2
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/475858
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Also need to update the spdk_nvmf_tcp_poll_group_poll.
Since if the tqpair recv state in wait_for_req,
we may already received the data, and there could be
not epoll event.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I9c5a202e47e57aaba63da143f954a20c135a98ae
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/473626
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Purpose: But if we use asynchronous writev
for pdu sending, the call_back of writev may occur
after the new data coming. So it means that the
free tcp request may not be available.
So we use the strategy to check the request status
in TCP_REQUEST_STATE_TRANSFERRING_CONTROLLER_TO_HOST.
So the strategy is checking the state_cntr of all the
reqs in TCP_REQUEST_STATE_TRANSFERRING_CONTROLLER_TO_HOST
state.
1 If the state_cntr > 0, we should queue
the new request.
2 If the statec_cntr == 0, it means that
there is no available slot for the new tcp request
, i.e., the new nvme command comming from the initiator.
If we receive this, it means that the initiator sends more
requests，and we should reject it.
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: Ifbeb510e669082cb7b80faf2e7987075af31d176
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/472912
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
To avoid the allocation of ttransport in the sub functions,
and it makes the code much efficient.
Change-Id: Ie4c5a1755ddbecf10dc364ff811f74a7af5f9c3b
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/473003
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
When we use async writev (e.g., lib io_uring), we find that
the callback of writev is executed after recving the new
data from the initiator, and this is possible.
For example, if the NVMe-oF TCP target receives the ic_req from the
initiator, and sendout the ic_resp, the state of tqpair will change from
invalid to running until the callback is executed. And the data of ic_resp
is already sent to the initiator, and we receive the new command later. However,
we may still not get the call back function executed
(i.e, spdk_nvmf_tcp_send_icresp_complete). And it is possible
for using lib io_uring, I faced this issue when using lib uring.
And this patch can fix this issue.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I7f4332522866d475e106ac6d36a8ec715133f0dc
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/472770
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
The loop is intended to accept multiple socks when
available, but once accept returns NULL, there's no
reason to keep trying.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I896908d276da35bc3fff172c1c17e22abd2a5343
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/473234
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
It's important to be able to recover full context from just
the PDU in the future.
Change-Id: I3d1f3c326299b1237b42dbe33d340a282c3bc5bb
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/470531
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
This is always the request pointer, so rename it for clarity.
Change-Id: Ifbda7db7787c65f0deb190a1e94f0676b2c0d99a
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/470530
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Use whatever size the socket layer thinks is best. In practice,
this is the same size as before.
Change-Id: I4820e16d8da6e566d1f8f078a75d345399f64ab5
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/470511
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Sasha Kotchubievsky <sashakot@mellanox.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Since we use big buffer to read the data, so the incoming
data may already be read when the req is waiting for the buffer.
So if we use the orginalstatement machine, there will be no
read event will be generated again.
The quick solution is to restore the original code, since
for req which has incapsule data, we not need to wait for the
buffer from the shared buffer pool.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: Ib195d57cc2969235203c34664115c3322d1c9eae
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/472047
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
It can be useful for passing additional information about nvmf
target to a handler for new nvmf connections. Context can be
stored in globals as it is currently done in nvmf code. However
in case of multiple targets or languages where accessing global
state is challenging (i.e. Rust), this becomes inconvenient.
Change-Id: Ia6a2fdba4601531822b3e5fda7ac5ab89d46f6c5
Signed-off-by: Jan Kryl <jan.kryl@mayadata.io>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/469263
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Seth Howell <seth.howell@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Sasha Kotchubievsky <sashakot@mellanox.com>
Run the following command:
pahole ./app/nvmf_tgt/nvmf_tgt -R -C spdk_nvmf_tcp_req
It tells me change the bool definition location
of dif_insert_or_strip.
Change-Id: Ia43ab62bcc223a07e6415b2c769fe4af2b097f18
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/470401
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
By passing the pointer to struct spdk_nvmf_transport_poll_group
to spdk_nvmf_tcp_req_parse_sgl(), we can remove spdk_nvmf_tcp_req_fill_iovs()
and inline spdk_nvmf_request_get_buffers() into spdk_nvmf_tcp_req_parse_sgl().
Pointers to struct spdk_nvmf_request are used in many lines of
spdk_nvmf_tcp_req_parse_sgl(). Caching and using them simplifies and
improves readability a little for spdk_nvmf_tcp_req_parse_sgl().
We can pass pointer to not struct spdk_nvmf_tcp_transport but struct
spdk_nvmf_transport to spdk_nvmf_tcp_req_parse_sgl().
Ordering the pointer to struct spdk_nvmf_tcp_req first in parameters
of spdk_nvmf_tcp_req_parse_sgl() matches the function name.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I9f0d33b48383800c3b0a738eb24b11ffed7e6e60
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/469640
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This patch is close to the end of the effort to unify buffer allocation
among NVMe-oF transports.
Merge each transport's fill_buffers() into common
spdk_nvmf_request_get_buffers() of the generic NVMe-oF transport.
One noticeable change is to set req->data_from_pool to true not in
each specific transport but in the generic transport.
The next patch will add spdk_nvmf_request_get_multi_buffers() for
multi SGL case of RDMA transport.
This relatively long patch series is a preparation to support
zcopy APIs in NVMe-oF target.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Icb04e3a1fa4f5a360b1b26d2ab7c67606ca7c9a0
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/469205
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The subsequent patches unifies getting buffers, filling iovecs, and
filling WRs in a single API. This is a preparation.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I077c4ea8957dcb3c7e4f4181f18b04b343e9927d
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/468953
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Seth Howell <seth.howell@intel.com>
This patch makes multi SGL case possible to call spdk_nvmf_request_get_buffers()
per WR.
This patch has an unrelated fix to clear req->iovcnt in
reset_nvmf_rdma_request() in UT. We can do the fix in a separate patch
but include it in this patch because it is very small.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: If6e5af0505fb199c95ef5d0522b579242a7cef29
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/468942
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Seth Howell <seth.howell@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
On a DIF verification error, fail the read command with a status code
of APPLICATION_TAG_CHECK_ERROR, GUARD_CHECK_ERROR, or
REFERENCE_TAG_CHECK_ERROR and a status code type of SCT_MEDIA_ERROR.
The state of the request is TCP_REQUEST_STATE_TRANSFERRING_CONTROLLER_TO_HOST
when a DIF verification error is detected. So dequeue the request
from C2H data queue, return the response PDU, and then send the command
response.
This was an item on the TODO list. RDMA transport do this right
behavior from the start and so TCP transport follows it by this patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I102bbd253cc8c1379d0937c9536bf2bfe04cbf6a
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/468911
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
tcp_req->orig_length had been set just before I/O submission but
the value is already fixed in spdk_nvmf_tcp_req_parse_sgl().
Hence move setting tcp_req->orig_length accordingly.
This follows the good practice of RDMA transport.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I99f6e266d8f7027bce810864314f3ee24a1af10c
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/468910
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
