In a similar manner for what we do for other per IO data-structures of cmds,
cpls and bufs, use the conventional huge-pages based spdk allocation scheme
for RDMA requests and receives.
Change-Id: I4c2e86e928106e78c053f24915e2a9ce1a200c78
Signed-off-by: Or Gerlitz <ogerlitz@nvidia.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12273
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
To maximize cache locality, use lifo and not fifo when managing objects
which are used per IO such as the RDMA receive elements queue.
Change-Id: Id8917558acc1bec29943fcbae6afe6b072bde6ac
Reported-by: Jim Harris <james.r.harris@intel.com>
Signed-off-by: Or Gerlitz <ogerlitz@nvidia.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12272
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
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>
RDMA READs can cause cmds to be submitted to the target
layer in a different order than they were received
from the host. Normally this is fine, but not for
fused commands.
So track fused commands as they reach
nvmf_rdma_request_process(). If we find a pair of
sequential commands that don't have valid FUSED settings
(i.e. NONE/SECOND, FIRST/NONE, FIRST/FIRST), we mark
the requests as "fused_failed" and will later fail them
just before they would be normally sent to the target
layer.
When we do find a pair of valid fused commands (FIRST
followed by SECOND), we will wait until both are
READY_TO_EXECUTE, and then submit them to the target
layer consecutively.
This fixes issue #2428 for RDMA transport.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I01ebf90e17761499fb6601456811f442dc2a2950
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12018
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The rdma buffer for stripping DIF metadata is added. CPU strips the DIF
metadata and copies it to the rdma buffer, improving the rdma write
bandwith. The network bandwidth during 4KB random read test is increased
from 79 Gbps to 99 Gbps, the IOPS is increased from 2075K to 2637K.
Fixes issue #2418
Signed-off-by: Chunsong Feng <fengchunsong@huawei.com>
Change-Id: If1c31256f0390f31d396812fa33cd650bf52b336
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/11861
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.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>
Extract the code for DIF from nvmf_rdma_fill_wr_sgl() into
nvmf_rdma_fill_wr_sgl_with_dif().
Then clean up nvmf_rdma_request_fill_iovs() and
nvmf_rdma_request_fill_iovs_multi_sgl().
Additionally, this patch has a bug fix. nvmf_rdma_fill_wr_sgl_with_dif()
returned false if spdk_rdma_get_translation() failed. However, the
type of return value of nvmf_rdma_fill_wr_sgl_with_dif() is not bool
but int. The boolean false is 0 in integer. Hence in this case,
nvmf_rdma_fill_wr_sgl_with_dif() returned 0 even if it failed.
Change nvmf_rdma_fill_wr_sgl_with_dif() to return rc as is if
spdk_rdma_get_translation() returns non-zero rc.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I71cc186458bfe8863964ab68e2d014c495312cd3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/11965
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: fengchunsong <fengchunsong@huawei.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
When rq is used, rnr_retry_count is not set. As a result, the initiator
cannot connect to the target after some error. Set the rnr_retry_count
to avoid infinite retries.
Fix issue #2429
Signed-off-by: Chunsong Feng <fengchunsong@huawei.com>
Change-Id: Iab03ea5ff10f5a9759c41f26589731aa92ef2f56
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/11946
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
If 0 - UINT32_MAX or UINT32_MAX - 0 is substituted into a int variable,
we cannot get any expected result.
Fix the bug and add unit test case to verify the fix.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Iad2ea681ad8ad234e70c7310b58785a999612156
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/11837
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>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
To speed up qpair lookup, an idea to use the DPDK hash library was
proposed. But we have some concerns for the idea, we may see performance
regression for small configuration, lib/nvmf is depending on the DPDK
feature, and so on.
On the other hand, RB tree is simple to use and is not likely to have
negative side effect.
To use RB tree easily, cache QP num into struct spdk_nvmf_rdma_qpair.
Change-Id: Id4ea4e75692f86335b3e984c3fc128e187f22c72
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/11798
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
nvmf_rdma_poll_group_add() added a rqpair to the poll_group->qpairs list
before its initialization.
If we change the poll_group->qpairs list from linked list to RB tree,
the qpair can be added to the poll_group_qpairs list only after its
initialization completes because QP number is necessary.
As a preparation, change nvmf_rdma_poll_group_add() to add a rqpair to
the poll_group->qpairs list after its initialization succeeds.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I42525f68c5efbd445f96da8f4f63d31f06f739e3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/11800
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
For the benefit of forthcoming vfio-user changes, register the poll
group poller prior to calling the transport create callback, and pass in
a pointer to the poll group itself.
Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: Idbc24126c9d46f8162e4ded07c5a0ecf074fc7dd
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10718
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
The dif_enable is not cleared in _nvmf_rdma_request_free. When the
rdma_req is used again, the dif_enable is true while
dif.dif_ctx->block_size is zero. As a result, an infinite loop occurs in
nvmf_rdma_fill_wr_sgl.
Fix issue: #2380
Signed-off-by: Chunsong Feng <fengchunsong@huawei.com>
Change-Id: Ic179855f7b257e39ed4a5f6705fbc9dea64210ae
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/11646
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
SPDK RDMA target reports msdbd=16, these addtitional
SGL descriptors are located in capsule. The user can
set ICD size lower than required for msdbd=16. This
patch verifies that ICD can hold all additional SGLs.
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: I875d40e14e6506c39169d084e56df7ca5d761209
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10686
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Acceptor poller is registered using rate value
from transport opts structure, but this structure is
initialized on generic transport layer when create()
function completes, so at this time acceptor poll rate
is 0.
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: I2138825f3ff9dd3cc0ccaa65e8d5c23aab338ad4
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/11095
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Anil Veerabhadrappa <anil.veerabhadrappa@broadcom.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
These definitions will be used in the next patch to check if
device is rxe
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: Icc073344103991ff24fc3bb88a1ceb9867de6f6a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10727
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Some devices may support SRQ depth lower than defaulut
value 4096
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: I58da0ac268a6d4c4a7e3b500ae37b8fad4810e17
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10654
Community-CI: Mellanox Build Bot
Community-CI: Broadcom 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: Jim Harris <james.r.harris@intel.com>
Not every transport requires accept poller - transport specific
layer can have its own policy and way of handling new connection.
APIs to notify generic layer are already in place
- spdk_nvmf_poll_group_add
- spdk_nvmf_tgt_new_qpair
Having accept poller removed should simplify interrupt mode impl
in transport specific layer.
Fixes issue #1876
Change-Id: Ia6cac0c2da67a298e88956734c50fb6e6b7521f1
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7268
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>
Add a parameter which determines the owner of the
map - target or initiator. It allows to set different
access flags when creating Memory Regions
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: I0016847fe116e193d0954db1c8e65066b4ff82bf
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10283
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This is to help with binding trace objects together and
for the convenience (all trace definitions are in one place
instad of being scattered accross multiple files).
Change-Id: Ib15bc9c2eeee9c4d0816bcee509ab69f3f558e19
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9574
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@gmail.com>
Now that the trace library can handle multiple arguments, there's no
point in passing 0 for tracepoints that don't have any arguments. This
patch removes all such instances. It allows us to to verify that
`spdk_trace_record()` was issued with the exact number of arguments as
specified in the definition of the tracepoint.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Idbdb6f5111bd6175e145a12c1f0c095b62d744a9
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8125
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
in-capsule data over 4KiB when using the Linux initiator.
This is fixed in the latest kernel. See
https://lists.infradead.org/pipermail/linux-nvme/2021-May/025641.htmlFixes#1823
Change-Id: Ie383ea774ee31ef8fe255119095b21603483c33f
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8424
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
The arguments of a tracepoint are formatted when they're printed now, so
there's no need to append ":" or pad it with spaces.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I74f5568f1982dacc079e3b80bd19a9cd740b48ce
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7955
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Both values should provide similar information, while the qpair can also
be matched to the traces from lib/nvmf allowing the user to track the
qpairs across these modules.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Iba9abdd3f41b93100c0403b1c90fc4549d39189e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7159
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: Paul Luse <paul.e.luse@intel.com>
Get all of the hot stuff to the first cache line.
* Shrink the xfer enum to one byte (it only has 3 values).
* Pull out the dif enabled flag form the dif structure so it
can be access separately
* Rearrange the members
Change-Id: Id4a2fe90a49c055a4672642faac0028671ebfae9
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7827
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: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Currently we allocate buffers perf each SGL descriptor.
That can lead to a problem when we use NVME bdev with
PRP controller and length of the 1st SGL descriptor is
not multiple of block size, i.e. the initiator may send
PRP1 (which is SGL[0]) which end address is page aligned
while start address is not aligned. This is allowed by
the spec. But when we read such a data to a local buffer,
start of the buffer is page aligned when its end is not.
That violates PRP requirements and we can't handle such
request. However if we use contig buffer to write both
PRP1 and PRP2 (SGL[0] and SGL[1]) then we won't meet
this problem.
Some existing unit tests were updated, 1 new was added.
Fixes github issue #1853
Change-Id: Ib2d56112b7b25e235d17bbc6df8dce4dc556e12d
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7259
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: Ben Walker <benjamin.walker@intel.com>
This patch reduces admount of changes in the next patch,
no functional changes added.
The next patch will add usage of contig IO buffers for
multi SGL payload. To support it we need to pass an
offset to fill_wr_sgl function. Also in the current
version we assume that for 1 iteration we fill 1 IO
buffer, the next patch will change it and we'll need
to swtich to the next IO buffer in special case. That
can't be done easily if we use fill_wr_sge function
Change-Id: Iee8209634637697f700f8fa9fe61ead156b6d622
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7258
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: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Note: a lot of the TCP and FC trace registers were
specifying '1' which means the arg type is a pointer,
but in reality it is always passing 0 for arg1. So
this patch just changes them to SPDK_TRACE_ARG_TYPE_INT.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I18d3cedd21e516f16cb2cd0a7f8c16670b1895d7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7763
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
This change refactors the way nvmf_get_stats RPC works.
The RPC layer passes JSON write context to custom dump function defined within transport ops.
The RPC layer no longer needs to know the structure of transport poll group statictics.
Functions and structures used in the previous flow have been deprecated and will be removed.
JSON returned for RDMA transport should be the same as before this change.
Signed-off-by: Maciej Szulik <maciej.szulik@intel.com>
Change-Id: I03308c45be120793d316bf79814a1295afd9fb95
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6681
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
This statistic is incremented when we don't reap
anything from the CQ. Together with the total number
of polls it can be useful to estimate idle percentage.
Change-Id: I61b51d049b0bc506fb8a896e225187e46e75a564
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6295
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
nvmf_create_transport rpc parameter to configure the CQ size helps
if the user is aware of CQ size needed as iWARP doesn't support CQ resize.
Fixes issue #1747
Signed-off-by: Monica Kenguva <monica.kenguva@intel.com>
Change-Id: Ia9ba2b5f612993be27ebfa3455fb4fefd80ae738
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6495
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>
Since iWARP doesn't support CQ resize, we can't
accept new qpairs when CQ size is insufficient.
In other case if we accept new qpairs we can
et CQ overrun and receive IBV_EVENT_CQ_ERR ibv
async event. In that case we will have to close
all qpairs that use this CQ.
Part of the fix for issue #1747
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: I30b9163a7c8abb8651f8813c6bbdb80000741c9c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6106
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: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
When SRQ is enabled, all qpairs share common pool of
rdma requests and we should check that rdma_req which
cid matches belongs to the correct qpair.
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: I2505ca39c5110e8e67a48abbe34b6df9ff00229f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6110
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Since now min supported DPDK version is 19.11, we can
remove handling of buffer that may be split over several
Memory Regions as it was specific for DPDK < 19.05
Change-Id: I61a79c80b864e3bdde134e8bff6622025ea578a7
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5937
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: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Transport specific request to abort is never null,
if we don't find a request by cid then we will abort the
last request.
Change-Id: Ie9eb55822b47381ae254e19266b9f3ffadbb67ea
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5196
Community-CI: Broadcom CI
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: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Originally introduced in:
"(8126509) rdma: replace improperly aligned buffers in requests."
This behaviour is for handling buffer split over two memory regions.
It is not unique to specific DPDK versions.
As such the comment and ERRLOG is not relevant here.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I31bec435d92d8271d4a040855cfc4ff97f3d9d50
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5345
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>
The code changes, and this is not the second priority.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I08fdac561f2d3f902763c39c37455727aa529b9f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5663
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>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
There are transport types like vfio-user or fc for which this field is
optional so there is no need to explicitly specify it for the add
listener rpc. Still validation of this filed is preserved for rdma and
tcp transport.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I6aef55bed45b3653472fd356e5a7cfead22c6f2b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5605
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>
With new spdk_nvmf_transport_listen function it should be possible to
add generic options without breaking API/ABI. For now it only delivers
json parameters which can be decoded on a transport specific layer.
This is similar to what was done for spdk_nvmf_transport_create and opts there.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: Iaf576248a0b10b408c4a3182785270be3e32ebe4
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5570
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>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This reverts commit e21de9a9cc.
Reason for revert: The public API spdk_nvmf_qpair_get_listen_trid() was
overlooked when this patch is created. Then the code to use the trid was
replaced by spdk_nvmf_qpair_get_listen_trid(). To avoid someone from using
the trid in future, revert this patch. struct spdk_nvmf_qpair is defined]
in the public header file, nvmf_transport.h but it is mainly for pluggable
custom transport.
Any custom transport will not use the trid, and so even if we remove it,
we will have no issue.
Change-Id: If4656edf3ac68456e758013d487eb582699750aa
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5614
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Anil Veerabhadrappa <anil.veerabhadrappa@broadcom.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jacek Kalwas <jacek.kalwas@intel.com>
iWARP doesn't support this ibv async event.
Probably some clean up of SRQ needs to be done, e.g.
post dummy drain WR to SRQ and wait for its completion
when qpair is being destroyed.
Change-Id: Ic1035dbd8192e665eedbaf89e5cf4e00cb0506c6
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5474
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
In some cases we need to check if we are working with
SoftRoCE driver (rxe) since it doesn't support several features
such as SEND_WITH_INVAL and LAST_WQE_REACHED ibv async event.
vendor_id of this driver has been changed from 0 (it was not
initialized at all) to 0XFFFFFF in commit
0184afd15a141d7ce24c32c0d86a1e3ba6bc0eb3
So now we should check for both values.
Change-Id: I62ce4d68c11612a07fe97606658a5b79f1db2149
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5307
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
As part of FC transport destroy, FC LLD (Low level Driver) needs to
to do its cleanup which cannot be completed synchronously. So
allow transport destroy to be asynchronous.
FC transport code to use this functionality will be pushed shortly.
Signed-off-by: Naresh Gottumukkala <raju.gottumukkala@broadcom.com>
Change-Id: I104cf7d131e18199abdcf0651df261fe41d666ef
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5180
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
As part of nvmf_transport_qpair_fini, FC transport needs to cleanup
all the resources used by that QPair on the shared hardware. This
hardware cleanup is asynchronous in nature.
FC transport code to use this functionality will be pushed shortly.
Change-Id: I5606a33dff45971badd74e0cc087b132b56af076
Signed-off-by: Naresh Gottumukkala <raju.gottumukkala@broadcom.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5100
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: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Anil Veerabhadrappa <anil.veerabhadrappa@broadcom.com>
