Commit Graph

8577 Commits

Author SHA1 Message Date
Konrad Sztyber
5263f0a12f nvme: extract qpair connect fail to a seprate function
In the next patch, the qpair is polled from a poll group and needs a
disconnect callback, which should also fail the qpair, so it makes sense
to have a separate function doing that.

Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ied76431520962b25220027be829a4609afb6bbda
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9157
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: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-08-13 07:27:07 +00:00
Monica Kenguva
771f65bb1f nvme: asynchronous create io qpair
async_mode option is currently supported in PCIe transport layer
to create io qpair asynchronously. User polls the io_qpair for
completions, after create cq and sq completes in order, pqpair
is set to READY state. I/O submitted before the qpair is ready
is queued internally. Currently other transports only support
synchronous io qpair creation.

Signed-off-by: Monica Kenguva <monica.kenguva@intel.com>
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: Ib2f9043872bd5602274e2508cf1fe9ff4211cabb
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8911
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: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-08-13 07:27:07 +00:00
Tomasz Zawadzki
a7b15178c2 lib/event: do not set interrupt mode when eventfd is not supported
spdk_reactor_set_interrupt_mode() writes/reads from fds created during
reactor_interrupt_init(). Since spdk_fd_group_create() depends
on eventfd, this will not work for systems that do not have it.

reactor_interrupt_init() handled lack of support for eventfd correctly,
while spdk_reactor_set_interrupt_mode() did not check for it.

Reported-by: Nick Connolly <nick.connolly@mayadata.io>
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I5181d436636c55cca3a06b1947e944502a9204ce
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9131
Reviewed-by: Nick Connolly <nick.connolly@mayadata.io>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2021-08-13 07:21:33 +00:00
Changpeng Liu
1afe4de585 nvmf/vfio-user: refactor map_admin_cmd_req()
PRP1 value 0 is a valid Guest physical address and it may use two vectors
for one page, so we need to set `iovcnt` as the IO commands.  Also
fix the calculation for Get Log Page command.

Change-Id: Ib036bcf1ca9f6644dfd8a7406fea37269cea5c73
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9122
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
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>
2021-08-13 01:25:07 +00:00
Changpeng Liu
0ac315f047 nvmf/vfio-user: inline ctrlr_interrupt_enabled()
This function is called in the IO processing context, also
rename vfio-user controller data structure with "vu_" prefix.

Change-Id: Icb8fd18772f64c1e633f4a8b8eac26d0bafeed16
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9121
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: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-08-13 01:25:07 +00:00
Changpeng Liu
2746d51aa7 nvmf/vfio-user: inline io_q_id()
This function is called in the IO processing context, also rename
internal queue pair variable with "vu_" prefix.

Change-Id: Iac6be1fddf9ee28e0485ff0c3a707de4e98be96a
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9120
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: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-08-13 01:25:07 +00:00
Changpeng Liu
40c447f843 nvmf/vfio-user: remove duplicated one line
The submission queue doorbell will be cleared to 0 at the end
of this function.

Change-Id: I0146f3dc208b8ca1d7fec4b2dae1d7395f304444
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8992
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: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Thanos Makatos <thanos.makatos@nutanix.com>
2021-08-13 01:25:07 +00:00
Changpeng Liu
42abbac90e nvmf/vfio-user: eliminate cq_next function
There is only one place to call it now, so just remove it.

Change-Id: Ib16d640c80dbe79df46c965c5ee8bcd092e57b0d
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9013
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: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-08-13 01:25:07 +00:00
Changpeng Liu
2f85499e43 nvmf/vfio-user: don't use Guest's phase bit for completion response
We should maintain phase bit in vfio-user target, it's not safe
to use Guest completion queue's phase bit.

Change-Id: I94856853784802f7695a0484b783d8de4042f8a6
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8988
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: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Thanos Makatos <thanos.makatos@nutanix.com>
2021-08-13 01:25:07 +00:00
Changpeng Liu
a93387f69f nvmf/vfio-user: remap admin queue during the memory hotplug if necessary
It's wrong to call enable_admin_queue() in the memory hotplug context,
as the function will change doorbells and clear queue pair memory.

We can remap the ADMIN queue pair's memory just same as IO queue pairs.

Change-Id: I090ea40ce0f0613c15319bfb18771a154ca788c7
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8987
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: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Thanos Makatos <thanos.makatos@nutanix.com>
2021-08-13 01:25:07 +00:00
Changpeng Liu
e9fd737a8e nvmf/vfio-user: add map_q function
The map_q function is used to map Guest's physical memory address
to Host virtual address for SQ/CQ.  This function can be used both
for initializing SQ/CQ and remap SQ/CQ, when used for remap SQ/CQ,
we don't need to unmap the related memory region.

Change-Id: Ia42e01a68482e5678dbaf8d0d4e8c04c1a94789d
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9012
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: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
2021-08-13 01:25:07 +00:00
Changpeng Liu
7d136a2088 nvmf: use INFOLOG for unsupported commands
Fix #2004.

Change-Id: I60899de4d369f50f159e6bbe1ba9eb546f367d87
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9035
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: Thanos Makatos <thanos.makatos@nutanix.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-08-12 12:57:50 +00:00
Changpeng Liu
a784122709 nvmf/vfio-user: use correct syslog level
vsnprintf() is called in the IO processing path if we set LOG_DEBUG
by default, which will impact the performance numbers. See below
perf record:

13.44%  reactor_0  [kernel.kallsyms]   [k] do_syscall_64
 7.59%  reactor_0  [kernel.kallsyms]   [k] syscall_return_via_sysret
 7.18%  reactor_0  libc-2.29.so        [.] __vfprintf_internal
 5.60%  reactor_0  [kernel.kallsyms]   [k] entry_SYSCALL_64
 4.74%  reactor_0  [kernel.kallsyms]   [k] vmx_deliver_posted_interrupt
 4.36%  reactor_0  nvmf_tgt            [.] post_completion

Change-Id: Icc12345c341d870375a1db729412acb7b3cd0211
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9079
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: John Levon <levon@movementarian.org>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-08-12 12:57:39 +00:00
Changpeng Liu
01936d2ed8 log: add an API to return syslog level based on SPDK's log level
Change-Id: I970119385989e7d97bda065f5aa2bbd82519c655
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9103
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: John Levon <levon@movementarian.org>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-08-12 12:57:39 +00:00
paul luse
7abf75c2b5 lib/idxd: pre translate the completion address for batch entries
Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: I9df51acc2bbd3deffcb172c0fd3e7012ec3d4b3b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9100
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: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2021-08-12 12:57:07 +00:00
paul luse
d09b5674a4 lib/idxd: reduce size of per channel batch pool
Was set at a pretty high number during early development.
Instead of a #define, lets use the same math we use to
determine the size of the operation and descriptor pools as
that's the max number of batches that will be needed.

Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: Ide5311b4b6c931010413d0f82baed0e05d5fcde7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9099
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: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2021-08-12 12:57:07 +00:00
paul luse
31e3ed5ccb lib/idxd: fix leak of DSA operation for non batched commands
When address translation failed we were not putting the op back
into the pool.  In one place also changed the return error from
translation from hardcoded value to the rc returned by the call
to be consistent with the rest of the code.

Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: I6d0c9c0f469a213721c6a5ce37c4e9dbdcef8183
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9097
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: Aleksey Marchuk <alexeymar@mellanox.com>
2021-08-12 12:57:07 +00:00
paul luse
d30e459e8d lib/idxd: pre translate completion record addresses
These are fixed, no reason to tranlate them with every IO.  This
patch is just for non-batched IO.  Batched IO will come later.

Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: Ifb7bdf147658a23714d94bb7f88d7fedd1749153
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9073
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: Ben Walker <benjamin.walker@intel.com>
2021-08-12 12:57:07 +00:00
Jim Harris
690783a3ae env_dpdk: tokenize env_context
DPDK requires each command line parameter to be
passed as a separate string in the args array.  So
we need to tokenize opts.env_context to handle the
case where a user passes multiple arguments in the
env_context string.

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Ifc72a7c25b31f1296d3aba3a3f7664ad8edf128f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9132
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: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-08-10 08:19:52 +00:00
Rui Chang
33acbe9caf spdk_top: fix vhost poller name issue
When run spdk_top against vhost, the poller name is displayed as
"bvdev->bdev ? vdev_worker : no_bdev_vdev_worker" which should be "vdev_worker".

Signed-off-by: Rui Chang <rui.chang@arm.com>
Change-Id: Ie97c69bd8cdd3fdafc5fe4dd80e1fde85a8e3238
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9130
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: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
2021-08-10 08:19:33 +00:00
Richael Zhuang
e258fe53a1 lib/event: add in_interrupt in rpc
add in_interrupt in framework_get_reactors.

Change-Id: I7137aa604129cfd26acd2dec4ae9ca545c7e4348
Signed-off-by: Richael Zhuang <richael.zhuang@arm.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8552
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: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2021-08-09 08:40:25 +00:00
yupeng
304153e5c7 blobstore: calculate correct num_md_clusters in bs_recover
The current num_md_clusters doesn't include the the part before
md_start. So the bs_recover will get more num_free_clusters than it
should be. This patch can fix it.

Signed-off-by: Peng Yu <yupeng0921@gmail.com>
Change-Id: I911926beb69aca677da508ba71f292496c917e7f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9034
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: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2021-08-09 08:39:42 +00:00
Monica Kenguva
455a5d7821 nvme/pcie: Create queue pairs asynchronously
The generic transport layer still does a busy wait, but at least
the logic in the PCIe transport now creates the queue pair
asynchronously.

Signed-off-by: Monica Kenguva <monica.kenguva@intel.com>
Change-Id: I9669ccb81a90ee0a36d3f5512bc49c503923b293
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8910
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: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-08-09 08:39:28 +00:00
paul luse
f93f3ee3f9 lib/idxd: move _vtophys() up
In prep for upcoming patch

Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: Iebc5bf5f6715c85cc09b404128338cd6f08c421e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9072
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2021-08-06 07:33:06 +00:00
paul luse
f3b423dc84 lib/idxd: remove batch_op element from op struct
We can instead use a combination of the op code and the batch
element in the op structure to determine if the op that is
completing is part of a batch or not so we know whether to
return it to the free list or not.

Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: I8e27c7b991f5a20e82394a25a52ece560cab5543
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9071
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2021-08-06 07:33:06 +00:00
paul luse
bc393d841c lib/idxd: translate the completion address for batched commands
Was simply missing.

Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: I111ad42977f75b8df89a7d63bba6936c5861f657
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9070
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: Jim Harris <james.r.harris@intel.com>
2021-08-06 07:33:06 +00:00
paul luse
5bfcd5b329 lib/idxd: wait to add operation to outstanding list until hw submit
No real material impact but easier to read and cleans up error
handling when submitting a batch. Before we put the operation on
the list of operations to poll when it was prepared, a few LOC
before it was submitted to HW.  Now we do it in the hw submission
function. For batch elements, they are added just before the
batch operation itself (as was before).

Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: Icfeaeeb650d5b63f81e13a8ffd27e349c7118089
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9069
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: Jim Harris <james.r.harris@intel.com>
2021-08-06 07:33:06 +00:00
paul luse
e5a19b3ea2 lib/idxd: replace bit arrays with lists
The bit arrays were used for dynamic flow control in a previous
implementation.  They are no longer needed as flow control is
now static and managed solely in the idxd plug in module. Use
simple lists of descriptors and completion records instead.

This is a simpler implementation and will allow for some future
clean up of structures as well.

Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: I8c4cce12e88ac5416e3fe29a416487759214ec9f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8922
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: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
2021-08-06 07:33:06 +00:00
Changpeng Liu
a8ca2e74f9 nvmf/vfio-user: rename map/unmap_admin_queue function
Actually we should not re-enable the ADMIN queue during
the memory region callback, similar with the IO queues,
we should just do the remap, but to avoid changing other
values such as doorbells.

This is a preparation patch just to rename it, and will fix
the issue in coming patchs.

Change-Id: Ic19654e944e4c58ce0fac9582c6bb187110fe571
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8986
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: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Thanos Makatos <thanos.makatos@nutanix.com>
2021-08-06 07:31:57 +00:00
Changpeng Liu
28a95bdaa7 nvmf/vfio-user: inline SQ/CQ operation function
Change-Id: I97d085d7ffcb01e4dae0bc6b686bffc129132aca
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8985
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: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Thanos Makatos <thanos.makatos@nutanix.com>
2021-08-06 07:31:57 +00:00
Changpeng Liu
77522c0a67 nvmf/vfio-user: fill SQID field when posting completion response
This field is required for Multiple SQs share one Completion queue.
Windows requires SQID field even not for the shared Completion
queue case, so we need to fill it.

Fix issue #2009

Change-Id: Ifb2f20196baeaf9b982154f0480f360bc97a8c2d
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8971
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Thanos Makatos <thanos.makatos@nutanix.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-08-06 07:31:57 +00:00
Rui Chang
d483d8a4f1 lib/util: optimize base64 encode and decode using ARM SVE intrinsics
Optimize base64 encode and decode using ARM SVE intrinsics.

Signed-off-by: Rui Chang <rui.chang@arm.com>
Change-Id: I63ae5b9105aef690e1fad8ec1ef98c8339758e9f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8858
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: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-08-06 07:31:06 +00:00
John Levon
849f05c9df nvmf/vfio-user: create queues directly
Currently, queue creation is delegated to nvmf_vfio_user_accept() for both I/O
and admin queues, by adding it to a ->new_qps list. But there's no good reason
for this indirection to exist, and it would make interrupt support for the
accept poller harder to implement. Instead, directly call
spdk_nvmf_tgt_new_qpair().

Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: I182bfac9c621eda2f746201db7022ab6fce1bf30
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9053
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2021-08-05 07:43:11 +00:00
John Levon
efd928e959 nvmf/vfio-user: fix comment for map_io_cmd_req()
The comment no longer reflected the code.

Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: I5903e9483f01dd9b7a61c8f725eba4c38e1b8459
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9008
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2021-08-02 09:05:34 +00:00
Jim Harris
24a8c04af0 nvmf: zero data buffer before building identify ns list
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I772823ae4a57d988ed186c3fba91871580e95d5a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8984
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: Aleksey Marchuk <alexeymar@mellanox.com>
2021-08-02 09:05:18 +00:00
Changpeng Liu
80b3b68841 nvmf/vfio-user: don't pass the original NVMe command pointer to post_completion()
Only the CID is required when posting a completion response to the
completion queue.

Change-Id: I6a009386994d19a2fa05a0f5ed8c3237263644f3
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8970
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: Thanos Makatos <thanos.makatos@nutanix.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-07-30 08:23:45 +00:00
Changpeng Liu
4216f7fff5 nvmf/vfio-user: fix shared CQ usage case
Multiple IO Submission Queue can share one Completion Queue, and
we use field 'cqid' to save it in Submission Queue, so when posting
completion response, we need to get the Submission Queue's CQID first,
then post the completion queue based on CQID.

Also rename vfio-user internal variables with 'vu_' prefix in this
function.

Change-Id: Ib73b91a86740d105b5fb5c73127484ebfb6b55ef
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8969
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: Thanos Makatos <thanos.makatos@nutanix.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-07-30 08:23:45 +00:00
Changpeng Liu
5c1f77f590 nvmf/vfio-user: optimize create_io_q function with a stack qid variable
No actual logic change for this patch.

Change-Id: I486e889499ee2e4bdf05d1f633ded1712333fa88
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8968
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: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Thanos Makatos <thanos.makatos@nutanix.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-07-30 08:23:45 +00:00
Ziye Yang
208b7016b3 accel_engine: Fix the crc destnation address
This issue is introduced by the refactoring, i.e.,
changed from dst to crc_dst. And this code
part is missed.

This patch can fix this issue.

Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: Iad03d08707167c957193f6101eb9166a30e5cdd3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8935
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-07-30 08:22:48 +00:00
Liu Xiaodong
7ff8688ceb thread: print notice on changing intr mode
Change-Id: I79fcef41adf672d2e9dc34a579a1882f8e964bd7
Signed-off-by: Liu Xiaodong <xiaodong.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8890
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-07-30 08:21:43 +00:00
John Levon
1d4a88104d vfio-user: correct return code for library context poller
Currently, the poller that calls vfu_run_ctx() always returns SPDK_POLLER_BUSY.
Update libvfio-user and adjust the API usage so that it can accurately
report SPDK_POLLER_IDLE when needed.

Additionally, renaming the poller to better reflect its meaning: it's not just
for mmio handlers, but libvfio-user handling in general.

Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: I5e598241ac0a692f03ee36242ff977c4e6f14987
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8568
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: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-07-30 08:20:49 +00:00
John Levon
923bab5f12 vfio-user: correct accept poller return code
Return the number of events handled as expected by the poller.

Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: I70c4d32bf091b2c1a293eaa41f00869a3a7303f2
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8563
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: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Thanos Makatos <thanos.makatos@nutanix.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-07-30 08:20:49 +00:00
John Levon
f1c8170632 vfio-user: correct poller return value
The transport poller is supposed to return the number of events handled to the
generic nvmf code; correct the vfio-user implementation so it does that.

Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: I804e1e8c75701c0b22ea6afd350b455c39908511
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8562
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>
2021-07-30 08:20:49 +00:00
John Levon
6ec5e75ae4 nvmf: fix keep alive poller
Return SPDK_POLLER_IDLE if we didn't disable any controllers.

Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: Ie16301c6e5f3f7e3f4b46906a95ef214674f4af8
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8560
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: Thanos Makatos <thanos.makatos@nutanix.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-07-30 08:20:49 +00:00
Changpeng Liu
09c0c9764a nvmf: return feature not changeable for Interrupt Coalescing feature
Windows will always sends a Set Feature Interrupt Coalescing even
SPDK reports we can't support it in Get Feature command.  Here
we return Feature Not Changeable instead of Invalid Field which
is more meaningful.

Change-Id: Ie08086c3eba1e2d790a7ae4976653b6f9085028c
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8923
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>
2021-07-29 07:04:27 +00:00
Ben Walker
ea0aaf5e85 nvme: Transports now set qpair state to NVME_QPAIR_CONNECTED inside
.ctrlr_connect_qpair

Previously this was assumed to be a synchronous process so the generic
layer transport code updated the state after .ctrlr_connect_qpair
returned. In preparation for making this support asynchronous mode,
shift that responsibility down into the individual transports.

While none of the transports actually do this asynchronously, insert a
busy wait in nvme_transport_ctrlr_connect_qpair to wait for the qpair to
exit from the CONNECTING state. None of the upper layer code can
actually correct handle a transport doing this asynchronously, so the
busy wait will cover that.

Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: I3c1a5c115264ffcb87e549765d891d796e0c81fe
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8909
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: Monica Kenguva <monica.kenguva@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-07-28 07:04:00 +00:00
Ziye Yang
50472c44cf idxd: Do not assign the results if there is error on completion.
If there is hardware issues, we do not need to assign
the result. Because we will report the error status to the uplayer.

Change-Id: I647ddd609a1d5d0d52cc4fee59699b9992da4fa4
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8864
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>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-07-28 07:02:22 +00:00
Ziye Yang
39827dd445 lib/idxd: Add assert check in each public API.
Those functions are exported publicly, so better to
add some assert functions to detect some null pointer
errors.

We do not use if/else check, because it is too heavy.

Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I16814efb84a5a41876657f0caf5f0a6d0c2db8f3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8863
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>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
2021-07-28 07:02:22 +00:00
Ziye Yang
4a670845f8 idxd: Add a check for the batches list.
This assert is used to make sure that there is no
active batch (spdk_accel_batch) task is used.

If there are active batches found, it means that
we did not handle this case in a good manner.

Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I7ea6247d2d5a40cf4f3f31cd8b1240fed4648d62
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8857
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: <dongx.yi@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-07-28 07:02:22 +00:00
Michal Mielewczyk
975852a079 lib/ocf_env: Force struct list 64 bytes alignment
After compiling SPDK with `--enable-ubsan` option, ocf tests fail with the
following error:

src/ocf/mngt/ocf_mngt_common.c:170:2: runtime error: member access within
misaligned address 0x200003800188 for type 'struct ocf_cache', which requires
64 byte alignment

The mentioned line of code is `list_for_each_entry()` macro used for iterating
lists. Forcing `struct list` alignment removes the issue.

Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Change-Id: I803dd962ff873679f42568e6f42fec7fed278f67
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8516
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Rafal Stefanowski <rafal.stefanowski@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2021-07-28 07:02:14 +00:00