NVME TCP driver supports up to 16 sge elements
while only 1 sge is reported - that leads to
unnecessary requests split which degrades perf.
Also pass correct iovcnt to nvme_tcp_build_iovs -
it should be 32. Otherwise, pdu header consumes
1 iov and data is written partially.
Add a check that at least data_len bytes were
appended to the socket iovs and fail request
otherwise.
Signed-off-by: Alexey Marchuk <alexeymar@nvidia.com>
Change-Id: Ie83c807dd3fec2c7e7cbcda1e493d6fd74ebe599
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17006
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Make it clear that number of entries might not be equal
to the number of recorded traces, as some of the latter
might occupy two entries due to their length.
Change-Id: I3099cfb719c38bdee48fbe20fccef3ef43e820a3
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16916
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
nvme_rdma_qpair_abort_reqs() and nvme_tcp_qpair_abort_reqs() did not
initialize cpl->sqid. Hence, unexpected message was printed by
spdk_nvme_print_completion(). Fix the bugs in this patch.
Fixes#2930
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I8b41166e58b26ce22c453ab85794b46dbe3dd3a2
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17067
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
nvme_ctrlr_disable_poll() continued to be called until it returned 0.
However, if the corresponding drive was unresponsive, the continuous
calls consumed CPU and affected other operations.
If the corresponding drive is unresponsive, we cannot complete disabling
the controller. Hence, call nvme_transport_ctrlr_disconnect_qpair_done()
if nvme_ctrlr_disable_poll() returned any value other than -EAGAIN.
Even before this patch, nvme_ctrlr_disable_poll() collected an error log
if it failed. Hence, we do not have to add more error logs.
Fixes issue #2931
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I26cabb94e5744e3a2d975670adbf2e4e48d5bd7a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17002
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
By the patch 736b9da034
nvme_qpair_abort_all_queued_reqs() was changed to be called after the
adminq is actually disconnected.
However, the patch ac31590b37
unexpectedly disabled to call nvme_qpair_abort_all_queued_reqs() for
adminq because qpair->active_proc is NULL for adminq.
Add one more condition to nvme_transport_ctrlr_disconnect_qpair_done().
Fixes issue #2928
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Ic65f4cd952e6e89275788ff4b86ceca050f624d5
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17001
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Add command dword 13 field to the extendable structure
spdk_nvme_ns_cmd_ext_io_opts. This now enables us to pass dspec
and dsm fields.
Signed-off-by: Ankit Kumar <ankit.kumar@samsung.com>
Change-Id: Id4d3dac14fdbf0e2a57e0bf287551dfd827dd503
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16945
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Community-CI: Mellanox Build Bot
Made cq_is_full() as wrapper around cq_free_slots()
Signed-off-by: Swapnil Ingle <swapnil.ingle@nutanix.com>
Change-Id: I392f62e959c7e23b4360e77759027ea55c2398b9
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16789
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: John Levon <levon@movementarian.org>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Linux host nvme driver processes all pending cqe's in one batch along with
completing backing blk_mq req's and later rings cq_doorbell once for all
processed cqes.
As blk_mq req's are completed there is room for more submissions
before ringing cq_doorbell.
This may race with vfio_user cq_is_full() which uses cq_doorbell to make final
decision and as host has not updated cq_doorbell we fail with cq_full error.
To mitigate this only process commands from sq which have free cq slot.
Signed-off-by: Swapnil Ingle <swapnil.ingle@nutanix.com>
Change-Id: I0cefb41df8099eb71de25923d05a9fcb28e4d124
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16788
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Add default copy command support in bdev layer for backing devices that
does not support copy command.
Signed-off-by: Rui Chang <rui.chang@arm.com>
Change-Id: I5632e25544e95ac0c53ff91c4cd135dac53323ae
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16638
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
When IBV_EVENT_DEVICE_FATAL & RDMA_CM_EVENT_DEVICE_REMOVAL occurs,
destroy qpair immediately and do no assume that no successful WQE will
be received after rdma_disconnect.
Signed-off-by: sijie.sun <sijie.sun@smartx.com>
Change-Id: I23e44dd32c8adea301e5251659b1be519f5dfdf7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16314
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Community-CI: Mellanox Build Bot
IB device may be unplugged & hotplugged when modifying slaves of bonding
IB devices. This patch will try to recreate ibv device contexts, poller
and listeners after IB devices come back.
Signed-off-by: sijie.sun <sijie.sun@smartx.com>
Change-Id: I3288174bad847edc2d9859cb34aa93c6af8c673b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15616
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
This is the beginning of support for external snapshots. An external
snapshot is a read-only blobstore device (struct spdk_bs_dev) that can
be used as a blob's back device. Normally a blob will have no back
device (a normal blob), a zeroes back device (a thin provisioned blob),
or a blob back device (a clone blob). When a blob has an external
snapshot ("esnap") as its back device, it is called an esnap clone.
With this patch, esnap clones can be created but they are not yet
useful. Subsequent patches in the series will plumb the IO path, enable
various features, and allow lvol bdevs to be esnap clones.
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Change-Id: I29206b628a2b03b6386a88532565e228df988e0e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14969
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Removed SPDK_IDXD_FLAG_PERSISTENT flag and associated code.
Change-Id: Ib4e038794792ae9866bdf344f1ec58dd04dbd483
Signed-off-by: Marcin Spiewak <marcin.spiewak@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16986
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This patch removes references to deprecated PMEM from accel library.
The code that was executed when ACCEL_FLAG_PERSISTENT flag is set,
is no longer needed and is removed.
_sw_accel_copy() function is removed and replaced with memcpy(), as
after PMEM removal its functionality is the same as memcpy().
_sw_accel_dualcast() is no longer needed, replaced with direct calls
to memcpy()
Removed 'flags' parameter - it is no longer needed
accel_ut.c: removed references to PMDK
deprecation.md updated
ACCEL_FLAG_PERSISTENT flag will be removed in next patch.
Change-Id: I86130466fe7a5f6ee547df1517b803035ff41a7a
Signed-off-by: Marcin Spiewak <marcin.spiewak@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16899
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
If a driver is registered and selected, it'll now be used to execute
sequences of accel operations. The driver has priority over accel
modules, so the modules will only be used to execute operations that the
driver cannot perform.
Once driver completes a task (or a number of tasks), it notifies accel
using standard spdk_accel_task_complete(). To let accel continue
processing a sequence, driver can call spdk_accel_sequence_continue().
This can be done when the driver executes all tasks (1), an error occurs
(2), or the driver doesn't know how to execute a given opcode (3). In
case of (3), that operation will be executed using appropriate accel
module and, while the rest of the sequence will be sent back to the
driver.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: If414c02073ffc731454e03d25c7ee02bef58463b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16548
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The following error was reported when running gpt_ut which is related
to crc32_update().
"load of misaligned address 0x001ffeff78cc for type 'const uint64_t',
which requires 8 byte alignment".
This patch preprocesses the first several bytes to make the buf address
passed to __crc32_d or__crc32_cd is 8 byte aligned. And finally process
the trailing bytes.
For function spdk_crc32c_update in crc32c.c, memcpy was used to avoid
misaligned load problem. Update it with above solution to reduce extra
overhead.
Signed-off-by: Richael Zhuang <richael.zhuang@arm.com>
Change-Id: I7c7aaa41e1c042a96668158818b06729fb3ceec6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16801
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>
Depending on the number of cores there are sporadic issues getting
elements of that pool although free elements are there during poll
group creation. Operation returns -ENOBUF. It results in odd notice
msg.
"nvmf_transport_poll_group_create: *NOTICE*: Unable to reserve the
full number of buffers for the pg buffer cache. Decrease the number of
cached buffers from 455 to 1366"
In this case 1366 is the actual number of available elements in the
pool. Few poll groups suceeds and few are ending up with the buffer
cache size set to 0.
Issue has been rootcaused as bug or behaviour change in DPDK v22.01.
Consider example:
We create DPDK mempool with 4K buffers, cache of 256. When first poll
group requests 512 buffers, DPDK mempool first looks in its per-core
cache, sees no buffers (mempool buffer cache doesn't get prepopulated)
and then requests 512 + 256 buffers from the backing pool. It returns
512 of the buffers to the user, and puts the other 256 buffers in the
cache ...it should only request 512 buffers total. For 8 cores and 512
buffers requested only 5 cores will get their buffers.
Disabling mempool cache seems to workaround the issue. More effective
cache is already implemented on nvmf generic layer.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I3149dea95a4f24a75dd0074eda9468c4856d901d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16913
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Add error check to unclaim mechanism. Issue #2920
showed that unclaiming CPU locks might fail and
we should catch errors to determine the cause.
Change-Id: Ifdfb7db2595d73f8bae13418ef145ad80e1d07ef
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16958
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>
Add implementation of uuid_generate_sha1() for systems
that do not have this function in their system libraries.
Use uuid_generate_sha1 from uuid.h inside a new function
spdk_uuid_generate_sha1(). The reason for this addition
is to prepare for UUID generation correction to conform
to standards.
First part of series addressing #2788.
Change-Id: Ib357aa1ee832e886288d176d8a47efdaa326f537
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16414
Reviewed-by: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Return the real spdk_fd_group object so it can later be nested.
Change-Id: I84c8a174c7d177799fa484b350269082c61b18a5
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15474
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: John Levon <levon@movementarian.org>
Don't assume spdk_thread_poll() will ever get called. Instead, send
a message to process the exit.
Change-Id: Idd98e7e8164c5efebd0d7c9287e62731e7cbc998
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15551
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: John Levon <levon@movementarian.org>
Move away from relying on spdk_thread_poll() to do clean up in interrupt
mode. In the future, we don't want to have spdk_thread_poll() called at
all.
Change-Id: I5318a7889601a3d3463e35419918b7305f68ee8d
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15550
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: John Levon <levon@movementarian.org>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Eventually, we want to allow merging of spdk_fd_groups, removing a level
of indirection. That means that some interrupt handlers won't
necessarily fire with the spdk_thread context already set. Set it in the
wrappers to ensure it's right.
Change-Id: Ief18d58cf3ee005c2969a9c0ee132b34b24cbd61
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15476
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: John Levon <levon@movementarian.org>
This unifies the poller fds with the interrupt mechanism internally.
Change-Id: I57a270260981ff54670365dddb33a1d9bdb56781
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15754
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Added initialization of the prev_crc variable to avoid compilation error:
idxd.c: In function ?spdk_idxd_submit_copy_crc32c?:
idxd.c:1138:51: error: ?prev_crc? may be used uninitialized [-Werror=maybe-uninitialized]
1138 | desc->crc32c.addr = prev_crc;
| ~~~~~~~~~~~~~~~~~~^~~~~~~~~~
idxd.c:1081:18: note: ?prev_crc? was declared here
1081 | uint64_t prev_crc;
| ^~~~~~~~
Change-Id: I6b93d5d85b52e20f8a2c313c41b740f66eebe1c7
Signed-off-by: Marcin Spiewak <marcin.spiewak@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16900
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Fix for number of dwords which is 0 based as per spec.
Use bitwise operators instead of division and modulus.
Change-Id: Ib315bf9394ef599317f41429742e7b8054069549
Signed-off-by: Ankit Kumar <ankit.kumar@samsung.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16814
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>
They aren't cleared before a task is submitted and might store pointers
from a previous operation. This can lead to issues if the previous
operation was using memory domains and we submit the task to a module
also supporting memory domains.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Icafb924c2e936ee6a83d921ae48e953b98f00841
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16848
Community-CI: Mellanox Build Bot
Reviewed-by: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Remove libuuid usage on FreeBSD and add dedicated implementation of
spdk_uuid API using functions from the standard library.
Fixes: #2878
Change-Id: Ie49ccb2842acad6064bffd789e4f64b7365b6e5c
Signed-off-by: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16558
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
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>
An example of async operation which can be handled on specific
transport layer could be creation of spdk thread followed by
a poller registration.
This change also aligns with transport destroy which is already
async operation.
Current transport create function is marked deprecated and is meant
for transports supporting sync create only to maintain backward
compatibility. Async version supports both create operations.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I1f5a477819e58f30983d26f81a1416bed1279ecf
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16463
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The spdk_bdev_ext_io_opts structure is used to pass extra options when
submitting a bdev IO request, without having to modify/add functions to
handle new options. Additionally, the structure has a size field to
allow adding new fields without breaking the ABI (and thus having to
bump up the major version of a library).
It is also a part of spdk_bdev_io and there are several reasons for
removing it from that structure:
1. The size field only makes sense in structures that are passed
through pointers. And spdk_bdev_ext_io_opts is indeed passed as a
pointer to spdk_bdev_{readv,writev}_blocks_ext(), however it is
also embedded in spdk_bdev_io (internal.ext_opts_copy), which is
also part of the API. It means that each time a new field is added
to spdk_bdev_ext_io_opts, the size of spdk_bdev_io will also
change, so we will need to bump the major version of libspdk_bdev
anyway, thus making spdk_bdev_ext_io_opts.size useless.
2. The size field also makes internal.ext_opts cumbersome to use, as
each time one of its fields is accessed, we need to check the size.
Currently the code doesn't do that, because all of the existing
spdk_bdev_ext_io_opts fields were present when this structure was
initially introduced, but we'd need to do check the size before
accessing any new fields.
3. spdk_bdev_ext_io_opts has a metadata field, while spdk_bdev_io
already has u.bdev.md_buf, which means that we store the same thing
in several different places in spdk_bdev_io (u.bdev.md_buf,
u.bdev.ext_opts->metadata, internal.ext_opts->metadata).
Therefore, this patch removes all references to spdk_bdev_ext_io_opts
from spdk_bdev_io and replaces them with fields (memory_domain,
memory_domain_ctx) that were missing in spdk_bdev_io. Unfortunately,
this change breaks the API and requires changes in bdev modules that
supported spdk_bdev_io.u.bdev.ext_opts.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I49b7524eb84d1d4d7f12b7ab025fec36da1ee01f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16773
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
TP4146 introduced support for two new IO commands,
IO management receive and send.
Signed-off-by: Ankit Kumar <ankit.kumar@samsung.com>
Change-Id: Iaf37310b84e278df043dcf71a0c2ef912c2fca8e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16520
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
TP4146 added support for 4 new log pages.
These are FDP configurations, reclaim unit handle usage,
FDP statistics and FDP events.
Updated the identify example file accordingly.
Signed-off-by: Ankit Kumar <ankit.kumar@samsung.com>
Change-Id: I5a20b728605257774d72bc184b50bc5008e142ea
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16518
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Format LBA size (FLBAS) is updated to have:
Bit 3:0 as least significant 4 bits for format index
Bit 6:5 as most significant 2 bits for format index
NVMe format command fields are updated accordingly.
Add a new helper function to fetch the correct format index.
Update examples and unit test files accordingly.
Signed-off-by: Ankit Kumar <ankit.kumar@samsung.com>
Change-Id: I2d6d9045b9d65ae91cb18843ca75b59cc27ed2f2
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16515
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
It isn't used in this function and the callers always pass NULL.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I07baa13a25b1e4e0b8832a093a53250392b10f10
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16682
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
This should help catching bugs when a failed sequence gets cleared its
failed state.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I9389a2610e94e766aaf4185445c36442c4d4a1f7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16545
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
This will allow a platform driver to allocate a buffer in case it cannot
execute the whole sequence and the destination buffer of the last
operation is a "virtual" accel buffer.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ia947cf553619828a170c5d0563b4c355d7b5ead5
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16377
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
This will allow drivers to check if a task is using buffers from accel
domain. This is just a helper, since the same can be achieved by
calling `spdk_memory_domain_get_first("SPDK_ACCEL_DMA_DEVICE")`, but
there's only a single accel domain and it is a bit special, so it makes
sense to have a dedicated helper function for getting it.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I07db7445ed9b109e66ecdbc0483a6a158a551070
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16376
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The goal of a platform driver is to execute chained accel operations in
the most efficient way possible. A driver is aware of the hardware
available on a platform and can execute several operations as a single
one. For instance, if we want to do DMA and then encrypt the data, the
driver can do both at the same time, if the hardware is capable of doing
that.
Platform drivers aren't required to support all operations. If a given
operation cannot be executed, the driver should notify accel to continue
processing a sequence, via spdk_accel_sequence_continue(), and that
operation will processed by a module assigned to its opcode.
It is required however, that all platform drivers support memory
domains, including the "virtual" accel domain. A method for allocating
those buffers will be added in the following patches.
This patch only adds methods to register and select platorm drivers, but
doesn't change the way a sequnce is executed (i.e. it doesn't use the
driver to execute it).
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I97a0b07e264601ab3cf980735319fe8cea54d38e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16375
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
If req->data is set, with all the previous changes, then req->iovcnt
should also be more than zero.
Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: I29b5f45541c9dba2dd896109dd43d2b5321ec467
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16274
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Also deprecate the existing spdk_nvmf_request_data() API, which is
incompatible with iovecs.
Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: I44df8ff30a431873a0c2f34b0cdb58df858fd7e3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16200
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Use req->iov instead of req->data in reservation handling code.
Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: I6d79711d03f45bd5e118c6324d22decad887a788
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16199
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
