spdk_blob_is_clone() should return true only for normal clones. To
detect esnap clones, use spdk_blob_is_esnap_clone(). This also clarifies
documentation of spdk_blob_is_esnap_clone() to match the implementation.
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Change-Id: I9993ab60c1a097531a46fb6760124a632f6857cd
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17544
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>
The health of clones of esnap clones depends on the health of the esnap
clone. This allows recursion through a chain of clones so that degraded
state propagates up from any back_bs_dev that is degraded.
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Change-Id: Iadd879d589f6ce4d0b654945db065d304b0c8357
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17517
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>
In preparation for supporting degraded lvols, spdk_blob_is_degraded() is
added. To support this, bs_dev gains an optional is_degraded() callback.
spdk_blob_is_degraded() returns false so long as no bs_dev that the blob
depends on is degraded. Depended upon bs_devs include the blobstore's
device and the blob's back_bs_dev.
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Change-Id: Ib02227f5735b00038ed30923813e1d5b57deb1ab
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17516
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>
While getting memory domains, vbdev_lvol will need to be able to access
the bdev that acts as the lvol's external snapshot. The introduction of
spdk_blob_get_esnap_bs_dev() facilitates this access.
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Change-Id: I604c957a468392d40b824c3d2afb00cbfe89cd21
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16429
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>
For operations that have differently sized input/output buffers (e.g.
compress, decompress), the size of the src buffer is recorded.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I1ee47a2e678ac1b5172ad3d8da6ab548e1aa3631
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17624
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@nvidia.com>
It is now possible to append an operation calculating crc32c to an accel
sequence. A crc32c operation needs special care when it's part of a
sequence, because it doesn't have a destination buffer. It means that
we can remove copy operations following crc32c only when it's possible
to change the dst buffer of the operation preceding crc32c.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I29204ce52d635162d2202136609f8f8f33db312d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17427
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
This will reduce the amount of changes in the following patch which
makes this function recursive.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: If8da6ae52d78358b66b2d9303413a9723687a767
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17568
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
g_stats_lock is an spdk_spin_lock that is initialized as the module is
loading. With this change, it is destroyed as the module finishes.
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Change-Id: I5263547f6d0e8981765d59665bd826cf07a6f83e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17681
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
This ensures that there are no more outstanding operations, so we can
safely free any global resources.
Fixes#2987
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Iac423b4f2a1183278d1db20f96c1a3b1bb657f85
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17767
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Mike Gerdts <mgerdts@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Issue #2983 shows a case where we seem to get a
device remove notification from DPDK (via vfio
path) after we have already detached the device
explicitly by SPDK.
This issue has proven difficult to reproduce
outside of the one observed failure so far, so
adding a couple of ERRLOGs into this path to help
confirm the this theory should it happen again.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I0fda4229fe150ca17417b227e8587cd7fbda6692
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17631
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
When using `__lsan_do_recoverable_leak_check` (e.g when fuzzing),
to check for leaks during runtime. Leak sanitizer can not follow
reference of memory that is allocated on heap (e.g. calloc)
and then stored on hugepage causing lsan to incorrectly report
direct leak.
Fixes#2967
Signed-off-by: Sebastian Brzezinka <sebastian.brzezinka@intel.com>
Change-Id: I3511e117a07ca8daa96f19bf1437c0d788b64cb1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17682
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Amir Haroush <amir.haroush@huawei.com>
This was arbitrarily picked as 2 seconds in commit
0e3dbd. But for extremely high connection count
use cases, such as nvme-perf with several cores
and high connection count per core, this 2 second
time window can get exceeded.
So increase this to 10 seconds, but only for qpairs
that are being connected asynchronously.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I906ca9e6561b778613c80b739a20bd72c807216c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17619
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
lvs_load() function verifies if options passed to it
are valid, but doesn't return, if they are not (only error
is logged and callback is called with -EINVAL code). Now
it is corrected and the function ends after the error
is reported.
Change-Id: I19b0b22466b6980345477f62084d27ef13414752
Signed-off-by: Marcin Spiewak <marcin.spiewak@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17582
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Mike Gerdts <mgerdts@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
When changing src/dst buffers, we copied memory domain pointers, but we
didn't copy memory domain context, which is obviously incorrect. It was
probably missed, because we never append a copy with non-NULL memory
domain. Added a unit test case to verify this behavior.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ic174e0e72c33d3f437f0faddd3405638049f0c74
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17425
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This file should be external to enable out-of-tree accel modules.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I2e973d0e88d7145d0fc9714f56db48486b00f3b7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17419
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jacek Kalwas <jacek.kalwas@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
It always returns 0 and any errors are reported in the callback. Making
it void simplifies error handling.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I0d4299a2789a688eae38d76de46d1baf27cbbd8f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17194
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
Deprecation notice for pmem was removed, as libreduce will
still use it until pmem is supported.
Change-Id: I7555dbf20a408a67fac8a6e7b2eaa23edf985eec
Signed-off-by: Marcin Spiewak <marcin.spiewak@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17538
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
To make sure we don't lose statistics of destroyed channels, they're now
added to a global stats structure when a channel is destroyed.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ic3b4d285b83267ac06fad1e83721c1b15cc8ec8a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17567
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
The RPC allows the user to retrieve accel framework's statistics.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I5cd1b45686504c08eda50513ad1dae2f8d65013b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17191
Community-CI: Mellanox Build Bot
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>
This patch adds support for collecting statistics in accel framework.
Currently, it counts the following events:
1. The number and the type of executed/failed operations.
2. The number of executed/failed accel sequences.
For now, these statistics are only collected and there's no way of
retrieving (or resetting) them - that will be added in the following
patches.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Id211067eb810e7b7d30c756a01b35eb5019c57e7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17190
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
It'll allow for setting accel-specific options. For now, it makes the
size of iobuf caches configurable.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Iaf505cc5e98dc6411453d9964250a4ba22267d79
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17188
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
In case SPDK is build with shared libraries and there's no
LD_LIBRARY_PATH around, linker will complain about missing .sos
similar to:
/usr/bin/ld.bfd: warning: librte_meter.so.23, needed by
/root/spdk/dpdk/build/lib/librte_ethdev.so, not found (try using -rpath
or -rpath-link)
We can't see that under CI since autotest_common.sh always makes sure
the LD_LIBRARY_PATH is properly set.
Add the -rpath to make the build less spammy.
Change-Id: I1d9d1775b2aa24e65cc4b776c2549457b0d7aac3
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17492
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: Ben Walker <benjamin.walker@intel.com>
In commit 23baa67, we will start virtio device only once,
and update the VQ's information in SET_VRING_KICK message
context, so when multi-queues are enabled, SPDK doesn't
register VQ's interrupt handler, here we add it when enable
VQ.
Fix issue #2940.
Change-Id: I29dbd7bf0b81b23c2e47e37c467952cc5887b5bf
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17354
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
bdev modules should have call spdk_bdev_io_complete twice
for the same IO. We can help find cases where this happens
by adding an assert in spdk_bdev_io_complete - confirming
that the current status is still PENDING, before changing
it to the status passed by the caller.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Id8a044a94113f1ac5e3c8d86e426654bfa8d5c5a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17330
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reset the status for a bdev_io that fails with NOMEM status
back to PENDING immediately when it is put on the nomem_io
list, instead of waiting until it gets submitted again.
This helps keep the bdev_io states consistent, so that if
we need to complete these IO for abort reasons later, that
we aren't completing IO that already have a non-PENDING
state.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I9532095141209ed6f7af362b52c689da62e755ce
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17335
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Mellanox Build Bot
Handling this callback is quite complex and may lead to
various problems. In most of places, the actual event
when qpair is dosconnected is not importnat for the
app logic. Only in shutdown path we need to be sure
that all qpairs are disconnected, it can be achieved
by checking poll_group::qpairs list
Signed-off-by: Alexey Marchuk <alexeymar@nvidia.com>
Change-Id: I453961299f67342c1193dc622685aefb46bfceb6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17165
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
If the qpair is already in the process of disconnect,
the spdk_nvmf_qpair_disconnect API now return -EINPROGRESS
and doesn't call the callback passed by the user.
Signed-off-by: Alexey Marchuk <alexeymar@nvidia.com>
Change-Id: If996b0496bf15729654d18771756b736e41812ae
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17164
Tested-by: SPDK CI Jenkins <sys_sgci@intel.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>
Current implementation of spdk_nvmf_qpair_disconnect
saves and calls user's callback correctly only on
the first call. If this function is called when
qpair is already in the process of disconnect, the
cb_fn is called immediately, that may lead to stack
overflow.
In most of places this function is called with
cb_fn = NULL, that means that the real qpair disconnect
is not important for the app logic. Only in several
places (nvmf tgt shutdown flow) that is important to
wait for all qpairs to be disconnected.
Taking into account complexity related to possible stack
overflow, do not pass the cb_fn to spdk_nvmf_qpair_disconnect.
Instead, wait until a list of qpairs is empty in shutdown path.
Next patches will change spdk_nvmf_qpair_disconnect behaviour
when disconnect is in progress and deprecate cb_fn and ctx
parameters.
Fixes issue #2765
Signed-off-by: Alexey Marchuk <alexeymar@nvidia.com>
Change-Id: Ie8d49c88cc009b774b45adab3e37c4dde4395549
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17163
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This patch removes usage of cb_fn argument of
spdk_nvmf_qpair_disconnect API. Instead of relying
on the callback, post a completion on delete SQ
command when transport qpair_fini is called.
Signed-off-by: Alexey Marchuk <alexeymar@nvidia.com>
Change-Id: I68dec97ea94e89f48a8667da82f88b5e24fc0d88
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17168
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Make sure UUID is present for every bdev, even ephemeral ones.
Furthermore, this change removes assumption that bdev UUID
may remain empty.
Change-Id: I924c1ba9dedfe88a05044bb1073f28085735b1c1
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17106
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Commit aaba5d introduced a build warning with some
compilers. While fixing it, I realized the function was
difficult to immediately understand. So in addition to fixing
the build warning, I also made the following changes:
* Improved names for local variables
* Use TAILQ_INIT for local TAILQ instead of TAILQ_HEAD_INITIALIZER.
* Add comments explaining more clearly what the nested loops are
doing.
* Use TAILQ_SWAP instead of a FOREACH + REMOVE + INSERT.
Fixes: aaba5d ("subsystem: Gather list changed conditions.")
Fixes issue #2978.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Ic8740b5706537938d62a0acfac62625b2424b85f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17496
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Mike Gerdts <mgerdts@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
rte_rcu is available on all versions of DPDK supported by SPDK. It is
also required by quite a few DPDK libraries. So just include
it always, it's a small library so let's not try to over-complicate by
trying to figure out exactly when it's needed.
This change fixes linking issue when crypto enabled (and vhost not).
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: Ibdd6acb5a25c401b462022bbd94bd380690640d0
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17514
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.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: Aleksey Marchuk <alexeymar@nvidia.com>
Fixed function is used to determine if it is possible to get iobuf
from the pool. To make sure that buf size alignment requirement is
satisifed value returned shall include alignment value but subtracted
by one.
e.g.
transaction size length = 64k
buffer alignment = 1 byte (no alignment requirement)
metadata length = 0
Without the fix the function returned 64k + 1, now it returns 64k
which is correct behavior and allows to proceed with further command
processing (if max buffer size limit is set to 64k only).
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I09104ad21b3652ba1aa5c3805a04b1c6549d04ac
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17513
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
In the presence of hardware offload (for data digest) we may not be
able to immediately release all PDUs to free a connection. Add a
state to wait for them to finish.
Fixes#2862
Change-Id: I5ecbdad394c0296af6f5c2310d7867dd9de154cb
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16637
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Fixes#2812
This patch adds support for '--lcores <map_list>'
parameter in spdk.
This parameter allow mapping of the lcores
to CPU IDs, if the system contains CPUs with IDs
greater or equal to 128 (RTE_MAX_LCORE). Such CPUs
can not be directly included in core mask
specified in '-m <mask>' parameter, as the dpdk
rejects cores if IDs are greater than 127.
The only way to use them in spdk is to map lcore
to CPU using --lcores parameters specified
in command line.
--lcores and -m parameters are mutually
exclusive. Please use only one of them.
Examples:
build/bin/nvmf_tgt --lcores 0@130
build/bin/nvmf_tgt --lcores 0@150,1@151
build/bin/nvmf_tgt --lcores "(5-7)@(10-12)"
build/bin/nvmf_tgt --lcores "(5-7)@(136,138,140)"
Change-Id: Ia92be4499c8daaa936b1a4357d52ae303d6f3eb1
Signed-off-by: Marcin Spiewak <marcin.spiewak@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17403
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
This patch adds support for lcore mapping list, which
is needed by spdk if someone wants to use CPUs with IDs
greater than RTE_MAX_LCORE (128). For such CPUs it
is impossible to include them in the core mask (passed
to dpdk as '-c <mask>') as the dpdk doesn't allow
IDs greater than RTE_MAX_LCORE. Therefore they
must be mapped to lower lcore values using
'--lcores <maping_list>' passed to dpdk
Change-Id: If68f15cef2bca9e42a3457bf35477793b58ec53d
Signed-off-by: Marcin Spiewak <marcin.spiewak@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17399
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
A host will consider a QID as reusable once it disconnects
from the target. But our target does not immediately
free the QID's bit from the ctrlr->qpair_mask - it waits
until after a message is sent to the ctrlr's thread.
So this opens up a small window where the host makes
a valid connection with a recently free QID, but the
target rejects it.
When this happens, we will now start a 100us poller, and
recheck again. This will give those messages time to
execute in this case, and avoid unnecessarily rejecting
the CONNECT command.
Tested with local patch that injects 10us delay before
clearing bit in qpair_mask, along with fused_ordering
test that allocates and frees qpair in quick succession.
Also tested with unit tests added in this patch.
Fixes issue #2955.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I850b895c29d86be9c5070a0e6126657e7a0578fe
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17362
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Just remove the duplicated Code and make the
Conditions for g_subsystems list to subsystems_list
together.
Signed-off-by: yidong0635 <dongx.yi@intel.com>
Change-Id: I011b550b83d32580bfd25130dab9e44bcbdc1daf
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/13753
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>
If user passes --no-huge as part of env_context, do
not add other huge-related options to the EAL command
line. Instead emit an error message and return failure, if
any of them were specified explicitly.
Fixes c833f6aa ("env_dpdk: unlink hugepages if shm_id is not specified")
Fixes issue #2973.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I7aa49e4af5f3c333fa1e7dec4e3f5b4b92e7d414
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17483
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
iSCSI connection closes its socket when it is terminated. After the
socket is closed, the connection cannot access to it. However, the iSCSI
fuzz test terminated a connection while processing a text command. The
connection aborted the text command and the corresponding completion
callback accessed the closed socket. This unexpected access caused a
NULL pointer access.
Add a check if conn->sock is not NULL to iscsi_conn_params_update()
to avoid such NULL pointer access. The return type of the most iSCSI
library functions are void. Here, it is enough not to return 0. Hence,
use -ENXIO simply to indicate there is no available socket.
Fixes the issue #2958
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I2c1f58a63ee0a40561a17f81d4b4264061f411f6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17353
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Michal Berger <michal.berger@intel.com>
When bs_create_blob() is creating the internal xattr for the esnap ID,
it errors out if the ID is too long. This error path neglected to set
the return value. It now returns -EINVAL in this case.
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Change-Id: I6d756da47f41fb554cd6782add63378e81735118
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17292
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: Ben Walker <benjamin.walker@intel.com>
Move cq_is_full() closer to its caller post_completion() and along with
fixing comments.
Signed-off-by: Swapnil Ingle <swapnil.ingle@nutanix.com>
Change-Id: I93262d1805f0f9f075c6946ed97cd3006ffba130
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16415
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
From the issue report in #2507
that comparing blob maybe be NULL.
So add assert, that in CI may catch this issue.
And other funtions add this also.
Signed-off-by: yidong0635 <dongx.yi@intel.com>
Change-Id: I98179ec76f2b6785b6921c37373204021c0669b6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12737
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>