Put the locks inside cleanup_pci_devices().
This serves as cleanup.
Change-Id: I040b28006e5584d1f33af26b63cafedbafe04fdb
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458934
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: John Kariuki <John.K.Kariuki@intel.com>
The global pci tailq is no longer modified on the dpdk
thread, so on the spdk thread we can access it safely
without any lock. The code is slightly more readable
then.
This shows that cleanup_pci_devices() is always wrapped
with lock/unlock. We'll put the locks inside this
function in the next patch.
Change-Id: Ia4d386b78a87078761df0a3b953bfc4ff44102f8
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458933
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
To safely access the global pci device list on an spdk
thread, we'll need not to modify this list on any other
thread. When device gets hotplugged on a dpdk thread,
it will be now inserted into a new global tailq that
can be accessed only under g_pci_mutex. Then any
subsequently called public pci function will add it to
the regular device tailq.
Change-Id: I9cb9d6b24fd731641fd764d0da71bedab38824c9
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458932
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
To safely access the global pci device list on an spdk
thread, we'll need not to modify this list on any other
thread. When device gets hotremoved on a dpdk thread,
it will now set a new per-device `removed` flag. Then
any subsequently called public pci function will remove
it from the list.
Change-Id: I0f16237617e0bea75b322ab402407780616424c3
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458931
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
For VMD driver we'll need to introduce some way of
iterating over all spdk pci device objects and we would
like to achieve that with simple spdk_pci_get_first_dev()/get_next_dev()
APIs. To make it thread safe though, we would have to
expose some public pci mutex to be locked around the
iteration and we don't want to do that, so we'll make
PCI APIs usable from only a single thread - this will
prevent any pci devices from being removed inbetween
subsequent get_first/get_next calls.
We currently have the following players accessing pci
device state:
1) public APIs, obviously (on any thread right now)
2) VFIO hotremove callback (dpdk interrupt thread)
3) rte_eal_alarm for detaching rte_pci_devices (dpdk
interrupt thread)
4) DPDK hotplug IPC (dpdk interrupt thread)
There is g_pci_mutex providing the thread safety, but
even today it doesn't protect #3 and #4, making the
entire pci layer prone to data corruption.
To make #3 and #4 safe, we would have to lock inside
device init/fini callbacks (spdk_pci_device_init/fini),
but those are called directly inside the public device
attach/detach functions which already lock.
So now, with the decision to drop thread safety from
public pci APIs, we narrow down the locks inside public
functions and introduce locks inside those lower-level
init/fini callbacks.
Change-Id: I5dcbc9cdcbab65ee76cd3c42890f596069ec9a8a
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458930
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Converting to the struct virtio_req is useless.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I141268314d28cf87bdef529808c8e18bd1b41c9d
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/459360
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Currently we are missing cleanup routine for case when
power failure interrupts creating snapshot. This patch
add such routine.
For the case where we find blob with a parent snapshot ID
matching newly created snapshot we can finish whole process
during recovery by processing forward with setting snpashot
as read only, removing xattr and syncing. We should remove
snapshot only if there is no blob with parent pointing at
snapshot.
Fixes github issue #760
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: I2f0e298164e07a2b4dfa5367e8878facef640702
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455216
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
By default QAT will be selected if available however a new RPC
can be used to either auto-select (default) or specify either
ISAL or QAT.
Change-Id: I37cf7640bbd8cef455583e1eccb8adb59cc419d8
Signed-off-by: paul luse <paul.e.luse@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456693
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Fixed issue when restoring from a dirty shutdown - sometimes end md
wasn't erased after a band was prepared from writing when a shutdown
happened. This resulted in inconsistency between the new head md and
old tail md, which was technically valid. Band sequence numbers would
then be reused, causing a failure on any subsequent restore.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: Ic3e968be02bb814d6c85f0a3279403fe99337b86
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/459287
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>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Changed to use 4k alignment in dirty shutdown I/Os. Otherwise the
scatter gather lists used in QEMU for underlying file/block device
would use an extra entry (e.g. 17 for 16 sector writes), and eventually
some I/Os would write to offset 0 in underlying file, corrupting head
metadata.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: If8c88ce708529b094a09c8ee952912cc22cd53b9
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458090
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Band's lba_map needs to be set to NULL before restore completes, as
it's not allocated on a per band basis and instead uses a pool from
restore struct itself. Without the fix initializing a band for writing
would hit an assert during proper allocation in ftl_band_alloc_lba_map.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: Icff4f54cbe722cb6030b9dfd55726b9b0d6c1e27
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458422
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This patch makes spdk_dif_set_md_interleave_iovs() and
spdk_dif_generate_stream() process unaligned start of data segment
properly by using ctx->data_offset.
Separating this patch into two may be required but this patch is
small and aggregating into a patch is good to test.
UT code demonstrates how it is realized.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Idb5250aba4e12a34102e5ce067d725c685681177
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458142
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
To process unaligned data segment properly when a whole data buffer
is splitted into multiple data segments and each data segment has
any alignment, we have to update only data offset of DIF context
according to the progress.
Hence this patch adds an new API spdk_dif_ctx_set_data_offset().
The API will be used in the next patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I346ab583518b80792ea40d34cf0c8536ecc3d904
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458141
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
There're outstanding requests in spdk_jsonrpc_parse_request which caused by
connection close.
There are methods to call spdk_jsonrpc_server_conn_close, including
spdk_jsonrpc_server_conn_remove and spdk_jsonrpc_server_shutdown,
Some rpc methods call these functions to terminate connections ,that leads to
memory leaks.
Try to free outstanding requests after deciding to terminate a connection.
And do this follwing with close(conn->sockfd).
Fix issue #784, and can resolve other similar memory leaks about this.
Signed-off-by: yidong0635 <dongx.yi@intel.com>
Change-Id: Icd287bd0c5670ee8ec32750b999f82b0fa89cf84
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458438
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
By now (5.1 is released), the Linux kernel initiator supports the
success optimization and further, the version that doesn't support
it (5.0) was EOL-ed. As such, lets open it up @ spdk by default.
Doing so provides a notable performance improvement: running perf with
iodepth of 64, randread, two threads and block size of 512 bytes for 60s
("-q 64 -w randread -o 512 -c 0x5000 -t 60") over the VMA socket acceleration
library and null backing store, we got 730K IOPS with the success
optimization vs 550K without it.
IOPS MiB/s Average min max
549274.10 268.20 232.99 93.23 3256354.96
728117.57 355.53 175.76 85.93 14632.16
To allow for interop with older kernel initiators, we added
a config knob under which the success optimization can be
enabled or disabled.
Change-Id: Ia4c79f607f82c3563523ae3e07a67eac95b56dbb
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457644
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
If users set the persist through power loss configuation file,
that means the Namespace has the capability to support ptpl
feature, here we added a ptpl_activated flag to indicate that
the users enable the feature or not. Users can use Set features
or Reservation Register commands to change the value.
Change-Id: Iae3fd44085c5be5bf9574e49efa567e8212dee20
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455906
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Vhost testing crashed from Nightly testing, because a member
access within null pointer of type 'struct ibv_send_wr'.
Change-Id: If8f34f23864883ea73516d2d1fe3b30137c04316
Signed-off-by: Hailiang Wang <hailiangx.e.wang@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458913
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
When JSON config was used, app layer was calling the
app start callback twice - once from internally-sent
"start_subsystem_init" RPC, and once from the app layer
itself.
In case of JSON configs, the callback from within the
RPC was actually called prematurely, as the real RPC
server was still starting in the background at that
point. We still need to start the app from that RPC in
case of `--wait-for-listen` option, but for JSON configs
it doesn't make sense. Just ignore it now and rely on
json config load completion callback to start the app.
Fixes#816
Change-Id: Ib54d624f3167137216c910b2d947bbd1dc5023b1
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458351
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
If reading the JSON config file has failed, we entered
spdk_app_json_config_load_done(-ERRNO) and tried to
close a client connection that was never initiated,
which resulted in NULL dereference.
To fix it, just check if client_conn != NULL before
attempting to close it.
Change-Id: I7340567c45e795f77110c2914e94ba83fa8d1bff
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458350
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Added new VMD subsystem to enumerate
devices behind VMD when event framewrok
is used.
To enable VMD, user need to provide Enable
flag via config file.
Change-Id: I89bfe22b127c00d358dac7336ffb44b0c0f426ea
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458443
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
NVMe/TCP target may split a whole data payload into multiple H2C
or C2H PDUs with any alignment. Hence to insert or strip DIF correctly
to the split H2C or C2H PDUs, we have to bring the interim guard
value of the last partial data block of the current H2C or C2H
PDU to the first partial data block of the next H2C or C2H PDU.
So we add last_guard to struct spdk_dif_ctx and use it in
spdk_dif_generate_stream().
API spdk_dif_generate_stream() is not changed and UT code should
pass without any change.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I12636c5ac7f619483402538faff4339a16c0e6b0
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457545
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
For NVMe/TCP target, data segments which correspond to H2C or C2H PDU
will have any alignment, and _dif_generate_split will have to process
partial data block, particularly the following types:
- start and end are both within a data block.
- start is within a data block, and end is at the end of a block
According to the refactoring done in the last patch, this patch
exposes offset_in_block, data_len, and guard as parameters of
_dif_generate_split() and make _dif_generate_split() process
the above two types of data block properly.
The next patch will utilize the updated _dif_generate_split in
spdk_dif_generate_stream().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I4211e65ead7fc256a40748412c670e46f83b1731
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457544
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
For NVMe/TCP target, data segments which correspond to H2C or C2H PDU
will have any alignment, and _dif_generate_split will have to process
partial data data block, particularly the following types:
- start and end are both within a data block.
- start is within a data block, and end is at the end of a block
On the other hand, _dif_generate_split had assumed that passed
block is always a complete block.
To process the above types, separating guard computation, DIF generation,
and DIF copy into three parts will be helpful and is done in this
patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I0171d9021837b9a4b425370293cef45dbe7500e8
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458225
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Four variables head_unalign, tail_unalign, num_blocks, and offset_blocks
became unuseful by the last patch. Hence reduce them to buf_len and
buf_offset in this patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I04fc1e442be6569a96533cdfe36b27fcc78e98d4
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457876
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
For NVMe/TCP target, data segments which correspond to H2C or C2H PDU
will have any alignment. spdk_dif_set_md_interleave_iovs() have allowed
reading data to have any alignment but had required data segment to
be a multiple of block size.
In other words, spdk_dif_set_md_interleave_iovs() had required that both
ctx->data_offset and (data_offset + data_len) must be a multiple of the
data block size.
This patch refines the algorithm to remove the latter requirement.
The update implies that spdk_dif_set_md_interleave_iovs support any
data buffer whose size is less than a single data block.
The update doesn't change parameters of spdk_dif_set_md_interleave_iovs
and existing UT should be passed.
This patch adds additional UT code to test these updates.
Change-Id: I88c7d2a80a8d92b54863b6ad1c3a9d2761a6195d
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457542
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
The subsequent patches will refine spdk_dif_set_md_interleave_iovs
to change the data_len parameter to be the remaining length, and
to remove alignment constraint completely. This patch is a
preparation to subsequent patches.
This patch doesn't change any behavior.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I0e9be7e66d313f3ec2bd8c55cce8bb18e4fff892
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457721
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
There is one existing example usage case to describe the issue:
Users(e.g. Vhost-blk target with Windows Guest) call spdk_bdev_readv_blocks()
to submit a 128KiB length data READ request, and the data buffer
provides by vhost isn't aligned, but the backend block device
requires aligned data buffer, so existing function call trace:
spdk_bdev_readv_blocks()-->
spdk_bdev_io_submit()-->
spdk_bdev_io_get_buf()
spdk_bdev_io_get_buf() will allocate buffer from large data
buffer pool for 128KiB length, of course, it will return error
with existing logic.
So here, no matter what the data length is, we can go through
the split process first for both READ and WRITE.
However, there is one scenario that for iSCSI READ request,
the iSCSI layer will not allocate data buffer for the request,
so for this case if the IO boundary is required we should keep
the logic as before.
Change-Id: I67661f5fa4c3c7c561b45c86146759aa3477adbf
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/453133
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
The completion cb of outstanding_tr may submit new requeset to
the outstanding_tr list of the qpair, it's an endless loop.
We only abort the remaining outstanding trackers.
Fix#819
Change-Id: I342f52f4d1836f8ef620ef9e3add0b1986727282
Signed-off-by: JinYu <jin.yu@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457755
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
The next patch in this series introduces and RPC that makes the
selection more flexible.
Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: I55617d3c37b51cf9474c358cb92f5218397c0c58
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457157
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
There were some TODO items in here that were incorrect if using
an underlying block device that had a 4K block size. The values
here will set the compress vol LBA size to match that of the
underlying lvol but the backing IO unit size, the min size that
reduce will use to perform backing device IO is set to 4K as
smaller than this makes little sense given chunk size is currently
at 16K.
Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: Ic6a9421a25d947ae7d4dce190d6599eec1d4a05a
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458703
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Module only supports shareable transforms, also adjusted the number
to what we need, just 2.
Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: Ieb77b5f1221878ac7a4be5176ed459d5165f8430
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458698
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
In _spdk_bs_snapshot_origblob_sync_cpl function on error
path we should not close snapshot as it will be closed during
volume closing when bs_dev is being destroyed.
This issue was found in unit test (see next patch in series).
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: I51c38d1f1f97b134679251b43109b1265e565a17
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455215
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
This bug was found by code inspection. When PDU read restarts
after data segment, iov_offset must be reduced by data length.
However iov_offset had been reduced by buffer length by mistake.
Lack of UT code was the main reason of this bug. Hence add UT
code to test the fix of the bug together in this patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I74f2f6ae8dca2e78a64bfdef8080c8031dfabb87
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458530
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
If bdev is configured to use separate buffers for metadata transfer,
but it doesn't support the write zeroes command, pass empty zeroed
buffer to make sure the metadata is cleared out as well.
Change-Id: If6f024266067e5764a28a276296f651d31da4792
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457628
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
The function spdk_bdev_desc_get_bdev() was added,
but remained unused. Couple places done the exact same thing,
so now they just use it directly.
Change-Id: I018f57b7b4bcfe22ec6a97cf5261b9a257ba93de
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456957
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Allow requests using separate buffer for metadata IO if base bdev
supports it. Copy base bdev's metadata configuration (separate /
interleaved, size, DIF type and location).
Change-Id: I89996f89cad805e696cacf0bba486b4cc1f617e6
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458407
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Starting with this patch it is possible to remove a
snapshot if there is only a one clone created from it.
In such case snapshot can be removed without any data
copying. This is achieved with following steps (case
with only one clone):
1. Open snapshot (Snapshot1) that shall be removed
2. Check if the Snapshot1 has no more than 1 clone (Clone1)
3. Remove Clone1 entry from Snapshot1
4. If the Snapshot1 has a parent snapshot (Snapshot2):
4a. Add Clone1 entry to the Snapshot2 clones list
4b. Remove Snapshot1 entry from Snapshot2 clones list
5. Open Clone1 blob
6. Freeze I/O operations on Clone1
7. Temporarily override md_ro flag for Snapshot1 and Clone1
for MD modification
8. Merge Snapshot1 and Clone1 clusters maps into Clone1
clusters map
9a. If Snapshot2 is present switch parent ID and backing
bs_dev on Clone1
9b. If Snapshot2 is not present set parent ID to
SPDK_BLOBID_INVALID and backing bs_dev to zeroes_dev
10. Sync MD on Clone1
11. Sync MD on Snapshot1
12. Restore MD flags for Clone1 and Snapshot1
13. Unfreeze I/O on Clone1
14. Close Clone1 blob
15. Remove Snapshot1
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: I800724b981af894e01e1912d0077c5b34a2ae634
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/445576
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Adds fields to structure spdk_nvme_io_qpair_opts.
These fields allow specifying the locations of memory buffers used
for the submission and/or completion queues.
By default, vaddr is set to NULL meaning SPDK will allocate the memory to be used.
If vaddr is NULL then paddr must be set to 0.
If vaddr is non-NULL, and paddr is zero, SPDK derives the physical
address for the NVMe device, in this case the memory must be registered.
If a paddr value is non-zero, SPDK uses the vaddr and paddr as passed.
SPDK assumes that the memory passed is both virtually and physically
contiguous.
If these fields are used, SPDK will NOT impose any restriction
on the number of elements in the queues.
The buffer sizes are in number of bytes, and are used to confirm
that the buffers are large enough to contain the appropriate queue.
These fields are only used by PCIe attached NVMe devices. They
are presently ignored for other transports.
Signed-off-by: James Bergsten <jamesx.bergsten@intel.com>
Change-Id: Ibfab3939eefe48109335f43a1167082dd4865e7c
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/454074
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
SQSIZE parameter validation in Connect command was broken because QID
field in qpair was used before intialization.
Signed-off-by: Evgeniy Kochetov <evgeniik@mellanox.com>
Change-Id: I8a0b359937d661df3b9888e6084e7d0b4a9056ea
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455667
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This patch changes the meaning of the data_len parameter of
spdk_dif_set_md_interleave_iovs from `Expected data length of the payload`
to `Expected length of the newly read data in the extended LBA payload`.
This change will align the parameters of spdk_dif_set_md_intereleave_iovs
to of spdk_dif_generate_stream.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I7f9c45e78be977625713acb79d2ae82d4375f419
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457543
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
As same as spdk_dif_set_md_interleave_iovs, no noticeable performance
difference was observed even if dif_generate_stream and
dif_generate_stream_split are merged into the latter.
spdk_dif_generate_stream will be updated in subsequent patches anyway.
Hence merge them into dif_generate_stream_split.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I77b588527bc7533dc38a36e450972758ce57d6bb
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457760
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: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Even if the code for single iovec case is removed, no performance
difference was observed when inserting/striping DIF for iSCSI target.
And the subsequent patches will add more complex operation and hence
simplifying the code before them will be valuable.
So this patch merges dif_set_md_interleave_iovs and
dif_set_md_interleave_iovs_split into the latter.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I784b9e04f0983ed91c7aa66bbc7ceaed4406e3ac
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457758
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: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
When setting the offset into pmem for the chunk mamps, the wrong
value was used for the length of the logical map resulting in
map corruption after extended fio run.
Fixes issue #813
Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: Icb51cf8f38a04f05f69595ebd510ced7f1b608f3
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457818
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Prints key info about a volume and its PMEM. As the metadata is
mapped, there's no real reason to dump all this to a file, it
can be accessed via gdb using the addresses printed out.
Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: I2c7b8c32b6142ba0fded623a660b4d7d0c5b19bc
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/458117
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reported by clang:
rte_vhost_compat.c:114:36: error: taking address
of packed member 'payload' of class or structure
'vhost_user_msg' may result in an unaligned
pointer value.
To fix it, just remove the extra unaligned pointer
and inline all its accesses.
Change-Id: I7e4ab536b87ab02a4ea12c55d55a6e495c3091ca
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457559
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
In several error cases, the status_class and status_detail
is not properly set.
Change-Id: Ie9cf8ec13b971d7295862872e8c7e834d08e7f14
Signed-off-by: GangCao <gang.cao@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457932
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
We shoud not always return SPDK_ISCSI_CONNECTION_FATAL
in spdk_iscsi_read_pdu function.
Reason: In iscsi_conn_handle_incoming_pdus, the loop
directly return only rc==SPDK_ISCSI_CONNECTION_FATAL.
But it masks all the necessary information. So we would like
to keep some information of the return value for spdk_iscsi_read_pdu,
and we can use error log to track those information.
Then we can return SPDK_ISCSI_CONNECTION_FATAL as the error
return value for iscsi_conn_handle_incoming_pdus function.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I4c40fcb27052b55cb92e06273701a881def18e12
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457078
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
NVMe bdev module support separate metadata now but NVMf subsystem
cannot process bdev with separate metadata yet.
Hence reject any bdev with separate metadata to be attached
explicitly by this patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I793c6c5f61deb766d7bf427ff67ccc57a48974cf
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457167
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
_spdk_bdev_io_split() will be used to requests that have IO buffers.
Change-Id: I5ba0b83e47b84066e6c5032de2e02404155896d3
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457549
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
struct ether_addr was renamed to struct rte_ether_addr
in latest DPDK master, but our internal fork of rte_vhost
still used the old name, which can be now a non-defined type.
Together with the struct, the RTE_ETHER_ADDR_LEN define
was renamed as well, so we'll now check if it's defined and
we'll manually define struct ether_addr to keep the old
rte_vhost working.
Change-Id: I78b8104ed3bfe03397881a94f0f8bee14f9efae8
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457609
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
For reservation feature in NVMoF, we can't support the persist through
power loss feature, now we will add the configuration file parameter
with Namespace, after users set the configuration file parameter with
one NS, then the PTPL feature can be enabled.
Change-Id: Id72699093f7e68318b9529f7bacc5c9804f7f86b
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455905
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This patch uses the change by the last patch to initialize DIF
context in SCSI layer. Besides this patch changes the name of a
parameter from offset to data_offset to clarify the meaning.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I54bf1168ec5959432aa15dae0360c0640138b033
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457541
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Data offset are intended to correspond to DATAO in NVMe/TCP and
Buffer Offset in iSCSI.
Previously for iSCSI, buffer offset had been merged to start block
address, but passing buffer offset separately from start block address
clarifies the logic more.
On the other hand, for NVMe/TCP, passing DATAO separately from start
block address will be critically important because DATAO will bave any
alignment and will be necessary to use for not only reference tag
but also guard computation.
This patch adds data_offset to struct spdk_dif_ctx and adds it to the
parameters of spdk_dif_ctx_init(). ref_tag_offset is also added to struct
spdk_dif_ctx and it is computed by dividing data_offset by data_block_size
and is used to compute reference tag.
The next patch will use this change when getting DIF context in SCSI.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Id0e12ca9b1dc75d0589787520feb0c2ee0f844a5
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457540
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
In lib/ftl PAGE_SIZE is used in many places. But I think there's no relation with
the host memory PAGE_SIZE.
In most changes, PAGE_SIZE is replaced by FTL_BLOCK_SIZE for ftl block size.
In ftl/ftl-anm.c, PAGE_SIZE is replaced by 4096 for alignment.
Change-Id: I72e7c65c83b9fe1a4e50944dc5d90b9459b4e593
Signed-off-by: Richael Zhuang <richael.zhuang@arm.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455347
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Unsetting this flag will decrease the number of WRs retrieved during CQ polling and will decrease
the oeverall processing time. Since RDMA_WRITE operations are always paired with RDMA_SEND (response),
it is possible to track the number of outstanding WRs relying on the completed response WR.
Completed WRs of type RDMA_WR_TYPE_DATA are now always RDMA_READ operations.
The patch shows %2 better peformance for read operations on x86 machine. The performance was measured using perf with the following parameters:
-q 16 -o 4096 -w read -t 300 -c 2
with nvme null device, each measurement was done 4 times
avg IOPS (with patch): 865861.71
avg IOPS (master): 847958.77
avg latency (with patch): 18.46 [us]
avg latency (master): 18.85 [us]
Change-Id: Ifd3329fbd0e45dd5f27213b36b9444308660fc8b
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Signed-off-by: Sasha Kotchubievsky <sashakot@mellanox.com>
Signed-off-by: Evgenii Kochetov <evgeniik@mellanox.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456469
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Seth Howell <seth.howell5141@gmail.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Vhost-scsi target will not set task's initiator port parameter,
so reservation commands sent from Guest will cause segment fault.
Change-Id: Ifc175effded5371eca08d5bfe7e4aa977dd4b1c6
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457550
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Added basic configuration details (transport type and address,
parallel units, cache info) as well as most important OCSSD geometry
data to get_bdevs driver_specific section.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: I00c50706bd9203bcef1701be1b7d87e93c10e57f
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456790
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Read requested range of lba map during relocation
instead whole lba map.
To read partial lba map range single move now has
three states: read, read_lba_map and write.
Free queue and write queue were merged to a single
move queue.
Change-Id: I86839f2286d42d4debf87cea40091370e5283b15
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/454747
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Keep track of read lba map segments to avoid
unnecessary traffic in case ANM events.
Lba map is divided on 4KB segments which can
store 512 lba entries.
In case multipe read request on same segment
keep pending request list and process it in
read_lba_map() completion callback.
Change-Id: I2661bdb716ab7c975140e0b37aebcb17aa23901d
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/453371
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
spdk_vmd_init() will attach all VMD devices that were
unbinded from system. There is not need to specify VMD
bdf in VMD public interface since it can be controlled
by setup.sh script.
Change-Id: Ifc45c32dc7e11b59429a41ddfdd596db30e27731
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456631
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
For the nvme device, I/Os are completed asynchronously. So we
need to check the outstanding I/Os before putting IO channel
when we hot remove the device. We should be sure that all the
I/Os have been completed when we change the sgroup->state to
PAUSED, so that we can update the subsystem.
Fix#615#755
Change-Id: I0f727a7bd0734fa9be1193e1f574892ab3e68b55
Signed-off-by: JinYu <jin.yu@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/452038
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
rte_vhost_vring_call() from upstream DPDK can read some
unitialized memory and crash if it's called on invalid
queue ids. The implementation in our internal rte_vhost
fork ends up wiritng to a random descriptor number, which
doesn't cause any crashes but is a bug nevertheless.
To fix it, just check if the queue is initialized before
interrupting it during the session start. It's not a hot
I/O path and there's no performance impact.
Change-Id: I830c1be98ef00d4ece9a6bd88cf79b9dfe29d2a9
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457247
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
We had "./configure --with-uring" , but actually, it wasn't compiled.
we need add uring to Makefile that code will add this.
then we can continue to verify this module.
Change-Id: I8b98825f6795eb9f9e7b4947d1c7c3a44a6f0f64
Signed-off-by: yidong0635 <dongx.yi@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457081
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Previously the connection was scheduled to another core just
after setting conn->full_feature to 1, but conn->state was still
ISCSI_CONN_STATE_INVALID. The connection started to run on another
core but could not run normally until conn->state became
ISCSI_CONN_STATE_RUNNING. conn->state was changed late to
ISCSI_CONN_STATE_RUNNING after sending the first login reponse.
This gap window had caused intermittent critical failures.
Based on this analysis, this patch changes to call
spdk_iscsi_conn_schedule() just after sending the final login response.
Whether any login response is final or not can be known by checking
conn->full_feature is not zero.
Fixes#785
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I8ea55fef27e2f332fcd789d32daf479a24c0588d
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457414
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Actually write/writev/write_zeroes/unmap are never be called, and we add
the error code here to keep it same style with snapshot bs_bdev.
Change-Id: I32ad051c1902bd7080b894e36f7c89f1c8d27434
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456924
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Gives the ability to change behavior of restore after dirty shutdown
without recompiling ftl library. User can define if partial recovery
or error should be returned after such a scenario.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: I6dda40df7b92d6a377957e4a70a3eab91a6ac4a9
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456185
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
If a band wasn't closed during shutdown (ie. after dirty shutdown),
the start md contains valid data, however end md does not (or it may
not be written at all).
The config gives the user the ability to specify if encountering
this case should result in an error, or if ftl should pad the band
with data and recover from closed bands only.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: If02f7ca8bc90bb61698fb710fee2274af6af01e4
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455513
Reviewed-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
If an open band is found during recovery (after dirty shutdown)
ftl will attempt to pad it, instead of returning a recovery error.
Any previously written data on that band is lost, as the line is
treated as being in free state after the pad.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: Ie48fa9fa37a3853f41921b9417c69301cf47e673
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455512
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The patch also changes, so the command calculates the needed
chunk offset based on PPA, instead of relying on the user to
provide one.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: Ic4eec1b86ded4eb71de860015403294ed0c8c266
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455973
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Direct wptr allows for writing to band with externally
provided PPA, rather than one assigned by the wptr. This
can be useful when padding open bands during recovery or
bdev creation, as ftl can find partially written end meta
with no way to close the band with valid data, that would
conform to an already written CRC value.
This does mean that I/O sender is responsible for calculating
and obeying per-chunk write pointer arithmetics.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: Ifcbc4e8ffb69c829469a571bed8e27b148d7a651
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455511
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Almost all I/O callbacks in ftl utilize data in ftl_io, which is
initialized as callback argument in ftl_io_init_internal. This patch
changes the behavior to always returning the ftl_io struct in addition
to an extra opaque buffer.
Signed-off-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Change-Id: I611ab1b33575f599798a2bb65c231a724c852c7f
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455831
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Improve handling of partial IOs for case when iovcnt == 1.
This case is pretty frqeuent because all data allocated by OCF
have iovcnt == 1 (see ctx->data->alloc() or vbdev_ocf_ctx_data_alloc())
Change-Id: I8d38ca9e9d2bd3e6ce298bf788ce6ed782b56594
Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456202
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This patch fixes submission of partial IOs in bottom adapter
Existing check (if offset > 0) was not sufficient to detect partial IOs
because there could be an IO with offset = 0 but length < total size of iovs.
This patch changes the check, but also free operation on completion
because now the old free does not cover all situations when we allocate iovs.
`Partial IOs` are the IOs handled by bottom adapter
which specify only part of the internal iovs vector.
So their length is less that the length of internal iovs vector.
They exist because sometimes parts of single IO from top adapter
need to be sent to different locations.
Also, in general, IOs initiated by OCF (such as cleaner IOs) are represented
as single big iov that is submitted by parts in form of 'Partial IOs'.
Change-Id: I8ae47659fb34904c593a696d74d683a418ac9962
Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455821
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Implement metadata probe functionality to load cache state
from disk.
During metadata probe, we inspect UUIDs of core devices
and create vbdev configurations based on them.
Then, to start vbdev, we use load path (loadq = true).
After this change persistent metadata is officially supported,
we can save and restore cache state from persistant storage.
WriteBack mode is now safe to use in respect to unexpected shutdowns,
because all information about dirty data is also restored during cache load.
Change-Id: I6cf86aabd68177b88638a68ea6a5b78a1068a4d0
Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455417
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This patch updates submodule and appropriate functions.
Cleaner poller is now registered only when needed.
Change-Id: Ic4ca7ce6f77b71ac12c19462f62ae7cd96c59006
Signed-off-by: Vitaliy Mysak <vitaliy.mysak@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455408
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
using data_offset and data_len is consistent with related APIs,
and is done in this patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ic9d216e25880f7b5ab33b764f45d332c090fa88c
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456291
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
In some test cases, unknown iSCSI PDU other than the login request
has been sent to the iSCSI target when the iSCSI target is not in
runnable state, and it has caused failure in iSCSI target.
To know what PDU was received by the iSCSI target, this patch changes
the iSCSI taget to collect dump of the PDU.
SPDK has already the SPDK_LOGDUMP macro but the SPDK_ERRLOGDUMP macro
will be appropriate in this use case and added. Then the SPDK_ERRLOGDUMP
is used in iSCSI library.
We can decode PDU and output any format easy to read by human, but
creating good format is not easy task and error prone. So this patch
uses simple dump.
Dump outputs like:
PDU
00000000 40 80 00 00 00 00 00 00 00 00 00 00 00 00 00 00 @...............
00000010 0b 06 00 10 ff ff ff ff 00 00 00 5a 00 00 00 03 ...........Z....
00000020 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I3ed2fca7fec24ccff17b89ba749a58c397b72c13
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456952
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
AIO backend requires aligned data buffers, and the maximum
IOVs supported in bdev module is defined to 32, there are
cases for Windows Guest which will send data segments more
than 32, SPDK can't process such cases, so here we can set
the 'seg_max' parameter based on bdev module capability.
Also set the maximum segment size for those requests.
Fix issue #625.
Change-Id: I0ff61e55872af17115c0b6b28425e70cb8769790
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/452378
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
iSCSI application may set the data iovec to NULL for READ requests.
Change-Id: Ic6b73f62f75dc2d927c18219fa1c2cb242f0528e
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/457079
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Keep lba map related fields in separate struct directly
in the band. Cleanup interfaces depended on ftl_md.
Lba map structure will be extended in next commit.
Change-Id: I1cfc2f2ff0c0e90bb63f39808780845673002e70
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/453370
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
As the name suggests, this function iterates through all elements of the
mempool invoking a callback function on each one. It's particularly
useful when deinitializing mempool that requires freeing resources tied
to each element (e.g. allocated through spdk_mempool_create_ctor).
Change-Id: I3da1fee527a36bf99f0b0e2dd3d6f9297422ff25
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455971
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Split the IO in case io->iov_cnt exceeds the FTL_IO_MAX_IOVEC. It allows
the user to pass any number of iovecs. Each IO request is now split if
necessary and submitted via ftl_io_call_foreach_child submitting each
child individually. Any resubmissions need to be handled on a child
level, as parent request doesn't track its children submission status.
Change-Id: If54249d54225b34191216a366c227f79215abc90
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455527
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
As the name suggests, the function calls a function passed as the
argument for all children IOs.
Change-Id: Idd7b6013e68fdd188db6e682bb9ba9ec2e6f52f1
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/455526
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Extended ftl_band_read_lba_map() interface to provide
range of lba map to read.
Added ftl_xfer_offset_from_ppa() function for
calculating offset of read portion of lba map
based on ppa address from IO. This offset is
used for coping read lba map portion to internal
ftl buffer.
Change-Id: I1c72a18c79eda8c33cd0b20ea36a5d9521a09d06
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/449435
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
SPDK_SUCCESS is a remnant and all other SPDK libraries have used
just 0. So this patch removes SPDK_SUCCESS from iSCSI library and
uses 0 instead.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ie33fd26238411e994ea9ea0c898da1abb502cad6
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456928
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This patch uses the newly added API spdk_dif_update_crc32c() when
DIF insert/strip is enabled.
Change-Id: Icf32a0ddef6cf92b0887b38495457e7fcac30987
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456172
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
iscsi_sgl_append_with_md had not increment _iscsi_sgl::iov
by appended count of iovecs. Hence if data digest is enabled,
data segment will be overwritten by data digest.
This patch fixes the bug.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ibcdacb883b2b97ad86cfc39a035c76264090401d
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456451
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
