ASAN will report memory leak if we didn't free the old thread context,
it's related with Rocksdb version, I didn't get this issue with Rocksdb
v5.6, but Rocksdb v5.14 and v5.18 will complain the memory leak.
Change-Id: I116cad6c1bc723a60284a5c7cf1e78d10efa9ddd
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/666
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
If set, SPDK will continue loading the JSON config even if
some commands caused an error. This can be useful when loading
RPC config from spdk_tgt into e.g. bdevperf, which supports
only a subset of RPC commands and would usually fail with
"Method not found" message.
Resolves#840
Change-Id: I070fea862fd99e5882d870e11e6a28dc9d0c8ba6
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/620
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>
In the reactor main loop, _spdk_reactor_run(), check if both
spdk_thread_is_exited() and spdk_thread_is_idle() return true,
and if they are true, then remove it from the queue and call
spdk_thread_destroy() to delete it.
By previous patches, the exited thread does not accept any
new message or poller, all pending messages are processed, and
all unregistering pollers are completed. Hence we only need to
have very simple check in the reactor main loop.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ib678245de98a5c050843e6435026bdcf8b6c75c7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/507
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This is a preparation to support voluntary thread termination by
calling spdk_thread_exit().
The comment in the header file has already said that all associated
I/O channels must be released before calling spdk_thread_exit().
This patch actually checks if it is satisfied in spdk_thread_exit().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I56ac50b561c6ca91d3dc2d60c21c8d91d38f081b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/823
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This is a preparation to support voluntary thread termination by
calling spdk_thread_exit().
Change spdk_thread_exit() to return -EBUSY if the thread has any
registered poller. We enforce all pollers including paused poller
are unresitered before the thread is marked as exited.
By this change, a bug was found in reactor_perf test tool. Fix it
by adding spdk_poller_unregister() and add the g_ prefix to avoid
future potential errors.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: If7f40357c9a6f4101b3998ea0da3cc46cc435031
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/487
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This is a preparation to support voluntary thread termination by
calling spdk_thread_exit().
Change spdk_thread_exit() to return -EINVAL if the thread is already
marked as exited. This will be helpful to detect wrong call sequence
of voluntary thread termination.
Besides, update reactor shutdown and unit test framework shutdown
to incorporate this change accordingly.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I2296c61e273bf4d9580656dcbc2da0e8a8f3bcf7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/671
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This is a preparation to support voluntary thread termination by
calling spdk_thread_exit().
One of the subsequent patches will allow thread to exit only if
all of its pollers are being unregistered. After the thread is
marked as exited, only unregistering pollers will remain if the
exited thread does not accept registering new poller.
Hence in this patch, we change spdk_poller_register() to fail if the
current thread is marked as exited first.
Then, in subsequent patches, if we remove break from poller
processing in spdk_thread_poll(), poller unregistration of the exited
thread will complete, and we will be able to support voluntary thread
termination easily by checking only the thread is exited and idle.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I5b843a07049ef01a5ff402eb521e294182ce2ae2
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/822
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This is a preparation to support voluntary thread termination by
calling spdk_thread_exit().
By the last patch, the asynchronous release of I/O channel will
complete even after spdk_thread_exit() because pending messages will
be reaped.
Then this patch stops new allocation of I/O channel after
spdk_thread_exit().
Hence we will be able to release all I/O channels for exiting
thread within finite time.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I48a45bcba7c4b2c62d8c9d398ac35a584b533627
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/821
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This is a preparation to support voluntary thread termination by
calling spdk_thread_exit().
Previously, the exiting thread had discarded all pending mesasges.
We change this to stop accepting any new message in spdk_thread_send_msg()
and reap pending messages in _spdk_msg_queue_run_batch().
Add unit test case for the new behavior. Adding g_ prefix to global
variables for clarification is done together.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ida78e7bb1b86357602aea6938dd514897b67edd6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/482
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Add spdk_thread_is_exited() to check if the thread is marked as
exited. This API will be used by reactor to destroy the thread
which exited voluntarily.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I2086984b7813e8a1f401852fde7ab263bcf8ef60
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/481
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Rename all variables and comments related with timed poller from
timer_poller to timed_poller.
This patch was originally to count poller and add the current
value to thread_get_stats RPC. This will be realized to add
thread_get_pollers RPC instead but renaming will be valuable itself.
So keep only renaming in this patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I73880b8df6ed35391383e0890f76b4e509690e06
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/479
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
To track the ZCopy IO stats like IOs and Bytes read and written.
Change-Id: Iab2c4d8397aee5d746db9491bd4e8ac4eea69529
Signed-off-by: GangCao <gang.cao@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/539
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>
We only have the generic transport wrapper now.
Change-Id: Ib1487e9c7b7ba3b573f5ea47597ad6eac1c5407e
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/734
Reviewed-by: Seth Howell <seth.howell5141@gmail.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
With the transport plugin system, this is no longer necessary.
Change-Id: Ia73878599658db84150603223ac811cb5a34ffba
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/713
Reviewed-by: Seth Howell <seth.howell5141@gmail.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
With the transport plugin system, this isn't used anymore.
Change-Id: Ib81c73f262d44edb6c937ca0056ac027b1e1ca75
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/712
Reviewed-by: Seth Howell <seth.howell5141@gmail.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
All of the code now goes through the transport plugin system,
so this isn't necessary.
While doing this, caught a bug that the get_registers function
wasn't being set for the PCIe transport.
Change-Id: If19a933e0c6f656bc55232b15d59052e22af3ee9
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/711
Reviewed-by: Seth Howell <seth.howell5141@gmail.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This is meaningless for network devices, but will be useful
when emulating the more complete register state of local devices.
Change-Id: I37052e514101c298a1f66cc72135a8c3dd669003
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/420
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>
This doesn't do anything for a network fabric, but it doesn't
hurt to allow these commands to set the emulated register
values for AQA. This will be more useful when emulating a
physical NVMe device.
Change-Id: I2891d7a07a5dceff50c6d66a8ce0b6b7c22a79f8
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/419
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
The custom command handlers are registered by outside software.
Move the implementation from lib/nvmf to the nvmf_tgt application
to match the intended usage.
Change-Id: Iedb7ae5356f195dfb5bb465975808c8749d16f32
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/416
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This is a public header that needs to be accessible to
code outside of the SPDK project. The spdk_internal/
directory does not end up getting packaged - it's just for
headers used by multiple libraries within SPDK.
Change-Id: I14e1ab4fda4b0ee779203d190a266240b10be6ae
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/413
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Michael Haeuptle <michaelhaeuptle@gmail.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This defines the official interface that NVMe-oF target
transports may use. For now, all code is just copied
from elsewhere. Eventually we'll want to add doxygen
comments.
Change-Id: I0cd9368607544be18c7c49188d071e38ceb59b8f
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/412
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jacek Kalwas <jacek.kalwas@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
init_ctx and fini_ctx from dev structyre was redundant to initialization
module context. This patch merge those structures and keep initialization
and deinitialization context private for init module.
Change-Id: Ic94bc813112b265642933bc8183acb1961a0c01a
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/549
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
ibv_query_qp can return nonzero value if e.g. we received
IBV_EVENT_DEVICE_FATAL. Remove assertion not to break SPDK
in debug mode
Change-Id: I00b3bef448a69e2f43ee90e5466b2d78b55d8a08
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/659
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This event can occur for either qpair or listening device. The
current implementation assumes that every event refers to a qpair
which is wrong. Fix: check if the event refers to a device and
disconnect all qpairs associated with the device and stop all
listeners.
Update spdk_nvmf_process_cm_event - break iteration if
rdma_get_cm_event returns a nonzero value to reduce the
indentation depth
Fixes#1184
Change-Id: I8c4244d030109ab33223057513674af69dcf2be2
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/574
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Seth Howell <seth.howell5141@gmail.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
ftl_wptr_ready() return value is true or false so
it should return bool instead int.
Change-Id: I653e4ee1271d54f2e8c7b6fa0144a468faec4607
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/555
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Zone state should be changed to "full" when zone is
fully written.
Change-Id: Ib9ad29d1a0e788fba8a7eae881e07d384fa90548
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/554
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Since zone reset is private for core thread it will be
always issued from core thread.
Change-Id: Ice03d339cb59548690ccf8f48d7c1606ee408a59
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/552
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Sine core module is responsible for handling IO it
should also handling zone reset logic.
Change-Id: Id8be4bb221cc85f207d44bd45761a72e263ea5ce
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/551
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Context:
Every-time the SPDK starts, it creates a large file under
/dev/shm (130 MB in my setup) which is never removed when
the application terminates.
The number of trace entries can be tuned using the SPDK runtime
option 'num_entries'. A value of 0 for this option considerably
reduces the size of the trace file (few MBs) but it is still
created.
This patch adds a special case in the code when num_entries=0:
the trace system would not be initialized, and the trace file
would not be created in this case.
The rest of the code properly handle the case where the trace
system is never initialized.
Signed-off-by: Sylvain Didelot <sdidelot@ddn.com>
Change-Id: If54a548de4a2ce2def8e57d527d0dc20bc55fe4d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/606
Reviewed-by: Jim Harris <james.r.harris@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>
fdump puts out buffer in following format:
"00000000 76 61 6c 00 val. "
Each buffer is displayed with hex values up to maximum, then
followed up with a value. This is done to keep starting value
with the same alignment between each consecutive line.
Value contents should end at its length instead of
adding trailing whitespace.
After this patch, output is changed to:
"00000000 76 61 6c 00 val."
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I26b327e83f296ba3865f1a337f4a70764a80e2b0
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/483706
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
the purpose of this commit is to avoid an entry into the error log in
the case that a thread does not exist, since there is no separate
function to check if a thread exists
Signed-off-by: Jesse Grodman <jgrodman@gmail.com>
Change-Id: Iec79b87e11e411bde59004baae1dc996864cec4d
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/481576
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
When reset operation is ongoing mark zone as busy
to avoid write operation on it.
This path fixes issue #1171.
Change-Id: I30097e1db1bfbbabebd7e6adeab6e8810dc8b2b1
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/483264
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Mateusz Kozlowski <mateusz.kozlowski@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
It fixes memory leak e.g. when add_listener rpc called twice with the
same trid on the same subsystem (ref = 2). In such case kill or
remove_listener decrements ref only once.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: Ib19f2e50838feff1c9108957ee82a42da66e54a2
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482446
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This allows to configure desired retry_count instead of using
hard coded value
Change-Id: I25c9601997ace916dfb735469a4b443c0cd2a96b
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482499
Reviewed-by: Seth Howell <seth.howell@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Some functions performed incorrect header/data digest
support check, align it with NVMEoF spec. Use a table
to check if PDU supports digest depending on its type.
Change-Id: I6170dd19ace017f37fda0a923f604732799460b9
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/483375
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Bdevperf tool had not called spdk_poller_unregister() called on the
same thread that called spdk_poller_register(). But it had not caused
any issue because spdk_poller_unregister() simply set the state to
unregistered. This design flaw has been fixed recently.
However the new pause feature has been added to poller and this
design flaw might cause any unexpected behavior if paused poller is
unregistered.
We do not know any other case such that spdk_poller_unregister()
is not called on the same thread that called spdk_poller_register(),
but we have no way to know it even if it exists. Hence let's add
assert for such cases.
Parsing poller lists managed by thread may be another option but
spdk_poller_unregister() is performance critical. So we do not check
list.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I9d91daaeb81fa33d5f042dbe7ddbd8ab6ea98d55
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479391
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
spdk_put_io_channel() was designed to be called on the same thread
that called spdk_get_io_channel(). spdk_put_io_channel() sends a
message to its own thread, to allow the context to unwind before
releasing the resources. This had the side effect to allow an
incorrect thread to call spdk_put_io_channel(). This patch will fix
that.
Bdevperf tool had a design flaw that needed the side effect, but
it was fixed recently. We do not know if we have any other case.
Hence add assert to spdk_put_io_channel() to find other case.
We found that unit test for blobstore had called
spdk_put_io_channel() and fix it together in this patch.
Besides, correct the comment for spdk_put_io_channel() in
include/spdk/thread.h not to create any other case in future.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I6ec7bf074818abef43b23ca40bc9385adac70a75
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479390
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
There is no need to support read operations on separate thread.
Change-Id: I10b595b8eeaf5fd0182f05913fdd5baa4b84961f
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/471910
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
"ocssd" pool name is no longer valid.
Change-Id: Ie9671d9598633ebf165570419a9bbe89e539b04e
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/481789
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: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Adding support for handling media management events.
Change-Id: I1b87143ddd360e314385e8dc70359aeb0d526b5a
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/481691
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: 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: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
In order to handle media management events spdk_bdev_open_ext()
should be used instead spdk_bdev_open(). Move this call to ftl lib
to keep media management events internal to the library.
Change-Id: If4c9382cc89fc537667923f00d3dae5df0ace248
Signed-off-by: Wojciech Malikowski <wojciech.malikowski@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/481503
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Added spdk_vmd_fini(), which detaches all PCI devices acquired by the
VMD subsystem.
Fixes#1148
Change-Id: I43218ef5f9a764546b655c28688897fb91b779cb
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482852
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
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: Jim Harris <james.r.harris@intel.com>
It allows to property set (e.g. cc) when subsystem and qpair are not
active.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I0b0d150fbdac5bdf0d20762337f0a811f4d6d243
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/481494
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This internal interface allows to create nvmf ctrlr and connect io
qpairs on add listener rpc request (i.e. when subsystem is stopped
and listener is not yet on subsystem's list).
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I998cb72ed773094faacc6668cf069ba9e2a6bf50
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/481409
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Had to remove one part of a unit test because the null
checking is moved to a different function.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I0a95d0a9a9a5708416fdc7efefb36e17b1ffe010
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/480008
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Seth Howell <seth.howell@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
trstring variable in spdk_nvme_trid_populate_transport is not
initialized, that can lead to snprintf() writes some garbage to
trid->trstring if the user passes SPDK_NVME_TRANSPORT_CUSTOM trtype
Add return statement and assert to CUSTOM/default switch
Change-Id: I6c6c37f9aa74d61b346f7be27fb890c7a34e9229
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/483469
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Seth Howell <seth.howell@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
When attempting to decode a base64 string, while there is a way to
calculate the maximum possible decode length, there isn't a way to
calculate the exact decode length without duplicating some base64
specific logic located in spdk_base64_decode. With this change, the
spdk_base64_decode function can now optionally calculate the exact
decode length without actually performing the decode by passing NULL
in as the dst argument.
Change-Id: Ice83db979f86a6fe9f39d236d3083102ca37ec68
Signed-off-by: Mike Carlin <mikefcarlin@protonmail.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479479
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Xiaodong Liu <xiaodong.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
For nvme_ctrlr_cmd_format command status should be used as
nvme_completion_poll_cb callback argument instead of pointer to
local variable.
Change-Id: Id65cb395d137c4e907c1ef019b131e8822ddfe34
Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/483513
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
With recent changes to extent on-disk metadata format,
new format (Extent Pages) is not backwards compatible.
Meanwhile old format (Extent RLE) is backwards
compatible with older SPDK applications.
Summing up:
Blobstore created pre SPDK 20.01 can only use Extent RLE.
Blobstore created starting with SPDK 20.01 can use both,
Extent Pages and Extent RLE specified by use_extent_table opts.
When use_extent_table is set to true, invalid flag for it is set.
SPDK application pre 20.01, will not load such blob.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: If14ebd03f19eb581d71dcb46191e099336655189
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/483220
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This allows us to avoid trying to map the same physical address to the
IOMMU in physical mode while still making sure that we don't
accidentally unmap that physical address before we are done referencing
it.
Change-Id: I947408411538b921bdc5a89ce8d5e40fd826e971
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/483133
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
The fuse command value is a two byte value, but we were only checking to
see if the fuse value was equal to SPDK_NVME_CMD_FUSE_FIRST or
SPDK_NVME_CMD_FUSE_SECOND in spdk_nvmf_ctrlr_process_io_fused_cmd. If a
haywire initiator sent a command with a fused value equal to
SPDK_NVME_CMD_FUSE_MASK, that would result in us skipping all checks and
dereferencing a null pointer in
spdk_nvmf_bdev_ctrlr_compare_and_write_cmd.
To fix this, add an extra condition to validate the cuse field.
Change-Id: I1ec4169ff5637562effd694f7046c6e3389627f1
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/483123
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
This was observed after running nighly tests on previous patch.
As part of it, autopackage.sh compiles SPDK
without debug flag set. Exposing the uninitialized var here.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Iedb1641f3c0d4a21f293c81cd4fcf35c6d1c7ae5
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482893
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Extent table and extent page descriptors are now
set to be default way clusters are serialized on disk.
With this patch UT are ran with and without
extent table.
Changed two asserts in test, since amount is dependent on
which type of serialization is used.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ica58fce6a4effd014d7dd40ee26edd0fa3196d0f
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/481901
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
ctx->extent_page signifies if page was allocated
for insertion.
1) It is possible for a thread to claim extent page
on its own thread, and put it in ctx->extent_page.
If conflicting thread allocates another ctx->extent_page,
then it should be freed. This does not mean failure
to insert cluster. As different threads could have
been trying to allocate different clusters,
so condition on line 6716 does not cover it.
If so then it shouldn't be an issue to release
the claimed ctx->extent_page and proceed with updating the
extent page which originally won the race.
NOTE: if clusters were conflicting, then extent_page is
freed in _spdk_blob_insert_cluster_cpl().
2) At this point of _spdk_blob_insert_cluster_msg()
we already verified that there already is
extent page allocated at "*extent_page".
In such case ctx->extent_page will be 0,
and should not be used.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Id5b57c88248890eee60d2e7dbecbd984c98b561b
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482867
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
sz is set to number of clusters that should be have been
in particular unallocated EP (remaining_clusters_in_et
up to SPDK_EXTENTS_PER_EP).
The cluster array should be set to 0 only in region
between original size (cluster_array_size) and new
total size (active.num_clusters).
It was incorrectly using sz in the memset.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ic43e89c17d53e9529e3ed0349aeb4fb7dc6593f2
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482858
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Previously part of function assumed that cluster count
1)means number of clusters in EP and another 2) that it is
following the active.num_clusters (akin to extent_rle).
This was incosistent and showed when using multiple
extent pages to serialize metadata.
This patch changes it to only go with 1), so it is clear
that it means number clusters within particular EP.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I155104cabc127ed47df04434032fb01e08948e13
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482848
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Following changes are done in this patch:
1) _spdk_blob_serialize_extent_table now persists
at least single extent page. When num_extent_pages == 0.
2) Minimum valid size of ET descriptor is even without
extent_pages. This is a case when there are no EP,
but we still want to persist num_clusters in ET.
3) Taking above points, redone the loop for serializing
extent pages.
4) Make sure to mark blob dirty if any new extent pages
were allocted.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I81dc6cf2de2722bb49927ed42f4b9f31292f78c5
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482847
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Changed assert checking if cluster array is allocated
when loading extent pages. This is true only for
the first extent page being loaded, of course after that
the cluster array can be already allocated.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I0f96294ede5a12ffd6bca73cbeadba8d94a35bac
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482857
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Size of a blob (thus size of clusters array in mutable data)
is known from extent table descriptor.
Extent pages were read sequentially in order they were
placed in extent table. This meant that cluster
array could have been filled up from beginning to end.
Yet reading extent pages in any other order,
would result in incorrect placement of clusters.
This patch adds first cluster index that is contained within
each extent page. This will allow to read/write
multiple extent pages in parallel, since
we will know where in clusters array to put the cluster idxs.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ib6b9332111cd93f990d057dc60624152907dd87f
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482701
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
A pointer to a stack variable is passed as an argument to
nvme_completion_poll_cb function, later this variable is used
to track completion in the spdk_nvme_wait_for_completion() function.
If normal scenario a request submitted to the admin queue will be completed
within the function which submitted the request.
spdk_nvme_wait_for_completion() calls nvme_transport_qpair_process_completions
which may return an error to the caller, the caller may exit from the
function which submitted the request and the pointer to the stack variable
will no longer be valid. Thereby the request may not be completed at that time
and completed later (e.g. when the controller/qpair are destroyed)
and that will lead to call to nvme_completion_poll_cb with the pointer
to invalid stack variable.
Fix - Dynamically allocate status structure to track the completion;
Add a new field to nvme_completion_poll_status structure to track status
objects that need to be freed in a completion callback
Fixes#1125
Change-Id: Ie0cd8316e1284d42a67439b056c48ab89f23e0d0
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/481530
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This is more adequate name, since this value if first read from
Extent Table descriptor. Then decreased when iterating over entries in
extent table and extent pages are read.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ib188c524b8488b38d4de063a9970dcfdf49c9acd
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482600
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
When replaying md chain for a blob, extent table
descriptor can be read. When it is present, all allocated pages
it points to are now being put into extent_pages array in ctx.
If multiple extent table descriptors are in single md chain,
the array is expanded accordingly.
After replaying single md chain is done, replay extent pages
starting from last one. Replaying extent pages, is similar to
extent_rle in that each allocated cluster is claimed and
number of free clusters in blobstore decreased.
When all extent pages are read, return to
_spdk_bs_load_replay_md_cpl() for continuing replaying
next valid md chain.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I4573226aff7d7b1bcdfd188518235c8d4b68a4c3
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/481621
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>
_spdk_bs_load_replay_md_parse_page() is only used in
replay path during blobstore load.
Next patch will expand the load ctx with array of
extent pages to be read. It is filled out when reading
in-chain metadata of extent table descriptors.
Passing the load ctx here will make it simpler to
fill out the array when processing extent table.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: If96e6670560c8c4a3610f33ece14c354d7d5da39
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482412
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
When EXTENT_TABLE descriptor is found when parsing metadata
that means there can be extent pages to read.
If extent page was not allocated, number of clusters can be
increased depending on the num_clusters_in_et.
Unallocated extent page contains either SPDK_EXTENTS_PER_EP
or remainder of num_clusters_in_et worth of clusters.
Depending which is less.
Added decreasing fo num_clusters_in_et to parsing
extent pages as well.
While here, remove ctx->seq = seq assignment as that is
done at beginning of blob load.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I57f54634b908ffb406f3e91e15841b7f36fd6de6
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/476429
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Add new serialization of changed extent pages before persisting md.
Iterate over active extent pages (not array !). When they are
allocated but not yet present on disk - write them out.
All extent pages in clean mutable data are assumed to be written out
already.
So there are two cases here:
1) Active mutable array is larger than clean
All allocated extent pages should be written out.
2) Cluster allocation created new extent page
Blob has to be thin provisioned and persist was called
as part of cluster allocation. New extent page needs to be
written out and EXTENT_TABLE allocated.
Iteration is done over num_extent_pages instead of extent_pages_array_size,
to prevent writting out too many extent pages when size of blob was
made smaller. The two values come back in sync at the end of persist
either way.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I780819fd7f3c44e4cf5d71c188c642536d3cc320
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479851
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Right now output from _spdk_bs_cluster_to_extent_page()
is used to determine whether the exten_table is used at all.
If NULL pointer was returned this meant that extent table
was not allocated, even if the code might suggest just
checking if we overran the array.
To make it more obvious, the _spdk_bs_cluster_to_extent_page()
now only asserts the extent_table_id.
blob->use_extent_table is now always used to determine the
serialization path.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I9d2630645213539bae5cd1d72e5f9b878f53c2bc
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482599
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This patch add single EXTENT_PAGE updates on cluster allocations.
There are three possible outcomes after inserting a cluster:
1) blob uses EXTENT_RLE
Proceed to usual sync_md.
2) blob uses EXTENT_TABLE and extent page was not yet written out
Update the active mutable data to contain the claimed md page,
write out the EXTENT_PAGE and sync_md to update EXTENT_TABLE.
3) blob uses EXTENT_TABLE and extent page was previously written out
Only serialize that single EXTENT_PAGE and write out the updated
cluster map for it.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ia057b074ad1466c0e1eb9c186d09d6e944d93d03
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/470015
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Force number of Extents to fit into Extent Page to
be power of 2, in order to simplify calculations
on cluster allocations.
At this time SPDK_BS_PAGE_SIZE is 4k, which would
results in SPDK_EXTENTS_PER_EP to be 512.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I7e09d92b00dfe5c12d7dd10ac0fc5a9a10d526ac
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/472041
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Similar to EXTENT_RLE, this descriptor holds LBA of clusters.
Difference is that EXTENT is kept in separate md pages,
and only single EXTENT will be updated on cluster allocation.
This patch adds the EXTENT processing, which is not used
until following patch.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ifbac23db7ca3e7c8c91cee01018f20071f0d5160
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/470014
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Added claiming the extent page.
Which is then followed by updates in updates
of mutable data on md thread.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: If511564f812685381c48924310105a4cb6f63cd1
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479850
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Functions to claim and release md pages were added.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I1c8ddc13c8a5806fb874e5c34dae2a327e1ff248
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482011
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I3e49c398d9bdf9f4eacba65061cc7fe4b300fb56
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479963
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
With this patch extent pages array will change it size accordingly
to size of the blob. Similar to clusters, only resizing up
is done on blob resize. Shrinking is done on persisting the blob.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Id7f7c81efbd96af414fce9fc4045cbb476cc93a6
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479962
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Extent Pages claim and insertion can be asynchronous
when cluster allocation happens due to writing to a new cluster.
In such case lowest free cluster and lowest free md page
is claimed, and message is passed to md_thread.
Where inserting both into the arrays and md_sycn happens.
This patch adds parameters to pass the Extent Page offset
in such case.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I46d8ace9cd5abc0bfe48174c2f2ec218145b9c75
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479849
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Added new descriptor SPDK_MD_DESCRIPTOR_TYPE_EXTENT_TABLE.
Extent Table will hold md page offsets for new Extent Page descriptor.
Entries in Extent Table are run-length encoded 0's as unallocated
Extent Page descriptors.
Additionally total number of clusters is persisted in each Extent
Table descriptor. This is because there is no guarantee that
last Extent Page of a blob will be allocated.
Even if number of Extents per Extent Page is always the same,
Extent Page can hold less Extents than that.
This patch does not add more metadata on disk right now.
Only added descriptor parsing/serialization and applicable fields
to store it in run time.
Following patches are going to implement TODO's added in this patch.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Iac5d8f00ddfc655c507bc26d69d7adf8495074e9
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/466920
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Lets get it removed ! :)
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I91b994a883a642d87ecc8c152c801b8a7676f33a
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482010
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Since further patches will be adding new descriptors
that are related to cluster layout throughout the blobstore,
add description for existing descriptor too.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I722eb633445685789d5185ed59dfc910f76b109f
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/481724
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This is an additional option that can be passed when creating
a blob.
When opts->enable_extent_pages is set to false (current default),
only EXTENT_RLE should be persisted on sync.
During blob load, when EXTENT_RLE is present in md,
blob->extent_rle_found is set to true.
When opts->enable_extent_pages is set to true,
only EXTENT_TABLE and EXTENT_PAGES should be persisted on sync.
During blob load, when EXTENT_TABLE is present in md,
blob->extent_table_found is set to true.
It is possible to find neither EXTENT_* descriptor when loading a blob.
This means that blob length is 0 and EXTENT_RLE was supposed to be used.
Yet none were persisted due to lack of clusters.
In such case blob->use_extent_table is set to true after finishing
blob load.
When parsing metadata ends, if extent_table_found is set - then
support for extent_table is enabled. All other cases disable it.
At this time path for Extent Pages is not implemented, so it should
not be used.
Later in the series, it will become the default path for serialization.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I2146da6130a0645e686ab02a3b5d2d86a7d35a1f
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479853
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
When a command arrives and no requests are available, the socket
recv state machine sits in the RECV_STATE_AWAIT_REQ state until another
network event occurs. If this I/O was the last one sent, this leaves the
target hung. To fix this, when a request is completed, kick the state
machine to make forward progress.
In practice, this can only occur once the pdu send acknowledgements are
asynchronous relative to arriving commands. That only begins happening
with the use of MSG_ZEROCOPY. When MSG_ZEROCOPY is turned on, it's
possible receive the next PDU in a chain for a command prior to seeing
the acknowledgement that the response that triggered that PDU actually
sent.
Change-Id: I556f31ad56970d36aa3538cfde375d35f3d4e551
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/480002
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Previously, the R2T was sent and if an H2C arrived prior
to seeing the R2T ack, it was processed anyway. Serialize
this process.
In practice, if the H2C arrives with a correctly functioning
initiator, that means the R2T already made it to the initiator.
But because the PDU hasn't been released yet, immediately processing the
PDU requires an extra PDU associated with the request. Basically, making
this change halves the worst-case number of PDUs required per
connection.
In the current sock layer implementations, it's not actually possible
for the R2T send ack to occur after that H2C arrives. But with the
upcoming addition of MSG_ZEROCOPY and other sock implementations, it's
best to fix this now.
Change-Id: Ifefaf48fcf2ff1dcc75e1686bbb9229b7ae3c219
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479906
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>
This function was only called from one spot.
Change-Id: I856f564d3ef6c6157be7a32a2cd812c702516a8d
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482003
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This seems like a more descriptive name
Change-Id: Ia616865b3fb36d8f9ccc5fb2ca6185bdd8543cf8
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482002
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
With our target design, there's no advantage to sending
multiple R2T PDUs per nvme command. This patch starts by
setting up the math so that at most 1 R2T PDU is required
per request. This can be guaranteed because the maximum
data transfer size (MDTS) is pre-negotiated in NVMe-oF
to a reasonable size at start up.
It then proceeds to simplify all of the logic around mapping
requests to PDUs. It turns out that the mapping is now always
1:1. There are two additional cases where there is no request
object at all but a PDU is still needed - the connection response
and termination request. Put an extra PDU on the queue object
for that purpose.
This is a major simplification.
Change-Id: I8d41f9bf95e70c354ece8fb786793624bec757ea
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479905
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
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: Alexey Marchuk <alexeymar@mellanox.com>
We can always accept up to the maximum I/O size in an H2C,
so eliminate the #define.
Change-Id: I349dab5f9b6ec482a7c580b1396e03c8d30a250b
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482278
Community-CI: 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: Alexey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
The resources allocated to a queue pair do not need to be directly
correlated to the queue size requested by the initiator in NVMe-oF, as
long as enough resources are present. The RDMA transport, for instance,
does complex pooling of the resources behind the scenes when using a
shared receive queue.
Simplify the resource allocation for a TCP qpair to just always allocate
the max allowed queue size right away. This is a configurable parameter,
so system administrators can adjust for their needs. The initiator may
then request a queue size less than or equal to that, which will only be
enforced by queue depth counting and not impact the actual number of
resources allocated on the target.
This change relies on the MaxC2HSize being equal to the Maximum Data
Transfer Size (MDTS) reported. That is the default configuration, but
MDTS is configurable. Changing the MDTS with this patch to a value
larger than 128k will cause the target to break. This is addressed in
the next patch in this series.
Change-Id: Ibd4723785c6a4d8d444f9b7bbfa89f98de2320f5
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479733
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
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: Alexey Marchuk <alexeymar@mellanox.com>
These values do not need to be negative.
Change-Id: Id9f798cf1c9da354448f9c6fbb90e599f877bb32
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482277
Community-CI: 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: Alexey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
By releasing the just-completed PDU prior to calling the callback,
for flows that immediately submit another PDU inside the callback,
the just-released PDU can be immediately reused. This reduces the number
of PDUs required in the pool to continue forward progress to half of the
previous value, while also making it more CPU cache friendly.
Change-Id: I8031b8f9f57ac05f261d96433d9899fe5e31d318
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479904
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Or Gerlitz <gerlitz.or@gmail.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
When we unlock a range, we remove the range from the
locked bdev list before doing the for_each_channel
iteration to remove the range from each channel.
But at the same time, right after removing from the
locked list, a new lock on that range could start.
In that case, we also do a for_each_channel to add
the range to each channel, and that will race with
the for_each_channel remove. When the lock start
wins, it finds the range already in the channel,
but doesn't set the owner_range which results in
a seg fault when the for_each_channel completes.
The fix is actually rather simple. We just add the
locked_ctx to the comparison when checking if the
range is already in the channel. If the locked_ctx
matches, then we know it was added as part of
initializing a new channel. If it doesn't, then
we create a new range object pointing to the new
locked_ctx. The first one will get removed when
the remove for_each_channel catches up.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I94f8b20376dd437f404add35744d42fc148303ff
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482620
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Maciej Szwed <maciej.szwed@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
If a locking operation has to wait because of an
existing lock, we queue the lock context. When the
existing lock finishes unlocking, we restart the
queued lock context. But we have to make sure
we restart the lock context on the same thread it
was originally submitted, since it has a channel
associated with it.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I555515f3adfc3c13a86584c601ed541d605980b7
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482463
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Maciej Szwed <maciej.szwed@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
For ACWU we always set value 1 because bdev holds
information specific for namespace only. This value
actually does not matter because we also set NACWU
which makes ACWU irrelevant. We set ACWU because
NVMe specs requires ACWU != 0 if fused commands
are supported.
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: Ida4357026d3b32677fc824b3cd878e7ad8ef2680
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/477915
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This function is required for NVMf implementation
for compare and write fused command.
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: If41611f5c0b8e4ed8eec66f09858c724f1800d59
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/477914
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Add call for spdk_nvmf_bdev_ctrlr_compare_and_write_cmd
function in spdk_nvmf_ctrlr_process_io_cmd function
when fused command is discovered.
This patch also removes redundant defines for fused flags.
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: I61971a56577ab32b52e1fde1e572f718a9a2d9aa
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/476621
Community-CI: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
