- Cumulative Moving Average (CMA): average of all data up to current
- Exponential Moving Average (EMA): weighted mean of the previous n
data and more weight is given to recent
- Simple Moving Average (SMA): unweighted mean of the previous n data
Currently bdevperf support only CMA to know average of IOPS. However
CMA may not be appropriate to know the trend of IOPS. EMA or SMA will
be better to take the role.
The formula of EMA utilized in bdevperf is the following
Multiplier = (2 / (Time periods + 1))
EMA: EMA[n+1] = (IOPS - EMA[n]} x multiplier + EMA[n].
One of the source of the formula is https://github.com/patharanordev/ema
Change-Id: Ia7a282f6d173ace470e05b19ee6dee5edf640617
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-on: https://review.gerrithub.io/398886
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Modifies behavior of spdk_app_start() and spdk_app_parse_args()
such that they return on failure instead of terminating with
exit().
Change-Id: I82566417f04e1ae2e3ca60a00c72e664db26c9e4
Signed-off-by: Lance Hartmann <lance.hartmann@oracle.com>
Reviewed-on: https://review.gerrithub.io/401243
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Automatically detect more whitespace errors.
All existing cases are fixed; only whitespace change (verify with
diff -w) except for one comment style fixup in include/spdk/nvme.h.
Change-Id: If750e54b9c8e3421ea6feda5f20184a31431631e
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/402360
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
This is the first step in reorganizing the spdk test heirarchy.
Change-Id: I83467653af2da2a53251649092bed6902d6cb86e
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/401707
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I7a65a5818f2b851962c2be391efedfb9bfc8398c
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/373841
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
It is not ideal to have to add this in a bunch of
different Makefiles, but further consolidation of
Makefiles is going to be a more arduous process.
With this change, rebuilding SPDK after no changes
will result in no output - all of the "Nothing to
be done for 'all'" messages have been removed. Note
that DPDK build output still remains - this can be
suppressed by either using an out-of-tree DPDK, or
using SKIP_DPDK_BUILD=1 when using the in-tree DPDK
submodule.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Ib1d45432f2baaa1d9c1b9509906ee1e77966d3ca
Reviewed-on: https://review.gerrithub.io/399918
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
This allows us to remove the test/lib/bdev/nbd directory.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I3941b68c9c36a0e934acf2a0570e69c4d9dba722
Reviewed-on: https://review.gerrithub.io/399762
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
This was probably overlooked when developing 57d174ff6.
Fixes: 57d174ff6 ("bdev: add spdk_bdev_open/close")
Change-Id: Idb30287e740ac0300a5c7dc9fad7e06693f58330
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/398585
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This patch is used to test bdevperf on one NVMe disk with 2 GPT partition.
Change-Id: I46f6b5be3f4951625678861ff540629265e8931f
Signed-off-by: Hailiang Wang <hailiangx.e.wang@intel.com>
Reviewed-on: https://review.gerrithub.io/390709
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Liang Yan <liang.z.yan@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Instead of inserting a task into the TAILQ and then
immediately taking it back off, just pass the task
to bdevperf_submit_single instead.
This reduces overhead of bdevperf compared to nvme/perf.
nvme/perf does not use a TAILQ at all, and does something
similar by passing the just completed task so it can be
reused by its submit function.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I574b459a32ebe2e91ac1351de360de86cbc4a86d
Reviewed-on: https://review.gerrithub.io/393833
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
nbd functionality now can be gotten by nbd rpc.
Change-Id: I85ebcb44d9ed263bcee19162dc060f3b268dddcd
Signed-off-by: Xiaodong Liu <xiaodong.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/391371
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Test nbd rpc methods:
start_nbd_disk
get_nbd_disk
stop_nbd_disk
Verify nbd data read and write by dd and cmp.
Change-Id: Ie32639d11727893ceac7adce0119b9c2085e74e1
Signed-off-by: Xiaodong Liu <xiaodong.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/391732
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
There used to be a fixed 1 second of real time performance
display with the "-S" parameter. Make it as a configurable
parameter, the minimum interval is still 1 second while user
can configure a larger value.
Change-Id: I159d867f2b6019b095d0291c436f01b110ff69d8
Signed-off-by: GangCao <gang.cao@intel.com>
Reviewed-on: https://review.gerrithub.io/393306
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The task resource is allocated one by one and when it fails
to allocate at some point, call the explicit destroy to release
the resource.
Change-Id: Ida916ed25a279806fee76104990f1f92daf3e1b6
Signed-off-by: GangCao <gang.cao@intel.com>
Reviewed-on: https://review.gerrithub.io/393009
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Only Makefiles for libraries that directly depend on DPDK (rather than
the SPDK env abstraction) should add $(ENV_CFLAGS).
Change-Id: Ifdf44d3ef8c42bbf7f20edd524b330d00658235b
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/392818
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
We can use spdk_env_get_core_count() to determine the
core count
Change-Id: I854ec86f10c5670b02295214d1d36fe5d57e31a4
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.gerrithub.io/392934
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Purpose: To avoid overflow, if the passed value is too
big.
Change-Id: I59d31c0baa742cab14e22dcd88a61cb5adca247a
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/392579
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Use double-precision floating point in the performance_dump() function,
since we can be dealing with large numbers (especially now that
io_completed is now uint64_t).
This isn't part of the I/O path, just used to print statistics, so
there should be no performance impact. Also, the arguments passed to
printf() were already implicitly being promoted to double, so this
actually removes an extra conversion step.
Change-Id: I12aae36bc42e83777eaa050561eb5e55060ae81c
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/392414
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reason: If test a very long time, it will overflow
when we use int
Change-Id: I729e8cb5862ab78808da5121666b41e599e1dddb
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/392122
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
This patch does the following things:
1 Set the time period 1 second into a global variable,
then if we modify that value, we will make sure that
we change it in the whole file.
2 Fix the period performance printing data . We should not pass
the time period, but we should pass time period * times, since
I/Os completed are caculated from the beginning.
Change-Id: I1eaaf1655389cbccefdcc879a51fca94028afdcb
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.gerrithub.io/391421
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: I619f2e772b38061e26e3c5f2b4014ee73a5f5d81
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/391894
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Since we use time_in_sec to judge < 0 case
Change-Id: I8283c8209359acfe1b1d5db8a9d4315e5722c5c6
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/391154
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Require braces around all conditional statements, e.g.:
if (cond)
statement();
becomes:
if (cond) {
statement();
}
This is the style used through most of the SPDK code, but several
exceptions crept in over time. Add the astyle option to make sure we
are consistent.
Change-Id: I5a71980147fe8dfb471ff42e8bc06db2124a1a7f
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/390914
Reviewed-by: <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
We also caculate the time for still dumping the
performance data.
Change-Id: I59b736268ca57a366666a4bf55458ec3f279f4e7
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.gerrithub.io/390702
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Add g_spdk_nbd to record multi-nbd-disk in the
future. nbd subsystem will be used to init and
finish g_spdk_nbd and its linked spdk_nbd_disk.
Change-Id: Ia67caa57b9920d0080f8b2fa1c8e87615803e5fe
Signed-off-by: Xiaodong Liu <xiaodong.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/390097
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
nbd poller is put into spdk_nbd_disk.
nbd test app no longer needs to process spdk_bdev.
This enables upcomming patches to move nbd test app
functionality to SPDK library.
Change-Id: I1b563a49dc9488e8dcc20706be82b17fdbd07ff1
Signed-off-by: Xiaodong Liu <xiaodong.liu@intel.com>
Reviewed-on: https://review.gerrithub.io/390093
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
This issue may be hit under some large IO and large queue
depth cases where much memory will be consumed and the channel
may be failed to create due to insufficient memory.
The general fix here is that we need to properly handle
the error case that the channel is not created due to this
or that reason.
Change-Id: Icf58a7c8a9cba625ef1b8da6b8bf68ef72ef5546
Signed-off-by: GangCao <gang.cao@intel.com>
Reviewed-on: https://review.gerrithub.io/386051
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Hailiang Wang <hailiangx.e.wang@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Make this generic and not directly dependent on
the event framework. That way our libraries can
register pollers without adding a dependency.
Change-Id: I7ad7a7ddc131596ca1791a7b0f43dabfda050f5f
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/387690
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
This should be removed.
Change-Id: I122ef2729d8389fab6cc8883bc70f221c48f574b
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.gerrithub.io/388884
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
This patch is used to remove tht task_pool allocated
by spdk_mempool_create.
Change-Id: I8d3a483477d7fd8b5893baa41b8658893c852f6c
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/388534
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
This patch could prevent the potential coredump of
accessing rte_mempool among multiple threads. We
allocate the tasks from the beginning.
BTW, this patch also refactors the bdevperf to remove
dpdk dependency
Change-Id: I841e6c8b43276923acaf6686377741c37e2e71ec
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/387871
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: I8ea8c59a0c67176c0c0c39abf807afad61ff3828
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/387689
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Change-Id: I864cd8a6c206dfbe62fcb3f72275c1ae51aa4ed7
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/387688
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
nbd_start(), nbd_poll(), and nbd_stop() will all be called
on the same core, so the poller unregister will complete
synchronously.
Change-Id: I4f83f3821102277704422ed63d4cde424324e0a4
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/387686
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
According to my analysis, the exact number is not
related with core number, which should be calucated
as g_target_count * g_queue_depth. Since each
target will be run I/O by differnet cores.
And it would solve the task number not enough issue.
Change-Id: I483f4609341766123f95b39bff057e95a9d19531
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/386801
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Hailiang Wang <hailiangx.e.wang@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
RPC is a default feature required for almost all usages,
so enable RPC by default, but with a UNIX domain socket
for security reasons.
-r can now be used from the command line to specify
an alternative RPC listen address from the default
/var/tmp/spdk.sock.
Remove the Enable parameter from the Rpc config section
but still allow specifying an alternative listen address
using the Listen parameter as an alternative to the
command line option. This keeps backward compatibility
for this release for anyone using the configuration file
still.
Remove the Rpc sections from all configuration files
that were using them, except for those that specified
alternate TCP ports for multi-process test cases. We
can fix these later to use an alternate UNIX domain
socket and to use the command line instead.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Ife0d03fcab638c67b659f1eb85348ddc2b55c4c4
Reviewed-on: https://review.gerrithub.io/386561
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
RPC is core functionality for SPDK applications and should always
be initialized last (after all subsystems have been initialized).
So make RPC a first class citizen and integrate it with the
app framework directly instead of making it an "optional" subsystem.
Then we initializing it after all subsystems have completed
initialization, and tear it down before tearing down subsystems.
We can also do some other cleanup while here - for example, reactors
are already started when spdk_rpc_initialize() is called, so remove
the extra event that was added during initialization since this is no
longer needed.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I4cc63586a6d55be68786629a2176c61a88979267
Reviewed-on: https://review.gerrithub.io/385914
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This patch does the following work:
1 Fix the performance display name issue. When doing
the io statistics, we call spdk_bdev_get_name, however
this time, we are already calling spdk_app_stop, so
we lost the name info. Currently code, we will not
see the correct name. And this patch will help to fix
this issue, by storing the name at the begining.
2 Fix memory leak issue. It seems that we do not free the target
space.
Change-Id: I716bda1a7340921e01f9f9ea28b2b908cd19e0af
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/385347
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Previously, we were waiting to find the target bdev as
registered before proceeding with nbd operations. But
there can be a delay between when the bdev is registered
and when the nbd device becomes ready for block I/O
operations. This delay has recently become longer as
lvol now performs I/O on registered bdevs to check for
an existing logical volume store.
So instead, check for the existence of the nbd device
in /proc/partitions output.
While here, also fix a bug in the nbd.c code - it needs
to wait for the poller to be unregistered before calling
spdk_nbd_stop(). Normally I would fix this in a separate
patch but because these issues are causing a lot of
failures in the test pool, I'm expediting this by putting
both fixes in one patch (so we can avoid a bunch of
re-runs).
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Ia297337338f7eeee9b4c56b80e941d373c1a965d
Reviewed-on: https://review.gerrithub.io/385687
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Fixed a typo introduced in patch 90440e1fe1.
Fixes: 90440e1fe1 ("test/bdevio: do unit tests on their own thread")
Change-Id: I15a6c5366435f5d84e171f443182f292d4f7a0e3
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/383431
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Pawel Wodkowski <pawelx.wodkowski@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
if write_zeroes is not supported by the block device, we can get the
same behavior by simply writing a buffer full of zeroes to the blocks
we want to erase. I also incorporate splitting into the bdev layer to
accomodate large i/o.
Change-Id: I8fa1bfaaf22d7bfc6e3afb6e89d22fa9f7767e55
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/373829
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Previously unit tests were run
on the same thread that called
the SPDK init. The bdevio app
had 2 threads: one for init&ut
and one for I/O processing.
When doing any asynchronous I/O,
the init&ut thread was waiting
via pthread condition variable
for I/O thread to finish it's
work.
With upcoming virtio-initiator
RESET implementation, the reset
I/O has to be processed by
a poller on the init thread.
When init&ut thread sends an
RESET requests the app deadlocks,
as pollers on the UT thread
won't be processed until I/O
thread wakes the condition
variable.
This patch separates ut thread
from init, so that init can
run it's tasks in background
even when ut thread is waiting
for condition variable.
Change-Id: Id4b6962f4288e6a07354dbf1d3aec0d0fd9de916
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/382663
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
This restriction causes bdevs examination/discovery to fail
when there are asynchronous operations. We should remove this
limitation for now. Future plan is to implement approach
similar to the one that is present in kernel.
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: Ibe5572297672022412d25a4a835dc9527ce97f3e
Reviewed-on: https://review.gerrithub.io/378758
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
bdevio used to do I/O at hardcoded offset 2048.
This obviously fails for any blocksize > 2048.
All 2048 offsets has been changed to 8192.
Change-Id: I70c057b8de539718175a8d29aa91422ba3fba70e
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/378298
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
