bdev: submit queued IO after disabling QoS

There's no reason to abort IO that have been queued
due to QoS limits, when QoS is switched from enabled
to disabled. Submit them to the bdev instead.

Fi

bdev: submit queued IO after disabling QoS

There's no reason to abort IO that have been queued
due to QoS limits, when QoS is switched from enabled
to disabled. Submit them to the bdev instead.

Fixes issue #357.

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: If5eafc53418ac686120e1d6a1da884b42cef845e

Reviewed-on: https://review.gerrithub.io/418128
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Seth Howell <seth.howell5141@gmail.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>

show more ...

test: Move poller test to the thread unit test

This is the more appropriate location for this test

Change-Id: If556b365975f5658a55924aeb7122186961972bd
Signed-off-by: Ben Walker <benjamin.walker@in

test: Move poller test to the thread unit test

This is the more appropriate location for this test

Change-Id: If556b365975f5658a55924aeb7122186961972bd
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/417365
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>

show more ...

test: Fix inaccurate mock of poller behavior

This exposes that one of the tests was not expecting
the correct behavior, so also fix that.

Change-Id: Idb73b3ea74950b2e6f959a40e5740375cb76b8c7
Signed

test: Fix inaccurate mock of poller behavior

This exposes that one of the tests was not expecting
the correct behavior, so also fix that.

Change-Id: Idb73b3ea74950b2e6f959a40e5740375cb76b8c7
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/417364
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>

show more ...

bdev: encapsulate private members of spdk_bdev

Change-Id: Ica5abcfe5f9b73217e2d91c33c2cd418e061cf96
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/416458

bdev: encapsulate private members of spdk_bdev

Change-Id: Ica5abcfe5f9b73217e2d91c33c2cd418e061cf96
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/416458
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>

show more ...

bdev: move link to spdk_bdev_io internal struct

I missed this one in the initial series.

Change-Id: Id4dc7574a04cd964455852f1a00084b65ab989b3
Signed-off-by: Seth Howell <seth.howell@intel.com>
Revi

bdev: move link to spdk_bdev_io internal struct

I missed this one in the initial series.

Change-Id: Id4dc7574a04cd964455852f1a00084b65ab989b3
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/416253
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>

show more ...

bdev: add INI config file options for bdev_io parameters

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I10bd5cab8789b8cf4d8fd9cb848cede8372be45e

Reviewed-on: https://review.gerrit

bdev: add INI config file options for bdev_io parameters

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I10bd5cab8789b8cf4d8fd9cb848cede8372be45e

Reviewed-on: https://review.gerrithub.io/414712
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>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>

show more ...

test/bdev: use DEFINE_STUB for spdk_conf_* stubs

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I2165bdb1c2dc08d790aa6dc63a8367c7a1a3bc4d

Reviewed-on: https://review.gerrithub.io/4

test/bdev: use DEFINE_STUB for spdk_conf_* stubs

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I2165bdb1c2dc08d790aa6dc63a8367c7a1a3bc4d

Reviewed-on: https://review.gerrithub.io/414711
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>
Reviewed-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>

show more ...

bdev/qos: add the bandwidth rate limit control

This patch is to add the core control policy for the
bandwidth rate limit (max bytes per second). Change
the existing functions for a common name and s

bdev/qos: add the bandwidth rate limit control

This patch is to add the core control policy for the
bandwidth rate limit (max bytes per second). Change
the existing functions for a common name and specially
handle the case when IOPS and bandwidth rate limiting
are both enabled.

Change-Id: I9f4565958d472559ef6d8bea52b1fe2a5f3c8969
Signed-off-by: GangCao <gang.cao@intel.com>
Reviewed-on: https://review.gerrithub.io/413821
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>

show more ...

bdev/qos: add the configuration parse of bandwidth rate limiting

To support the bandwidth rate limiting besides the IOPS rate
limiting, this patch is to add the support of the configuration
parse. T

bdev/qos: add the configuration parse of bandwidth rate limiting

To support the bandwidth rate limiting besides the IOPS rate
limiting, this patch is to add the support of the configuration
parse. The format will be as following to have a 10 (10M)
on the Malloc0 bdev. The default unit is in MB.

Limit_BWPS Malloc0 10

Change-Id: I62d70391ccad7804e6673ec56a3ed1cb0a4fbbd4
Signed-off-by: GangCao <gang.cao@intel.com>
Reviewed-on: https://review.gerrithub.io/413652
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: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>

show more ...

bdev: Make the QoS channel use a regular channel

This makes a number of things much simpler and eliminates
the need for upcoming proposed reset handling.

Change-Id: I23a6badd0873f6dcf38ba1e55bf18d8

bdev: Make the QoS channel use a regular channel

This makes a number of things much simpler and eliminates
the need for upcoming proposed reset handling.

Change-Id: I23a6badd0873f6dcf38ba1e55bf18d846c2843df
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/407357
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>

show more ...

bdev: rename bdev_module_channel to bdev_shared_resource

It doesn't represent an io_channel, so lets
rename it to something less confusing.

Change-Id: I730eb0ff0eb7737a59c190812a25b327ccd8fc14
Sugg

bdev: rename bdev_module_channel to bdev_shared_resource

It doesn't represent an io_channel, so lets
rename it to something less confusing.

Change-Id: I730eb0ff0eb7737a59c190812a25b327ccd8fc14
Suggested-by: Ben Walker <benjamin.walker@intel.com>
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/409998
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>

show more ...

ut/bdev_mt: add test case for ENOMEM handling with multiple io targets

Bdev layer has been recently refactored to share
internal module channels for *all* bdevs of the
same bdev module. This makes u

ut/bdev_mt: add test case for ENOMEM handling with multiple io targets

Bdev layer has been recently refactored to share
internal module channels for *all* bdevs of the
same bdev module. This makes us return ENOMEM for I/O
of one bdev if another bdev of the same module (nvme, split, etc)
is entirely saturated. This is not the behavior we want,
as these bdevs may not have anything in common.

This `issue` has been fixed just now, but to prevent
it coming back again lets add a unit test for it.

This test case creates two bdevs using two separate
io_devices and checks if ENOMEM handling is somehow common
for these two. It must not be.

Change-Id: Ic32ef4b4347b8856d4b9f91107e6f188ad67978e
Suggested-by: Ben Walker <benjamin.walker@intel.com>
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/409997
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
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>

show more ...

bdev: Make QoS structure a separate allocation

Future changes will rely on the ability to atomically
swap this data structure out.

Change-Id: I77b93252df63155692271b5665cc040cb83a33d0
Signed-off-by

bdev: Make QoS structure a separate allocation

Future changes will rely on the ability to atomically
swap this data structure out.

Change-Id: I77b93252df63155692271b5665cc040cb83a33d0
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/409224
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>

show more ...

bdev/qos: Add unit tests for spdk_bdev_set_qos_limit_iops

These won't cover race conditions across threads, but
at least we have something to test the behavior.

Change-Id: I8e620d2076fe7a3d95df668f

bdev/qos: Add unit tests for spdk_bdev_set_qos_limit_iops

These won't cover race conditions across threads, but
at least we have something to test the behavior.

Change-Id: I8e620d2076fe7a3d95df668fda4bee49b6d0afa7
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/409343
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: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>

show more ...

bdev: Move QoS parameters into separate struct

The QoS parameters are only used from the QoS thread,
so gather them together in a struct pointed to by the
bdev.

Change-Id: I91005f1d00205217a189d1eb

bdev: Move QoS parameters into separate struct

The QoS parameters are only used from the QoS thread,
so gather them together in a struct pointed to by the
bdev.

Change-Id: I91005f1d00205217a189d1ebacdb55782814f247
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/407356
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>

show more ...

bdev: Fix race condition when testing whether QoS is enabled

When testing whether QoS is enabled, the code previously
checked mutable values in the bdev itself. Instead, it needs
to check the flag i

bdev: Fix race condition when testing whether QoS is enabled

When testing whether QoS is enabled, the code previously
checked mutable values in the bdev itself. Instead, it needs
to check the flag in the channel.

Right now, QoS can only be configured statically when the
bdev is created. This means that no channels will exist
prior to QoS being turned on, which simplifies setting
the per-channel flag (only need to set it when a channel
is created).

Change-Id: I59e56c64c18c262cc2a7f71a6dde8329edb35db7
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/407354
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>

show more ...

bdev: Remove implementation assumptions from QoS unit test

This will allow a bit of internal refactoring to occur
without breaking the unit test.

Change-Id: Id7da6b14e9cd4cab5fc4dc004b5858dbbb34bc3

bdev: Remove implementation assumptions from QoS unit test

This will allow a bit of internal refactoring to occur
without breaking the unit test.

Change-Id: Id7da6b14e9cd4cab5fc4dc004b5858dbbb34bc3a
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/407366
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>

show more ...

bdev: add new optional bdev i/f entry point

Add new optional bdev module interface function, init_complete, to notify bdev modules
when the bdev subsystem initialization is complete. Useful for virt

bdev: add new optional bdev i/f entry point

Add new optional bdev module interface function, init_complete, to notify bdev modules
when the bdev subsystem initialization is complete. Useful for virtual bdevs that require
notification that the set of initialization examine() calls is complete.

Change-Id: I0997fb5749d430f2fd3a40172ec8a1d5caa96964
Signed-off-by: paul luse <paul.e.luse@intel.com>
Reviewed-on: https://review.gerrithub.io/407222
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>

show more ...

test/bdev/mt free unregistered bdevs after polling.

This unit test operated under the assumption that one could immediately
free a bdev after calling spdk_bdev_unregister. This assumption is
broken

test/bdev/mt free unregistered bdevs after polling.

This unit test operated under the assumption that one could immediately
free a bdev after calling spdk_bdev_unregister. This assumption is
broken when we moved the actual destroy call to an asynchronous callback
to spdk_io_device_unregister.

Change-Id: I92d34f7e2e2993bfe9391f9bb72e08128dec74f4
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/406429
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>
Reviewed-by: Jim Harris <james.r.harris@intel.com>

show more ...

bdev/qos: add the QoS setting in conf file

For application like vhost/iSCSI target/NVMe-oF target/etc,
a new section as below can be added for the QoS rate limiting.

[QoS]
Limit_IOPS Malloc0 1000

bdev/qos: add the QoS setting in conf file

For application like vhost/iSCSI target/NVMe-oF target/etc,
a new section as below can be added for the QoS rate limiting.

[QoS]
Limit_IOPS Malloc0 100000
Limit_IOPS Nvme0n1 500000

Also added a sample change at test/lib/bdev/bdev.conf.in

Change-Id: I7c7d951fbe1352ca2571f135c657bc4fa43b56c7
Signed-off-by: GangCao <gang.cao@intel.com>
Reviewed-on: https://review.gerrithub.io/393221
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>

show more ...

bdev/qos: add the support to reset QoS bdev channel

As there is no I/O channel associated with the QoS
bdev channel, add the specific functions to handle
the reset operation on the QoS bdev channel.

bdev/qos: add the support to reset QoS bdev channel

As there is no I/O channel associated with the QoS
bdev channel, add the specific functions to handle
the reset operation on the QoS bdev channel. The
reset operation will be conducted on the QoS thread.

Related UT code on QoS with reset operation was also
included.

Change-Id: Ibba68ddb132fa926fec6327829157b43ac806713
Signed-off-by: GangCao <gang.cao@intel.com>
Reviewed-on: https://review.gerrithub.io/393181
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>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>

show more ...

test: move ut_multithread.c to common/lib

Change-Id: Ied7da08503b52470dbb8bb77eee1a711f41e7f28
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/404977
Test

test: move ut_multithread.c to common/lib

Change-Id: Ied7da08503b52470dbb8bb77eee1a711f41e7f28
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/404977
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>

show more ...

test: move test_env.c under test/common/lib

This file is accessed by files in multiple directories under test.

Change-Id: I634481fb58eab5c097aaece5289f88e531954fcb
Signed-off-by: Seth Howell <seth.

test: move test_env.c under test/common/lib

This file is accessed by files in multiple directories under test.

Change-Id: I634481fb58eab5c097aaece5289f88e531954fcb
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.gerrithub.io/404976
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>

show more ...

bdev: introduce 'get_bdevs_config' RPC call

Write information needed to recreate each bdev.

Change-Id: I3d2b24fd4aaa8b98ec558d864cc28b55a899452f
Signed-off-by: Pawel Wodkowski <pawelx.wodkowski@int

bdev: introduce 'get_bdevs_config' RPC call

Write information needed to recreate each bdev.

Change-Id: I3d2b24fd4aaa8b98ec558d864cc28b55a899452f
Signed-off-by: Pawel Wodkowski <pawelx.wodkowski@intel.com>
Reviewed-on: https://review.gerrithub.io/401217
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>

show more ...

bdev: make QoS channel management thread safe

We must hold bdev->mutex around all QoS channel manipulations, not just
channel_count; otherwise, there are race conditions.

Change-Id: I6183aef83f4d57

bdev: make QoS channel management thread safe

We must hold bdev->mutex around all QoS channel manipulations, not just
channel_count; otherwise, there are race conditions.

Change-Id: I6183aef83f4d5789bded426a1832e3faaa688363
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/403367
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>

show more ...