Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Change-Id: I54093bef0fc34947dce161bacee7a75721ff937f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8712
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Michal Berger <michalx.berger@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This patch is used to add the support for users to configure
use kernel or userspace idxd library.
Change-Id: Ie159b897bc9595894ad8f333168efaea6c2a3d78
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7332
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
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: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This patch is used to add the kernel idxd support.
Without this patch, we can use userspace idxd driver
under accel_engine library (module/accel/idxd/accel_engine).
With this patch, we can also kernel idxd driver under the
accel_engine library.
Our approach is implementing a wrapper library to use IDXD
device by leveraging the kernel DSA driver in SPDK idxd library
(lib/idxd).
Then users can leverage the RPC later to configure how to
use the DSA device by user space driver or kernel driver.
In this patch, our approach is to use the idxd-config library
to export the WQs (Working Queues) exported by the kernel.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I3a25a4fe0327bd626bf6883dfbe54437d3209e51
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7331
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
There is no need to map the PRP/SGL list RW since this memory is never written
to. In fact, SeaBIOS might submit a request where the PRP list resides on
read-only memory, so attempting to map it RW can break things.
Change-Id: I7e4e90b1fa7e33e81b8d5cd8dcb9568c038938ec
Signed-off-by: Thanos Makatos <thanos.makatos@nutanix.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7288
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Nvmf/vfio-user uses this API to map NVMe command sent from
VM from Guest Physical Address to Host Virtual Address, so
now we moved this API from the nvme library to nvmf/vfio-user
as an internal API.
UT code will be added back in coming patch.
Change-Id: I54817fc9811ccd9ddd97b3aa6762a2fce4bbdda6
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8574
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
This patch is used to improve the performance when
we need to reorder the list.
PS: Bring the similar operations from posix implementation.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I7031b35ddb597730ee160690e8ab9caf9b2b64b7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8675
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Just remove this function pointer and add a new one,i.e.,
dump_sw_error.
Because this function pointer is only used to
read a sw err info. We can hide it in the detailed
idxd implemenation.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I42fe2220dae85df307b5af64e37acfd7f748915b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8707
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
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: Changpeng Liu <changpeng.liu@intel.com>
Latest nvme-cli (>= 1.13) fails to issue commands towards SPDK's cuse
ctrl device, e.g.:
$ nvme get-feature /dev/spdk/nvme0 -f 1 -s 1 -l 100
nvme_cuse.c: 654:cuse_ctrlr_ioctl: *ERROR*: Unsupported IOCTL 0x4E40.
get-namespace-id: Invalid argument
The reason is because nvme-cli now also sends NVME_IOCTL_ID to the
target device to determine if it's indeed a controller or a ns. In
case kernel returns ENOTTY then nvme-cli considers the device to be
a controller. Since cuse_ctrlr_ioctl() returns EINVAL in such a case
the nvme-cli fails.
To avoid this simply replace EINVAL with ENOTTY for the ioctls that
may be not supported by ctrl or ns device.
nvme-cli commit in question:
fa2b91da74
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Change-Id: I29003864bc2a5c1a8906d6d01beba3d6f4e31b0e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8531
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
gcc 11 started to complain about this condition. Ignore it.
Spotted under fedora34.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Change-Id: I472e5e88a30feaacc7ea96d5787544f49eaceb16
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8530
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: John Kariuki <John.K.Kariuki@intel.com>
Reviewed-by: Pawel Piatek <pawelx.piatek@intel.com>
Reviewed-by: Maciej Wawryk <maciejx.wawryk@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
When the application is not set with the interrupt mode,
we should fail the rpc.
Fixes issue #2023
Change-Id: I4157f220948b7c96f474682c41ac9d78977b1c40
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8676
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Xiaodong Liu <xiaodong.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: <dongx.yi@intel.com>
POSIX defines PRId64/PRIu64/PRIx64 for printing 64-bit values in a
portable way. Replace a reference to %lu to remove the assumption
about the size of a long.
Signed-off-by: Nick Connolly <nick.connolly@mayadata.io>
Change-Id: I5696bcdb3add3d8b9d5ea91d9bc50b1dad1fbc95
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8685
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Undefined references to the ncurses library APIs
_nc_panelhook_sp, getmouse, ungetmouse, copywin etc. while
cross-compiling with aarch64-none-linux-gnu/10.2.1 compiler.
Hence putting the libraries in the right link order.
Signed-off-by: Krishna Kanth Reddy <krish.reddy@samsung.com>
Change-Id: If019744cf6b7bf79c072de57232bcf412d919a2b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8701
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
The RPC method create_pmem_pool has been changed to bdev_pmem_create_pool.
Change-Id: I8ad1fb98d8e03240a4d982e04e9270cf90e198fc
Signed-off-by: wanghailiangx <hailiangx.e.wang@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8708
Community-CI: Broadcom 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: Changpeng Liu <changpeng.liu@intel.com>
Now that we've deprecated the RPCs for a release, we can remove the whole
library.
Change-Id: I0f1a357fcfb3404efac39aa021928841c2f22ff1
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4305
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
bdev_nvme_reset() will be used by JSON RPC and we will have to call
the callback to JSON RPC at bdev_nvme_reset_complete(). To do it
easily, register the current completion function for nvme_bdev_io
in bdev_nvme_reset_complete() into nvme_ctrlr as a generic callback.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ie59551dc343215a95bfa5b22f234fc153c9db1b5
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8589
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
nvme_ctrlr will be registered as io_device even when multipath is
supported. Hence while spdk_for_each_channel() is executed in reset
processing, we can get nvme_ctrlr both spdk_io_channel_iter_get_ctx()
and spdk_io_channel_iter_get_io_device(). This duplication is not
necessary. Use spdk_io_channel_iter_get_io_device() and pass NULL
to the context parameter of spdk_for_each_channel() for clarification.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ifdbd0af4274081c4be7ab0735eb8bf9ae10e3493
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8588
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
The current nvme_ctrlr will be registered as io_device even when
multipath is supported. Then we do not have to differentiate completion
processing between reset from bdev_io and internal reset. Hence
inline bdev_nvme_reset_io_complete() into bdev_nvme_reset_complete().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ife2c4c93d423da3953174ac860485a6e095a28bd
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8587
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
When we support multipath, reset_io will hold the controller currently
being reset to reset all underlying controllers sequentially.
bdev_nvme_submit_request() basically passes nvme_bdev_io to each I/O
type, and we have a convenient helper function bdev_nvme_io_complete()
which has nvme_bdev_io as a parametetr.
So revert the previous change to bring nvme_bdev_io as context
for reset processing.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I19697e8252505bab519a42889d1a88d967932f22
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8586
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reset requests from the upper layer will reset the underlying
ctrlrs of a bdev ctrlr but internal reset requests will reset only
the specified ctrlr.
To clarify such difference, rename bdev_nvme_reset() by
bdev_nvme_reset_io() and remove the underscore prefix from
_bdev_nvme_reset() and related functions.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I9a2d124f6e2039bfecfdd6599827354d6c373f2e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8492
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This will reduce the size of the following patches and improve the
readability.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I2ff5a10622141466d6a0352b2a6387b1412cfbf8
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8641
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Create an nvme_bdev by adding namespace to an attached controller,
and use bdev_nvme_submit_request() in the test case test_pending_reset().
This will reduce the size of the following patches and will increase
the test coverage.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ifb05952a53949e12562e9be3f6526b0f8ad57696
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8640
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
nvme_bdev_first_ctrlr() and nvme_bdev_next_ctrlr() were not possible
to hold mutex correctly, and nvme_ctrlr_get() and nvme_ctrlr_get_by_name()
had not held mutex.
nvme_bdev_first_ctrlr() and nvme_bdev_next_ctrlr() were replaced by
nvme_ctrlr_for_each() in the last patch.
In this patch, add mutex to three helper functions, nvme_ctrlr_get(),
nvme_ctrlr_get_by_name(), and nvme_ctrlr_for_each().
Add mutex to nvme_ctrlr_create() but it will be removed in the
following patches because nvme_ctrlr will be added to not global
ctrlr list but ctrlr list per nvme_bdev_ctrlr.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ife27066d2dcac82db0616b0afeaf68e5705d7da1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8722
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Replace two helper functions, nvme_bdev_first_ctrlr() and
nvme_bdev_next_ctrlr() by an new helper function nvme_ctrlr_for_each().
This will make us easier to guard data structure correctly in the
following patches.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ibd81286e454fd6127fd150a7d48d8381bd1b89d3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8721
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This object aggregates multiple I/O qpairs for their completion
operations and may be a higher layer object. However, the
aggregation is only to poll completions efficiently. Hence if we
follow the new naming rule, nvme_poll_group is better than
nvme_ctrlr_poll_group and nvme_bdev_poll_group, and rename
nvme_bdev_poll_group by nvme_poll_group.
Besides, many functions in NVMe bdev module have a naming rule,
bdev_nvme + verb + objective
Follow this rule for a few functions related with nvme_poll_group.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I5e5cb6001d4a862c2121b7265cbbffe0c2109785
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8720
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This object is per I/O path and will be aggregated by an new upper
layer object.
Hence rename nvme_bdev_ctrlr by nvme_ctrlr. Then the following patches
will add nvme_bdev_ctrlr as a different upper layer object.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ieed634447785cc98140b3d49c52a2c753988ece7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8381
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This object is used for failover and per I/O path. A controller may
have multiple of this object. A controller is per path and may be
aggregated by an new object. Hence this object is a lower layer
object.
Based on the new naming rule, rename nvme_bdev_ctrlr_trid by
nvme_ctrlr_trid.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I0d5e5812560a6947a0c25af05dea168e8745130e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8380
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
This object will be aggregated by the upper layer object nvme_bdev.
Hence based on the new naming rule, rename nvme_bdev_ns by nvme_ns.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I96a70213b29fb53437acd080a0787ec9f5a6759a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8379
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
We will name the lower level objects starting with nvme_* and the
upper level objects starting with nvme_bdev_*.
This object is a channel per ctrlr and another new channel will be
added on top of this object.
Rename nvme_io_path by nvme_ctrlr_channel based on the new naming rule.
nvme_io_path will be used for a new object which is used to find an
optimal I/O path and to reset multiple ctrlrs sequentially when
multipath is supported.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I1d4fa6d4625de3413d629a1ff412e00de12dfaf4
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8378
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Bit 1 in the CMIC of the Identify Controller Data Structure specifies
if the NVM subsystem may have multiple controllers or not.
However, multi_host indicated a particular use case such that the NVM
subsystem is used by multiple hosts.
multi_ctrlr will be more appropriate.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I0246096a5cc44721aeff3ff6f96473a2abe11964
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8719
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Related to #1950
In the issue, CPU idle time measured from /proc/stats make it
seem like no reactor goes idle.
If true, then it would help for debugging to check stats from reactors.
Their busy/idle time and number of threads on them.
This patch reorders the checks in interrupt test,
to first show reactor stats, then confirm with idle time from
/proc/stats.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ia1770833828ef688110473e45ed5998d213059f3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8223
Community-CI: Broadcom 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: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
For cases where cpumask for a thread was not set,
all bits were turned on for whole length of cpuset structure.
This resulted in JSON RPC reponses with way too long cpumask
for what is useful.
Now the response is limited to the applications core mask,
as that makes sense so long as number of cores cannot change.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ib5cf271d3b219ba679f1abe498516796693a87dd
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8288
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Maciej Szwed <maciej.szwed@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
The round-robin logic is no longer necessary to spread
the threads around the cores. Starting from core other
than first is even counter-productive to bunching up
threads.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I5fcee2bacc2d0b4af26336caf381ed954814d731
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8085
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Before this patch _find_optimal_core() returned
1) any core that could fit the thread
2) if current core was over the limit, the least busy core
3) current core if no better candidate was found
Combined with _get_next_target_core() round-robining
the first core to consider, resulted in threads being
unnecessarily spread over the cores.
This patch only places threads on lower lcore id,
or when current core is over limit then any core that can fit it.
Next patch will remove round-robin logic to always start with
lowest lcore id.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I54e373d3ca02a5633607d22978305baa1142f8bd
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8112
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Maciej Szwed <maciej.szwed@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Before this patch the idle time of a core was increased
by the amount of busy time of thread that was moved out.
No assumption was made as to how the remaining threads,
would behave during next scheduling period.
This approach is fine, as over multiple scheduling periods
we'd arrive at a point where threads could do no more work
or all cores would be busy.
Yet this requires multiple scheduling periods to sort out
the threads.
Later in the series core_load will be used to determine,
when to start moving threads out of the core. So changing
this assumption will allow for faster responses to thread load,
at cost of sometimes spreading threads too much briefly.
With this patch, we are assuming that threads remaining
on the core will do proportionally the same amount of work
during next scheduling period.
See an example illustrating the change:
Before moving Thread1
Thread1 Busy 80 Idle 20 Load 80%
Thread2 Busy 60 Idle 40 Load 60%
Core Busy 140 Idle 60 Load 70%
After moving Thread1 out (original code)
Core Busy 140-80=60 Idle 60+80=140 Load 30%
After moving Thread1 out (this patch)
Core Busy 140-80=60 Idle 60-20=40 Load 60%
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I1f347983449b2fde476dab360c4df689965ca3ea
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8279
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Maciej Szwed <maciej.szwed@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
So far the test application was mostly used for cases with
single active thread on a core.
Setting activity as percentage of all poller operations worked
fine for those cases.
This patch changes the thread activity to mean percentage of
a set time period.
Which equals to the maximum load a thread will see when being
alone on the core.
Right now g_core_time_period is set to 100ms, and is the
maximum execution time for all threads on a core when
thread activity is set to 100%.
g_timeslice is set to 100us and is the execution time of
single thread poll.
Consider following scenario:
Before the patch
Thread1 Activity 70% Busy 70 Idle 30 Load 70%
Thread2 Activity 60% Busy 60 Idle 40 Load 60%
Core Busy 130 Idle 70 Load 65%
After the patch
Thread1 Activity 70% Busy 100 Idle 0 Load 100%
Thread2 Activity 60% Busy 100 Idle 0 Load 100%
Core Busy 200 Idle 0 Load 100%
Additionally increased period of sleeping from 1us to 100us,
to offset the time needed for additional operations
during thread poll.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: I027432db1bea570dd58829eeccf7dd8bb06bf249
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7474
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
In cases when all cores are already doing too much work
to fit a thread, active threads should still be balanced
over all cores.
When current core is overloaded, place the thread
on another that is less busy.
The core limit is set to 95% to catch only ones that are
fully busy.
Decreasing that value would make spreading out the threads
move aggressive.
Changed thread load in one of the unit tests to reflect the
95% limit.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I3b3bc5f7fbd22725441fa811d61446950000cc46
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8113
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Maciej Szwed <maciej.szwed@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
We had not held mutex while removing bdev name or alias from bdev
name tree for most cases. Fix these in this patch.
spdk_bdev_unregister() already holds g_bdev_mgr.mutex when removing
name, and so we do not need to change it.
spdk_bdev_close() had not held g_bdev_mgr.mutex. What we want to lock
is only when removing name from name tree, that is, calling
bdev_name_del() in bdev_unregister_unsafe(). However, we need to
keep hierarchical lock ordering. Hence get and free g_bdev_mgr.mutex
outside of bdev->internal.mutex.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I4e2c8604e27c8603725efa9bc0bee2013eccb2ac
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8527
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
We had not held mutex when adding bdev name to global bdev name tree
in bdev_name_add(). Fix these in this patch.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I33813638f11da85263ec0c8849e566d247a45d43
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8524
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
