fi_provider(7) @VERSION@ fi_provider(7)NAMEfi_provider - Fabric Interface Providers
OVERVIEW
Conceptually, a fabric provider may be viewed as a local hardware NIC
driver, though a provider is not limited by this definition. The first
component of libfabric is a general purpose framework that is capable
of handling different types of fabric hardware. All fabric hardware
devices and their software drivers are required to support this frame‐
work. Devices and the drivers that plug into the libfabric framework
are referred to as fabric providers, or simply providers.
This distribution of libfabric contains the following providers
(although more may be available via run-time plug-ins):
Core providers
GNI : A provider for the Aries interconnect in Cray XC(TM) systems uti‐
lizing the user-space Generic Networking Interface. See fi_gni(7) for
more information.
PSM : High-speed InfiniBand networking from Intel. See fi_psm(7) for
more information.
Sockets : A general purpose provider that can be used on any network
that supports TCP/UDP sockets. This provider is not intended to pro‐
vide performance improvements over regular TCP/UDP sockets, but rather
to allow developers to write, test, and debug application code even on
platforms that do not have high-speed networking. See fi_sockets(7)
for more information.
usNIC : Ultra low latency Ethernet networking over Cisco userspace VIC
adapters. See fi_usnic(7) for more information.
Verbs : This provider uses the Linux Verbs API for network transport.
Application performance is, obviously expected to be similar to that of
the native Linux Verbs API. Analogous to the Sockets provider, the
Verbs provider is intended to enable developers to write, test, and
debug application code on platforms that only have Linux Verbs-based
networking. See fi_verbs(7) for more information.
Blue Gene/Q : See fi_bgq(7) for more information.
Utility providers
RxM : The RxM provider (ofi_rxm) is an utility provider that supports
RDM endpoints emulated over MSG endpoints of a core provider. See
fi_rxm(7) for more information.
CORE VERSUS UTILITY PROVIDERS
Core providers implement the libfabric interfaces directly over
low-level hardware and software interfaces. They are designed to sup‐
port a specific class of hardware, and may be limited to supporting a
single NIC. Core providers often only support libfabric features and
interfaces that map efficiently to their underlying hardware.
Utility providers are distinct from core providers in that they are not
associated with specific classes of devices. They instead work with
core providers to expand their features, and interact with core
providers through libfabric interfaces internally. Utility providers
are often used to support a specific endpoint type over a simpler end‐
point type. For example, the RXD provider implements reliability over
unreliable datagram endpoints. The utility providers will not layer
over the sockets provider unless it is explicitly requested.
Utility providers show up as a component in the core provider's compo‐
nent list. See fi_fabric(3). Utility providers are enabled automati‐
cally for core providers that do not support the feature set requested
by an application.
PROVIDER REQUIREMENTS
Libfabric provides a general framework for supporting multiple types of
fabric objects and their related interfaces. Fabric providers have a
large amount of flexibility in selecting which components they are able
and willing to support, based on specific hardware constraints.
Provider developers should refer to docs/provider for information on
functionaliy supplied by the framework to assist in provider implemen‐
tation. To assist in the development of applications, libfabric speci‐
fies the following requirements that must be met by any fabric
provider, if requested by an application.
Note that the instantiation of a specific fabric object is subject to
application configuration parameters and need not meet these require‐
ments.
· A fabric provider must support at least one endpoint type.
· All endpoints must support the message queue data transfer interface
(fi_ops_msg).
· An endpoint that advertises support for a specific endpoint capabil‐
ity must support the corresponding data transfer interface.
· FI_ATOMIC - fi_ops_atomic
· FI_RMA - fi_ops_rma
· FI_TAGGED - fi_ops_tagged
· Endpoints must support all transmit and receive operations for any
data transfer interface that they support.
· Exception: If an operation is only usable for an operation that the
provider does not support, and support for that operation is conveyed
using some other mechanism, the operation may return
· FI_ENOSYS. For example, if the provider does not support injected
data, it can set the attribute inject_size = 0, and fail all
fi_inject operations.
· The framework supplies wrappers around the 'msg' operations that can
be used. For example, the framework implements the sendv() msg oper‐
ation by calling sendmsg(). Providers may reference the general
operation, and supply on the sendmsg() implementation.
· Providers must set all operations to an implementation. Function
pointers may not be left NULL or uninitialized. The framework sup‐
plies empty functions that return -FI_ENOSYS which can be used for
this purpose.
· Endpoints must support the CM interface as follows:
· FI_EP_MSG endpoints must support all CM operations.
· FI_EP_DGRAM endpoints must support CM getname and setname.
· FI_EP_RDM endpoints must support CM getname and setname.
· Providers that support connectionless endpoints must support all AV
operations (fi_ops_av).
· Providers that support memory registration, must support all MR oper‐
ations (fi_ops_mr).
· Providers should support both completion queues and counters.
· If FI_RMA_EVENT is not supported, counter support is limited to local
events only.
· Completion queues must support the FI_CQ_FORMAT_CONTEXT and
FI_CQ_FORMAT_MSG.
· Providers that support FI_REMOTE_CQ_DATA shall support FI_CQ_FOR‐
MAT_DATA.
· Providers that support FI_TAGGED shall support FI_CQ_FORMAT_TAGGED.
· A provider is expected to be forward compatible, and must be able to
be compiled against expanded fi_xxx_ops structures that define new
functions added after the provider was written. Any unknown func‐
tions must be set to NULL.
· Providers shall document in their man page which features they sup‐
port, and any missing requirements.
Future versions of libfabric will automatically enable a more complete
set of features for providers that focus their implementation on a nar‐
row subset of libfabric capabilities.
LOGGING INTERFACE
Logging is performed using the FI_ERR, FI_LOG, and FI_DEBUG macros.
DEFINITIONS
#define FI_ERR(prov_name, subsystem, ...)
#define FI_LOG(prov_name, prov, level, subsystem, ...)
#define FI_DEBUG(prov_name, subsystem, ...)
ARGUMENTS
prov_name : String representing the provider name.
prov : Provider context structure.
level : Log level associated with log statement.
subsystem : Subsystem being logged from.
DESCRIPTION
FI_ERR : Always logged.
FI_LOG : Logged if the intended provider, log level, and subsystem
parameters match the user supplied values.
FI_DEBUG : Logged if configured with the --enable-debug flag.
SEE ALSOfi_gni(7), fi_psm(7), fi_sockets(7), fi_usnic(7), fi_verbs(7),
fi_bgq(7),
AUTHORS
OpenFabrics.
Libfabric Programmer's Manual 2017-07-24 fi_provider(7)
