Network File System Version 4                              C. Lever, Ed.
Internet-Draft                                                    Oracle
Obsoletes: 5666 (if approved)                                 W. Simpson
Intended status: Standards Track                              DayDreamer
Expires: June 16, July 14, 2016                                         T. Talpey
                                                               Microsoft
                                                       December 14, 2015
                                                        January 11, 2016

    Remote Direct Memory Access Transport for Remote Procedure Call
                     draft-ietf-nfsv4-rfc5666bis-01
                     draft-ietf-nfsv4-rfc5666bis-02

Abstract

   This document specifies a protocol for conveying Remote Procedure
   Call (RPC) messages on physical transports capable of Remote Direct
   Memory Access (RDMA).  The RDMA transport binding enables efficient
   bulk-data transport over high-speed networks with minimal change to
   RPC applications.  It requires no revision to application RPC
   protocols or the RPC protocol itself.  This document obsoletes RFC
   5666.

Status of This Memo

   This Internet-Draft is submitted in full conformance with the
   provisions of BCP 78 and BCP 79.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF).  Note that other groups may also distribute
   working documents as Internet-Drafts.  The list of current Internet-
   Drafts is at http://datatracker.ietf.org/drafts/current/.

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time.  It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress."

   This Internet-Draft will expire on June 16, July 14, 2016.

Copyright Notice

   Copyright (c) 2015 2016 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (http://trustee.ietf.org/license-info) in effect on the date of
   publication of this document.  Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.  Code Components extracted from this document must
   include Simplified BSD License text as described in Section 4.e of
   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.

Table of Contents

   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . .   3
     1.1.  Requirements Language . . . . . . . . . . . . . . . . . .   3
     1.2.  RPC Over On RDMA Transports  . . . . . . . . . . . . . . . . .   3
   2.  Changes Since RFC 5666  . . . . . . . . . . . . . . . . . . .   4
     2.1.  Changes To The Specification  . . . . . . . . . . . . . .   4
     2.2.  Changes To The Protocol . . . . . . . . . . . . . . . . .   5   4
   3.  Terminology . . . . . . . . . . . . . . . . . . . . . . . . .   5
     3.1.  Remote Procedure Calls  . . . . . . . . . . . . . . . . .   5
     3.2.  Remote Direct Memory Access . . . . . . . . . . . . . . .   8
   4.  RPC-Over-RDMA Protocol Framework  . . . . . . . . . . . . . .  10
     4.1.  Transfer Models . . . . . . . . . . . . . . . . . . . . .  10
     4.2.  RPC Message Framing . . . . . . . . . . . . . . . . . . .  11
     4.3.  Flow Control  . . . . . . . . . . . . . . . . . . . . . .  11
     4.4.  XDR Encoding With Chunks  . . . . . . . . . . . . . . . .  13
     4.5.  Data Exchange . . . . . . . . . . . . . . . . . . . . . .  19
     4.6.  Message Size  . . . . . . . . . . . . . . . . . . . . . .  21  19
   5.  RPC-Over-RDMA In Operation  . . . . . . . . . . . . . . . . .  23  20
     5.1.  Fixed Header Fields . .  XDR Protocol Definition . . . . . . . . . . . . . . . . .  23  21
     5.2.  Chunk Lists .  Fixed Header Fields . . . . . . . . . . . . . . . . . . .  23
     5.3.  Chunk Lists . . .  24
     5.3.  Forming Messages . . . . . . . . . . . . . . . . . . . .  26  25
     5.4.  Memory Registration . . . . . . . . . . . . . . . . . . .  29  26
     5.5.  Error Handling Errors  . . . . . . . . . . . . . . . . . . . . .  30  28
     5.6.  XDR Language Description  Protocol Elements No Longer Supported . . . . . . . . . .  30
     5.7.  XDR Examples  . . . . . .  31
     5.7.  Deprecated Protocol Elements . . . . . . . . . . . . . .  34
   6.  Upper Layer Binding Specifications . .  31
   6.  RPC Bind Parameters . . . . . . . . . . .  34
     6.1.  Determining DDP-Eligibility . . . . . . . . . .  32
   7.  Bi-Directional RPC-Over-RDMA  . . . . .  35
     6.2.  Write List Ordering . . . . . . . . . . .  34
     7.1.  RPC Direction . . . . . . . .  36
     6.3.  DDP-Eligibility Violation . . . . . . . . . . . . . .  34
     7.2.  Backward Direction Flow Control . .  36
     6.4.  Other Binding Information . . . . . . . . . . .  35
     7.3.  Conventions For Backward Operation  . . . . .  37
   7.  RPC Bind Parameters . . . . . .  36
     7.4.  Backward Direction Upper Layer Binding  . . . . . . . . .  38
   8.  Upper Layer Binding Specifications  . . . . . .  37
   8.  Bi-Directional RPC-Over-RDMA . . . . . . .  39
     8.1.  DDP-Eligibility . . . . . . . . .  38
     8.1.  RPC Direction . . . . . . . . . . . .  39
     8.2.  Maximum Reply Size  . . . . . . . . . .  39
     8.2.  Backward Direction Flow Control . . . . . . . . .  41
     8.3.  Additional Considerations . . . .  40
     8.3.  Conventions For Backward Operation . . . . . . . . . . .  41 .  42
     8.4.  Backward Direction  Upper Layer Binding Protocol Extensions . . . . . . . . .  43 . . . .  42
   9.  Transport Protocol Extensibility  . . . . . . . . . . . . . .  44  42
     9.1.  Bumping The  RPC-over-RDMA Version Numbering . . . . . . . . . . . .  44 .  43
   10. Security Considerations . . . . . . . . . . . . . . . . . . .  45  43
     10.1.  Memory Protection  . . . . . . . . . . . . . . . . . . .  45  43
     10.2.  Using GSS With RPC-Over-RDMA . . . . . . . . . . . . . .  45  44
   11. IANA Considerations . . . . . . . . . . . . . . . . . . . . .  46  45
   12. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . .  47  46
   13. Appendices  . . . . . . . . . . . . . . . . . . . . . . . . .  47
     13.1.  Appendix 1: XDR Examples . . . . . . . . . . . . . . . .  47
   14. References  . . . . . . . . . . . . . . . . . . . . . . . . .  49
     14.1.  46
     13.1.  Normative References . . . . . . . . . . . . . . . . . .  49
     14.2.  46
     13.2.  Informative References . . . . . . . . . . . . . . . . .  50  47
   Authors' Addresses  . . . . . . . . . . . . . . . . . . . . . . .  51  48

1.  Introduction

   This document obsoletes RFC 5666; however, the protocol specified by
   this document is based on existing interoperating implementations of
   the RPC-over-RDMA Version One protocol.  The new specification
   clarifies text that is subject to multiple interpretations and
   eliminates
   removes support for unimplemented RPC-over-RDMA Version One protocol
   elements.  This document makes the role of Upper Layer Bindings an
   explicit part of the specification.  In addition, it this document
   introduces conventions that enable bi-directional RPC-over-RDMA operation.
   operation to allow operation of NFSv4.1 [RFC5661] on RDMA transports.

1.1.  Requirements Language

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
   document are to be interpreted as described in [RFC2119].

1.2.  RPC Over On RDMA Transports

   Remote Direct Memory Access (RDMA) [RFC5040] [RFC5041] [IB] is a
   technique for moving data efficiently between end nodes.  By
   directing data into destination buffers as it is sent on a network,
   and placing it via direct memory access by hardware, the benefits of
   faster transfers and reduced host overhead are obtained.

   Open Network Computing Remote Procedure Call (ONC RPC, or simply,
   RPC) [RFC5531] is a remote procedure call protocol that runs over a
   variety of transports.  Most RPC implementations today use UDP or
   TCP.  On UDP, RPC messages are encapsulated inside datagrams, while
   on a TCP byte stream, RPC messages are delineated by a record marking
   protocol.  An RDMA transport also conveys RPC messages in a specific
   fashion that must be fully described if RPC implementations are to
   interoperate.

   RDMA transports present semantics different from either UDP or TCP.
   They retain message delineations like UDP, but provide a reliable and
   sequenced data transfer like TCP.  They also provide an efficient
   bulk-transfer offloaded
   bulk transfer service not provided by UDP or TCP.  RDMA transports
   are therefore appropriately viewed as a new transport type by RPC.

   RDMA as a transport can enhance the performance of RPC protocols that
   move large quantities of data, since RDMA hardware excels at moving
   data efficiently between host memory and a high-speed network with
   little host CPU involvement.

   In this context, the Network File System (NFS) protocols as described
   in [RFC1094], [RFC1813], [RFC7530], [RFC5661], and future NFSv4 minor
   verions are obvious beneficiaries of RDMA transports.  A complete
   problem statement is discussed in [RFC5532], and NFSv4-related issues
   are discussed in [RFC5661].  Many other RPC-based protocols can also
   benefit.

   Although the RDMA transport described here can provide relatively
   transparent support for any RPC application, this document also
   describes mechanisms that can optimize data transfer further, given
   more active participation by RPC applications.

2.  Changes Since RFC 5666

2.1.  Changes To The Specification

   The following alterations have been made to the RPC-over-RDMA Version
   One specification: specification.  The section numbers below refer to [RFC5666].

   o  Section 2 has been expanded to introduce and explain key RPC, XDR,
      and RDMA terminology.  These terms are now used consistently
      throughout the specification.  This change was necesssary because
      implementers familiar with RDMA are often not familiar with the
      mechanics of RPC, and vice versa.

   o  Section 3 has been re-organized and split into sub-sections to
      facilitate locating
      help implementers locate specific requirements and definitions.

   o  Section  Sections 4 and 5 have been combined for clarity and to improve the
      organization of this information.

   o  The XDR definition of RPC-over-RDMA Version One has been updated
      (without on-the-wire changes) to align with the terms and concepts
      introduced in this specification.

   o  The specification of the optional Connection Configuration
      Protocol has been removed from the specification, as there are no
      known implementations of the protocol.

   o  Sections discussing  A section outlining requirements for Upper Layer Bindings have has been
      added.

   o  A section discussing RPC-over-RDMA protocol extensibility has been
      added.

2.2.  Changes To The Protocol

   While the protocol described herein interoperates with existing
   implementations of [RFC5666], the following changes have been made
   relative to the protocol described in that document:

   o  Support for the Read-Read transfer model has been removed.  Read-
      Read is a slower transfer model than Read-Write, thus implementers
      have chosen not to support it.

   o  Support  This simplifies explanatory text,
      and support for sending the RDMA_MSGP RDMA_DONE message type has been
      deprecated.  This document instructs senders not to use it, but
      receivers must continue to recognize it.

      RDMA_MSGP has is no benefit for RPC programs that place bulk payload
      items at positions other than at the end longer necessary.

   o  The specification of RDMA_MSGP in [RFC5666] and current
      implementations of their argument or
      result lists, as is common with NFSv4 COMPOUND RPCs [RFC7530].
      Similarly it is not beneficial when a connection's inline
      threshold is significantly smaller than are incomplete.  Therefore the system page size, as RDMA_MSGP
      message type is typical for no longer supported.

   o  Technical errors with regard to handling RPC-over-RDMA Version One implementations. header
      errors have been corrected.

   o  Specific requirements related to handling XDR round-up and
      abstract data types have been added.  Responders are now forbidden
      from writing Write chunk round-up bytes.

   o  Clear guidance about Send and Receive buffer size has been added.
      This enables better decisions about when to provide and use the
      Reply chunk.

   o  A section specifying bi-directional RPC operation on RPC-over-RDMA
      has been added.  This enables the NFSv4.1 backchannel [RFC5661] backchannel on
      RPC-over-RDMA Version One transports when both endpoints support
      the new functionality.

   The protocol version number has not been changed because the protocol
   specified in this document fully interoperates with implementations
   of the RPC-over-RDMA Version One protocol specified in [RFC5666].

3.  Terminology

3.1.  Remote Procedure Calls

   This section introduces key elements of the Remote Procedure Call
   [RFC5531] and External Data Representation [RFC4506] protocols protocols, upon
   which RPC-over-RDMA Version One is constructed.

3.1.1.  Upper Layer Protocols

   Remote Procedure Calls are an abstraction used to implement the
   operations of an "upper layer protocol," "Upper Layer Protocol," sometimes referred to as a
   ULP.  The term Upper Layer Protocol refers to an RPC Program and
   Version tuple, which is a versioned set of procedure calls that
   comprise a single well-defined API.  One example of such a protocol an Upper Layer
   Protocol is the Network File System Version 4.0 [RFC7530].

3.1.2.  Requesters And Responders

   Like a local procedure call, every Remote Procedure Call has a set of
   "arguments" and a set of "results".  A calling context is not allowed
   to proceed until the procedure's results are available to it.  Unlike
   a local procedure call, the called procedure is executed remotely
   rather than in the local application's context.

   The RPC protocol as described in [RFC5531] is fundamentally a
   message-passing protocol between one server and one or more clients.
   ONC RPC transactions are made up of two types of messages:

   CALL Message
      A CALL message, or "Call", requests that work be done.  A Call is
      designated by the value CALL in the message's msg_type field.  An
      arbitrary unique value is placed in the message's xid field.

   REPLY Message
      A REPLY message, or "Reply", reports the results of work requested
      by a Call.  A Reply is designated by the value REPLY in the
      message's msg_type field.  The value contained in the message's
      xid field is copied from the Call whose results are being
      reported.

   An RPC client endpoint, or "requester", serializes an RPC call's
   arguments and conveys them to a server endpoint via an RPC call
   message.  This message contains an RPC protocol header, a header
   describing the requested upper layer operation, and all arguments.

   The server endpoint, or "responder", deserializes the arguments and
   processes the requested operation.  It then serializes the
   operation's results into another byte stream.  This byte stream is
   conveyed back to the requester via an RPC reply message.  This
   message contains an RPC protocol header, a header describing the
   upper layer reply, and all results.

   The requester deserializes the results and allows the original caller
   to proceed.

   RPC-over-RDMA is a connection-oriented  At this point the RPC transport.  When a
   connection-oriented transport transaction designated by the xid
   in the call message is used, ONC terminated and the xid is retired.

3.1.3.  RPC client endpoints are
   responsible for initiating Transports

   The role of an "RPC transport" is to mediate the exchange of RPC
   messages between requesters and responders.  An RPC transport bridges
   the gap between the RPC message abstraction and the native operations
   of a particular network transport.

   RPC-over-RDMA is a connection-oriented RPC transport.  When a
   connection-oriented transport is used, ONC RPC client endpoints are
   responsible for initiating transport connections, while ONC RPC
   service endpoints wait passively for incoming connection requests.

3.1.3.

3.1.4.  External Data Representation

   In a heterogenous environment, one cannot assume that all requesters and
   responders represent data the same way.  RPC uses eXternal Data
   Representation, or XDR, to translate data types and serialize
   arguments and results. results [RFC4506].

   The XDR protocol encodes data independent of the endianness or size
   of host-native data types, allowing unambiguous decoding of data on
   the receiving end.  RPC programs are specified by writing an XDR
   definition of their procedures, argument data types, and result data
   types.

   XDR assumes that the number of bits in a byte (octet) and their order
   are the same on both endpoints and on the physical network.  The
   smallest indivisible unit of XDR encoding is a group of four octets
   in little-endian order.  XDR also flattens lists, arrays, and other
   abstract
   complex data types so they can be conveyed as a simple stream of bytes.

   A serialized stream of bytes that is the result of XDR encoding is
   referred to as an "XDR stream."  A sending endpoint encodes native
   data into an XDR stream and then transmits that stream to a receiver.
   A receiving endpoint decodes incoming XDR byte streams into its
   native data representation format.

   The function of an RPC transport is to convey RPC messages, each
   encoded as a separate XDR stream, from one endpoint to another.

3.1.3.1.

3.1.4.1.  XDR Opaque Data

   Sometimes a data item must be transferred as-is, without encoding or
   decoding.  Such a data item is referred to as "opaque data."  XDR
   encoding places opaque data items directly into an XDR stream without
   altering its content in any way.  Upper Layer Protocols or
   applications perform any needed data translation in this case.
   Examples of opaque data items include the contents of files, and
   generic byte strings.

3.1.3.2.

3.1.4.2.  XDR Round-up

   The number of octets in a variable-size data item precedes that item
   in the encoding stream.  If the size of an encoded data item is not a
   multiple of four octets, octets containing zero are added to the end
   of the item as it is encoded so that the next encoded data item
   starts on a four-octet boundary.  The encoded size of the item is not
   changed by the addition of the extra octets. octets, and the zero bytes are
   not exposed to the Upper Layer.

   This technique is referred to as "XDR round-up," and the extra octets
   are referred to as "XDR padding".  The content of XDR pad octets is
   ignored by receivers.

3.2.  Remote Direct Memory Access

   RPC requesters and responders can be made more efficient if large RPC
   messages are transferred by a third party such as intelligent network
   interface hardware (data movement offload), and placed in the
   receiver's memory so that no additional adjustment of data alignment
   has to be made (direct data placement).  Remote Direct Memory Access
   enables both optimizations.

3.2.1.  Direct Data Placement

   Very often,

   Typically, RPC implementations copy the contents of RPC messages into
   a buffer before being sent.  An efficient RPC implementation sends
   bulk data without copying it into a separate send buffer first.

   However, socket-based RPC implementations are often unable to receive
   data directly into its final place in memory.  Receivers often need
   to copy incoming data to finish an RPC operation; sometimes, only to
   adjust data alignment.

   In this document, "RDMA" refers to the physical mechanism an RDMA
   transport utilizes when moving data.  Although it this may not be
   efficient, before an RDMA transfer a sender may copy data into an
   intermediate buffer before an RDMA transfer.  After an RDMA transfer,
   a receiver may copy that data again to its final destination.

   This document uses the term "direct data placement" (or DDP) to refer
   specifically to an optimized data transfer where it is unnecessary
   for a receiving host's CPU to copy transferred data to another
   location after it has been received.  Not all RDMA-based data
   transfer qualifies as Direct Data Placement, and DDP can be achieved
   using non-RDMA mechanisms.

3.2.2.  RDMA Transport Requirements

   The RPC-over-RDMA Version One protocol assumes the physical transport
   provides the following abstract operations.  A more complete
   discussion of these operations is found in [RFC5040].

   Registered Memory
      Registered memory is a segment of memory that is assigned a
      steering tag that temporarily permits access by the RDMA provider
      to perform data transfer operations.  The RPC-over-RDMA Version
      One protocol assumes that each segment of registered memory MUST
      be identified with a steering tag of no more than 32 bits and
      memory addresses of up to 64 bits in length.

   RDMA Send
      The RDMA provider supports an RDMA Send operation, with completion
      signaled on the receiving peer after data has been placed in a
      pre-posted memory segment.  Sends complete at the receiver in the
      order they were issued at the sender.  The amount of data
      transferred by an RDMA Send operation is limited by the size of
      the remote pre-posted memory segment.

   RDMA Receive
      The RDMA provider supports an RDMA Receive operation to receive
      data conveyed by incoming RDMA Send operations.  To reduce the
      amount of memory that must remain pinned awaiting incoming Sends,
      the amount of pre-posted memory is limited.  Flow-control to
      prevent overrunning receiver resources is provided by the RDMA
      consumer (in this case, the RPC-over-RDMA Version One protocol).

   RDMA Write
      The RDMA provider supports an RDMA Write operation to directly
      place data in remote memory.  The local host initiates an RDMA
      Write, and completion is signaled there; no there.  No completion is
      signaled on the remote.  The local host provides a steering tag,
      memory address, and length of the remote's memory segment.

      RDMA Writes are not necessarily ordered with respect to one
      another, but are ordered with respect to RDMA Sends.  A subsequent
      RDMA Send completion obtained at the write initiator guarantees
      that prior RDMA Write data has been successfully placed in the
      remote peer's memory.

   RDMA Read
      The RDMA provider supports an RDMA Read operation to directly
      place peer source data in the read initiator's memory.  The local
      host initiates an RDMA Read, and completion is signaled there; no
      completion is signaled on the remote.  The local host provides
      steering tags, memory addresses, and a length for the remote
      source and local destination memory segments.

      The remote peer receives no notification of RDMA Read completion.
      The local host signals completion as part of an RDMA Send message
      so that the remote peer can release steering tags and subsequently
      free associated source memory segments.

   The RPC-over-RDMA Version One protocol is designed to be carried over
   RDMA transports that support the above abstract operations.  This
   protocol conveys to the RPC peer information sufficient for that RPC
   peer to direct an RDMA layer to perform transfers containing RPC data
   and to communicate their result(s).  For example, it is readily
   carried over RDMA transports such as Internet Wide Area RDMA Protocol
   (iWARP) [RFC5040] [RFC5041].

4.  RPC-Over-RDMA Protocol Framework

4.1.  Transfer Models

   A "transfer model" designates which endpoint is responsible for
   performing RDMA Read and Write operations.  To enable these
   operations, the peer endpoint first exposes segments of its memory to
   the endpoint performing the RDMA Read and Write operations.

   Read-Read
      Requesters expose their memory to the responder, and the responder
      exposes its memory to requesters.  The responder employs RDMA Read
      operations to convey pull RPC arguments or whole RPC calls. calls from the
      requester.  Requesters employ RDMA Read operations to convey pull RPC
      results or whole RPC
      relies. relies from the responder.

   Write-Write
      Requesters expose their memory to the responder, and the responder
      exposes its memory to requesters.  Requesters employ RDMA Write
      operations to convey push RPC arguments or whole RPC calls. calls to the
      responder.  The responder employs RDMA Write operations to convey push
      RPC results or whole RPC relies. relies to the requester.

   Read-Write
      Requesters expose their memory to the responder, but the responder
      does not expose its memory.  The responder employs RDMA Read
      operations to convey pull RPC arguments or whole RPC calls. calls from the
      requester.  The responder employs RDMA Write operations to convey push
      RPC results or whole RPC relies. relies to the requester.

   Write-Read
      The responder exposes its memory to requesters, but requesters do
      not expose their memory.  Requesters employ RDMA Write operations
      to convey push RPC arguments or whole RPC calls. calls to the responder.
      Requesters employ RDMA Read operations to convey pull RPC results or
      whole RPC relies. relies from the responder.

   [RFC5666] specifies the use of both the Read-Read and the Read-Write
   Transfer Model.  All current RPC-over-RDMA Version One
   implementations use the Read-Write Transfer Model.  Use of the Read-
   Read Transfer Model by RPC-over-RDMA Version One implementations is
   no longer supported.  Other Transfer Models may be used by a future
   version of RPC-over-RDMA.

4.2.  RPC Message Framing

   During transmission, the XDR stream containing

   On an RPC-over-RDMA transport, each RPC message is
   preceded encapsulated by an
   RPC-over-RDMA header. message.  An RPC-over-RDMA message consists of two XDR
   streams.

   Transport-Specific Stream
      The "transport-specific XDR stream," or "Transport stream,"
      contains an RPC-over-RDMA header that describes and controls the
      transfer of the Payload stream in this RPC-over-RDMA message.
      This header is analogous to the record marking used for RPC over
      TCP but is more extensive, since RDMA transports support several
      modes of data transfer.

   All transfers of an

   RPC message begin with an RDMA Send that
   transfers an RPC-over-RDMA header and part Payload XDR Stream
      The "RPC payload stream," or all of "Payload stream", contains the accompanying
      encapsulated RPC message being transferred by this RPC-over-RDMA
      message.  Because the size

   In its simplest form, an RPC-over-RDMA message consists of what may be transmitted via RDMA
   Send is limited a
   Transport stream followed immediately by the size of the receiver's pre-posted buffers, the
   RPC-over-RDMA transport provides a number of methods to reduce the
   amount transferred via Payload stream conveyed
   together in a single RDMA Send.  Parts of  To transmit large RPC messages not
   transferred via messages, a
   combination of one RDMA Send are transferred using operation and one or more RDMA Read or RDMA
   Write operations. operations is employed.

   RPC-over-RDMA framing replaces all other RPC framing (such as TCP
   record marking) when used atop an RPC-over-RDMA association, even
   when the underlying RDMA protocol may itself be layered atop a
   transport with a defined RPC framing (such as TCP).

   It is however possible for RPC-over-RDMA to be dynamically enabled in
   the course of negotiating the use of RDMA via an Upper Layer Protocol
   exchange.  Because RPC framing delimits an entire RPC request or
   reply, the resulting shift in framing must occur between distinct RPC
   messages, and in concert with the underlying transport.

4.3.  Flow Control

   It is critical to provide RDMA Send flow control for an RDMA
   connection.  RDMA receive operations can fail if a pre-posted receive
   buffer is not available to accept an incoming RDMA Send, and repeated
   occurrences of such errors can be fatal to the connection.  This is a
   departure from conventional TCP/IP networking where buffers are
   allocated dynamically as part of receiving messages.

   Flow control for RDMA Send operations directed to the responder is
   implemented as a simple request/grant protocol in the RPC-over-RDMA
   header associated with each RPC message (Section 5.1.3 5.2.3 has details).

   o  The RPC-over-RDMA header for RPC call messages contains a
      requested credit value for the responder.  This is the maximum
      number of RPC replies the requester can handle at once,
      independent of how many RPCs are in flight at that moment.  The
      requester MAY dynamically adjust the requested credit value to
      match its expected needs.

   o  The RPC-over-RDMA header for RPC reply messages provides the
      granted result.  This is the maximum number of RPC calls the
      responder can handle at once, without regard to how many RPCs are
      in flight at that moment.  The granted value MUST NOT be zero,
      since such a value would result in deadlock.  The responder MAY
      dynamically adjust the granted credit value to match its needs or
      policies (e.g. to accommodate the available resources in a shared
      receive queue).

   The requester MUST NOT send unacknowledged requests in excess of this
   granted responder credit limit.  If the limit is exceeded, the RDMA
   layer may signal an error, possibly terminating the connection.  Even
   if an RDMA layer error does not occur, the responder MAY handle
   excess requests or return an RPC layer error to the requester.

   While RPC calls complete in any order, the current flow control limit
   at the responder is known to the requester from the Send ordering
   properties.  It is always the lower of the requested and granted
   credit values, minus the number of requests in flight.  Advertised
   credit values are not altered because when individual RPCs are started or
   completed.

   On occasion a requester or responder may need to adjust the amount of
   resources available to a connection.  When this happens, the
   responder needs to ensure that a credit increase is effected (i.e.
   receives are posted) before the next reply is sent.

   Certain RDMA implementations may impose additional flow control
   restrictions, such as limits on RDMA Read operations in progress at
   the responder.  Accommodation of such restrictions is considered the
   responsibility of each RPC-over-RDMA Version One implementation.

4.3.1.  Initial Connection State

   There are two operational parameters for each connection:

   Credit Limit
      As described above, the total number of responder receive buffers
      is sometimes referred to as a connection's credit limit.  The
      credit limit is advertised in the RPC-over-RDMA header in each RPC
      message, and can change during the lifetime of a connection.

   Inline Threshold
      The size of the
      A receiver's smallest posted receive buffer "inline threshold" value is the largest size message size
      (in bytes) that a sender can convey with be conveyed via an RDMA Send
      operation, and is known as a connection's "inline threshold." Send/Receive
      combination.  Each connection has two inline threshold values, one
      for each peer receiver.

      Unlike the connection's credit limit, the inline threshold value
      is values are
      not advertised to peers via the RPC-over-RDMA Version One
      protocol, and there is no provision for the inline threshold value
      to change during the lifetime of an RPC-over-RDMA Version One
      connection.  Connection peers MAY have different inline
      thresholds.

   The longevity of a transport connection requires that sending
   endpoints respect the resource limits of peer receivers.  However,
   when a connection is first established, peers cannot know how many
   receive buffers the other has, nor how large the buffers are.

   As a basis for an initial exchange of RPC requests, each RPC-over-
   RDMA Version One connection provides the ability to exchange at least
   one RPC message at a time that is 1024 bytes in size.  A responder
   MAY exceed this basic level of configuration, but a requester MUST
   NOT assume more than one credit is available, and MUST receive a
   valid reply from the responder carrying the actual number of
   available credits, prior to sending its next request.

   Implementations

   Receiver implementations MUST support an inline threshold of 1024
   bytes, but MAY support larger inline thresholds.  In the absense of a thresholds values.  A mechanism
   for discovering a peer's inline threshold, threshold value before a connection
   is established may be used to optimize Send operations.  In the
   absense of such a mechanism, senders MUST assume a receiver's inline
   threshold is 1024 bytes.

4.4.  XDR Encoding With Chunks

   On traditional RPC transports,

   XDR data items in an RPC message are encoded as a contiguous sequence
   of bytes for network transmission.
   However, in  This sequence of bytes is known
   as an XDR stream.  In the case of an RDMA transport, during XDR
   encoding it can be determined that (for instance) an opaque byte array XDR data item is large enough to
   that it might be moved via an RDMA Read or RDMA Write operation. more efficient if the transport placed the content
   of the data item directly in the receiver's memory.

4.4.1.  Reducing An XDR Stream

   RPC-over-RDMA Version One provides a mechanism for moving part of an
   RPC message via a data transfer separate from an RDMA Send/Receive.
   The sender removes one or more XDR data transfer. items from the Payload
   stream.  They are conveyed via one or more RDMA Read or Write
   operations.  The receiver inserts the data items into the Payload
   stream before passing it to the Upper Layer.

   A contiguous piece of an
   XDR a Payload stream that is split out and moved
   via a separate RDMA operation operations is known as a "chunk."  The sender removes the chunk data out from
   the XDR stream conveyed via RDMA Send, and the receiver inserts it
   before handing the reconstructed  A Payload stream
   after chunks have been removed is referred to the Upper Layer.

4.4.1. as a "reduced" Payload
   stream.

4.4.2.  DDP-Eligibility

   Only an XDR data item that might benefit from Direct Data Placement
   should
   may be moved to a chunk. reduced.  The eligibility of specific particular XDR data items to be moved as a chunk, as opposed to being left in the XDR
   stream,
   reduced is not specified by this document.  A

   To maintain interoperability on an RPC-over-RDMA transport, a
   determination must be
   made for of which XDR data items in each Upper Layer Protocol which items in its XDR definition
   are allowed to use Direct Data Placement.  Therefore an additional
   specification is needed that describes how an Upper Layer Protocol
   enables Direct Data Placement.  The set of requirements for a ULP an Upper
   Layer Protocol to use an RDMA RPC-over-RDMA transport is known as an
   "Upper Layer Binding"
   specification, Binding specification," or ULB.

   An Upper Layer Binding specification states which specific individual
   XDR data items in an Upper Layer Protocol MAY be transferred via
   Direct Data Placement.  This document will refer to such XDR data items
   that are permitted to be reduced as "DDP-
   eligible". "DDP-eligible".  All other XDR
   data items MUST NOT be placed in a chunk. reduced.  RPC-over-RDMA Version One uses RDMA
   Read and Write operations to transfer DDP-eligible data that has been placed in chunks.

   The details and
   reduced.

   Detailed requirements for Upper Layer Bindings are discussed in full
   in Section 6.

4.4.2. 8.

4.4.3.  RDMA Segments

   When encoding an RPC message a Payload stream that contains a DDP-eligible data
   item,
   the RPC-over-RDMA transport a sender may choose to reduce that data item.  It does not
   place the item into the RPC
   message's XDR Payload stream.  Instead, it the sender records
   in the RPC-over-RDMA header the actual address and size of the memory
   region containing the that data item.

   The requester sends this provides location information for DDP-eligible data
   items in both RPC calls and replies.  The responder uses this
   information to initiate RDMA Read and Write operations on to retrieve or
   update the memory
   regions. content of the requester's memory.

   An "RDMA segment", or just "segment", is an RPC-over-RDMA header data
   object that contain contains the precise co-ordinates of a contiguous memory
   region that is to be conveyed via one or more RDMA Read or RDMA Write
   operations.  The following fields are contained in a each segment:

   Handle
      Steering tag or handle obtained when the segment's memory is
      registered for RDMA.  Sometimes known as an R_key.

   Length
      The length of the segment in bytes.

   Offset
      The offset or beginning memory address of the segment.

   See [RFC5040] for further discussion of the meaning of these fields.

4.4.3.

4.4.4.  Chunks

   A

   In RPC-over-RDMA Version One, a "chunk" refers to a portion of XDR the
   Payload stream data that is moved via RDMA Read or Write operations.
   Chunk data is removed from the sender's XDR Payload stream, transferred
   by separate RDMA operations, and then re-inserted into the receiver's XDR
   Payload stream.

   Each chunk consists of one or more RDMA segments.  Each segment
   represents a single contiguous piece of that chunk.

   Except in special cases, a chunk contains exactly one XDR data item.
   This makes it straightforward to remove chunks from an XDR stream
   without affecting XDR alignment.

      +----------------+ +----------------+------------------
      | RPC-over-RDMA  | |                |
      |    header w/   | |   RPC Header   | Non-chunk args/results
      |    segments    | |                |
      +----------------+ +----------------+------------------
               |
               +-> Chunk A
               +-> Chunk B
               +-> Chunk C
                    . . .

                 Block diagram of an RPC-over-RDMA message  Not every message has chunks
   associated with it.  The structure of
   the RPC-over-RDMA header is covered in Section 5.

4.4.3.1.

4.4.4.1.  Counted Arrays

   If a chunk is to move contains a counted array data type, the count of array
   elements MUST remain in the XDR Payload stream, while the array elements
   MUST be moved to the chunk.  For example, when encoding an opaque
   byte array as a chunk, the count of bytes stays in the XDR Payload
   stream, while the bytes in the array are removed from the XDR Payload
   stream and transferred via within the chunk.

   Any byte count left in the XDR Payload stream MUST match the sum of the
   lengths of the segments making up the chunk.  If they do not agree,
   an RPC protocol encoding error results.

   Individual array elements appear in the a chunk in their entirety.  For
   example, when encoding an array of arrays as a chunk, the count of
   items in the enclosing array stays in the XDR Payload stream, but each
   enclosed array, including its item count, is transferred as part of
   the chunk.

4.4.3.2.

4.4.4.2.  Optional-data And Unions

   If a chunk is to move contains an optional-data data type, the "is present"
   field MUST remain in the XDR Payload stream, while the data, if present,
   MUST be moved to the chunk.

4.4.4.3.  XDR Unions

   A union data type should never be made DDP-eligible, but one or more
   of its arms may be DDP-eligible.

4.4.4.

4.4.5.  Read Chunks

   A "Read chunk" represents an XDR data item that is to be pulled from
   the requester to the responder using RDMA Read operations.

   A Read chunk is a list of one or more RDMA segments.  Each RDMA
   segment in a Read chunk has an additional Position field.

   Position
      For data that is to be encoded, the
      The byte offset in the RPC message
      XDR Payload stream where the receiver re-inserts re-
      inserts the chunk data. data item conveyed in a chunk.  The byte
      offset Position value
      MUST be computed from the beginning of the RPC message, not
      the beginning of the RPC-over-RDMA header. Payload stream, which
      begins at Position zero.  All segments belonging to the same Read
      chunk have the same value in their Position field.

   While constructing the RPC call, the requester registers memory an RPC-over-RDMA Call message, a requester
   registers memory segments containing data in Read chunks.  It
   advertises these chunks in the RPC-over-RDMA header of the RPC call.

   After receiving the an RPC call sent via an RDMA Send operation, the a
   responder transfers the chunk data from the requester using RDMA Read
   operations.  The responder reconstructs the transferred chunk data by
   concatenating the contents of each segment, in list order, into the
   RPC call XDR stream.  The first segment begins
   received Payload stream at the XDR position Position value recorded in the
   segment.

   Put another way, a receiver inserts the first segment in a Read chunk
   into the Payload stream at the byte offset indicated by its Position field, and subsequent segments
   field.  Segments whose Position field value match this offset are
   concatenated
   afterwards afterwards, until there are no more segments left at that
   Position value.  The next XDR
   Position.

4.4.4.1. data item in the Payload stream
   follows.

4.4.5.1.  Read Chunk Round-up

   XDR requires each encoded data item to start on four-byte alignment.
   When an odd-length data item is marshaled, its length is encoded
   literally, while the data is padded so the next data item in the XDR
   stream can start on a four-byte boundary in the XDR stream. boundary.  Receivers ignore the
   content of the pad bytes.

   Data

   After an XDR data item has been reduced, all data items remaining in
   the XDR Payload stream must all continue to adhere to the above these padding
   requirements.  When a Read chunk is removed from  Thus when an XDR
   stream, data item is moved from the Payload
   stream into a Read chunk, the requester MUST remove any needed XDR padding for
   that
   chunk as well.  Alignment of the items remaining in data item from the Payload stream is
   unaffected. as well.

   The length of a Read chunk is the sum of the lengths of the segments
   that comprise it.  If this sum is not a multiple of four, the
   requester MAY choose to send a Read chunk without any XDR padding.
   The responder MUST be prepared to provide appropriate round-up in its the
   reconstructed call XDR stream if the requester provides no actual
   round-up in a Read chunk.

   The Position field in read segments indicates where the containing
   Read chunk starts in the RPC message XDR stream.  The value in this
   field MUST be a multiple of four.  Moreover, all segments in the same
   Read chunk share the same Position value, even if one or more of the
   segments have a non-four-byte aligned length.

4.4.4.2.

4.4.5.2.  Decoding Read Chunks

   XDR decoding moves data from an XDR stream into a data structure
   provided by an RPC application.  Where elements of the destination
   data structure are buffers or strings, the RPC application can either
   pre-allocate storage to receive the data, or leave the string or
   buffer fields null and allow the XDR decode stage of RPC processing
   to automatically allocate storage of sufficient size.

   When decoding a message from an RDMA transport, RPC-over-RDMA message, the receiver responder first decodes
   the chunk lists from the RPC-over-RDMA header, then proceeds to
   decode the body of the RPC message. Payload stream.  Whenever the XDR offset in the decode Payload
   stream matches that of a Read chunk, the transport initiates an RDMA
   Read to bring over the chunk data into locally registered memory for
   the destination buffer.

   When processing an RPC request, the

   The responder acknowledges its completion of use of the Read chunk source
   buffers by simply replying when it replies to the requester.  The requester may then free all source buffers
   release Read chunks advertised
   by in the request.

4.4.5.

4.4.6.  Write Chunks

   A "Write chunk" represents an XDR data item that is to be pushed from
   the
   a responder to the a requester using RDMA Write operations.

   A Write chunk is an array of one or more RDMA segments.  Segments in
   a Write chunk do not have a Position field because Write chunks are
   provided by a requester long before the responder prepares has prepared the
   reply
   XDR Payload stream.

   While constructing the an RPC call, the call message, a requester also sets up prepares
   memory
   segments regions to catch DDP-eligible reply data.  The data items.  A requester provides as
   many segments as needed to accommodate
   does not know the largest possible size actual length of the result data item in each to be
   returned, thus it MUST register a Write chunk.

   The responder transfers the chunk data long enough to
   accommodate the requester using RDMA
   Write operations.  The maximum possible size of the returned data item.

   A responder copies the responder's requester-provided Write chunk segments into
   the RPC-over-RDMA header to be sent that it returns with the reply.  The
   responder updates the segment length fields to reflect the actual
   amount of data that is being returned in the Write chunk.  The
   updated length of a Write chunk segment MAY be zero if the segment
   was not filled by the responder.  However the responder MUST NOT
   change the number of segments in the Write chunk.

   The responder then sends the RPC reply via an RDMA Send operation.
   After receiving the RPC reply, the requester reconstructs the
   transferred data by concatenating the contents of each segment, in
   array order, into RPC reply XDR stream.

4.4.5.1.

4.4.6.1.  Unused Write Chunks

   There are occasions when a requester provides a Write chunk but the
   responder does not use it.  For example, an Upper Layer Protocol may
   have
   define a union result where some arms of the union contain a DDP-
   eligible data item, and other arms do not.  To return an unused Write
   chunk, the responder MUST set the length of all segments in the chunk
   to zero.

   Unused write chunks, or unused bytes in write chunk segments, are not
   returned as results and their memory is returned to the Upper Layer
   as part of RPC completion.  However, the RPC layer MUST NOT assume
   that the buffers have not been modified.

4.4.5.2.

4.4.6.2.  Write Chunk Round-up

   XDR requires each encoded data item to start on four-byte alignment.
   When an odd-length data item is marshaled, its length is encoded
   literally, while the data is padded so the next data item in the XDR
   stream can start on a four-byte boundary in the XDR stream. boundary.  Receivers ignore the
   content of the pad bytes.

   Data

   After a data item is reduced, data items remaining in the XDR Payload
   stream must all continue to adhere to the above these padding requirements.  When a Write chunk is removed from  Thus
   when an XDR
   stream, data item is moved from a reply Payload stream into a
   Write chunk, the requester responder MUST remove any needed XDR padding for that
   chunk as well.  Alignment of the items remaining in data item
   from the reply Payload stream is
   unaffected.

   The length of a Write chunk is the sum of the lengths of the segments
   that comprise it.  If this sum is not a multiple of four, the
   responder MAY choose not to write XDR padding.  The as well.

   A requester does
   not know the actual SHOULD NOT provide extra length of in a Write chunk when it is prepared, but
   it SHOULD provide enough segments to
   accommodate any needed XDR
   padding.  The requester pad bytes.  A responder MUST be prepared to provide appropriate
   round-up in its reconstructed NOT write XDR stream if the responder provides no
   actual round-up in pad bytes
   for a Write chunk.

4.5.  Data Exchange

   In summary, there are three mechanisms for moving data between
   requester and responder.

   Inline
      Data is moved between requester and responder via an  Message Size

   A receiver of RDMA Send
      operation.

   RDMA Read
      Data is moved between requester and responder via an RDMA Read
      operation.  Address and offset are obtained from a Read chunk in
      the requester's RPC call message.

   RDMA Write
      Data is moved from responder to requester via an RDMA Write
      operation.  Address and offset are obtained from a Write chunk in
      the requester's RPC call message.

   Many combinations are possible.  For instance, an RPC call may
   contain some inline data along with Read or Write chunks.  The reply
   to that call may have chunks that the responder RDMA Writes back to
   the requester.  The following diagrams illustrate RPC calls that use
   these methods to move RPC message data.

        Requester                             Responder
            |               RPC Call              |
       Send |   ------------------------------>   |
            |                                     |
            |               RPC Reply             |
            |   <------------------------------   | Send

            An RPC with no chunks in the call or reply messages

       Requester                             Responder
           |        RPC Call + Write chunks      |
      Send |   ------------------------------>   |
           |                                     |
           |               Chunk 1               |
           |   <------------------------------   | Write
           |                  :                  |
           |               Chunk n               |
           |   <------------------------------   | Write
           |                                     |
           |               RPC Reply             |
           |   <------------------------------   | Send

               An RPC with write chunks in the call message

   In the presence of write chunks, RDMA ordering guarantees that all
   data in the RDMA Write operations has been placed in memory prior to
   the requester's RPC reply processing.

       Requester                             Responder
           |        RPC Call + Read chunks       |
      Send |   ------------------------------>   |
           |                                     |
           |               Chunk 1               |
           |   +------------------------------   | Read
           |   v----------------------------->   |
           |                  :                  |
           |               Chunk n               |
           |   +------------------------------   | Read
           |   v----------------------------->   |
           |                                     |
           |               RPC Reply             |
           |   <------------------------------   | Send

                An RPC with read chunks in the call message

       Requester                             Responder
           |   RPC Call + Read and Write chunks  |
      Send |   ------------------------------>   |
           |                                     |
           |             Read chunk 1            |
           |   +------------------------------   | Read
           |   v----------------------------->   |
           |                  :                  |
           |             Read chunk n            |
           |   +------------------------------   | Read
           |   v----------------------------->   |
           |                                     |
           |             Write chunk 1           |
           |   <------------------------------   | Write
           |                  :                  |
           |             Write chunk n           |
           |   <------------------------------   | Write
           |                                     |
           |               RPC Reply             |
           |   <------------------------------   | Send

           An RPC with read and write chunks in the call message

4.6.  Message Size

   The receiver of RDMA Send operations operations is required by RDMA to have
   previously posted one or more adequately sized buffers (see
   Section 4.3.1). buffers.  Memory
   savings can be achieved on both requesters and responders by leaving
   the inline threshold small.

4.6.1.

4.5.1.  Short Messages

   RPC messages are frequently smaller than the connection's typical inline
   threshold. thresholds.
   For example, the NFS version 3 GETATTR request is only 56 bytes: 20
   bytes of RPC header, plus a 32-byte file handle argument and 4 bytes
   for its length.  The reply to this common request is about 100 bytes.

   Since all RPC messages conveyed via RPC-over-RDMA require an RDMA
   Send operation, the most efficient way to send an RPC message that is
   smaller than the connection's receiver's inline threshold is to append its XDR the Payload
   stream directly to the buffer carrying the RPC-over-RDMA header. Transport stream.  An RPC-over-RDMA header
   with a small RPC call or reply message immediately following is
   transferred using a single RDMA Send operation.  No RDMA Read or
   Write operations are needed.

4.6.2.

4.5.2.  Chunked Messages

   If DDP-eligible data items are present in an RPC message, a Payload stream, a sender
   MAY remove them from reduce the RPC message, Payload stream and use RDMA Read or Write operations
   to move that data.  The RPC-over-RDMA header with the
   shortened RPC call or reply message reduced data items.  The Transport stream with the
   reduced Payload stream immediately following is transferred using a
   single RDMA Send operation.  Removed DDP-eligible
   data items are conveyed using

   After receiving the Transport and Payload streams of a Chunked RPC-
   over-RDMA Call message, the responder uses RDMA Read or Write operations using
   additional information provided to
   move reduced data items in Read chunks.  Before sending the Transport
   and Payload streams of a Chunked RPC-over-RDMA header.

4.6.3. Reply message, the
   responder uses RDMA Write operations to move reduced data items in
   Write and Reply chunks.

4.5.3.  Long Messages

   When an RPC message a Payload stream is larger than the connection's receiver's inline threshold,
   the Payload stream is reduced by removing DDP-eligible data items are removed from the message and placed
   placing them in chunks and to be moved separately.  If there are no DDP-eligible DDP-
   eligible data items in the message, Payload stream, or the message Payload stream is
   still too large after DDP-eligible
   items have it has been removed, reduced, the RDMA transport MUST
   use RDMA Read or Write operations to convey the RPC message body Payload stream
   itself.  This mechanism is referred to as a "Long Message."

   To send an RPC message as transmit a Long Message, the sender conveys only the
   RPC-over-RDMA header Transport
   stream with an RDMA Send operation.  The RPC message
   itself Payload stream is not
   included in the Send buffer. buffer in this instance.  Instead, the requester
   provides chunks that the responder uses to move the whole RPC
   message. Payload stream.

   Long RPC call
      To handle an RPC request using send a Long Message, RPC-over-RDMA Call message, the requester provides
      a special Read chunk that contains the RPC call's XDR Payload stream.
      Every segment in this Read chunk MUST contain zero in its Position
      field.  This  Thus this chunk is known as a "Position Zero Read chunk."

   Long RPC reply
      To handle an RPC reply using send a Long Message, RPC-over-RDMA Reply message, the requester provides
      a single special Write chunk, chunk in advance, known as the "Reply
      chunk", that contains will contain the RPC reply's XDR Payload stream.  The
      requester sizes the Reply chunk to accommodate the largest possible maximum
      expected reply size for that Upper Layer operation.

   Though the purpose of a Long Message is to handle large RPC messages,
   requesters MAY use a Long Message at any time to convey an RPC call.
   Responders SHOULD use MUST send a Long Message reply whenever a Reply chunk has been
   provided by a requester.  Both types of special chunk MAY be present
   in the same RPC message.

   Because these special chunks contain a whole RPC message, any XDR
   data item MAY appear in one of these special chunks without regard to
   its DDP-eligibility.  DDP-eligible data items MAY be removed from
   these special chunks and conveyed via normal chunks, but non-eligible
   data items MUST NOT appear in normal chunks.

5.  RPC-Over-RDMA In Operation

   An

   Every RPC-over-RDMA Version One message has a header precedes all RPC messages
   conveyed across an RDMA transport.  This header that includes a
   copy of the message's transaction ID, data for managing RDMA flow
   control credits, and lists of memory addresses RDMA segments used for RDMA Read and
   Write operations.  All RPC-over-RDMA header content is contained in
   the Transport stream, and thus MUST be XDR encoded.

   RPC message layout is unchanged from that described in [RFC5531]
   except for the possible removal reduction of data items that are moved by
   RDMA Read or Write operations.  If an RPC message (along with its RPC-
   over-RDMA header) is larger than the connection's inline threshold
   even after any large chunks are removed, then

5.1.  XDR Protocol Definition

   Code components extracted from this document must include the RPC message MAY be
   moved separately as a chunk, leaving just
   following license boilerplate.

   <CODE BEGINS>

      /*
       * Copyright (c) 2010, 2015 IETF Trust and the RPC-over-RDMA header in persons
       * identified as authors of the RDMA Send.

5.1.  Fixed Header Fields code.  All rights reserved.
       *
       * The RPC-over-RDMA header begins with four fixed 32-bit fields that
   MUST be present and that control authors of the RDMA interaction including RDMA-
   specific flow control.  These four fields code are:

5.1.1.  Transaction ID (XID)

   The XID generated for the RPC call
       * B. Callaghan, T. Talpey, and reply.  Having the XID at a
   fixed location C. Lever.
       *
       * Redistribution and use in source and binary forms, with
       * or without modification, are permitted provided that the header makes it easy for the receiver to
   establish context as soon as the message arrives.  This XID MUST be
       * following conditions are met:
       *
       * - Redistributions of source code must retain the same as above
       *   copyright notice, this list of conditions and the XID
       *   following disclaimer.
       *
       * - Redistributions in binary form must reproduce the RPC header.  The receiver MAY perform its
   processing based solely on above
       *   copyright notice, this list of conditions and the XID
       *   following disclaimer in the RPC-over-RDMA header, and
   thereby ignore documentation and/or other
       *   materials provided with the XID in distribution.
       *
       * - Neither the RPC header, if it so chooses.

5.1.2.  Version number

   For RPC-over-RDMA Version One, this field MUST contain name of Internet Society, IETF or IETF
       *   Trust, nor the value 1
   (one).  Further discussion names of protocol extensibility can specific contributors, may be found in
   Section 9.

5.1.3.  Flow control credit value

   When sent in an RPC call message, the requested credit value is
   provided.  When sent in an RPC reply message, the granted credit
   value is returned.  RPC calls SHOULD NOT be sent in excess of the
   currently granted limit.  Further discussion of flow control can be
   found in Section 4.3.

5.1.4.  Message type

   o  RDMA_MSG = 0 indicates that chunk lists and an RPC message follow.
      The format of the chunk lists is discussed below.

   o  RDMA_NOMSG = 1 indicates that after the chunk lists there is no
      RPC message.  In this case, the chunk lists provide information to
      allow the responder to transfer the RPC message using RDMA Read or
      Write operations.

   o  RDMA_MSGP = 2 is reserved, and no longer used.

   o  RDMA_DONE = 3 is reserved, and no longer used.

   o  RDMA_ERROR = 4 is
       *   used to signal a responder-detected error in
      RDMA chunk encoding.

   For a message of type RDMA_MSG, the four fixed fields are followed by
   the Read and Write lists and the Reply chunk (though any or all three
   MAY be marked as not present), then an RPC message, beginning with
   its XID field.  The Send buffer holds two separate XDR streams: the
   first XDR stream contains the RPC-over-RDMA header, and the second
   XDR stream contains the RPC message.

   For a message of type RDMA_NOMSG, the four fixed fields are followed
   by the Read and Write chunk lists and the Reply chunk (though any endorse or
   all three MAY be marked as not present).  The Send buffer holds one
   XDR stream which contains the RPC-over-RDMA header.

   For a message of type RDMA_ERROR, the four fixed fields are followed
   by formatted error information.

   The above content (the fixed fields, the chunk lists, and the RPC
   message, when present) MUST be conveyed via a single RDMA Send
   operation. promote products derived from this
       *   software without specific prior written permission.
       *
       *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS
       *   AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED
       *   WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
       *   IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
       *   FOR A gather operation on the Send can be used to marshal the
   separate RPC-over-RDMA header, the chunk lists, and the RPC message
   itself.  However, the total length of the gathered send buffers
   cannot exceed the peer's inline threshold. PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO
       *   EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
       *   LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
       *   EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
       *   NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
       *   SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
       *   INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
       *   LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
       *   OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
       *   IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
       *   ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
       */

      struct rpcrdma1_segment {
              uint32 rdma_handle;
              uint32 rdma_length;
              uint64 rdma_offset;
      };

      struct rpcrdma1_read_segment {
              uint32                  rdma_position;
              struct rpcrdma1_segment rdma_target;
      };

      struct rpcrdma1_read_list {
              struct rpcrdma1_read_segment rdma_entry;
              struct rpcrdma1_read_list    *rdma_next;
      };

      struct rpcrdma1_write_chunk {
              struct rpcrdma1_segment rdma_target<>;
      };

      struct rpcrdma1_write_list {
              struct rpcrdma1_write_chunk rdma_entry;
              struct rpcrdma1_write_list  *rdma_next;
      };

      struct rpcrdma1_header {
              uint32        rdma_xid;
              uint32        rdma_vers;
              uint32        rdma_credit;
              rpcrdma1_body rdma_body;
      };

      enum rpcrdma1_proc {
              RDMA_MSG = 0,
              RDMA_NOMSG = 1,
              RDMA_MSGP = 2,  /* Reserved */
              RDMA_DONE = 3,  /* Reserved */
              RDMA_ERROR = 4
      };
      struct rpcrdma1_chunks {
              struct rpcrdma1_read_list   *rdma_reads;
              struct rpcrdma1_write_list  *rdma_writes;
              struct rpcrdma1_write_chunk *rdma_reply;
      };

      enum rpcrdma1_errcode {
              RDMA_ERR_VERS = 1,
              RDMA_ERR_CHUNK = 2
      };

      union rpcrdma1_error switch (rpcrdma1_errcode rdma_err) {
              case RDMA_ERR_VERS:
                uint32 rdma_vers_low;
                uint32 rdma_vers_high;
              case RDMA_ERR_CHUNK:
                void;
      };

      union rdma_body switch (rpcrdma1_proc rdma_proc) {
              case RDMA_MSG:
              case RDMA_NOMSG:
                rpcrdma1_chunks rdma_chunks;
              case RDMA_MSGP:
                uint32          rdma_align;
                uint32          rdma_thresh;
                rpcrdma1_chunks rdma_achunks;
              case RDMA_DONE:
                void;
              case RDMA_ERROR:
                rpcrdma1_error rdma_error;
      };

   <CODE ENDS>

5.2.  Chunk Lists  Fixed Header Fields

   The chunk lists in an RPC-over-RDMA Version One header are three XDR
   optional-data fields that MUST follow the begins with four fixed header fields in
   RDMA_MSG and RDMA_NOMSG type messages.  Read Section 4.19 of
   [RFC4506] carefully to understand how optional-data fields work.
   Examples of XDR encoded chunk lists are provided in Section 13.1 to
   aid understanding.

5.2.1.  Read List

   Each RPC-over-RDMA Version One header has one "Read list."  The Read
   list is a list of zero or more Read segments, provided by the
   requester, that are grouped by their Position 32-bit fields into Read
   chunks.  Each Read chunk advertises the locations of data that
   MUST be present and that control the
   responder is to pull via RDMA Read operations. interaction including RDMA-
   specific flow control.  These four fields are:

5.2.1.  Transaction ID (XID)

   The requester SHOULD
   sort XID generated for the chunks in RPC Call and Reply.  Having the Read list in Position order.

   Via a Position Zero Read Chunk, XID at a requester may provide part or all
   of an entire RPC call
   fixed location in the header makes it easy for the receiver to
   establish context as soon as the message arrives.  This XID MUST be
   the same as the first chunk XID in this list. the RPC message.  The Read list receiver MAY be empty if perform its
   processing based solely on the RPC call has no argument data that
   is DDP-eligible XID in the RPC-over-RDMA header, and
   thereby ignore the Position Zero Read Chunk is not being used. XID in the RPC message, if it so chooses.

5.2.2.  Write List

   Each  Version number

   For RPC-over-RDMA Version One header has one "Write list."  The
   Write list One, this field MUST contain the value 1
   (one).  Further discussion of protocol extensibility can be found in
   Section 9.

5.2.3.  Flow control credit value

   When sent in an RPC Call message, the requested credit value is a list
   provided.  When sent in an RPC Reply message, the granted credit
   value is returned.  RPC Calls SHOULD NOT be sent in excess of zero or more Write chunks, provided by the
   requester.  Each Write
   currently granted limit.  Further discussion of flow control can be
   found in Section 4.3.

5.2.4.  Message type

   o  RDMA_MSG = 0 indicates that chunk is lists and an array RPC message follow.
      The format of RDMA segments, thus the
   Write list is a list of counted arrays.  Each Write chunk advertises
   receptacles for DDP-eligible data to be pushed by the responder.

   When a Write list lists is provided for the results of discussed below.

   o  RDMA_NOMSG = 1 indicates that after the chunk lists there is no
      RPC call, message.  In this case, the
   responder MUST chunk lists provide any corresponding data via RDMA Write information to
      allow the
   memory referenced in the chunk's segments.  The Write list MAY be
   empty if responder to transfer the RPC operation has no DDP-eligible result data.

   When multiple message using RDMA Read or
      Write chunks are present, the responder fills operations.

   o  RDMA_MSGP = 2 is reserved.

   o  RDMA_DONE = 3 is reserved.

   o  RDMA_ERROR = 4 is used to signal an error in each
   Write RDMA chunk with a DDP-eligible result until either there are no more
   results or no more Write chunks. encoding.

   An Upper Layer Binding MUST
   determine how Write list entries are mapped to procedure arguments
   for each Upper Layer procedure.  For details, see Section 6.

   The RPC reply RDMA_MSG type message conveys the size of result data by returning the Write
   list to the requester with the lengths rewritten to match the actual
   transfer.  Decoding the reply therefore performs no local data
   transfer but merely returns Transport stream and the length obtained from Payload
   stream via an RDMA Send operation.  The Transport stream contains the reply.

   Each decoded result consumes one entry in
   four fixed fields, followed by the Read and Write list, which in
   turn consists of an array of RDMA segments. lists and the Reply
   chunk, though any or all three MAY be marked as not present.  The length of
   Payload stream then follows, beginning with its XID field.  If a Read
   or Write chunk list is therefore the sum present, a portion of all returned lengths in all segments
   comprising the corresponding list entry.  As each Payload stream has
   been excised and is conveyed separately via RDMA Read or Write
   operations.

   An RDMA_NOMSG type message conveys the Transport stream via an RDMA
   Send operation.  The Transport stream contains the four fixed fields,
   followed by the Read and Write chunk is
   decoded, lists and the entire entry is consumed.

5.2.3. Reply Chunk

   Each RPC-over-RDMA Version One header has chunk.
   Though any MAY be marked as not present, one "Reply Chunk."  The
   Reply Chunk is MUST be present and MUST
   hold the Payload stream for this RPC-over-RDMA message, beginning
   with its XID field.  If a Read or Write chunk, provided by the requester.  The Reply
   Chunk chunk list is present, a single counted array
   portion of the Payload stream has been excised and is conveyed
   separately via RDMA segments.  A responder MAY
   convey part Read or all of an entire RPC reply Write operations.

   An RDMA_ERROR type message in this chunk.

   A requester provides the Reply chunk whenever it predicts conveys the
   responder's reply might not fit in Transport stream via an RDMA
   Send operation.  A
   requester MAY choose to provide the Reply chunk even when  The Transport stream contains the
   responder can return only a small reply.

5.3.  Forming Messages

5.3.1.  Short Messages

   A Short Message without chunks four fixed fields,
   followed by formatted error information.  No Payload stream is contained entirely within a single
   conveyed in this type of RPC-over-RDMA message.

   A gather operation on each RDMA Send Operation.  Since operation can be used to marshal
   the RPC call message immediately follows Transport and Payload streams separately.  However, the RPC-over-RDMA header in total
   length of the gathered send buffer, the requester buffers MUST set NOT exceed the message type to RDMA_MSG.

   <------------------ RPC-over-RDMA header --------------->
   +--------+---------+---------+------------+-------------+ +----------
   |        |         |         |            |     NULL    | | Whole
   |  XID   | Version | Credits |  RDMA_MSG  | peer
   receiver's inline threshold.

5.3.  Chunk Lists | |  RPC
   |        |         |         |            |             | | Message
   +--------+---------+---------+------------+-------------+ +----------

5.3.2.  Chunked Messages

   A Chunked Message is similar to a Short Message, but the RPC message
   does not contain the

   The chunk data.  Since the RPC call message
   immediately follows the lists in an RPC-over-RDMA Version One header in the send buffer, the
   requester are three XDR
   optional-data fields that MUST set follow the message type to RDMA_MSG.

   <------------------ RPC-over-RDMA fixed header --------------->
   +--------+---------+---------+------------+-------------+ +----------
   |        |         |         |            |             | | Modified
   |  XID   | Version | Credits | fields in
   RDMA_MSG  | Chunk Lists | |   RPC
   |        |         |         |            |             | | Message
   +--------+---------+---------+------------+-------------+ +----------
                                                |
                                                |  +----------
                                                |  |
                                                +->| Chunks
                                                   |
                                                   +----------

5.3.3.  Long Call Messages

   To send a Long Call Message, the requester registers the memory
   containing the RPC call message and adds a RDMA_NOMSG type messages.  Read Section 4.19 of
   [RFC4506] carefully to understand how optional-data fields work.
   Examples of XDR encoded chunk lists are provided in Section 5.7 as an
   aid to the understanding.

5.3.1.  Read List at
   Position Zero.  Since the RPC call message does not follow the RPC-
   over-RDMA header in the send buffer, the requester MUST set the
   message type to RDMA_NOMSG.

      <------------------ RPC-over-RDMA header --------------->
      +--------+---------+---------+------------+-------------+
      |        |         |         |            |             |
      |  XID   | Version | Credits |

   Each RDMA_MSG or RDMA_NOMSG | Chunk Lists |
      |        |         |         |            |             |
      +--------+---------+---------+------------+-------------+
                                                   |
                                                   |  +----------
                                                   |  | RPC Call
                                                   +->|
                                                      | Message
                                                      +----------

   If a responder gets an RPC-over-RDMA header with a message type of
   RDMA_NOMSG and finds an initial message has one "Read list."  The
   Read list entry with is a list of zero XDR
   position, it allocates a registered buffer and issues an RDMA or more Read segments, provided by the
   requester, that are grouped by their Position fields into Read
   chunks.  Each Read chunk advertises the location of data the RPC message into it.  The
   responder then proceeds is to XDR decode
   the retrieve via RDMA Read operations.

   Via a Position Zero Read Chunk, a requester may provide an RPC Call
   message as a chunk in the Read list.

   The Read list is empty if it had received it with the Send data.  Further
   decoding may issue additional RDMA Reads to bring over additional
   chunks.

       Requester                             Responder
           |        RDMA-over-RPC Header         |
      Send |   ------------------------------>   |
           |                                     |
           |          Long RPC Call Msg          |
           |   <------------------------------   | Read
           |   ------------------------------>   |
           |                                     |
           |         RDMA-over-RPC Reply         |
           |   <------------------------------   | Send

            A long call RPC with request supplied via RDMA Read

5.3.4.  Long Reply Messages

   To send a Long Reply Message, has no argument data that is
   DDP-eligible, and the requester MAY register Position Zero Read Chunk is not being used.

5.3.2.  Write List

   Each RDMA_MSG or RDMA_NOMSG type message has one "Write list."  The
   Write list is a large
   buffer into which list of zero or more Write chunks, provided by the responder can write
   requester.  Each Write chunk is an RPC reply.  This buffer array of RDMA segments, thus the
   Write list is passed a list of counted arrays.  Each Write chunk advertises
   receptacles for DDP-eligible data to be pushed by the responder in via
   RDMA Write operations.

   When a Write list is provided for the results of an RPC call message as Call, the Reply
   chunk.

   If
   responder MUST provide any corresponding data via RDMA Write to the responder's reply message
   memory referenced in the chunk's segments.  The Write list is too long to return with an RDMA
   Send operation, even after empty
   if the RPC operation has no DDP-eligible result data.

   When multiple Write chunks are removed, then present, the responder performs an RDMA fills in each
   Write chunk with a DDP-eligible result until either there are no more
   results or no more Write of the chunks.

   The RPC reply message into conveys the
   buffer indicated size of result data by returning the Reply chunk.  Since Write
   list to the requester with the lengths rewritten to match the actual
   transfer.  Decoding the RPC reply message
   does not follow therefore performs no local data
   transfer but merely returns the RPC-over-RDMA header length obtained from the reply.

   Each decoded result consumes one entry in the send buffer, Write list, which in
   turn consists of an array of RDMA segments.  The length of a Write
   chunk is therefore the
   responder MUST set sum of all returned lengths in all segments
   comprising the message type to RDMA_NOMSG.

      <------------------ RPC-over-RDMA header --------------->
      +--------+---------+---------+------------+-------------+
      |        |         |         |            |             |
      |  XID   | Version | Credits | RDMA_NOMSG | Chunk Lists |
      |        |         |         |            |             |
      +--------+---------+---------+------------+-------------+
                                                   |
                                                   |  +----------
                                                   |  | RPC corresponding list entry.  As each Write chunk is
   decoded, the entire entry is consumed.

5.3.3.  Reply
                                                   +->|
                                                      | Message
                                                      +----------

       Requester                             Responder
           |      RPC Call with Chunk

   Each RDMA_MSG or RDMA_NOMSG type message has one "Reply chunk."  The
   Reply chunk      |
      Send |   ------------------------------>   |
           |                                     |
           |          Long RPC Reply Msg         |
           |   <------------------------------   | Write
           |         RDMA-over-RPC Header        |
           |   <------------------------------   | Send

              An RPC with long reply returned via RDMA is a Write chunk, provided by the requester.  The use Reply
   chunk is a single counted array of RDMA Write to return long replies requires that the segments.

   A requester anticipates MUST provide a long reply and has some knowledge of its size
   so that an adequately sized buffer can be allocated.  Typically the
   Upper Layer Protocol can limit Reply chunk whenever the maximum possible
   size of RPC replies appropriately.

   It the reply is possible for a single RPC procedure to employ both a long call
   for larger than its arguments and own inline threshold.  The Reply
   chunk MUST be large enough to contain a long reply for its results.  However, such an
   operation Payload stream (RPC message)
   of this maximum size.

   When a Reply chunk is atypical, as few upper layers define such exchanges. provided, a responder MUST convey the RPC reply
   message in this chunk.

5.4.  Memory Registration

   RDMA requires that data is transferred only between only registered memory
   segments at the source and destination.  All protocol headers as well
   as separately transferred data chunks use must reside in registered
   memory.

   Since the cost of registering and de-registering memory can be a
   significant proportion of the RDMA transaction cost, it is important
   to minimize registration activity.  This is easily can be achieved within
   RPC-controlled RPC-
   controlled memory by allocating chunk list data and RPC headers in a
   reusable way from pre-registered pools.

5.4.1.  Registration Longevity

   Data chunks transferred via RDMA Read and Write MAY reside in a
   memory allocation that persists outside the bounds of the RPC
   transaction.  Hence, the default behavior of an RPC-over-RDMA
   transport is to register and invalidate these chunks on every RPC
   transaction.

   The requester endpoint must ensure that these memory segments are
   properly fenced from the responder before allowing Upper Layer access
   to the data contained in them.  The data in such segments must be at
   rest while a responder has access to that memory.

   This includes segments that are associated with canceled RPCs.  A
   responder cannot know that the requester is no longer waiting for a
   reply, and might proceed to read or even update memory that the
   requester has released for other use.

5.4.2.  Communicating DDP-Eligibility

   The interface by which an Upper Layer Protocol implementation
   communicates the eligibility of a data item locally to its local RPC-
   over-RDMA endpoint is out of scope for not described by this specification.

   Depending on the implementation and constraints imposed by Upper
   Layer Bindings, it is possible to implement an RPC chunking facility
   that is transparent reduction transparently
   to upper layers.  Such implementations may lead to inefficiencies,
   either because they require the RPC layer to perform expensive
   registration and de-registration of memory "on the fly", or they may
   require using RDMA chunks in reply messages, along with the resulting
   additional handshaking with the RPC-over-RDMA peer.

   However, these issues are internal and generally confined to the
   local interface between RPC and its upper layers, one in which
   implementations are free to innovate.  The only requirement is that
   the resulting RPC-over-RDMA protocol sent to the peer is valid for
   the upper layer.

5.4.3.  Registration Strategies

   The choice of which memory registration strategies to employ is left
   to requester and responder implementers.  To support the widest array
   of RDMA implementations, as well as the most general steering tag
   scheme, an Offset field is included in each segment.

   While zero-based offset schemes are available in many RDMA
   implementations, their use by RPC requires individual registration of
   each segment.  For such implementations, this can be a significant
   overhead.  By providing an offset in each chunk, many pre-
   registration or region-based registrations can be readily supported.
   By using a single, universal chunk representation, the RPC-over-RDMA
   protocol implementation is simplified to its most general form.

5.5.  Error Handling Errors

   When a peer receives an RPC-over-RDMA message, it MUST perform

   A receiver performs basic validity checks on the RPC-over-RDMA header
   and chunk contents.  If such errors are
   detected in the request, an RDMA_ERROR reply MUST be generated.

   Two types of errors are defined, version mismatch and invalid chunk
   format.

   o  When a responder detects an RPC-over-RDMA header version that it
      does not support (currently this document defines only Version
      One), contents before it replies with an error code of ERR_VERS, and provides passes the
      low and high inclusive version numbers it does, in fact, support.
      The version number in this reply MUST be any value otherwise valid
      at RPC message to the receiver.

   o  When a responder detects other decoding RPC
   consumer.  If errors are detected in the header or
      chunks, one of the following errors MUST be returned: either an
      RPC decode error such as RPC_GARBAGEARGS, or the RPC-over-RDMA
      error code ERR_CHUNK.

   When a requester cannot parse a responder's reply, the requester
   SHOULD drop the RPC request and return an error to the application to
   prevent retransmission of header, an operation that can never complete.

   A requester might not provide a responder with enough resources to
   reply.  For example, if a requester's receive buffer is too small,
   the responder's Send operation completes with a Local Length Error,
   and the connection is dropped.  Or, if the requester's Reply chunk is
   too small to accommodate the whole RPC reply,
   RDMA_ERROR type message MUST be generated.  Because the responder can tell
   as it is constructing transport
   layer may not be aware of the reply.  The responder SHOULD send direction of a reply
   with problematic RPC message,
   an RDMA_ERROR to signal to the type message MAY be generated by either a requester that no RPC-level reply is
   possible, and or
   a responder.

   To form an RDMA_ERROR type message: The rdma_xid field MUST contain
   the same XID should not be retried.

   It is assumed that was in the link itself will provide some degree of error
   detection and retransmission.  iWARP's Marker PDU Aligned (MPA) layer
   (when used over TCP), Stream Control Transmission Protocol (SCTP), as
   well as the InfiniBand link layer all provide Cyclic Redundancy Check
   (CRC) protection of rdma_xid field in the RDMA payload, and CRC-class protection is a
   general attribute of such transports.

   Additionally, failing request;
   The rdma_vers field MUST contain the RPC layer itself can accept errors from same version that was in the link
   level and recover via retransmission.  RPC recovery can handle
   complete loss and re-establishment of
   rdma_vers field in the link. failing request; The details of
   reporting and recovery from RDMA link layer errors are outside rdma_proc field MUST
   contain the
   scope of this protocol specification.

   See Section 10 for further discussion of value RDMA_ERROR; The rdma_err field contains a value
   that reflects the use type of RPC-level
   integrity schemes error that occurred, as described below.

   An RDMA_ERROR type message indicates a permanent error.  When
   receiving an RDMA_ERROR type message, a requester should attempt to detect errors and related efficiency issues.

5.6.  XDR Language Description

   Code components extracted from this document must include
   terminate the
   following license boilerplate.

   <CODE BEGINS>
      /*
       * Copyright (c) 2010, 2015 IETF Trust RPC transaction if it recognizes the XID in the reply's
   rdma_xid field, and return an error to the persons
       * identified as authors of application to prevent
   retrying the code.  All rights reserved.
       *
       * The authors failed RPC transaction.

   To avoid an infinite loop, a receiver should drop an RDMA_ERROR type
   message that is malformed.

5.5.1.  Header Version Mismatch

   When a receiver detects an RPC-over-RDMA header version that it does
   not support (currently this document defines only Version One), it
   MUST reply with an rdma_err value of ERR_VERS, providing the code are:
       * B. Callaghan, T. Talpey, and C. Lever.
       *
       * Redistribution low and use
   high inclusive version numbers it does, in source and binary forms, with
       * or without modification, are permitted provided fact, support.

5.5.2.  XDR Errors

   A receiver might encounter an XDR parsing error that prevents it from
   processing the
       * following conditions are met:
       *
       * - Redistributions incoming Transport stream.  Examples of source code must retain such errors
   include an invalid value in the above
       *   copyright notice, this list rdma_proc field, an RDMA_NOMSG
   message that has no chunk lists, or the contents of conditions and the
       *   following disclaimer.
       *
       * - Redistributions in binary form must reproduce rdma_xid
   field might not match the above
       *   copyright notice, this list contents of conditions and the
       *   following disclaimer XID field in the documentation and/or other
       *   materials provided
   accompanying RPC message.  In such cases, the responder MUST reply
   with an rdma_err value of ERR_CHUNK.

   When a responder receives a valid RPC-over-RDMA header but the distribution.
       *
       * - Neither
   responder's Upper Layer Protocol implementation cannot parse the name of Internet Society, IETF or IETF
       *   Trust, nor RPC
   arguments in the names RPC Call message, the responder SHOULD return a
   RPC_GARBAGEARGS reply, using an RDMA_MSG type message.  This type of specific contributors, may
   parsing failure might be
       *   used due to endorse mismatches between chunk sizes or promote products derived from this
       *   software without specific prior written permission.
       *
       *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS
       *   AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED
       *   WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
       *   IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
       *   FOR
   offsets and the contents of the Payload stream, for example.  A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO
       *   EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
       *   LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
       *   EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
       *   NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
       *   SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
       *   INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
       *   LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
       *   OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
       *   IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
       *   ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
       */

      struct rpcrdma1_segment {
              uint32 rdma_handle;
              uint32 rdma_length;
              uint64 rdma_offset;
      };
      struct rpcrdma1_read_segment {
              uint32                  rdma_position;
              struct rpcrdma1_segment rdma_target;
      };

      struct rpcrdma1_read_list {
              struct rpcrdma1_read_segment rdma_entry;
              struct rpcrdma1_read_list    *rdma_next;
      };

      struct rpcrdma1_write_chunk {
              struct rpcrdma1_segment rdma_target<>;
      };

      struct rpcrdma1_write_list {
              struct rpcrdma1_write_chunk rdma_entry;
              struct rpcrdma1_write_list  *rdma_next;
      };

      struct rpcrdma1_msg {
              uint32        rdma_xid;
              uint32        rdma_vers;
              uint32        rdma_credit;
              rpcrdma1_body rdma_body;
      };

      enum rpcrdma1_proc {
              RDMA_MSG = 0,
              RDMA_NOMSG = 1,
              RDMA_MSGP = 2,  /* Reserved */
              RDMA_DONE = 3,  /* Reserved */
   responder MAY also report the presence of a non-DDP-eligible data
   item in a Read or Write chunk using RPC_GARBAGEARGS.

5.5.3.  Responder Operational Errors

   Problems can arise as a responder attempts to use requester-provided
   resources for RDMA Read or Write operations.  For example:

   o  Chunks can be validated only by using their contents to form RDMA
      Read or Write operations.  If chunk contents are invalid (say, a
      segment is no longer registered, or a chunk length is too long), a
      Remote Access error occurs.

   o  If a requester's receive buffer is too small, the responder's Send
      operation completes with a Local Length Error.

   o  If the requester-provided Reply chunk is too small to accommodate
      a large RPC reply, a Remote Access error occurs.  A responder can
      detect this problem before attempting to write past the end of the
      Reply chunk.

   Operational errors are typically fatal to the connection.  To avoid a
   retransmission loop and repeated connection loss that deadlocks the
   connection, once the requester has re-established a connection, the
   responder should send an RDMA_ERROR = 4
      };

      struct rpcrdma1_chunks {
              struct rpcrdma1_read_list   *rdma_reads;
              struct rpcrdma1_write_list  *rdma_writes;
              struct rpcrdma1_write_chunk *rdma_reply;
      };

      enum rpcrdma1_errcode {
              RDMA_ERR_VERS = 1,
              RDMA_ERR_CHUNK = 2
      };

      union rpcrdma1_error switch (rpcrdma1_errcode err) {
              case RDMA_ERR_VERS:
                uint32 rdma_vers_low;
                uint32 rdma_vers_high;
              case RDMA_ERR_CHUNK:
                void;
      };

      union rdma_body switch (rpcrdma1_proc proc) {
              case RDMA_MSG:
              case RDMA_NOMSG:
                rpcrdma1_chunks rdma_chunks;
              case RDMA_MSGP:
                uint32          rdma_align;
                uint32          rdma_thresh;
                rpcrdma1_chunks rdma_achunks;
              case RDMA_DONE:
                void;
              case RDMA_ERROR:
                rpcrdma1_error rdma_error;
      };

   <CODE ENDS>

5.7.  Deprecated reply with an rdma_err value of
   ERR_CHUNK to indicate that no RPC-level reply is possible for that
   XID.

5.5.4.  RDMA Transport Errors

   The RDMA connection and physical link provide some degree of error
   detection and retransmission.  iWARP's Marker PDU Aligned (MPA) layer
   (when used over TCP), Stream Control Transmission Protocol (SCTP), as
   well as the InfiniBand link layer all provide Cyclic Redundancy Check
   (CRC) protection of the RDMA payload, and CRC-class protection is a
   general attribute of such transports.

   Additionally, the RPC layer itself can accept errors from the link
   level and recover via retransmission.  RPC recovery can handle
   complete loss and re-establishment of the link.

   The details of reporting and recovery from RDMA link layer errors are
   outside the scope of this protocol specification.  See Section 10 for
   further discussion of the use of RPC-level integrity schemes to
   detect errors.

5.6.  Protocol Elements

5.7.1. No Longer Supported

   The following protocol elements are no longer supported in RPC-over-
   RDMA Version One.  Related enum values and structure definitions
   remain in the RPC-over-RDMA Version One protocol for backwards
   compatibility.

5.6.1.  RDMA_MSGP

   The specification of RDMA_MSGP in Section 3.9 of [RFC5666] is
   incomplete.  To fully specify RDMA_MSGP would require:

   o  Updating the definition of DDP-eligibility to include data items
      that may be transferred, with padding, via RDMA_MSGP type messages

   o  Adding full operational descriptions of the alignment and
      threshold fields

   o  Discussing how alignment preferences are communicated between two
      peers without using CCP

   o  Describing the treatment of RDMA_MSGP type messages that convey
      Read or Write chunks

   The RDMA_MSGP message type is beneficial only when the padded data
   payload is at the end of an RPC message's argument or result list.
   This is not typical for NFSv4 COMPOUND RPCs, which often include a
   GETATTR operation as the final element of the compound operation
   array.

   Without a full specification of RDMA_MSGP, there has been no fully
   implemented prototype of it.  Without a complete prototype of
   RDMA_MSGP support, it is difficult to assess whether this protocol
   element has benefit, or can even be made to work interoperably.

   Therefore, senders MUST NOT send RDMA_MSGP type messages.  When
   receiving an RDMA_MSGP

   Implementers type message, receivers SHOULD reply with an
   RDMA_ERROR type message, setting the rdma_err field to ERR_CHUNK.

5.6.2.  RDMA_DONE

   Because no implementation of RPC-over-RDMA Version One have neglected to make use uses the Read-
   Read transfer model, there is never a need to send an RDMA_DONE type
   message.

   Therefore, senders MUST NOT send RDMA_DONE messages.  When receiving
   an RDMA_DONE type message, receivers SHOULD reply with an RDMA_ERROR
   type message, setting the rdma_err field to ERR_CHUNK.

5.7.  XDR Examples

   RPC-over-RDMA chunk lists are complex data types.  In this appendix,
   illustrations are provided to help readers grasp how chunk lists are
   represented inside an RPC-over-RDMA header.

   An RDMA segment is the simplest component, being made up of a 32-bit
   handle (H), a 32-bit length (L), and 64-bits of offset (OO).  Once
   flattened into an XDR stream, RDMA segments appear as

      HLOO

   A Read segment has an additional 32-bit position field.  Read
   segments appear as

      PHLOO

   A Read chunk is a list of Read segments.  Each segment is preceded by
   a 32-bit word containing a one if there is a segment, or a zero if
   there are no more segments (optional-data).  In XDR form, this would
   look like

      1 PHLOO 1 PHLOO 1 PHLOO 0

   where P would hold the same value for each segment belonging to the
   same Read chunk.

   The Read List is also a list of Read segments.  In XDR form, this
   would look like a Read chunk, except that the P values could vary
   across the list.  An empty Read List is encoded as a single 32-bit
   zero.

   One Write chunk is a counted array of segments.  In XDR form, the RDMA_MSGP message type.  Therefore RDMA_MSGP is deprecated.

   Senders SHOULD NOT send RDMA_MSGP type messages.  Receivers MUST
   treat received RDMA_MSGP type messages as they
   count would treat RDMA_MSG
   type messages.  The additional alignment information is appear as the first 32-bit word, followed by an
   optimization hint that may be ignored.

5.7.2.  RDMA_DONE

   Because implementations HLOO for
   each element of RPC-over-RDMA Version One do not use the
   Read-Read transfer model, there should never be any need to send an
   RDMA_DONE type message.  Therefore RDMA_DONE array.  For instance, a Write chunk with three
   elements would look like

      3 HLOO HLOO HLOO

   The Write List is deprecated.

   Receivers MUST drop RDMA_DONE type messages without additional
   processing.

6.  Upper Layer Binding Specifications

   Each RPC program and version tuple that operates on an RDMA transport
   MUST have an Upper Layer Binding specification.  A ULB may be part of
   another protocol specification, or it may be a stand-alone document,
   similar to [RFC5667].

   A ULB specification MUST provide four important pieces list of
   information:

   o  Which counted arrays.  In XDR data items in the RPC program are eligible for Direct
      Data Placement

   o  How form, this is a responder utilizes chunks provided in
   combination of optional-data and counted arrays.  To represent a
   Write list

   o  How DDP-eligibility violations are reported to peers

   o  An rpcbind port assignment for operation of the RPC program on
      RDMA transports

6.1.  Determining DDP-Eligibility

   A DDP-eligible List containing a Write chunk with three segments and a Write
   chunk with two segments, XDR data item would encode

      1 3 HLOO HLOO HLOO 1 2 HLOO HLOO 0

   An empty Write List is encoded as a single 32-bit zero.

   The Reply chunk is one that MAY be moved in a Write chunk.
   All other XDR data items MUST NOT be moved  Since it is an optional-data
   field, however, there is a 32-bit field in front of it that contains
   a one if the Reply chunk that is part of
   a Short present, or Chunked Message, nor as a separate zero if it is not.  After
   encoding, a Reply chunk in with 2 segments would look like

      1 2 HLOO HLOO

   Frequently a Long
   Message.

   Only an XDR data item requester does not provide any chunks.  In that might benefit from Direct Data Placement
   should be transferred case,
   after the four fixed fields in the RPC-over-RDMA header, there are
   simply three 32-bit fields that contain zero.

6.  RPC Bind Parameters

   In setting up a chunk.  An Upper Layer Binding
   specification should consider an XDR data item for DDP-eligibility if new RDMA connection, the data item can be larger than first action by a Send buffer, and at least one of requester
   is to obtain a transport address for the following:

   o  Is sensitive responder.  The mechanism
   used to page alignment (eg. it would require pullup obtain this address, and to open an RDMA connection is
   dependent on the
      receiver before it can be used)

   o  Is not translated or marshaled when it type of RDMA transport, and is XDR encoded (eg. the responsibility of
   each RPC protocol binding and its local implementation.

   RPC services normally register with a portmap or rpcbind [RFC1833]
   service, which associates an
      opaque array)

   o  Is not immediately used by applications (eg. RPC Program number with a service
   address.  (In the case of UDP or TCP, the service address for NFS is
   normally port 2049.)  This policy is part no different with RDMA
   transports, although it may require the allocation of data
      backup or replication)

   The port numbers
   appropriate to each Upper Layer Protocol implementation or that uses the RDMA RPC framing
   defined here.

   When mapped atop the iWARP transport
   implementation decide when [RFC5040] [RFC5041], which uses
   IP port addressing due to move a DDP-eligible data item into a
   chunk instead its layering on TCP and/or SCTP, port
   mapping is trivial and consists merely of leaving issuing the item port in the RPC message's XDR stream.
   connection process.  The interface NFS/RDMA protocol service address has been
   assigned port 20049 by IANA, for both iWARP/TCP and iWARP/SCTP.

   When mapped atop InfiniBand [IB], which an Upper Layer implementation communicates the
   chunk eligibility of uses a data item locally to the RPC transport is out
   of scope for this specification.  The only requirement Group Identifier
   (GID)-based service endpoint naming scheme, a translation MUST be
   employed.  One such translation is that the
   resulting RPC-over-RDMA protocol sent to defined in the peer InfiniBand Port
   Addressing Annex [IBPORT], which is valid appropriate for translating IP
   port addressing to the
   Upper Layer.

   The XDR language definition of DDP-eligible data items is not
   decorated InfiniBand network.  Therefore, in any way.

   It is the responsibility of the protocol's Upper Layer Binding to
   specify DDP-eligibity rules so that if a DDP-eligible XDR data item
   is embedded within another, only one of these two objects is to this case,
   IP port addressing may be
   represented by a chunk.  This ensures that the mapping from XDR
   position to the XDR object represented is unambiguous.

6.2.  Write List Ordering

   A requester constructs the Write list for an RPC transaction before
   the responder has formulated readily employed by the transaction's reply. upper layer.

   When a mapping standard or convention exists for IP ports on an RDMA
   interconnect, there are several possibilities for each upper layer to
   consider:

   o  One possibility is only one result data item that is DDP-eligible, to have responder register its mapped IP port
      with the rpcbind service, under the netid (or netid's) defined
      here.  An RPC-over-RDMA-aware requester appends only can then resolve its
      desired service to a single Write chunk mappable port, and proceed to that Write list.  If connect.  This
      is the responder populates most flexible and compatible approach, for those upper
      layers that chunk with data, there is no ambiguity
   about which result is contained in it.

   However, an Upper Layer Binding MAY permit a reply where more than
   one result data item is DDP-eligible.  For example, an NFSv4 COMPOUND
   reply are defined to use the rpcbind service.

   o  A second possibility is composed of individual NFSv4 operations, more than one of
   which can contain to have the responder's portmapper
      register itself on the RDMA interconnect at a DDP-eligible result. "well known" service
      address (on UDP or TCP, this corresponds to port 111).  A
      requester provides multiple Write chunks when it expects could connect to this service address and use the RPC
   reply
      portmap protocol to contain more than one data item that is DDP-elegible.
   Ambiguities can arise when replies contain XDR unions or arrays of
   complex data types which allow a responder options about whether obtain a
   DDP-eligible data item is included or not.

   Requester and responder must agree beforehand which data items appear service address in which Write chunk.  Therefore response to a
      program number, e.g., an iWARP port number, or an Upper Layer Binding MUST
   determine how Write list entries are mapped InfiniBand GID.

   o  Alternatively, the requester could simply connect to procedure arguments the mapped
      well-known port for each Upper Layer procedure.

6.3.  DDP-Eligibility Violation

   If the Upper Layer on a receiver service itself, if it is not aware of appropriately
      defined.  By convention, the presence and
   operation of an RPC-over-RDMA transport under it, it could be
   challenging to discover NFS/RDMA service, when a sender has violated operating atop
      such an InfiniBand fabric, will use the same 20049 assignment as
      for iWARP.

   Historically, different RPC protocols have taken different approaches
   to their port assignment; therefore, the specific method is left to
   each RPC-over-RDMA-enabled Upper Layer
   Binding rule.

   If a violation does occur, RFC 5666 does binding, and not define an unambiguous
   mechanism addressed
   here.

   In Section 11, this specification defines two new "netid" values, to
   be used for reporting the violation.  The violation registration of Binding
   rules upper layers atop iWARP [RFC5040]
   [RFC5041] and (when a suitable port translation service is an Upper Layer Protocol issue, but it available)
   InfiniBand [IB].  Additional RDMA-capable networks MAY define their
   own netids, or if they provide a port translation, MAY share the one
   defined here.

7.  Bi-Directional RPC-Over-RDMA

7.1.  RPC Direction

7.1.1.  Forward Direction

   A traditional ONC RPC client is likely that there always a requester.  A traditional
   ONC RPC service is always a responder.  This traditional form of ONC
   RPC message passing is nothing referred to as operation in the Upper Layer can do but reply with "forward
   direction."

   During forward direction operation, the ONC RPC client is responsible
   for establishing transport connections.

7.1.2.  Backward Direction

   The ONC RPC standard does not forbid passing messages in the equivalent of
   BAD XDR.

   When an erroneously-constructed reply reaches other
   direction.  An ONC RPC service endpoint can act as a requester, there in
   which case an ONC RPC client endpoint acts as a responder.  This form
   of message passing is
   no recourse but referred to drop the reply, and perhaps the transport
   connection as well.

   Policing DDP-eligibility must be done operation in co-operation with the Upper
   Layer Protocol by its receive endpoint implementation.

   It "backward
   direction."

   During backward direction operation, the ONC RPC client is
   responsible for establishing transport connections, even though ONC
   RPC Calls come from the Upper Layer Binding's responsibility ONC RPC server.

7.1.3.  Bi-direction

   A pair of endpoints may choose to specify how use only forward or only backward
   direction operations on a
   responder must reply if particular transport.  Or, the endpoints
   may send operations in both directions concurrently on the same
   transport.

   Bi-directional operation occurs when both transport endpoints act as
   a requester violates and a DDP-eligibilty rule.
   The Binding specification should provide similar guidance for
   requesters about handling invalid RPC-over-RDMA replies.

6.4.  Other Binding Information

   An Upper Layer Binding may recommend inline threshold values responder at the same time.  As above, the ONC RPC
   client is responsible for RPC-
   over-RDMA Version One connections bearing that Upper Layer Protocol.
   However, note that RPC-over-RDMA connections can be shared by more
   than one Upper Layer Protocol, establishing transport connections.

7.1.4.  XIDs with Bi-direction

   During bi-directional operation, the forward and that an implementation may backward directions
   use independent xid spaces.

   In other words, a forward direction requester MAY use the same inline threshold for all connections and Protocols that flow
   between two peers.

   If an Upper Layer Protocol specifies a method for exchanging inline
   threshold information, xid
   value at the same time as a backward direction requester on the same
   transport connection, but such concurrent requests represent distinct
   ONC RPC transactions.

7.2.  Backward Direction Flow Control

7.2.1.  Backward RPC-over-RDMA Credits

   Credits work the same way in the sender can find out backward direction as they do in the receiver's
   threshold
   forward direction.  However, forward direction credits and backward
   direction credits are accounted separately.

   In other words, the forward direction credit value only subsequent to establishing is the same
   whether or not there are backward direction resources associated with
   an RPC-over-RDMA transport connection.  The new threshold backward direction credit
   value can take effect only when a new
   connection is established.

7.  RPC Bind Parameters

   In setting up a new RDMA connection, MAY be different than the first action by forward direction credit value.  The
   rdma_credit field in a backward direction RPC-over-RDMA message MUST
   NOT contain the value zero.

   A backward direction requester
   is to obtain a transport address for (an RPC-over-RDMA service endpoint)
   requests credits from the responder. responder (an RPC-over-RDMA client
   endpoint).  The mechanism
   used to obtain this address, and to open an RDMA connection responder reports how many credits it can grant.
   This is
   dependent on the type number of RDMA transport, and is backward direction Calls the responsibility of
   each RPC protocol binding and its local implementation.

   RPC services normally register with a portmap or rpcbind [RFC1833]
   service, which associates responder is
   prepared to handle at once.

   When an RPC program number with RPC-over-RDMA server endpoint is operating correctly, it
   sends no more outstanding requests at a time than the client
   endpoint's advertised backward direction credit value.

7.2.2.  Receive Buffer Management

   An RPC-over-RDMA transport endpoint must pre-post receive buffers
   before it can receive and process incoming RPC-over-RDMA messages.
   If a service
   address.  (In the case of UDP or TCP, the service address sender transmits a message for NFS is
   normally port 2049.)  This policy is a receiver which has no different with posted
   receive buffer, the RDMA
   transports, although it may require provider MAY drop the allocation of port numbers
   appropriate RDMA connection.

7.2.2.1.  Client Receive Buffers

   Typically an RPC-over-RDMA caller posts only as many receive buffers
   as there are outstanding RPC Calls.  A client endpoint without
   backward direction support might therefore at times have no pre-
   posted receive buffers.

   To receive incoming backward direction Calls, an RPC-over-RDMA client
   endpoint must pre-post enough additional receive buffers to each Upper Layer Protocol that uses the match its
   advertised backward direction credit value.  Each outstanding forward
   direction RPC framing
   defined here. requires an additional receive buffer above this
   minimum.

   When mapped atop the iWARP an RDMA transport [RFC5040] [RFC5041], which uses
   IP port addressing due to its layering on TCP and/or SCTP, port
   mapping is trivial and consists merely of issuing the port in the connection process.  The NFS/RDMA protocol service address has been
   assigned port 20049 by IANA, for both iWARP/TCP is lost, all active receive buffers
   are flushed and iWARP/SCTP. are no longer available to receive incoming messages.
   When mapped atop InfiniBand [IB], which uses a Group Identifier
   (GID)-based service fresh transport connection is established, a client endpoint naming scheme,
   must re-post a translation MUST be
   employed.  One such translation is defined in receive buffer to handle the InfiniBand Port
   Addressing Annex [IBPORT], which is appropriate Reply for translating IP
   port addressing each
   retransmitted forward direction Call, and a full set of receive
   buffers to handle backward direction Calls.

7.2.2.2.  Server Receive Buffers

   A forward direction RPC-over-RDMA service endpoint posts as many
   receive buffers as it expects incoming forward direction Calls.  That
   is, it posts no fewer buffers than the InfiniBand network.  Therefore, number of RPC-over-RDMA
   credits it advertises in this case,
   IP port addressing may be readily employed by the upper layer.

   When a mapping standard or convention exists for IP ports on rdma_credit field of forward direction
   RPC replies.

   To receive incoming backward direction replies, an RDMA
   interconnect, there are several possibilities RPC-over-RDMA
   server endpoint must pre-post a receive buffer for each upper layer to
   consider:

   o  One possibility is to have responder register its mapped IP port
      with the rpcbind service, under backward
   direction Call it sends.

   When the netid (or netid's) defined
      here.  An RPC-over-RDMA-aware requester can then resolve its
      desired service to a mappable port, and proceed to connect.  This existing transport connection is the most flexible lost, all active receive
   buffers are flushed and compatible approach, for those upper
      layers that are defined no longer available to use the rpcbind service.

   o  A second possibility receive incoming
   messages.  When a fresh transport connection is to have the responder's portmapper
      register itself on the RDMA interconnect at established, a "well known" service
      address.  (On UDP or TCP, this corresponds to port 111.)  A
      requester could connect server
   endpoint must re-post a receive buffer to this service address and use handle the
      portmap protocol to obtain Reply for each
   retransmitted backward direction Call, and a service address full set of receive
   buffers for receiving forward direction Calls.

7.3.  Conventions For Backward Operation

7.3.1.  In the Absense of Backward Direction Support

   An RPC-over-RDMA transport endpoint might not support backward
   direction operation.  There might be no mechanism in response the transport
   implementation to a
      program number, e.g., an iWARP port number, do so, or an InfiniBand GID.

   o  Alternatively, the requester could simply connect to Upper Layer Protocol consumer might
   not yet have configured the mapped
      well-known port for transport to handle backward direction
   traffic.

   A loss of the service itself, RDMA connection may result if it is appropriately
      defined.  By convention, the NFS/RDMA service, when operating atop
      such receiver is not
   prepared to receive an InfiniBand fabric, will use the same 20049 assignment as
      for iWARP.

   Historically, different RPC protocols have taken different approaches incoming message.  Thus a denial-of-service
   could result if a sender continues to their port assignment; therefore, the specific method send backchannel messages after
   every transport reconnect to an endpoint that is left not prepared to
   each RPC-over-RDMA-enabled
   receive them.

   For RPC-over-RDMA Version One transports, the Upper Layer binding, and not addressed
   here.

   In Section 12, "IANA Considerations", this specification defines two
   new "netid" values, to be used Protocol is
   responsible for registration of upper layers atop
   iWARP [RFC5040] [RFC5041] and (when informing its peer when it has established a suitable port translation
   service is available) InfiniBand [IB].  Additional RDMA-capable
   networks MAY define their own netids, or if they provide backward
   direction capability.  Otherwise even a port
   translation, MAY share the one defined here.

8.  Bi-Directional RPC-Over-RDMA
8.1.  RPC Direction

8.1.1.  Forward Direction

   A traditional simple backward direction
   NULL probe from a peer would result in a lost connection.

   An Upper Layer Protocol consumer MUST NOT perform backward direction
   ONC RPC client operations unless the peer consumer has indicated it is always a requester.
   prepared to handle them.  A traditional description of Upper Layer Protocol
   mechanisms used for this indication is outside the scope of this
   document.

7.3.2.  Backward Direction Retransmission

   In rare cases, an ONC RPC service is always transaction cannot be completed within a responder.
   certain time.  This traditional form of can be because the transport connection was lost,
   the Call or Reply message was dropped, or because the Upper Layer
   consumer delayed or dropped the ONC RPC message passing request.  Typically, the
   requester sends the transaction again, reusing the same RPC XID.
   This is referred to known as operation in an "RPC retransmission".

   In the "forward
   direction."

   During forward direction operation, direction, the Caller is the ONC RPC client.  The
   client is always responsible for establishing a transport connections.

8.1.2.  Backward Direction

   The connection
   before sending again.

   In the backward direction, the Caller is the ONC RPC standard server.  Because
   an ONC RPC server does not forbid passing messages in establish transport connections with
   clients, it cannot send a retransmission if there is no transport
   connection.  It must wait for the other
   direction.  An ONC RPC service endpoint client to re-establish the
   transport connection before it can act as a requester, retransmit ONC RPC transactions in
   which case
   the backward direction.

   If an ONC RPC client endpoint acts as has no work to do, it may be some time before it
   re-establishes a responder.  This form
   of message passing is referred transport connection.  Backward direction Callers
   must be prepared to as operation in the "backward
   direction."

   During wait indefinitely before a connection is
   established before a pending backward direction ONC RPC Call can be
   retransmitted.

7.3.3.  Backward Direction Message Size

   RPC-over-RDMA backward direction operation, messages are transmitted and
   received using the ONC RPC client is
   responsible for establishing transport connections, even though ONC
   RPC Calls come from same buffers as messages in the ONC RPC server.

8.1.3.  Bi-direction

   A pair of endpoints may choose forward direction.
   Therefore they are constrained to use only be no larger than receive buffers
   posted for forward messages.

   It is expected that the Upper Layer Protocol consumer establishes an
   appropriate payload size limit for backward direction operations,
   either by advertising that size limit to its peers, or only by convention.
   If that is done, backward direction operations on a particular transport.  Or, the endpoints
   may send operations in both directions concurrently on messages do not exceed the same
   transport.

   Bi-directional operation occurs when both transport endpoints act as size
   of receive buffers at either endpoint.

   If a requester and sender transmits a responder at the same time.  As above, backward direction message that is larger
   than the ONC RPC
   client receiver is responsible for establishing transport connections.

8.1.4.  XIDs with Bi-direction

   During bi-directional operation, prepared for, the forward and backward directions
   use independent xid spaces.

   In other words, a forward direction requester MAY use RDMA provider drops the same xid
   value at
   message and the same time as RDMA connection.

7.3.4.  Sending A Backward Direction Call

   To form a backward direction requester RPC-over-RDMA Call message on the same
   transport connection, but such concurrent requests use represent
   distinct an RPC-
   over-RDMA Version One transport, an ONC RPC transactions.

8.2.  Backward Direction Flow Control

8.2.1.  Backward service endpoint
   constructs an RPC-over-RDMA Credits

   Credits work header containing a fresh RPC XID in the
   rdma_xid field.

   The rdma_vers field MUST contain the same way value one.  The number of
   requested credits is placed in the backward direction as they do rdma_credit field.

   The rdma_proc field in the
   forward direction.  However, forward direction credits and backward
   direction credits are accounted separately.

   In other words, RPC-over-RDMA header MUST contain the forward direction credit
   value is RDMA_MSG.  All three chunk lists MUST be empty.

   The ONC RPC Call header MUST follow immediately, starting with the
   same
   whether or not there are backward direction resources associated with
   an RPC-over-RDMA transport connection.  The backward direction credit XID value MAY be different than that is present in the forward direction credit value. RPC-over-RDMA header.  The
   rdma_credit Call
   header's msg_type field in a backward direction RPC-over-RDMA message MUST
   NOT contain the value zero. CALL.

7.3.5.  Sending A Backward Direction Reply

   To form a backward direction requester (an RPC-over-RDMA service endpoint)
   requests credits from the Responder (an RPC-over-RDMA Reply message on an RPC-
   over-RDMA Version One transport, an ONC RPC client
   endpoint). endpoint
   constructs an RPC-over-RDMA header containing a copy of the matching
   ONC RPC Call's RPC XID in the rdma_xid field.

   The Responder reports how many credits it can grant.
   This is rdma_vers field MUST contain the value one.  The number of backward direction Calls the Responder is
   prepared to handle at once.

   When an RPC-over-RDMA server endpoint
   granted credits is operating correctly, it
   sends no more outstanding requests at a time than placed in the client
   endpoint's advertised backward direction credit value.

8.2.2.  Receive Buffer Management

   An RPC-over-RDMA transport endpoint must pre-post receive buffers
   before it can receive and process incoming RPC-over-RDMA messages.
   If a sender transmits a message for a receiver which has no posted
   receive buffer, rdma_credit field.

   The rdma_proc field in the RDMA provider RPC-over-RDMA header MUST contain the
   value RDMA_MSG.  All three chunk lists MUST be empty.

   The ONC RPC Reply header MUST follow immediately, starting with the
   same XID value that is allowed to drop present in the RDMA
   connection.

8.2.2.1.  Client Receive Buffers

   Typically an RPC-over-RDMA caller posts only as many receive buffers
   as there are outstanding header.  The
   Reply header's msg_type field MUST contain the value REPLY.

7.4.  Backward Direction Upper Layer Binding

   RPC Calls.  A client endpoint without
   backward direction support might therefore at times have no pre-
   posted receive buffers.

   To receive incoming programs that operate on RPC-over-RDMA Version One only in the
   backward direction Calls, do not require an Upper Layer Binding
   specification.  Because RPC-over-RDMA client
   endpoint must pre-post enough additional receive buffers to match its
   advertised Version One operation in the
   backward direction credit value.  Each outstanding forward does not allow reduction, there can be no DDP-
   eligible data items in such a program.  Backward direction RPC requires an additional receive buffer above this
   minimum.

   When operation
   occurs on an RDMA transport connection already-established connection, thus there is lost, all active receive buffers
   are flushed and are no longer available need to receive incoming messages.

   When a fresh
   specify RPC bind parameters.

8.  Upper Layer Binding Specifications

   Each RPC program and version tuple that operates on an RDMA transport connection is established, a client endpoint
   must re-post
   MUST have an Upper Layer Binding (ULB) specification.  An Upper Layer
   Binding specification can be part of another protocol specification
   document, or it might be a receive buffer stand-alone document, similar to handle
   [RFC5667].

   An Upper Layer Protocol is typically defined independently of a
   particular RPC transport.  An Upper Layer Binding specification
   provides guidance that helps the Reply for each
   retransmitted forward direction Call, Upper Layer Protocol interoperate
   correctly and efficiently over a full set of receive
   buffers to handle backward direction Calls.

8.2.2.2.  Server Receive Buffers

   A forward direction RPC-over-RDMA service endpoint posts as many
   receive buffers particular transport, such as RPC-
   over-RDMA Version One.  In particular, it expects incoming forward direction Calls.  That
   is, it posts no fewer buffers than the number provides:

   o  A taxonomy of RPC-over-RDMA
   credits it advertises XDR data items that are eligible for Direct Data
      Placement

   o  Clarifications on how to compute the maximum reply size for
      operations in the rdma_credit field Upper Layer Protocol

   o  An rpcbind port assignment for operation of forward direction the RPC replies.

   To receive incoming backward direction replies, Program and
      Version on an RPC-over-RDMA
   server endpoint must pre-post a receive buffer for each backward
   direction Call it sends.

   When the existing transport connection is lost, all active receive
   buffers are flushed and are no longer available to receive incoming
   messages.  When a fresh transport connection

8.1.  DDP-Eligibility

   To optimize the use of an RDMA transport, an Upper Layer Binding
   designates some XDR data items as eligible for Direct Data Placement.
   A data item is established, a server
   endpoint must re-post a receive buffer to handle the Reply candidate for each
   retransmitted backward direction Call, and eligibility if there is a full set of receive
   buffers clear
   benefit for receiving forward direction Calls.

8.3.  Conventions For Backward Operation

8.3.1.  In moving the Absense contents of Backward Direction Support

   An RPC-over-RDMA transport endpoint might not support backward
   direction operation.  There might be no mechanism in the transport
   implementation to do so, or the Upper Layer Protocol consumer might
   not yet have configured item directly from the transport to handle backward direction
   traffic.

   A loss
   sender's memory into the receiver's memory.  Criteria for DDP-
   eligibility include:

   1.  The size of the RDMA connection may result if XDR data item is frequently much larger than the receiver
       inline threshold.

   2.  Transport-level processing of the XDR data item is not
   prepared to receive an incoming message.  Thus a denial-of-service
   could result if a sender continues to send backchannel messages after
   every transport reconnect to needed.
       For example, the data item is an endpoint that opaque byte array, which
       requires no XDR encoding and decoding of its content.

   3.  The content of the XDR data item is not prepared sensitive to
   receive them. address
       alignment.  For RPC-over-RDMA Version One transports, example, pullup would be required on the Upper Layer Protocol receiver
       before the content of the item can be used.

   As RPC-over-RDMA messages are formed, DDP-eligible data items are
   treated specially.  A DDP-eligible XDR data item is
   responsible for informing its peer when it has established a backward
   direction capability.  Otherwise even a simple backward direction
   NULL probe from a peer would result one that MAY be
   conveyed by itself in a lost connection.

   An separate chunk.  The Upper Layer Protocol consumer MUST NOT perform backward direction
   ONC RPC operations unless
   implementation or the peer consumer has indicated it is
   prepared RDMA transport implementation decides when to handle them.  A description
   move a DDP-eligible data item into a chunk instead of Upper Layer Protocol
   mechanisms used for this indication is outside leaving the
   item in the scope of this
   document.

8.3.2.  Backward Direction Retransmission

   In rare cases, an ONC RPC transaction cannot message's XDR stream.

   All other XDR data items are considered non-DDP-eligible, and MUST
   NOT be completed within moved in a
   certain time.  This can separate chunk.  They MAY, however, be because the transport connection was lost,
   the Call moved as
   part of a Position Zero Read Chunk or a Reply message was dropped, or because the chunk.

   The interface by which an Upper Layer
   consumer delayed or dropped the ONC RPC request.  Typically, the
   requester sends implementation indicates the transaction again, reusing
   DDP-eligibility of a data item to the same RPC XID.
   This transport is known as an "RPC retransmission".

   In the forward direction, not described
   by this specification.  The only requirements are that the Caller is receiver
   can re-assemble the ONC RPC client.  The
   client is always responsible for establishing transmitted RPC-over-RDMA message into a transport connection
   before sending again.

   In the backward direction, valid
   XDR stream, and that DDP-eligibility rules specified by the Caller Upper
   Layer Binding are respected.

   There is no provision to express DDP-eligibility within the ONC RPC server.  Because
   an ONC RPC server does not establish transport connections with
   clients, it cannot send a retransmission if there XDR
   language.  The only definitive specification of DDP-eligibility is no transport
   connection.  It must wait for
   the ONC RPC client to re-establish Upper Layer Binding itself.

   It is the
   transport connection before it can retransmit ONC RPC transactions in responsibility of the backward direction.

   If an ONC RPC client has no work to do, it may be some time before it
   re-establishes a transport connection.  Backward direction Callers
   must be prepared protocol's Upper Layer Binding to wait indefinitely before
   specify DDP-eligibity rules so that if a connection DDP-eligible XDR data item
   is embedded within another, only one of these two objects is
   established before a pending backward direction ONC RPC Call can to be
   retransmitted.

8.3.3.  Backward Direction Message Size

   RPC-over-RDMA backward direction messages are transmitted and
   received using
   represented by a chunk.  This ensures that the same buffers as messages in mapping from XDR
   position to the forward direction.
   Therefore they XDR object represented is unambiguous.  Note however
   that such complex data types are constrained unlikely to be no larger than receive buffers
   posted good candidates for forward messages.  Typical implementations have chosen to
   use 1024-byte buffers.

   It
   Direct Data Placement.

8.1.1.  Write List Ordering Ambiguity

   A requester constructs the Write list for an RPC transaction before
   the responder has formulated its reply.  When there is expected only one
   result data item that is DDP-eligible, the requester appends only a
   single Write chunk to that Write list.  If the responder populates
   that chunk with data, the requester knows with certainty which result
   is contained in it.

   However, Upper Layer Protocol consumer establishes procedures may allow replies where more
   than one result data item is DDP-eligible.  For example, an
   appropriate payload size limit for backward direction NFSv4
   COMPOUND is composed of individual NFSv4 operations, more than one of
   which may have a reply containing a DDP-eligible result.  As stated
   in Section 5.3.2, when multiple Write chunks are present, the
   responder fills in each Write chunk with a DDP-eligible result until
   either by advertising that size limit to its peers, there are no more results or by convention.
   If that is done, backward direction messages do not exceed the size no more Write chunks.

   Ambiguities can arise when replies contain XDR unions or arrays of receive buffers at either endpoint.

   If
   complex data types which allow a sender transmits responder options about whether a backward direction message that
   DDP-eligible data item is larger
   than the receiver included or not.  It is prepared for, the RDMA provider drops responsibility
   of the
   message and Upper Layer Binding to avoid situations where there is
   ambiguity about which result is in which chunk in the RDMA connection. Write list.  If a sender transmits
   an RDMA message that ambiguity is too small unavoidable, the Upper Layer Binding MUST specify how
   Write list entries are mapped to convey DDP-eligible results.

8.1.2.  DDP-Eligibility Violation

   A DDP-eligibility violation occurs when a
   complete and valid RPC-over-RDMA and RPC requester forms a Call
   message with a non-DDP-eligible data item in either direction,
   the receiver MUST NOT use any value a Read chunk, or
   provides a Write list when there are no DDP-eligible items allowed in
   the fields that were
   transmitted.  Namely, operation's reply.  A violation occurs when a responder forms a
   Reply message without reducing a DDP-eligible data item when there is
   a Write list provided by the rdma_credit field MUST be ignored, and requester.

   In the
   message dropped.

8.3.4.  Sending A Backward Direction Call

   To form first case, a backward direction RPC-over-RDMA responder might attempt to parse and process the
   Call message on an RPC-
   over-RDMA Version One transport, an ONC RPC service endpoint
   constructs an RPC-over-RDMA header containing a fresh RPC XID in anyway.  If the
   rdma_xid field.

   The rdma_vers field MUST contain responder cannot process the value one.  The number of
   requested credits is placed in Call, it
   MUST report this either via an RDMA_ERROR type message with the rdma_credit field.

   The rdma_proc
   rdma_err field in set to ERR_CHUNK, or via an RPC-level RPC_GARBAGEARGS
   message.

   In the RPC-over-RDMA header MUST contain second case, the
   value RDMA_MSG.  All three responder is in a bind: when a Write chunk lists MUST be empty.

   The ONC RPC Call header is
   provided, it MUST follow immediately, starting with use it, but the
   same XID value that ULB specification does not say what
   result is present expected in that chunk.  This is considered a transport-
   level error, and MUST be reported to the RPC-over-RDMA header.  The Call
   header's msg_type requester via an RDMA_ERROR
   type message with the rdma_err field MUST contain set to ERR_CHUNK.

   In the value CALL.

8.3.5.  Sending A Backward Direction Reply

   To form third case, a backward direction RPC-over-RDMA requester might attempt to parse and process the
   Reply message on an RPC-
   over-RDMA Version One transport, an ONC RPC client endpoint
   constructs an RPC-over-RDMA header containing a copy of anyway.  If the matching
   ONC RPC Call's RPC XID in requester cannot process the rdma_xid field.

   The rdma_vers field Reply, it
   MUST contain the value one.  The number of
   granted credits is placed in report this via an RDMA_ERROR type message with the rdma_credit field.

   The rdma_proc rdma_err
   field in the RPC-over-RDMA header MUST contain the
   value RDMA_MSG.  All three chunk lists MUST be empty.

   The ONC RPC set to ERR_CHUNK.

8.2.  Maximum Reply header MUST follow immediately, starting with Size

   A requester provides resources for both a Call message and its
   matching Reply message.  A requester forms the
   same XID value that is present in Call message itself,
   thus can compute the RPC-over-RDMA header.  The
   Reply header's msg_type field MUST contain exact resources needed for it.

   A requester must allocate resources for the value REPLY.

8.4.  Backward Direction Upper Layer Binding

   RPC programs that operate on RPC-over-RDMA Version One only in Reply message (an RPC-
   over-RDMA credit, a Receive buffer, and possibly a Write list and
   Reply chunk) before the responder has formed the actual reply.  To
   accommodate all possible replies for the
   backward direction do not require an Upper Layer Binding
   specification.  Because RPC-over-RDMA Version One operation in the
   backward direction does not allow chunking, there can be no DDP-
   eligible data items in such Call
   message, a program.  Backward direction operation
   occurs requester must allocate reply resources based on an already-established connection, thus the
   maximum possible size of the expected reply.

   If there is no need to
   specify RPC bind parameters.

9.  Transport Protocol Extensibility

   RPC programs are defined solely by their XDR definitions.  They are
   independent of operations in the transport mechanism used to convey base RPC
   messages.  Protocols defined by XDR often have signifcant
   extensibility restrictions placed on them.

   Not all extensibility restrictions on RPC-based Upper Layer Protocols
   may be appropriate for an RPC transport protocol.  TCP [RFC0793], Protocol for
   example, which there
   is no clear payload maximum, an RPC transport protocol that has been extended many
   times independently of the RPC and XDR standards.

   RPC-over-RDMA might be considered as an extension of the RPC protocol
   rather than a separate transport, however.

   o  The mechanisms that TCP uses to move data are opaque to the RPC
      implementation and RPC programs using it. Upper Layer Protocols
      are often aware that RPC-over-RDMA is present, as they identify
      data items that Binding MUST provide a
   mechanism a requester implementation can be moved via direct data placement.

   o  RPC-over-RDMA is used only for moving RPC messages, and not ever use to determine the
   resources needed for generic data transfer. these operations.

8.3.  Additional Considerations

   There may be other details provided in an Upper Layer Binding.

   o  An Upper Layer Binding may recommend an inline threshold value or
      other transport-related parameters for RPC-over-RDMA relies on Version One
      connections bearing that Upper Layer Protocol.

   o  An Upper Layer Protocol may provide a more sophisticated set of base transport
      operations than traditional socket-based transports.
      Interoperability depends on RPC-over-RDMA implementations using means to communicate these operations in
      transport-related parameters between peers.  Note that RPC-over-
      RDMA Version One does not specify any mechanism for changing any
      transport-related parameter after a predictable way. connection has been
      established.

   o  The RPC-over-RDMA header is specified using XDR, unlike other RPC
      transport protocols.

9.1.  Bumping The  Multiple Upper Layer Protocols may share a single RPC-over-RDMA
      Version

   Place holder section.

   Because One connection when their Upper Layer Bindings allow the version number is encoded as part
      use of the RPC-over-RDMA
   header and the RDMA_ERROR message type is used to indicate errors,
   these first four fields Version One and the start of rpcbind port assignments
      for the chunk lists MUST always
   remain aligned at Protocols allow connection sharing.  In this case, the
      same fixed offsets for transport parameters (such as inline threshold) apply to all versions of the RPC-
   over-RDMA header.

   The value of
      Protocols using that connection.

   Given the RPC-over-RDMA header's version field MUST above, Upper Layer Bindings and Upper Layer Protocols must
   be changed
   o  Whenever the on-the-wire format of the RPC-over-RDMA header is
      changed designed to interoperate correctly no matter what connection
   parameters are in effect on a way that prevents interoperability with current
      implementations

   o  Whenever the set of abstract RDMA operations that connection.

8.4.  Upper Layer Protocol Extensions

   An RPC Program and Version tuple may be used is
      changed

   o  Whenever the set of allowable transfer models is altered

10.  Security Considerations

10.1.  Memory Protection

   A primary consideration extensible.  For instance,
   there may be a minor versioning scheme that is not reflected in the protection of
   RPC version number.  Or, the integrity Upper Layer Protocol may allow
   additional features to be specified after the original RPC program
   specification was ratified.  Upper Layer Bindings are provided for
   interoperable programs and
   privacy of local memory versions by an RPC-over-RDMA transport.  The use of
   RPC-over-RDMA MUST NOT introduce any vulnerabilities to system memory
   contents, nor extending existing Upper Layer
   Bindings to memory owned reflect the changes made necessary by user processes.

   It is REQUIRED that any RDMA provider used for RPC transport be
   conformant each addition to
   the requirements of [RFC5042] in order to satisfy these
   protections.  These protections existing XDR.

9.  Transport Protocol Extensibility

   Upper Layer RPC Protocols are provided defined solely by the RDMA layer
   specifications, and specifically their security models.

   o  The use XDR
   definitions.  They are independent of Protection Domains to limit the exposure of memory
      segments transport mechanism used to a single connection is critical.  Any attempt
   convey base RPC messages.  Protocols defined by a
      host not participating in that connection to re-use handles will
      result in a connection failure.  Because XDR often have
   signifcant extensibility restrictions placed on them.

   Not all extensibility restrictions on RPC-based Upper Layer Protocol
      security mechanisms rely on this aspect of Reliable Connection
      behavior, strong authentication Protocols
   may be appropriate for an RPC transport protocol.  TCP [RFC0793], for
   example, is an RPC transport protocol that has been extended many
   times independently of the remote is recommended.

   o  Unpredictable memory handles should RPC and XDR standards.

   RPC-over-RDMA might be used for any operation
      requiring advertised memory segments.  Advertising a continuously
      registered memory region allows considered as an extension of the RPC protocol
   rather than a remote host separate transport, however.

   o  The mechanisms that TCP uses to read or write move data are opaque to
      that region even when an the RPC involving
      implementation and RPC programs using it.  Upper Layer Protocols
      are often aware that memory RPC-over-RDMA is not under
      way.  Therefore advertising persistently registered memory should present, as they identify
      data items that can be avoided. moved via direct data placement.

   o  Advertised memory segments should be invalidated as soon as
      related  RPC-over-RDMA is used only for moving RPC operations are complete.  Invalidation messages, and DMA
      unmapping not ever
      for generic data transfer.

   o  RPC-over-RDMA relies on a more sophisticated set of segments should be complete before an RPC application base transport
      operations than traditional socket-based transports.
      Interoperability depends on RPC-over-RDMA implementations using
      these operations in a predictable way.

   o  The RPC-over-RDMA header is allowed to continue execution and use specified using XDR, unlike other RPC
      transport protocols.

9.1.  RPC-over-RDMA Version Numbering

   Because the contents version number is encoded as part of a memory
      region.

10.2.  Using GSS With RPC-Over-RDMA

   ONC RPC provides its own security via the RPCSEC_GSS framework
   [RFC2203].  RPCSEC_GSS can provide message authentication, integrity
   checking, RPC-over-RDMA
   header and privacy.  This security mechanism is unaffected by the
   RDMA transport.  However, there RDMA_ERROR message type is much data movement associated with
   computation used to indicate errors,
   these first four fields and verification the start of integrity, or encryption/decryption,
   so certain performance advantages may be lost.

   For efficiency, a more appropriate security mechanism the chunk lists MUST always
   remain aligned at the same fixed offsets for RDMA links
   may be link-level protection, such as certain configurations all versions of
   IPsec, which may be co-located in the RDMA hardware. RPC-
   over-RDMA header.

   The use value of
   link-level protection MAY the RPC-over-RDMA header's version field MUST be negotiated through changed

   o  Whenever the use on-the-wire format of the
   RPCSEC_GSS mechanism defined in [RFC5403] RPC-over-RDMA header is
      changed in conjunction a way that prevents interoperability with current
      implementations

   o  Whenever the
   Channel Binding mechanism [RFC5056] and IPsec Channel Connection
   Latching [RFC5660].  Use set of such mechanisms abstract RDMA operations that may be used is REQUIRED where
   integrity and/or privacy
      changed

   o  Whenever the set of allowable transfer models is desired, and where efficiency altered

10.  Security Considerations

10.1.  Memory Protection

   A primary consideration is
   required.

   Once delivered securely by the RDMA provider, any RDMA-exposed memory
   will contain only RPC payloads in the chunk lists, transferred under the protection of RPCSEC_GSS the integrity and privacy.  By these means,
   the data will be protected end-to-end, as required by the RPC layer
   security model.

11.  IANA Considerations

   Three new assignments are specified by this document:

   - A new set
   privacy of RPC "netids" for resolving local memory by an RPC-over-RDMA services

   - Optional service port assignments for Upper Layer Bindings

   - An RPC program number assignment for the configuration protocol

   These assignments have been established, as below. transport.  The new RPC transport has been assigned an RPC "netid", which is an
   rpcbind [RFC1833] string used to describe the underlying protocol in
   order for RPC to select the appropriate transport framing, as well as
   the format use of the service addresses and ports.

   The following "Netid" registry strings are defined for this purpose:

      NC_RDMA "rdma"
      NC_RDMA6 "rdma6"

   These netids MAY be used for
   RPC-over-RDMA MUST NOT introduce any vulnerabilities to system memory
   contents, nor to memory owned by user processes.

   It is REQUIRED that any RDMA network satisfying provider used for RPC transport be
   conformant to the requirements of Section 2, and able to identify service endpoints
   using IP port addressing, possibly through use of a translation
   service as described above [RFC5042] in Section 10, "RPC Binding".  The "rdma"
   netid is order to be used when IPv4 addressing is employed satisfy these
   protections.  These protections are provided by the
   underlying transport, RDMA layer
   specifications, and "rdma6" for IPv6 addressing. specifically their security models.

10.1.1.  Protection Domains

   The netid assignment policy and registry are defined in [RFC5665].

   As a new RPC transport, this protocol has no effect on RPC program
   numbers or existing registered port numbers.  However, new port
   numbers MAY be registered for use by RPC-over-RDMA-enabled services,
   as appropriate of Protection Domains to limit the new networks over which the services will
   operate.

   For example, the NFS/RDMA service defined exposure of memory
   segments to a single connection is critical.  Any attempt by a host
   not participating in [RFC5667] has been
   assigned the port 20049, that connection to re-use handles will result in
   a connection failure.  Because Upper Layer Protocol security
   mechanisms rely on this aspect of Reliable Connection behavior,
   strong authentication of the IANA registry:

      nfsrdma 20049/tcp Network File System (NFS) over RDMA
      nfsrdma 20049/udp Network File System (NFS) over RDMA
      nfsrdma 20049/sctp Network File System (NFS) over RDMA

   The remote is recommended.

10.1.2.  Handle Predictability

   Unpredictable memory handles should be used for any operation
   requiring advertised memory segments.  Advertising a continuously
   registered memory region allows a remote host to read or write to
   that region even when an RPC program number assignment policy and registry involving that memory is not under way.
   Therefore implementations should avoid advertising persistently
   registered memory.

10.1.3.  Memory Fencing

   Advertised memory segments should be invalidated as soon as related
   RPC operations are defined in
   [RFC5531].

12.  Acknowledgments

   The editor gratefully acknowledges the work complete.  Invalidation and DMA unmapping of Brent Callaghan
   segments should be complete before an RPC application is allowed to
   continue execution and
   Tom Talpey on use or alter the original RPC-over-RDMA Version One specification
   [RFC5666].

   The comments and contributions contents of Karen Deitke, Dai Ngo, Chunli
   Zhang, Dominique Martinet, a memory region.

10.2.  Using GSS With RPC-Over-RDMA

   ONC RPC provides its own security via the RPCSEC_GSS framework
   [RFC2203].  RPCSEC_GSS can provide message authentication, integrity
   checking, and Mahesh Siddheshwar are accepted privacy.  This security mechanism is unaffected by the
   RDMA transport.  However, there is much host data movement associated
   with
   many and great thanks.  The editor also wishes to thank Dave Noveck the computation and Bill Baker for their unwavering support verification of this work.

   Special thanks go to nfsv4 Working Group Chair Spencer Shepler integrity and
   nfsv4 Working Group Secretary Thomas Haynes with
   encryption/decryption, so performance advantages can be lost.

   For efficiency, a more appropriate security mechanism for their support.

13.  Appendices

13.1.  Appendix 1: XDR Examples

   RPC-over-RDMA chunk lists are complex data types.  In this appendix,
   illustrations are provided to help readers grasp how chunk lists are
   represented inside an RPC-over-RDMA header.

   An RDMA segment is links
   may be link-level protection, such as certain configurations of
   IPsec, which may be co-located in the RDMA hardware.  The use of
   link-level protection MAY be negotiated through the simplest component, being made up use of a 32-bit
   handle (H), a 32-bit length (L), the
   RPCSEC_GSS mechanism defined in [RFC5403] in conjunction with the
   Channel Binding mechanism [RFC5056] and 64-bits IPsec Channel Connection
   Latching [RFC5660].  Use of offset (OO). such mechanisms is REQUIRED where
   integrity and/or privacy is desired, and where efficiency is
   required.

   Once
   flattened into an XDR stream, delivered securely by the RDMA segments appear as
      HLOO

   A Read segment has an additional 32-bit position field.  Read
   segments appear as

      PHLOO

   A Read provider, any RDMA-exposed memory
   will contain only RPC payloads in the chunk is a list lists, transferred under
   the protection of Read segments.  Each segment is preceded RPCSEC_GSS integrity and privacy.  By these means,
   the data will be protected end-to-end, as required by
   a 32-bit word containing a one if there is a segment, or a zero if
   there the RPC layer
   security model.

11.  IANA Considerations

   Three new assignments are no more segments (optional-data).  In XDR form, specified by this would
   look like

      1 PHLOO 1 PHLOO 1 PHLOO 0

   where P would hold the same value document:

   o  A new set of RPC "netids" for resolving RPC-over-RDMA services

   o  Optional service port assignments for Upper Layer Bindings

   o  An RPC program number assignment for each segment belonging to the
   same Read chunk. configuration protocol

   These assignments have been established, as below.

   The Read List new RPC transport has been assigned an RPC "netid", which is also a list of Read segments.  In XDR form, this
   would look a lot like a Read chunk, except that an
   rpcbind [RFC1833] string used to describe the P values could
   vary across underlying protocol in
   order for RPC to select the list.  An empty Read List is encoded appropriate transport framing, as a single
   32-bit zero.

   One Write chunk is a counted array of segments.  In XDR form, the
   count would appear well as
   the first 32-bit word, followed by an HLOO for
   each element format of the array.  For instance, a Write chunk with three
   elements would look like

      3 HLOO HLOO HLOO service addresses and ports.

   The Write List is a list of counted arrays.  In XDR form, following "Netid" registry strings are defined for this is a
   combination purpose:

      NC_RDMA "rdma"
      NC_RDMA6 "rdma6"

   These netids MAY be used for any RDMA network satisfying the
   requirements of optional-data and counted arrays.  To represent a
   Write List containing a Write chunk with three segments Section 2, and able to identify service endpoints
   using IP port addressing, possibly through use of a Write
   chunk with two segments, XDR would encode

      1 3 HLOO HLOO HLOO 1 2 HLOO HLOO 0

   An empty Write List is encoded translation
   service as a single 32-bit zero. described above in Section 6.  The Reply chunk "rdma" netid is to be
   used when IPv4 addressing is employed by the underlying transport,
   and "rdma6" for IPv6 addressing.

   The netid assignment policy and registry are defined in [RFC5665].

   As a new RPC transport, this protocol has no effect on RPC Program
   numbers or existing registered port numbers.  However, new port
   numbers MAY be registered for use by RPC-over-RDMA-enabled services,
   as appropriate to the new networks over which the services will
   operate.

   For example, the same as a Write chunk.  Since it is an
   optional-data field, however, there is a 32-bit field NFS/RDMA service defined in front of it
   that contains a one if [RFC5667] has been
   assigned the Reply chunk is present, or a zero if it is
   not.  After encoding, a Reply chunk with 2 segments would look like

      1 2 HLOO HLOO

   Frequently a requester does not provide any chunks.  In that case,
   after port 20049, in the four fixed fields IANA registry:

      nfsrdma 20049/tcp Network File System (NFS) over RDMA
      nfsrdma 20049/udp Network File System (NFS) over RDMA
      nfsrdma 20049/sctp Network File System (NFS) over RDMA

   The RPC program number assignment policy and registry are defined in
   [RFC5531].

12.  Acknowledgments

   The editor gratefully acknowledges the work of Brent Callaghan and
   Tom Talpey on the original RPC-over-RDMA header, there Version One specification
   [RFC5666].

   Dave Noveck provided excellent review, constructive suggestions, and
   consistent navigational guidance throughout the process of drafting
   this document.

   The comments and contributions of Karen Deitke, Dai Ngo, Chunli
   Zhang, Dominique Martinet, and Mahesh Siddheshwar are
   simply three 32-bit fields that contain zero.

14. accepted with
   many and great thanks.  The editor also wishes to thank Bill Baker
   for his unwavering support of this work.

   Special thanks go to nfsv4 Working Group Chair Spencer Shepler and
   nfsv4 Working Group Secretary Thomas Haynes for their support.

13.  References

14.1.

13.1.  Normative References

   [RFC1833]  Srinivasan, R., "Binding Protocols for ONC RPC Version 2",
              RFC 1833, DOI 10.17487/RFC1833, August 1995,
              <http://www.rfc-editor.org/info/rfc1833>.

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/
              RFC2119, March 1997,
              <http://www.rfc-editor.org/info/rfc2119>.

   [RFC2203]  Eisler, M., Chiu, A., and L. Ling, "RPCSEC_GSS Protocol
              Specification", RFC 2203, DOI 10.17487/RFC2203, September
              1997, <http://www.rfc-editor.org/info/rfc2203>.

   [RFC4506]  Eisler, M., Ed., "XDR: External Data Representation
              Standard", STD 67, RFC 4506, DOI 10.17487/RFC4506, May
              2006, <http://www.rfc-editor.org/info/rfc4506>.

   [RFC5042]  Pinkerton, J. and E. Deleganes, "Direct Data Placement
              Protocol (DDP) / Remote Direct Memory Access Protocol
              (RDMAP) Security", RFC 5042, DOI 10.17487/RFC5042, October
              2007, <http://www.rfc-editor.org/info/rfc5042>.

   [RFC5056]  Williams, N., "On the Use of Channel Bindings to Secure
              Channels", RFC 5056, DOI 10.17487/RFC5056, November 2007,
              <http://www.rfc-editor.org/info/rfc5056>.

   [RFC5403]  Eisler, M., "RPCSEC_GSS Version 2", RFC 5403, DOI
              10.17487/RFC5403, February 2009,
              <http://www.rfc-editor.org/info/rfc5403>.

   [RFC5531]  Thurlow, R., "RPC: Remote Procedure Call Protocol
              Specification Version 2", RFC 5531, DOI 10.17487/RFC5531,
              May 2009, <http://www.rfc-editor.org/info/rfc5531>.

   [RFC5660]  Williams, N., "IPsec Channels: Connection Latching", RFC
              5660, DOI 10.17487/RFC5660, October 2009,
              <http://www.rfc-editor.org/info/rfc5660>.

   [RFC5665]  Eisler, M., "IANA Considerations for Remote Procedure Call
              (RPC) Network Identifiers and Universal Address Formats",
              RFC 5665, DOI 10.17487/RFC5665, January 2010,
              <http://www.rfc-editor.org/info/rfc5665>.

   [RFC5666]  Talpey, T. and B. Callaghan, "Remote Direct Memory Access
              Transport for Remote Procedure Call", RFC 5666, DOI
              10.17487/RFC5666, January 2010,
              <http://www.rfc-editor.org/info/rfc5666>.

14.2.

13.2.  Informative References

   [IB]       InfiniBand Trade Association, "InfiniBand Architecture
              Specifications", <http://www.infinibandta.org>.

   [IBPORT]   InfiniBand Trade Association, "IP Addressing Annex",
              <http://www.infinibandta.org>.

   [RFC0793]  Postel, J., "Transmission Control Protocol", STD 7, RFC
              793, DOI 10.17487/RFC0793, September 1981,
              <http://www.rfc-editor.org/info/rfc793>.

   [RFC1094]  Nowicki, B., "NFS: Network File System Protocol
              specification", RFC 1094, DOI 10.17487/RFC1094, March
              1989, <http://www.rfc-editor.org/info/rfc1094>.

   [RFC1813]  Callaghan, B., Pawlowski, B., and P. Staubach, "NFS
              Version 3 Protocol Specification", RFC 1813, DOI 10.17487/
              RFC1813, June 1995,
              <http://www.rfc-editor.org/info/rfc1813>.

   [RFC5040]  Recio, R., Metzler, B., Culley, P., Hilland, J., and D.
              Garcia, "A Remote Direct Memory Access Protocol
              Specification", RFC 5040, DOI 10.17487/RFC5040, October
              2007, <http://www.rfc-editor.org/info/rfc5040>.

   [RFC5041]  Shah, H., Pinkerton, J., Recio, R., and P. Culley, "Direct
              Data Placement over Reliable Transports", RFC 5041, DOI
              10.17487/RFC5041, October 2007,
              <http://www.rfc-editor.org/info/rfc5041>.

   [RFC5532]  Talpey, T. and C. Juszczak, "Network File System (NFS)
              Remote Direct Memory Access (RDMA) Problem Statement", RFC
              5532, DOI 10.17487/RFC5532, May 2009,
              <http://www.rfc-editor.org/info/rfc5532>.

   [RFC5661]  Shepler, S., Ed., Eisler, M., Ed., and D. Noveck, Ed.,
              "Network File System (NFS) Version 4 Minor Version 1
              Protocol", RFC 5661, DOI 10.17487/RFC5661, January 2010,
              <http://www.rfc-editor.org/info/rfc5661>.

   [RFC5666]  Talpey, T. and B. Callaghan, "Remote Direct Memory Access
              Transport for Remote Procedure Call", RFC 5666, DOI
              10.17487/RFC5666, January 2010,
              <http://www.rfc-editor.org/info/rfc5666>.

   [RFC5667]  Talpey, T. and B. Callaghan, "Network File System (NFS)
              Direct Data Placement", RFC 5667, DOI 10.17487/RFC5667,
              January 2010, <http://www.rfc-editor.org/info/rfc5667>.

   [RFC7530]  Haynes, T., Ed. and D. Noveck, Ed., "Network File System
              (NFS) Version 4 Protocol", RFC 7530, DOI 10.17487/RFC7530,
              March 2015, <http://www.rfc-editor.org/info/rfc7530>.

Authors' Addresses

   Charles Lever (editor)
   Oracle Corporation
   1015 Granger Avenue
   Ann Arbor, MI  48104
   USA

   Phone: +1 734 274 2396
   Email: chuck.lever@oracle.com
   William Allen Simpson
   DayDreamer
   1384 Fontaine
   Madison Heights, MI  48071
   USA

   Email: william.allen.simpson@gmail.com

   Tom Talpey
   Microsoft Corp.
   One Microsoft Way
   Redmond, WA  98052
   USA

   Phone: +1 425 704-9945
   Email: ttalpey@microsoft.com