<?xml version="1.0"encoding="utf-8"?> <?xml-stylesheet type="text/xsl" href="lib/rfc2629.xslt"?> <?rfc toc="yes" ?> <?rfc symrefs="yes" ?> <?rfc sortrefs="yes" ?> <?rfc compact="yes"?> <?rfc subcompact="no" ?> <?rfc linkmailto="no" ?> <?rfc editing="no" ?> <?rfc comments="yes" ?> <?rfc inline="yes"?> <?rfc rfcedstyle="yes"?> <?rfc-ext allow-markup-in-artwork="yes" ?> <?rfc-ext include-index="no" ?> <!DOCTYPE rfc SYSTEM "rfc2629-xhtml.ent">encoding="UTF-8"?> <rfc xmlns:xi="http://www.w3.org/2001/XInclude"xmlns:x="http://purl.org/net/xml2rfc/ext"ipr="trust200902" category="std"docName="draft-ietf-httpbis-http2bis-07"docName="draft-ietf-httpbis-http2bis-latest" tocInclude="true" symRefs="true" sortRefs="true" version="3" submissionType="IETF" consensus="true" number="9113" obsoletes="7540,8740"><!-- <x:feedback template="mailto:ietf-http-wg@w3.org?subject={docname},%20%22{section}%22&body=<{ref}>:"/> --><front> <title>HTTP/2</title> <seriesInfoname="Internet-Draft" value="draft-ietf-httpbis-http2bis-07"/>name="RFC" value="9113"/> <author initials="M." surname="Thomson" fullname="Martin Thomson" role="editor"> <organization>Mozilla</organization> <address> <postal> <country>Australia</country> </postal> <email>mt@lowentropy.net</email> </address> </author> <author initials="C." surname="Benfield" fullname="Cory Benfield" role="editor"> <organization>Apple Inc.</organization> <address> <email>cbenfield@apple.com</email> </address> </author> <date month="June" year="2022"/> <area>Applications and Real-Time</area> <workgroup>HTTPbis</workgroup> <keyword>HTTP</keyword> <keyword>SPDY</keyword> <keyword>Web</keyword> <abstract><t> This<t>This specification describes an optimized expression of the semantics of the Hypertext Transfer Protocol (HTTP), referred to as HTTP version 2 (HTTP/2). HTTP/2 enables a more efficient use of network resources and a reduced latency by introducing field compression and allowing multiple concurrent exchanges on the sameconnection. </t> <t> Thisconnection.</t> <t>This document obsoletesRFCRFCs 7540 andRFC 8740. </t>8740.</t> </abstract><note title="Discussion Venues" removeInRFC="true"> <t> Discussion of this document takes place on the HTTPBIS Working Group mailing list (ietf-http-wg@w3.org), which is archived at <eref target="https://lists.w3.org/Archives/Public/ietf-http-wg/"/>. </t> <t> Source for this draft and an issue tracker can be found at <eref target="https://github.com/httpwg/http2-spec"/>. </t> </note></front> <middle> <section anchor="intro"> <name>Introduction</name><t> The<t>The performance of applications using the Hypertext Transfer Protocol (HTTP, <xreftarget="HTTP"/>)target="RFC9110"/>) is linked to how each version of HTTP uses the underlying transport, and the conditions under which the transportoperates. </t> <t> Makingoperates.</t> <t>Making multiple concurrent requests can reduce latency and improve application performance. HTTP/1.0 allowed only one request to be outstanding at a time on a given TCP <xreftarget="TCP"/>target="RFC0793"/> connection. HTTP/1.1(<xref target="HTTP11"/>)<xref target="RFC9112"/> added request pipelining, but this only partially addressed request concurrency and still suffers from application-layer head-of-line blocking. Therefore, HTTP/1.0 and HTTP/1.1 clients use multiple connections to a server to make concurrentrequests. </t> <t> Furthermore,requests.</t> <t>Furthermore, HTTP fields are often repetitive and verbose, causing unnecessary network traffic as well as causing the initial TCP congestion window to quickly fill. This can result in excessive latency when multiple requests are made on a new TCPconnection. </t> <t> HTTP/2connection.</t> <t>HTTP/2 addresses these issues by defining an optimized mapping of HTTP's semantics to an underlying connection. Specifically, it allows interleaving of messages on the same connection and uses an efficient coding for HTTP fields. It also allows prioritization of requests, letting more important requests complete more quickly, further improvingperformance. </t> <t> Theperformance.</t> <t>The resulting protocol is more friendly to the network because fewer TCP connections can be used in comparison to HTTP/1.x. This means less competition with other flows and longer-lived connections, which in turn lead to better utilization of available network capacity. Note, however, that TCP head-of-line blocking is not addressed by thisprotocol. </t> <t> Finally,protocol.</t> <t>Finally, HTTP/2 also enables more efficient processing of messages through use of binary messageframing. </t> <t> Thisframing.</t> <t>This document obsoletes<xref target="RFC7540">RFC 7540</xref>RFCs 7540 and<xref target="RFC8740">RFC 8740</xref>.8740. <xref target="revision-updates"/> lists notablechanges. </t>changes.</t> </section> <section anchor="Overview"> <name>HTTP/2 Protocol Overview</name><t> HTTP/2<t>HTTP/2 provides an optimized transport for HTTP semantics. HTTP/2 supports all of the core features of HTTP but aims to be more efficient thanHTTP/1.1. </t> <t> HTTP/2HTTP/1.1.</t> <t>HTTP/2 is a connection-oriented application-layer protocol that runs over a TCP connection (<xreftarget="TCP"/>).target="RFC0793"/>). The client is the TCP connectioninitiator. </t> <t> Theinitiator.</t> <t>The basic protocol unit in HTTP/2 is a <xref target="FrameHeader">frame</xref>. Each frame type serves a different purpose. For example, <xref target="HEADERS" format="none">HEADERS</xref> and <xref target="DATA" format="none">DATA</xref> frames form the basis of <xref target="HttpFraming">HTTP requests and responses</xref>; other frame types like <xref target="SETTINGS" format="none">SETTINGS</xref>, <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref>, and <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> are used in support of other HTTP/2features. </t> <t> Multiplexingfeatures.</t> <t>Multiplexing of requests is achieved by having each HTTP request/response exchange associated with its own <xref target="StreamsLayer">stream</xref>. Streams are largely independent of each other, so a blocked or stalled request or response does not prevent progress on otherstreams. </t> <t> Effectivestreams.</t> <t>Effective use of multiplexing depends on flow control and prioritization. <xref target="FlowControl">Flow control</xref> ensures that it is possible to efficiently use multiplexed streams by restricting data that is transmitted to what the receiver is able to handle. <xref target="StreamPriority">Prioritization</xref> ensures that limited resources are used most effectively. This revision of HTTP/2 deprecates the priority signaling scheme from <xreftarget="RFC7540"/>. </t> <t> Becausetarget="RFC7540"/>.</t> <t>Because HTTP fields used in a connection can contain large amounts of redundant data, frames that contain them are <xref target="FieldBlock">compressed</xref>. This has especially advantageous impact upon request sizes in the common case, allowing many requests to be compressed into onepacket. </t> <t> Finally,packet.</t> <t>Finally, HTTP/2 adds a new, optional interaction mode whereby a server can <xref target="PushResources">push responses to a client</xref>. This is intended to allow a server to speculatively send data to a client that the server anticipates the client will need, trading off some network usage against a potential latency gain. The server does this by synthesizing a request, which it sends as a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame. The server is then able to send a response to the synthetic request on a separatestream. </t>stream.</t> <section> <name>Document Organization</name><t> The<t>The HTTP/2 specification is split into fourparts: </t>parts:</t> <ul spacing="normal"> <li><xref target="starting">Starting HTTP/2</xref> covers how an HTTP/2 connection isinitiated. </li> <li> Theinitiated.</li> <li>The <xref target="FramingLayer">frame</xref> and <xref target="StreamsLayer">stream</xref> layers describe the way HTTP/2 frames are structured and formed into multiplexedstreams. </li>streams.</li> <li><xref target="FrameTypes">Frame</xref> and <xref target="ErrorCodes">error</xref> definitions include details of the frame and error types used inHTTP/2. </li>HTTP/2.</li> <li><xref target="HttpLayer">HTTP mappings</xref> and <xref target="HttpExtra">additional requirements</xref> describe how HTTP semantics are expressed using frames andstreams. </li>streams.</li> </ul><t> While<t>While some of theframeframe- andstream layerstream-layer concepts are isolated from HTTP, this specification does not define a completely generic frame layer. The frame and stream layers are tailored to the needs ofthe HTTP protocol. </t>HTTP.</t> </section> <section> <name>Conventions and Terminology</name> <t>The key words "<bcp14>MUST</bcp14>", "<bcp14>MUST NOT</bcp14>", "<bcp14>REQUIRED</bcp14>", "<bcp14>SHALL</bcp14>", "<bcp14>SHALL NOT</bcp14>", "<bcp14>SHOULD</bcp14>", "<bcp14>SHOULD NOT</bcp14>", "<bcp14>RECOMMENDED</bcp14>", "<bcp14>NOT RECOMMENDED</bcp14>", "<bcp14>MAY</bcp14>", and "<bcp14>OPTIONAL</bcp14>" in this document are to be interpreted as described in BCP 14 <xreftarget="RFC2119" format="default"/>target="RFC2119"/> <xreftarget="RFC8174" format="default"/>target="RFC8174"/> when, and only when, they appear in all capitals, as shown here.</t><t> All<t>All numeric values are in network byte order. Values are unsigned unless otherwise indicated. Literal values are provided in decimal or hexadecimal as appropriate. Hexadecimal literals are prefixed with "<tt>0x</tt>" to distinguish them from decimalliterals. </t> <t> Thisliterals.</t> <t>This specification describes binary formats using theconventionconventions described in <xreftarget="QUIC"target="RFC9000" section="1.3">RFC 9000</xref>. Note that this format uses network byte order and that high-valued bits are listed before low-valuedbits. </t> <t> Thebits.</t> <t>The following terms areused: </t>used:</t> <dl newline="false" spacing="normal"> <dt>client:</dt><dd> The<dd>The endpoint that initiates an HTTP/2 connection. Clients send HTTP requests and receive HTTPresponses. </dd>responses.</dd> <dt>connection:</dt><dd> A<dd>A transport-layer connection between twoendpoints. </dd>endpoints.</dd> <dt>connection error:</dt><dd> An<dd>An error that affects the entire HTTP/2connection. </dd>connection.</dd> <dt>endpoint:</dt><dd> Either<dd>Either the client or server of theconnection. </dd>connection.</dd> <dt>frame:</dt><dd> The<dd>The smallest unit of communication within an HTTP/2 connection, consisting of a header and a variable-length sequence of octets structured according to the frametype. </dd>type.</dd> <dt>peer:</dt><dd> An<dd>An endpoint. When discussing a particular endpoint, "peer" refers to the endpoint that is remote to the primary subject ofdiscussion. </dd>discussion.</dd> <dt>receiver:</dt><dd> An<dd>An endpoint that is receivingframes. </dd>frames.</dd> <dt>sender:</dt><dd> An<dd>An endpoint that is transmittingframes. </dd>frames.</dd> <dt>server:</dt><dd> The<dd>The endpoint that accepts an HTTP/2 connection. Servers receive HTTP requests and send HTTPresponses. </dd>responses.</dd> <dt>stream:</dt><dd> A<dd>A bidirectional flow of frames within the HTTP/2connection. </dd>connection.</dd> <dt>stream error:</dt><dd> An<dd>An error on the individual HTTP/2stream. </dd>stream.</dd> </dl><t> Finally,<t>Finally, the terms "gateway", "intermediary", "proxy", and "tunnel" are defined in <xreftarget="HTTP"target="RFC9110" section="3.7"/>. Intermediaries act as both client and server at differenttimes. </t> <t> Thetimes.</t> <t>The term "content" as it applies to message bodies is defined in <xreftarget="HTTP" section="6.4"/>. </t>target="RFC9110" section="6.4"/>.</t> </section> </section> <section anchor="starting"> <name>Starting HTTP/2</name><t> Implementations<t>Implementations that generate HTTP requests need to discover whether a server supportsHTTP/2. </t> <t> HTTP/2HTTP/2.</t> <t>HTTP/2 uses the"http""<tt>http</tt>" and"https""<tt>https</tt>" URI schemes defined in <xreftarget="HTTP"target="RFC9110" section="4.2"/>, with the same default port numbers as HTTP/1.1(<xref target="HTTP11"/>).<xref target="RFC9112"/>. These URIs do not include any indication about what HTTP versions an upstream server (the immediate peer to which the client wishes to establish a connection)supports. </t> <t> Thesupports.</t> <t>The means by which support for HTTP/2 is determined is different for"http""<tt>http</tt>" and"https""<tt>https</tt>" URIs. Discovery for"https""<tt>https</tt>" URIs is described in <xref target="discover-https"/>. HTTP/2 support for"http""<tt>http</tt>" URIs can only be discovered by out-of-bandmeans,means and requires prior knowledge of the support as described in <xreftarget="known-http"/>. </t>target="known-http"/>.</t> <section anchor="versioning"> <name>HTTP/2 Version Identification</name><t> The<t>The protocol defined in this document has two identifiers. Creating a connection based on either implies the use of the transport, framing, and message semantics described in thisdocument. </t>document.</t> <ul spacing="normal"> <li><t> The<t>The string "h2" identifies the protocol where HTTP/2 uses Transport Layer Security (TLS); see <xref target="TLSUsage"/>. This identifier is used in the <xreftarget="TLS-ALPN">TLS application-layer protocol negotiationtarget="RFC7301">TLS Application-Layer Protocol Negotiation (ALPN) extension</xref> field and in any place where HTTP/2 over TLS isidentified. </t> <t> Theidentified.</t> <t>The "h2" string is serialized into an ALPN protocol identifier as the two-octet sequence: 0x68,0x32. </t>0x32.</t> </li> <li><t> The<t>The "h2c" string was previously used as a token for use in the HTTP Upgrade mechanism's Upgrade header field (<xreftarget="HTTP"target="RFC9110" section="7.8"/>). This usage was never widely deployed and is deprecated by this document. The sameapplesapplies to the HTTP2-Settings header field, which was used with the upgrade to"h2c". </t>"h2c".</t> </li> </ul> </section> <section anchor="discover-https"> <name>Starting HTTP/2 for"https""<tt>https</tt>" URIs</name><t> A<t>A client that makes a request to an"https""<tt>https</tt>" URI uses <xreftarget="TLS13">TLS</xref>target="RFC8446">TLS</xref> with the <xreftarget="TLS-ALPN">application-layer protocol negotiation (ALPN) extension</xref>. </t> <t> HTTP/2target="RFC7301">ALPN extension</xref>.</t> <t>HTTP/2 over TLS uses the "h2" protocol identifier. The "h2c" protocol identifierMUST NOT<bcp14>MUST NOT</bcp14> be sent by a client or selected by a server; the "h2c" protocol identifier describes a protocol that does not useTLS. </t> <t> OnceTLS.</t> <t>Once TLS negotiation is complete, both the client and the serverMUST<bcp14>MUST</bcp14> send a <xref target="preface">connectionpreface</xref>. </t>preface</xref>.</t> </section> <section anchor="known-http"> <name>Starting HTTP/2 with Prior Knowledge</name><t> A<t>A client can learn that a particular server supports HTTP/2 by other means. For example, a client could be configured with knowledge that a server supportsHTTP/2. </t> <t> AHTTP/2.</t> <t>A client that knows that a server supports HTTP/2 can establish a TCP connection and send the <xref target="preface">connection preface</xref> followed by HTTP/2 frames. Servers can identify these connections by the presence of the connection preface. This only affects the establishment of HTTP/2 connections over cleartext TCP; HTTP/2 connections over TLSMUST<bcp14>MUST</bcp14> use <xreftarget="TLS-ALPN">protocoltarget="RFC7301">protocol negotiation inTLS</xref>. </t> <t> Likewise,TLS</xref>.</t> <t>Likewise, the serverMUST<bcp14>MUST</bcp14> send a <xref target="preface">connectionpreface</xref>. </t> <t> Withoutpreface</xref>.</t> <t>Without additional information, prior support for HTTP/2 is not a strong signal that a given server will support HTTP/2 for future connections. For example, it is possible for server configurations to change, for configurations to differ between instances in clustered servers, or for network conditions tochange. </t>change.</t> </section> <section anchor="preface"> <name>HTTP/2 Connection Preface</name><t> In<t>In HTTP/2, each endpoint is required to send a connection preface as a final confirmation of the protocol in use and to establish the initial settings for the HTTP/2 connection. The client and server each send a different connectionpreface. </t> <t> Thepreface.</t> <t>The client connection preface starts with a sequence of 24 octets, which in hex notationis: </t>is:</t> <artwork type="inline"><![CDATA[ 0x505249202a20485454502f322e300d0a0d0a534d0d0a0d0a ]]></artwork><t> That<t>That is, the connection preface starts with the string "<tt>PRI * HTTP/2.0\r\n\r\nSM\r\n\r\n</tt>". This sequenceMUST<bcp14>MUST</bcp14> be followed by a <xref target="SETTINGS" format="none">SETTINGS</xref> frame (<xref target="SETTINGS"/>), whichMAY<bcp14>MAY</bcp14> be empty. The client sends the client connection preface as the first application data octets of aconnection. </t>connection.</t> <aside> <t>Note: The client connection preface is selected so that a large proportion of HTTP/1.1 or HTTP/1.0 servers and intermediaries do not attempt to process further frames. Note that this does not address the concerns raised in <xreftarget="TALKING"/>. </t>target="TALKING"/>.</t> </aside><t> The<t>The server connection preface consists of a potentially empty <xref target="SETTINGS" format="none">SETTINGS</xref> frame (<xref target="SETTINGS"/>) thatMUST<bcp14>MUST</bcp14> be the first frame the server sends in the HTTP/2connection. </t> <t> Theconnection.</t> <t>The <xref target="SETTINGS" format="none">SETTINGS</xref> frames received from a peer as part of the connection prefaceMUST<bcp14>MUST</bcp14> be acknowledged (see <xref target="SettingsSync"/>) after sending the connectionpreface. </t> <t> Topreface.</t> <t>To avoid unnecessary latency, clients are permitted to send additional frames to the server immediately after sending the client connection preface, without waiting to receive the server connection preface. It is important to note, however, that the server connection preface <xref target="SETTINGS" format="none">SETTINGS</xref> frame might include settings that necessarily alter how a client is expected to communicate with the server. Upon receiving the <xref target="SETTINGS" format="none">SETTINGS</xref> frame, the client is expected to honor any settings established. In some configurations, it is possible for the server to transmit <xref target="SETTINGS" format="none">SETTINGS</xref> before the client sends additional frames, providing an opportunity to avoid thisissue. </t> <t> Clientsissue.</t> <t>Clients and serversMUST<bcp14>MUST</bcp14> treat an invalid connection preface as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref>. A <xref target="GOAWAY" format="none">GOAWAY</xref> frame (<xref target="GOAWAY"/>)MAY<bcp14>MAY</bcp14> be omitted in this case, since an invalid preface indicates that the peer is not usingHTTP/2. </t>HTTP/2.</t> </section> </section> <section anchor="FramingLayer"> <name>HTTP Frames</name><t> Once<t>Once the HTTP/2 connection is established, endpoints can begin exchangingframes. </t>frames.</t> <section anchor="FrameHeader"> <name>Frame Format</name><t> All<t>All frames begin with a fixed 9-octet header followed by a variable-length framepayload. </t>payload.</t> <figure anchor="FrameLayout"> <name>Frame Layout</name> <artwork type="inline"><![CDATA[ HTTP Frame { Length (24), Type (8), Flags (8), Reserved (1), Stream Identifier (31), Frame Payload (..), } ]]></artwork> </figure><t> The<t>The fields of the frame header are definedas: </t>as:</t> <dl newline="false" spacing="normal"> <dt>Length:</dt> <dd><t> The<t>The length of the frame payload expressed as an unsigned 24-bit integer in units of octets. Values greater than 2<sup>14</sup> (16,384)MUST NOT<bcp14>MUST NOT</bcp14> be sent unless the receiver has set a larger value for <xref target="SETTINGS_MAX_FRAME_SIZE"format="none">SETTINGS_MAX_FRAME_SIZE</xref>. </t> <t> Theformat="none">SETTINGS_MAX_FRAME_SIZE</xref>.</t> <t>The 9 octets of the frame header are not included in thisvalue. </t>value.</t> </dd> <dt>Type:</dt> <dd><t> The<t>The 8-bit type of the frame. The frame type determines the format and semantics of the frame. Frames defined in this document are listed in <xref target="FrameTypes"/>. ImplementationsMUST<bcp14>MUST</bcp14> ignore and discard frames of unknowntypes. </t>types.</t> </dd> <dt>Flags:</dt> <dd><t> An<t>An 8-bit field reserved for boolean flags specific to the frametype. </t> <t> Flagstype.</t> <t>Flags are assigned semantics specific to the indicated frame type. Unused flags are those that have no defined semantics for a particular frame type. Unused flagsMUST<bcp14>MUST</bcp14> be ignored on receipt andMUST<bcp14>MUST</bcp14> be left unset(0x0)(0x00) whensending. </t>sending.</t> </dd> <dt>Reserved:</dt> <dd><t> A<t>A reserved 1-bit field. The semantics of this bit are undefined, and the bitMUST<bcp14>MUST</bcp14> remain unset(0x0)(0x00) when sending andMUST<bcp14>MUST</bcp14> be ignored whenreceiving. </t>receiving.</t> </dd> <dt>Stream Identifier:</dt> <dd><t> A<t>A stream identifier (see <xref target="StreamIdentifiers"/>) expressed as an unsigned 31-bit integer. The value0x00x00 is reserved for frames that are associated with the connection as a whole as opposed to an individualstream. </t>stream.</t> </dd> </dl><t> The<t>The structure and content of the frame payloadisare dependent entirely on the frametype. </t>type.</t> </section> <section anchor="FrameSize"> <name>Frame Size</name><t> The<t>The size of a frame payload is limited by the maximum size that a receiver advertises in the <xref target="SETTINGS_MAX_FRAME_SIZE" format="none">SETTINGS_MAX_FRAME_SIZE</xref> setting. This setting can have any value between 2<sup>14</sup> (16,384) and 2<sup>24</sup>-1 (16,777,215) octets,inclusive. </t> <t> Allinclusive.</t> <t>All implementationsMUST<bcp14>MUST</bcp14> be capable of receiving and minimally processing frames up to 2<sup>14</sup> octets in length, plus the 9-octet <xref target="FrameHeader">frame header</xref>. The size of the frame header is not included when describing framesizes. </t>sizes.</t> <aside> <t>Note: Certain frame types, such as <xref target="PING">PING</xref>, impose additional limits on the amount of frame payload dataallowed. </t>allowed.</t> </aside><t> An<t>An endpointMUST<bcp14>MUST</bcp14> send an error code of <xref target="FRAME_SIZE_ERROR" format="none">FRAME_SIZE_ERROR</xref> if a frame exceeds the size defined in <xref target="SETTINGS_MAX_FRAME_SIZE" format="none">SETTINGS_MAX_FRAME_SIZE</xref>, exceeds any limit defined for the frame type, or is too small to contain mandatory frame data. A frame size error in a frame that could alter the state of the entire connectionMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref>; this includes any frame carrying a <xref target="FieldBlock">field block</xref> (that is, <xref target="HEADERS" format="none">HEADERS</xref>, <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>, and <xref target="CONTINUATION" format="none">CONTINUATION</xref>), a <xref target="SETTINGS"format="none">SETTINGS</xref>,format="none">SETTINGS</xref> frame, and any frame with a stream identifier of0. </t> <t> Endpoints0.</t> <t>Endpoints are not obligated to use all available space in a frame. Responsiveness can be improved by using frames that are smaller than the permitted maximum size. Sending large frames can result in delays in sending time-sensitive frames (such as <xref target="RST_STREAM" format="none">RST_STREAM</xref>, <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref>, or <xref target="PRIORITY" format="none">PRIORITY</xref>), which, if blocked by the transmission of a large frame, could affectperformance. </t>performance.</t> </section> <section anchor="FieldBlock"> <name>Field Section Compression and Decompression</name><t> Field<t>Field section compression is the process of compressing a set of field lines (<xreftarget="HTTP"target="RFC9110" section="5.2"/>) to form a field block. Field section decompression is the process of decoding a field block into a set of field lines. Details of HTTP/2 field section compression and decompressionisare defined in <xreftarget="COMPRESSION"/>,target="RFC7541"/>, which, for historical reasons, refers to these processes as header compression anddecompression. </t> <t> Eachdecompression.</t> <t>Each field block carries all of the compressed field lines of a single field section. Header sections also include control data associated with the message in the form of <xref target="PseudoHeaderFields">pseudo-header fields</xref> that use the same format as a fieldline. </t>line.</t> <aside><t> Note:<t>Note: <xref target="RFC7540">RFC 7540</xref> used the term "header block" in place of the more generic "fieldblock". </t>block".</t> </aside><t> Field<t>Field blocks carry control data and header sections for requests, responses, promised requests, and pushed responses (see <xref target="PushResources"/>). All these messages, except for interim responses and requests contained in <xref target="PUSH_PROMISE">PUSH_PROMISE</xref> frames, can optionally include a field block that carries a trailersection. </t> <t> Asection.</t> <t>A field section is a collection of field lines. Each of the field lines in a field blockcarrycarries a single value. The serialized field block is then divided into one or more octet sequences, called field block fragments. The first field block fragment is transmitted within the frame payload of <xref target="HEADERS">HEADERS</xref> or <xref target="PUSH_PROMISE">PUSH_PROMISE</xref>, each of which could be followed by <xref target="CONTINUATION">CONTINUATION</xref> frames to carry subsequent field blockfragments. </t> <t> Thefragments.</t> <t>The <xreftarget="COOKIE">Cookietarget="RFC6265">Cookie header field</xref> is treated specially by the HTTP mapping (see <xreftarget="CompressCookie"/>). </t> <t> Atarget="CompressCookie"/>).</t> <t>A receiving endpoint reassembles the field block by concatenating its fragments and then decompresses the block to reconstruct the fieldsection. </t> <t> Asection.</t> <t>A complete field section consists ofeither: </t>either:</t> <ul spacing="normal"><li> a<li>a single <xref target="HEADERS" format="none">HEADERS</xref> or <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame, with the END_HEADERS flag set,or </li> <li> aor</li> <li>a <xref target="HEADERS" format="none">HEADERS</xref> or <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame with the END_HEADERS flag unset and one or more <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames, where the last <xref target="CONTINUATION" format="none">CONTINUATION</xref> frame has the END_HEADERS flagset. </li>set.</li> </ul><t> Each<t>Each field block is processed as a discrete unit. Field blocksMUST<bcp14>MUST</bcp14> be transmitted as a contiguous sequence of frames, with no interleaved frames of any other type or from any other stream. The last frame in a sequence of <xref target="HEADERS" format="none">HEADERS</xref> or <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames has the END_HEADERS flag set. The last frame in a sequence of <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> or <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames has the END_HEADERS flag set. This allows a field block to be logically equivalent to a singleframe. </t> <t> Fieldframe.</t> <t>Field block fragments can only be sent as the frame payload of <xref target="HEADERS" format="none">HEADERS</xref>, <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>, or <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames because these frames carry data that can modify the compression context maintained by a receiver. An endpoint receiving <xref target="HEADERS" format="none">HEADERS</xref>, <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>, or <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames needs to reassemble field blocks and perform decompression even if the frames are to be discarded. A receiverMUST<bcp14>MUST</bcp14> terminate the connection with a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="COMPRESSION_ERROR" format="none">COMPRESSION_ERROR</xref> if it does not decompress a fieldblock. </t> <t> Ablock.</t> <t>A decoding error in a field blockMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="COMPRESSION_ERROR"format="none">COMPRESSION_ERROR</xref>. </t>format="none">COMPRESSION_ERROR</xref>.</t> <section anchor="dynamic-table"> <name>Compression State</name><t> Field<t>Field compression is stateful. Each endpoint has an HPACK encoder context and an HPACK decoder context that are used for encoding and decoding all field blocks on a connection. <xreftarget="COMPRESSION"target="RFC7541" section="4"/> defines the dynamic table, which is the primary state for eachcontext. </t> <t> Thecontext.</t> <t>The dynamic table has a maximum size that is set by an HPACK decoder. An endpoint communicates the size chosen by its HPACK decoder context using the SETTINGS_HEADER_TABLE_SIZE setting; see <xref target="SettingValues"/>. When a connection is established, the dynamic table size for the HPACK decoder and encoder at both endpoints starts at 4,096 bytes, the initial value of the SETTINGS_HEADER_TABLE_SIZEsetting. </t> <t> Anysetting.</t> <t>Any change to the maximum value set using SETTINGS_HEADER_TABLE_SIZE takes effect when the endpoint <xref target="SettingsSync">acknowledges settings</xref>. The HPACK encoder at that endpoint can set the dynamic table to any size up to the maximum value set by the decoder. An HPACK encoder declares the size of the dynamic table with a Dynamic Table Size Update instruction (<xreftarget="COMPRESSION" section="6.3"/>). </t> <t> Oncetarget="RFC7541" section="6.3"/>).</t> <t>Once an endpoint acknowledges a change to SETTINGS_HEADER_TABLE_SIZE that reduces the maximum below the current size of the dynamic table, its HPACK encoderMUST<bcp14>MUST</bcp14> start the next field block with a Dynamic Table Size Update instruction that sets the dynamic table to a size that is less than or equal to the reduced maximum; see <xreftarget="COMPRESSION"target="RFC7541" section="4.2"/>. An endpointMUST<bcp14>MUST</bcp14> treat a field block that follows an acknowledgment of the reduction to the maximum dynamic table size as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="COMPRESSION_ERROR" format="none">COMPRESSION_ERROR</xref> if it does not start with a conformant Dynamic Table Size Updateinstruction. </t>instruction.</t> <aside><t> Implementers<t>Implementers are advised that reducing the value of SETTINGS_HEADER_TABLE_SIZE is not widely interoperable. Use of the connection preface to reduce the value below the initial value of 4,096 is somewhat better supported, but this might fail with someimplementations. </t>implementations.</t> </aside> </section> </section> </section> <section anchor="StreamsLayer"> <name>Streams and Multiplexing</name><t> A<t>A "stream" is an independent, bidirectional sequence of frames exchanged between the client and server within an HTTP/2 connection. Streams have several importantcharacteristics: </t>characteristics:</t> <ul spacing="normal"><li> A<li>A single HTTP/2 connection can contain multiple concurrently open streams, with either endpoint interleaving frames from multiplestreams. </li> <li> Streamsstreams.</li> <li>Streams can be established and used unilaterally or shared by eitherendpoint. </li> <li> Streamsendpoint.</li> <li>Streams can be closed by eitherendpoint. </li> <li> Theendpoint.</li> <li>The order in which frames are sent is significant. Recipients process frames in the order they are received. In particular, the order of <xref target="HEADERS" format="none">HEADERS</xref> and <xref target="DATA" format="none">DATA</xref> frames is semanticallysignificant. </li> <li> Streamssignificant.</li> <li>Streams are identified by an integer. Stream identifiers are assigned to streams by the endpoint initiating thestream. </li>stream.</li> </ul> <section anchor="StreamStates"> <name>Stream States</name><t> The<t>The lifecycle of a stream is shown in <xreftarget="StreamStatesFigure"/>. </t>target="StreamStatesFigure"/>.</t> <figure anchor="StreamStatesFigure"> <name>Stream States</name> <artset> <artwork type="svg"> <svg xmlns="http://www.w3.org/2000/svg" version="1.1" height="480" width="488" viewBox="0 0 488 480" class="diagram" text-anchor="middle" font-family="monospace" font-size="13px"> <g transform="translate(8,16)"> <path d="M 0,112 L 0,416" fill="none" stroke="black"/> <path d="M 56,80 L 56,144" fill="none" stroke="black"/> <path d="M 56,240 L 56,304" fill="none" stroke="black"/> <path d="M 88,144 L 88,232" fill="none" stroke="black"/> <path d="M 96,304 L 96,384" fill="none" stroke="black"/> <path d="M 144,80 L 144,144" fill="none" stroke="black"/> <path d="M 144,240 L 144,304" fill="none" stroke="black"/> <path d="M 200,0 L 200,64" fill="none" stroke="black"/> <path d="M 200,160 L 200,224" fill="none" stroke="black"/> <path d="M 200,368 L 200,432" fill="none" stroke="black"/> <path d="M 232,64 L 232,152" fill="none" stroke="black"/> <path d="M 232,224 L 232,360" fill="none" stroke="black"/> <path d="M 272,0 L 272,64" fill="none" stroke="black"/> <path d="M 272,160 L 272,224" fill="none" stroke="black"/> <path d="M 272,368 L 272,432" fill="none" stroke="black"/> <path d="M 328,80 L 328,144" fill="none" stroke="black"/> <path d="M 328,240 L 328,304" fill="none" stroke="black"/> <path d="M 376,304 L 376,384" fill="none" stroke="black"/> <path d="M 384,144 L 384,232" fill="none" stroke="black"/> <path d="M 416,80 L 416,144" fill="none" stroke="black"/> <path d="M 416,240 L 416,304" fill="none" stroke="black"/> <path d="M 472,112 L 472,416" fill="none" stroke="black"/> <path d="M 200,0 L 272,0" fill="none" stroke="black"/> <path d="M 128,32 L 200,32" fill="none" stroke="black"/> <path d="M 272,32 L 344,32" fill="none" stroke="black"/> <path d="M 200,64 L 272,64" fill="none" stroke="black"/> <path d="M 56,80 L 144,80" fill="none" stroke="black"/> <path d="M 328,80 L 416,80" fill="none" stroke="black"/> <path d="M 0,112 L 56,112" fill="none" stroke="black"/> <path d="M 416,112 L 472,112" fill="none" stroke="black"/> <path d="M 56,144 L 144,144" fill="none" stroke="black"/> <path d="M 328,144 L 416,144" fill="none" stroke="black"/> <path d="M 200,160 L 272,160" fill="none" stroke="black"/> <path d="M 136,192 L 200,192" fill="none" stroke="black"/> <path d="M 272,192 L 336,192" fill="none" stroke="black"/> <path d="M 200,224 L 272,224" fill="none" stroke="black"/> <path d="M 56,240 L 144,240" fill="none" stroke="black"/> <path d="M 328,240 L 416,240" fill="none" stroke="black"/> <path d="M 56,304 L 144,304" fill="none" stroke="black"/> <path d="M 328,304 L 416,304" fill="none" stroke="black"/> <path d="M 200,368 L 272,368" fill="none" stroke="black"/> <path d="M 96,384 L 192,384" fill="none" stroke="black"/> <path d="M 280,384 L 376,384" fill="none" stroke="black"/> <path d="M 0,416 L 192,416" fill="none" stroke="black"/> <path d="M 280,416 L 472,416" fill="none" stroke="black"/> <path d="M 200,432 L 272,432" fill="none" stroke="black"/> <path d="M 336,192 L 356,232" fill="none" stroke="black"/> <path d="M 344,32 L 364,72" fill="none" stroke="black"/> <path d="M 108,72 L 128,32" fill="none" stroke="black"/> <path d="M 116,232 L 136,192" fill="none" stroke="black"/> <polygon class="arrowhead" points="392,232 380,226.4 380,237.6 " fill="black" transform="rotate(90,384,232)"/> <polygon class="arrowhead" points="372,72 360,66.4 360,77.6 " fill="black" transform="rotate(63.43494882292201,364,72)"/> <polygon class="arrowhead" points="364,232 352,226.4 352,237.6 " fill="black" transform="rotate(63.43494882292201,356,232)"/> <polygon class="arrowhead" points="288,416 276,410.4 276,421.6 " fill="black" transform="rotate(180,280,416)"/> <polygon class="arrowhead" points="288,384 276,378.4 276,389.6 " fill="black" transform="rotate(180,280,384)"/> <polygon class="arrowhead" points="240,360 228,354.4 228,365.6 " fill="black" transform="rotate(90,232,360)"/> <polygon class="arrowhead" points="240,152 228,146.4 228,157.6 " fill="black" transform="rotate(90,232,152)"/> <polygon class="arrowhead" points="200,416 188,410.4 188,421.6 " fill="black" transform="rotate(0,192,416)"/> <polygon class="arrowhead" points="200,384 188,378.4 188,389.6 " fill="black" transform="rotate(0,192,384)"/> <polygon class="arrowhead" points="124,232 112,226.4 112,237.6 " fill="black" transform="rotate(116.56505117707799,116,232)"/> <polygon class="arrowhead" points="116,72 104,66.4 104,77.6 " fill="black" transform="rotate(116.56505117707799,108,72)"/> <polygon class="arrowhead" points="96,232 84,226.4 84,237.6 " fill="black" transform="rotate(90,88,232)"/> <g class="text"> <text x="160" y="20">send PP</text> <text x="312" y="20">recv PP</text> <text x="236" y="36">idle</text> <text x="276" y="100">send H /</text> <text x="100" y="116">reserved</text> <text x="268" y="116">recv H</text> <text x="372" y="116">reserved</text> <text x="96" y="132">(local)</text> <text x="372" y="132">(remote)</text> <text x="160" y="180">recv ES</text> <text x="312" y="180">send ES</text> <text x="52" y="196">send H</text> <text x="236" y="196">open</text> <text x="420" y="196">recv H</text> <text x="100"y="260">half</text>y="260">half-</text> <text x="372"y="260">half</text>y="260">half-</text> <text x="100" y="276">closed</text> <text x="276" y="276">send R /</text> <text x="372" y="276">closed</text> <text x="100" y="292">(remote)</text> <text x="268" y="292">recv R</text> <text x="368" y="292">(local)</text> <text x="144" y="340">send ES /</text> <text x="328" y="340">recv ES /</text> <text x="148" y="356">send R /</text> <text x="332" y="356">send R /</text> <text x="140" y="372">recv R</text> <text x="324" y="372">recv R</text> <text x="44" y="388">send R /</text> <text x="428" y="388">send R /</text> <text x="36" y="404">recv R</text> <text x="236" y="404">closed</text> <text x="420" y="404">recv R</text> </g> </g> </svg> </artwork> <artwork type="ascii-art"><![CDATA[ +--------+ send PP | | recv PP ,--------+ idle +--------. / | | \ v +--------+ v +----------+ | +----------+ | | | send H / | | ,------+ reserved | | recv H | reserved +------. | | (local) | | | (remote) | | | +---+------+ v +------+---+ | | | +--------+ | | | | recv ES | | send ES | | | send H | ,-------+ open +-------. | recv H | | | / | | \ | | | v v +---+----+ v v | | +----------+ | +----------+ | | |halfhalf- | | |halfhalf- | | | | closed | | send R / | closed | | | | (remote) | | recv R | (local) | | | +----+-----+ | +-----+----+ | | | | | | | | send ES / | recv ES / | | | | send R / v send R / | | | | recv R +--------+ recv R | | | send R / `----------->| |<-----------' send R / | | recv R | closed | recv R | `----------------------->| |<-----------------------' +--------+ ]]></artwork> </artset> </figure> <dl spacing="compact"> <dt><tt>send</tt>:</dt> <dd>endpoint sends this frame</dd> <dt><tt>recv</tt>:</dt> <dd>endpoint receives this frame</dd> <dt><tt>H</tt>:</dt> <dd><xref target="HEADERS" format="none">HEADERS</xref> frame (with implied <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames)</dd> <dt><tt>ES</tt>:</dt> <dd>END_STREAM flag</dd> <dt><tt>R</tt>:</dt> <dd><xref target="RST_STREAM" format="none">RST_STREAM</xref> frame</dd> <dt><tt>PP</tt>:</dt> <dd><xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame (with implied <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames); state transitions are for the promised stream</dd> </dl><t> Note<t>Note that this diagram shows stream state transitions and the frames and flags that affect those transitions only. In this regard, <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames do not result in state transitions; they are effectively part of the <xref target="HEADERS" format="none">HEADERS</xref> or <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> that they follow. For the purpose of state transitions, the END_STREAM flag is processed as a separate event to the frame that bears it; a <xref target="HEADERS" format="none">HEADERS</xref> frame with the END_STREAM flag set can cause two statetransitions. </t> <t> Bothtransitions.</t> <t>Both endpoints have a subjective view of the state of a stream that could be different when frames are in transit. Endpoints do not coordinate the creation of streams; they are created unilaterally by either endpoint. The negative consequences of a mismatch in states are limited to the "closed" state after sending <xref target="RST_STREAM" format="none">RST_STREAM</xref>, where frames might be received for some time afterclosing. </t> <t> Streamsclosing.</t> <t>Streams have the followingstates: </t>states:</t> <dl newline="false" spacing="normal"> <dt>idle:</dt> <dd><t> All<t>All streams start in the "idle"state. </t> <t> Thestate.</t> <t>The following transitions are valid from thisstate: </t>state:</t> <ul spacing="normal"><li> Sending<li>Sending a <xref target="HEADERS" format="none">HEADERS</xref> frame as a client, or receiving a HEADERS frame as a server, causes the stream to become "open". The stream identifier is selected as described in <xref target="StreamIdentifiers"/>. The same <xref target="HEADERS" format="none">HEADERS</xref> frame can also cause a stream to immediately become"half-closed". </li> <li> Sending"half-closed".</li> <li>Sending a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame on another stream reserves the idle stream that is identified for later use. The stream state for the reserved stream transitions to "reserved (local)". Only a server may send <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>frames. </li> <li> Receivingframes.</li> <li>Receiving a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame on another stream reserves an idle stream that is identified for later use. The stream state for the reserved stream transitions to "reserved (remote)". Only a client may receive <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>frames. </li> <li> Noteframes.</li> <li>Note that the <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame is not sent on the idle stream but references the newly reserved stream in the Promised Stream IDfield. </li> <li> Openingfield.</li> <li>Opening a stream with a higher-valued stream identifier causes the stream to transition immediately to a "closed" state; note that this transition is not shown in thediagram. </li>diagram.</li> </ul><t> Receiving<t>Receiving any frame other than <xref target="HEADERS" format="none">HEADERS</xref> or <xref target="PRIORITY" format="none">PRIORITY</xref> on a stream in this stateMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref>. If this stream isserver-initiated,initiated by the server, as described in <xref target="StreamIdentifiers"/>, then receiving a <xref target="HEADERS" format="none">HEADERS</xref> frameMUST<bcp14>MUST</bcp14> also be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> </dd> <dt>reserved (local):</dt> <dd><t> A<t>A stream in the "reserved (local)" state is one that has been promised by sending a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame. A <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame reserves an idle stream by associating the stream with an open stream that was initiated by the remote peer (see <xreftarget="PushResources"/>). </t> <t> Intarget="PushResources"/>).</t> <t>In this state, only the following transitions arepossible: </t>possible:</t> <ul spacing="normal"><li> The<li>The endpoint can send a <xref target="HEADERS" format="none">HEADERS</xref> frame. This causes the stream to open in a "half-closed (remote)"state. </li> <li> Eitherstate.</li> <li>Either endpoint can send a <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame to cause the stream to become "closed". This releases the streamreservation. </li>reservation.</li> </ul><t> An<t>An endpointMUST NOT<bcp14>MUST NOT</bcp14> send any type of frame other than <xref target="HEADERS" format="none">HEADERS</xref>, <xref target="RST_STREAM" format="none">RST_STREAM</xref>, or <xref target="PRIORITY" format="none">PRIORITY</xref> in thisstate. </t> <t> Astate.</t> <t>A <xref target="PRIORITY" format="none">PRIORITY</xref> or <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref> frameMAY<bcp14>MAY</bcp14> be received in this state. Receiving any type of frame other than <xref target="RST_STREAM" format="none">RST_STREAM</xref>, <xref target="PRIORITY" format="none">PRIORITY</xref>, or <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref> on a stream in this stateMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> </dd> <dt>reserved (remote):</dt> <dd><t> A<t>A stream in the "reserved (remote)" state has been reserved by a remotepeer. </t> <t> Inpeer.</t> <t>In this state, only the following transitions arepossible: </t>possible:</t> <ul spacing="normal"><li> Receiving<li>Receiving a <xref target="HEADERS" format="none">HEADERS</xref> frame causes the stream to transition to "half-closed(local)". </li> <li> Either(local)".</li> <li>Either endpoint can send a <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame to cause the stream to become "closed". This releases the streamreservation. </li>reservation.</li> </ul><t> An<t>An endpointMUST NOT<bcp14>MUST NOT</bcp14> send any type of frame other than <xref target="RST_STREAM" format="none">RST_STREAM</xref>, <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref>, or <xref target="PRIORITY" format="none">PRIORITY</xref> in thisstate. </t> <t> Receivingstate.</t> <t>Receiving any type of frame other than <xref target="HEADERS" format="none">HEADERS</xref>, <xref target="RST_STREAM" format="none">RST_STREAM</xref>, or <xref target="PRIORITY" format="none">PRIORITY</xref> on a stream in this stateMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> </dd> <dt>open:</dt> <dd><t> A<t>A stream in the "open" state may be used by both peers to send frames of any type. In this state, sending peers observe advertised <xref target="FlowControl">stream-level flow-controllimits</xref>. </t> <t> Fromlimits</xref>.</t> <t>From this state, either endpoint can send a frame with an END_STREAM flag set, which causes the stream to transition into one of the "half-closed" states. An endpoint sending an END_STREAM flag causes the stream state to become "half-closed (local)"; an endpoint receiving an END_STREAM flag causes the stream state to become "half-closed(remote)". </t> <t> Either(remote)".</t> <t>Either endpoint can send a <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame from this state, causing it to transition immediately to"closed". </t>"closed".</t> </dd> <dt>half-closed (local):</dt> <dd><t> A<t>A stream that is in the "half-closed (local)" state cannot be used for sending frames other than <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref>, <xref target="PRIORITY" format="none">PRIORITY</xref>, and <xref target="RST_STREAM"format="none">RST_STREAM</xref>. </t> <t> Aformat="none">RST_STREAM</xref>.</t> <t>A stream transitions from this state to "closed" when a frame is received with the END_STREAM flag set or when either peer sends a <xref target="RST_STREAM" format="none">RST_STREAM</xref>frame. </t> <t> Anframe.</t> <t>An endpoint can receive any type of frame in this state. Providing flow-control credit using <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref> frames is necessary to continue receiving flow-controlled frames. In this state, a receiver can ignore <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref> frames, which might arrive for a short period after a frame with the END_STREAM flag set issent. </t>sent.</t> <t><xref target="PRIORITY" format="none">PRIORITY</xref> frames can be received in thisstate. </t>state.</t> </dd> <dt>half-closed (remote):</dt> <dd><t> A<t>A stream that is "half-closed (remote)" is no longer being used by the peer to send frames. In this state, an endpoint is no longer obligated to maintain a receiver flow-controlwindow. </t> <t> Ifwindow.</t> <t>If an endpoint receives additional frames, other than <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref>, <xref target="PRIORITY" format="none">PRIORITY</xref>, or <xref target="RST_STREAM" format="none">RST_STREAM</xref>, for a stream that is in this state, itMUST<bcp14>MUST</bcp14> respond with a <xref target="StreamErrorHandler">stream error</xref> of type <xref target="STREAM_CLOSED"format="none">STREAM_CLOSED</xref>. </t> <t> Aformat="none">STREAM_CLOSED</xref>.</t> <t>A stream that is "half-closed (remote)" can be used by the endpoint to send frames of any type. In this state, the endpoint continues to observe advertised <xref target="FlowControl">stream-level flow-controllimits</xref>. </t> <t> Alimits</xref>.</t> <t>A stream can transition from this state to "closed" by sending a frame with the END_STREAM flag set or when either peer sends a <xref target="RST_STREAM" format="none">RST_STREAM</xref>frame. </t>frame.</t> </dd> <dt>closed:</dt> <dd><t> The<t>The "closed" state is the terminalstate. </t> <t> Astate.</t> <t>A stream enters the "closed" state after an endpoint both sends and receives a frame with an END_STREAM flag set. A stream also enters the "closed" state after an endpoint either sends or receives a <xref target="RST_STREAM" format="none">RST_STREAM</xref>frame. </t> <t> Anframe.</t> <t>An endpointMUST NOT<bcp14>MUST NOT</bcp14> send frames other than <xref target="PRIORITY" format="none">PRIORITY</xref> on a closed stream. An endpointMAY<bcp14>MAY</bcp14> treat receipt of any other type of frame on a closed stream as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="STREAM_CLOSED" format="none">STREAM_CLOSED</xref>, except as notedbelow. </t> <t> Anbelow.</t> <t>An endpoint that sends a frame with the END_STREAM flag set or a <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame might receive a <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref> or <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame from its peer in the time before the peer receives and processes the frame that closes thestream. </t> <t> Anstream.</t> <t>An endpoint that sends a <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame on a stream that is in the "open" or "half-closed (local)" state could receive any type of frame. The peer might have sent or enqueued for sending these frames before processing the <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame. An endpointMUST<bcp14>MUST</bcp14> minimally process and then discard any frames it receives in this state. This means updating header compression state for <xref target="HEADERS" format="none">HEADERS</xref> and <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>frames;frames. Receiving a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>framesframe alsocausecauses the promised stream to become"reserved","reserved (remote)", even when the <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame is received on a closedstream; and,stream. Additionally, the content of <xref target="DATA" format="none">DATA</xref> framescountcounts toward the connection flow-controlwindow. </t> <t> Anwindow.</t> <t>An endpoint can perform this minimal processing for all streams that are in the "closed" state. EndpointsMAY<bcp14>MAY</bcp14> use other signals to detect that a peer has received the frames that caused the stream to enter the "closed" state and treat receipt of any frame other than <xref target="PRIORITY" format="none">PRIORITY</xref> as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref>. Endpoints can use frames that indicate that the peer has received the closing signal to drive this. EndpointsSHOULD NOT<bcp14>SHOULD NOT</bcp14> use timers for this purpose. For example, an endpoint that sends a <xref target="SETTINGS" format="none">SETTINGS</xref> frame after closing a stream can safely treat receipt of a <xref target="DATA" format="none">DATA</xref> frame on that stream as an error after receiving an acknowledgment of the settings. Other things that might be used are <xref target="PING" format="none">PING</xref> frames, receiving data on streams that were created after closing the stream, or responses to requests created after closing thestream. </t>stream.</t> </dd> </dl><t> In<t>In the absence of more specific rules, implementationsSHOULD<bcp14>SHOULD</bcp14> treat the receipt of a frame that is not expressly permitted in the description of a state as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref>. Note that <xref target="PRIORITY" format="none">PRIORITY</xref> can be sent and received in any streamstate. </t> <t> Thestate.</t> <t>The rules in this section only apply to frames defined in this document. Receipt of frames for which the semantics are unknown cannot be treated as anerrorerror, as the conditions for sending and receiving those frames are also unknown; see <xreftarget="extensibility"/>. </t> <t> Antarget="extensibility"/>.</t> <t>An example of the state transitions for an HTTP request/response exchange can be found in <xref target="HttpExamples"/>. An example of the state transitions for server push can be found inSections <xrefSections <xref target="PushRequests" format="counter"/> and <xref target="PushResponses"format="counter"/>. </t>format="counter"/>.</t> <section anchor="StreamIdentifiers"> <name>Stream Identifiers</name><t> Streams<t>Streams are identified by an unsigned 31-bit integer. Streams initiated by a clientMUST<bcp14>MUST</bcp14> use odd-numbered stream identifiers; those initiated by the serverMUST<bcp14>MUST</bcp14> use even-numbered stream identifiers. A stream identifier of zero(0x0)(0x00) is used for connection control messages; the stream identifier of zero cannot be used to establish a newstream. </t> <t> Thestream.</t> <t>The identifier of a newly established streamMUST<bcp14>MUST</bcp14> be numerically greater than all streams that the initiating endpoint has opened or reserved. This governs streams that are opened using a <xref target="HEADERS" format="none">HEADERS</xref> frame and streams that are reserved using <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>. An endpoint that receives an unexpected stream identifierMUST<bcp14>MUST</bcp14> respond with a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Aformat="none">PROTOCOL_ERROR</xref>.</t> <t>A <xref target="HEADERS" format="none">HEADERS</xref> frame will transition the client-initiated stream identified by the stream identifier in the frame header from "idle" to "open". A <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame will transition the server-initiated stream identified by the"PromisedPromised StreamID"ID field in the frame payload from "idle" to"reserved"."reserved (local)" or "reserved (remote)". When a stream transitions out of the "idle" state, all"idle"streams in the "idle" state that might have been opened by the peer with a lower-valued stream identifier immediately transition to "closed". That is, an endpoint may skip a stream identifier, with the effect being that the skipped stream is immediatelyclosed. </t> <t> Streamclosed.</t> <t>Stream identifiers cannot be reused. Long-lived connections can result in an endpoint exhausting the available range of stream identifiers. A client that is unable to establish a new stream identifier can establish a new connection for new streams. A server that is unable to establish a new stream identifier can send a <xref target="GOAWAY" format="none">GOAWAY</xref> frame so that the client is forced to open a new connection for newstreams. </t>streams.</t> </section> <section> <name>Stream Concurrency</name><t> A<t>A peer can limit the number of concurrently active streams using the <xref target="SETTINGS_MAX_CONCURRENT_STREAMS" format="none">SETTINGS_MAX_CONCURRENT_STREAMS</xref> parameter (see <xref target="SettingValues"/>) within a <xref target="SETTINGS" format="none">SETTINGS</xref> frame. The maximum concurrent streams setting is specific to each endpoint and applies only to the peer that receives the setting. That is, clients specify the maximum number of concurrent streams the server can initiate, and servers specify the maximum number of concurrent streams the client caninitiate. </t> <t> Streamsinitiate.</t> <t>Streams that are in the "open" state or in either of the "half-closed" states count toward the maximum number of streams that an endpoint is permitted to open. Streams in any of these three states count toward the limit advertised in the <xref target="SETTINGS_MAX_CONCURRENT_STREAMS" format="none">SETTINGS_MAX_CONCURRENT_STREAMS</xref> setting. Streams in either of the "reserved" states do not count toward the streamlimit. </t> <t> Endpoints MUST NOTlimit.</t> <t>Endpoints <bcp14>MUST NOT</bcp14> exceed the limit set by their peer. An endpoint that receives a <xref target="HEADERS" format="none">HEADERS</xref> frame that causes its advertised concurrent stream limit to be exceededMUST<bcp14>MUST</bcp14> treat this as a <xref target="StreamErrorHandler">stream error</xref> of type <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref> or <xref target="REFUSED_STREAM" format="none">REFUSED_STREAM</xref>. The choice of error code determines whether the endpoint wishes to enable automatic retry (see <xref target="Reliability"/> fordetails). </t> <t> Andetails).</t> <t>An endpoint that wishes to reduce the value of <xref target="SETTINGS_MAX_CONCURRENT_STREAMS" format="none">SETTINGS_MAX_CONCURRENT_STREAMS</xref> to a value that is below the current number of open streams can either close streams that exceed the new value or allow streams tocomplete. </t>complete.</t> </section> </section> <section anchor="FlowControl"> <name>Flow Control</name><t> Using<t>Using streams for multiplexing introduces contention over use of the TCP connection, resulting in blocked streams. A flow-control scheme ensures that streams on the same connection do not destructively interfere with each other. Flow control is used for both individual streams andforthe connection as awhole. </t> <t> HTTP/2whole.</t> <t>HTTP/2 provides for flow control through use of the <xref target="WINDOW_UPDATE">WINDOW_UPDATEframe</xref>. </t>frame</xref>.</t> <section anchor="fc-principles"> <name>Flow-Control Principles</name><t> HTTP/2<t>HTTP/2 stream flow control aims to allow a variety of flow-control algorithms to be used without requiring protocol changes. Flow control in HTTP/2 has the followingcharacteristics: </t>characteristics:</t> <ol spacing="normal" type="1"><li> Flow<li>Flow control is specific to a connection. HTTP/2 flow control operates between the endpoints of a single hop and not over the entire end-to-endpath. </li> <li> Flowpath.</li> <li>Flow control is based on <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref> frames. Receivers advertise how many octets they are prepared to receive on a stream and for the entire connection. This is a credit-basedscheme. </li> <li> Flowscheme.</li> <li>Flow control is directional with overall control provided by the receiver. A receiverMAY<bcp14>MAY</bcp14> choose to set any window size that it desires for each stream and for the entire connection. A senderMUST<bcp14>MUST</bcp14> respect flow-control limits imposed by a receiver. Clients, servers, and intermediaries all independently advertise their flow-control window as a receiver and abide by the flow-control limits set by their peer whensending. </li> <li> Thesending.</li> <li>The initial value for the flow-control window is 65,535 octets for both new streams and the overallconnection. </li> <li> Theconnection.</li> <li>The frame type determines whether flow control applies to a frame. Of the frames specified in this document, only <xref target="DATA" format="none">DATA</xref> frames are subject to flow control; all other frame types do not consume space in the advertised flow-control window. This ensures that important control frames are not blocked by flowcontrol. </li> <li> Ancontrol.</li> <li>An endpoint can choose to disable its own flow control, but an endpoint cannot ignoreflow controlflow-control signals from itspeer. </li> <li> HTTP/2peer.</li> <li>HTTP/2 defines only the format and semantics of the <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref> frame (<xref target="WINDOW_UPDATE"/>). This document does not stipulate how a receiver decides when to send this frame or the value that it sends, nor does it specify how a sender chooses to send packets. Implementations are able to select any algorithm that suits theirneeds. </li>needs.</li> </ol><t> Implementations<t>Implementations are also responsible for prioritizing the sending of requests and responses, choosing how to avoid head-of-line blocking for requests, and managing the creation of new streams. Algorithm choices for these could interact with any flow-controlalgorithm. </t>algorithm.</t> </section> <section anchor="DisableFlowControl"> <name>Appropriate Use of Flow Control</name><t> Flow<t>Flow control is defined to protect endpoints that are operating under resource constraints. For example, a proxy needs to share memory between many connections and also might have a slow upstream connection and a fast downstream one. Flow control addresses cases where the receiver is unable to process data on one stream yet wants to continue to process other streams in the sameconnection. </t> <t> Deploymentsconnection.</t> <t>Deployments that do not require this capability can advertise a flow-control window of the maximum size (2<sup>31</sup>-1) and can maintain this window by sending a <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref> frame when any data is received. This effectively disables flow control for that receiver. Conversely, a sender is always subject to the flow-control window advertised by thereceiver. </t> <t> Deploymentsreceiver.</t> <t>Deployments with constrained resources (for example, memory) can employ flow control to limit the amount of memory a peer can consume. Note, however, that this can lead to suboptimal use of available network resources if flow control is enabled without knowledge of thebandwidth-delaybandwidth * delay product (see <xreftarget="RFC7323"/>). </t> <t> Eventarget="RFC7323"/>).</t> <t>Even with full awareness of the currentbandwidth-delaybandwidth * delay product, implementation of flow control can be difficult. EndpointsMUST<bcp14>MUST</bcp14> read and process HTTP/2 frames from the TCP receive buffer as soon as data is available. Failure to read promptly could lead to a deadlock when critical frames, such as <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref>, are not read and acted upon. Reading frames promptly does not expose endpoints to resource exhaustionattacksattacks, as HTTP/2 flow control limits resourcecommitments. </t>commitments.</t> </section> <section anchor="FlowControlPerformance"><name>Flow Control<name>Flow-Control Performance</name><t> If<t>If an endpoint cannot ensure that its peer always has availableflow controlflow-control window space that is greater than the peer'sbandwidth-delaybandwidth * delay product on this connection, its receive throughput will be limited by HTTP/2 flow control. This will result in degradedperformance. </t> <t> Sendingperformance.</t> <t>Sending timely <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref> frames can improve performance. Endpoints will want to balance the need to improve receive throughput with the need to manage resource exhaustionrisks,risks and should take careful note of <xref target="dos"/> in defining their strategy to manage windowsizes. </t>sizes.</t> </section> </section> <section anchor="StreamPriority"> <name>Prioritization</name><t> In<t>In a multiplexed protocol like HTTP/2, prioritizing allocation of bandwidth and computation resources to streams can be critical to attaining good performance. A poor prioritization scheme can result in HTTP/2 providing poor performance. With no parallelism at the TCP layer, performance could be significantly worse thanHTTP/1.1. </t> <t> AHTTP/1.1.</t> <t>A good prioritization scheme benefits from the application of contextual knowledge such as the content of resources, how resources are interrelated, and how those resources will be used by a peer. In particular, clients can possess knowledge about the priority of requests that is relevant to server prioritization. In those cases, having clients provide priority information can improveperformance. </t>performance.</t> <section anchor="PriorityHistory"> <name>Background on Priority in RFC 7540</name><t> RFC<t>RFC 7540 defined a rich system for signaling priority of requests. However, this system proved to becomplexcomplex, and it was not uniformlyimplemented. </t> <t> Theimplemented.</t> <t>The flexible scheme meant that it was possible for clients to express priorities in very different ways, with little consistency in the approaches that were adopted. For servers, implementing generic support for the scheme was complex. Implementation of priorities was uneven in both clients and servers. Many server deployments ignored client signals when prioritizing their handling ofrequests. </t> <t> Inrequests.</t> <t>In short, the prioritization signaling in <xreftarget="RFC7540">RFC7540</xref>target="RFC7540">RFC 7540</xref> was notsuccessful. </t>successful.</t> </section> <section anchor="PriorityHere"> <name>Priority Signaling inthis Document</name> <t>This Document</name> <t>This update to HTTP/2 deprecates the priority signaling defined in <xref target="RFC7540">RFC 7540</xref>. The bulk of the text related to priority signals is not included in this document. The description of frame fields and some of the mandatory handling is retained to ensure that implementations of this document remain interoperable with implementations that use the priority signaling described in RFC7540. </t> <t> A7540.</t> <t>A thorough description of the RFC 7540 priority scheme remains in <xref target="RFC7540"section="5.3"/>. </t> <t> Signalingsection="5.3"/>.</t> <t>Signaling priority information is necessary to attain good performance in many cases. Where signaling priority information is important, endpoints are encouraged to use an alternative scheme, such as the scheme described in <xreftarget="I-D.ietf-httpbis-priority"/>. </t> <t> Thoughtarget="RFC9218"/>.</t> <t>Though the priority signaling from RFC 7540 was not widely adopted, the information it provides can still be useful in the absence of better information. Endpoints that receive priority signals in <xref target="HEADERS" format="none">HEADERS</xref> or <xref target="PRIORITY" format="none">PRIORITY</xref> frames can benefit from applying that information. In particular, implementations that consume these signals would not benefit from discarding these priority signals in the absence ofalternatives. </t> <t> Servers SHOULDalternatives.</t> <t>Servers <bcp14>SHOULD</bcp14> use other contextual information in determining priority of requests in the absence of any priority signals. ServersMAY<bcp14>MAY</bcp14> interpret the complete absence of signals as an indication that the client has not implemented the feature. The defaults described in <xref target="RFC7540" section="5.3.5"/> are known to have poor performance under mostconditionsconditions, and their use is unlikely to bedeliberate. </t>deliberate.</t> </section> </section> <section anchor="ErrorHandler"> <name>Error Handling</name><t> HTTP/2<t>HTTP/2 framing permits two classes oferror: </t>errors:</t> <ul spacing="normal"><li> An<li>An error condition that renders the entire connection unusable is a connectionerror. </li> <li> Anerror.</li> <li>An error in an individual stream is a streamerror. </li>error.</li> </ul><t> A<t>A list of error codes is included in <xreftarget="ErrorCodes"/>. </t> <t> Ittarget="ErrorCodes"/>.</t> <t>It is possible that an endpoint will encounter frames that would cause multiple errors. ImplementationsMAY<bcp14>MAY</bcp14> discover multiple errors during processing, but theySHOULD<bcp14>SHOULD</bcp14> report at most one stream and one connection error as aresult. </t> <t> Theresult.</t> <t>The first stream error reported for a given stream prevents any other errors on that stream from being reported. In comparison, the protocol permits multiple <xref target="GOAWAY" format="none">GOAWAY</xref> frames, though an endpointSHOULD<bcp14>SHOULD</bcp14> report just one type of connection error unless an error is encountered during graceful shutdown. If this occurs, an endpointMAY<bcp14>MAY</bcp14> send an additional GOAWAY frame with the new error code, in addition to any prior GOAWAY that contained <xref target="NO_ERROR"format="none">NO_ERROR</xref>. </t> <t> Ifformat="none">NO_ERROR</xref>.</t> <t>If an endpoint detects multiple different errors, itMAY<bcp14>MAY</bcp14> choose to report any one of those errors. If a frame causes a connection error, that errorMUST<bcp14>MUST</bcp14> be reported. Additionally, an endpointMAY<bcp14>MAY</bcp14> use any applicable error code when it detects an error condition; a generic error code (such as <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref> or <xref target="INTERNAL_ERROR" format="none">INTERNAL_ERROR</xref>) can always be used in place of more specific errorcodes. </t>codes.</t> <section anchor="ConnectionErrorHandler"> <name>Connection Error Handling</name><t> A<t>A connection error is any error that prevents further processing of the frame layer or corrupts any connectionstate. </t> <t> Anstate.</t> <t>An endpoint that encounters a connection errorSHOULD<bcp14>SHOULD</bcp14> first send a <xref target="GOAWAY" format="none">GOAWAY</xref> frame (<xref target="GOAWAY"/>) with the stream identifier of the last stream that it successfully received from its peer. The <xref target="GOAWAY" format="none">GOAWAY</xref> frame includes an <xref target="ErrorCodes">error code</xref> that indicates why the connection is terminating. After sending the <xref target="GOAWAY" format="none">GOAWAY</xref> frame for an error condition, the endpointMUST<bcp14>MUST</bcp14> close the TCPconnection. </t> <t> Itconnection.</t> <t>It is possible that the <xref target="GOAWAY" format="none">GOAWAY</xref> will not be reliably received by the receiving endpoint. In the event of a connection error, <xref target="GOAWAY" format="none">GOAWAY</xref> only provides a best-effort attempt to communicate with the peer about why the connection is beingterminated. </t> <t> Anterminated.</t> <t>An endpoint can end a connection at any time. In particular, an endpointMAY<bcp14>MAY</bcp14> choose to treat a stream error as a connection error. EndpointsSHOULD<bcp14>SHOULD</bcp14> send a <xref target="GOAWAY" format="none">GOAWAY</xref> frame when ending a connection, providing that circumstances permitit. </t>it.</t> </section> <section anchor="StreamErrorHandler"> <name>Stream Error Handling</name><t> A<t>A stream error is an error related to a specific stream that does not affect processing of otherstreams. </t> <t> Anstreams.</t> <t>An endpoint that detects a stream error sends a <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame (<xref target="RST_STREAM"/>) that contains the stream identifier of the stream where the error occurred. The <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame includes an error code that indicates the type oferror. </t> <t> Aerror.</t> <t>A <xref target="RST_STREAM" format="none">RST_STREAM</xref> is the last frame that an endpoint can send on a stream. The peer that sends the <xref target="RST_STREAM" format="none">RST_STREAM</xref> frameMUST<bcp14>MUST</bcp14> be prepared to receive any frames that were sent or enqueued for sending by the remote peer. These frames can be ignored, except where they modify connection state (such as the state maintained for <xref target="FieldBlock">field section compression</xref> or flowcontrol). </t> <t> Normally,control).</t> <t>Normally, an endpointSHOULD NOT<bcp14>SHOULD NOT</bcp14> send more than one <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame for any stream. However, an endpointMAY<bcp14>MAY</bcp14> send additional <xref target="RST_STREAM" format="none">RST_STREAM</xref> frames if it receives frames on a closed stream after more than a round-trip time. This behavior is permitted to deal with misbehavingimplementations. </t> <t> Toimplementations.</t> <t>To avoid looping, an endpointMUST NOT<bcp14>MUST NOT</bcp14> send a <xref target="RST_STREAM" format="none">RST_STREAM</xref> in response to a <xref target="RST_STREAM" format="none">RST_STREAM</xref>frame. </t>frame.</t> </section> <section> <name>Connection Termination</name><t> If<t>If the TCP connection is closed or reset while streams remain in the "open" or "half-closed" states, then the affected streams cannot be automatically retried (see <xref target="Reliability"/> fordetails). </t>details).</t> </section> </section> <section anchor="extensibility"> <name>Extending HTTP/2</name><t> HTTP/2<t>HTTP/2 permits extension of the protocol. Within the limitations described in this section, protocol extensions can be used to provide additional services or alter any aspect of the protocol. Extensions are effective only within the scope of a single HTTP/2connection. </t> <t> Thisconnection.</t> <t>This applies to the protocol elements defined in this document. This does not affect the existing options for extending HTTP, such as defining new methods, status codes, or fields (see <xreftarget="HTTP" section="16"/>). </t> <t> Extensionstarget="RFC9110" section="16"/>).</t> <t>Extensions are permitted to use new <xref target="FrameHeader">frame types</xref>, new <xref target="SETTINGS">settings</xref>, or new <xref target="ErrorCodes">error codes</xref>. Registries for managing these extension points are defined in <xref section="11"target="RFC7540"/>. </t> <t> Implementations MUSTtarget="RFC7540"/>.</t> <t>Implementations <bcp14>MUST</bcp14> ignore unknown or unsupported values in all extensible protocol elements. ImplementationsMUST<bcp14>MUST</bcp14> discard frames that have unknown or unsupported types. This means that any of these extension points can be safely used by extensions without prior arrangement or negotiation. However, extension frames that appear in the middle of a <xref target="FieldBlock">field block</xref> are not permitted; theseMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Extensions SHOULDformat="none">PROTOCOL_ERROR</xref>.</t> <t>Extensions <bcp14>SHOULD</bcp14> avoid changing protocol elements defined in this document or elements for which no extension mechanism is defined. This includes changes to the layout of frames, additions or changes to the way that frames are composed into <xref target="HttpFraming">HTTP messages</xref>, the definition of pseudo-header fields, or changes to any protocol element that a compliant endpoint might treat as a <xref target="ConnectionErrorHandler">connectionerror</xref>. </t> <t> Anerror</xref>.</t> <t>An extension that changes existing protocol elements or stateMUST<bcp14>MUST</bcp14> be negotiated before being used. For example, an extension that changes the layout of the <xref target="HEADERS" format="none">HEADERS</xref> frame cannot be used until the peer has given a positive signal that this is acceptable. In this case, it could also be necessary to coordinate when the revised layout comes into effect. For example, treating frames other than <xref target="DATA" format="none">DATA</xref> frames as flow controlled requires a change in semantics that both endpoints need to understand, so this can only be done throughnegotiation. </t> <t> Thisnegotiation.</t> <t>This document doesn't mandate a specific method for negotiating the use of an extension but notes that a <xref target="SettingValues">setting</xref> could be used for that purpose. If both peers set a value that indicates willingness to use the extension, then the extension can be used. If a setting is used for extension negotiation, the initial valueMUST<bcp14>MUST</bcp14> be defined in such a fashion that the extension is initiallydisabled. </t>disabled.</t> </section> </section> <section anchor="FrameTypes"> <name>Frame Definitions</name><t> This<t>This specification defines a number of frame types, each identified by a unique 8-bit type code. Each frame type serves a distinct purpose in the establishment and managementeitherof either the connection as a whole orofindividualstreams. </t> <t> Thestreams.</t> <t>The transmission of specific frame types can alter the state of a connection. If endpoints fail to maintain a synchronized view of the connection state, successful communication within the connection will no longer be possible. Therefore, it is important that endpoints have a shared comprehension of how the state is affected by the use of any givenframe. </t>frame.</t> <section anchor="DATA"> <name>DATA</name><t> DATA<t>DATA frames(type=0x0)(type=0x00) convey arbitrary, variable-length sequences of octets associated with a stream. One or more DATA frames are used, for instance, to carry HTTP request or response messagecontents. </t> <t> DATAcontents.</t> <t>DATA framesMAY<bcp14>MAY</bcp14> also contain padding. Padding can be added to DATA frames to obscure the size of messages. Padding is a security feature; see <xreftarget="padding"/>. </t>target="padding"/>.</t> <figure anchor="DATAFrameFormat"> <name>DATA Frame Format</name> <artwork type="inline"><![CDATA[ DATA Frame { Length (24), Type (8) =0x0,0x00, Unused Flags (4), PADDED Flag (1), Unused Flags (2), END_STREAM Flag (1), Reserved (1), Stream Identifier (31), [Pad Length (8)], Data (..), Padding (..2040), } ]]></artwork> </figure><t> The<t>The Length, Type, Unused Flag(s), Reserved, and Stream Identifier fields are described in <xref target="FramingLayer"/>. The DATA frame contains the following additionalfields: </t>fields:</t> <dl newline="false" spacing="normal"> <dt>Pad Length:</dt><dd> An<dd>An 8-bit field containing the length of the frame padding in units of octets. This field is conditional and is only present if the PADDED flag isset. </dd>set.</dd> <dt>Data:</dt><dd> Application<dd>Application data. The amount of data is the remainder of the frame payload after subtracting the length of the other fields that arepresent. </dd>present.</dd> <dt>Padding:</dt><dd> Padding<dd>Padding octets that contain no application semantic value. Padding octetsMUST<bcp14>MUST</bcp14> be set to zero when sending. A receiver is not obligated to verify padding butMAY<bcp14>MAY</bcp14> treat non-zero padding as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </dd>format="none">PROTOCOL_ERROR</xref>.</dd> </dl><t> The<t>The DATA frame defines the followingflags: </t>flags:</t> <dl newline="false" spacing="normal"> <dt>PADDED(0x8):</dt> <dd> When(0x08):</dt> <dd>When set, the PADDED flag indicates that the Pad Length field and any padding that it describes arepresent. </dd>present.</dd> <dt>END_STREAM(0x1):</dt> <dd> When(0x01):</dt> <dd>When set, the END_STREAM flag indicates that this frame is the last that the endpoint will send for the identified stream. Setting this flag causes the stream to enter one of <xref target="StreamStates">the "half-closed" states or the "closed"state</xref>. </dd>state</xref>.</dd> </dl> <aside><t> Note:<t>Note: An endpoint that learns of stream closure after sending all data can close a stream by sending a STREAM frame with a zero-length Data field and the END_STREAM flag set. This is only possible if the endpoint does not sendtrailerstrailers, as the END_STREAM flag appears on a HEADERS frame in that case; see <xreftarget="HttpFraming"/>. </t>target="HttpFraming"/>.</t> </aside><t> DATA<t>DATA framesMUST<bcp14>MUST</bcp14> be associated with a stream. If a DATA frame is received whosestream identifierStream Identifier field is0x0,0x00, the recipientMUST<bcp14>MUST</bcp14> respond with a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> DATAformat="none">PROTOCOL_ERROR</xref>.</t> <t>DATA frames are subject to flow control and can only be sent when a stream is in the "open" or "half-closed (remote)" state. The entire DATA frame payload is included in flow control, including the Pad Length and Padding fields if present. If a DATA frame is received whose stream is not in the "open" or "half-closed (local)" state, the recipientMUST<bcp14>MUST</bcp14> respond with a <xref target="StreamErrorHandler">stream error</xref> of type <xref target="STREAM_CLOSED"format="none">STREAM_CLOSED</xref>. </t> <t> Theformat="none">STREAM_CLOSED</xref>.</t> <t>The total number of padding octets is determined by the value of the Pad Length field. If the length of the padding is the length of the frame payload or greater, the recipientMUST<bcp14>MUST</bcp14> treat this as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> <aside> <t>Note: A frame can be increased in size by one octet by including a Pad Length field with a value ofzero. </t>zero.</t> </aside> </section> <section anchor="HEADERS"> <name>HEADERS</name><t> The<t>The HEADERS frame(type=0x1)(type=0x01) is used to <xref target="StreamStates">open a stream</xref>, and additionally carries a field block fragment. Despite the name, a HEADERS frame can carry a header section or a trailer section. HEADERS frames can be sent on a stream in the "idle", "reserved (local)", "open", or "half-closed (remote)"state. </t>state.</t> <figure anchor="HEADERSFrameFormat"> <name>HEADERS Frame Format</name> <artwork type="inline"><![CDATA[ HEADERS Frame { Length (24), Type (8) =0x1,0x01, Unused Flags (2), PRIORITY Flag (1), Unused Flag (1), PADDED Flag (1), END_HEADERS Flag (1), Unused Flag (1), END_STREAM Flag (1), Reserved (1), Stream Identifier (31), [Pad Length (8)], [Exclusive (1)], [Stream Dependency (31)], [Weight (8)], Field Block Fragment (..), Padding (..2040), } ]]></artwork> </figure><t> The<t>The Length, Type, Unused Flag(s), Reserved, and Stream Identifier fields are described in <xref target="FramingLayer"/>. The HEADERS frame payload has the following additionalfields: </t>fields:</t> <dl newline="false" spacing="normal"> <dt>Pad Length:</dt><dd> An<dd>An 8-bit field containing the length of the frame padding in units of octets. This field is only present if the PADDED flag isset. </dd>set.</dd> <dt>Exclusive:</dt><dd> A<dd>A single-bit flag. This field is only present if the PRIORITY flag is set. Priority signals in HEADERS frames are deprecated; see <xreftarget="PriorityHere"/>. </dd>target="PriorityHere"/>.</dd> <dt>Stream Dependency:</dt><dd> A<dd>A 31-bit stream identifier. This field is only present if the PRIORITY flag isset. </dd>set.</dd> <dt>Weight:</dt><dd> An<dd>An unsigned 8-bit integer. This field is only present if the PRIORITY flag isset. </dd>set.</dd> <dt>Field Block Fragment:</dt><dd> A<dd>A <xref target="FieldBlock">field blockfragment</xref>. </dd>fragment</xref>.</dd> <dt>Padding:</dt><dd> Padding<dd>Padding octets that contain no application semantic value. Padding octetsMUST<bcp14>MUST</bcp14> be set to zero when sending. A receiver is not obligated to verify padding butMAY<bcp14>MAY</bcp14> treat non-zero padding as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </dd>format="none">PROTOCOL_ERROR</xref>.</dd> </dl><t> The<t>The HEADERS frame defines the followingflags: </t>flags:</t> <dl newline="false" spacing="normal"> <dt>PRIORITY (0x20):</dt> <dd><t> When<t>When set, the PRIORITY flag indicates that the Exclusive, Stream Dependency, and Weight fields arepresent. </t>present.</t> </dd> <dt>PADDED(0x8):</dt>(0x08):</dt> <dd><t> When<t>When set, the PADDED flag indicates that the Pad Length field and any padding that it describes arepresent. </t>present.</t> </dd> <dt>END_HEADERS(0x4):</dt>(0x04):</dt> <dd><t> When<t>When set, the END_HEADERS flag indicates that this frame contains an entire <xref target="FieldBlock">field block</xref> and is not followed by any <xref target="CONTINUATION" format="none">CONTINUATION</xref>frames. </t> <t> Aframes.</t> <t>A HEADERS frame without the END_HEADERS flag setMUST<bcp14>MUST</bcp14> be followed by a <xref target="CONTINUATION" format="none">CONTINUATION</xref> frame for the same stream. A receiverMUST<bcp14>MUST</bcp14> treat the receipt of any other type of frame or a frame on a different stream as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> </dd> <dt>END_STREAM(0x1):</dt>(0x01):</dt> <dd><t> When<t>When set, the END_STREAM flag indicates that the <xref target="FieldBlock">field block</xref> is the last that the endpoint will send for the identifiedstream. </t> <t> Astream.</t> <t>A HEADERS frame with the END_STREAM flag set signals the end of a stream. However, a HEADERS frame with the END_STREAM flag set can be followed by <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames on the same stream. Logically, the <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames are part of the HEADERSframe. </t>frame.</t> </dd> </dl><t> The<t>The frame payload of a HEADERS frame contains a <xref target="FieldBlock">field block fragment</xref>. A field block that does not fit within a HEADERS frame is continued in a <xref target="CONTINUATION">CONTINUATIONframe</xref>. </t> <t> HEADERSframe</xref>.</t> <t>HEADERS framesMUST<bcp14>MUST</bcp14> be associated with a stream. If a HEADERS frame is received whosestream identifierStream Identifier field is0x0,0x00, the recipientMUST<bcp14>MUST</bcp14> respond with a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Theformat="none">PROTOCOL_ERROR</xref>.</t> <t>The HEADERS frame changes the connection state as described in <xreftarget="FieldBlock"/>. </t> <t> Thetarget="FieldBlock"/>.</t> <t>The total number of padding octets is determined by the value of the Pad Length field. If the length of the padding is the length of the frame payload or greater, the recipientMUST<bcp14>MUST</bcp14> treat this as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> <aside> <t>Note: A frame can be increased in size by one octet by including a Pad Length field with a value ofzero. </t>zero.</t> </aside> </section> <section anchor="PRIORITY"> <name>PRIORITY</name><t> The<t>The PRIORITY frame(type=0x2)(type=0x02) is deprecated; see <xref target="PriorityHere"/>. A PRIORITY frame can be sent in any stream state, including idle or closedstreams. </t>streams.</t> <figure anchor="PRIORITYFrameFormat"> <name>PRIORITY Frame Format</name> <artwork type="inline"><![CDATA[ PRIORITY Frame { Length (24) =0x5,0x05, Type (8) =0x2,0x02, Unused Flags (8), Reserved (1), Stream Identifier (31), Exclusive (1), Stream Dependency (31), Weight (8), } ]]></artwork> </figure><t> The<t>The Length, Type, Unused Flag(s), Reserved, and Stream Identifier fields are described in <xref target="FramingLayer"/>. The frame payload of a PRIORITY frame contains the following additionalfields: </t>fields:</t> <dl newline="false" spacing="normal"> <dt>Exclusive:</dt><dd> A<dd>A single-bitflag. </dd>flag.</dd> <dt>Stream Dependency:</dt><dd> A<dd>A 31-bit streamidentifier. </dd>identifier.</dd> <dt>Weight:</dt><dd> An<dd>An unsigned 8-bitinteger. </dd>integer.</dd> </dl><t> The<t>The PRIORITY frame does not define anyflags. </t> <t> Theflags.</t> <t>The PRIORITY frame always identifies a stream. If a PRIORITY frame is received with a stream identifier of0x0,0x00, the recipientMUST<bcp14>MUST</bcp14> respond with a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Sendingformat="none">PROTOCOL_ERROR</xref>.</t> <t>Sending or receiving a PRIORITY frame does not affect the state of any stream (<xref target="StreamStates"/>). The PRIORITY frame can be sent on a stream in any state, including "idle" or "closed". A PRIORITY frame cannot be sent between consecutive frames that comprise a single <xref target="FieldBlock">fieldblock</xref>. </t> <t> Ablock</xref>.</t> <t>A PRIORITY frame with a length other than 5 octetsMUST<bcp14>MUST</bcp14> be treated as a <xref target="StreamErrorHandler">stream error</xref> of type <xref target="FRAME_SIZE_ERROR"format="none">FRAME_SIZE_ERROR</xref>. </t>format="none">FRAME_SIZE_ERROR</xref>.</t> </section> <section anchor="RST_STREAM"> <name>RST_STREAM</name><t> The<t>The RST_STREAM frame(type=0x3)(type=0x03) allows for immediate termination of a stream. RST_STREAM is sent to request cancellation of a stream or to indicate that an error condition hasoccurred. </t>occurred.</t> <figure anchor="RST_STREAMFrameFormat"> <name>RST_STREAM Frame Format</name> <artwork type="inline"><![CDATA[ RST_STREAM Frame { Length (24) =0x4,0x04, Type (8) =0x3,0x03, Unused Flags (8), Reserved (1), Stream Identifier (31), Error Code (32), } ]]></artwork> </figure><t> The<t>The Length, Type, Unused Flag(s), Reserved, and Stream Identifier fields are described in <xref target="FramingLayer"/>. Additionally, the RST_STREAM frame contains a single unsigned, 32-bit integer identifying the <xref target="ErrorCodes">error code</xref>. The error code indicates why the stream is beingterminated. </t> <t> Theterminated.</t> <t>The RST_STREAM frame does not define anyflags. </t> <t> Theflags.</t> <t>The RST_STREAM frame fully terminates the referenced stream and causes it to enter the "closed" state. After receiving a RST_STREAM on a stream, the receiverMUST NOT<bcp14>MUST NOT</bcp14> send additional frames for that stream, except for <xref target="PRIORITY" format="none">PRIORITY</xref>. However, after sending the RST_STREAM, the sending endpointMUST<bcp14>MUST</bcp14> be prepared to receive and process additional frames sent on the stream that might have been sent by the peer prior to the arrival of theRST_STREAM. </t> <t> RST_STREAMRST_STREAM.</t> <t>RST_STREAM framesMUST<bcp14>MUST</bcp14> be associated with a stream. If a RST_STREAM frame is received with a stream identifier of0x0,0x00, the recipientMUST<bcp14>MUST</bcp14> treat this as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> RST_STREAMformat="none">PROTOCOL_ERROR</xref>.</t> <t>RST_STREAM framesMUST NOT<bcp14>MUST NOT</bcp14> be sent for a stream in the "idle" state. If a RST_STREAM frame identifying an idle stream is received, the recipientMUST<bcp14>MUST</bcp14> treat this as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Aformat="none">PROTOCOL_ERROR</xref>.</t> <t>A RST_STREAM frame with a length other than 4 octetsMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="FRAME_SIZE_ERROR"format="none">FRAME_SIZE_ERROR</xref>. </t>format="none">FRAME_SIZE_ERROR</xref>.</t> </section> <section anchor="SETTINGS"> <name>SETTINGS</name><t> The<t>The SETTINGS frame(type=0x4)(type=0x04) conveys configuration parameters that affect how endpoints communicate, such as preferences and constraints on peer behavior. The SETTINGS frame is also used to acknowledge the receipt of those settings. Individually, a configuration parameter from a SETTINGS frame is referred to as a"setting". </t> <t> Settings"setting".</t> <t>Settings are not negotiated; they describe characteristics of the sending peer, which are used by the receiving peer. Different values for the same setting can be advertised by each peer. For example, a client might set a high initial flow-control window, whereas a server might set a lower value to conserveresources. </t> <t> Aresources.</t> <t>A SETTINGS frameMUST<bcp14>MUST</bcp14> be sent by both endpoints at the start of a connection andMAY<bcp14>MAY</bcp14> be sent at any other time by either endpoint over the lifetime of the connection. ImplementationsMUST<bcp14>MUST</bcp14> support all of the settings defined by thisspecification. </t> <t> Eachspecification.</t> <t>Each parameter in a SETTINGS frame replaces any existing value for that parameter. Settings are processed in the order in which they appear, and a receiver of a SETTINGS frame does not need to maintain any state other than the current value of each setting. Therefore, the value of a SETTINGS parameter is the last value that is seen by areceiver. </t> <t> SETTINGSreceiver.</t> <t>SETTINGS frames are acknowledged by the receiving peer. To enable this, the SETTINGS frame defines the ACKflag: </t>flag:</t> <dl newline="false" spacing="normal"> <dt>ACK(0x1):</dt> <dd> When(0x01):</dt> <dd>When set, the ACK flag indicates that this frame acknowledges receipt and application of the peer's SETTINGS frame. When this bit is set, the frame payload of the SETTINGS frameMUST<bcp14>MUST</bcp14> be empty. Receipt of a SETTINGS frame with the ACK flag set and a length field value other than 0MUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="FRAME_SIZE_ERROR" format="none">FRAME_SIZE_ERROR</xref>. For more information, see <xref target="SettingsSync"/> ("<xref target="SettingsSync"format="title"/>"). </dd>format="title"/>").</dd> </dl><t> SETTINGS<t>SETTINGS frames always apply to a connection, never a single stream. The stream identifier for a SETTINGS frameMUST<bcp14>MUST</bcp14> be zero(0x0).(0x00). If an endpoint receives a SETTINGS frame whosestream identifierStream Identifier field is anything other than0x0,0x00, the endpointMUST<bcp14>MUST</bcp14> respond with a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Theformat="none">PROTOCOL_ERROR</xref>.</t> <t>The SETTINGS frame affects connection state. A badly formed or incomplete SETTINGS frameMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Aformat="none">PROTOCOL_ERROR</xref>.</t> <t>A SETTINGS frame with a length other than a multiple of 6 octetsMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="FRAME_SIZE_ERROR"format="none">FRAME_SIZE_ERROR</xref>. </t>format="none">FRAME_SIZE_ERROR</xref>.</t> <section anchor="SettingFormat"> <name>SETTINGS Format</name><t> The<t>The frame payload of a SETTINGS frame consists of zero or more settings, each consisting of an unsigned 16-bit setting identifier and an unsigned 32-bitvalue. </t>value.</t> <figure anchor="SettingFrameFormat"> <name>SETTINGS Frame Format</name> <artwork type="inline"><![CDATA[ SETTINGS Frame { Length (24), Type (8) =0x4,0x04, Unused Flags (7), ACK Flag (1), Reserved (1), Stream Identifier (31) = 0, Setting (48) ..., } Setting { Identifier (16), Value (32), } ]]></artwork> </figure><t> The<t>The Length, Type, Unused Flag(s), Reserved, and Stream Identifier fields are described in <xref target="FramingLayer"/>. The frame payload of a SETTINGS frame contains any number of Setting fields, each of which consistsof: </t>of:</t> <dl newline="false" spacing="normal"> <dt>Identifier:</dt><dd> A<dd>A 16-bit setting identifier; see <xreftarget="SettingValues"/>. </dd>target="SettingValues"/>.</dd> <dt>Value:</dt><dd> A<dd>A 32-bit value for thesetting. </dd>setting.</dd> </dl> </section> <section anchor="SettingValues"> <name>Defined Settings</name><t> The<t>The following settings aredefined: </t>defined:</t> <dl newline="false" spacing="normal"> <dt anchor="SETTINGS_HEADER_TABLE_SIZE">SETTINGS_HEADER_TABLE_SIZE(0x1):</dt>(0x01):</dt> <dd><t> Allows<t>This setting allows the sender to inform the remote endpoint of the maximum size of the compression table used to decode field blocks, in units of octets. The encoder can select any size equal to or less than this value by using signaling specific to the compression format inside a field block (see <xreftarget="COMPRESSION"/>).target="RFC7541"/>). The initial value is 4,096octets. </t>octets.</t> </dd> <dt anchor="SETTINGS_ENABLE_PUSH">SETTINGS_ENABLE_PUSH(0x2):</dt>(0x02):</dt> <dd><t> This<t>This setting can be used to enable or disable<xref target="PushResources">server push</xref>.server push. A serverMUST NOT<bcp14>MUST NOT</bcp14> send a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame if it receives this parameter set to a value of0.0; see <xref target="PushResources"/>. A client that has both set this parameter to 0 and had it acknowledgedMUST<bcp14>MUST</bcp14> treat the receipt of a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Theformat="none">PROTOCOL_ERROR</xref>.</t> <t>The initial value of SETTINGS_ENABLE_PUSH is 1. For aclientclient, this value indicates that it is willing to receive PUSH_PROMISE frames. For aserverserver, this initial value has no effect, and is equivalent to the value 0. Any value other than 0 or 1MUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Aformat="none">PROTOCOL_ERROR</xref>.</t> <t>A serverMUST NOT<bcp14>MUST NOT</bcp14> explicitly set this value to 1. A serverMAY<bcp14>MAY</bcp14> choose to omit this setting when it sends a SETTINGS frame, but if a server does include avaluevalue, itMUST<bcp14>MUST</bcp14> be 0. A clientMUST<bcp14>MUST</bcp14> treat receipt of a SETTINGS frame with SETTINGS_ENABLE_PUSH set to 1 as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> </dd> <dt anchor="SETTINGS_MAX_CONCURRENT_STREAMS">SETTINGS_MAX_CONCURRENT_STREAMS(0x3):</dt>(0x03):</dt> <dd><t> Indicates<t>This setting indicates the maximum number of concurrent streams that the sender will allow. This limit is directional: it applies to the number of streams that the sender permits the receiver to create. Initially, there is no limit to this value. It is recommended that this value be no smaller than 100, so as to not unnecessarily limitparallelism. </t> <t> Aparallelism.</t> <t>A value of 0 for SETTINGS_MAX_CONCURRENT_STREAMSSHOULD NOT<bcp14>SHOULD NOT</bcp14> be treated as special by endpoints. A zero value does prevent the creation of new streams; however, this can also happen for any limit that is exhausted with active streams. ServersSHOULD<bcp14>SHOULD</bcp14> only set a zero value for short durations; if a server does not wish to accept requests, closing the connection is moreappropriate. </t>appropriate.</t> </dd> <dt anchor="SETTINGS_INITIAL_WINDOW_SIZE">SETTINGS_INITIAL_WINDOW_SIZE(0x4):</dt>(0x04):</dt> <dd><t> Indicates<t>This setting indicates the sender's initial window size (in units of octets) for stream-level flow control. The initial value is 2<sup>16</sup>-1 (65,535)octets. </t> <t> Thisoctets.</t> <t>This setting affects the window size of all streams (see <xreftarget="InitialWindowSize"/>). </t> <t> Valuestarget="InitialWindowSize"/>).</t> <t>Values above the maximum flow-control window size of 2<sup>31</sup>-1MUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="FLOW_CONTROL_ERROR"format="none">FLOW_CONTROL_ERROR</xref>. </t>format="none">FLOW_CONTROL_ERROR</xref>.</t> </dd> <dt anchor="SETTINGS_MAX_FRAME_SIZE">SETTINGS_MAX_FRAME_SIZE(0x5):</dt>(0x05):</dt> <dd><t> Indicates<t>This setting indicates the size of the largest frame payload that the sender is willing to receive, in units ofoctets. </t> <t> Theoctets.</t> <t>The initial value is 2<sup>14</sup> (16,384) octets. The value advertised by an endpointMUST<bcp14>MUST</bcp14> be between this initial value and the maximum allowed frame size (2<sup>24</sup>-1 or 16,777,215 octets), inclusive. Values outside this rangeMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> </dd> <dt anchor="SETTINGS_MAX_HEADER_LIST_SIZE">SETTINGS_MAX_HEADER_LIST_SIZE(0x6):</dt>(0x06):</dt> <dd><t> This<t>This advisory setting informs a peer of the maximumsize offield section size that the sender is prepared to accept, in units of octets. The value is based on the uncompressed size of field lines, including the length of the name and value in units of octets plus an overhead of 32 octets for each fieldline. </t> <t> Forline.</t> <t>For any given request, a lower limit than what is advertisedMAY<bcp14>MAY</bcp14> be enforced. The initial value of this setting isunlimited. </t>unlimited.</t> </dd> </dl><t> An<t>An endpoint that receives a SETTINGS frame with any unknown or unsupported identifierMUST<bcp14>MUST</bcp14> ignore thatsetting. </t>setting.</t> </section> <section anchor="SettingsSync"> <name>Settings Synchronization</name><t> Most<t>Most values in SETTINGS benefit from or require an understanding of when the peer has received and applied the changed parameter values. In order to provide such synchronization timepoints, the recipient of a SETTINGS frame in which the ACK flag is not setMUST<bcp14>MUST</bcp14> apply the updated settings as soon as possible upon receipt. SETTINGS frames are acknowledged in the order in which they arereceived. </t> <t> Thereceived.</t> <t>The values in the SETTINGS frameMUST<bcp14>MUST</bcp14> be processed in the order they appear, with no other frame processing between values. Unsupported settingsMUST<bcp14>MUST</bcp14> be ignored. Once all values have been processed, the recipientMUST<bcp14>MUST</bcp14> immediately emit a SETTINGS frame with the ACK flag set. Upon receiving a SETTINGS frame with the ACK flag set, the sender of the altered settings can rely on the values from the oldest unacknowledged SETTINGS frame having beenapplied. </t> <t> Ifapplied.</t> <t>If the sender of a SETTINGS frame does not receive an acknowledgment within a reasonable amount of time, itMAY<bcp14>MAY</bcp14> issue a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="SETTINGS_TIMEOUT" format="none">SETTINGS_TIMEOUT</xref>. In setting a timeout, some allowance needs to be made for processing delays at the peer; a timeout that is solely based on theround tripround-trip time between endpoints might result in spuriouserrors. </t>errors.</t> </section> </section> <section anchor="PUSH_PROMISE"> <name>PUSH_PROMISE</name><t> The<t>The PUSH_PROMISE frame(type=0x5)(type=0x05) is used to notify the peer endpoint in advance of streams the sender intends to initiate. The PUSH_PROMISE frame includes the unsigned 31-bit identifier of the stream the endpoint plans to create along with a field section that provides additional context for the stream. <xref target="PushResources"/> contains a thorough description of the use of PUSH_PROMISEframes. </t>frames.</t> <figure anchor="PUSH_PROMISEFrameFormat"> <name>PUSH_PROMISE Frame Format</name> <artwork type="inline"><![CDATA[ PUSH_PROMISE Frame { Length (24), Type (8) =0x5,0x05, Unused Flags (4), PADDED Flag (1), END_HEADERS Flag (1), Unused Flags (2), Reserved (1), Stream Identifier (31), [Pad Length (8)], Reserved (1), Promised Stream ID (31), Field Block Fragment (..), Padding (..2040), } ]]></artwork> </figure><t> The<t>The Length, Type, Unused Flag(s), Reserved, and Stream Identifier fields are described in <xref target="FramingLayer"/>. The PUSH_PROMISE frame payload has the following additionalfields: </t>fields:</t> <dl newline="false" spacing="normal"> <dt>Pad Length:</dt><dd> An<dd>An 8-bit field containing the length of the frame padding in units of octets. This field is only present if the PADDED flag isset. </dd> <dt>Reserved:</dt> <dd> A single reserved bit. </dd>set.</dd> <dt>Promised Stream ID:</dt><dd> An<dd>An unsigned 31-bit integer that identifies the stream that is reserved by the PUSH_PROMISE. The promised stream identifierMUST<bcp14>MUST</bcp14> be a valid choice for the next stream sent by the sender (see "new stream identifier" in <xreftarget="StreamIdentifiers"/>). </dd>target="StreamIdentifiers"/>).</dd> <dt>Field Block Fragment:</dt><dd> A<dd>A <xref target="FieldBlock">field block fragment</xref> containing the request control data and a headersection. </dd>section.</dd> <dt>Padding:</dt><dd> Padding<dd>Padding octets that contain no application semantic value. Padding octetsMUST<bcp14>MUST</bcp14> be set to zero when sending. A receiver is not obligated to verify padding butMAY<bcp14>MAY</bcp14> treat non-zero padding as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </dd>format="none">PROTOCOL_ERROR</xref>.</dd> </dl><t> The<t>The PUSH_PROMISE frame defines the followingflags: </t>flags:</t> <dl newline="false" spacing="normal"> <dt>PADDED(0x8):</dt>(0x08):</dt> <dd><t> When<t>When set, the PADDED flag indicates that the Pad Length field and any padding that it describes arepresent. </t>present.</t> </dd> <dt>END_HEADERS(0x4):</dt>(0x04):</dt> <dd><t> When<t>When set, the END_HEADERS flag indicates that this frame contains an entire <xref target="FieldBlock">field block</xref> and is not followed by any <xref target="CONTINUATION" format="none">CONTINUATION</xref>frames. </t> <t> Aframes.</t> <t>A PUSH_PROMISE frame without the END_HEADERS flag setMUST<bcp14>MUST</bcp14> be followed by a CONTINUATION frame for the same stream. A receiverMUST<bcp14>MUST</bcp14> treat the receipt of any other type of frame or a frame on a different stream as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> </dd> </dl><t> PUSH_PROMISE<t>PUSH_PROMISE framesMUST<bcp14>MUST</bcp14> only be sent on a peer-initiated stream that is in either the "open" or "half-closed (remote)" state. The stream identifier of a PUSH_PROMISE frame indicates the stream it is associated with. If thestream identifierStream Identifier field specifies the value0x0,0x00, a recipientMUST<bcp14>MUST</bcp14> respond with a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Promisedformat="none">PROTOCOL_ERROR</xref>.</t> <t>Promised streams are not required to be used in the order they are promised. The PUSH_PROMISE only reserves stream identifiers for lateruse. </t> <t> PUSH_PROMISE MUST NOTuse.</t> <t>PUSH_PROMISE <bcp14>MUST NOT</bcp14> be sent if the <xref target="SETTINGS_ENABLE_PUSH" format="none">SETTINGS_ENABLE_PUSH</xref> setting of the peer endpoint is set to 0. An endpoint that has set this setting and has received acknowledgmentMUST<bcp14>MUST</bcp14> treat the receipt of a PUSH_PROMISE frame as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Recipientsformat="none">PROTOCOL_ERROR</xref>.</t> <t>Recipients of PUSH_PROMISE frames can choose to reject promised streams by returning a <xref target="RST_STREAM" format="none">RST_STREAM</xref> referencing the promised stream identifier back to the sender of thePUSH_PROMISE. </t> <t> APUSH_PROMISE.</t> <t>A PUSH_PROMISE frame modifies the connection state in two ways. First, the inclusion of a <xref target="FieldBlock">field block</xref> potentially modifies the state maintained for field section compression. Second, PUSH_PROMISE also reserves a stream for later use, causing the promised stream to enter the"reserved""reserved (local)" or "reserved (remote)" state. A senderMUST NOT<bcp14>MUST NOT</bcp14> send a PUSH_PROMISE on a stream unless that stream is either "open" or "half-closed (remote)"; the senderMUST<bcp14>MUST</bcp14> ensure that the promised stream is a valid choice for a <xref target="StreamIdentifiers">new stream identifier</xref> (that is, the promised streamMUST<bcp14>MUST</bcp14> be in the "idle"state). </t> <t> Sincestate).</t> <t>Since PUSH_PROMISE reserves a stream, ignoring a PUSH_PROMISE frame causes the stream state to become indeterminate. A receiverMUST<bcp14>MUST</bcp14> treat the receipt of a PUSH_PROMISE on a stream that is neither "open" nor "half-closed (local)" as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref>. However, an endpoint that has sent <xref target="RST_STREAM" format="none">RST_STREAM</xref> on the associated streamMUST<bcp14>MUST</bcp14> handle PUSH_PROMISE frames that might have been created before the <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame is received andprocessed. </t> <t> Aprocessed.</t> <t>A receiverMUST<bcp14>MUST</bcp14> treat the receipt of a PUSH_PROMISE that promises an <xref target="StreamIdentifiers">illegal stream identifier</xref> as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref>. Note that an illegal stream identifier is an identifier for a stream that is not currently in the "idle"state. </t> <t> Thestate.</t> <t>The total number of padding octets is determined by the value of the Pad Length field. If the length of the padding is the length of the frame payload or greater, the recipientMUST<bcp14>MUST</bcp14> treat this as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> <aside> <t>Note: A frame can be increased in size by one octet by including a Pad Length field with a value ofzero. </t>zero.</t> </aside> </section> <section anchor="PING"> <name>PING</name><t> The<t>The PING frame(type=0x6)(type=0x06) is a mechanism for measuring a minimal round-trip time from the sender, as well as determining whether an idle connection is still functional. PING frames can be sent from anyendpoint. </t>endpoint.</t> <figure anchor="PINGFrameFormat"> <name>PING Frame Format</name> <artwork type="inline"><![CDATA[ PING Frame { Length (24) =0x8,0x08, Type (8) =0x6,0x06, Unused Flags (7), ACK Flag (1), Reserved (1), Stream Identifier (31) = 0, Opaque Data (64), } ]]></artwork> </figure><t> The<t>The Length, Type, Unused Flag(s), Reserved, and Stream Identifier fields are described in <xreftarget="FramingLayer"/>. </t> <t> Intarget="FramingLayer"/>.</t> <t>In addition to the frame header, PING framesMUST<bcp14>MUST</bcp14> contain 8 octets of opaque data in the frame payload. A sender can include any value it chooses and use those octets in anyfashion. </t> <t> Receiversfashion.</t> <t>Receivers of a PING frame that does not include an ACK flagMUST<bcp14>MUST</bcp14> send a PING frame with the ACK flag set in response, with an identical frame payload. PING responsesSHOULD<bcp14>SHOULD</bcp14> be given higher priority than any otherframe. </t> <t> Theframe.</t> <t>The PING frame defines the followingflags: </t>flags:</t> <dl newline="false" spacing="normal"> <dt>ACK(0x1):</dt> <dd> When(0x01):</dt> <dd>When set, the ACK flag indicates that this PING frame is a PING response. An endpointMUST<bcp14>MUST</bcp14> set this flag in PING responses. An endpointMUST NOT<bcp14>MUST NOT</bcp14> respond to PING frames containing thisflag. </dd>flag.</dd> </dl><t> PING<t>PING frames are not associated with any individual stream. If a PING frame is received with astream identifierStream Identifier field value other than0x0,0x00, the recipientMUST<bcp14>MUST</bcp14> respond with a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Receiptformat="none">PROTOCOL_ERROR</xref>.</t> <t>Receipt of a PING frame with a length field value other than 8MUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="FRAME_SIZE_ERROR"format="none">FRAME_SIZE_ERROR</xref>. </t>format="none">FRAME_SIZE_ERROR</xref>.</t> </section> <section anchor="GOAWAY"> <name>GOAWAY</name><t> The<t>The GOAWAY frame(type=0x7)(type=0x07) is used to initiate shutdown of a connection or to signal serious error conditions. GOAWAY allows an endpoint to gracefully stop accepting new streams while still finishing processing of previously established streams. This enables administrative actions, like servermaintenance. </t> <t> Theremaintenance.</t> <t>There is an inherent race condition between an endpoint starting new streams and the remote peer sending a GOAWAY frame. To deal with this case, the GOAWAY contains the stream identifier of the last peer-initiated stream that was or might be processed on the sending endpoint in this connection. For instance, if the server sends a GOAWAY frame, the identified stream is the highest-numbered stream initiated by theclient. </t> <t> Onceclient.</t> <t>Once the GOAWAY is sent, the sender will ignore frames sent on streams initiated by the receiver if the stream has an identifier higher than the included last stream identifier. Receivers of a GOAWAY frameMUST NOT<bcp14>MUST NOT</bcp14> open additional streams on the connection, although a new connection can be established for newstreams. </t> <t> Ifstreams.</t> <t>If the receiver of the GOAWAY has sent data on streams with a higher stream identifier than what is indicated in the GOAWAY frame, those streams are not or will not be processed. The receiver of the GOAWAY frame can treat the streams as though they had never been created at all, thereby allowing those streams to be retried later on a newconnection. </t> <t> Endpoints SHOULDconnection.</t> <t>Endpoints <bcp14>SHOULD</bcp14> always send a GOAWAY frame before closing a connection so that the remote peer can know whether a stream has been partially processed or not. For example, if an HTTP client sends a POST at the same time that a server closes a connection, the client cannot know if the server started to process that POST request if the server does not send a GOAWAY frame to indicate what streams it might have actedon. </t> <t> Anon.</t> <t>An endpoint might choose to close a connection without sending a GOAWAY for misbehavingpeers. </t> <t> Apeers.</t> <t>A GOAWAY frame might not immediately precede closing of the connection; a receiver of a GOAWAY that has no more use for the connectionSHOULD<bcp14>SHOULD</bcp14> still send a GOAWAY frame before terminating theconnection. </t>connection.</t> <figure anchor="GOAWAYFrameFormat"> <name>GOAWAY Frame Format</name> <artwork type="inline"><![CDATA[ GOAWAY Frame { Length (24), Type (8) =0x7,0x07, Unused Flags (8), Reserved (1), Stream Identifier (31) = 0, Reserved (1), Last-Stream-ID (31), Error Code (32), Additional Debug Data (..), } ]]></artwork> </figure><t> The<t>The Length, Type, Unused Flag(s), Reserved, and Stream Identifier fields are described in <xreftarget="FramingLayer"/>. </t> <t> Thetarget="FramingLayer"/>.</t> <t>The GOAWAY frame does not define anyflags. </t> <t> Theflags.</t> <t>The GOAWAY frame applies to the connection, not a specific stream. An endpointMUST<bcp14>MUST</bcp14> treat a <xref target="GOAWAY" format="none">GOAWAY</xref> frame with a stream identifier other than0x00x00 as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Theformat="none">PROTOCOL_ERROR</xref>.</t> <t>The last stream identifier in the GOAWAY frame contains the highest-numbered stream identifier for which the sender of the GOAWAY frame might have taken some action on or might yet take action on. All streams up to and including the identified stream might have been processed in some way. The last stream identifier can be set to 0 if no streams wereprocessed. </t>processed.</t> <aside> <t>Note: In this context, "processed" means that some data from the stream was passed to some higher layer of software that might have taken some action as aresult. </t>result.</t> </aside><t> If<t>If a connection terminates without a GOAWAY frame, the last stream identifier is effectively the highest possible streamidentifier. </t> <t> Onidentifier.</t> <t>On streams with lower- or equal-numbered identifiers that were not closed completely prior to the connection being closed, reattempting requests, transactions, or any protocol activity is not possible, except for idempotent actions like HTTP GET, PUT, or DELETE. Any protocol activity that uses higher-numbered streams can be safely retried using a newconnection. </t> <t> Activityconnection.</t> <t>Activity on streams numbered lower than or equal to the last stream identifier might still complete successfully. The sender of a GOAWAY frame might gracefully shut down a connection by sending a GOAWAY frame, maintaining the connection in an "open" state until all in-progress streamscomplete. </t> <t> Ancomplete.</t> <t>An endpointMAY<bcp14>MAY</bcp14> send multiple GOAWAY frames if circumstances change. For instance, an endpoint that sends GOAWAY with <xref target="NO_ERROR" format="none">NO_ERROR</xref> during graceful shutdown could subsequently encounter a condition that requires immediate termination of the connection. The last stream identifier from the last GOAWAY frame received indicates which streams could have been acted upon. EndpointsMUST NOT<bcp14>MUST NOT</bcp14> increase the value they send in the last stream identifier, since the peers might already have retried unprocessed requests on anotherconnection. </t> <t> Aconnection.</t> <t>A client that is unable to retry requests loses all requests that are in flight when the server closes the connection. This is especially true for intermediaries that might not be serving clients using HTTP/2. A server that is attempting to gracefully shut down a connectionSHOULD<bcp14>SHOULD</bcp14> send an initial GOAWAY frame with the last stream identifier set to 2<sup>31</sup>-1 and a <xref target="NO_ERROR" format="none">NO_ERROR</xref> code. This signals to the client that a shutdown is imminent and that initiating further requests is prohibited. After allowing time for any in-flight stream creation (at least one round-trip time), the serverMAY<bcp14>MAY</bcp14> send another GOAWAY frame with an updated last stream identifier. This ensures that a connection can be cleanly shut down without losingrequests. </t> <t> Afterrequests.</t> <t>After sending a GOAWAY frame, the sender can discard frames for streams initiated by the receiver with identifiers higher than the identified last stream. However, any frames that alter connection state cannot be completely ignored. For instance, <xref target="HEADERS" format="none">HEADERS</xref>, <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>, and <xref target="CONTINUATION" format="none">CONTINUATION</xref> framesMUST<bcp14>MUST</bcp14> be minimally processed to ensure that the state maintained for field section compression is consistent (see <xref target="FieldBlock"/>); similarly, DATA framesMUST<bcp14>MUST</bcp14> be counted toward the connection flow-control window. Failure to process these frames can cause flow control or field section compression state to becomeunsynchronized. </t> <t> Theunsynchronized.</t> <t>The GOAWAY frame also contains a 32-bit <xref target="ErrorCodes">error code</xref> that contains the reason for closing theconnection. </t> <t> Endpoints MAYconnection.</t> <t>Endpoints <bcp14>MAY</bcp14> append opaque data to the frame payload of any GOAWAY frame. Additional debug data is intended for diagnostic purposes only and carries no semantic value. Debug information could contain security- or privacy-sensitive data. Logged or otherwise persistently stored debug dataMUST<bcp14>MUST</bcp14> have adequate safeguards to prevent unauthorizedaccess. </t>access.</t> </section> <section anchor="WINDOW_UPDATE"> <name>WINDOW_UPDATE</name><t> The<t>The WINDOW_UPDATE frame(type=0x8)(type=0x08) is used to implement flow control; see <xref target="FlowControl"/> for anoverview. </t> <t> Flowoverview.</t> <t>Flow control operates at two levels: on each individual stream and on the entireconnection. </t> <t> Bothconnection.</t> <t>Both types of flow control are hop by hop, that is, only between the two endpoints. Intermediaries do not forward WINDOW_UPDATE frames between dependent connections. However, throttling of data transfer by any receiver can indirectly cause the propagation of flow-control information toward the originalsender. </t> <t> Flowsender.</t> <t>Flow control only applies to frames that are identified as being subject to flow control. Of the frame types defined in this document, this includes only <xref target="DATA" format="none">DATA</xref> frames. Frames that are exempt from flow controlMUST<bcp14>MUST</bcp14> be accepted and processed, unless the receiver is unable to assign resources to handling the frame. A receiverMAY<bcp14>MAY</bcp14> respond with a <xref target="StreamErrorHandler">stream error</xref> or <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="FLOW_CONTROL_ERROR" format="none">FLOW_CONTROL_ERROR</xref> if it is unable to accept aframe. </t>frame.</t> <figure anchor="WINDOW_UPDATEFrameFormat"> <name>WINDOW_UPDATE Frame Format</name> <artwork type="inline"><![CDATA[ WINDOW_UPDATE Frame { Length (24) =0x4,0x04, Type (8) =0x8,0x08, Unused Flags (8), Reserved (1), Stream Identifier (31), Reserved (1), Window Size Increment (31), } ]]></artwork> </figure><t> The<t>The Length, Type, Unused Flag(s), Reserved, and Stream Identifier fields are described in <xref target="FramingLayer"/>. The frame payload of a WINDOW_UPDATE frame is one reserved bit plus an unsigned 31-bit integer indicating the number of octets that the sender can transmit in addition to the existing flow-control window. The legal range for the increment to the flow-control window is 1 to 2<sup>31</sup>-1 (2,147,483,647)octets. </t> <t> Theoctets.</t> <t>The WINDOW_UPDATE frame does not define anyflags. </t> <t> Theflags.</t> <t>The WINDOW_UPDATE frame can be specific to a stream or to the entire connection. In the former case, the frame's stream identifier indicates the affected stream; in the latter, the value "0" indicates that the entire connection is the subject of theframe. </t> <t> Aframe.</t> <t>A receiverMUST<bcp14>MUST</bcp14> treat the receipt of a WINDOW_UPDATE frame with a flow-control window increment of 0 as a <xref target="StreamErrorHandler">stream error</xref> of type <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref>; errors on the connection flow-control windowMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connectionerror</xref>. </t> <t> WINDOW_UPDATEerror</xref>.</t> <t>WINDOW_UPDATE can be sent by a peer that has sent a frame with the END_STREAM flag set. This means that a receiver could receive a WINDOW_UPDATE frame on a stream in a "half-closed (remote)" or "closed"stream.state. A receiverMUST NOT<bcp14>MUST NOT</bcp14> treat this as an error (see <xreftarget="StreamStates"/>). </t> <t> Atarget="StreamStates"/>).</t> <t>A receiver that receives a flow-controlled frameMUST<bcp14>MUST</bcp14> always account for its contribution against the connection flow-control window, unless the receiver treats this as a <xref target="ConnectionErrorHandler">connection error</xref>. This is necessary even if the frame is in error. The sender counts the frame toward the flow-control window, but if the receiver does not, the flow-control window at the sender and receiver can becomedifferent. </t> <t> Adifferent.</t> <t>A WINDOW_UPDATE frame with a length other than 4 octetsMUST<bcp14>MUST</bcp14> be treated as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="FRAME_SIZE_ERROR"format="none">FRAME_SIZE_ERROR</xref>. </t>format="none">FRAME_SIZE_ERROR</xref>.</t> <section> <name>The Flow-Control Window</name><t> Flow<t>Flow control in HTTP/2 is implemented using a window kept by each sender on every stream. The flow-control window is a simple integer value that indicates how many octets of data the sender is permitted to transmit; as such, its size is a measure of the buffering capacity of thereceiver. </t> <t> Tworeceiver.</t> <t>Two flow-control windows are applicable: the stream flow-control window and the connection flow-control window. The senderMUST NOT<bcp14>MUST NOT</bcp14> send a flow-controlled frame with a length that exceeds the space available in either of the flow-control windows advertised by the receiver. Frames with zero length with the END_STREAM flag set (that is, an empty <xref target="DATA" format="none">DATA</xref> frame)MAY<bcp14>MAY</bcp14> be sent if there is no available space in either flow-controlwindow. </t> <t> Forwindow.</t> <t>For flow-control calculations, the 9-octet frame header is notcounted. </t> <t> Aftercounted.</t> <t>After sending a flow-controlled frame, the sender reduces the space available in both windows by the length of the transmittedframe. </t> <t> Theframe.</t> <t>The receiver of a frame sends a WINDOW_UPDATE frame as it consumes data and frees up space in flow-control windows. Separate WINDOW_UPDATE frames are sent for the stream- and connection-level flow-control windows. Receivers are advised to have mechanisms in place to avoid sending WINDOW_UPDATE frames with very small increments; see <xref target="RFC1122"section="4.2.3.3"/>. </t> <t> Asection="4.2.3.3"/>.</t> <t>A sender that receives a WINDOW_UPDATE frame updates the corresponding window by the amount specified in theframe. </t> <t> Aframe.</t> <t>A senderMUST NOT<bcp14>MUST NOT</bcp14> allow a flow-control window to exceed 2<sup>31</sup>-1 octets. If a sender receives a WINDOW_UPDATE that causes a flow-control window to exceed this maximum, itMUST<bcp14>MUST</bcp14> terminate either the stream or the connection, as appropriate. For streams, the sender sends a <xref target="RST_STREAM" format="none">RST_STREAM</xref> with an error code of <xref target="FLOW_CONTROL_ERROR" format="none">FLOW_CONTROL_ERROR</xref>; for the connection, a <xref target="GOAWAY" format="none">GOAWAY</xref> frame with an error code of <xref target="FLOW_CONTROL_ERROR" format="none">FLOW_CONTROL_ERROR</xref> issent. </t> <t> Flow-controlledsent.</t> <t>Flow-controlled frames from the sender and WINDOW_UPDATE frames from the receiver are completely asynchronous with respect to each other. This property allows a receiver to aggressively update the window size kept by the sender to prevent streams fromstalling. </t>stalling.</t> </section> <section anchor="InitialWindowSize"> <name>Initial Flow-Control Window Size</name><t> When<t>When an HTTP/2 connection is first established, new streams are created with an initial flow-control window size of 65,535 octets. The connection flow-control window is also 65,535 octets. Both endpoints can adjust the initial window size for new streams by including a value for <xref target="SETTINGS_INITIAL_WINDOW_SIZE" format="none">SETTINGS_INITIAL_WINDOW_SIZE</xref> in the <xref target="SETTINGS" format="none">SETTINGS</xref> frame. The connection flow-control window can only be changed using WINDOW_UPDATEframes. </t> <t> Priorframes.</t> <t>Prior to receiving a <xref target="SETTINGS" format="none">SETTINGS</xref> frame that sets a value for <xref target="SETTINGS_INITIAL_WINDOW_SIZE" format="none">SETTINGS_INITIAL_WINDOW_SIZE</xref>, an endpoint can only use the default initial window size when sending flow-controlled frames. Similarly, the connection flow-control window is set based on the default initial window size until a WINDOW_UPDATE frame isreceived. </t> <t> Inreceived.</t> <t>In addition to changing the flow-control window for streams that are not yet active, a <xref target="SETTINGS" format="none">SETTINGS</xref> frame can alter the initial flow-control window size for streams with active flow-control windows (that is, streams in the "open" or "half-closed (remote)" state). When the value of <xref target="SETTINGS_INITIAL_WINDOW_SIZE" format="none">SETTINGS_INITIAL_WINDOW_SIZE</xref> changes, a receiverMUST<bcp14>MUST</bcp14> adjust the size of all stream flow-control windows that it maintains by the difference between the new value and the oldvalue. </t> <t> Avalue.</t> <t>A change to <xref target="SETTINGS_INITIAL_WINDOW_SIZE" format="none">SETTINGS_INITIAL_WINDOW_SIZE</xref> can cause the available space in a flow-control window to become negative. A senderMUST<bcp14>MUST</bcp14> track the negative flow-control window andMUST NOT<bcp14>MUST NOT</bcp14> send new flow-controlled frames until it receives WINDOW_UPDATE frames that cause the flow-control window to becomepositive. </t> <t> Forpositive.</t> <t>For example, if the client sends 60 KB immediately on connection establishment and the server sets the initial window size to be 16 KB, the client will recalculate the available flow-control window to be -44 KB on receipt of the <xref target="SETTINGS" format="none">SETTINGS</xref> frame. The client retains a negative flow-control window until WINDOW_UPDATE frames restore the window to being positive, after which the client can resumesending. </t> <t> Asending.</t> <t>A <xref target="SETTINGS" format="none">SETTINGS</xref> frame cannot alter the connection flow-controlwindow. </t> <t> Anwindow.</t> <t>An endpointMUST<bcp14>MUST</bcp14> treat a change to <xref target="SETTINGS_INITIAL_WINDOW_SIZE" format="none">SETTINGS_INITIAL_WINDOW_SIZE</xref> that causes any flow-control window to exceed the maximum size as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="FLOW_CONTROL_ERROR"format="none">FLOW_CONTROL_ERROR</xref>. </t>format="none">FLOW_CONTROL_ERROR</xref>.</t> </section> <section> <name>Reducing the Stream Window Size</name><t> A<t>A receiver that wishes to use a smaller flow-control window than the current size can send a new <xref target="SETTINGS" format="none">SETTINGS</xref> frame. However, the receiverMUST<bcp14>MUST</bcp14> be prepared to receive data that exceeds this window size, since the sender might send data that exceeds the lower limit prior to processing the <xref target="SETTINGS" format="none">SETTINGS</xref>frame. </t> <t> Afterframe.</t> <t>After sending a SETTINGS frame that reduces the initial flow-control window size, a receiverMAY<bcp14>MAY</bcp14> continue to process streams that exceed flow-control limits. Allowing streams to continue does not allow the receiver to immediately reduce the space it reserves for flow-control windows. Progress on these streams can also stall, since <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref> frames are needed to allow the sender to resume sending. The receiverMAY<bcp14>MAY</bcp14> instead send a <xref target="RST_STREAM" format="none">RST_STREAM</xref> with an error code of <xref target="FLOW_CONTROL_ERROR" format="none">FLOW_CONTROL_ERROR</xref> for the affectedstreams. </t>streams.</t> </section> </section> <section anchor="CONTINUATION"> <name>CONTINUATION</name><t> The<t>The CONTINUATION frame(type=0x9)(type=0x09) is used to continue a sequence of <xref target="FieldBlock">field block fragments</xref>. Any number of CONTINUATION frames can be sent, as long as the preceding frame is on the same stream and is a <xref target="HEADERS" format="none">HEADERS</xref>, <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>, or CONTINUATION frame without the END_HEADERS flagset. </t>set.</t> <figure anchor="CONTINUATIONFrameFormat"> <name>CONTINUATION Frame Format</name> <artwork type="inline"><![CDATA[ CONTINUATION Frame { Length (24), Type (8) =0x9,0x09, Unused Flags (5), END_HEADERS Flag (1), Unused Flags (2), Reserved (1), Stream Identifier (31), Field Block Fragment (..), } ]]></artwork> </figure><t> The<t>The Length, Type, Unused Flag(s), Reserved, and Stream Identifier fields are described in <xref target="FramingLayer"/>. The CONTINUATION frame payload contains a <xref target="FieldBlock">field blockfragment</xref>. </t> <t> Thefragment</xref>.</t> <t>The CONTINUATION frame defines the followingflag: </t>flag:</t> <dl newline="false" spacing="normal"> <dt>END_HEADERS(0x4):</dt>(0x04):</dt> <dd><t> When<t>When set, the END_HEADERS flag indicates that this frame ends a <xref target="FieldBlock">fieldblock</xref>. </t> <t> Ifblock</xref>.</t> <t>If the END_HEADERS flag is not set, this frameMUST<bcp14>MUST</bcp14> be followed by another CONTINUATION frame. A receiverMUST<bcp14>MUST</bcp14> treat the receipt of any other type of frame or a frame on a different stream as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> </dd> </dl><t> The<t>The CONTINUATION frame changes the connection state as defined in <xreftarget="FieldBlock"/>. </t> <t> CONTINUATIONtarget="FieldBlock"/>.</t> <t>CONTINUATION framesMUST<bcp14>MUST</bcp14> be associated with a stream. If a CONTINUATION frame is receivedwhose stream identifierwith a Stream Identifier fieldis 0x0,of 0x00, the recipientMUST<bcp14>MUST</bcp14> respond with a <xref target="ConnectionErrorHandler">connection error</xref> of typePROTOCOL_ERROR. </t> <t> APROTOCOL_ERROR.</t> <t>A CONTINUATION frameMUST<bcp14>MUST</bcp14> be preceded by a <xref target="HEADERS" format="none">HEADERS</xref>, <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> or CONTINUATION frame without the END_HEADERS flag set. A recipient that observes violation of this ruleMUST<bcp14>MUST</bcp14> respond with a <xreftarget="ConnectionErrorHandler"> connectiontarget="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t>format="none">PROTOCOL_ERROR</xref>.</t> </section> </section> <section anchor="ErrorCodes"> <name>Error Codes</name><t> Error<t>Error codes are 32-bit fields that are used in <xref target="RST_STREAM" format="none">RST_STREAM</xref> and <xref target="GOAWAY" format="none">GOAWAY</xref> frames to convey the reasons for the stream or connectionerror. </t> <t> Errorerror.</t> <t>Error codes share a common code space. Some error codes apply only to either streams or the entire connection and have no defined semantics in the othercontext. </t> <t> Thecontext.</t> <t>The following error codes aredefined: </t>defined:</t> <dl newline="false" spacing="normal"> <dt>NO_ERROR(0x0):</dt>(0x00):</dt> <ddanchor="NO_ERROR"> Theanchor="NO_ERROR">The associated condition is not a result of an error. For example, a <xref target="GOAWAY" format="none">GOAWAY</xref> might include this code to indicate graceful shutdown of aconnection. </dd>connection.</dd> <dt>PROTOCOL_ERROR(0x1):</dt>(0x01):</dt> <ddanchor="PROTOCOL_ERROR"> Theanchor="PROTOCOL_ERROR">The endpoint detected an unspecific protocol error. This error is for use when a more specific error code is notavailable. </dd>available.</dd> <dt>INTERNAL_ERROR(0x2):</dt>(0x02):</dt> <ddanchor="INTERNAL_ERROR"> Theanchor="INTERNAL_ERROR">The endpoint encountered an unexpected internalerror. </dd>error.</dd> <dt>FLOW_CONTROL_ERROR(0x3):</dt>(0x03):</dt> <ddanchor="FLOW_CONTROL_ERROR"> Theanchor="FLOW_CONTROL_ERROR">The endpoint detected that its peer violated the flow-controlprotocol. </dd>protocol.</dd> <dt>SETTINGS_TIMEOUT(0x4):</dt>(0x04):</dt> <ddanchor="SETTINGS_TIMEOUT"> Theanchor="SETTINGS_TIMEOUT">The endpoint sent a <xref target="SETTINGS" format="none">SETTINGS</xref> frame but did not receive a response in a timely manner. See <xref target="SettingsSync"/> ("SettingsSynchronization"). </dd>Synchronization").</dd> <dt>STREAM_CLOSED(0x5):</dt>(0x05):</dt> <ddanchor="STREAM_CLOSED"> Theanchor="STREAM_CLOSED">The endpoint received a frame after a stream washalf-closed. </dd>half-closed.</dd> <dt>FRAME_SIZE_ERROR(0x6):</dt>(0x06):</dt> <ddanchor="FRAME_SIZE_ERROR"> Theanchor="FRAME_SIZE_ERROR">The endpoint received a frame with an invalidsize. </dd>size.</dd> <dt>REFUSED_STREAM(0x7):</dt>(0x07):</dt> <ddanchor="REFUSED_STREAM"> Theanchor="REFUSED_STREAM">The endpoint refused the stream prior to performing any application processing (see <xref target="Reliability"/> fordetails). </dd>details).</dd> <dt>CANCEL(0x8):</dt>(0x08):</dt> <ddanchor="CANCEL"> Used by theanchor="CANCEL">The endpoint uses this error code to indicate that the stream is no longerneeded. </dd>needed.</dd> <dt>COMPRESSION_ERROR(0x9):</dt>(0x09):</dt> <ddanchor="COMPRESSION_ERROR"> Theanchor="COMPRESSION_ERROR">The endpoint is unable to maintain the field section compression context for theconnection. </dd>connection.</dd> <dt>CONNECT_ERROR(0xa):</dt>(0x0a):</dt> <ddanchor="CONNECT_ERROR"> Theanchor="CONNECT_ERROR">The connection established in response to a <xref target="CONNECT">CONNECT request</xref> was reset or abnormallyclosed. </dd>closed.</dd> <dt>ENHANCE_YOUR_CALM(0xb):</dt>(0x0b):</dt> <ddanchor="ENHANCE_YOUR_CALM"> Theanchor="ENHANCE_YOUR_CALM">The endpoint detected that its peer is exhibiting a behavior that might be generating excessiveload. </dd>load.</dd> <dt>INADEQUATE_SECURITY(0xc):</dt>(0x0c):</dt> <ddanchor="INADEQUATE_SECURITY"> Theanchor="INADEQUATE_SECURITY">The underlying transport has properties that do not meet minimum security requirements (see <xreftarget="TLSUsage"/>). </dd>target="TLSUsage"/>).</dd> <dt>HTTP_1_1_REQUIRED(0xd):</dt>(0x0d):</dt> <ddanchor="HTTP_1_1_REQUIRED"> Theanchor="HTTP_1_1_REQUIRED">The endpoint requires that HTTP/1.1 be used instead ofHTTP/2. </dd>HTTP/2.</dd> </dl><t> Unknown<t>Unknown or unsupported error codesMUST NOT<bcp14>MUST NOT</bcp14> trigger any special behavior. TheseMAY<bcp14>MAY</bcp14> be treated by an implementation as being equivalent to <xref target="INTERNAL_ERROR"format="none">INTERNAL_ERROR</xref>. </t>format="none">INTERNAL_ERROR</xref>.</t> </section> <section anchor="HttpLayer"> <name>Expressing HTTP Semantics in HTTP/2</name><t> HTTP/2<t>HTTP/2 is an instantiation of the HTTP message abstraction (<xreftarget="HTTP" section="6"/>). </t>target="RFC9110" section="6"/>).</t> <section anchor="HttpFraming"> <name>HTTP Message Framing</name><t> A<t>A client sends an HTTP request on a new stream, using a previously unused <xref target="StreamIdentifiers">stream identifier</xref>. A server sends an HTTP response on the same stream as therequest. </t> <t> Anrequest.</t> <t>An HTTP message (request or response) consistsof: </t>of:</t> <ol spacing="normal" type="1"><li> one<li>one <xref target="HEADERS" format="none">HEADERS</xref> frame (followed by zero or more <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames) containing the header section (see <xreftarget="HTTP" section="6.3"/>), </li> <li> zerotarget="RFC9110" section="6.3"/>),</li> <li>zero or more <xref target="DATA" format="none">DATA</xref> frames containing the message content (see <xreftarget="HTTP"target="RFC9110" section="6.4"/>),and </li> <li> optionally,and</li> <li>optionally, one <xref target="HEADERS" format="none">HEADERS</xref> frame (followed by zero or more <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames) containing the trailer section, if present (see <xreftarget="HTTP" section="6.5"/>). </li>target="RFC9110" section="6.5"/>).</li> </ol><t> For<t>For a response only, a serverMAY<bcp14>MAY</bcp14> send any number of interim responses before the <xref target="HEADERS" format="none">HEADERS</xref> frame containing a final response. An interim response consists of a <xref target="HEADERS" format="none">HEADERS</xref> frame (which might be followed by zero or more <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames) containing the control data and header section of an interim (1xx) HTTP response (see <xreftarget="HTTP"target="RFC9110" section="15"/>). A <xref target="HEADERS" format="none">HEADERS</xref> frame with the END_STREAM flag set that carries an informational status code is <xreftarget="malformed">malformed</xref>. </t> <t> Thetarget="malformed">malformed</xref>.</t> <t>The last frame in the sequence bears an END_STREAM flag, noting that a <xref target="HEADERS" format="none">HEADERS</xref> frame with the END_STREAM flag set can be followed by <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames that carry any remaining fragments of the fieldblock. </t> <t> Otherblock.</t> <t>Other frames (from any stream)MUST NOT<bcp14>MUST NOT</bcp14> occur between the <xref target="HEADERS" format="none">HEADERS</xref> frame and any <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames that mightfollow. </t> <t> HTTP/2follow.</t> <t>HTTP/2 uses DATA frames to carry message content. The <tt>chunked</tt> transfer encoding defined in <xreftarget="HTTP11"target="RFC9112" section="7.1"/> cannot be used in HTTP/2; see <xreftarget="ConnectionSpecific"/>. </t> <t> Trailertarget="ConnectionSpecific"/>.</t> <t>Trailer fields are carried in a field block that also terminates the stream. That is, trailer fields comprise a sequence starting with a <xref target="HEADERS" format="none">HEADERS</xref> frame, followed by zero or more <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames, where the <xref target="HEADERS" format="none">HEADERS</xref> frame bears an END_STREAM flag. TrailersMUST NOT<bcp14>MUST NOT</bcp14> include <xref target="PseudoHeaderFields">pseudo-header fields</xref>. An endpoint that receives pseudo-header fields in trailersMUST<bcp14>MUST</bcp14> treat the request or response as <xreftarget="malformed">malformed</xref>. </t> <t> Antarget="malformed">malformed</xref>.</t> <t>An endpoint that receives a <xref target="HEADERS" format="none">HEADERS</xref> frame without the END_STREAM flag set after receiving the <xref target="HEADERS" format="none">HEADERS</xref> frame that opens a request or after receiving a final (non-informational) status codeMUST<bcp14>MUST</bcp14> treat the corresponding request or response as <xreftarget="malformed">malformed</xref>. </t> <t> Antarget="malformed">malformed</xref>.</t> <t>An HTTP request/response exchange fully consumes a single stream. A request starts with the <xref target="HEADERS" format="none">HEADERS</xref> frame that puts the stream into the "open" state. The request ends with a frame with the END_STREAM flag set, which causes the stream to become "half-closed (local)" for the client and "half-closed (remote)" for the server. A response stream starts with zero or more interim responses in <xref target="HEADERS" format="none">HEADERS</xref> frames, followed by a <xref target="HEADERS" format="none">HEADERS</xref> frame containing a final statuscode. </t> <t> Ancode.</t> <t>An HTTP response is complete after the server sends -- or the client receives -- a frame with the END_STREAM flag set (including any <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames needed to complete a field block). A server can send a complete response prior to the client sending an entire request if the response does not depend on any portion of the request that has not been sent and received. When this is true, a serverMAY<bcp14>MAY</bcp14> request that the client abort transmission of a request without error by sending a <xref target="RST_STREAM" format="none">RST_STREAM</xref> with an error code of <xref target="NO_ERROR" format="none">NO_ERROR</xref> after sending a complete response (i.e., a frame with the END_STREAM flag set). ClientsMUST NOT<bcp14>MUST NOT</bcp14> discard responses as a result of receiving such a <xref target="RST_STREAM" format="none">RST_STREAM</xref>, though clients can always discard responses at their discretion for otherreasons. </t>reasons.</t> <section anchor="malformed"> <name>Malformed Messages</name><t> A<t>A malformed request or response is one that is an otherwise valid sequence of HTTP/2 frames but is invalid due to the presence of extraneous frames, prohibited fields or pseudo-header fields, the absence of mandatory pseudo-header fields, the inclusion of uppercase field names, or invalid field names and/or values (in certain circumstances; see <xreftarget="HttpHeaders"/>). </t> <t> Atarget="HttpHeaders"/>).</t> <t>A request or response that includes message content can include a <tt>content-length</tt> header field. A request or response is also malformed if the value of a <tt>content-length</tt> header field does not equal the sum of the <xref target="DATA" format="none">DATA</xref> frame payload lengths that form the content, unless the message is defined as having no content. For example, 204 or 304 responses contain no content, as does the response to a HEAD request. A response that is defined to have no content, as described in <xreftarget="HTTP"target="RFC9110" section="6.4.1"/>,MAY<bcp14>MAY</bcp14> have a non-zero <tt>content-length</tt> header field, even though no content is included in <xref target="DATA" format="none">DATA</xref>frames. </t> <t> Intermediariesframes.</t> <t>Intermediaries that process HTTP requests or responses (i.e., any intermediary not acting as a tunnel)MUST NOT<bcp14>MUST NOT</bcp14> forward a malformed request or response. Malformed requests or responses that are detectedMUST<bcp14>MUST</bcp14> be treated as a <xref target="StreamErrorHandler">stream error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Forformat="none">PROTOCOL_ERROR</xref>.</t> <t>For malformed requests, a serverMAY<bcp14>MAY</bcp14> send an HTTP response prior to closing or resetting the stream. ClientsMUST NOT<bcp14>MUST NOT</bcp14> accept a malformedresponse. </t> <t> Endpointsresponse.</t> <t>Endpoints that progressively process messages might have performed some processing before identifying a request or response as malformed. For instance, it might be possible to generate an informational or 404 status code without having received a complete request. Similarly, intermediaries might forward incomplete messages before detecting errors. A serverMAY<bcp14>MAY</bcp14> generate a final response before receiving an entire request when the response does not depend on the remainder of the request beingcorrect. </t> <t> Thesecorrect.</t> <t>These requirements are intended to protect against several types of common attacks against HTTP; they are deliberately strict because being permissive can expose implementations to thesevulnerabilities. </t>vulnerabilities.</t> </section> </section> <section anchor="HttpHeaders"> <name>HTTP Fields</name><t> HTTP<t>HTTP fields (<xreftarget="HTTP"target="RFC9110" section="5"/>) are conveyed by HTTP/2 in the HEADERS, CONTINUATION, and PUSH_PROMISE frames, compressed with <xreftarget="COMPRESSION">HPACK</xref>. </t> <t> Fieldtarget="RFC7541">HPACK</xref>.</t> <t>Field namesMUST<bcp14>MUST</bcp14> be converted to lowercase when constructing an HTTP/2message. </t>message.</t> <section> <name>Field Validity</name><t> The<t>The definitions of field names and values in HTTP prohibit some characters that HPACK might be able to convey. HTTP/2 implementationsSHOULD<bcp14>SHOULD</bcp14> validate field names and values according to their definitions in Sections <xreftarget="HTTP"target="RFC9110" section="5.1" sectionFormat="bare"/> and <xreftarget="HTTP"target="RFC9110" section="5.5" sectionFormat="bare"/> of <xreftarget="HTTP"/> respectivelytarget="RFC9110"/>, respectively, and treat messages that contain prohibited characters as <xreftarget="malformed">malformed</xref>. </t> <t> Failuretarget="malformed">malformed</xref>.</t> <t>Failure to validate fields can be exploited for request smuggling attacks. In particular, unvalidated fields might enable attacks when messages are forwarded using <xreftarget="HTTP11">HTTP 1.1</xref>,target="RFC9112">HTTP/1.1</xref>, where characters such asCR, LF,carriage return (CR), line feed (LF), and COLON are used as delimiters. ImplementationsMUST<bcp14>MUST</bcp14> perform the following minimal validation of field names andvalues: </t>values:</t> <ul><li> A<li>A field nameMUST NOT<bcp14>MUST NOT</bcp14> contain characters in the ranges 0x00-0x20, 0x41-0x5a, or 0x7f-0xff (all ranges inclusive). This specifically excludes all non-visible ASCII characters, ASCII SP (0x20), and uppercase characters ('A' to 'Z', ASCII 0x41 to0x5a). </li> <li> With0x5a).</li> <li>With the exception of <xref target="PseudoHeaderFields">pseudo-header fields</xref>, which have a name that starts with a single colon, field namesMUST NOT<bcp14>MUST NOT</bcp14> include a colon (ASCII COLON,0x3a). </li> <li> A0x3a).</li> <li>A field valueMUST NOT<bcp14>MUST NOT</bcp14> contain the zero value (ASCII NUL,0x0),0x00), line feed (ASCII LF,0xa),0x0a), or carriage return (ASCII CR,0xd)0x0d) at anyposition. </li> <li> Aposition.</li> <li>A field valueMUST NOT<bcp14>MUST NOT</bcp14> start or end with an ASCII whitespace character (ASCII SP or HTAB, 0x20 or0x9). </li>0x09).</li> </ul> <aside><t> Note:<t>Note: An implementation that validates fields according to the definitions in Sections <xreftarget="HTTP"target="RFC9110" section="5.1" sectionFormat="bare"/> and <xreftarget="HTTP"target="RFC9110" section="5.5" sectionFormat="bare"/> of <xreftarget="HTTP"/>target="RFC9110"/> only needs an additional check that field names do not include uppercasecharacters. </t>characters.</t> </aside><t> A<t>A request or response that contains a field that violates any of these conditionsMUST<bcp14>MUST</bcp14> be treated as <xref target="malformed">malformed</xref>. In particular, an intermediary that does not process fields when forwarding messagesMUST NOT<bcp14>MUST NOT</bcp14> forward fields that contain any of the values that are listed as prohibitedabove. </t> <t> Whenabove.</t> <t>When a request message violates one of these requirements, an implementationSHOULD<bcp14>SHOULD</bcp14> generate a 400 (Bad Request) status code (see <xreftarget="HTTP"target="RFC9110" section="15.5.1"/>), unless a more suitable status code is defined or the status code cannot be sent (e.g., because the error occurs in a trailerfield). </t>field).</t> <aside><t> Note:<t>Note: Field values that are not valid according to the definition of the corresponding field do not cause a request to be <xref target="malformed" format="none">malformed</xref>; the requirements above only apply to the generic syntax for fields as defined in <xreftarget="HTTP" section="5"/>. </t>target="RFC9110" section="5"/>.</t> </aside> </section> <section anchor="ConnectionSpecific"> <name>Connection-Specific Header Fields</name><t> HTTP/2<t>HTTP/2 does not use the <tt>Connection</tt> header field (<xreftarget="HTTP"target="RFC9110" section="7.6.1"/>) to indicate connection-specific header fields; in this protocol, connection-specific metadata is conveyed by other means. An endpointMUST NOT<bcp14>MUST NOT</bcp14> generate an HTTP/2 message containing connection-specific header fields. This includes the <tt>Connection</tt> header field and those listed as having connection-specific semantics in <xreftarget="HTTP"target="RFC9110" section="7.6.1"/> (that is, <tt>Proxy-Connection</tt>, <tt>Keep-Alive</tt>, <tt>Transfer-Encoding</tt>, and <tt>Upgrade</tt>). Any message containing connection-specific header fieldsMUST<bcp14>MUST</bcp14> be treated as <xreftarget="malformed">malformed</xref>. </t> <t> Thetarget="malformed">malformed</xref>.</t> <t>The only exception to this is the TE header field, whichMAY<bcp14>MAY</bcp14> be present in an HTTP/2 request; when it is, itMUST NOT<bcp14>MUST NOT</bcp14> contain any value other than"trailers". </t> <t> An"trailers".</t> <t>An intermediary transforming an HTTP/1.x message to HTTP/2MUST<bcp14>MUST</bcp14> remove connection-specific header fields as discussed in <xreftarget="HTTP"target="RFC9110" section="7.6.1"/>, or their messages will be treated by other HTTP/2 endpoints as <xreftarget="malformed">malformed</xref>. </t>target="malformed">malformed</xref>.</t> <aside> <t>Note: HTTP/2 purposefully does not support upgrade to another protocol. The handshake methods described in <xref target="starting"/> are believed sufficient to negotiate the use of alternativeprotocols. </t>protocols.</t> </aside> </section> <section anchor="CompressCookie"> <name>Compressing the Cookie Header Field</name><t> The<t>The <xreftarget="COOKIE">Cookietarget="RFC6265">Cookie header field</xref> uses a semicolon (";") to delimit cookie-pairs (or "crumbs"). This header field contains multiple values, but does not use a COMMA (",") as a separator,which preventsthereby preventing cookie-pairs from being sent on multiple field lines (see <xreftarget="HTTP"target="RFC9110" section="5.2"/>). This can significantly reduce compressionefficiencyefficiency, as updates to individual cookie-pairs would invalidate any field lines that are stored in the HPACKtable. </t> <t> Totable.</t> <t>To allow for better compression efficiency, the Cookie header fieldMAY<bcp14>MAY</bcp14> be split into separate header fields, each with one or more cookie-pairs. If there are multiple Cookie header fields after decompression, theseMUST<bcp14>MUST</bcp14> be concatenated into a single octet string using the two-octet delimiter of 0x3b, 0x20 (the ASCII string "; ") before being passed into a non-HTTP/2 context, such as an HTTP/1.1 connection, or a generic HTTP serverapplication. </t>application.</t> <tkeepWithNext="true"> Therefore,keepWithNext="true">Therefore, the following two lists of Cookie header fields are semanticallyequivalent. </t>equivalent.</t> <artwork type="inline"><![CDATA[ cookie: a=b; c=d; e=f cookie: a=b cookie: c=d cookie: e=f ]]></artwork> </section> </section> <section anchor="PseudoHeaderFields"> <name>HTTP Control Data</name><t> HTTP/2<t>HTTP/2 uses special pseudo-header fields beginning with a ':' character (ASCII 0x3a) to convey message control data (see <xreftarget="HTTP" section="6.2"/>). </t> <t> Pseudo-headertarget="RFC9110" section="6.2"/>).</t> <t>Pseudo-header fields are not HTTP header fields. EndpointsMUST NOT<bcp14>MUST NOT</bcp14> generate pseudo-header fields other than those defined in this document. Note that an extension could negotiate the use of additional pseudo-header fields; see <xreftarget="extensibility"/>. </t> <t> Pseudo-headertarget="extensibility"/>.</t> <t>Pseudo-header fields are only valid in the context in which they are defined. Pseudo-header fields defined for requestsMUST NOT<bcp14>MUST NOT</bcp14> appear in responses; pseudo-header fields defined for responsesMUST NOT<bcp14>MUST NOT</bcp14> appear in requests. Pseudo-header fieldsMUST NOT<bcp14>MUST NOT</bcp14> appear in a trailer section. EndpointsMUST<bcp14>MUST</bcp14> treat a request or response that contains undefined or invalid pseudo-header fields as <xreftarget="malformed">malformed</xref>. </t> <t> Alltarget="malformed">malformed</xref>.</t> <t>All pseudo-header fieldsMUST<bcp14>MUST</bcp14> appear in a field block before all regular field lines. Any request or response that contains a pseudo-header field that appears in a field block after a regular field lineMUST<bcp14>MUST</bcp14> be treated as <xreftarget="malformed">malformed</xref>. </t> <t> Thetarget="malformed">malformed</xref>.</t> <t>The same pseudo-header field nameMUST NOT<bcp14>MUST NOT</bcp14> appear more than once in a field block. A field block for an HTTP request or response that contains a repeated pseudo-header field nameMUST<bcp14>MUST</bcp14> be treated as <xreftarget="malformed">malformed</xref>. </t>target="malformed">malformed</xref>.</t> <section anchor="HttpRequest"> <name>Request Pseudo-Header Fields</name><t> The<t>The following pseudo-header fields are defined for HTTP/2requests: </t>requests:</t> <ul spacing="normal"> <li><t> The <tt>:method</tt><t>The "<tt>:method</tt>" pseudo-header field includes the HTTP method (<xreftarget="HTTP" section="9"/>). </t>target="RFC9110" section="9"/>).</t> </li> <li><t> The <tt>:scheme</tt><t>The "<tt>:scheme</tt>" pseudo-header field includes the scheme portion of the request target. The scheme is taken from the target URI (<xref target="RFC3986" section="3.1"/>) when generating a request directly, or from the scheme of a translated request (for example, see <xreftarget="HTTP11"target="RFC9112" section="3.3"/>). Scheme is omitted for <xref target="CONNECT">CONNECTrequests</xref>. </t> <t><tt>:scheme</tt>requests</xref>.</t> <t>"<tt>:scheme</tt>" is not restricted to<tt>http</tt>"<tt>http</tt>" and<tt>https</tt>"<tt>https</tt>" schemed URIs. A proxy or gateway can translate requests for non-HTTP schemes, enabling the use of HTTP to interact with non-HTTPservices. </t>services.</t> </li> <li><t> The <tt>:authority</tt><t>The "<tt>:authority</tt>" pseudo-header field conveys the authority portion (<xref target="RFC3986" section="3.2"/>) of the target URI (<xreftarget="HTTP"target="RFC9110" section="7.1"/>). The recipient ofaan HTTP/2 requestMUST NOT<bcp14>MUST NOT</bcp14> use the <tt>Host</tt> header field to determine the target URI if<tt>:authority</tt>"<tt>:authority</tt>" ispresent. </t> <t> Clientspresent.</t> <t>Clients that generate HTTP/2 requests directlyMUST<bcp14>MUST</bcp14> use the<tt>:authority</tt>"<tt>:authority</tt>" pseudo-header field to convey authority information, unless there is no authority information to convey (in which case itMUST NOT<bcp14>MUST NOT</bcp14> generate:authority). </t> <t> Clients MUST NOT"<tt>:authority</tt>").</t> <t>Clients <bcp14>MUST NOT</bcp14> generate a request with a <tt>Host</tt> header field that differs from the<tt>:authority</tt>"<tt>:authority</tt>" pseudo-header field. A serverSHOULD<bcp14>SHOULD</bcp14> treat a request as malformed if it contains a <tt>Host</tt> header field that identifiesa differentan entitytothat differs from the<tt>:authority</tt>entity in the "<tt>:authority</tt>" pseudo-header field. The values of fields need to be normalized to compare them (see <xref target="RFC3986" section="6.2"/>). An origin server can apply any normalization method, whereas other serversMUST<bcp14>MUST</bcp14> perform scheme-based normalization (see <xref target="RFC3986" section="6.2.3"/>) of the twofields. </t> <t> Anfields.</t> <t>An intermediary that forwards a request over HTTP/2MUST<bcp14>MUST</bcp14> construct an<tt>:authority</tt>"<tt>:authority</tt>" pseudo-header field using the authority information from the control data of the original request, unless the original request's target URI does not contain authority information (in which case itMUST NOT<bcp14>MUST NOT</bcp14> generate<tt>:authority</tt>)."<tt>:authority</tt>"). Note that the <tt>Host</tt> header field is not the sole source of this information; see <xreftarget="HTTP" section="7.2"/>. </t> <t> Antarget="RFC9110" section="7.2"/>.</t> <t>An intermediary that needs to generate a <tt>Host</tt> header field (which might be necessary to construct an HTTP/1.1 request)MUST<bcp14>MUST</bcp14> use the value from the<tt>:authority</tt>"<tt>:authority</tt>" pseudo-header field as the value of the <tt>Host</tt> field, unless the intermediary also changes the request target. This replaces any existing <tt>Host</tt> field to avoid potential vulnerabilities in HTTProuting. </t> <t> Anrouting.</t> <t>An intermediary that forwards a request over HTTP/2MAY<bcp14>MAY</bcp14> retain any <tt>Host</tt> headerfield. </t> <t> Notefield.</t> <t>Note that request targets for CONNECT or asterisk-form OPTIONS requests never include authority information; see Sections <xref target="RFC9110" section="7.1" sectionFormat="bare"/> and <xreftarget="HTTP" section="7.1"/>. </t> <t> <tt>:authority</tt> MUST NOTtarget="RFC9110" section="7.2" sectionFormat="bare"/> of <xref target="RFC9110"/>.</t> <t>"<tt>:authority</tt>" <bcp14>MUST NOT</bcp14> include the deprecated userinfo subcomponent for<tt>http</tt>"<tt>http</tt>" or<tt>https</tt>"<tt>https</tt>" schemedURIs. </t>URIs.</t> </li> <li><t> The <tt>:path</tt><t>The "<tt>:path</tt>" pseudo-header field includes the path and query parts of the target URI (the <tt>absolute-path</tt> productionand optionallyand, optionally, a '?' character followed by the <tt>query</tt> production; see <xreftarget="HTTP"target="RFC9110" section="4.1"/>). A request in asterisk form (for OPTIONS) includes the value '*' for the<tt>:path</tt>"<tt>:path</tt>" pseudo-headerfield. </t> <t> Thisfield.</t> <t>This pseudo-header fieldMUST NOT<bcp14>MUST NOT</bcp14> be empty for<tt>http</tt>"<tt>http</tt>" or<tt>https</tt>"<tt>https</tt>" URIs;<tt>http</tt>"<tt>http</tt>" or<tt>https</tt>"<tt>https</tt>" URIs that do not contain a path componentMUST<bcp14>MUST</bcp14> include a value of '/'. The exceptions to this ruleare: </t>are:</t> <ul><li> an<li>an OPTIONS request for an<tt>http</tt>"<tt>http</tt>" or<tt>https</tt>"<tt>https</tt>" URI that does not include a path component; theseMUST<bcp14>MUST</bcp14> include a<tt>:path</tt>"<tt>:path</tt>" pseudo-header field with a value of '*' (see <xreftarget="HTTP" section="7.1"/>) </li> <li> <xreftarget="RFC9110" section="7.1"/>).</li> <li><xref target="CONNECT">CONNECT requests</xref>, where the<tt>:path</tt>"<tt>:path</tt>" pseudo-header field isomitted. </li>omitted.</li> </ul> </li> </ul><t> All<t>All HTTP/2 requestsMUST<bcp14>MUST</bcp14> include exactly one valid value for the<tt>:method</tt>, <tt>:scheme</tt>,"<tt>:method</tt>", "<tt>:scheme</tt>", and<tt>:path</tt>"<tt>:path</tt>" pseudo-header fields, unlessit is athey are <xref target="CONNECT">CONNECTrequest</xref>.requests</xref>. An HTTP request that omits mandatory pseudo-header fields is <xreftarget="malformed">malformed</xref>. </t> <t> Individualtarget="malformed">malformed</xref>.</t> <t>Individual HTTP/2 requests do not carry an explicit indicator of protocol version. All HTTP/2 requests implicitly have a protocol version of "2.0" (see <xreftarget="HTTP" section="6.2"/>). </t>target="RFC9110" section="6.2"/>).</t> </section> <section anchor="HttpResponse"> <name>Response Pseudo-Header Fields</name><t> For<t>For HTTP/2 responses, a single<tt>:status</tt>"<tt>:status</tt>" pseudo-header field is defined that carries the HTTP status code field (see <xreftarget="HTTP"target="RFC9110" section="15"/>). This pseudo-header fieldMUST<bcp14>MUST</bcp14> be included in all responses, including interim responses; otherwise, the response is <xreftarget="malformed">malformed</xref>. </t> <t> HTTP/2target="malformed">malformed</xref>.</t> <t>HTTP/2 responses implicitly have a protocol version of"2.0". </t>"2.0".</t> </section> </section> <section anchor="PushResources"> <name>Server Push</name><t> HTTP/2<t>HTTP/2 allows a server to preemptively send (or "push") responses (along with corresponding "promised" requests) to a client in association with a previous client-initiatedrequest. </t> <t> Serverrequest.</t> <t>Server push was designed to allow a server to improve client-perceived performance by predicting what requests will follow those that it receives, thereby removing a round trip for them. For example, a request for HTML is often followed by requests for stylesheets and scripts referenced by that page. When these requests are pushed, the client does not need to wait to receive the references to them in the HTML and issue separaterequests. </t> <t> Inrequests.</t> <t>In practice, server push is difficult to use effectively, because it requires the server to correctly anticipate the additional requests the client will make, taking into account factors such as caching, content negotiation, and user behavior. Errors in prediction can lead to performance degradation, due to the opportunity cost that the additional data on the wire represents. In particular, pushing any significant amount of data can cause contention issues withmore-important responses. </t> <t> Aresponses that are more important.</t> <t>A client can request that server push be disabled, though this is negotiated for each hop independently. The <xref target="SETTINGS_ENABLE_PUSH" format="none">SETTINGS_ENABLE_PUSH</xref> setting can be set to 0 to indicate that server push isdisabled. </t> <t> Promiseddisabled.</t> <t>Promised requestsMUST<bcp14>MUST</bcp14> be safe (see <xreftarget="HTTP"target="RFC9110" section="9.2.1"/>) and cacheable (see <xreftarget="HTTP"target="RFC9110" section="9.2.3"/>). Promised requests cannot include any content or a trailer section. Clients that receive a promised request that is not cacheable, that is not known to be safe, or that indicates the presence of request contentMUST<bcp14>MUST</bcp14> reset the promised stream with a <xref target="StreamErrorHandler">stream error</xref> of type <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref>. Note that this could result in the promised stream being reset if the client does not recognize a newly defined method as beingsafe. </t> <t> Pushedsafe.</t> <t>Pushed responses that are cacheable (see <xreftarget="CACHE"target="RFC9111" section="3"/>) can be stored by the client, if it implements an HTTP cache. Pushed responses are considered successfully validated on the origin server (e.g., if the "no-cache" cache response directive is present; see <xreftarget="CACHE"target="RFC9111" section="5.2.2.4"/>) while the stream identified by the promised streamIDidentifier is stillopen. </t> <t> Pushedopen.</t> <t>Pushed responses that are not cacheableMUST NOT<bcp14>MUST NOT</bcp14> be stored by any HTTP cache. TheyMAY<bcp14>MAY</bcp14> be made available to the applicationseparately. </t> <t> Theseparately.</t> <t>The serverMUST<bcp14>MUST</bcp14> include a value in the<tt>:authority</tt>"<tt>:authority</tt>" pseudo-header field for which the server is authoritative (see <xref target="authority"/>). A clientMUST<bcp14>MUST</bcp14> treat a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> for which the server is not authoritative as a <xref target="StreamErrorHandler">stream error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Anformat="none">PROTOCOL_ERROR</xref>.</t> <t>An intermediary can receive pushes from the server and choose not to forward them on to the client. In other words, how to make use of the pushed information is up to that intermediary. Equally, the intermediary might choose to make additional pushes to the client, without any action taken by theserver. </t> <t> Aserver.</t> <t>A client cannot push. Thus, serversMUST<bcp14>MUST</bcp14> treat the receipt of a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref>. A server cannot set the <xref target="SETTINGS_ENABLE_PUSH" format="none">SETTINGS_ENABLE_PUSH</xref> setting to a value other than 0 (see <xreftarget="SettingValues"/>). </t>target="SettingValues"/>).</t> <section anchor="PushRequests"> <name>Push Requests</name><t> Server<t>Server push is semantically equivalent to a server responding to a request; however, in this case, that request is also sent by the server, as a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>frame. </t> <t> Theframe.</t> <t>The <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame includes a field block that contains control data and a complete set of request header fields that the server attributes to the request. It is not possible to push a response to a request that includes messagecontent. </t> <t> Promisedcontent.</t> <t>Promised requests are always associated with an explicit request from the client. The <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frames sent by the server are sent on that explicit request's stream. The <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame also includes a promised stream identifier, chosen from the stream identifiers available to the server (see <xreftarget="StreamIdentifiers"/>). </t> <t> Thetarget="StreamIdentifiers"/>).</t> <t>The header fields in <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> and any subsequent <xref target="CONTINUATION" format="none">CONTINUATION</xref> framesMUST<bcp14>MUST</bcp14> be a valid and complete set of <xref target="HttpRequest">request header fields</xref>. The serverMUST<bcp14>MUST</bcp14> include a method in the<tt>:method</tt>"<tt>:method</tt>" pseudo-header field that is safe and cacheable. If a client receives a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> that does not include a complete and valid set of header fields or the<tt>:method</tt>"<tt>:method</tt>" pseudo-header field identifies a method that is not safe, itMUST<bcp14>MUST</bcp14> respond on the promised stream with a <xref target="StreamErrorHandler">stream error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Theformat="none">PROTOCOL_ERROR</xref>.</t> <t>The serverSHOULD<bcp14>SHOULD</bcp14> send <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> (<xref target="PUSH_PROMISE"/>) frames prior to sending any frames that reference the promised responses. This avoids a race where clients issue requests prior to receiving any <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>frames. </t> <t> Forframes.</t> <t>For example, if the server receives a request for a document containing embedded links to multiple image files and the server chooses to push those additional images to the client, sending <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frames before the <xref target="DATA" format="none">DATA</xref> frames that contain the image links ensures that the client is able to see that a resource will be pushed before discovering embedded links. Similarly, if the server pushes resources referenced by the field block (for instance, in Link header fields), sending a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> before sending the header ensures that clients do not request thoseresources. </t>resources.</t> <t><xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> framesMUST NOT<bcp14>MUST NOT</bcp14> be sent by theclient. </t>client.</t> <t><xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frames can be sent by the server on any client-initiated stream, but the streamMUST<bcp14>MUST</bcp14> be in either the "open" or "half-closed (remote)" state with respect to the server. <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frames are interspersed with the frames that comprise a response, though they cannot be interspersed with <xref target="HEADERS" format="none">HEADERS</xref> and <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames that comprise a single fieldblock. </t> <t> Sendingblock.</t> <t>Sending a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame creates a new stream and puts the stream into the "reserved (local)" state for the server and the "reserved (remote)" state for theclient. </t>client.</t> </section> <section anchor="PushResponses"> <name>Push Responses</name><t> After<t>After sending the <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame, the server can begin delivering the pushed response as a <xref target="HttpResponse">response</xref> on a server-initiated stream that uses the promised stream identifier. The server uses this stream to transmit an HTTP response, using the same sequence of frames as that defined in <xref target="HttpFraming"/>. This stream becomes <xref target="StreamStates">"half-closed" to the client</xref> after the initial <xref target="HEADERS" format="none">HEADERS</xref> frame issent. </t> <t> Oncesent.</t> <t>Once a client receives a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frame and chooses to accept the pushed response, the clientSHOULD NOT<bcp14>SHOULD NOT</bcp14> issue any requests for the promised response until after the promised stream hasclosed. </t> <t> Ifclosed.</t> <t>If the client determines, for any reason, that it does not wish to receive the pushed response from the server or if the server takes too long to begin sending the promised response, the client can send a <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame, using either the <xref target="CANCEL" format="none">CANCEL</xref> or <xref target="REFUSED_STREAM" format="none">REFUSED_STREAM</xref> code and referencing the pushed stream'sidentifier. </t> <t> Aidentifier.</t> <t>A client can use the <xref target="SETTINGS_MAX_CONCURRENT_STREAMS" format="none">SETTINGS_MAX_CONCURRENT_STREAMS</xref> setting to limit the number of responses that can be concurrently pushed by a server. Advertising a <xref target="SETTINGS_MAX_CONCURRENT_STREAMS" format="none">SETTINGS_MAX_CONCURRENT_STREAMS</xref> value of zero prevents the server from opening the streams necessary to push responses. However, this does not prevent the server from reserving streams using <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frames, because"reserved"reserved streams do not count toward the concurrent stream limit. Clients that do not wish to receive pushed resources need to reset any unwanted reserved streams or set <xref target="SETTINGS_ENABLE_PUSH" format="none">SETTINGS_ENABLE_PUSH</xref> to0. </t> <t> Clients0.</t> <t>Clients receiving a pushed responseMUST<bcp14>MUST</bcp14> validate that either the server is authoritative (see <xref target="authority"/>) or the proxy that provided the pushed response is configured for the corresponding request. For example, a server that offers a certificate for only the <tt>example.com</tt> DNS-ID (see <xref target="RFC6125"/>) is not permitted to push a response for<tt>https://www.example.org/doc</tt>. </t> <t> The<<tt>https://www.example.org/doc</tt>>.</t> <t>The response for a <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> stream begins with a <xref target="HEADERS" format="none">HEADERS</xref> frame, which immediately puts the stream into the "half-closed (remote)" state for the server and "half-closed (local)" state for the client, and ends with a frame with the END_STREAM flag set, which places the stream in the "closed"state. </t>state.</t> <aside> <t>Note: The client never sends a frame with the END_STREAM flag set for a serverpush. </t>push.</t> </aside> </section> </section> <section anchor="CONNECT"> <name>The CONNECT Method</name><t> The<t>The CONNECT method (<xreftarget="HTTP"target="RFC9110" section="9.3.6"/>) is used to convert an HTTP connection into a tunnel to a remote host. CONNECT is primarily used with HTTP proxies to establish a TLS session with an origin server for the purposes of interacting with<tt>https</tt> resources. </t> <t> In"<tt>https</tt>" resources.</t> <t>In HTTP/2, the CONNECT method establishes a tunnel over a single HTTP/2 stream to a remote host, rather than converting the entire connection to a tunnel. A CONNECT header section is constructed as defined in <xref target="HttpRequest"/> ("<xref target="HttpRequest" format="title"/>"), with a few differences.Specifically: </t>Specifically:</t> <ul spacing="normal"><li> The <tt>:method</tt><li>The "<tt>:method</tt>" pseudo-header field is set to<tt>CONNECT</tt>. </li> <li> The <tt>:scheme</tt><tt>CONNECT</tt>.</li> <li>The "<tt>:scheme</tt>" and<tt>:path</tt>"<tt>:path</tt>" pseudo-header fieldsMUST<bcp14>MUST</bcp14> beomitted. </li> <li> The <tt>:authority</tt>omitted.</li> <li>The "<tt>:authority</tt>" pseudo-header field contains the host and port to connect to (equivalent to the authority-form of the request-target of CONNECT requests; see <xreftarget="HTTP11" section="3.2.3"/>). </li>target="RFC9112" section="3.2.3"/>).</li> </ul><t> A<t>A CONNECT request that does not conform to these restrictions is <xreftarget="malformed">malformed</xref>. </t> <t> Atarget="malformed">malformed</xref>.</t> <t>A proxy that supports CONNECT establishes a <xreftarget="TCP">TCPtarget="RFC0793">TCP connection</xref> to the host and port identified in the<tt>:authority</tt>"<tt>:authority</tt>" pseudo-header field. Once this connection is successfully established, the proxy sends a <xref target="HEADERS" format="none">HEADERS</xref> frame containing a2xx series2xx-series status code to the client, as defined in <xreftarget="HTTP" section="9.3.6"/>. </t> <t> Aftertarget="RFC9110" section="9.3.6"/>.</t> <t>After the initial <xref target="HEADERS" format="none">HEADERS</xref> frame sent by each peer, all subsequent <xref target="DATA" format="none">DATA</xref> frames correspond to data sent on the TCP connection. The frame payload of any <xref target="DATA" format="none">DATA</xref> frames sent by the client is transmitted by the proxy to the TCP server; data received from the TCP server is assembled into <xref target="DATA" format="none">DATA</xref> frames by the proxy. Frame types other than <xref target="DATA" format="none">DATA</xref> or stream management frames (<xref target="RST_STREAM" format="none">RST_STREAM</xref>, <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref>, and <xref target="PRIORITY" format="none">PRIORITY</xref>)MUST NOT<bcp14>MUST NOT</bcp14> be sent on a connected stream andMUST<bcp14>MUST</bcp14> be treated as a <xref target="StreamErrorHandler">stream error</xref> ifreceived. </t> <t> Thereceived.</t> <t>The TCP connection can be closed by either peer. The END_STREAM flag on a <xref target="DATA" format="none">DATA</xref> frame is treated as being equivalent to the TCP FIN bit. A client is expected to send a <xref target="DATA" format="none">DATA</xref> frame with the END_STREAM flag set after receiving a frame with the END_STREAM flag set. A proxy that receives a <xref target="DATA" format="none">DATA</xref> frame with the END_STREAM flag set sends the attached data with the FIN bit set on the last TCP segment. A proxy that receives a TCP segment with the FIN bit set sends a <xref target="DATA" format="none">DATA</xref> frame with the END_STREAM flag set. Note that the final TCP segment or <xref target="DATA" format="none">DATA</xref> frame could beempty. </t> <t> Aempty.</t> <t>A TCP connection error is signaled with <xref target="RST_STREAM" format="none">RST_STREAM</xref>. A proxy treats any error in the TCP connection, which includes receiving a TCP segment with the RST bit set, as a <xref target="StreamErrorHandler">stream error</xref> of type <xref target="CONNECT_ERROR" format="none">CONNECT_ERROR</xref>. Correspondingly, a proxyMUST<bcp14>MUST</bcp14> send a TCP segment with the RST bit set if it detects an error with the stream or the HTTP/2connection. </t>connection.</t> </section> <section anchor="informational-responses"> <name>The Upgrade Header Field</name><t> HTTP/2<t>HTTP/2 does not support the 101 (Switching Protocols) informational status code (<xreftarget="HTTP" section="15.2.2"/>). </t> <t> Thetarget="RFC9110" section="15.2.2"/>).</t> <t>The semantics of 101 (Switching Protocols) aren't applicable to a multiplexed protocol. Similar functionality might be enabled through the use of <xref target="RFC8441">extendedCONNECT</xref>CONNECT</xref>, and other protocols are able to use the same mechanisms that HTTP/2 uses to negotiate their use (see <xreftarget="starting"/>). </t>target="starting"/>).</t> </section> <section anchor="Reliability"> <name>Request Reliability</name><t> In<t>In general, an HTTP client is unable to retry a non-idempotent request when an error occurs because there is no means to determine the nature of the error (see <xreftarget="HTTP"target="RFC9110" section="9.2.2"/>). It is possible that some server processing occurred prior to the error, which could result in undesirable effects if the request werereattempted. </t> <t> HTTP/2reattempted.</t> <t>HTTP/2 provides two mechanisms for providing a guarantee to a client that a request has not beenprocessed: </t>processed:</t> <ul spacing="normal"><li> The<li>The <xref target="GOAWAY" format="none">GOAWAY</xref> frame indicates the highest stream number that might have been processed. Requests on streams with higher numbers are therefore guaranteed to be safe toretry. </li> <li> Theretry.</li> <li>The <xref target="REFUSED_STREAM" format="none">REFUSED_STREAM</xref> error code can be included in a <xref target="RST_STREAM" format="none">RST_STREAM</xref> frame to indicate that the stream is being closed prior to any processing having occurred. Any request that was sent on the reset stream can be safelyretried. </li>retried.</li> </ul><t> Requests<t>Requests that have not been processed have not failed; clientsMAY<bcp14>MAY</bcp14> automatically retry them, even those with non-idempotentmethods. </t> <t> Amethods.</t> <t>A serverMUST NOT<bcp14>MUST NOT</bcp14> indicate that a stream has not been processed unless it can guarantee that fact. If frames that are on a stream are passed to the application layer for any stream, then <xref target="REFUSED_STREAM" format="none">REFUSED_STREAM</xref>MUST NOT<bcp14>MUST NOT</bcp14> be used for that stream, and a <xref target="GOAWAY" format="none">GOAWAY</xref> frameMUST<bcp14>MUST</bcp14> include a stream identifier that is greater than or equal to the given streamidentifier. </t> <t> Inidentifier.</t> <t>In addition to these mechanisms, the <xref target="PING" format="none">PING</xref> frame provides a way for a client to easily test a connection. Connections that remain idle can becomebroken asbroken, because some middleboxes (for instance, network address translators or load balancers) silently discard connection bindings. The <xref target="PING" format="none">PING</xref> frame allows a client to safely test whether a connection is still active without sending arequest. </t>request.</t> </section> <section anchor="HttpExamples"> <name>Examples</name><t> This<t>This section shows HTTP/1.1 requests and responses, with illustrations of equivalent HTTP/2 requests andresponses. </t> <section><name>Simpleresponses.</t> <section> <name>Simple Request</name><t> An<t>An HTTP GET request includes control data and a request header with no message content and is therefore transmitted as a single <xref target="HEADERS" format="none">HEADERS</xref> frame, followed by zero or more <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames containing the serialized block of request header fields. The <xref target="HEADERS" format="none">HEADERS</xref> frame in the following has both the END_HEADERS and END_STREAM flags set; no <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames aresent. </t>sent.</t> <artwork type="inline"><![CDATA[ GET /resource HTTP/1.1 HEADERS Host: example.org ==> + END_STREAM Accept: image/jpeg + END_HEADERS :method = GET :scheme = https :authority = example.org :path = /resource host = example.org accept = image/jpeg ]]></artwork> </section><section><name>Simple<section> <name>Simple Response</name><t> Similarly,<t>Similarly, a response that includes only control data and a response header is transmitted as a <xref target="HEADERS" format="none">HEADERS</xref> frame (again, followed by zero or more <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames) containing the serialized block of response headerfields. </t>fields.</t> <artwork type="inline"><![CDATA[ HTTP/1.1 304 Not Modified HEADERS ETag: "xyzzy" ==> + END_STREAM Expires: Thu, 23 Jan ... + END_HEADERS :status = 304 etag = "xyzzy" expires = Thu, 23 Jan ... ]]></artwork> </section><section><name>Complex<section> <name>Complex Request</name><t> An<t>An HTTP POST request that includes control data and a request headerandwith message content is transmitted as one <xref target="HEADERS" format="none">HEADERS</xref> frame, followed by zero or more <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames containing the request header, followed by one or more <xref target="DATA" format="none">DATA</xref> frames, with the last <xref target="CONTINUATION" format="none">CONTINUATION</xref> (or <xref target="HEADERS" format="none">HEADERS</xref>) frame having the END_HEADERS flag set and the final <xref target="DATA" format="none">DATA</xref> frame having the END_STREAM flagset: </t>set:</t> <artwork type="inline"><![CDATA[ POST /resource HTTP/1.1 HEADERS Host: example.org ==> - END_STREAM Content-Type: image/jpeg - END_HEADERS Content-Length: 123 :method = POST :authority = example.org :path = /resource {binary data} :scheme = https CONTINUATION + END_HEADERS content-type = image/jpeg host = example.org content-length = 123 DATA + END_STREAM {binary data} ]]></artwork> <tkeepWithPrevious="true"> NotekeepWithPrevious="true">Note that data contributing to any given field line could be spread between field block fragments. The allocation of field lines to frames in this example is illustrativeonly. </t>only.</t> </section><section><name>Response<section> <name>Response with Body</name><t> A<t>A response that includes control data and a response headerandwith message content is transmitted as a <xref target="HEADERS" format="none">HEADERS</xref> frame, followed by zero or more <xref target="CONTINUATION" format="none">CONTINUATION</xref> frames, followed by one or more <xref target="DATA" format="none">DATA</xref> frames, with the last <xref target="DATA" format="none">DATA</xref> frame in the sequence having the END_STREAM flagset: </t>set:</t> <artwork type="inline"><![CDATA[ HTTP/1.1 200 OK HEADERS Content-Type: image/jpeg ==> - END_STREAM Content-Length: 123 + END_HEADERS :status = 200 {binary data} content-type = image/jpeg content-length = 123 DATA + END_STREAM {binary data} ]]></artwork> </section><section><name>Informational<section> <name>Informational Responses</name><t> An<t>An informational response using a 1xx status code other than 101 is transmitted as a <xref target="HEADERS" format="none">HEADERS</xref> frame, followed by zero or more <xref target="CONTINUATION" format="none">CONTINUATION</xref>frames. </t> <t> Aframes.</t> <t>A trailer section is sent as a field block after both the request or response field block and all the <xref target="DATA" format="none">DATA</xref> frames have been sent. The <xref target="HEADERS" format="none">HEADERS</xref> frame starting the field block that comprises the trailer section has the END_STREAM flagset. </t>set.</t> <tkeepWithNext="true"> ThekeepWithNext="true">The following example includes both a 100 (Continue) status code, which is sent in response to a request containing a "100-continue" token in the Expect header field, and a trailersection: </t>section:</t> <artwork type="inline"><![CDATA[ HTTP/1.1 100 Continue HEADERS Extension-Field: bar ==> - END_STREAM + END_HEADERS :status = 100 extension-field = bar HTTP/1.1 200 OK HEADERS Content-Type: image/jpeg ==> - END_STREAM Transfer-Encoding: chunked + END_HEADERS Trailer: Foo :status = 200 content-type = image/jpeg 123 trailer = Foo {binary data} 0 DATA Foo: bar - END_STREAM {binary data} HEADERS + END_STREAM + END_HEADERS foo = bar ]]></artwork> </section> </section> </section> <section anchor="HttpExtra"> <name>HTTP/2 Connections</name><t> This<t>This section outlines attributes oftheHTTPprotocolthat improve interoperability, reduce exposure to known security vulnerabilities, or reduce the potential for implementationvariation. </t>variation.</t> <section> <name>Connection Management</name><t> HTTP/2<t>HTTP/2 connections are persistent. For best performance, it is expected that clients will not close connections until it is determined that no further communication with a server is necessary (for example, when a user navigates away from a particular web page) or until the server closes theconnection. </t> <t> Clients SHOULD NOTconnection.</t> <t>Clients <bcp14>SHOULD NOT</bcp14> open more than one HTTP/2 connection to a given host and port pair, where the host is derived from a URI, a selected <xreftarget="ALT-SVC">alternativetarget="RFC7838">alternative service</xref>, or a configuredproxy. </t> <t> Aproxy.</t> <t>A client can create additional connections as replacements, either to replace connections that are near to exhausting the available <xref target="StreamIdentifiers">stream identifier space</xref>, to refresh the keying material for a TLS connection, or to replace connections that have encountered <xreftarget="ConnectionErrorHandler">errors</xref>. </t> <t> Atarget="ConnectionErrorHandler">errors</xref>.</t> <t>A clientMAY<bcp14>MAY</bcp14> open multiple connections to the same IP address and TCP port using different <xreftarget="TLS-EXT">Servertarget="RFC6066">Server Name Indication</xref> values or to provide different TLS client certificates butSHOULD<bcp14>SHOULD</bcp14> avoid creating multiple connections with the sameconfiguration. </t> <t> Serversconfiguration.</t> <t>Servers are encouraged to maintain open connections for as long as possible but are permitted to terminate idle connections if necessary. When either endpoint chooses to close the transport-layer TCP connection, the terminating endpointSHOULD<bcp14>SHOULD</bcp14> first send a <xref target="GOAWAY" format="none">GOAWAY</xref> (<xref target="GOAWAY"/>) frame so that both endpoints can reliably determine whether previously sent frames have been processed and gracefully complete or terminate any necessary remainingtasks. </t>tasks.</t> <section anchor="reuse"> <name>Connection Reuse</name><t> Connections<t>Connections that are made to an origin server, either directly or through a tunnel created using the <xref target="CONNECT">CONNECT method</xref>,MAY<bcp14>MAY</bcp14> be reused for requests with multiple different URI authority components. A connection can be reused as long as the origin server is <xref target="authority">authoritative</xref>. For TCP connections without TLS, this depends on the host having resolved to the same IPaddress. </t> <t> For <tt>https</tt>address.</t> <t>For "<tt>https</tt>" resources, connection reuse additionally depends on having a certificate that is valid for the host in the URI. The certificate presented by the serverMUST<bcp14>MUST</bcp14> satisfy any checks that the client would perform when forming a new TLS connection for the host in the URI. A single certificate can be used to establish authority for multiple origins. <xreftarget="HTTP"target="RFC9110" section="4.3"/> describes how a client determines whether a server is authoritative for aURI. </t> <t> InURI.</t> <t>In some deployments, reusing a connection for multiple origins can result in requests being directed to the wrong origin server. For example, TLS termination might be performed by a middlebox that uses the TLS <xreftarget="TLS-EXT">Servertarget="RFC6066">Server NameIndication (SNI)</xref>Indication</xref> extension to select an origin server. This means that it is possible for clients to send requests to servers that might not be the intended target for the request, even though the server is otherwiseauthoritative. </t> <t> Aauthoritative.</t> <t>A server that does not wish clients to reuse connections can indicate that it is not authoritative for a request by sending a 421 (Misdirected Request) status code in response to the request (see <xreftarget="HTTP" section="15.5.20"/>). </t> <t> Atarget="RFC9110" section="15.5.20"/>).</t> <t>A client that is configured to use a proxy over HTTP/2 directs requests to that proxy through a single connection. That is, all requests sent via a proxy reuse the connection to theproxy. </t>proxy.</t> </section> </section> <section anchor="TLSUsage"> <name>Use of TLS Features</name><t> Implementations<t>Implementations of HTTP/2MUST<bcp14>MUST</bcp14> use <xreftarget="TLS12">TLStarget="RFC5246">TLS version 1.2</xref> or higher for HTTP/2 over TLS. The general TLS usage guidance in <xreftarget="TLSBCP"/> SHOULDtarget="RFC7525"/> <bcp14>SHOULD</bcp14> be followed, with some additional restrictions that are specific toHTTP/2. </t> <t> TheHTTP/2.</t> <t>The TLS implementationMUST<bcp14>MUST</bcp14> support the <xreftarget="TLS-EXT">Servertarget="RFC6066">Server Name Indication (SNI)</xref> extension to TLS. If the server is identified by a <xreftarget="DNS-TERMS">domaintarget="RFC8499">domain name</xref>, clientsMUST<bcp14>MUST</bcp14> send the server_name TLS extension unless an alternative mechanism to indicate the target host isused. </t> <t> Requirementsused.</t> <t>Requirements for deployments of HTTP/2 that negotiate <xreftarget="TLS13">TLStarget="RFC8446">TLS 1.3</xref> are included in <xref target="tls13features"/>. Deployments of TLS 1.2 are subject to the requirements in<xref target="tls12features"/>Sections <xref target="tls12features" format="counter"/> and <xreftarget="tls12ciphers"/>.target="tls12ciphers" format="counter"/>. Implementations are encouraged to provide defaults that comply, but it is recognized that deployments are ultimately responsible forcompliance. </t>compliance.</t> <section anchor="tls12features"> <name>TLS 1.2 Features</name><t> This<t>This section describes restrictions on the TLS 1.2 feature set that can be used with HTTP/2. Due to deployment limitations, it might not be possible to fail TLS negotiation when these restrictions are not met. An endpointMAY<bcp14>MAY</bcp14> immediately terminate an HTTP/2 connection that does not meet these TLS requirements with a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="INADEQUATE_SECURITY"format="none">INADEQUATE_SECURITY</xref>. </t> <t> Aformat="none">INADEQUATE_SECURITY</xref>.</t> <t>A deployment of HTTP/2 over TLS 1.2MUST<bcp14>MUST</bcp14> disable compression. TLS compression can lead to the exposure of information that would not otherwise be revealed <xref target="RFC3749"/>. Generic compression isunnecessaryunnecessary, since HTTP/2 provides compression features that are more aware of context and therefore likely to be more appropriate for use for performance, security, or otherreasons. </t> <t> Areasons.</t> <t>A deployment of HTTP/2 over TLS 1.2MUST<bcp14>MUST</bcp14> disable renegotiation. An endpointMUST<bcp14>MUST</bcp14> treat a TLS renegotiation as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR" format="none">PROTOCOL_ERROR</xref>. Note that disabling renegotiation can result in long-lived connections becoming unusable due to limits on the number of messages the underlying cipher suite canencipher. </t> <t> Anencipher.</t> <t>An endpointMAY<bcp14>MAY</bcp14> use renegotiation to provide confidentiality protection for client credentials offered in the handshake, but any renegotiationMUST<bcp14>MUST</bcp14> occur prior to sending the connection preface. A serverSHOULD<bcp14>SHOULD</bcp14> request a client certificate if it sees a renegotiation request immediately after establishing aconnection. </t> <t> Thisconnection.</t> <t>This effectively prevents the use of renegotiation in response to a request for a specific protected resource. A future specification might provide a way to support this use case. Alternatively, a server might use an <xreftarget="ErrorHandler"> error</xref>target="ErrorHandler">error</xref> of type <xref target="HTTP_1_1_REQUIRED" format="none">HTTP_1_1_REQUIRED</xref> to request that the client use a protocol that supportsrenegotiation. </t> <t> Implementations MUSTrenegotiation.</t> <t>Implementations <bcp14>MUST</bcp14> support ephemeral key exchange sizes of at least 2048 bits for cipher suites that use ephemeral finite field Diffie-Hellman (DHE) (<xreftarget="TLS12" section="8.1.2"/>target="RFC5246" section="8.1.2"/>) and 224 bits for cipher suites that use ephemeral elliptic curve Diffie-Hellman (ECDHE) <xref target="RFC8422"/>. ClientsMUST<bcp14>MUST</bcp14> accept DHE sizes of up to 4096 bits. EndpointsMAY<bcp14>MAY</bcp14> treat negotiation of key sizes smaller than the lower limits as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="INADEQUATE_SECURITY"format="none">INADEQUATE_SECURITY</xref>. </t>format="none">INADEQUATE_SECURITY</xref>.</t> </section> <section anchor="tls12ciphers"> <name>TLS 1.2 Cipher Suites</name><t> A<t>A deployment of HTTP/2 over TLS 1.2SHOULD NOT<bcp14>SHOULD NOT</bcp14> use any of the prohibited cipher suitesthat arelisted inthe<xreftarget="BadCipherSuites">list of prohibited cipher suites</xref>. </t> <t> Endpoints MAYtarget="BadCipherSuites"/>.</t> <t>Endpoints <bcp14>MAY</bcp14> choose to generate a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="INADEQUATE_SECURITY" format="none">INADEQUATE_SECURITY</xref> if one of the prohibited cipher suites is negotiated. A deployment that chooses to use a prohibited cipher suite risks triggering a connection error unless the set of potential peers is known to accept that ciphersuite. </t> <t> Implementations MUST NOTsuite.</t> <t>Implementations <bcp14>MUST NOT</bcp14> generate this error in reaction to the negotiation of a cipher suite that is not prohibited. Consequently, when clients offer a cipher suite that is not prohibited, they have to be prepared to use that cipher suite withHTTP/2. </t> <t> TheHTTP/2.</t> <t>The list of prohibited cipher suites includes the cipher suite that TLS 1.2 makes mandatory, which means that TLS 1.2 deployments could have non-intersecting sets of permitted cipher suites. To avoid thisproblem causingproblem, which causes TLS handshake failures, deployments of HTTP/2 that use TLS 1.2MUST<bcp14>MUST</bcp14> support TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 <xreftarget="TLS-ECDHE"/>target="RFC5289"/> with the P-256 elliptic curve <xreftarget="RFC8422"/>. </t> <t> Notetarget="RFC8422"/>.</t> <t>Note that clients might advertise support of cipher suites that are prohibited in order to allow for connection to servers that do not support HTTP/2. This allows servers to select HTTP/1.1 with a cipher suite that is prohibited in HTTP/2. However, this can result in HTTP/2 being negotiated with a prohibited cipher suite if the application protocol and cipher suite are independentlyselected. </t>selected.</t> </section> <section anchor="tls13features"> <name>TLS 1.3 Features</name><t> TLS<t>TLS 1.3 includes a number of features not available in earlier versions. This section discusses the use of thesefeatures. </t> <t> HTTP/2features.</t> <t>HTTP/2 serversMUST NOT<bcp14>MUST NOT</bcp14> send post-handshake TLS 1.3 CertificateRequest messages. HTTP/2 clientsMUST<bcp14>MUST</bcp14> treat a TLS post-handshake CertificateRequest message as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="PROTOCOL_ERROR"format="none">PROTOCOL_ERROR</xref>. </t> <t> Theformat="none">PROTOCOL_ERROR</xref>.</t> <t>The prohibition on post-handshake authentication applies even if the client offered the "post_handshake_auth" TLS extension. Post-handshake authentication support might be advertised independently of <xreftarget="TLS-ALPN">ALPN</xref>.target="RFC7301">ALPN</xref>. Clients might offer the capability for use in other protocols, but inclusion of the extension cannot imply support withinHTTP/2. </t>HTTP/2.</t> <t><xreftarget="TLS13"/>target="RFC8446"/> defines other post-handshake messages, NewSessionTicket and KeyUpdate, which can be used as they have no direct interaction with HTTP/2. Unless the use of a new type of TLS message depends on an interaction with the application-layer protocol, that TLS message can be sent after the handshakecompletes. </t> <t> TLScompletes.</t> <t>TLS early dataMAY<bcp14>MAY</bcp14> be used to send requests, provided that the guidance in <xref target="RFC8470"/> is observed. Clients send requests in early data assuming initial values for all serversettings. </t>settings.</t> </section> </section> </section> <section anchor="security"> <name>Security Considerations</name><t> The<t>The use of TLS is necessary to provide many of the security properties of this protocol. Many of the claims in this section do not hold unless TLS is used as described in <xreftarget="TLSUsage"/>. </t>target="TLSUsage"/>.</t> <section anchor="authority"> <name>Server Authority</name><t> HTTP/2<t>HTTP/2 relies on the HTTP definition of authority for determining whether a server is authoritative in providing a given response (see <xreftarget="HTTP"target="RFC9110" section="4.3"/>). This relies on local name resolution for the"http""<tt>http</tt>" URI scheme and the authenticated server identity for the"https" scheme. </t>"<tt>https</tt>" scheme.</t> </section> <section> <name>Cross-Protocol Attacks</name><t> In<t>In a cross-protocol attack, an attacker causes a client to initiate a transaction in one protocol toward a server that understands a different protocol. An attacker might be able to cause the transaction to appear as a valid transaction in the second protocol. In combination with the capabilities of the web context, this can be used to interact with poorly protected servers in privatenetworks. </t> <t> Completingnetworks.</t> <t>Completing a TLS handshake with an ALPN identifier for HTTP/2 can be considered sufficient protection against cross-protocol attacks. ALPN provides a positive indication that a server is willing to proceed with HTTP/2, which prevents attacks on other TLS-basedprotocols. </t> <t> Theprotocols.</t> <t>The encryption in TLS makes it difficult for attackers to control the data that could be used in a cross-protocol attack on a cleartextprotocol. </t> <t> Theprotocol.</t> <t>The cleartext version of HTTP/2 has minimal protection against cross-protocol attacks. The <xref target="preface">connection preface</xref> contains a string that is designed to confuse HTTP/1.1 servers, but no special protection is offered for otherprotocols. </t>protocols.</t> </section> <section> <name>Intermediary Encapsulation Attacks</name><t> HPACK<t>HPACK permits encoding of field names and values that might be treated as delimiters in other HTTP versions. An intermediary that translates an HTTP/2 request or responseMUST<bcp14>MUST</bcp14> validate fields according to the rules in <xref target="HttpHeaders"/> before translating a message to another HTTP version. Translating a field that includes invalid delimiters could be used to cause recipients to incorrectly interpret a message, which could be exploited by anattacker. </t> <t> <xrefattacker.</t> <t><xref target="HttpHeaders"/> does not include specific rules for validation of pseudo-header fields. If the values of these fields are used, additional validation is necessary. This is particularly important where<tt>:scheme</tt>, <tt>:authority</tt>,"<tt>:scheme</tt>", "<tt>:authority</tt>", and<tt>:path</tt>"<tt>:path</tt>" are combined to form a single URI string(<xref target="RFC3986"/>).<xref target="RFC3986"/>. Similar problems might occur when that URI or just<tt>:path</tt> are"<tt>:path</tt>" is combined with<tt>:method</tt>"<tt>:method</tt>" to construct a request line (as in <xreftarget="HTTP11"target="RFC9112" section="3"/>). Simple concatenation is not secure unless the input values are fullyvalidated. </t> <t> Anvalidated.</t> <t>An intermediary can reject fields that contain invalid field names or values for otherreasons,reasons -- inparticularparticular, those fields that do not conform to the HTTP ABNF grammar from <xreftarget="HTTP"target="RFC9110" section="5"/>. Intermediaries that do not perform any validation of fields other than the minimum required by <xref target="HttpHeaders"/> could forward messages that contain invalid field names orvalues. </t> <t> Anvalues.</t> <t>An intermediary that receives anyfieldfields thatrequiresrequire removal before forwarding (see <xreftarget="HTTP"target="RFC9110" section="7.6.1"/>)MUST<bcp14>MUST</bcp14> remove or replace those header fields when forwarding messages. Additionally, intermediaries should take care when forwarding messages containingContent-Length<tt>Content-Length</tt> fields to ensure that the message is <xref target="malformed">well-formed</xref>. This ensures that if the message is translated into HTTP/1.1 at anypointpoint, the framing will becorrect. </t>correct.</t> </section> <section> <name>Cacheability of Pushed Responses</name><t> Pushed<t>Pushed responses do not have an explicit request from the client; the request is provided by the server in the <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref>frame. </t> <t> Cachingframe.</t> <t>Caching responses that are pushed is possible based on the guidance provided by the origin server in the Cache-Control header field. However, this can cause issues if a single server hosts more than one tenant. For example, a server might offer multiple users each a small portion of its URIspace. </t> <t> Wherespace.</t> <t>Where multiple tenants share space on the same server, that serverMUST<bcp14>MUST</bcp14> ensure that tenants are not able to push representations of resources that they do not have authority over. Failure to enforce this would allow a tenant to provide a representation that would be served out of cache, overriding the actual representation that the authoritative tenantprovides. </t> <t> Pushedprovides.</t> <t>Pushed responses for which an origin server is not authoritative (see <xref target="authority"/>)MUST NOT<bcp14>MUST NOT</bcp14> be used orcached. </t>cached.</t> </section> <section anchor="dos"> <name>Denial-of-Service Considerations</name><t> An<t>An HTTP/2 connection can demand a greater commitment of resources to operate than an HTTP/1.1 connection.The use ofBoth field section compression and flow control depend on a commitment ofresources for storinga greater amount of state. Settings for these features ensure that memory commitments for these features are strictlybounded. </t> <t> Thebounded.</t> <t>The number of <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frames is not constrained in the same fashion. A client that accepts server pushSHOULD<bcp14>SHOULD</bcp14> limit the number of streams it allows to be in the "reserved (remote)" state. An excessive number of server push streams can be treated as a <xref target="StreamErrorHandler">stream error</xref> of type <xref target="ENHANCE_YOUR_CALM"format="none">ENHANCE_YOUR_CALM</xref>. </t> <t> Aformat="none">ENHANCE_YOUR_CALM</xref>.</t> <t>A number of HTTP/2 implementations were found to be vulnerable to denial of service <xref target="NFLX-2019-002"/>.The following listsBelow is a list of known ways that implementations might be subject todenial of service attack: </t>denial-of-service attacks:</t> <ul> <li><t> Inefficient<t>Inefficient tracking of outstanding outbound frames can lead to overload if an adversary can cause large numbers of frames to be enqueued for sending. A peer could use one of several techniques to cause large numbers of frames to begenerated: </t>generated:</t> <ul><li> Providing<li>Providing tiny increments to flow control in <xref target="WINDOW_UPDATE" format="none">WINDOW_UPDATE</xref> frames can cause a sender to generate a large number of <xref target="DATA" format="none">DATA</xref>frames. </li> <li> Anframes.</li> <li>An endpoint is required to respond to a <xref target="PING" format="none">PING</xref>frame. </li> <li> Eachframe.</li> <li>Each <xref target="SETTINGS" format="none">SETTINGS</xref> frame requiresacknowledgment. </li> <li> Anacknowledgment.</li> <li>An invalid request (or server push) can cause a peer to send <xref target="RST_STREAM" format="none">RST_STREAM</xref> frames inresponse. </li>response.</li> </ul> </li><li> An<li>An attacker can provide large amounts offlow controlflow-control credit at the HTTP/2layer,layer but withhold credit at the TCP layer, preventing frames from being sent. An endpoint that constructs and remembers frames for sending without considering TCP limits might be subject to resourceexhaustion. </li> <li> Largeexhaustion.</li> <li>Large numbers of small or empty frames can be abused to cause a peer to expend time processing frame headers. Caution is required here as some uses of small frames are entirely legitimate, such as the sending of an empty <xref target="DATA" format="none">DATA</xref> or <xref target="CONTINUATION" format="none">CONTINUATION</xref> frame at the end of astream. </li> <li> Thestream.</li> <li>The <xref target="SETTINGS" format="none">SETTINGS</xref> frame might also be abused to cause a peer to expend additional processing time. This might be done by pointlessly changing settings, sending multiple undefined settings, or changing the same setting multiple times in the sameframe. </li> <li> Handlingframe.</li> <li>Handling reprioritization with <xref target="PRIORITY" format="none">PRIORITY</xref> frames can require significant processing time and can lead to overload if many <xref target="PRIORITY" format="none">PRIORITY</xref> frames aresent. </li> <li> Fieldsent.</li> <li>Field section compression alsooffers someprovides opportunities for an attacker to waste processing resources; see <xreftarget="COMPRESSION"target="RFC7541" section="7"/> for more details on potentialabuses. </li> <li> Limitsabuses.</li> <li>Limits in <xref target="SETTINGS" format="none">SETTINGS</xref> cannot be reduced instantaneously, which leaves an endpoint exposed to behavior from a peer that could exceed the new limits. In particular, immediately after establishing a connection, limits set by a server are not known to clients and could be exceeded without being an obvious protocolviolation. </li>violation.</li> </ul><t> Most<t>Most of the features that might be exploited for denial of service --i.e.,such as <xref target="SETTINGS" format="none">SETTINGS</xref> changes, small frames, field section compression -- have legitimate uses. These features become a burden only when they are used unnecessarily or toexcess. </t> <t> Anexcess.</t> <t>An endpoint that doesn't monitor use of these features exposes itself to a risk of denial of service. ImplementationsSHOULD<bcp14>SHOULD</bcp14> track the use of these features and set limits on their use. An endpointMAY<bcp14>MAY</bcp14> treat activity that is suspicious as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="ENHANCE_YOUR_CALM"format="none">ENHANCE_YOUR_CALM</xref>. </t>format="none">ENHANCE_YOUR_CALM</xref>.</t> <section anchor="MaxFieldBlock"> <name>Limits on Field Block Size</name><t> A<t>A large <xref target="FieldBlock">field block</xref> can cause an implementation to commit a large amount of state. Field lines that are critical for routing can appear toward the end of a field block, which prevents streaming of fields to their ultimate destination. This ordering and other reasons, such as ensuring cache correctness, mean that an endpoint might need to buffer the entire field block. Since there is no hard limit to the size of a field block, some endpoints could be forced to commit a large amount of available memory for fieldblocks. </t> <t> Anblocks.</t> <t>An endpoint can use the <xref target="SETTINGS_MAX_HEADER_LIST_SIZE" format="none">SETTINGS_MAX_HEADER_LIST_SIZE</xref> to advise peers of limits that might apply on the size of uncompressed field blocks. This setting is only advisory, so endpointsMAY<bcp14>MAY</bcp14> choose to send field blocks that exceed this limit and risk the request or response being treated as malformed. This setting is specific to a connection, so any request or response could encounter a hop with a lower, unknown limit. An intermediary can attempt to avoid this problem by passing on values presented by different peers, but they are not obliged to doso. </t> <t> Aso.</t> <t>A server that receives a larger field block than it is willing to handle can send an HTTP 431 (Request Header Fields Too Large) status code <xref target="RFC6585"/>. A client can discard responses that it cannot process. The field blockMUST<bcp14>MUST</bcp14> be processed to ensure a consistent connection state, unless the connection isclosed. </t>closed.</t> </section> <section anchor="connectDos"> <name>CONNECT Issues</name><t> The<t>The CONNECT method can be used to create disproportionate load on a proxy, since stream creation is relatively inexpensive when compared to the creation and maintenance of a TCP connection. A proxy might also maintain some resources for a TCP connection beyond the closing of the stream that carries the CONNECT request, since the outgoing TCP connection remains in the TIME_WAIT state. Therefore, a proxy cannot rely on <xref target="SETTINGS_MAX_CONCURRENT_STREAMS" format="none">SETTINGS_MAX_CONCURRENT_STREAMS</xref> alone to limit the resources consumed by CONNECTrequests. </t>requests.</t> </section> </section> <section> <name>Use of Compression</name><t> Compression<t>Compression can allow an attacker to recover secret data when it is compressed in the same context as data under attacker control. HTTP/2 enables compression of field lines (<xref target="FieldBlock"/>); the following concerns also apply to the use of HTTP compressed content-codings (<xreftarget="HTTP" section="8.4.1"/>). </t> <t> Theretarget="RFC9110" section="8.4.1"/>).</t> <t>There are demonstrable attacks on compression that exploit the characteristics of thewebWeb (e.g., <xref target="BREACH"/>). The attacker induces multiple requests containing varying plaintext, observing the length of the resulting ciphertext in each, which reveals a shorter length when a guess about the secret iscorrect. </t> <t> Implementationscorrect.</t> <t>Implementations communicating on a secure channelMUST NOT<bcp14>MUST NOT</bcp14> compress content that includes both confidential and attacker-controlled data unless separate compression dictionaries are used for each source of data. CompressionMUST NOT<bcp14>MUST NOT</bcp14> be used if the source of data cannot be reliably determined. Generic stream compression, such as that provided by TLS,MUST NOT<bcp14>MUST NOT</bcp14> be used with HTTP/2 (see <xreftarget="TLSUsage"/>). </t> <t> Furthertarget="TLSUsage"/>).</t> <t>Further considerations regarding the compression of header fields are described in <xreftarget="COMPRESSION"/>. </t>target="RFC7541"/>.</t> </section> <section anchor="padding"> <name>Use of Padding</name><t> Padding<t>Padding within HTTP/2 is not intended as a replacement for general purpose padding, such as that provided by <xreftarget="TLS13">TLS</xref>.target="RFC8446">TLS</xref>. Redundant padding could even be counterproductive. Correct application can depend on having specific knowledge of the data that is beingpadded. </t> <t> Topadded.</t> <t>To mitigate attacks that rely on compression, disabling or limiting compression might be preferable to padding as acountermeasure. </t> <t> Paddingcountermeasure.</t> <t>Padding can be used to obscure the exact size of frame content and is provided to mitigate specific attacks withinHTTP,HTTP -- for example, attacks where compressed content includes both attacker-controlled plaintext and secret data (e.g., <xreftarget="BREACH"/>). </t> <t> Usetarget="BREACH"/>).</t> <t>Use of padding can result in less protection than might seem immediately obvious. At best, padding only makes it more difficult for an attacker to infer length information by increasing the number of frames an attacker has to observe. Incorrectly implemented padding schemes can be easily defeated. In particular, randomized padding with a predictable distribution provides very little protection; similarly, padding frame payloads to a fixed size exposes information as frame payload sizes cross the fixed-sized boundary, which could be possible if an attacker can controlplaintext. </t> <t> Intermediaries SHOULDplaintext.</t> <t>Intermediaries <bcp14>SHOULD</bcp14> retain padding for <xref target="DATA" format="none">DATA</xref>frames,frames butMAY<bcp14>MAY</bcp14> drop padding for <xref target="HEADERS" format="none">HEADERS</xref> and <xref target="PUSH_PROMISE" format="none">PUSH_PROMISE</xref> frames. A valid reason for an intermediary to change the amount of padding of frames is to improve the protections that paddingprovides. </t>provides.</t> </section> <section> <name>Privacy Considerations</name><t> Several<t>Several characteristics of HTTP/2 provide an observer an opportunity to correlate actions of a single client or server over time. These include thevaluevalues of settings, the manner in which flow-control windows are managed, the way priorities are allocated to streams, the timing of reactions to stimulus, and the handling of any features that are controlled bysettings. </t> <t> Assettings.</t> <t>As far as these create observable differences in behavior, they could be used as a basis for fingerprinting a specific client, as defined in <xreftarget="PRIVACY" section="3.2"/>. </t> <t> HTTP/2'starget="RFC6973" section="3.2"/>.</t> <t>HTTP/2's preference for using a single TCP connection allows correlation of a user's activity on a site. Reusing connections for different origins allows tracking across thoseorigins. </t> <t> Becauseorigins.</t> <t>Because the PING and SETTINGS frames solicit immediate responses, they can be used by an endpoint to measure latency to their peer. This might have privacy implications in certainscenarios. </t>scenarios.</t> </section> <section> <name>Remote Timing Attacks</name><t> Remote<t>Remote timing attacks extract secrets from servers by observing variations in the time that servers take when processing requests that use secrets. HTTP/2 enables concurrent request creation and processing, which can give attackers better control over when request processing commences. Multiple HTTP/2 requests can be included in the same IP packet or TLS record. HTTP/2 can therefore make remote timing attacks more efficient by eliminating variability in request delivery, leaving only request order and the delivery of responses as sources of timingvariability. </t> <t> Ensuringvariability.</t> <t>Ensuring that processing time is not dependent on the value ofsecretsa secret is the best defense against any form of timingattack. </t>attack.</t> </section> </section> <section anchor="iana"> <name>IANA Considerations</name><t> This<t>This revision ofthe documentHTTP/2 marks the <tt>HTTP2-Settings</tt> header field and the <tt>h2c</tt>Upgradeupgrade token, both defined in <xref target="RFC7540"/>, asobsolete. </t> <t> <xrefobsolete.</t> <t><xref target="RFC7540" section="11"/> registered the <tt>h2</tt> and <tt>h2c</tt> ALPN identifiers along with the <tt>PRI</tt> HTTP method. RFC 7540 also established a registry for frame types, settings, and error codes. These registrations and registries apply to HTTP/2, but are not redefined in thisdocument. </t> <t> IANA is requested to updatedocument.</t> <t>IANA has updated references to RFC 7540 in the following registries to refer to this document: "TLS Application-Layer Protocol Negotiation (ALPN) ProtocolIDs, HTTP/2IDs", "HTTP/2 FrameType, HTTP/2 Settings, HTTP/2Type", "HTTP/2 Settings", "HTTP/2 ErrorCode,Code", andHTTP"HTTP MethodRegistry.Registry". The registration of the <tt>PRI</tt> methodneeds to behas been updated to refer to <xref target="preface"/>; all other section numbers have notchanged. </t> <t> IANA is requested to changechanged.</t> <t>IANA has changed the policy on those portions of the "HTTP/2 Frame Type" and "HTTP/2 Settings" registries that were reserved for Experimental Use in RFC 7540. These portions of theregistryregistries shall operate on the same policy as the remainder of eachregistry. </t>registry.</t> <section anchor="HTTP2-Settings"> <name>HTTP2-Settings Header Field Registration</name><t> This<t>This section marks the <tt>HTTP2-Settings</tt> header field registered by <xref target="RFC7540" section="11.5"/> in theHypertext"Hypertext Transfer Protocol (HTTP) Field NameRegistryRegistry" as obsolete. This capability has been removed: see <xref target="versioning"/>. The registration is updated to include the details as required by <xreftarget="HTTP" section="18.4"/>: </t>target="RFC9110" section="18.4"/>:</t> <dl newline="false" spacing="normal"> <dt>Field Name:</dt> <dd>HTTP2-Settings</dd> <dt>Status:</dt><dd>Standard</dd> <dt>Ref.:</dt> <dd> <xref<dd>obsoleted</dd> <dt>Reference:</dt> <dd><xref target="RFC7540"section="3.2.1"/> </dd>section="3.2.1"/></dd> <dt>Comments:</dt> <dd>Obsolete; see <xref target="HTTP2-Settings"/> of thisdocument</dd>document.</dd> </dl> </section> <section anchor="iana-h2c"> <name>The h2c Upgrade Token</name><t> This<t>This section records the <tt>h2c</tt> upgrade token registered by <xref target="RFC7540" section="11.8"/> in theHypertext"Hypertext Transfer Protocol (HTTP) Upgrade TokenRegistryRegistry" as obsolete. This capability has been removed: see <xref target="versioning"/>. The registration is updated asfollows: </t>follows:</t> <dl><dt>Value:</dt><dd>h2c</dd> <dt>Description:</dt><dd>Hypertext<dt>Value:</dt> <dd>h2c</dd> <dt>Description:</dt> <dd>(OBSOLETE) Hypertext Transfer Protocol version 2 (HTTP/2)</dd> <dt>Expected VersionTokens:</dt><dd>None</dd> <dt>Reference:</dt><dd><xrefTokens:</dt> <dd>None</dd> <dt>Reference:</dt> <dd><xref target="versioning"/> of this document</dd> </dl> </section> </section> </middle> <back> <displayreference target="RFC0793" to="TCP"/> <displayreference target="RFC5246" to="TLS12"/> <displayreference target="RFC5289" to="TLS-ECDHE"/> <displayreference target="RFC6066" to="TLS-EXT"/> <displayreference target="RFC6265" to="COOKIE"/> <displayreference target="RFC6973" to="PRIVACY"/> <displayreference target="RFC7301" to="TLS-ALPN"/> <displayreference target="RFC7525" to="TLSBCP"/> <displayreference target="RFC7541" to="COMPRESSION"/> <displayreference target="RFC7838" to="ALT-SVC"/> <displayreference target="RFC8446" to="TLS13"/> <displayreference target="RFC8499" to="DNS-TERMS"/> <displayreference target="RFC9000" to="QUIC"/> <displayreference target="RFC9110" to="HTTP"/> <displayreference target="RFC9111" to="CACHING"/> <displayreference target="RFC9112" to="HTTP/1.1"/> <displayreference target="RFC9218" to="HTTP-PRIORITY"/> <references> <name>References</name> <references> <name>Normative References</name> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.0793.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.2119.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.3986.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5246.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5289.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.6066.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.6265.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7301.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7525.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7541.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8174.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8422.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8446.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8470.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.9000.xml"/> <!-- draft-ietf-httpbis-semantics (RFC 9110) --> <referenceanchor="COMPRESSION"> <front> <title>HPACK: Header Compression for HTTP/2</title> <seriesInfo name="RFC" value="7541"/> <seriesInfo name="DOI" value="10.17487/RFC7541"/> <author initials="R." surname="Peon" fullname="Roberto Peon"/> <author initials="H." surname="Ruellan" fullname="Herve Ruellan"/> <date month="May" year="2015"/> </front> </reference> <reference anchor="TCP">anchor="RFC9110" target="https://www.rfc-editor.org/info/rfc9110"> <front><title abbrev="Transmission Control Protocol"> Transmission Control Protocol </title> <seriesInfo name="STD" value="7"/> <seriesInfo name="RFC" value="793"/> <seriesInfo name="DOI" value="10.17487/RFC793"/><title>HTTP Semantics</title> <authorinitials="J." surname="Postel" fullname="Jon Postel"> <organization>University of Southern California (USC)/Information Sciences Institute</organization>initials="R" surname="Fielding" fullname="Roy Fielding" role="editor"> <organization/> </author><date year="1981" month="September"/> </front> </reference> <reference anchor="TLSBCP"> <front> <title>Recommendations for Secure Use of Transport Layer Security (TLS) and Datagram Transport Layer Security (DTLS)</title> <seriesInfo name="BCP" value="195"/> <seriesInfo name="RFC" value="7525"/> <seriesInfo name="DOI" value="10.17487/RFC7525"/> <author initials="Y." surname="Sheffer" fullname="Yaron Sheffer"/> <author initials="R." surname="Holz" fullname="Ralph Holz"/> <author initials="P." surname="Saint-Andre" fullname="Peter Saint-Andre"/> <date month="May" year="2015"/> </front> </reference> <reference anchor="RFC2119"> <front> <title> Key words for use in RFCs to Indicate Requirement Levels </title> <seriesInfo name="BCP" value="14"/> <seriesInfo name="RFC" value="2119"/> <seriesInfo name="DOI" value="10.17487/RFC2119"/><authorinitials="S." surname="Bradner" fullname="Scott Bradner"> <organization>Harvard University</organization> <address> <email>sob@harvard.edu</email> </address>initials="M" surname="Nottingham" fullname="Mark Nottingham" role="editor"> <organization/> </author><date month="March" year="1997"/> </front> </reference> <reference anchor="RFC8174"> <front> <title>Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words</title><authorfullname="B. Leiba" initials="B." surname="Leiba">initials="J" surname="Reschke" fullname="Julian Reschke" role="editor"> <organization/> </author> <datemonth="May" year="2017"/>year="2022" month="June"/> </front> <seriesInfoname="BCP" value="14"/> <seriesInfo name="RFC" value="8174"/> <seriesInfo name="DOI" value="10.17487/RFC8174"/> </reference> <reference anchor="RFC3986"> <front> <title abbrev="URI Generic Syntax">Uniform Resource Identifier (URI): Generic Syntax</title> <seriesInfoname="STD"value="66"/> <seriesInfo name="RFC" value="3986"/> <seriesInfo name="DOI" value="10.17487/RFC3986"/> <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee"/> <author initials="R." surname="Fielding" fullname="Roy T. Fielding"/> <author initials="L." surname="Masinter" fullname="Larry Masinter"/> <date year="2005" month="January"/> </front> </reference> <reference anchor="TLS12"> <front> <title>The Transport Layer Security (TLS) Protocol Version 1.2</title>value="97"/> <seriesInfo name="RFC"value="5246"/>value="9110"/> <seriesInfo name="DOI"value="10.17487/RFC5246"/> <author initials="T." surname="Dierks" fullname="Tim Dierks"/> <author initials="E." surname="Rescorla" fullname="Eric Rescorla"/> <date year="2008" month="August"/> </front> </reference> <reference anchor="TLS13"> <front> <title>The Transport Layer Security (TLS) Protocol Version 1.3</title> <seriesInfo name="RFC" value="8446"/> <seriesInfo name="DOI" value="10.17487/RFC8446"/> <author initials="E." surname="Rescorla" fullname="E. Rescorla"/> <date year="2018" month="August"/> </front> </reference> <reference anchor="RFC8470"> <front> <title>Using Early Data in HTTP</title> <seriesInfo name="RFC" value="8470"/> <seriesInfo name="DOI" value="10.17487/RFC8470"/> <author initials="M." surname="Thomson" fullname="M. Thomson"/> <author initials="M." surname="Nottingham" fullname="M. Nottingham"/> <author initials="W." surname="Tarreau" fullname="W. Tarreau"/> <date year="2018" month="September"/> </front> </reference> <reference anchor="TLS-EXT"> <front> <title> Transport Layer Security (TLS) Extensions: Extension Definitions </title> <seriesInfo name="RFC" value="6066"/> <seriesInfo name="DOI" value="10.17487/RFC6066"/> <author initials="D." surname="Eastlake 3rd" fullname="D. Eastlake 3rd"/> <date year="2011" month="January"/> </front> </reference> <reference anchor="TLS-ALPN"> <front> <title>Transport Layer Security (TLS) Application-Layer Protocol Negotiation Extension</title> <seriesInfo name="RFC" value="7301"/> <seriesInfo name="DOI" value="10.17487/RFC7301"/> <author initials="S." surname="Friedl" fullname="Stephan Friedl"/> <author initials="A." surname="Popov" fullname="Andrei Popov"/> <author initials="A." surname="Langley" fullname="Adam Langley"/> <author initials="E." surname="Stephan" fullname="Emile Stephan"/> <date month="July" year="2014"/> </front> </reference> <reference anchor="TLS-ECDHE"> <front> <title> TLS Elliptic Curve Cipher Suites with SHA-256/384 and AES Galois Counter Mode (GCM) </title> <seriesInfo name="RFC" value="5289"/> <seriesInfo name="DOI" value="10.17487/RFC5289"/> <author initials="E." surname="Rescorla" fullname="E. Rescorla"/> <date year="2008" month="August"/> </front> </reference> <reference anchor="RFC8422" target="https://www.rfc-editor.org/info/rfc8422"> <front> <title>Elliptic Curve Cryptography (ECC) Cipher Suites for Transport Layer Security (TLS) Versions 1.2 and Earlier</title> <seriesInfo name="RFC" value="8422"/> <seriesInfo name="DOI" value="10.17487/RFC8422"/> <author initials="Y." surname="Nir" fullname="Y. Nir"/> <author initials="S." surname="Josefsson" fullname="S. Josefsson"/> <author initials="M." surname="Pegourie-Gonnard" fullname="M. Pegourie-Gonnard"/> <date year="2018" month="August" /> </front> </reference> <reference anchor="COOKIE"> <front> <title>HTTP State Management Mechanism</title> <seriesInfo name="RFC" value="6265"/> <seriesInfo name="DOI" value="10.17487/RFC6265"/> <author initials="A." surname="Barth" fullname="A. Barth"/> <date year="2011" month="April"/> </front> </reference> <reference anchor="HTTP"> <front> <title>HTTP Semantics</title> <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-semantics-18"/> <author fullname="Roy T. Fielding" initials="R." surname="Fielding" role="editor"/> <author fullname="Mark Nottingham" initials="M." surname="Nottingham" role="editor"/> <author fullname="Julian Reschke" initials="J." surname="Reschke" role="editor"/> <date year="2021" month="August" day="18"/> </front>value="10.17487/RFC9110"/> </reference> <!-- draft-ietf-httpbis-cache (RFC 9111) --> <referenceanchor="CACHE">anchor="RFC9111" target="https://www.rfc-editor.org/info/rfc9111"> <front> <title>HTTP Caching</title><seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-cache-18"/><author initials="R" surname="Fielding" fullname="RoyT.Fielding"initials="R." surname="Fielding" role="editor"/>role="editor"> <organization/> </author> <author initials="M" surname="Nottingham" fullname="Mark Nottingham"initials="M." surname="Nottingham" role="editor"/> <author fullname="Julian Reschke" initials="J." surname="Reschke" role="editor"/> <date year="2021" month="August" day="18"/> </front> </reference> <reference anchor="QUIC" target="https://www.rfc-editor.org/info/rfc9000"> <front> <title>QUIC: A UDP-Based Multiplexed and Secure Transport</title> <author initials="J." surname="Iyengar" fullname="J. Iyengar"role="editor"> <organization/> </author> <authorinitials="M." surname="Thomson" fullname="M. Thomson"initials="J" surname="Reschke" fullname="Julian Reschke" role="editor"> <organization/> </author> <dateyear="2021" month="May"/>year="2022" month="June"/> </front> <seriesInfo name="STD" value="98"/> <seriesInfo name="RFC"value="9000"/>value="9111"/> <seriesInfo name="DOI"value="10.17487/RFC9000"/>value="10.17487/RFC9111"/> </reference> </references> <references> <name>Informative References</name> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.1122.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.3749.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.6125.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.6585.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.6973.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7323.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7540.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7838.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8441.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8499.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8740.xml"/> <!-- draft-ietf-httpbis-messaging (RFC 9112) --> <referenceanchor="RFC7540"> <front> <title>Hypertext Transfer Protocol Version 2 (HTTP/2)</title> <seriesInfo name="RFC" value="7540"/> <seriesInfo name="DOI" value="10.17487/RFC7540"/> <author initials="M." surname="Belshe" fullname="M. Belshe"/> <author initials="R." surname="Peon" fullname="R. Peon"/> <author initials="M." surname="Thomson" fullname="M. Thomson" role="editor"/> <date year="2015" month="May"/> </front> </reference> <reference anchor="RFC8740"> <front> <title>Using TLS 1.3 with HTTP/2</title> <seriesInfo name="RFC" value="8740"/> <seriesInfo name="DOI" value="10.17487/RFC8740"/> <author initials="D." surname="Benjamin" fullname="D. Benjamin"/> <date year="2020" month="February"/> </front> </reference> <reference anchor="HTTP11">anchor="RFC9112" target="https://www.rfc-editor.org/info/rfc9112"> <front> <title>HTTP/1.1</title><seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-messaging-18"/><author initials="R" surname="Fielding" fullname="RoyT.Fielding"initials="R." surname="Fielding" role="editor"/>role="editor"> <organization/> </author> <author initials="M" surname="Nottingham" fullname="Mark Nottingham"initials="M." surname="Nottingham" role="editor"/>role="editor"> <organization/> </author> <author initials="J" surname="Reschke" fullname="Julian Reschke"initials="J." surname="Reschke" role="editor"/> <date year="2021" month="August" day="18"/> </front> </reference> <reference anchor="RFC8441"> <front> <title>Bootstrapping WebSockets with HTTP/2</title> <seriesInfo name="RFC" value="8441"/> <seriesInfo name="DOI" value="10.17487/RFC8441"/> <author initials="P." surname="McManus" fullname="P. McManus"/> <date year="2018" month="September" /> </front> </reference> <reference anchor="RFC1122"> <front> <title>Requirements for Internet Hosts - Communication Layers</title> <seriesInfo name="RFC" value="1122"/> <seriesInfo name="DOI" value="10.17487/RFC1122"/> <author initials="R." surname="Braden" fullname="Robert T. Braden" role="editor"/> <date year="1989" month="October"/> </front> </reference> <reference anchor="RFC7323"> <front> <title>TCP Extensions for High Performance</title> <seriesInfo name="RFC" value="7323"/> <seriesInfo name="DOI" value="10.17487/RFC7323"/> <author initials="D." surname="Borman" fullname="Dave Borman"/> <author initials="B." surname="Braden" fullname="Bob Braden"/> <author initials="V." surname="Jacobson" fullname="Van Jacobson"/> <author initials="R." surname="Scheffenegger" fullname="Richard Scheffenegger" role="editor"/> <date year="2014" month="September"/> </front> </reference> <reference anchor="RFC6125" target="https://www.rfc-editor.org/info/rfc6125"> <front> <title>Representation and Verification of Domain-Based Application Service Identity within Internet Public Key Infrastructure Using X.509 (PKIX) Certificates in the Context of Transport Layer Security (TLS)</title> <seriesInfo name="RFC" value="6125"/> <seriesInfo name="DOI" value="10.17487/RFC6125"/> <author initials="P." surname="Saint-Andre" fullname="P. Saint-Andre"/> <author initials="J." surname="Hodges" fullname="J. Hodges"/> <date year="2011" month="March" /> </front> </reference> <reference anchor="RFC3749"> <front> <title>Transport Layer Security Protocol Compression Methods</title> <seriesInfo name="RFC" value="3749"/> <seriesInfo name="DOI" value="10.17487/RFC3749"/> <author initials="S." surname="Hollenbeck" fullname="S. Hollenbeck"/>role="editor"> <organization/> </author> <dateyear="2004" month="May"/>month="June" year="2022"/> </front></reference> <reference anchor="RFC6585"> <front> <title>Additional HTTP Status Codes</title> <seriesInfo name="RFC" value="6585"/><seriesInfoname="DOI" value="10.17487/RFC6585"/> <author initials="M." surname="Nottingham" fullname="Mark Nottingham"/> <author initials="R." surname="Fielding" fullname="Roy Fielding"/> <date year="2012" month="April"/> </front> </reference> <reference anchor="PRIVACY"> <front> <title>Privacy Considerations for Internet Protocols</title>name="STD" value="99"/> <seriesInfo name="RFC"value="6973"/>value="9112"/> <seriesInfo name="DOI"value="10.17487/RFC6973"/> <author initials="A." surname="Cooper" fullname="A. Cooper"/> <author initials="H." surname="Tschofenig" fullname="H. Tschofenig"/> <author initials="B." surname="Aboba" fullname="B. Aboba"/> <author initials="J." surname="Peterson" fullname="J. Peterson"/> <author initials="J." surname="Morris" fullname="J. Morris"/> <author initials="M." surname="Hansen" fullname="M. Hansen"/> <author initials="R." surname="Smith" fullname="R. Smith"/> <date year="2013" month="July"/> </front>value="10.17487/RFC9112"/> </reference> <reference anchor="TALKING" target="https://www.adambarth.com/papers/2011/huang-chen-barth-rescorla-jackson.pdf"> <front><title> Talking<title>Talking to Yourself for Fun andProfit </title>Profit</title> <author initials="L." surname="Huang"/> <author initials="E." surname="Chen"/> <author initials="A." surname="Barth"/> <author initials="E." surname="Rescorla"/> <author initials="C." surname="Jackson"/> <date year="2011"/> </front> </reference> <reference anchor="BREACH" target="https://breachattack.com/resources/BREACH%20-%20SSL,%20gone%20in%2030%20seconds.pdf"> <front><title> BREACH:<title>BREACH: Reviving the CRIMEAttack </title>Attack</title> <author initials="Y." surname="Gluck"/> <author initials="N." surname="Harris"/> <author initials="A." surname="Prado"/> <date year="2013" month="July" day="12"/> </front> </reference> <referenceanchor="ALT-SVC"> <front> <title> HTTP Alternative Services </title> <seriesInfo name="RFC" value="7838"/> <seriesInfo name="DOI" value="10.17487/RFC7838"/> <author initials="M." surname="Nottingham" fullname="Mark Nottingham"> <organization>Akamai</organization> </author> <author initials="P." surname="McManus" fullname="Patrick McManus"> <organization>Mozilla</organization> </author> <author initials="J." surname="Reschke" fullname="Julian Reschke"> <organization>greenbytes</organization> </author> <date year="2016" month="April"/> </front> </reference> <reference anchor="DNS-TERMS"> <front> <title>DNS Terminology</title> <seriesInfo name="BCP" value="219"/> <seriesInfo name="RFC" value="8499"/> <seriesInfo name="DOI" value="10.17487/RFC8499"/> <author initials="P." surname="Hoffman" fullname="P. Hoffman"/> <author initials="A." surname="Sullivan" fullname="A. Sullivan"/> <author initials="K." surname="Fujiwara" fullname="K. Fujiwara"/> <date year="2019" month="January"/> </front> </reference> <referenceanchor="NFLX-2019-002" target="https://github.com/Netflix/security-bulletins/blob/master/advisories/third-party/2019-002.md"> <front> <title>HTTP/2 Denial of Service Advisory</title> <author> <organization>Netflix</organization> </author> <date year="2019" month="August" day="13"/> </front> </reference><xi:include href="https://datatracker.ietf.org/doc/bibxml3/draft-ietf-httpbis-priority/xml"/><!-- draft-ietf-httpbis-priority (RFC 9218) --> <reference anchor="RFC9218" target="https://www.rfc-editor.org/info/rfc9218"> <front> <title>Extensible Prioritization Scheme for HTTP</title> <author initials="K" surname="Oku" asciiFullname="Kazuho Oku" fullname="奥 一穂"/> <author initials="L" surname="Pardue" fullname="Lucas Pardue"/> <date year="2022" month="June"/> </front> <seriesInfo name="RFC" value="9218"/> <seriesInfo name="DOI" value="10.17487/RFC9218"/> </reference> </references> </references> <section anchor="BadCipherSuites"> <name>Prohibited TLS 1.2 Cipher Suites</name><t> An<t>An HTTP/2 implementationMAY<bcp14>MAY</bcp14> treat the negotiation of any of the following cipher suites with TLS 1.2 as a <xref target="ConnectionErrorHandler">connection error</xref> of type <xref target="INADEQUATE_SECURITY"format="none">INADEQUATE_SECURITY</xref>: </t>format="none">INADEQUATE_SECURITY</xref>:</t> <ul spacing="compact"> <li>TLS_NULL_WITH_NULL_NULL</li> <li>TLS_RSA_WITH_NULL_MD5</li> <li>TLS_RSA_WITH_NULL_SHA</li> <li>TLS_RSA_EXPORT_WITH_RC4_40_MD5</li> <li>TLS_RSA_WITH_RC4_128_MD5</li> <li>TLS_RSA_WITH_RC4_128_SHA</li> <li>TLS_RSA_EXPORT_WITH_RC2_CBC_40_MD5</li> <li>TLS_RSA_WITH_IDEA_CBC_SHA</li> <li>TLS_RSA_EXPORT_WITH_DES40_CBC_SHA</li> <li>TLS_RSA_WITH_DES_CBC_SHA</li> <li>TLS_RSA_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_DH_DSS_EXPORT_WITH_DES40_CBC_SHA</li> <li>TLS_DH_DSS_WITH_DES_CBC_SHA</li> <li>TLS_DH_DSS_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_DH_RSA_EXPORT_WITH_DES40_CBC_SHA</li> <li>TLS_DH_RSA_WITH_DES_CBC_SHA</li> <li>TLS_DH_RSA_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_DHE_DSS_EXPORT_WITH_DES40_CBC_SHA</li> <li>TLS_DHE_DSS_WITH_DES_CBC_SHA</li> <li>TLS_DHE_DSS_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_DHE_RSA_EXPORT_WITH_DES40_CBC_SHA</li> <li>TLS_DHE_RSA_WITH_DES_CBC_SHA</li> <li>TLS_DHE_RSA_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_DH_anon_EXPORT_WITH_RC4_40_MD5</li> <li>TLS_DH_anon_WITH_RC4_128_MD5</li> <li>TLS_DH_anon_EXPORT_WITH_DES40_CBC_SHA</li> <li>TLS_DH_anon_WITH_DES_CBC_SHA</li> <li>TLS_DH_anon_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_KRB5_WITH_DES_CBC_SHA</li> <li>TLS_KRB5_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_KRB5_WITH_RC4_128_SHA</li> <li>TLS_KRB5_WITH_IDEA_CBC_SHA</li> <li>TLS_KRB5_WITH_DES_CBC_MD5</li> <li>TLS_KRB5_WITH_3DES_EDE_CBC_MD5</li> <li>TLS_KRB5_WITH_RC4_128_MD5</li> <li>TLS_KRB5_WITH_IDEA_CBC_MD5</li> <li>TLS_KRB5_EXPORT_WITH_DES_CBC_40_SHA</li> <li>TLS_KRB5_EXPORT_WITH_RC2_CBC_40_SHA</li> <li>TLS_KRB5_EXPORT_WITH_RC4_40_SHA</li> <li>TLS_KRB5_EXPORT_WITH_DES_CBC_40_MD5</li> <li>TLS_KRB5_EXPORT_WITH_RC2_CBC_40_MD5</li> <li>TLS_KRB5_EXPORT_WITH_RC4_40_MD5</li> <li>TLS_PSK_WITH_NULL_SHA</li> <li>TLS_DHE_PSK_WITH_NULL_SHA</li> <li>TLS_RSA_PSK_WITH_NULL_SHA</li> <li>TLS_RSA_WITH_AES_128_CBC_SHA</li> <li>TLS_DH_DSS_WITH_AES_128_CBC_SHA</li> <li>TLS_DH_RSA_WITH_AES_128_CBC_SHA</li> <li>TLS_DHE_DSS_WITH_AES_128_CBC_SHA</li> <li>TLS_DHE_RSA_WITH_AES_128_CBC_SHA</li> <li>TLS_DH_anon_WITH_AES_128_CBC_SHA</li> <li>TLS_RSA_WITH_AES_256_CBC_SHA</li> <li>TLS_DH_DSS_WITH_AES_256_CBC_SHA</li> <li>TLS_DH_RSA_WITH_AES_256_CBC_SHA</li> <li>TLS_DHE_DSS_WITH_AES_256_CBC_SHA</li> <li>TLS_DHE_RSA_WITH_AES_256_CBC_SHA</li> <li>TLS_DH_anon_WITH_AES_256_CBC_SHA</li> <li>TLS_RSA_WITH_NULL_SHA256</li> <li>TLS_RSA_WITH_AES_128_CBC_SHA256</li> <li>TLS_RSA_WITH_AES_256_CBC_SHA256</li> <li>TLS_DH_DSS_WITH_AES_128_CBC_SHA256</li> <li>TLS_DH_RSA_WITH_AES_128_CBC_SHA256</li> <li>TLS_DHE_DSS_WITH_AES_128_CBC_SHA256</li> <li>TLS_RSA_WITH_CAMELLIA_128_CBC_SHA</li> <li>TLS_DH_DSS_WITH_CAMELLIA_128_CBC_SHA</li> <li>TLS_DH_RSA_WITH_CAMELLIA_128_CBC_SHA</li> <li>TLS_DHE_DSS_WITH_CAMELLIA_128_CBC_SHA</li> <li>TLS_DHE_RSA_WITH_CAMELLIA_128_CBC_SHA</li> <li>TLS_DH_anon_WITH_CAMELLIA_128_CBC_SHA</li> <li>TLS_DHE_RSA_WITH_AES_128_CBC_SHA256</li> <li>TLS_DH_DSS_WITH_AES_256_CBC_SHA256</li> <li>TLS_DH_RSA_WITH_AES_256_CBC_SHA256</li> <li>TLS_DHE_DSS_WITH_AES_256_CBC_SHA256</li> <li>TLS_DHE_RSA_WITH_AES_256_CBC_SHA256</li> <li>TLS_DH_anon_WITH_AES_128_CBC_SHA256</li> <li>TLS_DH_anon_WITH_AES_256_CBC_SHA256</li> <li>TLS_RSA_WITH_CAMELLIA_256_CBC_SHA</li> <li>TLS_DH_DSS_WITH_CAMELLIA_256_CBC_SHA</li> <li>TLS_DH_RSA_WITH_CAMELLIA_256_CBC_SHA</li> <li>TLS_DHE_DSS_WITH_CAMELLIA_256_CBC_SHA</li> <li>TLS_DHE_RSA_WITH_CAMELLIA_256_CBC_SHA</li> <li>TLS_DH_anon_WITH_CAMELLIA_256_CBC_SHA</li> <li>TLS_PSK_WITH_RC4_128_SHA</li> <li>TLS_PSK_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_PSK_WITH_AES_128_CBC_SHA</li> <li>TLS_PSK_WITH_AES_256_CBC_SHA</li> <li>TLS_DHE_PSK_WITH_RC4_128_SHA</li> <li>TLS_DHE_PSK_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_DHE_PSK_WITH_AES_128_CBC_SHA</li> <li>TLS_DHE_PSK_WITH_AES_256_CBC_SHA</li> <li>TLS_RSA_PSK_WITH_RC4_128_SHA</li> <li>TLS_RSA_PSK_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_RSA_PSK_WITH_AES_128_CBC_SHA</li> <li>TLS_RSA_PSK_WITH_AES_256_CBC_SHA</li> <li>TLS_RSA_WITH_SEED_CBC_SHA</li> <li>TLS_DH_DSS_WITH_SEED_CBC_SHA</li> <li>TLS_DH_RSA_WITH_SEED_CBC_SHA</li> <li>TLS_DHE_DSS_WITH_SEED_CBC_SHA</li> <li>TLS_DHE_RSA_WITH_SEED_CBC_SHA</li> <li>TLS_DH_anon_WITH_SEED_CBC_SHA</li> <li>TLS_RSA_WITH_AES_128_GCM_SHA256</li> <li>TLS_RSA_WITH_AES_256_GCM_SHA384</li> <li>TLS_DH_RSA_WITH_AES_128_GCM_SHA256</li> <li>TLS_DH_RSA_WITH_AES_256_GCM_SHA384</li> <li>TLS_DH_DSS_WITH_AES_128_GCM_SHA256</li> <li>TLS_DH_DSS_WITH_AES_256_GCM_SHA384</li> <li>TLS_DH_anon_WITH_AES_128_GCM_SHA256</li> <li>TLS_DH_anon_WITH_AES_256_GCM_SHA384</li> <li>TLS_PSK_WITH_AES_128_GCM_SHA256</li> <li>TLS_PSK_WITH_AES_256_GCM_SHA384</li> <li>TLS_RSA_PSK_WITH_AES_128_GCM_SHA256</li> <li>TLS_RSA_PSK_WITH_AES_256_GCM_SHA384</li> <li>TLS_PSK_WITH_AES_128_CBC_SHA256</li> <li>TLS_PSK_WITH_AES_256_CBC_SHA384</li> <li>TLS_PSK_WITH_NULL_SHA256</li> <li>TLS_PSK_WITH_NULL_SHA384</li> <li>TLS_DHE_PSK_WITH_AES_128_CBC_SHA256</li> <li>TLS_DHE_PSK_WITH_AES_256_CBC_SHA384</li> <li>TLS_DHE_PSK_WITH_NULL_SHA256</li> <li>TLS_DHE_PSK_WITH_NULL_SHA384</li> <li>TLS_RSA_PSK_WITH_AES_128_CBC_SHA256</li> <li>TLS_RSA_PSK_WITH_AES_256_CBC_SHA384</li> <li>TLS_RSA_PSK_WITH_NULL_SHA256</li> <li>TLS_RSA_PSK_WITH_NULL_SHA384</li> <li>TLS_RSA_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_DH_DSS_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_DH_RSA_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_DHE_DSS_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_DHE_RSA_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_DH_anon_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_RSA_WITH_CAMELLIA_256_CBC_SHA256</li> <li>TLS_DH_DSS_WITH_CAMELLIA_256_CBC_SHA256</li> <li>TLS_DH_RSA_WITH_CAMELLIA_256_CBC_SHA256</li> <li>TLS_DHE_DSS_WITH_CAMELLIA_256_CBC_SHA256</li> <li>TLS_DHE_RSA_WITH_CAMELLIA_256_CBC_SHA256</li> <li>TLS_DH_anon_WITH_CAMELLIA_256_CBC_SHA256</li> <li>TLS_EMPTY_RENEGOTIATION_INFO_SCSV</li> <li>TLS_ECDH_ECDSA_WITH_NULL_SHA</li> <li>TLS_ECDH_ECDSA_WITH_RC4_128_SHA</li> <li>TLS_ECDH_ECDSA_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_ECDH_ECDSA_WITH_AES_128_CBC_SHA</li> <li>TLS_ECDH_ECDSA_WITH_AES_256_CBC_SHA</li> <li>TLS_ECDHE_ECDSA_WITH_NULL_SHA</li> <li>TLS_ECDHE_ECDSA_WITH_RC4_128_SHA</li> <li>TLS_ECDHE_ECDSA_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA</li> <li>TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA</li> <li>TLS_ECDH_RSA_WITH_NULL_SHA</li> <li>TLS_ECDH_RSA_WITH_RC4_128_SHA</li> <li>TLS_ECDH_RSA_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_ECDH_RSA_WITH_AES_128_CBC_SHA</li> <li>TLS_ECDH_RSA_WITH_AES_256_CBC_SHA</li> <li>TLS_ECDHE_RSA_WITH_NULL_SHA</li> <li>TLS_ECDHE_RSA_WITH_RC4_128_SHA</li> <li>TLS_ECDHE_RSA_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA</li> <li>TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA</li> <li>TLS_ECDH_anon_WITH_NULL_SHA</li> <li>TLS_ECDH_anon_WITH_RC4_128_SHA</li> <li>TLS_ECDH_anon_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_ECDH_anon_WITH_AES_128_CBC_SHA</li> <li>TLS_ECDH_anon_WITH_AES_256_CBC_SHA</li> <li>TLS_SRP_SHA_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_SRP_SHA_RSA_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_SRP_SHA_DSS_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_SRP_SHA_WITH_AES_128_CBC_SHA</li> <li>TLS_SRP_SHA_RSA_WITH_AES_128_CBC_SHA</li> <li>TLS_SRP_SHA_DSS_WITH_AES_128_CBC_SHA</li> <li>TLS_SRP_SHA_WITH_AES_256_CBC_SHA</li> <li>TLS_SRP_SHA_RSA_WITH_AES_256_CBC_SHA</li> <li>TLS_SRP_SHA_DSS_WITH_AES_256_CBC_SHA</li> <li>TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256</li> <li>TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA384</li> <li>TLS_ECDH_ECDSA_WITH_AES_128_CBC_SHA256</li> <li>TLS_ECDH_ECDSA_WITH_AES_256_CBC_SHA384</li> <li>TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256</li> <li>TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA384</li> <li>TLS_ECDH_RSA_WITH_AES_128_CBC_SHA256</li> <li>TLS_ECDH_RSA_WITH_AES_256_CBC_SHA384</li> <li>TLS_ECDH_ECDSA_WITH_AES_128_GCM_SHA256</li> <li>TLS_ECDH_ECDSA_WITH_AES_256_GCM_SHA384</li> <li>TLS_ECDH_RSA_WITH_AES_128_GCM_SHA256</li> <li>TLS_ECDH_RSA_WITH_AES_256_GCM_SHA384</li> <li>TLS_ECDHE_PSK_WITH_RC4_128_SHA</li> <li>TLS_ECDHE_PSK_WITH_3DES_EDE_CBC_SHA</li> <li>TLS_ECDHE_PSK_WITH_AES_128_CBC_SHA</li> <li>TLS_ECDHE_PSK_WITH_AES_256_CBC_SHA</li> <li>TLS_ECDHE_PSK_WITH_AES_128_CBC_SHA256</li> <li>TLS_ECDHE_PSK_WITH_AES_256_CBC_SHA384</li> <li>TLS_ECDHE_PSK_WITH_NULL_SHA</li> <li>TLS_ECDHE_PSK_WITH_NULL_SHA256</li> <li>TLS_ECDHE_PSK_WITH_NULL_SHA384</li> <li>TLS_RSA_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_RSA_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_DH_DSS_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_DH_DSS_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_DH_RSA_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_DH_RSA_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_DHE_DSS_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_DHE_DSS_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_DHE_RSA_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_DHE_RSA_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_DH_anon_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_DH_anon_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_ECDHE_ECDSA_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_ECDHE_ECDSA_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_ECDH_ECDSA_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_ECDH_ECDSA_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_ECDHE_RSA_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_ECDHE_RSA_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_ECDH_RSA_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_ECDH_RSA_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_RSA_WITH_ARIA_128_GCM_SHA256</li> <li>TLS_RSA_WITH_ARIA_256_GCM_SHA384</li> <li>TLS_DH_RSA_WITH_ARIA_128_GCM_SHA256</li> <li>TLS_DH_RSA_WITH_ARIA_256_GCM_SHA384</li> <li>TLS_DH_DSS_WITH_ARIA_128_GCM_SHA256</li> <li>TLS_DH_DSS_WITH_ARIA_256_GCM_SHA384</li> <li>TLS_DH_anon_WITH_ARIA_128_GCM_SHA256</li> <li>TLS_DH_anon_WITH_ARIA_256_GCM_SHA384</li> <li>TLS_ECDH_ECDSA_WITH_ARIA_128_GCM_SHA256</li> <li>TLS_ECDH_ECDSA_WITH_ARIA_256_GCM_SHA384</li> <li>TLS_ECDH_RSA_WITH_ARIA_128_GCM_SHA256</li> <li>TLS_ECDH_RSA_WITH_ARIA_256_GCM_SHA384</li> <li>TLS_PSK_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_PSK_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_DHE_PSK_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_DHE_PSK_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_RSA_PSK_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_RSA_PSK_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_PSK_WITH_ARIA_128_GCM_SHA256</li> <li>TLS_PSK_WITH_ARIA_256_GCM_SHA384</li> <li>TLS_RSA_PSK_WITH_ARIA_128_GCM_SHA256</li> <li>TLS_RSA_PSK_WITH_ARIA_256_GCM_SHA384</li> <li>TLS_ECDHE_PSK_WITH_ARIA_128_CBC_SHA256</li> <li>TLS_ECDHE_PSK_WITH_ARIA_256_CBC_SHA384</li> <li>TLS_ECDHE_ECDSA_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_ECDHE_ECDSA_WITH_CAMELLIA_256_CBC_SHA384</li> <li>TLS_ECDH_ECDSA_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_ECDH_ECDSA_WITH_CAMELLIA_256_CBC_SHA384</li> <li>TLS_ECDHE_RSA_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_ECDHE_RSA_WITH_CAMELLIA_256_CBC_SHA384</li> <li>TLS_ECDH_RSA_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_ECDH_RSA_WITH_CAMELLIA_256_CBC_SHA384</li> <li>TLS_RSA_WITH_CAMELLIA_128_GCM_SHA256</li> <li>TLS_RSA_WITH_CAMELLIA_256_GCM_SHA384</li> <li>TLS_DH_RSA_WITH_CAMELLIA_128_GCM_SHA256</li> <li>TLS_DH_RSA_WITH_CAMELLIA_256_GCM_SHA384</li> <li>TLS_DH_DSS_WITH_CAMELLIA_128_GCM_SHA256</li> <li>TLS_DH_DSS_WITH_CAMELLIA_256_GCM_SHA384</li> <li>TLS_DH_anon_WITH_CAMELLIA_128_GCM_SHA256</li> <li>TLS_DH_anon_WITH_CAMELLIA_256_GCM_SHA384</li> <li>TLS_ECDH_ECDSA_WITH_CAMELLIA_128_GCM_SHA256</li> <li>TLS_ECDH_ECDSA_WITH_CAMELLIA_256_GCM_SHA384</li> <li>TLS_ECDH_RSA_WITH_CAMELLIA_128_GCM_SHA256</li> <li>TLS_ECDH_RSA_WITH_CAMELLIA_256_GCM_SHA384</li> <li>TLS_PSK_WITH_CAMELLIA_128_GCM_SHA256</li> <li>TLS_PSK_WITH_CAMELLIA_256_GCM_SHA384</li> <li>TLS_RSA_PSK_WITH_CAMELLIA_128_GCM_SHA256</li> <li>TLS_RSA_PSK_WITH_CAMELLIA_256_GCM_SHA384</li> <li>TLS_PSK_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_PSK_WITH_CAMELLIA_256_CBC_SHA384</li> <li>TLS_DHE_PSK_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_DHE_PSK_WITH_CAMELLIA_256_CBC_SHA384</li> <li>TLS_RSA_PSK_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_RSA_PSK_WITH_CAMELLIA_256_CBC_SHA384</li> <li>TLS_ECDHE_PSK_WITH_CAMELLIA_128_CBC_SHA256</li> <li>TLS_ECDHE_PSK_WITH_CAMELLIA_256_CBC_SHA384</li> <li>TLS_RSA_WITH_AES_128_CCM</li> <li>TLS_RSA_WITH_AES_256_CCM</li> <li>TLS_RSA_WITH_AES_128_CCM_8</li> <li>TLS_RSA_WITH_AES_256_CCM_8</li> <li>TLS_PSK_WITH_AES_128_CCM</li> <li>TLS_PSK_WITH_AES_256_CCM</li> <li>TLS_PSK_WITH_AES_128_CCM_8</li> <li>TLS_PSK_WITH_AES_256_CCM_8</li> </ul><aside><t>Note:<aside> <t>Note: This list was assembled from the set of registered TLS cipher suites when <xref target="RFC7540"/> was developed. This list includes those cipher suites that do not offer an ephemeral key exchange and those that are based on the TLS null, stream, or block cipher type (as defined in <xreftarget="TLS12"target="RFC5246" section="6.2.3"/>). Additional cipher suites with these properties could be defined; these would not be explicitlyprohibited.</t></aside> <t> Forprohibited.</t> </aside> <t>For more details, see <xreftarget="tls12ciphers"/> </t>target="tls12ciphers"/>.</t> </section> <section anchor="revision-updates"> <name>Changes from RFC 7540</name><t> This<t>This revision includes the following substantivechanges: </t>changes:</t> <ul spacing="normal"><li> Use<li>Use of TLS 1.3 was defined based onRFC 8740,<xref target="RFC8740"/>, which this documentobsoletes. </li> <li> Theobsoletes.</li> <li>The priority scheme defined in RFC 7540 is deprecated. Definitions for the format of the <xref target="PRIORITY" format="none">PRIORITY</xref> frame and the priority fields in the <xref target="HEADERS" format="none">HEADERS</xref> frame have been retained, plus the rules governing when <xref target="PRIORITY" format="none">PRIORITY</xref> frames can be sent and received, but the semantics of these fields are only described in RFC 7540. The priority signaling scheme from RFC 7540 was not successful. Using the simpler signaling in <xreftarget="I-D.ietf-httpbis-priority">successor signaling</xref>target="RFC9218"/> isrecommended. </li> <li> Therecommended.</li> <li>The HTTP/1.1 Upgrade mechanism is deprecated and no longer specified in this document. It was never widely deployed, with plaintext HTTP/2 users choosing to use the prior-knowledge implementationinstead. </li> <li> Validationinstead.</li> <li>Validation for field names and values has been narrowed. The validation that is mandatory for intermediaries is preciselydefineddefined, and error reporting for requests has been amended to encourage sending 400-series statuscodes. </li> <li> Thecodes.</li> <li>The ranges of codepoints for settings and frame types that were reserved for"Experimental Use"Experimental Use are now available for generaluse. </li> <li> Connection-specificuse.</li> <li>Connection-specific header fields--- which are prohibited--- are more precisely and comprehensivelyidentified. </li> <li> <tt>Host</tt>identified.</li> <li><tt>Host</tt> and<tt>:authority</tt>"<tt>:authority</tt>" are no longer permitted todisagree. </li> <li> Rulesdisagree.</li> <li>Rules for sending Dynamic Table Size Update instructions after changes in settings have been clarified in <xreftarget="dynamic-table"/>. </li>target="dynamic-table"/>.</li> </ul><t> Editorial<t>Editorial changes are also included. In particular, changes to terminology and document structure are in response to updates to <xreftarget="HTTP">coretarget="RFC9110">core HTTP semantics</xref>. Those documents now include some concepts that were first defined in RFC 7540, such as the 421 status code or connectioncoalescing. </t> </section> <section numbered="false"> <name>Contributors</name> <t> The previous version of this document was authored by Mike Belshe and Roberto Peon. </t>coalescing.</t> </section> <section numbered="false"> <name>Acknowledgments</name><t> Credit<t>Credit for non-trivial input to this document is owed to a large number of people who have contributed to the HTTPworking groupWorking Group over the years. <xref target="RFC7540"/> contains a more extensive list of people that deserve acknowledgment for their contributions.</t> </section> <section numbered="false"> <name>Contributors</name> <t><contact fullname="Mike Belshe"/> and <contact fullname="Roberto Peon"/> authored the text that this document is based on.</t> </section> </back> </rfc>