From: Stefan Eissing Date: Mon, 11 Dec 2023 10:52:26 +0000 (+0100) Subject: CONNECTION-FILTERS: update documentation X-Git-Tag: curl-8_6_0~232 X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=02d81c5a6808e189cd20418274abccbd9bb1385d;p=thirdparty%2Fcurl.git CONNECTION-FILTERS: update documentation Closes #12497 --- diff --git a/docs/CONNECTION-FILTERS.md b/docs/CONNECTION-FILTERS.md index c4b11dbb01..d973bc79b1 100644 --- a/docs/CONNECTION-FILTERS.md +++ b/docs/CONNECTION-FILTERS.md @@ -1,7 +1,7 @@ # curl connection filters Connection filters is a design in the internals of curl, not visible in its public API. They were added -in curl v7.xx.x. This document describes the concepts, its high level implementation and the motivations. +in curl v7.87.0. This document describes the concepts, its high level implementation and the motivations. ## Filters @@ -110,18 +110,113 @@ The `recv` implementation is equivalent. ## Filter Types -The (currently) existing filter types are: SOCKET, SOCKET-ACCEPT, SSL, HTTP-PROXY and SOCKS-PROXY. Vital to establishing and read/writing a connection. But filters are also a good way to implement tasks for *managing* a connection: +The currently existing filter types (curl 8.5.0) are: -* **Statistics**: a filter that counts the number of bytes sent/received. Place one in front of SOCKET and one higher up and get the number of raw and "easy" bytes transferred. They may track the speed as well, or number of partial writes, etc. -* **Timeout**: enforce timeouts, e.g. fail if a connection cannot be established in a certain amount of time. -* **Progress**: report progress on a connection. -* **Pacing**: limit read/write rates. -* **Testing**: simulate network condition or failures. +* `TCP`, `UDP`, `UNIX`: filters that operate on a socket, providing raw I/O. +* `SOCKET-ACCEPT`: special TCP socket that has a socket that has been `accept()`ed in a `listen()` +* `SSL`: filter that applies TLS en-/decryption and handshake. Manages the underlying TLS backend implementation. +* `HTTP-PROXY`, `H1-PROXY`, `H2-PROXY`: the first manages the connection to a HTTP proxy server and uses the other depending on which ALPN protocol has been negotiated. +* `SOCKS-PROXY`: filter for the various SOCKS proxy protocol variations +* `HAPROXY`: filter for the protocol of the same name, providing client IP information to a server. +* `HTTP/2`: filter for handling multiplexed transfers over a HTTP/2 connection +* `HTTP/3`: filter for handling multiplexed transfers over a HTTP/3+QUIC connection +* `HAPPY-EYEBALLS`: meta filter that implements IPv4/IPv6 "happy eyeballing". It creates up to 2 sub-filters that race each other for a connection. +* `SETUP`: meta filter that manages the creation of sub-filter chains for a specific transport (e.g. TCP or QUIC). +* `HTTPS-CONNECT`: meta filter that races a TCP+TLS and a QUIC connection against each other to determine if HTTP/1.1, HTTP/2 or HTTP/3 shall be used for a transfer. -As you see, filters are a good way to add functionality to curl's internal handling of transfers without impact on other code. +Meta filters are combining other filters for a specific purpose, mostly during connection establishment. Other filters like `TCP`, `UDP` and `UNIX` are only to be found at the end of filter chains. SSL filters provide encryption, of course. Protocol filters change the bytes sent and received. -## Easy Filters? +## Filter Flags -Some things that curl needs to manage are not directly tied to a specific connection but the property of the `Curl_easy` handle, e.g. a particular transfer. When using HTTP/2 or HTTP/3, many transfers can use the same connection. If one wants to monitor of the transfer itself or restricting its speed alone, a connection filter is not the right place to do this. +Filter types carry flags that inform what they do. These are (for now): + +* `CF_TYPE_IP_CONNECT`: this filter type talks directly to a server. This does not have to be the server the transfer wants to talk to. For example when a proxy server is used. +* `CF_TYPE_SSL`: this filter type provides encryption. +* `CF_TYPE_MULTIPLEX`: this filter type can manage multiple transfers in parallel. + +Filter types can combine these flags. For example, the HTTP/3 filter types have `CF_TYPE_IP_CONNECT`, `CF_TYPE_SSL` and `CF_TYPE_MULTIPLEX` set. + +Flags are useful to extrapolate properties of a connection. To check if a connection is encrypted, libcurl inspect the filter chain in place, top down, for `CF_TYPE_SSL`. If it finds `CF_TYPE_IP_CONNECT` before any `CF_TYPE_SSL`, the connection is not encrypted. + +For example, `conn1` is for a `http:` request using a tunnel through a HTTP/2 `https:` proxy. `conn2` is a `https:` HTTP/2 connection to the same proxy. `conn3` uses HTTP/3 without proxy. The filter chains would look like this (simplified): + +``` +conn1 --> `HTTP-PROXY` --> `H2-PROXY` --> `SSL` --> `TCP` +flags: `IP_CONNECT` `SSL` `IP_CONNECT` + +conn2 --> `HTTP/2` --> `SSL` --> `HTTP-PROXY` --> `H2-PROXY` --> `SSL` --> `TCP` +flags: `SSL` `IP_CONNECT` `SSL` `IP_CONNECT` + +conn3 --> `HTTP/3` +flags: `SSL|IP_CONNECT` +``` + +Inspecting the filter chains, `conn1` is seen as unencrypted, since it contains an `IP_CONNECT` filter before any `SSL`. `conn2` is clearly encrypted as an `SSL` flagged filter is seen first. `conn3` is also encrypted as the `SSL` flag is checked before the presence of `IP_CONNECT`. + +Similar checks can determine if a connection is multiplexed or not. + +## Filter Tracing + +Filters may make use of special trace macros like `CURL_TRC_CF(data, cf, msg, ...)`. With `data` being the transfer and `cf` being the filter instance. These traces are normally not active and their execution is guarded so that they are cheap to ignore. + +Users of `curl` may activate them by adding the name of the filter type to the `--trace-config` argument. For example, in order to get more detailed tracing of a HTTP/2 request, invoke curl with: + +``` +> curl -v --trace-config ids,time,http/2 https://curl.se +``` +Which will give you trace output with time information, transfer+connection ids and details from the `HTTP/2` filter. Filter type names in the trace config are case insensitive. You may use `all` to enable tracing for all filter types. When using `libcurl` you may call `curl_global_trace(config_string)` at the start of your application to enable filter details. + +## Meta Filters + +Meta filters is a catch-all name for filter types that do not change the transfer data in any way but provide other important services to curl. In general, it is possible to do all sorts of silly things with them. One of the commonly used, important things is "eyeballing". + +The `HAPPY-EYEBALLS` filter is involved in the connect phase. It's job is to try the various IPv4 and IPv6 addresses that are known for a server. If only one address family is known (or configured), it tries the addresses one after the other with timeouts calculated from the amount of addresses and the overall connect timeout. + +When more than one address family is to be tried, it splits the address list into IPv4 and IPv6 and makes parallel attempts. The connection filter chain will look like this: + +``` +* create connection for http://curl.se +conn[curl.se] --> SETUP[TCP] --> HAPPY-EYEBALLS --> NULL +* start connect +conn[curl.se] --> SETUP[TCP] --> HAPPY-EYEBALLS --> NULL + - ballerv4 --> TCP[151.101.1.91]:443 + - ballerv6 --> TCP[2a04:4e42:c00::347]:443 +* v6 answers, connected +conn[curl.se] --> SETUP[TCP] --> HAPPY-EYEBALLS --> TCP[2a04:4e42:c00::347]:443 +* transfer +``` + +The modular design of connection filters and that we can plug them into each other is used to control the parallel attempts. When a `TCP` filter does not connect (in time), it is torn down and another one is created for the next address. This keeps the `TCP` filter simple. + +The `HAPPY-EYEBALLS` on the other hand stays focused on its side of the problem. We can use it also to make other type of connection by just giving it another filter type to try and have happy eyeballing for QUIC: + +``` +* create connection for --http3-only https://curl.se +conn[curl.se] --> SETUP[QUIC] --> HAPPY-EYEBALLS --> NULL +* start connect +conn[curl.se] --> SETUP[QUIC] --> HAPPY-EYEBALLS --> NULL + - ballerv4 --> HTTP/3[151.101.1.91]:443 + - ballerv6 --> HTTP/3[2a04:4e42:c00::347]:443 +* v6 answers, connected +conn[curl.se] --> SETUP[QUIC] --> HAPPY-EYEBALLS --> HTTP/3[2a04:4e42:c00::347]:443 +* transfer +``` + +And when we plug these two variants together, we get the `HTTPS-CONNECT` filter type that is used for `--http3` when **both** HTTP/3 and HTTP/2 or HTTP/1.1 shall be attempted: + +``` +* create connection for --http3 https://curl.se +conn[curl.se] --> HTTPS-CONNECT --> NULL +* start connect +conn[curl.se] --> HTTPS-CONNECT --> NULL + - SETUP[QUIC] --> HAPPY-EYEBALLS --> NULL + - ballerv4 --> HTTP/3[151.101.1.91]:443 + - ballerv6 --> HTTP/3[2a04:4e42:c00::347]:443 + - SETUP[TCP] --> HAPPY-EYEBALLS --> NULL + - ballerv4 --> TCP[151.101.1.91]:443 + - ballerv6 --> TCP[2a04:4e42:c00::347]:443 +* v4 QUIC answers, connected +conn[curl.se] --> HTTPS-CONNECT --> SETUP[QUIC] --> HAPPY-EYEBALLS --> HTTP/3[151.101.1.91]:443 +* transfer +``` -So we might add "easy filters" one day. Who knows?