Waitress is a Web Server Gateway Interface server for Python 2 and 3. When using Waitress versions 2.1.0 and prior behind a proxy that does not properly validate the incoming HTTP request matches the RFC7230 standard, Waitress and the frontend proxy may disagree on where one request starts and where it ends. This would allow requests to be smuggled via the front-end proxy to waitress and later behavior. There are two classes of vulnerability that may lead to request smuggling that are addressed by this advisory: The use of Python's `int()` to parse strings into integers, leading to `+10` to be parsed as `10`, or `0x01` to be parsed as `1`, where as the standard specifies that the string should contain only digits or hex digits; and Waitress does not support chunk extensions, however it was discarding them without validating that they did not contain illegal characters. This vulnerability has been patched in Waitress 2.1.1. A workaround is available. When deploying a proxy in front of waitress, turning on any and all functionality to make sure that the request matches the RFC7230 standard. Certain proxy servers may not have this functionality though and users are encouraged to upgrade to the latest version of waitress instead.

Puma is a simple, fast, multi-threaded, parallel HTTP 1.1 server for Ruby/Rack applications. When using Puma behind a proxy that does not properly validate that the incoming HTTP request matches the RFC7230 standard, Puma and the frontend proxy may disagree on where a request starts and ends. This would allow requests to be smuggled via the front-end proxy to Puma. The vulnerability has been fixed in 5.6.4 and 4.3.12. Users are advised to upgrade as soon as possible. Workaround: when deploying a proxy in front of Puma, turning on any and all functionality to make sure that the request matches the RFC7230 standard.

Apache HTTP Server 2.4.52 and earlier fails to close inbound connection when errors are encountered discarding the request body, exposing the server to HTTP Request Smuggling

Netty is an open-source, asynchronous event-driven network application framework for rapid development of maintainable high performance protocol servers & clients. In Netty (io.netty:netty-codec-http2) before version 4.1.60.Final there is a vulnerability that enables request smuggling. If a Content-Length header is present in the original HTTP/2 request, the field is not validated by `Http2MultiplexHandler` as it is propagated up. This is fine as long as the request is not proxied through as HTTP/1.1. If the request comes in as an HTTP/2 stream, gets converted into the HTTP/1.1 domain objects (`HttpRequest`, `HttpContent`, etc.) via `Http2StreamFrameToHttpObjectCodec `and then sent up to the child channel's pipeline and proxied through a remote peer as HTTP/1.1 this may result in request smuggling. In a proxy case, users may assume the content-length is validated somehow, which is not the case. If the request is forwarded to a backend channel that is a HTTP/1.1 connection, the Content-Length now has meaning and needs to be checked. An attacker can smuggle requests inside the body as it gets downgraded from HTTP/2 to HTTP/1.1. For an example attack refer to the linked GitHub Advisory. Users are only affected if all of this is true: `HTTP2MultiplexCodec` or `Http2FrameCodec` is used, `Http2StreamFrameToHttpObjectCodec` is used to convert to HTTP/1.1 objects, and these HTTP/1.1 objects are forwarded to another remote peer. This has been patched in 4.1.60.Final As a workaround, the user can do the validation by themselves by implementing a custom `ChannelInboundHandler` that is put in the `ChannelPipeline` behind `Http2StreamFrameToHttpObjectCodec`.

An issue was discovered in Squid through 4.13 and 5.x through 5.0.4. Due to improper input validation, it allows a trusted client to perform HTTP Request Smuggling and access services otherwise forbidden by the security controls. This occurs for certain uri_whitespace configuration settings.

In Eclipse Jetty, versions 9.2.x and older, 9.3.x (all configurations), and 9.4.x (non-default configuration with RFC2616 compliance enabled), HTTP/0.9 is handled poorly. An HTTP/1 style request line (i.e. method space URI space version) that declares a version of HTTP/0.9 was accepted and treated as a 0.9 request. If deployed behind an intermediary that also accepted and passed through the 0.9 version (but did not act on it), then the response sent could be interpreted by the intermediary as HTTP/1 headers. This could be used to poison the cache if the server allowed the origin client to generate arbitrary content in the response.

In Eclipse Jetty, versions 9.2.x and older, 9.3.x (all configurations), and 9.4.x (non-default configuration with RFC2616 compliance enabled), transfer-encoding chunks are handled poorly. The chunk length parsing was vulnerable to an integer overflow. Thus a large chunk size could be interpreted as a smaller chunk size and content sent as chunk body could be interpreted as a pipelined request. If Jetty was deployed behind an intermediary that imposed some authorization and that intermediary allowed arbitrarily large chunks to be passed on unchanged, then this flaw could be used to bypass the authorization imposed by the intermediary as the fake pipelined request would not be interpreted by the intermediary as a request.

In Apache Tomcat 9.0.0.M1 to 9.0.30, 8.5.0 to 8.5.50 and 7.0.0 to 7.0.99 the HTTP header parsing code used an approach to end-of-line parsing that allowed some invalid HTTP headers to be parsed as valid. This led to a possibility of HTTP Request Smuggling if Tomcat was located behind a reverse proxy that incorrectly handled the invalid Transfer-Encoding header in a particular manner. Such a reverse proxy is considered unlikely.

Apache HTTP Server versions 2.4.20 to 2.4.43. A specially crafted value for the 'Cache-Digest' header in a HTTP/2 request would result in a crash when the server actually tries to HTTP/2 PUSH a resource afterwards. Configuring the HTTP/2 feature via "H2Push off" will mitigate this vulnerability for unpatched servers.

There is a vulnerability in Apache Traffic Server 6.0.0 to 6.2.3, 7.0.0 to 7.1.8, and 8.0.0 to 8.0.5 with a smuggling attack and Transfer-Encoding and Content length headers. Upgrade to versions 7.1.9 and 8.0.6 or later versions.

An issue was discovered in Squid before 4.13 and 5.x before 5.0.4. Due to incorrect data validation, HTTP Request Smuggling attacks may succeed against HTTP and HTTPS traffic. This leads to cache poisoning. This allows any client, including browser scripts, to bypass local security and poison the proxy cache and any downstream caches with content from an arbitrary source. When configured for relaxed header parsing (the default), Squid relays headers containing whitespace characters to upstream servers. When this occurs as a prefix to a Content-Length header, the frame length specified will be ignored by Squid (allowing for a conflicting length to be used from another Content-Length header) but relayed upstream.

Apache HTTP Server versions 2.4.20 to 2.4.43 When trace/debug was enabled for the HTTP/2 module and on certain traffic edge patterns, logging statements were made on the wrong connection, causing concurrent use of memory pools. Configuring the LogLevel of mod_http2 above "info" will mitigate this vulnerability for unpatched servers.

In Puma (RubyGem) before 4.3.5 and 3.12.6, a client could smuggle a request through a proxy, causing the proxy to send a response back to another unknown client. If the proxy uses persistent connections and the client adds another request in via HTTP pipelining, the proxy may mistake it as the first request's body. Puma, however, would see it as two requests, and when processing the second request, send back a response that the proxy does not expect. If the proxy has reused the persistent connection to Puma to send another request for a different client, the second response from the first client will be sent to the second client. This is a similar but different vulnerability from CVE-2020-11076. The problem has been fixed in Puma 3.12.6 and Puma 4.3.5.

In Twisted Web through 19.10.0, there was an HTTP request splitting vulnerability. When presented with a content-length and a chunked encoding header, the content-length took precedence and the remainder of the request body was interpreted as a pipelined request.

Netty is an open-source, asynchronous event-driven network application framework for rapid development of maintainable high performance protocol servers & clients. In Netty (io.netty:netty-codec-http2) before version 4.1.61.Final there is a vulnerability that enables request smuggling. The content-length header is not correctly validated if the request only uses a single Http2HeaderFrame with the endStream set to to true. This could lead to request smuggling if the request is proxied to a remote peer and translated to HTTP/1.1. This is a followup of GHSA-wm47-8v5p-wjpj/CVE-2021-21295 which did miss to fix this one case. This was fixed as part of 4.1.61.Final.

Netty 4.1.43.Final allows HTTP Request Smuggling because it mishandles Transfer-Encoding whitespace (such as a [space]Transfer-Encoding:chunked line) and a later Content-Length header. This issue exists because of an incomplete fix for CVE-2019-16869.

Symantec Advanced Secure Gateway (ASG) and ProxySG are susceptible to an HTTP desync vulnerability. When a remote unauthenticated attacker and other web clients communicate through the proxy with the same web server, the attacker can send crafted HTTP requests and cause the proxy to forward web server responses to unintended clients. Severity/CVSSv3: High / 8.1 AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N

In JetBrains Ktor before 1.4.1, HTTP request smuggling was possible.

An error in the evaluation of the fetch metadata headers could allow a bypass of the CSRF protection in Apache Wicket. This issue affects Apache Wicket: from 9.1.0 through 9.16.0, and the milestone releases for the 10.0 series. Apache Wicket 8.x does not support CSRF protection via the fetch metadata headers and as such is not affected. Users are recommended to upgrade to version 9.17.0 or 10.0.0, which fixes the issue.

Puma is a Ruby/Rack web server built for parallelism. Prior to versions 6.3.1 and 5.6.7, puma exhibited incorrect behavior when parsing chunked transfer encoding bodies and zero-length Content-Length headers in a way that allowed HTTP request smuggling. Severity of this issue is highly dependent on the nature of the web site using puma is. This could be caused by either incorrect parsing of trailing fields in chunked transfer encoding bodies or by parsing of blank/zero-length Content-Length headers. Both issues have been addressed and this vulnerability has been fixed in versions 6.3.1 and 5.6.7. Users are advised to upgrade. There are no known workarounds for this vulnerability.

A vulnerability in the web-based interface of Cisco Unified Communications Manager and Cisco Unified Communications Manager Session Management Edition (SME) could allow an unauthenticated, remote attacker to bypass security restrictions. The vulnerability is due to improper handling of malformed HTTP methods. An attacker could exploit this vulnerability by sending a crafted HTTP request to the affected system. A successful exploit could allow the attacker to gain unauthorized access to the system.

This affects all versions before 10.1.14 and from 10.2.0 to 10.2.4 of package com.typesafe.akka:akka-http-core. It allows multiple Transfer-Encoding headers.

An issue was discovered in the tiny_http crate through 2020-06-16 for Rust. HTTP Request smuggling can occur via a malformed Transfer-Encoding header.

Acceptance of some invalid Transfer-Encoding headers in the HTTP/1 client in net/http before Go 1.17.12 and Go 1.18.4 allows HTTP request smuggling if combined with an intermediate server that also improperly fails to reject the header as invalid.

Pound before 2.8 allows HTTP request smuggling, a related issue to CVE-2016-10711.

aiohttp is an asynchronous HTTP client/server framework for asyncio and Python. Affected versions of aiohttp have a security vulnerability regarding the inconsistent interpretation of the http protocol. HTTP/1.1 is a persistent protocol, if both Content-Length(CL) and Transfer-Encoding(TE) header values are present it can lead to incorrect interpretation of two entities that parse the HTTP and we can poison other sockets with this incorrect interpretation. A possible Proof-of-Concept (POC) would be a configuration with a reverse proxy(frontend) that accepts both CL and TE headers and aiohttp as backend. As aiohttp parses anything with chunked, we can pass a chunked123 as TE, the frontend entity will ignore this header and will parse Content-Length. The impact of this vulnerability is that it is possible to bypass any proxy rule, poisoning sockets to other users like passing Authentication Headers, also if it is present an Open Redirect an attacker could combine it to redirect random users to another website and log the request. This vulnerability has been addressed in release 3.8.0 of aiohttp. Users are advised to upgrade. There are no known workarounds for this vulnerability.

A flaw was found in Undertow in versions before 2.1.1.Final, regarding the processing of invalid HTTP requests with large chunk sizes. This flaw allows an attacker to take advantage of HTTP request smuggling.

HTTP request/response smuggling vulnerability in HAProxy version 2.7.0, and 2.6.1 to 2.6.7 allows a remote attacker to alter a legitimate user's request. As a result, the attacker may obtain sensitive information or cause a denial-of-service (DoS) condition.