HTTP vulnerabilities

An issue was discovered in OpenResty before 1.15.8.4. ngx_http_lua_subrequest.c allows HTTP request smuggling, as demonstrated by the ngx.location.capture API.

In Apache HTTP Server 2.4.0 to 2.4.41, mod_proxy_ftp may use uninitialized memory when proxying to a malicious FTP server.

When using the Apache JServ Protocol (AJP), care must be taken when trusting incoming connections to Apache Tomcat. Tomcat treats AJP connections as having higher trust than, for example, a similar HTTP connection. If such connections are available to an attacker, they can be exploited in ways that may be surprising. In Apache Tomcat 9.0.0.M1 to 9.0.0.30, 8.5.0 to 8.5.50 and 7.0.0 to 7.0.99, Tomcat shipped with an AJP Connector enabled by default that listened on all configured IP addresses. It was expected (and recommended in the security guide) that this Connector would be disabled if not required. This vulnerability report identified a mechanism that allowed: - returning arbitrary files from anywhere in the web application - processing any file in the web application as a JSP Further, if the web application allowed file upload and stored those files within the web application (or the attacker was able to control the content of the web application by some other means) then this, along with the ability to process a file as a JSP, made remote code execution possible. It is important to note that mitigation is only required if an AJP port is accessible to untrusted users. Users wishing to take a defence-in-depth approach and block the vector that permits returning arbitrary files and execution as JSP may upgrade to Apache Tomcat 9.0.31, 8.5.51 or 7.0.100 or later. A number of changes were made to the default AJP Connector configuration in 9.0.31 to harden the default configuration. It is likely that users upgrading to 9.0.31, 8.5.51 or 7.0.100 or later will need to make small changes to their configurations.

Zimbra Collaboration Suite (ZCS) before 8.8.15 Patch 7 allows SSRF when WebEx zimlet is installed and zimlet JSP is enabled.

A vulnerability was found in Hibernate-Validator. The SafeHtml validator annotation fails to properly sanitize payloads consisting of potentially malicious code in HTML comments and instructions. This vulnerability can result in an XSS attack.

Heap buffer overflow in the TFTP protocol handler in cURL 7.19.4 to 7.65.3.

A non-privileged user or program can put code and a config file in a known non-privileged path (under C:/usr/local/) that will make curl <= 7.65.1 automatically run the code (as an openssl "engine") on invocation. If that curl is invoked by a privileged user it can do anything it wants.

In libexpat in Expat before 2.2.7, XML input including XML names that contain a large number of colons could make the XML parser consume a high amount of RAM and CPU resources while processing (enough to be usable for denial-of-service attacks).

In Apache HTTP Server 2.4 releases 2.4.17 to 2.4.38, with MPM event, worker or prefork, code executing in less-privileged child processes or threads (including scripts executed by an in-process scripting interpreter) could execute arbitrary code with the privileges of the parent process (usually root) by manipulating the scoreboard. Non-Unix systems are not affected.

Akka HTTP versions <= 10.0.5 Illegal Media Range in Accept Header Causes StackOverflowError Leading to Denial of Service

Apache httpd allows remote attackers to read secret data from process memory if the Limit directive can be set in a user's .htaccess file, or if httpd.conf has certain misconfigurations, aka Optionsbleed. This affects the Apache HTTP Server through 2.2.34 and 2.4.x through 2.4.27. The attacker sends an unauthenticated OPTIONS HTTP request when attempting to read secret data. This is a use-after-free issue and thus secret data is not always sent, and the specific data depends on many factors including configuration. Exploitation with .htaccess can be blocked with a patch to the ap_limit_section function in server/core.c.

Apache HTTP Server, in all releases prior to 2.2.32 and 2.4.25, was liberal in the whitespace accepted from requests and sent in response lines and headers. Accepting these different behaviors represented a security concern when httpd participates in any chain of proxies or interacts with back-end application servers, either through mod_proxy or using conventional CGI mechanisms, and may result in request smuggling, response splitting and cache pollution.

In Apache HTTP Server versions 2.4.0 to 2.4.23, malicious input to mod_auth_digest can cause the server to crash, and each instance continues to crash even for subsequently valid requests.

In Apache HTTP Server versions 2.4.0 to 2.4.23, mod_session_crypto was encrypting its data/cookie using the configured ciphers with possibly either CBC or ECB modes of operation (AES256-CBC by default), hence no selectable or builtin authenticated encryption. This made it vulnerable to padding oracle attacks, particularly with CBC.

A maliciously constructed HTTP/2 request could cause mod_http2 in Apache HTTP Server 2.4.24, 2.4.25 to dereference a NULL pointer and crash the server process.

When under stress, closing many connections, the HTTP/2 handling code in Apache httpd 2.4.26 would sometimes access memory after it has been freed, resulting in potentially erratic behaviour.

In Apache httpd before 2.2.34 and 2.4.x before 2.4.27, the value placeholder in [Proxy-]Authorization headers of type 'Digest' was not initialized or reset before or between successive key=value assignments by mod_auth_digest. Providing an initial key with no '=' assignment could reflect the stale value of uninitialized pool memory used by the prior request, leading to leakage of potentially confidential information, and a segfault in other cases resulting in denial of service.

Nginx versions since 0.5.6 up to and including 1.13.2 are vulnerable to integer overflow vulnerability in nginx range filter module resulting into leak of potentially sensitive information triggered by specially crafted request.

In Apache httpd 2.2.x before 2.2.33 and 2.4.x before 2.4.26, mod_mime can read one byte past the end of a buffer when sending a malicious Content-Type response header.

The HTTP strict parsing changes added in Apache httpd 2.2.32 and 2.4.24 introduced a bug in token list parsing, which allows ap_find_token() to search past the end of its input string. By maliciously crafting a sequence of request headers, an attacker may be able to cause a segmentation fault, or to force ap_find_token() to return an incorrect value.

In Apache httpd 2.2.x before 2.2.33 and 2.4.x before 2.4.26, mod_ssl may dereference a NULL pointer when third-party modules call ap_hook_process_connection() during an HTTP request to an HTTPS port.

In Apache httpd 2.2.x before 2.2.33 and 2.4.x before 2.4.26, use of the ap_get_basic_auth_pw() by third-party modules outside of the authentication phase may lead to authentication requirements being bypassed.

The ASN1_TFLG_COMBINE implementation in crypto/asn1/tasn_dec.c in OpenSSL before 0.9.8zh, 1.0.0 before 1.0.0t, 1.0.1 before 1.0.1q, and 1.0.2 before 1.0.2e mishandles errors caused by malformed X509_ATTRIBUTE data, which allows remote attackers to obtain sensitive information from process memory by triggering a decoding failure in a PKCS#7 or CMS application.

Stack-based buffer overflow in the Administration Server in IBM HTTP Server 6.1.0.x through 6.1.0.47, 7.0.0.x before 7.0.0.39, 8.0.0.x before 8.0.0.12, and 8.5.x before 8.5.5.7, as used in WebSphere Application Server and other products, allows remote authenticated users to execute arbitrary code via unspecified vectors.

The ap_some_auth_required function in server/request.c in the Apache HTTP Server 2.4.x before 2.4.14 does not consider that a Require directive may be associated with an authorization setting rather than an authentication setting, which allows remote attackers to bypass intended access restrictions in opportunistic circumstances by leveraging the presence of a module that relies on the 2.2 API behavior.

The chunked transfer coding implementation in the Apache HTTP Server before 2.4.14 does not properly parse chunk headers, which allows remote attackers to conduct HTTP request smuggling attacks via a crafted request, related to mishandling of large chunk-size values and invalid chunk-extension characters in modules/http/http_filters.c.

The read_request_line function in server/protocol.c in the Apache HTTP Server 2.4.12 does not initialize the protocol structure member, which allows remote attackers to cause a denial of service (NULL pointer dereference and process crash) by sending a request that lacks a method to an installation that enables the INCLUDES filter and has an ErrorDocument 400 directive specifying a local URI.

The RC4 algorithm, as used in the TLS protocol and SSL protocol, does not properly combine state data with key data during the initialization phase, which makes it easier for remote attackers to conduct plaintext-recovery attacks against the initial bytes of a stream by sniffing network traffic that occasionally relies on keys affected by the Invariance Weakness, and then using a brute-force approach involving LSB values, aka the "Bar Mitzvah" issue.

The lua_websocket_read function in lua_request.c in the mod_lua module in the Apache HTTP Server through 2.4.12 allows remote attackers to cause a denial of service (child-process crash) by sending a crafted WebSocket Ping frame after a Lua script has called the wsupgrade function.

mod_lua.c in the mod_lua module in the Apache HTTP Server 2.3.x and 2.4.x through 2.4.10 does not support an httpd configuration in which the same Lua authorization provider is used with different arguments within different contexts, which allows remote attackers to bypass intended access restrictions in opportunistic circumstances by leveraging multiple Require directives, as demonstrated by a configuration that specifies authorization for one group to access a certain directory, and authorization for a second group to access a second directory.

The STARTTLS implementation in mail/ngx_mail_smtp_handler.c in the SMTP proxy in nginx 1.5.x and 1.6.x before 1.6.1 and 1.7.x before 1.7.4 does not properly restrict I/O buffering, which allows man-in-the-middle attackers to insert commands into encrypted SMTP sessions by sending a cleartext command that is processed after TLS is in place, related to a "plaintext command injection" attack, a similar issue to CVE-2011-0411.

The handle_headers function in mod_proxy_fcgi.c in the mod_proxy_fcgi module in the Apache HTTP Server 2.4.10 allows remote FastCGI servers to cause a denial of service (buffer over-read and daemon crash) via long response headers.

nginx 0.5.6 through 1.7.4, when using the same shared ssl_session_cache or ssl_session_ticket_key for multiple servers, can reuse a cached SSL session for an unrelated context, which allows remote attackers with certain privileges to conduct "virtual host confusion" attacks.

The cache_merge_headers_out function in modules/cache/cache_util.c in the mod_cache module in the Apache HTTP Server before 2.4.11 allows remote attackers to cause a denial of service (NULL pointer dereference and application crash) via an empty HTTP Content-Type header.

Memory leak in the winnt_accept function in server/mpm/winnt/child.c in the WinNT MPM in the Apache HTTP Server 2.4.x before 2.4.10 on Windows, when the default AcceptFilter is enabled, allows remote attackers to cause a denial of service (memory consumption) via crafted requests.

The mod_cgid module in the Apache HTTP Server before 2.4.10 does not have a timeout mechanism, which allows remote attackers to cause a denial of service (process hang) via a request to a CGI script that does not read from its stdin file descriptor.

Race condition in the mod_status module in the Apache HTTP Server before 2.4.10 allows remote attackers to cause a denial of service (heap-based buffer overflow), or possibly obtain sensitive credential information or execute arbitrary code, via a crafted request that triggers improper scoreboard handling within the status_handler function in modules/generators/mod_status.c and the lua_ap_scoreboard_worker function in modules/lua/lua_request.c.

The deflate_in_filter function in mod_deflate.c in the mod_deflate module in the Apache HTTP Server before 2.4.10, when request body decompression is enabled, allows remote attackers to cause a denial of service (resource consumption) via crafted request data that decompresses to a much larger size.

The mod_proxy module in the Apache HTTP Server 2.4.x before 2.4.10, when a reverse proxy is enabled, allows remote attackers to cause a denial of service (child-process crash) via a crafted HTTP Connection header.

The cache_invalidate function in modules/cache/cache_storage.c in the mod_cache module in the Apache HTTP Server 2.4.6, when a caching forward proxy is enabled, allows remote HTTP servers to cause a denial of service (NULL pointer dereference and daemon crash) via vectors that trigger a missing hostname value.

The SPDY implementation in the ngx_http_spdy_module module in nginx 1.5.10 before 1.5.11, when running on a 32-bit platform, allows remote attackers to execute arbitrary code via a crafted request.

The mod_headers module in the Apache HTTP Server 2.2.22 allows remote attackers to bypass "RequestHeader unset" directives by placing a header in the trailer portion of data sent with chunked transfer coding. NOTE: the vendor states "this is not a security issue in httpd as such."

Heap-based buffer overflow in the SPDY implementation in nginx 1.3.15 before 1.4.7 and 1.5.x before 1.5.12 allows remote attackers to execute arbitrary code via a crafted request.

The log_cookie function in mod_log_config.c in the mod_log_config module in the Apache HTTP Server before 2.4.8 allows remote attackers to cause a denial of service (segmentation fault and daemon crash) via a crafted cookie that is not properly handled during truncation.

The dav_xml_get_cdata function in main/util.c in the mod_dav module in the Apache HTTP Server before 2.4.8 does not properly remove whitespace characters from CDATA sections, which allows remote attackers to cause a denial of service (daemon crash) via a crafted DAV WRITE request.

The RC4 algorithm, as used in the TLS protocol and SSL protocol, has many single-byte biases, which makes it easier for remote attackers to conduct plaintext-recovery attacks via statistical analysis of ciphertext in a large number of sessions that use the same plaintext.

The ap_read_request function in server/protocol.c in the Apache HTTP Server 2.2.x before 2.2.15, when a multithreaded MPM is used, does not properly handle headers in subrequests in certain circumstances involving a parent request that has a body, which might allow remote attackers to obtain sensitive information via a crafted request that triggers access to memory locations associated with an earlier request.

The ap_proxy_ajp_request function in mod_proxy_ajp.c in mod_proxy_ajp in the Apache HTTP Server 2.2.x before 2.2.15 does not properly handle certain situations in which a client sends no request body, which allows remote attackers to cause a denial of service (backend server outage) via a crafted request, related to use of a 500 error code instead of the appropriate 400 error code.

The Apache HTTP Server 2.0.44, when DNS resolution is enabled for client IP addresses, allows remote attackers to inject arbitrary text into log files via an HTTP request in conjunction with a crafted DNS response, as demonstrated by injecting XSS sequences, related to an "Inverse Lookup Log Corruption (ILLC)" issue.

The Apache HTTP Server 2.0.44, when DNS resolution is enabled for client IP addresses, uses a logging format that does not identify whether a dotted quad represents an unresolved IP address, which allows remote attackers to spoof IP addresses via crafted DNS responses containing numerical top-level domains, as demonstrated by a forged 123.123.123.123 domain name, related to an "Inverse Lookup Log Corruption (ILLC)" issue.