Apache : Security Vulnerabilities, CVEs, Published In 2007

The login method in LoginModule implementations in Apache Geronimo 2.0 does not throw FailedLoginException for failed logins, which allows remote attackers to bypass authentication requirements, deploy arbitrary modules, and gain administrative access by sending a blank username and password with the command line deployer in the deployment module.

Integer overflow in the TIFF parser in OpenOffice.org (OOo) before 2.3; and Sun StarOffice 6, 7, and 8 Office Suite (StarSuite); allows remote attackers to execute arbitrary code via a TIFF file with crafted values of unspecified length fields, which triggers allocation of an incorrect amount of memory, resulting in a heap-based buffer overflow.

The AJP connector in Apache Tomcat 5.5.15 uses an incorrect length for chunks, which can cause a buffer over-read in the ajp_process_callback in mod_jk, which allows remote attackers to read portions of sensitive memory.

The Apache HTTP Server, when accessed through a TCP connection with a large window size, allows remote attackers to cause a denial of service (network bandwidth consumption) via a Range header that specifies multiple copies of the same fragment. NOTE: the severity of this issue has been disputed by third parties, who state that the large window size required by the attack is not normally supported or configured by the server, or that a DDoS-style attack would accomplish the same goal

Stack-based buffer overflow in the map_uri_to_worker function (native/common/jk_uri_worker_map.c) in mod_jk.so for Apache Tomcat JK Web Server Connector 1.2.19 and 1.2.20, as used in Tomcat 4.1.34 and 5.5.20, allows remote attackers to execute arbitrary code via a long URL that triggers the overflow in a URI worker map routine.

SQLLoginModule in Apache Geronimo 2.0 through 2.1 does not throw an exception for a nonexistent username, which allows remote attackers to bypass authentication via a login attempt with any username not contained in the database.

The default catalina.policy in the JULI logging component in Apache Tomcat 5.5.9 through 5.5.25 and 6.0.0 through 6.0.15 does not restrict certain permissions for web applications, which allows attackers to modify logging configuration options and overwrite arbitrary files, as demonstrated by changing the (1) level, (2) directory, and (3) prefix attributes in the org.apache.juli.FileHandler handler.

Multiple race conditions in suexec in Apache HTTP Server (httpd) 2.2.3 between directory and file validation, and their usage, allow local users to gain privileges and execute arbitrary code by renaming directories or performing symlink attacks. NOTE: the researcher, who is reliable, claims that the vendor disputes the issue because "the attacks described rely on an insecure server configuration" in which the user "has write access to the document root."

Directory traversal vulnerability in Apache HTTP Server and Tomcat 5.x before 5.5.22 and 6.x before 6.0.10, when using certain proxy modules (mod_proxy, mod_rewrite, mod_jk), allows remote attackers to read arbitrary files via a .. (dot dot) sequence with combinations of (1) "/" (slash), (2) "\" (backslash), and (3) URL-encoded backslash (%5C) characters in the URL, which are valid separators in Tomcat but not in Apache.

PerlRun.pm in Apache mod_perl before 1.30, and RegistryCooker.pm in mod_perl 2.x, does not properly escape PATH_INFO before use in a regular expression, which allows remote attackers to cause a denial of service (resource consumption) via a crafted URI.

mod_jk in Apache Tomcat JK Web Server Connector 1.2.x before 1.2.23 decodes request URLs within the Apache HTTP Server before passing the URL to Tomcat, which allows remote attackers to access protected pages via a crafted prefix JkMount, possibly involving double-encoded .. (dot dot) sequences and directory traversal, a related issue to CVE-2007-0450.

The recall_headers function in mod_mem_cache in Apache 2.2.4 does not properly copy all levels of header data, which can cause Apache to return HTTP headers containing previously used data, which could be used by remote attackers to obtain potentially sensitive information.

cache_util.c in the mod_cache module in Apache HTTP Server (httpd), when caching is enabled and a threaded Multi-Processing Module (MPM) is used, allows remote attackers to cause a denial of service (child processing handler crash) via a request with the (1) s-maxage, (2) max-age, (3) min-fresh, or (4) max-stale Cache-Control headers without a value.

Apache Axis 1.0 allows remote attackers to obtain sensitive information by requesting a non-existent WSDL file, which reveals the installation path in the resulting exception message.

The date handling code in modules/proxy/proxy_util.c (mod_proxy) in Apache 2.3.0, when using a threaded MPM, allows remote origin servers to cause a denial of service (caching forward proxy process crash) via crafted date headers that trigger a buffer over-read.

Unspecified vulnerability in the management EJB (MEJB) in Apache Geronimo before 2.0.2 allows remote attackers to bypass authentication and obtain "access to Geronimo internals" via unspecified vectors.

Apache httpd 2.0.59 and 2.2.4, with the Prefork MPM module, allows local users to cause a denial of service via certain code sequences executed in a worker process that (1) stop request processing by killing all worker processes and preventing creation of replacements or (2) hang the system by forcing the master process to fork an arbitrarily large number of worker processes. NOTE: This might be an inherent design limitation of Apache with respect to worker processes in hosted environments.

Apache httpd 1.3.37, 2.0.59, and 2.2.4 with the Prefork MPM module, allows local users to cause a denial of service by modifying the worker_score and process_score arrays to reference an arbitrary process ID, which is sent a SIGUSR1 signal from the master process, aka "SIGUSR1 killer."

suexec in Apache HTTP Server (httpd) 2.2.3 does not verify combinations of user and group IDs on the command line, which might allow local users to leverage other vulnerabilities to create arbitrary UID/GID owned files if /proc is mounted. NOTE: the researcher, who is reliable, claims that the vendor disputes the issue because "the attacks described rely on an insecure server configuration" in which the user "has write access to the document root." In addition, because this is dependent on other vulnerabilities, perhaps this is resultant and should not be included in CVE.

Cross-site scripting (XSS) vulnerability in mod_status.c in the mod_status module in Apache HTTP Server (httpd), when ExtendedStatus is enabled and a public server-status page is used, allows remote attackers to inject arbitrary web script or HTML via unspecified vectors involving charsets with browsers that perform "charset detection" when the content-type is not specified.

Cross-site scripting (XSS) vulnerability in implicit-objects.jsp in Apache Tomcat 5.0.0 through 5.0.30 and 5.5.0 through 5.5.17 allows remote attackers to inject arbitrary web script or HTML via certain header values.

Cross-site scripting (XSS) vulnerability in the calendar application example in Apache Tomcat 4.0.0 through 4.0.6, 4.1.0 through 4.1.31, 5.0.0 through 5.0.30, and 5.5.0 through 5.5.15 allows remote attackers to inject arbitrary web script or HTML via the time parameter to cal2.jsp and possibly unspecified other vectors. NOTE: this may be related to CVE-2006-0254.1.

Apache SpamAssassin before 3.1.8 allows remote attackers to cause a denial of service via long URLs in malformed HTML, which triggers "massive memory usage."

Multiple cross-site scripting (XSS) vulnerabilities in the appdev/sample/web/hello.jsp example application in Tomcat 4.0.0 through 4.0.6, 4.1.0 through 4.1.36, 5.0.0 through 5.0.30, 5.5.0 through 5.5.23, and 6.0.0 through 6.0.10 allow remote attackers to inject arbitrary web script or HTML via the test parameter and unspecified vectors.

Multiple cross-site scripting (XSS) vulnerabilities in certain JSP files in the examples web application in Apache Tomcat 4.0.0 through 4.0.6, 4.1.0 through 4.1.36, 5.0.0 through 5.0.30, 5.5.0 through 5.5.24, and 6.0.0 through 6.0.13 allow remote attackers to inject arbitrary web script or HTML via the portion of the URI after the ';' character, as demonstrated by a URI containing a "snp/snoop.jsp;" sequence.