- Arch-security - lists.archlinux.org

[arch-security] [ASA-201611-17] libgit2: denial of service
by Levente Polyak 16 Nov '16

16 Nov '16

Arch Linux Security Advisory ASA-201611-17 ========================================== Severity: Low Date : 2016-11-16 CVE-ID : CVE-2016-8568 CVE-2016-8569 Package : libgit2 Type : denial of service Remote : Yes Link : https://wiki.archlinux.org/index.php/CVE Summary ======= The package libgit2 before version 1:0.24.3-1 is vulnerable to denial of service. Resolution ========== Upgrade to 1:0.24.3-1. # pacman -Syu "libgit2>=1:0.24.3-1" The problems have been fixed upstream in version 0.24.3. Workaround ========== None. Description =========== - CVE-2016-8568 (denial of service) A heap-based read out-of-bounds access has been discovered while parsing a malformed object file. - CVE-2016-8569 (denial of service) A null pointer dereference has been discovered while showing a malformed object file. Impact ====== A remote attacker is able to create specially crafted object files that lead to an application crash resulting in denial of service. References ========== http://seclists.org/oss-sec/2016/q4/64 https://github.com/libgit2/libgit2/issues/3936 https://github.com/libgit2/libgit2/issues/3937 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-8568 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-8569

1 0

[arch-security] [ASA-201611-16] firefox: multiple issues
by Levente Polyak 16 Nov '16

16 Nov '16

Arch Linux Security Advisory ASA-201611-16 ========================================== Severity: Critical Date : 2016-11-16 CVE-ID : CVE-2016-5289 CVE-2016-5290 CVE-2016-5291 CVE-2016-5292 CVE-2016-5296 CVE-2016-5297 CVE-2016-9063 CVE-2016-9064 CVE-2016-9066 CVE-2016-9067 CVE-2016-9068 CVE-2016-9070 CVE-2016-9071 CVE-2016-9073 CVE-2016-9075 CVE-2016-9076 CVE-2016-9077 Package : firefox Type : multiple issues Remote : Yes Link : https://wiki.archlinux.org/index.php/CVE Summary ======= The package firefox before version 50.0-1 is vulnerable to multiple issues including arbitrary code execution, information disclosure, insufficient validation, privilege escalation, content spoofing, same- origin policy bypass and sandbox escape. Resolution ========== Upgrade to 50.0-1. # pacman -Syu "firefox>=50.0-1" The problems have been fixed upstream in version 50.0. Workaround ========== None. Description =========== - CVE-2016-5289 (arbitrary code execution) Mozilla developers and community members Christian Holler, Andrew McCreight, Dan Minor, Tyson Smith, Jon Coppeard, Jan-Ivar Bruaroey, Jesse Ruderman, and Markus Stange reported memory safety bugs present in Firefox 49. Some of these bugs showed evidence of memory corruption and we presume that with enough effort that some of these could be exploited to run arbitrary code. - CVE-2016-5290 (arbitrary code execution) Mozilla developers and community members Olli Pettay, Christian Holler, Ehsan Akhgari, Jon Coppeard, Gary Kwong, Tooru Fujisawa, Philipp, and Randell Jesup reported memory safety bugs present in Firefox 49 and Firefox ESR 45.4. Some of these bugs showed evidence of memory corruption and we presume that with enough effort that some of these could be exploited to run arbitrary code. - CVE-2016-5291 (same-origin policy bypass) A same-origin policy bypass with local shortcut files to load arbitrary local content from disk. - CVE-2016-5292 (arbitrary code execution) During URL parsing, a maliciously crafted URL can cause a potentially exploitable crash. - CVE-2016-5296 (arbitrary code execution) A heap-buffer-overflow in Cairo when processing SVG content caused by compiler optimization, resulting in a potentially exploitable crash. - CVE-2016-5297 (arbitrary code execution) An error in argument length checking in JavaScript, leading to potential integer overflows or other bounds checking issues. - CVE-2016-9063 (arbitrary code execution) An integer overflow during the parsing of XML using the Expat library. - CVE-2016-9064 (insufficient validation) Add-on updates failed to verify that the add-on ID inside the signed package matched the ID of the add-on being updated. An attacker who could perform a man-in-the-middle attack on the user's connection to the update server and defeat the certificate pinning protection could provide a malicious signed add-on instead of a valid update. - CVE-2016-9066 (arbitrary code execution) A buffer overflow resulting in a potentially exploitable crash due to memory allocation issues when handling large amounts of incoming data. - CVE-2016-9067 (arbitrary code execution) Two heap-use-after-free errors during DOM operations in nsINode::ReplaceOrInsertBefore resulting in potentially exploitable crashes. - CVE-2016-9068 (arbitrary code execution) A heap-use-after-free in nsRefreshDriver during web animations when working with timelines resulting in a potentially exploitable crash. - CVE-2016-9070 (same-origin policy bypass) A maliciously crafted page loaded to the sidebar through a bookmark can reference a privileged chrome window and engage in limited JavaScript operations violating cross-origin protections. - CVE-2016-9071 (information disclosure) Content Security Policy combined with HTTP to HTTPS redirection can be used by malicious server to verify whether a known site is within a user's browser history. - CVE-2016-9073 (sandbox escape) WebExtensions can bypass security checks to load privileged URLs and potentially escape the WebExtension sandbox. - CVE-2016-9075 (privilege escalation) An issue where WebExtensions can use the mozAddonManager API to elevate privilege due to privileged pages being allowed in the permissions list. This allows a malicious extension to then install additional extensions without explicit user permission. - CVE-2016-9076 (content spoofing) An issue where a <select> dropdown menu can be used to cover location bar content, resulting in potential spoofing attacks. This attack requires e10s to be enabled in order to function. - CVE-2016-9077 (information disclosure) Canvas allows the use of the feDisplacementMap filter on images loaded cross-origin. The rendering by the filter is variable depending on the input pixel, allowing for timing attacks when the images are loaded from third party locations. Impact ====== A remote attacker is able to execute arbitrary code, disclose sensitive information, escape the WebExtension sandbox, spoof content and bypass the same-origin policy via various vectors. References ========== https://www.mozilla.org/en-US/security/advisories/mfsa2016-89 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-5289 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-5290 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-5291 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-5292 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-5296 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-5297 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-9063 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-9064 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-9066 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-9067 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-9068 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-9070 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-9071 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-9073 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-9075 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-9076 https://www.mozilla.org/en-US/security/advisories/mfsa2016-89/#CVE-2016-9077 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-5289 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-5290 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-5291 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-5292 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-5296 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-5297 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-9063 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-9064 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-9066 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-9067 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-9068 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-9070 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-9071 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-9073 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-9075 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-9076 https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-9077

1 0

[arch-security] [ASA-201611-15] python-django: multiple issues
by Levente Polyak 16 Nov '16

16 Nov '16

Arch Linux Security Advisory ASA-201611-15 ========================================== Severity: High Date : 2016-11-16 CVE-ID : CVE-2016-9013 CVE-2016-9014 Package : python-django Type : multiple issues Remote : Yes Link : https://wiki.archlinux.org/index.php/CVE Summary ======= The package python-django before version 1.10.3-1 is vulnerable to multiple issues including access restriction bypass and authentication bypass. Resolution ========== Upgrade to 1.10.3-1. # pacman -Syu "python-django>=1.10.3-1" The problems have been fixed upstream in version 1.10.3. Workaround ========== - CVE-2016-9013 Ensure the debug mode is deactivated via: settings.DEBUG=False Description =========== - CVE-2016-9013 (authentication bypass) When running tests with an Oracle database, Django creates a temporary database user. In older versions, if a password isn't manually specified in the database settings TEST dictionary, a hardcoded password is used. This could allow an attacker with network access to the database server to connect. This user is usually dropped after the test suite completes, but not when using the manage.py test --keepdb option or if the user has an active session (such as an attacker's connection). - CVE-2016-9014 (access restriction bypass) Older versions of Django don't validate the Host header against settings.ALLOWED_HOSTS when settings.DEBUG=True. This makes them vulnerable to a DNS rebinding attack. While Django doesn't ship a module that allows remote code execution, this is at least a cross-site scripting vector, which could be quite serious if developers load a copy of the production database in development or connect to some production services for which there's no development instance, for example. If a project uses a package like the django-debug-toolbar, then the attacker could execute arbitrary SQL, which could be especially bad if the developers connect to the database with a superuser account. settings.ALLOWED_HOSTS is now validated regardless of DEBUG. For convenience, if ALLOWED_HOSTS is empty and DEBUG=True, the following variations of localhost are allowed ['localhost', '127.0.0.1', '::1']. If your local settings file has your production ALLOWED_HOSTS value, you must now omit it to get those fallback values. Impact ====== A remote attacker is able to bypass certain access restrictions that, depending on used django modules, is possibly leading to cross-site scripting or arbitrary SQL execution. Furthermore a remote attacker with network access to the database server is able to bypass the authentication. References ========== https://www.djangoproject.com/weblog/2016/nov/01/security-releases/ https://access.redhat.com/security/cve/CVE-2016-9013 https://access.redhat.com/security/cve/CVE-2016-9014

1 0

[arch-security] [ASA-201611-14] python2-django: multiple issues
by Levente Polyak 16 Nov '16

16 Nov '16

Arch Linux Security Advisory ASA-201611-14 ========================================== Severity: High Date : 2016-11-16 CVE-ID : CVE-2016-9013 CVE-2016-9014 Package : python2-django Type : multiple issues Remote : Yes Link : https://wiki.archlinux.org/index.php/CVE Summary ======= The package python2-django before version 1.10.3-1 is vulnerable to multiple issues including access restriction bypass and authentication bypass. Resolution ========== Upgrade to 1.10.3-1. # pacman -Syu "python2-django>=1.10.3-1" The problems have been fixed upstream in version 1.10.3. Workaround ========== - CVE-2016-9013 Ensure the debug mode is deactivated via: settings.DEBUG=False Description =========== - CVE-2016-9013 (authentication bypass) When running tests with an Oracle database, Django creates a temporary database user. In older versions, if a password isn't manually specified in the database settings TEST dictionary, a hardcoded password is used. This could allow an attacker with network access to the database server to connect. This user is usually dropped after the test suite completes, but not when using the manage.py test --keepdb option or if the user has an active session (such as an attacker's connection). - CVE-2016-9014 (access restriction bypass) Older versions of Django don't validate the Host header against settings.ALLOWED_HOSTS when settings.DEBUG=True. This makes them vulnerable to a DNS rebinding attack. While Django doesn't ship a module that allows remote code execution, this is at least a cross-site scripting vector, which could be quite serious if developers load a copy of the production database in development or connect to some production services for which there's no development instance, for example. If a project uses a package like the django-debug-toolbar, then the attacker could execute arbitrary SQL, which could be especially bad if the developers connect to the database with a superuser account. settings.ALLOWED_HOSTS is now validated regardless of DEBUG. For convenience, if ALLOWED_HOSTS is empty and DEBUG=True, the following variations of localhost are allowed ['localhost', '127.0.0.1', '::1']. If your local settings file has your production ALLOWED_HOSTS value, you must now omit it to get those fallback values. Impact ====== A remote attacker is able to bypass certain access restrictions that, depending on used django modules, is possibly leading to cross-site scripting or arbitrary SQL execution. Furthermore a remote attacker with network access to the database server is able to bypass the authentication. References ========== https://www.djangoproject.com/weblog/2016/nov/01/security-releases/ https://access.redhat.com/security/cve/CVE-2016-9013 https://access.redhat.com/security/cve/CVE-2016-9014

1 0

[arch-security] [ASA-201611-13] shutter: arbitrary code execution
by Levente Polyak 14 Nov '16

14 Nov '16

Arch Linux Security Advisory ASA-201611-13 ========================================== Severity: Medium Date : 2016-11-14 CVE-ID : CVE-2015-0854 Package : shutter Type : arbitrary code execution Remote : No Link : https://wiki.archlinux.org/index.php/CVE Summary ======= The package shutter before version 0.93.1-3 is vulnerable to arbitrary code execution. Resolution ========== Upgrade to 0.93.1-3. # pacman -Syu "shutter>=0.93.1-3" The problem has been fixed upstream but no release is available yet. Workaround ========== None. Description =========== A vulnerability has been discovered in shutter. Using the "Show in folder" menu option while viewing a file with a specially-crafted path allows arbitrary code execution with the permissions of the user running shutter. Impact ====== An attacker is able to use a specially crafted image file to execute arbitrary code by tricking the user into opening it with a specific option. References ========== https://bugs.archlinux.org/task/50735 http://seclists.org/oss-sec/2015/q3/541 https://access.redhat.com/security/cve/CVE-2015-0854

1 0

[arch-security] [ASA-201611-12] lib32-gdk-pixbuf2: arbitrary code execution
by Levente Polyak 03 Nov '16

03 Nov '16

Arch Linux Security Advisory ASA-201611-12 ========================================== Severity: Critical Date : 2016-11-03 CVE-ID : CVE-2016-6352 Package : lib32-gdk-pixbuf2 Type : arbitrary code execution Remote : Yes Link : https://wiki.archlinux.org/index.php/CVE Summary ======= The package lib32-gdk-pixbuf2 before version 2.36.0+2+ga7c869a-1 is vulnerable to arbitrary code execution. Resolution ========== Upgrade to 2.36.0+2+ga7c869a-1. # pacman -Syu "lib32-gdk-pixbuf2>=2.36.0+2+ga7c869a-1" The problem has been fixed upstream in version 2.35.3. Workaround ========== None. Description =========== An out-of-bounds write has been discovered in the OneLine32() function while parsing an ico file. A maliciously crafted file can cause the application to crash or possibly execute arbitrary code. Impact ====== A remote attacker is able to use a specially crafted ico file that, when loaded, is leading to arbitrary code execution. References ========== https://bugzilla.redhat.com/show_bug.cgi?id=1349751 https://git.gnome.org/browse/gdk-pixbuf/commit/?id=88af50a864195da1a4f7bda5… https://access.redhat.com/security/cve/CVE-2016-6352

1 0

[arch-security] [ASA-201611-11] tar: arbitrary file overwrite
by Levente Polyak 03 Nov '16

03 Nov '16

Arch Linux Security Advisory ASA-201611-11 ========================================== Severity: Medium Date : 2016-11-03 CVE-ID : CVE-2016-6321 Package : tar Type : arbitrary file overwrite Remote : Yes Link : https://wiki.archlinux.org/index.php/CVE Summary ======= The package tar before version 1.29-2 is vulnerable to arbitrary file overwrite. Resolution ========== Upgrade to 1.29-2. # pacman -Syu "tar>=1.29-2" The problem has been fixed upstream but no release is available yet. Workaround ========== None. Description =========== The GNU tar archiver attempts to avoid path traversal attacks by removing offending parts of the element name at extract. This sanitizing leads to a vulnerability where the attacker can bypass the path name(s) specified on the command line leading to arbitrary overwrite of files and directories inside the target directory. Impact ====== A remote attacker is able to use a specially crafted tar archive that, when extracted by the victim, replaces files and directories regardless of the path name(s) specified. References ========== https://bugs.archlinux.org/task/51563 http://seclists.org/fulldisclosure/2016/Oct/96 http://git.savannah.gnu.org/cgit/tar.git/commit/?id=7340f67b9860ea053 https://sintonen.fi/advisories/tar-extract-pathname-bypass.txt https://access.redhat.com/security/cve/CVE-2016-6321

1 0

[arch-security] [ASA-201611-10] lib32-libcurl-gnutls: multiple issues
by Levente Polyak 03 Nov '16

03 Nov '16

Arch Linux Security Advisory ASA-201611-10 ========================================== Severity: High Date : 2016-11-03 CVE-ID : CVE-2016-8615 CVE-2016-8616 CVE-2016-8617 CVE-2016-8618 CVE-2016-8619 CVE-2016-8621 CVE-2016-8623 CVE-2016-8624 CVE-2016-8625 Package : lib32-libcurl-gnutls Type : multiple issues Remote : Yes Link : https://wiki.archlinux.org/index.php/CVE Summary ======= The package lib32-libcurl-gnutls before version 7.51.0-1 is vulnerable to multiple issues including arbitrary code execution, content spoofing, information disclosure, insufficient validation and authentication bypass. Resolution ========== Upgrade to 7.51.0-1. # pacman -Syu "lib32-libcurl-gnutls>=7.51.0-1" The problems have been fixed upstream in version 7.51.0. Workaround ========== None. Description =========== - CVE-2016-8615 (content spoofing) If cookie state is written into a cookie jar file that is later read back and used for subsequent requests, a malicious HTTP server can inject new cookies for arbitrary domains into said cookie jar. The issue pertains to the function that loads cookies into memory, which reads the specified file into a fixed-size buffer in a line-by- line manner using the fgets() function. If an invocation of fgets() cannot read the whole line into the destination buffer due to it being too small, it truncates the output. This way, a very long cookie (name + value) sent by a malicious server would be stored in the file and subsequently that cookie could be read partially and crafted correctly, it could be treated as a different cookie for another server. - CVE-2016-8616 (authentication bypass) When re-using a connection, curl was doing case insensitive comparisons of user name and password with the existing connections. This means that if an unused connection with proper credentials exists for a protocol that has connection-scoped credentials, an attacker can cause that connection to be reused if s/he knows the case-insensitive version of the correct password. - CVE-2016-8617 (arbitrary code execution) In libcurl's base64 encode function, the output buffer is allocated as follows without any checks on insize: malloc( insize * 4 / 3 + 4 ) On systems with 32-bit addresses in userspace (e.g. x86, ARM, x32), the multiplication in the expression wraps around if insize is at least 1GB of data. If this happens, an undersized output buffer will be allocated, but the full result will be written, thus causing the memory behind the output buffer to be overwritten. If a username is set directly via CURLOPT_USERNAME (or curl's -u, --user option), this vulnerability can be triggered. The name has to be at least 512MB big in a 32bit system. Systems with 64 bit versions of the size_t type are not affected by this issue. - CVE-2016-8618 (arbitrary code execution) The libcurl API function called curl_maprintf() can be tricked into doing a double-free due to an unsafe size_t multiplication, on systems using 32 bit size_t variables. The function is also used internally in numerous situations. The function doubles an allocated memory area with realloc() and allows the size to wrap and become zero and when doing so realloc() returns NULL and frees the memory - in contrary to normal realloc() fails where it only returns NULL - causing libcurl to free the memory again in the error path. This behavior is triggerable using the publicly exposed function. Systems with 64 bit versions of the size_t type are not affected by this issue. - CVE-2016-8619 (arbitrary code execution) In curl's implementation of the Kerberos authentication mechanism, the function read_data() in security.c is used to fill the necessary krb5 structures. When reading one of the length fields from the socket, it fails to ensure that the length parameter passed to realloc() is not set to 0. This would lead to realloc() getting called with a zero size and when doing so realloc() returns NULL and frees the memory - in contrary to normal realloc() fails where it only returns NULL - causing libcurl to free the memory again in the error path. This flaw could be triggered by a malicious or just otherwise ill- behaving server. - CVE-2016-8621 (information disclosure) The curl_getdate converts a given date string into a numerical timestamp and it supports a range of different formats and possibilites to express a date and time. The underlying date parsing function is also used internally when parsing for example HTTP cookies (possibly received from remote servers) and it can be used when doing conditional HTTP requests. The date parser function uses the libc sscanf() function at two places, with the parsing strings "%02d:%02d" and ""%02d:%02d:%02d". The intent being that it would parse either a string with HH:MM (two digits colon two digits) or HH:MM:SS (two digits colon two digits colon two digits). If instead the piece of time that was sent in had the final digit cut off, thus ending with a single-digit, the date parser code would advance its read pointer one byte too much and end up reading out of bounds. - CVE-2016-8623 (arbitrary code execution) libcurl explicitly allows users to share cookies between multiple easy handles that are concurrently employed by different threads. When cookies to be sent to a server are collected, the matching function collects all cookies to send and the cookie lock is released immediately afterwards. That function however only returns a list with references back to the original strings for name, value, path and so on. Therefore, if another thread quickly takes the lock and frees one of the original cookie structs together with its strings, a use-after- free can occur possibly leading to arbitrary code execution. Another thread can also replace the contents of the cookies from separate HTTP responses or API calls. - CVE-2016-8624 (insufficient validation) curl doesn't parse the authority component of the URL correctly when the host name part ends with a '#' character, and could instead be tricked into connecting to a different host. This may have security implications if you for example use a URL parser that follows the RFC to check for allowed domains before using curl to request them. Passing in http://example.com#@evil.com/x.txt would wrongly make curl send a request to evil.com while your browser would connect to example.com given the same URL. The problem exists for most protocol schemes. - CVE-2016-8625 (insufficient validation) When curl is built with libidn to handle International Domain Names (IDNA), it translates them to puny code for DNS resolving using the IDNA 2003 standard, while IDNA 2008 is the modern and up-to-date IDNA standard. This misalignment causes problems with for example domains using the German ß character (known as the Unicode Character 'LATIN SMALL LETTER SHARP S') which is used at times in the .de TLD and is translated differently in the two IDNA standards, leading to users potentially and unknowingly issuing network transfer requests to the wrong host. For example, straße.de is translated into strasse.de using IDNA 2003 but is translated into xn--strae-oqa.de using IDNA 2008. Needless to say, those host names could very well resolve to different addresses and be two completely independent servers. IDNA 2008 is mandatory for .de domains. This name problem exists for DNS-using protocols in curl, but only when built to use libidn. Impact ====== A remote attacker is able to execute arbitrary code, inject cookies for arbitrary domains and disclose sensitive information via various vectors. References ========== https://curl.haxx.se/changes.html#7_51_0 https://curl.haxx.se/docs/adv_20161102A.html https://curl.haxx.se/docs/adv_20161102B.html https://curl.haxx.se/docs/adv_20161102C.html https://curl.haxx.se/docs/adv_20161102D.html https://curl.haxx.se/docs/adv_20161102E.html https://curl.haxx.se/docs/adv_20161102G.html https://curl.haxx.se/docs/adv_20161102I.html https://curl.haxx.se/docs/adv_20161102J.html https://curl.haxx.se/docs/adv_20161102K.html https://access.redhat.com/security/cve/CVE-2016-8615 https://access.redhat.com/security/cve/CVE-2016-8616 https://access.redhat.com/security/cve/CVE-2016-8617 https://access.redhat.com/security/cve/CVE-2016-8618 https://access.redhat.com/security/cve/CVE-2016-8619 https://access.redhat.com/security/cve/CVE-2016-8621 https://access.redhat.com/security/cve/CVE-2016-8623 https://access.redhat.com/security/cve/CVE-2016-8624 https://access.redhat.com/security/cve/CVE-2016-8625

1 0

[arch-security] [ASA-201611-9] libcurl-gnutls: multiple issues
by Levente Polyak 03 Nov '16

03 Nov '16

Arch Linux Security Advisory ASA-201611-9 ========================================= Severity: High Date : 2016-11-03 CVE-ID : CVE-2016-8615 CVE-2016-8616 CVE-2016-8617 CVE-2016-8619 CVE-2016-8621 CVE-2016-8622 CVE-2016-8623 CVE-2016-8624 CVE-2016-8625 Package : libcurl-gnutls Type : multiple issues Remote : Yes Link : https://wiki.archlinux.org/index.php/CVE Summary ======= The package libcurl-gnutls before version 7.51.0-1 is vulnerable to multiple issues including arbitrary code execution, content spoofing, information disclosure, insufficient validation and authentication bypass. Resolution ========== Upgrade to 7.51.0-1. # pacman -Syu "libcurl-gnutls>=7.51.0-1" The problems have been fixed upstream in version 7.51.0. Workaround ========== None. Description =========== - CVE-2016-8615 (content spoofing) If cookie state is written into a cookie jar file that is later read back and used for subsequent requests, a malicious HTTP server can inject new cookies for arbitrary domains into said cookie jar. The issue pertains to the function that loads cookies into memory, which reads the specified file into a fixed-size buffer in a line-by- line manner using the fgets() function. If an invocation of fgets() cannot read the whole line into the destination buffer due to it being too small, it truncates the output. This way, a very long cookie (name + value) sent by a malicious server would be stored in the file and subsequently that cookie could be read partially and crafted correctly, it could be treated as a different cookie for another server. - CVE-2016-8616 (authentication bypass) When re-using a connection, curl was doing case insensitive comparisons of user name and password with the existing connections. This means that if an unused connection with proper credentials exists for a protocol that has connection-scoped credentials, an attacker can cause that connection to be reused if s/he knows the case-insensitive version of the correct password. - CVE-2016-8617 (arbitrary code execution) In libcurl's base64 encode function, the output buffer is allocated as follows without any checks on insize: malloc( insize * 4 / 3 + 4 ) On systems with 32-bit addresses in userspace (e.g. x86, ARM, x32), the multiplication in the expression wraps around if insize is at least 1GB of data. If this happens, an undersized output buffer will be allocated, but the full result will be written, thus causing the memory behind the output buffer to be overwritten. If a username is set directly via CURLOPT_USERNAME (or curl's -u, --user option), this vulnerability can be triggered. The name has to be at least 512MB big in a 32bit system. Systems with 64 bit versions of the size_t type are not affected by this issue. - CVE-2016-8619 (arbitrary code execution) In curl's implementation of the Kerberos authentication mechanism, the function read_data() in security.c is used to fill the necessary krb5 structures. When reading one of the length fields from the socket, it fails to ensure that the length parameter passed to realloc() is not set to 0. This would lead to realloc() getting called with a zero size and when doing so realloc() returns NULL and frees the memory - in contrary to normal realloc() fails where it only returns NULL - causing libcurl to free the memory again in the error path. This flaw could be triggered by a malicious or just otherwise ill- behaving server. - CVE-2016-8621 (information disclosure) The curl_getdate converts a given date string into a numerical timestamp and it supports a range of different formats and possibilites to express a date and time. The underlying date parsing function is also used internally when parsing for example HTTP cookies (possibly received from remote servers) and it can be used when doing conditional HTTP requests. The date parser function uses the libc sscanf() function at two places, with the parsing strings "%02d:%02d" and ""%02d:%02d:%02d". The intent being that it would parse either a string with HH:MM (two digits colon two digits) or HH:MM:SS (two digits colon two digits colon two digits). If instead the piece of time that was sent in had the final digit cut off, thus ending with a single-digit, the date parser code would advance its read pointer one byte too much and end up reading out of bounds. - CVE-2016-8622 (arbitrary code execution) The URL percent-encoding decode function in libcurl is called curl_easy_unescape. Internally, even if this function would be made to allocate a unscape destination buffer larger than 2GB, it would return that new length in a signed 32 bit integer variable, thus the length would get either just truncated or both truncated and turned negative. That could then lead to libcurl writing outside of its heap based buffer. This can be triggered by a user on a 64bit system if the user can send in a custom (very large) URL to a libcurl using program. - CVE-2016-8623 (arbitrary code execution) libcurl explicitly allows users to share cookies between multiple easy handles that are concurrently employed by different threads. When cookies to be sent to a server are collected, the matching function collects all cookies to send and the cookie lock is released immediately afterwards. That function however only returns a list with references back to the original strings for name, value, path and so on. Therefore, if another thread quickly takes the lock and frees one of the original cookie structs together with its strings, a use-after- free can occur possibly leading to arbitrary code execution. Another thread can also replace the contents of the cookies from separate HTTP responses or API calls. - CVE-2016-8624 (insufficient validation) curl doesn't parse the authority component of the URL correctly when the host name part ends with a '#' character, and could instead be tricked into connecting to a different host. This may have security implications if you for example use a URL parser that follows the RFC to check for allowed domains before using curl to request them. Passing in http://example.com#@evil.com/x.txt would wrongly make curl send a request to evil.com while your browser would connect to example.com given the same URL. The problem exists for most protocol schemes. - CVE-2016-8625 (insufficient validation) When curl is built with libidn to handle International Domain Names (IDNA), it translates them to puny code for DNS resolving using the IDNA 2003 standard, while IDNA 2008 is the modern and up-to-date IDNA standard. This misalignment causes problems with for example domains using the German ß character (known as the Unicode Character 'LATIN SMALL LETTER SHARP S') which is used at times in the .de TLD and is translated differently in the two IDNA standards, leading to users potentially and unknowingly issuing network transfer requests to the wrong host. For example, straße.de is translated into strasse.de using IDNA 2003 but is translated into xn--strae-oqa.de using IDNA 2008. Needless to say, those host names could very well resolve to different addresses and be two completely independent servers. IDNA 2008 is mandatory for .de domains. This name problem exists for DNS-using protocols in curl, but only when built to use libidn. Impact ====== A remote attacker is able to execute arbitrary code, inject cookies for arbitrary domains and disclose sensitive information via various vectors. References ========== https://curl.haxx.se/changes.html#7_51_0 https://curl.haxx.se/docs/adv_20161102A.html https://curl.haxx.se/docs/adv_20161102B.html https://curl.haxx.se/docs/adv_20161102C.html https://curl.haxx.se/docs/adv_20161102E.html https://curl.haxx.se/docs/adv_20161102G.html https://curl.haxx.se/docs/adv_20161102H.html https://curl.haxx.se/docs/adv_20161102I.html https://curl.haxx.se/docs/adv_20161102J.html https://curl.haxx.se/docs/adv_20161102K.html https://access.redhat.com/security/cve/CVE-2016-8615 https://access.redhat.com/security/cve/CVE-2016-8616 https://access.redhat.com/security/cve/CVE-2016-8617 https://access.redhat.com/security/cve/CVE-2016-8619 https://access.redhat.com/security/cve/CVE-2016-8621 https://access.redhat.com/security/cve/CVE-2016-8622 https://access.redhat.com/security/cve/CVE-2016-8623 https://access.redhat.com/security/cve/CVE-2016-8624 https://access.redhat.com/security/cve/CVE-2016-8625

1 0

[arch-security] [ASA-201611-8] libcurl-compat: multiple issues
by Levente Polyak 03 Nov '16

03 Nov '16

Arch Linux Security Advisory ASA-201611-8 ========================================= Severity: High Date : 2016-11-03 CVE-ID : CVE-2016-8615 CVE-2016-8616 CVE-2016-8617 CVE-2016-8619 CVE-2016-8621 CVE-2016-8622 CVE-2016-8623 CVE-2016-8624 CVE-2016-8625 Package : libcurl-compat Type : multiple issues Remote : Yes Link : https://wiki.archlinux.org/index.php/CVE Summary ======= The package libcurl-compat before version 7.51.0-1 is vulnerable to multiple issues including arbitrary code execution, content spoofing, information disclosure, insufficient validation and authentication bypass. Resolution ========== Upgrade to 7.51.0-1. # pacman -Syu "libcurl-compat>=7.51.0-1" The problems have been fixed upstream in version 7.51.0. Workaround ========== None. Description =========== - CVE-2016-8615 (content spoofing) If cookie state is written into a cookie jar file that is later read back and used for subsequent requests, a malicious HTTP server can inject new cookies for arbitrary domains into said cookie jar. The issue pertains to the function that loads cookies into memory, which reads the specified file into a fixed-size buffer in a line-by- line manner using the fgets() function. If an invocation of fgets() cannot read the whole line into the destination buffer due to it being too small, it truncates the output. This way, a very long cookie (name + value) sent by a malicious server would be stored in the file and subsequently that cookie could be read partially and crafted correctly, it could be treated as a different cookie for another server. - CVE-2016-8616 (authentication bypass) When re-using a connection, curl was doing case insensitive comparisons of user name and password with the existing connections. This means that if an unused connection with proper credentials exists for a protocol that has connection-scoped credentials, an attacker can cause that connection to be reused if s/he knows the case-insensitive version of the correct password. - CVE-2016-8617 (arbitrary code execution) In libcurl's base64 encode function, the output buffer is allocated as follows without any checks on insize: malloc( insize * 4 / 3 + 4 ) On systems with 32-bit addresses in userspace (e.g. x86, ARM, x32), the multiplication in the expression wraps around if insize is at least 1GB of data. If this happens, an undersized output buffer will be allocated, but the full result will be written, thus causing the memory behind the output buffer to be overwritten. If a username is set directly via CURLOPT_USERNAME (or curl's -u, --user option), this vulnerability can be triggered. The name has to be at least 512MB big in a 32bit system. Systems with 64 bit versions of the size_t type are not affected by this issue. - CVE-2016-8619 (arbitrary code execution) In curl's implementation of the Kerberos authentication mechanism, the function read_data() in security.c is used to fill the necessary krb5 structures. When reading one of the length fields from the socket, it fails to ensure that the length parameter passed to realloc() is not set to 0. This would lead to realloc() getting called with a zero size and when doing so realloc() returns NULL and frees the memory - in contrary to normal realloc() fails where it only returns NULL - causing libcurl to free the memory again in the error path. This flaw could be triggered by a malicious or just otherwise ill- behaving server. - CVE-2016-8621 (information disclosure) The curl_getdate converts a given date string into a numerical timestamp and it supports a range of different formats and possibilites to express a date and time. The underlying date parsing function is also used internally when parsing for example HTTP cookies (possibly received from remote servers) and it can be used when doing conditional HTTP requests. The date parser function uses the libc sscanf() function at two places, with the parsing strings "%02d:%02d" and ""%02d:%02d:%02d". The intent being that it would parse either a string with HH:MM (two digits colon two digits) or HH:MM:SS (two digits colon two digits colon two digits). If instead the piece of time that was sent in had the final digit cut off, thus ending with a single-digit, the date parser code would advance its read pointer one byte too much and end up reading out of bounds. - CVE-2016-8622 (arbitrary code execution) The URL percent-encoding decode function in libcurl is called curl_easy_unescape. Internally, even if this function would be made to allocate a unscape destination buffer larger than 2GB, it would return that new length in a signed 32 bit integer variable, thus the length would get either just truncated or both truncated and turned negative. That could then lead to libcurl writing outside of its heap based buffer. This can be triggered by a user on a 64bit system if the user can send in a custom (very large) URL to a libcurl using program. - CVE-2016-8623 (arbitrary code execution) libcurl explicitly allows users to share cookies between multiple easy handles that are concurrently employed by different threads. When cookies to be sent to a server are collected, the matching function collects all cookies to send and the cookie lock is released immediately afterwards. That function however only returns a list with references back to the original strings for name, value, path and so on. Therefore, if another thread quickly takes the lock and frees one of the original cookie structs together with its strings, a use-after- free can occur possibly leading to arbitrary code execution. Another thread can also replace the contents of the cookies from separate HTTP responses or API calls. - CVE-2016-8624 (insufficient validation) curl doesn't parse the authority component of the URL correctly when the host name part ends with a '#' character, and could instead be tricked into connecting to a different host. This may have security implications if you for example use a URL parser that follows the RFC to check for allowed domains before using curl to request them. Passing in http://example.com#@evil.com/x.txt would wrongly make curl send a request to evil.com while your browser would connect to example.com given the same URL. The problem exists for most protocol schemes. - CVE-2016-8625 (insufficient validation) When curl is built with libidn to handle International Domain Names (IDNA), it translates them to puny code for DNS resolving using the IDNA 2003 standard, while IDNA 2008 is the modern and up-to-date IDNA standard. This misalignment causes problems with for example domains using the German ß character (known as the Unicode Character 'LATIN SMALL LETTER SHARP S') which is used at times in the .de TLD and is translated differently in the two IDNA standards, leading to users potentially and unknowingly issuing network transfer requests to the wrong host. For example, straße.de is translated into strasse.de using IDNA 2003 but is translated into xn--strae-oqa.de using IDNA 2008. Needless to say, those host names could very well resolve to different addresses and be two completely independent servers. IDNA 2008 is mandatory for .de domains. This name problem exists for DNS-using protocols in curl, but only when built to use libidn. Impact ====== A remote attacker is able to execute arbitrary code, inject cookies for arbitrary domains and disclose sensitive information via various vectors. References ========== https://curl.haxx.se/changes.html#7_51_0 https://curl.haxx.se/docs/adv_20161102A.html https://curl.haxx.se/docs/adv_20161102B.html https://curl.haxx.se/docs/adv_20161102C.html https://curl.haxx.se/docs/adv_20161102E.html https://curl.haxx.se/docs/adv_20161102G.html https://curl.haxx.se/docs/adv_20161102H.html https://curl.haxx.se/docs/adv_20161102I.html https://curl.haxx.se/docs/adv_20161102J.html https://curl.haxx.se/docs/adv_20161102K.html https://access.redhat.com/security/cve/CVE-2016-8615 https://access.redhat.com/security/cve/CVE-2016-8616 https://access.redhat.com/security/cve/CVE-2016-8617 https://access.redhat.com/security/cve/CVE-2016-8619 https://access.redhat.com/security/cve/CVE-2016-8621 https://access.redhat.com/security/cve/CVE-2016-8622 https://access.redhat.com/security/cve/CVE-2016-8623 https://access.redhat.com/security/cve/CVE-2016-8624 https://access.redhat.com/security/cve/CVE-2016-8625

1 0