Compound Element ID: 689 (Compound Element Base: Composite)
Status: Draft
Description
Description Summary
The product, while copying or cloning a resource, does not set the resource's permissions or access control until the copy is complete, leaving the resource exposed to other spheres while the copy is taking place.
Time of Introduction
Implementation
Applicable Platforms
Languages
C
Perl
Common Consequences
Scope
Effect
Confidentiality
Integrity
Technical Impact: Read application
data; Modify application
data
database product creates files world-writable
before initializing the setuid bits, leading to modification of
executables.
Other Notes
This is a general issue, although few subtypes are currently known. The
most common examples occur in file archive extraction, in which the product
begins the extraction with insecure default permissions, then only sets the
final permissions (as specified in the archive) once the copy is complete.
The larger the archive, the larger the timing window for the race condition.
This weakness has also occurred in some operating system utilities that
perform copies of deeply nested directories containing a large number of
files.
Weakness Ordinalities
Ordinality
Description
Primary
(where
the weakness exists independent of other weaknesses)
Under-studied. It seems likely that this weakness could occur in any
situation in which a complex or large copy operation occurs, when the
resource can be made available to other spheres as soon as it is created,
but before its initialization is complete.
Concurrent Execution using Shared Resource with Improper Synchronization ('Race Condition')
Definition in a New Window
Weakness ID: 362 (Weakness Class)
Status: Draft
Description
Description Summary
The program contains a code sequence that can run concurrently with other code, and the code sequence requires temporary, exclusive access to a shared resource, but a timing window exists in which the shared resource can be modified by another code sequence that is operating concurrently.
Extended Description
This can have security implications when the expected synchronization is in security-critical code, such as recording whether a user is authenticated or modifying important state information that should not be influenced by an outsider.
A race condition occurs within concurrent environments, and is effectively a property of a code sequence. Depending on the context, a code sequence may be in the form of a function call, a small number of instructions, a series of program invocations, etc.
A race condition violates these properties, which are closely related:
Exclusivity - the code sequence is given exclusive access to the shared resource, i.e., no other code sequence can modify properties of the shared resource before the original sequence has completed execution.
Atomicity - the code sequence is behaviorally atomic, i.e., no other thread or process can concurrently execute the same sequence of instructions (or a subset) against the same resource.
A race condition exists when an "interfering code sequence" can still access the shared resource, violating exclusivity. Programmers may assume that certain code sequences execute too quickly to be affected by an interfering code sequence; when they are not, this violates atomicity. For example, the single "x++" statement may appear atomic at the code layer, but it is actually non-atomic at the instruction layer, since it involves a read (the original value of x), followed by a computation (x+1), followed by a write (save the result to x).
The interfering code sequence could be "trusted" or "untrusted." A trusted interfering code sequence occurs within the program; it cannot be modified by the attacker, and it can only be invoked indirectly. An untrusted interfering code sequence can be authored directly by the attacker, and typically it is external to the vulnerable program.
Time of Introduction
Architecture and Design
Implementation
Applicable Platforms
Languages
C: (Sometimes)
C++: (Sometimes)
Java: (Sometimes)
Language-independent
Architectural Paradigms
Concurrent Systems Operating on Shared Resources: (Often)
When a race condition makes it possible to bypass a resource cleanup routine or trigger multiple initialization routines, it may lead to resource exhaustion (CWE-400).
When a race condition allows multiple control flows to access a
resource simultaneously, it might lead the program(s) into unexpected
states, possibly resulting in a crash.
Confidentiality
Integrity
Technical Impact: Read files or
directories; Read application
data
When a race condition is combined with predictable resource names and loose permissions, it may be possible for an attacker to overwrite or access confidential data (CWE-59).
Likelihood of Exploit
Medium
Detection Methods
Black Box
Black box methods may be able to identify evidence of race conditions
via methods such as multiple simultaneous connections, which may cause
the software to become instable or crash. However, race conditions with
very narrow timing windows would not be detectable.
White Box
Common idioms are detectable in white box analysis, such as time-of-check-time-of-use (TOCTOU) file operations (CWE-367), or double-checked locking (CWE-609).
Automated Dynamic Analysis
This weakness can be detected using dynamic tools and techniques that
interact with the software using large test suites with many diverse
inputs, such as fuzz testing (fuzzing), robustness testing, and fault
injection. The software's operation may slow down, but it should not
become unstable, crash, or generate incorrect results.
Race conditions may be detected with a stress-test by calling the
software simultaneously from a large number of threads or processes, and
look for evidence of any unexpected behavior.
Insert breakpoints or delays in between relevant code statements to
artificially expand the race window so that it will be easier to
detect.
Effectiveness: Moderate
Demonstrative Examples
Example 1
This code could be used in an e-commerce application that supports
transfers between accounts. It takes the total amount of the transfer, sends
it to the new account, and deducts the amount from the original
account.
(Bad Code)
Example
Language: Perl
$transfer_amount = GetTransferAmount();
$balance = GetBalanceFromDatabase();
if ($transfer_amount < 0) {
FatalError("Bad Transfer Amount");
}
$newbalance = $balance - $transfer_amount;
if (($balance - $transfer_amount) < 0) {
FatalError("Insufficient Funds");
}
SendNewBalanceToDatabase($newbalance);
NotifyUser("Transfer of $transfer_amount succeeded.");
NotifyUser("New balance: $newbalance");
A race condition could occur between the calls to
GetBalanceFromDatabase() and SendNewBalanceToDatabase().
Suppose the balance is initially 100.00. An attack could be
constructed as follows:
(Attack)
Example
Language: PseudoCode
The attacker makes two simultaneous calls of the program, CALLER-1
and CALLER-2. Both callers are for the same user account.
CALLER-1 (the attacker) is associated with PROGRAM-1 (the instance
that handles CALLER-1). CALLER-2 is associated with
PROGRAM-2.
CALLER-1 makes a transfer request of 80.00.
PROGRAM-1 calls GetBalanceFromDatabase and sets $balance to
100.00
PROGRAM-1 calculates $newbalance as 20.00, then calls
SendNewBalanceToDatabase().
Due to high server load, the PROGRAM-1 call to
SendNewBalanceToDatabase() encounters a delay.
CALLER-2 makes a transfer request of 1.00.
PROGRAM-2 calls GetBalanceFromDatabase() and sets $balance to
100.00. This happens because the previous PROGRAM-1 request was not
processed yet.
PROGRAM-2 determines the new balance as 99.00.
After the initial delay, PROGRAM-1 commits its balance to the
database, setting it to 20.00.
PROGRAM-2 sends a request to update the database, setting the
balance to 99.00
At this stage, the attacker should have a balance of 19.00 (due to
81.00 worth of transfers), but the balance is 99.00, as recorded in the
database.
To prevent this weakness, the programmer has several options,
including using a lock to prevent multiple simultaneous requests to the
web application, or using a synchronization mechanism that includes all
the code between GetBalanceFromDatabase() and
SendNewBalanceToDatabase().
Example 2
The following function attempts to acquire a lock in order to
perform operations on a shared resource.
(Bad Code)
Example
Language: C
void f(pthread_mutex_t *mutex) {
pthread_mutex_lock(mutex);
/* access shared resource */
pthread_mutex_unlock(mutex);
}
However, the code does not check the value returned by
pthread_mutex_lock() for errors. If pthread_mutex_lock() cannot acquire
the mutex for any reason, the function may introduce a race condition
into the program and result in undefined behavior.
In order to avoid data races, correctly written programs must check
the result of thread synchronization functions and appropriately handle
all errors, either by attempting to recover from them or reporting it to
higher levels.
chain: time-of-check time-of-use (TOCTOU) race
condition in program allows bypass of protection mechanism that was designed
to prevent symlink attacks.
chain: time-of-check time-of-use (TOCTOU) race
condition in program allows bypass of protection mechanism that was designed
to prevent symlink attacks.
chain: race condition might allow resource to be
released before operating on it, leading to NULL dereference
Potential Mitigations
Phase: Architecture and Design
In languages that support it, use synchronization primitives. Only
wrap these around critical code to minimize the impact on
performance.
Phase: Architecture and Design
Use thread-safe capabilities such as the data access abstraction in
Spring.
Phase: Architecture and Design
Minimize the usage of shared resources in order to remove as much
complexity as possible from the control flow and to reduce the
likelihood of unexpected conditions occurring.
Additionally, this will minimize the amount of synchronization necessary and may even help to reduce the likelihood of a denial of service where an attacker may be able to repeatedly trigger a critical section (CWE-400).
Phase: Implementation
When using multithreading and operating on shared variables, only use
thread-safe functions.
Phase: Implementation
Use atomic operations on shared variables. Be wary of innocent-looking
constructs such as "x++". This may appear atomic at the code layer, but
it is actually non-atomic at the instruction layer, since it involves a
read, followed by a computation, followed by a write.
Phase: Implementation
Use a mutex if available, but be sure to avoid related weaknesses such as CWE-412.
Phase: Implementation
Avoid double-checked locking (CWE-609) and other implementation errors that arise when trying to avoid the overhead of synchronization.
Phase: Implementation
Disable interrupts or signals over critical parts of the code, but
also make sure that the code does not go into a large or infinite
loop.
Phase: Implementation
Use the volatile type modifier for critical variables to avoid
unexpected compiler optimization or reordering. This does not
necessarily solve the synchronization problem, but it can help.
Phases: Architecture and Design; Operation
Strategy: Environment Hardening
Run your code using the lowest privileges that are required to
accomplish the necessary tasks. If possible, create isolated accounts
with limited privileges that are only used for a single task. That way,
a successful attack will not immediately give the attacker access to the
rest of the software or its environment. For example, database
applications rarely need to run as the database administrator,
especially in day-to-day operations.
Race conditions in web applications are under-studied and probably
under-reported. However, in 2008 there has been growing interest in this
area.
Much of the focus of race condition research has been in Time-of-check Time-of-use (TOCTOU) variants (CWE-367), but many race conditions are related to synchronization problems that do not necessarily require a time-of-check.
Taxonomy Mappings
Mapped Taxonomy Name
Node ID
Fit
Mapped Node Name
PLOVER
Race Conditions
CERT C Secure Coding
FIO31-C
Do not simultaneously open the same file multiple
times
CERT Java Secure Coding
VNA03-J
Do not assume that a group of calls to independently atomic
methods is atomic
CERT C++ Secure Coding
FIO31-CPP
Do not simultaneously open the same file multiple
times
The relationship between race conditions and synchronization problems (CWE-662) needs to be further developed. They are not necessarily two perspectives of the same core concept, since synchronization is only one technique for avoiding race conditions, and synchronization can be used for other purposes besides race condition prevention.
Incorrect Permission Assignment for Critical Resource
Definition in a New Window
Weakness ID: 732 (Weakness Class)
Status: Draft
Description
Description Summary
The software specifies permissions for a security-critical resource in a way that allows that resource to be read or modified by unintended actors.
Extended Description
When a resource is given a permissions setting that provides access to a wider range of actors than required, it could lead to the exposure of sensitive information, or the modification of that resource by unintended parties. This is especially dangerous when the resource is related to program configuration, execution or sensitive user data.
Time of Introduction
Architecture and Design
Implementation
Installation
Operation
Applicable Platforms
Languages
Language-independent
Modes of Introduction
The developer may set loose permissions in order to minimize problems when
the user first runs the program, then create documentation stating that
permissions should be tightened. Since system administrators and users do
not always read the documentation, this can result in insecure permissions
being left unchanged.
The developer might make certain assumptions about the environment in
which the software runs - e.g., that the software is running on a
single-user system, or the software is only accessible to trusted
administrators. When the software is running in a different environment, the
permissions become a problem.
Common Consequences
Scope
Effect
Confidentiality
Technical Impact: Read application
data; Read files or
directories
An attacker may be able to read sensitive information from the
associated resource, such as credentials or configuration information
stored in a file.
Access Control
Technical Impact: Gain privileges / assume
identity
An attacker may be able to modify critical properties of the
associated resource to gain privileges, such as replacing a
world-writable executable with a Trojan horse.
Integrity
Other
Technical Impact: Modify application
data; Other
An attacker may be able to destroy or corrupt critical data in the
associated resource, such as deletion of records from a database.
Likelihood of Exploit
Medium to High
Detection Methods
Automated Static Analysis
Automated static analysis may be effective in detecting permission
problems for system resources such as files, directories, shared memory,
device interfaces, etc. Automated techniques may be able to detect the
use of library functions that modify permissions, then analyze function
calls for arguments that contain potentially insecure values.
However, since the software's intended security policy might allow
loose permissions for certain operations (such as publishing a file on a
web server), automated static analysis may produce some false positives
- i.e., warnings that do not have any security consequences or require
any code changes.
When custom permissions models are used - such as defining who can
read messages in a particular forum in a bulletin board system - these
can be difficult to detect using automated static analysis. It may be
possible to define custom signatures that identify any custom functions
that implement the permission checks and assignments.
Automated Dynamic Analysis
Automated dynamic analysis may be effective in detecting permission
problems for system resources such as files, directories, shared memory,
device interfaces, etc.
However, since the software's intended security policy might allow
loose permissions for certain operations (such as publishing a file on a
web server), automated dynamic analysis may produce some false positives
- i.e., warnings that do not have any security consequences or require
any code changes.
When custom permissions models are used - such as defining who can
read messages in a particular forum in a bulletin board system - these
can be difficult to detect using automated dynamic analysis. It may be
possible to define custom signatures that identify any custom functions
that implement the permission checks and assignments.
Manual Analysis
This weakness can be detected using tools and techniques that require
manual (human) analysis, such as penetration testing, threat modeling,
and interactive tools that allow the tester to record and modify an
active session.
These may be more effective than strictly automated techniques. This
is especially the case with weaknesses that are related to design and
business rules.
Manual Static Analysis
Manual static analysis may be effective in detecting the use of custom
permissions models and functions. The code could then be examined to
identifying usage of the related functions. Then the human analyst could
evaluate permission assignments in the context of the intended security
model of the software.
Manual Dynamic Analysis
Manual dynamic analysis may be effective in detecting the use of
custom permissions models and functions. The program could then be
executed with a focus on exercising code paths that are related to the
custom permissions. Then the human analyst could evaluate permission
assignments in the context of the intended security model of the
software.
Fuzzing
Fuzzing is not effective in detecting this weakness.
Black Box
Use monitoring tools that examine the software's process as it
interacts with the operating system and the network. This technique is
useful in cases when source code is unavailable, if the software was not
developed by you, or if you want to verify that the build phase did not
introduce any new weaknesses. Examples include debuggers that directly
attach to the running process; system-call tracing utilities such as
truss (Solaris) and strace (Linux); system activity monitors such as
FileMon, RegMon, Process Monitor, and other Sysinternals utilities
(Windows); and sniffers and protocol analyzers that monitor network
traffic.
Attach the monitor to the process and watch for library functions or
system calls on OS resources such as files, directories, and shared
memory. Examine the arguments to these calls to infer which permissions
are being used.
Note that this technique is only useful for permissions issues related
to system resources. It is not likely to detect application-level
business rules that are related to permissions, such as if a user of a
blog system marks a post as "private," but the blog system inadvertently
marks it as "public."
Demonstrative Examples
Example 1
The following code sets the umask of the process to 0 before
creating a file and writing "Hello world" into the file.
(Bad Code)
Example
Language: C
#define OUTFILE "hello.out"
umask(0);
FILE *out;
/* Ignore CWE-59 (link following) for brevity */
out = fopen(OUTFILE, "w");
if (out) {
fprintf(out, "hello world!\n");
fclose(out);
}
After running this program on a UNIX system, running the "ls -l"
command might return the following output:
(Result)
-rw-rw-rw- 1 username 13 Nov 24 17:58 hello.out
The "rw-rw-rw-" string indicates that the owner, group, and world (all
users) can read the file and write to it.
Example 2
This code creates a home directory for a new user, and makes that
user the owner of the directory. If the new directory cannot be owned by the
user, the directory is deleted.
(Bad Code)
Example
Language: PHP
function createUserDir($username){
$path = '/home/'.$username;
if(!mkdir($path)){
return false;
}
if(!chown($path,$username)){
rmdir($path);
return false;
}
return true;
}
Because the optional "mode" argument is omitted from the call to
mkdir(), the directory is created with the default permissions 0777.
Simply setting the new user as the owner of the directory does not
explicitly change the permissions of the directory, leaving it with the
default. This default allows any user to read and write to the
directory, allowing an attack on the user's files. The code also fails
to change the owner group of the directory, which may result in access
by unexpected groups.
This code may also be vulnerable to Path Traversal (CWE-22) attacks if an attacker supplies a non alphanumeric username.
Example 3
The following code snippet might be used as a monitor to
periodically record whether a web site is alive. To ensure that the file can
always be modified, the code uses chmod() to make the file
world-writable.
The first time the program runs, it might create a new file that
inherits the permissions from its environment. A file listing might look
like:
(Result)
-rw-r--r-- 1 username 13 Nov 24 17:58 secretFile.out
This listing might occur when the user has a default umask of 022,
which is a common setting. Depending on the nature of the file, the user
might not have intended to make it readable by everyone on the
system.
The next time the program runs, however - and all subsequent
executions - the chmod will set the file's permissions so that the
owner, group, and world (all users) can read the file and write to
it:
(Result)
-rw-rw-rw- 1 username 13 Nov 24 17:58 secretFile.out
Perhaps the programmer tried to do this because a different process
uses different permissions that might prevent the file from being
updated.
Example 4
The following command recursively sets world-readable permissions
for a directory and all of its children:
(Bad Code)
Example
Language: Shell
chmod -R ugo+r DIRNAME
If this command is run from a program, the person calling the program
might not expect that all the files under the directory will be
world-readable. If the directory is expected to contain private data,
this could become a security problem.
Anti-virus product sets insecure "Everyone: Full
Control" permissions for files under the "Program Files" folder, allowing
attackers to replace executables with Trojan
horses.
Library function copies a file to a new target and
uses the source file's permissions for the target, which is incorrect when
the source file is a symbolic link, which typically has 0777 permissions.
Product uses "Everyone: Full Control" permissions
for memory-mapped files (shared memory) in inter-process communication,
allowing attackers to tamper with a session.
Chain: database product contains buffer overflow
that is only reachable through a .ini configuration file - which has
"Everyone: Full Control" permissions.
Potential Mitigations
Phase: Implementation
When using a critical resource such as a configuration file, check to see if the resource has insecure permissions (such as being modifiable by any regular user) [R.732.1], and generate an error or even exit the software if there is a possibility that the resource could have been modified by an unauthorized party.
Phase: Architecture and Design
Divide your application into anonymous, normal, privileged, and administrative areas. Reduce the attack surface by carefully defining distinct user groups, privileges, and/or roles. Map these against data, functionality, and the related resources. Then set the permissions accordingly. This will allow you to maintain more fine-grained control over your resources. [R.732.2]
Effectiveness: Moderate
This can be an effective strategy. However, in practice, it may be
difficult or time consuming to define these areas when there are many
different resources or user types, or if the applications features
change rapidly.
Phases: Architecture and Design; Operation
Strategy: Sandbox or Jail
Run your code in a "jail" or similar sandbox environment that enforces
strict boundaries between the process and the operating system. This may
effectively restrict which files can be accessed in a particular
directory or which commands can be executed by your software.
OS-level examples include the Unix chroot jail, AppArmor, and SELinux.
In general, managed code may provide some protection. For example,
java.io.FilePermission in the Java SecurityManager allows you to specify
restrictions on file operations.
This may not be a feasible solution, and it only limits the impact to
the operating system; the rest of your application may still be subject
to compromise.
Be careful to avoid CWE-243 and other weaknesses related to jails.
Effectiveness: Moderate
The effectiveness of this mitigation depends on the prevention
capabilities of the specific sandbox or jail being used and might only
help to reduce the scope of an attack, such as restricting the attacker
to certain system calls or limiting the portion of the file system that
can be accessed.
Phases: Implementation; Installation
During program startup, explicitly set the default permissions or
umask to the most restrictive setting possible. Also set the appropriate
permissions during program installation. This will prevent you from
inheriting insecure permissions from any user who installs or runs the
program.
Effectiveness: High
Phase: System Configuration
For all configuration files, executables, and libraries, make sure
that they are only readable and writable by the software's
administrator.
Effectiveness: High
Phase: Documentation
Do not suggest insecure configuration changes in your documentation,
especially if those configurations can extend to resources and other
software that are outside the scope of your own software.
Phase: Installation
Do not assume that the system administrator will manually change the
configuration to the settings that you recommend in the manual.
Phases: Operation; System Configuration
Strategy: Environment Hardening
Ensure that your software runs properly under the Federal Desktop Core Configuration (FDCC) [R.732.4] or an equivalent hardening configuration guide, which many organizations use to limit the attack surface and potential risk of deployed software.
[R.732.1] Mark Dowd, John McDonald
and Justin Schuh. "The Art of Software Security Assessment". Chapter 9, "File Permissions." Page 495.. 1st Edition. Addison Wesley. 2006.
[R.732.2] John Viega and
Gary McGraw. "Building Secure Software". Chapter 8, "Access Control." Page 194.. 1st Edition. Addison-Wesley. 2002.
The relationships between privileges, permissions, and actors (e.g. users and groups) need further refinement within the Research view. One complication is that these concepts apply to two different pillars, related to control of resources (CWE-664) and protection mechanism failures (CWE-396).
Description
