Mutexes are often used for critical resources to prevent multiple threads from accessing them at the same time. Sometimes, when locking mutexes, deadlock will happen when multiple threads hold each other's lock and the program consequently comes to a halt. There are four requirements for deadlock:
- Mutual Exclusion
- Hold and Wait
- No Preemption
- Circular Wait
Each deadlock requires all four conditions. Therefore, to prevent deadlock, prevent any one of the four conditions from being satisfied. This guideline recommends locking the mutexes in a predefined order to prevent circular wait.
Noncompliant Code Example
The following code has behavior which is dependend on the runtime environment and the platform's scheduler. However, with proper timing, the main()
function will deadlock when running thr1
and thr2
in which thr1
tries to lock ba2
's mutex while thr2
tries to lock on ba1
's mutex in the deposit()
function and the program will not progress.
typedef struct { int balance; pthread_mutex_t balance_mutex; } bank_account; typedef struct { bank_account *from; bank_account *to; int amount; } deposit_thr_args; void create_bank_account(bank_account **ba, int initial_amount) { int result; bank_account *nba = malloc(sizeof(bank_account)); if (nba == NULL) { /* Handle Error */ } nba->balance = initial_amount; result = pthread_mutex_init(&nba->balance_mutex, NULL); if (result) { /* Handle Error */ } *ba = nba; } void *deposit(void *ptr) { int result; deposit_thr_args *args = (deposit_thr_args *)ptr; if ((result = pthread_mutex_lock(&(args->from->balance_mutex))) != 0) { /* Handle Error */ } /* not enough balance to transfer */ if (args->from->balance < args->amount) { if ((result = pthread_mutex_unlock(&(args->from->balance_mutex))) != 0) { /* Handle Error */ } return NULL; } if ((result = pthread_mutex_lock(&(args->to->balance_mutex))) != 0) { /* Handle Error */ } args->from->balance -= args->amount; args->to->balance += args->amount; if ((result = pthread_mutex_unlock(&(args->from->balance_mutex))) != 0) { /* Handle Error */ } if ((result = pthread_mutex_unlock(&(args->to->balance_mutex))) != 0) { /* Handle Error */ } free(ptr); return NULL; } int main(void) { pthread_t thr1, thr2; int result; bank_account *ba1; bank_account *ba2; create_bank_account(&ba1, 1000); create_bank_account(&ba2, 1000); deposit_thr_args *arg1 = malloc(sizeof(deposit_thr_args)); if (arg1 == NULL) { /* Handle Error */ } deposit_thr_args *arg2 = malloc(sizeof(deposit_thr_args)); if (arg2 == NULL) { /* Handle Error */ } arg1->from = ba1; arg1->to = ba2; arg1->amount = 100; arg2->from = ba2; arg2->to = ba1; arg2->amount = 100; /* perform the deposits */ if ((result = pthread_create(&thr1, NULL, deposit, (void *)arg1)) != 0) { /* Handle Error */ } if ((result = pthread_create(&thr2, NULL, deposit, (void *)arg2)) != 0) { /* Handle Error */ } pthread_exit(NULL); return 0; }
Compliant Solution
The solution to the deadlock problem is to use a predefined order for the locks in the deposit()
function. In the following compliant solution, each thread will lock based on the id of bank_account
id defined in the struct initialization. This prevents the circular wait problem.
typedef struct { int balance; pthread_mutex_t balance_mutex; unsigned int id; /* should never be changed after initialized */ } bank_account; unsigned int global_id = 1; void create_bank_account(bank_account **ba, int initial_amount) { int result; bank_account *nba = malloc(sizeof(bank_account)); if (nba == NULL) { /* Handle Error */ } nba->balance = initial_amount; result = pthread_mutex_init(&nba->balance_mutex, NULL); if (result != 0) { /* Handle Error */ } nba->id = global_id++; *ba = nba; } void *deposit(void *ptr) { deposit_thr_args *args = (deposit_thr_args *)ptr; int result; if (args->from->id == args->to->id) return; /* ensure proper ordering for locking */ if (args->from->id < args->to->id) { if ((result = pthread_mutex_lock(&(args->from->balance_mutex))) != 0) { /* Handle Error */ } if ((result = pthread_mutex_lock(&(args->to->balance_mutex))) != 0) { /* Handle Error */ } } else { if ((result = pthread_mutex_lock(&(args->to->balance_mutex))) != 0) { /* Handle Error */ } if ((result = pthread_mutex_lock(&(args->from->balance_mutex))) != 0) { /* Handle Error */ } } /* not enough balance to transfer */ if (args->from->balance < args->amount) { if ((result = pthread_mutex_unlock(&(args->from->balance_mutex))) != 0) { /* Handle Error */ } if ((result = pthread_mutex_unlock(&(args->to->balance_mutex))) != 0) { /* Handle Error */ } return; } args->from->balance -= args->amount; args->to->balance += args->amount; if ((result = pthread_mutex_unlock(&(args->from->balance_mutex))) != 0) { /* Handle Error */ } if ((result = pthread_mutex_unlock(&(args->to->balance_mutex))) != 0) { /* Handle Error */ } free(ptr); return; }
Risk Assessment
Deadlock causes multiple threads to become unable to progress and thus halts the executing program. This is a potential denial-of-service attack because the attacker can force deadlock situations. It is likely for deadlock to occur in multi-threaded programs that manage multiple shared resources.
Recommendation |
Severity |
Likelihood |
Remediation Cost |
Level |
Priority |
---|---|---|---|---|---|
CON35-C |
low |
probable |
medium |
L3 |
P3 |
Other Languages
This rule appears in the Java Secure Coding Standard as LCK07-J. Avoid deadlock by requesting and releasing locks in the same order.
References
[pthread_mutex ] pthread_mutex tutorial
[MITRE CWE:764 ] Multiple Locks of Critical Resources
[[Bryant 03]] Chapter 13, Concurrent Programming