Many functions require the allocation of multiple resources. Failing and returning somewhere in the middle of this function without freeing all of the allocated resources could produce a memory leak. It is a common error to forget to free one (or all) of the resources in this manner, so a goto chain is the simplest and cleanest way to organize exits while preserving the order of freed resources.
Noncompliant Code Example
In this noncompliant example, exit code is written for every instance in which the function can terminate prematurely. Notice how failing to close fin2 produces a resource leak, leaving an open file descriptor.
Please note that these examples assume errno_t and NOERR to be defined, as recommended in DCL09-C. Declare functions that return errno with a return type of errno_t. An equivalent compatible example would define errno_t as an int and NOERR as zero.
| Code Block | ||||
|---|---|---|---|---|
| ||||
typedef struct object { /* Generic struct: contents don't matter */
int propertyA, propertyB, propertyC;
} object_t;
errno_t do_something(void){
FILE *fin1, *fin2;
object_t *obj;
errno_t ret_val;
fin1 = fopen("some_file", "r");
if (fin1 == NULL) {
return errno;
}
fin2 = fopen("some_other_file", "r");
if (fin2 == NULL) {
fclose(fin1);
return errno;
}
obj = malloc(sizeof(object_t));
if (obj == NULL) {
ret_val = errno;
fclose(fin1);
return ret_val; /* Forgot to close fin2!! */
}
/* ... More code ... */
fclose(fin1);
fclose(fin2);
free(obj);
return NOERR;
}
|
This is just a small example; in much larger examples, errors like this are even harder to detect.
Compliant Solution
In this revised version, a goto chain replaces each individual return segment. If no error occurs, control flow falls through to the SUCCESS label, releases all of the resources, and returns NOERR. If an error occurs, the return value is set to errno, control flow jumps to the proper failure label, and the appropriate resources are released before returning.
| Code Block | ||||
|---|---|---|---|---|
| ||||
/* ... Assume the same struct used previously ... */
errno_t do_something(void) {
FILE *fin1, *fin2;
object_t *obj;
errno_t ret_val = NOERR; /* Initially assume a successful return value */
fin1 = fopen("some_file", "r");
if (fin == NULL) {
ret_val = errno;
goto FAIL_FIN1;
}
fin2 = fopen("some_other_file", "r");
if (fin2 == NULL) {
ret_val = errno;
goto FAIL_FIN2;
}
obj = malloc(sizeof(object_t));
if (obj == NULL) {
ret_val = errno;
goto FAIL_OBJ;
}
/* ... More code ... */
SUCCESS: /* Clean up everything */
free(obj);
FAIL_OBJ: /* Otherwise, close only the resources we opened */
fclose(fin2);
FAIL_FIN2:
fclose(fin1);
FAIL_FIN1:
return ret_val;
}
|
This method is beneficial because the code is cleaner, and the programmer does not need to rewrite similar code upon every function error.
Compliant Solution (copy_process() from Linux kernel)
Some effective examples of goto chains are quite large. This compliant solution is an excerpt from the Linux kernel. This is the copy_process function from kernel/fork.c from version 2.6.29 of the kernel.
The function uses 17 goto labels (not all displayed here) to perform cleanup code should any internal function yield an error code. If no errors occur, the program returns a pointer to the new process p. If any error occurs, the program diverts control to a particular goto label, which performs cleanup for sections of the function that have currently been successfully executed but not for sections that have not yet been executed. Consequently, only resources that were successfully opened are actually closed.
All comments in this excerpt were added to indicate additional code in the kernel not displayed here.
| Code Block | ||||
|---|---|---|---|---|
| ||||
static struct task_struct *copy_process(unsigned long clone_flags,
unsigned long stack_start,
struct pt_regs *regs,
unsigned long stack_size,
int __user *child_tidptr,
struct pid *pid,
int trace)
{
int retval;
struct task_struct *p;
int cgroup_callbacks_done = 0;
if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS))
return ERR_PTR(-EINVAL);
/* ... */
retval = security_task_create(clone_flags);
if (retval)
goto fork_out;
retval = -ENOMEM;
p = dup_task_struct(current);
if (!p)
goto fork_out;
/* ... */
/* Copy all the process information */
if ((retval = copy_semundo(clone_flags, p)))
goto bad_fork_cleanup_audit;
if ((retval = copy_files(clone_flags, p)))
goto bad_fork_cleanup_semundo;
if ((retval = copy_fs(clone_flags, p)))
goto bad_fork_cleanup_files;
if ((retval = copy_sighand(clone_flags, p)))
goto bad_fork_cleanup_fs;
if ((retval = copy_signal(clone_flags, p)))
goto bad_fork_cleanup_sighand;
if ((retval = copy_mm(clone_flags, p)))
goto bad_fork_cleanup_signal;
if ((retval = copy_namespaces(clone_flags, p)))
goto bad_fork_cleanup_mm;
if ((retval = copy_io(clone_flags, p)))
goto bad_fork_cleanup_namespaces;
retval = copy_thread(0, clone_flags, stack_start, stack_size, p, regs);
if (retval)
goto bad_fork_cleanup_io;
/* ... */
return p;
/* ... Cleanup code starts here ... */
bad_fork_cleanup_io:
put_io_context(p->io_context);
bad_fork_cleanup_namespaces:
exit_task_namespaces(p);
bad_fork_cleanup_mm:
if (p->mm)
mmput(p->mm);
bad_fork_cleanup_signal:
cleanup_signal(p);
bad_fork_cleanup_sighand:
__cleanup_sighand(p->sighand);
bad_fork_cleanup_fs:
exit_fs(p); /* Blocking */
bad_fork_cleanup_files:
exit_files(p); /* Blocking */
bad_fork_cleanup_semundo:
exit_sem(p);
bad_fork_cleanup_audit:
audit_free(p);
/* ... More cleanup code ... */
fork_out:
return ERR_PTR(retval);
}
|
Risk Assessment
Failure to free allocated memory or close opened files results in a memory leak and possibly unexpected results.
Recommendation | Severity | Likelihood | Remediation Cost | Priority | Level |
|---|---|---|---|---|---|
MEM12-C | Low | Probable | Medium | P4 | L3 |
Bibliography
| Linux Kernel Sourcecode (v2.6.xx) | 2.6.29, kernel/fork.c, the copy_process() Function |
| [Seacord 2013] | Chapter 4, "Dynamic Memory Management" |