Re: + proc-export-a-processes-resource-limits-via-proc-pid.patch added to -mm tree

[Oleg Nesterov]

On 08/18, Neil Horman wrote:
>
> +static int proc_pid_limits(struct task_struct *task, char *buffer)
> +{
> +	unsigned int i;
> +	int count = 0;
> +	unsigned long flags;
> +	char *bufptr = buffer;
> +
> +	struct rlimit rlim[RLIM_NLIMITS];
> +
> +	rcu_read_lock();
> +	lock_task_sighand(task,&flags);
> +	if (task->signal == NULL){
> +		unlock_task_sighand(task, &flags);
> +		rcu_read_unlock();
> +		return 0;
> +	}
> +	memcpy(rlim, task->signal->rlim, sizeof(struct rlimit) * RLIM_NLIMITS);
> +	unlock_task_sighand(task, &flags);
> +	rcu_read_unlock();

No, no. If lock_task_sighand() fails, ->signal _should be_ == NULL, but the
"if (task->signal == NULL)" check is not reliable, we don't have any barriers
to serialize with __exit_signal().

More importantly, it is very wrong to do unlock_task_sighand() in that case,
this means NULL pointer dereference.

What we need is:

	rcu_read_lock();
	if (!lock_task_sighand(task, &flags)) {
		rcu_read_unlock();
		return 0;
	}
	memcpy(rlim, task->signal->rlim, sizeof(rlim));
	unlock_task_sighand(task, &flags);
	rcu_read_unlock();

No need to check ->signal != NULL if lock_task_sighand() succeeds.

Oleg.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/