[PATCH] pid: sys_wait... fixes (v2)

[Eric W. Biederman]

Oleg Nesterov <oleg at tv-sign.ru> writes:

> On 12/06, Eric W. Biederman wrote:
>> 
>> +static struct pid *task_pid_type(struct task_struct *task, enum pid_type
> type)
>> +{
>> +	struct pid *pid = NULL;
>> +	if (type == PIDTYPE_PID)
>> +		pid = task->pids[type].pid;
>> +	else if (type < PIDTYPE_MAX)
>> +		pid = task->group_leader->pids[type].pid;
>> +	return pid;
>> +}
>
> Looks like a useful helper regardless.

Sort of.  We don't have any place else that could use it yet.

>> +static int eligible_child(enum pid_type type, struct pid *pid, int options,
>> +			  struct task_struct *p)
>>  {
>>  	int err;
>> -	struct pid_namespace *ns;
>>  
>> -	ns = current->nsproxy->pid_ns;
>> -	if (pid > 0) {
>> -		if (task_pid_nr_ns(p, ns) != pid)
>> -			return 0;
>> -	} else if (!pid) {
>> -		if (task_pgrp_nr_ns(p, ns) != task_pgrp_vnr(current))
>> -			return 0;
>> -	} else if (pid != -1) {
>> -		if (task_pgrp_nr_ns(p, ns) != -pid)
>> +	if (type < PIDTYPE_MAX) {
>> +		if (task_pid_type(p, type) != pid)
>>  			return 0;
>>  	}
>
> Minor, minor nit.
>
> If type == PIDTYPE_MAX, pid must be == NULL. but task_pid_type(PIDTYPE_PID)
> also returns NULL. So we can just do
>
> 	if (task_pid_type(p, type) != pid)
> 		return 0;
>
> Not that it really makes a difference though.

No.

Your comments are good but I figure this patch is now correct and thus
good enough for me.

Eric