Re: [PATCH] cpuset: Add a missing unlock in cpuset_write_resmask()

[Li Zefan]

>> @@ -1561,8 +1561,10 @@ static int cpuset_write_resmask(struct cgroup *cgrp, struct cftype *cft,
>>  		return -ENODEV;
>>  
>>  	trialcs = alloc_trial_cpuset(cs);
>> -	if (!trialcs)
>> +	if (!trialcs) {
>> +		cgroup_unlock();
>>  		return -ENOMEM;
>> +	}
>>  
>>  	switch (cft->private) {
>>  	case FILE_CPULIST:
> 
> It would be better to avoid multiple returns - it leads to more
> maintainable code and often shorter code:
> 

I have no strong opinion on this.

> --- a/kernel/cpuset.c~cpuset-add-a-missing-unlock-in-cpuset_write_resmask-fix
> +++ a/kernel/cpuset.c
> @@ -1562,8 +1562,8 @@ static int cpuset_write_resmask(struct c
>  
>  	trialcs = alloc_trial_cpuset(cs);
>  	if (!trialcs) {
> -		cgroup_unlock();
> -		return -ENOMEM;
> +		retval = -ENOMEM;
> +		goto out;
>  	}
>  
>  	switch (cft->private) {
> @@ -1579,6 +1579,7 @@ static int cpuset_write_resmask(struct c
>  	}
>  
>  	free_trial_cpuset(trialcs);
> +out:
>  	cgroup_unlock();
>  	return retval;
>  }
> _
> 
> also, alloc_trial_cpuset() is a fairly slow-looking function. 
> cpuset_write_resmask() could run alloc_trial_cpuset() before running
> cgroup_lock_live_group(), thereby reducing lock hold times.
> 

Nope. alloc_trial_cpuset() will read 'cs', so it must be protected by
the lock.
--
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/