Re: [PATCH] quota: propagate error from __dquot_initialize

[Chao Yu]

On 2017/11/21 21:18, Jan Kara wrote:
> On Fri 17-11-17 10:07:53, Chao Yu wrote:
>> In commit 6184fc0b8dd7 ("quota: Propagate error from ->acquire_dquot()"),
>> we have propagated error from __dquot_initialize to caller, but we forgot
>> to handle such error in add_dquot_ref(), so, currently, during quota
>> accounting information initialization flow, if we failed for some of
>> inodes, we just ignore such error, and do account for others, which is
>> not a good implementation.
>>
>> In this patch, we choose to let user be aware of such error, so after
>> turning on quota successfully, we can make sure all inodes disk usage
>> can be accounted, which will be more reasonable.
> 
> Thanks for the patch! One comment below:
> 
>> @@ -2371,10 +2377,18 @@ static int vfs_load_quota_inode(struct inode *inode, int type, int format_id,
>>  	dqopt->flags |= dquot_state_flag(flags, type);
>>  	spin_unlock(&dq_state_lock);
>>  
>> -	add_dquot_ref(sb, type);
>> +	error = add_dquot_ref(sb, type);
>> +	if (error)
>> +		goto out_dquot_flags;
>>  
>>  	return 0;
>> -
>> +out_dquot_flags:
>> +	spin_lock(&dq_data_lock);
>> +	dqopt->info[type].dqi_flags &= ~DQF_SYS_FILE;
>> +	spin_unlock(&dq_data_lock);
>> +	spin_lock(&dq_state_lock);
>> +	dqopt->flags &= ~(dquot_state_flag(flags, type));
>> +	spin_unlock(&dq_state_lock);
>>  out_file_init:
>>  	dqopt->files[type] = NULL;
>>  	iput(inode);
> 
> This bail out path is not correct. You have to go through full quota off at
> this point (dquot_disable() function) as some inodes had already quotas
> initialized and can be using them...

Yes, you're right, have updated in v2, please help to check that.

Thanks,

> 
> 								Honza
>