Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <8a239157-de28-44ae-edef-336092c34a15@linux.alibaba.com>
Date:   Fri, 9 Sep 2022 18:28:04 +0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:91.0)
 Gecko/20100101 Thunderbird/91.13.0
Subject: Re: [PATCH V2 4/5] erofs: remove duplicated unregister_cookie
Content-Language: en-US
From:   JeffleXu <jefflexu@linux.alibaba.com>
To:     Jia Zhu <zhujia.zj@bytedance.com>, linux-erofs@lists.ozlabs.org,
        xiang@kernel.org, chao@kernel.org
Cc:     linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
        yinxin.x@bytedance.com, huyue2@coolpad.com
References: <20220902105305.79687-1-zhujia.zj@bytedance.com>
 <20220902105305.79687-5-zhujia.zj@bytedance.com>
 <3f75d266-7ccd-be6d-657c-fe0633b25687@linux.alibaba.com>
In-Reply-To: <3f75d266-7ccd-be6d-657c-fe0633b25687@linux.alibaba.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 9/9/22 5:55 PM, JeffleXu wrote:
> 
> 
> On 9/2/22 6:53 PM, Jia Zhu wrote:
>> In erofs umount scenario, erofs_fscache_unregister_cookie() is called
>> twice in kill_sb() and put_super().
>>
>> It works for original semantics, cause 'ctx' will be set to NULL in
>> put_super() and will not be unregister again in kill_sb().
>> However, in shared domain scenario, we use refcount to maintain the
>> lifecycle of cookie. Unregister the cookie twice will cause it to be
>> released early.

Sorry, why can't we also set sbi->s_fscache to NULL after decrementing
the refcount in shared domain mode, to avoid the refcount being
decremented twice?

>>
>> For the above reasons, this patch removes duplicate unregister_cookie
>> and move fscache_unregister_* before shotdown_super() to prevent busy
>> inode(ctx->inode) when umount.
>>
>> Signed-off-by: Jia Zhu <zhujia.zj@bytedance.com>
>> ---
>>  fs/erofs/super.c | 16 ++++++++--------
>>  1 file changed, 8 insertions(+), 8 deletions(-)
>>
>> diff --git a/fs/erofs/super.c b/fs/erofs/super.c
>> index 69de1731f454..667a78f0ee70 100644
>> --- a/fs/erofs/super.c
>> +++ b/fs/erofs/super.c
>> @@ -919,19 +919,20 @@ static void erofs_kill_sb(struct super_block *sb)
>>  		kill_litter_super(sb);
>>  		return;
>>  	}
>> -	if (erofs_is_fscache_mode(sb))
>> -		generic_shutdown_super(sb);
>> -	else
>> -		kill_block_super(sb);
>> -
>>  	sbi = EROFS_SB(sb);
>>  	if (!sbi)
>>  		return;
>>  
>> +	if (erofs_is_fscache_mode(sb)) {
>> +		erofs_fscache_unregister_cookie(&sbi->s_fscache);
>> +		erofs_fscache_unregister_fs(sb);
>> +		generic_shutdown_super(sb);
> 
> Generally we can't do clean ups before generic_shutdown_super(), since
> generic_shutdown_super() may trigger IO, e.g. in sync_filesystem(),
> though it's not the case for erofs (read-only).
> 
> How about embedding erofs_fscache_unregister_cookie() into
> erofs_fscache_unregister_fs(), and thus we can check domain_id in
> erofs_fscache_unregister_fs()?
> 
>> +	} else {
>> +		kill_block_super(sb);
>> +	}
>> +
>>  	erofs_free_dev_context(sbi->devs);
>>  	fs_put_dax(sbi->dax_dev, NULL);
>> -	erofs_fscache_unregister_cookie(&sbi->s_fscache);
>> -	erofs_fscache_unregister_fs(sb);
>>  	kfree(sbi->opt.fsid);
>>  	kfree(sbi->opt.domain_id);
>>  	kfree(sbi);
>> @@ -951,7 +952,6 @@ static void erofs_put_super(struct super_block *sb)
>>  	iput(sbi->managed_cache);
>>  	sbi->managed_cache = NULL;
>>  #endif
>> -	erofs_fscache_unregister_cookie(&sbi->s_fscache);
>>  }
>>  
>>  struct file_system_type erofs_fs_type = {
> 

-- 
Thanks,
Jingbo