Message-ID: <59351CB4.1010100@163.com>
Date: Mon, 05 Jun 2017 16:56:20 +0800
From: Jia-Ju Bai <baijiaju1990@163.com>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:11.0) Gecko/20120410 Thunderbird/11.0.1
MIME-Version: 1.0
To: Shan Hai <shan.hai@oracle.com>
CC: darrick.wong@oracle.com, linux-xfs@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH] fs: xfs: Fix a lock-twice and sleep-in-atomic bug in
 xfs_iget
References: <1496650649-2296-1-git-send-email-baijiaju1990@163.com> <43a9b80f-fd7d-2df3-f726-207742c7924f@oracle.com>
In-Reply-To: <43a9b80f-fd7d-2df3-f726-207742c7924f@oracle.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1344
Lines: 45

On 06/05/2017 04:32 PM, Shan Hai wrote:
>
>
> On 2017年06月05日 16:17, Jia-Ju Bai wrote:
>> The driver may sleep under a read rcu lock, and function call path is:
>> xfs_iget (acquire the lock by rcu_read_lock)
>>    "goto out_error_or_again" after xfs_iget_cache_hit
>>      delay
>>        schedule_timeout_uninterruptible --> may sleep
>> Meanwhile, the rcu_read_lock will be called twice in this situation.
>>
>> To fix it, the lock is released before "goto".
>>
>> Signed-off-by: Jia-Ju Bai <baijiaju1990@163.com>
>> ---
>>   fs/xfs/xfs_icache.c |    4 +++-
>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c
>> index f61c84f8..c2a4722 100644
>> --- a/fs/xfs/xfs_icache.c
>> +++ b/fs/xfs/xfs_icache.c
>> @@ -600,8 +600,10 @@ struct xfs_inode *
>>         if (ip) {
>>           error = xfs_iget_cache_hit(pag, ip, ino, flags, lock_flags);
>> -        if (error)
>> +        if (error) {
>> +            rcu_read_unlock();
>
> Seems you are going to double unlock by doing this, since it is 
> unlocked in the xfs_iget_cache_hit.
>
> Thanks
> Shan Hai
>
>>               goto out_error_or_again;
>> +        }
>>       } else {
>>           rcu_read_unlock();
>>           XFS_STATS_INC(mp, xs_ig_missed);
>
I think you are right. Please ignore my patch.

Thanks,
Jia-Ju Bai