Re: Huge race in lockd for async lock requests?

[Rob Gardner <rob.gardner@hp.com>]

J. Bruce Fields wrote:
>>>>> At 04:43 PM 5/19/2009, Rob Gardner wrote:
>>>>>       
>>>>>           
>>>>>> I've got a question about lockd in conjunction with a filesystem 
>>>>>> that provides its own (async) locking.
>>>>>>
>>>>>> After nlmsvc_lock() calls vfs_lock_file(), it seems to be that we 
>>>>>> might get the async callback (nlmsvc_grant_deferred) at any time. 
>>>>>> What's to stop it from arriving before we even put the block on 
>>>>>> the nlm_block list? If this happens, then nlmsvc_grant_deferred() 
>>>>>> will print "grant for unknown block" and then we'll wait forever 
>>>>>> for a grant that will never come.
>>>>>>     
>> dealing with a filesystem that provides its own locking functions via  
>> file->f_op->lock(). Such a filesystem might easily defer a non-blocking  
>> lock request and invoke the callback later. At least I don't know of any  
>> rule that says that it can't do this, and clearly the code expects this  
>> possibility:
>>
>>              case FILE_LOCK_DEFERRED:
>>                        if (wait)
>>                                break;
>>                        /* Filesystem lock operation is in progress
>>                           Add it to the queue waiting for callback */
>>                        ret = nlmsvc_defer_lock_rqst(rqstp, block);
>>
>>     
>> It looks to me like a bug in the server. The server must be able to deal  
>> with async filesystem callbacks happening at any time, however 
>> inconvenient.
>>     
>
> Absolutely, if that's possible then it's a server bug.
>
> --b.
>   

It's definitely possible for the async filesystem callback to occur at 
any time. I think at the very least, nlmsvc_lock() ought to put the 
block on the nlm_block list *before* calling vfs_lock_file(), and then 
remove it immediately if the lock is granted synchronously. I would like 
to develop and submit a patch for this, but I am currently working with 
a much older kernel and it will take some time before I get to work with 
newer bits.

Rob Gardner