On Wed, Feb 7, 2018 at 4:35 PM, NeilBrown <neilb@xxxxxxxx> wrote:
>> On Wed, Feb 7, 2018 at 8:58 AM, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
>> [...]
>>>
>>>> Lockdep reports this issue when GFP_FS is infact set, and we enter
>>>> this path and acquire the lock. So lockdep seems to be doing the right
>>>> thing however by design it is reporting a false-positive.
>>>
>>> So I'm not seeing how its a false positive. fs/inode.c sets a different
>>> lock class per filesystem type. So recursing on an i_mutex within a
>>> filesystem does sound dodgy.
>>
>> But directory inodes and file inodes in the same filesystem share the
>> same lock class right?
>
> Not since v2.6.24
> Commit: 14358e6ddaed ("lockdep: annotate dir vs file i_mutex")
>
> You were using 4.9.60. so they should be separate....
>
> Maybe shmem_get_inode() needs to call unlock_new_inode() or just
> lockdep_annotate_inode_mutex_key() after inode_init_owner().
>
> Maybe inode_init_owner() should call lockdep_annotate_inode_mutex_key()
> directly.
Thanks for the ideas! I will test lockdep_annotate_inode_mutex_key
after inode_init_owner in shmem and let you know if the issue goes
away. It seems hugetlbfs does this too (I think for similar reasons).
- Joel
