Hello,
My static analysis tool reports a possible ABBA deadlock in the ocfs2
module in Linux 5.10:
ocfs2_dio_end_io_write()
down_write(&oi->ip_alloc_sem); --> Line 2322 (Lock A)
ocfs2_del_inode_from_orphan()
inode_lock(orphan_dir_inode); --> Line 2701 (Lock B)
ocfs2_move_extents()
inode_lock(inode); --> Line 916 (Lock B)
down_write(&OCFS2_I(inode)->ip_alloc_sem); --> Line 936 (Lock A)
When ocfs2_dio_end_io_write() and ocfs2_move_extents() are concurrently
executed, the deadlocks can occur.
I am not quite sure whether this possible deadlock is real and how to
fix it if it is real.
Any feedback would be appreciated, thanks :)
Reported-by: TOTE Robot <[email protected]>
Best wishes,
Jia-Ju Bai
Hi,
Thanks for the report. But I don't think this is a possible deadlock in
practice. Please see my comments below.
On 12/10/21 4:38 PM, Jia-Ju Bai wrote:
> Hello,
>
> My static analysis tool reports a possible ABBA deadlock in the ocfs2 module in Linux 5.10:
>
> ocfs2_dio_end_io_write()
> down_write(&oi->ip_alloc_sem); --> Line 2322 (Lock A)
> ocfs2_del_inode_from_orphan()
> inode_lock(orphan_dir_inode); --> Line 2701 (Lock B)
Just as the variable name indicates, it is orphan dir inode, which is a
system file (something hidden inside the filesystem) and not visible for
end user.
>
> ocfs2_move_extents()
> inode_lock(inode); --> Line 916 (Lock B)
This is called from a ioctl from user, it a normal file inode.
Thanks,
Joseph
> down_write(&OCFS2_I(inode)->ip_alloc_sem); --> Line 936 (Lock A)
>
> When ocfs2_dio_end_io_write() and ocfs2_move_extents() are concurrently executed, the deadlocks can occur.
>
> I am not quite sure whether this possible deadlock is real and how to fix it if it is real.
> Any feedback would be appreciated, thanks :)
>
> Reported-by: TOTE Robot <[email protected]>
>
>
> Best wishes,
> Jia-Ju Bai
On 2021/12/14 15:08, Joseph Qi wrote:
> Hi,
>
> Thanks for the report. But I don't think this is a possible deadlock in
> practice. Please see my comments below.
>
> On 12/10/21 4:38 PM, Jia-Ju Bai wrote:
>> Hello,
>>
>> My static analysis tool reports a possible ABBA deadlock in the ocfs2 module in Linux 5.10:
>>
>> ocfs2_dio_end_io_write()
>> down_write(&oi->ip_alloc_sem); --> Line 2322 (Lock A)
>> ocfs2_del_inode_from_orphan()
>> inode_lock(orphan_dir_inode); --> Line 2701 (Lock B)
> Just as the variable name indicates, it is orphan dir inode, which is a
> system file (something hidden inside the filesystem) and not visible for
> end user.
>
>> ocfs2_move_extents()
>> inode_lock(inode); --> Line 916 (Lock B)
> This is called from a ioctl from user, it a normal file inode.
Okay, thanks for the explanation :)
Best wishes,
Jia-Ju Bai