Re: [BUG] fs: btrfs: several possible ABBA deadlocks

[Nikolay Borisov <nborisov@suse.com>]

On 29.11.21 г. 2:34, Josef Bacik wrote:
> On Sat, Nov 27, 2021 at 04:23:37PM +0800, Jia-Ju Bai wrote:
>> Hello,
>>
>> My static analysis tool reports several possible ABBA deadlocks in the btrfs
>> module in Linux 5.10:
>>
>> # DEADLOCK 1:
>> __clear_extent_bit()
>>   spin_lock(&tree->lock); --> Line 733 (Lock A)
>>   split_state()
>>     btrfs_split_delalloc_extent()
>>       spin_lock(&BTRFS_I(inode)->lock); --> Line 1870 (Lock B)
>>
>> btrfs_inode_safe_disk_i_size_write()
>>   spin_lock(&BTRFS_I(inode)->lock); --> Line 53 (Lock B)
>>   find_contiguous_extent_bit()
>>     spin_lock(&tree->lock); --> Line 1620 (Lock A)
>>
>> When __clear_extent_bit() and btrfs_inode_safe_disk_i_size_write() are
>> concurrently executed, the deadlock can occur.
>>
>> # DEADLOCK 2:
>> __set_extent_bit()
>>   spin_lock(&tree->lock); --> Line 995 (Lock A)
>>   set_state_bits()
>>     btrfs_set_delalloc_extent()
>>       spin_lock(&BTRFS_I(inode)->lock); --> Line 2007 or 2017 or 2029 (Lock
>> B)
>>
>> btrfs_inode_safe_disk_i_size_write()
>>   spin_lock(&BTRFS_I(inode)->lock); --> Line 53 (Lock B)
>>   find_contiguous_extent_bit()
>>     spin_lock(&tree->lock); --> Line 1620 (Lock A)
>>
>> When __set_extent_bit() and btrfs_inode_safe_disk_i_size_write() are
>> concurrently executed, the deadlock can occur.
>>
>> # DEADLOCK 3:
>> convert_extent_bit()
>>   spin_lock(&tree->lock); --> Line 1241 (Lock A)
>>   set_state_bits()
>>     btrfs_set_delalloc_extent()
>>       spin_lock(&BTRFS_I(inode)->lock); --> Line 2007 or 2017 or 2029 (Lock
>> B)
>>
>> btrfs_inode_safe_disk_i_size_write()
>>   spin_lock(&BTRFS_I(inode)->lock); --> Line 53 (Lock B)
>>   find_contiguous_extent_bit()
>>     spin_lock(&tree->lock); --> Line 1620 (Lock A)
>>
>> When convert_extent_bit() and btrfs_inode_safe_disk_i_size_write() are
>> concurrently executed, the deadlock can occur.
>>
>> I am not quite sure whether these possible deadlocks are real and how to fix
>> them if they are real.
>> Any feedback would be appreciated, thanks :)
>>
> 
> Hey Jia-Ju,
> 
> This is pretty good work, unfortunately it's wrong but it's in a subtle way that
> a tool wouldn't be able to catch.  The btrfs_inode_safe_disk_i_size_write()
> helper only messes with BTRFS_I(inode)->file_extent_tree, which is separate from
> the BTRFS_I(inode)->io_tree.  io_tree gets the btrfs_set_delalloc_extent() stuff
> called on it, but the file_extent_tree does not.  The file_extent_tree has
> inode->lock -> tree->lock as the locking order, whereas the file_extent_tree has
> inode->lock -> tree->lock as the locking order.  Thanks,

nit: did you mean to reverse tree->lock ->inode->lock for the
file_extent_tree?

> 
> Josef
>