On Fri, 9 Aug 2019, Dave Chinner wrote: > And, FWIW, there's an argument to be made here that the underlying > bug is dm_bufio_shrink_scan() blocking kswapd by waiting on IO > completions while holding a mutex that other IO-level reclaim > contexts require to make progress. > > Cheers, > > Dave. The IO-level reclaim contexts should use GFP_NOIO. If the dm-bufio shrinker is called with GFP_NOIO, it cannot be blocked by kswapd, because: * it uses trylock to acquire the mutex * it doesn't attempt to free buffers that are dirty or under I/O (see __try_evict_buffer) I think that this logic is sufficient to prevent the deadlock. Mikulas
