On Fri, Mar 18, 2022 at 5:36 PM Trond Myklebust <trondmy@xxxxxxxxxxxxxxx> wrote:
>
> Isn't that stack trace showing a slab direct reclaim, and not a
> filesystem writeback situation?
>
> Does memalloc_nofs_save()/restore() really fix this problem? It seems
> to me that it cannot, particularly since knfsd is not a filesystem, and
> so does not ever handle writeback of dirty pages.
Ah you're right. An alternative would be delaying the destroy_mark,
which I am noticing now that, on mainline, the shrinker calls
nfsd_file_dispose_list_delayed, something missing from the version of
5.4 I was looking at.
To my reading 9542e6a643fc6 ("nfsd: Containerise filecache
laundrette") should have the effect of fixing this deadlock (and the
message does explicitly call out notifier deadlocks) - maybe something
to send towards stable?
>
>
> Cc: the linux-mm mailing list in search of answers to the above 2
> questions.
>
> --
> Trond Myklebust
> Linux NFS client maintainer, Hammerspace
> trond.myklebust@xxxxxxxxxxxxxxx
>
>
Attachment:
smime.p7s
Description: S/MIME Cryptographic Signature
