MIME-Version: 1.0
References: <20190422163424.19402-1-jlayton@kernel.org> <20190422163424.19402-2-jlayton@kernel.org>
 <87wojl61s5.fsf@notabene.neil.brown.name> <20190424135854.GB20542@fieldses.org>
 <2b131f18-1df4-790c-a548-3155a0677c34@RedHat.com> <20190424154735.GC20542@fieldses.org>
In-Reply-To: <20190424154735.GC20542@fieldses.org>
From:   Pavel Shilovsky <piastryyy@gmail.com>
Date:   Wed, 24 Apr 2019 12:09:41 -0700
Message-ID: <CAKywueQBSwecUpr7CerLSzFT07F55sLC3vF=xFMV13uCn5_iOA@mail.gmail.com>
Subject: Re: [PATCH v2 1/2] nfsd: wake waiters blocked on file_lock before
 deleting it
To:     "J. Bruce Fields" <bfields@fieldses.org>
Cc:     Steve Dickson <SteveD@redhat.com>, NeilBrown <neilb@suse.com>,
        Jeff Layton <jlayton@kernel.org>, slawek1211@gmail.com,
        Linux NFS Mailing list <linux-nfs@vger.kernel.org>,
        linux-cifs <linux-cifs@vger.kernel.org>,
        Steve French <smfrench@gmail.com>,
        Ronnie Sahlberg <lsahlber@redhat.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-nfs-owner@vger.kernel.org
Precedence: bulk

Yes, I think there is a bug here, thanks!

If cinode->can_cache_brlcks is false we should return 1 but the code
will return 0 if coming from the "try_again" label. We also need to
call locks_delete_block unconditionally at the end of the function.

--
Best regards,
Pavel Shilovsky

=D1=81=D1=80, 24 =D0=B0=D0=BF=D1=80. 2019 =D0=B3. =D0=B2 08:48, J. Bruce Fi=
elds <bfields@fieldses.org>:
>
> On Wed, Apr 24, 2019 at 11:29:59AM -0400, Steve Dickson wrote:
> >
> >
> > On 4/24/19 9:58 AM, J. Bruce Fields wrote:
> > > Steve, see Neil's comment, is there a cifs bug here?
> > Looking into it...
>
> I was thinking Steve French, though I'm sure he wouldn't mind if you
> fixed cifs bugs.  Too many Steves!
>
> --b.
>
> >
> > steved.
> > >
> > > --b.
> > >
> > > On Tue, Apr 23, 2019 at 09:47:06AM +1000, NeilBrown wrote:
> > >> On Mon, Apr 22 2019, Jeff Layton wrote:
> > >>
> > >>> After a blocked nfsd file_lock request is deleted, knfsd will send =
a
> > >>> callback to the client and then free the request. Commit 16306a61d3=
b7
> > >>> ("fs/locks: always delete_block after waiting.") changed it such th=
at
> > >>> locks_delete_block is always called on a request after it is awoken=
,
> > >>> but that patch missed fixing up blocked nfsd request handling.
> > >>>
> > >>> Call locks_delete_block on the block to wake up any locks still blo=
cked
> > >>> on the nfsd lock request before freeing it. Some of its callers alr=
eady
> > >>> do this however, so just remove those calls.
> > >>>
> > >>> URL: https://bugzilla.kernel.org/show_bug.cgi?id=3D203363
> > >>> Fixes: 16306a61d3b7 ("fs/locks: always delete_block after waiting."=
)
> > >>> Reported-by: Slawomir Pryczek <slawek1211@gmail.com>
> > >>> Cc: Neil Brown <neilb@suse.com>
> > >>> Cc: stable@vger.kernel.org
> > >>> Signed-off-by: Jeff Layton <jlayton@kernel.org>
> > >>> ---
> > >>>  fs/nfsd/nfs4state.c | 3 +--
> > >>>  1 file changed, 1 insertion(+), 2 deletions(-)
> > >>>
> > >>> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> > >>> index 6a45fb00c5fc..e87e15df2044 100644
> > >>> --- a/fs/nfsd/nfs4state.c
> > >>> +++ b/fs/nfsd/nfs4state.c
> > >>> @@ -265,6 +265,7 @@ find_or_allocate_block(struct nfs4_lockowner *l=
o, struct knfsd_fh *fh,
> > >>>  static void
> > >>>  free_blocked_lock(struct nfsd4_blocked_lock *nbl)
> > >>>  {
> > >>> + locks_delete_block(&nbl->nbl_lock);
> > >>>   locks_release_private(&nbl->nbl_lock);
> > >>
> > >> Thanks for tracking this down.
> > >>
> > >> An implication of this bug and fix is that we need to be particularl=
y
> > >> careful to make sure locks_delete_block() is called on all relevant
> > >> paths.
> > >> Can we make that easier?  My first thought was to include the call i=
n
> > >> locks_release_private, but lockd calls the two quite separately and =
it
> > >> certainly seems appropriate that locks_delete_block should be called
> > >> asap, but locks_release_private() can be delayed.
> > >>
> > >> Also cifs calls locks_delete_block, but never calls
> > >> locks_release_private, so it wouldn't help there.
> > >>
> > >> Looking at cifs, I think there is a call missing there too.
> > >> cifs_posix_lock_set() *doesn't* always call locks_delete_block() aft=
er
> > >> waiting.  In particular, if ->can_cache_brlcks becomes true while
> > >> waiting then I don't think the behaviour is right.... though I'm not
> > >> sure it is right for other reasons.  It looks like the return value
> > >> should be 1 in that case, but it'll be zero.
> > >>
> > >> But back to my question about making it easier, move the BUG_ON()
> > >> calls from locks_free_lock() into locks_release_private().
> > >>
> > >> ??
> > >>
> > >> Thanks,
> > >> NeilBrown
> > >>
> > >>
> > >>>   kfree(nbl);
> > >>>  }
> > >>> @@ -293,7 +294,6 @@ remove_blocked_locks(struct nfs4_lockowner *lo)
> > >>>           nbl =3D list_first_entry(&reaplist, struct nfsd4_blocked_=
lock,
> > >>>                                   nbl_lru);
> > >>>           list_del_init(&nbl->nbl_lru);
> > >>> -         locks_delete_block(&nbl->nbl_lock);
> > >>>           free_blocked_lock(nbl);
> > >>>   }
> > >>>  }
> > >>> @@ -4863,7 +4863,6 @@ nfs4_laundromat(struct nfsd_net *nn)
> > >>>           nbl =3D list_first_entry(&reaplist,
> > >>>                                   struct nfsd4_blocked_lock, nbl_lr=
u);
> > >>>           list_del_init(&nbl->nbl_lru);
> > >>> -         locks_delete_block(&nbl->nbl_lock);
> > >>>           free_blocked_lock(nbl);
> > >>>   }
> > >>>  out:
> > >>> --
> > >>> 2.20.1
> > >
> > >