1 From foo@baz Wed Dec 6 17:39:55 CET 2017
2 From: Benjamin Coddington <bcodding@redhat.com>
3 Date: Thu, 5 Jan 2017 10:20:16 -0500
4 Subject: nfs: Don't take a reference on fl->fl_file for LOCK operation
6 From: Benjamin Coddington <bcodding@redhat.com>
9 [ Upstream commit 4b09ec4b14a168bf2c687e1f598140c3c11e9222 ]
11 I have reports of a crash that look like __fput() was called twice for
12 a NFSv4.0 file. It seems possible that the state manager could try to
13 reclaim a lock and take a reference on the fl->fl_file at the same time the
14 file is being released if, during the close(), a signal interrupts the wait
15 for outstanding IO while removing locks which then skips the removal
18 Since 83bfff23e9ed ("nfs4: have do_vfs_lock take an inode pointer") has
19 removed the need to traverse fl->fl_file->f_inode in nfs4_lock_done(),
20 taking that reference is no longer necessary.
22 Signed-off-by: Benjamin Coddington <bcodding@redhat.com>
23 Reviewed-by: Jeff Layton <jlayton@redhat.com>
24 Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
25 Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
26 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
28 fs/nfs/nfs4proc.c | 3 ---
29 1 file changed, 3 deletions(-)
31 --- a/fs/nfs/nfs4proc.c
32 +++ b/fs/nfs/nfs4proc.c
35 #include <linux/delay.h>
36 #include <linux/errno.h>
37 -#include <linux/file.h>
38 #include <linux/string.h>
39 #include <linux/ratelimit.h>
40 #include <linux/printk.h>
41 @@ -6006,7 +6005,6 @@ static struct nfs4_lockdata *nfs4_alloc_
43 atomic_inc(&lsp->ls_count);
44 p->ctx = get_nfs_open_context(ctx);
45 - get_file(fl->fl_file);
46 memcpy(&p->fl, fl, sizeof(p->fl));
49 @@ -6119,7 +6117,6 @@ static void nfs4_lock_release(void *call
50 nfs_free_seqid(data->arg.lock_seqid);
51 nfs4_put_lock_state(data->lsp);
52 put_nfs_open_context(data->ctx);
53 - fput(data->fl.fl_file);
55 dprintk("%s: done!\n", __func__);