2017-06-28 01:46:53

by Trond Myklebust

[permalink] [raw]
Subject: [PATCH] NFSv4.1: Fix a race in nfs4_proc_layoutget

If the task calling layoutget is signalled, then it is possible for the
calls to nfs4_sequence_free_slot() and nfs4_layoutget_prepare() to race,
in which case we leak a slot.
The fix is to move the call to nfs4_sequence_free_slot() into the
nfs4_layoutget_release() so that it gets called at task teardown time.

Fixes: 2e80dbe7ac51 ("NFSv4.1: Close callback races for OPEN, LAYOUTGET...")
Cc: [email protected] # v4.8+
Signed-off-by: Trond Myklebust <[email protected]>
---
fs/nfs/nfs4proc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 4e43fd1270c4..dbfa18900e25 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -8417,6 +8417,7 @@ static void nfs4_layoutget_release(void *calldata)
size_t max_pages = max_response_pages(server);

dprintk("--> %s\n", __func__);
+ nfs4_sequence_free_slot(&lgp->res.seq_res);
nfs4_free_pages(lgp->args.layout.pages, max_pages);
pnfs_put_layout_hdr(NFS_I(inode)->layout);
put_nfs_open_context(lgp->args.ctx);
@@ -8491,7 +8492,6 @@ nfs4_proc_layoutget(struct nfs4_layoutget *lgp, long *timeout, gfp_t gfp_flags)
/* if layoutp->len is 0, nfs4_layoutget_prepare called rpc_exit */
if (status == 0 && lgp->res.layoutp->len)
lseg = pnfs_layout_process(lgp);
- nfs4_sequence_free_slot(&lgp->res.seq_res);
rpc_put_task(task);
dprintk("<-- %s status=%d\n", __func__, status);
if (status)
--
2.9.4