NFSv4.x recover from pre-mature loss of openstateid
authorOlga Kornievskaia <kolga@netapp.com>
Wed, 18 Dec 2019 21:50:42 +0000 (16:50 -0500)
committerAnna Schumaker <Anna.Schumaker@Netapp.com>
Wed, 15 Jan 2020 15:54:33 +0000 (10:54 -0500)
commitd826e5b827641ae1bebb33d23a774f4e9bb8e94f
treea767e974f0c3098d8fec2a6b623cbba258a0307d
parent62a1573fcf844a559a79bec2eafc4309f2be5c5b
NFSv4.x recover from pre-mature loss of openstateid

Ever since the commit 0e0cb35b417f, it's possible to lose an open stateid
while retrying a CLOSE due to ERR_OLD_STATEID. Once that happens,
operations that require openstateid fail with EAGAIN which is propagated
to the application then tests like generic/446 and generic/168 fail with
"Resource temporarily unavailable".

Instead of returning this error, initiate state recovery when possible to
recover the open stateid and then try calling nfs4_select_rw_stateid()
again.

Fixes: 0e0cb35b417f ("NFSv4: Handle NFS4ERR_OLD_STATEID in CLOSE/OPEN_DOWNGRADE")
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
fs/nfs/nfs42proc.c
fs/nfs/nfs4proc.c
fs/nfs/pnfs.c