[PATCH 0/4] Address several issues with TEST_STATEID

[Chuck Lever <chuck.lever@oracle.com>]

Hi-

While testing the recently introduced PURGE_STATE logic, I found some
issues with the error handling logic under nfs41_open_expired().

After fixing these issues, I am able to get Connectathon basic,
general, and special tests to pass with a server rigged to assert
SEQ4_STATUS_ADMIN_STATE_REVOKED at regular intervals.  The
Connectathon lock tests encounter a deadlock, which I've already
reported, that appears to pre-date PURGE_STATE.

---

Chuck Lever (4):
      NFS: Refactor nfs41_check_expired_stateid()
      NFS: state reclaim clears OPEN state
      NFS: Properly sort TEST_STATEID results
      NFS: Clean up TEST_STATEID and FREE_STATEID proc error reporting


 fs/nfs/nfs4proc.c |  144 ++++++++++++++++++++++++++++++++++++++++++-----------
 1 files changed, 115 insertions(+), 29 deletions(-)

-- 
Chuck Lever