On Wed, 2022-02-02 at 04:09 +0000, inoguchi.yuki@xxxxxxxxxxx wrote:
> > I don't think that case adds anything interesting. When the file
> > is
> > closed, the lock is dropped. If there were any writes without a
> > delegation, then the changeid isn't a reliable indication that no
> > other
> > client wrote. So the cache must be dropped.
>
> I've understood it.
>
> I've made the patch based on your idea. It invalidates the cache
> after a client without write-delegation sends CLOSE call.
> My Open MPI test program confirmed that this fix resolves the
> problem.
>
> The patch is as follows. What do you think?
> -----
> diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
> index b18f31b2c9e7..67021786034d 100644
> --- a/fs/nfs/nfs4proc.c
> +++ b/fs/nfs/nfs4proc.c
> @@ -3711,7 +3711,8 @@ static const struct rpc_call_ops nfs4_close_ops
> = {
> */
> int nfs4_do_close(struct nfs4_state *state, gfp_t gfp_mask, int
> wait)
> {
> - struct nfs_server *server = NFS_SERVER(state->inode);
> + struct inode *inode = state->inode;
> + struct nfs_server *server = NFS_SERVER(inode);
> struct nfs_seqid *(*alloc_seqid)(struct nfs_seqid_counter *,
> gfp_t);
> struct nfs4_closedata *calldata;
> struct nfs4_state_owner *sp = state->owner;
> @@ -3773,6 +3774,15 @@ int nfs4_do_close(struct nfs4_state *state,
> gfp_t gfp_mask, int wait)
> status = 0;
> if (wait)
> status = rpc_wait_for_completion_task(task);
> +
> + if (status >= 0 && !NFS_PROTO(inode)->have_delegation(inode,
> FMODE_WRITE)) {
> + nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR
> + | NFS_INO_INVALID_DATA
> + | NFS_INO_INVALID_ACCESS
> + | NFS_INO_INVALID_ACL
> + | NFS_INO_REVAL_PAGECACHE);
> + }
> +
> rpc_put_task(task);
> return status;
> out_free_calldata:
> -----
NACK. That is completely unacceptable.
--
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trond.myklebust@xxxxxxxxxxxxxxx
