> On May 23, 2023, at 2:02 AM, Dai Ngo <dai.ngo@xxxxxxxxxx> wrote: > > > On 5/22/23 7:43 PM, Chuck Lever wrote: >> On Mon, May 22, 2023 at 04:52:39PM -0700, Dai Ngo wrote: >>> If the GETATTR request on a file that has write delegation in effect >>> and the request attributes include the change info and size attribute >>> then the write delegation is recalled and NFS4ERR_DELAY is returned >>> for the GETATTR. >> Isn't this yet another case where the server should send the >> CB_RECALL and wait for it briefly, before resorting to >> NFS4ERR_DELAY? > > Think about this more, I don't think we even need to recall the > delegation at all. The Linux client does not flush the dirty file > data before returning the delegation so the GETATTR still get stale > attributes at the server. And the spec is not explicitly requires > the delegation to be recalled. I'm trying to get some feedback from other server implementers about this. I'm willing to consider dropping this patch for now. -- Chuck Lever
