On Wed, Oct 31, 2018 at 02:54:51PM -0400, Olga Kornievskaia wrote:
> On Wed, Oct 31, 2018 at 2:40 PM J. Bruce Fields <bfields@xxxxxxxxxxxx> wrote:
> >
> > Just to pick one thing that I don't understand yet:
> >
> > On Fri, Oct 19, 2018 at 11:28:56AM -0400, Olga Kornievskaia wrote:
> > > +EXPORT_SYMBOL_GPL(nfs42_ssc_open);
> > > +void nfs42_ssc_close(struct file *filep)
> > > +{
> > > + struct nfs_open_context *ctx = nfs_file_open_context(filep);
> > > +
> > > + ctx->state->flags = 0;
> >
> > Why is this needed?
>
> This is needed so that CLOSE isn't going on the wire but closed internally.
>
> > Also, given the name and the pairing with nfs42_ssc_open(), would it be
> > more logical for it to do the fput() as well?
>
> I'd like to keep that fput() in the nfsd to make it consistent with
> the "intra". There are fput()s for intra but intra doesn't call into
> nfs42_ssc_close().
OK, I think.
I'll need an ACK from Trond and/or Anna on this one in any case.
--b.
