On Wed, 2024-02-07 at 14:58 -0500, Olga Kornievskaia wrote:
> On Wed, Feb 7, 2024 at 2:51 PM Olga Kornievskaia
> <olga.kornievskaia@xxxxxxxxx> wrote:
> >
> > On Wed, Feb 7, 2024 at 2:12 PM Trond Myklebust
> > <trondmy@xxxxxxxxxxxxxxx> wrote:
> > >
> > > On Wed, 2024-02-07 at 13:29 -0500, Olga Kornievskaia wrote:
> > > > From: Olga Kornievskaia <kolga@xxxxxxxxxx>
> > > >
> > > > Currently, if the server returns a partial layout, the client
> > > > gets
> > > > stuck asking for a layout indefinitely. Until we add support
> > > > for
> > > > partial layouts, treat partial layout as layout unavailable
> > > > error.
> > > >
> > > > Signed-off-by: Olga Kornievskaia <kolga@xxxxxxxxxx>
> > > > ---
> > > > fs/nfs/nfs4proc.c | 6 ++++++
> > > > 1 file changed, 6 insertions(+)
> > > >
> > > > diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
> > > > index dae4c1b6cc1c..108bc7f3e8c2 100644
> > > > --- a/fs/nfs/nfs4proc.c
> > > > +++ b/fs/nfs/nfs4proc.c
> > > > @@ -9790,6 +9790,12 @@ nfs4_proc_layoutget(struct
> > > > nfs4_layoutget
> > > > *lgp,
> > > > if (status != 0)
> > > > goto out;
> > > >
> > > > + /* Since client does not support partial pnfs layout,
> > > > then
> > > > treat
> > > > + * getting a partial layout as LAYOUTUNAVAILABLE error
> > > > + */
> > > > + if (lgp->args.range.length != lgp->res.range.length)
> > > > + task->tk_status = -NFS4ERR_LAYOUTUNAVAILABLE;
> > >
> > >
> > > I think this case is better handled by allowing the caller to set
> > > lgp-
> > > > args.minlength to an appropriate minimum value.
> >
> > I do not understand what this suggestion means. What I can think of
> > is
> > that the caller would set an appropriate minimum value and the code
> > here would check that the result is at least as large?
>
> A follow up question on a "minimum value". It seems that since the
> client would then need to set it to the same value as the "length"
> (ie
> whole file layout value), yes? So it shifts the responsibility to the
> server, disallowing it from returning a partial layout.
>
> > If so, can you explain why that's more desirable? Seems to me it'd
> > be
> > more lines for something that would be removed later?
>
What I'm saying is that the protocol expects the client to send the
minimal acceptable layout length as a separate argument from the
desired length. Right now, we set the minimal length in
pnfs_alloc_init_layoutget_args() to be the smaller of PAGE_SIZE or the
length of the I/O segment.
The expectation is that all the pnfs drivers should be able to handle
that.
If you're telling me that there are drivers that do not handle being
given a layout with the minimal length that is set in
pnfs_alloc_init_layoutget_args(), then we should give them control over
that value.
> > >
> > > > +
> > > > if (task->tk_status < 0) {
> > > > exception->retry = 1;
> > > > status = nfs4_layoutget_handle_exception(task,
> > > > lgp,
> > > > exception);
> > >
> > >
--
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trond.myklebust@xxxxxxxxxxxxxxx
