Re: [PATCH 1/2] nfsd4: fix cached replies to solo SEQUENCE compounds

["J. Bruce Fields" <bfields@redhat.com>]

On Fri, Oct 20, 2017 at 02:55:00PM -0400, Olga Kornievskaia wrote:
> On Fri, Oct 20, 2017 at 1:47 PM, J. Bruce Fields <bfields@redhat.com> wrote:
> > On Thu, Oct 19, 2017 at 05:19:40PM -0400, Olga Kornievskaia wrote:
> >> On Thu, Oct 19, 2017 at 5:04 PM, Olga Kornievskaia <aglo@umich.edu> wrote:
> >> > I got this crash upon receiving the SEQUENCE that reused the slot that
> >> > was used by the ctrl-c-ed COPY.
> >
> > Yeah, I was being inconsistent about what "NFSD4_SLOT_CACHED" meant--in
> > replay_matches_cache I'm assuming opcnt and cred are still stored, but I
> > wasn't storing them in nfsd4_store_cache_entry.  Not sure why that
> > didn't triger in my tests.
> >
> > Does this help?
> 
> Yes. No oops and reply is SEQ_FALSE_RETRY.

Thanks!--b.

> 
> > --b.
> >
> > diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> > index ace89bf45a0b..4b2107d11d5e 100644
> > --- a/fs/nfsd/nfs4state.c
> > +++ b/fs/nfsd/nfs4state.c
> > @@ -2296,14 +2296,15 @@ nfsd4_store_cache_entry(struct nfsd4_compoundres *resp)
> >
> >         slot->sl_flags |= NFSD4_SLOT_INITIALIZED;
> >         free_svc_cred(&slot->sl_cred);
> > +       copy_cred(&slot->sl_cred, &resp->rqstp->rq_cred);
> > +       slot->sl_opcnt = resp->opcnt;
> > +       slot->sl_status = resp->cstate.status;
> > +
> >         if (!nfsd4_cache_this(resp)) {
> >                 slot->sl_flags &= ~NFSD4_SLOT_CACHED;
> >                 return;
> >         }
> >         slot->sl_flags |= NFSD4_SLOT_CACHED;
> > -       copy_cred(&slot->sl_cred, &resp->rqstp->rq_cred);
> > -       slot->sl_opcnt = resp->opcnt;
> > -       slot->sl_status = resp->cstate.status;
> >
> >         base = resp->cstate.data_offset;
> >         slot->sl_datalen = buf->len - base;