On Wed, Apr 1, 2009 at 12:12 AM, J. Bruce Fields <bfields@xxxxxxxxxxxx> wrote:
> On Sat, Mar 28, 2009 at 11:33:11AM +0300, Benny Halevy wrote:
>> From: Andy Adamson <andros@xxxxxxxxxx>
>>
>> A session inactivity time compound (lease renewal) or a compound where the
>> sequence operation has sa_cachethis set to FALSE do not require any pages
>> to be held in the v4.1 DRC. This is because struct nfsd4_slot is already
>> caching the session information.
>>
>> Add logic to the nfs41 server to not cache response pages for solo sequence
>> responses.
>>
>> Return nfserr_replay_uncached_rep on the operation following the sequence
>> operation when sa_cachethis is FALSE.
>>
>> Signed-off-by: Andy Adamson <andros@xxxxxxxxxx>
>> Signed-off-by: Benny Halevy <bhalevy@xxxxxxxxxxx>
>> ---
>> fs/nfsd/nfs4proc.c | 34 +++++++++++++++++++++++++++++-
>> fs/nfsd/nfs4state.c | 47 ++++++++++++++++++++++++++++++++++++++-----
>> fs/nfsd/nfs4xdr.c | 5 ++-
>> include/linux/nfsd/state.h | 1 +
>> include/linux/nfsd/xdr4.h | 15 +++++++++++++-
>> 5 files changed, 91 insertions(+), 11 deletions(-)
>>
>> diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c
>> index bdbeb87..a273023 100644
>> --- a/fs/nfsd/nfs4proc.c
>> +++ b/fs/nfsd/nfs4proc.c
>> @@ -828,6 +828,34 @@ static struct nfsd4_operation nfsd4_ops[];
>> static const char *nfsd4_op_name(unsigned opnum);
>>
>> /*
>> + * This is a replay of a compound for which no cache entry pages
>> + * were used. Encode the sequence operation, and if cachethis is FALSE
>> + * encode the uncache rep error on the next operation.
>> + */
>> +static __be32
>> +nfsd4_enc_no_page_replay(struct nfsd4_compoundargs *args,
>> + struct nfsd4_compoundres *resp)
>> +{
>> + struct nfsd4_op *op;
>> +
>> + dprintk("--> %s resp->opcnt %d ce_cachethis %u \n", __func__,
>> + resp->opcnt, resp->cstate.slot->sl_cache_entry.ce_cachethis);
>> +
>> + /* Encode the replayed sequence operation */
>> + BUG_ON(resp->opcnt != 1);
>> + op = &args->ops[resp->opcnt - 1];
>> + nfsd4_encode_operation(resp, op);
>> +
>> + /*return nfserr_retry_uncached_rep in next operation. */
>> + if (resp->cstate.slot->sl_cache_entry.ce_cachethis == 0) {
>> + op = &args->ops[resp->opcnt++];
>> + op->status = nfserr_retry_uncached_rep;
>> + nfsd4_encode_operation(resp, op);
>
> Encoding both operations here makes me very nervous, but I haven't
> thought it through.
We only encode the sequence operation here, and set the error in the
next operation which gets encoded in the next nfsd4_proc_compound
loop.
>
>> + }
>> + return op->status;
>> +}
>> +
>> +/*
>> * COMPOUND call.
>> */
>> static __be32
>> @@ -879,7 +907,6 @@ nfsd4_proc_compound(struct svc_rqst *rqstp,
>> dprintk("nfsv4 compound op #%d/%d: %d (%s)\n",
>> resp->opcnt, args->opcnt, op->opnum,
>> nfsd4_op_name(op->opnum));
>> -
>> /*
>> * The XDR decode routines may have pre-set op->status;
>> * for example, if there is a miscellaneous XDR error
>> @@ -923,7 +950,10 @@ encode_op:
>> /* Only from SEQUENCE or CREATE_SESSION */
>> if (resp->cstate.status == nfserr_replay_cache) {
>> dprintk("%s NFS4.1 replay from cache\n", __func__);
>> - status = op->status;
>> + if (nfsd4_no_page_in_cache(resp))
>
> Why not just call that nfsd4_not_cached()?
OK
>
>> + status = nfsd4_enc_no_page_replay(args, resp);
>
> and nfsd4_enc_uncached_replay()? (The "no_page" this is a technical
> detail of the current caching implementation.)
OK
>
>> + else
>> + status = op->status;
>> goto out;
>> }
>> if (op->status == nfserr_replay_me) {
>> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
>> index 61af434..f42cda9 100644
>> --- a/fs/nfsd/nfs4state.c
>> +++ b/fs/nfsd/nfs4state.c
>> @@ -1068,17 +1068,31 @@ nfsd4_set_cache_entry(struct nfsd4_compoundres *resp)
>> /* Don't cache a failed OP_SEQUENCE. */
>> if (resp->opcnt == 1 && op->opnum == OP_SEQUENCE && resp->cstate.status)
>> return;
>> +
>> nfsd4_release_respages(entry->ce_respages, entry->ce_resused);
>> + entry->ce_opcnt = resp->opcnt;
>> + entry->ce_status = resp->cstate.status;
>> +
>> + /*
>> + * Don't need a page to cache just the sequence operation - the slot
>> + * does this for us!
>> + */
>> +
>> + if (nfsd4_no_page_in_cache(resp)) {
>> + entry->ce_resused = 0;
>> + entry->ce_rpchdrlen = 0;
>> + dprintk("%s Just cache SEQUENCE. ce_cachethis %d\n", __func__,
>> + resp->cstate.slot->sl_cache_entry.ce_cachethis);
>> + return;
>> + }
>
> Do we *ever* actually need to cache the initial sequence op? Should we
> only be storing subsequent ops in the reply cache?
We cache encoded pages, or not. If we did not do this, then we would
be mucking with each and every reply just to cache it. e.g. we would
be doing a copy of all the other subsequent ops out of the page into
some other storage. We also need the rpc header of the reply so that
we have the choice later on to compare the original principal with the
incoming replay principal.
To me it makes a lot more sense to do the least amount of work in the
common no-replay case, which is to simply store the page or two of the
reply.
>
>> entry->ce_resused = rqstp->rq_resused;
>> if (entry->ce_resused > NFSD_PAGES_PER_SLOT + 1)
>> entry->ce_resused = NFSD_PAGES_PER_SLOT + 1;
>> nfsd4_move_pages(entry->ce_respages, rqstp->rq_respages,
>> entry->ce_resused);
>> - entry->ce_status = resp->cstate.status;
>> entry->ce_datav.iov_base = resp->cstate.statp;
>> entry->ce_datav.iov_len = resv->iov_len - ((char *)resp->cstate.statp -
>> (char *)page_address(rqstp->rq_respages[0]));
>> - entry->ce_opcnt = resp->opcnt;
>> /* Current request rpc header length*/
>> entry->ce_rpchdrlen = (char *)resp->cstate.statp -
>> (char *)page_address(rqstp->rq_respages[0]);
>> @@ -1117,13 +1131,28 @@ nfsd41_copy_replay_data(struct nfsd4_compoundres *resp,
>> * cached page. Replace any futher replay pages from the cache.
>> */
>> __be32
>> -nfsd4_replay_cache_entry(struct nfsd4_compoundres *resp)
>> +nfsd4_replay_cache_entry(struct nfsd4_compoundres *resp,
>> + struct nfsd4_sequence *seq)
>> {
>> struct nfsd4_cache_entry *entry = &resp->cstate.slot->sl_cache_entry;
>> __be32 status;
>>
>> dprintk("--> %s entry %p\n", __func__, entry);
>>
>> + /*
>> + * If this is just the sequence operation, we did not keep
>> + * a page in the cache entry because we can just use the
>> + * slot info stored in struct nfsd4_sequence that was checked
>> + * against the slot in nfsd4_sequence().
>> + *
>> + * This occurs when seq->cachethis is FALSE, or when the client
>> + * session inactivity timer fires and a solo sequence operation
>> + * is sent (lease renewal).
>> + */
>> + if (seq && nfsd4_no_page_in_cache(resp)) {
>> + seq->maxslots = resp->cstate.slot->sl_session->se_fnumslots;
>> + return nfs_ok;
>> + }
>>
>> if (!nfsd41_copy_replay_data(resp, entry)) {
>> /*
>> @@ -1347,7 +1376,7 @@ nfsd4_create_session(struct svc_rqst *rqstp,
>> cstate->slot = slot;
>> cstate->status = status;
>> /* Return the cached reply status */
>> - status = nfsd4_replay_cache_entry(resp);
>> + status = nfsd4_replay_cache_entry(resp, NULL);
>> goto out;
>> } else if (cr_ses->seqid != conf->cl_slot.sl_seqid + 1) {
>> status = nfserr_seq_misordered;
>> @@ -1397,6 +1426,8 @@ nfsd4_create_session(struct svc_rqst *rqstp,
>>
>> slot->sl_inuse = true;
>> cstate->slot = slot;
>> + /* Ensure a page is used for the cache */
>> + slot->sl_cache_entry.ce_cachethis = 1;
>> out:
>> nfs4_unlock_state();
>> dprintk("%s returns %d\n", __func__, ntohl(status));
>> @@ -1441,8 +1472,8 @@ nfsd4_sequence(struct svc_rqst *rqstp,
>> if (status == nfserr_replay_cache) {
>> cstate->slot = slot;
>> /* Return the cached reply status and set cstate->status
>> - * for nfsd4_svc_encode_compoundres processing*/
>> - status = nfsd4_replay_cache_entry(resp);
>> + * for nfsd4_svc_encode_compoundres processing */
>
> The comment typo-fix doesn't belong in this patch.
OK
>
>> + status = nfsd4_replay_cache_entry(resp, seq);
>> cstate->status = nfserr_replay_cache;
>> goto replay_cache;
>> }
>> @@ -1452,6 +1483,10 @@ nfsd4_sequence(struct svc_rqst *rqstp,
>> /* Success! bump slot seqid */
>> slot->sl_inuse = true;
>> slot->sl_seqid = seq->seqid;
>> + slot->sl_cache_entry.ce_cachethis = seq->cachethis;
>> + /* Always set the cache entry cachethis for solo sequence */
>> + if (nfsd4_is_solo_sequence(resp))
>> + slot->sl_cache_entry.ce_cachethis = 1;
>>
>> cstate->slot = slot;
>>
>> diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
>> index 60db854..a8bb04a 100644
>> --- a/fs/nfsd/nfs4xdr.c
>> +++ b/fs/nfsd/nfs4xdr.c
>> @@ -2984,7 +2984,7 @@ nfsd4_encode_destroy_session(struct nfsd4_compoundres *resp, int nfserr,
>> return nfserr;
>> }
>>
>> -static __be32
>> +__be32
>> nfsd4_encode_sequence(struct nfsd4_compoundres *resp, int nfserr,
>> struct nfsd4_sequence *seq)
>> {
>> @@ -3204,7 +3204,8 @@ nfs4svc_encode_compoundres(struct svc_rqst *rqstp, __be32 *p, struct nfsd4_compo
>> BUG_ON(iov->iov_len > PAGE_SIZE);
>> #ifdef CONFIG_NFSD_V4_1
>> if (resp->cstate.slot != NULL) {
>> - if (resp->cstate.status == nfserr_replay_cache) {
>> + if (resp->cstate.status == nfserr_replay_cache &&
>> + !nfsd4_no_page_in_cache(resp)) {
>> iov->iov_len = resp->cstate.iovlen;
>> } else {
>> nfsd4_set_cache_entry(resp);
>> diff --git a/include/linux/nfsd/state.h b/include/linux/nfsd/state.h
>> index 49d89fd..47c7836 100644
>> --- a/include/linux/nfsd/state.h
>> +++ b/include/linux/nfsd/state.h
>> @@ -110,6 +110,7 @@ struct nfsd4_cache_entry {
>> __be32 ce_status;
>> struct kvec ce_datav; /* encoded NFSv4.1 data in rq_res.head[0] */
>> struct page *ce_respages[NFSD_PAGES_PER_SLOT + 1];
>> + int ce_cachethis;
>> short ce_resused;
>> int ce_opcnt;
>> int ce_rpchdrlen;
>> diff --git a/include/linux/nfsd/xdr4.h b/include/linux/nfsd/xdr4.h
>> index c7bf0a1..641e5cd 100644
>> --- a/include/linux/nfsd/xdr4.h
>> +++ b/include/linux/nfsd/xdr4.h
>> @@ -482,6 +482,18 @@ struct nfsd4_compoundres {
>> struct nfsd4_compound_state cstate;
>> };
>>
>> +static inline u32 nfsd4_is_solo_sequence(struct nfsd4_compoundres *resp)
>> +{
>> + struct nfsd4_compoundargs *args = resp->rqstp->rq_argp;
>> + return args->opcnt == 1 ? 1 : 0;
>
> Drop the redundant "? 1: 0", and make the return int (or boolean, if you
> want).
OK
>
>> +}
>> +
>> +static inline u32 nfsd4_no_page_in_cache(struct nfsd4_compoundres *resp)
>
> Ditto on the return type.
OK
>
>> +{
>> + return (resp->cstate.slot->sl_cache_entry.ce_cachethis == 0 ||
>> + nfsd4_is_solo_sequence(resp));
>
> Drop the extra parentheses.
OK
>
> --b.
>
>> +}
>> +
>> #define NFS4_SVC_XDRSIZE sizeof(struct nfsd4_compoundargs)
>>
>> static inline void
>> @@ -513,7 +525,8 @@ extern __be32 nfsd4_setclientid_confirm(struct svc_rqst *rqstp,
>> struct nfsd4_setclientid_confirm *setclientid_confirm);
>> #if defined(CONFIG_NFSD_V4_1)
>> extern void nfsd4_set_cache_entry(struct nfsd4_compoundres *resp);
>> -extern __be32 nfsd4_replay_cache_entry(struct nfsd4_compoundres *resp);
>> +extern __be32 nfsd4_replay_cache_entry(struct nfsd4_compoundres *resp,
>> + struct nfsd4_sequence *seq);
>> extern __be32 nfsd4_exchange_id(struct svc_rqst *rqstp,
>> struct nfsd4_compound_state *,
>> struct nfsd4_exchange_id *);
>> --
>> 1.6.2.1
>>
> _______________________________________________
> pNFS mailing list
> pNFS@xxxxxxxxxxxxx
> http://linux-nfs.org/cgi-bin/mailman/listinfo/pnfs
>
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
