by J. Bruce Fields

[permalink] [raw]

Subject: Re: [PATCH 13/15] knfsd: Simplify NFS duplicate replay cache

On Wed, Oct 03, 2018 at 02:11:50PM -0400, [email protected] wrote:
> On Wed, Oct 03, 2018 at 06:01:36PM +0000, Trond Myklebust wrote:
> > On Wed, 2018-10-03 at 13:14 -0400, J . Bruce Fields wrote:
> > > On Mon, Oct 01, 2018 at 10:41:55AM -0400, Trond Myklebust wrote:
> > > > Simplify the duplicate replay cache by initialising the
> > > > preallocated
> > > > cache entry, so that we can use it as a key for the cache lookup.
> > > >
> > > > Note that the 99.999% case we want to optimise for is still the one
> > > > where the lookup fails, and we have to add this entry to the cache,
> > > > so preinitialising should not cause a performance penalty.
> > > >
> > > > Signed-off-by: Trond Myklebust <[email protected]>
> > > > ---
> > > > fs/nfsd/cache.h | 6 +--
> > > > fs/nfsd/nfscache.c | 94 ++++++++++++++++++++++------------------
> > > > ------
> > > > 2 files changed, 47 insertions(+), 53 deletions(-)
> > > >
> > > > diff --git a/fs/nfsd/cache.h b/fs/nfsd/cache.h
> > > > index b7559c6f2b97..bd77d5f6fe53 100644
> > > > --- a/fs/nfsd/cache.h
> > > > +++ b/fs/nfsd/cache.h
> > > > @@ -24,13 +24,13 @@ struct svc_cacherep {
> > > > unsigned char c_state, /* unused, inprog,
> > > > done */
> > > > c_type, /* status, buffer
> > > > */
> > > > c_secure : 1; /* req came from
> > > > port < 1024 */
> > > > - struct sockaddr_in6 c_addr;
> > > > __be32 c_xid;
> > > > - u32 c_prot;
> > > > + __wsum c_csum;
> > > > u32 c_proc;
> > > > + u32 c_prot;
> > > > u32 c_vers;
> > > > unsigned int c_len;
> > > > - __wsum c_csum;
> > > > + struct sockaddr_in6 c_addr;
> > > > unsigned long c_timestamp;
> > > > union {
> > > > struct kvec u_vec;
> > >
> > > Unless I've missed something subtle--I'll move this chunk into the
> > > next
> > > patch.--b.
> >
> > Nothing too subtle. The only reason for keeping it in this patch is
> > because even with the current code, most of the comparisons hit the
> > c_xid and possibly sometimes the c_csum, so those are the main fields
> > that you want to try to keep in the same cache line.
>
> That could use a comment.

I'm adding just

struct svc_cacherep {
struct {
+ /* Keep often-read xid, csum in the same cache line: */
__be32 k_xid;
__wsum k_csum;
u32 k_proc;

--b.

2018-10-04 07:35:18

by J. Bruce Fields

[permalink] [raw]

Subject: Re: [PATCH 15/15] knfsd: Improve lookup performance in the duplicate reply cache using an rbtree

On Mon, Oct 01, 2018 at 10:41:57AM -0400, Trond Myklebust wrote:
> Use an rbtree to ensure the lookup/insert of an entry in a DRC bucket is
> O(log(N)).

So the naming on those hash chain statistics are off now, but I guess
that's harmless.

All looks good to me, thanks--applying.

--b.

>
> Signed-off-by: Trond Myklebust <[email protected]>
> ---
> fs/nfsd/cache.h | 1 +
> fs/nfsd/nfscache.c | 37 ++++++++++++++++++++++++++-----------
> 2 files changed, 27 insertions(+), 11 deletions(-)
>
> diff --git a/fs/nfsd/cache.h b/fs/nfsd/cache.h
> index cbcdc15753b7..ef1c8aa89ca4 100644
> --- a/fs/nfsd/cache.h
> +++ b/fs/nfsd/cache.h
> @@ -29,6 +29,7 @@ struct svc_cacherep {
> struct sockaddr_in6 k_addr;
> } c_key;
>
> + struct rb_node c_node;
> struct list_head c_lru;
> unsigned char c_state, /* unused, inprog, done */
> c_type, /* status, buffer */
> diff --git a/fs/nfsd/nfscache.c b/fs/nfsd/nfscache.c
> index 230cc83921ad..e2fe0e9ce0df 100644
> --- a/fs/nfsd/nfscache.c
> +++ b/fs/nfsd/nfscache.c
> @@ -30,6 +30,7 @@
> #define TARGET_BUCKET_SIZE 64
>
> struct nfsd_drc_bucket {
> + struct rb_root rb_head;
> struct list_head lru_head;
> spinlock_t cache_lock;
> };
> @@ -129,6 +130,7 @@ nfsd_reply_cache_alloc(struct svc_rqst *rqstp, __wsum csum)
> if (rp) {
> rp->c_state = RC_UNUSED;
> rp->c_type = RC_NOCACHE;
> + RB_CLEAR_NODE(&rp->c_node);
> INIT_LIST_HEAD(&rp->c_lru);
>
> memset(&rp->c_key, 0, sizeof(rp->c_key));
> @@ -145,13 +147,14 @@ nfsd_reply_cache_alloc(struct svc_rqst *rqstp, __wsum csum)
> }
>
> static void
> -nfsd_reply_cache_free_locked(struct svc_cacherep *rp)
> +nfsd_reply_cache_free_locked(struct nfsd_drc_bucket *b, struct svc_cacherep *rp)
> {
> if (rp->c_type == RC_REPLBUFF && rp->c_replvec.iov_base) {
> drc_mem_usage -= rp->c_replvec.iov_len;
> kfree(rp->c_replvec.iov_base);
> }
> if (rp->c_state != RC_UNUSED) {
> + rb_erase(&rp->c_node, &b->rb_head);
> list_del(&rp->c_lru);
> atomic_dec(&num_drc_entries);
> drc_mem_usage -= sizeof(*rp);
> @@ -163,7 +166,7 @@ static void
> nfsd_reply_cache_free(struct nfsd_drc_bucket *b, struct svc_cacherep *rp)
> {
> spin_lock(&b->cache_lock);
> - nfsd_reply_cache_free_locked(rp);
> + nfsd_reply_cache_free_locked(b, rp);
> spin_unlock(&b->cache_lock);
> }
>
> @@ -219,7 +222,7 @@ void nfsd_reply_cache_shutdown(void)
> struct list_head *head = &drc_hashtbl[i].lru_head;
> while (!list_empty(head)) {
> rp = list_first_entry(head, struct svc_cacherep, c_lru);
> - nfsd_reply_cache_free_locked(rp);
> + nfsd_reply_cache_free_locked(&drc_hashtbl[i], rp);
> }
> }
>
> @@ -258,7 +261,7 @@ prune_bucket(struct nfsd_drc_bucket *b)
> if (atomic_read(&num_drc_entries) <= max_drc_entries &&
> time_before(jiffies, rp->c_timestamp + RC_EXPIRE))
> break;
> - nfsd_reply_cache_free_locked(rp);
> + nfsd_reply_cache_free_locked(b, rp);
> freed++;
> }
> return freed;
> @@ -349,17 +352,29 @@ static struct svc_cacherep *
> nfsd_cache_insert(struct nfsd_drc_bucket *b, struct svc_cacherep *key)
> {
> struct svc_cacherep *rp, *ret = key;
> - struct list_head *rh = &b->lru_head;
> + struct rb_node **p = &b->rb_head.rb_node,
> + *parent = NULL;
> unsigned int entries = 0;
> + int cmp;
>
> - list_for_each_entry(rp, rh, c_lru) {
> + while (*p != NULL) {
> ++entries;
> - if (nfsd_cache_key_cmp(key, rp) == 0) {
> + parent = *p;
> + rp = rb_entry(parent, struct svc_cacherep, c_node);
> +
> + cmp = nfsd_cache_key_cmp(key, rp);
> + if (cmp < 0)
> + p = &parent->rb_left;
> + else if (cmp > 0)
> + p = &parent->rb_right;
> + else {
> ret = rp;
> - break;
> + goto out;
> }
> }
> -
> + rb_link_node(&key->c_node, parent, p);
> + rb_insert_color(&key->c_node, &b->rb_head);
> +out:
> /* tally hash chain length stats */
> if (entries > longest_chain) {
> longest_chain = entries;
> @@ -414,7 +429,7 @@ nfsd_cache_lookup(struct svc_rqst *rqstp)
> spin_lock(&b->cache_lock);
> found = nfsd_cache_insert(b, rp);
> if (found != rp) {
> - nfsd_reply_cache_free_locked(rp);
> + nfsd_reply_cache_free_locked(NULL, rp);
> rp = found;
> goto found_entry;
> }
> @@ -462,7 +477,7 @@ nfsd_cache_lookup(struct svc_rqst *rqstp)
> break;
> default:
> printk(KERN_WARNING "nfsd: bad repcache type %d\n", rp->c_type);
> - nfsd_reply_cache_free_locked(rp);
> + nfsd_reply_cache_free_locked(b, rp);
> }
>
> goto out;
> --
> 2.17.1