netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Josh Hunt <johunt@akamai.com>
To: Pablo Neira Ayuso <pablo@netfilter.org>,
	Patrick McHardy <kaber@trash.net>, Thomas Graf <tgraf@suug.ch>
Cc: netfilter-devel@vger.kernel.org, netdev@vger.kernel.org,
	Daniel Borkmann <daniel@iogearbox.net>,
	Josh Hunt <johunt@akamai.com>
Subject: [PATCH 3/3] nft_hash: introduce init_size set parameter
Date: Mon,  9 Feb 2015 19:48:31 -0500	[thread overview]
Message-ID: <1423529311-26050-4-git-send-email-johunt@akamai.com> (raw)
In-Reply-To: <1423529311-26050-1-git-send-email-johunt@akamai.com>

Currently nft_hash is using desc->size to both set a ceiling on the number of
entries a set can have, and as the nelem_hint for rhashtable. This not correct
since nelem_hint defines the # of initial buckets for the set to use. It is not
used to enforce a maximum size on the table. That's done through max_shift.

This creates a new parameter 'init_size' to pass as the nelem_hint to rhashtable
as the # of buckets you would like to initialize the table with.

Signed-off-by: Josh Hunt <johunt@akamai.com>
---
 include/net/netfilter/nf_tables.h        |    4 +++-
 include/uapi/linux/netfilter/nf_tables.h |    2 ++
 net/netfilter/nf_tables_api.c            |    4 ++++
 net/netfilter/nft_hash.c                 |    2 +-
 4 files changed, 10 insertions(+), 2 deletions(-)

diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index 9eaaa78..2c9130d 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -153,12 +153,14 @@ struct nft_set_iter {
  *
  *	@klen: key length
  *	@dlen: data length
- *	@size: number of set elements
+ *	@size: max number of set elements
+ *	@init_size: initial set size
  */
 struct nft_set_desc {
 	unsigned int		klen;
 	unsigned int		dlen;
 	unsigned int		size;
+	unsigned int		init_size;
 };
 
 /**
diff --git a/include/uapi/linux/netfilter/nf_tables.h b/include/uapi/linux/netfilter/nf_tables.h
index 832bc46..63e53eb 100644
--- a/include/uapi/linux/netfilter/nf_tables.h
+++ b/include/uapi/linux/netfilter/nf_tables.h
@@ -230,10 +230,12 @@ enum nft_set_policies {
  * enum nft_set_desc_attributes - set element description
  *
  * @NFTA_SET_DESC_SIZE: number of elements in set (NLA_U32)
+ * @NFTA_SET_DESC_INIT_SIZE: initial set size (NLA_U32)
  */
 enum nft_set_desc_attributes {
 	NFTA_SET_DESC_UNSPEC,
 	NFTA_SET_DESC_SIZE,
+	NFTA_SET_DESC_INIT_SIZE,
 	__NFTA_SET_DESC_MAX
 };
 #define NFTA_SET_DESC_MAX	(__NFTA_SET_DESC_MAX - 1)
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 199fd0f..275f41b 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -2211,6 +2211,7 @@ static const struct nla_policy nft_set_policy[NFTA_SET_MAX + 1] = {
 
 static const struct nla_policy nft_set_desc_policy[NFTA_SET_DESC_MAX + 1] = {
 	[NFTA_SET_DESC_SIZE]		= { .type = NLA_U32 },
+	[NFTA_SET_DESC_INIT_SIZE]	= { .type = NLA_U32 },
 };
 
 static int nft_ctx_init_from_setattr(struct nft_ctx *ctx,
@@ -2552,6 +2553,9 @@ static int nf_tables_set_desc_parse(const struct nft_ctx *ctx,
 	if (da[NFTA_SET_DESC_SIZE] != NULL)
 		desc->size = ntohl(nla_get_be32(da[NFTA_SET_DESC_SIZE]));
 
+	if (da[NFTA_SET_DESC_INIT_SIZE] != NULL)
+		desc->init_size = ntohl(nla_get_be32(da[NFTA_SET_DESC_INIT_SIZE]));
+
 	return 0;
 }
 
diff --git a/net/netfilter/nft_hash.c b/net/netfilter/nft_hash.c
index 08ec179..e03d1c6 100644
--- a/net/netfilter/nft_hash.c
+++ b/net/netfilter/nft_hash.c
@@ -190,7 +190,7 @@ static int nft_hash_init(const struct nft_set *set,
 {
 	struct rhashtable *priv = nft_set_priv(set);
 	struct rhashtable_params params = {
-		.nelem_hint = desc->size ? : NFT_HASH_ELEMENT_HINT,
+		.nelem_hint = desc->init_size ? : NFT_HASH_ELEMENT_HINT,
 		.head_offset = offsetof(struct nft_hash_elem, node),
 		.key_offset = offsetof(struct nft_hash_elem, key),
 		.key_len = set->klen,
-- 
1.7.9.5

      parent reply	other threads:[~2015-02-10  0:48 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-02-10  0:48 [PATCH 0/3] nft hash set expansion fixes Josh Hunt
2015-02-10  0:48 ` [PATCH 1/3] rhashtable: require max_shift definition Josh Hunt
2015-02-10  0:58   ` Thomas Graf
2015-02-10  8:30     ` Daniel Borkmann
2015-02-10 15:56       ` Josh Hunt
2015-02-10 17:06         ` Daniel Borkmann
2015-02-10 17:22           ` Thomas Graf
2015-02-10 17:44             ` Patrick McHardy
2015-02-10 21:18               ` Josh Hunt
2015-02-10  0:48 ` [PATCH 2/3] nft_hash: define max_shift rhashtable param Josh Hunt
2015-02-10  0:48 ` Josh Hunt [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1423529311-26050-4-git-send-email-johunt@akamai.com \
    --to=johunt@akamai.com \
    --cc=daniel@iogearbox.net \
    --cc=kaber@trash.net \
    --cc=netdev@vger.kernel.org \
    --cc=netfilter-devel@vger.kernel.org \
    --cc=pablo@netfilter.org \
    --cc=tgraf@suug.ch \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).