summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorDavid S. Miller <davem@davemloft.net>2015-05-04 00:09:09 -0400
committerDavid S. Miller <davem@davemloft.net>2015-05-04 00:09:09 -0400
commit7c9a2eeac9b9caafb806e624fd2b75d30ecaee64 (patch)
tree6a96e0a238759aa880c17b3efbe2ba616eaad86e
parent6a21165480a066a27c1f1dbd32aec581c612ba23 (diff)
parent2e99403d28c182aa7ffb2d4ef34472df4873a4dd (diff)
Merge branch 'flow_keys_digest'
Tom Herbert says: ==================== net: Eliminate calls to flow_dissector and introduce flow_keys_digest In this patch set we add skb_get_hash_perturb which gets the skbuff hash for a packet and perturbs it using a provided key and jhash1. This function is used in serveral qdiscs and eliminates many calls to flow_dissector and jhash3 to get a perturbed hash for a packet. To handle the sch_choke issue (passes flow_keys in skbuff cb) we add flow_keys_digest which is a digest of a flow constructed from a flow_keys structure. This is the second version of these patches I posted a while ago, and is prerequisite work to increasing the size of the flow_keys structure and hashing over it (full IPv6 address, flow label, VLAN ID, etc.). Version 2: - Add keyval parameter to __flow_hash_from_keys which allows caller to set the initval for jhash - Perturb always does flow dissection and creates hash based on input perturb value which acts as the keyval to __flow_hash_from_keys - Added a _flow_keys_digest_data which is used in make_flow_keys_digest. This fills out the digest by populating individual fields instead of copying the whole structure. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
-rw-r--r--include/linux/skbuff.h2
-rw-r--r--include/net/flow_keys.h16
-rw-r--r--net/core/flow_dissector.c65
-rw-r--r--net/sched/sch_choke.c14
-rw-r--r--net/sched/sch_fq_codel.c11
-rw-r--r--net/sched/sch_hhf.c19
-rw-r--r--net/sched/sch_sfb.c24
-rw-r--r--net/sched/sch_sfq.c27
8 files changed, 92 insertions, 86 deletions
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 66e374d62f64..acb83e249e3f 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -927,6 +927,8 @@ static inline __u32 skb_get_hash(struct sk_buff *skb)
return skb->hash;
}
+__u32 skb_get_hash_perturb(const struct sk_buff *skb, u32 perturb);
+
static inline __u32 skb_get_hash_raw(const struct sk_buff *skb)
{
return skb->hash;
diff --git a/include/net/flow_keys.h b/include/net/flow_keys.h
index dc8fd81412bf..6d6ef626811a 100644
--- a/include/net/flow_keys.h
+++ b/include/net/flow_keys.h
@@ -42,4 +42,20 @@ static inline __be32 skb_flow_get_ports(const struct sk_buff *skb, int thoff, u8
u32 flow_hash_from_keys(struct flow_keys *keys);
unsigned int flow_get_hlen(const unsigned char *data, unsigned int max_len,
__be16 protocol);
+
+/* struct flow_keys_digest:
+ *
+ * This structure is used to hold a digest of the full flow keys. This is a
+ * larger "hash" of a flow to allow definitively matching specific flows where
+ * the 32 bit skb->hash is not large enough. The size is limited to 16 bytes so
+ * that it can by used in CB of skb (see sch_choke for an example).
+ */
+#define FLOW_KEYS_DIGEST_LEN 16
+struct flow_keys_digest {
+ u8 data[FLOW_KEYS_DIGEST_LEN];
+};
+
+void make_flow_keys_digest(struct flow_keys_digest *digest,
+ const struct flow_keys *flow);
+
#endif
diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
index 2c35c02a931e..d3acc4dff4ae 100644
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c
@@ -267,13 +267,12 @@ static __always_inline void __flow_hash_secret_init(void)
net_get_random_once(&hashrnd, sizeof(hashrnd));
}
-static __always_inline u32 __flow_hash_3words(u32 a, u32 b, u32 c)
+static __always_inline u32 __flow_hash_3words(u32 a, u32 b, u32 c, u32 keyval)
{
- __flow_hash_secret_init();
- return jhash_3words(a, b, c, hashrnd);
+ return jhash_3words(a, b, c, keyval);
}
-static inline u32 __flow_hash_from_keys(struct flow_keys *keys)
+static inline u32 __flow_hash_from_keys(struct flow_keys *keys, u32 keyval)
{
u32 hash;
@@ -287,7 +286,8 @@ static inline u32 __flow_hash_from_keys(struct flow_keys *keys)
hash = __flow_hash_3words((__force u32)keys->dst,
(__force u32)keys->src,
- (__force u32)keys->ports);
+ (__force u32)keys->ports,
+ keyval);
if (!hash)
hash = 1;
@@ -296,10 +296,47 @@ static inline u32 __flow_hash_from_keys(struct flow_keys *keys)
u32 flow_hash_from_keys(struct flow_keys *keys)
{
- return __flow_hash_from_keys(keys);
+ __flow_hash_secret_init();
+ return __flow_hash_from_keys(keys, hashrnd);
}
EXPORT_SYMBOL(flow_hash_from_keys);
+static inline u32 ___skb_get_hash(const struct sk_buff *skb,
+ struct flow_keys *keys, u32 keyval)
+{
+ if (!skb_flow_dissect(skb, keys))
+ return 0;
+
+ return __flow_hash_from_keys(keys, keyval);
+}
+
+struct _flow_keys_digest_data {
+ __be16 n_proto;
+ u8 ip_proto;
+ u8 padding;
+ __be32 ports;
+ __be32 src;
+ __be32 dst;
+};
+
+void make_flow_keys_digest(struct flow_keys_digest *digest,
+ const struct flow_keys *flow)
+{
+ struct _flow_keys_digest_data *data =
+ (struct _flow_keys_digest_data *)digest;
+
+ BUILD_BUG_ON(sizeof(*data) > sizeof(*digest));
+
+ memset(digest, 0, sizeof(*digest));
+
+ data->n_proto = flow->n_proto;
+ data->ip_proto = flow->ip_proto;
+ data->ports = flow->ports;
+ data->src = flow->src;
+ data->dst = flow->dst;
+}
+EXPORT_SYMBOL(make_flow_keys_digest);
+
/*
* __skb_get_hash: calculate a flow hash based on src/dst addresses
* and src/dst port numbers. Sets hash in skb to non-zero hash value
@@ -309,8 +346,12 @@ EXPORT_SYMBOL(flow_hash_from_keys);
void __skb_get_hash(struct sk_buff *skb)
{
struct flow_keys keys;
+ u32 hash;
- if (!skb_flow_dissect(skb, &keys))
+ __flow_hash_secret_init();
+
+ hash = ___skb_get_hash(skb, &keys, hashrnd);
+ if (!hash)
return;
if (keys.ports)
@@ -318,10 +359,18 @@ void __skb_get_hash(struct sk_buff *skb)
skb->sw_hash = 1;
- skb->hash = __flow_hash_from_keys(&keys);
+ skb->hash = hash;
}
EXPORT_SYMBOL(__skb_get_hash);
+__u32 skb_get_hash_perturb(const struct sk_buff *skb, u32 perturb)
+{
+ struct flow_keys keys;
+
+ return ___skb_get_hash(skb, &keys, perturb);
+}
+EXPORT_SYMBOL(skb_get_hash_perturb);
+
/*
* Returns a Tx hash based on the given packet descriptor a Tx queues' number
* to be used as a distribution range.
diff --git a/net/sched/sch_choke.c b/net/sched/sch_choke.c
index c009eb9045ce..dfe3da75594c 100644
--- a/net/sched/sch_choke.c
+++ b/net/sched/sch_choke.c
@@ -133,16 +133,10 @@ static void choke_drop_by_idx(struct Qdisc *sch, unsigned int idx)
--sch->q.qlen;
}
-/* private part of skb->cb[] that a qdisc is allowed to use
- * is limited to QDISC_CB_PRIV_LEN bytes.
- * As a flow key might be too large, we store a part of it only.
- */
-#define CHOKE_K_LEN min_t(u32, sizeof(struct flow_keys), QDISC_CB_PRIV_LEN - 3)
-
struct choke_skb_cb {
u16 classid;
u8 keys_valid;
- u8 keys[QDISC_CB_PRIV_LEN - 3];
+ struct flow_keys_digest keys;
};
static inline struct choke_skb_cb *choke_skb_cb(const struct sk_buff *skb)
@@ -177,18 +171,18 @@ static bool choke_match_flow(struct sk_buff *skb1,
if (!choke_skb_cb(skb1)->keys_valid) {
choke_skb_cb(skb1)->keys_valid = 1;
skb_flow_dissect(skb1, &temp);
- memcpy(&choke_skb_cb(skb1)->keys, &temp, CHOKE_K_LEN);
+ make_flow_keys_digest(&choke_skb_cb(skb1)->keys, &temp);
}
if (!choke_skb_cb(skb2)->keys_valid) {
choke_skb_cb(skb2)->keys_valid = 1;
skb_flow_dissect(skb2, &temp);
- memcpy(&choke_skb_cb(skb2)->keys, &temp, CHOKE_K_LEN);
+ make_flow_keys_digest(&choke_skb_cb(skb2)->keys, &temp);
}
return !memcmp(&choke_skb_cb(skb1)->keys,
&choke_skb_cb(skb2)->keys,
- CHOKE_K_LEN);
+ sizeof(choke_skb_cb(skb1)->keys));
}
/*
diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c
index 1e52decb7b59..a6fc53d69513 100644
--- a/net/sched/sch_fq_codel.c
+++ b/net/sched/sch_fq_codel.c
@@ -23,7 +23,6 @@
#include <linux/vmalloc.h>
#include <net/netlink.h>
#include <net/pkt_sched.h>
-#include <net/flow_keys.h>
#include <net/codel.h>
/* Fair Queue CoDel.
@@ -68,15 +67,9 @@ struct fq_codel_sched_data {
};
static unsigned int fq_codel_hash(const struct fq_codel_sched_data *q,
- const struct sk_buff *skb)
+ struct sk_buff *skb)
{
- struct flow_keys keys;
- unsigned int hash;
-
- skb_flow_dissect(skb, &keys);
- hash = jhash_3words((__force u32)keys.dst,
- (__force u32)keys.src ^ keys.ip_proto,
- (__force u32)keys.ports, q->perturbation);
+ u32 hash = skb_get_hash_perturb(skb, q->perturbation);
return reciprocal_scale(hash, q->flows_cnt);
}
diff --git a/net/sched/sch_hhf.c b/net/sched/sch_hhf.c
index 15d3aabfe250..9d15cb6b8cb1 100644
--- a/net/sched/sch_hhf.c
+++ b/net/sched/sch_hhf.c
@@ -9,7 +9,6 @@
#include <linux/module.h>
#include <linux/skbuff.h>
#include <linux/vmalloc.h>
-#include <net/flow_keys.h>
#include <net/pkt_sched.h>
#include <net/sock.h>
@@ -176,22 +175,6 @@ static u32 hhf_time_stamp(void)
return jiffies;
}
-static unsigned int skb_hash(const struct hhf_sched_data *q,
- const struct sk_buff *skb)
-{
- struct flow_keys keys;
- unsigned int hash;
-
- if (skb->sk && skb->sk->sk_hash)
- return skb->sk->sk_hash;
-
- skb_flow_dissect(skb, &keys);
- hash = jhash_3words((__force u32)keys.dst,
- (__force u32)keys.src ^ keys.ip_proto,
- (__force u32)keys.ports, q->perturbation);
- return hash;
-}
-
/* Looks up a heavy-hitter flow in a chaining list of table T. */
static struct hh_flow_state *seek_list(const u32 hash,
struct list_head *head,
@@ -280,7 +263,7 @@ static enum wdrr_bucket_idx hhf_classify(struct sk_buff *skb, struct Qdisc *sch)
}
/* Get hashed flow-id of the skb. */
- hash = skb_hash(q, skb);
+ hash = skb_get_hash_perturb(skb, q->perturbation);
/* Check if this packet belongs to an already established HH flow. */
flow_pos = hash & HHF_BIT_MASK;
diff --git a/net/sched/sch_sfb.c b/net/sched/sch_sfb.c
index 5819dd82630d..4b815193326c 100644
--- a/net/sched/sch_sfb.c
+++ b/net/sched/sch_sfb.c
@@ -26,7 +26,6 @@
#include <net/ip.h>
#include <net/pkt_sched.h>
#include <net/inet_ecn.h>
-#include <net/flow_keys.h>
/*
* SFB uses two B[l][n] : L x N arrays of bins (L levels, N bins per level)
@@ -285,9 +284,9 @@ static int sfb_enqueue(struct sk_buff *skb, struct Qdisc *sch)
int i;
u32 p_min = ~0;
u32 minqlen = ~0;
- u32 r, slot, salt, sfbhash;
+ u32 r, sfbhash;
+ u32 slot = q->slot;
int ret = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS;
- struct flow_keys keys;
if (unlikely(sch->q.qlen >= q->limit)) {
qdisc_qstats_overlimit(sch);
@@ -309,22 +308,17 @@ static int sfb_enqueue(struct sk_buff *skb, struct Qdisc *sch)
fl = rcu_dereference_bh(q->filter_list);
if (fl) {
+ u32 salt;
+
/* If using external classifiers, get result and record it. */
if (!sfb_classify(skb, fl, &ret, &salt))
goto other_drop;
- keys.src = salt;
- keys.dst = 0;
- keys.ports = 0;
+ sfbhash = jhash_1word(salt, q->bins[slot].perturbation);
} else {
- skb_flow_dissect(skb, &keys);
+ sfbhash = skb_get_hash_perturb(skb, q->bins[slot].perturbation);
}
- slot = q->slot;
- sfbhash = jhash_3words((__force u32)keys.dst,
- (__force u32)keys.src,
- (__force u32)keys.ports,
- q->bins[slot].perturbation);
if (!sfbhash)
sfbhash = 1;
sfb_skb_cb(skb)->hashes[slot] = sfbhash;
@@ -356,10 +350,8 @@ static int sfb_enqueue(struct sk_buff *skb, struct Qdisc *sch)
if (unlikely(p_min >= SFB_MAX_PROB)) {
/* Inelastic flow */
if (q->double_buffering) {
- sfbhash = jhash_3words((__force u32)keys.dst,
- (__force u32)keys.src,
- (__force u32)keys.ports,
- q->bins[slot].perturbation);
+ sfbhash = skb_get_hash_perturb(skb,
+ q->bins[slot].perturbation);
if (!sfbhash)
sfbhash = 1;
sfb_skb_cb(skb)->hashes[slot] = sfbhash;
diff --git a/net/sched/sch_sfq.c b/net/sched/sch_sfq.c
index b877140beda5..7d1492663360 100644
--- a/net/sched/sch_sfq.c
+++ b/net/sched/sch_sfq.c
@@ -23,7 +23,6 @@
#include <linux/vmalloc.h>
#include <net/netlink.h>
#include <net/pkt_sched.h>
-#include <net/flow_keys.h>
#include <net/red.h>
@@ -156,30 +155,10 @@ static inline struct sfq_head *sfq_dep_head(struct sfq_sched_data *q, sfq_index
return &q->dep[val - SFQ_MAX_FLOWS];
}
-/*
- * In order to be able to quickly rehash our queue when timer changes
- * q->perturbation, we store flow_keys in skb->cb[]
- */
-struct sfq_skb_cb {
- struct flow_keys keys;
-};
-
-static inline struct sfq_skb_cb *sfq_skb_cb(const struct sk_buff *skb)
-{
- qdisc_cb_private_validate(skb, sizeof(struct sfq_skb_cb));
- return (struct sfq_skb_cb *)qdisc_skb_cb(skb)->data;
-}
-
static unsigned int sfq_hash(const struct sfq_sched_data *q,
const struct sk_buff *skb)
{
- const struct flow_keys *keys = &sfq_skb_cb(skb)->keys;
- unsigned int hash;
-
- hash = jhash_3words((__force u32)keys->dst,
- (__force u32)keys->src ^ keys->ip_proto,
- (__force u32)keys->ports, q->perturbation);
- return hash & (q->divisor - 1);
+ return skb_get_hash_perturb(skb, q->perturbation) & (q->divisor - 1);
}
static unsigned int sfq_classify(struct sk_buff *skb, struct Qdisc *sch,
@@ -196,10 +175,8 @@ static unsigned int sfq_classify(struct sk_buff *skb, struct Qdisc *sch,
return TC_H_MIN(skb->priority);
fl = rcu_dereference_bh(q->filter_list);
- if (!fl) {
- skb_flow_dissect(skb, &sfq_skb_cb(skb)->keys);
+ if (!fl)
return sfq_hash(q, skb) + 1;
- }
*qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS;
result = tc_classify(skb, fl, &res);