[PATCH openEuler-1.0-LTS v3 0/7] backport udp-4tuple-hash

backport udp-4tuple-hash Paolo Abeni (1): udp: fix l4 hash after reconnect Philo Lu (5): net/udp: Add a new struct for hash2 slot net/udp: Add 4-tuple hash list basis ipv4/udp: Add 4-tuple hash for connected socket ipv6/udp: Add 4-tuple hash for connected socket udp: Make rehash4 independent in udp_lib_rehash() Yue Haibing (1): inet6: Remove unused function declaration udpv6_connect() include/linux/udp.h | 12 ++ include/net/transp_v6.h | 2 - include/net/udp.h | 139 +++++++++++++++++++++- net/ipv4/datagram.c | 8 +- net/ipv4/udp.c | 251 ++++++++++++++++++++++++++++++++++++---- net/ipv6/udp.c | 121 ++++++++++++++++--- 6 files changed, 484 insertions(+), 49 deletions(-) -- 2.34.1

反馈: 您发送到kernel@openeuler.org的补丁/补丁集,已成功转换为PR! PR链接地址: https://gitee.com/openeuler/kernel/pulls/16874 邮件列表地址:https://mailweb.openeuler.org/archives/list/kernel@openeuler.org/message/NKY... FeedBack: The patch(es) which you have sent to kernel@openeuler.org mailing list has been converted to a pull request successfully! Pull request link: https://gitee.com/openeuler/kernel/pulls/16874 Mailing list address: https://mailweb.openeuler.org/archives/list/kernel@openeuler.org/message/NKY...

From: Yue Haibing <yuehaibing@huawei.com> mainline inclusion from mainline-v6.6-rc1 commit 999d0863ff6416dd33c904f496bf363c8afb6540 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/ICIAC2 CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i... ------------------------------------------------- This is never implemented since the beginning of git history. Signed-off-by: Yue Haibing <yuehaibing@huawei.com> Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com> Link: https://lore.kernel.org/r/20230731140437.37056-1-yuehaibing@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Liu Jian <liujian56@huawei.com> --- include/net/transp_v6.h | 2 -- 1 file changed, 2 deletions(-) diff --git a/include/net/transp_v6.h b/include/net/transp_v6.h index a8f6020f1196..b7fa8e179dec 100644 --- a/include/net/transp_v6.h +++ b/include/net/transp_v6.h @@ -31,8 +31,6 @@ void udplitev6_exit(void); int tcpv6_init(void); void tcpv6_exit(void); -int udpv6_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len); - /* this does all the common and the specific ctl work */ void ip6_datagram_recv_ctl(struct sock *sk, struct msghdr *msg, struct sk_buff *skb); -- 2.34.1

From: Philo Lu <lulie@linux.alibaba.com> mainline inclusion from mainline-v6.13-rc1 commit accdd51dc74ff65b7b7be1961b11723d228fbbbd category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/ICIAC2 CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i... ------------------------------------------------- Preparing for udp 4-tuple hash (uhash4 for short). To implement uhash4 without cache line missing when lookup, hslot2 is used to record the number of hashed sockets in hslot4. Thus adding a new struct udp_hslot_main with field hash4_cnt, which is used by hash2. The new struct is used to avoid doubling the size of udp_hslot. Before uhash4 lookup, firstly checking hash4_cnt to see if there are hashed sks in hslot4. Because hslot2 is always used in lookup, there is no cache line miss. Related helpers are updated, and use the helpers as possible. uhash4 is implemented in following patches. Signed-off-by: Philo Lu <lulie@linux.alibaba.com> Acked-by: Willem de Bruijn <willemb@google.com> Acked-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net> Conflicts: net/ipv4/udp.c net/ipv6/udp.c include/net/udp.h [Mainly caused by OLK did not backport 9804985bf27f, 5788b3a07fc5, 4cdeeee9252a.] Signed-off-by: Liu Jian <liujian56@huawei.com> --- include/net/udp.h | 40 ++++++++++++++++++++++++++++++++++++---- net/ipv4/udp.c | 31 +++++++++++++++---------------- net/ipv6/udp.c | 17 +++++++---------- 3 files changed, 58 insertions(+), 30 deletions(-) diff --git a/include/net/udp.h b/include/net/udp.h index bd27e9dd6b84..5a55d280f517 100644 --- a/include/net/udp.h +++ b/include/net/udp.h @@ -52,7 +52,7 @@ struct udp_skb_cb { #define UDP_SKB_CB(__skb) ((struct udp_skb_cb *)((__skb)->cb)) /** - * struct udp_hslot - UDP hash slot + * struct udp_hslot - UDP hash slot used by udp_table.hash * * @head: head of list of sockets * @count: number of sockets in 'head' list @@ -62,7 +62,22 @@ struct udp_hslot { struct hlist_head head; int count; spinlock_t lock; -} __attribute__((aligned(2 * sizeof(long)))); +} __aligned(2 * sizeof(long)); + +/** + * struct udp_hslot_main - UDP hash slot used by udp_table.hash2 + * + * @hslot: basic hash slot + * @hash4_cnt: number of sockets in hslot4 of the same + * (local port, local address) + */ +struct udp_hslot_main { + struct udp_hslot hslot; /* must be the first member */ +#if !IS_ENABLED(CONFIG_BASE_SMALL) + u32 hash4_cnt; +#endif +} __aligned(2 * sizeof(long)); +#define UDP_HSLOT_MAIN(__hslot) ((struct udp_hslot_main *)(__hslot)) /** * struct udp_table - UDP table @@ -74,7 +89,7 @@ struct udp_hslot { */ struct udp_table { struct udp_hslot *hash; - struct udp_hslot *hash2; + struct udp_hslot_main *hash2; unsigned int mask; unsigned int log; }; @@ -85,6 +100,7 @@ static inline struct udp_hslot *udp_hashslot(struct udp_table *table, { return &table->hash[udp_hashfn(net, num, table->mask)]; } + /* * For secondary hash, net_hash_mix() is performed before calling * udp_hashslot2(), this explains difference with udp_hashslot() @@ -92,8 +108,24 @@ static inline struct udp_hslot *udp_hashslot(struct udp_table *table, static inline struct udp_hslot *udp_hashslot2(struct udp_table *table, unsigned int hash) { - return &table->hash2[hash & table->mask]; + return &table->hash2[hash & table->mask].hslot; +} + +#if IS_ENABLED(CONFIG_BASE_SMALL) +static inline void udp_table_hash4_init(struct udp_table *table) +{ +} +#else /* !CONFIG_BASE_SMALL */ + +/* Must be called with table->hash2 initialized */ +static inline void udp_table_hash4_init(struct udp_table *table) +{ + int i; + + for (i = 0; i <= table->mask; i++) + table->hash2[i].hash4_cnt = 0; } +#endif /* CONFIG_BASE_SMALL */ extern struct proto udp_prot; diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 61af6d3388b6..4784ecd9d336 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -468,7 +468,7 @@ struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr, { struct sock *sk, *result; unsigned short hnum = ntohs(dport); - unsigned int hash2, slot2, slot = udp_hashfn(net, hnum, udptable->mask); + unsigned int hash2, slot = udp_hashfn(net, hnum, udptable->mask); struct udp_hslot *hslot2, *hslot = &udptable->hash[slot]; bool exact_dif = udp_lib_exact_dif_match(net, skb); int score, badness; @@ -476,8 +476,7 @@ struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr, if (hslot->count > 10) { hash2 = ipv4_portaddr_hash(net, daddr, hnum); - slot2 = hash2 & udptable->mask; - hslot2 = &udptable->hash2[slot2]; + hslot2 = udp_hashslot2(udptable, hash2); if (hslot->count < hslot2->count) goto begin; @@ -485,14 +484,13 @@ struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr, daddr, hnum, dif, sdif, exact_dif, hslot2, skb); if (!result) { - unsigned int old_slot2 = slot2; + void *old_hslot2 = hslot2; hash2 = ipv4_portaddr_hash(net, htonl(INADDR_ANY), hnum); - slot2 = hash2 & udptable->mask; + hslot2 = udp_hashslot2(udptable, hash2); /* avoid searching the same slot again. */ - if (unlikely(slot2 == old_slot2)) + if (unlikely(hslot2 == old_hslot2)) return result; - hslot2 = &udptable->hash2[slot2]; if (hslot->count < hslot2->count) goto begin; @@ -2083,7 +2081,7 @@ static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb, udptable->mask; hash2 = ipv4_portaddr_hash(net, daddr, hnum) & udptable->mask; start_lookup: - hslot = &udptable->hash2[hash2]; + hslot = &udptable->hash2[hash2].hslot; offset = offsetof(typeof(*sk), __sk_common.skc_portaddr_node); } @@ -2338,8 +2336,7 @@ static struct sock *__udp4_lib_demux_lookup(struct net *net, { unsigned short hnum = ntohs(loc_port); unsigned int hash2 = ipv4_portaddr_hash(net, loc_addr, hnum); - unsigned int slot2 = hash2 & udp_table.mask; - struct udp_hslot *hslot2 = &udp_table.hash2[slot2]; + struct udp_hslot *hslot2 = udp_hashslot2(&udp_table, hash2); INET_ADDR_COOKIE(acookie, rmt_addr, loc_addr); const __portpair ports = INET_COMBINED_PORTS(rmt_port, hnum); struct sock *sk; @@ -2918,10 +2915,11 @@ __setup("uhash_entries=", set_uhash_entries); void __init udp_table_init(struct udp_table *table, const char *name) { - unsigned int i; + unsigned int i, slot_size; + slot_size = sizeof(struct udp_hslot) + sizeof(struct udp_hslot_main); table->hash = alloc_large_system_hash(name, - 2 * sizeof(struct udp_hslot), + slot_size, uhash_entries, 21, /* one slot per 2 MB */ 0, @@ -2930,17 +2928,18 @@ void __init udp_table_init(struct udp_table *table, const char *name) UDP_HTABLE_SIZE_MIN, 64 * 1024); - table->hash2 = table->hash + (table->mask + 1); + table->hash2 = (void *)(table->hash + (table->mask + 1)); for (i = 0; i <= table->mask; i++) { INIT_HLIST_HEAD(&table->hash[i].head); table->hash[i].count = 0; spin_lock_init(&table->hash[i].lock); } for (i = 0; i <= table->mask; i++) { - INIT_HLIST_HEAD(&table->hash2[i].head); - table->hash2[i].count = 0; - spin_lock_init(&table->hash2[i].lock); + INIT_HLIST_HEAD(&table->hash2[i].hslot.head); + table->hash2[i].hslot.count = 0; + spin_lock_init(&table->hash2[i].hslot.lock); } + udp_table_hash4_init(table); } u32 udp_flow_hashrnd(void) diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index bc10b241da84..48b48be8b167 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -215,7 +215,7 @@ struct sock *__udp6_lib_lookup(struct net *net, { struct sock *sk, *result; unsigned short hnum = ntohs(dport); - unsigned int hash2, slot2, slot = udp_hashfn(net, hnum, udptable->mask); + unsigned int hash2, slot = udp_hashfn(net, hnum, udptable->mask); struct udp_hslot *hslot2, *hslot = &udptable->hash[slot]; bool exact_dif = udp6_lib_exact_dif_match(net, skb); int score, badness; @@ -223,8 +223,7 @@ struct sock *__udp6_lib_lookup(struct net *net, if (hslot->count > 10) { hash2 = ipv6_portaddr_hash(net, daddr, hnum); - slot2 = hash2 & udptable->mask; - hslot2 = &udptable->hash2[slot2]; + hslot2 = udp_hashslot2(udptable, hash2); if (hslot->count < hslot2->count) goto begin; @@ -232,14 +231,13 @@ struct sock *__udp6_lib_lookup(struct net *net, daddr, hnum, dif, sdif, exact_dif, hslot2, skb); if (!result) { - unsigned int old_slot2 = slot2; + void *old_hslot2 = hslot2; hash2 = ipv6_portaddr_hash(net, &in6addr_any, hnum); - slot2 = hash2 & udptable->mask; + hslot2 = udp_hashslot2(udptable, hash2); /* avoid searching the same slot again. */ - if (unlikely(slot2 == old_slot2)) + if (unlikely(hslot2 == old_hslot2)) return result; - hslot2 = &udptable->hash2[slot2]; if (hslot->count < hslot2->count) goto begin; @@ -713,7 +711,7 @@ static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb, udptable->mask; hash2 = ipv6_portaddr_hash(net, daddr, hnum) & udptable->mask; start_lookup: - hslot = &udptable->hash2[hash2]; + hslot = &udptable->hash2[hash2].hslot; offset = offsetof(typeof(*sk), __sk_common.skc_portaddr_node); } @@ -908,8 +906,7 @@ static struct sock *__udp6_lib_demux_lookup(struct net *net, { unsigned short hnum = ntohs(loc_port); unsigned int hash2 = ipv6_portaddr_hash(net, loc_addr, hnum); - unsigned int slot2 = hash2 & udp_table.mask; - struct udp_hslot *hslot2 = &udp_table.hash2[slot2]; + struct udp_hslot *hslot2 = udp_hashslot2(&udp_table, hash2); const __portpair ports = INET_COMBINED_PORTS(rmt_port, hnum); struct sock *sk; -- 2.34.1

From: Philo Lu <lulie@linux.alibaba.com> mainline inclusion from mainline-v6.13-rc1 commit dab78a1745ab3c6001e1e4d50a9d09efef8e260d category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/ICIAC2 CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i... ------------------------------------------------- Add a new hash list, hash4, in udp table. It will be used to implement 4-tuple hash for connected udp sockets. This patch adds the hlist to table, and implements helpers and the initialization. 4-tuple hash is implemented in the following patch. hash4 uses hlist_nulls to avoid moving wrongly onto another hlist due to concurrent rehash, because rehash() can happen with lookup(). Co-developed-by: Cambda Zhu <cambda@linux.alibaba.com> Signed-off-by: Cambda Zhu <cambda@linux.alibaba.com> Co-developed-by: Fred Chen <fred.cc@alibaba-inc.com> Signed-off-by: Fred Chen <fred.cc@alibaba-inc.com> Co-developed-by: Yubing Qiu <yubing.qiuyubing@alibaba-inc.com> Signed-off-by: Yubing Qiu <yubing.qiuyubing@alibaba-inc.com> Signed-off-by: Philo Lu <lulie@linux.alibaba.com> Acked-by: Willem de Bruijn <willemb@google.com> Acked-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net> Conflicts: include/linux/udp.h include/net/udp.h net/ipv4/udp.c [Mainly caused by OLK did not backport 9804985bf27f.] Signed-off-by: Liu Jian <liujian56@huawei.com> --- include/linux/udp.h | 12 +++++++ include/net/udp.h | 85 +++++++++++++++++++++++++++++++++++++++++++-- net/ipv4/udp.c | 3 +- 3 files changed, 96 insertions(+), 4 deletions(-) diff --git a/include/linux/udp.h b/include/linux/udp.h index 320d49d85484..8f75f3f6eda5 100644 --- a/include/linux/udp.h +++ b/include/linux/udp.h @@ -50,6 +50,13 @@ struct udp_sock { __u8 encap_type; /* Is this an Encapsulation socket? */ unsigned char no_check6_tx:1,/* Send zero UDP6 checksums on TX? */ no_check6_rx:1;/* Allow zero UDP6 checksums on RX? */ + +#if !IS_ENABLED(CONFIG_BASE_SMALL) + /* For UDP 4-tuple hash */ + __u16 udp_lrpa_hash; + struct hlist_nulls_node udp_lrpa_node; +#endif + /* * Following member retains the information to create a UDP header * when the socket is uncorked. @@ -121,6 +128,11 @@ static inline bool udp_get_no_check6_rx(struct sock *sk) #define udp_portaddr_for_each_entry_rcu(__sk, list) \ hlist_for_each_entry_rcu(__sk, list, __sk_common.skc_portaddr_node) +#if !IS_ENABLED(CONFIG_BASE_SMALL) +#define udp_lrpa_for_each_entry_rcu(__up, node, list) \ + hlist_nulls_for_each_entry_rcu(__up, node, list, udp_lrpa_node) +#endif + #define IS_UDPLITE(__sk) (__sk->sk_protocol == IPPROTO_UDPLITE) #endif /* _LINUX_UDP_H */ diff --git a/include/net/udp.h b/include/net/udp.h index 5a55d280f517..ad1c855ecb5f 100644 --- a/include/net/udp.h +++ b/include/net/udp.h @@ -52,14 +52,21 @@ struct udp_skb_cb { #define UDP_SKB_CB(__skb) ((struct udp_skb_cb *)((__skb)->cb)) /** - * struct udp_hslot - UDP hash slot used by udp_table.hash + * struct udp_hslot - UDP hash slot used by udp_table.hash/hash4 * * @head: head of list of sockets + * @nulls_head: head of list of sockets, only used by hash4 * @count: number of sockets in 'head' list * @lock: spinlock protecting changes to head/count */ struct udp_hslot { - struct hlist_head head; + union { + struct hlist_head head; + /* hash4 uses hlist_nulls to avoid moving wrongly onto another + * hlist, because rehash() can happen with lookup(). + */ + struct hlist_nulls_head nulls_head; + }; int count; spinlock_t lock; } __aligned(2 * sizeof(long)); @@ -84,12 +91,17 @@ struct udp_hslot_main { * * @hash: hash table, sockets are hashed on (local port) * @hash2: hash table, sockets are hashed on (local port, local address) + * @hash4: hash table, connected sockets are hashed on + * (local port, local address, remote port, remote address) * @mask: number of slots in hash tables, minus 1 * @log: log2(number of slots in hash table) */ struct udp_table { struct udp_hslot *hash; struct udp_hslot_main *hash2; +#if !IS_ENABLED(CONFIG_BASE_SMALL) + struct udp_hslot *hash4; +#endif unsigned int mask; unsigned int log; }; @@ -115,6 +127,36 @@ static inline struct udp_hslot *udp_hashslot2(struct udp_table *table, static inline void udp_table_hash4_init(struct udp_table *table) { } + +static inline struct udp_hslot *udp_hashslot4(struct udp_table *table, + unsigned int hash) +{ + BUILD_BUG(); + return NULL; +} + +static inline bool udp_hashed4(const struct sock *sk) +{ + return false; +} + +static inline unsigned int udp_hash4_slot_size(void) +{ + return 0; +} + +static inline bool udp_has_hash4(const struct udp_hslot *hslot2) +{ + return false; +} + +static inline void udp_hash4_inc(struct udp_hslot *hslot2) +{ +} + +static inline void udp_hash4_dec(struct udp_hslot *hslot2) +{ +} #else /* !CONFIG_BASE_SMALL */ /* Must be called with table->hash2 initialized */ @@ -122,8 +164,45 @@ static inline void udp_table_hash4_init(struct udp_table *table) { int i; - for (i = 0; i <= table->mask; i++) + table->hash4 = (void *)(table->hash2 + (table->mask + 1)); + for (i = 0; i <= table->mask; i++) { table->hash2[i].hash4_cnt = 0; + + INIT_HLIST_NULLS_HEAD(&table->hash4[i].nulls_head, i); + table->hash4[i].count = 0; + spin_lock_init(&table->hash4[i].lock); + } +} + +static inline struct udp_hslot *udp_hashslot4(struct udp_table *table, + unsigned int hash) +{ + return &table->hash4[hash & table->mask]; +} + +static inline bool udp_hashed4(const struct sock *sk) +{ + return !hlist_nulls_unhashed(&udp_sk(sk)->udp_lrpa_node); +} + +static inline unsigned int udp_hash4_slot_size(void) +{ + return sizeof(struct udp_hslot); +} + +static inline bool udp_has_hash4(const struct udp_hslot *hslot2) +{ + return UDP_HSLOT_MAIN(hslot2)->hash4_cnt; +} + +static inline void udp_hash4_inc(struct udp_hslot *hslot2) +{ + UDP_HSLOT_MAIN(hslot2)->hash4_cnt++; +} + +static inline void udp_hash4_dec(struct udp_hslot *hslot2) +{ + UDP_HSLOT_MAIN(hslot2)->hash4_cnt--; } #endif /* CONFIG_BASE_SMALL */ diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 4784ecd9d336..d008ded329fa 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -2917,7 +2917,8 @@ void __init udp_table_init(struct udp_table *table, const char *name) { unsigned int i, slot_size; - slot_size = sizeof(struct udp_hslot) + sizeof(struct udp_hslot_main); + slot_size = sizeof(struct udp_hslot) + sizeof(struct udp_hslot_main) + + udp_hash4_slot_size(); table->hash = alloc_large_system_hash(name, slot_size, uhash_entries, -- 2.34.1

From: Philo Lu <lulie@linux.alibaba.com> mainline inclusion from mainline-v6.13-rc1 commit 78c91ae2c6deb5d236a5a93ff2995cdd05514380 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/ICIAC2 CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i... ------------------------------------------------- Currently, the udp_table has two hash table, the port hash and portaddr hash. Usually for UDP servers, all sockets have the same local port and addr, so they are all on the same hash slot within a reuseport group. In some applications, UDP servers use connect() to manage clients. In particular, when firstly receiving from an unseen 4 tuple, a new socket is created and connect()ed to the remote addr:port, and then the fd is used exclusively by the client. Once there are connected sks in a reuseport group, udp has to score all sks in the same hash2 slot to find the best match. This could be inefficient with a large number of connections, resulting in high softirq overhead. To solve the problem, this patch implement 4-tuple hash for connected udp sockets. During connect(), hash4 slot is updated, as well as a corresponding counter, hash4_cnt, in hslot2. In __udp4_lib_lookup(), hslot4 will be searched firstly if the counter is non-zero. Otherwise, hslot2 is used like before. Note that only connected sockets enter this hash4 path, while un-connected ones are not affected. hlist_nulls is used for hash4, because we probably move to another hslot wrongly when lookup with concurrent rehash. Then we check nulls at the list end to see if we should restart lookup. Because udp does not use SLAB_TYPESAFE_BY_RCU, we don't need to touch sk_refcnt when lookup. Stress test results (with 1 cpu fully used) are shown below, in pps: (1) _un-connected_ socket as server [a] w/o hash4: 1,825176 [b] w/ hash4: 1,831750 (+0.36%) (2) 500 _connected_ sockets as server [c] w/o hash4: 290860 (only 16% of [a]) [d] w/ hash4: 1,889658 (+3.1% compared with [b]) With hash4, compute_score is skipped when lookup, so [d] is slightly better than [b]. Co-developed-by: Cambda Zhu <cambda@linux.alibaba.com> Signed-off-by: Cambda Zhu <cambda@linux.alibaba.com> Co-developed-by: Fred Chen <fred.cc@alibaba-inc.com> Signed-off-by: Fred Chen <fred.cc@alibaba-inc.com> Co-developed-by: Yubing Qiu <yubing.qiuyubing@alibaba-inc.com> Signed-off-by: Yubing Qiu <yubing.qiuyubing@alibaba-inc.com> Signed-off-by: Philo Lu <lulie@linux.alibaba.com> Acked-by: Willem de Bruijn <willemb@google.com> Acked-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net> Conflicts: net/ipv4/udp.c [Mainly caused by OLK did not backport 9804985bf27f and eda090c31fe9.] Signed-off-by: Liu Jian <liujian56@huawei.com> --- include/net/udp.h | 16 +++- net/ipv4/udp.c | 202 ++++++++++++++++++++++++++++++++++++++++++++-- net/ipv6/udp.c | 2 +- 3 files changed, 213 insertions(+), 7 deletions(-) diff --git a/include/net/udp.h b/include/net/udp.h index ad1c855ecb5f..ffa122601e23 100644 --- a/include/net/udp.h +++ b/include/net/udp.h @@ -311,13 +311,27 @@ static inline int udp_lib_hash(struct sock *sk) } void udp_lib_unhash(struct sock *sk); -void udp_lib_rehash(struct sock *sk, u16 new_hash); +void udp_lib_rehash(struct sock *sk, u16 new_hash, u16 new_hash4); static inline void udp_lib_close(struct sock *sk, long timeout) { sk_common_release(sk); } +/* hash4 routines shared between UDPv4/6 */ +#if IS_ENABLED(CONFIG_BASE_SMALL) +static inline void udp_lib_hash4(struct sock *sk, u16 hash) +{ +} + +static inline void udp4_hash4(struct sock *sk) +{ +} +#else /* !CONFIG_BASE_SMALL */ +void udp_lib_hash4(struct sock *sk, u16 hash); +void udp4_hash4(struct sock *sk); +#endif /* CONFIG_BASE_SMALL */ + int udp_lib_get_port(struct sock *sk, unsigned short snum, unsigned int hash2_nulladdr); diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index d008ded329fa..e765697b8497 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -459,6 +459,159 @@ static struct sock *udp4_lib_lookup2(struct net *net, return result; } +#if IS_ENABLED(CONFIG_BASE_SMALL) +static struct sock *udp4_lib_lookup4(const struct net *net, + __be32 saddr, __be16 sport, + __be32 daddr, unsigned int hnum, + int dif, int sdif, + struct udp_table *udptable) +{ + return NULL; +} + +static void udp_rehash4(struct udp_table *udptable, struct sock *sk, + u16 newhash4) +{ +} + +static void udp_unhash4(struct udp_table *udptable, struct sock *sk) +{ +} +#else /* !CONFIG_BASE_SMALL */ +static struct sock *udp4_lib_lookup4(const struct net *net, + __be32 saddr, __be16 sport, + __be32 daddr, unsigned int hnum, + int dif, int sdif, + struct udp_table *udptable) +{ + const __portpair ports = INET_COMBINED_PORTS(sport, hnum); + const struct hlist_nulls_node *node; + struct udp_hslot *hslot4; + unsigned int hash4, slot; + struct udp_sock *up; + struct sock *sk; + + hash4 = udp_ehashfn(net, daddr, hnum, saddr, sport); + slot = hash4 & udptable->mask; + hslot4 = &udptable->hash4[slot]; + INET_ADDR_COOKIE(acookie, saddr, daddr); + +begin: + /* SLAB_TYPESAFE_BY_RCU not used, so we don't need to touch sk_refcnt */ + udp_lrpa_for_each_entry_rcu(up, node, &hslot4->nulls_head) { + sk = (struct sock *)up; + if (INET_MATCH(sk, net, acookie, saddr, daddr, ports, dif, sdif)) + return sk; + } + + /* if the nulls value we got at the end of this lookup is not the + * expected one, we must restart lookup. We probably met an item that + * was moved to another chain due to rehash. + */ + if (get_nulls_value(node) != slot) + goto begin; + + return NULL; +} + +/* In hash4, rehash can happen in connect(), where hash4_cnt keeps unchanged. */ +static void udp_rehash4(struct udp_table *udptable, struct sock *sk, + u16 newhash4) +{ + struct udp_hslot *hslot4, *nhslot4; + + hslot4 = udp_hashslot4(udptable, udp_sk(sk)->udp_lrpa_hash); + nhslot4 = udp_hashslot4(udptable, newhash4); + udp_sk(sk)->udp_lrpa_hash = newhash4; + + if (hslot4 != nhslot4) { + spin_lock_bh(&hslot4->lock); + hlist_nulls_del_init_rcu(&udp_sk(sk)->udp_lrpa_node); + hslot4->count--; + spin_unlock_bh(&hslot4->lock); + + spin_lock_bh(&nhslot4->lock); + hlist_nulls_add_head_rcu(&udp_sk(sk)->udp_lrpa_node, + &nhslot4->nulls_head); + nhslot4->count++; + spin_unlock_bh(&nhslot4->lock); + } +} + +static void udp_unhash4(struct udp_table *udptable, struct sock *sk) +{ + struct udp_hslot *hslot2, *hslot4; + + if (udp_hashed4(sk)) { + hslot2 = udp_hashslot2(udptable, udp_sk(sk)->udp_portaddr_hash); + hslot4 = udp_hashslot4(udptable, udp_sk(sk)->udp_lrpa_hash); + + spin_lock(&hslot4->lock); + hlist_nulls_del_init_rcu(&udp_sk(sk)->udp_lrpa_node); + hslot4->count--; + spin_unlock(&hslot4->lock); + + spin_lock(&hslot2->lock); + udp_hash4_dec(hslot2); + spin_unlock(&hslot2->lock); + } +} + +void udp_lib_hash4(struct sock *sk, u16 hash) +{ + struct udp_hslot *hslot, *hslot2, *hslot4; + struct net *net = sock_net(sk); + struct udp_table *udptable; + + /* Connected udp socket can re-connect to another remote address, + * so rehash4 is needed. + */ + udptable = &udp_table; + if (udp_hashed4(sk)) { + udp_rehash4(udptable, sk, hash); + return; + } + + hslot = udp_hashslot(udptable, net, udp_sk(sk)->udp_port_hash); + hslot2 = udp_hashslot2(udptable, udp_sk(sk)->udp_portaddr_hash); + hslot4 = udp_hashslot4(udptable, hash); + udp_sk(sk)->udp_lrpa_hash = hash; + + spin_lock_bh(&hslot->lock); + if (rcu_access_pointer(sk->sk_reuseport_cb)) + reuseport_detach_sock(sk); + + spin_lock(&hslot4->lock); + hlist_nulls_add_head_rcu(&udp_sk(sk)->udp_lrpa_node, + &hslot4->nulls_head); + hslot4->count++; + spin_unlock(&hslot4->lock); + + spin_lock(&hslot2->lock); + udp_hash4_inc(hslot2); + spin_unlock(&hslot2->lock); + + spin_unlock_bh(&hslot->lock); +} +EXPORT_SYMBOL(udp_lib_hash4); + +/* call with sock lock */ +void udp4_hash4(struct sock *sk) +{ + struct net *net = sock_net(sk); + unsigned int hash; + + if (sk_unhashed(sk) || sk->sk_rcv_saddr == htonl(INADDR_ANY)) + return; + + hash = udp_ehashfn(net, sk->sk_rcv_saddr, sk->sk_num, + sk->sk_daddr, sk->sk_dport); + + udp_lib_hash4(sk, hash); +} +EXPORT_SYMBOL(udp4_hash4); +#endif /* CONFIG_BASE_SMALL */ + /* UDP is nearly always wildcards out the wazoo, it makes no sense to try * harder than this. -DaveM */ @@ -474,9 +627,17 @@ struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr, int score, badness; u32 hash = 0; + hash2 = ipv4_portaddr_hash(net, daddr, hnum); + hslot2 = udp_hashslot2(udptable, hash2); + + if (udp_has_hash4(hslot2)) { + result = udp4_lib_lookup4(net, saddr, sport, daddr, hnum, + dif, sdif, udptable); + if (result) /* udp4_lib_lookup4 return sk or NULL */ + return result; + } + if (hslot->count > 10) { - hash2 = ipv4_portaddr_hash(net, daddr, hnum); - hslot2 = udp_hashslot2(udptable, hash2); if (hslot->count < hslot2->count) goto begin; @@ -1782,6 +1943,18 @@ int udp_pre_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) } EXPORT_SYMBOL(udp_pre_connect); +static int udp_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) +{ + int res; + + lock_sock(sk); + res = __ip4_datagram_connect(sk, uaddr, addr_len); + if (!res) + udp4_hash4(sk); + release_sock(sk); + return res; +} + int __udp_disconnect(struct sock *sk, int flags) { struct inet_sock *inet = inet_sk(sk); @@ -1837,6 +2010,8 @@ void udp_lib_unhash(struct sock *sk) hlist_del_init_rcu(&udp_sk(sk)->udp_portaddr_node); hslot2->count--; spin_unlock(&hslot2->lock); + + udp_unhash4(udptable, sk); } spin_unlock_bh(&hslot->lock); } @@ -1846,7 +2021,7 @@ EXPORT_SYMBOL(udp_lib_unhash); /* * inet_rcv_saddr was changed, we must rehash secondary hash */ -void udp_lib_rehash(struct sock *sk, u16 newhash) +void udp_lib_rehash(struct sock *sk, u16 newhash, u16 newhash4) { if (sk_hashed(sk)) { struct udp_table *udptable = sk->sk_prot->h.udp_table; @@ -1878,6 +2053,19 @@ void udp_lib_rehash(struct sock *sk, u16 newhash) spin_unlock(&nhslot2->lock); } + if (udp_hashed4(sk)) { + udp_rehash4(udptable, sk, newhash4); + + if (hslot2 != nhslot2) { + spin_lock(&hslot2->lock); + udp_hash4_dec(hslot2); + spin_unlock(&hslot2->lock); + + spin_lock(&nhslot2->lock); + udp_hash4_inc(nhslot2); + spin_unlock(&nhslot2->lock); + } + } spin_unlock_bh(&hslot->lock); } } @@ -1889,7 +2077,11 @@ static void udp_v4_rehash(struct sock *sk) u16 new_hash = ipv4_portaddr_hash(sock_net(sk), inet_sk(sk)->inet_rcv_saddr, inet_sk(sk)->inet_num); - udp_lib_rehash(sk, new_hash); + u16 new_hash4 = udp_ehashfn(sock_net(sk), + sk->sk_rcv_saddr, sk->sk_num, + sk->sk_daddr, sk->sk_dport); + + udp_lib_rehash(sk, new_hash, new_hash4); } static int __udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb) @@ -2694,7 +2886,7 @@ struct proto udp_prot = { .owner = THIS_MODULE, .close = udp_lib_close, .pre_connect = udp_pre_connect, - .connect = ip4_datagram_connect, + .connect = udp_connect, .disconnect = udp_disconnect, .ioctl = udp_ioctl, .init = udp_init_sock, diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index 48b48be8b167..8436635e5e77 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -120,7 +120,7 @@ static void udp_v6_rehash(struct sock *sk) &sk->sk_v6_rcv_saddr, inet_sk(sk)->inet_num); - udp_lib_rehash(sk, new_hash); + udp_lib_rehash(sk, new_hash, 0); /* 4-tuple hash not implemented */ } static int compute_score(struct sock *sk, struct net *net, -- 2.34.1

From: Philo Lu <lulie@linux.alibaba.com> mainline inclusion from mainline-v6.13-rc1 commit 1b29a730ef8b6fd3aa3e11c2f6d409cf201cd913 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/ICIAC2 CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i... ------------------------------------------------- Implement ipv6 udp hash4 like that in ipv4. The major difference is that the hash value should be calculated with udp6_ehashfn(). Besides, ipv4-mapped ipv6 address is handled before hash() and rehash(). Export udp_ehashfn because now we use it in udpv6 rehash. Core procedures of hash/unhash/rehash are same as ipv4, and udpv4 and udpv6 share the same udptable, so some functions in ipv4 hash4 can also be shared. Co-developed-by: Cambda Zhu <cambda@linux.alibaba.com> Signed-off-by: Cambda Zhu <cambda@linux.alibaba.com> Co-developed-by: Fred Chen <fred.cc@alibaba-inc.com> Signed-off-by: Fred Chen <fred.cc@alibaba-inc.com> Co-developed-by: Yubing Qiu <yubing.qiuyubing@alibaba-inc.com> Signed-off-by: Yubing Qiu <yubing.qiuyubing@alibaba-inc.com> Signed-off-by: Philo Lu <lulie@linux.alibaba.com> Acked-by: Willem de Bruijn <willemb@google.com> Acked-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net> Conflicts: net/ipv4/udp.c net/ipv6/udp.c [Mainly caused by OLK did not backport 0f495f761722 and 9804985bf27f.] Signed-off-by: Liu Jian <liujian56@huawei.com> --- include/net/udp.h | 2 + net/ipv4/udp.c | 7 +-- net/ipv6/udp.c | 106 ++++++++++++++++++++++++++++++++++++++++++++-- 3 files changed, 108 insertions(+), 7 deletions(-) diff --git a/include/net/udp.h b/include/net/udp.h index ffa122601e23..fdcd6dcfae1d 100644 --- a/include/net/udp.h +++ b/include/net/udp.h @@ -312,6 +312,8 @@ static inline int udp_lib_hash(struct sock *sk) void udp_lib_unhash(struct sock *sk); void udp_lib_rehash(struct sock *sk, u16 new_hash, u16 new_hash4); +u32 udp_ehashfn(const struct net *net, const __be32 laddr, const __u16 lport, + const __be32 faddr, const __be16 fport); static inline void udp_lib_close(struct sock *sk, long timeout) { diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index e765697b8497..db3c0f67c8fc 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -410,9 +410,9 @@ static int compute_score(struct sock *sk, struct net *net, return score; } -static u32 udp_ehashfn(const struct net *net, const __be32 laddr, - const __u16 lport, const __be32 faddr, - const __be16 fport) +u32 udp_ehashfn(const struct net *net, const __be32 laddr, + const __u16 lport, const __be32 faddr, + const __be16 fport) { static u32 udp_ehash_secret __read_mostly; @@ -421,6 +421,7 @@ static u32 udp_ehashfn(const struct net *net, const __be32 laddr, return __inet_ehashfn(laddr, lport, faddr, fport, udp_ehash_secret + net_hash_mix(net)); } +EXPORT_SYMBOL(udp_ehashfn); /* called with rcu_read_lock() */ static struct sock *udp4_lib_lookup2(struct net *net, diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index 8436635e5e77..e66ef7ff0912 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -119,8 +119,19 @@ static void udp_v6_rehash(struct sock *sk) u16 new_hash = ipv6_portaddr_hash(sock_net(sk), &sk->sk_v6_rcv_saddr, inet_sk(sk)->inet_num); + u16 new_hash4; - udp_lib_rehash(sk, new_hash, 0); /* 4-tuple hash not implemented */ + if (ipv6_addr_v4mapped(&sk->sk_v6_rcv_saddr)) { + new_hash4 = udp_ehashfn(sock_net(sk), + sk->sk_rcv_saddr, sk->sk_num, + sk->sk_daddr, sk->sk_dport); + } else { + new_hash4 = udp6_ehashfn(sock_net(sk), + &sk->sk_v6_rcv_saddr, sk->sk_num, + &sk->sk_v6_daddr, sk->sk_dport); + } + + udp_lib_rehash(sk, new_hash, new_hash4); } static int compute_score(struct sock *sk, struct net *net, @@ -206,6 +217,74 @@ static struct sock *udp6_lib_lookup2(struct net *net, return result; } +#if IS_ENABLED(CONFIG_BASE_SMALL) +static struct sock *udp6_lib_lookup4(const struct net *net, + const struct in6_addr *saddr, __be16 sport, + const struct in6_addr *daddr, + unsigned int hnum, int dif, int sdif, + struct udp_table *udptable) +{ + return NULL; +} + +static void udp6_hash4(struct sock *sk) +{ +} +#else /* !CONFIG_BASE_SMALL */ +static struct sock *udp6_lib_lookup4(const struct net *net, + const struct in6_addr *saddr, __be16 sport, + const struct in6_addr *daddr, + unsigned int hnum, int dif, int sdif, + struct udp_table *udptable) +{ + const __portpair ports = INET_COMBINED_PORTS(sport, hnum); + const struct hlist_nulls_node *node; + struct udp_hslot *hslot4; + unsigned int hash4, slot; + struct udp_sock *up; + struct sock *sk; + + hash4 = udp6_ehashfn(net, daddr, hnum, saddr, sport); + slot = hash4 & udptable->mask; + hslot4 = &udptable->hash4[slot]; + +begin: + udp_lrpa_for_each_entry_rcu(up, node, &hslot4->nulls_head) { + sk = (struct sock *)up; + if (INET6_MATCH(sk, net, saddr, daddr, ports, dif, sdif)) + return sk; + } + + /* if the nulls value we got at the end of this lookup is not the + * expected one, we must restart lookup. We probably met an item that + * was moved to another chain due to rehash. + */ + if (get_nulls_value(node) != slot) + goto begin; + + return NULL; +} + +static void udp6_hash4(struct sock *sk) +{ + struct net *net = sock_net(sk); + unsigned int hash; + + if (ipv6_addr_v4mapped(&sk->sk_v6_rcv_saddr)) { + udp4_hash4(sk); + return; + } + + if (sk_unhashed(sk) || ipv6_addr_any(&sk->sk_v6_rcv_saddr)) + return; + + hash = udp6_ehashfn(net, &sk->sk_v6_rcv_saddr, sk->sk_num, + &sk->sk_v6_daddr, sk->sk_dport); + + udp_lib_hash4(sk, hash); +} +#endif /* CONFIG_BASE_SMALL */ + /* rcu_read_lock() must be held */ struct sock *__udp6_lib_lookup(struct net *net, const struct in6_addr *saddr, __be16 sport, @@ -221,9 +300,16 @@ struct sock *__udp6_lib_lookup(struct net *net, int score, badness; u32 hash = 0; + hash2 = ipv6_portaddr_hash(net, daddr, hnum); + hslot2 = udp_hashslot2(udptable, hash2); + if (udp_has_hash4(hslot2)) { + result = udp6_lib_lookup4(net, saddr, sport, daddr, hnum, + dif, sdif, udptable); + if (result) /* udp6_lib_lookup4 return sk or NULL */ + return result; + } + if (hslot->count > 10) { - hash2 = ipv6_portaddr_hash(net, daddr, hnum); - hslot2 = udp_hashslot2(udptable, hash2); if (hslot->count < hslot2->count) goto begin; @@ -1003,6 +1089,18 @@ static int udpv6_pre_connect(struct sock *sk, struct sockaddr *uaddr, return BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr); } +static int udpv6_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) +{ + int res; + + lock_sock(sk); + res = __ip6_datagram_connect(sk, uaddr, addr_len); + if (!res) + udp6_hash4(sk); + release_sock(sk); + return res; +} + /** * udp6_hwcsum_outgoing - handle outgoing HW checksumming * @sk: socket we are sending on @@ -1601,7 +1699,7 @@ struct proto udpv6_prot = { .owner = THIS_MODULE, .close = udp_lib_close, .pre_connect = udpv6_pre_connect, - .connect = ip6_datagram_connect, + .connect = udpv6_connect, .disconnect = udp_disconnect, .ioctl = udp_ioctl, .init = udpv6_init_sock, -- 2.34.1

From: Paolo Abeni <pabeni@redhat.com> mainline inclusion from mainline-v6.13-rc3 commit 51a00be6a0994da2ba6b4ace3b7a0d9373b4b25e category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/ICIAC2 CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i... ------------------------------------------------- After the blamed commit below, udp_rehash() is supposed to be called with both local and remote addresses set. Currently that is already the case for IPv6 sockets, but for IPv4 the destination address is updated after rehashing. Address the issue moving the destination address and port initialization before rehashing. Fixes: 1b29a730ef8b ("ipv6/udp: Add 4-tuple hash for connected socket") Reviewed-by: Eric Dumazet <edumazet@google.com> Link: https://patch.msgid.link/4761e466ab9f7542c68cdc95f248987d127044d2.1733499715... Signed-off-by: Paolo Abeni <pabeni@redhat.com> Conflicts: net/ipv4/datagram.c [Caused by OLK did not backport 69421bf98482.] Signed-off-by: Liu Jian <liujian56@huawei.com> --- net/ipv4/datagram.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/net/ipv4/datagram.c b/net/ipv4/datagram.c index dc8517c2cdc7..7f2d12d384fb 100644 --- a/net/ipv4/datagram.c +++ b/net/ipv4/datagram.c @@ -65,15 +65,17 @@ int __ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len err = -EACCES; goto out; } + + /* Update addresses before rehashing */ + inet->inet_daddr = fl4->daddr; + inet->inet_dport = usin->sin_port; if (!inet->inet_saddr) - inet->inet_saddr = fl4->saddr; /* Update source address */ + inet->inet_saddr = fl4->saddr; if (!inet->inet_rcv_saddr) { inet->inet_rcv_saddr = fl4->saddr; if (sk->sk_prot->rehash) sk->sk_prot->rehash(sk); } - inet->inet_daddr = fl4->daddr; - inet->inet_dport = usin->sin_port; reuseport_has_conns(sk, true); sk->sk_state = TCP_ESTABLISHED; sk_set_txhash(sk); -- 2.34.1

From: Philo Lu <lulie@linux.alibaba.com> mainline inclusion from mainline-v6.13 commit 644f9108f3a505022ef43510e5143cb985e0cf8b category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/ICIAC2 CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i... ------------------------------------------------- As discussed in [0], rehash4 could be missed in udp_lib_rehash() when udp hash4 changes while hash2 doesn't change. This patch fixes this by moving rehash4 codes out of rehash2 checking, and then rehash2 and rehash4 are done separately. By doing this, we no longer need to call rehash4 explicitly in udp_lib_hash4(), as the rehash callback in __ip4_datagram_connect takes it. Thus, now udp_lib_hash4() returns directly if the sk is already hashed. Note that uhash4 may fail to work under consecutive connect(<dst address>) calls because rehash() is not called with every connect(). To overcome this, connect(<AF_UNSPEC>) needs to be called after the next connect to a new destination. [0] https://lore.kernel.org/all/4761e466ab9f7542c68cdc95f248987d127044d2.1733499... Fixes: 78c91ae2c6de ("ipv4/udp: Add 4-tuple hash for connected socket") Suggested-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Philo Lu <lulie@linux.alibaba.com> Link: https://patch.msgid.link/20250110010810.107145-1-lulie@linux.alibaba.com Signed-off-by: Paolo Abeni <pabeni@redhat.com> Conflicts: net/ipv4/udp.c [Caused by OLK did not backport 9804985bf27f.] Signed-off-by: Liu Jian <liujian56@huawei.com> --- net/ipv4/udp.c | 46 +++++++++++++++++++++++++++------------------- 1 file changed, 27 insertions(+), 19 deletions(-) diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index db3c0f67c8fc..3f3a10f9a32d 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -515,7 +515,7 @@ static struct sock *udp4_lib_lookup4(const struct net *net, return NULL; } -/* In hash4, rehash can happen in connect(), where hash4_cnt keeps unchanged. */ +/* udp_rehash4() only checks hslot4, and hash4_cnt is not processed. */ static void udp_rehash4(struct udp_table *udptable, struct sock *sk, u16 newhash4) { @@ -564,15 +564,13 @@ void udp_lib_hash4(struct sock *sk, u16 hash) struct net *net = sock_net(sk); struct udp_table *udptable; - /* Connected udp socket can re-connect to another remote address, - * so rehash4 is needed. + /* Connected udp socket can re-connect to another remote address, which + * will be handled by rehash. Thus no need to redo hash4 here. */ - udptable = &udp_table; - if (udp_hashed4(sk)) { - udp_rehash4(udptable, sk, hash); + if (udp_hashed4(sk)) return; - } + udptable = &udp_table; hslot = udp_hashslot(udptable, net, udp_sk(sk)->udp_port_hash); hslot2 = udp_hashslot2(udptable, udp_sk(sk)->udp_portaddr_hash); hslot4 = udp_hashslot4(udptable, hash); @@ -2028,14 +2026,14 @@ void udp_lib_rehash(struct sock *sk, u16 newhash, u16 newhash4) struct udp_table *udptable = sk->sk_prot->h.udp_table; struct udp_hslot *hslot, *hslot2, *nhslot2; + hslot = udp_hashslot(udptable, sock_net(sk), + udp_sk(sk)->udp_port_hash); hslot2 = udp_hashslot2(udptable, udp_sk(sk)->udp_portaddr_hash); nhslot2 = udp_hashslot2(udptable, newhash); udp_sk(sk)->udp_portaddr_hash = newhash; if (hslot2 != nhslot2 || rcu_access_pointer(sk->sk_reuseport_cb)) { - hslot = udp_hashslot(udptable, sock_net(sk), - udp_sk(sk)->udp_port_hash); /* we must lock primary chain too */ spin_lock_bh(&hslot->lock); if (rcu_access_pointer(sk->sk_reuseport_cb)) @@ -2054,19 +2052,29 @@ void udp_lib_rehash(struct sock *sk, u16 newhash, u16 newhash4) spin_unlock(&nhslot2->lock); } - if (udp_hashed4(sk)) { - udp_rehash4(udptable, sk, newhash4); + spin_unlock_bh(&hslot->lock); + } - if (hslot2 != nhslot2) { - spin_lock(&hslot2->lock); - udp_hash4_dec(hslot2); - spin_unlock(&hslot2->lock); + /* Now process hash4 if necessary: + * (1) update hslot4; + * (2) update hslot2->hash4_cnt. + * Note that hslot2/hslot4 should be checked separately, as + * either of them may change with the other unchanged. + */ + if (udp_hashed4(sk)) { + spin_lock_bh(&hslot->lock); - spin_lock(&nhslot2->lock); - udp_hash4_inc(nhslot2); - spin_unlock(&nhslot2->lock); - } + udp_rehash4(udptable, sk, newhash4); + if (hslot2 != nhslot2) { + spin_lock(&hslot2->lock); + udp_hash4_dec(hslot2); + spin_unlock(&hslot2->lock); + + spin_lock(&nhslot2->lock); + udp_hash4_inc(nhslot2); + spin_unlock(&nhslot2->lock); } + spin_unlock_bh(&hslot->lock); } } -- 2.34.1
participants (2)
-
Liu Jian
-
patchwork bot