From eaa72dc47488d599439cd0fd0f8c4f1bcb3906bb Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 29 Aug 2017 15:16:01 -0700 Subject: [PATCH] neigh: increase queue_len_bytes to match wmem_default Florian reported UDP xmit drops that could be root caused to the too small neigh limit. Current limit is 64 KB, meaning that even a single UDP socket would hit it, since its default sk_sndbuf comes from net.core.wmem_default (~212992 bytes on 64bit arches). Once ARP/ND resolution is in progress, we should allow a little more packets to be queued, at least for one producer. Once neigh arp_queue is filled, a rogue socket should hit its sk_sndbuf limit and either block in sendmsg() or return -EAGAIN. Signed-off-by: Eric Dumazet Reported-by: Florian Fainelli Signed-off-by: David S. Miller --- Documentation/networking/ip-sysctl.txt | 7 +++++-- include/net/sock.h | 10 ++++++++++ net/core/sock.c | 10 ---------- net/decnet/dn_neigh.c | 2 +- net/ipv4/arp.c | 2 +- net/ipv4/tcp_input.c | 2 +- net/ipv6/ndisc.c | 2 +- 7 files changed, 19 insertions(+), 16 deletions(-) diff --git a/Documentation/networking/ip-sysctl.txt b/Documentation/networking/ip-sysctl.txt index 6b0bc0f71534..b3345d0fe0a6 100644 --- a/Documentation/networking/ip-sysctl.txt +++ b/Documentation/networking/ip-sysctl.txt @@ -109,7 +109,10 @@ neigh/default/unres_qlen_bytes - INTEGER queued for each unresolved address by other network layers. (added in linux 3.3) Setting negative value is meaningless and will return error. - Default: 65536 Bytes(64KB) + Default: SK_WMEM_MAX, (same as net.core.wmem_default). + Exact value depends on architecture and kernel options, + but should be enough to allow queuing 256 packets + of medium size. neigh/default/unres_qlen - INTEGER The maximum number of packets which may be queued for each @@ -119,7 +122,7 @@ neigh/default/unres_qlen - INTEGER unexpected packet loss. The current default value is calculated according to default value of unres_qlen_bytes and true size of packet. - Default: 31 + Default: 101 mtu_expires - INTEGER Time, in seconds, that cached PMTU information is kept. diff --git a/include/net/sock.h b/include/net/sock.h index 1c2912d433e8..03a362568357 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -2368,6 +2368,16 @@ bool sk_net_capable(const struct sock *sk, int cap); void sk_get_meminfo(const struct sock *sk, u32 *meminfo); +/* Take into consideration the size of the struct sk_buff overhead in the + * determination of these values, since that is non-constant across + * platforms. This makes socket queueing behavior and performance + * not depend upon such differences. + */ +#define _SK_MEM_PACKETS 256 +#define _SK_MEM_OVERHEAD SKB_TRUESIZE(256) +#define SK_WMEM_MAX (_SK_MEM_OVERHEAD * _SK_MEM_PACKETS) +#define SK_RMEM_MAX (_SK_MEM_OVERHEAD * _SK_MEM_PACKETS) + extern __u32 sysctl_wmem_max; extern __u32 sysctl_rmem_max; diff --git a/net/core/sock.c b/net/core/sock.c index dfdd14cac775..9b7b6bbb2a23 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -307,16 +307,6 @@ static struct lock_class_key af_wlock_keys[AF_MAX]; static struct lock_class_key af_elock_keys[AF_MAX]; static struct lock_class_key af_kern_callback_keys[AF_MAX]; -/* Take into consideration the size of the struct sk_buff overhead in the - * determination of these values, since that is non-constant across - * platforms. This makes socket queueing behavior and performance - * not depend upon such differences. - */ -#define _SK_MEM_PACKETS 256 -#define _SK_MEM_OVERHEAD SKB_TRUESIZE(256) -#define SK_WMEM_MAX (_SK_MEM_OVERHEAD * _SK_MEM_PACKETS) -#define SK_RMEM_MAX (_SK_MEM_OVERHEAD * _SK_MEM_PACKETS) - /* Run time adjustable parameters. */ __u32 sysctl_wmem_max __read_mostly = SK_WMEM_MAX; EXPORT_SYMBOL(sysctl_wmem_max); diff --git a/net/decnet/dn_neigh.c b/net/decnet/dn_neigh.c index 21dedf6fd0f7..22bf0b95d6ed 100644 --- a/net/decnet/dn_neigh.c +++ b/net/decnet/dn_neigh.c @@ -94,7 +94,7 @@ struct neigh_table dn_neigh_table = { [NEIGH_VAR_BASE_REACHABLE_TIME] = 30 * HZ, [NEIGH_VAR_DELAY_PROBE_TIME] = 5 * HZ, [NEIGH_VAR_GC_STALETIME] = 60 * HZ, - [NEIGH_VAR_QUEUE_LEN_BYTES] = 64*1024, + [NEIGH_VAR_QUEUE_LEN_BYTES] = SK_WMEM_MAX, [NEIGH_VAR_PROXY_QLEN] = 0, [NEIGH_VAR_ANYCAST_DELAY] = 0, [NEIGH_VAR_PROXY_DELAY] = 0, diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c index 8b52179ddc6e..7c45b8896709 100644 --- a/net/ipv4/arp.c +++ b/net/ipv4/arp.c @@ -171,7 +171,7 @@ struct neigh_table arp_tbl = { [NEIGH_VAR_BASE_REACHABLE_TIME] = 30 * HZ, [NEIGH_VAR_DELAY_PROBE_TIME] = 5 * HZ, [NEIGH_VAR_GC_STALETIME] = 60 * HZ, - [NEIGH_VAR_QUEUE_LEN_BYTES] = 64 * 1024, + [NEIGH_VAR_QUEUE_LEN_BYTES] = SK_WMEM_MAX, [NEIGH_VAR_PROXY_QLEN] = 64, [NEIGH_VAR_ANYCAST_DELAY] = 1 * HZ, [NEIGH_VAR_PROXY_DELAY] = (8 * HZ) / 10, diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 568ccfd6dd37..7616cd76f6f6 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -6086,9 +6086,9 @@ int tcp_conn_request(struct request_sock_ops *rsk_ops, struct tcp_sock *tp = tcp_sk(sk); struct net *net = sock_net(sk); struct sock *fastopen_sk = NULL; - struct dst_entry *dst = NULL; struct request_sock *req; bool want_cookie = false; + struct dst_entry *dst; struct flowi fl; /* TW buckets are converted to open requests without diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c index 5e338eb89509..266a530414d7 100644 --- a/net/ipv6/ndisc.c +++ b/net/ipv6/ndisc.c @@ -127,7 +127,7 @@ struct neigh_table nd_tbl = { [NEIGH_VAR_BASE_REACHABLE_TIME] = ND_REACHABLE_TIME, [NEIGH_VAR_DELAY_PROBE_TIME] = 5 * HZ, [NEIGH_VAR_GC_STALETIME] = 60 * HZ, - [NEIGH_VAR_QUEUE_LEN_BYTES] = 64 * 1024, + [NEIGH_VAR_QUEUE_LEN_BYTES] = SK_WMEM_MAX, [NEIGH_VAR_PROXY_QLEN] = 64, [NEIGH_VAR_ANYCAST_DELAY] = 1 * HZ, [NEIGH_VAR_PROXY_DELAY] = (8 * HZ) / 10, -- 2.20.1