Eric Dumazet
fa76ce7328
inet: get rid of central tcp/dccp listener timer ...
One of the major issue for TCP is the SYNACK rtx handling,
done by inet_csk_reqsk_queue_prune(), fired by the keepalive
timer of a TCP_LISTEN socket.
This function runs for awful long times, with socket lock held,
meaning that other cpus needing this lock have to spin for hundred of ms.
SYNACK are sent in huge bursts, likely to cause severe drops anyway.
This model was OK 15 years ago when memory was very tight.
We now can afford to have a timer per request sock.
Timer invocations no longer need to lock the listener,
and can be run from all cpus in parallel.
With following patch increasing somaxconn width to 32 bits,
I tested a listener with more than 4 million active request sockets,
and a steady SYNFLOOD of ~200,000 SYN per second.
Host was sending ~830,000 SYNACK per second.
This is ~100 times more what we could achieve before this patch.
Later, we will get rid of the listener hash and use ehash instead.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-03-20 12:40:25 -04:00
netfilter: fix sparse warnings in reject handling
2015-03-10 15:01:32 +01:00
net: Remove iocb argument from sendmsg and recvmsg
2015-03-02 13:06:31 -05:00
ipsec: Remove obsolete MAX_AH_AUTH_LEN
2014-09-18 10:54:36 +02:00
neigh: Factor out ___neigh_lookup_noref
2015-03-04 00:23:23 -05:00
cipso: don't use IPCB() to locate the CIPSO IP option
2015-02-11 14:46:37 -05:00
net: Save TX flow hash in sock and set in skbuf on xmit
2014-07-07 21:14:21 -07:00
ipv4, ipv6: kill ip_mc_{join, leave}_group and ipv6_sock_mc_{join, drop}
2015-03-18 22:05:09 -04:00
net: esp: Convert NETDEBUG to pr_info
2014-11-06 15:11:10 -05:00
fib_trie: Avoid NULL pointer if local table is not allocated
2015-03-12 18:26:51 -04:00
ipv4: FIB Local/MAIN table collapse
2015-03-11 16:22:14 -04:00
ipv4: FIB Local/MAIN table collapse
2015-03-11 16:22:14 -04:00
net: Kill hold_net release_net
2015-03-12 14:39:40 -04:00
fib_trie: Provide a deterministic order for fib_alias w/ tables merged
2015-03-12 18:26:51 -04:00
gue: Use checksum partial with remote checksum offload
2015-02-11 15:12:13 -08:00
openvswitch: Add support for checksums on UDP tunnels.
2015-01-28 23:04:15 -08:00
net: Fix GRE RX to use skb_transport_header for GRE header offset
2014-09-08 15:23:05 -07:00
gre: Set inner mac header in gro complete
2014-12-05 21:18:34 -08:00
ipv4: icmp: use percpu allocation
2015-01-31 17:48:18 -08:00
ipv4, ipv6: kill ip_mc_{join, leave}_group and ipv6_sock_mc_{join, drop}
2015-03-18 22:05:09 -04:00
inet: get rid of central tcp/dccp listener timer
2015-03-20 12:40:25 -04:00
inet: get rid of central tcp/dccp listener timer
2015-03-20 12:40:25 -04:00
net: Convert LIMIT_NETDEBUG to net_dbg_ratelimited
2014-11-11 14:10:31 -05:00
inet: get rid of last __inet_hash_connect() argument
2015-03-18 22:00:35 -04:00
lro: remove dead code
2013-12-29 16:34:25 -05:00
inet: add a schedule point in inet_twsk_purge()
2015-03-18 22:38:13 -04:00
inet: remove dead inetpeer sequence code
2014-09-08 16:42:42 -07:00
ipv4: try to cache dst_entries which would cause a redirect
2015-01-26 17:28:27 -08:00
ipv4: ip_check_defrag should not assume that skb_network_offset is zero
2015-03-05 21:43:48 -05:00
gre/ipip: use be16 variants of netlink functions
2015-02-08 16:28:06 -08:00
net: Fix memory leak if TPROXY used with TCP early demux
2014-01-27 16:22:11 -08:00
ipv4: rename ip_options_echo to __ip_options_echo()
2014-09-28 16:35:42 -04:00
udp: only allow UFO for packets from SOCK_DGRAM sockets
2015-03-02 22:19:29 -05:00
ipv4, ipv6: kill ip_mc_{join, leave}_group and ipv6_sock_mc_{join, drop}
2015-03-18 22:05:09 -04:00
tunnels: advertise link netns via netlink
2015-01-19 14:32:03 -05:00
ipv4: fix a potential use after free in ip_tunnel_core.c
2014-10-17 23:45:26 -04:00
tunnels: advertise link netns via netlink
2015-01-19 14:32:03 -05:00
ipcomp4: Use the IPsec protocol multiplexer API
2014-02-25 07:04:17 +01:00
net: ipv4: handle DSA enabled master network devices
2015-01-19 15:45:10 -05:00
gre/ipip: use be16 variants of netlink functions
2015-02-08 16:28:06 -08:00
net: Introduce possible_net_t
2015-03-12 14:39:40 -04:00
net: Move fou_build_header into fou.c and refactor
2014-11-05 16:30:02 -05:00
net: Add Geneve tunneling protocol driver
2014-10-06 00:32:20 -04:00
netfilter: remove double colon
2014-02-19 11:41:25 +01:00
netns: constify net_hash_mix() and various callers
2015-03-18 22:00:34 -04:00
tcp: helpers to mitigate ACK loops by rate-limiting out-of-window dupacks
2015-02-08 01:03:12 -08:00
net: Export inet_offloads and inet6_offloads
2014-09-19 17:15:31 -04:00
net: Remove iocb argument from sendmsg and recvmsg
2015-03-02 13:06:31 -05:00
net: Remove protocol from struct dst_ops
2015-03-09 16:06:10 -04:00
inet: get rid of central tcp/dccp listener timer
2015-03-20 12:40:25 -04:00
ipv4: Create probe timer for tcp PMTU as per RFC4821
2015-03-06 14:57:42 -05:00
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
2015-03-09 23:38:02 -04:00
tcp: stretch ACK fixes prep
2015-01-28 22:18:37 -08:00
tcp: silence registration message
2015-02-20 15:04:03 -05:00
tcp: fix timing issue in CUBIC slope calculation
2015-01-28 22:18:38 -08:00
net: tcp: add DCTCP congestion control algorithm
2014-09-29 00:13:10 -04:00
inet_diag: add const to inet_diag_req_v2
2015-03-10 13:45:28 -04:00
inet: get rid of central tcp/dccp listener timer
2015-03-20 12:40:25 -04:00
tcp: whitespace fixes
2014-09-01 18:12:45 -07:00
tcp: whitespace fixes
2014-09-01 18:12:45 -07:00
tcp: whitespace fixes
2014-09-01 18:12:45 -07:00
tcp: whitespace fixes
2014-09-01 18:12:45 -07:00
inet: drop prev pointer handling in request sock
2015-03-20 12:40:25 -04:00
inet: get rid of central tcp/dccp listener timer
2015-03-20 12:40:25 -04:00
tcp: remove in_flight parameter from cong_avoid() methods
2014-05-03 19:23:07 -04:00
memcg: cleanup static keys decrement
2015-02-12 18:54:10 -08:00
tcp_metrics: fix wrong lockdep annotations
2015-03-16 16:32:23 -04:00
inet: get rid of central tcp/dccp listener timer
2015-03-20 12:40:25 -04:00
tcp: cleanup static functions
2015-02-28 16:56:51 -05:00
ipv4: Create probe timer for tcp PMTU as per RFC4821
2015-03-06 14:57:42 -05:00
tcp: whitespace fixes
2014-09-01 18:12:45 -07:00
tcp: stretch ACK fixes prep
2015-01-28 22:18:37 -08:00
inet: get rid of central tcp/dccp listener timer
2015-03-20 12:40:25 -04:00
tcp: whitespace fixes
2014-09-01 18:12:45 -07:00
net: ipv4/ipv6: Remove extern from function prototypes
2013-10-19 19:12:11 -04:00
tcp: stretch ACK fixes prep
2015-01-28 22:18:37 -08:00
net: tcp: split ack slow/fast events from cwnd_event
2014-09-29 00:13:10 -04:00
tcp: stretch ACK fixes prep
2015-01-28 22:18:37 -08:00
netns: constify net_hash_mix() and various callers
2015-03-18 22:00:34 -04:00
inet_diag: add const to inet_diag_req_v2
2015-03-10 13:45:28 -04:00
net: Remove iocb argument from sendmsg and recvmsg
2015-03-02 13:06:31 -05:00
udp: Set SKB_GSO_UDP_TUNNEL* in UDP GRO path
2015-02-11 15:12:10 -08:00
udp: Do not require sock in udp_tunnel_xmit_skb
2015-01-24 23:15:40 -08:00
net: Eliminate no_check from protosw
2014-05-23 16:28:53 -04:00
xfrm4: Add IPsec protocol multiplexer
2014-02-25 07:04:16 +01:00
ipv4: ERROR: code indent should use tabs where possible
2013-12-26 13:43:21 -05:00
inetpeer: get rid of ip_id_count
2014-06-02 11:00:41 -07:00
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
2014-05-24 00:32:30 -04:00
net: Remove protocol from struct dst_ops
2015-03-09 16:06:10 -04:00
xfrm4: Remove duplicate semicolon
2014-06-30 07:49:47 +02:00
inet: make no_pmtu_disc per namespace and kill ipv4_config
2013-12-18 16:58:20 -05:00
sit: add IPv4 over IPv4 support
2013-05-31 17:19:05 -07:00
fa76ce7328