kernel-fxtec-pro1x/net/dccp
Eric Dumazet 3b24d854cb tcp/dccp: do not touch listener sk_refcnt under synflood
When a SYNFLOOD targets a non SO_REUSEPORT listener, multiple
cpus contend on sk->sk_refcnt and sk->sk_wmem_alloc changes.

By letting listeners use SOCK_RCU_FREE infrastructure,
we can relax TCP_LISTEN lookup rules and avoid touching sk_refcnt

Note that we still use SLAB_DESTROY_BY_RCU rules for other sockets,
only listeners are impacted by this change.

Peak performance under SYNFLOOD is increased by ~33% :

On my test machine, I could process 3.2 Mpps instead of 2.4 Mpps

Most consuming functions are now skb_set_owner_w() and sock_wfree()
contending on sk->sk_wmem_alloc when cooking SYNACK and freeing them.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-04-04 22:11:20 -04:00
..
ccids
ackvec.c dccp: drop null test before destroy functions 2015-09-15 16:49:43 -07:00
ackvec.h
ccid.c dccp: drop null test before destroy functions 2015-09-15 16:49:43 -07:00
ccid.h
dccp.h tcp/dccp: fix hashdance race for passive sessions 2015-10-23 05:42:21 -07:00
diag.c sock_diag: specify info_size per inet protocol 2015-06-15 19:49:22 -07:00
feat.c dccp: kerneldoc warning fixes 2014-11-18 15:26:31 -05:00
feat.h
input.c dccp: spelling s/reseting/resetting 2014-11-18 15:26:32 -05:00
ipv4.c tcp/dccp: do not touch listener sk_refcnt under synflood 2016-04-04 22:11:20 -04:00
ipv6.c tcp/dccp: do not touch listener sk_refcnt under synflood 2016-04-04 22:11:20 -04:00
ipv6.h
Kconfig
Makefile
minisocks.c tcp/dccp: fix hashdance race for passive sessions 2015-10-23 05:42:21 -07:00
options.c
output.c net: Generalise wq_has_sleeper helper 2015-11-30 14:47:33 -05:00
probe.c Use 64-bit timekeeping 2015-11-01 17:01:16 -05:00
proto.c net: rename SOCK_ASYNC_NOSPACE and SOCK_ASYNC_WAITDATA 2015-12-01 15:45:05 -05:00
qpolicy.c
sysctl.c
timer.c inet: get rid of central tcp/dccp listener timer 2015-03-20 12:40:25 -04:00