linux/net/core
Eric Dumazet 217f697436 net: busy-poll: allow preemption in sk_busy_loop()
After commit 4cd13c21b2 ("softirq: Let ksoftirqd do its job"),
sk_busy_loop() needs a bit of care :
softirqs might be delayed since we do not allow preemption yet.

This patch adds preemptiom points in sk_busy_loop(),
and makes sure no unnecessary cache line dirtying
or atomic operations are done while looping.

A new flag is added into napi->state : NAPI_STATE_IN_BUSY_POLL

This prevents napi_complete_done() from clearing NAPIF_STATE_SCHED,
so that sk_busy_loop() does not have to grab it again.

Similarly, netpoll_poll_lock() is done one time.

This gives about 10 to 20 % improvement in various busy polling
tests, especially when many threads are busy polling in
configurations with large number of NIC queues.

This should allow experimenting with bigger delays without
hurting overall latencies.

Tested:
 On a 40Gb mlx4 NIC, 32 RX/TX queues.

 echo 70 >/proc/sys/net/core/busy_read
 for i in `seq 1 40`; do echo -n $i: ; ./super_netperf $i -H lpaa24 -t UDP_RR -- -N -n; done

    Before:      After:
 1:   90072   92819
 2:  157289  184007
 3:  235772  213504
 4:  344074  357513
 5:  394755  458267
 6:  461151  487819
 7:  549116  625963
 8:  544423  716219
 9:  720460  738446
10:  794686  837612
11:  915998  923960
12:  937507  925107
13: 1019677  971506
14: 1046831 1113650
15: 1114154 1148902
16: 1105221 1179263
17: 1266552 1299585
18: 1258454 1383817
19: 1341453 1312194
20: 1363557 1488487
21: 1387979 1501004
22: 1417552 1601683
23: 1550049 1642002
24: 1568876 1601915
25: 1560239 1683607
26: 1640207 1745211
27: 1706540 1723574
28: 1638518 1722036
29: 1734309 1757447
30: 1782007 1855436
31: 1724806 1888539
32: 1717716 1944297
33: 1778716 1869118
34: 1805738 1983466
35: 1815694 2020758
36: 1893059 2035632
37: 1843406 2034653
38: 1888830 2086580
39: 1972827 2143567
40: 1877729 2181851

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Cc: Adam Belay <abelay@google.com>
Cc: Tariq Toukan <tariqt@mellanox.com>
Cc: Yuval Mintz <Yuval.Mintz@cavium.com>
Cc: Ariel Elior <ariel.elior@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-16 13:40:57 -05:00
..
datagram.c udp: do fwd memory scheduling on dequeue 2016-11-07 13:24:41 -05:00
dev_addr_lists.c
dev_ioctl.c
dev.c net: busy-poll: allow preemption in sk_busy_loop() 2016-11-16 13:40:57 -05:00
devlink.c genetlink: mark families as __ro_after_init 2016-10-27 16:16:09 -04:00
drop_monitor.c genetlink: mark families as __ro_after_init 2016-10-27 16:16:09 -04:00
dst_cache.c net: dst_cache_per_cpu_dst_set() can be static 2016-03-18 17:45:08 -04:00
dst.c net: add dst_cache to ovs vxlan lwtunnel 2016-02-16 20:21:48 -05:00
ethtool.c sctp: Add GSO support 2016-06-03 19:37:21 -04:00
fib_rules.c net: core: add missing check for uid_range in rule_exists. 2016-11-09 13:28:10 -05:00
filter.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2016-11-15 10:54:36 -05:00
flow_dissector.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2016-11-15 10:54:36 -05:00
flow.c flowcache: Avoid OOM condition under preasure 2016-03-17 10:28:42 +01:00
gen_estimator.c net: sched: do not acquire qdisc spinlock in qdisc/class stats dump 2016-06-07 16:37:14 -07:00
gen_stats.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2016-06-10 11:52:24 -07:00
hwbm.c net: hwbm: Fix unbalanced spinlock in error case 2016-05-25 12:35:09 -07:00
link_watch.c
lwtunnel.c ipv6: sr: add support for SRH encapsulation and injection with lwtunnels 2016-11-09 20:40:06 -05:00
Makefile net: add a hardware buffer management helper API 2016-03-14 12:19:46 -04:00
neighbour.c neigh: allow admin to set NUD_STALE 2016-08-08 15:36:38 -07:00
net_namespace.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2016-10-30 12:42:58 -04:00
net-procfs.c net: remove NETDEV_TX_LOCKED support 2016-04-26 15:53:05 -04:00
net-sysfs.c net: Add support for XPS with QoS via traffic classes 2016-10-31 15:00:48 -04:00
net-sysfs.h
net-traces.c net: IPv6 fib lookup tracepoint 2015-11-22 11:54:10 -05:00
netclassid_cgroup.c core: remove unneded headers for net cgroup controllers. 2016-02-17 15:31:27 -05:00
netevent.c
netpoll.c net: tracepoint napi:napi_poll add work and budget 2016-07-09 18:05:02 -04:00
netprio_cgroup.c core: remove unneded headers for net cgroup controllers. 2016-02-17 15:31:27 -05:00
pktgen.c net: pktgen: remove rcu locking in pktgen_change_name() 2016-10-17 10:52:59 -04:00
ptp_classifier.c ptp: Change ptp_class to a proper bitmask 2015-11-03 11:08:22 -05:00
request_sock.c tcp: restore fastopen operations 2015-10-05 03:19:06 -07:00
rtnetlink.c rtnl: reset calcit fptr in rtnl_unregister() 2016-11-09 20:18:19 -05:00
scm.c unix: correctly track in-flight fds in sending process user_struct 2016-02-08 10:30:42 -05:00
secure_seq.c
skbuff.c sock: do not set sk_err in sock_dequeue_err_skb 2016-11-07 20:29:10 -05:00
sock_diag.c sock_diag: align nlattr properly when needed 2016-04-26 12:00:48 -04:00
sock_reuseport.c soreuseport: do not export reuseport_add_sock() 2016-10-18 14:18:23 -04:00
sock.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2016-11-15 10:54:36 -05:00
stream.c net: fix sleeping for sk_wait_event() 2016-11-14 13:17:21 -05:00
sysctl_net_core.c bpf: add generic constant blinding for use in jits 2016-05-16 13:49:32 -04:00
timestamping.c net: skb_defer_rx_timestamp should check for phydev before setting up classify 2015-07-09 14:17:15 -07:00
tso.c net: tso: add support for IPv6 2015-10-26 22:24:22 -07:00
utils.c net: the space is required before the open parenthesis '(' 2016-06-29 05:15:14 -04:00