A while back, I sent one RFC patch using lockless stats gathering
on 64bit arches.
This patch series does it more cleanly, using a seqcount.
Since qdisc/class stats are written at dequeue() time,
we can ask the dequeue to change the seqcount, so that
stats readers can avoid taking the root qdisc lock,
and instead the typical read_seqcount_{begin|retry} guarded
loop.
This does not change fast path costs, as the seqcount
increments are not more expensive than the bit manipulation,
and allows readers to not freeze the fast path anymore.
Eric Dumazet (2):
net_sched: transform qdisc running bit into a seqcount
net: sched: do not acquire qdisc spinlock in qdisc/class stats dump
Documentation/networking/gen_stats.txt | 2 +-
drivers/net/bonding/bond_main.c | 2 ++
drivers/net/ppp/ppp_generic.c | 3 +++
drivers/net/team/team.c | 2 ++
include/linux/netdevice.h | 1 +
include/net/gen_stats.h | 12 ++++++++----
include/net/sch_generic.h | 23 ++++++++++++-----------
net/bluetooth/6lowpan.c | 2 ++
net/core/dev.c | 2 +-
net/core/gen_estimator.c | 24 ++++++++++++++++--------
net/core/gen_stats.c | 34 +++++++++++++++++++++++-----------
net/ieee802154/6lowpan/core.c | 3 +++
net/l2tp/l2tp_eth.c | 4 ++++
net/netfilter/xt_RATEEST.c | 2 +-
net/sched/act_api.c | 4 ++--
net/sched/act_police.c | 3 ++-
net/sched/sch_api.c | 21 +++++++++++----------
net/sched/sch_atm.c | 3 ++-
net/sched/sch_cbq.c | 9 ++++++---
net/sched/sch_drr.c | 9 ++++++---
net/sched/sch_fq_codel.c | 15 +++++++++++----
net/sched/sch_generic.c | 14 ++++++++++----
net/sched/sch_hfsc.c | 10 +++++-----
net/sched/sch_htb.c | 11 ++++++-----
net/sched/sch_mq.c | 2 +-
net/sched/sch_mqprio.c | 11 +++++++----
net/sched/sch_multiq.c | 3 ++-
net/sched/sch_prio.c | 3 ++-
net/sched/sch_qfq.c | 9 ++++++---
29 files changed, 158 insertions(+), 85 deletions(-)
--
2.8.0.rc3.226.g39d4020
