Re: [PATCH net-next 00/14] net: annotate data-races around sysctl knobs

From
Jason Xing <hidden>
Date
2026-11-09 03:12:07
Also in
bridge, linux-rdma, linux-s390, linux-sctp, lkml, lvs-devel, mptcp, netdev, netfilter-devel
Reply to
Jiayuan Chen, 2026-11-09
On Mon, Nov 9, 2026 at 10:55 AM Jiayuan Chen [off-list ref] wrote:
While testing the recent sysctl range fixes I noticed that a lot of our
sysctl knobs are read with plain loads while the handlers update them
with WRITE_ONCE() and no lock, so KCSAN reports them as data races. I
then used an LLM to go through every ctl_table under net/ and list the
readers that were left behind. This series annotates them with
READ_ONCE(), and adds WRITE_ONCE() to a few custom handlers and netlink
paths that store to the same fields with plain writes. Most of it is a
plain READ_ONCE() at each read; where a function reads the same knob
twice to make one decision, it now reads it once into a local, so the
comparison and the value it acts on cannot come from two different
writes.

There are no Fixes tags. Done one knob per patch this would be well over
a hundred commits, so I grouped them by subsystem and feature instead to
keep the review manageable. If maintainers prefer, I can split it up
further with a Fixes tag on each one, as several smaller series.
Over the last year, I've been thinking about the annotation stuff
especially around sysctls, which, I suppose, shouldn't actually affect
the real workload but does affect the outcome of the syzbot.

Out of sheer curiosity, my question would be what is the point of
trying to change all of them? Or what kind of benefits can we get from
the big series?

Thanks,
Jason
No functional change. KCSAN does hit these, e.g. toggling
net.ipv6.icmp.echo_ignore_all while receiving ICMPv6 echo requests:

  BUG: KCSAN: data-race in icmpv6_rcv / proc_dou8vec_minmax

  write (marked) to 0xffffffff8ee2940d of 1 bytes by task 504 on cpu 0:
   proc_dou8vec_minmax+0x1ab/0x1f0
   proc_sys_call_handler+0x282/0x3a0
   proc_sys_write+0x13/0x20
   ...

  read to 0xffffffff8ee2940d of 1 bytes by interrupt on cpu 3:
   icmpv6_rcv+0x99d/0xd40
   ip6_protocol_deliver_rcu+0x787/0x9a0
   ip6_input_finish+0xb0/0x1b0
   ip6_input+0xa8/0x250
   ipv6_rcv+0x14b/0x1c0
   ...

  value changed: 0x01 -> 0x00

Jiayuan Chen (14):
  ipv4: annotate data-races around ip_rt_* sysctls
  inet: frags: annotate data-races around fqdir->timeout and
    fqdir->max_dist
  ipv6: annotate data-races around addrconf sysctls
  ipv6: annotate data-races around icmpv6 reply sysctls
  calipso: annotate data-races around cache sysctls
  sctp: annotate data-races around sctp sysctls
  mptcp: annotate data-races around pernet sysctls
  rds: annotate data-races around core rds sysctls
  netfilter: conntrack: annotate data-races around conntrack sysctls
  netfilter: bridge: annotate data-races around br_netfilter and nf_log
    sysctls
  ipvs: annotate data-races around lblc_expiration and lblcr_expiration
  net/smc: annotate data-races around smc sysctls
  netfilter: conntrack: annotate data-races around timeout and tcp
    sysctls
  netfilter: cttimeout: annotate data-races around the default timeouts

 include/net/ipv6.h                            |   2 +-
 include/net/netfilter/nf_conntrack_acct.h     |   6 +-
 .../net/netfilter/nf_conntrack_timestamp.h    |   2 +-
 net/bridge/br_netfilter_hooks.c               |  20 ++--
 net/bridge/netfilter/ebt_log.c                |   2 +-
 net/ipv4/inet_fragment.c                      |   2 +-
 net/ipv4/ip_fragment.c                        |   6 +-
 net/ipv4/route.c                              |  34 +++---
 net/ipv6/addrconf.c                           |  22 ++--
 net/ipv6/calipso.c                            |   6 +-
 net/ipv6/icmp.c                               |  10 +-
 net/ipv6/ndisc.c                              |   2 +-
 net/ipv6/route.c                              |   8 +-
 net/mptcp/ctrl.c                              |  18 +--
 net/netfilter/ipvs/ip_vs_lblc.c               |   2 +-
 net/netfilter/ipvs/ip_vs_lblcr.c              |   2 +-
 net/netfilter/nf_conntrack_core.c             |   8 +-
 net/netfilter/nf_conntrack_ecache.c           |   2 +-
 net/netfilter/nf_conntrack_expect.c           |   2 +-
 net/netfilter/nf_conntrack_netlink.c          |   2 +-
 net/netfilter/nf_conntrack_proto.c            |   6 +-
 net/netfilter/nf_conntrack_proto_generic.c    |   8 +-
 net/netfilter/nf_conntrack_proto_gre.c        |  20 ++--
 net/netfilter/nf_conntrack_proto_icmp.c       |  12 +-
 net/netfilter/nf_conntrack_proto_icmpv6.c     |  12 +-
 net/netfilter/nf_conntrack_proto_sctp.c       |  13 ++-
 net/netfilter/nf_conntrack_proto_tcp.c        | 110 +++++++++---------
 net/netfilter/nf_conntrack_proto_udp.c        |  24 ++--
 net/netfilter/nf_flow_table_core.c            |   4 +-
 net/netfilter/nf_log_syslog.c                 |   2 +-
 net/rds/recv.c                                |   2 +-
 net/rds/send.c                                |   8 +-
 net/rds/threads.c                             |   4 +-
 net/sctp/associola.c                          |   2 +-
 net/sctp/bind_addr.c                          |   2 +-
 net/sctp/endpointola.c                        |  14 +--
 net/sctp/input.c                              |   2 +-
 net/sctp/sm_make_chunk.c                      |   6 +-
 net/sctp/sm_sideeffect.c                      |   2 +-
 net/sctp/sm_statefuns.c                       |   6 +-
 net/sctp/socket.c                             |  46 ++++----
 net/sctp/sysctl.c                             |  16 +--
 net/sctp/transport.c                          |   6 +-
 net/smc/af_smc.c                              |   8 +-
 net/smc/smc_clc.c                             |   8 +-
 net/smc/smc_core.c                            |   2 +-
 net/smc/smc_tx.c                              |   2 +-
 47 files changed, 259 insertions(+), 246 deletions(-)

--
2.43.0
Thread (51 messages) 51 messages, 7 authors, last reply 2d ago
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help