From: Eric Dumazet <edumazet@google.com>
To: "David S . Miller" <davem@davemloft.net>,
Jakub Kicinski <kuba@kernel.org>,
Paolo Abeni <pabeni@redhat.com>
Cc: netdev@vger.kernel.org, eric.dumazet@gmail.com,
Eric Dumazet <edumazet@google.com>
Subject: [PATCH net-next 6/8] rtnetlink: do not depend on RTNL in rtnl_fill_proto_down()
Date: Fri, 3 May 2024 19:20:57 +0000 [thread overview]
Message-ID: <20240503192059.3884225-7-edumazet@google.com> (raw)
In-Reply-To: <20240503192059.3884225-1-edumazet@google.com>
Change dev_change_proto_down() and dev_change_proto_down_reason()
to write once on dev->proto_down and dev->proto_down_reason.
Then rtnl_fill_proto_down() can use READ_ONCE() annotations
and run locklessly.
rtnl_proto_down_size() should assume worst case,
because readng dev->proto_down_reason multiple
times would be racy without RTNL in the future.
Signed-off-by: Eric Dumazet <edumazet@google.com>
---
net/core/dev.c | 11 +++++++----
net/core/rtnetlink.c | 8 ++++----
2 files changed, 11 insertions(+), 8 deletions(-)
diff --git a/net/core/dev.c b/net/core/dev.c
index 35ce603ffc57fa209dc9a57e60981a2bca5e6a29..06304f07146c7f672573a977b069e7f2c031122e 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -9223,7 +9223,7 @@ int dev_change_proto_down(struct net_device *dev, bool proto_down)
netif_carrier_off(dev);
else
netif_carrier_on(dev);
- dev->proto_down = proto_down;
+ WRITE_ONCE(dev->proto_down, proto_down);
return 0;
}
@@ -9237,18 +9237,21 @@ int dev_change_proto_down(struct net_device *dev, bool proto_down)
void dev_change_proto_down_reason(struct net_device *dev, unsigned long mask,
u32 value)
{
+ u32 proto_down_reason;
int b;
if (!mask) {
- dev->proto_down_reason = value;
+ proto_down_reason = value;
} else {
+ proto_down_reason = dev->proto_down_reason;
for_each_set_bit(b, &mask, 32) {
if (value & (1 << b))
- dev->proto_down_reason |= BIT(b);
+ proto_down_reason |= BIT(b);
else
- dev->proto_down_reason &= ~BIT(b);
+ proto_down_reason &= ~BIT(b);
}
}
+ WRITE_ONCE(dev->proto_down_reason, proto_down_reason);
}
struct bpf_xdp_link {
diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index 242c24e857ec4c799f0239be3371fd589a8ed191..6af7f00503b43d4989d0aaafc8b968216a6e77f5 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -1036,8 +1036,8 @@ static size_t rtnl_proto_down_size(const struct net_device *dev)
{
size_t size = nla_total_size(1);
- if (dev->proto_down_reason)
- size += nla_total_size(0) + nla_total_size(4);
+ /* Assume dev->proto_down_reason is not zero. */
+ size += nla_total_size(0) + nla_total_size(4);
return size;
}
@@ -1737,10 +1737,10 @@ static int rtnl_fill_proto_down(struct sk_buff *skb,
struct nlattr *pr;
u32 preason;
- if (nla_put_u8(skb, IFLA_PROTO_DOWN, dev->proto_down))
+ if (nla_put_u8(skb, IFLA_PROTO_DOWN, READ_ONCE(dev->proto_down)))
goto nla_put_failure;
- preason = dev->proto_down_reason;
+ preason = READ_ONCE(dev->proto_down_reason);
if (!preason)
return 0;
--
2.45.0.rc1.225.g2a3ae87e7f-goog
next prev parent reply other threads:[~2024-05-03 19:21 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-03 19:20 [PATCH net-next 0/8] rtnetlink: more rcu conversions for rtnl_fill_ifinfo() Eric Dumazet
2024-05-03 19:20 ` [PATCH net-next 1/8] rtnetlink: do not depend on RTNL for IFLA_QDISC output Eric Dumazet
2024-05-05 14:48 ` Simon Horman
2024-05-03 19:20 ` [PATCH net-next 2/8] rtnetlink: do not depend on RTNL for IFLA_IFNAME output Eric Dumazet
2024-05-05 14:49 ` Simon Horman
2024-05-03 19:20 ` [PATCH net-next 3/8] rtnetlink: do not depend on RTNL for IFLA_TXQLEN output Eric Dumazet
2024-05-05 14:43 ` Simon Horman
2024-05-07 9:26 ` Paolo Abeni
2024-05-07 9:35 ` Eric Dumazet
2024-05-07 16:32 ` Simon Horman
2024-05-03 19:20 ` [PATCH net-next 4/8] net: write once on dev->allmulti and dev->promiscuity Eric Dumazet
2024-05-05 14:54 ` Simon Horman
2024-05-03 19:20 ` [PATCH net-next 5/8] rtnetlink: do not depend on RTNL for many attributes Eric Dumazet
2024-05-05 14:46 ` Simon Horman
2024-05-05 15:00 ` Eric Dumazet
2024-05-05 15:06 ` Simon Horman
2024-05-05 15:14 ` Eric Dumazet
2024-05-07 16:38 ` Simon Horman
2024-05-07 16:39 ` Eric Dumazet
2024-05-07 16:55 ` Simon Horman
2024-05-03 19:20 ` Eric Dumazet [this message]
2024-05-05 14:52 ` [PATCH net-next 6/8] rtnetlink: do not depend on RTNL in rtnl_fill_proto_down() Simon Horman
2024-05-03 19:20 ` [PATCH net-next 7/8] rtnetlink: do not depend on RTNL in rtnl_xdp_prog_skb() Eric Dumazet
2024-05-05 14:53 ` Simon Horman
2024-05-03 19:20 ` [PATCH net-next 8/8] rtnetlink: allow rtnl_fill_link_netnsid() to run under RCU protection Eric Dumazet
2024-05-05 14:54 ` Simon Horman
2024-05-07 9:40 ` [PATCH net-next 0/8] rtnetlink: more rcu conversions for rtnl_fill_ifinfo() patchwork-bot+netdevbpf
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240503192059.3884225-7-edumazet@google.com \
--to=edumazet@google.com \
--cc=davem@davemloft.net \
--cc=eric.dumazet@gmail.com \
--cc=kuba@kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).