From: Stanislav Fomichev <stfomichev@gmail.com>
To: Jakub Kicinski <kuba@kernel.org>
Cc: Kohei Enju <enjuk@amazon.com>,
netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
bpf@vger.kernel.org, "David S . Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@google.com>,
Paolo Abeni <pabeni@redhat.com>, Simon Horman <horms@kernel.org>,
Kuniyuki Iwashima <kuniyu@amazon.com>,
Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
Ahmed Zaki <ahmed.zaki@intel.com>,
Stanislav Fomichev <sdf@fomichev.me>,
Alexander Lobakin <aleksander.lobakin@intel.com>,
Kohei Enju <kohei.enju@gmail.com>
Subject: Re: [PATCH net-next v1] dev: remove netdev_lock() and netdev_lock_ops() in register_netdevice().
Date: Sat, 8 Mar 2025 14:41:41 -0800 [thread overview]
Message-ID: <Z8zHpf6JPfjkC_Sv@mini-arch> (raw)
In-Reply-To: <20250308131813.4f8c8f0d@kernel.org>
On 03/08, Jakub Kicinski wrote:
> On Sun, 9 Mar 2025 05:37:18 +0900 Kohei Enju wrote:
> > Both netdev_lock() and netdev_lock_ops() are called before
> > list_netdevice() in register_netdevice().
> > No other context can access the struct net_device, so we don't need these
> > locks in this context.
That's technically true, but it will set off a bunch of lockdep
warnings :-(
> Doesn't sysfs get registered earlier?
> I'm afraid not being able to take the lock from the registration
> path ties our hands too much. Maybe we need to make a more serious
> attempt at letting the caller take the lock?
This looks like another case of upper/lower :-( So maybe we need to solve
it for real? With an extra upper_lock pointer in the netdev?
Untested patch to convey the general idea:
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index d3c549f73909..9c85179431e6 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -2520,6 +2520,7 @@ struct net_device {
* Ordering: take after rtnl_lock.
*/
struct mutex lock;
+ struct mutex *upper_lock;
#if IS_ENABLED(CONFIG_NET_SHAPER)
/**
diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index 90597bf84e3d..3d0fda6e9bca 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -3022,6 +3022,9 @@ static int do_setlink(const struct sk_buff *skb, struct net_device *dev,
char ifname[IFNAMSIZ];
int err;
+ /* TODO: add another wrapper for this */
+ if (dev->upper_lock)
+ mutex_lock(dev->upper_lock);
netdev_lock_ops(dev);
err = validate_linkmsg(dev, tb, extack);
@@ -3394,6 +3397,8 @@ static int do_setlink(const struct sk_buff *skb, struct net_device *dev,
}
netdev_unlock_ops(dev);
+ if (dev->upper_lock)
+ mutex_unlock(dev->upper_lock);
return err;
}
diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c
index b0423046028c..818ff487b363 100644
--- a/net/mac80211/iface.c
+++ b/net/mac80211/iface.c
@@ -304,7 +304,7 @@ static int ieee80211_change_mac(struct net_device *dev, void *addr)
if (!dev->ieee80211_ptr->registered)
return 0;
- guard(wiphy)(local->hw.wiphy);
+ /* TODO: remove guard from other places */
return _ieee80211_change_mac(sdata, addr);
}
@@ -2227,6 +2227,8 @@ int ieee80211_if_add(struct ieee80211_local *local, const char *name,
free_netdev(ndev);
return ret;
}
+
+ ndev->upper_lock = &local->hw.wiphy.mtx;
}
mutex_lock(&local->iflist_mtx);
next prev parent reply other threads:[~2025-03-08 22:41 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-03-08 20:37 [PATCH net-next v1] dev: remove netdev_lock() and netdev_lock_ops() in register_netdevice() Kohei Enju
2025-03-08 21:18 ` Jakub Kicinski
2025-03-08 22:41 ` Stanislav Fomichev [this message]
2025-03-09 21:41 ` Stanislav Fomichev
2025-03-08 22:41 ` Jakub Kicinski
2025-03-08 22:59 ` Stanislav Fomichev
2025-03-08 23:51 ` Stanislav Fomichev
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=Z8zHpf6JPfjkC_Sv@mini-arch \
--to=stfomichev@gmail.com \
--cc=ahmed.zaki@intel.com \
--cc=aleksander.lobakin@intel.com \
--cc=bigeasy@linutronix.de \
--cc=bpf@vger.kernel.org \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=enjuk@amazon.com \
--cc=horms@kernel.org \
--cc=kohei.enju@gmail.com \
--cc=kuba@kernel.org \
--cc=kuniyu@amazon.com \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=sdf@fomichev.me \
/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).