From: Nicolai Buchwitz <nb@tipi-net.de>
To: Jakub Kicinski <kuba@kernel.org>
Cc: davem@davemloft.net, netdev@vger.kernel.org, edumazet@google.com,
pabeni@redhat.com, andrew+netdev@lunn.ch, horms@kernel.org,
jakub@cloudflare.com, maxime.chevallier@bootlin.com,
lee@kernel.org, linux-leds@vger.kernel.org, pavel@kernel.org,
jv@jvosburgh.net, michael.chan@broadcom.com, jhs@mojatatu.com,
vinicius.gomes@intel.com, idosch@nvidia.com, razor@blackwall.org,
hare@suse.de, jhasan@marvell.com, danieller@nvidia.com
Subject: Re: [PATCH net-next v2 07/11] net: bridge: don't recurse on the port's netdev ops lock
Date: Wed, 03 Jun 2026 09:16:39 +0200 [thread overview]
Message-ID: <b9a0430bb83cc76096488ce4bf239368@tipi-net.de> (raw)
In-Reply-To: <20260603012840.2254293-8-kuba@kernel.org>
Hi Jakub
On 3.6.2026 03:28, Jakub Kicinski wrote:
> port_cost() calls __ethtool_get_link_ksettings() on the port device,
> which will soon take the port's ops lock. br_port_carrier_check()
> is reached via the NETDEV_CHANGE notifier from linkwatch, which
> already holds the port's ops lock, so the call would deadlock.
>
> Make port_cost() expect the port's ops lock held and switch to
> netif_get_link_ksettings(). The only other caller is new_nbp(),
> make sure it takes the lock explicitly.
>
> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
> ---
> net/bridge/br_if.c | 7 +++++--
> 1 file changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c
> index d39571e13744..049d1d25bc26 100644
> --- a/net/bridge/br_if.c
> +++ b/net/bridge/br_if.c
> @@ -19,6 +19,7 @@
> #include <linux/if_ether.h>
> #include <linux/slab.h>
> #include <net/dsa.h>
> +#include <net/netdev_lock.h>
> #include <net/sock.h>
> #include <linux/if_vlan.h>
> #include <net/switchdev.h>
> @@ -30,13 +31,13 @@
> * Determine initial path cost based on speed.
> * using recommendations from 802.1d standard
> *
> - * Since driver might sleep need to not be holding any locks.
> + * Since driver might sleep, we need to not be holding any bridge
> spinlocks.
> */
> static int port_cost(struct net_device *dev)
> {
> struct ethtool_link_ksettings ecmd;
>
> - if (!__ethtool_get_link_ksettings(dev, &ecmd)) {
> + if (!netif_get_link_ksettings(dev, &ecmd)) {
> switch (ecmd.base.speed) {
> case SPEED_10000:
> return 2;
> @@ -436,7 +437,9 @@ static struct net_bridge_port *new_nbp(struct
> net_bridge *br,
> p->br = br;
> netdev_hold(dev, &p->dev_tracker, GFP_KERNEL);
> p->dev = dev;
> + netdev_lock_ops(dev);
> p->path_cost = port_cost(dev);
> + netdev_unlock_ops(dev);
> p->priority = 0x8000 >> BR_PORT_BITS;
> p->port_no = index;
> p->flags = BR_LEARNING | BR_FLOOD | BR_MCAST_FLOOD | BR_BCAST_FLOOD;
Reviewed-by: Nicolai Buchwitz <nb@tipi-net.de>
Thanks
Nicolai
next prev parent reply other threads:[~2026-06-03 7:16 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-03 1:28 [PATCH net-next v2 00/11] net: ethtool: make sure __ethtool_get_link_ksettings() is ops-locked Jakub Kicinski
2026-06-03 1:28 ` [PATCH net-next v2 01/11] net: rename netdev_ops_assert_locked() Jakub Kicinski
2026-06-03 7:29 ` Nicolai Buchwitz
2026-06-03 1:28 ` [PATCH net-next v2 02/11] net: ethtool: cmis_cdb: hold instance lock for ops locked devices Jakub Kicinski
2026-06-03 1:28 ` [PATCH net-next v2 03/11] net: document NETDEV_CHANGENAME as ops locked Jakub Kicinski
2026-06-03 1:28 ` [PATCH net-next v2 04/11] net: ethtool: add netif_get_link_ksettings() for correct ops-locked use Jakub Kicinski
2026-06-03 6:40 ` Maxime Chevallier
2026-06-03 1:28 ` [PATCH net-next v2 05/11] net: bonding: don't recurse on the slave's netdev ops lock Jakub Kicinski
2026-06-03 7:20 ` Nicolai Buchwitz
2026-06-03 1:28 ` [PATCH net-next v2 06/11] net: team: don't recurse on the port's " Jakub Kicinski
2026-06-03 1:28 ` [PATCH net-next v2 07/11] net: bridge: " Jakub Kicinski
2026-06-03 7:16 ` Nicolai Buchwitz [this message]
2026-06-03 7:54 ` Nikolay Aleksandrov
2026-06-03 1:28 ` [PATCH net-next v2 08/11] net: sched: don't recurse on the netdev ops lock in qdiscs Jakub Kicinski
2026-06-03 1:28 ` [PATCH net-next v2 09/11] leds: trigger: netdev: don't recurse on the netdev ops lock Jakub Kicinski
2026-06-03 1:28 ` [PATCH net-next v2 10/11] scsi: fcoe: don't recurse on the netdev's " Jakub Kicinski
2026-06-03 1:28 ` [PATCH net-next v2 11/11] net: ethtool: make sure __ethtool_get_link_ksettings() is ops-locked Jakub Kicinski
2026-06-03 7:18 ` Nicolai Buchwitz
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=b9a0430bb83cc76096488ce4bf239368@tipi-net.de \
--to=nb@tipi-net.de \
--cc=andrew+netdev@lunn.ch \
--cc=danieller@nvidia.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=hare@suse.de \
--cc=horms@kernel.org \
--cc=idosch@nvidia.com \
--cc=jakub@cloudflare.com \
--cc=jhasan@marvell.com \
--cc=jhs@mojatatu.com \
--cc=jv@jvosburgh.net \
--cc=kuba@kernel.org \
--cc=lee@kernel.org \
--cc=linux-leds@vger.kernel.org \
--cc=maxime.chevallier@bootlin.com \
--cc=michael.chan@broadcom.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=pavel@kernel.org \
--cc=razor@blackwall.org \
--cc=vinicius.gomes@intel.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