Re: [PATCH iwl-net 3/3] ice: do not block on the aux device lock in ndo_setup_tc

[Petr Oros]

On 10/21/26 1:22 PM, Loktionov, Aleksandr wrote:
> > -----Original Message-----
> > From: Petr Oros <poros@xxxxxxxxxx>
> > Sent: Monday, October 19, 2026 8:53 AM
> > To: netdev@xxxxxxxxxxxxxxx
> > Cc: Oros, Petr <poros@xxxxxxxxxx>; Nguyen, Anthony L
> > <anthony.l.nguyen@xxxxxxxxx>; Kitszel, Przemyslaw
> > <przemyslaw.kitszel@xxxxxxxxx>; Andrew Lunn <andrew+netdev@xxxxxxx>;
> > David S . Miller <davem@xxxxxxxxxxxxx>; Eric Dumazet
> > <edumazet@xxxxxxxxxx>; Jakub Kicinski <kuba@xxxxxxxxxx>; Paolo Abeni
> > <pabeni@xxxxxxxxxx>; Ertman, David M <david.m.ertman@xxxxxxxxx>; Shiraz
> > Saleem <shiraz.saleem@xxxxxxxxx>; Leon Romanovsky <leon@xxxxxxxxxx>;
> > Maziarz, Kamil <kamil.maziarz@xxxxxxxxx>; Rafal Rogalski
> > <rafalx.rogalski@xxxxxxxxx>; Loktionov, Aleksandr
> > <aleksandr.loktionov@xxxxxxxxx>; Schmidt, Michal <mschmidt@xxxxxxxxxx>;
> > linux-rdma@xxxxxxxxxxxxxxx; intel-wired-lan@xxxxxxxxxxxxxxxx; linux-
> > kernel@xxxxxxxxxxxxxxx; stable@xxxxxxxxxxxxxxx; Sebastian Andrzej
> > Siewior <bigeasy@xxxxxxxxxxxxx>; Clark Williams <clrkwllms@xxxxxxxxxx>;
> > Steven Rostedt <rostedt@xxxxxxxxxxx>; Mateusz Palczewski
> > <mateusz.palczewski@xxxxxxxxx>; linux-rt-devel@xxxxxxxxxxxxxxx
> > Subject: [PATCH iwl-net 3/3] ice: do not block on the aux device lock in
> > ndo_setup_tc
> >
> > The TC_SETUP_QDISC_MQPRIO path of ice_setup_tc() takes the auxiliary
> > device lock to test whether the RDMA driver is bound, and ndo_setup_tc()
> > runs with the RTNL lock held. This closes the same cycle against the
> > RDMA auxiliary probe as the two preceding patches.
> >
> > Use device_trylock() and report the device as busy when it fails, with a
> > distinct message for the same reason as in the previous patch.
> >
> > Fixes: 4b31fd4d77ff ("ice: Fix RDMA VSI removal during queue rebuild")
> > Cc: stable@xxxxxxxxxxxxxxx
> > Signed-off-by: Petr Oros <poros@xxxxxxxxxx>
> > ---
> >   drivers/net/ethernet/intel/ice/ice_main.c | 6 +++++-
> >   1 file changed, 5 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/net/ethernet/intel/ice/ice_main.c
> > b/drivers/net/ethernet/intel/ice/ice_main.c
> > index edb62606d2aae3..7b251136149615 100644
> > --- a/drivers/net/ethernet/intel/ice/ice_main.c
> > +++ b/drivers/net/ethernet/intel/ice/ice_main.c
> > @@ -9448,7 +9448,11 @@ ice_setup_tc(struct net_device *netdev, enum
> > tc_setup_type type,
> >   		cdev = pf->cdev_info;
> >   		if (cdev && cdev->adev) {
> >   			mutex_lock(&pf->adev_mutex);
> > -			device_lock(&cdev->adev->dev);
> > +			if (!device_trylock(&cdev->adev->dev)) {
> > +				mutex_unlock(&pf->adev_mutex);
> > +				netdev_dbg(netdev, "RDMA auxiliary device is
> > busy, try again\n");
> > +				return -EBUSY;
> You return error here, but it looks like being silently dropped in mqprio_disable_offload() isn't it ?

Yes, mqprio_disable_offload() ignores it, the same as it ignores the
existing -EBUSY for a bound RDMA driver. Blocking there brings the
deadlock back and skipping the lock races with the probe, so v2 keeps
the refusal and warns about it instead of netdev_dbg().

> > +			}
> >   			locked = true;
> >   			if (cdev->adev->dev.driver) {
> >   				netdev_err(netdev, "Cannot change qdisc when
> > RDMA is active\n");
> > --
> > 2.56.0