Re: [PATCH net-next V2 2/2] net: sched: generalize check for no-op qdisc

[Simon Horman <horms@kernel.org>]

On Tue, Apr 08, 2025 at 05:31:24PM +0200, Jesper Dangaard Brouer wrote:
> Several drivers (e.g., veth, vrf) contain open-coded checks to determine
> whether a TX queue has a real qdisc attached - typically by testing if
> qdisc->enqueue is non-NULL.
> 
> These checks are functionally equivalent to comparing the queue's qdisc
> pointer against &noop_qdisc (qdisc named "noqueue"). This equivalence
> stems from noqueue_init(), which explicitly clears the enqueue pointer
> for the "noqueue" qdisc. As a result, __dev_queue_xmit() treats the qdisc
> as a no-op only when enqueue == NULL.
> 
> This patch introduces a common helper, qdisc_txq_is_noop() to standardize
> this check. The helper is added in sch_generic.h and replaces open-coded
> logic in both the veth and vrf drivers.
> 
> This is a non-functional change.
> 
> Signed-off-by: Jesper Dangaard Brouer <hawk@kernel.org>

...

> diff --git a/drivers/net/vrf.c b/drivers/net/vrf.c
> index 7168b33adadb..d4fe36c55f29 100644
> --- a/drivers/net/vrf.c
> +++ b/drivers/net/vrf.c
> @@ -349,9 +349,8 @@ static bool qdisc_tx_is_default(const struct net_device *dev)
>  		return false;
>  
>  	txq = netdev_get_tx_queue(dev, 0);
> -	qdisc = rcu_access_pointer(txq->qdisc);

nit: the qdisc variable is now unused in this function and can be removed.

Flagged by W=1 builds.

>  
> -	return !qdisc->enqueue;
> +	return qdisc_txq_is_noop(txq);
>  }
>  
>  /* Local traffic destined to local address. Reinsert the packet to rx

...