Re: [PATCH net-next v2 3/4] net: core: resegment oversized TCP GSO skbs

[Wang Zhan <wang.zhan@smartx.com>]

On Sat, 19 Sep 2026 11:37:45 -0400 Willem de Bruijn wrote:
> >  	return skb_shinfo(skb)->gso_segs <= READ_ONCE(dev->gso_max_segs) &&
> > -	       skb->len < netif_get_gso_max_size(dev, skb->protocol);
> > +	       skb->len < netif_get_gso_max_size(dev, vlan_get_protocol(skb));
>
> If this change is needed, it is not new for this feature and should be
> a separate commit.

Okay, will do it in v3 as patch 1/5, with a Fixes tag.

> > +	if (!skb_is_gso(skb) || !skb_is_gso_tcp(skb) ||
> > +	    skb->encapsulation || mss == GSO_BY_FRAGS ||
> > +	    !skb_mac_header_was_set(skb) ||
> > +	    !skb_transport_header_was_set(skb) ||
> > +	    !skb_can_gso_resegment(skb, features))
> > +		return 0;
>
> Is this duplicating/extending skb_can_gso_resegment

Okay, v3 folds skb_can_gso_resegment() into skb_gso_resegment_max_segs():
one caller, so the checks stay in one list.

> > +	if (skb_is_gso(skb) && skb_is_gso_tcp(skb) && !skb->encapsulation &&
> > +	    !gso_within_device_limits(skb, dev)) {
> > +		netdev_features_t offload = __netif_skb_features(skb, false);
> > +
> > +		resegment_max_segs =
> > +			skb_gso_resegment_max_segs(skb, dev, offload);
> > +		if (resegment_max_segs)
> > +			features = offload;
> > +	}
>
> This is a lot to put in the hot path for a rare use case. Consider how
> to make this less expensive.

Okay, v3 enters on

	if (unlikely(skb_is_gso(skb) && !(features & NETIF_F_GSO_MASK))) {

features is what netif_skb_features() returned a few lines above, and
gso_features_check() has already cleared the GSO bits for an over-limit
skb, so the common path reads two bits of a value which is already loaded.
An oversized skb then pays one __netif_skb_features(skb, false) call.
Could you take a look whether that is cheap enough?