On Wed, 2021-05-19 at 12:47 +0200, Philipp Borgers wrote:
> On Tue, May 18, 2021 at 01:19:46PM +0200, Johannes Berg wrote:
> > On Tue, 2021-05-18 at 13:17 +0200, Felix Fietkau wrote:
> > >
> > > Frames with IEEE80211_TX_CTL_HW_80211_ENCAP set have no 802.11 header,
> > > so please change the check something like this:
> > >
> > > if ((info->flags & IEEE80211_TX_CTL_NO_ACK) &&
> > > ((info->flags & IEEE80211_TX_CTL_HW_80211_ENCAP) ||
> > > ieee80211_is_data(hdr->frame_control)))
> >
> > Maybe we should consider some kind of inline helper?
> >
> > static inline bool ieee80211_is_tx_data(struct sk_buff *skb)
> > {
> > ... *info = ...
> > ... *hdr = (void *)skb->data;
> >
> > return (info->flags & IEEE80211_TX_CTL_HW_80211_ENCAP) ||
> > ieee80211_is_data(hdr->frame_control);
> > }
>
> A frame with IEEE80211_TX_CTL_HW_80211_ENCAP set is always a data frame?
Yes, other frames can't be HW-encap'ed, nor would it make sense to
offload that.
> Should I put the definition of the function into include/net/mac80211.h?
>
Seems reasonable, yeah.
johannes
