On Tue, 2008-05-20 at 10:58 -0400, Stefanik Gábor wrote:
> On Tue, May 20, 2008 at 10:27 AM, Johannes Berg
> <johannes@xxxxxxxxxxxxxxxx> wrote:
> >
> >> - /* drop all data frame if we are not associated */
> >> - if ((!iwl3945_is_associated(priv) ||
> >> + /* drop all data frame if we are not associated and not monitoring */
> >> + if (((!iwl3945_is_associated(priv) && priv->iw_mode !=
> >> IEEE80211_IF_TYPE_MNTR) ||
> >> ((priv->iw_mode == IEEE80211_IF_TYPE_STA) && !priv->assoc_id)) &&
> >> ((fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA)) {
> >> - IWL_DEBUG_DROP("Dropping - !iwl3945_is_associated\n");
> >> + IWL_DEBUG_DROP("Dropping - !iwl3945_is_associated outside monitor mode\n");
> >> goto drop_unlock;
> >> }
> >
> > This makes injection not work when you're not exclusively in monitor
> > mode. Yes, it may break the spec to be sending frames to an arbitrary
> > station, but it's useful for all kinds of things including testing to
> > have a monitor and a regular interface working at the same time.
> >
> > johannes
> >
>
> What do you think is a better way to check if we have a monitor
> interface? && priv->iw_mode == IEEE80211_IF_TYPE_STA? Or should this
> be removed altogether?
I tend to think it should be removed, in regular operation mac80211
would never create such a data frame, and when injecting you actually
want it.
johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
