Re: BUG(?): dev->wl->if_type is 0 for monitor interfaces (should be 5); b43_is_mode(dev->wl; IEEE80211_IF_TYPE_MNTR) is always false

Re: BUG(?): dev->wl->if_type is 0 for monitor interfaces (should be 5); b43_is_mode(dev->wl; IEEE80211_IF_TYPE_MNTR) is always false

[Stefanik Gábor]

On Wed, Jun 25, 2008 at 10:22 PM, Stefanik G=E1bor
<netrolller.3d@gmail.com> wrote:
> In both b43 and b43legacy, dev->wl->if_type is 0
> (IEEE80211_IF_TYPE_INVALID) when injecting/transmitting packets on a
> monitor interface (not sure about receiving), while the logical resul=
t
> would be  5 (IEEE80211_IF_TYPE_MNTR). Due to this, the check
> b43(legacy)_is_mode(dev->wl; IEEE80211_IF_TYPE_MNTR) always returns 0=
=2E
>
> I discovered this when updating the b43/b43legacy injection patches.
> This is the latest patch I made, working around this bug by checking
> for both 0 and 5 (I added a quick printk("B43 DEBUG: dev->wl->if_type
> =3D %d", dev->wl->if_type) after the /* MAC control */ comment, which
> helped me find this bug):
>
> diff --git a/drivers/net/wireless/b43/xmit.c b/drivers/net/wireless/b=
43/xmit.c
> index f9e1cff..b7553ab 100644
> --- a/drivers/net/wireless/b43/xmit.c
> +++ b/drivers/net/wireless/b43/xmit.c
> @@ -315,9 +315,11 @@ int b43_generate_txhdr(struct b43_wldev *dev,
>        }
>
>        /* MAC control */
> -       if (!(info->flags & IEEE80211_TX_CTL_NO_ACK))
> +       if (dev->wl->if_type !=3D 5 && dev->wl->if_type !=3D 0 &&
> +             !(info->flags & IEEE80211_TX_CTL_NO_ACK))
>                mac_ctl |=3D B43_TXH_MAC_ACK;
> -       if (!(((fctl & IEEE80211_FCTL_FTYPE) =3D=3D IEEE80211_FTYPE_C=
TL) &&
> +       if (dev->wl->if_type !=3D 5 && dev->wl->if_type !=3D 0 &&
> +             !(((fctl & IEEE80211_FCTL_FTYPE) =3D=3D IEEE80211_FTYPE=
_CTL) &&
>              ((fctl & IEEE80211_FCTL_STYPE) =3D=3D IEEE80211_STYPE_PS=
POLL)))
>                mac_ctl |=3D B43_TXH_MAC_HWSEQ;
>        if (info->flags & IEEE80211_TX_CTL_FIRST_FRAGMENT)
>
> Is this intentional? I don't know, but it does looks strange for sure=
=2E
>
> --G=E1bor
>
> --
> Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-)
>

Oops... also CCing linux.wireless.

--=20
Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-)
--
To unsubscribe from this list: send the line "unsubscribe linux-wireles=
s" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: BUG(?): dev->wl->if_type is 0 for monitor interfaces (should be 5); b43_is_mode(dev->wl; IEEE80211_IF_TYPE_MNTR) is always false

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 662 bytes --]

On Wed, 2008-06-25 at 22:23 +0200, Stefanik Gábor wrote:
> On Wed, Jun 25, 2008 at 10:22 PM, Stefanik Gábor
> <netrolller.3d@gmail.com> wrote:
> > In both b43 and b43legacy, dev->wl->if_type is 0
> > (IEEE80211_IF_TYPE_INVALID) when injecting/transmitting packets on a
> > monitor interface (not sure about receiving), while the logical result
> > would be  5 (IEEE80211_IF_TYPE_MNTR).

I should answer on the list too then:

No, 5 wouldn't be the logical result. Try to understand how monitor mode
and virtual interfaces in mac80211 work.

b43 only sees what mode it is operating in, but pure monitor mode isn't
"operating" at all.

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]