Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <f717c60c24bf3dcec7cd3b0308c6bd804683276a.camel@sipsolutions.net>
Subject: Re: [PATCH] cfg80211: fix sband iftype data lookup for AP_VLAN
From:   Johannes Berg <johannes@sipsolutions.net>
To:     Felix Fietkau <nbd@nbd.name>, linux-wireless@vger.kernel.org
Date:   Thu, 22 Jun 2023 18:41:17 +0200
In-Reply-To: <2815ff55-5b5d-1412-5694-7692337bc473@nbd.name>
References: <20230622160501.40666-1-nbd@nbd.name>
         <adf8c1b8ea96c0c6ddc12579eacb8d9948440dcf.camel@sipsolutions.net>
         <2815ff55-5b5d-1412-5694-7692337bc473@nbd.name>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.48.3 (3.48.3-1.fc38) 
MIME-Version: 1.0
Precedence: bulk

On Thu, 2023-06-22 at 18:25 +0200, Felix Fietkau wrote:
> On 22.06.23 18:07, Johannes Berg wrote:
> > On Thu, 2023-06-22 at 18:05 +0200, Felix Fietkau wrote:
> > > Since AP_VLAN interfaces are not pushed to the driver,=20
> > >=20
> > That's a mac80211 thing though.
> >=20
> > > the driver should not
> > > be expected to register iftype data for them.
> > > Map them to the regular AP iftype on lookup.
> >=20
> > And this is in cfg80211 - not sure that seems right?
> >=20
> > OTOH I'd expect no callers with VLAN here, it doesn't really make sense
> > since they're not a standalone mode that actually has HE/EHT, but still=
,
> > seems odd this way?
> >=20
> > What's actually calling it? I'm guessing somewhere in mac80211?
>=20
> Yes, I guess only mac80211 is affected. I put in the cfg80211 prefix
> because that's what the header file belongs to.
>=20
> I made the patch in response to this:
> https://patchwork.kernel.org/project/linux-wireless/patch/20230605152141.=
17434-4-shayne.chen@mediatek.com/

OK, sure, that also doesn't really make sense.

> I found that there are several calls to ieee80211_get_he_iftype_cap and
> ieee80211_get_eht_iftype_cap, which could be affected by this issue.
> I thought dealing with this in a single place would be better than playin=
g
> whac-a-mole by fixing it at the call sites.
>=20

I replaced almost all of them with ieee80211_get_he_iftype_cap_vif() so
it shouldn't be that bad? Looks like I forgot some though.

johannes