Message-ID: <6528ceb761b7be9862dfc5e404031e832f086cb3.camel@sipsolutions.net>
Subject: Re: [PATCH] wifi: mac80211: check the control channel before
 downgrading the bandwidth
From:   Johannes Berg <johannes@sipsolutions.net>
To:     Michael-cy Lee =?UTF-8?Q?=28=E6=9D=8E=E5=B3=BB=E5=AE=87=29?= 
        <Michael-cy.Lee@mediatek.com>,
        "linux-wireless@vger.kernel.org" <linux-wireless@vger.kernel.org>
Cc:     Evelyn Tsai =?UTF-8?Q?=28=E8=94=A1=E7=8F=8A=E9=88=BA=29?= 
        <Evelyn.Tsai@mediatek.com>,
        "linux-mediatek@lists.infradead.org" 
        <linux-mediatek@lists.infradead.org>,
        Shayne Chen =?UTF-8?Q?=28=E9=99=B3=E8=BB=92=E4=B8=9E=29?= 
        <Shayne.Chen@mediatek.com>, "nbd@nbd.name" <nbd@nbd.name>,
        Money Wang =?UTF-8?Q?=28=E7=8E=8B=E4=BF=A1=E5=AE=89=29?= 
        <Money.Wang@mediatek.com>,
        "lorenzo@kernel.org" <lorenzo@kernel.org>,
        Ryder Lee <Ryder.Lee@mediatek.com>
Date:   Tue, 14 Mar 2023 09:40:52 +0100
In-Reply-To: <c9a31a6c06fe7a48df1d7871ac563b6f71cd4436.camel@mediatek.com>
References: <20221222091354.14050-1-michael-cy.lee@mediatek.com>
         <277f8db1f790931aaac0f94ba0ea69efcb6ba3ac.camel@sipsolutions.net>
         <70f19f86043750c8f164c316496a8690709fc062.camel@mediatek.com>
         <386c21d475215c95966bbac6b0853e2715081789.camel@sipsolutions.net>
         <c9a31a6c06fe7a48df1d7871ac563b6f71cd4436.camel@mediatek.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.46.4 (3.46.4-1.fc37) 
MIME-Version: 1.0
Precedence: bulk

Hi,

> It matters because Extender's STA interface's ifmgd->flags is changed
> during bandwidth downgrading and remains changed to the next
> association.

OK but then _that_ seems like the problem here, no?

> Since there is no way to reset ifmgd->flags after bandwidth
> downgrading, we check the control channel before downgrading the
> bandwidth. If the control channels are different, we directly fail the
> association.
>=20

Right, but what if you had an association to some other AP, or two
chanctxes were allowed, or something else?

I feel we should probably instead just reset these ifmgd->flags before
association, or so?

johannes