Received-SPF: pass (google.com: domain of linux-kernel+bounces-163737-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Message-ID: <e2f20484fb1f4607d099d2184c1d74c6a39febc1.camel@sipsolutions.net>
Subject: Re: [PATCH] wifi: nl80211: Avoid address calculations via out of
 bounds array indexing
From: Johannes Berg <johannes@sipsolutions.net>
To: Kees Cook <keescook@chromium.org>
Cc: Nathan Chancellor <nathan@kernel.org>, "David S. Miller"
 <davem@davemloft.net>, Eric Dumazet <edumazet@google.com>, Jakub Kicinski
 <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>, 
 linux-wireless@vger.kernel.org, netdev@vger.kernel.org, Jeff Johnson
 <quic_jjohnson@quicinc.com>, "Gustavo A. R. Silva" <gustavoars@kernel.org>,
  linux-kernel@vger.kernel.org, linux-hardening@vger.kernel.org
Date: Tue, 30 Apr 2024 12:01:44 +0200
In-Reply-To: <20240424220057.work.819-kees@kernel.org>
References: <20240424220057.work.819-kees@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.50.4 (3.50.4-1.fc39) 
Precedence: bulk
MIME-Version: 1.0

On Wed, 2024-04-24 at 15:01 -0700, Kees Cook wrote:
> Before request->channels[] can be used, request->n_channels must be set.
> Additionally, address calculations for memory after the "channels" array
> need to be calculated from the allocation base ("request") rather than
> via the first "out of bounds" index of "channels", otherwise run-time
> bounds checking will throw a warning.
>=20
> Reported-by: Nathan Chancellor <nathan@kernel.org>
> Fixes: e3eac9f32ec0 ("wifi: cfg80211: Annotate struct cfg80211_scan_reque=
st with __counted_by")

While I was weighing whether or not to apply this for 6.9 still ...

> +	request->n_channels =3D n_channels;
> =20
>  	if (n_ssids)
> -		request->ssids =3D (void *)&request->channels[n_channels];
> +		request->ssids =3D (void *)request + ssids_offset;

This really doesn't even seem right, shouldn't do pointer arithmetic on
void pointers. Same applies below too.

And also if you set n_channels before, perhaps it's actually OK to get a
pointer to *after*? Not sure though.

johannes