Re: [PATCH 1/3] wifi: cfg80211: validate rx/tx MLME callback frame lengths before access

[Zhao Li <enderaoelyther@gmail.com>]

On Mon, 2026-07-06 at 18:37 +0800, Johannes Berg wrote:
> > Side effects of this change:
> >  - The WARN_ON(len < 2) is replaced by a silent early return, since
> >    these cfg80211 callbacks can legitimately receive short frames from
> >    drivers.
>
> Can they? How?

You're right, that wording is wrong; no in-tree caller passes len < 2.
mac80211 gates every path well above that (ieee80211_rx_mgmt_auth() at
len < 24 + 6, ieee80211_rx_mgmt_deauth()/disassoc() at len < 24 + 2), and
the locally built deauth/disassoc frames are always full size.

The frames that do reach these callbacks undersized are short relative
to their subtype, not shorter than 2 bytes. mwifiex accepts a 4-address
ieee80211_hdr plus the 2-byte firmware length prefix. After it strips
the prefix and removes addr4, pkt_len can be exactly 24: a bare
3-address management header with no reason-code body.
WARN_ON(len < 2) does not fire on that, and cfg80211_process_deauth()
then reads u.deauth.reason_code as a two-byte access starting at offset
24, immediately past the 24-byte buffer.

So the len >= 2 check only guards the frame_control read; the per-subtype
offsetofend() checks are the actual fix. I'll drop the WARN rather than
keep it, because these are exported callbacks and a malformed frame from
a driver should be dropped silently instead of backtraced.

v2 rewrites the commit message to describe the mwifiex path instead of
the inaccurate "legitimately receive short frames" claim; no code change.

Thanks,
Zhao Li