net/mac80211/tdls.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
NL80211_TDLS_ENABLE_LINK can be called without completing the TDLS
setup handshake, or after the TDLS_PEER_SETUP_TIMEOUT (15 seconds)
has expired and tdls_peer_del_work has zeroed out the peer address.
In both cases the WARN_ON_ONCE triggers because tdls_peer is either
zero or mismatched with the requested peer.
Replace the WARN_ON_ONCE with a proper check that returns -ENOLINK
to enforce that ENABLE_LINK can only proceed after a successful TDLS
handshake.
Reported-by: syzbot+56b6a844a4ea74487b7b@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=56b6a844a4ea74487b7b
Tested-by: syzbot+56b6a844a4ea74487b7b@syzkaller.appspotmail.com
Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com>
---
net/mac80211/tdls.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/net/mac80211/tdls.c b/net/mac80211/tdls.c
index dbbfe2d6842f..d7d8b2fcc3ee 100644
--- a/net/mac80211/tdls.c
+++ b/net/mac80211/tdls.c
@@ -1457,8 +1457,9 @@ int ieee80211_tdls_oper(struct wiphy *wiphy, struct net_device *dev,
set_sta_flag(sta, WLAN_STA_TDLS_PEER_AUTH);
- WARN_ON_ONCE(is_zero_ether_addr(sdata->u.mgd.tdls_peer) ||
- !ether_addr_equal(sdata->u.mgd.tdls_peer, peer));
+ if (is_zero_ether_addr(sdata->u.mgd.tdls_peer) ||
+ !ether_addr_equal(sdata->u.mgd.tdls_peer, peer))
+ return -ENOLINK;
break;
case NL80211_TDLS_DISABLE_LINK:
/*
--
2.43.0
On Tue, 2026-03-10 at 21:30 +0530, Deepanshu Kartikey wrote:
>
> - WARN_ON_ONCE(is_zero_ether_addr(sdata->u.mgd.tdls_peer) ||
> - !ether_addr_equal(sdata->u.mgd.tdls_peer, peer));
> + if (is_zero_ether_addr(sdata->u.mgd.tdls_peer) ||
> + !ether_addr_equal(sdata->u.mgd.tdls_peer, peer))
> + return -ENOLINK;
>
I think that check needs to be earlier, otherwise side effects happen
(TDLS_PEER_AUTH flag).
Also, I'm a bit confused, how is it possible the sta_info_get() worked,
but there's no TDLS? Maybe really what it needs is
sta = sta_info_get(sdata, peer);
- if (!sta)
+ if (!sta || !sta->sta.tdls)
return -ENOLINK;
instead?
johannes
On Fri, Mar 13, 2026 at 12:42 PM Johannes Berg <johannes@sipsolutions.net> wrote: > On Tue, 2026-03-10 at 21:30 +0530, Deepanshu Kartikey wrote: > I think that check needs to be earlier, otherwise side effects happen > (TDLS_PEER_AUTH flag). > > Also, I'm a bit confused, how is it possible the sta_info_get() worked, > but there's no TDLS? Maybe really what it needs is > > sta = sta_info_get(sdata, peer); > - if (!sta) > + if (!sta || !sta->sta.tdls) > return -ENOLINK; > > instead? > > johannes Thanks for the clarification. I have sent the v2 patch. Deepanshu
© 2016 - 2026 Red Hat, Inc.