Skip to contentSkip to conversation

mac80211: prevent out of bounds access in ieee80211_rx_h_action()

PATCH In review Dan Carpenter 2 participants 4 messages Jan 19, 2022 – Jan 22, 2022

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mac80211: prevent out of bounds access in ieee80211_rx_h_action()

Smatch complains that status->band comes from the skb->data so it cannot be trusted. It's a u8 so it can be in the 0-255 range but the rx->local->hw.wiphy->bands[] array only has 6 elements.

Fixes 1d8d3dec5fbb ("mac80211: handle SMPS action frames")
---
 net/mac80211/rx.c | 3 +++
 1 file changed, 3 insertions(+)
net/mac80211/rx.c +3 -0
index 93680af62c47..9b0cf9e3b7bd 100644
@@ -3287,6 +3287,9 @@ ieee80211_rx_h_action(struct ieee80211_rx_data *rx)
 	if (len < IEEE80211_MIN_ACTION_SIZE)
 		return RX_DROP_UNUSABLE;
 
+	if (status->band >= NUM_NL80211_BANDS)
+		goto invalid;
+
 	if (!rx->sta && mgmt->u.action.category != WLAN_CATEGORY_PUBLIC &&
 	    mgmt->u.action.category != WLAN_CATEGORY_SELF_PROTECTED &&
 	    mgmt->u.action.category != WLAN_CATEGORY_SPECTRUM_MGMT)
Signature · 1 line
2.20.1
maintainer
#1 lore raw
Quoted 2 lines · Dan Carpenter

Smatch complains that status->band comes from the skb->data

Hmm. How does it come to that conclusion? It's not really true? It comes from skb->cb, and the driver should fill it. Also, we have:

void ieee80211_rx_list(struct ieee80211_hw *hw, struct ieee80211_sta *pubsta,
struct sk_buff *skb, struct list_head *list)
{
...

        if (WARN_ON(status->band >= NUM_NL80211_BANDS))
                goto drop;

so I really don't think this patch is needed?

johannes

#2 lore raw
Quoted 6 lines · Johannes Berg
Quoted 2 lines · Dan Carpenter

Smatch complains that status->band comes from the skb->data

Hmm. How does it come to that conclusion? It's not really true? It comes from skb->cb, and the driver should fill it.

Ugh... Sorry. I misread the code. I spent some time trying to figure this out as well, but I still didn't figure it out.

So, yeah. It's skb->cb and Smatch for some reason thinks skb->cb holds user data... I will look into this.

Quoted 11 lines

Also, we have:

void ieee80211_rx_list(struct ieee80211_hw *hw, struct ieee80211_sta *pubsta,
struct sk_buff *skb, struct list_head *list)
{
...

        if (WARN_ON(status->band >= NUM_NL80211_BANDS))
                goto drop;

so I really don't think this patch is needed?

The problem with that is that ->cb is an array of char so Smatch doesn't track status->band across function boundaries.

regards,
dan carpenter

#3 lore raw
Quoted 2 lines · Dan Carpenter

So, yeah. It's skb->cb and Smatch for some reason thinks skb->cb holds user data... I will look into this.

I fixed this bug. I'll test the fix for a while before pushing it.

regards,
dan carpenter