2021-04-30 14:48:03

by wangyunjian

Subject: [PATCH net] rsi: Add a NULL check in rsi_core_xmit

From: Yunjian Wang <[email protected]>

The skb may be NULL in rsi_core_xmit(). Add a check to avoid
dereferencing null pointer.

Addresses-Coverity: ("Dereference after null check")
Fixes: dad0d04fa7ba ("rsi: Add RS9113 wireless driver")
Signed-off-by: Yunjian Wang <[email protected]>
---
drivers/net/wireless/rsi/rsi_91x_core.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/wireless/rsi/rsi_91x_core.c b/drivers/net/wireless/rsi/rsi_91x_core.c
index a48e616e0fb9..436e7b30d159 100644
--- a/drivers/net/wireless/rsi/rsi_91x_core.c
+++ b/drivers/net/wireless/rsi/rsi_91x_core.c
@@ -492,5 +492,6 @@ void rsi_core_xmit(struct rsi_common *common, struct sk_buff *skb)
xmit_fail:
rsi_dbg(ERR_ZONE, "%s: Failed to queue packet\n", __func__);
/* Dropping pkt here */
- ieee80211_free_txskb(common->priv->hw, skb);
+ if (skb)
+ ieee80211_free_txskb(common->priv->hw, skb);
}
--
2.19.1

2021-04-30 15:04:47

by Johannes Berg

Subject: Re: [PATCH net] rsi: Add a NULL check in rsi_core_xmit

On Fri, 2021-04-30 at 22:46 +0800, wangyunjian wrote:
> From: Yunjian Wang <[email protected]>
>
> The skb may be NULL in rsi_core_xmit().

How so?

Static checkers are good. Coverity is one of the better ones, in my
experience. But blindly believing static checkers still isn't good.

I see why the static checker is confused, but really, _you_ should have
done that work, not me.

johannes