[PATCH] wifi: ath11k: Fix error handling in ath11k_wmi_p2p_noa_event()

[PATCH] wifi: ath11k: Fix error handling in ath11k_wmi_p2p_noa_event()

[Christophe JAILLET]

if (noa_descriptors > WMI_P2P_MAX_NOA_DESCRIPTORS), there is a mix of
return and goto. in such a case, 'ret' should be assigned an error code and
the 'td' should be freed to avoid a memory leak.

While at it, return 'ret' instead of 0 in case of error.
This is actually harmless, because the only caller does not handle the
return value.

Fixes: 2408379f15a1 ("wifi: ath11k: implement handling of P2P NoA event")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
---
 drivers/net/wireless/ath/ath11k/wmi.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/net/wireless/ath/ath11k/wmi.c b/drivers/net/wireless/ath/ath11k/wmi.c
index c74aa3f95658..589a1b5ec517 100644
--- a/drivers/net/wireless/ath/ath11k/wmi.c
+++ b/drivers/net/wireless/ath/ath11k/wmi.c
@@ -8657,7 +8657,7 @@ static int ath11k_wmi_p2p_noa_event(struct ath11k_base *ab,
 	const struct wmi_p2p_noa_event *ev;
 	const struct ath11k_wmi_p2p_noa_info *noa;
 	struct ath11k *ar;
-	int ret, vdev_id;
+	int ret = 0, vdev_id;
 	u8 noa_descriptors;
 
 	tb = ath11k_wmi_tlv_parse_alloc(ab, skb, GFP_ATOMIC);
@@ -8682,7 +8682,7 @@ static int ath11k_wmi_p2p_noa_event(struct ath11k_base *ab,
 	if (noa_descriptors > WMI_P2P_MAX_NOA_DESCRIPTORS) {
 		ath11k_warn(ab, "invalid descriptor num %d in P2P NoA event\n",
 			    noa_descriptors);
-		return -EINVAL;
+		ret = -EINVAL;
 		goto out;
 	}
 
@@ -8705,7 +8705,7 @@ static int ath11k_wmi_p2p_noa_event(struct ath11k_base *ab,
 	rcu_read_unlock();
 out:
 	kfree(tb);
-	return 0;
+	return ret;
 }
 
 static void ath11k_wmi_tlv_op_rx(struct ath11k_base *ab, struct sk_buff *skb)
-- 
2.44.0

Re: [PATCH] wifi: ath11k: Fix error handling in ath11k_wmi_p2p_noa_event()

[Jeff Johnson]

On 4/15/2024 4:23 AM, Christophe JAILLET wrote:
> if (noa_descriptors > WMI_P2P_MAX_NOA_DESCRIPTORS), there is a mix of
> return and goto. in such a case, 'ret' should be assigned an error code and
> the 'td' should be freed to avoid a memory leak.
> 
> While at it, return 'ret' instead of 0 in case of error.
> This is actually harmless, because the only caller does not handle the
> return value.

in that case it would be preferable to change this to be a void function and
not return anything. that would be consistent with most, if not all, of the
other event handlers