[PATCH] mac80211: fix memory leak in ieee80211_send_probe_req

[PATCH] mac80211: fix memory leak in ieee80211_send_probe_req

[tom.leiming]

From: Ming Lei <tom.leiming@gmail.com>

This patch fixes memory leak in ieee80211_send_probe_req, which
is introduced in 7c12ce8b854df346388ea56d684784e3484012cf:

	mac80211: use Probe Request template when sending a direct scan

The patch is against the latest wireless-test tree.

Signed-off-by: Ming Lei <tom.leiming@gmail.com>
---
 net/mac80211/util.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/net/mac80211/util.c b/net/mac80211/util.c
index e278f97..ca170b4 100644
--- a/net/mac80211/util.c
+++ b/net/mac80211/util.c
@@ -1028,6 +1028,7 @@ void ieee80211_send_probe_req(struct ieee80211_sub_if_data *sdata, u8 *dst,
 
 	IEEE80211_SKB_CB(skb)->flags |= IEEE80211_TX_INTFL_DONT_ENCRYPT;
 	ieee80211_tx_skb(sdata, skb);
+	kfree(buf);
 }
 
 u32 ieee80211_sta_get_rates(struct ieee80211_local *local,
-- 
1.6.2.5

Re: [PATCH] mac80211: fix memory leak in ieee80211_send_probe_req

[Kalle Valo]

"ext tom.leiming@gmail.com" <tom.leiming@gmail.com> writes:

> From: Ming Lei <tom.leiming@gmail.com>
>
> This patch fixes memory leak in ieee80211_send_probe_req, which
> is introduced in 7c12ce8b854df346388ea56d684784e3484012cf:
>
> 	mac80211: use Probe Request template when sending a direct scan


Sorry, I introduced this bug. Thank you finding this.

> The patch is against the latest wireless-test tree.

You should add comments like this after the "---" line.

> @@ -1028,6 +1028,7 @@ void ieee80211_send_probe_req(struct ieee80211_sub_if_data *sdata, u8 *dst,
>  
>  	IEEE80211_SKB_CB(skb)->flags |= IEEE80211_TX_INTFL_DONT_ENCRYPT;
>  	ieee80211_tx_skb(sdata, skb);
> +	kfree(buf);

Better to use dev_kfree_skb() here.

Can you resend, please?

-- 
Kalle Valo

Re: [PATCH] mac80211: fix memory leak in ieee80211_send_probe_req

[Ming Lei]

2010/1/15 Kalle Valo <kalle.valo@iki.fi>:
> "ext tom.leiming@gmail.com" <tom.leiming@gmail.com> writes:
>
>> From: Ming Lei <tom.leiming@gmail.com>
>>
>> This patch fixes memory leak in ieee80211_send_probe_req, which
>> is introduced in 7c12ce8b854df346388ea56d684784e3484012cf:
>>
>>       mac80211: use Probe Request template when sending a direct scan
>
>
> Sorry, I introduced this bug. Thank you finding this.
>
>> The patch is against the latest wireless-test tree.
>
> You should add comments like this after the "---" line.
>
>> @@ -1028,6 +1028,7 @@ void ieee80211_send_probe_req(struct ieee80211_sub_if_data *sdata, u8 *dst,
>>
>>       IEEE80211_SKB_CB(skb)->flags |= IEEE80211_TX_INTFL_DONT_ENCRYPT;
>>       ieee80211_tx_skb(sdata, skb);
>> +     kfree(buf);
>
> Better to use dev_kfree_skb() here.

Why is dev_kfree_skb better?
It is 'buf' allocated in ieee80211_send_probe_req to be freed, not skb.

>
> Can you resend, please?

OK.

-- 
Lei Ming

Re: [PATCH] mac80211: fix memory leak in ieee80211_send_probe_req

[Kalle Valo]

Ming Lei <tom.leiming@gmail.com> writes:

>>>       IEEE80211_SKB_CB(skb)->flags |= IEEE80211_TX_INTFL_DONT_ENCRYPT;
>>>       ieee80211_tx_skb(sdata, skb);
>>> +     kfree(buf);
>>
>> Better to use dev_kfree_skb() here.
>
> Why is dev_kfree_skb better?
> It is 'buf' allocated in ieee80211_send_probe_req to be freed, not skb.

Yes, I misread your patch first. Please forget that comment :)

-- 
Kalle Valo

Re: [PATCH] mac80211: fix memory leak in ieee80211_send_probe_req

[Kalle Valo]

Kalle Valo <kalle.valo@iki.fi> writes:

> "ext tom.leiming@gmail.com" <tom.leiming@gmail.com> writes:
>
>> From: Ming Lei <tom.leiming@gmail.com>
>>
>> This patch fixes memory leak in ieee80211_send_probe_req, which
>> is introduced in 7c12ce8b854df346388ea56d684784e3484012cf:
>>
>> 	mac80211: use Probe Request template when sending a direct scan
>

[...]

>> @@ -1028,6 +1028,7 @@ void ieee80211_send_probe_req(struct ieee80211_sub_if_data *sdata, u8 *dst,
>>  
>>  	IEEE80211_SKB_CB(skb)->flags |= IEEE80211_TX_INTFL_DONT_ENCRYPT;
>>  	ieee80211_tx_skb(sdata, skb);
>> +	kfree(buf);
>
> Better to use dev_kfree_skb() here.

I don't know what I was thinking when I sent the previous email  :)

After second look, this looks good.

Reviewed-by: Kalle Valo <kalle.valo@nokia.com>

-- 
Kalle Valo