Subject: Re: [PATCH] net: mac80211: Add NULL checks for sta->sdata
To:     Simon Horman <simon.horman@corigine.com>
Cc:     johannes@sipsolutions.net, davem@davemloft.net,
        edumazet@google.com, kuba@kernel.org, pabeni@redhat.com,
        linux-wireless@vger.kernel.org, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
References: <ZBlsh0JFBagmPNCA@corigine.com>
From:   Jia-Ju Bai <baijiaju@buaa.edu.cn>
Message-ID: <a8b12c2a-67d8-d932-7505-542f4dcfefd7@buaa.edu.cn>
Date:   Tue, 21 Mar 2023 17:37:06 +0800
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:78.0) Gecko/20100101
 Thunderbird/78.0
MIME-Version: 1.0
In-Reply-To: <ZBlsh0JFBagmPNCA@corigine.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Content-Language: en-US
Precedence: bulk


On 2023/3/21 16:36, Simon Horman wrote:
>>>> However, in the same call stack, sta->sdata is also used in the
>>>> following functions:
>>>>
>>>> ieee80211_ba_session_work()
>>>>     ___ieee80211_stop_rx_ba_session(sta)
>>>>       ht_dbg(sta->sdata, ...); -> No check
>>>>       sdata_info(sta->sdata, ...); -> No check
>>>>       ieee80211_send_delba(sta->sdata, ...) -> No check
>>>>     ___ieee80211_start_rx_ba_session(sta)
>>>>       ht_dbg(sta->sdata, ...); -> No check
>>>>       ht_dbg_ratelimited(sta->sdata, ...); -> No check
>>>>     ieee80211_tx_ba_session_handle_start(sta)
>>>>       sdata = sta->sdata; if (!sdata) -> Add check by previous commit
>>>>     ___ieee80211_stop_tx_ba_session(sdata)
>>>>       ht_dbg(sta->sdata, ...); -> No check
>>>>     ieee80211_start_tx_ba_cb(sdata)
>>>>       sdata = sta->sdata; local = sdata->local -> No check
>>>>     ieee80211_stop_tx_ba_cb(sdata)
>>>>       ht_dbg(sta->sdata, ...); -> No check
>>> I wonder if it would be better to teach ht_* do do nothing
>>> if the first argument is NULL.
>> Okay, I will use this way in patch v2.
> Maybe it is not a good idea.
> But I think it is worth trying, at least locally, to see how it goes.

After checking the code, I find that ht_* is actually a macro from _sdata_*.
Many code points use ht_* and _sdata_*, and thus I am not sure it is 
fine to try this way.

For simplification, I still think checking sdata before the calls to 
ht_* or _sdata_* should be more proper :)

>
>>> Also, are these theoretical bugs?
>>> Or something that has been observed?
>>> And has a reproducer?
>> These bugs are found by my static analysis tool, by extending a known bug
>> fixed in a previous commit 69403bad97aa.
>> Thus, they could be theoretical bugs.
> Thanks, understood.
> I think it would be worth making that a bit clearer in the
> patch description (commit message).

Okay.

I have sent the v2 patch, please have a look.
Thanks a lot :)


Best wishes,
Jia-Ju Bai