linux-kernel - Re: [PATCH v2] staging: rtl8723bs: remove possible deadlock when disconnect

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Sat, 11 Sep 2021 12:53:08 +0200
From:   Fabio Aiuto <fabioaiuto83@...il.com>
To:     Hans de Goede <hdegoede@...hat.com>
Cc:     gregkh@...uxfoundation.org, Larry.Finger@...inger.net,
        linux-staging@...ts.linux.dev, linux-kernel@...r.kernel.org,
        Martin Kaiser <martin@...ser.cx>,
        Phillip Potter <phil@...lpotter.co.uk>,
        Michael Straube <straube.linux@...il.com>,
        "Fabio M. De Francesco" <fmdefrancesco@...il.com>,
        Pavel Skripkin <paskripkin@...il.com>
Subject: Re: [PATCH v2] staging: rtl8723bs: remove possible deadlock when
 disconnect

Hello Hans,

On Fri, Sep 10, 2021 at 11:06:58PM +0200, Hans de Goede wrote:
> Hi,

<snip>

> 
> Thank you for your work on this. Overall this looks good.
> 
> I have one remark, since now you are relying on the
> sta->sleep_q.lock to protect the sleep_q data, you also
> need to update the sleep_q accesses in rtw_free_stainfo()
> specifically you need to add a spin_{lock,unlock}_bh(psta->sleep_q.lock)
> around these lines:
> 
> 
>         rtw_free_xmitframe_queue(pxmitpriv, &psta->sleep_q);
>         psta->sleepq_len = 0;
> 
> 
> Note there also is a:
> 
>         spin_lock_bh(&pxmitpriv->lock);
> 
> Just above this which needs to be pushed down to below the
> block which takes the psta->sleep_q.lock, so that the entire
> thing ends up looking like this:
> 
> 	spin_lock_bh(&psta->sleep_q.lock);
>         rtw_free_xmitframe_queue(pxmitpriv, &psta->sleep_q);
>         psta->sleepq_len = 0;
> 	spin_unlock_bh(&psta->sleep_q.lock);
> 
>         spin_lock_bh(&pxmitpriv->lock);
> 
> Other then that this patch looks good, thanks.
> 
> Regards,
> 
> Hans
> 

good catch, thanks a lot for review.

Greg, I see that this one is not in staging-next,
would you drop it so I can send a new one?
Or shall I send a Fixes: patch?

thank you all,

fabio