Re: [PATCH 1/2] staging: r8188eu: call _cancel_timer_ex from _rtw_free_recv_priv

From: Vihas Makwana <makvihas@gmail.com>
To: Dan Carpenter <dan.carpenter@oracle.com>
Cc: Larry Finger <Larry.Finger@lwfinger.net>,
	Phillip Potter <phil@philpotter.co.uk>,
	 Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Michael Straube <straube.linux@gmail.com>,
	 Martin Kaiser <martin@kaiser.cx>,
	Pavel Skripkin <paskripkin@gmail.com>,
	linux-staging@lists.linux.dev,  linux-kernel@vger.kernel.org
Subject: Re: [PATCH 1/2] staging: r8188eu: call _cancel_timer_ex from _rtw_free_recv_priv
Date: Wed, 9 Mar 2022 15:21:20 +0530	[thread overview]
Message-ID: <CAH1kMwT67mVn=UvwjtK28PvWy2CBddVx=CwvWRvqSV8_E81usw@mail.gmail.com> (raw)
In-Reply-To: <20220309075951.GQ3315@kadam>

> *If* then timer_setup() belongs in _rtw_init_recv_priv() then this is
> where the _cancel_timer_ex() belongs, yes.  But what about if the devs
> hid it in a different wrong place?
>
> Right the del_timer is in rtw_cancel_all_timer(), which is called
> from rtw_usb_if1_deinit() when we remove the USB device.  So something
> more complicated is wrong.  I would prefer to just note this as a bug
> until we can investigate more completely.

Sounds fair.

>
> I believe we can del_timer() twice without creating a bug, but I'm not
> positive.

Yes, we can. I tried it yesterday and it didn't create any bug.

On Wed, Mar 9, 2022 at 1:30 PM Dan Carpenter <dan.carpenter@oracle.com> wrote:
>
> On Wed, Mar 09, 2022 at 02:25:09AM +0530, Vihas Makwana wrote:
> > The _rtw_init_recv_priv() initializes precvpriv->signal_stat_timer and
> > sets it's timeout interval to 1000 ms. But _rtw_free_recv_priv()
> > doesn't cancel the timer and we need to explicitly call
> > _cancel_timer_ex() after we call _rtw_free_recv_priv() to cancel the
> > timer.
> > Call _cancel_timer_ex() from inside _rtw_free_recv_priv() as every init
> > function needs a matching free function.
> >
> >
> > Signed-off-by: Vihas Makwana <makvihas@gmail.com>
> > ---
> >  drivers/staging/r8188eu/core/rtw_recv.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/staging/r8188eu/core/rtw_recv.c b/drivers/staging/r8188eu/core/rtw_recv.c
> > index d77d98351..61308eb39 100644
> > --- a/drivers/staging/r8188eu/core/rtw_recv.c
> > +++ b/drivers/staging/r8188eu/core/rtw_recv.c
> > @@ -103,6 +103,7 @@ void _rtw_free_recv_priv(struct recv_priv *precvpriv)
> >       vfree(precvpriv->pallocated_frame_buf);
> >
> >       rtl8188eu_free_recv_priv(padapter);
> > +     _cancel_timer_ex(&precvpriv->signal_stat_timer);
> >  }
>
> *If* then timer_setup() belongs in _rtw_init_recv_priv() then this is
> where the _cancel_timer_ex() belongs, yes.  But what about if the devs
> hid it in a different wrong place?
>
> Right the del_timer is in rtw_cancel_all_timer(), which is called
> from rtw_usb_if1_deinit() when we remove the USB device.  So something
> more complicated is wrong.  I would prefer to just note this as a bug
> until we can investigate more completely.
>
> I believe we can del_timer() twice without creating a bug, but I'm not
> positive.
>
> regards,
> dan carpenter
>

-- 
Thanks,
Vihas