Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Authenticated-By: 
From:   Hayes Wang <hayeswang@realtek.com>
To:     Jakub Kicinski <jakub.kicinski@netronome.com>
CC:     "netdev@vger.kernel.org" <netdev@vger.kernel.org>,
        nic_swsd <nic_swsd@realtek.com>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "linux-usb@vger.kernel.org" <linux-usb@vger.kernel.org>
Subject: RE: [PATCH net-next 2/5] r8152: replace array with linking list for rx information
Thread-Topic: [PATCH net-next 2/5] r8152: replace array with linking list
 for rx information
Thread-Index: AQHVTEi8dqICcI3sDEmqKF0S4wiBPqbuAp4AgAEQ7MA=
Date:   Wed, 7 Aug 2019 04:34:24 +0000
Message-ID: <0835B3720019904CB8F7AA43166CEEB2F18D04FA@RTITMBSVM03.realtek.com.tw>
References: <1394712342-15778-289-albertk@realtek.com>
        <1394712342-15778-291-albertk@realtek.com>
 <20190806125342.4620a94f@cakuba.netronome.com>
In-Reply-To: <20190806125342.4620a94f@cakuba.netronome.com>
Accept-Language: zh-TW, en-US
Content-Language: zh-TW
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 8BIT
MIME-Version: 1.0
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Jakub Kicinski [mailto:jakub.kicinski@netronome.com]
[...]
> >  static int rtl_stop_rx(struct r8152 *tp)
> >  {
> > -	int i;
> > +	struct list_head *cursor, *next, tmp_list;
> > +	unsigned long flags;
> > +
> > +	INIT_LIST_HEAD(&tmp_list);
> >
> > -	for (i = 0; i < RTL8152_MAX_RX; i++)
> > -		usb_kill_urb(tp->rx_info[i].urb);
> > +	/* The usb_kill_urb() couldn't be used in atomic.
> > +	 * Therefore, move the list of rx_info to a tmp one.
> > +	 * Then, list_for_each_safe could be used without
> > +	 * spin lock.
> > +	 */
> 
> Would you mind explaining in a little more detail why taking the
> entries from the list for a brief period of time is safe?

Usually, it needs the spin lock before accessing the entry
of the list "tp->rx_info". However, for some reasons,
if we want to access the entry without spin lock, we
cloud move all entries to a local list temporally. Then,
we could make sure no other one could access the entries
included in the temporal local list.

For this case, when I move all entries to a temporal 
local list, no other one could access them. Therefore,
I could access the entries included in the temporal local
list without spin lock.


Best Regards,
Hayes