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] [day] [month] [year] [list]
Message-ID: <e3oew536p4eghgtryz7luciuzg5wnwg27b6d3xn5btynmbjaes@dz46we4z4pzv>
Date: Thu, 7 Aug 2025 12:03:39 +0800
From: Xu Yang <xu.yang_2@....com>
To: Andrew Lunn <andrew@...n.ch>
Cc: andrew+netdev@...n.ch, davem@...emloft.net, edumazet@...gle.com, 
	kuba@...nel.org, pabeni@...hat.com, max.schulze@...ine.de, khalasa@...p.pl, 
	o.rempel@...gutronix.de, linux-usb@...r.kernel.org, netdev@...r.kernel.org, 
	linux-kernel@...r.kernel.org, imx@...ts.linux.dev, jun.li@....com
Subject: Re: [PATCH] net: usb: asix: avoid to call phylink_stop() a second
 time

Hi Andrew,

Thanks for your comments!

On Wed, Aug 06, 2025 at 05:58:18PM +0200, Andrew Lunn wrote:
> On Wed, Aug 06, 2025 at 04:30:17PM +0800, Xu Yang wrote:
> > The kernel will have below dump when system resume if the USB net device
> > was already disconnected during system suspend.
> 
> By disconnected, you mean pulled out?

Yes.

> 
> > It's because usb_resume_interface() will be skipped if the USB core found
> > the USB device was already disconnected. In this case, asix_resume() will
> > not be called anymore. So asix_suspend/resume() can't be balanced. When
> > ax88772_stop() is called, the phy device was already stopped. To avoid
> > calling phylink_stop() a second time, check whether usb net device is
> > already in suspend state.
> > 
> > Fixes: e0bffe3e6894 ("net: asix: ax88772: migrate to phylink")
> > Cc: stable@...r.kernel.org
> > Signed-off-by: Xu Yang <xu.yang_2@....com>
> > ---
> >  drivers/net/usb/asix_devices.c | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/net/usb/asix_devices.c b/drivers/net/usb/asix_devices.c
> > index 9b0318fb50b5..ac28f5fe7ac2 100644
> > --- a/drivers/net/usb/asix_devices.c
> > +++ b/drivers/net/usb/asix_devices.c
> > @@ -932,7 +932,8 @@ static int ax88772_stop(struct usbnet *dev)
> >  {
> >  	struct asix_common_private *priv = dev->driver_priv;
> >  
> > -	phylink_stop(priv->phylink);
> > +	if (!dev->suspend_count)
> > +		phylink_stop(priv->phylink);
> 
> Looking at ax88172a.c, lan78xx.c and smsc95xx.c, they don't have
> anything like this. Is asix special, or are all the others broken as
> well?

I have limited USB net devices. So I can't test others now.

But based on the error path, only below driver call phy_stop() or phylink_stop()
in their stop() callback:

drivers/net/usb/asix_devices.c
  ax88772_stop()
    phylink_stop()

drivers/net/usb/ax88172a.c
  ax88172a_stop()
    phy_stop()

drivers/net/usb/lan78xx.c
  lan78xx_stop()
    phylink_stop()

drivers/net/usb/smsc95xx.c
  smsc95xx_stop()
    phy_stop()

However, only asix_devices.c and lan78xx.c call phylink_suspend() in suspend()
callback. So I think lan78xx.c has this issue too.

Should I change usbnet common code like below?

diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c
index c39dfa17813a..44a8d325dfb1 100644
--- a/drivers/net/usb/usbnet.c
+++ b/drivers/net/usb/usbnet.c
@@ -839,7 +839,7 @@ int usbnet_stop (struct net_device *net)
        pm = usb_autopm_get_interface(dev->intf);
        /* allow minidriver to stop correctly (wireless devices to turn off
         * radio etc) */
-       if (info->stop) {
+       if (info->stop && !dev->suspend_count) {
                retval = info->stop(dev);
                if (retval < 0)
                        netif_info(dev, ifdown, dev->net,

Thanks,
Xu Yang

> 
> 	Andrew

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ