smsc9511: Register access happens after unregistration

smsc9511: Register access happens after unregistration

[Fabio Estevam]

Hi,

On a imx8mm-iotgateway board running kernel 5.17-rc6 I am getting the following
smsc95xx errors after a 'reboot' command:

ci_hdrc ci_hdrc.1: remove, state 1
usb usb2: USB disconnect, device number 1
usb 2-1: USB disconnect, device number 2
usb 2-1.1: USB disconnect, device number 3
smsc95xx 2-1.1:1.0 eth1: unregister 'smsc95xx' usb-ci_hdrc.1-1.1,
smsc95xx USB 2.0 Ethernet
smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
smsc95xx 2-1.1:1.0 eth1: hardware isn't capable of remote wakeup
usb 2-1.4: USB disconnect, device number 4
ci_hdrc ci_hdrc.1: USB bus 2 deregistered
ci_hdrc ci_hdrc.0: remove, state 4
usb usb1: USB disconnect, device number 1
ci_hdrc ci_hdrc.0: USB bus 1 deregistered
imx2-wdt 30280000.watchdog: Device shutdown: Expect reboot!
reboot: Restarting system

smsc95xx register accesses are happening after the device is unregistered.

What would be the correct way to prevent such smsc95xx register accesses
to happen?

Hi Frieder,

I see that imx8mm-kontron-n801x-s.dts also uses an SMSC9514 device.

Do you also see such errors after running "reboot"?

Thanks,

Fabio Estevam

Re: smsc9511: Register access happens after unregistration

[Fabio Estevam]

Hi Martyn,

On Thu, Mar 3, 2022 at 10:14 AM Fabio Estevam <festevam@gmail.com> wrote:
>
> Hi,
>
> On a imx8mm-iotgateway board running kernel 5.17-rc6 I am getting the following
> smsc95xx errors after a 'reboot' command:
>
> ci_hdrc ci_hdrc.1: remove, state 1
> usb usb2: USB disconnect, device number 1
> usb 2-1: USB disconnect, device number 2
> usb 2-1.1: USB disconnect, device number 3
> smsc95xx 2-1.1:1.0 eth1: unregister 'smsc95xx' usb-ci_hdrc.1-1.1,
> smsc95xx USB 2.0 Ethernet
> smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
> smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
> smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
> smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
> smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
> smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
> smsc95xx 2-1.1:1.0 eth1: hardware isn't capable of remote wakeup

The same error has been previously reported by Ferry Toth on Intel
Edison-Arduino:
https://lore.kernel.org/netdev/5cd6fc87-0f8d-0b9b-42be-8180540a94e7@gmail.com/

It happens after commit a049a30fc27c ("net: usb: Correct PHY handling
of smsc95xx")

A simple way to reproduce it:

echo -n "2-1" > /sys/bus/usb/drivers/usb/unbind

It is not clear to me why register 0x114 (MII_ADDR) is accessed after
smsc95xx unregistration.

Any ideas?

Thanks

Re: smsc9511: Register access happens after unregistration

[Andrew Lunn]

On Fri, Mar 04, 2022 at 10:58:47AM -0300, Fabio Estevam wrote:
> Hi Martyn,
> 
> On Thu, Mar 3, 2022 at 10:14 AM Fabio Estevam <festevam@gmail.com> wrote:
> >
> > Hi,
> >
> > On a imx8mm-iotgateway board running kernel 5.17-rc6 I am getting the following
> > smsc95xx errors after a 'reboot' command:
> >
> > ci_hdrc ci_hdrc.1: remove, state 1
> > usb usb2: USB disconnect, device number 1
> > usb 2-1: USB disconnect, device number 2
> > usb 2-1.1: USB disconnect, device number 3
> > smsc95xx 2-1.1:1.0 eth1: unregister 'smsc95xx' usb-ci_hdrc.1-1.1,
> > smsc95xx USB 2.0 Ethernet
> > smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
> > smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
> > smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
> > smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
> > smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
> > smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
> > smsc95xx 2-1.1:1.0 eth1: hardware isn't capable of remote wakeup
> 
> The same error has been previously reported by Ferry Toth on Intel
> Edison-Arduino:
> https://lore.kernel.org/netdev/5cd6fc87-0f8d-0b9b-42be-8180540a94e7@gmail.com/
> 
> It happens after commit a049a30fc27c ("net: usb: Correct PHY handling
> of smsc95xx")
> 
> A simple way to reproduce it:
> 
> echo -n "2-1" > /sys/bus/usb/drivers/usb/unbind
> 
> It is not clear to me why register 0x114 (MII_ADDR) is accessed after
> smsc95xx unregistration.

I suggest you look at the call sequences. When does phy_stop() and
phy_disconnect() get called? phylib has a work queue which runs once
per second to poll the PHY and get its status. That polling is stopped
by calling phy_stop(). It is also assumed you can still talk to the
PHY while performing phy_stop().

    Andrew

Re: smsc9511: Register access happens after unregistration

[Martyn Welch]

On Fri, 2022-03-04 at 15:11 +0100, Andrew Lunn wrote:
> On Fri, Mar 04, 2022 at 10:58:47AM -0300, Fabio Estevam wrote:
> > Hi Martyn,
> > 
> > On Thu, Mar 3, 2022 at 10:14 AM Fabio Estevam <festevam@gmail.com>
> > wrote:
> > > 
> > > Hi,
> > > 
> > > On a imx8mm-iotgateway board running kernel 5.17-rc6 I am getting
> > > the following
> > > smsc95xx errors after a 'reboot' command:
> > > 
> > > ci_hdrc ci_hdrc.1: remove, state 1
> > > usb usb2: USB disconnect, device number 1
> > > usb 2-1: USB disconnect, device number 2
> > > usb 2-1.1: USB disconnect, device number 3
> > > smsc95xx 2-1.1:1.0 eth1: unregister 'smsc95xx' usb-ci_hdrc.1-1.1,
> > > smsc95xx USB 2.0 Ethernet
> > > smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
> > > smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
> > > smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
> > > smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
> > > smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
> > > smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
> > > smsc95xx 2-1.1:1.0 eth1: hardware isn't capable of remote wakeup
> > 
> > The same error has been previously reported by Ferry Toth on Intel
> > Edison-Arduino:
> > https://lore.kernel.org/netdev/5cd6fc87-0f8d-0b9b-42be-8180540a94e7@gmail.com/
> > 
> > It happens after commit a049a30fc27c ("net: usb: Correct PHY
> > handling
> > of smsc95xx")
> > 
> > A simple way to reproduce it:
> > 
> > echo -n "2-1" > /sys/bus/usb/drivers/usb/unbind
> > 
> > It is not clear to me why register 0x114 (MII_ADDR) is accessed
> > after
> > smsc95xx unregistration.
> 
> I suggest you look at the call sequences. When does phy_stop() and
> phy_disconnect() get called? phylib has a work queue which runs once
> per second to poll the PHY and get its status. That polling is
> stopped
> by calling phy_stop(). It is also assumed you can still talk to the
> PHY while performing phy_stop().
> 
> 

I was just trying to remember the details from when I looked at this
driver. IIRC, that's broadly the conclusion I came to and couldn't
really see a way around it. Access to the phy was being lost before
phy_stop() was called.

Martyn

>     Andrew

Re: smsc9511: Register access happens after unregistration

[Andrew Lunn]

> > I suggest you look at the call sequences. When does phy_stop() and
> > phy_disconnect() get called? phylib has a work queue which runs once
> > per second to poll the PHY and get its status. That polling is
> > stopped
> > by calling phy_stop(). It is also assumed you can still talk to the
> > PHY while performing phy_stop().
> > 
> > 
> 
> I was just trying to remember the details from when I looked at this
> driver. IIRC, that's broadly the conclusion I came to and couldn't
> really see a way around it. Access to the phy was being lost before
> phy_stop() was called.

This i don't really understand. This is a reboot(1), so a controlled
shutdown. Why is access being lost to the PHY? A hot unplug you cannot
do anything about, but this is not the case here.

      Andrew

Re: smsc9511: Register access happens after unregistration

[Fabio Estevam]

Hi Andrew,

On Fri, Mar 4, 2022 at 11:11 AM Andrew Lunn <andrew@lunn.ch> wrote:

> I suggest you look at the call sequences. When does phy_stop() and
> phy_disconnect() get called? phylib has a work queue which runs once
> per second to poll the PHY and get its status. That polling is stopped
> by calling phy_stop(). It is also assumed you can still talk to the
> PHY while performing phy_stop().

With this debug patch:

--- a/drivers/net/phy/phy.c
+++ b/drivers/net/phy/phy.c
@@ -1034,6 +1034,8 @@ void phy_stop(struct phy_device *phydev)
 {
        struct net_device *dev = phydev->attached_dev;

+       pr_err("*************** %s\n", __func__);
+
        if (!phy_is_started(phydev) && phydev->state != PHY_DOWN) {
                WARN(1, "called from state %s\n",
                     phy_state_to_str(phydev->state));
diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
index ce0bb5951b81..e26d9edf32f3 100644
--- a/drivers/net/phy/phy_device.c
+++ b/drivers/net/phy/phy_device.c
@@ -1108,8 +1108,11 @@ EXPORT_SYMBOL(phy_connect);
  */
 void phy_disconnect(struct phy_device *phydev)
 {
-       if (phy_is_started(phydev))
+       pr_err("*********** %s: 1\n", __func__);
+       if (phy_is_started(phydev)) {
+               pr_err("*********** %s: 2\n", __func__);
                phy_stop(phydev);
+       }

        if (phy_interrupt_is_valid(phydev))
                phy_free_interrupt(phydev);
diff --git a/drivers/net/usb/smsc95xx.c b/drivers/net/usb/smsc95xx.c
index bc1e3dd67c04..6ed674636961 100644
--- a/drivers/net/usb/smsc95xx.c
+++ b/drivers/net/usb/smsc95xx.c
@@ -1190,6 +1190,7 @@ static void smsc95xx_unbind(struct usbnet *dev,
struct usb_interface *intf)
 {
        struct smsc95xx_priv *pdata = dev->driver_priv;

+       pr_err("*********** %s: calling phy_disconnect\n", __func__);
        phy_disconnect(dev->net->phydev);
        mdiobus_unregister(pdata->mdiobus);
        mdiobus_free(pdata->mdiobus);
@@ -1206,6 +1207,7 @@ static int smsc95xx_start_phy(struct usbnet *dev)

 static int smsc95xx_stop(struct usbnet *dev)
 {
+       pr_err("*********** %s\n", __func__);
        if (dev->net->phydev)
                phy_stop(dev->net->phydev);

The result is:

 # echo -n "2-1" > /sys/bus/usb/drivers/usb/unbind
usb 2-1.1: USB disconnect, device number 3
smsc95xx 2-1.1:1.0 eth1: unregister 'smsc95xx' usb-ci_hdrc.1-1.1,
smsc95xx USB 2.0 Ethernet
smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
*********** smsc95xx_unbind: calling phy_disconnect
smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
libphy: *********** phy_disconnect: 1
libphy: *********** phy_disconnect: 2
*************** phy_stop
smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
*********** smsc95xx_stop
smsc95xx 2-1.1:1.0 eth1: hardware isn't capable of remote wakeup
usb 2-1.4: USB disconnect, device number 4

Maybe the -19 errors should be ignored?

Re: smsc9511: Register access happens after unregistration

[Andrew Lunn]

>  # echo -n "2-1" > /sys/bus/usb/drivers/usb/unbind
> usb 2-1.1: USB disconnect, device number 3
> smsc95xx 2-1.1:1.0 eth1: unregister 'smsc95xx' usb-ci_hdrc.1-1.1,
> smsc95xx USB 2.0 Ethernet
> smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
> smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
> *********** smsc95xx_unbind: calling phy_disconnect
> smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
> libphy: *********** phy_disconnect: 1
> libphy: *********** phy_disconnect: 2
> *************** phy_stop
> smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
> smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
> smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
> smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
> smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
> smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
> *********** smsc95xx_stop
> smsc95xx 2-1.1:1.0 eth1: hardware isn't capable of remote wakeup
> usb 2-1.4: USB disconnect, device number 4
> 
> Maybe the -19 errors should be ignored?

But why does it return ENODEV? It seems to me, ignoring it is papering
over the cracks. Why cannot we access to the PHY?

     Andrew

Re: smsc9511: Register access happens after unregistration

[Fabio Estevam]

On Fri, Mar 4, 2022 at 11:51 AM Andrew Lunn <andrew@lunn.ch> wrote:

> But why does it return ENODEV? It seems to me, ignoring it is papering
> over the cracks. Why cannot we access to the PHY?

The -ENODEV is returned by usb_control_msg():

__smsc95xx_read_reg: -19
         usbnet_read_cmd: -19
              usb_control_msg: -19

 # echo -n "2-1" > /sys/bus/usb/drivers/usb/unbind
usb 2-1.1: USB disconnect, device number 3
smsc95xx 2-1.1:1.0 eth1: unregister 'smsc95xx' usb-ci_hdrc.1-1.1,
smsc95xx USB 2.0 Ethernet
libphy: *********** phy_disconnect: 1
libphy: *********** phy_disconnect: 2
********** returning -ENODEV from usb_control_msg
*********** returning -ENODEV from __usbnet_read_cmd
smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy

Re: smsc9511: Register access happens after unregistration

[Fabio Estevam]

On Fri, Mar 4, 2022 at 12:57 PM Fabio Estevam <festevam@gmail.com> wrote:
>
> On Fri, Mar 4, 2022 at 11:51 AM Andrew Lunn <andrew@lunn.ch> wrote:
>
> > But why does it return ENODEV? It seems to me, ignoring it is papering
> > over the cracks. Why cannot we access to the PHY?
>
> The -ENODEV is returned by usb_control_msg():
>
> __smsc95xx_read_reg: -19
>          usbnet_read_cmd: -19
>               usb_control_msg: -19

I added a WARN_ON() inside usb_control_msg() that triggers when
usb_control_msg() returns -ENODEV.

Here are the results on two unbind commands:

root@iotgate:~# echo -n "2-1" > /sys/bus/usb/drivers/usb/unbind
[   19.989294] usb 2-1.1: USB disconnect, device number 3
[   19.994484] ********** returning -ENODEV from usb_control_msg
[   20.000555] ------------[ cut here ]------------
[   20.005172] WARNING: CPU: 0 PID: 7 at
drivers/usb/core/message.c:163 usb_control_msg+0x178/0x180
[   20.013970] Modules linked in:
[   20.017028] CPU: 0 PID: 7 Comm: kworker/u8:0 Not tainted
5.17.0-rc6-dirty #581
[   20.024254] Hardware name: CompuLab i.MX8MM IoT Gateway (DT)
[   20.029913] Workqueue: events_power_efficient phy_state_machine
[   20.035842] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[   20.042807] pc : usb_control_msg+0x178/0x180
[   20.047080] lr : usb_control_msg+0x178/0x180
[   20.051357] sp : ffff80000965bb10
[   20.054670] x29: ffff80000965bb10 x28: 0000000000000040 x27: 00000000000000a0
[   20.061811] x26: 0000000000000000 x25: 0000000000000114 x24: 0000000000000004
[   20.068955] x23: 0000000080000300 x22: ffff000006812000 x21: ffff000000095400
[   20.076098] x20: 00000000ffffffed x19: ffff00000f26ea80 x18: 0000000000000010
[   20.083242] x17: ffff8000768f5000 x16: ffff800008004000 x15: 000011acfbd3184b
[   20.090385] x14: 0000000000000159 x13: 0000000000000001 x12: 0000000000000000
[   20.097527] x11: 0000000000000000 x10: 00000000000009a0 x9 : ffff80000965b8b0
[   20.104671] x8 : ffff000000095e00 x7 : ffff00007fb70b40 x6 : 0000000000000107
[   20.111814] x5 : ffff00007fb6a998 x4 : 0000000000000000 x3 : 0000000000000027
[   20.118957] x2 : 0000000000000000 x1 : 0000000000000000 x0 : 0000000000000031
[   20.126104] Call trace:
[   20.128549]  usb_control_msg+0x178/0x180
[   20.132479]  __usbnet_write_cmd.isra.0+0x8c/0xd8
[   20.137098]  usbnet_write_cmd+0x64/0xa0
[   20.140936]  __smsc95xx_write_reg+0x6c/0xc8
[   20.145125]  __smsc95xx_mdio_read+0x78/0x160
[   20.149400]  smsc95xx_mdiobus_read+0x18/0x28
[   20.153674]  mdiobus_read+0x40/0x80
[   20.157167]  lan87xx_read_status+0xe4/0x158
[   20.161355]  phy_check_link_status+0x48/0xc0
[   20.165628]  phy_state_machine+0x1a8/0x228
[   20.169729]  process_one_work+0x1c0/0x338
[   20.173742]  worker_thread+0x50/0x420
[   20.177407]  kthread+0x110/0x120
[   20.180641]  ret_from_fork+0x10/0x20
[   20.184221] ---[ end trace 0000000000000000 ]---
[   20.188952] smsc95xx 2-1.1:1.0 eth1: unregister 'smsc95xx'
usb-ci_hdrc.1-1.1, smsc95xx USB 2.0 Ethernet
[   20.246723] smsc95xx 2-1.1:1.0 eth1: Failed to write reg index
0x00000114: -19
[   20.253972] smsc95xx 2-1.1:1.0 eth1: Error writing MII_ADDR
[   20.259678] ********** returning -ENODEV from usb_control_msg
[   20.265480] ------------[ cut here ]------------
[   20.270099] WARNING: CPU: 0 PID: 257 at
drivers/usb/core/message.c:163 usb_control_msg+0x178/0x180
[   20.279075] Modules linked in:
[   20.282135] CPU: 0 PID: 257 Comm: sh Tainted: G        W
5.17.0-rc6-dirty #581
[   20.290055] Hardware name: CompuLab i.MX8MM IoT Gateway (DT)
[   20.295715] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[   20.302681] pc : usb_control_msg+0x178/0x180
[   20.306956] lr : usb_control_msg+0x178/0x180
[   20.311234] sp : ffff800009c3b310
[   20.314546] x29: ffff800009c3b310 x28: 00000000000000c0 x27: 00000000000000a1
[   20.321692] x26: 0000000000000000 x25: 0000000000000114 x24: 0000000000000004
[   20.328837] x23: 0000000080000380 x22: ffff000006812000 x21: ffff00000d950e00
[   20.335982] x20: 00000000ffffffed x19: ffff00000f26ea00 x18: 0000000000000010
[   20.343125] x17: 0000000000000000 x16: 0000000000000000 x15: ffff00000d951280
[   20.350270] x14: 0000000000000169 x13: ffff00000d951280 x12: 00000000ffffffea
[   20.357417] x11: ffff8000094a1b28 x10: ffff800009489ae8 x9 : ffff800009489b40
[   20.364559] x8 : 0000000000017fe8 x7 : c0000000ffffefff x6 : 0000000000000001
[   20.371703] x5 : 0000000000000000 x4 : ffff00007fb6a998 x3 : ffff00007fb6da08
[   20.378848] x2 : 0000000000000000 x1 : 0000000000000000 x0 : 0000000000000031
[   20.385995] Call trace:
[   20.388441]  usb_control_msg+0x178/0x180
[   20.392371]  __usbnet_read_cmd.isra.0+0x8c/0x130
[   20.396993]  usbnet_read_cmd+0x64/0xa0
[   20.400744]  __smsc95xx_read_reg+0x6c/0xd8
[   20.404845]  __smsc95xx_phy_wait_not_busy+0x60/0xc8
[   20.409729]  __smsc95xx_mdio_read+0x54/0x160
[   20.414007]  smsc95xx_mdiobus_read+0x18/0x28
[   20.418282]  __mdiobus_modify_changed+0x44/0xe0
[   20.422817]  phy_modify+0x48/0x78
[   20.426140]  genphy_suspend+0x1c/0x28
[   20.429804]  phy_suspend+0xa8/0xe8
[   20.433210]  phy_state_machine+0x19c/0x228
[   20.437310]  phy_stop+0x5c/0x198
[   20.440542]  phy_disconnect+0x58/0x60
[   20.444207]  smsc95xx_unbind+0x20/0x48
[   20.447958]  usbnet_disconnect+0x40/0xc8
[   20.451887]  usb_unbind_interface+0x74/0x248
[   20.456161]  device_release_driver_internal+0x1fc/0x200
[   20.461395]  device_release_driver+0x18/0x28
[   20.465669]  bus_remove_device+0x128/0x138
[   20.469768]  device_del+0x16c/0x3a8
[   20.473261]  usb_disable_device+0x8c/0x158
[   20.477363]  usb_disconnect+0xb8/0x2a0
[   20.481116]  hub_quiesce+0x78/0xd0
[   20.484523]  hub_disconnect+0x50/0x190
[   20.488276]  usb_unbind_interface+0x74/0x248
[   20.492551]  device_release_driver_internal+0x1fc/0x200
[   20.497781]  device_release_driver+0x18/0x28
[   20.502055]  bus_remove_device+0x128/0x138
[   20.506157]  device_del+0x16c/0x3a8
[   20.509646]  usb_disable_device+0x8c/0x158
[   20.513750]  usb_set_configuration+0x538/0x898
[   20.518202]  usb_generic_driver_disconnect+0x2c/0x40
[   20.523176]  usb_unbind_device+0x30/0x98
[   20.527100]  device_release_driver_internal+0x1fc/0x200
[   20.532333]  device_driver_detach+0x18/0x28
[   20.536524]  unbind_store+0xe0/0x108
[   20.540101]  drv_attr_store+0x24/0x38
[   20.543769]  sysfs_kf_write+0x44/0x58
[   20.547434]  kernfs_fop_write_iter+0x118/0x1a8
[   20.551881]  new_sync_write+0xe8/0x190
[   20.555638]  vfs_write+0x234/0x378
[   20.559048]  ksys_write+0x6c/0x100
[   20.562457]  __arm64_sys_write+0x1c/0x28
[   20.566388]  invoke_syscall+0x44/0x108
[   20.570140]  el0_svc_common.constprop.0+0xcc/0xf0
[   20.574848]  do_el0_svc+0x24/0x88
[   20.578165]  el0_svc+0x20/0x60
[   20.581228]  el0t_64_sync_handler+0xb0/0xb8
[   20.585415]  el0t_64_sync+0x1a0/0x1a4
[   20.589083] ---[ end trace 0000000000000000 ]---
[   20.594738] smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
[   20.601903] smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
[   20.607687] smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
[   20.618737] ********** returning -ENODEV from usb_control_msg
[   20.624562] ------------[ cut here ]------------
[   20.629180] WARNING: CPU: 3 PID: 257 at
drivers/usb/core/message.c:163 usb_control_msg+0x178/0x180
[   20.638154] Modules linked in:
[   20.641213] CPU: 3 PID: 257 Comm: sh Tainted: G        W
5.17.0-rc6-dirty #581
[   20.649133] Hardware name: CompuLab i.MX8MM IoT Gateway (DT)
[   20.654794] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[   20.661759] pc : usb_control_msg+0x178/0x180
[   20.666032] lr : usb_control_msg+0x178/0x180
[   20.670311] sp : ffff800009c3b370
[   20.673623] x29: ffff800009c3b370 x28: 00000000000000c0 x27: 00000000000000a1
[   20.680766] x26: 0000000000000000 x25: 0000000000000114 x24: 0000000000000004
[   20.687911] x23: 0000000080000380 x22: ffff000006812000 x21: ffff00000d950e00
[   20.695058] x20: 00000000ffffffed x19: ffff00000f0c6780 x18: 0000000000000010
[   20.702199] x17: ffff80007692e000 x16: ffff8000095e4000 x15: ffff00000d951280
[   20.709345] x14: 00000000000001b4 x13: ffff00000d951280 x12: 00000000ffffffea
[   20.716490] x11: ffff8000094a1b28 x10: ffff800009489ae8 x9 : ffff800009489b40
[   20.723634] x8 : 0000000000017fe8 x7 : c0000000ffffefff x6 : 0000000000000001
[   20.730778] x5 : 0000000000000000 x4 : ffff00007fba3998 x3 : ffff00007fba6a08
[   20.737921] x2 : 0000000000000000 x1 : 0000000000000000 x0 : 0000000000000031
[   20.745068] Call trace:
[   20.747514]  usb_control_msg+0x178/0x180
[   20.751445]  __usbnet_read_cmd.isra.0+0x8c/0x130
[   20.756064]  usbnet_read_cmd+0x64/0xa0
[   20.759816]  __smsc95xx_read_reg+0x6c/0xd8
[   20.763920]  __smsc95xx_phy_wait_not_busy+0x60/0xc8
[   20.768803]  __smsc95xx_mdio_read+0x54/0x160
[   20.773079]  smsc95xx_mdiobus_read+0x18/0x28
[   20.777356]  __mdiobus_modify_changed+0x44/0xe0
[   20.781891]  phy_modify+0x48/0x78
[   20.785210]  genphy_suspend+0x1c/0x28
[   20.788878]  phy_suspend+0xa8/0xe8
[   20.792280]  phy_detach+0x30/0x110
[   20.795686]  phy_disconnect+0x44/0x60
[   20.799349]  smsc95xx_unbind+0x20/0x48
[   20.803102]  usbnet_disconnect+0x40/0xc8
[   20.807030]  usb_unbind_interface+0x74/0x248
[   20.811305]  device_release_driver_internal+0x1fc/0x200
[   20.816535]  device_release_driver+0x18/0x28
[   20.820811]  bus_remove_device+0x128/0x138
[   20.824909]  device_del+0x16c/0x3a8
[   20.828400]  usb_disable_device+0x8c/0x158
[   20.832504]  usb_disconnect+0xb8/0x2a0
[   20.836256]  hub_quiesce+0x78/0xd0
[   20.839663]  hub_disconnect+0x50/0x190
[   20.843416]  usb_unbind_interface+0x74/0x248
[   20.847689]  device_release_driver_internal+0x1fc/0x200
[   20.852921]  device_release_driver+0x18/0x28
[   20.857195]  bus_remove_device+0x128/0x138
[   20.861297]  device_del+0x16c/0x3a8
[   20.864786]  usb_disable_device+0x8c/0x158
[   20.868888]  usb_set_configuration+0x538/0x898
[   20.873340]  usb_generic_driver_disconnect+0x2c/0x40
[   20.878312]  usb_unbind_device+0x30/0x98
[   20.882236]  device_release_driver_internal+0x1fc/0x200
[   20.887469]  device_driver_detach+0x18/0x28
[   20.891658]  unbind_store+0xe0/0x108
[   20.895235]  drv_attr_store+0x24/0x38
[   20.898903]  sysfs_kf_write+0x44/0x58
[   20.902568]  kernfs_fop_write_iter+0x118/0x1a8
[   20.907015]  new_sync_write+0xe8/0x190
[   20.910773]  vfs_write+0x234/0x378
[   20.914181]  ksys_write+0x6c/0x100
[   20.917590]  __arm64_sys_write+0x1c/0x28
[   20.921521]  invoke_syscall+0x44/0x108
[   20.925273]  el0_svc_common.constprop.0+0xcc/0xf0
[   20.929983]  do_el0_svc+0x24/0x88
[   20.933302]  el0_svc+0x20/0x60
[   20.936365]  el0t_64_sync_handler+0xb0/0xb8
[   20.940557]  el0t_64_sync+0x1a0/0x1a4
[   20.944223] ---[ end trace 0000000000000000 ]---
[   20.954744] smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
[   20.961913] smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
[   20.967696] smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
[   20.979207] smsc95xx 2-1.1:1.0 eth1: hardware isn't capable of remote wakeup
[   21.030745] usb 2-1.4: USB disconnect, device number 4


root@iotgate:~# echo -n "2-1" > /sys/bus/usb/drivers/usb/bind
[   48.591950] hub 2-1:1.0: USB hub found
[   48.595909] hub 2-1:1.0: 5 ports detected
root@iotgate:~# [   48.886892] usb 2-1.1: new high-speed USB device
number 5 using ci_hdrc
[   49.002509] smsc95xx v2.0.0
[   49.098891] SMSC LAN8700 usb-002:005:01: attached PHY driver
(mii_bus:phy_addr=usb-002:005:01, irq=POLL)
[   49.110400] smsc95xx 2-1.1:1.0 eth1: register 'smsc95xx' at
usb-ci_hdrc.1-1.1, smsc95xx USB 2.0 Ethernet, 00:01:c0:2f:36:3d
[   49.190808] smsc95xx 2-1.1:1.0 eth1: hardware isn't capable of remote wakeup
[   49.202758] usb 2-1.4: new full-speed USB device number 6 using ci_hdrc
[   49.212264] smsc95xx 2-1.1:1.0 eth1: Link is Down

root@iotgate:~#
root@iotgate:~#
root@iotgate:~# echo -n "2-1" > /sys/bus/usb/drivers/usb/unbind
[   52.302633] usb 2-1.1: USB disconnect, device number 5
[   52.310139] smsc95xx 2-1.1:1.0 eth1: unregister 'smsc95xx'
usb-ci_hdrc.1-1.1, smsc95xx USB 2.0 Ethernet
[   52.319574] ********** returning -ENODEV from usb_control_msg
[   52.325388] ------------[ cut here ]------------
[   52.330016] WARNING: CPU: 1 PID: 257 at
drivers/usb/core/message.c:163 usb_control_msg+0x178/0x180
[   52.338989] Modules linked in:
[   52.342050] CPU: 1 PID: 257 Comm: sh Tainted: G        W
5.17.0-rc6-dirty #581
[   52.349969] Hardware name: CompuLab i.MX8MM IoT Gateway (DT)
[   52.355629] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[   52.362595] pc : usb_control_msg+0x178/0x180
[   52.366868] lr : usb_control_msg+0x178/0x180
[   52.371144] sp : ffff800009c3b310
[   52.374456] x29: ffff800009c3b310 x28: 00000000000000c0 x27: 00000000000000a1
[   52.381602] x26: 0000000000000000 x25: 0000000000000114 x24: 0000000000000004
[   52.388747] x23: 0000000080000580 x22: ffff000006814800 x21: ffff00000d950e00
[   52.395892] x20: 00000000ffffffed x19: ffff00000d021d80 x18: 0000000000000010
[   52.403037] x17: 312e312d312e6372 x16: 64685f69632d6273 x15: ffff00000d951280
[   52.410180] x14: 000000000000020b x13: ffff00000d951280 x12: 00000000ffffffea
[   52.417324] x11: ffff8000094a1b28 x10: ffff800009489ae8 x9 : ffff800009489b40
[   52.424470] x8 : 0000000000017fe8 x7 : c0000000ffffefff x6 : 0000000000000001
[   52.431612] x5 : 0000000000000000 x4 : ffff00007fb7d998 x3 : ffff00007fb80a08
[   52.438757] x2 : 0000000000000000 x1 : 0000000000000000 x0 : 0000000000000031
[   52.445904] Call trace:
[   52.448351]  usb_control_msg+0x178/0x180
[   52.452283]  __usbnet_read_cmd.isra.0+0x8c/0x130
[   52.456903]  usbnet_read_cmd+0x64/0xa0
[   52.460654]  __smsc95xx_read_reg+0x6c/0xd8
[   52.464757]  __smsc95xx_phy_wait_not_busy+0x60/0xc8
[   52.469641]  __smsc95xx_mdio_read+0x54/0x160
[   52.473918]  smsc95xx_mdiobus_read+0x18/0x28
[   52.478192]  __mdiobus_modify_changed+0x44/0xe0
[   52.482728]  phy_modify+0x48/0x78
[   52.486050]  genphy_suspend+0x1c/0x28
[   52.489714]  phy_suspend+0xa8/0xe8
[   52.493118]  phy_state_machine+0x19c/0x228
[   52.497218]  phy_stop+0x5c/0x198
[   52.500450]  phy_disconnect+0x58/0x60
[   52.504115]  smsc95xx_unbind+0x20/0x48
[   52.507870]  usbnet_disconnect+0x40/0xc8
[   52.511797]  usb_unbind_interface+0x74/0x248
[   52.516072]  device_release_driver_internal+0x1fc/0x200
[   52.521303]  device_release_driver+0x18/0x28
[   52.525579]  bus_remove_device+0x128/0x138
[   52.529679]  device_del+0x16c/0x3a8
[   52.533173]  usb_disable_device+0x8c/0x158
[   52.537276]  usb_disconnect+0xb8/0x2a0
[   52.541029]  hub_quiesce+0x78/0xd0
[   52.544436]  hub_disconnect+0x50/0x190
[   52.548189]  usb_unbind_interface+0x74/0x248
[   52.552463]  device_release_driver_internal+0x1fc/0x200
[   52.557695]  device_release_driver+0x18/0x28
[   52.561969]  bus_remove_device+0x128/0x138
[   52.566071]  device_del+0x16c/0x3a8
[   52.569561]  usb_disable_device+0x8c/0x158
[   52.573665]  usb_set_configuration+0x538/0x898
[   52.578117]  usb_generic_driver_disconnect+0x2c/0x40
[   52.583088]  usb_unbind_device+0x30/0x98
[   52.587011]  device_release_driver_internal+0x1fc/0x200
[   52.592244]  device_driver_detach+0x18/0x28
[   52.596436]  unbind_store+0xe0/0x108
[   52.600012]  drv_attr_store+0x24/0x38
[   52.603678]  sysfs_kf_write+0x44/0x58
[   52.607349]  kernfs_fop_write_iter+0x118/0x1a8
[   52.611800]  new_sync_write+0xe8/0x190
[   52.615555]  vfs_write+0x234/0x378
[   52.618962]  ksys_write+0x6c/0x100
[   52.622370]  __arm64_sys_write+0x1c/0x28
[   52.626300]  invoke_syscall+0x44/0x108
[   52.630053]  el0_svc_common.constprop.0+0xcc/0xf0
[   52.634763]  do_el0_svc+0x24/0x88
[   52.638080]  el0_svc+0x20/0x60
[   52.641142]  el0t_64_sync_handler+0xb0/0xb8
[   52.645335]  el0t_64_sync+0x1a0/0x1a4
[   52.649001] ---[ end trace 0000000000000000 ]---
[   52.653704] smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
[   52.660885] smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
[   52.666664] smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
[   52.673255] ********** returning -ENODEV from usb_control_msg
[   52.679076] ------------[ cut here ]------------
[   52.683697] WARNING: CPU: 1 PID: 257 at
drivers/usb/core/message.c:163 usb_control_msg+0x178/0x180
[   52.692671] Modules linked in:
[   52.695731] CPU: 1 PID: 257 Comm: sh Tainted: G        W
5.17.0-rc6-dirty #581
[   52.703651] Hardware name: CompuLab i.MX8MM IoT Gateway (DT)
[   52.709311] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[   52.716276] pc : usb_control_msg+0x178/0x180
[   52.720551] lr : usb_control_msg+0x178/0x180
[   52.724827] sp : ffff800009c3b370
[   52.728139] x29: ffff800009c3b370 x28: 00000000000000c0 x27: 00000000000000a1
[   52.735285] x26: 0000000000000000 x25: 0000000000000114 x24: 0000000000000004
[   52.742428] x23: 0000000080000580 x22: ffff000006814800 x21: ffff00000d950e00
[   52.749573] x20: 00000000ffffffed x19: ffff00000d021d80 x18: 0000000000000010
[   52.756718] x17: 312d203a34313130 x16: 3030303078302078 x15: ffff00000d951280
[   52.763862] x14: 0000000000000256 x13: ffff00000d951280 x12: 00000000ffffffea
[   52.771005] x11: ffff8000094a1b28 x10: ffff800009489ae8 x9 : ffff800009489b40
[   52.778150] x8 : 0000000000017fe8 x7 : c0000000ffffefff x6 : 0000000000000001
[   52.785294] x5 : 0000000000000000 x4 : ffff00007fb7d998 x3 : ffff00007fb80a08
[   52.792440] x2 : 0000000000000000 x1 : 0000000000000000 x0 : 0000000000000031
[   52.799583] Call trace:
[   52.802030]  usb_control_msg+0x178/0x180
[   52.805960]  __usbnet_read_cmd.isra.0+0x8c/0x130
[   52.810579]  usbnet_read_cmd+0x64/0xa0
[   52.814330]  __smsc95xx_read_reg+0x6c/0xd8
[   52.818432]  __smsc95xx_phy_wait_not_busy+0x60/0xc8
[   52.823312]  __smsc95xx_mdio_read+0x54/0x160
[   52.827589]  smsc95xx_mdiobus_read+0x18/0x28
[   52.831867]  __mdiobus_modify_changed+0x44/0xe0
[   52.836401]  phy_modify+0x48/0x78
[   52.839721]  genphy_suspend+0x1c/0x28
[   52.843389]  phy_suspend+0xa8/0xe8
[   52.846791]  phy_detach+0x30/0x110
[   52.850195]  phy_disconnect+0x44/0x60
[   52.853859]  smsc95xx_unbind+0x20/0x48
[   52.857615]  usbnet_disconnect+0x40/0xc8
[   52.861544]  usb_unbind_interface+0x74/0x248
[   52.865819]  device_release_driver_internal+0x1fc/0x200
[   52.871049]  device_release_driver+0x18/0x28
[   52.875326]  bus_remove_device+0x128/0x138
[   52.879426]  device_del+0x16c/0x3a8
[   52.882918]  usb_disable_device+0x8c/0x158
[   52.887023]  usb_disconnect+0xb8/0x2a0
[   52.890774]  hub_quiesce+0x78/0xd0
[   52.894181]  hub_disconnect+0x50/0x190
[   52.897934]  usb_unbind_interface+0x74/0x248
[   52.902206]  device_release_driver_internal+0x1fc/0x200
[   52.907438]  device_release_driver+0x18/0x28
[   52.911714]  bus_remove_device+0x128/0x138
[   52.915815]  device_del+0x16c/0x3a8
[   52.919304]  usb_disable_device+0x8c/0x158
[   52.923409]  usb_set_configuration+0x538/0x898
[   52.927861]  usb_generic_driver_disconnect+0x2c/0x40
[   52.932832]  usb_unbind_device+0x30/0x98
[   52.936755]  device_release_driver_internal+0x1fc/0x200
[   52.941986]  device_driver_detach+0x18/0x28
[   52.946173]  unbind_store+0xe0/0x108
[   52.949750]  drv_attr_store+0x24/0x38
[   52.953418]  sysfs_kf_write+0x44/0x58
[   52.957086]  kernfs_fop_write_iter+0x118/0x1a8
[   52.961533]  new_sync_write+0xe8/0x190
[   52.965291]  vfs_write+0x234/0x378
[   52.968697]  ksys_write+0x6c/0x100
[   52.972104]  __arm64_sys_write+0x1c/0x28
[   52.976033]  invoke_syscall+0x44/0x108
[   52.979785]  el0_svc_common.constprop.0+0xcc/0xf0
[   52.984498]  do_el0_svc+0x24/0x88
[   52.987818]  el0_svc+0x20/0x60
[   52.990877]  el0t_64_sync_handler+0xb0/0xb8
[   52.995067]  el0t_64_sync+0x1a0/0x1a4
[   52.998736] ---[ end trace 0000000000000000 ]---
[   53.003408] smsc95xx 2-1.1:1.0 eth1: Failed to read reg index 0x00000114: -19
[   53.010584] smsc95xx 2-1.1:1.0 eth1: Error reading MII_ACCESS
[   53.016367] smsc95xx 2-1.1:1.0 eth1: __smsc95xx_mdio_read: MII is busy
[   53.023423] smsc95xx 2-1.1:1.0 eth1: hardware isn't capable of remote wakeup
[   53.059462] usb 2-1.4: USB disconnect, device number 6

Re: smsc9511: Register access happens after unregistration

[Andrew Lunn]

On Fri, Mar 04, 2022 at 03:01:28PM -0300, Fabio Estevam wrote:
> On Fri, Mar 4, 2022 at 12:57 PM Fabio Estevam <festevam@gmail.com> wrote:
> >
> > On Fri, Mar 4, 2022 at 11:51 AM Andrew Lunn <andrew@lunn.ch> wrote:
> >
> > > But why does it return ENODEV? It seems to me, ignoring it is papering
> > > over the cracks. Why cannot we access to the PHY?
> >
> > The -ENODEV is returned by usb_control_msg():
> >
> > __smsc95xx_read_reg: -19
> >          usbnet_read_cmd: -19
> >               usb_control_msg: -19
> 
> I added a WARN_ON() inside usb_control_msg() that triggers when
> usb_control_msg() returns -ENODEV.

https://elixir.bootlin.com/linux/v5.17-rc6/source/include/linux/usb.h#L1126

say:

 * @disconnect: Called when the interface is no longer accessible, usually
 *	because its device has been (or is being) disconnected or the
 *	driver module is being unloaded.

So i guess the USB core has disconnected the device, and is blocking
further control messages. I guess it handles shutdown the same as hot
unplug. So there are a couple things that can be done to make this
better.

Make __smsc95xx_phy_wait_not_busy(), __smsc95xx_mdio_read() and
__smsc95xx_mdio_write() not print a message on -ENODEV. I would
continue printing the warning for other error codes. The phylib and
phy drivers should be O.K. if they get -ENODEV.

Non-USB Ethernet drivers follow one of two patterns:

1) The PHY is connected in probe, started in open, stopped in close,
   and disconnected in remove.

2) The PHY is both connected and started in open, and stopped and
   disconnected in close.

Depending on what userspace you are using, i think it is normal for
userspace to ifdown interfaces during shutdown. So if you make use of
2) the PHY should be disconnected by the time the USB subsystem calls
the USB disconnect callback. However, for hot unplug, there will not
be an ifdown, so the disconnect callback needs to look at state
information and stop and disconnect the PHY if it is currently
connected and started.

	Andrew