Re: [PATCHv2 net] usbnet: modern method to get random MAC

[Eric Dumazet]

On Tue, Oct 15, 2024 at 12:24 AM Oliver Neukum <oneukum@xxxxxxxx> wrote:
>
> On 14.10.24 21:59, Eric Dumazet wrote:
>
> > As diagnosed by John Sperbeck :
> >
> > This patch implies all ->bind() method took care of populating net->dev_addr ?
> >
> > Otherwise the following existing heuristic is no longer working
> >
> > // heuristic:  "usb%d" for links we know are two-host,
> > // else "eth%d" when there's reasonable doubt.  userspace
> > // can rename the link if it knows better.
> > if ((dev->driver_info->flags & FLAG_ETHER) != 0 &&
> >      ((dev->driver_info->flags & FLAG_POINTTOPOINT) == 0 ||
> >       (net->dev_addr [0] & 0x02) == 0))
> > strscpy(net->name, "eth%d", sizeof(net->name));
> >
>
> Hi,
>
> you need to have a MAC to be an ethernet device, don't you?

Before or after your patch, there was/is a MAC address, eventually random.

The problem is about the test, which is now done while dev->dev_addr
is full of zeroes, which is not a valid address,
as shown by :

diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c
index e4775fb5a2f6..1a316773319f 100644
--- a/drivers/net/usb/usbnet.c
+++ b/drivers/net/usb/usbnet.c
@@ -1750,7 +1750,8 @@ usbnet_probe (struct usb_interface *udev, const
struct usb_device_id *prod)
                // can rename the link if it knows better.
                if ((dev->driver_info->flags & FLAG_ETHER) != 0 &&
                    ((dev->driver_info->flags & FLAG_POINTTOPOINT) == 0 ||
-                    (net->dev_addr [0] & 0x02) == 0))
+                    (is_valid_ether_addr(net->dev_addr) &&
+                     (net->dev_addr [0] & 0x02) == 0)))
                        strscpy(net->name, "eth%d", sizeof(net->name));
                /* WLAN devices should always be named "wlan%d" */
                if ((dev->driver_info->flags & FLAG_WLAN) != 0)

To be clear : We are hitting a regression after your patch was
backported to stable versions.