Re: [RFC] usbnet: assign unique random MAC

[Benjamin Poirier <benjamin.poirier@gmail.com>]

On 2023-11-16 15:05 +0100, Oliver Neukum wrote:
> The old method had the bug of issuing the same
> random MAC over and over even to every device.
> This bug is as old as the driver.
> 
> This new method generates each device whose minidriver
> does not provide its own MAC its own unique random
> MAC.
> 
> Signed-off-by: Oliver Neukum <oneukum@suse.com>
> ---
>  drivers/net/usb/usbnet.c | 15 +++++++--------
>  1 file changed, 7 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c
> index 2d14b0d78541..37e3bb2170bc 100644
> --- a/drivers/net/usb/usbnet.c
> +++ b/drivers/net/usb/usbnet.c
> @@ -61,9 +61,6 @@
>  
>  /*-------------------------------------------------------------------------*/
>  
> -// randomly generated ethernet address
> -static u8	node_id [ETH_ALEN];
> -
>  /* use ethtool to change the level for any given device */
>  static int msg_level = -1;
>  module_param (msg_level, int, 0);
> @@ -1731,7 +1728,6 @@ usbnet_probe (struct usb_interface *udev, const struct usb_device_id *prod)
>  
>  	dev->net = net;
>  	strscpy(net->name, "usb%d", sizeof(net->name));
> -	eth_hw_addr_set(net, node_id);
>  
>  	/* rx and tx sides can use different message sizes;
>  	 * bind() should set rx_urb_size in that case.
> @@ -1805,9 +1801,13 @@ usbnet_probe (struct usb_interface *udev, const struct usb_device_id *prod)
>  		goto out4;
>  	}
>  
> -	/* let userspace know we have a random address */
> -	if (ether_addr_equal(net->dev_addr, node_id))
> -		net->addr_assign_type = NET_ADDR_RANDOM;
> +	/*
> +	 * if the device does not come with a MAC

The patch's formatting has some problems. Please run checkpatch before
resubmitting.

> +	 * we ask the network core to generate us one
> +	 * and flag the device accordingly
> +	 */
> +	if (!is_valid_ether_addr(net->dev_addr))
> +			eth_hw_addr_random(net);

Before initialization, dev_addr is null (00:00:00:00:00:00). Since this
patch moves the fallback address initialization after the 
	if (info->bind) {
block, if the bind() did not initialize the address, this patch changes
the result of the
		     (net->dev_addr [0] & 0x02) == 0))
test within the block, no? The test now takes place on an uninitialized
address and the result goes from false to true.