`dev_addr` in `struct net_device` has now been made `const` about ldd3

[pauljherring]

Had to revert my changes (because, reasons,) but still needed apply the use of eth_hw_addr_set - the diff as a result:

diff --git a/snull/snull.c b/snull/snull.c
index 2efc2b5..adc5899 100644
--- a/snull/snull.c
+++ b/snull/snull.c
@@ -213,9 +213,10 @@ int snull_open(struct net_device *dev)
         * x is 0 or 1. The first byte is '\0' to avoid being a multicast
         * address (the first byte of multicast addrs is odd).
         */
-       memcpy(dev->dev_addr, "\0SNUL0", ETH_ALEN);
-       if (dev == snull_devs[1])
-               dev->dev_addr[ETH_ALEN-1]++; /* \0SNUL1 */
+       if (dev == snull_devs[0])
+               eth_hw_addr_set(dev, "\0SNUL0");
+       else
+               eth_hw_addr_set(dev, "\0SNUL1");
        if (use_napi) {
                struct snull_priv *priv = netdev_priv(dev);
                napi_enable(&priv->napi);

[dwalkes]

Thanks @pauljherring do you want to create a pull request for this?
Should probably check for kernel 5.15 and later based on torvalds/linux@48eab83

See

ldd3/sbull/sbull.c

Line 93 in bf49bb4

#if (LINUX_VERSION_CODE < KERNEL_VERSION(5, 9, 0))

for an example use of KERNEL_VERSION

[pauljherring]

.. I'm not at all sure that I'm the right person to be doing a pull request on this project, or for this in particular:

1. it's not clear it's sufficient - even after compiling, the driver doesn't "work" as expected for me (pings go to the first interface, but don't make it out the other end, for example); this merely "gets it to compile"
2. whether anything else in the repo needs looking at; I'm only concentrating on the network driver because I need to deal with an in-house network driver and this is part of my self-education on such matters.