Re: [PATCH v3 7/9] net: sh_eth: Adjust RZ/A1, add RZ/A2 support

[Marek Vasut]

On 7/6/25 1:29 PM, Magnus Damm wrote:

+++ work/drivers/net/sh_eth.c   2025-07-05 17:45:07.333754799 +0900
@@ -144,10 +144,10 @@ static int sh_eth_reset(struct sh_eth_in
  {
  #if defined(SH_ETH_TYPE_GETHER) || defined(SH_ETH_TYPE_RZ)
        int ret = 0, i;
-

Does something like this ...

if (!device_is_compatible(dev, "..."))

... work instead of the ifdef ?

If yes, it might be even improved into dedicated function:

static bool device_is_rza2(struct udevice *dev)
{
  if (!IS_ENABLED(CONFIG_RZA2))
    return false;
  else
    return device_is_compatible(dev, "...");
}

That should be compile-time checked and then optimized out, and in case 
of non-RZA2, this should be optimized out entirely.

+#if !defined(CONFIG_RZA2)
        /* Start e-dmac transmitter and receiver */
        sh_eth_write(port_info, EDSR_ENALL, EDSR);
-

[...]

@@ -722,8 +726,11 @@ static int sh_ether_probe(struct udevice
                goto err_mdio_register;
  
  	priv->bus = mdiodev;
-
+#ifdef BASE_IO_ADDR
        port_info->iobase = (void __iomem *)(uintptr_t)BASE_IO_ADDR;
+#else
+       port_info->iobase = (void __iomem *)pdata->iobase;
+#endif

Can the base address be pulled from DT?

[...]