Re: [PATCH net-next 2/2] net: sgi: ioc3-eth: fall back to the platform MAC address

From: Imre Kaloz

Date: Mon Nov 16 2026 - 05:43:05 EST


> Can eth_platform_get_mac_address() ever succeed for this device?

With the IP35 series it does: that series provides
arch_get_platform_mac_address(), which reads the address from the L1
system controller. Without it, as you say, it never does.

> Does this ordering let the platform address win before the 1-wire
> PROM has had a chance to appear?

Yes, and it matters for an IOC3 PCI card such as the CAD Duo, which
carries its own MAC PROM. The system-board IOC3 on these machines has
no PROM; a card in a PCI slot does.

> Could IP27/IP30 boards that do have a DS250x MAC PROM end up with a
> random MAC here?

Yes, with the w1 drivers as modules loaded after the deferred probe
timeout.

v2 takes the address from a device property instead. The IOC3 MFD
driver will attach it only to the Ethernet cell of the system-board
IOC3 that has no PROM, so every other IOC3 keeps waiting for its
PROM as before. The random fallback is gone:

https://lore.kernel.org/all/20261116102115.27574-1-kaloz@xxxxxxxxxx/

> ioc3eth_probe() never calls platform_set_drvdata(pdev, dev)
> [...]
> Is the shared IRQ handler safe with the current ordering in
> ioc3eth_probe()?

Both are real. I reproduced them on a Fuel by unbinding the device
through sysfs, the second one with CONFIG_DEBUG_SHIRQ, and the fixes
are posted separately to net:

https://lore.kernel.org/all/20261116101934.25910-1-kaloz@xxxxxxxxxx/

pw-bot: cr


Thanks,
Imre