Re: [PATCH v2 2/2] Add TAINT_DRIVER_OVERRIDE for usage of driver_override

On Sun Sep 27, 2026 at 6:50 PM CEST, Greg Kroah-Hartman wrote:

Huh?  That feels wrong, can't we fix s390 and spi instead?  Ok, maybe
not s390, but why is spi doing that?

Solely to preserve SPI's existing output; when driver_override is unset, SPI has
always printed "\n", whereas most other buses print "(null)\n".

I.e. SPI previously did

	sysfs_emit(buf, "%s\n", spi->driver_override ? : "");

now it does

	sysfs_emit(buf, "%s\n", dev->driver_override.name ?: "");

and the driver core does

	sysfs_emit(buf, "%s\n", dev->driver_override.name);

The driver-core behavior matches most other buses (except SPI and AP).
Switching SPI to it would have changed the userspace-visible value.