On Mon, Dec 13, 2021 at 3:50 PM Rob Herring robh+dt@kernel.org wrote:
On Mon, Dec 13, 2021 at 5:30 AM Javier Martinez Canillas javier@dowhile0.org wrote:
[snip]
You are right that passing NULL is a safe code path for now due the of_device_is_available(node) check, but that seems fragile to me since just adding a similar debug output to of_platform_device_create() could trigger the NULL pointer dereference.
All/most DT functions work with a NULL node ptr, so why should this one be different?
If you are OK with the patch as is, then I won't object :)
Rob
Best regards, Javier