On 6/15/25 10:54 PM, Fedor Pchelkin wrote:
So it looks like ar->registered being false here is a "correct" failure condition, i.e. it can be expected when the certain phase of the driver initialization fails and should be handled without any WARNs.
Looking through Documentation/process/coding-style.rst, it may be better to use pr_warn_once() instead; anyway I would prefer leave the final decision to the maintainer. Dmitry
