Re: driver(9)

On Sat, Mar 28, 2026 at 11:54:50 +0000, Patrick Welche wrote:

> On Sat, Mar 28, 2026 at 02:50:35PM +0300, Valery Ushakov wrote:
>
> > You are a bit confused here.  device_t is a _pointer_ to the struct
> > device, not inline struct device (as it used to be).  You can check
> > cvs history for commits that (among other things) do:
> > 
> > -	struct device foo_dev;
> > +	device_t foo_dev;
> > 
> > usually with something like "split softc and device_t" in the log
> > message.
> 
> oh - an entire struct device was contained in each struct softc(!)

As the log message you quote in your first mail says (emphasis mine):

  for drivers that don't expect _struct device_ as first field of softc

:)

-uwe