Re: [PATCH net-next 2/4] tools: ynl: user space helpers

[Simon Horman <simon.horman@corigine.com>]

On Sun, Jun 04, 2023 at 11:00:34AM -0700, Jakub Kicinski wrote:
> On Sat, 3 Jun 2023 16:47:10 +0200 Simon Horman wrote:
> > > +/**
> > > + * struct ynl_error - error encountered by YNL
> > > + * Users should interact with the err member of struct ynl_sock directly.
> > > + * The main exception to that rule is ynl_sock_create().  
> > 
> > nit: As this is a kernel doc, maybe document the structure members here.
> > 
> > > + */
> > > +struct ynl_error {
> > > +	enum ynl_error_code code;
> > > +	unsigned int attr_offs;
> > > +	char msg[512];
> > > +};
> > > +
> > > +/**
> > > + * struct ynl_family - YNL family info
> > > + * Family description generated by codegen.  
> > 
> > And here.
> > 
> > > + */
> > > +struct ynl_family {
> > > +	const char *name;
> > > +	const struct ynl_ntf_info *ntf_info;
> > > +	unsigned int ntf_info_size;
> > > +};  
> > 
> > ...
> > 
> > > +/**
> > > + * ynl_has_ntf() - check if socket has *parsed* notifications
> > > + * Note that this does not take into account notifications sitting
> > > + * in netlink socket, just the notifications which have already been
> > > + * read and parsed (e.g. during a ynl_ntf_check() call).  
> > 
> > And the parameter of this function here.
> 
> Thanks, not sure why my brain considered this "not really kernel code
> so I don't have to obey the rules" :S I marked the innards of the
> family as private, and tossed in the other descriptions. v2 posted.

Thanks, I assumed it was something like that.