Re: [PATCH] init: add comments to keep initcall-names in sync with initcall levels

[Paul Bolle <pebolle@tiscali.nl>]

On Thu, 2012-06-14 at 14:32 -0600, Jim Cromie wrote:
> On Thu, Jun 14, 2012 at 1:23 PM, Paul Bolle <pebolle@tiscali.nl> wrote:
> > On Thu, 2012-06-14 at 12:51 -0600, Jim Cromie wrote:
> >>
> >> -/* Don't use these in modules, but some people do... */
> >> +/* Don't use these in loadable modules, but some people do... */
> >
> > What problem does this solve?
> 
> kernel/params.c  and other builtins are also modules - at least wrt
> how theyre reported by dynamic_debug:
> 
> kernel/params.c:121 [params]parse_one =_ "Unknown argument `%s'\012"
> kernel/params.c:117 [params]parse_one =_ "Unknown argument: calling %p\012"
> kernel/params.c:108 [params]parse_one =_ "They are equal!  Calling %p\012"
> kernel/params.c:188 [params]parse_args =_ "Parsing ARGS: %s\012"
> 
> The advice to avoid those macros does not apply to builtin "modules"

I don't think I use dynamic_debug, but still, a pair of square brackets
doesn't make that some part of the kernel is considered to be a module,
does it? And more importantly, even if there's a difference between
"module" and "loadable module", which I rather doubt, aren't the people
who are expected to read this comment also expected to understand the
relevance of the preceding 
    #else /* MODULE */

line?


Paul Bolle