Re: [PATCH 07/13] mm: meminit: Initialise a subset of struct pages if CONFIG_DEFERRED_STRUCT_PAGE_INIT is set

[Mel Gorman]

On Wed, Apr 29, 2015 at 02:19:01PM -0700, Andrew Morton wrote:
> On Tue, 28 Apr 2015 15:37:04 +0100 Mel Gorman <mgorman@xxxxxxx> wrote:
> 
> > +/*
> > + * Deferred struct page initialisation requires some early init functions that
> > + * are removed before kswapd is up and running. The feature depends on memory
> > + * hotplug so put the data and code required by deferred initialisation into
> > + * the __meminit section where they are preserved.
> > + */
> > +#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
> > +#define __defermem_init __meminit
> > +#define __defer_init    __meminit
> > +#else
> > +#define __defermem_init
> > +#define __defer_init __init
> > +#endif
> 
> I still don't get it :(
> 

This version was sent out at roughly the same minute you asked the time
before so the comment was not updated. I suggested this as a possible
alternative.

/*
 * Deferred struct page initialisation requires init functions that are freed
 * before kswapd is available. Reuse the memory hotplug section annotation
 * to mark the required code.
 *
 * __defermem_init is code that always exists but is annotated __meminit * to
 *      avoid section warnings.
 * __defer_init code gets marked __meminit when deferring struct page
 *      initialistion but is otherwise in the init section.
 */

Suggestions on better names are welcome.

> __defermem_init:
> 
> 	if (CONFIG_DEFERRED_STRUCT_PAGE_INIT) {
> 		if (CONFIG_MEMORY_HOTPLUG)
> 			retain
> 	} else {
> 		retain
> 	}
> 
>     but CONFIG_DEFERRED_STRUCT_PAGE_INIT depends on
>     CONFIG_MEMORY_HOTPLUG, so this becomes
> 
> 	if (CONFIG_DEFERRED_STRUCT_PAGE_INIT) {
> 		retain
> 	} else {
> 		retain
> 	}
> 
>     which becomes
> 
> 	retain
> 
>     so why does __defermem_init exist?
> 

It suppresses section warnings. Another possibility is that I get rid of
it entirely and use __refok but I feared that it might hide a real problem
in the future.

-- 
Mel Gorman
SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/