Re: [PATCH] mm: fix race between kmem_cache destroy, create and deactivate

[Christopher Lameter]

On Mon, 21 May 2018, Andrew Morton wrote:

> The patch seems depressingly complex.
>
> And a bit underdocumented...

Maybe separate out the bits that rename refcount to alias_count?

> > +	refcount_t refcount;
> > +	int alias_count;
>
> The semantic meaning of these two?  What locking protects alias_count?

slab_mutex

>
> >  	int object_size;
> >  	int align;
> >
> > diff --git a/include/linux/slub_def.h b/include/linux/slub_def.h
> > index 3773e26c08c1..532d4b6f83ed 100644
> > --- a/include/linux/slub_def.h
> > +++ b/include/linux/slub_def.h
> > @@ -97,7 +97,8 @@ struct kmem_cache {
> >  	struct kmem_cache_order_objects max;
> >  	struct kmem_cache_order_objects min;
> >  	gfp_t allocflags;	/* gfp flags to use on each alloc */
> > -	int refcount;		/* Refcount for slab cache destroy */
> > +	refcount_t refcount;	/* Refcount for slab cache destroy */
> > +	int alias_count;	/* Number of root kmem caches merged */
>
> "merged" what with what in what manner?

That is a basic SLUB feature.