Re: [PATCH v6 2/4] mm/slub: only zero the requested size of buffer for kzalloc

[Andrey Konovalov]

On Tue, Sep 27, 2022 at 4:42 AM Feng Tang <feng.tang@xxxxxxxxx> wrote:
>
> > > @@ -746,7 +747,7 @@ static inline void slab_post_alloc_hook(struct kmem_cache *s,
> > >         for (i = 0; i < size; i++) {
> > >                 p[i] = kasan_slab_alloc(s, p[i], flags, init);
> > >                 if (p[i] && init && !kasan_has_integrated_init())
> > > -                       memset(p[i], 0, s->object_size);
> > > +                       memset(p[i], 0, orig_size);
> >
> > Note that when KASAN is enabled and has integrated init, it will
> > initialize the whole object, which leads to an inconsistency with this
> > change.
>
> Do you mean for kzalloc() only? or there is some kasan check newly added?

Hi Feng,

I mean that when init is true and kasan_has_integrated_init() is true
(with HW_TAGS mode), kasan_slab_alloc() initializes the whole object.
Which is inconsistent with the memset() of only orig_size when
!kasan_has_integrated_init(). But I think this is fine assuming SLAB
poisoning happens later. But please add a comment.

Thanks!