Date:   Mon, 26 Sep 2022 21:11:24 +0200
From:   Andrey Konovalov <andreyknvl@...il.com>
To:     Feng Tang <feng.tang@...el.com>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Vlastimil Babka <vbabka@...e.cz>,
        Christoph Lameter <cl@...ux.com>,
        Pekka Enberg <penberg@...nel.org>,
        David Rientjes <rientjes@...gle.com>,
        Joonsoo Kim <iamjoonsoo.kim@....com>,
        Roman Gushchin <roman.gushchin@...ux.dev>,
        Hyeonggon Yoo <42.hyeyoo@...il.com>,
        Dmitry Vyukov <dvyukov@...gle.com>,
        Jonathan Corbet <corbet@....net>,
        Dave Hansen <dave.hansen@...el.com>,
        Linux Memory Management List <linux-mm@...ck.org>,
        LKML <linux-kernel@...r.kernel.org>,
        kasan-dev <kasan-dev@...glegroups.com>,
        Kees Cook <keescook@...omium.org>
Subject: Re: [PATCH v6 2/4] mm/slub: only zero the requested size of buffer
 for kzalloc

On Tue, Sep 13, 2022 at 8:54 AM Feng Tang <feng.tang@...el.com> wrote:
>

Hi Feng,

> kzalloc/kmalloc will round up the request size to a fixed size
> (mostly power of 2), so the allocated memory could be more than
> requested. Currently kzalloc family APIs will zero all the
> allocated memory.
>
> To detect out-of-bound usage of the extra allocated memory, only
> zero the requested part, so that sanity check could be added to
> the extra space later.

I still don't like the idea of only zeroing the requested memory and
not the whole object. Considering potential info-leak vulnerabilities.

Can we only do this when SLAB_DEBUG is enabled?

> Performance wise, smaller zeroing length also brings shorter
> execution time, as shown from test data on various server/desktop
> platforms.
>
> For kzalloc users who will call ksize() later and utilize this
> extra space, please be aware that the space is not zeroed any
> more.

CC Kees

>
> Signed-off-by: Feng Tang <feng.tang@...el.com>
> ---
>  mm/slab.c |  7 ++++---
>  mm/slab.h |  5 +++--
>  mm/slub.c | 10 +++++++---
>  3 files changed, 14 insertions(+), 8 deletions(-)
>
> diff --git a/mm/slab.c b/mm/slab.c
> index a5486ff8362a..4594de0e3d6b 100644
> --- a/mm/slab.c
> +++ b/mm/slab.c
> @@ -3253,7 +3253,8 @@ slab_alloc_node(struct kmem_cache *cachep, struct list_lru *lru, gfp_t flags,
>         init = slab_want_init_on_alloc(flags, cachep);
>
>  out:
> -       slab_post_alloc_hook(cachep, objcg, flags, 1, &objp, init);
> +       slab_post_alloc_hook(cachep, objcg, flags, 1, &objp, init,
> +                               cachep->object_size);
>         return objp;
>  }
>
> @@ -3506,13 +3507,13 @@ int kmem_cache_alloc_bulk(struct kmem_cache *s, gfp_t flags, size_t size,
>          * Done outside of the IRQ disabled section.
>          */
>         slab_post_alloc_hook(s, objcg, flags, size, p,
> -                               slab_want_init_on_alloc(flags, s));
> +                       slab_want_init_on_alloc(flags, s), s->object_size);
>         /* FIXME: Trace call missing. Christoph would like a bulk variant */
>         return size;
>  error:
>         local_irq_enable();
>         cache_alloc_debugcheck_after_bulk(s, flags, i, p, _RET_IP_);
> -       slab_post_alloc_hook(s, objcg, flags, i, p, false);
> +       slab_post_alloc_hook(s, objcg, flags, i, p, false, s->object_size);
>         kmem_cache_free_bulk(s, i, p);
>         return 0;
>  }
> diff --git a/mm/slab.h b/mm/slab.h
> index d0ef9dd44b71..3cf5adf63f48 100644
> --- a/mm/slab.h
> +++ b/mm/slab.h
> @@ -730,7 +730,8 @@ static inline struct kmem_cache *slab_pre_alloc_hook(struct kmem_cache *s,
>
>  static inline void slab_post_alloc_hook(struct kmem_cache *s,
>                                         struct obj_cgroup *objcg, gfp_t flags,
> -                                       size_t size, void **p, bool init)
> +                                       size_t size, void **p, bool init,
> +                                       unsigned int orig_size)
>  {
>         size_t i;
>
> @@ -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.

>                 kmemleak_alloc_recursive(p[i], s->object_size, 1,
>                                          s->flags, flags);
>         }
> diff --git a/mm/slub.c b/mm/slub.c
> index c8ba16b3a4db..6f823e99d8b4 100644
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -3376,7 +3376,11 @@ static __always_inline void *slab_alloc_node(struct kmem_cache *s, struct list_l
>         init = slab_want_init_on_alloc(gfpflags, s);
>
>  out:
> -       slab_post_alloc_hook(s, objcg, gfpflags, 1, &object, init);
> +       /*
> +        * When init equals 'true', like for kzalloc() family, only
> +        * @orig_size bytes will be zeroed instead of s->object_size
> +        */
> +       slab_post_alloc_hook(s, objcg, gfpflags, 1, &object, init, orig_size);
>
>         return object;
>  }
> @@ -3833,11 +3837,11 @@ int kmem_cache_alloc_bulk(struct kmem_cache *s, gfp_t flags, size_t size,
>          * Done outside of the IRQ disabled fastpath loop.
>          */
>         slab_post_alloc_hook(s, objcg, flags, size, p,
> -                               slab_want_init_on_alloc(flags, s));
> +                       slab_want_init_on_alloc(flags, s), s->object_size);
>         return i;
>  error:
>         slub_put_cpu_ptr(s->cpu_slab);
> -       slab_post_alloc_hook(s, objcg, flags, i, p, false);
> +       slab_post_alloc_hook(s, objcg, flags, i, p, false, s->object_size);
>         kmem_cache_free_bulk(s, i, p);
>         return 0;
>  }
> --
> 2.34.1
>

Powered by blists - more mailing lists