Re: [PATCH] mm/slab_common.c: fix possible spectre-v1 in kmalloc_slab()

From: Dianzhang Chen <dianzhangchen0@gmail.com>
To: Michal Hocko <mhocko@kernel.org>
Cc: cl@linux.com, penberg@kernel.org, rientjes@google.com,
	iamjoonsoo.kim@lge.com, akpm@linux-foundation.org,
	linux-mm@kvack.org, LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] mm/slab_common.c: fix possible spectre-v1 in kmalloc_slab()
Date: Thu, 30 May 2019 00:39:53 +0800	[thread overview]
Message-ID: <CAFbcbMDJB0uNjTa9xwT9npmTdqMJ1Hez3CyeOCjjrLF2W0Wprw@mail.gmail.com> (raw)
In-Reply-To: <20190529162532.GG18589@dhcp22.suse.cz>

It's come from `192+1`.

The more code fragment is:

if (size <= 192) {

    if (!size)

        return ZERO_SIZE_PTR;

    size = array_index_nospec(size, 193);

    index = size_index[size_index_elem(size)];

}

Sine array_index_nospec(index, size) can clamp the index within the
range of [0, size), so in order to make the `size<=192`, need to clamp
the index in the range of [0, 192+1) .

On Thu, May 30, 2019 at 12:25 AM Michal Hocko <mhocko@kernel.org> wrote:
>
> On Wed 29-05-19 20:37:28, Dianzhang Chen wrote:
> [...]
> > @@ -1056,6 +1057,7 @@ struct kmem_cache *kmalloc_slab(size_t size, gfp_t flags)
> >               if (!size)
> >                       return ZERO_SIZE_PTR;
> >
> > +             size = array_index_nospec(size, 193);
> >               index = size_index[size_index_elem(size)];
>
> What is this 193 magic number?
> --
> Michal Hocko
> SUSE Labs