Re: [PATCH RFC v2] Randomized slab caches for kmalloc()

[Hyeonggon Yoo <42.hyeyoo@gmail.com>]

[Resending this email after noticing I did not reply-to-all]

On Fri, May 12, 2023 at 7:11 PM Gong Ruiqi <gongruiqi1@huawei.com> wrote:
>
>
>
> On 2023/05/11 2:43, Hyeonggon Yoo wrote:
> > I dont think adding a hardening feature by sacrificing one digit
> > percent performance
> > (and additional complexity) is worth. Heap spraying can only occur
> > when the kernel contains
> > security vulnerabilities, and if there is no known ways of performing
> > such an attack,
> > then we would simply be paying a consistent cost.
> >
> > Any opinions from hardening folks?
>
> I did a more throughout performance test on the same machine in the same
> way, and here are the results:
>
>               sched/  sched/  syscall/       mem/         mem/
>            messaging    pipe     basic     memcpy       memset
> control1       0.019   5.459     0.733  15.258789    51.398026
> control2       0.019   5.439     0.730  16.009221    48.828125
> control3       0.019   5.282     0.735  16.009221    48.828125
> control_avg    0.019   5.393     0.733  15.759077    49.684759
>
> exp1           0.019   5.374     0.741  15.500992    46.502976
> exp2           0.019   5.440     0.746  16.276042    51.398026
> exp3           0.019   5.242     0.752  15.258789    51.398026
> exp_avg        0.019   5.352     0.746  15.678608    49.766343
>
> I believe the results show only minor differences and normal
> fluctuation, and no substantial performance degradation.
>
> As Pedro points out in his reply, unfortunately there are always
> security vulnerabilities in the kernel, which is a fact that we have to
> admit. Having a useful mitigation mechanism at the expense of a little
> performance loss would be, in my opinion, quite a good deal in many
> circumstances. And people can still choose not to have it by setting the
> config to n.

Okay, now I don't think I need to tackle it from a performance
perspective anymore, at least it looks like a good tradeoff.

I had few design level concerns (i.e. in ARM64 instructions are 4-byte
aligned) before switching to hash_64(^ random sequence), but looks
good to me now.

> >> +#ifdef CONFIG_RANDOM_KMALLOC_CACHES
> >> +# define SLAB_RANDOMSLAB       ((slab_flags_t __force)0x01000000U)
> >> +#else
> >> +# define SLAB_RANDOMSLAB       0
> >> +#endif

There is already the SLAB_KMALLOC flag that indicates if a cache is a
kmalloc cache. I think that would be enough for preventing merging
kmalloc caches?