Re: [PATCH] mm/kfence: select random number before taking raw lock

[Marco Elver <elver@google.com>]

On Thu, 9 Jun 2022 at 14:17, Jason A. Donenfeld <Jason@zx2c4.com> wrote:
>
> The RNG uses vanilla spinlocks, not raw spinlocks, so kfence should pick
> its random numbers before taking its raw spinlocks. This also has the
> nice effect of doing less work inside the lock. It should fix a splat
> that Geert saw with CONFIG_PROVE_RAW_LOCK_NESTING:
>
>      dump_backtrace.part.0+0x98/0xc0
>      show_stack+0x14/0x28
>      dump_stack_lvl+0xac/0xec
>      dump_stack+0x14/0x2c
>      __lock_acquire+0x388/0x10a0
>      lock_acquire+0x190/0x2c0
>      _raw_spin_lock_irqsave+0x6c/0x94
>      crng_make_state+0x148/0x1e4
>      _get_random_bytes.part.0+0x4c/0xe8
>      get_random_u32+0x4c/0x140
>      __kfence_alloc+0x460/0x5c4
>      kmem_cache_alloc_trace+0x194/0x1dc
>      __kthread_create_on_node+0x5c/0x1a8
>      kthread_create_on_node+0x58/0x7c
>      printk_start_kthread.part.0+0x34/0xa8
>      printk_activate_kthreads+0x4c/0x54
>      do_one_initcall+0xec/0x278
>      kernel_init_freeable+0x11c/0x214
>      kernel_init+0x24/0x124
>      ret_from_fork+0x10/0x20
>
> Cc: John Ogness <john.ogness@linutronix.de>
> Cc: Alexander Potapenko <glider@google.com>
> Cc: Marco Elver <elver@google.com>
> Cc: Dmitry Vyukov <dvyukov@google.com>
> Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
> Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
> ---
>  mm/kfence/core.c | 7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/mm/kfence/core.c b/mm/kfence/core.c
> index 4e7cd4c8e687..6322b7729b50 100644
> --- a/mm/kfence/core.c
> +++ b/mm/kfence/core.c
> @@ -360,6 +360,9 @@ static void *kfence_guarded_alloc(struct kmem_cache *cache, size_t size, gfp_t g
>         unsigned long flags;
>         struct slab *slab;
>         void *addr;
> +       bool random_right_allocate = prandom_u32_max(2);
> +       bool random_fault = CONFIG_KFENCE_STRESS_TEST_FAULTS &&
> +                           !prandom_u32_max(CONFIG_KFENCE_STRESS_TEST_FAULTS);
>
>         /* Try to obtain a free object. */
>         raw_spin_lock_irqsave(&kfence_freelist_lock, flags);
> @@ -404,7 +407,7 @@ static void *kfence_guarded_alloc(struct kmem_cache *cache, size_t size, gfp_t g
>          * is that the out-of-bounds accesses detected are deterministic for
>          * such allocations.
>          */
> -       if (prandom_u32_max(2)) {
> +       if (random_right_allocate) {
>                 /* Allocate on the "right" side, re-calculate address. */
>                 meta->addr += PAGE_SIZE - size;
>                 meta->addr = ALIGN_DOWN(meta->addr, cache->align);
> @@ -444,7 +447,7 @@ static void *kfence_guarded_alloc(struct kmem_cache *cache, size_t size, gfp_t g
>         if (cache->ctor)
>                 cache->ctor(addr);
>
> -       if (CONFIG_KFENCE_STRESS_TEST_FAULTS && !prandom_u32_max(CONFIG_KFENCE_STRESS_TEST_FAULTS))
> +       if (random_fault)

The compiler should elide this branch entirely if
CONFIG_KFENCE_STRESS_TEST_FAULTS=0, but not sure it'll always do so
now. My suggestion is to make both new bools consts, to help out the
compiler a little.

Otherwise looks good, thanks for the quick fix!