Re: [PATCH RFC 3/4] mm, slab: add static key for should_failslab()

[Alexei Starovoitov <alexei.starovoitov@gmail.com>]

On Fri, May 31, 2024 at 2:33 AM Vlastimil Babka <vbabka@suse.cz> wrote:
>
> Since commit 4f6923fbb352 ("mm: make should_failslab always available for
> fault injection") should_failslab() is unconditionally a noinline
> function. This adds visible overhead to the slab allocation hotpath,
> even if the function is empty. With CONFIG_FAILSLAB=y there's additional
> overhead when the functionality is not enabled by a boot parameter or
> debugfs.
>
> The overhead can be eliminated with a static key around the callsite.
> Fault injection and error injection frameworks can now be told that the
> this function has a static key associated, and are able to enable and
> disable it accordingly.
>
> Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
> ---
>  mm/failslab.c |  2 +-
>  mm/slab.h     |  3 +++
>  mm/slub.c     | 10 +++++++---
>  3 files changed, 11 insertions(+), 4 deletions(-)
>
> diff --git a/mm/failslab.c b/mm/failslab.c
> index ffc420c0e767..878fd08e5dac 100644
> --- a/mm/failslab.c
> +++ b/mm/failslab.c
> @@ -9,7 +9,7 @@ static struct {
>         bool ignore_gfp_reclaim;
>         bool cache_filter;
>  } failslab = {
> -       .attr = FAULT_ATTR_INITIALIZER,
> +       .attr = FAULT_ATTR_INITIALIZER_KEY(&should_failslab_active.key),
>         .ignore_gfp_reclaim = true,
>         .cache_filter = false,
>  };
> diff --git a/mm/slab.h b/mm/slab.h
> index 5f8f47c5bee0..792e19cb37b8 100644
> --- a/mm/slab.h
> +++ b/mm/slab.h
> @@ -11,6 +11,7 @@
>  #include <linux/memcontrol.h>
>  #include <linux/kfence.h>
>  #include <linux/kasan.h>
> +#include <linux/jump_label.h>
>
>  /*
>   * Internal slab definitions
> @@ -160,6 +161,8 @@ static_assert(IS_ALIGNED(offsetof(struct slab, freelist), sizeof(freelist_aba_t)
>   */
>  #define slab_page(s) folio_page(slab_folio(s), 0)
>
> +DECLARE_STATIC_KEY_FALSE(should_failslab_active);
> +
>  /*
>   * If network-based swap is enabled, sl*b must keep track of whether pages
>   * were allocated from pfmemalloc reserves.
> diff --git a/mm/slub.c b/mm/slub.c
> index 0809760cf789..3bb579760a37 100644
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -3874,13 +3874,15 @@ static __always_inline void maybe_wipe_obj_freeptr(struct kmem_cache *s,
>                         0, sizeof(void *));
>  }
>
> +DEFINE_STATIC_KEY_FALSE(should_failslab_active);
> +
>  noinline int should_failslab(struct kmem_cache *s, gfp_t gfpflags)
>  {
>         if (__should_failslab(s, gfpflags))
>                 return -ENOMEM;
>         return 0;
>  }
> -ALLOW_ERROR_INJECTION(should_failslab, ERRNO);
> +ALLOW_ERROR_INJECTION_KEY(should_failslab, ERRNO, &should_failslab_active);
>
>  static __fastpath_inline
>  struct kmem_cache *slab_pre_alloc_hook(struct kmem_cache *s, gfp_t flags)
> @@ -3889,8 +3891,10 @@ struct kmem_cache *slab_pre_alloc_hook(struct kmem_cache *s, gfp_t flags)
>
>         might_alloc(flags);
>
> -       if (unlikely(should_failslab(s, flags)))
> -               return NULL;
> +       if (static_branch_unlikely(&should_failslab_active)) {
> +               if (should_failslab(s, flags))
> +                       return NULL;
> +       }

makes sense.
Acked-by: Alexei Starovoitov <ast@kernel.org>

Do you have any microbenchmark numbers before/after this optimization?