[PATCH V4 0/8] mm/slab: reduce slab accounting memory overhead by allocating slabobj_ext metadata within unsed slab space

[Harry Yoo <harry.yoo@oracle.com>]

RFC V3: https://lore.kernel.org/linux-mm/20251027122847.320924-1-harry.yoo@oracle.com

I believe I addressed all comments in RFC V3 (except handling lazy
allocation of slabobj_exts, which I would prefer to do as future work).
Please let me know if I missed your comments.

If there is no major drawbacks or concerns coming up, I would like to push
this forward for 7.0 merge window after some review & testing.

Have a wonderful end of the year!

RFC V3 -> V4:
- Rebased onto the latest slab/for-next, dropped RFC
- The metadata alignment (after orig_size) fix is now included as patch 1
  of this series
- Patch 2: Document that use_freeptr_offset can be used for caches with
  constructor (Suren, Vlastimil)
- Patch 6: use get/put_slab_obj_exts() instead of
  metadata_access_enable/disable (Suren)
- Patch 7: Change !mem_cgroup_disabled() check to memcg_kmem_online()
  (Andrey Ryabinin)
- Added Reviewed-by, Suggested-by tags, thanks!

When CONFIG_MEMCG and CONFIG_MEM_ALLOC_PROFILING are enabled,
the kernel allocates two pointers per object: one for the memory cgroup
(obj_cgroup) to which it belongs, and another for the code location
that requested the allocation.

In two special cases, this overhead can be eliminated by allocating
slabobj_ext metadata from unused space within a slab:

  Case 1. The "leftover" space after the last slab object is larger than
          the size of an array of slabobj_ext.

  Case 2. The per-object alignment padding is larger than
          sizeof(struct slabobj_ext).

For these two cases, one or two pointers can be saved per slab object.
Examples: ext4 inode cache (case 1) and xfs inode cache (case 2).
That's approximately 0.7-0.8% (memcg) or 1.5-1.6%% (memcg + mem profiling)
of the total inode cache size.

Implementing case 2 is not straightforward, because the existing code
assumes that slab->obj_exts is an array of slabobj_ext, while case 2
breaks the assumption.

As suggested by Vlastimil, abstract access to individual slabobj_ext
metadata via a new helper named slab_obj_ext():

static inline struct slabobj_ext *slab_obj_ext(struct slab *slab,
                                               unsigned long obj_exts,
                                               unsigned int index)
{
        return (struct slabobj_ext *)(obj_exts + slab_get_stride(slab) * index);
} 

In the normal case (including case 1), slab->obj_exts points to an array
of slabobj_ext, and the stride is sizeof(struct slabobj_ext).

In case 2, the stride is s->size and
slab->obj_exts = slab_address(slab) + s->red_left_pad + (offset of slabobj_ext)

With this approach, the memcg charging fastpath doesn't need to care the
storage method of slabobj_ext.

Harry Yoo (8):
  mm/slab: use unsigned long for orig_size to ensure proper metadata
    align
  mm/slab: allow specifying free pointer offset when using constructor
  ext4: specify the free pointer offset for ext4_inode_cache
  mm/slab: abstract slabobj_ext access via new slab_obj_ext() helper
  mm/slab: use stride to access slabobj_ext
  mm/memcontrol,alloc_tag: handle slabobj_ext access under KASAN poison
  mm/slab: save memory by allocating slabobj_ext array from leftover
  mm/slab: place slabobj_ext metadata in unused space within s->size

 fs/ext4/super.c      |  20 ++-
 include/linux/slab.h |  39 +++--
 mm/memcontrol.c      |  31 +++-
 mm/slab.h            | 120 ++++++++++++++-
 mm/slab_common.c     |   8 +-
 mm/slub.c            | 345 +++++++++++++++++++++++++++++++++++--------
 6 files changed, 466 insertions(+), 97 deletions(-)

-- 
2.43.0