[0/2] slab bulk alloc/free cleanups

Message

Hyeonggon Yoo June 14, 2022, 3:26 p.m. UTC

This makes SLUB and SLAB use their own kmem_cache_free_bulk()
(instead of __kmem_cache_free_bulk()) when kmem_cache_alloc_bulk() failed.
And then it moves generic __kmem_cache_{alloc,free}_bulk to SLOB.

This is suggested by Matthew WilCox.

Hyeonggon Yoo (2):
  mm/sl[au]b: use own bulk free function when bulk alloc failed
  mm/slab_common: move generic bulk alloc/free functions to SLOB

 mm/slab.c        |  2 +-
 mm/slab.h        |  9 ---------
 mm/slab_common.c | 27 ---------------------------
 mm/slob.c        | 25 +++++++++++++++++++++----
 mm/slub.c        |  4 ++--
 5 files changed, 24 insertions(+), 43 deletions(-)

Comments

Vlastimil Babka (SUSE) July 20, 2022, 2:16 p.m. UTC | #1

On 6/14/22 17:26, Hyeonggon Yoo wrote:
> This makes SLUB and SLAB use their own kmem_cache_free_bulk()
> (instead of __kmem_cache_free_bulk()) when kmem_cache_alloc_bulk() failed.
> And then it moves generic __kmem_cache_{alloc,free}_bulk to SLOB.
> 
> This is suggested by Matthew WilCox.
> 
> Hyeonggon Yoo (2):
>   mm/sl[au]b: use own bulk free function when bulk alloc failed
>   mm/slab_common: move generic bulk alloc/free functions to SLOB

Thanks, pushed to slab/for-5.20/bulk_cleanup

>  mm/slab.c        |  2 +-
>  mm/slab.h        |  9 ---------
>  mm/slab_common.c | 27 ---------------------------
>  mm/slob.c        | 25 +++++++++++++++++++++----
>  mm/slub.c        |  4 ++--
>  5 files changed, 24 insertions(+), 43 deletions(-)
>