Like I did in an earlier series for 6.18
(https://lore.kernel.org/linux-hardening/20260709043301.142931-1-ebiggers@kernel.org/),
this series backports kmalloc_obj() et al to 6.12 so that more upstream
commits will cherry-pick cleanly.  As in 6.18, the flex counter
integration isn't included.

Patches 6-7 then backport two fscrypt fixes (which, like lots of other
kernel code, happen to use kzalloc_obj()).  Patch 7 is a clean
cherry-pick, while patch 6 required resolving some other conflicts.

Eric Biggers (2):
  fscrypt: Fix key setup in edge case with multiple data unit sizes
  fscrypt: Replace mk_users keyring with simple list

Kees Cook (2):
  slab: Introduce kmalloc_obj() and family
  slab: Introduce kmalloc_flex() and family

Linus Torvalds (2):
  add default_gfp() helper macro and use it in the new *alloc_obj()
    helpers
  default_gfp(): avoid using the "newfangled" __VA_OPT__ trick

Randy Dunlap (1):
  slab: recognize @GFP parameter as optional in kernel-doc

 Documentation/process/deprecated.rst |  31 ++++
 fs/crypto/fscrypt_private.h          |  87 ++++++----
 fs/crypto/inline_crypt.c             |   8 +-
 fs/crypto/keyring.c                  | 239 ++++++++++++---------------
 fs/crypto/keysetup.c                 | 119 ++++++++-----
 include/linux/gfp.h                  |   4 +
 include/linux/slab.h                 | 100 +++++++++++
 7 files changed, 373 insertions(+), 215 deletions(-)


base-commit: 296aabce459470a4c1b68ffd0c0c0920e563aaad
-- 
2.55.0

On Thu, Jul 16, 2026 at 09:42:56PM -0700, Eric Biggers wrote:
> Like I did in an earlier series for 6.18
> (https://lore.kernel.org/linux-hardening/20260709043301.142931-1-ebiggers@kernel.org/),
> this series backports kmalloc_obj() et al to 6.12 so that more upstream
> commits will cherry-pick cleanly.  As in 6.18, the flex counter
> integration isn't included.
> 
> Patches 6-7 then backport two fscrypt fixes (which, like lots of other
> kernel code, happen to use kzalloc_obj()).  Patch 7 is a clean
> cherry-pick, while patch 6 required resolving some other conflicts.
> 
> Eric Biggers (2):
>   fscrypt: Fix key setup in edge case with multiple data unit sizes
>   fscrypt: Replace mk_users keyring with simple list
> 
> Kees Cook (2):
>   slab: Introduce kmalloc_obj() and family
>   slab: Introduce kmalloc_flex() and family
> 
> Linus Torvalds (2):
>   add default_gfp() helper macro and use it in the new *alloc_obj()
>     helpers
>   default_gfp(): avoid using the "newfangled" __VA_OPT__ trick
> 
> Randy Dunlap (1):
>   slab: recognize @GFP parameter as optional in kernel-doc

All now queued up, thanks!

greg k-h