[0/3] supplyment of slab allocator removal

Message

Xiongwei Song Dec. 1, 2023, 3:15 a.m. UTC

From: Xiongwei Song <[e-mail]>

Hi,

Patch 1 is to remove an unused parameter. This patch acutually is v3, but
it is not reasonable to add v3 tag in the cover letter, so I put the
change history inside the patch.

---
Patch 2 is to replace slub_$params with slab_$params.
Vlastimil Babka pointed out we should use "slab_$param" as the primary
prefix for long-term plan. Please see [1] for more infomation.

This patch is to do that. However, the patch is big, I'm not sure if
everything is proper in it, so I added "RFC" in the patch title. For more
information please see the commit message of patch.

I did the basic tests with qemu, which passed values by sl[au]b_max_order,
sl[au]b_min_order, sl[au]b_min_objects and sl[au]b_debug in command line.
The values looks correct by printing them out before calculating orders.

One thing I'm not sure about the forth parameter of __setup_param(),
Is it correct to set the parameter to 0 directly?

---
Patch 3 is not related to slab allocator removal. It's to correct the
description of default value of slub_min_objects in
Documentation/mm/slub.rst. 

---
Regards,
Xiongwei

[1] https://lore.kernel.org/linux-mm/[e-mail]/

Xiongwei Song (3):
  Documentation: kernel-parameters: remove noaliencache
  mm/slub: unify all slab/slub parameters with "slab_$param"
  mm/slub: correct the default slub_min_objects value in doc

 .../admin-guide/kernel-parameters.txt         |  48 +++---
 Documentation/mm/slub.rst                     |   2 +-
 drivers/misc/lkdtm/heap.c                     |   2 +-
 mm/Kconfig.debug                              |   6 +-
 mm/slab.h                                     |  16 +-
 mm/slab_common.c                              |   8 +-
 mm/slub.c                                     | 142 +++++++++---------
 7 files changed, 110 insertions(+), 114 deletions(-)