[PATCH V2 0/2] fix lockdep warnings with kmalloc_nolock()

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

Hi, I've observed two lockdep warnings while testing
kmalloc_nolock() in NMI:

  1. Accessing current->mems_allowed_seq seqlock in NMI isn't safe
     and lockdep complains.

  2. w/ CONFIG_SLAB_FREELIST_RANDOM, get_random_u32() acquires
     a local_lock, which isn't safe in NMI and could cause a deadlock.

Let's fix them.

Note: This is based on the latest slab/for-next. It wasn't clear to me
if I should base it on slab/for-next or slab/for-next-fixes,
because the merge window has started, this series needs some exposure
in -next, and the patches in slab/for-next might land in mainline
in the meantime.

But the conflict resolution should be trivial even if it should have
been based on slab/for-next-fixes.

v1 -> v2:
  - Patch 1: per Vlastimil's suggestion, do not access
    current->mems_allowed_seq and avoid retry if !allow_spin,
    rather than returning NULL.

v1: https://lore.kernel.org/linux-mm/20260206171348.35886-1-harry.yoo@oracle.com

Harry Yoo (2):
  mm/slab: do not access current->mems_allowed_seq if !allow_spin
  mm/slab: use prandom if !allow_spin

 mm/slub.c | 41 +++++++++++++++++++++++++++++++++++------
 1 file changed, 35 insertions(+), 6 deletions(-)


base-commit: f6ed7e47c1fc78e78c9bfeb668b1ad9ba5c58120

-- 
2.43.0