[PATCH v6 0/8] futex: Use runtime constants for futex_hash computation

[K Prateek Nayak <kprateek.nayak@amd.com>]

tl;dr

This series introduces runtime_const_mask_32() and uses runtime
constants for __ro_after_init data in futex_hash() hot path. More
information can be found on v2 at
https://lore.kernel.org/lkml/20260316052401.18910-1-kprateek.nayak@amd.com/

Notable changes in v5
=====================

Intel test robot flagged a case of compilers re-ordering the
runtime_const usage before their initialization in futex_init() at
https://lore.kernel.org/lkml/202607091538.58e3f39-lkp@intel.com/.
This was fixed by adding a barrier() after the runtime_const_init() bits
in futex_init() to prevent this re-ordering.

Sashiko flagged that __fls(val), when val is 0, can be optimized by the
compiler to consider val as non-zero and silently optimize the
BUG_ON(!val && ...) to only consider the latter bits, skipping the !val
check. Fix this by explicitly checking val before __fls().

Testing
=======

Apart from x86, which was build and boot tested on baremetal, all the
other architectures have been build and boot tested with cross-compile +
QEMU with some light sanity testing on each.

Patches are based on:

  git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git master

at commit 8d0b0ef04514 ("Merge branch into tip/master: 'x86/tdx'")
(27-07-2026)

Diffstat from v5 is essentially 5+++++-- with effectively 3 lines of
code change and 2 newlines; Resending for convenience since adding
parital diffs to v5 was getting too finnicky and this is easier for
pickup.

Charlie's optimization for RISC-V bits to patch slli + srli with
andi + nop if mask fits the immediate encoding is left to a later series
once this is merged.
(https://lore.kernel.org/lkml/178366995930.1208691.2993932866462893112.b4-review@b4/)

Few comments from checkpatch.pl have been ignored to adhere to the style
of the particular file. If something needs addressing, please let me
know and I'll address it with a v6.X fixups unless there is a larger
change that will require a re-spin

Everyone has been Cc'd on the cover-letter and the futex bits for the
context. Respective arch maintainers, reviewers, and whoever got lucky
with get_maintainer.pl have been Cc'd on their respective arch specific
changes. Futex maintainers and the lists will be receiving the whole
series (sorry in advance!)
---
changelog v5..v6:

o Added a barrier after runtime_const_init() to prevent compilers from
  re-ordering their access before the initialization in futex_init().
  (Sashiko, Intel test robot)

o Added a zero check before __fls() to prevent smart compiler
  optimizations of a later BUG_ON(). (Sashiko)

o Picked up Peter's S-o-b from his tree.

o Collected tags from Charlie. (Thanks a ton!)

o Rebased on latest tip.

v5: https://lore.kernel.org/lkml/20260630045531.3939-1-kprateek.nayak@amd.com/
---
K Prateek Nayak (5):
  arm64/runtime-const: Use aarch64_insn_patch_text_nosync() for patching
  arm64/runtime-const: Introduce runtime_const_mask_32()
  riscv/runtime-const: Replace open-coded placeholder with RUNTIME_MAGIC
  riscv/runtime-const: Introduce runtime_const_mask_32()
  s390/runtime-const: Introduce runtime_const_mask_32()

Peter Zijlstra (3):
  x86/runtime-const: Introduce runtime_const_mask_32()
  asm-generic/runtime-const: Add dummy runtime_const_mask_32()
  futex: Use runtime constants for __futex_hash() hot path

 arch/arm64/include/asm/runtime-const.h | 63 ++++++++++++++++----
 arch/riscv/include/asm/asm.h           |  1 +
 arch/riscv/include/asm/runtime-const.h | 82 ++++++++++++++++++++------
 arch/s390/include/asm/runtime-const.h  | 22 ++++++-
 arch/x86/include/asm/runtime-const.h   | 14 +++++
 include/asm-generic/runtime-const.h    |  1 +
 include/asm-generic/vmlinux.lds.h      |  5 +-
 kernel/futex/core.c                    | 44 ++++++++------
 8 files changed, 181 insertions(+), 51 deletions(-)


base-commit: 8d0b0ef045143e7ebf2b4c3ad5347bc1e0e45a80
-- 
2.34.1