The following patch series improves x86 arch_atomic*()
family of functions and merges x86_32 and x86_64
arch_atomic64_fetch_{and,or,xor}() functions.

The patch series enables impressive assembly code
reductions for x86_32 target and lowers future maintenace
burden and technical debt of the source code by unifying
several functions between x86_32 and x86_32 targets.

Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Peter Zijlstra <peterz@infradead.org>

Uros Bizjak (6):
  locking/atomic/x86: Introduce arch_atomic64_try_cmpxchg to x86_32
  locking/atomic/x86: Rewrite x86_32
    arch_atomic64_{,fetch}_{and,or,xor}() functions
  locking/atomic/x86: Use READ_ONCE before atomic{,64}_try_cmpxchg loops
  locking/atomic/x86: Merge x86_32 and x86_64
    arch_atomic64_fetch_{and,or,xor}() functions
  locking/atomic/x86: Define arch_atomic_sub() family using
    arch_atomic_add() functions
  locking/atomic/x86: Reorder a couple of arch_atomic64 functions

 arch/x86/include/asm/atomic.h      |  50 ++++++++++----
 arch/x86/include/asm/atomic64_32.h | 103 ++++++++++-------------------
 arch/x86/include/asm/atomic64_64.h |  44 +-----------
 3 files changed, 75 insertions(+), 122 deletions(-)

-- 
2.44.0