[PATCH v2 0/5] Rework the pte handling for hardware AF/DBM

On Fri, Aug 18, 2017 at 06:29:36PM +0100, Will Deacon wrote:

On Fri, Aug 18, 2017 at 06:25:57PM +0100, Catalin Marinas wrote:

quoted

That's the second version of the pte accessors reworking series (first
version at [1]).

Changes since v1:

- READ_ONCE moved out of the cmpxchg loops since cmpxchg already returns
  the old value (probably more efficient with LSE atomics)

- The first patch from the first series has been merged upstream (cc
  stable)

- Kconfig entry kept just to disable the hardware feature but the code
  paths all assume hardware AF/DBM

- PAGE_COPY(_EXEC) removed as they just duplicate PAGE_READONLY(_EXEC)

I kept the review-by tags in place as the logic is unchanged but feel
free to look at the patches again.

Thanks, this looks good to me now:

Reviewed-by: Will Deacon <redacted>

Thanks. I merged them into for-next/core.

-- 
Catalin