[0/2] arm64: memory: VA_START fixups

Message

Mark Rutland Aug. 14, 2019, 1:28 p.m. UTC

Hi all,

These patches address my concerns with the new VA_START semantic as I
spotted while reviewing Will's 52-bit VA cleanup. The first patch
corrects the newly broken usage of VA_START, and the second renames
VA_START to PAGE_END to make the new semantic clearer.

Both patches are based on the arm64 for-next/52-bit-kva branch, and I've
given a 52-bit VA configuration a build+boot test (on HW without 52-bit
VA support).

Thanks,
Mark.

Mark Rutland (2):
  arm64: memory: fix flipped VA space fallout
  arm64: memory: rename VA_START to PAGE_END

 arch/arm64/include/asm/memory.h  | 20 ++++++++++----------
 arch/arm64/include/asm/pgtable.h |  4 ++--
 arch/arm64/kernel/hibernate.c    |  2 +-
 arch/arm64/mm/dump.c             |  8 ++++----
 arch/arm64/mm/fault.c            |  2 +-
 arch/arm64/mm/kasan_init.c       |  2 +-
 arch/arm64/mm/mmu.c              |  4 ++--
 7 files changed, 21 insertions(+), 21 deletions(-)

Comments

Steve Capper Aug. 14, 2019, 2:51 p.m. UTC | #1

On Wed, Aug 14, 2019 at 02:28:46PM +0100, Mark Rutland wrote:
> Hi all,

Hi Mark,

> 
> These patches address my concerns with the new VA_START semantic as I
> spotted while reviewing Will's 52-bit VA cleanup. The first patch
> corrects the newly broken usage of VA_START, and the second renames
> VA_START to PAGE_END to make the new semantic clearer.
> 
> Both patches are based on the arm64 for-next/52-bit-kva branch, and I've
> given a 52-bit VA configuration a build+boot test (on HW without 52-bit
> VA support).
> 

A big thank you for this!
I have applied this series and tested it with CONFIG_DEBUG_VIRTUAL,
CONFIG_DEBUG_VM and KASAN SW TAGS.

I've tested this with 52-bit and have given the kernel page table dumper
a go too.

FWIW:
Tested-by: Steve Capper <[e-mail]>

Cheers,