On Tue, Nov 05, 2019 at 03:26:46AM +0530, Bhupesh Sharma wrote: > commit 9b31cf493ffa ("arm64: mm: Introduce MAX_USER_VA_BITS definition") > introduced the MAX_USER_VA_BITS definition, which was used to support > the arm64 mm use-cases where the user-space could use 52-bit virtual > addresses whereas the kernel-space would still could a maximum of 48-bit > virtual addressing. > > But, now with commit b6d00d47e81a ("arm64: mm: Introduce 52-bit Kernel > VAs"), we removed the 52-bit user/48-bit kernel kconfig option and hence > there is no longer any scenario where user VA != kernel VA size > (even with CONFIG_ARM64_FORCE_52BIT enabled, the same is true). > > Hence we can do away with the MAX_USER_VA_BITS macro as it is equal to > VA_BITS (maximum VA space size) in all possible use-cases. Note that > even though the 'vabits_actual' value would be 48 for arm64 hardware > which don't support LVA-8.2 extension (even when CONFIG_ARM64_VA_BITS_52 > is enabled), VA_BITS would still be set to a value 52. Hence this change > would be safe in all possible VA address space combinations. > > Cc: James Morse <james.morse@xxxxxxx> > Cc: Mark Rutland <mark.rutland@xxxxxxx> > Cc: Will Deacon <will@xxxxxxxxxx> > Cc: Steve Capper <steve.capper@xxxxxxx> > Cc: Catalin Marinas <catalin.marinas@xxxxxxx> > Cc: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx> > Cc: linux-kernel@xxxxxxxxxxxxxxx > Cc: kexec@xxxxxxxxxxxxxxxxxxx > Signed-off-by: Bhupesh Sharma <bhsharma@xxxxxxxxxx> Queued for 5.5. Thanks. -- Catalin _______________________________________________ kexec mailing list kexec@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/kexec