Re: [PATCH v2 6/6] ARM: add CONFIG_VMSPLIT_2G_OPT option

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Low] ARM: Kconfig: missing PAGE_OFFSET=0x70000000 in
  ARCH_MMAP_RND_BITS_MAX
--

commit 174ed2566532b79fdcb5633054674e730fcadf44
Author: Arnd Bergmann <arnd@arndb.de>

ARM: add CONFIG_VMSPLIT_2G_OPT option

This commit adds CONFIG_VMSPLIT_2G_OPT with PAGE_OFFSET=0x70000000 to
provide a full 2 GiB lowmem region on non-LPAE builds, and makes it the
default when CONFIG_HIGHMEM is enabled.

> diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
> index fee6b7435b7ca..17b4da137a8ea 100644
> --- a/arch/arm/Kconfig
> +++ b/arch/arm/Kconfig
[ ... ]
> @@ -1052,6 +1056,7 @@ config PAGE_OFFSET
>  	hex
>  	default PHYS_OFFSET if !MMU
>  	default 0x40000000 if VMSPLIT_1G
> +	default 0x70000000 if VMSPLIT_2G_OPT
>  	default 0x80000000 if VMSPLIT_2G
>  	default 0xB0000000 if VMSPLIT_3G_OPT
>  	default 0xC0000000

[Severity: Low]
Should ARCH_MMAP_RND_BITS_MAX also handle PAGE_OFFSET=0x70000000?

arch/arm/Kconfig:
config ARCH_MMAP_RND_BITS_MAX
	default 14 if PAGE_OFFSET=0x40000000
	default 15 if PAGE_OFFSET=0x80000000
	default 16

With VMSPLIT_2G_OPT, PAGE_OFFSET=0x70000000 falls through to 16 bits
(up to 256 MiB of mmap randomization), which is twice the maximum
allowed for the larger user address space under PAGE_OFFSET=0x80000000.
When KASAN is enabled (shrinking TASK_SIZE to 0x5ce00000) and the stack
gap reaches MAX_GAP, a 16-bit random offset can exceed STACK_TOP -
MAX_GAP in mmap_base() and cause top-down mmap allocations to fall back
to bottom-up.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261112112528.1116405-1-arnd@kernel.org?part=6