Re: [PATCH v3 20/37] mm/mprotect: Exclude shadow stack from preserve_write

[Peter Zijlstra]

On Fri, Nov 04, 2022 at 03:35:47PM -0700, Rick Edgecombe wrote:
> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
> index 73b9b78f8cf4..7643a4db1b50 100644
> --- a/mm/huge_memory.c
> +++ b/mm/huge_memory.c
> @@ -1803,6 +1803,13 @@ int change_huge_pmd(struct mmu_gather *tlb, struct vm_area_struct *vma,
>  		return 0;
>  
>  	preserve_write = prot_numa && pmd_write(*pmd);
> +
> +	/*
> +	 * Preserve only normal writable huge PMD, but not shadow
> +	 * stack (RW=0, Dirty=1).
> +	 */
> +	if (vma->vm_flags & VM_SHADOW_STACK)
> +		preserve_write = false;
>  	ret = 1;
>  
>  #ifdef CONFIG_ARCH_ENABLE_THP_MIGRATION
> diff --git a/mm/mprotect.c b/mm/mprotect.c
> index 668bfaa6ed2a..ea82ce5f38fe 100644
> --- a/mm/mprotect.c
> +++ b/mm/mprotect.c
> @@ -115,6 +115,13 @@ static unsigned long change_pte_range(struct mmu_gather *tlb,
>  			pte_t ptent;
>  			bool preserve_write = prot_numa && pte_write(oldpte);
>  
> +			/*
> +			 * Preserve only normal writable PTE, but not shadow
> +			 * stack (RW=0, Dirty=1).
> +			 */
> +			if (vma->vm_flags & VM_SHADOW_STACK)
> +				preserve_write = false;
> +
>  			/*
>  			 * Avoid trapping faults against the zero or KSM
>  			 * pages. See similar comment in change_huge_pmd.

These comments lack a why component; someone is going to wonder wtf this
code is doing in the near future -- that someone might be you.