Re: [PATCH v23 12/28] x86/mm: Update ptep_set_wrprotect() and pmdp_set_wrprotect() for transition from _PAGE_DIRTY to _PAGE_COW

[Kirill A. Shutemov]

On Tue, Mar 16, 2021 at 08:10:38AM -0700, Yu-cheng Yu wrote:
> When Shadow Stack is introduced, [R/O + _PAGE_DIRTY] PTE is reserved for
> shadow stack.  Copy-on-write PTEs have [R/O + _PAGE_COW].
> 
> When a PTE goes from [R/W + _PAGE_DIRTY] to [R/O + _PAGE_COW], it could
> become a transient shadow stack PTE in two cases:
> 
> The first case is that some processors can start a write but end up seeing
> a read-only PTE by the time they get to the Dirty bit, creating a transient
> shadow stack PTE.  However, this will not occur on processors supporting
> Shadow Stack, and a TLB flush is not necessary.
> 
> The second case is that when _PAGE_DIRTY is replaced with _PAGE_COW non-
> atomically, a transient shadow stack PTE can be created as a result.
> Thus, prevent that with cmpxchg.
> 
> Dave Hansen, Jann Horn, Andy Lutomirski, and Peter Zijlstra provided many
> insights to the issue.  Jann Horn provided the cmpxchg solution.
> 
> Signed-off-by: Yu-cheng Yu <yu-cheng.yu@xxxxxxxxx>
> Reviewed-by: Kees Cook <keescook@xxxxxxxxxxxx>
> ---
>  arch/x86/include/asm/pgtable.h | 36 ++++++++++++++++++++++++++++++++++
>  1 file changed, 36 insertions(+)
> 
> diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
> index e1739f590ca6..46d9394b884f 100644
> --- a/arch/x86/include/asm/pgtable.h
> +++ b/arch/x86/include/asm/pgtable.h
> @@ -1306,6 +1306,24 @@ static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm,
>  static inline void ptep_set_wrprotect(struct mm_struct *mm,
>  				      unsigned long addr, pte_t *ptep)
>  {
> +	/*
> +	 * If Shadow Stack is enabled, pte_wrprotect() moves _PAGE_DIRTY
> +	 * to _PAGE_COW (see comments at pte_wrprotect()).
> +	 * When a thread reads a RW=1, Dirty=0 PTE and before changing it
> +	 * to RW=0, Dirty=0, another thread could have written to the page
> +	 * and the PTE is RW=1, Dirty=1 now.  Use try_cmpxchg() to detect
> +	 * PTE changes and update old_pte, then try again.
> +	 */
> +	if (cpu_feature_enabled(X86_FEATURE_SHSTK)) {
> +		pte_t old_pte, new_pte;
> +
> +		old_pte = READ_ONCE(*ptep);
> +		do {
> +			new_pte = pte_wrprotect(old_pte);
> +		} while (!try_cmpxchg(&ptep->pte, &old_pte.pte, new_pte.pte));

I think this is wrong. You need to update old_pte on every loop iteration,
otherwise you can get in to endless loop.

The same issue for pmdp_set_wrprotect().

> +
> +		return;
> +	}
>  	clear_bit(_PAGE_BIT_RW, (unsigned long *)&ptep->pte);
>  }
>  
> @@ -1350,6 +1368,24 @@ static inline pud_t pudp_huge_get_and_clear(struct mm_struct *mm,
>  static inline void pmdp_set_wrprotect(struct mm_struct *mm,
>  				      unsigned long addr, pmd_t *pmdp)
>  {
> +	/*
> +	 * If Shadow Stack is enabled, pmd_wrprotect() moves _PAGE_DIRTY
> +	 * to _PAGE_COW (see comments at pmd_wrprotect()).
> +	 * When a thread reads a RW=1, Dirty=0 PMD and before changing it
> +	 * to RW=0, Dirty=0, another thread could have written to the page
> +	 * and the PMD is RW=1, Dirty=1 now.  Use try_cmpxchg() to detect
> +	 * PMD changes and update old_pmd, then try again.
> +	 */
> +	if (cpu_feature_enabled(X86_FEATURE_SHSTK)) {
> +		pmd_t old_pmd, new_pmd;
> +
> +		old_pmd = READ_ONCE(*pmdp);
> +		do {
> +			new_pmd = pmd_wrprotect(old_pmd);
> +		} while (!try_cmpxchg((pmdval_t *)pmdp, (pmdval_t *)&old_pmd, pmd_val(new_pmd)));
> +
> +		return;
> +	}
>  	clear_bit(_PAGE_BIT_RW, (unsigned long *)pmdp);
>  }
>  
> -- 
> 2.21.0
> 

-- 
 Kirill A. Shutemov