Re: [RFC PATCH v3 12/24] x86/mm: Modify ptep_set_wrprotect and pmdp_set_wrprotect for _PAGE_DIRTY_SW

[Yu-cheng Yu]

On Thu, 2018-08-30 at 12:59 -0700, Randy Dunlap wrote:
> On 08/30/2018 07:38 AM, Yu-cheng Yu wrote:
> > 
> > When Shadow Stack is enabled, the read-only and PAGE_DIRTY_HW PTE
> > setting is reserved only for the Shadow Stack.ÂÂTo track dirty of
> > non-Shadow Stack read-only PTEs, we use PAGE_DIRTY_SW.
> > 
> > Update ptep_set_wrprotect() and pmdp_set_wrprotect().
> > 
> > Signed-off-by: Yu-cheng Yu <yu-cheng.yu@xxxxxxxxx>
> > ---
> > Âarch/x86/include/asm/pgtable.h | 42
> > ++++++++++++++++++++++++++++++++++
> > Â1 file changed, 42 insertions(+)
> > 
> > diff --git a/arch/x86/include/asm/pgtable.h
> > b/arch/x86/include/asm/pgtable.h
> > index 4d50de77ea96..556ef258eeff 100644
> > --- a/arch/x86/include/asm/pgtable.h
> > +++ b/arch/x86/include/asm/pgtable.h
> > @@ -1203,7 +1203,28 @@ static inline pte_t
> > ptep_get_and_clear_full(struct mm_struct *mm,
> > Âstatic inline void ptep_set_wrprotect(struct mm_struct *mm,
> > Â				ÂÂÂÂÂÂunsigned long addr, pte_t
> > *ptep)
> > Â{
> > +	pte_t pte;
> > +
> > Â	clear_bit(_PAGE_BIT_RW, (unsigned long *)&ptep->pte);
> > +	pte = *ptep;
> > +
> > +	/*
> > +	Â* Some processors can start a write, but ending up
> > seeing
> 	ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂbut end up seeing
> 
> > 
> > +	Â* a read-only PTE by the time they get to the Dirty bit.
> > +	Â* In this case, they will set the Dirty bit, leaving a
> > +	Â* read-only, Dirty PTE which looks like a Shadow Stack
> > PTE.
> > +	Â*
> > +	Â* However, this behavior has been improved and will not
> > occur
> > +	Â* on processors supporting Shadow Stacks.ÂÂWithout this
> > +	Â* guarantee, a transition to a non-present PTE and flush
> > the
> > +	Â* TLB would be needed.
> > +	Â*
> > +	Â* When change a writable PTE to read-only and if the PTE
> > has
> 	ÂÂÂÂÂÂÂÂchanging
> 
> > 
> > +	Â* _PAGE_DIRTY_HW set, we move that bit to _PAGE_DIRTY_SW
> > so
> > +	Â* that the PTE is not a valid Shadow Stack PTE.
> > +	Â*/
> > +	pte = pte_move_flags(pte, _PAGE_DIRTY_HW,
> > _PAGE_DIRTY_SW);
> > +	set_pte_at(mm, addr, ptep, pte);
> > Â}
> > Â
> > Â#define flush_tlb_fix_spurious_fault(vma, address) do { } while
> > (0)
> > @@ -1266,7 +1287,28 @@ static inline pud_t
> > pudp_huge_get_and_clear(struct mm_struct *mm,
> > Âstatic inline void pmdp_set_wrprotect(struct mm_struct *mm,
> > Â				ÂÂÂÂÂÂunsigned long addr, pmd_t
> > *pmdp)
> > Â{
> > +	pmd_t pmd;
> > +
> > Â	clear_bit(_PAGE_BIT_RW, (unsigned long *)pmdp);
> > +	pmd = *pmdp;
> > +
> > +	/*
> > +	Â* Some processors can start a write, but ending up
> > seeing
> 	ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂbut end up seeing
> 
> > 
> > +	Â* a read-only PTE by the time they get to the Dirty bit.
> > +	Â* In this case, they will set the Dirty bit, leaving a
> > +	Â* read-only, Dirty PTE which looks like a Shadow Stack
> > PTE.
> > +	Â*
> > +	Â* However, this behavior has been improved and will not
> > occur
> > +	Â* on processors supporting Shadow Stacks.ÂÂWithout this
> > +	Â* guarantee, a transition to a non-present PTE and flush
> > the
> > +	Â* TLB would be needed.
> > +	Â*
> > +	Â* When change a writable PTE to read-only and if the PTE
> > has
> 	ÂÂÂÂÂÂÂÂchanging
> 
> > 
> > +	Â* _PAGE_DIRTY_HW set, we move that bit to _PAGE_DIRTY_SW
> > so
> > +	Â* that the PTE is not a valid Shadow Stack PTE.
> > +	Â*/
> > +	pmd = pmd_move_flags(pmd, _PAGE_DIRTY_HW,
> > _PAGE_DIRTY_SW);
> > +	set_pmd_at(mm, addr, pmdp, pmd);
> > Â}
> > Â
> > Â#define pud_write pud_write
> > 
> 

Thanks, I will fix it!