Re: [PATCH -next v2 2/4] mm: page_table_check: add hooks to public helpers

[Tong Tiangen <tongtiangen@xxxxxxxxxx>]

在 2022/3/24 10:12, Pasha Tatashin 写道:
> On Wed, Mar 23, 2022 at 10:07 PM Tong Tiangen <tongtiangen@xxxxxxxxxx> wrote:
> > 在 2022/3/24 1:42, Pasha Tatashin 写道:
> > > On Tue, Mar 22, 2022 at 10:25 AM Tong Tiangen <tongtiangen@xxxxxxxxxx> wrote:
> > > > Move ptep_clear() to the include/linux/pgtable.h and add page table check
> > > > relate hooks to some helpers, it's prepare for support page table check
> > > > feature on new architecture.
> > > >
> > > > Signed-off-by: Tong Tiangen <tongtiangen@xxxxxxxxxx>
> > > > ---
> > > >    arch/x86/include/asm/pgtable.h | 10 ----------
> > > >    include/linux/pgtable.h        | 27 +++++++++++++++++++--------
> > > >    2 files changed, 19 insertions(+), 18 deletions(-)
> > > >
> > > > diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
> > > > index 8cd6514e3052..8c85f2eabbaa 100644
> > > > --- a/arch/x86/include/asm/pgtable.h
> > > > +++ b/arch/x86/include/asm/pgtable.h
> > > > @@ -1077,16 +1077,6 @@ static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm,
> > > >           return pte;
> > > >    }
> > > >
> > > > -#define __HAVE_ARCH_PTEP_CLEAR
> > > > -static inline void ptep_clear(struct mm_struct *mm, unsigned long addr,
> > > > -                             pte_t *ptep)
> > > > -{
> > > > -       if (IS_ENABLED(CONFIG_PAGE_TABLE_CHECK))
> > > > -               ptep_get_and_clear(mm, addr, ptep);
> > > > -       else
> > > > -               pte_clear(mm, addr, ptep);
> > > > -}
> > > > -
> > > >    #define __HAVE_ARCH_PTEP_SET_WRPROTECT
> > > >    static inline void ptep_set_wrprotect(struct mm_struct *mm,
> > > >                                         unsigned long addr, pte_t *ptep)
> > > > diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h
> > > > index f4f4077b97aa..d27fd0ed84a9 100644
> > > > --- a/include/linux/pgtable.h
> > > > +++ b/include/linux/pgtable.h
> > > > @@ -12,6 +12,7 @@
> > > >    #include <linux/bug.h>
> > > >    #include <linux/errno.h>
> > > >    #include <asm-generic/pgtable_uffd.h>
> > > > +#include <linux/page_table_check.h>
> > > >
> > > >    #if 5 - defined(__PAGETABLE_P4D_FOLDED) - defined(__PAGETABLE_PUD_FOLDED) - \
> > > >           defined(__PAGETABLE_PMD_FOLDED) != CONFIG_PGTABLE_LEVELS
> > > > @@ -259,14 +260,6 @@ static inline int pmdp_clear_flush_young(struct vm_area_struct *vma,
> > > >    #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
> > > >    #endif
> > > >
> > > > -#ifndef __HAVE_ARCH_PTEP_CLEAR
> > > > -static inline void ptep_clear(struct mm_struct *mm, unsigned long addr,
> > > > -                             pte_t *ptep)
> > > > -{
> > > > -       pte_clear(mm, addr, ptep);
> > > > -}
> > > > -#endif
> > > > -
> > > >    #ifndef __HAVE_ARCH_PTEP_GET_AND_CLEAR
> > > >    static inline pte_t ptep_get_and_clear(struct mm_struct *mm,
> > > >                                          unsigned long address,
> > > > @@ -274,10 +267,23 @@ static inline pte_t ptep_get_and_clear(struct mm_struct *mm,
> > > >    {
> > > >           pte_t pte = *ptep;
> > > >           pte_clear(mm, address, ptep);
> > > > +       page_table_check_pte_clear(mm, address, pte);
> > > >           return pte;
> > > >    }
> > > >    #endif
> > > >
> > > > +#ifndef __HAVE_ARCH_PTEP_CLEAR
> > > > +static inline void ptep_clear(struct mm_struct *mm, unsigned long addr,
> > > > +                             pte_t *ptep)
> > > > +{
> > > > +#ifdef CONFIG_PAGE_TABLE_CHECK
> > > > +       ptep_get_and_clear(mm, addr, ptep);
> > > > +#else
> > > > +       pte_clear(mm, addr, ptep);
> > > > +#endif
> > >
> > > I have a preference to use if (IS_ENABLED(CONFIG_PAGE_TABLE_CHECK))
> > > instead of #ifdef. The end result is the same. Otherwise it looks
> > > good.
> > >
> > > Thanks,
> > > Pasha
> > > .
> >
> > I have a little hesitation when making this change , in theory, add if
> > here may affect the performance a little in some scenarios. However, the
> > impact on the whole call path should be small.
>
> I do not think so, the compiler should optimize out IS_ENABLED() when
> not enabled, no?

You are right.

https://www.kernel.org/doc/Documentation/process/coding-style.rst

The compiler will constant-fold the conditional away, and include or 
exclude the block of code just as with an #ifdef, so this will **not add 
any runtime overhead**.

Thanks :)

> > I will send v3 using if (IS_ENABLED(CONFIG_PAGE_TABLE_CHECK)).
> >
> > Thanks.
> > Tong
> > .
> .