Re: [PATCH v28 09/32] x86/mm: Introduce _PAGE_COW

[Yu, Yu-cheng]

On 8/16/2021 3:43 AM, Borislav Petkov wrote:
> On Thu, Jul 22, 2021 at 01:51:56PM -0700, Yu-cheng Yu wrote:
> > @@ -153,13 +178,23 @@ static inline int pud_young(pud_t pud)
> >   
> >   static inline int pte_write(pte_t pte)
> >   {
> > -	return pte_flags(pte) & _PAGE_RW;
> > +	/*
> > +	 * Shadow stack pages are always writable - but not by normal
> > +	 * instructions, and only by shadow stack operations.  Therefore,
> > +	 * the W=0,D=1 test with pte_shstk().
> > +	 */
> > +	return (pte_flags(pte) & _PAGE_RW) || pte_shstk(pte);
>
> Well, this is weird: if some kernel code queries a shstk page and this
> here function says it is writable but then goes and tries to write into
> it and that write fails, then it'll confuse the user.
>
> IOW, from where I'm standing, that should be:
>
> 	return (pte_flags(pte) & _PAGE_RW) && !pte_shstk(pte);
>
> as in, a writable page is one which has _PAGE_RW and it is *not* a
> shadow stack page because latter is special and not really writable.
> > Hmmm?

Indeed, this can be looked at in a few ways.  We can visualize 
pte_write() as 'CPU can write to it with MOV' or 'CPU can write to it 
with any opcodes'.  Depending on whatever pte_write() is, copy-on-write 
code can be adjusted accordingly.

Yu-cheng