Re: How to handle PTE tables with non contiguous entries ?

[Christophe LEROY]

Le 10/09/2018 Ã 23:06, Nicholas Piggin a ÃcritÂ:
> On Mon, 10 Sep 2018 14:34:37 +0000
> Christophe Leroy <christophe.leroy@xxxxxx> wrote:
>
> > Hi,
> >
> > I'm having a hard time figuring out the best way to handle the following
> > situation:
> >
> > On the powerpc8xx, handling 16k size pages requires to have page tables
> > with 4 identical entries.
> >
> > Initially I was thinking about handling this by simply modifying
> > pte_index() which changing pte_t type in order to have one entry every
> > 16 bytes, then replicate the PTE value at *ptep, *ptep+1,*ptep+2 and
> > *ptep+3 both in set_pte_at() and pte_update().
> >
> > However, this doesn't work because many many places in the mm core part
> > of the kernel use loops on ptep with single ptep++ increment.
> >
> > Therefore did it with the following hack:
> >
> >    /* PTE level */
> > +#if defined(CONFIG_PPC_8xx) && defined(CONFIG_PPC_16K_PAGES)
> > +typedef struct { pte_basic_t pte, pte1, pte2, pte3; } pte_t;
> > +#else
> >    typedef struct { pte_basic_t pte; } pte_t;
> > +#endif
> >
> > @@ -181,7 +192,13 @@ static inline unsigned long pte_update(pte_t *p,
> >           : "cc" );
> >    #else /* PTE_ATOMIC_UPDATES */
> >           unsigned long old = pte_val(*p);
> > -       *p = __pte((old & ~clr) | set);
> > +       unsigned long new = (old & ~clr) | set;
> > +
> > +#if defined(CONFIG_PPC_8xx) && defined(CONFIG_PPC_16K_PAGES)
> > +       p->pte = p->pte1 = p->pte2 = p->pte3 = new;
> > +#else
> > +       *p = __pte(new);
> > +#endif
> >    #endif /* !PTE_ATOMIC_UPDATES */
> >
> >    #ifdef CONFIG_44x
> >
> >
> > @@ -161,7 +161,11 @@ static inline void __set_pte_at(struct mm_struct
> > *mm, unsigned long addr,
> >           /* Anything else just stores the PTE normally. That covers all
> > 64-bit
> >            * cases, and 32-bit non-hash with 32-bit PTEs.
> >            */
> > +#if defined(CONFIG_PPC_8xx) && defined(CONFIG_PPC_16K_PAGES)
> > +       ptep->pte = ptep->pte1 = ptep->pte2 = ptep->pte3 = pte_val(pte);
> > +#else
> >           *ptep = pte;
> > +#endif
> >
> >
> >
> > But I'm not too happy with it as it means pte_t is not a single type
> > anymore so passing it from one function to the other is quite heavy.
> >
> >
> > Would someone have an idea of an elegent way to handle that ?
>
> I can't think of anything better. Do we pass pte by value to a lot of
> non inlined functions? Possible to inline the important ones?

Good question, I need to check that.

> Other option, try to get an iterator like pte = pte_next(pte) into core
> code.

Yes I've been thinking about that, but it looks like a huge job to 
identify all places, as some drivers are also playing with it.
I'm not sure it is only to find all 'pte++' and 'ptep++', I fear there 
might be places with more unexpected names.

Thanks
Christophe

> Thanks,
> Nick