Re: [PATCH v9 11/21] mm: pagewalk: Add p4d_entry() and pgd_entry()

[Mark Rutland]

On Wed, Jul 24, 2019 at 02:53:04PM +0100, Steven Price wrote:
> On 23/07/2019 11:14, Mark Rutland wrote:
> > On Mon, Jul 22, 2019 at 04:42:00PM +0100, Steven Price wrote:
> >> pgd_entry() and pud_entry() were removed by commit 0b1fbfe50006c410
> >> ("mm/pagewalk: remove pgd_entry() and pud_entry()") because there were
> >> no users. We're about to add users so reintroduce them, along with
> >> p4d_entry() as we now have 5 levels of tables.
> >>
> >> Note that commit a00cc7d9dd93d66a ("mm, x86: add support for
> >> PUD-sized transparent hugepages") already re-added pud_entry() but with
> >> different semantics to the other callbacks. Since there have never
> >> been upstream users of this, revert the semantics back to match the
> >> other callbacks. This means pud_entry() is called for all entries, not
> >> just transparent huge pages.
> >>
> >> Signed-off-by: Steven Price <steven.price@xxxxxxx>
> >> ---
> >>  include/linux/mm.h | 15 +++++++++------
> >>  mm/pagewalk.c      | 27 ++++++++++++++++-----------
> >>  2 files changed, 25 insertions(+), 17 deletions(-)
> >>
> >> diff --git a/include/linux/mm.h b/include/linux/mm.h
> >> index 0334ca97c584..b22799129128 100644
> >> --- a/include/linux/mm.h
> >> +++ b/include/linux/mm.h
> >> @@ -1432,15 +1432,14 @@ void unmap_vmas(struct mmu_gather *tlb, struct vm_area_struct *start_vma,
> >>  
> >>  /**
> >>   * mm_walk - callbacks for walk_page_range
> >> - * @pud_entry: if set, called for each non-empty PUD (2nd-level) entry
> >> - *	       this handler should only handle pud_trans_huge() puds.
> >> - *	       the pmd_entry or pte_entry callbacks will be used for
> >> - *	       regular PUDs.
> >> - * @pmd_entry: if set, called for each non-empty PMD (3rd-level) entry
> >> + * @pgd_entry: if set, called for each non-empty PGD (top-level) entry
> >> + * @p4d_entry: if set, called for each non-empty P4D entry
> >> + * @pud_entry: if set, called for each non-empty PUD entry
> >> + * @pmd_entry: if set, called for each non-empty PMD entry
> > 
> > How are these expected to work with folding?
> > 
> > For example, on arm64 with 64K pages and 42-bit VA, you can have 2-level
> > tables where the PGD is P4D, PUD, and PMD. IIUC we'd invoke the
> > callbacks for each of those levels where we found an entry in the pgd.
> > 
> > Either the callee handle that, or we should inhibit the callbacks when
> > levels are folded, and I think that needs to be explcitly stated either
> > way.
> > 
> > IIRC on x86 the p4d folding is dynamic depending on whether the HW
> > supports 5-level page tables. Maybe that implies the callee has to
> > handle that.
> 
> Yes, my assumption is that it has to be up to the callee to handle that
> because folding can be dynamic. I believe this also was how these
> callbacks work before they were removed. However I'll add a comment
> explaining that here as it's probably non-obvious.

That sounds good to me.

Thanks,
Mark.