On Fri, Sep 12, 2025 at 04:28:09PM +0200, David Hildenbrand wrote:

On 12.09.25 15:47, David Hildenbrand wrote:

quoted

On 12.09.25 14:19, Kiryl Shutsemau wrote:

quoted

On Thu, Sep 11, 2025 at 09:27:55PM -0600, Nico Pache wrote:

quoted

The following series provides khugepaged with the capability to collapse
anonymous memory regions to mTHPs.

To achieve this we generalize the khugepaged functions to no longer depend
on PMD_ORDER. Then during the PMD scan, we use a bitmap to track individual
pages that are occupied (!none/zero). After the PMD scan is done, we do
binary recursion on the bitmap to find the optimal mTHP sizes for the PMD
range. The restriction on max_ptes_none is removed during the scan, to make
sure we account for the whole PMD range. When no mTHP size is enabled, the
legacy behavior of khugepaged is maintained. max_ptes_none will be scaled
by the attempted collapse order to determine how full a mTHP must be to be
eligible for the collapse to occur. If a mTHP collapse is attempted, but
contains swapped out, or shared pages, we don't perform the collapse. It is
now also possible to collapse to mTHPs without requiring the PMD THP size
to be enabled.

When enabling (m)THP sizes, if max_ptes_none >= HPAGE_PMD_NR/2 (255 on
4K page size), it will be automatically capped to HPAGE_PMD_NR/2 - 1 for
mTHP collapses to prevent collapse "creep" behavior. This prevents
constantly promoting mTHPs to the next available size, which would occur
because a collapse introduces more non-zero pages that would satisfy the
promotion condition on subsequent scans.

Hm. Maybe instead of capping at HPAGE_PMD_NR/2 - 1 we can count
all-zeros 4k as none_or_zero? It mirrors the logic of shrinker.

BTW, I thought further about this and I agree: if we count zero-filled
pages towards none_or_zero one we can avoid the "creep" problem.

The scanning-for-zero part is rather nasty, though.

Aaand, thinking again from the other direction, this would mean that just
because pages became zero after some time that we would no longer collapse
because none_or_zero would then be higher. Hm ....

How I hate all of this so very very much :)

This is not new. Shrinker has the same problem: it cannot distinguish
between hot 4k that happened to be zero from the 4k that is there just
because of we faulted in 2M a time.

-- 
  Kiryl Shutsemau / Kirill A. Shutemov