On Wed, Aug 18, 2021 at 12:28:02PM +0800, Muchun Song wrote:

On Wed, Aug 18, 2021 at 11:18 AM Roman Gushchin [off-list ref] wrote:

quoted

On Sat, Aug 14, 2021 at 01:25:10PM +0800, Muchun Song wrote:

quoted

The diagram below shows how to make the folio lruvec lock safe when LRU
pages are reparented.

folio_lruvec_lock(folio)
    retry:
      lruvec = folio_lruvec(folio);

        // The folio is reparented at this time.
        spin_lock(&lruvec->lru_lock);

        if (unlikely(lruvec_memcg(lruvec) != folio_memcg(folio)))
            // Acquired the wrong lruvec lock and need to retry.
            // Because this folio is on the parent memcg lruvec list.
            goto retry;

        // If we reach here, it means that folio_memcg(folio) is stable.

memcg_reparent_objcgs(memcg)
    // lruvec belongs to memcg and lruvec_parent belongs to parent memcg.
    spin_lock(&lruvec->lru_lock);
    spin_lock(&lruvec_parent->lru_lock);

    // Move all the pages from the lruvec list to the parent lruvec list.

    spin_unlock(&lruvec_parent->lru_lock);
    spin_unlock(&lruvec->lru_lock);

After we acquire the lruvec lock, we need to check whether the folio is
reparented. If so, we need to reacquire the new lruvec lock. On the
routine of the LRU pages reparenting, we will also acquire the lruvec
lock (will be implemented in the later patch). So folio_memcg() cannot
be changed when we hold the lruvec lock.

Since lruvec_memcg(lruvec) is always equal to folio_memcg(folio) after
we hold the lruvec lock, lruvec_memcg_debug() check is pointless. So
remove it.

This is a preparation for reparenting the LRU pages.

Signed-off-by: Muchun Song <redacted>
Acked-by: Roman Gushchin <redacted>

Maybe it's mostly s/page/folio, but the patch looks quite differently
in comparison to the version I did ack. In general, please, drop acks
when there are significant changes between versions.

Got it. I'll drop all acks in the next version. Thanks.

Thank you!

The code look correct though. But the locking scheme becomes very complicated
(it was already complex). I wonder if there are better ideas. No ideas out
of my head, need to think more. If not, your approach looks ok.