Re: [PATCH] mm/swapfile: drop a large unmapped swap-in folio in unuse_pte()

[Mani Tofigh]

On 10/31/26 1:54 AM, Kairui Song wrote:
> On Sat, Oct 31, 2026 at 6:54 AM Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
> > Also, won't this change cause try_to_unuse() to keep spinning away until
> > someone kills it?  Might be better to return -EBUSY and fail the
> > swapoff?
>
> It seems fine? The folio is removed from the swap cache, so the next pass
> will go on. But you're right about the dirty and WARN case, it spins, so
> maybe:
>
>     if (unlikely(folio_test_large(folio) && !folio_test_anon(folio))) {
>        if (WARN_ON_ONCE(folio_test_dirty(folio)))
>           return -EBUSY;
>        swap_cache_del_folio(folio);
>        return 0;
>     }


Right, in the normal case the next pass of try_to_unuse() just swaps the
entry in as a small folio, so only the dirty case would spin.

David suggested a VM_WARN_ON_ONCE() since a dirty folio here would mean
a bug elsewhere, so for v2 I'm thinking of combining the two.  It'd be
a VM_WARN_ON_ONCE() for the dirty case, and the folio stays in the swap
cache with -EBUSY returned.

Thanks for taking a look!

Mani