Re: [PATCH v2] mm/gup.c: Simplify and fix check_and_migrate_movable_pages() return codes

[Alistair Popple <apopple@nvidia.com>]

Pasha Tatashin <pasha.tatashin@soleen.com> writes:

> On Mon, Aug 1, 2022 at 8:32 PM Alistair Popple <apopple@nvidia.com> wrote:
>>
>> When pinning pages with FOLL_LONGTERM check_and_migrate_movable_pages()
>> is called to migrate pages out of zones which should not contain any
>> longterm pinned pages.
>>
>> When migration succeeds all pages will have been unpinned so pinning
>> needs to be retried. This is indicated by returning zero. When all pages
>> are in the correct zone the number of pinned pages is returned.
>>
>> However migration can also fail, in which case pages are unpinned and
>> -ENOMEM is returned. However if the failure was due to not being unable
>> to isolate a page zero is returned. This leads to indefinite looping in
>> __gup_longterm_locked().
>
> Hi Alistair,
>
> During prohibiting pinning movable zone development, there was a
> discussion where we figured that isolation errors should be transient
> [1]. What isolation errors are you seeing that lead to infinite loop?
> Why do they happen?

Thanks for the pointer Pasha. There were reports of qemu running into
the same zero page problem you reported there, see
https://lore.kernel.org/linux-mm/165490039431.944052.12458624139225785964.stgit@omen/

This doesn't directly fix that problem as we need to allow pinning of
the zero page, but it does prevent the infinite loop. I was going to
re-spin this patch to retry instead of instant failure however reading
that thread it seems the infinite loop is desired behaviour. So will
re-spin this to leave that in-place.

 - Alistair

> Pasha
>
> [1] https://lore.kernel.org/linux-mm/20201218104655.GW32193@dhcp22.suse.cz