[PATCH v1 09/10] mm/memory_hotplug: Don't mark pages PG_reserved when initializing the memmap

[Boris Ostrovsky]

On 11/5/19 5:18 AM, David Hildenbrand wrote:
> On 04.11.19 23:44, Boris Ostrovsky wrote:
>> On 10/24/19 8:09 AM, David Hildenbrand wrote:
>>> diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c
>>> index 4f2e78a5e4db..af69f057913a 100644
>>> --- a/drivers/xen/balloon.c
>>> +++ b/drivers/xen/balloon.c
>>> @@ -374,6 +374,13 @@ static void xen_online_page(struct page *page,
>>> unsigned int order)
>>>       mutex_lock(&balloon_mutex);
>>>       for (i = 0; i < size; i++) {
>>>           p = pfn_to_page(start_pfn + i);
>>> +        /*
>>> +         * TODO: The core used to mark the pages reserved. Most
>>> probably
>>> +         * we can stop doing that now. However, especially
>>> +         * alloc_xenballooned_pages() left PG_reserved set
>>> +         * on pages that can get mapped to user space.
>>> +         */
>>> +        __SetPageReserved(p);
>>
>> I suspect this is not needed. Pages can get into balloon either from
>> here or from non-hotplug path (e.g. decrease_reservation()) and so when
>> we get a page from the balloon we would get a random page that may or
>> may not have Reserved bit set.
>
> Yeah, I also think it is fine. If you agree, I'll drop this hunk and
> add details to the patch description.
>
>


Yes, let's do that. Thanks.

-boris