On 14 May 2025, at 10:12, David Hildenbrand wrote:
> On 14.05.25 15:45, Zi Yan wrote:
>> On 14 May 2025, at 7:15, David Hildenbrand wrote:
>>
>>> This is a requirement for making PageOffline pages not have a refcount
>>> in the long future ("frozen"), and for reworking non-folio page migration
>>> in the near future.
>>>
>>> I have patches mostly ready to go to handle the latter. For turning all
>>> PageOffline() pages frozen, the non-folio page migration and memory
>>> ballooning drivers will have to be reworked first, to no longer rely on
>>> the refcount of PageOffline pages.
>>>
>>> Introduce PG_offline_skippable that only applies to PageOffline() pages --
>>> of course, reusing one of the existing PG_ flags for now -- and convert
>>> virtio-mem to make use of the new way: to allow for skipping PageOffline
>>> pages during memory offlining, treating them as if they would not be
>>> allocated.
>>
>
> Thanks for taking a look!
>
>> IIUC, based on Documentation/admin-guide/mm/memory-hotplug.rst,
>> to offline a page, the page first needs to be set PageOffline() to be
>
> PageOffline is not mentioned in there. :)
Sorry, I was mixing the code with the documentation as I was reading
both.
>
> Note that PageOffline() is a bit confusing because it's "Memory block online but page is logically offline (e.g., has a memmap that can be touched, but the page content should not be touched)".
So PageOffline() is before memory block offline, which is the first phase of
memory hotunplug.
>
> (memory block offline -> all pages offline and have effectively no state because the memmap is stale)
What do you mean by memmap is stale? When a memory block is offline, memmap is
still present, so pfn scanner can see these pages. pfn scanner checks memmap
to know that it should not touch these pages, right?
>
>> removed from page allocator.
>
> Usually, all pages are freed back to the buddy (isolated pageblock -> put onto the isolated list). Memory offlining code can then simply grab these "free" pages from the buddy -- no PageOffline involved.
>
> If something fails during memory offlining, these isolated pages are simply put back on the appropriate migratetype list and become ordinary free pages that can be allocated immediately.
I am familiar with this part. Then, when PageOffline is used?
