mbox series

[v2,0/2] mm: remove the memory isolate notifier

Message ID 20191114131911.11783-1-david@redhat.com (mailing list archive)
Headers show
Series mm: remove the memory isolate notifier | expand

Message

David Hildenbrand Nov. 14, 2019, 1:19 p.m. UTC 
This is the MM part of
	https://lkml.org/lkml/2019/10/31/487

"We can get rid of the memory isolate notifier by switching to balloon
compaction in powerpc's CMM (Collaborative Memory Management). The memory
isolate notifier was only necessary to allow to offline memory blocks that
contain inflated/"loaned" pages - which also possible when the inflated
pages are movable (via balloon compaction). [...]"

Michael queued the POWERPC bits that remove the single user, but I am
missing ACKs for the MM bits. I think it makes sense to let these two
patches also go via Michael's tree, to avoid collissions. Thoughts?

v1 -> v2: (MM bits)
- "mm: remove the memory isolate notifier"
-- Remove another stale comment
-- Minor code cleanup

David Hildenbrand (2):
  mm: remove the memory isolate notifier
  mm: remove "count" parameter from has_unmovable_pages()

 drivers/base/memory.c          | 19 -----------------
 include/linux/memory.h         | 27 ------------------------
 include/linux/page-isolation.h |  4 ++--
 mm/memory_hotplug.c            |  2 +-
 mm/page_alloc.c                | 21 +++++++------------
 mm/page_isolation.c            | 38 ++++------------------------------
 6 files changed, 14 insertions(+), 97 deletions(-)

Comments

David Hildenbrand Nov. 15, 2019, 1 p.m. UTC | #1 
On 14.11.19 14:19, David Hildenbrand wrote:
> This is the MM part of
> 	https://lkml.org/lkml/2019/10/31/487
> 
> "We can get rid of the memory isolate notifier by switching to balloon
> compaction in powerpc's CMM (Collaborative Memory Management). The memory
> isolate notifier was only necessary to allow to offline memory blocks that
> contain inflated/"loaned" pages - which also possible when the inflated
> pages are movable (via balloon compaction). [...]"
> 
> Michael queued the POWERPC bits that remove the single user, but I am
> missing ACKs for the MM bits. I think it makes sense to let these two
> patches also go via Michael's tree, to avoid collissions. Thoughts?

@Michael, the ACKs from Michal should be sufficient to take these two
patches via your tree. Fine with you? Thanks!
David Hildenbrand Dec. 2, 2019, 10:27 a.m. UTC | #2 
On 14.11.19 14:19, David Hildenbrand wrote:
> This is the MM part of
> 	https://lkml.org/lkml/2019/10/31/487
> 
> "We can get rid of the memory isolate notifier by switching to balloon
> compaction in powerpc's CMM (Collaborative Memory Management). The memory
> isolate notifier was only necessary to allow to offline memory blocks that
> contain inflated/"loaned" pages - which also possible when the inflated
> pages are movable (via balloon compaction). [...]"
> 
> Michael queued the POWERPC bits that remove the single user, but I am
> missing ACKs for the MM bits. I think it makes sense to let these two
> patches also go via Michael's tree, to avoid collissions. Thoughts?

The prereqs (powerpc bits) are upstream - I assume Michael didn't want
to mess with MM patches. @Andrew, please pick these up once you feel
like time for them has come. :)
Michael Ellerman Dec. 9, 2019, 11:15 a.m. UTC | #3 
David Hildenbrand <david@redhat.com> writes:
> On 14.11.19 14:19, David Hildenbrand wrote:
>> This is the MM part of
>> 	https://lkml.org/lkml/2019/10/31/487
>> 
>> "We can get rid of the memory isolate notifier by switching to balloon
>> compaction in powerpc's CMM (Collaborative Memory Management). The memory
>> isolate notifier was only necessary to allow to offline memory blocks that
>> contain inflated/"loaned" pages - which also possible when the inflated
>> pages are movable (via balloon compaction). [...]"
>> 
>> Michael queued the POWERPC bits that remove the single user, but I am
>> missing ACKs for the MM bits. I think it makes sense to let these two
>> patches also go via Michael's tree, to avoid collissions. Thoughts?
>
> The prereqs (powerpc bits) are upstream - I assume Michael didn't want
> to mess with MM patches.

Yes, sorry I meant to send you a mail saying so.

cheers