[PATCH] Page migration documentation update

Copyright (C) 2004 BULL SA.

Written by Simon.Derr@bull.net

Portions Copyright (c) 2004 Silicon Graphics, Inc.

Portions Copyright (c) 2004-2006 Silicon Graphics, Inc.

Modified by Paul Jackson <pj@sgi.com>

Modified by Christoph Lameter <clameter@sgi.com>

CONTENTS:

=========

These subsets, or "soft partitions" must be able to be dynamically

adjusted, as the job mix changes, without impacting other concurrently

executing jobs.

executing jobs. The location of the running jobs pages may also be moved

when the memory locations are changed.

The kernel cpuset patch provides the minimum essential kernel

mechanisms required to efficiently implement such subsets.  It

1.3 How are cpusets implemented ?

---------------------------------

Cpusets provide a Linux kernel (2.6.7 and above) mechanism to constrain

which CPUs and Memory Nodes are used by a process or set of processes.

Cpusets provide a Linux kernel mechanism to constrain which CPUs and

Memory Nodes are used by a process or set of processes.

The Linux kernel already has a pair of mechanisms to specify on which

CPUs a task may be scheduled (sched_setaffinity) and on which Memory

If the cpuset flag file 'memory_migrate' is set true, then when

tasks are attached to that cpuset, any pages that task had

allocated to it on nodes in its previous cpuset are migrated

to the tasks new cpuset.  Depending on the implementation,

this migration may either be done by swapping the page out,

so that the next time the page is referenced, it will be paged

into the tasks new cpuset, usually on the node where it was

referenced, or this migration may be done by directly copying

the pages from the tasks previous cpuset to the new cpuset,

where possible to the same node, relative to the new cpuset,

as the node that held the page, relative to the old cpuset.

to the tasks new cpuset. The relative placement of the page within

the cpuset is preserved during these migration operations if possible.

For example if the page was on the second valid node of the prior cpuset

then the page will be placed on the second valid node of the new cpuset.

Also if 'memory_migrate' is set true, then if that cpusets

'mems' file is modified, pages allocated to tasks in that

cpuset, that were on nodes in the previous setting of 'mems',

will be moved to nodes in the new setting of 'mems.'  Again,

depending on the implementation, this might be done by swapping,

or by direct copying.  In either case, pages that were not in

the tasks prior cpuset, or in the cpusets prior 'mems' setting,

will not be moved.

will be moved to nodes in the new setting of 'mems.'

Pages that were not in the tasks prior cpuset, or in the cpusets

prior 'mems' setting, will not be moved.

There is an exception to the above.  If hotplug functionality is used

to remove all the CPUs that are currently assigned to a cpuset,

  # The next line should display '/Charlie'

  cat /proc/self/cpuset

In the case that a change of cpuset includes wanting to move already

allocated memory pages, consider further the work of IWAMOTO

Toshihiro <iwamoto@valinux.co.jp> for page remapping and memory

hotremoval, which can be found at:

  http://people.valinux.co.jp/~iwamoto/mh.html

The integration of cpusets with such memory migration is not yet

available.

In the future, a C library interface to cpusets will likely be

available.  For now, the only way to query or modify cpusets is

via the cpuset file system, using the various cd, mkdir, echo, cat,

Page migration allows a process to manually relocate the node on which its

pages are located through the MF_MOVE and MF_MOVE_ALL options while setting

a new memory policy. The pages of process can also be relocated

a new memory policy via mbind(). The pages of process can also be relocated

from another process using the sys_migrate_pages() function call. The

migrate_pages function call takes two sets of nodes and moves pages of a

process that are located on the from nodes to the destination nodes.

Manual migration is very useful if for example the scheduler has relocated

Page migration functions are provided by the numactl package by Andi Kleen

(a version later than 0.9.3 is required. Get it from

ftp://ftp.suse.com/pub/people/ak). numactl provided libnuma which

provides an interface similar to other numa functionality for page migration.

cat /proc/<pid>/numa_maps allows an easy review of where the pages of

a process are located. See also the numa_maps manpage in the numactl package.

Manual migration is useful if for example the scheduler has relocated

a process to a processor on a distant node. A batch scheduler or an

administrator may detect the situation and move the pages of the process

nearer to the new processor. At some point in the future we may have

Larger installations usually partition the system using cpusets into

sections of nodes. Paul Jackson has equipped cpusets with the ability to

move pages when a task is moved to another cpuset. This allows automatic

control over locality of a process. If a task is moved to a new cpuset

then also all its pages are moved with it so that the performance of the

process does not sink dramatically (as is the case today).

move pages when a task is moved to another cpuset (See ../cpusets.txt).

Cpusets allows the automation of process locality. If a task is moved to

a new cpuset then also all its pages are moved with it so that the

performance of the process does not sink dramatically. Also the pages

of processes in a cpuset are moved if the allowed memory nodes of a

cpuset are changed.

Page migration allows the preservation of the relative location of pages

within a group of nodes for all migration techniques which will preserve a

Processes will run with similar performance after migration.

Page migration occurs in several steps. First a high level

description for those trying to use migrate_pages() and then

a low level description of how the low level details work.

description for those trying to use migrate_pages() from the kernel

(for userspace usage see the Andi Kleen's numactl package mentioned above)

and then a low level description of how the low level details work.

A. Use of migrate_pages()

-------------------------

A. In kernel use of migrate_pages()

-----------------------------------

1. Remove pages from the LRU.

   Lists of pages to be migrated are generated by scanning over

   pages and moving them into lists. This is done by

   calling isolate_lru_page() or __isolate_lru_page().

   calling isolate_lru_page().

   Calling isolate_lru_page increases the references to the page

   so that it cannot vanish under us.

   so that it cannot vanish while the page migration occurs.

   It also prevents the swapper or other scans to encounter

   the page.

2. Generate a list of newly allocates page to move the contents

   of the first list to.

2. Generate a list of newly allocates page. These pages will contain the

   contents of the pages from the first list after page migration is

   complete.

3. The migrate_pages() function is called which attempts

   to do the migration. It returns the moved pages in the

4. The leftover pages of various types are returned

   to the LRU using putback_to_lru_pages() or otherwise

   disposed of. The pages will still have the refcount as

   increased by isolate_lru_pages()!

   increased by isolate_lru_pages() if putback_to_lru_pages() is not

   used! The kernel may want to handle the various cases of failures in

   different ways.

B. Operation of migrate_pages()

--------------------------------

B. How migrate_pages() works

----------------------------

migrate_pages does several passes over its list of pages. A page is moved

if all references to a page are removable at the time.

migrate_pages() does several passes over its list of pages. A page is moved

if all references to a page are removable at the time. The page has

already been removed from the LRU via isolate_lru_page() and the refcount

is increased so that the page cannot be freed while page migration occurs.

Steps:

3. Make sure that the page has assigned swap cache entry if

   it is an anonyous page. The swap cache reference is necessary

   to preserve the information contain in the page table maps.

   to preserve the information contain in the page table maps while

   page migration occurs.

4. Prep the new page that we want to move to. It is locked

   and set to not being uptodate so that all accesses to the new

   page immediately lock while we are moving references.

   page immediately lock while the move is in progress.

5. All the page table references to the page are either dropped (file backed)

   or converted to swap references (anonymous pages). This should decrease the

   reference count.

5. All the page table references to the page are either dropped (file

   backed pages) or converted to swap references (anonymous pages).

   This should decrease the reference count.

6. The radix tree lock is taken

6. The radix tree lock is taken. This will cause all processes trying

   to reestablish a pte to block on the radix tree spinlock.

7. The refcount of the page is examined and we back out if references remain

   otherwise we know that we are the only one referencing this page.

8. The radix tree is checked and if it does not contain the pointer to this

   page then we back out.

   page then we back out because someone else modified the mapping first.

9. The mapping is checked. If the mapping is gone then a truncate action may

   be in progress and we back out.

10. The new page is prepped with some settings from the old page so that accesses

   to the new page will be discovered to have the correct settings.

10. The new page is prepped with some settings from the old page so that

   accesses to the new page will be discovered to have the correct settings.

11. The radix tree is changed to point to the new page.

12. The reference count of the old page is dropped because the reference has now

    been removed.

12. The reference count of the old page is dropped because the radix tree

    reference is gone.

13. The radix tree lock is dropped.

13. The radix tree lock is dropped. With that lookups become possible again

    and other processes will move from spinning on the tree lock to sleeping on

    the locked new page.

14. The page contents are copied to the new page.

17. Queued up writeback on the new page is triggered.

18. If swap pte's were generated for the page then remove them again.

18. If swap pte's were generated for the page then replace them with real

    ptes. This will reenable access for processes not blocked by the page lock.

19. The page locks are dropped from the old and new page.

    Processes waiting on the page lock can continue.

20. The new page is moved to the LRU and can be scanned by the swapper

    etc again.

TODO list

---------

- Page migration requires the use of swap handles to preserve the

  information of the anonymous page table entries. This means that swap

  space is reserved but never used. The maximum number of swap handles used

  is determined by CHUNK_SIZE (see mm/mempolicy.c) per ongoing migration.

  Reservation of pages could be avoided by having a special type of swap

  handle that does not require swap space and that would only track the page

  references. Something like that was proposed by Marcelo Tosatti in the

  past (search for migration cache on lkml or linux-mm@kvack.org).

19. The locks are dropped from the old and new page.

- Page migration unmaps ptes for file backed pages and requires page

  faults to reestablish these ptes. This could be optimized by somehow

  recording the references before migration and then reestablish them later.

  However, there are several locking challenges that have to be overcome

  before this is possible.

20. The new page is moved to the LRU.

- Page migration generates read ptes for anonymous pages. Dirty page

  faults are required to make the pages writable again. It may be possible

  to generate a pte marked dirty if it is known that the page is dirty and

  that this process has the only reference to that page.

Christoph Lameter, December 19, 2005.

Christoph Lameter, March 8, 2006.