2006-02-01 12:05:38 +01:00
|
|
|
Page migration
|
|
|
|
--------------
|
|
|
|
|
|
|
|
Page migration allows the moving of the physical location of pages between
|
|
|
|
nodes in a numa system while the process is running. This means that the
|
|
|
|
virtual addresses that the process sees do not change. However, the
|
|
|
|
system rearranges the physical location of those pages.
|
|
|
|
|
|
|
|
The main intend of page migration is to reduce the latency of memory access
|
|
|
|
by moving pages near to the processor where the process accessing that memory
|
|
|
|
is running.
|
|
|
|
|
|
|
|
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
|
2006-03-15 04:50:20 +01:00
|
|
|
a new memory policy via mbind(). The pages of process can also be relocated
|
2006-02-01 12:05:38 +01:00
|
|
|
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.
|
2006-03-15 04:50:20 +01:00
|
|
|
Page migration functions are provided by the numactl package by Andi Kleen
|
|
|
|
(a version later than 0.9.3 is required. Get it from
|
2008-08-15 09:40:18 +02:00
|
|
|
ftp://oss.sgi.com/www/projects/libnuma/download/). numactl provides 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 documentation in the
|
|
|
|
proc(5) man page.
|
2006-03-15 04:50:20 +01:00
|
|
|
|
|
|
|
Manual migration is useful if for example the scheduler has relocated
|
2006-02-01 12:05:38 +01:00
|
|
|
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
|
2006-06-23 11:03:55 +02:00
|
|
|
nearer to the new processor. The kernel itself does only provide
|
|
|
|
manual page migration support. Automatic page migration may be implemented
|
|
|
|
through user space processes that move pages. A special function call
|
|
|
|
"move_pages" allows the moving of individual pages within a process.
|
|
|
|
A NUMA profiler may f.e. obtain a log showing frequent off node
|
|
|
|
accesses and may use the result to move pages to more advantageous
|
|
|
|
locations.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
|
|
|
Larger installations usually partition the system using cpusets into
|
|
|
|
sections of nodes. Paul Jackson has equipped cpusets with the ability to
|
2009-02-04 10:12:08 +01:00
|
|
|
move pages when a task is moved to another cpuset (See
|
|
|
|
Documentation/cgroups/cpusets.txt).
|
2006-03-15 04:50:20 +01:00
|
|
|
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.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
|
|
|
Page migration allows the preservation of the relative location of pages
|
|
|
|
within a group of nodes for all migration techniques which will preserve a
|
|
|
|
particular memory allocation pattern generated even after migrating a
|
|
|
|
process. This is necessary in order to preserve the memory latencies.
|
|
|
|
Processes will run with similar performance after migration.
|
|
|
|
|
|
|
|
Page migration occurs in several steps. First a high level
|
2006-03-15 04:50:20 +01:00
|
|
|
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.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-03-15 04:50:20 +01:00
|
|
|
A. In kernel use of migrate_pages()
|
|
|
|
-----------------------------------
|
2006-02-01 12:05:38 +01:00
|
|
|
|
|
|
|
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
|
2006-03-15 04:50:20 +01:00
|
|
|
calling isolate_lru_page().
|
2006-02-01 12:05:38 +01:00
|
|
|
Calling isolate_lru_page increases the references to the page
|
2006-03-15 04:50:20 +01:00
|
|
|
so that it cannot vanish while the page migration occurs.
|
|
|
|
It also prevents the swapper or other scans to encounter
|
|
|
|
the page.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-06-23 11:03:55 +02:00
|
|
|
2. We need to have a function of type new_page_t that can be
|
|
|
|
passed to migrate_pages(). This function should figure out
|
|
|
|
how to allocate the correct new page given the old page.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
|
|
|
3. The migrate_pages() function is called which attempts
|
2006-06-23 11:03:55 +02:00
|
|
|
to do the migration. It will call the function to allocate
|
|
|
|
the new page for each page that is considered for
|
|
|
|
moving.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-03-15 04:50:20 +01:00
|
|
|
B. How migrate_pages() works
|
|
|
|
----------------------------
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-03-15 04:50:20 +01:00
|
|
|
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.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
|
|
|
Steps:
|
|
|
|
|
|
|
|
1. Lock the page to be migrated
|
|
|
|
|
|
|
|
2. Insure that writeback is complete.
|
|
|
|
|
2006-06-23 11:03:39 +02:00
|
|
|
3. Prep the new page that we want to move to. It is locked
|
2006-02-01 12:05:38 +01:00
|
|
|
and set to not being uptodate so that all accesses to the new
|
2006-03-15 04:50:20 +01:00
|
|
|
page immediately lock while the move is in progress.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-06-23 11:03:39 +02:00
|
|
|
4. The new page is prepped with some settings from the old page so that
|
|
|
|
accesses to the new page will discover a page with the correct settings.
|
|
|
|
|
|
|
|
5. All the page table references to the page are converted
|
|
|
|
to migration entries or dropped (nonlinear vmas).
|
|
|
|
This decrease the mapcount of a page. If the resulting
|
|
|
|
mapcount is not zero then we do not migrate the page.
|
|
|
|
All user space processes that attempt to access the page
|
|
|
|
will now wait on the page lock.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-03-15 04:50:20 +01:00
|
|
|
6. The radix tree lock is taken. This will cause all processes trying
|
2006-06-23 11:03:39 +02:00
|
|
|
to access the page via the mapping to block on the radix tree spinlock.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
|
|
|
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
|
2006-06-23 11:03:39 +02:00
|
|
|
page then we back out because someone else modified the radix tree.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-06-23 11:03:39 +02:00
|
|
|
9. The radix tree is changed to point to the new page.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-06-23 11:03:39 +02:00
|
|
|
10. The reference count of the old page is dropped because the radix tree
|
|
|
|
reference is gone. A reference to the new page is established because
|
|
|
|
the new page is referenced to by the radix tree.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-06-23 11:03:39 +02:00
|
|
|
11. The radix tree lock is dropped. With that lookups in the mapping
|
|
|
|
become possible again. Processes will move from spinning on the tree_lock
|
|
|
|
to sleeping on the locked new page.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-06-23 11:03:39 +02:00
|
|
|
12. The page contents are copied to the new page.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-06-23 11:03:39 +02:00
|
|
|
13. The remaining page flags are copied to the new page.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-06-23 11:03:39 +02:00
|
|
|
14. The old page flags are cleared to indicate that the page does
|
|
|
|
not provide any information anymore.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-06-23 11:03:39 +02:00
|
|
|
15. Queued up writeback on the new page is triggered.
|
2006-02-01 12:05:38 +01:00
|
|
|
|
2006-06-23 11:03:39 +02:00
|
|
|
16. If migration entries were page then replace them with real ptes. Doing
|
|
|
|
so will enable access for user space processes not already waiting for
|
|
|
|
the page lock.
|
2006-03-15 04:50:20 +01:00
|
|
|
|
|
|
|
19. The page locks are dropped from the old and new page.
|
2006-06-23 11:03:39 +02:00
|
|
|
Processes waiting on the page lock will redo their page faults
|
|
|
|
and will reach the new page.
|
2006-03-15 04:50:20 +01:00
|
|
|
|
|
|
|
20. The new page is moved to the LRU and can be scanned by the swapper
|
|
|
|
etc again.
|
|
|
|
|
2006-06-23 11:03:39 +02:00
|
|
|
Christoph Lameter, May 8, 2006.
|
2006-02-01 12:05:38 +01:00
|
|
|
|