lkml.org 
[lkml]   [2010]   [Jan]   [7]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 1/7] Allow CONFIG_MIGRATION to be set without CONFIG_NUMA
On Wed, 6 Jan 2010, Mel Gorman wrote:

> CONFIG_MIGRATION currently depends on CONFIG_NUMA. The current users of
> page migration such as sys_move_pages(), sys_migrate_pages() and cpuset
> process migration are ordinarily only beneficial on NUMA.
>
> As memory compaction will operate within a zone and is useful on both NUMA
> and non-NUMA systems, this patch allows CONFIG_MIGRATION to be set if the
> user selects CONFIG_COMPACTION as an option.
>
> TODO
> o After this patch is applied, the migration core is available but it
> also makes NUMA-specific features available. This is too much
> exposure so revisit this.
>

CONFIG_MIGRATION is no longer strictly dependent on CONFIG_NUMA since
ARCH_ENABLE_MEMORY_HOTREMOVE has allowed it to be configured for UMA
machines. All strictly NUMA features in the migration core should be
isolated under its #ifdef CONFIG_NUMA (sys_move_pages()) in mm/migrate.c
or by simply not compiling mm/mempolicy.c (sys_migrate_pages()), so this
patch looks fine as is (although the "help" text for CONFIG_MIGRATION
could be updated to reflect that it's useful for both memory hot-remove
and now compaction).

> Signed-off-by: Mel Gorman <mel@csn.ul.ie>
> ---
> mm/Kconfig | 12 +++++++++++-
> 1 files changed, 11 insertions(+), 1 deletions(-)
>
> diff --git a/mm/Kconfig b/mm/Kconfig
> index 17b8947..1d8e2b2 100644
> --- a/mm/Kconfig
> +++ b/mm/Kconfig
> @@ -168,12 +168,22 @@ config SPLIT_PTLOCK_CPUS
> default "4"
>
> #
> +# support for memory compaction
> +config COMPACTION
> + bool "Allow for memory compaction"
> + def_bool y
> + select MIGRATION
> + depends on EXPERIMENTAL && HUGETLBFS
> + help
> + Allows the compaction of memory for the allocation of huge pages.
> +
> +#
> # support for page migration
> #
> config MIGRATION
> bool "Page migration"
> def_bool y
> - depends on NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE
> + depends on NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION
> help
> Allows the migration of the physical location of pages of processes
> while the virtual addresses are not changed. This is useful for


\
 
 \ /
  Last update: 2010-01-07 22:49    [from the cache]
©2003-2011 Jasper Spaans