Reducing fragmentation using zones v6

This is V6 of the zone-based anti-fragmentation patches. These patches require the architecture-independent zone-sizing patches posted under the subject "[PATCH 0/7] Sizing zones and holes in an architecture independent manner V5" at http://marc.theaimsgroup.com/?l=linux-kernel&m=114649066331962&w=2 .

In this version, all the zone sizing is done in one place rather than per-architecture like older versions. Any architecture that uses add_active_range() and free_area_init_nodes() to initialise it's zones can trivially support ZONE_EASYRCLM by parsing a kernelcore= command-line parameter and calling set_required_kernelcore().

As these patches require the zone-sizing merged first, I am not looking to merge these now. I just wanted to show that zone-based anti-fragmentation is a bit nicer looking when the zone sizing is not architecture-specific. Some of this patch is a bit rough and ready :)

Changelog since v5
o Rebase on top of arch-independent zone and hole sizing code
o Calculate ZONE_EASYRCLM boundaries in an arch-independent manner

Changelog since v4
o Move x86_64 from one patch to another
o Fix for oops bug on ppc64

Changelog since v3
o Minor bugs
o ppc64 can specify kernelcore
o Ability to disable use of ZONE_EASYRCLM at boot time
o HugeTLB uses ZONE_EASYRCLM
o Add drain-percpu caches for testing
o boot-parameter documentation added

This is a zone-based approach to anti-fragmentation. This is posted in light of the discussions related to the list-based (sometimes dubbed as sub-zones) approach where the prevailing opinion was that zones were the answer. The patches have been boot-tested based on linux-2.6.17-rc3-mm1 with x86, x86_64, ppc64 in a variety of different configurations. It was boot-tested on ia64 where it blew up before the serial console was initialised. Lacking early_printk, I have not figured out what is going wrong there yet. If anyone has physical access to an IA64 that can send me a bug report, I'd appreciate it.

Ordinarily, I would include performance regressions, but I'm not looking to merge this time.

The diffstat for the all the patches is;

 Documentation/kernel-parameters.txt |   16 ++
 arch/i386/kernel/setup.c            |   12 ++
 arch/i386/kernel/srat.c             |    2
 arch/ia64/kernel/efi.c              |    5
 arch/powerpc/kernel/prom.c          |    9 +
 arch/ppc/mm/init.c                  |    9 +
 arch/x86_64/kernel/setup.c          |    6 +
 arch/x86_64/mm/init.c               |    2
 fs/compat.c                         |    2
 fs/exec.c                           |    2
 fs/inode.c                          |   11 ++
 fs/ntfs/malloc.h                    |    3
 include/asm-i386/page.h             |    3
 include/linux/gfp.h                 |    3
 include/linux/highmem.h             |    2
 include/linux/mm.h                  |    1
 include/linux/mmzone.h              |   12 +-
 mm/hugetlb.c                        |    4
 mm/mem_init.c                       |  198 +++++++++++++++++++++++++++++++++---
 mm/memory.c                         |    4
 mm/page_alloc.c                     |   10 +
 mm/shmem.c                          |    4
 mm/swap_state.c                     |    2
 23 files changed, 288 insertions(+), 34 deletions(-)

-- 
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>