The patch titled
Subject: mm/compaction: enable compacting >0 order folios.
has been added to the -mm mm-unstable branch. Its filename is
mm-compaction-enable-compacting-0-order-folios.patch
This patch will shortly appear at
https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/mm-compaction-enable-compacting-0-order-folios.patch
This patch will later appear in the mm-unstable branch at
git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***
The -mm tree is included into linux-next via the mm-everything
branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
and is updated there every 2-3 working days
------------------------------------------------------
From: Zi Yan <ziy@xxxxxxxxxx>
Subject: mm/compaction: enable compacting >0 order folios.
Date: Mon, 22 Jan 2024 22:46:33 -0500
Patch series "Enable >0 order folio memory compaction", v2.
This patchset enables >0 order folio memory compaction, which is one of
the prerequisitions for large folio support[1].
I am aware that split free pages is necessary for folio migration in
compaction, since if >0 order free pages are never split and no order-0
free page is scanned, compaction will end prematurely due to migration
returns -ENOMEM. Free page split becomes a must instead of an
optimization.
Some applications from vm-scalability show different performance trends on
default LRU and CONFIG_LRU_GEN from patch 1 (split folio during
compaction), to patch 2 (folio migration during compaction), to patch 3
(folio migration during compaction with free page split). I am looking
into it.
lkp ncompare results (with >5% delta) for default LRU and CONFIG_LRU_GEN
are shown at the bottom (on a 8-CPU (Intel Xeon E5-2650 v4 @ 2.20GHz) 16G
VM).
Overview
========
To support >0 order folio compaction, the patchset changes how free pages
used for migration are kept during compaction. Free pages used to be
split into order-0 pages that are post allocation processed (i.e.,
PageBuddy flag cleared, page order stored in page->private is zeroed, and
page reference is set to 1). Now all free pages are kept in a MAX_ORDER+1
array of page lists based on their order without post allocation process.
When migrate_pages() asks for a new page, one of the free pages, based on
the requested page order, is then processed and given out.
[1] https://lore.kernel.org/linux-mm/20230912162815.440749-1-zi.yan@xxxxxxxx/
[2] https://lore.kernel.org/linux-mm/20231113170157.280181-1-zi.yan@xxxxxxxx/
vm-scalability results on CONFIG_LRU_GEN
===
=========================================================================================
compiler/kconfig/rootfs/runtime/tbox_group/test/testcase:
gcc-13/defconfig/debian/300s/qemu-vm/small-allocs/vm-scalability
commit:
6.7.0-rc4+
6.7.0-rc4-split-folio-in-compaction+
6.7.0-rc4-folio-migration-in-compaction+
6.7.0-rc4-folio-migration-free-page-split+
6.7.0-rc4+ 6.7.0-rc4-split-folio-in-co 6.7.0-rc4-folio-migration-i 6.7.0-rc4-folio-migration-f
---------------- --------------------------- --------------------------- ---------------------------
%stddev %change %stddev %change %stddev %change %stddev
\ | \ | \ | \
2024326 +35.5% 2743772 ± 41% +364.0% 9392198 ± 35% +31.0% 2651634 vm-scalability.throughput
=========================================================================================
compiler/kconfig/rootfs/runtime/tbox_group/test/testcase:
gcc-13/defconfig/debian/300s/qemu-vm/small-allocs-mt/vm-scalability
commit:
6.7.0-rc4+
6.7.0-rc4-split-folio-in-compaction+
6.7.0-rc4-folio-migration-in-compaction+
6.7.0-rc4-folio-migration-free-page-split+
6.7.0-rc4+ 6.7.0-rc4-split-folio-in-co 6.7.0-rc4-folio-migration-i 6.7.0-rc4-folio-migration-f
---------------- --------------------------- --------------------------- ---------------------------
%stddev %change %stddev %change %stddev %change %stddev
\ | \ | \ | \
1450189 +0.9% 1463418 +30.4% 1891610 ± 22% +0.3% 1454100 vm-scalability.throughput
=========================================================================================
compiler/kconfig/rootfs/runtime/tbox_group/test/testcase:
gcc-13/defconfig/debian/300s/qemu-vm/mmap-xread-seq-mt/vm-scalability
commit:
6.7.0-rc4+
6.7.0-rc4-split-folio-in-compaction+
6.7.0-rc4-folio-migration-in-compaction+
6.7.0-rc4-folio-migration-free-page-split+
6.7.0-rc4+ 6.7.0-rc4-split-folio-in-co 6.7.0-rc4-folio-migration-i 6.7.0-rc4-folio-migration-f
---------------- --------------------------- --------------------------- ---------------------------
%stddev %change %stddev %change %stddev %change %stddev
\ | \ | \ | \
14428848 ± 27% -51.7% 6963308 ± 73% +13.5% 16372621 +11.2% 16046511 vm-scalability.throughput
=========================================================================================
compiler/kconfig/rootfs/runtime/tbox_group/test/testcase:
gcc-13/defconfig/debian/300s/qemu-vm/mmap-pread-seq/vm-scalability
commit:
6.7.0-rc4+
6.7.0-rc4-split-folio-in-compaction+
6.7.0-rc4-folio-migration-in-compaction+
6.7.0-rc4-folio-migration-free-page-split+
6.7.0-rc4+ 6.7.0-rc4-split-folio-in-co 6.7.0-rc4-folio-migration-i 6.7.0-rc4-folio-migration-f
---------------- --------------------------- --------------------------- ---------------------------
%stddev %change %stddev %change %stddev %change %stddev
\ | \ | \ | \
13569502 ± 24% -45.9% 7340064 ± 59% +12.3% 15240531 +10.4% 14983705 vm-scalability.throughput
=========================================================================================
compiler/kconfig/rootfs/runtime/tbox_group/test/testcase:
gcc-13/defconfig/debian/300s/qemu-vm/mmap-pread-seq-mt/vm-scalability
commit:
6.7.0-rc4+
6.7.0-rc4-split-folio-in-compaction+
6.7.0-rc4-folio-migration-in-compaction+
6.7.0-rc4-folio-migration-free-page-split+
6.7.0-rc4+ 6.7.0-rc4-split-folio-in-co 6.7.0-rc4-folio-migration-i 6.7.0-rc4-folio-migration-f
---------------- --------------------------- --------------------------- ---------------------------
%stddev %change %stddev %change %stddev %change %stddev
\ | \ | \ | \
13305823 ± 24% -45.1% 7299664 ± 56% +12.5% 14974725 +10.4% 14695963 vm-scalability.throughput
=========================================================================================
compiler/kconfig/rootfs/runtime/tbox_group/test/testcase:
gcc-13/defconfig/debian/300s/qemu-vm/lru-file-readtwice/vm-scalability
commit:
6.7.0-rc4+
6.7.0-rc4-split-folio-in-compaction+
6.7.0-rc4-folio-migration-in-compaction+
6.7.0-rc4-folio-migration-free-page-split+
6.7.0-rc4+ 6.7.0-rc4-split-folio-in-co 6.7.0-rc4-folio-migration-i 6.7.0-rc4-folio-migration-f
---------------- --------------------------- --------------------------- ---------------------------
%stddev %change %stddev %change %stddev %change %stddev
\ | \ | \ | \
13244376 ± 28% +54.2% 20425838 ± 23% -4.4% 12660113 ± 3% -9.0% 12045809 ± 3% vm-scalability.throughput
=========================================================================================
compiler/kconfig/rootfs/runtime/tbox_group/test/testcase:
gcc-13/defconfig/debian/300s/qemu-vm/lru-file-mmap-read/vm-scalability
commit:
6.7.0-rc4+
6.7.0-rc4-split-folio-in-compaction+
6.7.0-rc4-folio-migration-in-compaction+
6.7.0-rc4-folio-migration-free-page-split+
6.7.0-rc4+ 6.7.0-rc4-split-folio-in-co 6.7.0-rc4-folio-migration-i 6.7.0-rc4-folio-migration-f
---------------- --------------------------- --------------------------- ---------------------------
%stddev %change %stddev %change %stddev %change %stddev
\ | \ | \ | \
7021425 ± 11% -20.9% 5556751 ± 19% +14.8% 8057811 ± 3% +9.4% 7678613 ± 4% vm-scalability.throughput
=========================================================================================
compiler/kconfig/rootfs/runtime/size/tbox_group/test/testcase:
gcc-13/defconfig/debian/300s/256G/qemu-vm/msync/vm-scalability
commit:
6.7.0-rc4+
6.7.0-rc4-split-folio-in-compaction+
6.7.0-rc4-folio-migration-in-compaction+
6.7.0-rc4-folio-migration-free-page-split+
6.7.0-rc4+ 6.7.0-rc4-split-folio-in-co 6.7.0-rc4-folio-migration-i 6.7.0-rc4-folio-migration-f
---------------- --------------------------- --------------------------- ---------------------------
%stddev %change %stddev %change %stddev %change %stddev
\ | \ | \ | \
1208994 ±137% +263.5% 4394683 ± 49% -49.4% 611204 ± 6% -48.1% 627937 ± 13% vm-scalability.throughput
vm-scalability results on default LRU (with -no-mglru suffix)
===
=========================================================================================
compiler/kconfig/rootfs/runtime/tbox_group/test/testcase:
gcc-13/defconfig/debian/300s/qemu-vm/lru-file-readtwice/vm-scalability
commit:
6.7.0-rc4-no-mglru+
6.7.0-rc4-split-folio-in-compaction-no-mglru+
6.7.0-rc4-folio-migration-in-compaction-no-mglru+
6.7.0-rc4-folio-migration-free-page-split-no-mglru+
6.7.0-rc4-no-mgl 6.7.0-rc4-split-folio-in-co 6.7.0-rc4-folio-migration-i 6.7.0-rc4-folio-migration-f
---------------- --------------------------- --------------------------- ---------------------------
%stddev %change %stddev %change %stddev %change %stddev
\ | \ | \ | \
8412072 ± 3% +32.1% 11114537 ± 41% +3.5% 8703491 ± 3% +1.5% 8536343 ± 3% vm-scalability.throughput
=========================================================================================
compiler/kconfig/rootfs/runtime/tbox_group/test/testcase:
gcc-13/defconfig/debian/300s/qemu-vm/lru-file-mmap-read/vm-scalability
commit:
6.7.0-rc4-no-mglru+
6.7.0-rc4-split-folio-in-compaction-no-mglru+
6.7.0-rc4-folio-migration-in-compaction-no-mglru+
6.7.0-rc4-folio-migration-free-page-split-no-mglru+
6.7.0-rc4-no-mgl 6.7.0-rc4-split-folio-in-co 6.7.0-rc4-folio-migration-i 6.7.0-rc4-folio-migration-f
---------------- --------------------------- --------------------------- ---------------------------
%stddev %change %stddev %change %stddev %change %stddev
\ | \ | \ | \
7095358 +10.8% 7863635 ± 16% +5.5% 7484110 +1.5% 7200666 ± 4% vm-scalability.throughput
This patch (of 3):
migrate_pages() supports >0 order folio migration and during compaction,
even if compaction_alloc() cannot provide >0 order free pages,
migrate_pages() can split the source page and try to migrate the base
pages from the split. It can be a baseline and start point for adding
support for compacting >0 order folios.
Link: https://lkml.kernel.org/r/20240123034636.1095672-1-zi.yan@xxxxxxxx
Link: https://lkml.kernel.org/r/20240123034636.1095672-2-zi.yan@xxxxxxxx
Signed-off-by: Zi Yan <ziy@xxxxxxxxxx>
Suggested-by: Huang Ying <ying.huang@xxxxxxxxx>
Cc: Adam Manzanares <a.manzanares@xxxxxxxxxxx>
Cc: Baolin Wang <baolin.wang@xxxxxxxxxxxxxxxxx>
Cc: David Hildenbrand <david@xxxxxxxxxx>
Cc: Johannes Weiner <hannes@xxxxxxxxxxx>
Cc: Kemeng Shi <shikemeng@xxxxxxxxxxxxxxx>
Cc: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
Cc: Luis Chamberlain <mcgrof@xxxxxxxxxx>
Cc: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
Cc: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>
Cc: Ryan Roberts <ryan.roberts@xxxxxxx>
Cc: Vishal Moola (Oracle) <vishal.moola@xxxxxxxxx>
Cc: Vlastimil Babka <vbabka@xxxxxxx>
Cc: Yin Fengwei <fengwei.yin@xxxxxxxxx>
Cc: Yu Zhao <yuzhao@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/compaction.c | 43 +++++++++++++++++++++++++++++++++++--------
1 file changed, 35 insertions(+), 8 deletions(-)
--- a/mm/compaction.c~mm-compaction-enable-compacting-0-order-folios
+++ a/mm/compaction.c
@@ -816,6 +816,21 @@ static bool too_many_isolated(struct com
return too_many;
}
+/*
+ * 1. if the page order is larger than or equal to target_order (i.e.,
+ * cc->order and when it is not -1 for global compaction), skip it since
+ * target_order already indicates no free page with larger than target_order
+ * exists and later migrating it will most likely fail;
+ *
+ * 2. compacting > pageblock_order pages does not improve memory fragmentation,
+ * skip them;
+ */
+static bool skip_isolation_on_order(int order, int target_order)
+{
+ return (target_order != -1 && order >= target_order) ||
+ order >= pageblock_order;
+}
+
/**
* isolate_migratepages_block() - isolate all migrate-able pages within
* a single pageblock
@@ -1010,7 +1025,7 @@ isolate_migratepages_block(struct compac
/*
* Regardless of being on LRU, compound pages such as THP and
* hugetlbfs are not to be compacted unless we are attempting
- * an allocation much larger than the huge page size (eg CMA).
+ * an allocation larger than the compound page size.
* We can potentially save a lot of iterations if we skip them
* at once. The check is racy, but we can consider only valid
* values and the only danger is skipping too much.
@@ -1018,11 +1033,18 @@ isolate_migratepages_block(struct compac
if (PageCompound(page) && !cc->alloc_contig) {
const unsigned int order = compound_order(page);
- if (likely(order <= MAX_PAGE_ORDER)) {
- low_pfn += (1UL << order) - 1;
- nr_scanned += (1UL << order) - 1;
+ /*
+ * Skip based on page order and compaction target order
+ * and skip hugetlbfs pages.
+ */
+ if (skip_isolation_on_order(order, cc->order) ||
+ PageHuge(page)) {
+ if (order <= MAX_PAGE_ORDER) {
+ low_pfn += (1UL << order) - 1;
+ nr_scanned += (1UL << order) - 1;
+ }
+ goto isolate_fail;
}
- goto isolate_fail;
}
/*
@@ -1165,10 +1187,11 @@ isolate_migratepages_block(struct compac
}
/*
- * folio become large since the non-locked check,
- * and it's on LRU.
+ * Check LRU folio order under the lock
*/
- if (unlikely(folio_test_large(folio) && !cc->alloc_contig)) {
+ if (unlikely(skip_isolation_on_order(folio_order(folio),
+ cc->order) &&
+ !cc->alloc_contig)) {
low_pfn += folio_nr_pages(folio) - 1;
nr_scanned += folio_nr_pages(folio) - 1;
folio_set_lru(folio);
@@ -1786,6 +1809,10 @@ static struct folio *compaction_alloc(st
struct compact_control *cc = (struct compact_control *)data;
struct folio *dst;
+ /* this makes migrate_pages() split the source page and retry */
+ if (folio_test_large(src) > 0)
+ return NULL;
+
if (list_empty(&cc->freepages)) {
isolate_freepages(cc);
_
Patches currently in -mm which might be from ziy@xxxxxxxxxx are
mm-compaction-enable-compacting-0-order-folios.patch
mm-compaction-add-support-for-0-order-folio-memory-compaction.patch
mm-compaction-optimize-0-order-folio-compaction-with-free-page-split.patch
