The patch titled
Subject: khugepaged: allow to collapse PTE-mapped compound pages
has been added to the -mm tree. Its filename is
khugepaged-allow-to-collapse-pte-mapped-compound-pages.patch
This patch should soon appear at
http://ozlabs.org/~akpm/mmots/broken-out/khugepaged-allow-to-collapse-pte-mapped-compound-pages.patch
and later at
http://ozlabs.org/~akpm/mmotm/broken-out/khugepaged-allow-to-collapse-pte-mapped-compound-pages.patch
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 and is updated
there every 3-4 working days
------------------------------------------------------
From: "Kirill A. Shutemov" <kirill.shutemov@xxxxxxxxxxxxxxx>
Subject: khugepaged: allow to collapse PTE-mapped compound pages
We can collapse PTE-mapped compound pages. We only need to avoid handling
them more than once: lock/unlock page only once if it's present in the PMD
range multiple times as it handled on compound level. The same goes for
LRU isolation and putback.
Link: http://lkml.kernel.org/r/20200416160026.16538-7-kirill.shutemov@xxxxxxxxxxxxxxx
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
Reviewed-by: William Kucharski <william.kucharski@xxxxxxxxxx>
Reviewed-by: Zi Yan <ziy@xxxxxxxxxx>
Tested-by: Zi Yan <ziy@xxxxxxxxxx>
Acked-by: Yang Shi <yang.shi@xxxxxxxxxxxxxxxxx>
Cc: Andrea Arcangeli <aarcange@xxxxxxxxxx>
Cc: John Hubbard <jhubbard@xxxxxxxxxx>
Cc: Mike Kravetz <mike.kravetz@xxxxxxxxxx>
Cc: Ralph Campbell <rcampbell@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/khugepaged.c | 99 ++++++++++++++++++++++++++++++----------------
1 file changed, 65 insertions(+), 34 deletions(-)
--- a/mm/khugepaged.c~khugepaged-allow-to-collapse-pte-mapped-compound-pages
+++ a/mm/khugepaged.c
@@ -512,17 +512,30 @@ void __khugepaged_exit(struct mm_struct
static void release_pte_page(struct page *page)
{
- dec_node_page_state(page, NR_ISOLATED_ANON + page_is_file_lru(page));
+ mod_node_page_state(page_pgdat(page),
+ NR_ISOLATED_ANON + page_is_file_lru(page),
+ -compound_nr(page));
unlock_page(page);
putback_lru_page(page);
}
-static void release_pte_pages(pte_t *pte, pte_t *_pte)
+static void release_pte_pages(pte_t *pte, pte_t *_pte,
+ struct list_head *compound_pagelist)
{
+ struct page *page, *tmp;
+
while (--_pte >= pte) {
pte_t pteval = *_pte;
- if (!pte_none(pteval) && !is_zero_pfn(pte_pfn(pteval)))
- release_pte_page(pte_page(pteval));
+
+ page = pte_page(pteval);
+ if (!pte_none(pteval) && !is_zero_pfn(pte_pfn(pteval)) &&
+ !PageCompound(page))
+ release_pte_page(page);
+ }
+
+ list_for_each_entry_safe(page, tmp, compound_pagelist, lru) {
+ list_del(&page->lru);
+ release_pte_page(page);
}
}
@@ -546,7 +559,8 @@ static bool is_refcount_suitable(struct
static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
unsigned long address,
- pte_t *pte)
+ pte_t *pte,
+ struct list_head *compound_pagelist)
{
struct page *page = NULL;
pte_t *_pte;
@@ -576,13 +590,21 @@ static int __collapse_huge_page_isolate(
goto out;
}
- /* TODO: teach khugepaged to collapse THP mapped with pte */
+ VM_BUG_ON_PAGE(!PageAnon(page), page);
+
if (PageCompound(page)) {
- result = SCAN_PAGE_COMPOUND;
- goto out;
- }
+ struct page *p;
+ page = compound_head(page);
- VM_BUG_ON_PAGE(!PageAnon(page), page);
+ /*
+ * Check if we have dealt with the compound page
+ * already
+ */
+ list_for_each_entry(p, compound_pagelist, lru) {
+ if (page == p)
+ goto next;
+ }
+ }
/*
* We can do it before isolate_lru_page because the
@@ -611,19 +633,15 @@ static int __collapse_huge_page_isolate(
result = SCAN_PAGE_COUNT;
goto out;
}
- if (pte_write(pteval)) {
- writable = true;
- } else {
- if (PageSwapCache(page) &&
- !reuse_swap_page(page, NULL)) {
- unlock_page(page);
- result = SCAN_SWAP_CACHE_PAGE;
- goto out;
- }
+ if (!pte_write(pteval) && PageSwapCache(page) &&
+ !reuse_swap_page(page, NULL)) {
/*
- * Page is not in the swap cache. It can be collapsed
- * into a THP.
+ * Page is in the swap cache and cannot be re-used.
+ * It cannot be collapsed into a THP.
*/
+ unlock_page(page);
+ result = SCAN_SWAP_CACHE_PAGE;
+ goto out;
}
/*
@@ -635,16 +653,23 @@ static int __collapse_huge_page_isolate(
result = SCAN_DEL_PAGE_LRU;
goto out;
}
- inc_node_page_state(page,
- NR_ISOLATED_ANON + page_is_file_lru(page));
+ mod_node_page_state(page_pgdat(page),
+ NR_ISOLATED_ANON + page_is_file_lru(page),
+ compound_nr(page));
VM_BUG_ON_PAGE(!PageLocked(page), page);
VM_BUG_ON_PAGE(PageLRU(page), page);
+ if (PageCompound(page))
+ list_add_tail(&page->lru, compound_pagelist);
+next:
/* There should be enough young pte to collapse the page */
if (pte_young(pteval) ||
page_is_young(page) || PageReferenced(page) ||
mmu_notifier_test_young(vma->vm_mm, address))
referenced++;
+
+ if (pte_write(pteval))
+ writable = true;
}
if (likely(writable)) {
if (likely(referenced)) {
@@ -658,7 +683,7 @@ static int __collapse_huge_page_isolate(
}
out:
- release_pte_pages(pte, _pte);
+ release_pte_pages(pte, _pte, compound_pagelist);
trace_mm_collapse_huge_page_isolate(page, none_or_zero,
referenced, writable, result);
return 0;
@@ -667,13 +692,14 @@ out:
static void __collapse_huge_page_copy(pte_t *pte, struct page *page,
struct vm_area_struct *vma,
unsigned long address,
- spinlock_t *ptl)
+ spinlock_t *ptl,
+ struct list_head *compound_pagelist)
{
+ struct page *src_page, *tmp;
pte_t *_pte;
for (_pte = pte; _pte < pte + HPAGE_PMD_NR;
_pte++, page++, address += PAGE_SIZE) {
pte_t pteval = *_pte;
- struct page *src_page;
if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
clear_user_highpage(page, address);
@@ -693,7 +719,8 @@ static void __collapse_huge_page_copy(pt
} else {
src_page = pte_page(pteval);
copy_user_highpage(page, src_page, address, vma);
- release_pte_page(src_page);
+ if (!PageCompound(src_page))
+ release_pte_page(src_page);
/*
* ptl mostly unnecessary, but preempt has to
* be disabled to update the per-cpu stats
@@ -710,6 +737,11 @@ static void __collapse_huge_page_copy(pt
free_page_and_swap_cache(src_page);
}
}
+
+ list_for_each_entry_safe(src_page, tmp, compound_pagelist, lru) {
+ list_del(&src_page->lru);
+ release_pte_page(src_page);
+ }
}
static void khugepaged_alloc_sleep(void)
@@ -968,6 +1000,7 @@ static void collapse_huge_page(struct mm
struct page **hpage,
int node, int referenced, int unmapped)
{
+ LIST_HEAD(compound_pagelist);
pmd_t *pmd, _pmd;
pte_t *pte;
pgtable_t pgtable;
@@ -1068,7 +1101,8 @@ static void collapse_huge_page(struct mm
mmu_notifier_invalidate_range_end(&range);
spin_lock(pte_ptl);
- isolated = __collapse_huge_page_isolate(vma, address, pte);
+ isolated = __collapse_huge_page_isolate(vma, address, pte,
+ &compound_pagelist);
spin_unlock(pte_ptl);
if (unlikely(!isolated)) {
@@ -1093,7 +1127,8 @@ static void collapse_huge_page(struct mm
*/
anon_vma_unlock_write(vma->anon_vma);
- __collapse_huge_page_copy(pte, new_page, vma, address, pte_ptl);
+ __collapse_huge_page_copy(pte, new_page, vma, address, pte_ptl,
+ &compound_pagelist);
pte_unmap(pte);
__SetPageUptodate(new_page);
pgtable = pmd_pgtable(_pmd);
@@ -1212,11 +1247,7 @@ static int khugepaged_scan_pmd(struct mm
goto out_unmap;
}
- /* TODO: teach khugepaged to collapse THP mapped with pte */
- if (PageCompound(page)) {
- result = SCAN_PAGE_COMPOUND;
- goto out_unmap;
- }
+ page = compound_head(page);
/*
* Record which node the original page is from and save this
_
Patches currently in -mm which might be from kirill.shutemov@xxxxxxxxxxxxxxx are
khugepaged-add-self-test.patch
khugepaged-do-not-stop-collapse-if-less-than-half-ptes-are-referenced.patch
khugepaged-drain-all-lru-caches-before-scanning-pages.patch
khugepaged-drain-lru-add-pagevec-after-swapin.patch
khugepaged-allow-to-collapse-a-page-shared-across-fork.patch
khugepaged-allow-to-collapse-pte-mapped-compound-pages.patch
thp-change-cow-semantics-for-anon-thp.patch
khugepaged-introduce-max_ptes_shared-tunable.patch
