From: Hugh Dickins Date: Sat, 13 Dec 2014 00:56:19 +0000 (-0800) Subject: mm: unmapped page migration avoid unmap+remap overhead X-Git-Tag: firefly_0821_release~176^2~2692^2~54 X-Git-Url: http://demsky.eecs.uci.edu/git/?a=commitdiff_plain;h=2ebba6b7e1d98724d266ae048d8af4f7ca95cafd;p=firefly-linux-kernel-4.4.55.git mm: unmapped page migration avoid unmap+remap overhead Page migration's __unmap_and_move(), and rmap's try_to_unmap(), were created for use on pages almost certainly mapped into userspace. But nowadays compaction often applies them to unmapped page cache pages: which may exacerbate contention on i_mmap_rwsem quite unnecessarily, since try_to_unmap_file() makes no preliminary page_mapped() check. Now check page_mapped() in __unmap_and_move(); and avoid repeating the same overhead in rmap_walk_file() - don't remove_migration_ptes() when we never inserted any. (The PageAnon(page) comment blocks now look even sillier than before, but clean that up on some other occasion. And note in passing that try_to_unmap_one() does not use a migration entry when PageSwapCache, so remove_migration_ptes() will then not update that swap entry to newpage pte: not a big deal, but something else to clean up later.) Davidlohr remarked in "mm,fs: introduce helpers around the i_mmap_mutex" conversion to i_mmap_rwsem, that "The biggest winner of these changes is migration": a part of the reason might be all of that unnecessary taking of i_mmap_mutex in page migration; and it's rather a shame that I didn't get around to sending this patch in before his - this one is much less useful after Davidlohr's conversion to rwsem, but still good. Signed-off-by: Hugh Dickins Cc: Davidlohr Bueso Cc: Rik van Riel Cc: Mel Gorman Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- diff --git a/mm/migrate.c b/mm/migrate.c index 01439953abf5..253474c22239 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -746,7 +746,7 @@ static int fallback_migrate_page(struct address_space *mapping, * MIGRATEPAGE_SUCCESS - success */ static int move_to_new_page(struct page *newpage, struct page *page, - int remap_swapcache, enum migrate_mode mode) + int page_was_mapped, enum migrate_mode mode) { struct address_space *mapping; int rc; @@ -784,7 +784,7 @@ static int move_to_new_page(struct page *newpage, struct page *page, newpage->mapping = NULL; } else { mem_cgroup_migrate(page, newpage, false); - if (remap_swapcache) + if (page_was_mapped) remove_migration_ptes(page, newpage); page->mapping = NULL; } @@ -798,7 +798,7 @@ static int __unmap_and_move(struct page *page, struct page *newpage, int force, enum migrate_mode mode) { int rc = -EAGAIN; - int remap_swapcache = 1; + int page_was_mapped = 0; struct anon_vma *anon_vma = NULL; if (!trylock_page(page)) { @@ -870,7 +870,6 @@ static int __unmap_and_move(struct page *page, struct page *newpage, * migrated but are not remapped when migration * completes */ - remap_swapcache = 0; } else { goto out_unlock; } @@ -910,13 +909,17 @@ static int __unmap_and_move(struct page *page, struct page *newpage, } /* Establish migration ptes or remove ptes */ - try_to_unmap(page, TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS); + if (page_mapped(page)) { + try_to_unmap(page, + TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS); + page_was_mapped = 1; + } skip_unmap: if (!page_mapped(page)) - rc = move_to_new_page(newpage, page, remap_swapcache, mode); + rc = move_to_new_page(newpage, page, page_was_mapped, mode); - if (rc && remap_swapcache) + if (rc && page_was_mapped) remove_migration_ptes(page, page); /* Drop an anon_vma reference if we took one */ @@ -1017,6 +1020,7 @@ static int unmap_and_move_huge_page(new_page_t get_new_page, { int rc = 0; int *result = NULL; + int page_was_mapped = 0; struct page *new_hpage; struct anon_vma *anon_vma = NULL; @@ -1047,12 +1051,16 @@ static int unmap_and_move_huge_page(new_page_t get_new_page, if (PageAnon(hpage)) anon_vma = page_get_anon_vma(hpage); - try_to_unmap(hpage, TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS); + if (page_mapped(hpage)) { + try_to_unmap(hpage, + TTU_MIGRATION|TTU_IGNORE_MLOCK|TTU_IGNORE_ACCESS); + page_was_mapped = 1; + } if (!page_mapped(hpage)) - rc = move_to_new_page(new_hpage, hpage, 1, mode); + rc = move_to_new_page(new_hpage, hpage, page_was_mapped, mode); - if (rc != MIGRATEPAGE_SUCCESS) + if (rc != MIGRATEPAGE_SUCCESS && page_was_mapped) remove_migration_ptes(hpage, hpage); if (anon_vma)