From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 0FAD215359A for ; Thu, 2 May 2024 16:11:54 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=10.30.226.201 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1714666315; cv=none; b=UqZbiyB5zAdAeccWWLrlBiQXXBK5dZ7CHY0V2YaHLDb3tpW8LD6jakgvFcnABgPcbLXKeng9ymS1G0TKoZ13+3IiSMNIVzJZgaymlpO+oG/4opwrv+Wx43/R+4smmj28fgafkqYGW2buPhvIM4KgVnZcmduEuMvjJrV2AgD1p1M= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1714666315; c=relaxed/simple; bh=/T+pkM8k1medLGOeJon8F6uhahjA6AN3/18A3lSel40=; h=Date:To:From:Subject:Message-Id; b=b9RIKyHhbxDJaOdUPDveNW0WX9UKTqrsPNNuznotnSN3yEq0qE1vYtclMxcGo0QeswwMOjGfAiyHyvvszGLru/pHDkPRX8TcXP1idifcpcBHi00XkjvrVeasGiQr5U0kBhJzlf4ooQmOSYPh+h7g5s7K7X4NDQoidVKurzT6WKI= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux-foundation.org header.i=@linux-foundation.org header.b=ic6FO0c8; arc=none smtp.client-ip=10.30.226.201 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux-foundation.org header.i=@linux-foundation.org header.b="ic6FO0c8" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 58608C32789; Thu, 2 May 2024 16:11:54 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linux-foundation.org; s=korg; t=1714666314; bh=/T+pkM8k1medLGOeJon8F6uhahjA6AN3/18A3lSel40=; h=Date:To:From:Subject:From; b=ic6FO0c8B8/gduLxawCcgRAKQSf/5v1+MPmGqdm7tdLoYYbfKEMuMVm74HIx7CpC9 QZAvGdwpKBM/Os94AiXRn5C8rZbaceP1ypiAyzqSrGF7u3b0EMQykJ92uqlBkb0/SR F0co0EpWwvFVReAfuuRVpXRbjk7G6TiQbImW2DjA= Date: Thu, 02 May 2024 09:11:53 -0700 To: mm-commits@vger.kernel.org,zokeefe@google.com,ziy@nvidia.com,xiehuan09@gmail.com,willy@infradead.org,wangkefeng.wang@huawei.com,songmuchun@bytedance.com,sj@kernel.org,shy828301@gmail.com,ryan.roberts@arm.com,peterx@redhat.com,minchan@kernel.org,mhocko@suse.com,maskray@google.com,fengwei.yin@intel.com,david@redhat.com,21cnbao@gmail.com,ioworker0@gmail.com,akpm@linux-foundation.org From: Andrew Morton Subject: + mm-vmscan-avoid-split-lazyfree-thp-during-shrink_folio_list.patch added to mm-unstable branch Message-Id: <20240502161154.58608C32789@smtp.kernel.org> Precedence: bulk X-Mailing-List: mm-commits@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: The patch titled Subject: mm/vmscan: avoid split lazyfree THP during shrink_folio_list() has been added to the -mm mm-unstable branch. Its filename is mm-vmscan-avoid-split-lazyfree-thp-during-shrink_folio_list.patch This patch will shortly appear at https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/mm-vmscan-avoid-split-lazyfree-thp-during-shrink_folio_list.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: Lance Yang Subject: mm/vmscan: avoid split lazyfree THP during shrink_folio_list() Date: Wed, 1 May 2024 12:27:00 +0800 When the user no longer requires the pages, they would use madvise(MADV_FREE) to mark the pages as lazy free. Subsequently, they typically would not re-write to that memory again. During memory reclaim, if we detect that the large folio and its PMD are both still marked as clean and there are no unexpected references (such as GUP), so we can just discard the memory lazily, improving the efficiency of memory reclamation in this case. On an Intel i5 CPU, reclaiming 1GiB of lazyfree THPs using mem_cgroup_force_empty() results in the following runtimes in seconds (shorter is better): -------------------------------------------- | Old | New | Change | -------------------------------------------- | 0.683426 | 0.049197 | -92.80% | -------------------------------------------- Link: https://lkml.kernel.org/r/20240501042700.83974-4-ioworker0@gmail.com Signed-off-by: Lance Yang Suggested-by: Zi Yan Suggested-by: David Hildenbrand Tested-by: SeongJae Park Cc: Barry Song <21cnbao@gmail.com> Cc: Fangrui Song Cc: Jeff Xie Cc: Kefeng Wang Cc: Matthew Wilcox (Oracle) Cc: Michal Hocko Cc: Minchan Kim Cc: Muchun Song Cc: Peter Xu Cc: Ryan Roberts Cc: Yang Shi Cc: Yin Fengwei Cc: Zach O'Keefe Signed-off-by: Andrew Morton --- include/linux/huge_mm.h | 9 ++++ mm/huge_memory.c | 73 ++++++++++++++++++++++++++++++++++++++ mm/rmap.c | 3 + 3 files changed, 85 insertions(+) --- a/include/linux/huge_mm.h~mm-vmscan-avoid-split-lazyfree-thp-during-shrink_folio_list +++ a/include/linux/huge_mm.h @@ -411,6 +411,8 @@ static inline bool thp_migration_support void split_huge_pmd_locked(struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, bool freeze, struct folio *folio); +bool unmap_huge_pmd_locked(struct vm_area_struct *vma, unsigned long addr, + pmd_t *pmdp, struct folio *folio); static inline void align_huge_pmd_range(struct vm_area_struct *vma, unsigned long *start, @@ -492,6 +494,13 @@ static inline void align_huge_pmd_range( unsigned long *start, unsigned long *end) {} +static inline bool unmap_huge_pmd_locked(struct vm_area_struct *vma, + unsigned long addr, pmd_t *pmdp, + struct folio *folio) +{ + return false; +} + #define split_huge_pud(__vma, __pmd, __address) \ do { } while (0) --- a/mm/huge_memory.c~mm-vmscan-avoid-split-lazyfree-thp-during-shrink_folio_list +++ a/mm/huge_memory.c @@ -2689,6 +2689,79 @@ static void unmap_folio(struct folio *fo try_to_unmap_flush(); } +static bool __discard_trans_pmd_locked(struct vm_area_struct *vma, + unsigned long addr, pmd_t *pmdp, + struct folio *folio) +{ + struct mm_struct *mm = vma->vm_mm; + int ref_count, map_count; + pmd_t orig_pmd = *pmdp; + struct mmu_gather tlb; + struct page *page; + + if (pmd_dirty(orig_pmd) || folio_test_dirty(folio)) + return false; + if (unlikely(!pmd_present(orig_pmd) || !pmd_trans_huge(orig_pmd))) + return false; + + page = pmd_page(orig_pmd); + if (unlikely(page_folio(page) != folio)) + return false; + + tlb_gather_mmu(&tlb, mm); + orig_pmd = pmdp_huge_get_and_clear(mm, addr, pmdp); + tlb_remove_pmd_tlb_entry(&tlb, pmdp, addr); + + /* + * Syncing against concurrent GUP-fast: + * - clear PMD; barrier; read refcount + * - inc refcount; barrier; read PMD + */ + smp_mb(); + + ref_count = folio_ref_count(folio); + map_count = folio_mapcount(folio); + + /* + * Order reads for folio refcount and dirty flag + * (see comments in __remove_mapping()). + */ + smp_rmb(); + + /* + * If the PMD or folio is redirtied at this point, or if there are + * unexpected references, we will give up to discard this folio + * and remap it. + * + * The only folio refs must be one from isolation plus the rmap(s). + */ + if (ref_count != map_count + 1 || folio_test_dirty(folio) || + pmd_dirty(orig_pmd)) { + set_pmd_at(mm, addr, pmdp, orig_pmd); + return false; + } + + folio_remove_rmap_pmd(folio, page, vma); + zap_deposited_table(mm, pmdp); + add_mm_counter(mm, MM_ANONPAGES, -HPAGE_PMD_NR); + folio_put(folio); + + return true; +} + +bool unmap_huge_pmd_locked(struct vm_area_struct *vma, unsigned long addr, + pmd_t *pmdp, struct folio *folio) +{ + VM_WARN_ON_FOLIO(!folio_test_pmd_mappable(folio), folio); + VM_WARN_ON_FOLIO(!folio_test_locked(folio), folio); + VM_WARN_ON_ONCE(!IS_ALIGNED(addr, HPAGE_PMD_SIZE)); + + if (folio_test_anon(folio) && !folio_test_swapbacked(folio)) + return __discard_trans_pmd_locked(vma, addr, pmdp, folio); + + return false; +} + static void remap_page(struct folio *folio, unsigned long nr) { int i = 0; --- a/mm/rmap.c~mm-vmscan-avoid-split-lazyfree-thp-during-shrink_folio_list +++ a/mm/rmap.c @@ -1676,6 +1676,9 @@ static bool try_to_unmap_one(struct foli } if (!pvmw.pte && (flags & TTU_SPLIT_HUGE_PMD)) { + if (unmap_huge_pmd_locked(vma, range.start, pvmw.pmd, + folio)) + goto walk_done; /* * We temporarily have to drop the PTL and start once * again from that now-PTE-mapped page table. _ Patches currently in -mm which might be from ioworker0@gmail.com are mm-madvise-introduce-clear_young_dirty_ptes-batch-helper.patch mm-arm64-override-clear_young_dirty_ptes-batch-helper.patch mm-memory-add-any_dirty-optional-pointer-to-folio_pte_batch.patch mm-madvise-optimize-lazyfreeing-with-mthp-in-madvise_free.patch mm-rmap-remove-duplicated-exit-code-in-pagewalk-loop.patch mm-rmap-integrate-pmd-mapped-folio-splitting-into-pagewalk-loop.patch mm-vmscan-avoid-split-lazyfree-thp-during-shrink_folio_list.patch