From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 8D553C46467 for ; Tue, 10 Jan 2023 23:54:29 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S234508AbjAJXy1 (ORCPT ); Tue, 10 Jan 2023 18:54:27 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:56458 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S233003AbjAJXyZ (ORCPT ); Tue, 10 Jan 2023 18:54:25 -0500 Received: from ams.source.kernel.org (ams.source.kernel.org [145.40.68.75]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id CB47C55663 for ; Tue, 10 Jan 2023 15:54:23 -0800 (PST) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ams.source.kernel.org (Postfix) with ESMTPS id 7ED19B80F92 for ; Tue, 10 Jan 2023 23:54:22 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 2AC89C433D2; Tue, 10 Jan 2023 23:54:21 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linux-foundation.org; s=korg; t=1673394861; bh=8MQdPbFctob4iSpbIa+4ZrgTfmFH6PArNToBdiolf2Q=; h=Date:To:From:Subject:From; b=JsUlMQ9d1oq1Op+bl9GLN+/brMok4dcFzUl7jONbSUC7a8ynAgjmX3w3ERV9DPym5 +CGcR2J3YaHOMgxMzDCgvfpwGsAdvU5pROf+fqIFgEck1STq95KKj/DJCjfRjq9pxi oTpTKlJJd8rLZ2lZaotGFiIO2tLrwqIse0GL33ig= Date: Tue, 10 Jan 2023 15:54:20 -0800 To: mm-commits@vger.kernel.org, ziy@nvidia.com, xhao@linux.alibaba.com, willy@infradead.org, shy828301@gmail.com, osalvador@suse.de, bharata@amd.com, baolin.wang@linux.alibaba.com, apopple@nvidia.com, ying.huang@intel.com, akpm@linux-foundation.org From: Andrew Morton Subject: + migrate_pages-batch-flushing-tlb.patch added to mm-unstable branch Message-Id: <20230110235421.2AC89C433D2@smtp.kernel.org> Precedence: bulk Reply-To: linux-kernel@vger.kernel.org List-ID: X-Mailing-List: mm-commits@vger.kernel.org The patch titled Subject: migrate_pages: batch flushing TLB has been added to the -mm mm-unstable branch. Its filename is migrate_pages-batch-flushing-tlb.patch This patch will shortly appear at https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/migrate_pages-batch-flushing-tlb.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: Huang Ying Subject: migrate_pages: batch flushing TLB Date: Tue, 10 Jan 2023 15:53:26 +0800 The TLB flushing will cost quite some CPU cycles during the folio migration in some situations. For example, when migrate a folio of a process with multiple active threads that run on multiple CPUs. After batching the _unmap and _move in migrate_pages(), the TLB flushing can be batched easily with the existing TLB flush batching mechanism. This patch implements that. We use the following test case to test the patch. On a 2-socket Intel server, - Run pmbench memory accessing benchmark - Run `migratepages` to migrate pages of pmbench between node 0 and node 1 back and forth. With the patch, the TLB flushing IPI reduces 99.1% during the test and the number of pages migrated successfully per second increases 291.7%. NOTE: TLB flushing is batched only for normal folios, not for THP folios. Because the overhead of TLB flushing for THP folios is much lower than that for normal folios (about 1/512 on x86 platform). Link: https://lkml.kernel.org/r/20230110075327.590514-9-ying.huang@intel.com Signed-off-by: "Huang, Ying" Cc: Zi Yan Cc: Yang Shi Cc: Baolin Wang Cc: Oscar Salvador Cc: Matthew Wilcox Cc: Bharata B Rao Cc: Alistair Popple Cc: haoxin Signed-off-by: Andrew Morton --- mm/migrate.c | 4 +++- mm/rmap.c | 20 +++++++++++++++++--- 2 files changed, 20 insertions(+), 4 deletions(-) --- a/mm/migrate.c~migrate_pages-batch-flushing-tlb +++ a/mm/migrate.c @@ -1247,7 +1247,7 @@ static int migrate_folio_unmap(new_page_ /* Establish migration ptes */ VM_BUG_ON_FOLIO(folio_test_anon(src) && !folio_test_ksm(src) && !anon_vma, src); - try_to_migrate(src, 0); + try_to_migrate(src, TTU_BATCH_FLUSH); page_was_mapped = 1; } @@ -1790,6 +1790,8 @@ retry: stats->nr_thp_failed += thp_retry; stats->nr_failed_pages += nr_retry_pages; move: + try_to_unmap_flush(); + retry = 1; for (pass = 0; pass < NR_MAX_MIGRATE_PAGES_RETRY && (retry || large_retry); --- a/mm/rmap.c~migrate_pages-batch-flushing-tlb +++ a/mm/rmap.c @@ -1960,7 +1960,21 @@ static bool try_to_migrate_one(struct fo } else { flush_cache_page(vma, address, pte_pfn(*pvmw.pte)); /* Nuke the page table entry. */ - pteval = ptep_clear_flush(vma, address, pvmw.pte); + if (should_defer_flush(mm, flags)) { + /* + * We clear the PTE but do not flush so potentially + * a remote CPU could still be writing to the folio. + * If the entry was previously clean then the + * architecture must guarantee that a clear->dirty + * transition on a cached TLB entry is written through + * and traps if the PTE is unmapped. + */ + pteval = ptep_get_and_clear(mm, address, pvmw.pte); + + set_tlb_ubc_flush_pending(mm, pte_dirty(pteval)); + } else { + pteval = ptep_clear_flush(vma, address, pvmw.pte); + } } /* Set the dirty flag on the folio now the pte is gone. */ @@ -2132,10 +2146,10 @@ void try_to_migrate(struct folio *folio, /* * Migration always ignores mlock and only supports TTU_RMAP_LOCKED and - * TTU_SPLIT_HUGE_PMD and TTU_SYNC flags. + * TTU_SPLIT_HUGE_PMD, TTU_SYNC, and TTU_BATCH_FLUSH flags. */ if (WARN_ON_ONCE(flags & ~(TTU_RMAP_LOCKED | TTU_SPLIT_HUGE_PMD | - TTU_SYNC))) + TTU_SYNC | TTU_BATCH_FLUSH))) return; if (folio_is_zone_device(folio) && _ Patches currently in -mm which might be from ying.huang@intel.com are migrate_pages-organize-stats-with-struct-migrate_pages_stats.patch migrate_pages-separate-hugetlb-folios-migration.patch migrate_pages-restrict-number-of-pages-to-migrate-in-batch.patch migrate_pages-split-unmap_and_move-to-_unmap-and-_move.patch migrate_pages-batch-_unmap-and-_move.patch migrate_pages-move-migrate_folio_unmap.patch migrate_pages-share-more-code-between-_unmap-and-_move.patch migrate_pages-batch-flushing-tlb.patch migrate_pages-move-thp-hugetlb-migration-support-check-to-simplify-code.patch