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 kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by smtp.lore.kernel.org (Postfix) with ESMTP id D6989C001E0 for ; Thu, 10 Aug 2023 11:32:10 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id 599426B0071; Thu, 10 Aug 2023 07:32:10 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id 548B36B0074; Thu, 10 Aug 2023 07:32:10 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 3E9916B0075; Thu, 10 Aug 2023 07:32:10 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0017.hostedemail.com [216.40.44.17]) by kanga.kvack.org (Postfix) with ESMTP id 2C60E6B0071 for ; Thu, 10 Aug 2023 07:32:10 -0400 (EDT) Received: from smtpin16.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay03.hostedemail.com (Postfix) with ESMTP id C8B7CA1057 for ; Thu, 10 Aug 2023 11:32:09 +0000 (UTC) X-FDA: 81107981178.16.4E0C90A Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) by imf17.hostedemail.com (Postfix) with ESMTP id 641E740013 for ; Thu, 10 Aug 2023 11:32:07 +0000 (UTC) Authentication-Results: imf17.hostedemail.com; dkim=pass header.d=redhat.com header.s=mimecast20190719 header.b=gao6qnNY; dmarc=pass (policy=none) header.from=redhat.com; spf=pass (imf17.hostedemail.com: domain of david@redhat.com designates 170.10.129.124 as permitted sender) smtp.mailfrom=david@redhat.com ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1691667127; h=from:from:sender:reply-to:subject:subject:date:date: message-id:message-id:to:to:cc:cc:mime-version:mime-version: content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:dkim-signature; bh=oLK9LGDlC20kprqlDXLL2d+DOs1eVcgviEjDmvnvMC4=; b=atVjBUjjhOTTksmIeBQwwutkhHplcA7/e1doF4EiD2dgWazerr9QPr31w6P344/sj7R4xA spjAPaGNNeyUb3l7xz6DBnhSuHfBv6jmAXxNyOaNLBVJdin+eMpHZfVgEJGmrI3AW8vGeK eGVN05wNCzhdzD6CzQiwdGgwv/fILnk= ARC-Authentication-Results: i=1; imf17.hostedemail.com; dkim=pass header.d=redhat.com header.s=mimecast20190719 header.b=gao6qnNY; dmarc=pass (policy=none) header.from=redhat.com; spf=pass (imf17.hostedemail.com: domain of david@redhat.com designates 170.10.129.124 as permitted sender) smtp.mailfrom=david@redhat.com ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1691667127; a=rsa-sha256; cv=none; b=Ewq3QaH2Kyb24RSEfT8bcS87hazudBP0JTggbB9uR6df1qQUT/Q+RAZE5Z+mAufmT3PEvk WTcY0oW/XPWetiEYUjXYnKRkJHE89wVaOVX0H1Vp81wYMxZBbqsSRWfMmQEIQNTbFLSLLV TNpOs0Qg04NWg84fb0tbw7seRctBK5o= DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1691667126; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=oLK9LGDlC20kprqlDXLL2d+DOs1eVcgviEjDmvnvMC4=; b=gao6qnNY7zeoArfM78Lsh+sncCFM0WSL3xxCeze+i440kKvOuiSbuDX4KSmTaWfs/6eY7n AGoMAVhXwwMEvRg7zugepJ2HaewNdlxv6sEZ7OqaoYOqLPJaW142ly3V8wxQb0HQe+om7r Cenh32lYUtXvAEwwBw9xJOuZ5DFEsF0= Received: from mail-wm1-f69.google.com (mail-wm1-f69.google.com [209.85.128.69]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-691-K_-bGyQyMOu6wv6sh9ApfA-1; Thu, 10 Aug 2023 07:32:05 -0400 X-MC-Unique: K_-bGyQyMOu6wv6sh9ApfA-1 Received: by mail-wm1-f69.google.com with SMTP id 5b1f17b1804b1-3fe4f953070so4606675e9.2 for ; Thu, 10 Aug 2023 04:32:04 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1691667124; x=1692271924; h=content-transfer-encoding:in-reply-to:organization:references:cc:to :from:content-language:subject:user-agent:mime-version:date :message-id:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=oLK9LGDlC20kprqlDXLL2d+DOs1eVcgviEjDmvnvMC4=; b=dYq8V8GdAA0nJnK/y+MqupMXWPh49LcPn6nofbaw5n2VafAxZ9jI27rboo77EosSm2 c7u4jvhNj4bssi1V1Z24P2B+HTKH3SdcllIBBVPabx7NBYfNSQiGh6mehSDGEymlZNgj EdwD/AwBqBVmvJZ+jwH/Io78KB6jObfba675x9cjYUyk1hwMdHXP932ywi1F+rEj9tqb ZMyJ2/DVorxxYVzrGeQaeSfRwKwiKkQRWSM6nErzEFp1r8MiC3Y8TzGIMtmhYmn3TOtH HTu8MIr4KmDUuf2RxzH9QFLe9x9utrgSW37SIqIf4EE082rJAWH10p+t9Q7bk4/4EVZw uXIg== X-Gm-Message-State: AOJu0YwRQhXT2SfPiHXdJoVndphVrLgDtyrr7bg/CzQbQ34DT8bi4Mx5 q9crq4ISRNlWhTFR+8GozD5L3Vafflv4AVh8m+Q0BJllPJdJYcubFJLFQu2X5lneIS69dhFzDB/ BKU9Npv2AzEc= X-Received: by 2002:a1c:721a:0:b0:3fb:9474:54cd with SMTP id n26-20020a1c721a000000b003fb947454cdmr1581072wmc.19.1691667123840; Thu, 10 Aug 2023 04:32:03 -0700 (PDT) X-Google-Smtp-Source: AGHT+IGHPtXUwWy92l9t+AjlXLiwltHlHkd1BeIuuLATg2YdyzJZlg5cxODixzZtEFV80fVLLXGeWA== X-Received: by 2002:a1c:721a:0:b0:3fb:9474:54cd with SMTP id n26-20020a1c721a000000b003fb947454cdmr1581050wmc.19.1691667123429; Thu, 10 Aug 2023 04:32:03 -0700 (PDT) Received: from ?IPV6:2a09:80c0:192:0:5dac:bf3d:c41:c3e7? ([2a09:80c0:192:0:5dac:bf3d:c41:c3e7]) by smtp.gmail.com with ESMTPSA id y1-20020a05600c364100b003fbb06af219sm1840009wmq.32.2023.08.10.04.32.02 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Thu, 10 Aug 2023 04:32:03 -0700 (PDT) Message-ID: <08f066e9-f137-f96c-df63-d5ccc8138470@redhat.com> Date: Thu, 10 Aug 2023 13:32:02 +0200 MIME-Version: 1.0 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101 Thunderbird/102.13.0 Subject: Re: [PATCH mm-unstable v1] mm: add a total mapcount for large folios From: David Hildenbrand To: Ryan Roberts , linux-kernel@vger.kernel.org Cc: linux-mm@kvack.org, linux-doc@vger.kernel.org, Andrew Morton , Jonathan Corbet , Mike Kravetz , Hugh Dickins , "Matthew Wilcox (Oracle)" , Yin Fengwei , Yang Shi , Zi Yan References: <20230809083256.699513-1-david@redhat.com> <181fcc79-b1c6-412f-9ca1-d1f21ef33e32@arm.com> <60b5b2a2-1d1d-661c-d61e-855178fff44d@redhat.com> Organization: Red Hat In-Reply-To: X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Language: en-US Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit X-Rspam-User: X-Stat-Signature: i98psa6bk9sjhmuks69m7zx4ctgwhft1 X-Rspamd-Server: rspam07 X-Rspamd-Queue-Id: 641E740013 X-HE-Tag: 1691667127-29552 X-HE-Meta: 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 DhM2itDq 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 X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: On 10.08.23 13:27, David Hildenbrand wrote: > On 10.08.23 13:14, David Hildenbrand wrote: >> On 09.08.23 21:17, David Hildenbrand wrote: >>> On 09.08.23 21:07, Ryan Roberts wrote: >>>> On 09/08/2023 09:32, David Hildenbrand wrote: >>>>> Let's track the total mapcount for all large folios in the first subpage. >>>>> >>>>> The total mapcount is what we actually want to know in folio_mapcount() >>>>> and it is also sufficient for implementing folio_mapped(). This also >>>>> gets rid of any "raceiness" concerns as expressed in >>>>> folio_total_mapcount(). >>>>> >>>>> With sub-PMD THP becoming more important and things looking promising >>>>> that we will soon get support for such anon THP, we want to avoid looping >>>>> over all pages of a folio just to calculate the total mapcount. Further, >>>>> we may soon want to use the total mapcount in other context more >>>>> frequently, so prepare for reading it efficiently and atomically. >>>>> >>>>> Make room for the total mapcount in page[1] of the folio by moving >>>>> _nr_pages_mapped to page[2] of the folio: it is not applicable to hugetlb >>>>> -- and with the total mapcount in place probably also not desirable even >>>>> if PMD-mappable hugetlb pages could get PTE-mapped at some point -- so we >>>>> can overlay the hugetlb fields. >>>>> >>>>> Note that we currently don't expect any order-1 compound pages / THP in >>>>> rmap code, and that such support is not planned. If ever desired, we could >>>>> move the compound mapcount to another page, because it only applies to >>>>> PMD-mappable folios that are definitely larger. Let's avoid consuming >>>>> more space elsewhere for now -- we might need more space for a different >>>>> purpose in some subpages soon. >>>>> >>>>> Maintain the total mapcount also for hugetlb pages. Use the total mapcount >>>>> to implement folio_mapcount(), total_mapcount(), folio_mapped() and >>>>> page_mapped(). >>>>> >>>>> We can now get rid of folio_total_mapcount() and >>>>> folio_large_is_mapped(), by just inlining reading of the total mapcount. >>>>> >>>>> _nr_pages_mapped is now only used in rmap code, so not accidentially >>>>> externally where it might be used on arbitrary order-1 pages. The remaining >>>>> usage is: >>>>> >>>>> (1) Detect how to adjust stats: NR_ANON_MAPPED and NR_FILE_MAPPED >>>>> -> If we would account the total folio as mapped when mapping a >>>>> page (based on the total mapcount), we could remove that usage. >>>>> >>>>> (2) Detect when to add a folio to the deferred split queue >>>>> -> If we would apply a different heuristic, or scan using the rmap on >>>>> the memory reclaim path for partially mapped anon folios to >>>>> split them, we could remove that usage as well. >>>>> >>>>> So maybe, we can simplify things in the future and remove >>>>> _nr_pages_mapped. For now, leave these things as they are, they need more >>>>> thought. Hugh really did a nice job implementing that precise tracking >>>>> after all. >>>>> >>>>> Note: Not adding order-1 sanity checks to the file_rmap functions for >>>>> now. For anon pages, they are certainly not required right now. >>>>> >>>>> Cc: Andrew Morton >>>>> Cc: Jonathan Corbet >>>>> Cc: Mike Kravetz >>>>> Cc: Hugh Dickins >>>>> Cc: "Matthew Wilcox (Oracle)" >>>>> Cc: Ryan Roberts >>>>> Cc: Yin Fengwei >>>>> Cc: Yang Shi >>>>> Cc: Zi Yan >>>>> Signed-off-by: David Hildenbrand >>>> >>>> Other than the nits and query on zeroing _total_mapcount below, LGTM. If zeroing >>>> is correct: >>>> >>>> Reviewed-by: Ryan Roberts >>> >>> Thanks for the review! >>> >>> [...] >>> >>>>> >>>>> static inline int total_mapcount(struct page *page) >>>> >>>> nit: couldn't total_mapcount() just be implemented as a wrapper around >>>> folio_mapcount()? What's the benefit of PageCompound() check instead of just >>>> getting the folio and checking if it's large? i.e: >>> >>> Good point, let me take a look tomorrow if the compiler can optimize in >>> both cases equally well. >> >> I checked by adjusting total_mapcount(): >> >> Before: >> >> if (PageTransHuge(page) && total_mapcount(page) > 1) >> ffffffff81411931: 4c 89 e7 mov %r12,%rdi >> ffffffff81411934: e8 f7 b1 ff ff call ffffffff8140cb30 >> ffffffff81411939: 85 c0 test %eax,%eax >> ffffffff8141193b: 74 29 je ffffffff81411966 >> ffffffff8141193d: 49 8b 04 24 mov (%r12),%rax >> return test_bit(PG_head, &page->flags) || >> ffffffff81411941: a9 00 00 01 00 test $0x10000,%eax >> ffffffff81411946: 0f 85 1f 01 00 00 jne ffffffff81411a6b >> READ_ONCE(page->compound_head) & 1; >> ffffffff8141194c: 49 8b 44 24 08 mov 0x8(%r12),%rax >> return test_bit(PG_head, &page->flags) || >> ffffffff81411951: a8 01 test $0x1,%al >> ffffffff81411953: 0f 85 12 01 00 00 jne ffffffff81411a6b >> ffffffff81411959: 41 8b 44 24 30 mov 0x30(%r12),%eax >> return atomic_read(&page->_mapcount) + 1; >> ffffffff8141195e: 83 c0 01 add $0x1,%eax >> ffffffff81411961: 83 f8 01 cmp $0x1,%eax >> ffffffff81411964: 7f 77 jg ffffffff814119dd >> >> So a total of 10 instructions after handling the mov/call/test/je for PageTransHuge(). >> >> After: >> >> if (PageTransHuge(page) && total_mapcount(page) > 1) >> ffffffff81411931: 4c 89 e7 mov %r12,%rdi >> ffffffff81411934: e8 f7 b1 ff ff call ffffffff8140cb30 >> ffffffff81411939: 85 c0 test %eax,%eax >> ffffffff8141193b: 74 2f je ffffffff8141196c >> unsigned long head = READ_ONCE(page->compound_head); >> ffffffff8141193d: 49 8b 44 24 08 mov 0x8(%r12),%rax >> if (unlikely(head & 1)) >> ffffffff81411942: a8 01 test $0x1,%al >> ffffffff81411944: 0f 85 fc 05 00 00 jne ffffffff81411f46 >> ffffffff8141194a: 0f 1f 44 00 00 nopl 0x0(%rax,%rax,1) >> return page; >> ffffffff8141194f: 4c 89 e0 mov %r12,%rax >> ffffffff81411952: 48 8b 10 mov (%rax),%rdx >> if (likely(!folio_test_large(folio))) >> ffffffff81411955: f7 c2 00 00 01 00 test $0x10000,%edx >> ffffffff8141195b: 0f 85 da 05 00 00 jne ffffffff81411f3b >> ffffffff81411961: 8b 40 30 mov 0x30(%rax),%eax >> return atomic_read(&folio->_mapcount) + 1; >> ffffffff81411964: 83 c0 01 add $0x1,%eax >> ffffffff81411967: 83 f8 01 cmp $0x1,%eax >> ffffffff8141196a: 7f 77 jg ffffffff814119e3 >> >> So a total of 11 (+ 1 NOP) instructions after handling the mov/call/test/je for PageTransHuge(). >> >> Essentially one more MOV instruction. >> >> I guess nobody really cares :) >> > > Also, let's simply do: > > static inline bool folio_mapped(struct folio *folio) > { > folio_mapcount(folio) > 1; > 0, obviously :) -- Cheers, David / dhildenb