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 X-Spam-Level: X-Spam-Status: No, score=-11.3 required=3.0 tests=BAYES_00,DKIM_ADSP_ALL, DKIM_INVALID,DKIM_SIGNED,HEADER_FROM_DIFFERENT_DOMAINS,HTML_MESSAGE, INCLUDES_CR_TRAILER,INCLUDES_PATCH,MAILING_LIST_MULTI,NICE_REPLY_A, SPF_HELO_NONE,SPF_PASS,USER_AGENT_SANE_1 autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 4B6EEC2B9F4 for ; Fri, 25 Jun 2021 16:27:08 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id AD85F6161C for ; Fri, 25 Jun 2021 16:27:07 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org AD85F6161C Authentication-Results: mail.kernel.org; dmarc=fail (p=quarantine dis=none) header.from=amazon.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 9A7456B0036; Fri, 25 Jun 2021 12:27:06 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id 956C86B005D; Fri, 25 Jun 2021 12:27:06 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 783CA8D0001; Fri, 25 Jun 2021 12:27:06 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0222.hostedemail.com [216.40.44.222]) by kanga.kvack.org (Postfix) with ESMTP id 1EE1B6B0036 for ; Fri, 25 Jun 2021 12:27:06 -0400 (EDT) Received: from smtpin24.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay04.hostedemail.com (Postfix) with ESMTP id 380161356F for ; Fri, 25 Jun 2021 16:27:06 +0000 (UTC) X-FDA: 78292775652.24.CCB1D0C Received: from smtp-fw-33001.amazon.com (smtp-fw-33001.amazon.com [207.171.190.10]) by imf16.hostedemail.com (Postfix) with ESMTP id 5C0488019365 for ; Fri, 25 Jun 2021 16:27:05 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=amazon.com; i=@amazon.com; q=dns/txt; s=amazon201209; t=1624638425; x=1656174425; h=subject:to:cc:references:from:message-id:date: mime-version:in-reply-to; bh=ppnugSqzVms/4dfWdf+aJ0KIQFcvAXbfh0156L0hmX0=; b=fmx4mHXakUx9wH/rWNuhFgwmnsFi4CVQCTUoNgmTr/OPUskVjUKwdmYG ec853hR+Z4n/VkEtmdlyqvXweYVRyg1EyrXGl7334BFNzIdcqHHfA3i77 eBb5tItDUqLa81MlAOhg6oZ4Z4fDua9uXnWB+2GJeazTt9PFGIGLA2CnF 0=; X-IronPort-AV: E=Sophos;i="5.83,299,1616457600"; d="scan'208,217";a="133002763" Received: from pdx4-co-svc-p1-lb2-vlan2.amazon.com (HELO email-inbound-relay-2a-6e2fc477.us-west-2.amazon.com) ([10.25.36.210]) by smtp-border-fw-33001.sea14.amazon.com with ESMTP; 25 Jun 2021 16:26:57 +0000 Received: from EX13D04EUB002.ant.amazon.com (pdx1-ws-svc-p6-lb9-vlan2.pdx.amazon.com [10.236.137.194]) by email-inbound-relay-2a-6e2fc477.us-west-2.amazon.com (Postfix) with ESMTPS id 08C10A15CA; Fri, 25 Jun 2021 16:26:54 +0000 (UTC) Received: from 147dda5e91c0.ant.amazon.com (10.43.162.118) by EX13D04EUB002.ant.amazon.com (10.43.166.51) with Microsoft SMTP Server (TLS) id 15.0.1497.18; Fri, 25 Jun 2021 16:26:38 +0000 Subject: Re: [PATCH v23 0/9] Free some vmemmap pages of HugeTLB page To: Muchun Song , , , , , , , , , , , , , , , , , , , , , , , , , , , CC: , , , , , , References: <20210510030027.56044-1-songmuchun@bytedance.com> From: "Wali, Rohit" Message-ID: <234d1a38-28a2-8438-30f5-101361a69ea4@amazon.com> Date: Fri, 25 Jun 2021 21:56:33 +0530 User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:78.0) Gecko/20100101 Thunderbird/78.9.0 MIME-Version: 1.0 In-Reply-To: <20210510030027.56044-1-songmuchun@bytedance.com> Content-Type: multipart/alternative; boundary="------------A189D1B5D083A421D2F1C052" Content-Language: en-US X-Originating-IP: [10.43.162.118] X-ClientProxiedBy: EX13D38UWC004.ant.amazon.com (10.43.162.204) To EX13D04EUB002.ant.amazon.com (10.43.166.51) Authentication-Results: imf16.hostedemail.com; dkim=pass header.d=amazon.com header.s=amazon201209 header.b=fmx4mHXa; dmarc=pass (policy=quarantine) header.from=amazon.com; spf=pass (imf16.hostedemail.com: domain of "prvs=803bd12da=rohiwali@amazon.com" designates 207.171.190.10 as permitted sender) smtp.mailfrom="prvs=803bd12da=rohiwali@amazon.com" X-Rspamd-Server: rspam02 X-Stat-Signature: 4gws1pyxswbxhesfzdysy45kakcabddj X-Rspamd-Queue-Id: 5C0488019365 X-HE-Tag: 1624638425-327490 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: --------------A189D1B5D083A421D2F1C052 Content-Type: text/plain; charset="utf-8"; format=flowed Content-Transfer-Encoding: quoted-printable Tested-by: Rohit Wali I have tested the patch set and the changes work as expected. Test steps - Had an instance with 64G memory and allocated 50G in=20 hugepages of size 2M and 1G. (Shamelessly copying the result format used by Chen Huang for easier=20 co-relation :) ) -------------------------------------------------------------------------= ------------------------------| 2M page | = 1G page | -----------------------|--------------------------|----------------------= ---|--------------------------| enable | disable | enable =C2=A0= | disable | -----------------------|--------------------------|----------------------= ---|--------------------------| total | used | free | total | used | free | total | used | fre= e | total | used | free | 64,312 | 51,576 |12,574| 63,514 | 51,368 | 11,956| 64,312 | 51,361 | 12,= 789| 63,514 | 51,362 | 11,962| -------------------------------------------------------------------------= ------------------------------| One thing to be aware of, though, is that the gain in memory from=20 freeing page struct after hugepage reservation shows up in 'used' memory=20 after unreserving hugepages As an example, with using 50G worth of hugepages of size 1G. Before hugepage reservation |-------------------------| | total | used | free | | 63,512 | 206 | 63,083| |-------------------------| After reserving 50 pages of 1G |----------------------------| | total | used | free | | 64,312 | 51,361 | 12,789 | |----------------------------| After unreserving hugepages in previous step |-------------------------| | total | used | free | | 64,312 | 966 | 63,184| |-------------------------| On 10/05/21 8:30 am, Muchun Song wrote: > Hi, > > Since Mike's patches (make hugetlb put_page safe for all calling contex= ts[1]) > applied into the next-20210412. We can move forward on this patch serie= s now. > > This patch series will free some vmemmap pages(struct page structures) > associated with each HugeTLB page when preallocated to save memory. > > In order to reduce the difficulty of the first version of code review. > >From this version, we disable PMD/huge page mapping of vmemmap if this > feature was enabled. This acutely eliminates a bunch of the complex cod= e > doing page table manipulation. When this patch series is solid, we cam = add > the code of vmemmap page table manipulation in the future. > > The struct page structures (page structs) are used to describe a physic= al > page frame. By default, there is an one-to-one mapping from a page fram= e to > it's corresponding page struct. > > The HugeTLB pages consist of multiple base page size pages and is suppo= rted > by many architectures. See hugetlbpage.rst in the Documentation directo= ry > for more details. On the x86 architecture, HugeTLB pages of size 2MB an= d 1GB > are currently supported. Since the base page size on x86 is 4KB, a 2MB > HugeTLB page consists of 512 base pages and a 1GB HugeTLB page consists= of > 4096 base pages. For each base page, there is a corresponding page stru= ct. > > Within the HugeTLB subsystem, only the first 4 page structs are used to > contain unique information about a HugeTLB page. HUGETLB_CGROUP_MIN_ORD= ER > provides this upper limit. The only 'useful' information in the remaini= ng > page structs is the compound_head field, and this field is the same for= all > tail pages. > > By removing redundant page structs for HugeTLB pages, memory can return= ed to > the buddy allocator for other uses. > > When the system boot up, every 2M HugeTLB has 512 struct page structs w= hich > size is 8 pages(sizeof(struct page) * 512 / PAGE_SIZE). > > HugeTLB struct pages(8 pages) page frame(= 8 pages) > +-----------+ ---virt_to_page---> +-----------+ mapping to +-----= ------+ > | | | 0 | -------------> | = 0 | > | | +-----------+ +-----= ------+ > | | | 1 | -------------> | = 1 | > | | +-----------+ +-----= ------+ > | | | 2 | -------------> | = 2 | > | | +-----------+ +-----= ------+ > | | | 3 | -------------> | = 3 | > | | +-----------+ +-----= ------+ > | | | 4 | -------------> | = 4 | > | 2MB | +-----------+ +-----= ------+ > | | | 5 | -------------> | = 5 | > | | +-----------+ +-----= ------+ > | | | 6 | -------------> | = 6 | > | | +-----------+ +-----= ------+ > | | | 7 | -------------> | = 7 | > | | +-----------+ +-----= ------+ > | | > | | > | | > +-----------+ > > The value of page->compound_head is the same for all tail pages. The fi= rst > page of page structs (page 0) associated with the HugeTLB page contains= the 4 > page structs necessary to describe the HugeTLB. The only use of the rem= aining > pages of page structs (page 1 to page 7) is to point to page->compound_= head. > Therefore, we can remap pages 2 to 7 to page 1. Only 2 pages of page st= ructs > will be used for each HugeTLB page. This will allow us to free the rema= ining > 6 pages to the buddy allocator. > > Here is how things look after remapping. > > HugeTLB struct pages(8 pages) page frame(= 8 pages) > +-----------+ ---virt_to_page---> +-----------+ mapping to +-----= ------+ > | | | 0 | -------------> | = 0 | > | | +-----------+ +-----= ------+ > | | | 1 | -------------> | = 1 | > | | +-----------+ +-----= ------+ > | | | 2 | ----------------^ ^ ^= ^ ^ ^ > | | +-----------+ | |= | | | > | | | 3 | ------------------+ |= | | | > | | +-----------+ |= | | | > | | | 4 | --------------------+= | | | > | 2MB | +-----------+ = | | | > | | | 5 | ---------------------= -+ | | > | | +-----------+ = | | > | | | 6 | ---------------------= ---+ | > | | +-----------+ = | > | | | 7 | ---------------------= -----+ > | | +-----------+ > | | > | | > | | > +-----------+ > > When a HugeTLB is freed to the buddy system, we should allocate 6 pages= for > vmemmap pages and restore the previous mapping relationship. > > Apart from 2MB HugeTLB page, we also have 1GB HugeTLB page. It is simil= ar > to the 2MB HugeTLB page. We also can use this approach to free the vmem= map > pages. > > In this case, for the 1GB HugeTLB page, we can save 4094 pages. This is= a > very substantial gain. On our server, run some SPDK/QEMU applications w= hich > will use 1024GB HugeTLB page. With this feature enabled, we can save ~1= 6GB > (1G hugepage)/~12GB (2MB hugepage) memory. > > Because there are vmemmap page tables reconstruction on the freeing/all= ocating > path, it increases some overhead. Here are some overhead analysis. > > 1) Allocating 10240 2MB HugeTLB pages. > > a) With this patch series applied: > # time echo 10240 > /proc/sys/vm/nr_hugepages > > real 0m0.166s > user 0m0.000s > sys 0m0.166s > > # bpftrace -e 'kprobe:alloc_fresh_huge_page { @start[tid] =3D nsecs= ; } > kretprobe:alloc_fresh_huge_page /@start[tid]/ { @latency =3D hist= (nsecs - > @start[tid]); delete(@start[tid]); }' > Attaching 2 probes... > > @latency: > [8K, 16K) 5476 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@= @@@@@@@@@@@| > [16K, 32K) 4760 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@= @@@@ | > [32K, 64K) 4 | = | > > b) Without this patch series: > # time echo 10240 > /proc/sys/vm/nr_hugepages > > real 0m0.067s > user 0m0.000s > sys 0m0.067s > > # bpftrace -e 'kprobe:alloc_fresh_huge_page { @start[tid] =3D nsecs= ; } > kretprobe:alloc_fresh_huge_page /@start[tid]/ { @latency =3D hist= (nsecs - > @start[tid]); delete(@start[tid]); }' > Attaching 2 probes... > > @latency: > [4K, 8K) 10147 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@= @@@@@@@@@@@| > [8K, 16K) 93 | = | > > Summarize: this feature is about ~2x slower than before. > > 2) Freeing 10240 2MB HugeTLB pages. > > a) With this patch series applied: > # time echo 0 > /proc/sys/vm/nr_hugepages > > real 0m0.213s > user 0m0.000s > sys 0m0.213s > > # bpftrace -e 'kprobe:free_pool_huge_page { @start[tid] =3D nsecs; = } > kretprobe:free_pool_huge_page /@start[tid]/ { @latency =3D hist(n= secs - > @start[tid]); delete(@start[tid]); }' > Attaching 2 probes... > > @latency: > [8K, 16K) 6 | = | > [16K, 32K) 10227 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@= @@@@@@@@@@@| > [32K, 64K) 7 | = | > > b) Without this patch series: > # time echo 0 > /proc/sys/vm/nr_hugepages > > real 0m0.081s > user 0m0.000s > sys 0m0.081s > > # bpftrace -e 'kprobe:free_pool_huge_page { @start[tid] =3D nsecs; = } > kretprobe:free_pool_huge_page /@start[tid]/ { @latency =3D hist(n= secs - > @start[tid]); delete(@start[tid]); }' > Attaching 2 probes... > > @latency: > [4K, 8K) 6805 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@= @@@@@@@@@@@| > [8K, 16K) 3427 |@@@@@@@@@@@@@@@@@@@@@@@@@@ = | > [16K, 32K) 8 | = | > > Summarize: The overhead of __free_hugepage is about ~2-3x slower th= an before. > > Although the overhead has increased, the overhead is not significant. L= ike Mike > said, "However, remember that the majority of use cases create HugeTLB = pages at > or shortly after boot time and add them to the pool. So, additional ove= rhead is > at pool creation time. There is no change to 'normal run time' operatio= ns of > getting a page from or returning a page to the pool (think page fault/u= nmap)". > > Despite the overhead and in addition to the memory gains from this seri= es. The > following data is obtained by Joao Martins. Very thanks to his effort. > > There's an additional benefit which is page (un)pinners will see an imp= rovement > and Joao presumes because there are fewer memmap pages and thus the tai= l/head > pages are staying in cache more often. > > Out of the box Joao saw (when comparing linux-next against linux-next += this series) > with gup_test and pinning a 16G HugeTLB file (with 1G pages): > > get_user_pages(): ~32k -> ~9k > unpin_user_pages(): ~75k -> ~70k > > Usually any tight loop fetching compound_head(), or reading tail pages = data (e.g. > compound_head) benefit a lot. There's some unpinning inefficiencies Joa= o was > fixing[2], but with that in added it shows even more: > > unpin_user_pages(): ~27k -> ~3.8k > > [1] https://lore.kernel.org/linux-mm/20210409205254.242291-1-mike.krave= tz@oracle.com/ > [2] https://lore.kernel.org/linux-mm/20210204202500.26474-1-joao.m.mart= ins@oracle.com/ > > Todo: > - Free all of the tail vmemmap pages > Now for the 2MB HugrTLB page, we only free 6 vmemmap pages. we rea= lly can > free 7 vmemmap pages. In this case, we can see 8 of the 512 struct= page > structures has beed set PG_head flag. If we can adjust compound_he= ad() > slightly and make compound_head() return the real head struct page= when > the parameter is the tail struct page but with PG_head flag set. > > In order to make the code evolution route clearer. This feature ca= n can be > a separate patch after this patchset is solid. > > - Support for other architectures (e.g. aarch64). > - Enable PMD/huge page mapping of vmemmap even if this feature was e= nabled. > > Changelog in v22 -> v23: > - Collect Reviewed-by/Acked-by tags from Mike. > - Fix HPG_vmemmap_optimized comment. > > Very thanks to Mike's review. > > Changelog in v21 -> v22: > - Fix issue in isolate_or_dissolve_huge_page(). > - Fix page->private is cleared in migrate_page_states() > > Very thanks to Mike. > > Changelog in v20 -> v21: > - Collect Reviewed-by tags. > - Rebase to next-20210421 which is on the top of Oscar's changes. > - Use size_to_hstate() directly in free_hpage_workfn() suggested by = Mike. > - Add a comment above alloc_huge_page_vmemmap() in dissolve_free_hug= e_page() > suggested by Mike. > - Trim mhp_supports_memmap_on_memory() suggested by Oscar. > > Thanks to Mike and Oscar's suggestions. > > Changelog in v19 -> v20: > - Rebase to next-20210412. > - Introduce workqueue to defer freeing HugeTLB pages. > - Remove all tags (Reviewed-by ot Tested-by) from patch 6. > - Disable memmap_on_memory when hugetlb_free_vmemmap enabled (patch = 8). > > Changelog in v18 -> v19: > - Fix compiler error on sparc arch. Thanks Stephen. > - Make patch "gather discrete indexes of tail page" prior to "free t= he vmemmap > pages associated with each HugeTLB page". > - Remove some BUG_ON from patch #4. > - Update patch #6 changelog. > - Update Documentation/admin-guide/mm/memory-hotplug.rst. > - Drop the patch of "optimize the code with the help of the compiler= ". > - Update Documentation/admin-guide/kernel-parameters.txt in patch #7= . > - Trim update_and_free_page. > > Thanks to Michal, Oscar and Mike's review and suggestions. > > Changelog in v17 -> v18: > - Add complete copyright to bootmem_info.c (Suggested by Balbir). > - Fix some issues (in patch #4) suggested by Mike. > > Thanks to Balbir and Mike's review. Also thanks to Chen Huang and > Bodeddula Balasubramaniam's test. > > Changelog in v16 -> v17: > - Fix issues suggested by Mike and Oscar. > - Update commit log suggested by Michal. > > Thanks to Mike, David H and Michal's suggestions and review. > > Changelog in v15 -> v16: > - Use GFP_KERNEL to allocate vmemmap pages. > > Thanks to Mike, David H and Michal's suggestions. > > Changelog in v14 -> v15: > - Fix some issues suggested by Oscar. Thanks to Oscar. > - Add numbers which Joao Martins tested to cover letter. Thanks to h= is effort. > > Changelog in v13 -> v14: > - Refuse to free the HugeTLB page when the system is under memory pr= essure. > - Use GFP_ATOMIC to allocate vmemmap pages instead of GFP_KERNEL. > - Rebase to linux-next 20210202. > - Fix and add some comments for vmemmap_remap_free(). > > Thanks to Oscar, Mike, David H and David R's suggestions and review. > > Changelog in v12 -> v13: > - Remove VM_WARN_ON_PAGE macro. > - Add more comments in vmemmap_pte_range() and vmemmap_remap_free(). > > Thanks to Oscar and Mike's suggestions and review. > > Changelog in v11 -> v12: > - Move VM_WARN_ON_PAGE to a separate patch. > - Call __free_hugepage() with hugetlb_lock (See patch #5.) to serial= ize > with dissolve_free_huge_page(). It is to prepare for patch #9. > - Introduce PageHugeInflight. See patch #9. > > Changelog in v10 -> v11: > - Fix compiler error when !CONFIG_HUGETLB_PAGE_FREE_VMEMMAP. > - Rework some comments and commit changes. > - Rework vmemmap_remap_free() to 3 parameters. > > Thanks to Oscar and Mike's suggestions and review. > > Changelog in v9 -> v10: > - Fix a bug in patch #11. Thanks to Oscar for pointing that out. > - Rework some commit log or comments. Thanks Mike and Oscar for the = suggestions. > - Drop VMEMMAP_TAIL_PAGE_REUSE in the patch #3. > > Thank you very much Mike and Oscar for reviewing the code. > > Changelog in v8 -> v9: > - Rework some code. Very thanks to Oscar. > - Put all the non-hugetlb vmemmap functions under sparsemem-vmemmap.= c. > > Changelog in v7 -> v8: > - Adjust the order of patches. > > Very thanks to David and Oscar. Your suggestions are very valuable. > > Changelog in v6 -> v7: > - Rebase to linux-next 20201130 > - Do not use basepage mapping for vmemmap when this feature is disab= led. > - Rework some patchs. > [PATCH v6 08/16] mm/hugetlb: Free the vmemmap pages associated wit= h each hugetlb page > [PATCH v6 10/16] mm/hugetlb: Allocate the vmemmap pages associated= with each hugetlb page > > Thanks to Oscar and Barry. > > Changelog in v5 -> v6: > - Disable PMD/huge page mapping of vmemmap if this feature was enabl= ed. > - Simplify the first version code. > > Changelog in v4 -> v5: > - Rework somme comments and code in the [PATCH v4 04/21] and [PATCH = v4 05/21]. > > Thanks to Mike and Oscar's suggestions. > > Changelog in v3 -> v4: > - Move all the vmemmap functions to hugetlb_vmemmap.c. > - Make the CONFIG_HUGETLB_PAGE_FREE_VMEMMAP default to y, if we want= to > disable this feature, we should disable it by a boot/kernel comman= d line. > - Remove vmemmap_pgtable_{init, deposit, withdraw}() helper function= s. > - Initialize page table lock for vmemmap through core_initcall mecha= nism. > > Thanks for Mike and Oscar's suggestions. > > Changelog in v2 -> v3: > - Rename some helps function name. Thanks Mike. > - Rework some code. Thanks Mike and Oscar. > - Remap the tail vmemmap page with PAGE_KERNEL_RO instead of PAGE_KE= RNEL. > Thanks Matthew. > - Add some overhead analysis in the cover letter. > - Use vmemap pmd table lock instead of a hugetlb specific global loc= k. > > Changelog in v1 -> v2: > - Fix do not call dissolve_compound_page in alloc_huge_page_vmemmap(= ). > - Fix some typo and code style problems. > - Remove unused handle_vmemmap_fault(). > - Merge some commits to one commit suggested by Mike. > > Muchun Song (9): > mm: memory_hotplug: factor out bootmem core functions to > bootmem_info.c > mm: hugetlb: introduce a new config HUGETLB_PAGE_FREE_VMEMMAP > mm: hugetlb: gather discrete indexes of tail page > mm: hugetlb: free the vmemmap pages associated with each HugeTLB pag= e > mm: hugetlb: defer freeing of HugeTLB pages > mm: hugetlb: alloc the vmemmap pages associated with each HugeTLB pa= ge > mm: hugetlb: add a kernel parameter hugetlb_free_vmemmap > mm: memory_hotplug: disable memmap_on_memory when hugetlb_free_vmemm= ap > enabled > mm: hugetlb: introduce nr_free_vmemmap_pages in the struct hstate > > Documentation/admin-guide/kernel-parameters.txt | 25 ++ > Documentation/admin-guide/mm/hugetlbpage.rst | 11 + > Documentation/admin-guide/mm/memory-hotplug.rst | 13 ++ > arch/sparc/mm/init_64.c | 1 + > arch/x86/mm/init_64.c | 13 +- > drivers/acpi/acpi_memhotplug.c | 1 + > fs/Kconfig | 5 + > include/linux/bootmem_info.h | 66 ++++++ > include/linux/hugetlb.h | 46 +++- > include/linux/hugetlb_cgroup.h | 19 +- > include/linux/memory_hotplug.h | 27 --- > include/linux/mm.h | 5 + > mm/Makefile | 2 + > mm/bootmem_info.c | 127 ++++++++++ > mm/hugetlb.c | 198 +++++++++++++--= - > mm/hugetlb_vmemmap.c | 297 +++++++++++++++= +++++++++ > mm/hugetlb_vmemmap.h | 45 ++++ > mm/memory_hotplug.c | 117 +--------- > mm/migrate.c | 5 +- > mm/sparse-vmemmap.c | 267 +++++++++++++++= ++++++ > mm/sparse.c | 1 + > 21 files changed, 1102 insertions(+), 189 deletions(-) > create mode 100644 include/linux/bootmem_info.h > create mode 100644 mm/bootmem_info.c > create mode 100644 mm/hugetlb_vmemmap.c > create mode 100644 mm/hugetlb_vmemmap.h > --------------A189D1B5D083A421D2F1C052 Content-Type: text/html; charset="utf-8" Content-Transfer-Encoding: quoted-printable

Tested-by: Rohit Wali<rohiwali@amazon.com>

I have tested the patch set and the changes work as expected.
Test steps - Had an instance with 64G memory and allocated 50G in hugepages of size 2M and 1G.

(Shamelessly copying the result format used by Chen Huang for easier co-relation :) )

----------------------------------------------------------------=
---------------------------------------|
                      2M page                     |                      =
 1G page                      |
-----------------------|--------------------------|----------------------=
---|--------------------------|
       enable          |        disable           |        enable =C2=A0 =
        |          disable         |
-----------------------|--------------------------|----------------------=
---|--------------------------|
total  |  used  | free | total   |  used  | free  | total  |  used  | fre=
e  |  total  |  used  | free  |
64,312 | 51,576 |12,574| 63,514  | 51,368 | 11,956| 64,312 | 51,361 | 12,=
789| 63,514  | 51,362 | 11,962|
-------------------------------------------------------------------------=
------------------------------|

One thing to be aware of, though, is that the gain in memory from freeing page struct after hugepage reservation shows up in 'used' memory after unreserving hugepages

As an example, with using 50G worth of hugepages of size 1G.
Before hugepage reservation

|-------------------------|
| total  |  used  |  free |
| 63,512 |  206   | 63,083|=20
|-------------------------|

After reserving 50 pages of 1G

|----------------------------|
| total  |  used    | free   |
| 64,312 |  51,361  | 12,789 |=20
|----------------------------|

After unreserving hugepages in previous step

|-------------------------|
| total  |  used  |  free |
| 64,312 |  966   | 63,184|=20
|-------------------------|


On 10/05/21 8:30 am, Muchun Song wrote= :
Hi,

Since Mike's patches (make hugetlb put_page safe for all calling contexts=
[1])
applied into the next-20210412. We can move forward on this patch series =
now.

This patch series will free some vmemmap pages(struct page structures)
associated with each HugeTLB page when preallocated to save memory.

In order to reduce the difficulty of the first version of code review.
>From this version, we disable PMD/huge page mapping of vmemmap if thi=
s
feature was enabled. This acutely eliminates a bunch of the complex code
doing page table manipulation. When this patch series is solid, we cam ad=
d
the code of vmemmap page table manipulation in the future.

The struct page structures (page structs) are used to describe a physical
page frame. By default, there is an one-to-one mapping from a page frame =
to
it's corresponding page struct.

The HugeTLB pages consist of multiple base page size pages and is support=
ed
by many architectures. See hugetlbpage.rst in the Documentation directory
for more details. On the x86 architecture, HugeTLB pages of size 2MB and =
1GB
are currently supported. Since the base page size on x86 is 4KB, a 2MB
HugeTLB page consists of 512 base pages and a 1GB HugeTLB page consists o=
f
4096 base pages. For each base page, there is a corresponding page struct=
.

Within the HugeTLB subsystem, only the first 4 page structs are used to
contain unique information about a HugeTLB page. HUGETLB_CGROUP_MIN_ORDER
provides this upper limit. The only 'useful' information in the remaining
page structs is the compound_head field, and this field is the same for a=
ll
tail pages.

By removing redundant page structs for HugeTLB pages, memory can returned=
 to
the buddy allocator for other uses.

When the system boot up, every 2M HugeTLB has 512 struct page structs whi=
ch
size is 8 pages(sizeof(struct page) * 512 / PAGE_SIZE).

    HugeTLB                  struct pages(8 pages)         page frame(8 p=
ages)
 +-----------+ ---virt_to_page---> +-----------+   mapping to   +-----=
------+
 |           |                     |     0     | -------------> |     =
0     |
 |           |                     +-----------+                +--------=
---+
 |           |                     |     1     | -------------> |     =
1     |
 |           |                     +-----------+                +--------=
---+
 |           |                     |     2     | -------------> |     =
2     |
 |           |                     +-----------+                +--------=
---+
 |           |                     |     3     | -------------> |     =
3     |
 |           |                     +-----------+                +--------=
---+
 |           |                     |     4     | -------------> |     =
4     |
 |    2MB    |                     +-----------+                +--------=
---+
 |           |                     |     5     | -------------> |     =
5     |
 |           |                     +-----------+                +--------=
---+
 |           |                     |     6     | -------------> |     =
6     |
 |           |                     +-----------+                +--------=
---+
 |           |                     |     7     | -------------> |     =
7     |
 |           |                     +-----------+                +--------=
---+
 |           |
 |           |
 |           |
 +-----------+

The value of page->compound_head is the same for all tail pages. The f=
irst
page of page structs (page 0) associated with the HugeTLB page contains t=
he 4
page structs necessary to describe the HugeTLB. The only use of the remai=
ning
pages of page structs (page 1 to page 7) is to point to page->compound=
_head.
Therefore, we can remap pages 2 to 7 to page 1. Only 2 pages of page stru=
cts
will be used for each HugeTLB page. This will allow us to free the remain=
ing
6 pages to the buddy allocator.

Here is how things look after remapping.

    HugeTLB                  struct pages(8 pages)         page frame(8 p=
ages)
 +-----------+ ---virt_to_page---> +-----------+   mapping to   +-----=
------+
 |           |                     |     0     | -------------> |     =
0     |
 |           |                     +-----------+                +--------=
---+
 |           |                     |     1     | -------------> |     =
1     |
 |           |                     +-----------+                +--------=
---+
 |           |                     |     2     | ----------------^ ^ ^ ^ =
^ ^
 |           |                     +-----------+                   | | | =
| |
 |           |                     |     3     | ------------------+ | | =
| |
 |           |                     +-----------+                     | | =
| |
 |           |                     |     4     | --------------------+ | =
| |
 |    2MB    |                     +-----------+                       | =
| |
 |           |                     |     5     | ----------------------+ =
| |
 |           |                     +-----------+                         =
| |
 |           |                     |     6     | ------------------------=
+ |
 |           |                     +-----------+                         =
  |
 |           |                     |     7     | ------------------------=
--+
 |           |                     +-----------+
 |           |
 |           |
 |           |
 +-----------+

When a HugeTLB is freed to the buddy system, we should allocate 6 pages f=
or
vmemmap pages and restore the previous mapping relationship.

Apart from 2MB HugeTLB page, we also have 1GB HugeTLB page. It is similar
to the 2MB HugeTLB page. We also can use this approach to free the vmemma=
p
pages.

In this case, for the 1GB HugeTLB page, we can save 4094 pages. This is a
very substantial gain. On our server, run some SPDK/QEMU applications whi=
ch
will use 1024GB HugeTLB page. With this feature enabled, we can save ~16G=
B
(1G hugepage)/~12GB (2MB hugepage) memory.

Because there are vmemmap page tables reconstruction on the freeing/alloc=
ating
path, it increases some overhead. Here are some overhead analysis.

1) Allocating 10240 2MB HugeTLB pages.

   a) With this patch series applied:
   # time echo 10240 > /proc/sys/vm/nr_hugepages

   real     0m0.166s
   user     0m0.000s
   sys      0m0.166s

   # bpftrace -e 'kprobe:alloc_fresh_huge_page { @start[tid] =3D nsecs; }
     kretprobe:alloc_fresh_huge_page /@start[tid]/ { @latency =3D hist(ns=
ecs -
     @start[tid]); delete(@start[tid]); }'
   Attaching 2 probes...

   @latency:
   [8K, 16K)           5476 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@=
@@@@@@@@|
   [16K, 32K)          4760 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@=
@       |
   [32K, 64K)             4 |                                            =
        |

   b) Without this patch series:
   # time echo 10240 > /proc/sys/vm/nr_hugepages

   real     0m0.067s
   user     0m0.000s
   sys      0m0.067s

   # bpftrace -e 'kprobe:alloc_fresh_huge_page { @start[tid] =3D nsecs; }
     kretprobe:alloc_fresh_huge_page /@start[tid]/ { @latency =3D hist(ns=
ecs -
     @start[tid]); delete(@start[tid]); }'
   Attaching 2 probes...

   @latency:
   [4K, 8K)           10147 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@=
@@@@@@@@|
   [8K, 16K)             93 |                                            =
        |

   Summarize: this feature is about ~2x slower than before.

2) Freeing 10240 2MB HugeTLB pages.

   a) With this patch series applied:
   # time echo 0 > /proc/sys/vm/nr_hugepages

   real     0m0.213s
   user     0m0.000s
   sys      0m0.213s

   # bpftrace -e 'kprobe:free_pool_huge_page { @start[tid] =3D nsecs; }
     kretprobe:free_pool_huge_page /@start[tid]/ { @latency =3D hist(nsec=
s -
     @start[tid]); delete(@start[tid]); }'
   Attaching 2 probes...

   @latency:
   [8K, 16K)              6 |                                            =
        |
   [16K, 32K)         10227 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@=
@@@@@@@@|
   [32K, 64K)             7 |                                            =
        |

   b) Without this patch series:
   # time echo 0 > /proc/sys/vm/nr_hugepages

   real     0m0.081s
   user     0m0.000s
   sys      0m0.081s

   # bpftrace -e 'kprobe:free_pool_huge_page { @start[tid] =3D nsecs; }
     kretprobe:free_pool_huge_page /@start[tid]/ { @latency =3D hist(nsec=
s -
     @start[tid]); delete(@start[tid]); }'
   Attaching 2 probes...

   @latency:
   [4K, 8K)            6805 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@=
@@@@@@@@|
   [8K, 16K)           3427 |@@@@@@@@@@@@@@@@@@@@@@@@@@                  =
        |
   [16K, 32K)             8 |                                            =
        |

   Summarize: The overhead of __free_hugepage is about ~2-3x slower than =
before.

Although the overhead has increased, the overhead is not significant. Lik=
e Mike
said, "However, remember that the majority of use cases create HugeTLB pa=
ges at
or shortly after boot time and add them to the pool. So, additional overh=
ead is
at pool creation time. There is no change to 'normal run time' operations=
 of
getting a page from or returning a page to the pool (think page fault/unm=
ap)".

Despite the overhead and in addition to the memory gains from this series=
. The
following data is obtained by Joao Martins. Very thanks to his effort.

There's an additional benefit which is page (un)pinners will see an impro=
vement
and Joao presumes because there are fewer memmap pages and thus the tail/=
head
pages are staying in cache more often.

Out of the box Joao saw (when comparing linux-next against linux-next + t=
his series)
with gup_test and pinning a 16G HugeTLB file (with 1G pages):

	get_user_pages(): ~32k -> ~9k
	unpin_user_pages(): ~75k -> ~70k

Usually any tight loop fetching compound_head(), or reading tail pages da=
ta (e.g.
compound_head) benefit a lot. There's some unpinning inefficiencies Joao =
was
fixing[2], but with that in added it shows even more:

	unpin_user_pages(): ~27k -> ~3.8k

[1] https://lore.ker=
nel.org/linux-mm/20210409205254.242291-1-mike.kravetz@oracle.com/
[2] https://lore.ke=
rnel.org/linux-mm/20210204202500.26474-1-joao.m.martins@oracle.com/

Todo:
  - Free all of the tail vmemmap pages
    Now for the 2MB HugrTLB page, we only free 6 vmemmap pages. we really=
 can
    free 7 vmemmap pages. In this case, we can see 8 of the 512 struct pa=
ge
    structures has beed set PG_head flag. If we can adjust compound_head(=
)
    slightly and make compound_head() return the real head struct page wh=
en
    the parameter is the tail struct page but with PG_head flag set.

    In order to make the code evolution route clearer. This feature can c=
an be
    a separate patch after this patchset is solid.

  - Support for other architectures (e.g. aarch64).
  - Enable PMD/huge page mapping of vmemmap even if this feature was enab=
led.

Changelog in v22 -> v23:
  - Collect Reviewed-by/Acked-by tags from Mike.
  - Fix HPG_vmemmap_optimized comment.

  Very thanks to Mike's review.

Changelog in v21 -> v22:
  - Fix issue in isolate_or_dissolve_huge_page().
  - Fix page->private is cleared in migrate_page_states()

  Very thanks to Mike.

Changelog in v20 -> v21:
  - Collect Reviewed-by tags.
  - Rebase to next-20210421 which is on the top of Oscar's changes.
  - Use size_to_hstate() directly in free_hpage_workfn() suggested by Mik=
e.
  - Add a comment above alloc_huge_page_vmemmap() in dissolve_free_huge_p=
age()
    suggested by Mike.
  - Trim mhp_supports_memmap_on_memory() suggested by Oscar.

  Thanks to Mike and Oscar's suggestions.

Changelog in v19 -> v20:
  - Rebase to next-20210412.
  - Introduce workqueue to defer freeing HugeTLB pages.
  - Remove all tags (Reviewed-by ot Tested-by) from patch 6.
  - Disable memmap_on_memory when hugetlb_free_vmemmap enabled (patch 8).

Changelog in v18 -> v19:
  - Fix compiler error on sparc arch. Thanks Stephen.
  - Make patch "gather discrete indexes of tail page" prior to "free the =
vmemmap
    pages associated with each HugeTLB page".
  - Remove some BUG_ON from patch #4.
  - Update patch #6 changelog.
  - Update Documentation/admin-guide/mm/memory-hotplug.rst.
  - Drop the patch of "optimize the code with the help of the compiler".
  - Update Documentation/admin-guide/kernel-parameters.txt in patch #7.
  - Trim update_and_free_page.

 Thanks to Michal, Oscar and Mike's review and suggestions.

Changelog in v17 -> v18:
  - Add complete copyright to bootmem_info.c (Suggested by Balbir).
  - Fix some issues (in patch #4) suggested by Mike.

  Thanks to Balbir and Mike's review. Also thanks to Chen Huang and
  Bodeddula Balasubramaniam's test.

Changelog in v16 -> v17:
  - Fix issues suggested by Mike and Oscar.
  - Update commit log suggested by Michal.

  Thanks to Mike, David H and Michal's suggestions and review.

Changelog in v15 -> v16:
  - Use GFP_KERNEL to allocate vmemmap pages.

  Thanks to Mike, David H and Michal's suggestions.

Changelog in v14 -> v15:
  - Fix some issues suggested by Oscar. Thanks to Oscar.
  - Add numbers which Joao Martins tested to cover letter. Thanks to his =
effort.

Changelog in v13 -> v14:
  - Refuse to free the HugeTLB page when the system is under memory press=
ure.
  - Use GFP_ATOMIC to allocate vmemmap pages instead of GFP_KERNEL.
  - Rebase to linux-next 20210202.
  - Fix and add some comments for vmemmap_remap_free().

  Thanks to Oscar, Mike, David H and David R's suggestions and review.

Changelog in v12 -> v13:
  - Remove VM_WARN_ON_PAGE macro.
  - Add more comments in vmemmap_pte_range() and vmemmap_remap_free().

  Thanks to Oscar and Mike's suggestions and review.

Changelog in v11 -> v12:
  - Move VM_WARN_ON_PAGE to a separate patch.
  - Call __free_hugepage() with hugetlb_lock (See patch #5.) to serialize
    with dissolve_free_huge_page(). It is to prepare for patch #9.
  - Introduce PageHugeInflight. See patch #9.

Changelog in v10 -> v11:
  - Fix compiler error when !CONFIG_HUGETLB_PAGE_FREE_VMEMMAP.
  - Rework some comments and commit changes.
  - Rework vmemmap_remap_free() to 3 parameters.

  Thanks to Oscar and Mike's suggestions and review.

Changelog in v9 -> v10:
  - Fix a bug in patch #11. Thanks to Oscar for pointing that out.
  - Rework some commit log or comments. Thanks Mike and Oscar for the sug=
gestions.
  - Drop VMEMMAP_TAIL_PAGE_REUSE in the patch #3.

  Thank you very much Mike and Oscar for reviewing the code.

Changelog in v8 -> v9:
  - Rework some code. Very thanks to Oscar.
  - Put all the non-hugetlb vmemmap functions under sparsemem-vmemmap.c.

Changelog in v7 -> v8:
  - Adjust the order of patches.

  Very thanks to David and Oscar. Your suggestions are very valuable.

Changelog in v6 -> v7:
  - Rebase to linux-next 20201130
  - Do not use basepage mapping for vmemmap when this feature is disabled=
.
  - Rework some patchs.
    [PATCH v6 08/16] mm/hugetlb: Free the vmemmap pages associated with e=
ach hugetlb page
    [PATCH v6 10/16] mm/hugetlb: Allocate the vmemmap pages associated wi=
th each hugetlb page

  Thanks to Oscar and Barry.

Changelog in v5 -> v6:
  - Disable PMD/huge page mapping of vmemmap if this feature was enabled.
  - Simplify the first version code.

Changelog in v4 -> v5:
  - Rework somme comments and code in the [PATCH v4 04/21] and [PATCH v4 =
05/21].

  Thanks to Mike and Oscar's suggestions.

Changelog in v3 -> v4:
  - Move all the vmemmap functions to hugetlb_vmemmap.c.
  - Make the CONFIG_HUGETLB_PAGE_FREE_VMEMMAP default to y, if we want to
    disable this feature, we should disable it by a boot/kernel command l=
ine.
  - Remove vmemmap_pgtable_{init, deposit, withdraw}() helper functions.
  - Initialize page table lock for vmemmap through core_initcall mechanis=
m.

  Thanks for Mike and Oscar's suggestions.

Changelog in v2 -> v3:
  - Rename some helps function name. Thanks Mike.
  - Rework some code. Thanks Mike and Oscar.
  - Remap the tail vmemmap page with PAGE_KERNEL_RO instead of PAGE_KERNE=
L.
    Thanks Matthew.
  - Add some overhead analysis in the cover letter.
  - Use vmemap pmd table lock instead of a hugetlb specific global lock.

Changelog in v1 -> v2:
  - Fix do not call dissolve_compound_page in alloc_huge_page_vmemmap().
  - Fix some typo and code style problems.
  - Remove unused handle_vmemmap_fault().
  - Merge some commits to one commit suggested by Mike.

Muchun Song (9):
  mm: memory_hotplug: factor out bootmem core functions to
    bootmem_info.c
  mm: hugetlb: introduce a new config HUGETLB_PAGE_FREE_VMEMMAP
  mm: hugetlb: gather discrete indexes of tail page
  mm: hugetlb: free the vmemmap pages associated with each HugeTLB page
  mm: hugetlb: defer freeing of HugeTLB pages
  mm: hugetlb: alloc the vmemmap pages associated with each HugeTLB page
  mm: hugetlb: add a kernel parameter hugetlb_free_vmemmap
  mm: memory_hotplug: disable memmap_on_memory when hugetlb_free_vmemmap
    enabled
  mm: hugetlb: introduce nr_free_vmemmap_pages in the struct hstate

 Documentation/admin-guide/kernel-parameters.txt |  25 ++
 Documentation/admin-guide/mm/hugetlbpage.rst    |  11 +
 Documentation/admin-guide/mm/memory-hotplug.rst |  13 ++
 arch/sparc/mm/init_64.c                         |   1 +
 arch/x86/mm/init_64.c                           |  13 +-
 drivers/acpi/acpi_memhotplug.c                  |   1 +
 fs/Kconfig                                      |   5 +
 include/linux/bootmem_info.h                    |  66 ++++++
 include/linux/hugetlb.h                         |  46 +++-
 include/linux/hugetlb_cgroup.h                  |  19 +-
 include/linux/memory_hotplug.h                  |  27 ---
 include/linux/mm.h                              |   5 +
 mm/Makefile                                     |   2 +
 mm/bootmem_info.c                               | 127 ++++++++++
 mm/hugetlb.c                                    | 198 +++++++++++++---
 mm/hugetlb_vmemmap.c                            | 297 ++++++++++++++++++=
++++++
 mm/hugetlb_vmemmap.h                            |  45 ++++
 mm/memory_hotplug.c                             | 117 +---------
 mm/migrate.c                                    |   5 +-
 mm/sparse-vmemmap.c                             | 267 ++++++++++++++++++=
+++
 mm/sparse.c                                     |   1 +
 21 files changed, 1102 insertions(+), 189 deletions(-)
 create mode 100644 include/linux/bootmem_info.h
 create mode 100644 mm/bootmem_info.c
 create mode 100644 mm/hugetlb_vmemmap.c
 create mode 100644 mm/hugetlb_vmemmap.h
--------------A189D1B5D083A421D2F1C052--