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=-5.8 required=3.0 tests=BAYES_00,DKIMWL_WL_HIGH, DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED autolearn=no 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 2BFBAC63777 for ; Fri, 20 Nov 2020 22:06:16 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 3BD0E22470 for ; Fri, 20 Nov 2020 22:06:15 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b="NPRAG36s" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 3BD0E22470 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=redhat.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 81CE06B005C; Fri, 20 Nov 2020 17:06:14 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 7CCA56B005D; Fri, 20 Nov 2020 17:06:14 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 6E30D6B0068; Fri, 20 Nov 2020 17:06:14 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0203.hostedemail.com [216.40.44.203]) by kanga.kvack.org (Postfix) with ESMTP id 4170F6B005C for ; Fri, 20 Nov 2020 17:06:14 -0500 (EST) Received: from smtpin05.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay01.hostedemail.com (Postfix) with ESMTP id DE091180BA423 for ; Fri, 20 Nov 2020 22:06:13 +0000 (UTC) X-FDA: 77506180626.05.heart72_170234e2734e Received: from filter.hostedemail.com (10.5.16.251.rfc1918.com [10.5.16.251]) by smtpin05.hostedemail.com (Postfix) with ESMTP id C316B180B7E34 for ; Fri, 20 Nov 2020 22:06:13 +0000 (UTC) X-HE-Tag: heart72_170234e2734e X-Filterd-Recvd-Size: 10186 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [63.128.21.124]) by imf03.hostedemail.com (Postfix) with ESMTP for ; Fri, 20 Nov 2020 22:06:12 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1605909972; 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=vRqZWjIekH8IKiPozravJN75Z5P7iJOMdRUsw2VB3VE=; b=NPRAG36slTz3APoSjl5fFx1iVnpFqlm86cuXGqhe3NkDkKWTFfIwaqUigOEir+vSqjzf+7 tcGojIn+sktLpUaZenvU8tzsMT856zXRekCxSjllBAWh0bTCTmi56MGrsQVdEkLc0Eptzg xop/KLr3ABYibqDBXl5QF7Utn86zL2Y= Received: from mail-wm1-f70.google.com (mail-wm1-f70.google.com [209.85.128.70]) (Using TLS) by relay.mimecast.com with ESMTP id us-mta-580-eP6pciXhMZGA9pC8bSvySA-1; Fri, 20 Nov 2020 17:06:08 -0500 X-MC-Unique: eP6pciXhMZGA9pC8bSvySA-1 Received: by mail-wm1-f70.google.com with SMTP id j62so4577512wma.4 for ; Fri, 20 Nov 2020 14:06:08 -0800 (PST) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:content-transfer-encoding:from:mime-version :subject:date:message-id:references:cc:in-reply-to:to; bh=FbFV32ygDz7qsaJYh4yraORMBM+FGxbd+GP8M8kteqw=; b=UjbSkcmTXwiFLZLXS0t1Rz9z9CbGBgpxUL1Z1F06jSK3qcNZ0oCBdh9Mn2mkQCRE/T ar1WFpb8R9aMvyo3i3mv/4l2SC6JCuZjOzNmdPJ0JiKPU69c+rsLVQAPVb3ZhJWQfpjV bfA8FIUKxTw+gGCj9HjnMyYoJQ56jungWBY3GWYivyb2SQ/Z8O5NK/yd91oz9tDW3E8I 08PpbCs0DH09ocFU5oqNcvWZ2oEoWjfmenTrCIa8ClxTXT/ZOPdqY69wGPGKhqgm9siU J4z+12px6zyzycM01Uwzl+XYI8p2H70pPXSmsRUKtpXa0eK4SwlUfawg/cn2mJo1hNtf YM3w== X-Gm-Message-State: AOAM533C9b21bmA/Qn/cwmeYTkSYG1PjHgjokLwUJr7KEZ3XzNyNVMmb 07Bi6KECnZCUB7JW5Xke/imiuZW8JZKcf2XXrlRkZz7biFCgByNLXdsAgnVRsdYyWrXW6rLpKtW hRPcT3T86G5Q= X-Received: by 2002:a7b:c309:: with SMTP id k9mr11829418wmj.14.1605909966971; Fri, 20 Nov 2020 14:06:06 -0800 (PST) X-Google-Smtp-Source: ABdhPJxEM98LDUf+EVjrqARV7ebgtFqhhBe1exDpHAefUWWWQXX/9FolmyIU3Y2xaV0me8hxWz9M/w== X-Received: by 2002:a7b:c309:: with SMTP id k9mr11829394wmj.14.1605909966648; Fri, 20 Nov 2020 14:06:06 -0800 (PST) Received: from [192.168.3.114] (p5b0c6104.dip0.t-ipconnect.de. [91.12.97.4]) by smtp.gmail.com with ESMTPSA id e6sm5879095wme.27.2020.11.20.14.06.05 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Fri, 20 Nov 2020 14:06:06 -0800 (PST) From: David Hildenbrand Mime-Version: 1.0 (1.0) Subject: Re: Pinning ZONE_MOVABLE pages Date: Fri, 20 Nov 2020 23:06:05 +0100 Message-Id: References: Cc: David Hildenbrand , linux-mm , Andrew Morton , Vlastimil Babka , LKML , Michal Hocko , Oscar Salvador , Dan Williams , Sasha Levin , Tyler Hicks , Joonsoo Kim , sthemmin@microsoft.com In-Reply-To: To: Pavel Tatashin X-Mailer: iPhone Mail (18A8395) Authentication-Results: relay.mimecast.com; auth=pass smtp.auth=CUSA124A263 smtp.mailfrom=david@redhat.com X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable 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: > Am 20.11.2020 um 22:58 schrieb Pavel Tatashin = : >=20 > =EF=BB=BFOn Fri, Nov 20, 2020 at 3:59 PM David Hildenbrand wrote: >>=20 >>=20 >>>> Am 20.11.2020 um 21:28 schrieb Pavel Tatashin : >>>=20 >>> =EF=BB=BFRecently, I encountered a hang that is happening during memory= hot >>> remove operation. It turns out that the hang is caused by pinned user >>> pages in ZONE_MOVABLE. >>>=20 >>> Kernel expects that all pages in ZONE_MOVABLE can be migrated, but >>> this is not the case if a user applications such as through dpdk >>> libraries pinned them via vfio dma map. Kernel keeps trying to >>> hot-remove them, but refcnt never gets to zero, so we are looping >>> until the hardware watchdog kicks in. >>>=20 >>> We cannot do dma unmaps before hot-remove, because hot-remove is a >>> slow operation, and we have thousands for network flows handled by >>> dpdk that we just cannot suspend for the duration of hot-remove >>> operation. >>>=20 >>=20 >> Hi! >>=20 >> It=E2=80=98s a known problem also for VMs using vfio. I thought about th= is some while ago an came to the same conclusion: before performing long-te= rm pinnings, we have to migrate pages off the movable zone. After that, it= =E2=80=98s too late. >>=20 >> What happens when we can=E2=80=98t migrate (OOM on !MOVABLE memory, shor= t-term pinning)? TBD. >>=20 >>> The solution is for dpdk to allocate pages from a zone below >>> ZONE_MOVAVLE, i.e. ZONE_NORMAL/ZONE_HIGHMEM, but this is not possible. >>> There is no user interface that we have that allows applications to >>> select what zone the memory should come from. >>>=20 >>> I've spoken with Stephen Hemminger, and he said that DPDK is moving in >>> the direction of using transparent huge pages instead of HugeTLBs, >>> which means that we need to allow at least anonymous, and anonymous >>> transparent huge pages to come from non-movable zones on demand. >>>=20 >>> Here is what I am proposing: >>> 1. Add a new flag that is passed through pin_user_pages_* down to >>> fault handlers, and allow the fault handler to allocate from a >>> non-movable zone. >>>=20 >>> Sample function stacks through which this info needs to be passed is th= is: >>>=20 >>> pin_user_pages_remote(gup_flags) >>> __get_user_pages_remote(gup_flags) >>> __gup_longterm_locked(gup_flags) >>> __get_user_pages_locked(gup_flags) >>> __get_user_pages(gup_flags) >>> faultin_page(gup_flags) >>> Convert gup_flags into fault_flags >>> handle_mm_fault(fault_flags) >>>=20 >>> From handle_mm_fault(), the stack diverges into various faults, >>> examples include: >>>=20 >>> Transparent Huge Page >>> handle_mm_fault(fault_flags) >>> __handle_mm_fault(fault_flags) >>> Create: struct vm_fault vmf, use fault_flags to specify correct gfp_mas= k >>> create_huge_pmd(vmf); >>> do_huge_pmd_anonymous_page(vmf); >>> mm_get_huge_zero_page(vma->vm_mm); -> flag is lost, so flag from >>> vmf.gfp_mask should be passed as well. >>>=20 >>> There are several other similar paths in a transparent huge page, also >>> there is a named path where allocation is based on filesystems, and >>> the flag should be honored there as well, but it does not have to be >>> added at the same time. >>>=20 >>> Regular Pages >>> handle_mm_fault(fault_flags) >>> __handle_mm_fault(fault_flags) >>> Create: struct vm_fault vmf, use fault_flags to specify correct gfp_mas= k >>> handle_pte_fault(vmf) >>> do_anonymous_page(vmf); >>> page =3D alloc_zeroed_user_highpage_movable(vma, vmf->address); -> >>> replace change this call according to gfp_mask. >>>=20 >>> The above only take care of the case if user application faults on the >>> page during pinning time, but there are also cases where pages already >>> exist. >>>=20 >>> 2. Add an internal move_pages_zone() similar to move_pages() syscall >>> but instead of migrating to a different NUMA node, migrate pages from >>> ZONE_MOVABLE to another zone. >>> Call move_pages_zone() on demand prior to pinning pages from >>> vfio_pin_map_dma() for instance. >>>=20 >>> 3. Perhaps, it also makes sense to add madvise() flag, to allocate >>> pages from non-movable zone. When a user application knows that it >>> will do DMA mapping, and pin pages for a long time, the memory that it >>> allocates should never be migrated or hot-removed, so make sure that >>> it comes from the appropriate place. >>> The benefit of adding madvise() flag is that we won't have to deal >>> with slow page migration during pin time, but the disadvantage is that >>> we would need to change the user interface. >>>=20 >>=20 >> Hm, I am not sure we want to expose these details. What would be the sem= antics? =E2=80=9EMight pin=E2=80=9C? Hm, not sure. >=20 > The semantic would be PA must not change, something that DPDK > currently excpects from huge pages, which by the way is not true, as > huge pages are migratable. >=20 >>=20 >> Assume you start a fresh VM via QEMU with vfio. When we start mapping gu= est memory via vfio, that=E2=80=98s usually the time memory will get popula= ted. Not really much has to be migrated. I think this is even true during l= ive migration. >>=20 >> I think selective DMA pinning (e.g., vIOMMU in QEMU) is different, where= we keep pinning/unpinning on demand. But I guess even here, we will often = reuse some pages over and over again. >>=20 >>=20 >>> Before I start working on the above approaches, I would like to get an >>> opinion from the community on an appropriate path forward for this >>> problem. If what I described sounds reasonable, or if there are other >>> ideas on how to address the problem that I am seeing. >>=20 >> At least 1 and 2 sound sane. 3 is TBD - but it=E2=80=98s a pure optimiza= tion, so it can wait. >=20 > Makes sense, I am also worried about 3, but most of madvise() flags > are for pure optimization purposes: MADV_HUGEPAGE, MADV_SEQUENTIAL, > MADV_WILLNEED etc. BTW, I assume we should also directly tackle migrating pages off CMA region= s when pinning, I guess quite some people will be interested in that as wel= l. Have a nice weekend and thanks for looking into this issue :)