Received: by 2002:ac0:a581:0:0:0:0:0 with SMTP id m1-v6csp4762090imm;
        Mon, 25 Jun 2018 23:33:04 -0700 (PDT)
X-Google-Smtp-Source: AAOMgpdzzDrYIxQfyt5X7MRQBXf7fBBIkmUTFPnXAz6XgDpxJvj9YuaM9F1MWDVIFbpvs9qgoxhW
X-Received: by 2002:a62:aa18:: with SMTP id e24-v6mr266668pff.72.1529994784180;
        Mon, 25 Jun 2018 23:33:04 -0700 (PDT)
ARC-Seal: i=1; a=rsa-sha256; t=1529994784; cv=none;
        d=google.com; s=arc-20160816;
        b=DeWJhr9awG4qWBGynqp10wTYf+rWF8IhQXpMVz386/0MxjQLGzRvRq2zkoPOLQjfdo
         hla5wz1jS7szfnk+4yGN6x2+tyrQ/sTOYFB9H1WvDVCMOedUtVcuEpE6iXTn0xeNkfTb
         T1k3lt4+n91rSqUaqZYe6ZnBUiLhuGguA/SSzoXqr7XL5zF8L7qhYSvQT+arTFZFQ9Qj
         tXI1eloKC3f0rCVdcy8FDEAgCUBR43gq9sbbei0HPm8bqroC3nDVcssUKxO6b0ANzj2i
         49zinQbadhAxKDBHF8GeecPfj5DVsHv1+NecDyqn8eb3g/iRLdGLPKcOp3hZrLp2GPI6
         5Dpw==
ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816;
        h=list-id:precedence:sender:content-transfer-encoding
         :content-language:in-reply-to:mime-version:user-agent:date
         :message-id:from:references:cc:to:subject:arc-authentication-results;
        bh=KtpsBoaAkF90avDwcJR9wJ7bDV1QsCcn5HM1A/cOqhw=;
        b=fGSBBOt0+6Ps8sHanv/6psRQVbm10q9mRocs8Z1HFVZcepQVnpcN4wNR0sjwNdtVhR
         SSX0siJONz+KHoOUjWgHFb8Uz1I31lMfP9bDjvI4R10AmE8HPM1NoIn8B97P4Tu3FCxb
         yFvKgbHmPPSgRu492G3zbNwllyOjoLWam+Jgbcyj9AJp2CWeL4/kznZUjTv59JcSl1mM
         uZkVuMunTa2HV7VcK0BHYmIdoK3/EDkhRlyWN821C1IEZGEouWvsgCWhDOYCJRI0iJyM
         PfAr5+3dK5KrpgISfdmBzr/BP9LakzfRBl0XivR/hlU3T7kPyCNj5SYWcnZeAz7NKZZq
         +2wQ==
ARC-Authentication-Results: i=1; mx.google.com;
       spf=pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org;
       dmarc=fail (p=NONE sp=NONE dis=NONE) header.from=nvidia.com
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: from vger.kernel.org (vger.kernel.org. [209.132.180.67])
        by mx.google.com with ESMTP id c18-v6si772054pgp.467.2018.06.25.23.32.49;
        Mon, 25 Jun 2018 23:33:04 -0700 (PDT)
Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Authentication-Results: mx.google.com;
       spf=pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org;
       dmarc=fail (p=NONE sp=NONE dis=NONE) header.from=nvidia.com
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S1751843AbeFZGcJ (ORCPT <rfc822;yxhlfx@gmail.com> + 99 others);
        Tue, 26 Jun 2018 02:32:09 -0400
Received: from hqemgate15.nvidia.com ([216.228.121.64]:9859 "EHLO
        hqemgate15.nvidia.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1751181AbeFZGcH (ORCPT
        <rfc822;linux-kernel@vger.kernel.org>);
        Tue, 26 Jun 2018 02:32:07 -0400
Received: from hqpgpgate102.nvidia.com (Not Verified[216.228.121.13]) by hqemgate15.nvidia.com (using TLS: TLSv1, AES128-SHA)
        id <B5b31ddca0002>; Mon, 25 Jun 2018 23:31:38 -0700
Received: from HQMAIL107.nvidia.com ([172.20.161.6])
  by hqpgpgate102.nvidia.com (PGP Universal service);
  Mon, 25 Jun 2018 23:32:08 -0700
X-PGP-Universal: processed;
        by hqpgpgate102.nvidia.com on Mon, 25 Jun 2018 23:32:08 -0700
Received: from [10.110.48.28] (10.110.48.28) by HQMAIL107.nvidia.com
 (172.20.187.13) with Microsoft SMTP Server (TLS) id 15.0.1347.2; Tue, 26 Jun
 2018 06:32:06 +0000
Subject: Re: [PATCH 2/2] mm: set PG_dma_pinned on get_user_pages*()
To:     Jan Kara <jack@suse.cz>
CC:     Matthew Wilcox <willy@infradead.org>,
        Dan Williams <dan.j.williams@intel.com>,
        Christoph Hellwig <hch@lst.de>, Jason Gunthorpe <jgg@ziepe.ca>,
        John Hubbard <john.hubbard@gmail.com>,
        Michal Hocko <mhocko@kernel.org>,
        Christopher Lameter <cl@linux.com>,
        Linux MM <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>,
        linux-rdma <linux-rdma@vger.kernel.org>
References: <3898ef6b-2fa0-e852-a9ac-d904b47320d5@nvidia.com>
 <CAPcyv4iRBzmwWn_9zDvqdfVmTZL_Gn7uA_26A1T-kJib=84tvA@mail.gmail.com>
 <0e6053b3-b78c-c8be-4fab-e8555810c732@nvidia.com>
 <20180619082949.wzoe42wpxsahuitu@quack2.suse.cz>
 <20180619090255.GA25522@bombadil.infradead.org>
 <20180619104142.lpilc6esz7w3a54i@quack2.suse.cz>
 <70001987-3938-d33e-11e0-de5b19ca3bdf@nvidia.com>
 <20180620120824.bghoklv7qu2z5wgy@quack2.suse.cz>
 <151edbf3-66ff-df0c-c1cc-5998de50111e@nvidia.com>
 <20180621163036.jvdbsv3t2lu34pdl@quack2.suse.cz>
 <20180625152150.jnf5suiubecfppcl@quack2.suse.cz>
X-Nvconfidentiality: public
From:   John Hubbard <jhubbard@nvidia.com>
Message-ID: <550aacd3-cfea-c99a-3b60-563dd1621d5c@nvidia.com>
Date:   Mon, 25 Jun 2018 23:31:06 -0700
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.8.0
MIME-Version: 1.0
In-Reply-To: <20180625152150.jnf5suiubecfppcl@quack2.suse.cz>
X-Originating-IP: [10.110.48.28]
X-ClientProxiedBy: HQMAIL105.nvidia.com (172.20.187.12) To
 HQMAIL107.nvidia.com (172.20.187.13)
Content-Type: text/plain; charset="utf-8"
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On 06/25/2018 08:21 AM, Jan Kara wrote:
> On Thu 21-06-18 18:30:36, Jan Kara wrote:
>> On Wed 20-06-18 15:55:41, John Hubbard wrote:
>>> On 06/20/2018 05:08 AM, Jan Kara wrote:
>>>> On Tue 19-06-18 11:11:48, John Hubbard wrote:
>>>>> On 06/19/2018 03:41 AM, Jan Kara wrote:
>>>>>> On Tue 19-06-18 02:02:55, Matthew Wilcox wrote:
>>>>>>> On Tue, Jun 19, 2018 at 10:29:49AM +0200, Jan Kara wrote:
>>> [...]
> I've spent some time on this. There are two obstacles with my approach of
> putting special entry into inode's VMA tree:
> 
> 1) If I want to place this special entry in inode's VMA tree, I either need
> to allocate full VMA, somehow initiate it so that it's clear it's a special
> "pinned" range, not a VMA => uses unnecessarily too much memory, it is
> ugly. Another solution I was hoping for was that I would factor out some
> common bits of vm_area_struct (pgoff, rb_node, ..) into a structure common
> for VMA and the locked range => doable but causes a lot of churn as VMAs
> are accessed (and modified!) at hundreds of places in the kernel. Some
> accessor functions would help to reduce the churn a bit but then stuff like
> vma_set_pgoff(vma, pgoff) isn't exactly beautiful either.
> 
> 2) Some users of GUP (e.g. direct IO) get a block of pages and then put
> references to these pages at different times and in random order -
> basically when IO for given page is completed, reference is dropped and one
> GUP call can acquire page references for pages which end up in multiple
> different bios (we don't know in advance). This makes is difficult to
> implement counterpart to GUP to 'unpin' a range of pages - we'd either have
> to support partial unpins (and splitting of pinned ranges and all such fun)
> or just have to track internally in how many pages are still pinned in the
> originally pinned range and release the pin once all individual pages are
> unpinned but then it's difficult to e.g. get to this internal structure
> from IO completion callback where we only have the bio.
> 
> So I think the Matthew's idea of removing pinned pages from LRU is
> definitely worth trying to see how complex that would end up being. Did you
> get to looking into it? If not, I can probably find some time to try that
> out.
> 
 
OK, so I looked into this some more.

As you implied in an earlier response, removing a page from LRU is probably the
easy part. It's *keeping* it off the LRU that worries me. I looked at SetPageLRU()
uses, there were only 5 call sites, and of those, I think only one might be difficult:

    __pagevec_lru_add()

It seems like the way to avoid __pagevec_lru_add() calls on these pages is to
first call lru_add_drain_all, then remove the pages from LRU (presumably via
isolate_page_lru). I think that should do it. But I'm a little concerned that 
maybe I'm overlooking something.

Here are the 5 search hits and my analysis. This may have mistakes in it, as
I'm pretty new to this area, which is why I'm spelling it out:

1. mm/memcontrol.c:2082: SetPageLRU(page); 

    This is in unlock_page_lru(). Caller: commit_charge(), and it's conditional on 
    lrucare, so we can just skip it if the new page flag is set.

2. mm/swap.c:831: SetPageLRU(page_tail);
    This is in lru_add_page_tail(), which is only called by __split_huge_page_tail, and
    there, we can also just skip the call for these pages.

3. mm/swap.c:866:  SetPageLRU(page);
    This is in __pagevec_lru_add_fn (sole caller: __pagevec_lru_add), and is
    discussed above.

4. mm/vmscan.c:1680: SetPageLRU(page);	
    This is in putback_inactive_pages(), which I think won't get called unless
    the page is already on an LRU.

5. mm/vmscan.c:1873: SetPageLRU(page);	//  (N/A)
    This is in move_active_pages_to_lru(), which I also think won't get called unless 
    the page is already on an LRU.


thanks,
-- 
John Hubbard
NVIDIA