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;
MIME-Version: 1.0
References: <20181029210716.212159-1-brho@google.com> <CAPcyv4gJUjuSKwy7i2wuKR=Vz-AkDrxnGya5qkg7XTFxuXbtzw@mail.gmail.com>
 <e00c5f69-f4d7-859c-cbde-0392b0a5195c@redhat.com>
In-Reply-To: <e00c5f69-f4d7-859c-cbde-0392b0a5195c@redhat.com>
From:   Dan Williams <dan.j.williams@intel.com>
Date:   Wed, 31 Oct 2018 14:16:01 -0700
Message-ID: <CAPcyv4hqOhCBvYjxY31kc+BD+MraCBtou=MogApp_5wpz3W6Cw@mail.gmail.com>
Subject: Re: [RFC PATCH] kvm: Use huge pages for DAX-backed files
To:     Paolo Bonzini <pbonzini@redhat.com>
Cc:     Barret Rhoden <brho@google.com>, Dave Jiang <dave.jiang@intel.com>,
        zwisler@kernel.org, Vishal L Verma <vishal.l.verma@intel.com>,
        rkrcmar@redhat.com, Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        linux-nvdimm <linux-nvdimm@lists.01.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        "H. Peter Anvin" <hpa@zytor.com>, X86 ML <x86@kernel.org>,
        KVM list <kvm@vger.kernel.org>,
        "Zhang, Yu C" <yu.c.zhang@intel.com>,
        "Zhang, Yi Z" <yi.z.zhang@intel.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Oct 31, 2018 at 1:52 AM Paolo Bonzini <pbonzini@redhat.com> wrote:
>
> On 29/10/2018 23:25, Dan Williams wrote:
> > I'm wondering if we're adding an explicit is_zone_device_page() check
> > in this path to determine the page mapping size if that can be a
> > replacement for the kvm_is_reserved_pfn() check. In other words, the
> > goal of fixing up PageReserved() was to preclude the need for DAX-page
> > special casing in KVM, but if we already need add some special casing
> > for page size determination, might as well bypass the
> > kvm_is_reserved_pfn() dependency as well.
>
> No, please don't.  The kvm_is_reserved_pfn() check is for correctness,
> the page-size check is for optimization.  In theory you could have a
> ZONE_DEVICE area that is smaller than 2MB and thus does not use huge pages.

To be clear, I was not suggesting that a ZONE_DEVICE check would be
sufficient to determine a 2MB page. I was suggesting that given there
is a debate about removing the "reserved" designation for dax pages
that debate is moot if kvm is going to add interrogation code to
figure out the size of dax mappings.