Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <159457116473.754248.7879464730875147365.stgit@dwillia2-desk3.amr.corp.intel.com>
 <159457120334.754248.12908401960465408733.stgit@dwillia2-desk3.amr.corp.intel.com>
 <20200713065801.GB11000@linux.ibm.com>
In-Reply-To: <20200713065801.GB11000@linux.ibm.com>
From:   Dan Williams <dan.j.williams@intel.com>
Date:   Mon, 13 Jul 2020 08:42:41 -0700
Message-ID: <CAPcyv4gBgCptZupxhzbDN3qscnuJ9HNWHHvfhH6z+z84VE1cPg@mail.gmail.com>
Subject: Re: [PATCH v2 07/22] numa: Introduce a generic memory_add_physaddr_to_nid()
To:     Mike Rapoport <rppt@linux.ibm.com>
Cc:     linux-nvdimm <linux-nvdimm@lists.01.org>,
        Jia He <justin.he@arm.com>, Will Deacon <will@kernel.org>,
        David Hildenbrand <david@redhat.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Peter Zijlstra <peterz@infradead.org>,
        Vishal L Verma <vishal.l.verma@intel.com>,
        Dave Hansen <dave.hansen@linux.intel.com>,
        Ard Biesheuvel <ard.biesheuvel@linaro.org>,
        Linux MM <linux-mm@kvack.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Linux ACPI <linux-acpi@vger.kernel.org>,
        Christoph Hellwig <hch@lst.de>,
        Joao Martins <joao.m.martins@oracle.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Sun, Jul 12, 2020 at 11:58 PM Mike Rapoport <rppt@linux.ibm.com> wrote:
>
> Hi Dan,
>
> On Sun, Jul 12, 2020 at 09:26:43AM -0700, Dan Williams wrote:
> > For architectures that opt into storing their numa data in memblock
> > (only ARM64 currently), add a memblock generic way to interrogate that
> > data for memory_add_physaddr_to_nid(). This requires ARCH_KEEP_MEMBLOCK
> > to keep memblock text and data around after boot.
>
> I afraid we are too far from using memblock as a generic placeholder for
> numa data. Although all architectures now have the numa info in
> memblock, only arm64 uses memblock as the primary source of that data.
>
> I'd rather prefer Jia's solution [1] to have a weak default for
> memory_add_physaddr_to_nid() and let architectures override it.

I'm ok with that as long as we do the same for phys_to_target_node().

Will had the concern about adding a generic numa-info facility the
last I tried this. I just don't see a practical way to get there in
the near term.