Re: [PATCH 0/7] ACPI HMAT memory sysfs representation

From: Anshuman Khandual
Date: Mon Nov 26 2018 - 10:52:15 EST




On 11/24/2018 02:43 AM, Dan Williams wrote:
> On Fri, Nov 23, 2018 at 11:21 AM Dave Hansen <dave.hansen@xxxxxxxxx> wrote:
>>
>> On 11/22/18 10:42 PM, Anshuman Khandual wrote:
>>> Are we willing to go in the direction for inclusion of a new system
>>> call, subset of it appears on sysfs etc ? My primary concern is not
>>> how the attribute information appears on the sysfs but lack of it's
>>> completeness.
>>
>> A new system call makes total sense to me. I have the same concern
>> about the completeness of what's exposed in sysfs, I just don't see a
>> _route_ to completeness with sysfs itself. Thus, the minimalist
>> approach as a first step.
>
> Outside of platform-firmware-id to Linux-numa-node-id what other
> userspace API infrastructure does the kernel need to provide? It seems
> userspace enumeration of memory attributes is fully enabled once the
> firmware-to-Linux identification is established.

Which is true if the user space is required to probe the memory attribute
values for the platform-firmware-id from the platform and then request
required memory from corresponding Linux-numa-node-id via standard mm
interfaces like mbind(). But in this patch series we are not mapping
platform-firmware-id to Linux-numa-node-id. We are exporting properties
applicable to Linux nodes (Linux-numa-node-id).

Even if platform-firmware-id to Linux-numa-node-id is required it can
be done through a new file like the following. Applications can just
take the platform_id node and query platform about it's properties.

/sys/devices/system/node/nodeX/platform_id

This above interface would have been okay as its just an extension of
the existing node information on sysfs. But thats not the case with
this proposal.