[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH v2 0/2] kvm "fake DAX" device
From: |
Dan Williams |
Subject: |
Re: [Qemu-devel] [PATCH v2 0/2] kvm "fake DAX" device |
Date: |
Sat, 13 Oct 2018 09:06:07 -0700 |
On Fri, Oct 12, 2018 at 10:00 PM Pankaj Gupta <address@hidden> wrote:
>
> This patch series has implementation for "fake DAX".
> "fake DAX" is fake persistent memory(nvdimm) in guest
> which allows to bypass the guest page cache. This also
> implements a VIRTIO based asynchronous flush mechanism.
Can we stop calling this 'fake DAX', because it isn't 'DAX' and it's
starting to confuse people. This enabling is effectively a
host-page-cache-passthrough mechanism not DAX. Let's call the whole
approach virtio-pmem, and leave DAX out of the name to hopefully
prevent people from wondering why some DAX features are disabled with
this driver. For example MAP_SYNC is not compatible with this
approach.
Additional enabling is need to disable MAP_SYNC in the presence of a
virtio-pmem device. See the rough proposal here:
https://lkml.org/lkml/2018/4/25/756
Re: [Qemu-devel] [PATCH v2 0/2] kvm "fake DAX" device,
Dan Williams <=