Re: [PATCH v4 0/3] nvdimm: Enable sync-dax property for nvdimm

From: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>
To: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: Peter Maydell <peter.maydell@linaro.org>,
	Shivaprasad G Bhat <sbhat@linux.ibm.com>,
	"Michael S. Tsirkin" <mst@redhat.com>,
	Qemu Developers <qemu-devel@nongnu.org>,
	linux-nvdimm <linux-nvdimm@lists.01.org>,
	Markus Armbruster <armbru@redhat.com>,
	bharata@linux.vnet.ibm.com,
	Haozhong Zhang <haozhong.zhang@intel.com>,
	Eduardo Habkost <ehabkost@redhat.com>,
	richard.henderson@linaro.org, Greg Kurz <groug@kaod.org>,
	kvm-ppc@vger.kernel.org, qemu-arm@nongnu.org,
	Stefan Hajnoczi <stefanha@redhat.com>,
	Igor Mammedov <imammedo@redhat.com>,
	Dan Williams <dan.j.williams@intel.com>,
	kwangwoo.lee@sk.com, David Gibson <david@gibson.dropbear.id.au>,
	Xiao Guangrong <xiaoguangrong.eric@gmail.com>,
	shameerali.kolothum.thodi@huawei.com, shivaprasadbhat@gmail.com,
	qemu-ppc@nongnu.org, Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [PATCH v4 0/3] nvdimm: Enable sync-dax property for nvdimm
Date: Tue, 4 May 2021 14:32:50 +0530	[thread overview]
Message-ID: <023e584a-6110-4d17-7fec-ca715226f869@linux.ibm.com> (raw)
In-Reply-To: <CAM9Jb+g8bKF0Z7za4sZpc2tZ01Sp4c4FEaV65He8w1+QOL3_yw@mail.gmail.com>

On 5/4/21 11:13 AM, Pankaj Gupta wrote:
....

>>
>> What this patch series did was to express that property via a device
>> tree node and guest driver enables a hypercall based flush mechanism to
>> ensure persistence.
> 
> Would VIRTIO (entirely asynchronous, no trap at host side) based
> mechanism is better
> than hyper-call based? Registering memory can be done any way. We
> implemented virtio-pmem
> flush mechanisms with below considerations:
> 
> - Proper semantic for guest flush requests.
> - Efficient mechanism for performance pov.
> 

sure, virio-pmem can be used as an alternative.

> I am just asking myself if we have platform agnostic mechanism already
> there, maybe
> we can extend it to suit our needs? Maybe I am missing some points here.
> 

What is being attempted in this series is to indicate to the guest OS 
that the backing device/file used for emulated nvdimm device cannot 
guarantee the persistence via cpu cache flush instructions.

>>>> On PPC, the default is "sync-dax=writeback" - so the ND_REGION_ASYNC
>>>>
>>>> is set for the region and the guest makes hcalls to issue fsync on the host.
>>>>
>>>>
>>>> Are you suggesting me to keep it "unsafe" as default for all architectures
>>>>
>>>> including PPC and a user can set it to "writeback" if desired.
>>>
>>> No, I am suggesting that "sync-dax" is insufficient to convey this
>>> property. This behavior warrants its own device type, not an ambiguous
>>> property of the memory-backend-file with implicit architecture
>>> assumptions attached.
>>>
>>
>> Why is it insufficient?  Is it because other architectures don't have an
>> ability express this detail to guest OS? Isn't that an arch limitations?

-aneesh