From: Vaibhav Jain <vaibhav@linux.ibm.com>
To: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>,
linux-nvdimm@lists.01.org
Cc: Michael Ellerman <mpe@ellerman.id.au>,
Shivaprasad G Bhat <sbhat@linux.ibm.com>,
stable@vger.kernel.org
Subject: Re: [PATCH v2] libnvdimm/region: Update nvdimm_has_flush() to handle ND_REGION_ASYNC
Date: Thu, 08 Apr 2021 16:18:03 +0530 [thread overview]
Message-ID: <87tuohgiho.fsf@vajain21.in.ibm.com> (raw)
In-Reply-To: <5e64778d-bf48-9f10-7d3d-5e530e5db590@linux.ibm.com>
"Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com> writes:
> On 4/6/21 5:07 PM, Vaibhav Jain wrote:
>> Hi Aneesh,
>> Thanks for looking into this patch.
>>
>> "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com> writes:
>>
>>> Vaibhav Jain <vaibhav@linux.ibm.com> writes:
>>>
>>>> In case a platform doesn't provide explicit flush-hints but provides an
>>>> explicit flush callback via ND_REGION_ASYNC region flag, then
>>>> nvdimm_has_flush() still returns '0' indicating that writes do not
>>>> require flushing. This happens on PPC64 with patch at [1] applied,
>>>> where 'deep_flush' of a region was denied even though an explicit
>>>> flush function was provided.
>>>>
>>>> Similar problem is also seen with virtio-pmem where the 'deep_flush'
>>>> sysfs attribute is not visible as in absence of any registered nvdimm,
>>>> 'nd_region->ndr_mappings == 0'.
>>>>
>>>> Fix this by updating nvdimm_has_flush() adding a condition to
>>>> nvdimm_has_flush() testing for ND_REGION_ASYNC flag on the region
>>>> and see if a 'region->flush' callback is assigned. Also remove
>>>> explicit test for 'nd_region->ndr_mapping' since regions can be marked
>>>> 'ND_REGION_SYNC' without any explicit mappings as in case of
>>>> virtio-pmem.
>>>
>>> Do we need to check for ND_REGION_ASYNC? What if the backend wants to
>>> provide a synchronous dax region but with different deep flush semantic
>>> than writing to wpq flush address?
>>> ie,
>>
>> For a synchronous dax region, writes arent expected to require any
>> flushing (or deep-flush) so this function should ideally return '0' in
>> such a case. Hence I had added the test for ND_REGION_ASYNC region flag.
>>
>
> that is not correct. For example, we could ideally move the wpq flush as
> an nd_region->flush callback for acpi or we could implement a deep flush
> for a synchronous dax region exposed by papr_scm driver that ensures
> stores indeed reached the media managed by the hypervisor.
Sure, makes sense now. I have sent out a v3 of this patch with this
addressed.
>
> -aneesh
--
Cheers
~ Vaibhav
_______________________________________________
Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org
To unsubscribe send an email to linux-nvdimm-leave@lists.01.org
prev parent reply other threads:[~2021-04-08 10:48 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-04-06 3:22 [PATCH v2] libnvdimm/region: Update nvdimm_has_flush() to handle ND_REGION_ASYNC Vaibhav Jain
2021-04-06 5:13 ` Aneesh Kumar K.V
2021-04-06 11:37 ` Vaibhav Jain
2021-04-06 12:05 ` Aneesh Kumar K.V
2021-04-08 10:48 ` Vaibhav Jain [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87tuohgiho.fsf@vajain21.in.ibm.com \
--to=vaibhav@linux.ibm.com \
--cc=aneesh.kumar@linux.ibm.com \
--cc=linux-nvdimm@lists.01.org \
--cc=mpe@ellerman.id.au \
--cc=sbhat@linux.ibm.com \
--cc=stable@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).