From: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>
To: dan.j.williams@intel.com
Cc: linux-mm@kvack.org, Vaibhav Jain <vaibhav@linux.ibm.com>,
linuxppc-dev@lists.ozlabs.org, linux-nvdimm@lists.01.org
Subject: Re: [PATCH] mm/nvdimm: Pick the right alignment default when creating dax devices
Date: Sun, 19 May 2019 14:25:33 +0530 [thread overview]
Message-ID: <87sgtaddru.fsf@linux.ibm.com> (raw)
In-Reply-To: <de5cbe7d-bd47-6793-1f1a-2274c5c59eb5@linux.ibm.com>
Hi Dan,
"Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com> writes:
> On 5/17/19 8:19 PM, Vaibhav Jain wrote:
>> Hi Aneesh,
>>
....
>>
>>> + /*
>>> + * Check whether the we support the alignment. For Dax if the
>>> + * superblock alignment is not matching, we won't initialize
>>> + * the device.
>>> + */
>>> + if (!nd_supported_alignment(align) &&
>>> + memcmp(pfn_sb->signature, DAX_SIG, PFN_SIG_LEN)) {
>> Suggestion to change this check to:
>>
>> if (memcmp(pfn_sb->signature, DAX_SIG, PFN_SIG_LEN) &&
>> !nd_supported_alignment(align))
>>
>> It would look a bit more natural i.e. "If the device has dax signature and alignment is
>> not supported".
>>
>
> I guess that should be !memcmp()? . I will send an updated patch with
> the hash failure details in the commit message.
>
We need clarification on what the expected failure behaviour should be.
The nd_pmem_probe doesn't really have a failure behaviour in this
regard. For example.
I created a dax device with 16M alignment
{
"dev":"namespace0.0",
"mode":"devdax",
"map":"dev",
"size":"9.98 GiB (10.72 GB)",
"uuid":"ba62ef22-ebdf-4779-96f5-e6135383ed22",
"raw_uuid":"7b2492f9-7160-4ee9-9c3d-2f547d9ef3ee",
"daxregion":{
"id":0,
"size":"9.98 GiB (10.72 GB)",
"align":16777216,
"devices":[
{
"chardev":"dax0.0",
"size":"9.98 GiB (10.72 GB)"
}
]
},
"align":16777216,
"numa_node":0,
"supported_alignments":[
65536,
16777216
]
}
Now what we want is to fail the initialization of the device when we
boot a kernel that doesn't support 16M page size. But with the
nd_pmem_probe failure behaviour we now end up with
[
{
"dev":"namespace0.0",
"mode":"fsdax",
"map":"mem",
"size":10737418240,
"uuid":"7b2492f9-7160-4ee9-9c3d-2f547d9ef3ee",
"blockdev":"pmem0"
}
]
So it did fallthrough the
/* if we find a valid info-block we'll come back as that personality */
if (nd_btt_probe(dev, ndns) == 0 || nd_pfn_probe(dev, ndns) == 0
|| nd_dax_probe(dev, ndns) == 0)
return -ENXIO;
/* ...otherwise we're just a raw pmem device */
return pmem_attach_disk(dev, ndns);
Is it ok if i update the code such that we don't do that default
pmem_atach_disk if we have a label area?
-aneesh
next prev parent reply other threads:[~2019-05-19 8:57 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-05-14 2:54 [PATCH] mm/nvdimm: Pick the right alignment default when creating dax devices Aneesh Kumar K.V
2019-05-17 14:49 ` Vaibhav Jain
2019-05-17 15:17 ` Aneesh Kumar K.V
2019-05-19 8:55 ` Aneesh Kumar K.V [this message]
2019-05-19 16:30 ` Dan Williams
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=87sgtaddru.fsf@linux.ibm.com \
--to=aneesh.kumar@linux.ibm.com \
--cc=dan.j.williams@intel.com \
--cc=linux-mm@kvack.org \
--cc=linux-nvdimm@lists.01.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=vaibhav@linux.ibm.com \
/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).