DPDK patches and discussions
 help / color / mirror / Atom feed
From: Maxime Coquelin <maxime.coquelin@redhat.com>
To: Yuanhan Liu <yliu@fridaylinux.org>
Cc: "Wang, Wei 5. (NSB - CN/Hangzhou)" <wei.5.wang@nokia-sbell.com>,
	"dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] The limitation of VHOST_MEMORY_MAX_NREGIONS in virtio-user with vhost-user backend
Date: Wed, 6 Dec 2017 19:22:45 +0100	[thread overview]
Message-ID: <b5c076e6-bd40-f66e-af9f-25aaa26a45f1@redhat.com> (raw)
In-Reply-To: <20171206140652.GB17112@yliu-dev>



On 12/06/2017 03:06 PM, Yuanhan Liu wrote:
> On Wed, Dec 06, 2017 at 12:23:19PM +0100, Maxime Coquelin wrote:
>> Hi,
>>
>> On 12/05/2017 07:58 AM, Wang, Wei 5. (NSB - CN/Hangzhou) wrote:
>>>
>>>
>>> Hi, all
>>>
>>> In DPDK document, it it described that Virtio in containers Cannot work when there are more than VHOST_MEMORY_MAX_NREGIONS(8) hugepages. In another word, do not use 2MB hugepage so far. Do you know the reason of this limitation?
>>
>> What document are you referring to?
>> Actually, a single region is composed of several hugepages, be it 1G or
>> 2M. I use 2M hugepages daily without issues.
> 
> He was talking about virtio-user.

Ok, thanks for the correction.

The lookup time for every translation will be very costly if we have one
region per 2MB page.

> 
> 	--yliu
>>
>> The limitation comes from the Vhost-user protocol spec. I know there are
>> discussions to enlarge it, but it would require a new protocol flag not
>> to break existing implementations, so no workaround possible.
>>
>>> In my envirionment, the pdpe1gb is not set in cpu flag, so hugepage can't be set to 1GB size. The hugepage number shall be more than 8
>>>
>>>
>>> So is there any solution or workaround to fix this limitation? Or change dpdk code to fix this limitation?
>>
>> Have you tried to use 2MB pages and faced the NREGIONS > 8 case?
>> If so, could you tell us more about your setup (QEMU version, VM RAM
>> size, etc...)?
>>
>> Cheers,
>> Maxime
>>>
>>>
>>> Vivian
>>> seat:      19009
>>> phone: 13738006921
>>>
>>>
>>>

      reply	other threads:[~2017-12-06 18:22 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-12-05  6:58 Wang, Wei 5. (NSB - CN/Hangzhou)
2017-12-05 14:41 ` Yuanhan Liu
2017-12-06  2:55   ` Tan, Jianfeng
2017-12-06  2:59     ` Wang, Wei 5. (NSB - CN/Hangzhou)
2017-12-06 11:23 ` Maxime Coquelin
2017-12-06 14:06   ` Yuanhan Liu
2017-12-06 18:22     ` Maxime Coquelin [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=b5c076e6-bd40-f66e-af9f-25aaa26a45f1@redhat.com \
    --to=maxime.coquelin@redhat.com \
    --cc=dev@dpdk.org \
    --cc=wei.5.wang@nokia-sbell.com \
    --cc=yliu@fridaylinux.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).