From: "Xie, Huawei" <huawei.xie@intel.com>
To: "Iremonger, Bernard" <bernard.iremonger@intel.com>,
Yuanhan Liu <yuanhan.liu@linux.intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] [PATCH v2 1/1] doc: correct Vhost Sample Application guide
Date: Thu, 10 Dec 2015 13:33:53 +0000 [thread overview]
Message-ID: <C37D651A908B024F974696C65296B57B4BB89AEE@SHSMSX103.ccr.corp.intel.com> (raw)
In-Reply-To: <8CEF83825BEC744B83065625E567D7C219F8ED27@IRSMSX108.ger.corp.intel.com>
> -----Original Message-----
> From: Iremonger, Bernard
> Sent: Thursday, December 10, 2015 9:20 PM
> To: Xie, Huawei; Yuanhan Liu
> Cc: dev@dpdk.org
> Subject: RE: [dpdk-dev] [PATCH v2 1/1] doc: correct Vhost Sample
> Application guide
>
> Hi Huawei,
>
> <snip>
>
> > > > I don't think that's a right "correction": vhost-swtich would fail
> > > > to
> > > > start:
> > > >
> > > > EAL: Error - exiting with code: 1
> > > > Cause: Cannot create mbuf pool
> > > >
> > > > As vhost-switch creates more mbuf than 1024M can hold.
> > > >
> > > > However, I do think that adding this option is necessary, otherwise,
> > > > all hugepages will be allocated to vhost-switch, leaving no memory
> > > > for
> > > starting
> > > > VM at all. (And this is kind of informatin you need to put into
> > > > commit
> > > log).
> > > >
> > > > And limiting it to "1024M" is reasonable as well, so that we can run
> > > > it
> > > on a
> > > > desktop (or laptop) without too many memories. You just need fix
> > > > vhost- switch example to not allocate too many mbufs by default.
> > > >
> > > > --yliu
> > >
> > > Yes the --socket-mem <number> option is necessary.
> > > I will add a note that a value of <number> may need to be greater
> > > than 1024.
> > > I will submit a patch to vhost-switch to reduce the number of mbufs.
> > >
> > I recall we have to allocate mbufs for each queue rather than used queues
> > only, so memory consumption depends on queue number. After that issue is
> > fixed, I think 1024MB is enough. For the time being, you could
> temporarily
> > use 2048M and add an explanation.
> >
>
> I have sent a v3 of this patch which includes a note that the value of 1024
> may have to be increased.
> I would prefer to keep the value of 1024.
>
For FVL with more queues, I recall 1024mb isn't enough, but it is ok with a note. :).
> Regards,
>
> Bernard.
next prev parent reply other threads:[~2015-12-10 13:34 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-12-09 12:35 [dpdk-dev] [PATCH " Bernard Iremonger
2015-12-09 13:36 ` Xie, Huawei
2015-12-09 14:43 ` Mcnamara, John
2015-12-09 16:10 ` Iremonger, Bernard
2015-12-09 14:50 ` Mcnamara, John
2015-12-09 15:09 ` Thomas Monjalon
2015-12-09 16:14 ` Iremonger, Bernard
2015-12-10 6:40 ` Yuanhan Liu
2015-12-09 17:18 ` [dpdk-dev] [PATCH v2 " Bernard Iremonger
2015-12-10 6:52 ` Yuanhan Liu
2015-12-10 9:58 ` Iremonger, Bernard
2015-12-10 12:10 ` Xie, Huawei
2015-12-10 13:19 ` Iremonger, Bernard
2015-12-10 13:33 ` Xie, Huawei [this message]
2015-12-10 11:23 ` [dpdk-dev] [PATCH v3] " Bernard Iremonger
2015-12-14 0:23 ` Thomas Monjalon
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=C37D651A908B024F974696C65296B57B4BB89AEE@SHSMSX103.ccr.corp.intel.com \
--to=huawei.xie@intel.com \
--cc=bernard.iremonger@intel.com \
--cc=dev@dpdk.org \
--cc=yuanhan.liu@linux.intel.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).