From: "Jia.Sui(贾睢)" <Jia.Sui@advantech.com.cn>
To: Antti Kantee <pooka@iki.fi>, Adeel Amin <adeel_amin@hotmail.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] DPDK on bare-metal machine
Date: Wed, 19 Jun 2013 16:34:23 +0800 [thread overview]
Message-ID: <581E2E1085FAEF45B48CF8A139824CF804E5D6D419@ACNMAIL1.ACN.ADVANTECH.CORP> (raw)
In-Reply-To: <51C16BA7.4060001@iki.fi>
Hi Antti
As the DPDK release note said:
The Intel®DPDK Bare-metal features will now be provided off-roadmap to a limited set of customers on a case-by-case basis at this time.
Please contact your Intel representative if you wish to be considered as an evaluation candidate for the Baremetal features.
I haven't use bare-metal before. Sorry..
BRs
Jia Sui
-----Original Message-----
From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Antti Kantee
Sent: Wednesday, June 19, 2013 4:28 PM
To: Adeel Amin
Cc: dev@dpdk.org
Subject: Re: [dpdk-dev] DPDK on bare-metal machine
On 19.06.2013 09:20, Adeel Amin wrote:
> Hello,
>
> Can anyone tell me that how I can run a DPDK application as a bare-metal (i.e. without Linux support). I've tried to run rump kernel TCP/IP stack by Antti Kantee but I'm unable to figure out that how I'll be using it to run stand alone on x86 CPU. Currently I'm running the rump kernel as a Linux application.
Hi,
As far as I've been able to gather, the open source version of DPDK does
not include bare-metal support. However, I'm not very knowledgeable on
DPDK, especially on code which I cannot read myself, so maybe someone
else can answer that better.
I do know something about rump kernels, though. One of main future use
cases I see is indeed to have them run on bare metal and therefore allow
to continue use essential kernel-only implemented features but still get
rid of the prehistoric OS overhead layer. I have good reason to suspect
that running rump kernels this would be easy, but as far as I know, no
one has done the necessary work, at least not for x86. Unless the
bare-metal version of DPDK provides a suitable portability layer, you'd
be looking at implementing the rump kernel hypervisor interface for bare
metal (or bare firmware ?-). In such a case, you might find using that
using also the device layer from a rump kernel is the shortest path to
initial success, with migration to DPDK only after reaching stability in
that setup.
But, that's getting quite off-topic for this list. Contact me off-list
if you are serious about putting effort into the above.
- antti
next prev parent reply other threads:[~2013-06-19 8:34 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <BAY177-W41B0EB5E47C4EE7099BDCF28D0@phx.gbl>
2013-06-19 7:20 ` Adeel Amin
2013-06-19 8:28 ` Antti Kantee
2013-06-19 8:34 ` Jia.Sui [this message]
2013-06-19 14:57 ` Stephen Hemminger
2013-06-20 5:32 ` Stephen Hemminger
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=581E2E1085FAEF45B48CF8A139824CF804E5D6D419@ACNMAIL1.ACN.ADVANTECH.CORP \
--to=jia.sui@advantech.com.cn \
--cc=adeel_amin@hotmail.com \
--cc=dev@dpdk.org \
--cc=pooka@iki.fi \
/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).