DPDK patches and discussions
 help / color / mirror / Atom feed
From: Jerin Jacob <jerinjacobk@gmail.com>
To: Prashant Upadhyaya <praupadhyaya@gmail.com>
Cc: Bruce Richardson <bruce.richardson@intel.com>, dev@dpdk.org
Subject: Re: PMD for non PCI device
Date: Tue, 6 Feb 2024 21:28:13 +0530	[thread overview]
Message-ID: <CALBAE1P1a+h-jmwcbgjohx10jxPSahLd9yRVsnoDMnzHTOtFEQ@mail.gmail.com> (raw)
In-Reply-To: <CAPBAu3VdX-Lhn1YgHN5M1VCFSfv8mFjgo88Zv8zJx5kZyDcv4Q@mail.gmail.com>

On Tue, Feb 6, 2024 at 9:20 PM Prashant Upadhyaya
<praupadhyaya@gmail.com> wrote:
>
> On Tue, 6 Feb 2024 at 19:43, Bruce Richardson
> <bruce.richardson@intel.com> wrote:
> >
> > On Tue, Feb 06, 2024 at 07:36:16PM +0530, Prashant Upadhyaya wrote:
> > > Hi,
> > >
> > > I have a usecase where I have to evaluate writing a DPDK PMD for a non
> > > PCI/e device doing the ethernet packet i/o.
> > >
> > > Wanted to know if the above usecase is supported by DPDK infra and any
> > > pointers on how one should go about writing a PMD for such a usecase
> > > if supported. Would appreciate any inputs.
> > >
> > Hi,
> >
> > yes, such a usecase is supported, but the specifics of how to go about it
> > will vary depending on the type of PMD it is. DPDK already supports a range
> > of other types of PMD, for emulated, or SW backed PMDs, e.g. net/pcap
> > driver, and drivers for various SoCs which don't use PCI. For the case
> > where the PMD is backed by real hardware (or an emulated device that
> > appears to a VM as a piece of hardware), you may want to consider writing a
> > "bus" driver for DPDK to support probing of the device. For non-HW
> > devices, the "vdev" bus may be what you want to use, where probing is not
> > done and devices are created in response to cmdline arguments on init, or
> > via C APIs later in the app.
> >
> > Regards,
> > /Bruce
>
> Thanks Bruce, this is helpful. Is there any PMD in DPDK code that you
> can refer me to for any SoC which does not use PCI (the usecase of
> backing by real hardware), that would be great to follow.


See drivers/bus/*

Based on your description, vfio-platorm will be the closest
match(driver/bus/platform)


>
> Regards
> -Prashant

  reply	other threads:[~2024-02-06 15:58 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-02-06 14:06 Prashant Upadhyaya
2024-02-06 14:13 ` Bruce Richardson
2024-02-06 15:50   ` Prashant Upadhyaya
2024-02-06 15:58     ` Jerin Jacob [this message]
2024-02-14 15:19       ` Prashant Upadhyaya

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=CALBAE1P1a+h-jmwcbgjohx10jxPSahLd9yRVsnoDMnzHTOtFEQ@mail.gmail.com \
    --to=jerinjacobk@gmail.com \
    --cc=bruce.richardson@intel.com \
    --cc=dev@dpdk.org \
    --cc=praupadhyaya@gmail.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).