I asked the following in the Slack channel, responding to Stephen and
Thomas so asking in this thread as well to confirm:

To confirm, the three steps in 3.1.1 Memory Setup: Reservering Hugepages at
http://170.249.220.94/getting_started_guide/system_setup.html#system-setup-for-linux
can be replaced with the following:

Memory Setup: Reserving Hugepages
To simplify the hugepages setup, use the `dpdk-hugepages` script:
sudo dpdk-hugepages --setup

To confirm, the instructions in 3.1.2 Device Setup: VFIO at
http://170.249.220.94/getting_started_guide/system_setup.html#device-setup-vfio
can be replaced with the following:

Device Setup: VFIO
For VFIO setup, please refer to the `Kernel Documentation for VFIO <
https://www.kernel.org/doc/html/latest/driver-api/vfio.html>`_.

Thanks!
David Young
Professional Copywriter/Technical Writer
Young Copy
+1 (678) 500-9550
https://www.youngcopy.com


On Mon, Sep 25, 2023 at 8:22 AM Bruce Richardson <bruce.richard...@intel.com>
wrote:

> On Mon, Sep 25, 2023 at 12:31:34PM +0100, Ferruh Yigit wrote:
> > On 9/20/2023 4:48 PM, David Young wrote:
> > > ---
> > >  .../getting_started_guide/system_setup.rst    | 195 ++++++++++++++++++
> > >  1 file changed, 195 insertions(+)
> > >  create mode 100644 doc/guides/getting_started_guide/system_setup.rst
> > >
> > > diff --git a/doc/guides/getting_started_guide/system_setup.rst
> b/doc/guides/getting_started_guide/system_setup.rst
> > > new file mode 100644
> > > index 0000000000..fa9d249ec7
> > > --- /dev/null
> > > +++ b/doc/guides/getting_started_guide/system_setup.rst
> > > @@ -0,0 +1,195 @@
> > > +..  SPDX-License-Identifier: BSD-3-Clause
> > > +    Copyright(c) 2010-2025 Intel Corporation.
> > > +
> > > +.. _memory_setup:
> > > +
> > > +.. |reg| unicode:: U+000AE
> > > +
> > > +Setting up a System to Run DPDK Applications
> > > +============================================
> > > +
> > > +This section provides step-by-step instructions for setting up your
> system to run DPDK applications. It covers system configurations for Linux,
> FreeBSD, and Windows. Each section details the necessary memory and device
> setups for these operating systems.
> > > +
> > > +Navigate to the section that corresponds to your operating system to
> begin the setup process.
> > > +
> >
> > Not sure above sentences adds value.
> >
> >
> >
> > > +.. contents:: Table of Contents
> > > +   :local:
> > > +
> > > +System Setup for Linux
> > > +----------------------
> > > +
> > > +Memory Setup: Reserving Hugepages
> > > +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > +
> > > +For Linux, DPDK requires hugepages be reserved for its use on the
> system. To check if hugepages are are on your system, you can run the
> following command::
> > > +
> > > +        grep HugePages_Total /proc/meminfo
> > > +
> > > +If hugepages are not reserved, you will need to reserve them by
> following these steps:
> > > +
> > > +1. Determine the number of hugepages you want to allocate. For
> example, to allocate 1024 hugepages of 2MB each, you can use the following
> command::
> > > +
> > > +        echo 1024 | sudo tee
> /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages
> > > +
> > > +2. To make the hugepages configuration persistent across reboots, add
> the following line to your `/etc/sysctl.conf` file, adjusting the number of
> hugepages as needed::
> > > +
> > > +        vm.nr_hugepages = 1024
> > > +
> > > +3. Most distributions make hugepages available via `/dev/hugepages`,
> so this step may not be necessary. If you need to manually mount the
> hugepages filesystem, add the following line to your `/etc/fstab` file::
> > > +
> > > +        nodev /mnt/huge hugetlbfs defaults 0 0
> > > +
> > > +   Then, create the mount directory and mount the filesystem::
> > > +
> > > +        mkdir -p /mnt/huge
> > > +        mount -a
> > > +
> > >
> >
> >
> > We have './usertools/dpdk-hugepages.py' script for this, which I am
> > using regularly.
> >
> > Script is wrapper to what described above, so I think good to explain
> > basics, but also may worth mentioning from script, it is more user
> > friendly than above instructions.
> >
> Actually, if script can do all the basics - something I believe Stephen and
> Thomas also pointed out on a slack review earlier - then we should just
> drop the manual steps from here. If we think the script is sufficiently
> robust, we can drop them from the doc entirely, or if not, move them to an
> appendix which we reference here.
>
> /Bruce
>

Reply via email to