Re: [PATCH] Documentation: networking: add UPF (User Plane Function in 5GC) description

[Simon Horman]

On Wed, Jun 28, 2023 at 04:53:51PM +0800, Ian Chen wrote:
> Add the content of User Plane Function (UPF),
> which was defined in 3GPP specifications since release 15.
> 
> Signed-off-by: Ian Chen <yi.chen@xxxxxxxxxx>

Hi Ian,

I'm assuming this is targeted at 'net-next', as opposed to 'net',
which is for fixes (I'm never sure with documentation). In any case,
the target tree should be included in the subject.

	Subject: [PATCH net-next v2] ...

If it is for net-next, then please repost when net-next reopens after July 10th.

Link: https://www.kernel.org/doc/html/next/process/maintainer-netdev.html#development-cycle


> ---
>  Documentation/networking/gtp.rst | 10 +++++-----
>  1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/Documentation/networking/gtp.rst b/Documentation/networking/gtp.rst
> index 9a7835cc1..c02aa34c4 100644
> --- a/Documentation/networking/gtp.rst
> +++ b/Documentation/networking/gtp.rst
> @@ -31,12 +31,12 @@ payload, such as LLC/SNDCP/RLC/MAC.
>  
>  At some network element inside the cellular operator infrastructure
>  (SGSN in case of GPRS/EGPRS or classic UMTS, hNodeB in case of a 3G
> -femtocell, eNodeB in case of 4G/LTE), the cellular protocol stacking
> +femtocell, eNodeB in case of 4G/LTE, gNobeB in case of 5G), the cellular protocol stacking
>  is translated into GTP *without breaking the end-to-end tunnel*.  So
>  intermediate nodes just perform some specific relay function.

I think it would be best to update the line-wrapping to keep lines
less than 80 columns wide.

>  
> -At some point the GTP packet ends up on the so-called GGSN (GSM/UMTS)
> -or P-GW (LTE), which terminates the tunnel, decapsulates the packet
> +At some point the GTP packet ends up on the so-called GGSN (GSM/UMTS),
> +P-GW (LTE), or UPF (5G), which terminates the tunnel, decapsulates the packet
>  and forwards it onto an external packet data network.  This can be
>  public internet, but can also be any private IP network (or even
>  theoretically some non-IP network like X.25).
> @@ -60,7 +60,7 @@ payload, called GTP-U.  It does not implement the 'control plane',
>  which is a signaling protocol used for establishment and teardown of
>  GTP tunnels (GTP-C).
>  
> -So in order to have a working GGSN/P-GW setup, you will need a
> +So in order to have a working GGSN/P-GW/UPF setup, you will need a
>  userspace program that implements the GTP-C protocol and which then
>  uses the netlink interface provided by the GTP-U module in the kernel
>  to configure the kernel module.
> @@ -162,7 +162,7 @@ Local GTP-U entity and tunnel identification
>  GTP-U uses UDP for transporting PDU's. The receiving UDP port is 2152
>  for GTPv1-U and 3386 for GTPv0-U.
>  
> -There is only one GTP-U entity (and therefore SGSN/GGSN/S-GW/PDN-GW
> +There is only one GTP-U entity (and therefore SGSN/GGSN/S-GW/PDN-GW/UPF
>  instance) per IP address. Tunnel Endpoint Identifier (TEID) are unique
>  per GTP-U entity.
>  
> -- 
> 2.38.1

-- 
pw-bot: changes-requested