Re: [PATCH net-next 08/13] net: dsa: hellcreek: deny tc-taprio changes to per-tc max SDU

[Vladimir Oltean]

On Wed, Sep 14, 2022 at 08:13:53PM +0200, Kurt Kanzenbach wrote:
> I'd rather like to see that feature implemented :-). Something like below.
> 
> From 3d44683979bf50960125fa3005b1142af61525c7 Mon Sep 17 00:00:00 2001
> From: Kurt Kanzenbach <kurt@xxxxxxxxxxxxx>
> Date: Wed, 14 Sep 2022 19:51:40 +0200
> Subject: [PATCH] net: dsa: hellcreek: Offload per-tc max SDU from tc-taprio
> 
> Add support for configuring the max SDU per priority and per port. If not
> specified, keep the default.
> 
> Signed-off-by: Kurt Kanzenbach <kurt@xxxxxxxxxxxxx>
> ---

Nice :) Do you also want the iproute2 patch, so you can test it?

>  drivers/net/dsa/hirschmann/hellcreek.c | 61 +++++++++++++++++++++++---
>  drivers/net/dsa/hirschmann/hellcreek.h |  7 +++
>  2 files changed, 61 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/net/dsa/hirschmann/hellcreek.c b/drivers/net/dsa/hirschmann/hellcreek.c
> index 5ceee71d9a25..1b22710e1641 100644
> --- a/drivers/net/dsa/hirschmann/hellcreek.c
> +++ b/drivers/net/dsa/hirschmann/hellcreek.c
> @@ -128,6 +128,16 @@ static void hellcreek_select_prio(struct hellcreek *hellcreek, int prio)
>  	hellcreek_write(hellcreek, val, HR_PSEL);
>  }
>  
> +static void hellcreek_select_port_prio(struct hellcreek *hellcreek, int port,
> +				       int prio)
> +{
> +	u16 val = port << HR_PSEL_PTWSEL_SHIFT;
> +
> +	val |= prio << HR_PSEL_PRTCWSEL_SHIFT;
> +
> +	hellcreek_write(hellcreek, val, HR_PSEL);
> +}
> +
>  static void hellcreek_select_counter(struct hellcreek *hellcreek, int counter)
>  {
>  	u16 val = counter << HR_CSEL_SHIFT;
> @@ -1537,6 +1547,42 @@ hellcreek_port_prechangeupper(struct dsa_switch *ds, int port,
>  	return ret;
>  }
>  
> +static void hellcreek_setup_maxsdu(struct hellcreek *hellcreek, int port,
> +				   const struct tc_taprio_qopt_offload *schedule)
> +{
> +	int tc;
> +
> +	for (tc = 0; tc < 8; ++tc) {
> +		u16 val;
> +
> +		if (!schedule->max_sdu[tc])
> +			continue;
> +
> +		hellcreek_select_port_prio(hellcreek, port, tc);
> +
> +		val = (schedule->max_sdu[tc] & HR_PTPRTCCFG_MAXSDU_MASK)
> +			<< HR_PTPRTCCFG_MAXSDU_SHIFT;
> +
> +		hellcreek_write(hellcreek, val, HR_PTPRTCCFG);

So the maxSDU hardware register tracks exactly the L2 payload size, like
the software variable does, or does it include the Ethernet header size
and/or FCS?

> +	}
> +}