From: Luke Howard <lukeh@padl.com>
To: Andrew Lunn <andrew@lunn.ch>
Cc: netdev@vger.kernel.org, vladimir.oltean@nxp.com,
kieran@sienda.com, jcschroeder@gmail.com,
Max Hunter <max@huntershome.org>
Subject: Re: [RFC net-next 3/5] net: dsa: mv88e6xxx: MQPRIO support
Date: Tue, 26 May 2026 17:05:17 +1000 [thread overview]
Message-ID: <F4CE6A53-C253-461A-A845-DCD620D7D97A@padl.com> (raw)
In-Reply-To: <DE4F4AC6-2783-4374-A4F8-3032B32DB848@padl.com>
Hi Andrew,
> On 28 Sep 2025, at 04:14, Luke Howard <lukeh@padl.com> wrote:
>
> I guess we need a new attribute or flag when adding a MDB or FDB entry, and also a way to indicate that a TC is SRP-managed. The kernel bridge could then be modified to reject or reclassify packets that are classified into a TC with the SRP flag set, unless a MDB/FDB entry exists for the DA that also has that flag set.
I have looked into this once more (disclosure: this time, LLM-aided).
* adding a MDB entry can now take a nested MDBE_ATTR_FLAGS attribute; the only one defined is MDB_FLAGS_STREAM_RESERVED
* this is propagated to MDB_PG_FLAGS_STREAM_RESERVED in br_mdb_config’s pg_flags
* MDB_PG_FLAGS_STREAM_RESERVED is propagated, in turn, to SWITCHDEV_MDB_F_STREAM_RESERVED
I intentionally avoided FDB entries as they’re not typically used in AVB/TSN, and this would have involved more intrusive API changes (port_fdb_add takes just the address/VID as distinct parameters).
If mv88e6xxx_port_mdb_add() is called with SWITCHDEV_MDB_F_STREAM_RESERVED set, it adds the ATU entry as MV88E6XXX_G1_ATU_DATA_STATE_MC_STATIC_AVB_NRL instead of MV88E6XXX_G1_ATU_DATA_STATE_MC_STATIC. For a given destination MAC, the AVB_NRL state must match across ports (otherwise -EINVAL is returned).
For the software bridge, if a frame arrives with PCP 3 or 2 (Class A/B traffic per IEEE 802.1Q-2018 Annex L) and the destination MDB entry has at least one port group marked STREAM_RESERVED, the unflagged port groups are skipped and router ports are suppressed. MDB entries with no flagged port groups follow the existing br_multicast_flood() path unchanged. The PCP set is hard-coded for now as an example and probably needs a configuration knob.
Cheers,
Luke
next prev parent reply other threads:[~2026-05-26 7:05 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-09-27 7:07 [RFC net-next 0/5] net: dsa: mv88e6xxx: support for 802.1Qav Luke Howard
2025-09-27 7:07 ` [RFC net-next 1/5] net: dsa: mv88e6xxx: add num_tx_queues to chip info structure Luke Howard
2025-09-27 7:07 ` [RFC net-next 2/5] net: dsa: mv88e6xxx: add MV88E6XXX_G1_ATU_CTL_MAC_AVB setter Luke Howard
2025-09-27 14:29 ` Andrew Lunn
2025-09-27 7:07 ` [RFC net-next 3/5] net: dsa: mv88e6xxx: MQPRIO support Luke Howard
2025-09-27 15:12 ` Andrew Lunn
2025-09-27 16:00 ` Luke Howard
2025-09-27 17:36 ` Andrew Lunn
2025-09-27 18:14 ` Luke Howard
2026-05-26 7:05 ` Luke Howard [this message]
2025-09-27 7:07 ` [RFC net-next 4/5] net: dsa: mv88e6xxx: CBS support Luke Howard
2025-09-27 7:07 ` [RFC net-next 5/5] dt-bindings: net: dsa: mv88e6xxx: add mv88e6xxx-avb-mode property Luke Howard
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=F4CE6A53-C253-461A-A845-DCD620D7D97A@padl.com \
--to=lukeh@padl.com \
--cc=andrew@lunn.ch \
--cc=jcschroeder@gmail.com \
--cc=kieran@sienda.com \
--cc=max@huntershome.org \
--cc=netdev@vger.kernel.org \
--cc=vladimir.oltean@nxp.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