From: Jakub Kicinski <kuba@kernel.org>
To: Kurt Kanzenbach <kurt@linutronix.de>
Cc: Maxime Chevallier <maxime.chevallier@bootlin.com>,
Andrew Lunn <andrew+netdev@lunn.ch>,
"David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@google.com>,
Paolo Abeni <pabeni@redhat.com>,
Maxime Coquelin <mcoquelin.stm32@gmail.com>,
Alexandre Torgue <alexandre.torgue@foss.st.com>,
netdev@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com,
linux-arm-kernel@lists.infradead.org,
Lorenzo Bianconi <lorenzo@kernel.org>
Subject: Re: [PATCH net-next] net: stmmac: stm32: Set TSO/TBS Tx queues default settings
Date: Tue, 18 Aug 2026 07:53:35 -0700 [thread overview]
Message-ID: <20260818075335.48936b08@kernel.org> (raw)
In-Reply-To: <87mrujzzjv.fsf@jax.kurt.home>
On Tue, 18 Aug 2026 08:57:08 +0200 Kurt Kanzenbach wrote:
> On Mon Aug 17 2026, Jakub Kicinski wrote:
> > On Wed, 12 Aug 2026 08:28:57 +0200 Kurt Kanzenbach wrote:
> >> TSO and TBS cannot coexist. Use the first queue with TSO and the rest for
> >> TBS. Tx queues with TBS can support etf qdisc hw offload. This is done
> >> similar to dwmac-imx and dwmac-intel.
> >
> > Can you explain your use case?
>
> I want to use ETF Qdisc with hardware offload, which is currently not
> possible on the stm32mp2.
Do you need multiple queues for ETF/normal traffic?
Or one ETF and one "normal"?
How do you sort the traffic between the queues?
> > And how many queues the device has in total?
>
> The stm32mp2 has four Tx and two Rx queues.
>
> > You say "TSO and TBS cannot coexist" but can any queue on your
> > platform be configured to support either feature?
>
> I think so. The data sheet says: "Do not enable time-based scheduling for
> channels on which the TSO feature is enabled.". But, I didn't find any
> limitations on what queue/channel can enable TBS.
👍️
> > If yes why are we configuring this statically instead of making
> > appropriate configuration based on qdisc or some other uAPI knob?
>
> It seems like a static configuration in the driver today. I basically
> followed the same convention as dwmac-imx, dwmac-intel, dwmac-mediatek
> and dwmac-qcom-ethqos. Only dwmac-socfpga does it differently.
>
> Tx Launch Time requires a different DMA descriptor layout. Currently
> tc_setup_etf() just returns -EINVAL if the DMA configuration is not
> setup appropriately. I guess a dynamic configuration requires to change
> the DMA config and perform a full release/open cycle.
release/open is not ideal but still better than hardcoding?
next prev parent reply other threads:[~2026-08-18 14:53 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-12 6:28 [PATCH net-next] net: stmmac: stm32: Set TSO/TBS Tx queues default settings Kurt Kanzenbach
2026-08-17 17:33 ` Jakub Kicinski
2026-08-18 6:57 ` Kurt Kanzenbach
2026-08-18 14:53 ` Jakub Kicinski [this message]
2026-08-19 7:35 ` Kurt Kanzenbach
2026-08-20 21:45 ` Jakub Kicinski
2026-08-21 7:22 ` Kurt Kanzenbach
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=20260818075335.48936b08@kernel.org \
--to=kuba@kernel.org \
--cc=alexandre.torgue@foss.st.com \
--cc=andrew+netdev@lunn.ch \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=kurt@linutronix.de \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-stm32@st-md-mailman.stormreply.com \
--cc=lorenzo@kernel.org \
--cc=maxime.chevallier@bootlin.com \
--cc=mcoquelin.stm32@gmail.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.