From: Akhila Kavi <akavi@axiado.com>
To: Jeremy Kerr <jk@codeconstruct.com.au>,
Sammai Chakravarthy <schakravarthy@axiado.com>,
Prasad Bolisetty <pbolisetty@axiado.com>,
Andrew Lunn <andrew+netdev@lunn.ch>,
"David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@google.com>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
Rob Herring <robh@kernel.org>,
Krzysztof Kozlowski <krzk+dt@kernel.org>,
Conor Dooley <conor+dt@kernel.org>, Tzu-Hao Wei <twei@axiado.com>,
Karthikeyan Mitran <kmitran@axiado.com>,
Matt Johnston <matt@codeconstruct.com.au>
Cc: netdev@vger.kernel.org, devicetree@vger.kernel.org,
linux-arm-kernel@lists.infradead.org,
linux-kernel@vger.kernel.org, openbmc@lists.ozlabs.org,
Adam Young <admiyo@amperemail.onmicrosoft.com>
Subject: Re: [PATCH net-next 0/2] net: mctp: Add MCTP transport for the Axiado AX3005 IRoT
Date: Mon, 16 Nov 2026 11:52:12 +0530 [thread overview]
Message-ID: <f752d203-8239-4203-aa04-884b4d7e5d80@axiado.com> (raw)
In-Reply-To: <4102c20ece135c4d94e94af4441a01425ad159fc.camel@codeconstruct.com.au>
On 02-11-2026 08:25, Jeremy Kerr wrote:
> CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.
>
>
> Hi Akhila,
>
>> The Axiado AX3005 SoC has an integrated root-of-trust (IRoT)
>> co-processor next to the application processor. Security protocols such
>> as SPDM and PLDM run between the two over MCTP. This series adds a
>> transport for that link, so the host side uses ordinary AF_MCTP
>> sockets.
> One overall thing: why not use the MCTP MMBI (DSP0284) transport spec
> for this? Sounds like this would be an exact match for the MMBI
> use-case.
Thanks for the review. To give some context first:
this link is entirely internal to the AX3005, and not for host communication.
The IRoT's MCTP endpoint is only reachable from the application subsystem on
the same die, through the mailbox and the shared-memory carve-out.
There is no hardware path to it from outside the chip:
no PCIe, SMBus/I2C, I3C or USB connection reaches the IRoT.
So this is strictly one point-to-point link between two fixed endpoints
on the same SoC.
You're right that the mechanism is close to MMBI: shared memory plus a doorbell.
We looked at DSP0282/DSP0284 and decided to keep the existing protocol,
for the reason: Interoperability, which is the main benefit of a standard binding,
doesn't apply here. Both ends are fixed on the same die, and no third-party MCTP
device can ever be on the other end of this link, so carrying it as MMBI would
add nothing.For that reason the driver stays a narrow vendor binding:
It uses MCTP_PHYS_BINDING_VENDOR and matches only the axiado,ax3005-mctp-irot compatible.
It isn't meant as a general shared-memory transport, and it doesn't claim to be one.
>
> [Also: CCing Adam who is working on PCC, which may have some overlap
> with the mailbox protocol used here]
Thanks. The two look alike: one shared-memory window per direction, a doorbell,
and a completion before the window is reused. The differences:
AX3005 is a DT platform with no ACPI, so PCC itself isn't an option here.
Discovery is in-band. At probe the IRoT replies to a mailbox message with its window
addresses and MTU limit.
Each packet carries a counter, so a resent doorbell is ACKed again without the packet
being delivered twice.
Adam, if you see something both drivers could share, I'm happy to look at it.
>
> And finally: how confident are you that the peer will never be able to
> reassemble? Even if the peer's own MCTP stack may not reassemble,
> there's no chance of it bridging to other endpoints in future?
>
> I have a few (minor) comments on the implementation, but would like to
> sort out the overall protocol usage first.
The IRoT can't bridge anything to this link. As above, nothing outside the chip has a
hardware path to the IRoT's MCTP endpoint, so the application processor is its only MCTP peer.
Defaulting the netdev MTU to the negotiated limit only makes sure that today's peer gets
each message in a single packet. It isn't meant to stop anyone from using a smaller MTU.
min_mtu stays at 68, so if the peer later reassembles or bridges,userspace can lower the link
MTU, or set a per-route MTU for any EID behind it, without changing the driver.
If you would prefer, v2 can default mtu to the 68-byte baseline and leave max_mtu at the
negotiated limit, so the platform config raises it explicitly. I'm fine with either option.
Please send the implementation comments when you're ready; I'll fold them into v2.
>
> Cheers,
>
>
> Jeremy
Regards,
Akhila Kavi
next prev parent reply other threads:[~2026-11-16 6:22 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-01 14:00 [PATCH net-next 0/2] net: mctp: Add MCTP transport for the Axiado AX3005 IRoT Akhila Kavi
2026-11-01 14:00 ` [PATCH net-next 1/2] dt-bindings: net: Add Axiado AX3005 IRoT MCTP transport Akhila Kavi
2026-11-02 11:40 ` Krzysztof Kozlowski
2026-11-01 14:00 ` [PATCH net-next 2/2] net: mctp: Add Axiado AX3005 IRoT MCTP transport driver Akhila Kavi
2026-11-02 14:03 ` sashiko-bot
2026-11-05 13:56 ` Uwe Kleine-König
2026-11-02 2:55 ` [PATCH net-next 0/2] net: mctp: Add MCTP transport for the Axiado AX3005 IRoT Jeremy Kerr
2026-11-16 6:22 ` Akhila Kavi [this message]
2026-11-16 14:08 ` Jeremy Kerr
2026-11-05 23:40 ` Jakub Kicinski
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=f752d203-8239-4203-aa04-884b4d7e5d80@axiado.com \
--to=akavi@axiado.com \
--cc=admiyo@amperemail.onmicrosoft.com \
--cc=andrew+netdev@lunn.ch \
--cc=conor+dt@kernel.org \
--cc=davem@davemloft.net \
--cc=devicetree@vger.kernel.org \
--cc=edumazet@google.com \
--cc=jk@codeconstruct.com.au \
--cc=kmitran@axiado.com \
--cc=krzk+dt@kernel.org \
--cc=kuba@kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=matt@codeconstruct.com.au \
--cc=netdev@vger.kernel.org \
--cc=openbmc@lists.ozlabs.org \
--cc=pabeni@redhat.com \
--cc=pbolisetty@axiado.com \
--cc=robh@kernel.org \
--cc=schakravarthy@axiado.com \
--cc=twei@axiado.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