linux-kernel - Re: [PATCH 1/2] dt-bindings: mailbox: cix: add device tree binding documentation.

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <d0d82010-e311-4a41-a6c4-027626e408ba@kernel.org>
Date: Fri, 14 Mar 2025 09:13:52 +0100
From: Krzysztof Kozlowski <krzk@...nel.org>
To: Guomin chen <guomin.chen@...tech.com>,
 Jassi Brar <jassisinghbrar@...il.com>, Rob Herring <robh@...nel.org>,
 Conor Dooley <conor+dt@...nel.org>
Cc: linux-kernel@...r.kernel.org, cix-kernel-upstream@...tech.com,
 Peter Chen <peter.chen@...tech.com>, Lihua Liu <Lihua.Liu@...tech.com>
Subject: Re: [PATCH 1/2] dt-bindings: mailbox: cix: add device tree binding
 documentation.

On 14/03/2025 09:09, Guomin chen wrote:
>>
>>> +
>>> +      This controller supports three types of unidirectional channels, they are
>>> +      1 register based channel, 1 fifo based channel and 8 fast channels.
>>> +      A total of 10 channels for each controller. Following types are
>>> +      supported:
>>> +      channel 0_7 - Fast channel with 32bit transmit register and IRQ support.
>>> +      channel 8   - Reg based channel with 32*32bit transsmit register and
>>> +                    Doorbell+transmit acknowledgment IRQ support
>>> +      channel 9   - Fifo based channel with 32*32bit depth fifo and IRQ support.
>>> +    const: 1
>>> +
>>> +  cix,mbox-dir:
>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>> +    description: Direction of the mailbox (0:TX or 1:RX)
>>> +    enum: [0, 1]
>>
>> I don't understand why do you need it. By not sending us driver patch,
>> you are not making it easier. Why would provider care how consumers use
>> the mbox channel? Maybe consumer should choose the direction?
>>
> 
> As for the mbox-dir property, my driver code has already been submitted. 

There is no thing in this thread, so next time be sure you use standard
submitting process, so your work will be threaded.

> On the Cixtech Soc platform, although each mbox is unidirectional, 
> there are multiple mboxes—some for reading and some for writing. 

So all mboxes in a controller have the same direction or not? Do you
configure it or is it defined by hardware?

> Therefore, the mbox controller has added the mbox-dir property. 
> 
> Consumers only need to reference the corresponding mbox controller node, 
> and whether it is for reading or writing is already determined by the mbox controller, 
> without needing to further understand the mbox-dir property.

Don't explain how DT works, I know. Why this property is needed in DT in
the first place?

Best regards,
Krzysztof