Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <4440a0b1-bf67-9c37-0526-1a447ff8568a@dram.page>
Date:   Fri, 16 Jun 2023 22:39:09 +0800
MIME-Version: 1.0
To:     Anup Patel <anup.patel@wdc.com>,
        Palmer Dabbelt <palmer@dabbelt.com>,
        Palmer Dabbelt <palmerdabbelt@google.com>,
        Paul Walmsley <paul.walmsley@sifive.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Marc Zyngier <maz@kernel.org>,
        Daniel Lezcano <daniel.lezcano@linaro.org>,
        Rob Herring <robh+dt@kernel.org>
Cc:     Atish Patra <atish.patra@wdc.com>,
        Alistair Francis <Alistair.Francis@wdc.com>,
        Anup Patel <anup@brainfault.org>,
        linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org,
        devicetree@vger.kernel.org, Bin Meng <bmeng.cn@gmail.com>,
        "Hongren (Zenithal) Zheng" <i@zenithal.me>
References: <20211007123632.697666-1-anup.patel@wdc.com>
 <20211007123632.697666-6-anup.patel@wdc.com>
Content-Language: en-US
From:   Vivian Wang <uwu@dram.page>
Subject: Re: [RFC PATCH v4 05/10] dt-bindings: interrupt-controller: Add
 ACLINT MSWI and SSWI bindings
In-Reply-To: <20211007123632.697666-6-anup.patel@wdc.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

Hi all,

We are working on devicetree generation for rocket-chip, specifically
generating ACLINT nodes instead of CLINT nodes. (WIP at [1].) We were
wondering if riscv,aclint-{m,s}swi should really be an interrupt-controller.

According to the devicetree specification (v0.3 found at [2]), an
interrupt-controller *receives* interrupts. The ACLINT devices only ever
generates interrupts, so they would be classified as interrupt
generating device rather than interrupt controller or interrupt nexus.

These bindings, as is, require the MSWI and SSWI devices to have the
interrupt-controller property and #interrupt-cells = <0>, which does not
reflect its functionality. It nonsensically implies that another device
may have an interrupt routed through an MSWI/SSWI as interrupt-parent.
Removing these requirements makes more sense.

I'm not sure about what other node name to use though. It seems that
these are more like mailboxes, but also not exactly. In any case a
clarification of the bindings would be appreciated.

Thanks,
dram

[1]: https://github.com/chipsalliance/rocket-chip/pull/3330
[2]: https://www.devicetree.org/specifications/