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;
MIME-Version: 1.0
References: <20230331182727.4062790-1-evan@rivosinc.com> <25837ca5-b709-4d20-912b-ab13e6353069@spud>
In-Reply-To: <25837ca5-b709-4d20-912b-ab13e6353069@spud>
From:   Evan Green <evan@rivosinc.com>
Date:   Fri, 31 Mar 2023 12:02:28 -0700
Message-ID: <CALs-Hss-W78x04LS7X_RiTmU19Xv6d3f4x5rgeEhV7TVy7_CEA@mail.gmail.com>
Subject: Re: [RFC PATCH] riscv: dts: nezha-d1: Add memory
To:     Conor Dooley <conor@kernel.org>
Cc:     linux-sunxi@lists.linux.dev, conor.dooley@microchip.com,
        Albert Ou <aou@eecs.berkeley.edu>,
        Chen-Yu Tsai <wens@csie.org>,
        Jernej Skrabec <jernej.skrabec@gmail.com>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
        Palmer Dabbelt <palmer@dabbelt.com>,
        Paul Walmsley <paul.walmsley@sifive.com>,
        Rob Herring <robh+dt@kernel.org>,
        Samuel Holland <samuel@sholland.org>,
        Trevor Woerner <twoerner@gmail.com>,
        devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
        linux-kernel@vger.kernel.org, linux-riscv@lists.infradead.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Fri, Mar 31, 2023 at 11:39=E2=80=AFAM Conor Dooley <conor@kernel.org> wr=
ote:
>
> On Fri, Mar 31, 2023 at 11:27:27AM -0700, Evan Green wrote:
> > Add memory info for the D1 Nezha, which seems to be required for it to
> > boot with the stock firmware. Note that this hardcodes 1GB, which is
> > not technically correct as they also make models with different amounts
> > of RAM.
>
> > Is the firmware supposed to populate this?
>
> Yup. Samuel intentionally did it this way:
> https://lore.kernel.org/linux-riscv/8a2194bf-93bd-de4d-8d39-0cd72aabb0a9@=
sholland.org/
>
> Since I don't like using the firmware provided stuff to more easily
> test I do this in my u-boot boot script:
>
> tftp 0x54000000 d1-fitImage.fit
> bootm start 0x54000000
> bootm loados 0x54000000
> bootm ramdisk;
> bootm prep;
> fdt memory 0x40000000 0x20000000
> bootm go;
>
> Hope either of the above are helpful,
> Conor.

Ah, I see, thanks for the pointers. I've got whatever firmware came
with the "dd and go" Fedora image I found on the internet somewhere,
which doesn't seem to do the memory node population for me. I've been
hesitant to rebuild the firmware as I don't want to break the Fedora
"safe OS", and if there's a way to get mired in some build
environment, I seem to have a nose for finding it.

I'll probably just keep the patch locally in my tree, and if I'm
forced to update the firmware for some other reason I'll try removing
it.

-Evan