Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20210611050808.2554431-1-vkoul@kernel.org>
In-Reply-To: <20210611050808.2554431-1-vkoul@kernel.org>
From:   John Stultz <john.stultz@linaro.org>
Date:   Thu, 10 Jun 2021 22:27:16 -0700
Message-ID: <CALAqxLW4n8ijkD6hw_xqa2tzkttmXS_LFn_yJo6cP+iYYt-+=A@mail.gmail.com>
Subject: Re: [PATCH] arm64: dts: qcom: sm8350-mtp: Use mdt files for firmware
To:     Vinod Koul <vkoul@kernel.org>
Cc:     Bjorn Andersson <bjorn.andersson@linaro.org>,
        linux-arm-msm <linux-arm-msm@vger.kernel.org>,
        Andy Gross <agross@kernel.org>,
        Rob Herring <robh+dt@kernel.org>,
        Dmitry Baryshkov <dmitry.baryshkov@linaro.org>,
        "open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS" 
        <devicetree@vger.kernel.org>, lkml <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Jun 10, 2021 at 10:08 PM Vinod Koul <vkoul@kernel.org> wrote:
>
> As discussed in [1], it makes it easy for everyone to use mdt firmware file
> name instead of mbn ones, so changes this for SM8350
>
> [1]: http://lore.kernel.org/r/CALAqxLXn6wFBAxRkThxWg5RvTuFEX80kHPt8BVja1CpAB-qzGA@mail.gmail.com
>
> Signed-off-by: Vinod Koul <vkoul@kernel.org>
> ---
>  arch/arm64/boot/dts/qcom/sm8350-mtp.dts | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/arch/arm64/boot/dts/qcom/sm8350-mtp.dts b/arch/arm64/boot/dts/qcom/sm8350-mtp.dts
> index 93740444dd1e..d859305f1f75 100644
> --- a/arch/arm64/boot/dts/qcom/sm8350-mtp.dts
> +++ b/arch/arm64/boot/dts/qcom/sm8350-mtp.dts
> @@ -40,7 +40,7 @@ vph_pwr: vph-pwr-regulator {
>
>  &adsp {
>         status = "okay";
> -       firmware-name = "qcom/sm8350/adsp.mbn";
> +       irmware-name = "qcom/sm8350/adsp.mdt";
>  };

Uhh, isn't this the opposite of [1]?  My apologies for butting in, and
I'd stay out of the discussion, except for my mail being linked as
justification :)

In [1] the case was db845c was switched from older mdt files to using
the upstream linux-firmware mbn files. This was a bit of a pain, as it
broke on our userland with mdt files, and since we use both old and
new kernels we had to have both filenames on the disk (via symlink) to
keep it working everywhere.

My argument in [1] was for new boards, go with the new conventions,
but we should avoid breaking those conventions casually on existing
devices.  That said, I know it's more complex, and I graciously defer
to Bjorn and RobC on the decision.

But your patch above seems to be switching from mbn (what I understand
to be the new convention) to mdt (what I thought was the old way). And
from the git blame, it looks like it was introduced as mbn (new board,
new convention - so all good, right?).

So is this really the right change? Or maybe just more exposition in
the commit message is needed (rather than pointing to my mail, which
seems to be arguing the opposite) to explain it?

thanks
-john