Date:   Tue, 21 Dec 2021 08:46:28 +0000
From:   David Heidelberg <david@ixit.cz>
To:     Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
CC:     Andy Gross <agross@kernel.org>,
        Bjorn Andersson <bjorn.andersson@linaro.org>,
        Rob Herring <robh+dt@kernel.org>,
        ~okias/devicetree@lists.sr.ht, linux-arm-msm@vger.kernel.org,
        devicetree@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: =?US-ASCII?Q?Re=3A_=5BPATCH=5D_arm64=3A_dts=3A_qcom=3A_sdm845=3A_?= =?US-ASCII?Q?add_missing_power-controller_compatible?=
In-Reply-To: <20211221065845.GC26872@thinkpad>
References: <20211220211443.106754-1-david@ixit.cz> <20211221065845.GC26872@thinkpad>
Message-ID: <231125F6-0A22-4042-969E-45347CBE8CB7@ixit.cz>
MIME-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

It's more about how documentation is written=2E Documentation expect "speci=
fic compatible", "generic compatible" in that order=2E Of course it can be =
changed to allow also  only specific compatible=2E

Also this syntax ensure that older kernel (without implemented device comp=
atible/support) will try to work when new dts is used=2E Kernel can fallbac=
k to generic one compatible and code and device will still work, if the dri=
ver is generic enough to provide at least some basic support without knowin=
g which exact device is handled=2E

If I missed some reason to use it, I guess Rob will correct me :)

David


-------- P=C5=AFvodn=C3=AD zpr=C3=A1va --------
Odes=C3=ADlatel: Manivannan Sadhasivam <manivannan=2Esadhasivam@linaro=2Eo=
rg>
Odesl=C3=A1no: 21=2E prosince 2021 6:58:45 UTC
Komu: David Heidelberg <david@ixit=2Ecz>
Kopie: Andy Gross <agross@kernel=2Eorg>, Bjorn Andersson <bjorn=2Eandersso=
n@linaro=2Eorg>, Rob Herring <robh+dt@kernel=2Eorg>, ~okias/devicetree@list=
s=2Esr=2Eht, linux-arm-msm@vger=2Ekernel=2Eorg, devicetree@vger=2Ekernel=2E=
org, linux-kernel@vger=2Ekernel=2Eorg
P=C5=99edm=C4=9Bt: Re: [PATCH] arm64: dts: qcom: sdm845: add missing power=
-controller compatible

On Mon, Dec 20, 2021 at 10:14:43PM +0100, David Heidelberg wrote:
> dt-schema expect to have fallback compatible, which is now in-place=2E
>=20
> Fixes warning generated by `make qcom/sdm845-oneplus-fajita=2Edtb`:
> arch/arm64/boot/dts/qcom/sdm845-oneplus-fajita=2Edt=2Eyaml: power-contro=
ller@c300000: compatible: ['qcom,sdm845-aoss-qmp'] is too short
>         From schema: Documentation/devicetree/bindings/soc/qcom/qcom,aos=
s-qmp=2Eyaml
>=20
> Signed-off-by: David Heidelberg <david@ixit=2Ecz>
> ---
>  arch/arm64/boot/dts/qcom/sdm845=2Edtsi | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>=20
> diff --git a/arch/arm64/boot/dts/qcom/sdm845=2Edtsi b/arch/arm64/boot/dt=
s/qcom/sdm845=2Edtsi
> index 92ab4513a08b=2E=2Edbdb4243499c 100644
> --- a/arch/arm64/boot/dts/qcom/sdm845=2Edtsi
> +++ b/arch/arm64/boot/dts/qcom/sdm845=2Edtsi
> @@ -4619,7 +4619,7 @@ aoss_reset: reset-controller@c2a0000 {
>  		};
> =20
>  		aoss_qmp: power-controller@c300000 {
> -			compatible =3D "qcom,sdm845-aoss-qmp";
> +			compatible =3D "qcom,sdm845-aoss-qmp", "qcom,aoss-qmp";

"qcom,sdm845-aoss-qmp" compatible is supported by the driver=2E So ideally=
 we
don't need a fallback here=2E

Is this something for DT backwards compatibility?

Thanks,
Mani

>  			reg =3D <0 0x0c300000 0 0x100000>;
>  			interrupts =3D <GIC_SPI 389 IRQ_TYPE_EDGE_RISING>;
>  			mboxes =3D <&apss_shared 0>;
> --=20
> 2=2E34=2E1
>=20