Received-SPF: pass (google.com: domain of linux-kernel+bounces-137190-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Date: Tue, 9 Apr 2024 10:10:41 -0500
From: Rob Herring <robh@kernel.org>
To: Luca Weiss <luca@z3ntu.xyz>
Cc: ~postmarketos/upstreaming@lists.sr.ht, phone-devel@vger.kernel.org,
	Bjorn Andersson <andersson@kernel.org>,
	Mathieu Poirier <mathieu.poirier@linaro.org>,
	Krzysztof Kozlowski <krzk+dt@kernel.org>,
	Conor Dooley <conor+dt@kernel.org>, linux-arm-msm@vger.kernel.org,
	linux-remoteproc@vger.kernel.org, devicetree@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH 0/3] Fix up qcom,halt-regs definition in various schemas
Message-ID: <20240409151041.GA1141635-robh@kernel.org>
References: <20240407-qcom-halt-regs-fixup-v1-0-a0ea4e2c178e@z3ntu.xyz>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240407-qcom-halt-regs-fixup-v1-0-a0ea4e2c178e@z3ntu.xyz>

On Sun, Apr 07, 2024 at 11:58:29AM +0200, Luca Weiss wrote:
> The original motivation is that a bunch of other schemas fail to
> validate qcom,halt-regs, for example like in the following examples:
> 
> arch/arm64/boot/dts/qcom/apq8016-sbc.dtb: remoteproc@4080000: qcom,halt-regs:0: [20] is too short
>         from schema $id: http://devicetree.org/schemas/remoteproc/qcom,msm8916-mss-pil.yaml#
> arch/arm64/boot/dts/qcom/apq8096-ifc6640.dtb: remoteproc@2080000: qcom,halt-regs:0: [82] is too short
>         from schema $id: http://devicetree.org/schemas/remoteproc/qcom,msm8996-mss-pil.yaml#
> arch/arm64/boot/dts/qcom/apq8039-t2.dtb: remoteproc@4080000: qcom,halt-regs:0: [32] is too short
>         from schema $id: http://devicetree.org/schemas/remoteproc/qcom,msm8916-mss-pil.yaml#
> 
> While I'm actually not quite sure why these patches fix this in
> the other schemas - feels like a bug/limitation in dt-schema maybe? -

Was this with v2024.02? It should be a bit better there. Though it 
may just have different errors. The limitation is that property 
types and in the case of matrix's (which phandle-array actually is) 
range for dimensions are global. So if there's not correct dimensions 
for a property, the tools aren't going to decode it properly.

Rob