Re: [PATCH] arm64: dts: fsd: Change the reg properties from 64-bit to 32-bit

From: Robin Murphy <robin.murphy@arm.com>
To: Arnd Bergmann <arnd@arndb.de>,
	Vivek Yadav <vivek.2311@samsung.com>,
	rcsekar@samsung.com, krzysztof.kozlowski+dt@linaro.org,
	Wolfgang Grandegger <wg@grandegger.com>,
	Marc Kleine-Budde <mkl@pengutronix.de>,
	"David S . Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	pankaj.dubey@samsung.com, ravi.patel@samsung.com,
	Alim Akhtar <alim.akhtar@samsung.com>,
	linux-fsd@tesla.com, Rob Herring <robh+dt@kernel.org>
Cc: linux-can@vger.kernel.org, Netdev <netdev@vger.kernel.org>,
	linux-kernel@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	linux-samsung-soc@vger.kernel.org, devicetree@vger.kernel.org,
	aswani.reddy@samsung.com, sriranjani.p@samsung.com,
	ajaykumar.rs@samsung.com
Subject: Re: [PATCH] arm64: dts: fsd: Change the reg properties from 64-bit to 32-bit
Date: Wed, 16 Nov 2022 12:38:19 +0000	[thread overview]
Message-ID: <d63b59c3-f67d-e5ee-6cbf-9f97eec0aeaa@arm.com> (raw)
In-Reply-To: <37d42235-1960-4001-9be9-20a85de54730@app.fastmail.com>

On 2022-11-16 11:16, Arnd Bergmann wrote:
> On Wed, Nov 16, 2022, at 10:12, Vivek Yadav wrote:
>> Change the reg properties from 64-bit to 32-bit for all IPs, as none of
>> the nodes are above 32-bit range in the fsd SoC.
>>
>> Since dma-ranges length does not fit into 32-bit size, keep it 64-bit
>> and move it to specific node where it is used instead of SoC section.
> 
> I don't think that works, the dma-ranges property is part of the
> bus, not a particular device:
> 
>   		mdma0: dma-controller@10100000 {
>   			compatible = "arm,pl330", "arm,primecell";
> -			reg = <0x0 0x10100000 0x0 0x1000>;
> +			reg = <0x10100000 0x1000>;
>   			interrupts = <GIC_SPI 424 IRQ_TYPE_LEVEL_HIGH>;
>   			#dma-cells = <1>;
>   			clocks = <&clock_imem IMEM_DMA0_IPCLKPORT_ACLK>;
>   			clock-names = "apb_pclk";
>   			iommus = <&smmu_imem 0x800 0x0>;
> +			#address-cells = <2>;
> +			#size-cells = <2>;
> +			dma-ranges = <0x0 0x0 0x0 0x10 0x0>;
>   		};
> 
> Since the dma-controller has no children, I don't see how this has
> any effect. Also, translating a 36-bit address into a 32-bit
> address just means it gets truncated anyway, so there is no
> point in making it appear to have a larger address range.

Yes, this is definitely bogus on both counts.

Thanks,
Robin.

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel