[PATCH v5 05/12] arm64: dts: broadcom: bcmbca: Add NAND controller node

[William Zhang]

Hi Rafal,

On 2/22/24 00:27, Rafał Miłecki wrote:
> On 2024-02-07 21:22, William Zhang wrote:
>> diff --git a/arch/arm64/boot/dts/broadcom/bcmbca/bcm4908.dtsi
>> b/arch/arm64/boot/dts/broadcom/bcmbca/bcm4908.dtsi
>> index 2f124b027bbf..336016e334d9 100644
>> --- a/arch/arm64/boot/dts/broadcom/bcmbca/bcm4908.dtsi
>> +++ b/arch/arm64/boot/dts/broadcom/bcmbca/bcm4908.dtsi
>> @@ -589,7 +589,7 @@ hsspi: spi at 1000 {
>>              status = "disabled";
>>          };
>>
>> -        nand-controller at 1800 {
>> +        nand_controller: nand-controller at 1800 {
>>              #address-cells = <1>;
>>              #size-cells = <0>;
>>              compatible = "brcm,nand-bcm63138", "brcm,brcmnand-v7.1", 
>> "brcm,brcmnand";
>> @@ -597,7 +597,7 @@ nand-controller at 1800 {
>>              reg-names = "nand", "nand-int-base";
>>              interrupts = <GIC_SPI 37 IRQ_TYPE_LEVEL_HIGH>;
>>              interrupt-names = "nand_ctlrdy";
>> -            status = "okay";
>> +            status = "disabled";
>>
>>              nandcs: nand at 0 {
>>                  compatible = "brcm,nandcs";
> 
> You disable NAND controller here breaking support for all dependant
> devices.
> 
> I see you fix it afterwards in PATCH 06/12 but it's still a poor way of
> handling such change and this issue can be hit e.g. during git bisect.
> 
> Also: given that 99% of BCM4908 devices use NAND controller can we just
> save ourselves efforts and keep it enabled in .dtsi?
> 
I was trying to fix this according to the dts convention per 
maintainers' requirement(only enable the node in the board dts, not soc 
include). We have 100% of BCMBCA devices use strap for ecc but we still 
have to put the property in board dts.
I understand if someone try to bisect the commits, it may hit the 
problem but this will be very obvious that controller is not even shown 
in the boot up so hopefully it is easy to figure out.  Anyway if 
maintainer(Minque and/or anyone from device tree side) is okay with your 
request, I can change it back to enable in the dtsi.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4212 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.infradead.org/pipermail/linux-mtd/attachments/20240222/6b9cbcfc/attachment.p7s>