diff mbox series

[2/2] arm64: dts: rockchip: remove duplicate SPI aliases for helios64

Message ID 20240109-rk3399-spi-aliases-v1-2-2009e44e734a@theobroma-systems.com
State New
Headers
Series arm64: dts: rockchip: add SPI controller aliases to RK3399 |

Commit Message

Quentin Schulz Jan. 9, 2024, 1:35 p.m. UTC 
  From: Quentin Schulz <quentin.schulz@theobroma-systems.com>

An earlier commit defined an alias for all SPI controllers found on the
RK3399, so there's no need to duplicate the aliases in helios64's device
tree.

Cc: Quentin Schulz <foss+kernel@0leil.net>
Signed-off-by: Quentin Schulz <quentin.schulz@theobroma-systems.com>
---
 arch/arm64/boot/dts/rockchip/rk3399-kobol-helios64.dts | 3 ---
 1 file changed, 3 deletions(-)


-- 
2.43.0

Comments

Krzysztof Kozlowski Jan. 9, 2024, 3:16 p.m. UTC | #1 
On 09/01/2024 14:35, Quentin Schulz wrote:
> From: Quentin Schulz <quentin.schulz@theobroma-systems.com>
> 
> An earlier commit defined an alias for all SPI controllers found on the
> RK3399, so there's no need to duplicate the aliases in helios64's device
> tree.

That's not what we want. Boards should define aliases for what is
available, according to the board labeling.

Best regards,
Krzysztof
Heiko Stübner Jan. 9, 2024, 3:23 p.m. UTC | #2 
Am Dienstag, 9. Januar 2024, 16:16:15 CET schrieb Krzysztof Kozlowski:
> On 09/01/2024 14:35, Quentin Schulz wrote:
> > From: Quentin Schulz <quentin.schulz@theobroma-systems.com>
> > 
> > An earlier commit defined an alias for all SPI controllers found on the
> > RK3399, so there's no need to duplicate the aliases in helios64's device
> > tree.
> 
> That's not what we want. Boards should define aliases for what is
> available, according to the board labeling.

and the board labeling for spi2 has always been spi2-miso, spi2-miso etc.
In the last 10 years of doing Rockchip stuff, I haven't seen any schematic
doing it differently.
Krzysztof Kozlowski Jan. 9, 2024, 7:14 p.m. UTC | #3 
On 09/01/2024 16:23, Heiko Stübner wrote:
> Am Dienstag, 9. Januar 2024, 16:16:15 CET schrieb Krzysztof Kozlowski:
>> On 09/01/2024 14:35, Quentin Schulz wrote:
>>> From: Quentin Schulz <quentin.schulz@theobroma-systems.com>
>>>
>>> An earlier commit defined an alias for all SPI controllers found on the
>>> RK3399, so there's no need to duplicate the aliases in helios64's device
>>> tree.
>>
>> That's not what we want. Boards should define aliases for what is
>> available, according to the board labeling.
> 
> and the board labeling for spi2 has always been spi2-miso, spi2-miso etc.
> In the last 10 years of doing Rockchip stuff, I haven't seen any schematic
> doing it differently.

OK, this could be a case... but then you add aliases for things which
are not labeled on the board.

Let's bring Arnd again:
https://lore.kernel.org/linux-rockchip/CAK8P3a0ALgbhTVJ7t3XRXALs9vBM=XBvkGhNKXxB+QTepo-3AQ@mail.gmail.com/

Best regards,
Krzysztof
Heiko Stübner Jan. 10, 2024, 11:18 a.m. UTC | #4 
Hi Krzysztof,

[I'll limit myself to this thread, as we're essentially discussing the same in both]

Am Dienstag, 9. Januar 2024, 20:14:38 CET schrieb Krzysztof Kozlowski:
> On 09/01/2024 16:23, Heiko Stübner wrote:
> > Am Dienstag, 9. Januar 2024, 16:16:15 CET schrieb Krzysztof Kozlowski:
> >> On 09/01/2024 14:35, Quentin Schulz wrote:
> >>> From: Quentin Schulz <quentin.schulz@theobroma-systems.com>
> >>>
> >>> An earlier commit defined an alias for all SPI controllers found on the
> >>> RK3399, so there's no need to duplicate the aliases in helios64's device
> >>> tree.
> >>
> >> That's not what we want. Boards should define aliases for what is
> >> available, according to the board labeling.
> > 
> > and the board labeling for spi2 has always been spi2-miso, spi2-miso etc.
> > In the last 10 years of doing Rockchip stuff, I haven't seen any schematic
> > doing it differently.
> 
> OK, this could be a case... but then you add aliases for things which
> are not labeled on the board.

On the other hand, we did this dance last year for some other SoCs.
And having to repeat the same list of never-changing aliases for every
board was a sore spot for a lot of people it seemed.
Most boards follow the reference schematics quite closely afterall.

Like very shortly after the last discussion I got patches moving the
aliases to their then agreed upon position. (numeric busses in the
soc dtsi, everything else in board dts).

And back in november we ended with [0] - Krzysztof saying that
"it is just generic guideline, so up to Heiko what to do with it" ;-) .


I also made sure to send the changes upwards way before christmas
to give soc maintainers time to complain if needed and especially did
point out that change in my pull request text [1] ;-) .


Heiko

> Let's bring Arnd again:
> https://lore.kernel.org/linux-rockchip/CAK8P3a0ALgbhTVJ7t3XRXALs9vBM=XBvkGhNKXxB+QTepo-3AQ@mail.gmail.com/


[0] https://lore.kernel.org/linux-rockchip/7b8a830d-42d0-4220-b9fe-3f5fff43e74d@linaro.org/
[1] https://patchwork.kernel.org/project/linux-soc/patch/3535836.iIbC2pHGDl@phil/
diff mbox series

Patch

diff --git a/arch/arm64/boot/dts/rockchip/rk3399-kobol-helios64.dts b/arch/arm64/boot/dts/rockchip/rk3399-kobol-helios64.dts
index 1eb287a3f8c03..d664ac18a941e 100644
--- a/arch/arm64/boot/dts/rockchip/rk3399-kobol-helios64.dts
+++ b/arch/arm64/boot/dts/rockchip/rk3399-kobol-helios64.dts
@@ -21,9 +21,6 @@  / {
 	aliases {
 		mmc0 = &sdmmc;
 		mmc1 = &sdhci;
-		spi1 = &spi1;
-		spi2 = &spi2;
-		spi5 = &spi5;
 	};
 
 	avdd_0v9_s0: avdd-0v9-s0 {