2023-05-05 08:15:13

by Maksim Kiselev

[permalink] [raw]
Subject: [PATCH] riscv: dts: allwinner: d1: Add SPI0 controller node

Some boards form the MangoPi family (MQ\MQ-Dual\MQ-R) may have
an optional SPI flash that connects to the SPI0 controller.
This controller is already supported by sun8i-h3-spi driver.
So let's add its DT node.

Signed-off-by: Maksim Kiselev <[email protected]>
---
.../boot/dts/allwinner/sunxi-d1s-t113.dtsi | 20 +++++++++++++++++++
1 file changed, 20 insertions(+)

diff --git a/arch/riscv/boot/dts/allwinner/sunxi-d1s-t113.dtsi b/arch/riscv/boot/dts/allwinner/sunxi-d1s-t113.dtsi
index 922e8e0e2c09..d2de211d67d7 100644
--- a/arch/riscv/boot/dts/allwinner/sunxi-d1s-t113.dtsi
+++ b/arch/riscv/boot/dts/allwinner/sunxi-d1s-t113.dtsi
@@ -108,6 +108,12 @@ rmii_pe_pins: rmii-pe-pins {
function = "emac";
};

+ /omit-if-no-ref/
+ spi0_pins: spi0-pins {
+ pins = "PC2", "PC3", "PC4", "PC5";
+ function = "spi0";
+ };
+
/omit-if-no-ref/
uart1_pg6_pins: uart1-pg6-pins {
pins = "PG6", "PG7";
@@ -447,6 +453,20 @@ mmc2: mmc@4022000 {
#size-cells = <0>;
};

+ spi0: spi@4025000 {
+ compatible = "allwinner,sun8i-h3-spi";
+ reg = <0x04025000 0x300>;
+ interrupts = <SOC_PERIPHERAL_IRQ(15) IRQ_TYPE_LEVEL_HIGH>;
+ clocks = <&ccu CLK_BUS_SPI0>, <&ccu CLK_SPI0>;
+ clock-names = "ahb", "mod";
+ dmas = <&dma 22>, <&dma 22>;
+ dma-names = "rx", "tx";
+ resets = <&ccu RST_BUS_SPI0>;
+ status = "disabled";
+ #address-cells = <1>;
+ #size-cells = <0>;
+ };
+
usb_otg: usb@4100000 {
compatible = "allwinner,sun20i-d1-musb",
"allwinner,sun8i-a33-musb";
--
2.39.2


2023-05-05 10:34:27

by Andre Przywara

[permalink] [raw]
Subject: Re: [PATCH] riscv: dts: allwinner: d1: Add SPI0 controller node

On Fri, 5 May 2023 10:46:51 +0300
Maksim Kiselev <[email protected]> wrote:

Hi Maksim,

thanks for sending a patch!

> Some boards form the MangoPi family (MQ\MQ-Dual\MQ-R) may have
> an optional SPI flash that connects to the SPI0 controller.
> This controller is already supported by sun8i-h3-spi driver.
> So let's add its DT node.

Interesting, I see SPI mentioned in the D1 platform support cover letter,
but indeed there is no DT node. From a quick glance at the manuals, it
looks like there are not quite the same, though: the D1/R528/T113s
mentions a SPI_SAMP_DL register @0x28, whereas the older IP has a SPI_CCR
register @0x24 - which is not mentioned in the newer manuals. The driver
relies on that clock control register, so it wouldn't really work
reliably, if that register is not there. It *might* work by pure chance
because of a particular setup or clock rate, though.

Samuel, did you investigate SPI support on the D1/T113s? I see it marked
as "WIP" in the Wiki status page, so where there any patches floating
around?


Regardless of that, one comment that would apply anyway:

> Signed-off-by: Maksim Kiselev <[email protected]>
> ---
> .../boot/dts/allwinner/sunxi-d1s-t113.dtsi | 20 +++++++++++++++++++
> 1 file changed, 20 insertions(+)
>
> diff --git a/arch/riscv/boot/dts/allwinner/sunxi-d1s-t113.dtsi b/arch/riscv/boot/dts/allwinner/sunxi-d1s-t113.dtsi
> index 922e8e0e2c09..d2de211d67d7 100644
> --- a/arch/riscv/boot/dts/allwinner/sunxi-d1s-t113.dtsi
> +++ b/arch/riscv/boot/dts/allwinner/sunxi-d1s-t113.dtsi
> @@ -108,6 +108,12 @@ rmii_pe_pins: rmii-pe-pins {
> function = "emac";
> };
>
> + /omit-if-no-ref/
> + spi0_pins: spi0-pins {
> + pins = "PC2", "PC3", "PC4", "PC5";
> + function = "spi0";
> + };
> +
> /omit-if-no-ref/
> uart1_pg6_pins: uart1-pg6-pins {
> pins = "PG6", "PG7";
> @@ -447,6 +453,20 @@ mmc2: mmc@4022000 {
> #size-cells = <0>;
> };
>
> + spi0: spi@4025000 {
> + compatible = "allwinner,sun8i-h3-spi";

Even if it would be compatible, we need to use a more specific compatible
first, with the H3 one as a fallback:

compatible = "allwinner,sun20i-d1-spi",
"allwinner,sun8i-h3-spi";

But that would require that the H3 is a strict subset of the D1 SPI IP.

Cheers,
Andre


> + reg = <0x04025000 0x300>;
> + interrupts = <SOC_PERIPHERAL_IRQ(15) IRQ_TYPE_LEVEL_HIGH>;
> + clocks = <&ccu CLK_BUS_SPI0>, <&ccu CLK_SPI0>;
> + clock-names = "ahb", "mod";
> + dmas = <&dma 22>, <&dma 22>;
> + dma-names = "rx", "tx";
> + resets = <&ccu RST_BUS_SPI0>;
> + status = "disabled";
> + #address-cells = <1>;
> + #size-cells = <0>;
> + };
> +
> usb_otg: usb@4100000 {
> compatible = "allwinner,sun20i-d1-musb",
> "allwinner,sun8i-a33-musb";

2023-05-05 14:07:27

by Maksim Kiselev

[permalink] [raw]
Subject: Re: [PATCH] riscv: dts: allwinner: d1: Add SPI0 controller node

Hi Andre,

> From a quick glance at the manuals, it
> looks like there are not quite the same, though: the D1/R528/T113s
> mentions a SPI_SAMP_DL register @0x28, whereas the older IP has a SPI_CCR
> register @0x24 - which is not mentioned in the newer manuals. The driver
> relies on that clock control register, so it wouldn't really work
> reliably, if that register is not there.

Thank you for pointing this out. I missed this difference.
I actually have a board with T113 SoC, and it looks like writing to
SPI_CCR@0x24 does nothing.
And it doesn't affect access to connected SPI NOR flash (read\write
operations are fine).
But I completely agree with you that this difference should be handled
by the spi driver.

2023-05-05 16:24:43

by Andre Przywara

[permalink] [raw]
Subject: Re: [PATCH] riscv: dts: allwinner: d1: Add SPI0 controller node

On Fri, 5 May 2023 17:27:07 +0400
Maxim Kiselev <[email protected]> wrote:

Hi Maxim,

> > From a quick glance at the manuals, it
> > looks like there are not quite the same, though: the D1/R528/T113s
> > mentions a SPI_SAMP_DL register @0x28, whereas the older IP has a SPI_CCR
> > register @0x24 - which is not mentioned in the newer manuals. The driver
> > relies on that clock control register, so it wouldn't really work
> > reliably, if that register is not there.
>
> Thank you for pointing this out. I missed this difference.
> I actually have a board with T113 SoC, and it looks like writing to
> SPI_CCR@0x24 does nothing.

According to the manual, the register doesn't exist, so this would support
this theory. Most Allwinner IP just implements non-existing registers as
RAZ/WI.

> And it doesn't affect access to connected SPI NOR flash (read\write
> operations are fine).

I guess it just works with *some* (default?) clock settings then. Might be
faster than configured, but the SPI flash chip might be fine with that.

> But I completely agree with you that this difference should be handled
> by the spi driver.

As Samuel pointed out on IRC, there are patches, for the R329, using the
same IP:
https://lore.kernel.org/lkml/BYAPR20MB2472E8B10BFEF75E7950BBC0BCF79@BYAPR20MB2472.namprd20.prod.outlook.com/
And he also seemed to have issues, at least with SPI-NAND on one board:
https://lore.kernel.org/lkml/[email protected]/

Can you please try these patches, and use:
compatible = "allwinner,sun20i-d1-spi", "allwinner,sun50i-r329-spi";
for the compatible string?
If possible, you could revive this series, and add the D1/T113s support on
top: just add the compatible string combo to the binding document.

Cheers,
Andre

2023-05-05 21:41:56

by Maksim Kiselev

[permalink] [raw]
Subject: Re: [PATCH] riscv: dts: allwinner: d1: Add SPI0 controller node

> Can you please try these patches, and use:
> compatible = "allwinner,sun20i-d1-spi", "allwinner,sun50i-r329-spi";
> for the compatible string?

I tested my patch (with compatible prop changed to
allwinner,sun50i-r329-spi) on top of these patches and it works fine
for me.

> If possible, you could revive this series, and add the D1/T113s support on
> top: just add the compatible string combo to the binding document.

Just one question. Should I do it now, or is it better to wait while
Icenowy Zheng finishes work on the next revision of this series?

2023-05-05 21:52:27

by Andre Przywara

[permalink] [raw]
Subject: Re: [PATCH] riscv: dts: allwinner: d1: Add SPI0 controller node

On Sat, 6 May 2023 01:25:02 +0400
Maxim Kiselev <[email protected]> wrote:

Hi Maxim,

> > Can you please try these patches, and use:
> > compatible = "allwinner,sun20i-d1-spi", "allwinner,sun50i-r329-spi";
> > for the compatible string?
>
> I tested my patch (with compatible prop changed to
> allwinner,sun50i-r329-spi) on top of these patches and it works fine
> for me.
>
> > If possible, you could revive this series, and add the D1/T113s support on
> > top: just add the compatible string combo to the binding document.
>
> Just one question. Should I do it now, or is it better to wait while
> Icenowy Zheng finishes work on the next revision of this series?

This series is over a year old, I doubt she will jump on this again any
time soon. So it would be great if you could rebase the patches, keeping
her authorship and Signed-off-by:, adding your Signed-off-by: and
posting them.
I might take a deep breath and try to solder some NOR SPI on my MQ-R to
test it later on....

Many thanks!
Andre