[PATCH 0/2] dt-bindings: Add SpacemiT K1 PDMA request-number header and use it in DT

[PATCH 0/2] dt-bindings: Add SpacemiT K1 PDMA request-number header and use it in DT

[Guodong Xu]

Currently, K1 device trees specify PDMA peripheral requests as raw numbers.
Add a dt-bindings header naming those request lines, point the binding's
dma-cells description at it, and convert the current user (the K1 SPI3
node) to the new K1_PDMA_* macros.

Patch 1 adds include/dt-bindings/dma/spacemit,k1-pdma.h with the
  K1_PDMA_* request-number macros and update the spacemit,k1-pdma binding's
  dma-cells description.
Patch 2 updates the current pdma user in k1.dtsi to use these request
  number macros.

Signed-off-by: Guodong Xu <docular.xu@gmail.com>
---
Guodong Xu (2):
      dt-bindings: dmaengine: Add SpacemiT K1 PDMA request numbers
      riscv: dts: spacemit: Use symbolic PDMA request numbers on K1

 .../devicetree/bindings/dma/spacemit,k1-pdma.yaml  |  4 +-
 arch/riscv/boot/dts/spacemit/k1.dtsi               |  3 +-
 include/dt-bindings/dma/spacemit,k1-pdma.h         | 56 ++++++++++++++++++++++
 3 files changed, 61 insertions(+), 2 deletions(-)
---
base-commit: 793cc54475b49b5b558902b5c13e4bfe66530a50
change-id: 20260607-b4-k1-pdma-req-macros-8d276d0126df

Best regards,
--  
Guodong Xu <docular.xu@gmail.com>

[PATCH 1/2] dt-bindings: dmaengine: Add SpacemiT K1 PDMA request numbers

[Guodong Xu]

Add a dt-bindings header that gives symbolic names to the SpacemiT K1
PDMA request lines of the non-secure peripherals. Device trees can use
these K1_PDMA_* macros instead of magic numbers.

Point the spacemit,k1-pdma binding's #dma-cells description at the new
header.

Signed-off-by: Guodong Xu <docular.xu@gmail.com>
---
 .../devicetree/bindings/dma/spacemit,k1-pdma.yaml  |  4 +-
 include/dt-bindings/dma/spacemit,k1-pdma.h         | 56 ++++++++++++++++++++++
 2 files changed, 59 insertions(+), 1 deletion(-)

diff --git a/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml b/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
index ec06235baf5ca..0d4ac9849e27b 100644
--- a/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
+++ b/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
@@ -35,7 +35,9 @@ properties:
   '#dma-cells':
     const: 1
     description:
-      The DMA request number for the peripheral device.
+      The single cell is the DMA request number for the peripheral device.
+      See <dt-bindings/dma/spacemit,k1-pdma.h> for the list of valid request
+      numbers.
 
 required:
   - compatible
diff --git a/include/dt-bindings/dma/spacemit,k1-pdma.h b/include/dt-bindings/dma/spacemit,k1-pdma.h
new file mode 100644
index 0000000000000..491976516550a
--- /dev/null
+++ b/include/dt-bindings/dma/spacemit,k1-pdma.h
@@ -0,0 +1,56 @@
+/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
+/*
+ * This header provides DMA request number for non-secure peripherals of
+ * SpacemiT K1 PDMA.
+ *
+ * Copyright (c) 2026 Guodong Xu <docular.xu@gmail.com>
+ */
+
+#ifndef _DT_BINDINGS_DMA_SPACEMIT_K1_PDMA_H_
+#define _DT_BINDINGS_DMA_SPACEMIT_K1_PDMA_H_
+
+#define K1_PDMA_UART0_TX	3
+#define K1_PDMA_UART0_RX	4
+#define K1_PDMA_UART2_TX	5
+#define K1_PDMA_UART2_RX	6
+#define K1_PDMA_UART3_TX	7
+#define K1_PDMA_UART3_RX	8
+#define K1_PDMA_UART4_TX	9
+#define K1_PDMA_UART4_RX	10
+#define K1_PDMA_I2C0_TX		11
+#define K1_PDMA_I2C0_RX		12
+#define K1_PDMA_I2C1_TX		13
+#define K1_PDMA_I2C1_RX		14
+#define K1_PDMA_I2C2_TX		15
+#define K1_PDMA_I2C2_RX		16
+#define K1_PDMA_I2C4_TX		17
+#define K1_PDMA_I2C4_RX		18
+#define K1_PDMA_SPI3_TX		19
+#define K1_PDMA_SPI3_RX		20
+#define K1_PDMA_I2S0_TX		21
+#define K1_PDMA_I2S0_RX		22
+#define K1_PDMA_I2S1_TX		23
+#define K1_PDMA_I2S1_RX		24
+#define K1_PDMA_UART5_TX	25
+#define K1_PDMA_UART5_RX	26
+#define K1_PDMA_UART6_TX	27
+#define K1_PDMA_UART6_RX	28
+#define K1_PDMA_UART7_TX	29
+#define K1_PDMA_UART7_RX	30
+#define K1_PDMA_UART8_TX	31
+#define K1_PDMA_UART8_RX	32
+#define K1_PDMA_UART9_TX	33
+#define K1_PDMA_UART9_RX	34
+#define K1_PDMA_I2C5_TX		35
+#define K1_PDMA_I2C5_RX		36
+#define K1_PDMA_I2C6_TX		37
+#define K1_PDMA_I2C6_RX		38
+#define K1_PDMA_I2C7_TX		39
+#define K1_PDMA_I2C7_RX		40
+#define K1_PDMA_I2C8_TX		41
+#define K1_PDMA_I2C8_RX		42
+#define K1_PDMA_CAN0_RX		43
+#define K1_PDMA_QSPI_RX		44
+#define K1_PDMA_QSPI_TX		45
+
+#endif /* _DT_BINDINGS_DMA_SPACEMIT_K1_PDMA_H_ */

-- 
2.43.0

[PATCH 2/2] riscv: dts: spacemit: Use symbolic PDMA request numbers on K1

[Guodong Xu]

The K1 SPI3 node's "dmas" property hard-codes its PDMA request numbers.
Include <dt-bindings/dma/spacemit,k1-pdma.h> and use the K1_PDMA_SPI3_RX/TX
macros instead, for better code readability and easy for future
maintenance.

No functional change.

Signed-off-by: Guodong Xu <docular.xu@gmail.com>
---
 arch/riscv/boot/dts/spacemit/k1.dtsi | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/arch/riscv/boot/dts/spacemit/k1.dtsi b/arch/riscv/boot/dts/spacemit/k1.dtsi
index 08a0f28d011fe..c413a64d5560c 100644
--- a/arch/riscv/boot/dts/spacemit/k1.dtsi
+++ b/arch/riscv/boot/dts/spacemit/k1.dtsi
@@ -4,6 +4,7 @@
  */
 
 #include <dt-bindings/clock/spacemit,k1-syscon.h>
+#include <dt-bindings/dma/spacemit,k1-pdma.h>
 #include <dt-bindings/phy/phy.h>
 
 /dts-v1/;
@@ -1094,7 +1095,7 @@ spi3: spi@d401c000 {
 				clock-names = "core", "bus";
 				resets = <&syscon_apbc RESET_SSP3>;
 				interrupts = <55>;
-				dmas = <&pdma 20>, <&pdma 19>;
+				dmas = <&pdma K1_PDMA_SPI3_RX>, <&pdma K1_PDMA_SPI3_TX>;
 				dma-names = "rx", "tx";
 				status = "disabled";
 			};

-- 
2.43.0

Re: [PATCH 1/2] dt-bindings: dmaengine: Add SpacemiT K1 PDMA request numbers

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 3681 bytes --]

On Sun, Jun 07, 2026 at 01:41:30PM -0400, Guodong Xu wrote:
> Add a dt-bindings header that gives symbolic names to the SpacemiT K1
> PDMA request lines of the non-secure peripherals. Device trees can use
> these K1_PDMA_* macros instead of magic numbers.
> 
> Point the spacemit,k1-pdma binding's #dma-cells description at the new
> header.
> 
> Signed-off-by: Guodong Xu <docular.xu@gmail.com>
> ---
>  .../devicetree/bindings/dma/spacemit,k1-pdma.yaml  |  4 +-
>  include/dt-bindings/dma/spacemit,k1-pdma.h         | 56 ++++++++++++++++++++++
>  2 files changed, 59 insertions(+), 1 deletion(-)
> 
> diff --git a/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml b/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
> index ec06235baf5ca..0d4ac9849e27b 100644
> --- a/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
> +++ b/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
> @@ -35,7 +35,9 @@ properties:
>    '#dma-cells':
>      const: 1
>      description:
> -      The DMA request number for the peripheral device.
> +      The single cell is the DMA request number for the peripheral device.
> +      See <dt-bindings/dma/spacemit,k1-pdma.h> for the list of valid request
> +      numbers.
>  
>  required:
>    - compatible
> diff --git a/include/dt-bindings/dma/spacemit,k1-pdma.h b/include/dt-bindings/dma/spacemit,k1-pdma.h

Why does this need to be in a binding when there is no use of this in
the driver? May as well be a header, particularly if these are numbers
with a set meaning that are lifted from the TRM, rather than made up
numbers to make a driver work. The former seems likely, given you're
indexing from 3 not 0.

> new file mode 100644
> index 0000000000000..491976516550a
> --- /dev/null
> +++ b/include/dt-bindings/dma/spacemit,k1-pdma.h
> @@ -0,0 +1,56 @@
> +/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
> +/*
> + * This header provides DMA request number for non-secure peripherals of
> + * SpacemiT K1 PDMA.
> + *
> + * Copyright (c) 2026 Guodong Xu <docular.xu@gmail.com>
> + */
> +
> +#ifndef _DT_BINDINGS_DMA_SPACEMIT_K1_PDMA_H_
> +#define _DT_BINDINGS_DMA_SPACEMIT_K1_PDMA_H_
> +
> +#define K1_PDMA_UART0_TX	3
> +#define K1_PDMA_UART0_RX	4
> +#define K1_PDMA_UART2_TX	5
> +#define K1_PDMA_UART2_RX	6
> +#define K1_PDMA_UART3_TX	7
> +#define K1_PDMA_UART3_RX	8
> +#define K1_PDMA_UART4_TX	9
> +#define K1_PDMA_UART4_RX	10
> +#define K1_PDMA_I2C0_TX		11
> +#define K1_PDMA_I2C0_RX		12
> +#define K1_PDMA_I2C1_TX		13
> +#define K1_PDMA_I2C1_RX		14
> +#define K1_PDMA_I2C2_TX		15
> +#define K1_PDMA_I2C2_RX		16
> +#define K1_PDMA_I2C4_TX		17
> +#define K1_PDMA_I2C4_RX		18
> +#define K1_PDMA_SPI3_TX		19
> +#define K1_PDMA_SPI3_RX		20
> +#define K1_PDMA_I2S0_TX		21
> +#define K1_PDMA_I2S0_RX		22
> +#define K1_PDMA_I2S1_TX		23
> +#define K1_PDMA_I2S1_RX		24
> +#define K1_PDMA_UART5_TX	25
> +#define K1_PDMA_UART5_RX	26
> +#define K1_PDMA_UART6_TX	27
> +#define K1_PDMA_UART6_RX	28
> +#define K1_PDMA_UART7_TX	29
> +#define K1_PDMA_UART7_RX	30
> +#define K1_PDMA_UART8_TX	31
> +#define K1_PDMA_UART8_RX	32
> +#define K1_PDMA_UART9_TX	33
> +#define K1_PDMA_UART9_RX	34
> +#define K1_PDMA_I2C5_TX		35
> +#define K1_PDMA_I2C5_RX		36
> +#define K1_PDMA_I2C6_TX		37
> +#define K1_PDMA_I2C6_RX		38
> +#define K1_PDMA_I2C7_TX		39
> +#define K1_PDMA_I2C7_RX		40
> +#define K1_PDMA_I2C8_TX		41
> +#define K1_PDMA_I2C8_RX		42
> +#define K1_PDMA_CAN0_RX		43
> +#define K1_PDMA_QSPI_RX		44
> +#define K1_PDMA_QSPI_TX		45
> +
> +#endif /* _DT_BINDINGS_DMA_SPACEMIT_K1_PDMA_H_ */
> 
> -- 
> 2.43.0
> 

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH 1/2] dt-bindings: dmaengine: Add SpacemiT K1 PDMA request numbers

[Guodong Xu]

Hi, Conor

On 2026-06-08 18:33, Conor Dooley wrote:
>On Sun, Jun 07, 2026 at 01:41:30PM -0400, Guodong Xu wrote:
>> Add a dt-bindings header that gives symbolic names to the SpacemiT K1
>> PDMA request lines of the non-secure peripherals. Device trees can use
>> these K1_PDMA_* macros instead of magic numbers.
>>
>> Point the spacemit,k1-pdma binding's #dma-cells description at the new
>> header.
>>
>> Signed-off-by: Guodong Xu <docular.xu@gmail.com>
>> ---
>>  .../devicetree/bindings/dma/spacemit,k1-pdma.yaml  |  4 +-
>>  include/dt-bindings/dma/spacemit,k1-pdma.h         | 56 ++++++++++++++++++++++
>>  2 files changed, 59 insertions(+), 1 deletion(-)
>>
>> diff --git a/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml b/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
>> index ec06235baf5ca..0d4ac9849e27b 100644
>> --- a/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
>> +++ b/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
>> @@ -35,7 +35,9 @@ properties:
>>    '#dma-cells':
>>      const: 1
>>      description:
>> -      The DMA request number for the peripheral device.
>> +      The single cell is the DMA request number for the peripheral device.
>> +      See <dt-bindings/dma/spacemit,k1-pdma.h> for the list of valid request
>> +      numbers.
>>
>>  required:
>>    - compatible
>> diff --git a/include/dt-bindings/dma/spacemit,k1-pdma.h b/include/dt-bindings/dma/spacemit,k1-pdma.h
>
>Why does this need to be in a binding when there is no use of this in
>the driver? May as well be a header, particularly if these are numbers

Thanks for the review. You are correct that these are not referenced in the
driver. My change to k1-pdma.yaml should be dropped.

>with a set meaning that are lifted from the TRM, rather than made up
>numbers to make a driver work. The former seems likely, given you're
>indexing from 3 not 0.

Yes, it is defined in the K1 manual [1], see 9.4.3 DMA Connectivity &
Assignments

Link: https://www.spacemit.com/community/document/info?lang=en&nodepath=hardware/key_stone/k1/k1_docs/k1_usermanual/9.Top_System.md [1]

I will fix that in v2.

BR,
Guodong

>
>> new file mode 100644
>> index 0000000000000..491976516550a
>> --- /dev/null
>> +++ b/include/dt-bindings/dma/spacemit,k1-pdma.h
>> @@ -0,0 +1,56 @@
>> +/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
>> +/*
>> + * This header provides DMA request number for non-secure peripherals of
>> + * SpacemiT K1 PDMA.
>> + *
>> + * Copyright (c) 2026 Guodong Xu <docular.xu@gmail.com>
>> + */
>> +
>> +#ifndef _DT_BINDINGS_DMA_SPACEMIT_K1_PDMA_H_
>> +#define _DT_BINDINGS_DMA_SPACEMIT_K1_PDMA_H_
>> +
>> +#define K1_PDMA_UART0_TX	3
>> +#define K1_PDMA_UART0_RX	4
>> +#define K1_PDMA_UART2_TX	5
>> +#define K1_PDMA_UART2_RX	6
>> +#define K1_PDMA_UART3_TX	7
>> +#define K1_PDMA_UART3_RX	8
>> +#define K1_PDMA_UART4_TX	9
>> +#define K1_PDMA_UART4_RX	10
>> +#define K1_PDMA_I2C0_TX		11
>> +#define K1_PDMA_I2C0_RX		12
>> +#define K1_PDMA_I2C1_TX		13
>> +#define K1_PDMA_I2C1_RX		14
>> +#define K1_PDMA_I2C2_TX		15
>> +#define K1_PDMA_I2C2_RX		16
>> +#define K1_PDMA_I2C4_TX		17
>> +#define K1_PDMA_I2C4_RX		18
>> +#define K1_PDMA_SPI3_TX		19
>> +#define K1_PDMA_SPI3_RX		20
>> +#define K1_PDMA_I2S0_TX		21
>> +#define K1_PDMA_I2S0_RX		22
>> +#define K1_PDMA_I2S1_TX		23
>> +#define K1_PDMA_I2S1_RX		24
>> +#define K1_PDMA_UART5_TX	25
>> +#define K1_PDMA_UART5_RX	26
>> +#define K1_PDMA_UART6_TX	27
>> +#define K1_PDMA_UART6_RX	28
>> +#define K1_PDMA_UART7_TX	29
>> +#define K1_PDMA_UART7_RX	30
>> +#define K1_PDMA_UART8_TX	31
>> +#define K1_PDMA_UART8_RX	32
>> +#define K1_PDMA_UART9_TX	33
>> +#define K1_PDMA_UART9_RX	34
>> +#define K1_PDMA_I2C5_TX		35
>> +#define K1_PDMA_I2C5_RX		36
>> +#define K1_PDMA_I2C6_TX		37
>> +#define K1_PDMA_I2C6_RX		38
>> +#define K1_PDMA_I2C7_TX		39
>> +#define K1_PDMA_I2C7_RX		40
>> +#define K1_PDMA_I2C8_TX		41
>> +#define K1_PDMA_I2C8_RX		42
>> +#define K1_PDMA_CAN0_RX		43
>> +#define K1_PDMA_QSPI_RX		44
>> +#define K1_PDMA_QSPI_TX		45
>> +
>> +#endif /* _DT_BINDINGS_DMA_SPACEMIT_K1_PDMA_H_ */
>>
>> --
>> 2.43.0
>>

Re: [PATCH 1/2] dt-bindings: dmaengine: Add SpacemiT K1 PDMA request numbers

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 2534 bytes --]

On Tue, Jun 09, 2026 at 02:55:59PM -0400, Guodong Xu wrote:
> Hi, Conor
> 
> On 2026-06-08 18:33, Conor Dooley wrote:
> > On Sun, Jun 07, 2026 at 01:41:30PM -0400, Guodong Xu wrote:
> > > Add a dt-bindings header that gives symbolic names to the SpacemiT K1
> > > PDMA request lines of the non-secure peripherals. Device trees can use
> > > these K1_PDMA_* macros instead of magic numbers.
> > > 
> > > Point the spacemit,k1-pdma binding's #dma-cells description at the new
> > > header.
> > > 
> > > Signed-off-by: Guodong Xu <docular.xu@gmail.com>
> > > ---
> > >  .../devicetree/bindings/dma/spacemit,k1-pdma.yaml  |  4 +-
> > >  include/dt-bindings/dma/spacemit,k1-pdma.h         | 56 ++++++++++++++++++++++
> > >  2 files changed, 59 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml b/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
> > > index ec06235baf5ca..0d4ac9849e27b 100644
> > > --- a/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
> > > +++ b/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
> > > @@ -35,7 +35,9 @@ properties:
> > >    '#dma-cells':
> > >      const: 1
> > >      description:
> > > -      The DMA request number for the peripheral device.
> > > +      The single cell is the DMA request number for the peripheral device.
> > > +      See <dt-bindings/dma/spacemit,k1-pdma.h> for the list of valid request
> > > +      numbers.
> > > 
> > >  required:
> > >    - compatible
> > > diff --git a/include/dt-bindings/dma/spacemit,k1-pdma.h b/include/dt-bindings/dma/spacemit,k1-pdma.h
> > 
> > Why does this need to be in a binding when there is no use of this in
> > the driver? May as well be a header, particularly if these are numbers
> 
> Thanks for the review. You are correct that these are not referenced in the
> driver. My change to k1-pdma.yaml should be dropped.
> 
> > with a set meaning that are lifted from the TRM, rather than made up
> > numbers to make a driver work. The former seems likely, given you're
> > indexing from 3 not 0.
> 
> Yes, it is defined in the K1 manual [1], see 9.4.3 DMA Connectivity &
> Assignments
> 
> Link: https://www.spacemit.com/community/document/info?lang=en&nodepath=hardware/key_stone/k1/k1_docs/k1_usermanual/9.Top_System.md [1]
> 
> I will fix that in v2.

Just in case I wasn't clear (and I think I wasn't), when I said "may as
well be a header" I meant a header in arch/riscv/boot/dts/spacemit.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH 1/2] dt-bindings: dmaengine: Add SpacemiT K1 PDMA request numbers

[Guodong Xu]

On 2026-06-09 20:54, Conor Dooley wrote:
>On Tue, Jun 09, 2026 at 02:55:59PM -0400, Guodong Xu wrote:
>> Hi, Conor
>>
>> On 2026-06-08 18:33, Conor Dooley wrote:
>> > On Sun, Jun 07, 2026 at 01:41:30PM -0400, Guodong Xu wrote:
>> > > Add a dt-bindings header that gives symbolic names to the SpacemiT K1
>> > > PDMA request lines of the non-secure peripherals. Device trees can use
>> > > these K1_PDMA_* macros instead of magic numbers.
>> > >
>> > > Point the spacemit,k1-pdma binding's #dma-cells description at the new
>> > > header.
>> > >
>> > > Signed-off-by: Guodong Xu <docular.xu@gmail.com>
>> > > ---
>> > >  .../devicetree/bindings/dma/spacemit,k1-pdma.yaml  |  4 +-
>> > >  include/dt-bindings/dma/spacemit,k1-pdma.h         | 56 ++++++++++++++++++++++
>> > >  2 files changed, 59 insertions(+), 1 deletion(-)
>> > >
>> > > diff --git a/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml b/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
>> > > index ec06235baf5ca..0d4ac9849e27b 100644
>> > > --- a/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
>> > > +++ b/Documentation/devicetree/bindings/dma/spacemit,k1-pdma.yaml
>> > > @@ -35,7 +35,9 @@ properties:
>> > >    '#dma-cells':
>> > >      const: 1
>> > >      description:
>> > > -      The DMA request number for the peripheral device.
>> > > +      The single cell is the DMA request number for the peripheral device.
>> > > +      See <dt-bindings/dma/spacemit,k1-pdma.h> for the list of valid request
>> > > +      numbers.
>> > >
>> > >  required:
>> > >    - compatible
>> > > diff --git a/include/dt-bindings/dma/spacemit,k1-pdma.h b/include/dt-bindings/dma/spacemit,k1-pdma.h
>> >
>> > Why does this need to be in a binding when there is no use of this in
>> > the driver? May as well be a header, particularly if these are numbers
>>
>> Thanks for the review. You are correct that these are not referenced in the
>> driver. My change to k1-pdma.yaml should be dropped.
>>
>> > with a set meaning that are lifted from the TRM, rather than made up
>> > numbers to make a driver work. The former seems likely, given you're
>> > indexing from 3 not 0.
>>
>> Yes, it is defined in the K1 manual [1], see 9.4.3 DMA Connectivity &
>> Assignments
>>
>> Link: https://www.spacemit.com/community/document/info?lang=en&nodepath=hardware/key_stone/k1/k1_docs/k1_usermanual/9.Top_System.md [1]
>>
>> I will fix that in v2.
>
>Just in case I wasn't clear (and I think I wasn't), when I said "may as
>well be a header" I meant a header in arch/riscv/boot/dts/spacemit.

Oh, got it. Makes sense. I will move it to
arch/riscv/boot/dts/spacemit/k1-pdma.h

Note that I already sent v2 before reading this, please disregard v2.

I will send v3.

BR,
Guodong