1. Patchew
  2. linux
  3. [PATCH v2 0/7] dmaengine: Add Peripheral DMA support for SpacemiT K3 SoC
  4. View patch

[PATCH v2 3/7] dt-bindings: dmaengine: Add SpacemiT K3 DMA request definitions

Troy Mitchell posted 7 patches 1 week ago
There is a newer version of this series
[PATCH v2 3/7] dt-bindings: dmaengine: Add SpacemiT K3 DMA request definitions
Posted by Troy Mitchell 1 week ago 
From: liyeshan <yeshan.li@spacemit.com>

Add device tree binding header for SpacemiT k3 DMA request numbers. This
defines the DMA request mapping for non-secure peripherals including UART,
I2C, SSP/SPI, CAN, and QSPI.

Signed-off-by: liyeshan <yeshan.li@spacemit.com>
Signed-off-by: Guodong Xu <guodong@riscstar.com>
Signed-off-by: Troy Mitchell <troy.mitchell@linux.spacemit.com>
---
 include/dt-bindings/dma/k3-pdma.h | 83 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 83 insertions(+)

diff --git a/include/dt-bindings/dma/k3-pdma.h b/include/dt-bindings/dma/k3-pdma.h
new file mode 100644
index 000000000000..05541a9a9973
--- /dev/null
+++ b/include/dt-bindings/dma/k3-pdma.h
@@ -0,0 +1,83 @@
+/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
+/*
+ * This header provides DMA request number for non-secure peripherals of
+ * SpacemiT K3 PDMA.
+ *
+ * Copyright (c) 2025 SpacemiT
+ * Copyright (c) 2025 Guodong Xu <guodong@riscstar.com>
+ */
+
+#ifndef __DT_BINDINGS_DMA_K3_PDMA_H__
+#define __DT_BINDINGS_DMA_K3_PDMA_H__
+
+/* UART DMA request numbers */
+#define K3_PDMA_UART0_TX	3
+#define K3_PDMA_UART0_RX	4
+#define K3_PDMA_UART2_TX	5
+#define K3_PDMA_UART2_RX	6
+#define K3_PDMA_UART3_TX	7
+#define K3_PDMA_UART3_RX	8
+#define K3_PDMA_UART4_TX	9
+#define K3_PDMA_UART4_RX	10
+#define K3_PDMA_UART5_TX	25
+#define K3_PDMA_UART5_RX	26
+#define K3_PDMA_UART6_TX	27
+#define K3_PDMA_UART6_RX	28
+#define K3_PDMA_UART7_TX	29
+#define K3_PDMA_UART7_RX	30
+#define K3_PDMA_UART8_TX	31
+#define K3_PDMA_UART8_RX	32
+#define K3_PDMA_UART9_TX	33
+#define K3_PDMA_UART9_RX	34
+#define K3_PDMA_UART10_TX	53
+#define K3_PDMA_UART10_RX	54
+
+/* I2C DMA request numbers */
+#define K3_PDMA_I2C0_TX	11
+#define K3_PDMA_I2C0_RX	12
+#define K3_PDMA_I2C1_TX	13
+#define K3_PDMA_I2C1_RX	14
+#define K3_PDMA_I2C2_TX	15
+#define K3_PDMA_I2C2_RX	16
+#define K3_PDMA_I2C4_TX	17
+#define K3_PDMA_I2C4_RX	18
+#define K3_PDMA_I2C5_TX	35
+#define K3_PDMA_I2C5_RX	36
+#define K3_PDMA_I2C6_TX	37
+#define K3_PDMA_I2C6_RX	38
+#define K3_PDMA_I2C8_TX	41
+#define K3_PDMA_I2C8_RX	42
+
+/* SSP/SPI DMA request numbers */
+#define K3_PDMA_SSP3_TX	19
+#define K3_PDMA_SSP3_RX	20
+#define K3_PDMA_SSPA0_TX	21
+#define K3_PDMA_SSPA0_RX	22
+#define K3_PDMA_SSPA1_TX	23
+#define K3_PDMA_SSPA1_RX	24
+#define K3_PDMA_SSPA2_TX	56
+#define K3_PDMA_SSPA2_RX	57
+#define K3_PDMA_SSPA3_TX	58
+#define K3_PDMA_SSPA3_RX	59
+#define K3_PDMA_SSPA4_TX	60
+#define K3_PDMA_SSPA4_RX	61
+#define K3_PDMA_SSPA5_TX	62
+#define K3_PDMA_SSPA5_RX	63
+
+/* CAN DMA request numbers */
+#define K3_PDMA_CAN0_RX	43
+#define K3_PDMA_CAN1_RX	44
+#define K3_PDMA_CAN2_RX	51
+#define K3_PDMA_CAN3_RX	52
+
+/* SSP0/1 DMA request numbers */
+#define K3_PDMA_SSP0_TX	64
+#define K3_PDMA_SSP0_RX	65
+#define K3_PDMA_SSP1_TX	66
+#define K3_PDMA_SSP1_RX	67
+
+/* QSPI DMA request numbers */
+#define K3_PDMA_QSPI_RX	84
+#define K3_PDMA_QSPI_TX	85
+
+#endif /* __DT_BINDINGS_DMA_K3_PDMA_H__ */

-- 
2.53.0
Re: [PATCH v2 3/7] dt-bindings: dmaengine: Add SpacemiT K3 DMA request definitions
Posted by Krzysztof Kozlowski 6 days, 14 hours ago 
On Thu, Mar 26, 2026 at 04:17:18PM +0800, Troy Mitchell wrote:
> From: liyeshan <yeshan.li@spacemit.com>
> 
> Add device tree binding header for SpacemiT k3 DMA request numbers. This

Why?

> defines the DMA request mapping for non-secure peripherals including UART,
> I2C, SSP/SPI, CAN, and QSPI.
> 
> Signed-off-by: liyeshan <yeshan.li@spacemit.com>

Name looks close to login name?

> Signed-off-by: Guodong Xu <guodong@riscstar.com>
> Signed-off-by: Troy Mitchell <troy.mitchell@linux.spacemit.com>
> ---
>  include/dt-bindings/dma/k3-pdma.h | 83 +++++++++++++++++++++++++++++++++++++++

I am already confused what is happening in this patchset - so which
device are you adding? K1 or K3?

>  1 file changed, 83 insertions(+)
> 
> diff --git a/include/dt-bindings/dma/k3-pdma.h b/include/dt-bindings/dma/k3-pdma.h
> new file mode 100644
> index 000000000000..05541a9a9973
> --- /dev/null
> +++ b/include/dt-bindings/dma/k3-pdma.h
> @@ -0,0 +1,83 @@
> +/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
> +/*
> + * This header provides DMA request number for non-secure peripherals of
> + * SpacemiT K3 PDMA.
> + *
> + * Copyright (c) 2025 SpacemiT
> + * Copyright (c) 2025 Guodong Xu <guodong@riscstar.com>
> + */
> +
> +#ifndef __DT_BINDINGS_DMA_K3_PDMA_H__
> +#define __DT_BINDINGS_DMA_K3_PDMA_H__
> +
> +/* UART DMA request numbers */
> +#define K3_PDMA_UART0_TX	3

This starts from 0 or 1.

Best regards,
Krzysztof
Re: [PATCH v2 3/7] dt-bindings: dmaengine: Add SpacemiT K3 DMA request definitions
Posted by Troy Mitchell 5 days, 12 hours ago 
On Fri Mar 27, 2026 at 3:30 PM CST, Krzysztof Kozlowski wrote:
> On Thu, Mar 26, 2026 at 04:17:18PM +0800, Troy Mitchell wrote:
>> From: liyeshan <yeshan.li@spacemit.com>
>> 
>> Add device tree binding header for SpacemiT k3 DMA request numbers. This
>
> Why?
The DMA request mappings are hardware-hardwired and differ between the K1 and K3 SoCs.
Therefore, separate header files are required to define these unique constants for each
platform to ensure correct DMA channel allocation in the device tree.
>
>> defines the DMA request mapping for non-secure peripherals including UART,
>> I2C, SSP/SPI, CAN, and QSPI.
>> 
>> Signed-off-by: liyeshan <yeshan.li@spacemit.com>
>
> Name looks close to login name?
I will contact her and confirm.
>
>> Signed-off-by: Guodong Xu <guodong@riscstar.com>
>> Signed-off-by: Troy Mitchell <troy.mitchell@linux.spacemit.com>
>> ---
>>  include/dt-bindings/dma/k3-pdma.h | 83 +++++++++++++++++++++++++++++++++++++++
>
> I am already confused what is happening in this patchset - so which
> device are you adding? K1 or K3?
Aside from this K1-related patch, the rest of the series focuses on adding support for
the K3 SoC. I included the K1 header because I noticed it was missing in the upstream
tree while working on the K3 support.

It seems my attempt to 'clean up' K1 while adding K3 has caused unnecessary confusion.
Would you prefer me to remove the K1 patch from this series and submit it as a separate
patch?

                                    - Troy
Re: [PATCH v2 3/7] dt-bindings: dmaengine: Add SpacemiT K3 DMA request definitions
Posted by Krzysztof Kozlowski 5 days, 10 hours ago 
On 28/03/2026 10:58, Troy Mitchell wrote:
> On Fri Mar 27, 2026 at 3:30 PM CST, Krzysztof Kozlowski wrote:
>> On Thu, Mar 26, 2026 at 04:17:18PM +0800, Troy Mitchell wrote:
>>> From: liyeshan <yeshan.li@spacemit.com>
>>>
>>> Add device tree binding header for SpacemiT k3 DMA request numbers. This
>>
>> Why?
> The DMA request mappings are hardware-hardwired and differ between the K1 and K3 SoCs.
> Therefore, separate header files are required to define these unique constants for each
> platform to ensure correct DMA channel allocation in the device tree.

Nothing explains why we want this as ABI. Header does not ensure correct
DMA channel allocation.


>>
>>> defines the DMA request mapping for non-secure peripherals including UART,
>>> I2C, SSP/SPI, CAN, and QSPI.
>>>
>>> Signed-off-by: liyeshan <yeshan.li@spacemit.com>
>>
>> Name looks close to login name?
> I will contact her and confirm.
>>
>>> Signed-off-by: Guodong Xu <guodong@riscstar.com>
>>> Signed-off-by: Troy Mitchell <troy.mitchell@linux.spacemit.com>
>>> ---
>>>  include/dt-bindings/dma/k3-pdma.h | 83 +++++++++++++++++++++++++++++++++++++++
>>
>> I am already confused what is happening in this patchset - so which
>> device are you adding? K1 or K3?
> Aside from this K1-related patch, the rest of the series focuses on adding support for
> the K3 SoC. I included the K1 header because I noticed it was missing in the upstream
> tree while working on the K3 support.

So k3 is the new device? Then this is not a separate commit, but belongs
to the bindings patch.

> 
> It seems my attempt to 'clean up' K1 while adding K3 has caused unnecessary confusion.
> Would you prefer me to remove the K1 patch from this series and submit it as a separate
> patch?

I expect explaining in commit msg why you are doing things.

So far none of these are bindings, so drop both headers from bindings.
Feel free to provide explanation what part of Linux SW you are binding here.

Best regards,
Krzysztof

Patchew 2.1-devel-ea86937

© 2016 - 2026 Red Hat, Inc.