Thread (13 messages) 13 messages, 4 authors, 2026-02-08

Re: [PATCH 1/3] dt-bindings: mmc: Add sdhci support for Canaan k230

From: Conor Dooley <conor@kernel.org>
Date: 2026-02-05 19:19:42
Also in: linux-mmc, linux-riscv, lkml

On Thu, Feb 05, 2026 at 03:13:00PM +0800, Jiayu Du wrote:
On Wed, Feb 04, 2026 at 06:10:25PM +0000, Conor Dooley wrote:
quoted
On Wed, Feb 04, 2026 at 04:29:06PM +0800, Jiayu Du wrote:
quoted
The Canaan k230 uses the SDHCI from Synopsys. Add compatible strings
to the k230. The k230 has two controllers. MMC0 supports eMMC, while
MMC1 supports SDIO.

Signed-off-by: Jiayu Du <redacted>
---
 .../bindings/mmc/snps,dwcmshc-sdhci.yaml      | 22 +++++++++++++++++++
 1 file changed, 22 insertions(+)
diff --git a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
index 7e7c55dc2440..cab33da3af7d 100644
--- a/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
+++ b/Documentation/devicetree/bindings/mmc/snps,dwcmshc-sdhci.yaml
@@ -23,6 +23,8 @@ properties:
           - const: sophgo,sg2044-dwcmshc
           - const: sophgo,sg2042-dwcmshc
       - enum:
+          - canaan,k230-emmc
+          - canaan,k230-sdio
I don't understand why there are two compatibles here, when the driver
is able to handle them both identically, using the common properties for
mmc controllers to differentiate. Is which is emmc and which sdio
actually just determined by how the k230 boards use them?

Or, if there are pinout differences, would pinctrl allow each to be used
for either?
Hi, Thanks for the review!

Actually, the two compatibles aren't just board choices. They come from
real SoC-level differences between the two MMC controllers.

From the K230 Hardware Guide [1]:
- MMC0 supports eMMC5.0 and SDIO3.0, usually for eMMC chips.
- MMC1 only does SDIO3.0 in 4/1-bit mode up to SDR104, and the manual
  clearly says it can't handle eMMC because of pin count and limits.

And the driver treats them a bit differently, such as whether there
is a phy and the initialization method. So I think keeping separate
compatibles makes sense to show these hardware differences clearly.
Ah, I didn't notice that there was some logic, I just checked the match
data and I missed that. What you have is fine then I think.
About pinctrl, MMC1 pins can mux to other things like GPIO, but MMC0
pins are fixed in hardware. In next version, I will add pinctrl node
for MMC1 in the board dts.

Link: https://github.com/kendryte/k230_docs/blob/main/en/00_hardware/K230_Hardware_Design_Guide.md#mmc-circuit [1]
quoted
quoted
           - rockchip,rk3568-dwcmshc
           - rockchip,rk3588-dwcmshc
           - snps,dwcmshc-sdhci
@@ -87,6 +89,26 @@ required:
 allOf:
   - $ref: mmc-controller.yaml#
 
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - canaan,k230-emmc
+              - canaan,k230-sdio
+    then:
+      properties:
+        clocks:
+          minItems: 2
+          maxItems: 5
Additionally, why does this have a range? Why is it not minItems: 5?
You've got only one instance per compatible, so the number of clocks for
each compatible is not variable.
quoted
quoted
+        clock-names:
+          items:
+            - const: core
+            - const: bus
+            - const: axi
+            - const: block
+            - const: timer
+
   - if:
       properties:
         compatible:
-- 
2.52.0

Attachments

Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help