Hi Maniyam,
On Thu, 25 Jun 2026 19:23:36 +0800, [email protected] wrote:
From: Dinesh Maniyam <[email protected]>
Introduce a separate DTS for NAND2 boot flow where MMC is not
enabled. This ensures that only NAND configures the ComboPHY,
avoiding conflicts caused by DT not matching handoff ownership.
New Agilex5 NAND2 DTS files not listed in board MAINTAINERS
Please add F: lines for arch/arm/dts/socfpga_agilex5_socdk_nand2.dts and
arch/arm/dts/socfpga_agilex5_socdk_nand2-u-boot.dtsi (or a shared
socfpga_agilex5* pattern) next to the existing nand2_defconfig entry.
diff --git a/arch/arm/dts/socfpga_agilex5_socdk_nand2-u-boot.dtsi
b/arch/arm/dts/socfpga_agilex5_socdk_nand2-u-boot.dtsi
@@ -0,0 +98,3 @@
+
+&sd_emmc_power {
+ bootph-all;
[...]
Disabled MMC still carries bootph-all and SD regulator bootph nodes
Please drop bootph-all from &mmc when status is disabled, and omit
&sd_emmc_power / &sd_io_1v8_reg bootph overlays unless another consumer
needs them on NAND2.
diff --git a/arch/arm/dts/socfpga_agilex5_socdk_nand2-u-boot.dtsi
b/arch/arm/dts/socfpga_agilex5_socdk_nand2-u-boot.dtsi
@@ -0,0 +158,10 @@
+
+&gmac2 {
+ status = "disabled";
+ phy-mode = "rgmii";
+ phy-handle = <&emac2_phy0>;
+
+ max-frame-size = <9000>;
+
+ mdio0 {
+ #address-cells = <1>;
[...]
gmac2 disabled vs socdk/emmc without commit-message rationale
If intentional for the NAND2 board, please mention the gmac2 disable in
the commit message; if accidental from the emmc/socdk fork, restore
status = "okay" to match siblings.
Best regards,
Tien Fong
