Skip to contentSkip to conversation

ARM: dts: Removing pdma for SSDK5440

PATCH New Kukjin Kim 2 participants 3 messages Jun 11, 2013

Subsystems (2)
lore
Share & tools
lore

Kukjin Kim

author
#0 lore raw

PATCH 1/1 ARM: dts: Removing pdma for SSDK5440

From: Giridhar Maruthy <[email protected]>

Since the pdma works only in secure mode, accessing the same in hypervisor mode gives an abort. As we are not using pdma anywhere (including spi), removing the same.

---
 arch/arm/boot/dts/exynos5440.dtsi |   24 ------------------------
 1 file changed, 24 deletions(-)
arch/arm/boot/dts/exynos5440.dtsi +0 -24
index 871ebad..b08449e 100644
@@ -105,8 +105,6 @@
 		compatible = "samsung,exynos4210-spi";
 		reg = <0xD0000 0x1000>;
 		interrupts = <0 4 0>;
-		tx-dma-channel = <&pdma0 5>; /* preliminary */
-		rx-dma-channel = <&pdma0 4>; /* preliminary */
 		#address-cells = <1>;
 		#size-cells = <0>;
 		clocks = <&clock 21>, <&clock 16>;
@@ -184,28 +182,6 @@
 		compatible = "arm,amba-bus";
 		interrupt-parent = <&gic>;
 		ranges;
-
-		pdma0: pdma at 00121000 {
-			compatible = "arm,pl330", "arm,primecell";
-			reg = <0x121000 0x1000>;
-			interrupts = <0 46 0>;
-			clocks = <&clock 8>;
-			clock-names = "apb_pclk";
-			#dma-cells = <1>;
-			#dma-channels = <8>;
-			#dma-requests = <32>;
-		};
-
-		pdma1: pdma at 00120000 {
-			compatible = "arm,pl330", "arm,primecell";
-			reg = <0x120000 0x1000>;
-			interrupts = <0 47 0>;
-			clocks = <&clock 8>;
-			clock-names = "apb_pclk";
-			#dma-cells = <1>;
-			#dma-channels = <8>;
-			#dma-requests = <32>;
-		};
 	};
 
 	rtc {
Signature · 1 line
1.7.10.4
#1 lore raw

PATCH 1/1 ARM: dts: Removing pdma for SSDK5440

Quoted 8 lines · Kukjin Kim

From: Giridhar Maruthy <[email protected]>

Since the pdma works only in secure mode, accessing the same in hypervisor mode gives an abort. As we are not using pdma anywhere (including spi), removing the same.

I was wondering, wouldn't it be sufficient to add status = "disabled"; in those pdma nodes, to avoid re-adding them in future ?
As for the DMA properties in the spi device nodes, they should indeed be removed, as generic DMA bindings would be used instead.

Regards,
Sylwester

Quoted 44 lines
---
 arch/arm/boot/dts/exynos5440.dtsi |   24 ------------------------
 1 file changed, 24 deletions(-)
arch/arm/boot/dts/exynos5440.dtsi +0 -24
index 871ebad..b08449e 100644
@@ -105,8 +105,6 @@
 		compatible = "samsung,exynos4210-spi";Hmm, is there really no way to use PDMA in non-secure mode at all ?
 		reg = <0xD0000 0x1000>;
 		interrupts = <0 4 0>;
-		tx-dma-channel = <&pdma0 5>; /* preliminary */
-		rx-dma-channel = <&pdma0 4>; /* preliminary */
 		#address-cells = <1>;
 		#size-cells = <0>;
 		clocks = <&clock 21>, <&clock 16>;
@@ -184,28 +182,6 @@
 		compatible = "arm,amba-bus";
 		interrupt-parent = <&gic>;
 		ranges;
-
-		pdma0: pdma at 00121000 {
-			compatible = "arm,pl330", "arm,primecell";
-			reg = <0x121000 0x1000>;
-			interrupts = <0 46 0>;
-			clocks = <&clock 8>;
-			clock-names = "apb_pclk";
-			#dma-cells = <1>;
-			#dma-channels = <8>;
-			#dma-requests = <32>;
-		};
-
-		pdma1: pdma at 00120000 {
-			compatible = "arm,pl330", "arm,primecell";
-			reg = <0x120000 0x1000>;
-			interrupts = <0 47 0>;
-			clocks = <&clock 8>;
-			clock-names = "apb_pclk";
-			#dma-cells = <1>;
-			#dma-channels = <8>;
-			#dma-requests = <32>;
-		};
 	};
author
#2 lore raw

PATCH 1/1 ARM: dts: Removing pdma for SSDK5440

Quoted 1 line · Sylwester Nawrocki

Hi Sylwester,

Quoted 7 lines · Kukjin Kim
Quoted 6 lines

From: Giridhar Maruthy <[email protected]>

Since the pdma works only in secure mode, accessing the same in hypervisor mode gives an abort. As we are not using pdma anywhere (including spi), removing the same.

Quoted 5 lines
Quoted 2 lines

I was wondering, wouldn't it be sufficient to add status = "disabled"; in those pdma nodes, to avoid re-adding them in future ?

Yeah, could be. But the pdma will not be used on exynos5440 in future and if some guy use it in non-secure mode, it will cause problem, so I want to remove that.

Quoted 3 lines

As for the DMA properties in the spi device nodes, they should indeed be removed, as generic DMA bindings would be used instead.

Yes, agreed. So I'd prefer to remove that :-)

Thanks,
Kukjin