Patchwork [RFC] imx6: imx6duallite/solo i2c4 clock

login
register
mail settings
Submitter Iain Paton
Date April 15, 2014, 9:15 a.m.
Message ID <534CF8A3.3010404@gmail.com>
Download mbox | patch
Permalink /patch/339210/
State New
Headers show

Comments

Iain Paton - April 15, 2014, 9:15 a.m.
I'm working on a board using the imx6solo that implements i2c4 and while 
building the devicetree it became clear that the devicetree fragment for 
i2c4 in imx6dl.dtsi it's missing the necessary clock.

Investigating further turned up that the ecspi5 clock in CCM_CCGR1 on the 
quad/dual is replaced with the i2c4 clock on duallite/solo.

Would the following patch be sufficient to allow correct use of this clock?

I'm not familiar enough with the clock framework to be sure I'm not missing
something here, but have tested this patch on the board and found it to work.

The board is element14's RIoTboard http://www.riotboard.org and I have dts
mostly ready to submit for it pending resolving this and a couple of other 
minor things, and assuming Eric doesn't beat me to it.

Iain
Shawn Guo - April 16, 2014, 7:47 a.m.
On Tue, Apr 15, 2014 at 10:15:15AM +0100, Iain Paton wrote:
> I'm working on a board using the imx6solo that implements i2c4 and while 
> building the devicetree it became clear that the devicetree fragment for 
> i2c4 in imx6dl.dtsi it's missing the necessary clock.
> 
> Investigating further turned up that the ecspi5 clock in CCM_CCGR1 on the 
> quad/dual is replaced with the i2c4 clock on duallite/solo.
> 
> Would the following patch be sufficient to allow correct use of this clock?

Yes, Iain.  It looks correct to me.  Care to send a formal patch for it?

Shawn

> 
> I'm not familiar enough with the clock framework to be sure I'm not missing
> something here, but have tested this patch on the board and found it to work.
> 
> The board is element14's RIoTboard http://www.riotboard.org and I have dts
> mostly ready to submit for it pending resolving this and a couple of other 
> minor things, and assuming Eric doesn't beat me to it.
> 
> Iain
> 
> diff --git a/arch/arm/mach-imx/clk-imx6q.c b/arch/arm/mach-imx/clk-imx6q.c
> index b0e7f9d..2961b16 100644
> --- a/arch/arm/mach-imx/clk-imx6q.c
> +++ b/arch/arm/mach-imx/clk-imx6q.c
> @@ -352,7 +352,13 @@ static void __init imx6q_clocks_init(struct device_node *ccm_node)
>  	clk[ecspi2]       = imx_clk_gate2("ecspi2",        "ecspi_root",        base + 0x6c, 2);
>  	clk[ecspi3]       = imx_clk_gate2("ecspi3",        "ecspi_root",        base + 0x6c, 4);
>  	clk[ecspi4]       = imx_clk_gate2("ecspi4",        "ecspi_root",        base + 0x6c, 6);
> -	clk[ecspi5]       = imx_clk_gate2("ecspi5",        "ecspi_root",        base + 0x6c, 8);
> +	if (cpu_is_imx6dl())
> +		/*
> +		 * ecspi5 is replaced with i2c4 on imx6dl & imx6s
> +		 */
> +		clk[ecspi5] = imx_clk_gate2("i2c4",        "ipg_per",           base + 0x6c, 8);
> +	else
> +		clk[ecspi5] = imx_clk_gate2("ecspi5",      "ecspi_root",        base + 0x6c, 8);
>  	clk[enet]         = imx_clk_gate2("enet",          "ipg",               base + 0x6c, 10);
>  	clk[esai]         = imx_clk_gate2("esai",          "esai_podf",         base + 0x6c, 16);
>  	clk[gpt_ipg]      = imx_clk_gate2("gpt_ipg",       "ipg",               base + 0x6c, 20);
> 
>
Iain Paton - April 16, 2014, 6:33 p.m.
On 16/04/14 08:47, Shawn Guo wrote:
> On Tue, Apr 15, 2014 at 10:15:15AM +0100, Iain Paton wrote:
>> I'm working on a board using the imx6solo that implements i2c4 and while 
>> building the devicetree it became clear that the devicetree fragment for 
>> i2c4 in imx6dl.dtsi it's missing the necessary clock.
>>
>> Investigating further turned up that the ecspi5 clock in CCM_CCGR1 on the 
>> quad/dual is replaced with the i2c4 clock on duallite/solo.
>>
>> Would the following patch be sufficient to allow correct use of this clock?
> 
> Yes, Iain.  It looks correct to me.  Care to send a formal patch for it?

Sure, it's on it's way.

Patch

diff --git a/arch/arm/mach-imx/clk-imx6q.c b/arch/arm/mach-imx/clk-imx6q.c
index b0e7f9d..2961b16 100644
--- a/arch/arm/mach-imx/clk-imx6q.c
+++ b/arch/arm/mach-imx/clk-imx6q.c
@@ -352,7 +352,13 @@  static void __init imx6q_clocks_init(struct device_node *ccm_node)
 	clk[ecspi2]       = imx_clk_gate2("ecspi2",        "ecspi_root",        base + 0x6c, 2);
 	clk[ecspi3]       = imx_clk_gate2("ecspi3",        "ecspi_root",        base + 0x6c, 4);
 	clk[ecspi4]       = imx_clk_gate2("ecspi4",        "ecspi_root",        base + 0x6c, 6);
-	clk[ecspi5]       = imx_clk_gate2("ecspi5",        "ecspi_root",        base + 0x6c, 8);
+	if (cpu_is_imx6dl())
+		/*
+		 * ecspi5 is replaced with i2c4 on imx6dl & imx6s
+		 */
+		clk[ecspi5] = imx_clk_gate2("i2c4",        "ipg_per",           base + 0x6c, 8);
+	else
+		clk[ecspi5] = imx_clk_gate2("ecspi5",      "ecspi_root",        base + 0x6c, 8);
 	clk[enet]         = imx_clk_gate2("enet",          "ipg",               base + 0x6c, 10);
 	clk[esai]         = imx_clk_gate2("esai",          "esai_podf",         base + 0x6c, 16);
 	clk[gpt_ipg]      = imx_clk_gate2("gpt_ipg",       "ipg",               base + 0x6c, 20);