[PATCH 0/2] OMAP4: McBSP: Fix clock reparenting

[PATCH 0/2] OMAP4: McBSP: Fix clock reparenting

[Peter Ujfalusi]

Hello,

OMAP4 has different clock names compared to OMAP2/3.
Restructure the clock reparenting code for omap2plus in order to be able to
suport OMAP4 clock source selection without breaking upper layer API.

Regards,
Peter
---
Peter Ujfalusi (2):
  OMAP2+: mcbsp: Restructure clk reparenting code
  OMAP4: mcbsp: Clock reparenting support

 arch/arm/mach-omap2/mcbsp.c |   66 +++++++++++++++++++++++++++++++-----------
 1 files changed, 48 insertions(+), 18 deletions(-)

-- 
1.7.8

[PATCH 1/2] OMAP2+: mcbsp: Restructure clk reparenting code

[Peter Ujfalusi]

In OMAP4 we have different clk names compared to OMAP2/3.
To be able to support OMAP4 clock reparenting the set_clk_src
function need to be separated into two function.
One for selecting the platform's clock names, the other
will do the reparenting.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
---
 arch/arm/mach-omap2/mcbsp.c |   37 ++++++++++++++++++++++---------------
 1 files changed, 22 insertions(+), 15 deletions(-)

diff --git a/arch/arm/mach-omap2/mcbsp.c b/arch/arm/mach-omap2/mcbsp.c
index 857860b..9129003 100644
--- a/arch/arm/mach-omap2/mcbsp.c
+++ b/arch/arm/mach-omap2/mcbsp.c
@@ -65,25 +65,16 @@ static int omap2_mcbsp1_mux_rx_clk(struct device *dev, const char *signal,
 	return 0;
 }
 
-/* McBSP CLKS source switching function */
-static int omap2_mcbsp_set_clk_src(struct device *dev, struct clk *clk,
-				   const char *src)
+static int omap2_mcbsp_reparent_clk(struct device *dev, struct clk *clk,
+				    char *fck_src_name)
 {
 	struct clk *fck_src;
-	char *fck_src_name;
 	int r;
 
-	if (!strcmp(src, "clks_ext"))
-		fck_src_name = "pad_fck";
-	else if (!strcmp(src, "clks_fclk"))
-		fck_src_name = "prcm_fck";
-	else
-		return -EINVAL;
-
 	fck_src = clk_get(dev, fck_src_name);
 	if (IS_ERR_OR_NULL(fck_src)) {
-		pr_err("omap-mcbsp: %s: could not clk_get() %s\n", "clks",
-		       fck_src_name);
+		dev_err(dev, "clk reparent: could not clk_get() %s\n",
+			fck_src_name);
 		return -EINVAL;
 	}
 
@@ -91,8 +82,8 @@ static int omap2_mcbsp_set_clk_src(struct device *dev, struct clk *clk,
 
 	r = clk_set_parent(clk, fck_src);
 	if (IS_ERR_VALUE(r)) {
-		pr_err("omap-mcbsp: %s: could not clk_set_parent() to %s\n",
-		       "clks", fck_src_name);
+		dev_err(dev, "clk reparent: could not clk_set_parent() to %s\n",
+			fck_src_name);
 		clk_put(fck_src);
 		return -EINVAL;
 	}
@@ -104,6 +95,22 @@ static int omap2_mcbsp_set_clk_src(struct device *dev, struct clk *clk,
 	return 0;
 }
 
+/* McBSP CLKS source switching for OMAP2/3 */
+static int omap2_mcbsp_set_clk_src(struct device *dev, struct clk *clk,
+				   const char *src)
+{
+	char *fck_src_name;
+
+	if (!strcmp(src, "clks_ext"))
+		fck_src_name = "pad_fck";
+	else if (!strcmp(src, "clks_fclk"))
+		fck_src_name = "prcm_fck";
+	else
+		return -EINVAL;
+
+	return omap2_mcbsp_reparent_clk(dev, clk, fck_src_name);
+}
+
 static int omap3_enable_st_clock(unsigned int id, bool enable)
 {
 	unsigned int w;
-- 
1.7.8

[PATCH 2/2] OMAP4: mcbsp: Clock reparenting support

[Peter Ujfalusi]

With OMAP4 the clock names are different compared to
OMAP2/3.
The internal fclk name depends on the McBSP instance
number.
In case of MCBSP_TYPE4 (OMAP4) we should use different
clock selection method compared to OMAP2/3.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
---
 arch/arm/mach-omap2/mcbsp.c |   29 ++++++++++++++++++++++++++---
 1 files changed, 26 insertions(+), 3 deletions(-)

diff --git a/arch/arm/mach-omap2/mcbsp.c b/arch/arm/mach-omap2/mcbsp.c
index 9129003..829f4c7 100644
--- a/arch/arm/mach-omap2/mcbsp.c
+++ b/arch/arm/mach-omap2/mcbsp.c
@@ -111,6 +111,23 @@ static int omap2_mcbsp_set_clk_src(struct device *dev, struct clk *clk,
 	return omap2_mcbsp_reparent_clk(dev, clk, fck_src_name);
 }
 
+/* McBSP CLKS source switching for OMAP4 */
+static int omap4_mcbsp_set_clk_src(struct device *dev, struct clk *clk,
+				   const char *src)
+{
+	struct omap_mcbsp *mcbsp = dev_get_drvdata(dev);
+	char fck_src_name[30];
+
+	if (!strcmp(src, "clks_ext"))
+		strcpy(fck_src_name, "pad_clks_ck");
+	else if (!strcmp(src, "clks_fclk"))
+		sprintf(fck_src_name, "mcbsp%d_sync_mux_ck", mcbsp->id);
+	else
+		return -EINVAL;
+
+	return omap2_mcbsp_reparent_clk(dev, clk, fck_src_name);
+}
+
 static int omap3_enable_st_clock(unsigned int id, bool enable)
 {
 	unsigned int w;
@@ -184,9 +201,15 @@ static int omap_init_mcbsp(struct omap_hwmod *oh, void *unused)
 					name, oh->name);
 		return PTR_ERR(pdev);
 	}
-	pdata->set_clk_src = omap2_mcbsp_set_clk_src;
-	if (id == 1)
-		pdata->mux_signal = omap2_mcbsp1_mux_rx_clk;
+
+	if (oh->class->rev == MCBSP_CONFIG_TYPE4)
+		pdata->set_clk_src = omap4_mcbsp_set_clk_src;
+	else {
+		pdata->set_clk_src = omap2_mcbsp_set_clk_src;
+		if (id == 1)
+			pdata->mux_signal = omap2_mcbsp1_mux_rx_clk;
+	}
+
 	omap_mcbsp_count++;
 	return 0;
 }
-- 
1.7.8

Re: [PATCH 2/2] OMAP4: mcbsp: Clock reparenting support

[Paul Walmsley]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 2234 bytes --]

Hi Péter

On Thu, 15 Dec 2011, Peter Ujfalusi wrote:

> With OMAP4 the clock names are different compared to OMAP2/3. The 
> internal fclk name depends on the McBSP instance number. In case of 
> MCBSP_TYPE4 (OMAP4) we should use different clock selection method 
> compared to OMAP2/3.
> 
> Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
> ---
>  arch/arm/mach-omap2/mcbsp.c |   29 ++++++++++++++++++++++++++---
>  1 files changed, 26 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/arm/mach-omap2/mcbsp.c b/arch/arm/mach-omap2/mcbsp.c
> index 9129003..829f4c7 100644
> --- a/arch/arm/mach-omap2/mcbsp.c
> +++ b/arch/arm/mach-omap2/mcbsp.c
> @@ -111,6 +111,23 @@ static int omap2_mcbsp_set_clk_src(struct device *dev, struct clk *clk,
>  	return omap2_mcbsp_reparent_clk(dev, clk, fck_src_name);
>  }
>  
> +/* McBSP CLKS source switching for OMAP4 */
> +static int omap4_mcbsp_set_clk_src(struct device *dev, struct clk *clk,
> +				   const char *src)
> +{
> +	struct omap_mcbsp *mcbsp = dev_get_drvdata(dev);
> +	char fck_src_name[30];
> +
> +	if (!strcmp(src, "clks_ext"))
> +		strcpy(fck_src_name, "pad_clks_ck");
> +	else if (!strcmp(src, "clks_fclk"))
> +		sprintf(fck_src_name, "mcbsp%d_sync_mux_ck", mcbsp->id);
> +	else
> +		return -EINVAL;
> +
> +	return omap2_mcbsp_reparent_clk(dev, clk, fck_src_name);
> +}

What do you think about using clkdev alias lines and hwmod optional clks 
for this instead?  With these mechanisms, you can give clocks a consistent 
"role name" for a given device, even if the underlying clock changes on 
different platforms.

So in the mach-omap2/clock*_data.c files, you'd do something like

	CLK("omap-mcbsp.1",	"pad_fck",	&mcbsp_clks,  CK_3XXX),

for OMAP3, and 

	CLK("omap-mcbsp.1",	"pad_fck",	&pad_clks_clk,  CK_44XX),

for OMAP4.

Then in the hwmod data files, you'd add them as optional clocks - 
something like:

static struct omap_hwmod_opt_clk mcbsp1_opt_clks[] = {
	{ .role = "pad_fck", .clk = "mcbsp_clks" },
};

for OMAP3, and

static struct omap_hwmod_opt_clk mcbsp1_opt_clks[] = {
	{ .role = "pad_fck", .clk = "pad_clks_clk" },
};

for OMAP4.

I think this might work for your needs?


- Paul

Re: [PATCH 2/2] OMAP4: mcbsp: Clock reparenting support

[Peter Ujfalusi]

Hi Paul,

On 12/17/2011 02:37 AM, Paul Walmsley wrote:
> What do you think about using clkdev alias lines and hwmod optional clks 
> for this instead?  With these mechanisms, you can give clocks a consistent 
> "role name" for a given device, even if the underlying clock changes on 
> different platforms.

I need to look into the clock framework, but sounds promising.
I can see quite big difference between the clock3xxx_data, and
clock44xx_data regarding to McBSPs.
Need to understand why, and what need to be changed to achieve what I want.

> So in the mach-omap2/clock*_data.c files, you'd do something like
> 
> 	CLK("omap-mcbsp.1",	"pad_fck",	&mcbsp_clks,  CK_3XXX),
> 
> for OMAP3, and 
> 
> 	CLK("omap-mcbsp.1",	"pad_fck",	&pad_clks_clk,  CK_44XX),
> 
> for OMAP4.
> 
> Then in the hwmod data files, you'd add them as optional clocks - 
> something like:
> 
> static struct omap_hwmod_opt_clk mcbsp1_opt_clks[] = {
> 	{ .role = "pad_fck", .clk = "mcbsp_clks" },
> };
> 
> for OMAP3, and
> 
> static struct omap_hwmod_opt_clk mcbsp1_opt_clks[] = {
> 	{ .role = "pad_fck", .clk = "pad_clks_clk" },
> };
> 
> for OMAP4.
> 
> I think this might work for your needs?

Is it OK if I do this as an incremental patch as soon as I figured out
the way to do it nicely?

Thank you,
Péter
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html